ees. +. > te > ros04u. ~ @ eeees ba6r0e bnontes bernntenneoeneed pb abaaeaenes onsebebnsnarene rent wes Pere reeere rte PrePre tril licelaleeer eee re ey Tere ~ pete sea dbeeoaharntadeebentiehertetisne nu ANE REE TE TNES Her ere +g benacéracsetossiierscersrsre secs a pe iveseeteeeeese ts tees tyi{leserys eet alestes iy prersere * triers oer tet: reseseat=eetent oa Setters sete reece res ore Ie prone wee tne endear poet > vereery ert tee TETeTeruatroreseahaeyi ‘ pear <" pee be peomerers oes tape toy rvreer Tistetretes rire steer very bisesensssrssececanizbesctcatenacatanterne: seacnsanssaceetone yeas ete bnen eetastacgstiseere she Myanepnanensnnns sts Ne titiberecateetenettes: tse eatehs teste trenameecsoaratscengetegtusransoe so sctsetee te Ste aebe tes seftt Wires alate: oo pesen ere car gsetece tone yzery spreatctacssseesseasgegrosscitesesasssssptseaucecsreaeatceasssatateserstrestssctresseagsstaasstye tasatetatthaceshne. tte Szrasacsstastaatyetsssecats eats Sta hatytatatehatere BASIE EMIS Stet TEh totals Setetacate tacetetataty essay sient st tatateteleteresgt gt ytate rats te jee ie tereo tate 9+ gear t Pr eatags Prtatntore beta y. 20° Ph tbey oon’ so eoarers rer tr " ” tit MIT, Wate The barnd hrtrt + Fryer eae re reoeo ees res ce te : eceet sevens we lhe tserhe eee Te seen rate wy etre aente errererttT tise atts pas ce ts mee ynemarbesreres ATP Reed by Pee ee ages Ho eeep tT Re PT hee Hh ree COR VR Pee TP Rese adap en peer hr ertee : —s. * estat Sts! HrdSeT eT Leesa eng on seagate nsegrerecs Peseta tes Set Ah eee ereseeenecte) MARA peste cipitrertscottinesespesseetaereserie gees Noort yale tetra thet reson ahennsnbrersentesstesteets. Pete se: $05 9: +, ore preenre: e8 4 980) + SePSPEDREe poem does or essay arene +. esnyensenasecmetneetescatanices sabeSecaryrirecece: Pe + pevesereel peas +5 vere eetae te teth oasasees ste tateers aenscateneseryipetces us tesaiatesetiees sSosrmorteressees| roar ud baat be vie tuned wererts Pree gn} ones ts ye tanpudue vont = ppansreseeireses sssacrel 2b. o a" ; orp trer hres aeeairte a r poceeeces ST eetettieser: i. : tr yo teatetes tote ees res +. aerate tire neesseaeaanateas ty sgtregeeey ort yaesotrebeiratretcecne se emese go pesos tetatetesasssatetetasest: setts ¢ hisecarasessats? spteetete Meet tysticiten ts te ayite races rsagaratalscalstetvere, cysz ti rerseetecseszafacscecerecycacytseees ase. selatetoigtecataty cate Satecogats tare te tesate gee tetnensts tate, isatebetacaca pieistatitetatetytacety taste tes piace stat etecssstana tarstangrecerdtorinatstisisatssate:stetatucheasitecstatysatargtecace: atest ten : HANEEBE Pet Geet h t eausaieee arate Ralatetafiatg tats platetaty tyre tate 06 Uanenye 9a qrenSl ee >: 2 2 Hd homeo ws ng “ovees eSesees : . .33 ° tydert* pore <% sacle eves et eae eter re hs spaphsagesesysiteatatata tacuepetteceagepeeceatsatics’ 3 eeaaenseteste enti ; i es statesyhees sp eytt reptrtinterecpistrterstersh Selah 3g rhee rashisasass Sinar sca! Pete te st pe re een sttiets ete yte te De vetoes satatats Setce Ht sae : . ¥ . . - ¥ sitet ptaeteceze con einapte lec tatatet etel ats phate tasetebetesctess tetetsuagiosbet paatatptatetesptesesscite: at {iets Westy i Bert eratstatstet sts es3 ee. or ear hy statetsigtetatetateteticaatats tetatye ceiaes a3 te3 goss batedstatititetysatsatatacetetessta tpt ana ues ~ * iT ” =oe8 om ~ 4 ~ - Pes “0 ittsrty Hea eS jetneracess. nde bbetenes 3 bwergas ng et ty } sabgessrigestsscss Sactaseetetes transtetensesstey veatatnts iaazszecassaebahsssas Pebitatarssa rinses " abet Stireshy ! Sieh} Serripcieittsttestpritstisiberte py gir sash tSentetetisetesss bee sistaees ates slegyr et tte rr te rarpeerere. = te pasneet.s ete. ra etantenstescatans tetatae sti “ baebeae sphester atte oho vo st pense ossateeteests begs prerpegmens 5353 os pete bmag its! sSezateeacisatcetea nyt + tet is soe: peasieasestaiepests cats sptesacene sa tteps teres + ra iirtetitatstpsece mregensfeesatgoacensmestepeneds ft recat eras? Dyhebasy loss ssheeleeeleselsly’ bee wentes eet eres Wisithectrensstieds ntvate rircrarisehits ie (iaarict Label Tee Peteebagstpemtes net) bs reas. $< pate 3 seetithittisghertieniss tresta bytes Perera hey : > MAUNA ~ raisitsssaysstsessersrtevedisdnseeneegsatiesciaarertiay: rpetadtgstetstaebesdiats 4 szat. 2s | i epee Mosessatioresssbatssesstenstess : ates } tie ets asatace palsinata tit Seater pear tyte nastigaase ecamaacte patheinntep tire siete sateearetecs cibatsesdyr tat sth snrecsg tt if isiettitetetstedaeaztazgs gebe aeies tess aftereite. eatsists peered: ts $ oresape taste cs a2 bene w ar. Sesacs ee tetatetace eiieestissuesseuestrnaa atin teeahas titesstecastenatata set HRN GS as uc is sista area tiestrutasetr sates 353) Ve a Dyaeh taste ceteate guesses eos + + sriarinerittat retieetetat sida ee * : tiitete ye sae ss : os i Soetes ope pene pode debs bwaee ‘4 + : : vbaeses tot este atte ati tae tats Srabitetstetriestioatetty barete) : aitaitens naan + re aonb ei Tih eres eis stetss M Seat pt ted apsiers oye neh: fey peaeeet pit asts ahs seueese: re} tet 3 9. syreeat sethheseart Spot Mitt tive tire teasty THassittites} 23? eye teseataeace: setmimsobes ehes eter sasesks bee pee - +4 aye ee qe at dapottessenatepebpaecen oy setrbchieteea teresa? Athy et ' $; eee) sigan tparits ° + rh hetewe te mie Seesestansteesss oe . eave eigbepge eset “4 ‘ : seme tt ttt . ; baschsbene aa + petits 4 ‘ ott, : . Voreha tebwset aye ia jatoredal eh?) atete rrrpscie ey 2. 3 sree etbrsestan te stectest it : PRS Pseraasgescesteecossee eco eceeea ana Ste inter i Reg rece ees RITE 143 ont 8 babestantesses sth ; : tee ty esesaitaeseatieae 7 arahiteers “sy wally ratrs Hartt) eae Perareteegt wri pen sit artistes respi rpytsts uate ro 4 sf aeteaas * 4 terete att i : Hi i uy i y “ 13 pideeetalitesriery + pene : regis tare. tee Teleseiet Si seenescbbsttrossessseyeyegeieee: ts abst eee sores in bebe es bebe teens yirees sae ttses tcc $ ae ats if i ‘ ee it Seteteiet ists sie tose +3 ite: ‘ . eprestitzes tele +95 be saeco tstecaa testtate Seti eeeceaa REE ERE EE EG t titshraeensaepetin tgs bahia idissentititgteeytanaaatar seenptete 3 i totithatstathte: sitet teats eee ohere + e S 4 * pase = oe “4 srastetanagsraseetesesretratsass asitreitirits 3 H statatitarpcgens tet atacat hy Wiehe betay Ve beset oeteceayts. breseasseedy ty etrieiay iteritetenes sh Siestiesissersetyeteses ties ath ot pyessates +3 ; castes re : 1 * + + seater sities aise ert : i? aH mite we Ssates breseeasteasett : + : ists dsitier: ae a3 epnsessneset : : i Site Seite seceristetissss sesettesceshes spreeieasgeassa} : it ; ) shh Herithsuess pies tiapestazee s : 1 tate 14 Ys setets + ees " mitttic fa th ‘1 rt riety seit Petshis i Mab, bots 43 + 5 it $s cierete ee hateee est ittepsepey ¥ t crf] des neariespessetete Stetettteriiay sises ; oy ite tithes: sreeetr erste eset thitie ies ¢ eeeseerssesetose seh s. oti eet! > 7 . reesreres + 4 = Fé. ee . esterase 1 : ett 1] aS Fer) ty paeah esses : : iH it airhie sahit wt reher 4 t SKIN )} 910 7/6/68 oo ee BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BO LANY: VOL. 3 1963—65 TRUS DEES. Or THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1965 DATES OF PUBLICATION OF THE PARTS No. I, May 1963 No. 2, July 1963 No. 3, September 1963 No. 4, September 1964 No. 5, May 1965 No. 6 October 1965 Contents and Index, October 1965 PRINTED IN GREAT BRITAIN AT THE BARTHOLOMEW PRESS DORKING BY ADLARD AND SON, LTD. CONTENTS BOTANY VOLUME 3 PAGE A revision of the genera Buchenavia and Ramatuella. By A. W. EXELL and C. A. STACE > The diatom genus Capartogramma and the identity of Schizostauron. By R. Ross 47 Angiosperms of the Cambridge Annobon Island Expedition. By A. W. EXELL 93 A revision of the genus Petrorhagia. By P. W. Baty and V. H. HEYWoopD I19Q Marine algae of Gough Island. By YVoNNE M. CHAMBERLAIN 173 The Ceylon species of Asplenium. By W. A. SLEDGE 233 Index to Volume 3 270 “¢! GS ati + uF, Geis i 2 FE + | atang Maier 4 es |) ae, ae : fp os ns ae ae se tna ot So eae peat eo : ae Renta. a) a rye mn ent ae i ay 03 Weis OE ae ar peri age = A REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA A. W. EXELL AND GA. 8 FACE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) _ BOTANY | Vol. 3 No. 1 Fig LONDON: 1963 © a REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA Bivs AL We EXELL ann: CoA. STACE ways Pp. 1-46 ; 5 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BOTANY Vol. 3 No. 1 LONDON : 1963 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), tstituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 3, No. 1 of the Botany series. © Trustees of the British Museum, 1963 PRINTED BY ORDER OF THE TRUSTEES OF THES BRIEISH MUSEUM Issued May 1963 Price Sixteen Shillings A REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA By A. W. EXELL and C. A. STACE! THE two genera Buchenavia Eichl. and Ramatuella Kunth are confined to tropical America, in common with two other genera of the family Combretaceae (Bucida and Thiloa). While one of us (C. A. S.) was undertaking a detailed study of the cuti- cular anatomy of Buchenavia and Ramatuella, as part of a comprehensive survey of the whole family, it became apparent that the taxonomy of the former genus was badly in need of revision. Although twenty-eight specific names had appeared under Buchenavia it had never been revised. Since certain changes in the taxonomy of Ramatuella also seemed advisable, a joint revision of both genera was undertaken. As many of the earlier, and some of the later, species of Buchenavia and Ramatuella have been very inadequately described, and the discovery of new species has revealed further characters of diagnostic significance, descriptions of all the species of these two genera are included in this paper. Our work is based mainly on the collections in the herbaria of the British Museum (Natural History) (BM) and the Royal Botanic Gardens, Kew (K). Loans were also received from the Chicago Natural History Museum (F), Botanical Institute of the Academy of Sciences of the U.S.S.R., Leningrad (LE), Botanische Staats- sammlung, Munich (M), New York Botanical Garden (NY), Muséum National d’Histoire Naturelle, Laboratoire de Phanérogamie, Paris (P), U.S. National Museum, Washington (US), and Naturhistorisches Museum, Vienna (W). We are grateful to the Directors of these Institutions, and especially to the Botanical Institute of the Academy of Sciences of the U.S.S.R., and to Dr. J. Cuatrecasas of Washington, for the donation to the British Museum of duplicate specimens. Concurrently with the commencement of this revision a fairly large and very useful collection of the two genera concerned was received at the British Museum from the Instituto Agrondémico do Norte, Belém, Brazil (IAN). The abbreviations indicated parenthetically above are used in the citation of specimens throughout the paper. Other institutions cited are: Natuurweten- schappelijk Museum, Antwerp (AWH) ; Botanisches Museum, Berlin (B) ; Jardin Botanique de l’Etat, Brussels (BR) ; Museu Paraense Emilio Goeldi, Belém (MG) ; Jardim Botanico, Rio de Janeiro (RB); Botanical Department, Naturhistoriska Riksmuseum, Stockholm (S). 1 During this investigation C. A. S. was receiving a grant from the Department of Scientific and Industrial Research, for which he is most grateful. BOT. 3, I. 1§ 4 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA BUCHENAVIA Eichl. BucHENAVIA Eichl. in Flora 49:164 (1866); in Mart., Fl. Brasil. 14 (2): 95 (1867). Type: B. capitata (Vahl) Eichl. Buchenavia belongs to the tribe Terminalieae of the family Combretaceae and is most closely related to Terminalia, from which it differs in that the filaments of the stamens are adnate to the back of the anthers and the calyx lobes are only very slightly or scarcely at all developed. In Tevminalia the anthers are versatile and the calyx lobes well developed, usually triangular in shape and acute at the apex. As in all genera of the Combretaceae except Strephonema, the “ receptacle ”’ of Buchenavia is composed of two morphological components: an upper portion (“upper receptacle’ or “ calyx tube ’’), cup-shaped in this genus (sometimes con- siderably elongated in Combretum and Quisqualis), and a lower portion (‘‘ lower receptacle ’’) surrounding and fused with the ovary and (in this genus) extended above it. As it is usually not easy or possible to ascertain from gross morpho- logical studies which part of the “‘ upper receptacle ”’ is in fact receptacular rather than appendicular (although in Laguncularia and Lumnitzera the presence of adnate bracteoles indicates the probable receptacular origin of at least part of that organ), we have used the term “ hypanthium ”’ rather than “ receptacle”’. In Buchenavia the cupuliform upper hypanthium is almost always broader than long and bears five scarcely developed calyx lobes, ten stamens, no petals, and a well-developed hairy disk. It is usually more elongate (i.e. campanulate) in Terminalia. The lower hypanthium is also characteristic in Buchenavia, consisting of a proximal portion surrounding the sessile ovary and a distal portion tapering from the apex of the ovary and simulating a pedicel, bearing the upper hypanthium at its apex. In most species the portion surrounding the ovary is pubescent, while the distal portion and the whole of the outside of the upper hypanthium are glabrous or almost so. In one or two species, however, the lower portion is also glabrous. In Terminalia the whole hypanthium is usually uniformly hairy or (less often) glab- rous, but a few species have the Buchenavia pattern of indumentum. Some species of Terminalia possess a lower hypanthium shaped as in Buchenavia, but the majority lack the long pedicel-like distal portion. Eichler (1866, 1867) included eight species in the genus: one (B. capitata (Vahl) Eichl.) previously known in Bucida and also in Terminaha; one (B. oxycarpa (Mart.) Eichl.) previously known in Terminalia ; and the other six newly described. Since then twenty additional names have appeared in the literature (three of these being nomina nuda, the rest validly published) but the genus has not been revised. Out of the 25 valid names we recognize 21 species, and we also describe three addi- tional ones. From various herbarium specimens we have seen it seems that there may be several more awaiting description but they are at present insufficiently known. The genus is confined to but fairly widespread in tropical America where it is centred on the Amazon basin (see Figs. 1, 2). It occurs throughout the West Indies (one species) from Cuba (within 150 kilometres south of the Tropic of Cancer) to REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 5 Trinidad ; in Panama (one species); Venezuela (five species) ; Colombia (five species) ; the Guianas (four species) ; Brazil (twenty species) ; Peru (two species) ; and Bolivia (one species). One outlying species (B. kleinii) occurs 500 kilometres Tropic of Cancer V — °9 aS ° Soe 9 _ a ee Puerto Rico AS Laer. Jamaica S——_—— Hispaniola ~— S.__ o ¢ Antigua Montserrat °@ Guadeloupe Dominica,° © Martinique 0 St. Vincent ° \o Panama Equator ee BRAZIL Tropic of Capricorn Fic. 1. Distribution of Buchenavia (part) and Ramatuella: 1, B. capitata (insular) ; 2, B. parvifolia; 3, B. kleinii; 4, B. ochroprumna; 5, Ramatuella (all species) ; C, B. capitata (continental) ; S, B. sericocarpa, 6 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA Equator Tropic of Capricorn Fic. 2. Distribution of Buchenavia (part): 1, B. reticulata, B. suaveolens and B. ptero- carpa; 2, B. tomentosa; 3, B. oxycarpa; 4, B. macrophylla; 5, B. fanshawei; 6, B. grandis; Ac, B. acuminata; Ca, B. callistachya; Co, B. congesta; D, B. discolor ; H, B. huberi; L, B. longibracteata; Me, B. megalophylia; O, B. oxycarpa (outlier) ; Pa, B. pallidovirens ; Pl, B. pulcherrima; Pn, B. punctata; R, B. reticulata (outlier) ; V, B. viridifilora. south of the Tropic of Capricorn in southern Brazil. This distribution pattern follows that of the family as a whole (in the New World) and is a typical pattern for a predominantly tropical group. The largest two genera, Combretum and Terminalia, are almost pantropical; the two small genera Laguncularnia and Conocarpus are found on both sides of the Atlantic ; Buchenavia is the largest genus of Combretaceae confined to one hemisphere, this being due no doubt to the well- known phenomenon of prolific speciation in the Amazonian forest region ; Bucida, Ramatuella and Thiloa, the other three American genera, are also confined to the New World. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 7 Eichler (1866) recognized two groups of species in Buchenavia : “‘ Flores fructusque capitati’’ and “ Flores fructusque vel saltem fructus spicati’’. It is still possible to recognize these two groups with ease, and we have used this character in the key. We have refrained, however, from giving the two groups taxonomic status, since there appear to be no other constant differences between them, and from a study of other genera in the Combretaceae (especially Combretum) it seems that the capitate inflorescence has arisen on several separate occasions in rather distantly related groups. Furthermore, the capitate B. sericocarpa is otherwise almost identical (as far as we can tell) with the spicate B. acuminata. In the flowering state it is usually very easy to distinguish between species of the two groups, as the inflorescences of one are long and slender, the flowers in many cases being well separated from each other, while in the other group the flowers are crowded together on a rhachis rarely as long as the peduncle. In the fruiting state the distinction is no more difficult, although usually very few fruits reach maturity in each infructescence. The presence of a spicate inflorescence is, however, always evident, even when the peduncle bears but a single fruit, since the scars of the fallen flowers may be seen at intervals along the rhachis. In this spicate group the rhachis often elongates considerably during ripening of the fruits. Much of the difficulty encountered in the genus is due to the incompleteness of our knowledge of several species which are known from fruits and leaves or from flowers and leaves only. The leaves are frequently of a very different appearance at flowering and fruiting stages. Fortunately, although the fruits of Buchenavia do not differ from those of Terminalia as a whole, all the American species of the latter genus apparently have conspicuously winged fruits, while all species of Buchenavia have fruits without wings or with very narrow ones. The greatest development of wings is shown by B. pterocarpa and B. punctata. In using this character to separate the two genera, the only possible source of error would be in the case of Terminalia catabpa, now widespread in tropical America, but this is an introduced Asiatic and Australasian species easily recognizable by its large thin obovate leaves usually with a cordate or subcordate base. Buchenavia species are shrubs or trees from about three to fifty metres tall. Ecologically they may apparently be roughly divided into two main groups: small trees or shrubs of lowland forest often inundated by water, frequently growing on river banks, by lakes and in swamps; and tall trees of more upland rain forest. The latter are said to flower “‘ very seldom” (Ducke in Bol. Técn. Inst. Agron. Norte 4 : 25 (1945)). KEY TO THE SPECIES In the following key we have used the characters which are most likely to be present on herbarium specimens: thus, unless absolutely necessary, fruit characters alone are not used. This has sometimes resulted in long couplets citing exceptions but is thought to be justified by the added usefulness in identifying incomplete specimens. The flowers unfortunately appear to be of very little taxonomic help in this genus. As in the descriptions, the terms “ tomentellous’”’ (a diminutive of “tomentose ”’) and “puberulous ’”’ (a diminutive ot “ pubescent ’’) are taken to BOT. 3, I. 1§§ 8 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA indicate not a lesser amount of indumentum but a shorter length of the hairs. ‘‘ Pseudostipe ”’ indicates a stalk-like basal portion of the fruit, a true pedicel being absent. Leaf length always includes the petiole (except, of course, when leaf- lamina is specified). In cases where the lamina is decurrent into the petiole it is not feasible to determine an exact point at which to distinguish between them. We have occasionally referred to the indumentum (or lack of indumentum) of the ‘ovary ’’, this being merely a short way of saying “‘ that part of the lower hypan- thium enclosing the ovary”’. This part usually differs in indumentum from the apical narrowed part. Flowers in short ovoid or capitate spikes usually less than 1-5 cm. long; peduncle usually longer than the rhachis ; spikes not elongating in fruit : Fruit glabrous or rarely very sparsely pubescent : Leaf-lamina coriaceous when mature, c. 2-8 X I-4 cm., narrowly to broadly obovate ; major lateral nerves rather conspicuous, c. 2-6 pairs; petiole c. 4-20 mm. long, usually rather stout ; peduncle slender or stout ; flowers with densely pubescent ovary; fruit 18-30 x Io-20 mm.; endocarp markedly longitudinally ridged ; smallest twigs rather stout (widespread throughout tropical America from Cuba to S. Brazil) . 1. B. capitata Leaf-lamina chartaceous when mature, c. I-5-4 X 0*5-2cm., narrowly to broadly obovate ; major lateral nerves rather inconspicuous, c. 5-12 pairs ; petiole c. 2-5 mm. long, slender ; peduncle slender ; flowers with glabrous ovary ; fruit 14-20 X 8-12 mm.; endocarp indistinctly longitudinally ridged ; smallest twigs very slender (Amazon and Orinoco basins only) 2. B. parvifolia Fruit pubescent or puberulous : Fruit densely but minutely puberulous, obtuse or acute at the apex; leaves obovate, usually broadly so, rounded, obtuse or rarely very shortly api- culate at the apex (S. Brazil only) : ; ; 3. B. klein Fruit densely tomentose ; leaves narrowly to broadly obovate, rounded to long- acuminate at the apex (N. Brazil only) : Fruit silvery- or grey-tomentose, obtuse at the apex, smooth-surfaced ; leaf-lamina oblong-obovate, 5-12 cm. long, apex abruptly caudate- acuminate with an acute acumen ; lateral nerves 7-9 pairs 4. B. sericocarpa Fruit ferrugineous-tomentose, long-beaked, markedly and irregularly ridged ; leaf-lamina obovate to oblanceolate, 4-8 (10) cm. long, apex rounded ; lateral nerves 3-5 pairs . : $ ; 5. B. ochroprumna Flowers in long narrow spikes usually 4 cm. iene or more, if in shorter spikes then the latter elongating in fruit : Leaf-lamina chartaceous, up to 15 cm. long, very narrowly obovate to oblanceolate, subacute to acute at the apex, conspicuously but minutely pellucid-punctate ; fruit very strongly and symmetrically 5 (6)-ridged, apiculate to shortly (to 3 mm.) acuminate at the apex, 27-31 mm. long . , 6. B. punctata Leaf-lamina various, but never pellucid-punctate ; fruit usually more or less REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 9 terete or obscurely ridged, or if strongly ridged then 2 ridges much more developed than the others or apiculus 3-9 mm. long : Leaves at anthesis and usually for some time afterwards densely rubiginous- to ferrugineous-pubescent to -tomentose below, pubescent and nearly always drying reddish-brown above; reticulation usually very con- spicuously raised below at maturity, connecting the 5-15 pairs of lateral nerves : Leaf-lamina (2) 5-26 (36) < I-7 (12) cm., narrowly elliptic to oblanceolate, obtuse to acute or shortly apiculate at the apex ; inflorescence 7-18 cm. long ; fruit ferrugineous- or rubiginous-tomentose or densely pubescent, apiculate to acuminate at the apex . : ; . 7. B. reticulata Leaf-lamina 3:°5-13°5 X I°5-7 cm., elliptic or obovate (narrowly elliptic or oblanceolate when very young), rounded at the apex (obtuse or more pointed when very young) ; inflorescence 25-13 cm. long; fruit sparsely ferrugineous-pubescent, rounded at the apex : Leaf-lamina rather conspicuously reticulate below when mature ; inflores- cence 3°5-I13 cm. long; flowers c. 4-5 X 2°5-4 mm.; fruit 22-31 x 15-25 mm., rather to very sparsely puberulous 8. B. tomentosa Leaf-lamina very conspicuously reticulate below when mature ; inflores- cence 2:5-5°5 cm. long; flowers c. 2:5-3 X I°75-3 mm. ; fruit 21-23 < 12-15 mm., puberulous . ; : an E B. callistachya Leaves becoming sparsely pubescent to almost ginbroue some time before an- thesis, rarely drying reddish-brown above ; reticulation never conspicuously raised (though lateral nerves often so) : Leaves almost or quite glabrous, up to 11-5 cm. long but often much shorter ; lateral nerves inconspicuous below, scarcely or not at all raised, domatia entirely absent from their axils; fruit flattened or terete, glabrous, with a distinct apiculus up to 4 mm. long : Leaf-lamina narrowly elliptic or oblanceolate to elliptic or obovate (usually obovate) ; fruit 24-30 mm. long, more or less terete, apiculus usually straight ; endocarp smooth, almost round in transverse section 10. B. suaveolens Leaf-lamina narrowly elliptic or oblanceolate to narrowly obovate ; fruit 14-28 mm. long, flattened with two narrow lateral wings and three additional] facial ridges, apiculus usually curved ; endocarp irregularly furrowed, with two lateral wings : . 11. B. pterocarpa Leaves various ; lateral nerves conspicuous below, the major nerves consider- ably or somewhat raised and very often with domatia in some or all of their axils ; fruit various, often long-beaked or rounded at the apex : Inflorescence with conspicuous bracts c. 7-13 mm. long subtending and exceeding the flowers; bracts persistent at least until anthesis or after; leaf-lamina 8-23 cm. long, oblanceolate, acuminate at the apex ; petiole up to 1:5 cm. long . . . 20. B. longibracteata Inflorescence with small narrow bracts subtending the flowers, bracts rarely equalling the length of the open flowers, up to c. 5 mm. long, io REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA usually caducous before anthesis; (flowering material unknown in B. pulcherrima, but this with petiole over 3 cm. long) : Fruit with an apical beak (3) 5-20 mm. long and abruptly constricted from the rest of the fruit : Fruit scarcely or not pseudostipitate at the base, beak up to 13 mm. long ; leaf-lamina up to 13 X 5 cm., lateral nerves 3-9 (12) pairs : Leaves with (4) 6-9 (12) pairs of lateral nerves ; fruit quite glabrous 12. B. oxycarpa Leaves with 3-4 pairs of lateral nerves ; fruit tomentellous 13. B. discolor Fruit with a pseudostipe 2-7 mm. long, beak 10-20 mm. long ; leaf- lamina c. (15) 20-32 X 5-10 cm., lateral nerves (g) 10-14 pairs 14. B. megalophylla Fruit rounded to acute or apiculate at the apex, apiculus when present very short (abnormally in young B. fanshawet apiculus up to 2 mm. long) : Leaf-lamina up to II X 4 cm., oblanceolate, abruptly and acutely apiculate or mucronate or shortly acuminate at the apex ; lateral nerves 5-7 pairs, arising at about 40° or less and then curving towards the leaf-apex at about 25° or less for most of their length ; reticulation conspicuous, very slightly raised below with very small inter-reticular spaces ; fruit glabrous, ellipsoid-obovoid 15. B. pallidovirens Leaf-lamina various, often much larger than Ir x 4 cm., if smaller rarely acutely apiculate or acuminate at the apex; lateral nerves 3-—numerous pairs, very rarely or never less than about 40° to the midrib for their proximal half ; fruit various, often pubescent : Leaf-lamina up to 10 X 5°5 cm. but rarely over 8 X 4 cm.; apex acute to rounded or retuse, never apiculate or acuminate ; lateral nerves 4-7 pairs; inflorescence 2~7 cm. long; fruit densely greyish-brown-tomentellous; domatia never as in B. fanshawet (q. v.) : ; . 16. B. viridiflora Leaf-lamina usually in all or most leaves over 8 cm. long, if less then apex apiculate or acuminate, except in B. fanshawei (with distinctive domatia, q.v.) and B. grandis (with glabrous fruit and ovary often glabrous) : Leaf-lamina up to 9°5 x 5 cm., elliptic or narrowly elliptic, apex obtusely or acutely acuminate ; domatia never as in B. fan- shawei (q.v.) ; fruit densely silvery-brown-tomentellous 17. B. acuminata Leaf-lamina in all or most leaves usually I0 cm. or more long, if less then apex not acuminate (but apiculate to rounded or retuse) except in some forms of B. macrophylla (with less hairy brownish fruit) and of B. fanshawei (with fruit similar to the latter and distinctive domatia) : REVISION @®F THE GENERA BUCHENAVIA AND RAMATUELLA 11 Leaf-lamina up to 14 X 6 cm., rounded to apiculate (acuminate) at the apex ; lateral nerves 3-5 pairs; petiole sometimes but usually not biglandular near the apex; distinctive domatia present in nerve axils on lower surface of lamina on some or most leaves of older or sometimes young branch- lets, opening of domatium narrower than its diameter ; fruit rather densely but minutely tomentellous 18. B. fanshawer Leaf-lamina various, often over 15 cm. long ; lateral nerves (4) 6-c. 14 pairs, if 5 or less then leaf apex apiculate or acumin- ate and petiole biglandular ; domatia absent or present and then simple and triangular, pocket-like, with the opening as wide as the diameter of the domatium ; fruit various : Petiole conspicuously biglandular in most Jeaves; fruit minutely but rather densely brownish-tomentellous 19. B. macrophylla Petiole without glands or petiolar glands inconspicuous ; fruit either glabrous or densely greyish-brown- or rufous- pubescent : Leaf-lamina 27-29 x 8-10 cm., apiculate at the apex ; petiole 3.cm. or more long ; inflorescence c. 10-15 cm. long ; fruit rounded to obtuse at both ends, densely rufous- velutinous (British Guiana only) 21. B. pulcherrima Leaf-lamina not over 16 cm. long; petiole up to about 3:2 cm. long; fruit glabrous or pubescent and then apiculate at the apex, usually shortly pseudostipitate at the base when mature (N. Brazil only) : Fruit densely pubescent or tomentose at maturity, indumentum wearing off when fruit very old but still visible in parts . : . 22. B. huben Fruit glabrous from an early stage; ovary sometimes also glabrous : Leaves not very densely clustered at branchlet tips, up to 14 cm. long; lateral nerves 6-9 pairs ; petiole up to 2 cm. long ; fruit with a pseudostipe c. 2-3 mm. long, acutely apiculate at the apex 23. B. grandis Leaves very densely clustered at the branchlet tips, to-18 cm. long; lateral nerves g—I3 pairs; petiole 2-3 cm. long; fruit with a pseudostipe under 2 mm. long or not pseudostipitate, rounded to more or less obtuse at the apex 24. B. congesta 12 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 1. Buchenavia capitata (Vahl) Eich]. in Flora 49: 165 (1866); in Mart., FI. Brasil. 14 (2) : 96 (1867). Bucida capitata Vahl, Eclog. Amer. 1: 50, t. 8 (1796). Terminalia obovata Cambess. in St.-Hil., Fl. Brasil. Merid. 2 : 241 (1829). Terminalia hilariana Steud., Nomencl. Bot., ed. 2, 2 : 668 (1841), nom. illegit. Bucida angustifolia Spruce ex Eichl., loc. cit. (1867), nom. syn.; non B. angustifolia DC. (1828). Buchenavia vaupesana Cuatrec. in Fieldiana, Bot. 27 (1) : 108 (1950). Buchenavia ptariensis Steyerm. in Fieldiana, Bot. 28 : 423 (1952). Leaf-lamina membranous when young, coriaceous when mature, very variable in size, narrowly to broadly obovate, retuse or rounded at the apex, sometimes obtuse or subacute when very young, cuneate and sometimes narrowed at the base, decurrent into the petiole, densely pubescent when very young but becoming glabrous except on the midrib and lateral nerves when mature, often ciliate at flowering time ; midrib and major lateral nerves usually conspicuous below, the latter variable in number, 2-6 pairs; reticulation usually distinct when mature, slightly raised below ; petiole 4-20 mm. long, usually rather stout. Peduncle slender or stout, c. 6-23 mm. long, almost glabrous to rufous-pubescent when young, almost glabrous at maturity ; rhachis under 10 mm. long, often only 3 mm. long ; flowers in sparse or dense capitula, Io-many in each inflorescence ; bracts very small and caducous before anthesis. Upper hypanthium 1-1-5 mm. long, 2-3 mm. across, glabrous or sparsely pubescent outside, pubescent within and ciliate on the margin; lower hypanthium pedicel-like with the ovary at the base, 1-5-3 mm. long, densely rufous- sericeous. Stamens exserted for c. 0-5-1°5 mm. Fruit ellipsoid, 5-ridged to sub- terete, 18-30 X IO-20 mm., apiculate or acute to broadly obtuse at the apex, nar- rowed below into a very short pseudostipe, glabrous or rarely very sparsely pubescent, apparently rarely more than one ripening in each infructescence ; endocarp markedly longitudinally ridged, usually about five ridges more prominent than the rest. B. capitata is by far the most variable, abundant and widely distributed species in the genus. It is found in the West Indies from Cuba (within 150 kilometres south of the Tropic of Cancer) to Trinidad ; Panama; Colombia (Vaupés) ; Venezuela (Amazonas and Bolivar) ; Dutch and French Guiana ; Brazil (Amazonas, Maranhao, Ceara, Pernambuco and Rio de Janeiro within 150 kilometres north of the Tropic of Capricorn) ; and Bolivia (Santa Cruz and La Paz). It is the only Buchenavia found in the West Indies, where it occurs as a common tree of river banks and low- land forest. It also appears to be frequent in similar conditions in north-east South America, becoming rare westwards in Colombia and Venezuela and southwards in southern Brazil and Bolivia (Fig. 1). Glaziou says (in Mém. Soc. Bot. France r (3) : 203 (1908)) that it is cultivated in Rio de Janeiro. As it is a well-known species, described in many Floras, we have not cited the numerous specimens preserved in herbaria. We consider that B. ptariensis Steyerm. and B. vaupesana Cuatrec. fall within the wide range of variation shown by B. capitata. In the West Indies plants of B. cap1- tata are characterized by the large number of flowers in each inflorescence, the rather small flower size and the relatively stout peduncles. Plants from the Amazon basin REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 13 tend to have fewer rather large flowers in each inflorescence and often more slender and more glabrous peduncles. Specimens from coastal areas of the mainland are often more or less intermediate in these characters. We are of the opinion that, in the present state of our knowledge, this variation does not lend itself to subdivision of the species into lower categories, but that it represents an ecocline from the West Indies through the coastal areas of the South American continent to the Amazon basin. B. vaupesana appears to be an extreme example of the Amazon type, with rather few large flowers and slender peduncles. We have seen the other characters by which this ‘‘ species’ is said to differ from B. capitata (narrow somewhat pointed leaves and long petioles) in other specimens of B. capitata, mostly due to immaturity. The single known gathering (Cuatrecasas 6828), of which we have seen an isotype (US), is the only record we have of B. capitata from Colombia. It is in flower. B. ptariensis, of which we have seen the only known gathering, Steyermark 60038 (F, holotype), is a very small-leaved plant lacking flowers and fruit. We have no evidence that it actually has a capitate inflorescence but it appears to fall at one extreme of the B. capitata range of variation and we can see no reason for keeping it separate. The leaves are of quite a different texture, shape and venation from those of B. parvifolia, as the author points out. It is one of only two specimens of B. capitata that we know of from Venezuela. It seems quite possible that at some future date, when the material of B. capitata is really representative throughout its wide range, it will become advisable to divide the species into infraspecific units. Vahl, when originally describing Bucida capttata, cited specimens collected by Ryan in Montserrat which, after being described by Vahl, were sent to Banks. They are now in the British Museum Herbarium (W. R. Dawson, Banks Letters : 727 (1958) ; letter of r May 1794). There are two specimens, one with flowers, young leaves and old fruits corresponding well with Vahl’s t. 8 [fig. 1], and the other with adult leaves and extremely young inflorescences fitting his fig. 2. Since Vahl (tom. cit. : 51) refers to the latter as ‘“‘ Aliam quoque in Montserrat legit Ryan ”’ and later (op. cit. 3 : 55 (1807)) indexes it as “ varietas ? ’’ it is reasonable to regard the former specimen, with flowers and old fruits, as the holotype of B. capitata. The vernacular name is given as “ Tanibuca’”’ by Glaziou (in Mém. Soc. Bot. France I (3) :203), and Exell (Fl. Suriname. 3 (1) :175 (1935)) quotes “‘ Katoelima”’, “ Toekoeli’’, ‘‘ Toekadi’’, “‘Matakki’’ and ‘‘Gemberhout ”’. 2. Buchenavia parvifolia Ducke in Arch. Jard. Bot. Rio Janeiro 4 : 150 (1925). Leaf-lamina chartaceous when mature, rather small, c. 1-5-4 0:5—2 cm., narrowly obovate to obovate, usually rounded but sometimes slightly retuse or very broadly obtuse at the apex, cuneate at the base, narrowed and decurrent into the petiole, almost glabrous except for sparse pubescence on the midrib and lateral nerves when mature ; major lateral nerves often not very distinct, rather numerous for small size of leaf and often with many minor lateral nerves almost as conspicuous, c. 5—I2 pairs ; 14 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA petiole slender and short, c. 2-5 mm. long. Inflorescences (fide Ducke) more or less as in B. capitata ; peduncle c. 10-18 mm. long, slender. Ovary completely glabrous. Fruit ellipsoid or less often obovoid, more or less terete, sulcate when dry, c. 14-20 mm. long, usually shortly apiculate, sometimes rounded or obtuse at the apex, roun- ded to acute at the base, glabrous; endocarp very slightly longitudinally ridged, very slightly flattened but without prominent lateral ridges. VENEZUELA: Amazonas: Cerro Sipapo, ‘“‘ buttressed tree 30 m. high, 40 cm. diam., dense low elevation forests ’’, 24 Nov. 1948, Maguire & Politi 27370 (BM). BraziL: Amazonas: Estrado do Aleixo, Manaos, in non-flooded forest, 14 June 1933, Ducke 25016 (K). Cachoeira Grande, Manaos, in non-flooded forest, 3 Jan. 1937, Ducke 374 (K). Mandaos, Rio Taruma, “terra firme, alta, floresta alta’’, 10 Sept. 1950, Frdes 26580 (BM). Parad: Vila Braga, Rio Tapajoz, medium-sized tree in non-flooded wood by waterfall, 23 Sept. 1922 and 24 May 1923, Ducke 17686 (K ; RB, lectotype, not seen). Bosque Rodrigues Alves, Belém, without date, Guedes 233 (BM). Belém, without date, Guedes 239 (BM). Ducke also recorded the species “inter flumina Cumina-Mirim et Ariramba affl. Rio Trombetas ”’, Ducke 13584 (RB), and as sterile trees near Belém, Breves and Faro, all in the state of Para. But for Maguire G& Politi 27370, which was originally named B. capitata, this species has been found only near the Amazon and its tributaries in northern Brazil. The Venezuela specimen is from a locality close to the Rio Orinoco some 1,250 kilo- metres from Manaos, the nearest Brazilian locality. The distinctive leaves, however, make its identification quite definite. 7 No specimen we have seen has flowers, those of Ducke 17686 (which we have chosen as lectotype) having just fallen. The ovary in this gathering is glabrous, and as Ducke also states this it seems that this character provides a further good means of distinguishing B. parvifolia from B. capitata. By far the greater number of Buchen- avia species have densely pubescent ovaries, the indumentum often disappearing to give quite glabrous fruits (e.g. B. oxycarpa). B. grandis appears to be the only other species with a glabrous ovary. 3. Buchenavia kleinii Exell in Ann. & Mag. Nat. Hist., Ser. 12, 6 : 400 (1953). Leaf-lamina chartaceous when young, subcoriaceous when mature, 2-9 X I°5-5 cm., obovate to broadly obovate, rounded or obtuse or subacute or apiculate at the apex, acute-cuneate but not decurrent at the base, densely pubescent when very young, becoming glabrous except on the midrib and lateral nerves when mature ; major lateral nerves 4-8 pairs ; petiole distinct, fairly long in most cases, 8-20 mm. long. Inflorescences as in B. capitata; peduncle rather stout, Io-20 mm. long. Fruit ellipsoid or less often obovoid, 17-28 x 11-14 mm., apiculate or less often rounded at the apex, densely but minutely rufous-puberulous, becoming less hairy when very old ; endocarp very similar in shape to that of B. parvifolia. BraZiIL: Santa Catarina: Mata do Hoffmann, Brusque, 13 Jan. 1951, Klein 22 (BM; S, holotype). Cunhas, Itajai, 10 m., 29 Nov. 1954, Klein 870 (BM) ; same REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 15 locality, 8 Feb. 1955, Klein 1144 (BM); same locality, 10 Mar. 1955, Klein 1182 (BM) ; same locality, 26 July 1955, Klein 1479 (BM). Bracgo Joaquim, Luis Alves, Itajai, 300 m., 14 Dec. 1954, Klein 924 (BM) ; same locality, 350 m., 13 Jan. 1955, Klein 1075 (BM). Horto Florestal I.N.P., Ibirama, 250 m., 27 Dec. 1954, Klein 940 (BM) ; same locality, 300 m., 4 Feb. 1956, Reitz & Klein 2568 (BM). Morro da Fazenda, Itajai, 50 m., 17 Mar. 1955, Kleim 1232 (BM). Piloes, Palhoga, 300 m., 1r Jan. 1957, Klein 2158 (BM). Mata da Companhia Hering, Bom Retiro, Blumenau, 250 m., 15 Dec. 1959, Klein 2342 (BM). Morro Spitzkopf, Blumenau, 700 m., 1r Mar. 1960, Klein 2424 (BM). Morro da Ressacada, Itajai, 250 m., 29 Mar. 1956, Reitz & Klein 2915 (BM). Matador, Rio do Sul, 350 m., 27 Jan. 1959, Reitz & Klein 8319 (BM). Alto Matador, Rio do Sul, 800 m., 30 Dec. 1958, Reitz 6126 (BM). Vargem Grande Lauro Miiller, 350 m., 20 Feb. 1959, Reitz & Klein 8494 (BM). ae species is known only from the south-eastern corner of Santa Catarina, southern Brazil, although it may well occur elsewhere. The area is south of the Tropic of Capicorn, B. kleini: being the only species of the genus (and one of the comparatively few members of the Combretaceae) known outside the tropics. Its vernacular name is given as ‘‘ Guarajuba ”’. B. kleinii appears to be most closely related to B. capitata, but its puberulous fruits and characteristic leaf-shape make it quite easily identifiable. In the shape of its endocarp it approaches B. parvifolia more closely than B. capitata. 4. Buchenavia sericocarpa Ducke in Bol. Técn. Inst. Agron. Norte 4 : 23 (1945). Leaf-lamina subcoriaceous, 3-12 X I-6 cm., narrowly elliptic to elliptic, short- to rather long-acuminate at the apex, cuneate and narrowed into the petiole at the base, very sparsely pubescent to almost glabrous when mature (except for appressed pubescence on the midrib and lateral nerves on both surfaces) ; midrib and lateral nerves conspicuous and raised below; major lateral nerves 5-10 pairs; petiole distinct, c. 7-30 mm. long, appressed-fuscous-pubescent. Peduncle fairly stout in fruit, c. 20-25 mm. long, rather sparsely appressed-pubescent ; rhachis extremely short, up to c. 8 mm. long. Fruit narrowly ellipsoid-oblong, more or less terete, Cc. 20-25 X 7-I0 mm. at maturity, rounded to very broadly obtuse at both ends, densely but shortly appressed-pale-tawny-pubescent. BrAzIL: Amazonas: Manaos, Estrada do Bombeamento, ‘‘ Capoeirdo, terra firme, arvore pequena ”’, 26 Nov. and 30 Dec. 1943, Ducke r48r (K ; RB, holotype, not seen). Known only from the type gatherings (presumably made from the same tree). This species is very different from all others of the genus with a capitate inflores- cence in its highly distinctive leaves. There is a remarkable similarity between this species and B. acuminata, so much so that there is scarcely any other difference that we can detect apart from the elongate inflorescences of the latter. This could be a striking example of parallel evolution but would seem to be more probably BOT; 35-3. 1§§§ 16 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA the comparatively recent evolution of a species by the appearance of a single but very distinctive character expressing one of the fundamental tendencies in the family towards massing of the flowers in a condensed inflorescence. A close parallel is shown in the relationship between Combretum punctatum subsp. sguamosum (Roxb. ex G. Don) Exell and C. punctatum Blume subsp. punctatum, but in that case the difference in the form of the inflorescence is not so clear-cut and is considered to be of no more than subspecific value. 5. Buchenavia ochroprumna Eichl. in Flora 49:165 (1866); in Mart., FI. Brasil. 14 (2) : 96 (1867). Leaf-lamina subcoriaceous when mature, c. 2-9°5 X I-4°5 cm., oblanceolate to obovate, usually rounded but sometimes very broadly obtuse or (when damaged?) retuse at the apex, cuneate and somewhat decurrent into the petiole at the base, almost glabrous when mature except on the midrib and lateral nerves; major lateral nerves 3-7 pairs; petiole rather short, 4-15 mm. long, usually rather stout. Peduncle 6-30 mm. long, slender at flowering time, stout in fruit. Fruit ellipsoid or ovoid, indistinctly 5-angled, markedly and irregularly ridged when dry, 18-25 x 7-15 mm. including beak, long-beaked at the apex, rounded at the base, densely but minutely ferrugineous-tomentose ; beak 4-8 mm. long, narrow, straight or somewhat curved; endocarp markedly longitudinally ridged (more so than in B. capitata), often about 5 ridges more prominent than the rest. BRAZIL : Amazonas: Manaos, Igarapé da Cachoeira Grande, “igapd profunde inundato, arbor parva’’, 28 Mar. 1937, Ducke 445 (K) ; 28 Mar. 1937, Ducke 35172 (K). Mandos, “‘igapo ad faucem fluminis Taruma Grande’’, 7 Apr. 1932, Ducke 25019 (K). Rio Negro, in the vicinity of Barcelos, 19 Apr. 1952, Frées 28438 (BM). Region of Parintins, Lago do Juruti, 18 Jan. 1957, Frdes 33069 (BM). Parad: Near Santarém, Jan. 1850, Spruce 309 (M, lectotype) and without n. (BM; K). Santarém, by river, “‘ bush of ro feet’, Jan. 1850, Spruce 619 (K). Lago Cugari, plateau of Santarém, 15 Apr. 1955, Frdes 31671 (BM). Igarapé Cugari, plateau of Santarém, 15 Apr. 1955, Frdes 31773 (BM). Mata da Pirelli, Ananindeua, 8 Jan. 1958, Silva 501 (BM). Rio Aramun, Almeirim, 2 Sept. 1918, Ducke 17287 (BM). This species is thus known only from the lower regions of the Amazon. Eichler cited only Spruce 309, which he saw in Herb. Leningrad and Herb. Munich. We have seen the latter specimen, which we select as the lectotype. Ducke 17287, labelled B. oxycarpa, is almost certainly B. ochroprumna as shown by the rather distinctive leaves. Besides their shape, the leaves of B. ochroprumna are usually characterized by a ferrugineous colour on their under surfaces when dry. This specimen is the only one we have seen in flower. The flowers do not appear to differ from those of B. capitata. B. ochroprumna differs from all other species of the genus in its highly character- istic fruits, and may also be distinguished by its leaves. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 7 6. Buchenavia punctata Eichl. in Flora 49: 166 (1866); in Mart., Fl. Brasil. 14 (2) : 98 (1867). Terminalia punctata Spruce ex Eichl., loc. cit. (1867), nom. syn.; non T. punctata Roth (1821) nec Eichl., tom. cit. : 85 (1867). Leaf-lamina chartaceous when mature, c. 6-I5 X 2:5-4:5 cm., conspicuously but minutely pellucid-punctate, very narrowly obovate to oblanceolate, acute to sub- acute at the apex, very narrowly cuneate but not decurrent into the petiole at the base, very sparsely pubescent above, pubescent on the midrib and nerves and sparsely pubescent elsewhere below ; midrib and major lateral nerves raised and very con- spicuous below; reticulation rather inconspicuous ; major lateral nerves c. 7—II pairs; petiole short and stout, 5-15 mm. long, eglandular, pubescent. Peduncle stout, c. 35 mm. long; rhachis long (over 25 mm., but material fragmentary). Fruit ellipsoid, radially symmetrical, markedly 5(6)-ridged along its length, 27-31 x 12-19 mm., conspicuously apiculate at the apex, narrowed into a very short but distinct pseudostipe at the base, very sparsely pubescent between the ridges, glabrous else- where ; apiculus 1-3 mm. long. PERU: San Martin: In rocky stream, near Tarapoto, ‘“‘ Arbor 25 pedalis ramis paucis longis validis brevi-ramulosis ’’, 1855-56, Spruce 4945 (BM; BR, lectotype, not seen; K). This species is very distinct from all others of the genus in both its fruits and leaves. Only B. pterocarpa has fruits as conspicuously ridged. Since the flowers are unknown we cannot, of course, be absolutely certain that this species belongs to Buchenavia, but as the fruit is distinctly less winged than that of any American species of Terminalia we have seen it seems safe to include it in Buchenavia. Its locality is completely isolated from that of any other Buchenavia yet discovered and is one of only two records of the genus for Peru. 7. Buchenavia reticulata Eichl. in Flora 49: 166 (1866); in Mart., Fl. Brasil. 14 (2) : 98 (1867). Leaf-lamina more or less chartaceous when mature, (2) 5-26 (36) x I-7 (12) cm., narrowly elliptic to oblanceolate, obtuse to acute or shortly apiculate at the apex, narrowly cuneate but not decurrent at the base, ferrugineous-tomentose over the whole of the lower surface at flowering time and during fruit maturation but tomen- tum wearing off after fruit-fall, ferrugineous- to rubiginous-tomentose or -pubescent on midrib and lateral nerves of upper surface at flowering time but indumentum wearing off during fruit maturation, sparsely pubescent and becoming glabrous elsewhere on upper surface ; midrib very conspicuous and markedly raised below ; major lateral nerves conspicuous and markedly raised below, 5-15 pairs, arising at c. 45° or less ; reticulation at and after flowering time prominent and raised below, mostly consisting of secondary venules connecting the major lateral nerves ; petiole distinct but usually short, c. 6-23 mm. long, stout, ferrugineous- or rubiginous- tomentose, most often conspicuously biglandular towards the apex. Peduncle 18 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA fairly stout, c. 20-30 (40) mm. long, ferrugineous- or rubiginous-tomentose ; rhachis c. 50-140 mm. long, ferrugineous- or rubiginous-tomentose ;_ flowers rather large, c. 4-5 mm. long. Lower hypanthium ferrugineous- to rubiginous-tomentose round the ovary, almost glabrous and c. 1-2 mm. long above it ; upper hypanthium c. I'5-2 X 2°5-3°5 mm., almost or quite glabrous outside. Fruit narrowly ellipsoid to ellipsoid, irregularly and rather slightly longitudinally furrowed or scarcely so, sometimes slightly flattened, c. 19-28 x 8-12 mm., apiculate to acuminate at the apex, shortly (1-2 mm.) pseudostipitate at the base, shortly ferrugineous- to rubi- ginous-tomentose and rather velvety all over ; apiculus 1-6 mm. long, straight or slightly curved ; endocarp variable, very slightly to conspicuously longitudinally furrowed, more or less circular in section to rather flattened, sometimes with 2 very narrow lateral wings. VENEZUELA: Amazonas: On the Rivers Casiquiare, Vasiva and Pacimoni, 1853-54, Spruce 3453 in part (BM ; BR, lectotype, not seen; K). Along the River Pacimoni, frequent, and also frequent along the Rivers Guainia and Vasiva, Feb. 1854, Spruce 3453 in part (K, isolectoparatype). Frequent on banks of the River Guainia, “‘ Arbor parva patula 18 pedalis. Fructus molles ovali-fusiformes acuti haud angulati. This and Parkia auriculata are perhaps the commonest riparial trees of the rivers Guainia, Pacimoni, Vasiva and Atabapo”’, May 1854, Spruce 3453 in part (K, isolectoparatype). Cafio Yapacana, Cerro Yapacana, Alto Rio Orinoco, “‘ tree 3-10 m. high, fruit brown, frequent at laguna edge’, 19 Mar. 1953, Maguire & Wurdack 34584 (BM). Frequent along Caio Yapacana, below port for Cerro Yapacana, Rio Orinoco, “ tree to 20 m. high, inflorescence brown ’’, 6 Jan. 1951, Maguire, Cowan & Wurdack 30760 (BM). BRAZIL: Amazonas: Above Kibaru, Rio Negro, 23 June 1874, Tvaill 216 (K). Above Santa Izabel, Rio Negro, 5 Mar. 1936, Ducke 35167 (K). Riosinho Juruema, Alto Rio Jutahy, Solimoes, “ Terra firme, igapés. Arvore’’, 4 June 1945, Foes 21039 (K). Igarapé Imuta, tributary of Rio Negro, ‘terra firme. Arvore de 8 m., ramos inclinados ; flores roseas’’, 18 Mar. 1952, F’vdes 27936 (BM). Rio Negro, May 1947, Froes 22323 (BM). From Spruce’s records for Rio Guainia and Rio Atabapo it seems probable that this species also occurs in Colombia (see footnote under Ramatuella argentea con- cerning Spruce 3498, p. 41 below). This very distinct species has very large or large leaves although it is a rather small tree. It is apparently abundant in the upper Rio Negro-Orinoco region (fide Spruce 3453), and thus entirely separated from the other two species with reticulate tomentose leaves (B. tomentosa and B. callistachya) which occur only in the lower regions of the Amazon basin and further south (Fig. 2). The fruits of this species show a remarkable range in the form of the endocarp. In some specimens it is round in section and only slightly longitudinally furrowed, while in others it is conspicuously flattened, markedly longitudinally ridged, and has two lateral wings, although these are not as well developed as in B. pterocarpa. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELEA 19 8. Buchenavia tomentosa Eichl. in Flora 49: 166 (1866); in Mart., Fl. Brasil. 14 (2) : 97 (1867).—Glaziou in Mém. Soc. Bot. France 1 (3) : 203 (1908). Terminalia tomentosa Mart. ex Eichl., loc. cit. (1867), nom. syn.; non T. tomentosa Wight & Arn. (1834). Buchenavia corrugata Ducke in Arch. Jard. Bot. Rio Janeiro 4 : 150 (1925). Leaf-lamina chartaceous at flowering time, coriaceous when mature, c. 3°5- 13°5 X I°5-7 cm., narrowly elJiptic or oblanceolate to (probably always at maturity) elliptic or obovate, rounded at the apex but obtuse or more pointed when very young, cuneate at the base, often inrolled at the margin, ferrugineous- or rubiginous- tomentose or densely pubescent all over both surfaces at flowering time, indumentum beginning to wear off soon afterwards ; midrib very conspicuous and markedly raised below ; major lateral nerves conspicuous and somewhat raised below, c. 9-14 pairs, arising at about 45°; reticulation not prominent before flowering time but rather conspicuously so afterwards, though less so than in B. reticulata and B. callistachya, less regular than in B. reticulata ; petiole distinct, c. 3-18 mm. long, stout, densely rubiginous- or ferrugineous-velutinous when young, sometimes biglandular. Peduncle c. 6-20 mm. long, usually stout, densely rubiginous-velutinous when young ; rhachis c. 3-11 cm. long, densely rubiginous-velutinous when young ; flowers very like those of B. reticulata, c. 4-5 mm. long, indumentum identical. Upper hy- panthium c. 2-5-4 mm. across. Fruit ellipsoid to more or less spherical, probably more or less terete when fresh, 22-31 X I5-25 mm., more or less rounded at both ends, rather sparsely ferrugineous-pubescent when mature. Brazi_: “ Brasilia’’, Pohl 83 (W, lectoparatypes). Pard: Serra Pontada, in the Jutahy region between Almeirim and Prainha, “silva, arbor magna’’, 11 Sept. 1923, Ducke 17677 (K, isotype of B. corrugata). Santa Cruz dos Martirios, regiao do Araguais, “‘ arvore de ro m., frutos comestiveis ’’, 15 June 1953, F'vdes 30018 (BM). Piaut: Banks Gurgia (= Rio Gurgeia ?), “a large tree called Biriba—fruit acrid and bitter—liked by deer ’’, Aug. 1839, Gardner 2657 (K). Goids : Between Alegre and Ponte do Sevorino, “ arbre de petite taille ’’, 2 Oct. 1895, Glaziou 21125 (K; P). In woods and plains, Serra da Chapada, “ arbor 15-20 ped.”’, Sept. 1827, Martius 1180 (LE, lectoparatype). Minas Gerais: Biribiry (fide Glaziou), 1892, Glaziou 19144 (K). On sandy plains between Alegres and Rio Sao Francisco, “‘ arbor 8-20 ped.”’, Sept. 1834, Riedel 2641 (BM; K; LE, lectotype and lectoparatype ; W). Glaziou 21125 is unlocalized in Herb. Kew, but the Paris specimen bears a label concurring with the locality cited by Glaziou (1908). This sheet also bears a label identical to that of B. macahensis (nomen nudum ; see p. 37 below), Glaziou 18218, but the actual specimen bears a tag numbered 21725 and other pieces of evidence conclusively show that the presence of the second label is due to an error in mounting. Eichler (1866) cited specimens of B. tomentosa in Herb. Munich, Herb. Martius, Herb. Vienna and Herb. Leningrad, but those in the first two herbaria do not appear to have survived. Of the three specimens in Herb. Vienna (W) one is a duplicate of one of the Leningrad specimens whilst the other two (both Pohl 83) are mounted on the same sheet. There are also three specimens in Leningrad : one collected by Martius and two (duplicates) by Riedel. We have selected the better Riedel 20 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA specimen as the lectotype as duplicates of this are also present in the British Museum, Kew and Vienna Herbaria. B. corrugata shows considerable similarity to B. callistachya, but the differences noted by Ducke and ourselves appear to be sufficient to keep the two separate, at least for the present. Comparison of them with the type specimens of B. tomentosa leads us to the opinion that B. corrugata is conspecific with the latter. B. tomentosa seems to be a species which does not extend up the Amazon, in con- trast with B. reticulata which is confined to the upper Amazon and Orinoco tributaries (Fig. 2). The vernacular name is noted as ‘‘ Tanebuco ”’ by Pohl. g. Buchenavia callistachya Ducke in Arch. Inst. Biol. Veg. Rio Janeiro 2 : 64 (1935). Leaf-lamina coriaceous when mature, c. 5*5-I0°5 X 2°5-5°5 cm., elliptic to slightly obovate, rounded at the apex, cuneate at the base, inrolled at the margin when young, lower surface at flowering time densely ferrugineous-pubescent except tomen- tose on the midrib and nerves, indumentum beginning to wear off soon afterwards but midrib and nerves densely pubescent below even after fruit-fall ; midrib very conspicuous and markedly raised below ; lateral nerves also markedly raised below, c. 8-11 pairs, arising at 45° or slightly more ; reticulation very prominent below after flowering time, less regular than in B. reticulata ; petiole distinct, stout, ferru- gineous- to rubiginous-tomentose at flowering time, c. 6-16 mm. long, sometimes bi- glandular. Peduncle c. 10 mm. long or less at flowering time, up to 15 mm. long in fruit, stout, densely rubiginous-velutinous at flowering time, densely pubescent in fruit ; rhachis 15-25 mm. long at flowering time, up to 45 mm. in fruit ; flowers smaller than in B. reticulata and B. tomentosa, c. 2-5-3 mm. long, indumentum as in the two preceding species. Upper hypanthium c. 1-75-3 mm. across. Anthers also smaller. Fruit ellipsoid wrinkled when dry, probably more or less terete when fresh, 21-23 X I2-I5 mm., rounded at the apex, rounded or very shortly pseudo- stipitate at the base, finely ferrugineous- to rubiginous-puberulous when mature. BraziL: Amazonas: Manaos, near Estrada do Aleixo, in non-flooded wood, ‘“ Arbor sat. magna, spicae anthesi pallide viridiflorae pedunculo brunneo-purpureo”’, g July and 26 Nov. 1932, Ducke 25021 (K; RB, holotype, not seen). Maranhao : Ilha dos Botes, Rio Tocantins, near Carolina, 25 May 1950, Pires & Black 2095 (BM). B. callistachya is clearly closely allied to B. tomentosa, from which it differs in a few characters which remain constant with regard to the rather few specimens of each species known. The recently collected fruiting specimen (Pires & Black 2095) almost certainly belongs to this species, its leaves being very similar to those of the type. It is just possible, however, that we have confused the fruiting specimens of B. callistachya and B. tomentosa, or that B. callistachya itself may prove to be conspecific with the latter. These points must await more and better material for elucidation. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 21 10. Buchenavia suaveolens Eichl. in Flora 49 : 166 (1866) ; in Mart., Fl. Brasil. 14 (2) : 97 (1867). (Fig. 3 e, f) Buchenavia oxycarpa Eichl. in Flora 49: 165 (1866) pro parte, quoad Schomburgk 854 ; in Mart., loc. cit. (1867) pro parte, quoad Schomburgk 854. Terminalia suaveolens Spruce ex Eichl., loc. cit. (1867), nom. syn. Terminalia vasivae Spruce ex Eichl., loc. cit. (1867), nom. syn. (Spruce also used a third manuscript name which was not cited by Eichler.) Leaf-lamina coriaceous when mature, 1-8-8 x 0-3-4:3 cm., narrowly elliptic or oblanceolate to elliptic or obovate, retuse to rounded at the apex, narrowly to very narrowly cuneate and somewhat decurrent into the petiole at the base, almost glabrous when mature except on the midrib; midrib fairly conspicuous, slightly raised below ; major lateral nerves inconspicuous to very inconspicuous, but usually over 6 pairs visible below, scarcely or not raised ; reticulation not visible ; petiole fairly long and distinct, c. 3-20 mm. long, very sparsely and minutely pubescent, eglandular. Peduncle usually rather slender, sparsely pubescent, up to c. 25 mm. long ; rhachis up to 65 mm. long at flowering time, sparsely pubescent; flowers scarcely different from those of B. capitata, but ovary usually rather more sparsely pubescent. Fruit ellipsoid to slightly obovoid, more or less terete when fresh, 24-30 X IO-I4 mm., conspicuously apiculate at the apex, narrowed into a short but distinct pseudostipe at the base, glabrous, the apiculus 1-3 mm. long, more or less straight ; endocarp ellipsoid, almost round in section, c. 20 X 9 mm., obtuse at each end, more or less smooth. VENEZUELA: Amazonas: On the rivers Casiquiare, Vasiva and Pacimoni, 1853- 54, Spruce 3198 (BM; K; LE, lectoparatype). On the banks of Lake Vasiva, also observed on Rio Pacimoni, Dec. 1853, Spruce 3190 (K). Rio Sanariapo, 120 m., 2 July 1942, Willams 15955 (F). VENEZUELA or COLOMBIA: On the banks of Rio Guainia, “‘ Drupa viridis carnosa, recens vix obsolete prismatica’”’, June 1854, Spruce 1887 in part (K). On Rio Guainia above its confluence with Rio Casiquiare, 1854, Spruce 1887 in part (K). (See footnote under Ramatuella argentea concerning Spruce 3498, p. 41 below.) BrAziL: Amazonas: Rio Negro, between Barra and Barcellos, Nov. 1851, Spruce 1887 in part (BM; K; LE, lectotype). Caloruca, Rio Preto, tributary of Rio Negro, 19 Apr. 1952, Frdes 28309 (BM). Rio Branco: ‘‘ Guiana ”’ (but actu- ally by Falls on the Rio Branco, fide Schomburgk, MS.), 1839, Schomburgk 854 (BM; K). “Guiana Britannica’’, 1839, Schomburgk 814 (BM) (probably error for 854). B. suaveolens is thus confined, like B. pterocarpa and B. reticulata, to the upper Amazon-Orinoco region (Fig. 2). The distinction between B. suaveolens and B. plerocarpa is made under the latter species. Eichler (1866) cited specimens in Herb. Leningrad and Herb. Martius, of which we have seen the former. The two gatherings are Spruce 1887 and Spruce 3198 (which Eichler gives in error as 3789), and there are duplicates of both at the British Museum and Kew. We have selected Spruce 1887 as the lectotype as there 22 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA LMR "4 Fic. 3. Buchenavia pterocarpa Exell & Stace (holotype) : a, habit (x %) ; 6, c, two views of fruit (x1); d, fruit in transverse section (x1). B. suaveolens Eichl. (Frdes 28309) : e, fruit (x1); f, fruit in transverse section (x 1). REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 23 is no confusion about the number and other specimens (from another locality) also with this number bear fruits. Schomburgk 854 is labelled ‘“‘ Guiana’’ but according to the collector’s note- books he was actually at the Falls of the Rio Branco, Brazil. Schomburgk 814 is similarly labelled and since his note-book states that this number was a herbaceous plant it seems most probable that 8r4 is an error for 854, which it resembles in every detail. B. suaveolens is otherwise not known from British Guiana. Eichler (1866, 1867) cites Schomburgk 854 under B. oxycarpa (in 1867 as an unnamed variety), but this is clearly a misidentification. 11. Buchenavia pterocarpa Exell & Stace, sp. nov. (Fig. 3 a-d.) Arbor parva ad c. Io m. alta; ramuli griseo-fusci, fere glabri, apice non clavati. Folia spiraliter ordinata ad ramulorum apices congesta; lamina subcoriacea, I*5-9°5 X 0°5-2°8 cm., anguste elliptica vel oblanceolata ad anguste obovata, in- tegra, apice obtusa rotundata vel retusa, basi acuta decurrens, fere glabra, costa media infra prominenti, nervis lateralibus inconspicuis; petiolus c. 3-18 mm. longus, sparse minuteque puberulus eglandulosus. IJnflorescentiae axillares, elongato- spicatae ; pedunculus gracilis, c. to-17 mm. longus, sparse minuteque puberulus ; rhachis c. 15-53 mm. longa, puberula ; flores ferrugineo-virides ; bracteae parvae, ante anthesin caducae. Hypanthium inferum gracile, I-2 mm. longum, ovario in parte basali puberula incluso, parte apicali sparse minuteque puberula ; superum Cc. I-I'5 X 2-3 mm., extus sparse puberulum, intus pubescens, margine saepe fere integrum. Stamina 10, ad o-5-r mm. exserta. Stylus ad c. I°5 mm. exsertus, filiformis, glaber. Fructus (plerumque singulus tantum pro infructescentia matur- escens) latissime applanato-ellipsoideus vel obovoideus, c. 14-28 * 10-16 mm. et 5-8 mm. crassus, apice abrupte apiculatus, basi rotundatus vel breviter (ad 3 mm.) pseudostipitatus, glaber, apiculo ad 4 mm. longo curvato, alis 2 lateralibus angustis et angulis 3 prominentibus instructus ; endocarpium lignosum, lateraliter bialatum, irregulariter longitudinaliter sulcatum. VENEZUELA: Amazonas: Occasional along Rio Yatua, near Laja Catipan, Casiquiare, 100-140 m., “ flat-topped riverine tree, 6-8 m. high, fruit green ’”’, 6 Feb. 1954, Maguire, Wurdack & Bunting 37543 (BM, holotype). Common and subdominant along Cana Catua, Cerro Yapacana, Rio Orinoco, 125 m., 19 Nov. 1953, Maguire, Wurdack & Bunting 36555 (BM). Cano Yapacana, below port to Cerro Yapacana, Rio Orinoco, 6 Jan. 1951, Maguire, Cowan & Wurdack 30763 (BM). BRAZIL: Amazonas : Santa Izabel, Rio Negro, in flooded riparian forest, 8 Oct. 1932, Ducke 25018 in part (K) ; same locality, 9 Mar. 1936, Ducke 25018 in part (K). Igarapé Imuta, tributary of Rio Negro, ‘‘ terra firme’”’, 12 Mar. 1952, Froes 27937 (BM). All the specimens we cite, other than Frydes 27937, were previously labelled B. suaveolens. In the absence of fruits it does not always seem to be possible to 24 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA distinguish between B. suaveolens and B. pterocarpa, but in most cases the leaves of the former are broader and more markedly obovate. The fruits of the two species are, however, quite different (see Fig. 3). The fruit of B. suaveolens is more or less terete, narrower in proportion to its length and with a usually straight apiculus abruptly delimited from the obtuse or rounded apex. In our new species on the other hand the fruit is considerably flattened with two narrow lateral wings and three extra ridges (one on one face and two on the other), is broader in comparison with its length and has a usually curved apiculus very abruptly delimited from the usually rounded or truncate or even retuse apex. The endocarp is also markedly different in the two fruits, that of B. plerocarpa having two lateral wings and an irregularly furrowed surface, while that of B. suaveolens is smooth and almost round in cross-section. The bilateral symmetry of the fruit of B. pterocarpa is not found so greatly developed elsewhere in the genus, although some species (notably B. reti- culata and less conspicuously B. parvifolia and others) show tendencies towards it. Comparison of recent collections of B. suaveolens and B. pterocarpa with Spruce’s collections of the former shows that it is the flattened-fruited species which needs to be described as new. Although Eichler did not see fruits of B. suaveolens (none of his cited specimens is in fruit, and he states “‘ Fructus ignotus ”’ in Fl. Brasil.) it seems clear that it was the terete-fruited species that he described. The evidence for this is that both Leningrad specimens are broad-leaved, and that other specimens of Spruce 1887 (not duplicates of those which Eichler cited from Barra and Bar- cellos) from the Rio Guainia which are in Herb. Kew possess the distinctive terete fruits. 12. Buchenavia oxycarpa (Mart.) Eichl. in Flora 49 : 164, t. 3 figs. 17, 18, 20, 21 (r866) excl. Schomburgk 854; in Mart., Fl. Brasil. 14 (2) : 97, t. 25 (1867) excl. Schomburgk 854. Terminalia oxycarpa Mart. in Flora 24, Beibl. 2 : 22 (1841). Leaf-lamina chartaceous, I-13 X 0°5—5 cm., very narrowly elliptic or oblanceolate to obovate, rounded to acute at the apex, narrowly cuneate but scarcely decurrent at the base, almost glabrous when mature except on the midrib ; midrib fairly con- spicuous, slightly raised below ; major lateral nerves fairly conspicuous, very slightly raised below, (4) 6-9 (12) pairs ; reticulation fine but fairly conspicuous ; petiole very distinct, 4-25 mm. long, sparsely and minutely pubescent to almost glabrous, eglan- dular. Peduncle slender, rather sparsely pubescent, up to 55 mm. long; flowers very similar to those of B. capitata but ovary pubescent or more usually very sparsely pubescent to almost glabrous, and always glabrous soon after flowering. Fruit ellipsoid, wrinkled and slightly flattened to conspicuously 5-6 (7)-angled when dry, 19-30 X 9-I5 mm., abruptly acuminate at the apex, rounded to obtuse at the base, glabrous ; acumen 3-9 mm. long, pointed, more or less straight or slightly curved ; endocarp strongly longitudinally ridged, 5 ridges stronger than the rest, more or less radially symmetrical but 2 lateral ridges sometimes the most prominent. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA_ 25 Brazi_: “Brasilia’’, without date, Martius (M, lectoparatype). ‘‘ Brazil’’, with- out date, Poeppig 2689 ex Herb. Van Heurck (BM). “‘ Brasilia’’, without date, Riedel 744 ex Herb. Hort. Petrop. (K). Amazonas: In inundations on the bank of the river Solimoes, near Ega and elsewhere in the province of Rio Negro, 1820, Martius (M, lectotype). ‘‘ Prov. Rio Negro prope Ega ad ripam fluv. Solimoes ’’, without date, Martius (K; M; lectoparatypes). ‘In ripa fluv. Solimdes haud infrequens’’, 1820, Martius (M, lectoparatype). Humayta, near Tres Casas, basin of Rio Madeira, “tree 40 ft. high’’, 14 Sept.—rz Oct. 1934, Krukoff 6174 (BM; K). Humayta, near Livramento, on Rio Livramento, basin of Rio Madeira, on immediate shore of river, “shrub 20 ft. high’, 12 Oct.—6 Nov. 1934, Krukoff 6733 (BM; F; K). Parana do Careiro (mouth of the Solimées), “‘ Arvore pequena ; flér pardo vermelho com estamos brancos’’, 26 Oct. 1946, Ducke 2023 (BM). S. Paulo de Olivenga, igapé de Camatia, 2 Feb. 1937, Ducke go2 in part (K). Para: Near Santarém and Paricatuba, “silvae ad fluv. Amazonum locis inundatis’’, without date, Martius 2852 in part (M, lectoparatype). Near Santarém and Paricatuba, “In sylvis ad fluv. Amazonum in locis inundatis”’, without date, Martius (M, lecto- paratype). Near Santarém, Nov. 1849—Mar. 1850, Spruce 127 (M) and without n. (BM ; K). Rio Tapajoz, Santarém, “‘ small tree, lvs. and fruits fastigiate on short ramuli”’, Mar. 1850, Spruce 729 (K). On varzia land near Cassipa, Tapajoz River region, Sept. 1931, Krukoff 1241 (BM; K). Pard/Minas Gerais : “ Para et Minas Geraes, in sylvis ad rip. Amazonum, nec non Rio Fermozo’’, without date, Martius 2852 in part (mixed with Combretum laurifoliwm Mart.) (M, lectoparatype). Peru: Loreto: Inundation belt of Rio Itaya, Iquitos, “‘ Shrub, flowers reddish- brown ”’, 4 Nov. 1946, Asplund 14302 (K). Out of Martius’s eight syntypes we have selected a specimen possessing leaves, flowers and fruits as the lectotype. B. oxycarpa is a very distinct species, although Eichler (1866, 1867) cited a gather- ing of B. suaveolens (Schomburgk 854) under it. The endocarp of the present species is markedly ridged and more or less radially symmetrical, a feature only as greatly developed elsewhere in B. punctata. The fruits are markedly ridged in some speci- mens (e.g. in Martius’s syntypes) and in Eichler’s figures. Spruce’s specimens show no sign of these ridges, having slightly flattened and minutely wrinkled fruits. Fruits of Ducke 402 are, however, more or less intermediate in this feature. Again, in Krukoff 1241 and Krukoff 6174 the ovaries are completely glabrous at anthesis or just after, while those of Asplund 14302 are ferrugineous-tomentose. The ovaries of the flowers of Riedel 744 are completely intermediate in pubescence. The in- florescences of Asflund 14302 are additionally unusual in that the whole peduncle and rhachis are tomentose ; the long pedicel-like distal portion of the lower hypan- thium and the outside of the upper hypanthium are, however, glabrous. Eichler also recorded the species from the state of Mato Grosso (on banks of Rio Paraguay), Brazil. Since he cites Riedel as the collector from this locality, the specimen in question may well be the Riedel specimen referred to above under un- localized records for Brazil. 26 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 13. Buchenavia discolor Diels in Verh. Bot. Ver. Prov. Brandenb. 48: 192 (1907). Leaf-lamina chartaceous or coriaceous, c. 6-7-5 x 4.cm., obovate, obtuse to emar- ginate at the apex, narrowly cuneate at the base, sparsely pubescent below ; major lateral nerves prominulous below, 3-4 pairs, venules inconspicuous; petiole 10-15 mm. long. Peduncle plus rhachis 20-30 mm. long. Fruit ellipsoid, 5-angled, c. 15 mm. long excluding acumen, abruptly acuminate at the apex, tomentellous ; acumen up to 10 mm. long. (Ex descr. orig.) BRAZIL: Amazonas: Near Mandaos, on the banks of Rio Negro, Dec. 1901, Ule 5979 (B, holotype, destroyed). The only known specimen, which was in the Berlin Herbarium, has been destroyed. The species appears to be very distinct. Diels states that B. discolor is close to B. oxycarpa but differs in the structure of the leaves (mainly in having only three or our pairs of lateral nerves) and in the tomentellous fruit. It is quite possible, if the inflorescences have been misinterpreted by Diels, that this species is the same as B. ochroprumna, which has capitate inflorescences. 14. Buchenavia megalophylla Van Heurck & Muell. Arg. in Van Heurck, Obs. Bot. > 2r1 (1870). Leaf-lamina subcoriaceous when mature, I4-3I X 5-10 cm., narrowly obovate to oblanceolate, obtusely apiculate to shortly acuminate at the apex, very narrowly cuneate at the base but not decurrent into the petiole, almost glabrous above except on the midrib and lateral nerves, shortly pubescent on the midrib and lateral nerves below, almost glabrous to very sparsely puberulous elsewhere below ; midrib very prominent and raised below, slightly sunken above; lateral nerves conspicuous below, (9) 10-14 pairs ; reticulation quite conspicuous, mostly connecting the lateral nerves ; petiole distinct and stout, 12-23 mm. long, usually conspicuously bi- glandular towards the apex, shortly pubescent. Peduncle rather stout, densely puberulous, c. 20-30 mm. long ; rhachis fairly stout, densely puberulous or tomen- tellous, c. 70-125 mm. long ; flowers rather large,c.4 x 3mm. Ovary tomentellous. Upper hypanthium almost glabrous outside. Fruit very large, 30-48 mm. long, rather abruptly narrowed into a long acumen at the apex and into a pseudostipe at the base, ferrugineous-tomentellous ; body of fruit ellipsoid, 18-26 * II-I13 mm., slightly compressed, conspicuously 5-ridged ; pseudostipe often relatively long and slender, 2~7 mm. long ; acumen rather various, straight or markedly curved, slender or rather stout (5, damaged ?) ro-20 mm. long. BRITISH GUIANA: Demerara: ‘‘ Demerara’’, without date, Parker (K, as Pamea guianensis). Essequibo: Mazaruni Station, “shrub or small tree 8 ft. high from sand bank by river ’’, rr Sept. 1942, Field No. F845 in part = Forest Dept. 3581 in part (K) ; same locality, rr Apr. 1943, Field No. F845 in part = Forest Dept. 3581 in part (K). The type of this species is said to be in “ herb. van Heurck ”’ (AWH) but we have been informed that it is not to be found there now. It was said (Van Heurck & REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 27 Muell. Arg., op. cit. : 212) to have been sent to Mlle. Hermine Reichenbach, probably from Demerara, under the name Pamea guianensis Aubl., but collector and date are not given. A specimen at Kew (cited above) bears the annotation ‘“‘ Mr. Parker Demerara. Allosohabbo yielding excellent timber. Pamea guianensis. Aubl. Terminalia Pamea. Aubl.” This was in Herb. Hooker, presented to Kew in 1867, three years before Van Heurck and Mueller’s publication. It seems very likely that the Kew specimen, with which Forest Dept. 3581 is obviously conspecific, is a duplicate of the specimen described by Van Heurck and Mueller. However, in the absence of confirmation of this and of the fact that Van Heurck’s specimen is actually missing we cannot give the Kew specimen the status of a type. A further difficulty is that in their lengthy description Van Heurck and Mueller state that the inflorescences of this species are capitate, whilst the specimens we have seen all have very long spicate inflorescences. The rest of the description, notably of the leaves and fruits, fits the Kew Demerara specimen exactly. Without seeing Van Heurck’s specimen it is not possible to account for this discrepancy with cer- tainty, but it seems probably that the specimen described had incomplete inflor- escences which had been broken off. We believe, like Sandwith (in sched.), that the inflorescences of this species were misinterpreted by Van Heurck and Mueller, and that they are in fact elongate spikes as stated in our description. There remains the unlikely possibility, however, that B. megalophylla has a truly capitate inflorescence and that the three Kew specimens represent a new species. 15. Buchenavia pallidovirens Cuatrec. in Fieldiana, Bot. 27 (1) : 107 (1950). Leaf-lamina subcoriaceous, 3-II xX I-3-7 cm., oblanceolate to narrowly elliptic or oblong-lanceolate, abruptly and sharply apiculate or mucronate or shortly acuminate at the apex, narrowly cuneate but scarcely decurrent into the petiole at the base, almost glabrous above except sparsely puberulous on the midrib and lateral nerves, very sparsely puberulous or almost glabrous below except sparsely pubescent or pubescent on the midrib and lateral nerves ; midrib conspicuous and raised below ; lateral nerves conspicuous and raised below, 5-7 pairs, very dis- tinctive, arising at about 40° or less to the midrib and then curving rather abruptly inwards towards the apex and running at about 25° or less to the midrib for most of their length and nearly parallel with the leaf margin at their apex ; reticulation con- spicuous, very slightly raised below, forming very small inter-reticular spaces ; petiole often long, 5-13 mm. long in flowering material, pubescent, obscurely bi- glandular at the extreme apex in most leaves, 10-26 mm. long in fruiting material, sparsely or very sparsely pubescent, not glandular or rather conspicuously bi- glandular at about the middle. Peduncle 25-33 mm. long in flowering material, 8-18 mm. long in fruiting material, slender, sparsely pubescent, rhachis c. 40-70 mm. long, slender, pubescent ; flowers c. 3 X 2:5 mm. Ovary densely pubescent. Upper hypanthium pubescent on the outside. Fruit (immature) ellipsoid-obovoid, probably terete when fresh, up to 12 * 6 mm., subacute to obtuse at both ends, glabrous. 28 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA CotomBIA: Valle: Costa del Pacifico, Bahia de Buenaventura, Quebrada de San Joaquin, 0-10 m., “‘ Arbol grande. Racimos péndulos verdoso amarillentos. Perianto verde amarillento claro. Anteras amarillas’’, 23 Feb. 1946, Cuatrecasas 19939 (BM; F, holotype, not seen). BraziL: Amazonas: Rio Vaupés, Panuré, catinga, “‘ arbusto”’, 15 Nov. 1947, Pires 1043 (BM). In the above description ‘‘ flowering material’ refers to Cuatrecasas 19939 and “ fruiting material ’’ to Pires 1043. These are the only two specimens known to us, and it is of interest to note that they are from rather widely separated areas. It is unfortunate that we have to compare a flowering specimen with a fruiting specimen from a distant locality, but Pives 1043 seems most probably to belong to the same species as the type. The peculiar nervation and reticulation of the leaves are com- mon to both specimens and we have not met with them elsewhere in the genus. The main point of difference is that the glands of Pives 1043 are, when present, near the middle of the long petiole, while in Cuatrecasas 19939 they are either absent or very inconspicuous and at the extreme apex of the petiole. 16. Buchenavia viridiflora Ducke in Arch. Inst. Biol. Veg. Rio Janeiro 2: 63 (1935). Leaf-lamina subcoriaceous, I-I0 X 0°5-5:5 cm. (but rarely over 8 x 4 cm.). oblanceolate or narrowly elliptic to obovate or elliptic, acute or subacute to rounded or retuse at the apex, narrowly cuneate and often somewhat decurrent into the petiole at the base, almost glabrous except sparsely to very sparsely pubescent on the midrib and sometimes on the lateral nerves above, almost glabrous to sparsely puberulous on the midrib and lateral nerves below ; midrib conspicuous and raised below ; lateral nerves conspicuous and raised below, 4—7 pairs ; reticulation rather inconspicuous ; petiole (4) 8-16 mm. long, pubescent, usually not glandular but some very small (not young) leaves markedly biglandular at the apex of the petiole. Peduncle 5-27 mm. long, slender, pubescent ; rhachis 12-50 mm. long, slender, pubescent ; flowers 3-4:5 X 2-3 mm. Lower hypanthium densely pubescent round the ovary, sparsely so above ; upper hypanthium almost glabrous outside. Fruit more or less terete, up to 24 X I3 mm., subacute to very shortly apiculate at the apex, rounded to subacute at the base, densely greyish-brown-tomentellous ; endocarp terete, slightly longitudinally ridged, acute at the apex, rounded at the base. BRAZIL: Amazonas: Near Cachoeira Grande, Manaos, forest in a high locality, “arbor mediocris flor. laete viridibus ’’, 31 July 1932 and 8 Jan. 1933, Ducke 25022 (K). Estrada do Aleixo, Manaos, in non-flooded forest, 15 July 1932, Ducke 25023 (K ; RB, lectotype, not seen). Estrada do Aleixo, Manaos, in non-flooded forest, 26 Dec. 1936 and 19 Mar. 1937, Ducke 426 (K). Rio Branco: Boa Vista, Rio Branco, non-flooded wood by Igarapé Caxanga, “ arbor mediocris floribus viridibus ”’, 24 Aug. (or July?) and 16 Sept. 1943, Ducke 1378 (F; K). Riverbank wood near Boa Vista, Rio Branco, ‘ Bl. rétlichgelb, Bm. 5-15 m.”’, Oct. 1908, Ule 7682 (K). Amapa: Rio Diapogue, Cachoeira Grande Rocha, “terra firme”, 19 Oct. 1950, Froes 22698 (BM). REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 29 The leaf-shape of this species is apparently variable. In Ducke 1378, 25022 and 25023 the lamina is usually oblanceolate with a rounded apex, but some leaves on the first of these have acute or subacute apices, showing a connexion with Ule 7682 which has larger, more or less elliptic acute leaves. Leaves on the fruiting branch of Ducke 25022 (fruits are actually lacking in the specimen at Kew) are obovate and rounded or retuse at the apex, which is the shape found in Frdes 22698. It seems, however, that all the specimens cited belong to this species. Some of the inflor- escences of the last specimen are very short, barely exceeding 3 cm. long with a floriferous part little over r cm. long. 17. Buchenavia acuminata Exell & Stace, sp. nov. Arbor ad 20 m. alta; ramuli griseo-fusci, sparse pubescentes, apice non clavati. Folia spiraliter ordinata ad ramulorum apices congesta ; lamina chartacea, 2-9'5 x 0-8-5 cm., anguste elliptica vel elliptica, integra, apice obtuse vel acute acuminata (acumine ad 15 mm. longo), basi acute cuneata, fere glabra costa media nervisque lateralibus pubescentibus exceptis, costa media infra prominenti, nervis laterali- bus infra prominentibus 5—8-paribus, venulis inconspicuis ; petiolus c. (4) 8-24 mm. longus, pubescens, plerumque apice conspicue biglandulosus. IJmnflorescentiae axillares, elongato-spicatae ; pedunculus gracilis, c. 15-30 mm. longus, sparse pubescens ; rhachis gracilis, c. I5-30 mm. longa, sparse pubescens ; flores parvi ; bracteae parvae, ante anthesin caducae. Hypanthium inferum gracile, c. 2 mm. longum, ovario in parte basali dense pubescenti incluso, parte apicali pubescenti ; superum Cc. I-I°5 X 2-2-5 mm., extus sparse puberulum, intus pubescens, margine saepe fere integrum. Stamina Io, ad 0-75 mm. exserta. Stylus inclusus (semper?). Fructus (pauci pro infructescentia maturescentes) ellipsoideus vel obovoideus, tere- tiusculus, immaturus ad II x 5 mm., apice rotundatus, basi breviter pseudostipi- tatus, dense argenteo-brunneo-tomentellus. _ Brazit: Amazonas: Humayta, near Livramento, on Rio Livramento, basin of Rio Madeira, on terra firma, “ tree 60 ft. high ’’, 12 Oct. -6 Nov. 1934, Krukoff 6916 (BM, holotype; F; K). Rio Urubu, Barreirinha, “terra firme’, ‘ Arvore de 7 m., frutos em espiguetas, amareladas ’’, 24 Aug. 1949, Frdes 25134 (BM). The rather small long-acuminate leaves, long slender petioles and densely tomentellous fruits make this species quite distinct from all others except B. serico- carpa, which has a capitate inflorescence. Apart from this single but very con- spicuous and easily recognizable difference we have been unable to separate these two species, and they occur in the same area of Brazil. It is possible that here is a case of a dimorphic species but until this is actually demonstrated it seems more advisable to treat the plant with the elongate inflorescence as a different (new) species. 18. Buchenavia fanshawei Exell & Maguire in Bull. Torr. Bot. Club 75 : 648 (1948). Leaf-lamina coriaceous at maturity, I-14 x 0-5-6 cm., narrowly elliptic or ob- lanceolate to elliptic or obovate, rounded to apiculate or less often acuminate or 30 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA acute at the apex, narrowly cuneate and often rather decurrent into the petiole at the base, when mature almost glabrous above except sparsely pubescent on the midrib, almost glabrous to very sparsely pubescent below except pubescent to sparsely pubescent on the midrib and lateral nerves ; midrib conspicuous and raised below ; lateral nerves conspicuous and raised below, 3-5 pairs ; reticulation rather — inconspicuous or fairly conspicuous, not close; petiole slender to rather stout, 6-24 mm. long, pubescent to sparsely pubescent, usually not glandular but very small leaves often conspicuously biglandular near top of petiole and some specimens with normal leaves also biglandular ; some or most leaves on mature-leaved speci- mens with conspicuous domatia (in the axils of the lateral nerves on the lower surface) with a conspicuous opening smaller than the diameter of the domatium itself. Peduncle slender at first, stouter in fruit, 6-30 mm. long, pubescent or puberulous ; rhachis 7-38 mm. long, puberulous, often densely so ; flowers yellowish-green, c. 2°5-4 X 2-4 mm., seen on a rhachis up to only 20 mm. long. Fruit rather variable, ovoid to ellipsoid- or oblong-ovoid but when ripe often becoming pressed almost circular in outline, more or less terete when fresh or slightly flattened, up to 24 x 14 mm., subacute to rounded at the apex or rarely apiculate or shortly acuminate at the apex with a beak up to 2 mm. long, rounded or shortly pseudostipitate at the base, very shortly but rather densely tomentellous, indumentum wearing off when very old; endocarp slightly flattened, slightly longitudinally ridged, acute to sub- acute at both ends. _ ‘ BRITISH GUIANA: Essequibo: Mazaruni Station, “ roo ft. tree 16 in. diam. from mixed forest .. . fr. oval to oblong, yellow when ripe—pulp whitish, slightly sweet— seed oval, whitish, veined, stone very hard’, 7 Sept. 1942, Field No. F844 = Forest Dept. 3580 (K). Mazaruni Station, “ 60 ft. tree 10 in. diam. from secondary mixed forest on brown sand... fls. in axill. revolute spikes, yellow-green—perianth shaped like a flat bowl, barely lobed ’’, 5 May 1943, Feld No. F1270 = Forest Dept. 4006 (K). Mazaruni Station, seedlings of Forest Dept. 3580 ten weeks old from forest nursery, no date, Field No. F'1616 = Forest Dept. 4352 (K). Mazaruni Station, seedlings (no source) three weeks and seven months old from forest nursery, no date, Field No. F2233 = Forest Dept. 4969 (K). Mazaruni Station, towards Labbakabra Creek, 27 Aug. 1937, Sandwith 1219 (BM; K). Bank of Potaro River, Tumatumari, 4-6 July 1921, Gleason gor (K). Riverside below Tukeit, Potaro River Gorge, rare, 16 May 1944, Maguire & Fanshawe 23499 (BM; K; NY, holotype). Mahdia River, Potaro River, 172 km. along Bartica-Potaro road, 16 Jan. 1943, Field No. F'1039 = Forest Dept. 3775 (K); same locality, 2r Jan. 1943, Field No. F 1064 = Forest Dept. 3800 (K). Basin of Kuyuwini River, tributary of Essequibo River, about 240 km. from mouth, 12 Feb. 1938, Smith 3033 (K). Berbice: New River, Courantyne River, 5 Oct. rg11, Anderson 749 (K). BRAZIL: Amazonas: Cachoeira Baixa, Rio Taruma, Mandaos, “ 4arvore mediana ; flr pardo avermelhado clara; fruto amarelado palido’’, May 1950, Ducke 2278 (BM). Although not recorded from the third British Guianan county, Demerara, or from Surinam, B. fanshawei probably occurs in both. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 31 B. fanshawet is a variable species, but is usually quite easily recognized in all of its states. It appears to be most closely related to B. macrophylla, although in their typical forms the two are extremely different in aspect. Some specimens of B. macrophylla, however, have very small leaves with few nerves and stout inflores- cences, while some specimens of B. fanshawei have acuminate leaves, biglandular petioles and undeveloped or no domatia. The two species apparently overlap only in the Mandos region of the middle Amazon (Fig. 2). Leaves on more juvenile branches of B. fanshawei (but not juvenile leaves on older branches) show characters more approaching those of B. macrophylla, since the leaf- apex is commonly acuminate and the petioles may be glandular. These branches flower freely, however, and do not appear to be suckers. The more mature branches are swollen at the tips with the leaves more densely tufted at the apex. The seed- lings (Forest Dept. 4352 and 4969) are interesting since the leaves have a long acute acumen, although they do not appear to be glandular. The cotyledons are about 2-2°5 by 3-4 cm., truncate or very shallowly retuse at the apex and very weakly cordate to rounded at the base. The domatia are of unique structure in the genus, although they are of a type commonly found in Tervminalia. On mature branches they are found usually on some or most of the leaves, in the axils of the major lateral nerves on the lower surface. They are often very small, but may be up to about 1:5 mm. across, and are fre- quently dome-shaped in general outline, with a pore either centrally placed or at the end of the domatium distal to the nerve axil. The pore is never as broad as the total diameter of the domatium. In all other known species of Buchenavia domatia, if present, are simple more or less triangular pocket-like structures formed by the epi- dermal and cortical layers of the midrib and nerve overarching the nerve axil. In these cases the pore is thus positioned at and therefore always as wide as the maxi- mum diameter of the domatium and faces parallel to the leaf surface. As in all species, the domatia of B. fanshawei are frequently marked by a persistent axillary tuft of hairs in the young stage. Very often, at an early stage, the domatium is a simple wart-like object, the pore appearing later by the rupture of the wart. This is not the case in the other types of domatium (e.g. in B. capitata). In a survey of domatia in the family Combretaceae, of thirty American species of Terminalia ex- amined twenty-three were found to lack domatia, three to possess domatia like those of B. capitata and four to have domatia like those of B. fanshawei. In Buchenavia, of the twenty-three species available for study (B. discolor being, as stated, unavail- able) nine (ns. 7-11, 14 and 19-21 of this revision) lack domatia, thirteen (ns. 1-6, 12, 15-17 and 22-24) possess domatia like those of B. capitata and none except B. fanshawet has the other type. B. fanshawet is not separable from B. macrophylla on any single character, although a number of characters are very nearly constant. A combination of characters, however, will give a good separation in all specimens we have encountered. The single Brazilian specimen of B. fanshawezi is in fact fairly typical of the species, not approaching B. macrophylla. It has very conspicuous domatia typical of B. fan- shawet. 32 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 19. Buchenavia macrophylla Eichl. in Flora 49 : 166 (1866) ; in Mart., Fl. Brasil. 14 (2) : 98 (1867). Terminalia macrophylla Spruce ex Eichl., loc. cit. (1867), nom. syn. Buchenavia stellae Cuatrec. in Fieldiana, Bot. 27 (1) : 109 (1950). Leaf-lamina subcoriaceous to coriaceous at maturity, 3-34 I-5-II cm., (narrowly elliptic) oblanceolate to narrowly obovate, minutely apiculate to long-acuminate (or rarely rounded) at the apex with an acumen up to 2 cm. long, acutely to very nar- rowly cuneate and often markedly narrowed and sometimes decurrent into the petiole at the base, almost glabrous except pubescent on the midrib and lateral nerves on both surfaces ; midrib very conspicuous and raised below ; lateral nerves very conspicuous and raised below, (4) 6-14 pairs ; reticulation rather inconspicu- ous or fairly conspicuous, not close ; petiole rather slender to stout, very variable in length, 6-32 mm. long, densely to sparsely pubescent, nearly always conspicuously biglandular from below the middle to at the top. Peduncle not stout, 6-32 mm. long, shortly pubescent ; rhachis 20-100 mm. long, densely puberulous ; flowers as in B. fanshawet. ¥ruit oblong-ellipsoid to ovoid-ellipsoid, terete or very slightly flattened when fresh, up to 23 X I2 mm., rounded to subacute (apiculate) at the apex, rounded to very shortly pseudostipitate at the base, very shortly but densely ferrugineous-tomentellous, indumentum wearing off when very old ; endocarp as in B. fanshawet. CoLoMBIA : Caqueta: Florencia, in the cerros La Estrella, 400 m., 30 Mar. 1940, Cuatrecasas 8863 (US, isotype of B. stellae). BRAZIL: Amazonas: Near Panuré on Rio Vaupés, Oct. 1852—Jan. 1853, Spruce 2507 in part (BM; BR, lectotype, not seen; K). Cjapo, Rio Vaupés, “ Small elegant tree 25 ft. Fls. reddish-brown. I was nearly too late for this’, Sept. 1852, Spruce 2507 in part (K, isolectoparatype). Rio Demeni, tributary of R. Negro, lake of black water, 12 Oct. 1952, Frdes 28888 (BM). Jaureté, Papury, bank of river, 16 Oct. 1945, Frdes 21156 (K). Mandaos, Pensador, riparian forest of the Igarapé, to Jan. 1944, Ducke 1542 (K). Mandaos, Pensador, humid spot in non-flooded forest, 20 Aug. 1935 and 3 Feb. 1936, Ducke 35168 (K). Manaos, banks of the Igarapé Mindu, 8 Sept. 1945, Ducke 1753 (K). Mandaos, along road to Aleixo, 12 Aug. I Sept. 1936, Krukoff 7921 (BM; K). Humayta, near Tres Casas, basin of Rio Madeira, on varzea land, 14 Sept.—11 Oct. 1934, Krukoff 6117 (BM; F; K), 6273 (BM; F; K). Parad: Faro, 4 Feb. 1910, Ducke 10649 (BM); same locality, 9 Oct. 1915, Ducke 15798 (BM). Lago de Faro, on banks in rarely flooded places, 22 June 1926, Ducke 13583 (K). Rio Tapajoz, region of the lower waterfalls, 25 June 1918, Ducke 17065 (BM). Rio Sao Manoel, Cachoeira do Calderao, on the boundary with Mato Grosso State, by Igarapé Preto, 6 Jan. 1952, Pires 3822 (BM). Rio Pixuna, 40 km. from the mouth of Rio Cupari, between Prainha and Agua Béa, 22 Dec. 1947, Black 47-1964 (BM). Amapd: Rio Araguari, near the Cachoeira do Paredao, “‘ terra firme’, 25 July 1951, Frées & Black 27695 (BM). B. macrophylla is apparently the most abundant species of the genus in its area of distribution. It has not been recorded from Venezuela, although it seems likely to occur there. It probably also occurs further up the Vaupés in Colombia. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 33 The fruits of B. macrophylla are distinct from those of all other species than B. fanshawet, the difference between B. macrophylla and that species being discussed under the latter. In their typical form the leaves are also quite characteristic and distinctive. We have little hesitation in including B. stellae, of which we have seen an isotype, under the present species. It is the only record of B. macrophylla which we know of from Colombia. All the characters by which B. stellae is supposed to differ (notably its longer petioles) are found quite frequently in various forms of B. macro- phylla. Although B. stellae was compared by the author with B. reticulata, it and B. macrophylla in general are very distinct from that species. 20. Buchenavia longibracteata Fr’es in Bol. Técn. Inst. Agron. Norte 20: 53, t. 16 (1950). Leaf-lamina chartaceous at anthesis, 8-23 x 4-8 cm., usually oblanceolate, shortly acuminate at the apex, narrowly cuneate but not decurrent into the petiole at the base, almost glabrous above except tomentose on the midrib and densely pubescent on the lateral nerves, conspicuously ciliate on the margin, sparsely pubescent below except densely pubescent on the midrib and lateral nerves (at flowering time) ; midrib conspicuously raised below; lateral nerves conspicuously raised below, c. I2-16 pairs ; petiole short and stout, c. 6-13 mm. long, shortly pubescent, con- spicuously biglandular. Rhachis c. 55-110 mm. long, densely pubescent ; flowers c. 3°5 X 3 mm., with a pubescent ovary but a glabrous upper hypanthium ; bracts relatively large and conspicuous, foliaceous, c. 7-13 X I-2 mm., with a conspicuous midrib, pubescent, exceeding the flowers. Fruit so far unknown. BRAZIL: Amazonas: Cachoeira das Araras, Rio Vaupés (near the Colombian frontier), ““Arvore, 5 m.”, 1 Nov. 1945, Frdes 21308 (IAN, holotype, not seen; K). This species is very distinct in its large leafy bracts which, as far as is known, are unique in the genus. Apart from this character, however, it scarcely seems possible to distinguish it from B. macrophylla, and indeed it may be an abnormal form of that species. As the fruits are unknown (the only gathering being in flower) we prefer to maintain the species for the present. The locality is within the range of B. macrophylla but we have not seen the slightest trace of large leafy bracts or of a ciliate leaf-margin in the latter species. 21. Buchenavia pulcherrima Exell & Stace, sp. nov. Arbor c. 20 m. alta ; ramuli griseo-fusci, fere glabri, apice clavati. Folia spiraliter ordinata ad ramulorum apices congesta ; lamina subcoriacea, 27:5-29 < 8-10 cm., anguste obovata vel oblanceolata, integra, in toto fere glabra, apice abrupte acute mucronata vel apiculata (apiculo ad 8 mm. longo), basi anguste cuneata, costa media infra valde prominenti; nervis lateralibus infra prominentibus 9—11-paribus, venulis conspicuis vix prominentibus; petiolus 30-38 mm. longus, triquetrus, inconspicuissime glandulosus vel eglandulosus. Inflorescentiae axillares, pendulae, 34. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA elongato-spicatae ; pedunculus gracilis, c. 30-42 mm. longus, sparse rufo-pubescens ; rhachis gracilis, c. 70-110 mm. longa, rufo-pubescens. Fructus ellipsoideus, teretius- culus, immaturus ad 12 x 8 mm., apice basique rotundatus vel latissime obtusus, pulchre dense rufo-velutinus. BRITISH GUIANA: Essequibo : 172 km. along Bartica-Potaro road, “60 ft. tree ro in. diam. basally swollen from Clump Wallaba forest—lvs. tufted to branch ends, thinly leathery, primary nerves prominent beneath—young fr. on drooping stalks from below lvs., oval, rusty-velvety pubescent’, 15 Nov. 1943, Field No. F1r485 = Forest Dept. 4221 (K, holotype) ; same locality, “ Seedlings of 422r with coty- ledons only—perhaps 10 days old, from below parent tree’’, 15 Nov. 1943, Field No. F1486 = Forest Dept. 4222 (K). This species is very different from the other two species of the genus known from British Guiana: B. fanshawei with small leaves and very shortly brownish-pubescent fruits; and B. megalophylla with large leaves and long-beaked fruits. It is also distinct from all other members of the genus in its fruits. The specimen Forest Dept. 4222 in the Kew Herbarium consists of two seedlings each of which has only a pair of cotyledons on a short hypocotyl. The cotyledons are very like those of B. fanshawei from the same area: 3:5-3°8 by 4°5-5°5 cm., truncate at the apex, obtusely cuneate to almost truncate at the base. We have not seen cotyledons of any other species of Buchenavia. The vernacular name for both B. fanshawei and B. pulcherrima is given as “ Fukadi ”’, 22. Buchenavia huberi Ducke in Bol. Técn. Inst. Agron. Norte 4 : 24 (1945). Leaf-lamina coriaceous, 2:5-14 Xx I-3-6-7 cm., narrowly obovate or obovate to elliptic-obovate, rounded or very shallowly retuse or broadly obtuse or very shortly apiculate (apiculus up to 4 mm. long) at the apex, acutely cuneate at the base and not or scarcely decurrent into the petiole, almost glabrous except rather sparsely pubescent on the midrib and lateral nerves on both surfaces ; midrib conspicuous and raised below ; major lateral nerves conspicuous and raised below, 5—Io pairs ; reticulation fine but quite conspicuous, close, forming very small inter-reticular spaces ; petiole distinct and stout, 14-25 mm. long, sparsely pubescent or pubescent, eglandular or very inconspicuously glandular near the apex. Peduncle slender, 12-24 mm. long, puberulous, indumentum greyish or rufous; rhachis slender, c. 35-95 mm. long, rufous- or cano-puberulous ; flowers with a densely canescent- sericeous ovary and a glabrous upper hypanthium. Fruit ellipsoid to ellipsoid- obovoid, more or less terete, 28-31 X 13-14 mm. (c. 20 mm. broad fide Ducke— fresh material?), abruptly and shortly apiculate (up to 2 mm.) at the apex, shortly pseudostipitate (2-3 mm.) at the base, cano- or rufous-tomentose at first, indumentum wearing off but still visible in very old fruits in wrinkles of drupe and at the ends; endocarp almost terete, slightly longitudinally ridged, c. 25-30 X 12-15 mm., subacute at the apex, more or less rounded at the base. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 35 BRAZIL: Amazonas: Manaos, vicinity of the Cachoeira do Mindu, on “ terra »” ce firme ”’, “ arvore grande ’”’, 3 Dec. 1943, Ducke 1450 (K ; RB, lectotype, not seen). Tabatinga, Estrada do Marco, on “ terra firme ’’, ‘‘ arvore muito grande, com sapo- pemas possantes’’, 26 Nov. 1945, Ducke 1803 (K). Humayta, near Tres Casas, basin of Rio Madeira, on low terra firma, 14 Sept.—11 Oct. 1934, Krukoff 6472 (BM ; K). Humayta, near Livramento, on Rio Livramento, basin of Rio Madeira, on terra firma, 12 Oct._6 Nov. 1934, Krukoff 6794 (BM; F; K). Ducke also mentions specimens (Ducke 1308 (RB)) collected from the same locality (same tree?) as the lectotype on 12 Aug. 1943 (subadult flowers) and on 5 Oct. 1943 (young fruits). Apparently the first specimen known was a tree in the gardens of Belém (Para) Museum (Para State) introduced in 1904 by Dr. J. Huber from the mid region of the Rio Perus (Amazonas State). There is a specimen from this tree (Ducke 1281 (RB)) with mature fruit. This species appears to be very distinct from all others in its characteristic fruits, perhaps being closest to the glabrous-fruited B. grandis. Very old fruits seem quite glabrous but on examination remains of the tomentum can be seen in the wrinkles and at the ends. In the original description Ducke says that the fruits have a grey tomentum, and this is true of Krukoff’s two gatherings. Ducke’s two gatherings, however, have a conspicuously rufous tomentum. Such a change of indumentum from a greyish to a reddish colour during drying is also known in leaves of Chrysophyllum species. We do not know of it elsewhere in Buchenavia : B. sericocarpa has retained its greyish-pubescent fruits in the herbarium and B. pulcherrima is described in a field note as having ‘‘rusty-velvety pubescent ”’ fruits in the fresh state. 23. Buchenavia grandis Ducke in Arch. Jard. Bot. Rio Janeiro 4 : 148 (1925). Leaf-lamina chartaceous or subcoriaceous, 4°5-I2 X 2-5 cm., obovate to narrowly obovate, rounded to subacute or rarely apiculate at the apex, narrowly cuneate but scarcely decurrent into the petiole at the base, almost glabrous (or sometimes very sparsely pubescent below) except sparsely puberulous to puberulous on the midrib and lateral nerves on both surfaces ; midrib conspicuous and raised below ; major lateral nerves fairly conspicuous and somewhat raised below, c. 6-9 pairs ; reticula- tion rather conspicuous and close, forming fairly small inter-reticular spaces ; petiole distinct, not stout, (5) 10-20 mm. long, puberulous to sparsely puberulous, eglandular. Peduncle slender, c. 13-24 mm. long, rufous-puberulous ; rhachis slender, c. 22—100 mm. long, rufous-puberulous; flowers c. 3-4 X 2-3 mm. Ovary glabrous to densely rufous-puberulous. Distal portion of lower hypanthium glabrous to sparsely rufous-puberulous ; upper hypanthium glabrous or very sparsely puberulous. Fruit ellipsoid, slightly 4—5-angled when dry, c. 24-27 X II-I3 mm. (17 mm. wide fide Ducke), rather gradually acutely apiculate (up to 2 mm.) at the apex, shortly pseudo- stipitate (c. 2-3 mm.) at the base, glabrous; endocarp (fide Ducke) compressed, longitudinally ridged, 20-22 * 12-13 mm., 8-10 mm. thick, acute at each end. 36 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA Brazit : Parad: Obidos, 9 Mar. 1909, Ducke 10235 (BM ; MG, lectotype, not seen); same locality, wood in elevated places, ‘‘ arbor magna fl. viridibus ”’, 15 Sept. 1927, Ducke 21349 (K). Oriximina, by Rio Trombetas, 5 Feb. 1918, Ducke 16976 (BM, lectoparatype). In forest in the Quataquara hills near the middle of Rio Tapajoz, “arbor magna ’”’, 13 Aug. 1923, Ducke 17687 (K, lectoparatype). Ducke has recorded the species elsewhere in the State of Para: between Mt. and Lake Curumu (Ducke 17682 (RB)) ; by the Rio Tapajoz waterfalls near Bella Vista (Ducke 17688 (RB)) ; and in the Serra de Santarém. He has also recorded it from Maranhao : near Codo on the Rio Itaperurti (Ducke 658 (MG)) ; and near Mirador on the same river (Lisbéa 2351 (MG)). Besides Ducke’s four gatherings cited above we have seen no other specimens of this species, although specimens of other species have been placed here (e.g. of B. pallidovirens and B. hubert). It is possible that the specimens mentioned represent more than one taxon, al- though there is no direct evidence for this besides the great variation in pubescence noted by Ducke. He states that specimens from Tapajoz have glabrous ovaries and less densely pubescent inflorescences (i.e. rhachides), specimens from Obidos and Trombetas have moderately pubescent ovaries and inflorescences ; and that specimens from Maranhao have very densely pubescent ovaries and inflorescences. We have, however, noted this range of pubescence in B. oxycarpa, where an increased pubescence is probably correlated with a drier habitat. This is also very likely the case in B. grandis, as noted by Ducke since he says that trees from the drier regions (in Maranhao) are smaller (20 m. as opposed to 30-45 m. high). The Maranhao specimens are also more pubescent. This is a plant from the lower regions of the Amazon and coastal areas, which may well account for the fact that we have seen no recent collections of it. It appears to be most closely related to B. hubert and B. congesta. Ducke’s specimens 10235 and 16976 (and probably others) were originally deter- mined (by J. Huber fide Ducke) as Terminalia lucida Hoffmanns. ex Mart., as there is some resemblance between the leaves of the two species. The fruits, however, are quite different, those of T. lucida having very broad wings and a dense indumen- tum. Two gatherings of B. oxycarpa (Krukoff 1241 and 6174) and one of B. fan- shawei were also originally determined as T. lucida, although the Termuinalia is quite unrelated and bears only the most superficial resemblance to any species of Buchen- avia. We do not know of any other Tevminalia species which has been confused in this way. 24. Buchenavia congesta Ducke in Trop. Woods 90 : 24 (1947). Leaf-lamina chartaceous or subcoriaceous, 8-15 xX 3-7 cm., obovate or narrowly obovate, rounded and very sharply and abruptly shortly apiculate (apiculus up to c. 3 mm.) at the apex, narrowly cuneate and slightly decurrent into the petiole at the base, very sparsely puberulous to almost glabrous except puberulous on the midrib and lateral nerves above, sparsely puberulous at flowering time to very sparsely puberulous or almost glabrous except puberulous on the midrib and lateral REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 37 nerves below at fruiting; midrib conspicuous and raised below; lateral nerves conspicuous and raised below, slightly raised above, c. 9-13 pairs; reticulation rather conspicuous, slightly raised below and above, forming small inter-reticular spaces ; petiole not stout, c. 20-30 mm. long, puberulous or densely puberulous, eglandular. Peduncle rather slender, 20-35 mm. long, densely rubiginous-puberulous at flowering time to rather sparsely so in fruit ; rhachis rather slender, c. 35-85 mm. long, rather densely rubiginous-puberulous ; flowers c. 3-4 xX 2:3 mm. Lower hypanthium densely rubiginous-puberulous; upper hypanthium rather to very sparsely rubiginous-puberulous. Fruit oblong-ellipsoid, very slightly flattened to more or less terete, 22-25 X 9-I2 mm., rounded to broadly obtuse at the apex, more or less pseudostipitate (pseudostipe under 2 mm. long) at the base, glabrous from a very young stage. BRAZIL: Amazonas: Manaos, near Cachoeira do Mindt in a non-flooded wood in a humid spot, ‘“‘ arbor sat. magna’’ (over 30 m. fide Ducke in descr.), 3 Dec. 1943, Ducke 1465 (F ; RB, lectotype, not seen) ; same locality, 4 Oct. 1946, Ducke 2003 (RB, lectoparatype, not seen) ; same locality, ‘‘ Arvore grande; flér verde brancacenta da arvore tipica’”’, 25 Sept. 1947, Ducke 2104 (BM). Apparently only one tree of this species is known and from it all three of the above gatherings were made. From these we select Ducke 1465 as lectotype since this bears mature fruits: Ducke 2003 has young fruits and Ducke 2104 (which was not cited by Ducke) has flowers. In its densely congested long-petioled leaves and oblong-ellipsoid glabrous fruits this species appears to be quite distinct from all others, perhaps approaching most closely B. grandis and to a lesser extent B. macrophylla. Insufficiently known species Buchenavia gracilis Glaziou in Mém. Soc. Bot. France 1 (3) : 203 (1908), nom. nud. Based on Glaziou 5855 from the State of Rio de Janeiro, Brazil. We have exam- ined the specimen in Herb. Paris and a fragment in the British Museum Her- barium from the specimen in Herb. Copenhagen. They both consist of young sterile material alone, which makes their determination quite impossible even to the extent of specifying Buchenavia or Terminalia as the genus. Buchenavia macahensis Glaziou in Mém. Soc. Bot. France r (3) : 203 (1908), nom. nud. Based on Glaziou 18218 from the State of Rio de Janeiro, Brazil. We have ex- amined the specimen in Herb. Paris and confirmed the genus as Buchenavia. The specimen consists only of leaves and very young inflorescences and this makes a comparison with known species extremely difficult. We consider the specimen too immature to warrant description as a new species or to allow its identification, although there are a number of similarities with B. viridiflora. A further possibility is that the inflorescences would have been capitate. A second sheet in Herb. Paris bears an identical label to the previous specimen, in addition to the correct collector’s ticket and tag with n. 27125. This specimen is clearly B. tomentosa and was recorded 38 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA as such by Glaziou (loc. cit.). It was doubtless erroneously mounted with a label of B. macahensis. Excluded species Buchenavia fluminensis Glaziou in Mém. Soc. Bot. France 1 (3) : 203 (1908), nom. nud. Based on Glaziou 6142 from the State of Rio de Janeiro, Brazil. We have ex- amined the two specimens in Herb. Paris and found them to be an unknown species of Terminalia. RAMATUELLA Kunth RAMATUELLA Kunth, Nov. Gen. & Sp. Pl. 7: 253 (1825) (‘‘ Ramatuela’’). Type: R. argentea Kunth. Kunth originally used the spelling Ramatuela, naming the genus after a French horticulturist whom he called “‘ Ramatuel”’. De Candolle (Prodr. 3 : 16 (1828) stated that the correct spelling of this Frenchman’s name was Ramatuelle (actually given as M. de Ramatuelle in a paper published by the latter in Journ. Hist. Nat. 2: 233 (1792)) and thus rightly corrected the spelling of the generic name to Ramatuella.+ Ramatuella also belongs to the Terminalieae, differing from the other six genera in the tribe by having 4—5-winged actinomorphic fruits crowded in spherical masses. It is confined to the headwaters of the Rio Orinoco and Rio Negro and some of their tributaries in the Amazonas district of Venezuela, the Vaupés district of Colombia and the Amazonas district of Brazil (see Fig. 1), the whole region consisting of inter- connecting waterways since the upper Orinoco divides in such a way that some of its waters, curiously enough, flow into the Amazon via the Rio Negro. The area of distribution of the genus lies within latitudes 1° S. and 4° N., and occupies only about 8° of longitude. There appear to be no ecological or geographical differences between the species. The genus is closely related to Terminalia and it is highly probable that it was derived from an original Teyminalia stock. As in Terminalia and Buchenavia (and indeed all the Tevminalieae) Ramatuella is apetalous. The lower hypanthium is constricted at the apex but not extended into a long neck as in Buchenavia. It and the cupuliform upper hypanthium, which bears 4 or 5 triangular calyx-lobes, are pubescent on the exterior. The 8 or ro stamens are in two whorls, one opposite to and one alternate with the calyx teeth, and are conspicuously exserted, as is the style. In bud the stamens are bent double, the anthers being well protected in the base of the upper hypanthium which possesses a very well-developed hairy disk. As in Terminalia the anthers are versatile. All species usually have tetramerous and pentamerous flowers in the same inflorescence. In Ramatuella, as is frequently the case in Buchenavia, male and bisexual flowers are mixed in the same inflorescence, in the present genus the former sometimes being the more abundant. The lower 1 Poiret (Dict. Sci. Nat. 44: 427 (1826)) had used the spelling Ramatuella two years previously, but he gave no reasons for this. REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA _ 39 flowers of the inflorescence are more usually male, so that the floriferous rhachis may be three or more cm. long but give rise to a more or less capitate infructescence. Ramatuella was surveyed fairly recently by Maguire (in Mem. New York Bot. Gard. 8 : 130-132 (1953)), who recognized four species, one consisting of two varieties. The acquisition of further material has now led us to recognize six species and, apart from the fact that Maguire gave no descriptions, is the only reason for revising the genus after such a comparatively short interval. These species all consist of small to medium-sized evergreen trees of riverine forest. KEY TO THE SPECIES Lower surface of leaves and fruit densely silvery-sericeous ; wings of fruit short and broad, very abruptly narrowed at the base into a pseudostipe and at the apex into a beak 2-4 mm. long ; : 1. R. argentea Lower surface of leaves for the greater part glabrous or only rather sparsely hairy, not silvery or sericeous ; fruit various, the wings relatively long ; beak more or less absent to long : Fruit under 10 mm. long, deltate-ovate in outline ; wings abruptly narrowed at the base and apex of the fruit-body ; beak o-5-r mm. long; leaves up to Poa CMe: : : : 2. R. maguirer Fruit 10-20 mm. long, iancealave to circular’ in outline, rounded or curved at the margin ; beak up to 1 mm. long, or 2 or more mm. long and then the wings gradually narrowing into it: Leaves glabrous or very sparsely pubescent over the whole of both surfaces ; fruit acuminate at the apex, beaked, the beak 2-5 mm. long at maturity: Leaf-lamina 2°5-1I x I-5 cm.; petiole 4-15 mm. long; fruit silvery, ovate in outline ; beak c. 2-5 mm. long, rather slender ; pseudostipe c. 2 mm. long. ; ‘ ; : , : . 3. R. virens Leaf-lamina 8-14 x oy cm. © petiole 20-30 mm. long; fruit fulvous, nar- rowly ovate to lanceolate ; in outline ; beak c. 2-4 mm. long, rather stout ; pseudostipe absent ‘ ‘ Fae ce latifolia Leaves rather densely pubescent at the base of the midrib on the lower surface ; fruit truncate to obtuse at the apex, beakless or with a beak not exceed- ing I mm. long: Leaf-lamina oblanceolate to narrowly oblong ; fruit broadly elliptic to cir- cular in outline, 12-18 mm. wide, truncate to rounded-retuse at the apex ; wings markedly crisped, rather thin : . 5.2. crispialata Leaf-lamina obovate to oblong ; fruit oblong to ovate in outline, 8-15 mm. wide, obtuse to almost truncate at the apex; wings undulate, rather thick . : : : : : ; ' . 6. R. obtusa 1. Ramatuella argentea Kunth, Nov. Gen. & Sp. Pl. 7 : 254, t. 656 (1825).—DC., Prodr. 3 : 16 (1828).—Eichl. in Mart., Fl. Brasil. 14 (2) : 99, t. 26 fig. 1 (1867).— Maguire in Mem. New York Bot. Gard. 8 : 131 (1953). (Fig. 4 ¢.) 40 REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA Fic. 4. Ramatuella maguirei Exell & Stace (holotype): a, habit (x %); 0, fruit (x2). R. argentea Kunth (Spruce 3498): c, fruit (x2). R. virens Spruce ex Eichl. (Maguire, Cowan & Wurdack 30764): d, fruit (x2). R. latifolia Maguire (holotype): e, fruit (x2). R. obtusa (Maguire) Exell & Stace (holotype): f, fruit (x2). R. crispialata Ducke (Frdes 28040) : g, fruit (x2). REVISION OF THE GENERA BUCHENAVIA AND RAMATUELLA 41 Leaf-lamina coriaceous, 2-Io X I-4 cm., oblanceolate to narrowly obovate, retuse to rounded at the apex, narrowly cuneate and slightly decurrent into the petiole at the base, almost glabrous above except sparsely or very sparsely puberulous on the midrib, densely silvery-sericeous below ; midrib conspicuous and raised below ; major lateral nerves inconspicuous, c. 5—II pairs; reticulation not visible; petiole usually distinct, 5-16 mm. long, appressed-puberulous. Peduncle (in fruit) rather slender to stout, 5-35 mm. long, appressed-puberulous ; rhachis up to 6 mm. long. Fruits crowded into spherical capitula, 8-12 oon Ss ae acoM00 S = . SS S 0008 —= Sapser 20 yo oo Son aS Soc DOS =00o0¢ —w 00> CID oor Ss, WS? Sar poo Oa” e7 19 Fics. 17-22. Achnanthes fimbriata (Grun.) Ross. Fig. 17, raphe valve (in B.M. 13079) (x 1,000) ; Fig. 18, detail of same valve as Fig. 17 (x 2,000) ; Fig. 19, raphe-less valve (in B.M. 13079) (x 1,000) ; Fig. 20, raphe valve (B.M. 37339, specim. n. 4) (x 1,000) ; Fig. 21, detail of same valve as Fig. 20 (x 2,000) ; Fig. 22, raphe-less valve (Coll. F. W. Adams n. Bess. 896, specim. n. 1) ( x 1,000). JuGosLAvIA : Porto Zubzamki, SuSak Island, Gulf of Kvarner, Reichardt (Coll. Grunow ns. 869, 869b in Herb. Vienna, syntypes). Mali LoSinj, LoSinj Island, Gulf of Kvarner, Reichardt (Coll. Grunow ns. 864a, 864b in Herb. Vienna). SPAIN : Balearic Islands (B.M. 36256 = Cleve & Moller, Diatoms, n. 154). ITALy : Capri (B.M. 33068). Norway : Grip (Cleve & Moller, Diatoms, n. 311 in Coll. F. W. Adams in Herb. Brit. Mus.). U.S.A.: Connecticut : Morris Cove (B.M. 14577 = Temp. & Perag., Diatomées, 74 CAPARTOGRAMMA AND SCHIZOSTAURON n. 323 ; B.M.69301 = Temp. & Perag., Diatomées, ed. 2, n. 949 ; isotypes of A. mantfera). Florida : Pensacola (B.M. 13079 = Cleve & Moller, Diatoms, n. 320). ADEN : Harbour mud, Vozgé (Coll. Voigt n. 34099). CEYLON : Colombo, Craven (B.M. 8986, syntype of Stauronets cornuta) ; same locality, harbour mud, Voigt (Coll. Voigt ns. 14099, 30056). INDONESIA : Celebes, from holothurians (Coll. Voigt n. 29018). CuiInA : Kwangtung : Hainan, from holothurians (Coll. Voigt ns. 36053, 36060). Chekiang : Chusan Archipelago, Voigt (Coll. Voigt n. 3040). JAPAN : Kyushu : Nagasaki, off shells, Vozgt (Coll. Voigt n. 25044). AUSTRALIA : Queensland : Flinder’s Passage (B.M. 31942). Fossil : Miocene : HunGARY : Nyermegy (B.M. 37338 ; B.M. 68520, 68521 = Temp. & Perag., Diatomées, ed. 2, ns. 174, 175). Pliocene : BuLGARIA : Balchik (B.M. 2314; B.M. 26856, 26857 = Van Heurck, Typ. Synops. Diat. Belg., ns. 545, 546 ; B.M. 30827, 31384, 32629, 32630, 32633, 32836, 37339, 42089 ; Coll. F. W. Adams ns. A 167, Bess. 895, Bess. 896 in Herb. Brit. Mus.). HUNGARY: Bremia, near Kavna (B.M. 14757, 14758 = Temp. & Perag., Diatomées, ns. 503, 504; B.M. 33302, 33303; B.M. 68694 = Temp. & Perag., Diatomeées, ed. 2, n. 348). Karand (B.M. 70893). As will be seen, the range of density of striation on both the raphe-less and the raphe valves of this species is very wide. It is found to vary continuously, however, when a considerable number of specimens from a variety of localities is examined, and the variation on the two valves is not closely correlated. There are nevertheless some differences between the ranges found in recent and fossil specimens, although there is much overlap. Recent specimens (Figs. 17-19) range in size from 20u by 7u to 75u by 26yu, whilst the fossil ones (Figs. 20-22) range from 24u by IIp to 120%. by 40 ; the striae on the raphe-less valve of recent specimens range from 5°5 to 14 in 10y, whilst those on fossil ones range only from 7:5 to 13 in Ioy ; on raphe valves the striae of recent specimens range from 18 to 30 in Io in the centre part of the valve whilst those of fossil specimens range from 12 to 21in rou. Another variable feature is the direction of the striae on the raphe-less valve ; in some speci- mens these are definitely although not strongly radiate and a few short striae are present in the centre of the valve, whilst others have almost parallel striae and no short ones. Short striae are present on all fossil specimens but only on some recent ones. These differences between fossil and recent specimens are too indefinite to serve as a basis for an infraspecific division, and accordingly Achnanthes danica var. maxima, proposed by Cleve for fossil specimens of this species from Balchik, is not maintained here as a separate variety. The striae on the raphe-less valve of this species are always separated by a distinct gap, usually equal to at least half their own width, but in the coarsest specimens occasionally somewhat less. On most of Hustedt’s figures of this valve of the species (viz.: A. Schmidt, Atl. Diat.: t. 415 figs. 9, 11, 13-15, 18 (1937).; Bull. Duke Univ. CAPARTOGRAMMA AND SCHIZOSTAURON 75 Mar. Sta. 6: t. 6 figs. 5, 7, 8 (1955)) they are represented as contiguous. Whilst this may to some extent be due to the drawing convention adopted, it is misleading. Achnanthes fimbriata is less common in the miocene and pliocene deposits from Hungary than it is in the pliocene deposit from Balchik, Bulgaria, and even there it is infrequent. It is not common, either, in any of the recent gatherings in which I have seen it, but nevertheless it has a wide distribution, from the Atlantic Coast of North America, through Europe and the Mediterranean to Ceylon, the East Indies and the western borders of the Pacific. Grunow expressed doubt about the correctness of referring this species to Schizostauron when first describing it, and by 1877 he had decided that it did not belong there ; the phrase in the original diagnosis of the genus which especially refers to it, “ (vel laciniato fimbriato?) ”’, was deleted from the version published in that year (Month. Microsc. Journ. 18 : 181 (1877)). Grunow’s notes show that when he first decided that the species was not a Schizostauron he considered that it should be placed in Navicula but that he then found whole frustules in the material from the Balearic Islands issued as Cleve & Moller, Diatoms, ns. 154, 155 (1878), and realized that it was an Achnanthes. He apparently considered “‘ fimbriata ”’ an inappropriate epithet for the species when it was placed in this genus. The list of species in the preparations forming the third series of Cleve and Mdller’s Diatoms, which includes ns. 154, 155, is headed “ All the slides have been examined by M. Grunow ’’, and the list for those numbers includes ‘‘ Achnanthes heteropsis Grun. (= Navicula fimbriata Grun.)’”’. Neither of these names had been published pre- viously and both must accordingly be regarded as nomina nuda ; the identity of epithet is insufficient to connect them with the previously published description of Schizostauron fimbriatum Grun., although Grunow’s manuscript notes show that they were new names for that species. Between 1878 and 1880 Grunow formed the opinion that this species was identical with Cocconets danica Flogel and he accordingly made the new combination Ach- nanthes danica, quoting A. heteropsis as a synonym. This mistake was probably the result of his receiving, apparently from Weissflog, material of a further gathering from the Balearic Islands in which there was true A. danica but no A. fimbriata. From his notes it appears that he did not at any stage associate the raphe valves he called Schizostauron fimbriatum in the material from the Gulf of Kvarner with the raphe-less valves also present in his preparations of these gatherings. When he found specimens in the Cleve and Moller material from the Balearic Islands that caused him to decide that his S. fimbriatum was the raphe valve of an Achnanthes, he made no drawing of the raphe-less valve. He did, however, make drawings of both valves of A. danica from specimens in the Weissflog material and, misled by the similarity of the raphe valves, failed to realize that he was dealing with a different species from that in Cleve and Moller’s material from the same locality and also in their material from Pensacola, Florida, which he saw at about the same time. Asa result of Grunow’s error, A. fimbriata and A. danica have been confused ever since 1880. Thus Cleve took specimens of A. fimbriata in the material from the Balearic Islands issued as Cleve & Moller, Diatoms, ns. 154, 155, as representative 76 CAPARTOGRAMMA AND SCHIZOSTAURON of A. danica, and consequently in the Synopsis of the Naviculoid Diatoms he applied that name to A. fimbriata and misidentified true A. danica with A. lorenziana (Grun.) Cleve (= A. baldjickit subsp. lorenziana (Grun.) Ross). The way in which the name A. danica has been applied is analysed in the account of that species below (pp. 79-81). When Brun found specimens of this species in Tempére’s material from Morris Cove, Connecticut, he realized that they were not conspecific with Cocconets danica and so described them as a separate species, Achnanthes manifera. Such other authors as have recognized that this species is distinct from A. danica have, in general, used Brun’s name for it, but European specimens have for the most part continued to be identified as A. danica. The exceptions are those from Norway described by Hustedt as the separate species A. stroemi1, and the two from the Balearic Islands figured by him (in A. Schmidt, Atl. Diat.: t. 415 figs. 15, 16 (1937)) as A. manifera. Cleve (in Nordensk., Vega-Exped. Vet. Iaktt. 3 : 512 (1883)) was the first to treat Stauronets cornuta Leud.-Fortm. as a synonym of A. danica. Leuduger-Fortmorel’s species is based on a raphe valve and both A. danica and A. fimbriata occur in the material from Ceylon on which he was reporting. It is therefore impossible to be completely certain to which of the two species his type belongs. Nevertheless, the central area shown in his figure strongly suggests A. fimbriata rather than A. danica and his inability to resolve the striae points in the same direction. Navicula vaszaryt Pant. from the Hungarian miocene is clearly based on a raphe valve of Achnanthes fimbriata, which occurs in the diatomite from Nyermegy from which Pantocsek’s type came. Cleve transferred the species to Achnanthes on the basis of the resemblance of the raphe valve to that of an Achnanthes from Oregon which has a very different raphe-less valve and is clearly unrelated. Frenguelli (in Chiov., Fl. Somala : 394, t. 49 fig. 14 (1929)) describes and figures a raphe valve from near Mogadiscio in Somalia under the name Achnanthes mantfera. As Hustedt (Bull. Duke Univ. Mar. Sta. 6 : 18 (1955)) has pointed out, this cannot be identified with absolute certainty in the absence of a raphe-less valve, but its central gibbosity, the shape of its central area, and the closeness of its striae together make it highly likely that it is A. fimbriata and not A. danica. MHustedt’s (loc. cit.) statement that according to Frenguelli the striae on his specimen are I0-II in Iou is an error ; Frenguelli says that this is the density of the striae on the raphe valve of A. baldjickiit (Brightw.) Grun., with which he compares A. manifera, and that those on his own specimen are 22 in I0uw. Achnanthes stroemii Hust. is founded on specimens of this species with rather close striae on the raphe-less valve and not very close ones on the raphe valve. His type came from Nordasot, Norway. Specimens closely resembling Hustedt’s description and figure occur in Cleve & Moller, Diatoms, n. 311, from Grip, Norway, which is listed as containing A. danica. ACHNANTHES DANICA (Flégel) Grun. in K. Svensk. Vetensk.-Akad. Handl., Ny Foéljd, 17 (2) : 21 (1880) excl. syn. Achnanthes heteropsis Grun. et specim. ex Mediterraneo (partim), Adria et Florida.—A. Schmidt, Atl. Diat.: t. 198 fig. 50 CAPARTOGRAMMA AND SCHIZOSTAURON 77 (1895).—H. & M. Perag., Diat. Mar. France : 7, t. 2 figs. 1, 2 (1897) excl. t. 2 fig. 3.—Hust. in Rabenh., Krypt.-Fl. Deutsch. 7 (2) : 396, fig. 845 (1933) excl. syn. Schizostauron fimbriatum Grun., Achnanthes heteropsis Grun., Stauroneis cornuta Leud.-Fortm. et specim. ex Cleve & Mller, Diatoms, et Temp. & Perag., Diatomées, ed. 2.—Cleve-Euler in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5) : 18, fig. 510 (1953).—Job, Suppl. 1 List Diat. Alexandria : [4] (1961). (Figs. 23-26.) Cocconeis danica Flégel in Jahresber. Comm. Wiss. Unters. Deutsch. Meere Kiel 1 : 91, fig. 14 (1873).—Van Heurck, Treat. Diat., transl. Baxter : 289 (1896) excl. fig.; in Microgr. Prep. 13 : 83 (1905). Achnanthes lorenziana Grun. ex A. Schmidt, Atl. Diat.: t. 198 fig. 59 (1895). Achnanthes lorenziana Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 27 (3) : 186 (1896) pro parte ; non Rhaphoneis lorenziana Grun.—Hust., tom. cit.: 395 (1933) quoad descr. valvae cum raphide.—Cleve-Euler, loc. cit. (1953) quoad descr. valvae cum raphide.—Job, List Diat. Alexandria : [3] (1954). 23 Fics. 23-26. Achnanthes danica (Flégel) Grun. Fig. 23, raphe valve (in B.M. 71125) (x 1,000) ; Fig. 24, detail of same valve as Fig. 23 (x 2,000) ; Fig. 25, raphe-less valve (in B.M. 71125) (x 1,000) ; Fig. 26, detail of same valve as Fig. 25 (x 2,000). Valves rhombic-lanceolate, subrostrate, 50-65u long, 20-25u broad. Raphe-less valve with the pseudoraphe very narrowly lanceolate and a small hyaline space at each apex ; striae distant, parallel or very slightly radiate, 4-5-6 in 10u, each con- sisting of a single alveolus with oblique or transverse, sometimes bifurcating or anastomosing, secondary costae about 18-20 in Iou. Raphe valve with a straight filiform raphe with approximate central pores ; terminal nodules small, terminal fissures hooked ; axial area narrow, tapering uniformly to the apices ; central area transverse, rhombic, indefinitely bounded at its outer angles ; striae strongly radiate, longer and shorter in the centre of the valve, 16-20 in I0y, somewhat closer towards the apices ; alveoli mostly punctiform and 13-17 in Ioy, but those adjacent to the margin 1-5-2u long ; striae near the apices consisting of a single alveolus. From the coasts of western Europe, the Mediterranean and Ceylon, and fossil from southern Sweden (fide Cleve-Euler ; a doubtful record, cf. p. 81). 78 CAPARTOGRAMMA AND SCHIZOSTAURON ITaLy : Capri (B.M. 33068, 33069, 33426, 71121, 71122, 71123, 71124, 71125). SPAIN : Balearic Islands (B.M. 57563, coll. F. W. Adams ns. E 230, E 231 in Herb. Brit. Mus.). Ecypt : Ramleh, near Alexandria, 24 Sept. 1952, H. S. Job (Coll. H. S. Job n. 438) ; same locality, 27 Nov. 1952, H. S. Job (Coll. H. S. Job n. 446) ; same locality, 5 Apr. 1953, H. S. Job (Coll. H. S. Job n. 468). Rushdy, near Alexandria, 19 Nov. 1958, H. S. Job (Coll. H. S. Job n. 253). CEYLON : Colombo, Craven (B.M. 8986). This species, although widespread, is far from common, a fact which probably accounts for the inaccuracies in the accounts of it in the literature and for the long confusion of it with Achnanthes fimbriata (Grun.) Ross. When the raphe-less valve is observed under the microscope with a lens of less aperture than N.A. 1-3, or when such a lens is used with an inadequate illuminating cone, the striae appear to have a double row of puncta, in places alternating, in others opposite. When a lens of N.A. 1°3 or more is used with an adequate illuminating cone, however, it becomes apparent that this appearance is an optical artefact and that the true structure is as shown in Fig. 26. Similarly, careful observation with a lens of adequate aperture shows that the striae of the raphe valve are not regularly punctate through- out, as depicted by Hustedt, but that there is a row of long alveoli at the margin (Fig. 24). | Although the raphe valve is very similar to that of A. fimbriata, the central area is smaller and does not definitely expand outward, as that of A. fimbriata always does. The way in which Grunow confused this species and A. fimbriata has been dis- cussed under that species (p. 75 above). It has also been confused with A. baldjickit subsp. lorenziana (Grun.) Ross (A. lorenziana (Grun.) Cleve), firstly by Pantocsek, who misidentified specimens of that subspecies as A. danica, then by Cleve, who did the reverse. His account of A. lorenziana in the Synopsis of the Naviculoid Diatoms is based almost entirely on A. danica. Thus he says of the raphe-less valve “‘ Intercostal spaces finely punctate (puncta about 23 in 0,or mm.), frequently appar- ently smooth ’’, and his description of the raphe valve would apply to that of A. danica or A. fimbriata. The localities he cites for the species are ‘‘ Barcelona, Balearic Islands! Adriatic (Grun.), Bab el Mandeb! Macassar Straits (Grove Coll.)! ’’ and he adds the remark : ‘‘ I have seen an entire frustule in Grove’s collection, but the coarse structure of the upper valve made the examination of the lower very difficult.” No specimen to which this remark could refer is to be found now in Grove’s collection. About 1890 Thum issued slides of material from the Balearic Islands containing both A. danica and A. baldjickii subsp. lorenziana, but no A. fimbriata, and Cleve is certain to have seen some of these. As has already been pointed out (p. 76 above), he applied the name A. danica to A. fimbriata and, at least as far as the Balearic Island material was concerned, he seems to have identified as A. lorenziana the raphe- less valves of both A. baldjickii subsp. lorenziana, in which the outer membrane to the single alveolus forming each stria is smooth, and A. danica, in which that mem- brane has anastomosing secondary thickening. With these raphe-less valves he CAPARTOGRAMMA AND SCHIZOSTAURON 79 associated the raphe valves of A. danica, and he apparently never saw a raphe valve of A. baldjickwt subsp. lorenziana in this material. Before part II of the Synopsis of the Naviculoid Diatoms appeared, but during its preparation, plate 198 of A. Schmidt’s Atlas was published. Figure 59 on this plate is undoubtedly of a raphe-less valve of A. danica ; its legend is “‘ Nordsee, nach Cleve A[chnanthes| Lorenziana Grunow’’. As the name Achnanthes lorenziana had not by then been published, and as the figure is sufficient to validate its publica- tion there, Achnanthes lorenziana Grun. ex A. Schmidt (Atl. Diat.: t. 198 fig. 59 (1895)) is a legitimate name, typified by the North Sea specimen figured by Schmidt, a synonym of A. danica (Flogel) Grun., and an earlier homonym of A. lorenziana (Grun.) Cleve (K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 27 (3) : 186 (1896)), whose basionym is Rhaphonets lorenziana Grun. In view of the way in which the name Achnanthes danica has been so frequently misapplied, it will probably be useful to present here an analysis of all the accounts and records which have been published under it : Grunow in K. Svensk. Vetensk.-Akad. Handl., Ny Foéljd, 17 (2) : 21 (1880). The synonym A. heteropsis Grun., the lower limit of size of the valves, and the records from the Adriatic and Florida, and in part that from the Mediterranean, all refer to A. fimbriata. The remainder of the account is based on A. danica. Cleve & Mdller, Diatoms 6 : 3, n. 311 ; 4, n. 320 (1882). Both are A. fimbriata. Cleve in Nordensk., Vega-Exped. Vetensk. Iaktt. 3 : 512 (1883). The synonym Stauroneis cornuta Leud.-Fortm. refers to A. fimbriata. There is no information about the identity of the specimen on which this record is based, but it is most probably A. fimbriata, since, as explained above, Cleve later misidentified A. fimbriata as A. danica and included A. danica in A. lorenziana. Grunow in Van Heurck, Typ. Synops. Diat. Belg.: 121, n. 545 (1887). This is A. Jimbriata. Pantocsek, Beitr. Kenntn. Foss. Bacill. Ung. 1 : 31 (1886). I have seen no specimen from Dolje to which this record might refer. Both A. fimbriata and A. baldjicki subsp. lorenziana, but not A. danica, occur in other Hungarian fossil deposits, and the record is probably based on one or other of these. Pantocsek, op. cit. 2 : 57, t. 4 fig. 66 (1889). The figure is of A. baldjicki1 subsp. lorenziana, but A. fimbriata also occurs in material from Bremia, Hungary, and it is probable that Pantocsek did not distinguish the two species and that the record applies to both. De Toni, Syll. Alg. 2 : 480 (1891). The lower limits of size of the valve, and the synonyms A. heteropsis Grun. and Stauroneis cornuta Leud.-Fortm., refer to A. fimbriata. Of the locality records, Knarrhoi, Jutland, refers to A. danica, and so in part do Mediterranean and Ceylon, whilst the others refer to A. fimbriata, except that both that species and A. baldjickii subsp. lorenziana are covered by the record from Bremia, Hungary. A. Schmidt, Atl. Diat.: t. 198 figs. 60, 61 (1895), as A. danica var. maxima. These figures depict A. fimbriata. A. Schmidt, tab. cit. fig. 50 (1895). This figure of a raphe valve is probably 80 CAPARTOGRAMMA AND SCHIZOSTAURON correctly identified as A. danica. Fig. 59 on the same plate, named by Cleve A. lorenziana, is certainly A. danica. Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 27 (3) : 186 (1896). The only element in the description which refers to A. danica is the statement : “costae [on the raphe-less valve] . . . alternating with fine lineolae, twice as close as the costae’’. The remainder of the description, including that of A. danica var. major Cleve, and all the synonyms except Cocconeis danica Flogel, refer to A. fimbriata. The locality records do so also, except “‘ Mediterranean Sea ”’, which is taken from Grunow (1880) and is based on gatherings from the Balearic Islands in which both A. danica and A. fimbriata are represented. H. & M. Peragallo, Diat. Mar. France : 7, t. 2 figs. 1-3 (1897). The statement that the costae on the raphe-less valve are 8-13 in 10u (which is inconsistent with fig. 2, correctly identified as A. danica, as also is fig. 1) and the description of the raphe valve refer to A. fimbriata, as do the synonym A. heteropsis Grun. and the record from “ Mer du Nord ”’, also in part that from “‘ Baléares’’. The fig. 3 also depicts A. fimbriata. Tempére & Peragallo, Diatomées, ed. 2 : 82, ns. 152, 153 (1908). There is no specimen of this or any related species on either of these two slides in the set in the British Museum Herbarium. Tempére & Peragallo, op. cit.: 185, ns. 346-349 (1909). According to the index (op. cit., Tables : 9 (1915)), the species is only found on ns. 348, 349. In the set in the British Museum Herbarium there is a specimen of A. fimbriata on n. 348 but none of this or any related species on any other of the slides. I have found specimens of A. baldjickit subsp. lorenziana on other preparations of material from the same fossil deposit at Bremia, near Kavna, Hungary. Tempére & Peragallo, op. cit.: 200, n. 371 (1910). A. baldjickit subsp. lorenziana is present on this slide in the set in the British Museum Herbarium. Tempére & Peragallo, op. cit.: 232, ns. 442, 443 (1910). According to the index (op. cit., Tables : g (1915)), the species is to be found on n. 443. There is no specimen of this or any related species on this slide in the set in the British Museum Herbarium. Hustedt in Rabenh., Krypt.-Fl. Deutsch. 7 (2) : 396, fig. 845 (1933). The descrip- tion is of A. danica, except that the statement that the striae in the centre of the raphe valve are 26 in 10 would fit A. fimbriata but not A. danica ; “ 26”’ is, however, probably an error for ‘‘ 20”’ since Hustedt’s figure of this valve shows only 20 striae in 10 in the centre of the valve. This figure and that of the raphe-less valve represent A. danica. The synonyms Schizostauron fimbriatum Grun., A. heteropsis Grun., and Stauroneis cornuta Leud.-Fortm. refer to A. fimbriata. Of the gatherings cited by Hustedt, Cleve & Moller, Diatoms, ns. 154, 155 contain A. fimbriata, Tempére & Peragallo, Diatomées, ed. 2, ns. 348, 349 contain A. fimbriata and probably also A. baldjickii subsp. lorenziana, n. 371 contains A. baldjickit subsp. lorenziana, and n. 443 probably contains one or other or both of these species. Cleve-Euler in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5) : 18, fig. 510 (1953). CAPARTOGRAMMA AND SCHIZOSTAURON 81 The description of the raphe-less valve refers to A. danica but that of the raphe valve repeats Hustedt’s error that the striae are 26 in 10 at the centre. The only original figure, fig. 510 c, is indeterminable, but the density of its striation and also, to some extent, its shape suggest A. fimbriata rather than A. danica. If this is so, the fossil record from Skagganias in Smaland also refers to A. fimbriata. Job, Suppl. 1 List Diat. Alexandria : [4] (1961). This record is based on correctly identified specimens. ACHNANTHES BALDJICKII (Brightw.) Grun. in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 17 (2) : 21 (1880).—De Toni, Syll. Alg. 2 : 481 (1891).—Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 27 (3) : 187, t. 3 figs. 4, 5 (1896). Odontidium baldjickit Brightw. in Quart. Journ. Microsc. Sci. 7 : 180, t. 9 fig. 10 (1859). Dimeregramma baldjickii (Brightw.) Ralfs in Pritch., Hist. Infus., ed. 4 : 791 (1861).— Walker & Chase, Not. New & Rare Diat. [1] : 1, t. 1 fig. 3 (1886). Rhaphoneis baldjickit (Brightw.) Grun. in Verh. Zool.-bot. Ges. Wien 12 : 379 (1862). Achnanthidium baldjickii (Brightw.) Grun. ex A. Schmidt, Atl. Diat.: t. 198 figs. 44-48 (1895). Valves rhombic to lanceolate, sometimes slightly gibbous at the centre, subrostrate to rostrate at the apices, 25—gou long, 13-40 broad. Raphe-less valve with a lanceolate or moderately broad linear area, often transversely expanded on one or both sides at the centre ; striae each consisting of a single hyaline alveolus, distant, slightly to moderately radiate, some in the centre part of the valve usually short. Raphe valve with an elevated marginal zone ; raphe filiform, straight except near the central nodule, where the two halves are slightly bent in opposite directions ; axial area narrow ; central area transversely expanded ; striae sigmoid, very strongly radiate, longer and shorter in the central part of the valve, consisting of a Fics. 27-29. Achnanthes baldjickii (Brightw.) Grun. subsp. baldjickii. Fig. 27, raphe valve (in B.M. 43404) (x 1,000) ; Fig. 28, detail of same valve as Fig. 27 ( x 2,000) ; Fig. 29, raphe-less valve of same specimen as Fig. 27 ( x 1,000). Fics. 30-31. Achnanthes baldjickii subsp. lorenziana (Grun.) Ross. Fig. 30, raphe valve (in Coll. Grunow n. 864a) ; Fig. 31, raphe-less valve (in Coll. F. W. Adams n. E 230). (Both x 1,000.) 82 CAPARTOGRAMMA AND SCHIZOSTAURON single row of fine puncta except on the elevated marginal zone, where they are double (? and treble) rows of puncta separated by costae. Subsp. BALDJICKU. (Figs. 27-29.) Achnanthes baldjickti var. subquadrata Pant., Beitr. Kenntn. Foss. Bacill. Ung. 3 : t. 13 fig. 202 (1892) ; Beschr. Neu. Bacill. Pars III Beitr. Kenntn. Foss. Bacill. Ung.: 4 (1905). Valves rhombic, usually subrostrate and gibbous at the centre, 25-goy long, 17°5-40u broad. Raphe-less valve with long and short striae in the centre part, striae radiate, gently curved, 3-6 in 10y, closer towards the apices. Raphe valve with shortened striae to about half-way to the apices ; striae g-12 in I0y at the inargin, 6-7-5 in 10u against the raphe. Known only from marine pliocene deposits in Bulgaria and Hungary. FossiL : Phocene : BULGARIA: Balchik (B.M. 2305, 2311, 3399, 7157, 9721, 21073, 21074, 21620 ; B.M. 26856, 26857 = Van Heurck, Typ. Synops. Diat. Belg. ns. 545, 546; B.M. 30827, 31384, 32629, 32630, 32836, 33305, 33323, 43407, 43408, 45881, 54190, 58689 ; Coll. F. W. Adams ns. A 167, Bess. 895, Bess. 1660, F 153 in Herb. Brit. Mus.). HunGary : Borostelek (B.M. 33350, 70882). Subsp. lorenziana (Grun.) Ross, stat. nov. (Figs. 30, 31.) Rhaphonets lorenziana Grun. in Verh. Zool.-bot. Ges. Wien 12 : 381, t. 4 fig. 5 a, b (1862) (R. fluminensis errore in expl. tab., tom. cit.: 471).—Pant., Beitr. Kenntn. Foss. Bacill. Ung. 2 : 63, t. 28 fig. 401 (1889) (R. lorenziana var. minor in expl. tab.). Rhaphoneis fluminensis sensu Van Heurck, Synops. Diat. Belg.: t. 36 fig. 34 (1881) ; non Grun. ? Rhaphoneis scutellum sensu Petit in Mém. Soc. Sci. Nat. Math. Cherbourg 23 : 205, t. 3 fig. 6 (1881) ; non Ehrenb. Achnanthes danica sensu Pant., tom. cit.: 57, t. 4 fig. 66 (1889) ; non Grun. Achnanthes baldjickii sensu Pant., op. cit. 3 : t. 13 fig. 204 (1892) ; Beschr. Neu. Bacill. Pars. III Beitr. Kenntn. Foss. Bacill. Ung.: 4 (1905) ; non Grun. sensu stricto. Achnanthes lorenziana (Grun.) Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Fdéljd, 27 (3) : 186 (1896) excl. parte maxima descr.; non A. lorenziana Grun. ex A. Schmidt (1895).—H. & M. Perag., Diat. Mar. France : 8, t. 2 fig. 4 (1897).—Hust. in Rabenh., Krypt.-Fl. Deutsch. 7 (2) : 395, fig. 844 (1933) excl. descr. valvae cum raphide.—Cleve- Euler in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5): 18, fig. 511 a (1953) excl. descr. valvae cum raphide. Actinoneis lorenziana (Grun.) Mereschk. in Mém. Soc. Nat. Nouv.-Russ. 24 (2) : 43, 65, figs. 12, 13 (1902) (? excl. descr., fig. et var. parva Mereschk.). Valves rhombic-lanceolate to lanceolate, subrostrate to rostrate, 25—5ou long, 13-20 broad. Raphe-less valve with one or two short striae opposite the central nodule on one side, usually without other short striae, sometimes with one or two more ; striae scarcely to slightly radiate, 6-10 in I0u. + Raphe valve with shortened striae only opposite the central area ; striae 13-16 in Iou. Mediterranean (and possibly other seas—see below), and fossil from marine miocene deposits in Hungary and marine pliocene deposits in Austria, Hungary and Bulgaria. SPAIN : Balearic Islands (Coll. F. W. Adams ns. E 230, E 231 in Herb. Brit. Mus.). ITALy : Capri (B.M. 33068, 71125). CAPARTOGRAMMA AND SCHIZOSTAURON 83 JuGosLaviA: Mali LoSinj, LoSinj Island, Gulf of Kvarner, Reichardt (Coll. Grunow ns. 864a, 864b in Herb. Vienna). Porto Zubzamki, SuSak Island, Gulf of Kvarner, Reichardt (Coll. Grunow n. 869b in Herb. Vienna). FossiL: Miocene : HunGary : Nyermegy (B.M. 68520, 68521 = Temp. & Perag., Diatomées, ed. 2, ns. 174, 175). Pliocene : AUSTRIA : Styria (B.M. 68717 = Temp. & Perag., Diatomées, ed. 2, n. 371). Buicaria : Balchik (B.M. 2311, 7157 ; B.M. 26856, 26857 = Van Heurck, Typ. Synops. Diat. Belg., ns. 545, 546 ; B.M. 36269, 36230, 33305). HunGary : Bory (B.M. 33342). Bremia, near Kavna (B.M. 33302). It is not possible to be completely certain that the striae on the raphe valve of Achnanthes baldjickit subsp. lorenziana consist of a double row of puncta near the margin. All the raphe valves seen have been parts of whole frustules and the structure of the striae in the marginal zone is too fine to be resolved with the optical microscope, at any rate in whole frustules where the presence of the other valve interferes with image formation. Nevertheless it is possible to see that the striae broaden as they approach the marginal zone in a way which suggests that they, like those of the type subspecies, are double there. The costae between the striae are certainly present in the marginal zone of both subspecies. The very characteristic structure of the raphe valve (Figs. 27, 28, 30) of this species has not previously been correctly described or adequately figured. The descriptions by Grunow (K. Svensk. Vetensk.-Akad. Handl., Ny Féljd, 17 (2) : 21 (1880)) and by Cleve later in the same periodical (27 (3) : 187 (1896)) do not mention the doubling of the puncta near the margin, nor the costae, and these two features are not shown in Cleve’s figure (tom. cit.: t. 3 fig. 5 (1896)) nor in those of A. Schmidt (Atl. Diat.: t. 198 figs. 45, 48 (1895)). It has already been pointed out (p. 78) that the description of the raphe valve of A. lorenziana given by Cleve in the Synopsis of the Naviculoid Diatoms and copied by Hustedt and Cleve-Euler is actually of the raphe valve of A. danica. In the pliocene deposit from Balchik, Bulgaria, in which both subspecies occur, there are a few individuals which are to some extent intermediate between them in that, whilst having the outline of A. baldjickii subsp. lorenziana and a raphe-less valve with closer and less radiate striae, they have one or two shortened striae on this valve other than the two on one side opposite the central nodule. I have assigned such specimens to A. baldjickii subsp. lorenziana. The two subordinate taxa within the species have very different ranges in time, and possibly also in space, and when they do occur together in the same fossil deposit intermediates are very few. There is thus adequate basis for distinguishing the two as subspecies, but their close resemblance in morphology and the few intermediates found in the Balchik material preclude separation at a higher taxonomic level. The type subspecies is only known from the pliocene of Balchik, Bulgaria, and of 84 CAPARTOGRAMMA AND SCHIZOSTAURON Borostelek, Hungary. A. baldjickit subsp. lorenziana, however, occurs in the mio- cene of Nyermegy, Hungary, and the pliocene of Bory and Bremia, both in Hungary, of Balchik and of Styria, and it is also found in recent gatherings. I have seen only one specimen from Bory and one broken one from Bremia, each consisting solely of a raphe-less valve. They both have 7-7°5 striae in 10u, and there is sufficient of the specimen from Bremia, which is broken across near the centre, to show that at the centre it has two short striae on one side. Whilst the identifications of these two specimens are made with confidence, there can be no absolute certainty in the absence of associated raphe valves. A. hauckiana Grun. is present in the Bremia material, but the specimens referred to A. baldjickit subsp. lorenziana have their striae so widely spaced and so narrow that it is unlikely that they are raphe-less valves of A. hauckiana. The species is rare in recent gatherings. I have only seen it from the Gulf of Kvarner, from Capri, and from the Balearic Islands. H. Peragallo’s (in H. & M. Perag., Diat. Mar. France : t. 2 fig. 4 (1897)) figure of a specimen from Barcelona is obviously correctly identified. Leuduger-Fortmorel (Bull. Soc. Bot. France 25 : 23 (1878)) records it as Rhaphoneits lorenziana from Verdelet, Cdtes-du-Nord, France, but gives no figure or description. Cleve records A. lorenziana from Bab el Mandeb and the Macassar Straits, but there can be no certainty as to whether these records are correct or are based on A. danica. He also considers that Petit’s record of Rhaphoneis scutellum from Ning-po, China, is based on this species, but Petit’s figure is such that this identification, although possible, is doubtful. The specimens from the Sea of Azov described and figured by Mereschkowsky as A ctinonets lorenziana var. parva can also be referred here only with considerable doubt. QOstrup (Danske Diat.: 126 (1910)) records A. lorenziana from Vesterhavsgerne, Denmark, and Cleve-Euler, whose figure is a copy of Grunow’s, records it from a marine deposit of unspecified age at Robertsfors, Sweden. There is therefore some possibility that the species may occur recent outside the Mediterranean, but, in view of the possibility of its being confused with others, this must remain uncertain in the absence of specimens. ACHNANTHES HUNGARICA (Grun.) Grun. in K. Svensk. Vetensk.-Akad. Handl., Ny Féljd, 17 (2) : 20 (1880).—@strup, Danske Diat.: 128, t. 3 fig. 78 (1910).—Hust. in Pascher, Siissw.-Fl. Mitteleur. 10, ed. 2 : 201, fig. 283 (1930) ; in Rabenh., Krypt.-Fl. Deutsch. 7 (2) : 383, fig. 829 (1933).—Cleve-Euler in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5) : 23, fig. 521 (1953). (Figs. 32-37.) Achnanthidium hungaricum Grun. in Verh. Zool.-bot. Ges. Wien 13 : 146, t. 4 fig. 8 (1863). ? Achnanthidium neglectum Schum. in Schr. Phys.-6kon Ges. K6nigsb. 8 : 54, t. 1 fig. 17 (1867). Schizostauron andicola Cleve in Ofvers K. Vetensk.-Akad. Férhandl. Stockh. 38 (10) : 12, t. 16 fig. 8 (1882) (‘‘ andicolum’’). Stauroneis andicola (Cleve) Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Féljd, 26 (2) : 151 (1894).? Stauroneis andicola (Cleve) M. Perag., Cat. Gén. Diat.: 829 (1897). 1 This combination, although used here by Cleve, was not validly published by him (see pp. 63-64). CAPARTOGRAMMA AND SCHIZOSTAURON 85 Cocconets hungarica (Grun.) Schoenf., Diat. Germ,: 126, t. 13 fig. 239 (1907). Achnanthes andicola (Cleve) Hust. in Abhandl. Naturwiss. Ver. Bremen 20 : 279, t. 2 figs. 23, 24 (1911) ; in Pascher, Siissw.-Fl. Mitteleur. 10, ed. 2 : 201, fig. 284 (1930) ; in Rabenh., Krypt.-Fl. Deutsch. 7 (2) : 384, fig. 830 (1933).—Cleve-Euler, loc. cit., fig. 522 (1953) excl. B producta. Microneis hungarica (Grun.) Meister in Beitr. Krypt.-Fl. Schweiz 4 (1) : 98, t. 13 figs. 5, 6 (1912). Oblong-lanceolate with cuneate ends, the largest specimens sometimes gibbous in the centre and with subrostrate apices. Raphe-less valve with very narrow, very slightly oblique pseudoraphe and small circular central area, or, in some of the largest specimens, with a narrow transverse fascia ; striae parallel in the centre, radiate towards the apices, 18-20 in I0y, punctate. Raphe valve with a straight filiform raphe with the terminal fissures turned in opposite directions ; axial area very narrow ; central area transverse and widened outwards on one or both sides of the valve ; large hyaline terminal areas on the largest specimens but not on others ; striae parallel in the centre, radiate towards the apices, 20-22 in Ion, punctate. 32 33 ‘ 35 3 % 37 Fics. 32-37. Achnanthes hungarica (Grun.) Grun. (in Coll. P. T. Cleve n. 1208). Figs. 32-35, raphe valves ; Figs. 36, 37, raphe-less valves. (All x 1,000.) Cosmopolitan in fresh waters, normally as an epiphyte on aquatic phanerogams, especially Lemnaceae, and fossil from a freshwater miocene deposit in France. Because of the wide distribution and frequent occurrence of this species the only specimens cited are the type of Schizostauron andicola and the slides in published sets in which the species occurs. ARGENTINA : Sierra de Velasco, Prov. Rioja (Coll. P. T. Cleve n. 1208 in Herb. Mus. Stockh., type of Schizostauron andicola). Ecuapor : Pallatanga (B.M. 68928 = Temp. & Perag., Diatomées, ed. 2, n. 579). FRANCE : St. Cloud (B.M. 14625 = Temp. & Perag., Diatomées, n. 371 ; B.M. 69016 = Temp. & Perag., Diatomées, ed. 2, n. 666). BELGIuM : Austruweel, near Antwerp (B.M. 26507 = Van Heurck, Typ. Synops. Diat. Belg., n. 196). 86 CAPARTOGRAMMA AND SCHIZOSTAURON Cleve records both Achnanthes hungarica and Schizostauron andicola from the type gathering of the latter. In this gathering there are abundant specimens of A. hungarica ranging in length from 15 to over 40ou. The tranverse fascia in the centre of the raphe valves over 40u long and a small proportion of those slightly shorter than this widens outwards on both sides (Figs. 32-34), but on the other raphe valves it widens outwards on one side only (Fig. 35). The raphe valves over 40u long have hyaline areas about 1:5 long at each apex, and most of these valves have a rather irregular outline, with a tendency to triundulate margins (Fig. 33). I interpret these as being post-auxospore valves of A. hungarica and find confirmation of this view from a valve 4ou long with a hyaline area at one apex but not the other (Fig. 34). Cleve gives the length of Schizostauron andicola as 36u and makes no mention of hyaline areas at the apices, nor does his figure show them. His type was accordingly one of the very few specimens with a transverse fascia widening outwards on both sides but no hyaline areas, such as that illustrated in Fig. 32. In a small proportion of the raphe-less valves over 40u long there is a moderately broad transverse fascia that widens outwards (Fig. 37) ; in the type preparation there are appreciably fewer of these valves than of raphe valves with hyaline apices. All the remaining raphe-less valves, even the largest, have no transverse fascia (Fig. 36). The material from St. Cloud distributed by Tempére and Peragallo also contains, among abundant A. hungarica, a number of raphe valves over 40u long with hyaline terminal areas. The corresponding raphe-less valves have a very narrow transverse fascia, but there is no fascia on the shorter ones. Ostrup figures three raphe valves under the name A. hungarica var. andicola ; all have the fascia expanding outwards on both sides and the two larger have hyaline terminal areas. The raphe-less valve figured by him under the same name has a very narrow transverse fascia similar to that on the largest raphe-less valves in the St. Cloud material. According to his description and figure the terminal fissures of the raphe are both turned in the same direction, but in the Sierra de Velasco and St. Cloud gatherings the long specimens with hyaline terminal areas, like the shorter ones without such areas, have the terminal fissures turned in opposite directions. According to Hustedt’s first (1911) description and figure of A. andicola, the terminal fissures are both bent in the same direction and the raphe-less valve has a narrow transverse fascia. In his later accounts (1930, 1933) the direction of the terminal fissures is neither mentioned nor shown in the figures, but the raphe- less valve is said to have a broad fascia widening outwards like that on the raphe valves. In a more recent paper, Hustedt (Abhandl. Naturwiss. Ver. Bremen 34 : 246 (1957)) comments on the similarity of A. andicola to A. hungarica and on the sporadic occurrence of the former. It thus appears that not only was Schizostauron andicola based on a large specimen of A. hungarica but also the various diatoms later identified with Cleve’s by Ostrup and Hustedt are post-auxospore specimens of the same species. This would account CAPARTOGRAMMA AND SCHIZOSTAURON 87 for their sporadic occurrence, their constant association with A. hungarica, and their comparative rarity in the gatherings in which they have been found. Cleve-Euler maintains A. hungarica and A. andicola as separate species and com- ments on the discrepancies in the accounts of the terminal fissures. She also suggests that a specimen found fossil at Osterbotten might be referred to A. andicola but belong to a distinct infraspecific taxon. Her description and figure are insufficient for positive identification but make it plain that the specimen in question does not belong to this species. NAVICULA PUPULA Kiitz., Kies. Bacill.: 93, t. 30 fig. 40 (1844).—Grun. in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 17 (2) : 45, t. 2 fig. 53 (1880).—Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 26 (2) : 131 (1894).—Hust. in Pasch., Siissw.-F]. Mitteleur. 10, ed. 2 : 281, fig. 467 (1930) ; in A. Schmidt, Atl. Diat.: t. 396 figs. 10-33 (1934) ; in Internation. Rev. Hydrobiol. Hydrogr. 42 : 64 (1942) ; in Rabenh., Krypt.-Fl. Deutsch. 7 (3) : 120, figs. 1254, 1255 (1961).— Cleve-Euler in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5) : 186, fig. 890 (1953). Stauronets rectangularis Greg. in Quart. Journ. Microsc. Sci. 2 : 99, t. 4 fig. 17 (1854). Navicula laevissima sensu W. Smith, Synops. Brit. Diat. 2 : 91 (1856) ; non Kiitz.—Grun. in Verh. Zool.-bot. Ges. Wien Io : 549, t. 4 fig. 5 (1860).—Rabenh., Fl. Eur. Alg. 1 : 188 (1864).—O’Meara in Proc. R. Irish Acad., Ser. 2, 2, Sci.: 376, t. 31 fig. 54 (1875).—Brun, Diat. Alp. Jura : 68, t. 7 fig. 32 (1880). Navicula granum Schum. in Schr. Phys.-dkon. Ges. K6nigsb. 8 : 59, t. 2 fig. 46 (1867). Stauroneis verbania De Not. in Erb. Critt. Ital., Ser. 2 : n. 434 (1871). Stauroneis tatrica Gutw. in Sprawozd. Kom. Fizyjogr. 25 : (24), t. 1 fig. 20 (1890). Schizostauron tatricum (Gutw.) De Toni in Nuova Notarisia [1] : 196 (1890). Schizostauron verbanium (De Not.) De Toni, Syll. Alg. 2 : 225 (1891). Sellaphora pupula (Kiitz.) Mereschk. in Ann. & Mag. Nat. Hist., Ser. 7, 9 : 187, t. 4 figs. 1-5 (1902). Navicula pseudopupula Krasske in Bot. Arch. 3 : 197, fig. 4 (1923).—Cleve-Euler, tom. cit.: 187, fig. 891 (1953). Navicula pseudopupula var. aqueductae Krasske, loc. cit., fig. 8 (1923). Navicula aqueductae (Krasske) Krasske in Abhandl. & Ber. Ver. Naturk. Cassel 54 : 44, t. 2 fig. 23 (1925). Navicula mutata Krasske in Bot. Arch. 27 : 354, fig. 16 (1929). For description and figure, see Hustedt (1930, 1961). A cosmopolitan freshwater species, and widespread in pleistocene fossil deposits. GERMANY : Nordhausen, Kiitzing (B.M. 17918 = Coll. Kiitzing n. 58, ? type). River Main, Miindung des [illegible] bach in dem Main, 11 Sept. 1847, A. Braun (B.M. 18726 = Coll. Kiitzing n. 925). FRANCE : Falaise, Brébisson (B.M. 18725 = Coll. Kiitzing n. 1468). ENGLAND : Lewes, Sussex, Mar. 1853, W. Smith (B.M. 23638). ScoTLAND : Lochend, near Edinburgh, 7 Mar. 1854, Gregory (B.M. 20507, 48023). ITaLy : Verbania, in a puddle near the mouth of the San Bernardino, autumn 1868, De Notaris (Erb. Critt. Ital., Ser. 2, n. 434, isotype of Stauronets verbanta). FossIL : Pleistocene : ScoTLAND : Knock, Isle of Mull, Duke of Argyll (B.M. 22320, isotype of Stauronets rectangularis). 88 CAPARTOGRAMMA AND SCHIZOSTAURON The outline of this species is very variable and there is quite a wide range in the density of its striation. In consequence it is usually regarded as having many varieties (cf. Hustedt, 1930, 1961), or as consisting of more than one species (cf. Cleve-Euler, 1953). When a large number of populations are examined, however, the various forms are found so to intergrade that there seems little justification for the taxonomic recognition of them at any level. In his protologue Kiitzing gives the locality of this species as “‘ Lebend in siissem Wasser bei Nordhausen’’. The only specimens in his herbarium labelled Navicula pupula are the one from the River Main collected by A.Braun and the one from Falaise collected by Brébisson which are cited above. Both these came into his possession after the publication of the Bacillarien. Accordingly, he must either have failed to preserve the gathering in which he found the species or have found it in one of the gatherings from Nordhausen which are preserved in his herbarium but have failed to record the identification on the label. I have examined these gatherings and on his n. 58 I found small numbers of NV. fupula as currently under- stood. These match his figure well, better than anything else present in any of his gatherings from Nordhausen. This gathering is labelled ‘‘ Fragilaria corrugata Kg. Nordhausen ”’ in Kiitzing’s hand and is the only one of that species in his herbarium labelled with its name. In his protologue of Fragilaria corrugata Kiitzing (Kies. Bacill.: 45 (1844)) says “In Siisswassergraben unter verschiedenen andern Diatomeen bei Nordhausen ’’. Further confirmation that the current concept of N. pupula corresponds to Kiitzing’s is provided by the two gatherings in his herbarium labelled with its name. Both these contain the species as now understood. The identification on the A. Braun specimen is in Kiitzing’s own hand. That on the Brébisson specimen is written in ink by Brébisson, but other identifications have been added in pencil in Kiitzing’s hand, and the words “ Navicula Pupula Kg.”’ have been underlined in pencil and the full stop after ‘“ Kg’ converted into an exclamation mark by a pencil stroke, presumably by Kiitzing. This species was independently described by Gregory in 1854 as Stauroneis rectangularis, and W. Smith in 1856 applied the name Navicula laevissima Kiitz. to it, citing S. rvectangularis asa synonym. W.Smith’s Lewes gathering and Gregory’s from Lochend, together with the original material of S. rectangularis, represent such of the material on which W. Smith’s account of N. laevissima is based as is preserved in the British Museum Herbarium. Until 1880 most authors followed W. Smith and applied the name N. laevissima to this species. In 1860, however, Grunow suggested that N. pupula might be this species, and in 1880, after seeing “specimina authentica ’’, he stated firmly that this was so. At the same time he expressed doubts about the true identity of N. laevissima, and ever since then it 1 At this time Grunow was collaborating with Van Heurck in the preparation of the latter’s Synopsis des Diatomées, and Van Heurck had acquired at Eulenstein’s death a duplicate set of Kiitzing’s diatoms removed from the original packets by Eulenstein when the collection was in his hands (cf. Miller & Zaunick, Friedrich Traugott Kiitzing 1807-1893 Aufzeichnungen & Erinnerungen : 20 (1960)). The authentic material Grunow saw was therefore probably some of A. Braun’s River Main gathering and certainly Brébisson’s Falaise gathering, for his figure of the species is of a specimen from “ Falaise, von Kiitzing selbst bestimmt ”’. CAPARTOGRAMMA AND SCHIZOSTAURON 89 has been treated as a doubtful species (cf. Hustedt in Pascher, Siissw.-Fl. Mitteleur. 10, ed. 2 : 279 (1930)). Its type, however, shows it to be the species normally known as N. bacilliformis Grun., for which it provides the correct name (see p. 90 below). This species is the type of the genus Sellaphora Mereschk. (Ann. & Mag. Nat. Hist., Ser. 7, 9 : 186 (1902)). There is little doubt that it and the type of Navicula Bory belong to different genera, but until a comprehensive review of this group of genera can be undertaken, it seems best to retain it in its currently accepted position. One of the names included by Hustedt (1961) in the synonymy of N. pupula is N. borscowit Mereschk. That combination has never been validly published ; at the place referred to, Mereschkowsky described Sellaphora borscow1i as a new species based on marine specimens from San Pedro, California, that lived in Schizonema- type tubes. There is nothing in his description or figure (Ann. & Mag. Nat. Hist., Ser. 7, 9 : 188, t. 4 figs. 6-10 (1902)) to suggest that it is this species. As Mayer (Denkschr. Regensb. Bot. Ges. 21 : 168 (1940)) points out, Krasske’s original descriptions and figures of his N. pseudopupula, N. aqueductae and N. mutata do not show the characteristic terminal nodules of this species, nor, in the case of N. mutata, the laterally expanded central area. Hustedt’s figures in A. Schmidt’s Adlas (t. 396, figs. 10-14, 29-31 (1934)) of Krasske’s original material, however, show these features. Krasske attributes the figures of N. pseudopupula and N. pseudopupula var. aqueductae accompanying his original accounts of these taxa to Hustedt, so there is every reason for believing that Hustedt’s later figures are of the same taxa. He has explained (in Pascher, Siissw.-Fl. Mitteleur. 10, ed. 2 : 282 (1930) ; Internation. Rev. Hydrobiol. Hydrogr. 42 : 65 (1942)) that the original failure to see the characteristic features of these diatoms was due to the use of a mounting medium of too low a refractive index. NAVICULA LAEVISSIMA Kiitz., Kies. Bacill.: 96, t. 21 fig. 14 (1844). Stauroneis wittrockit Lagerst. in Bih. K. Svensk. Vetensk.-Akad. Handl. 1 (14) : 38, t. 2 fig. 15 (1873). Navicula bacilliformis Grun. in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 17 (2) : 44, t. 2 fig. 51 (1880).—Hust. in Pasch., Siissw.-Fl. Mitteleur. 10, ed. 2 : 273, fig. 446 (1930) ; in A. Schmidt, Atl. Diat.: t. 396, figs. 7-9 (1934).—Cleve-Euler in K. Svensk. Vetensk.- Akad. Handl., Ser. 4, 4 (5) : 188, fig. 892 (1953). Navicula fusticulus Ostr., Danske Diat.: 36, t. 1 fig. 19 (1910). Navicula wittrockii (Lagerst.) Cleve-Euler in Comment. Biol. Soc. Sci. Fenn. 4 (14) : 86, t. 5 fig. 145 (1934) ; in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5) : 188, fig. 893 (1953).—Hust. in Rabenh., Krypt.-Fl. Deutsch. 7 (3) : 124, figs. 1256, 1257 (1961). For description and figure, see Hustedt (1930 as N. bacilliformis, 1961 as N. wittrockit). A cosmopolitan freshwater species, and widespread in pleistocene fossil deposits. FossiIL: Pleistocene : ITALY : Santa Fiore (B.M. 18747, 18748 = Coll. Kiitzing n. gb ; B.M. 18820 = Coll. Kiitzing n. gII ; isotypes). This species is based on specimens that were found “ fossil im Bergmehl von San Fiore’. Among the specimens in the British Museum Herbarium labelled “ Coll. fol) CAPARTOGRAMMA AND SCHIZOSTAURON Kitz. Diat.”” there are a number of packets labelled in Eulenstein’s hand with a species name, the locality ‘‘ St. Fiore’”’ and the number “ g1r’’. One of these is labelled “‘ Navicula laevissima Kg.” and the slide n. B.M. 18820 was prepared from this material. In other cases where such packets occur, Kiitzing’s original packet with the same number is present under another species and bears the name of more than one species. No original packet of n. gII is present and this number is too high for a packet that was in Kiitzing’s possession at the time when he wrote the Baczl- larien. There is, however, a specimen labelled “ St. Fiora Campylodiscus Eunotia granulata Nav. binodis gb”’ in Kiitzing’s hand, and this packet contains considerable quantities of the diatomite. It is clearly part of Kiitzing’s original sample of this fossil deposit. Comparison of the microscope slides made from the material in this packet with those made from the packets numbered g11 by Eulenstein shows that all were prepared from the same fossil sample. Whether Kiitzing received a second batch of it at some stage or whether Eulenstein mis-read Kiitzing’s ‘‘ gb ”’ as “‘ gti ”’ (this is possible since Kiitzing’s figures are written very small) one cannot now tell. In this material the species under discussion here, i.e. that called Navicula bacilliformis by all authors from 1880 until 1961 and N. wittrocki by Hustedt in 1961, is not uncommon and is the only one matching Kiitzing’s description and figure. N. pupula also occurs, but much more rarely, and the specimens of it have not the gibbous centre mentioned in Kiitzing’s description (‘‘ medio leviter ventricosa ’’) and shown in his figure. The name UN. laevissima appears never to have been correctly applied by any subsequent author. W. Smith (Synops. Brit. Diat. 2 : 91 (1856)) used it for N. pupula and in this he was followed by a number of authors (cf. p. 88 above). Donkin (Nat. Hist. Brit. Diat.: 28, t. 5 fig. 2 (1871)), however, used it for a form which probably belongs to N. bacillum Ehrenb. In 1880 Grunow (K. Svensk. Vetensk.- Akad. Handl., Ny Foljd, 17 (2) : 45 (1880)), who had seen authentic material of N. pupula but not of N. laevissima, gave an accurate account of N. pupula but considered the identity of N. Jaevissima doubtful ; he gave a description under the name but with a query, and further down the same page said that N. laevissima might be identical with N. ventricosa sensu Donkin (= Calonets stlicula var. ventricosa Cleve), and that the species he had described under the name N. Jaevissima should be given the new name N. pseudobacillum. At the same time he described N. bacilliformis as a new species and suggested that N. laevissima sensu Donkin might belong to it. In Van Heurck’s Synopsis, published later in the same year, Grunow was responsible for two figures on the same plate, one labelled N. pseudobacillum (t. 13 fig. 9) and the other N. laevissima (t. 13 fig. 13). According to Hustedt (in Pasch., Siissw.-Fl. Mitteleur. 10, ed. 2 : 280 (1930)), N. pseudobacillum is a synonym of N. bacillum Ehrenb., and both these figures in the Synopsis seem to represent that species. Except for this uniting of N. bacillum and N. pseudobacillum, Grunow’s treatment of these species in the Avrctischen Diatomeen has normally been followed until very recently ; the name N. pupula has been correctly applied, the identity of N. laevissima has been regarded as doubtful, and the species under discussion has been CAPARTOGRAMMA AND SCHIZOSTAURON 91 called N. bacilliformis. H. and M. Peragallo (Diat. Mar. France : 67, t. 8 fig. 25 (1897)), however, used the name N. Jaevissima for a marine form quite unrelated to any of the species to which it had previously been applied, or misapplied, and Pantocsek (Res. Wiss. Erforsch. Balatonsees 2 (2, 1, Anhang) : 47, t. 5 fig. 112 (1902)) used the name in yet another sense. In 1961 Hustedt called this species N. wittrockii (Lagerst.) Cleve-Euler since he considered that the type material of Stawroneis wittrockii Lagerst. was conspecific with the type material of N. bacilliformis, an opinion which seems to me to be justified by the original accounts of the two. Cleve-Euler herself, on the other hand, maintains that there are two species involved and that N. wittrockii is distinguished from N. bacilliformis by having straight striae. Navicula caucasica Ross, nom. nov. (Fig. 38.) Schizostauron rhombicum M. Perag. in Temp. & Perag., Diatomées, ed. 2 : 398, n. 825 (1913). Valves narrowly rhombic-lanceolate with obtuse, rounded apices, 47-5—60y long, 12-15u broad. Raphe slightly sinuous, somewhat oblique, central pores close, terminal fissures hooked, both turned in the same direction. Axial area rather narrow. Central area three-quarters of the breadth of the valve, expanded out- wards. Terminal areas prominent and rather large. Striae radiate at the centre, convergent at the apices, somewhat curved, very finely punctate, 14-15 in Ioy ; short striae opposite the central nodule of uneven length. 38 Fic. 38. Navicula caucasica Ross (in B.M. 69176) (x 1,000). Known from a single freshwater gathering from the Caucasus. U.S.S.R.: Kisstib, Caucasus (B.M. 69176 = Temp. & Perag., Diatomées, ed. 2, n. 825, isotype). This species resembles Stauroneis amphioxys Greg. but the central area does not extend so far across the cell and widens outwards more markedly, and the striae are wider spaced, more strongly radiate at the centre, and convergent at the apices. 92 CAPARTOGRAMMA AND SCHIZOSTAURON In spite of this resemblance, I place this species in Navicula, not Stauronets, since the central area is not a stauros and the striae do not appear under the light microscope to have the structure typical of true Stawroneis. STAURONEIS SMITHII var. SAGITTA (Cleve) Hust. in Rabenh., Krypt.-Fl. Deutsch. 7 (2) : 811, fig. 1158 (1959). Pleurostauron sagitta Cleve in Cleve & Mller, Diatoms 5 : 6, n. 261 (1879), nom. nud. Stauroneis sagitta Cleve in K. Svensk. Vetensk.-Akad. Handl., Ny Foljd, 18 (5) : 15, t. 3 fig. 45 (1881) ; op. cit. 26 (2) : 151 (1894).—Cleve-Euler in K. Svensk. Vetensk.-Akad. Handl., Ser. 4, 4 (5) : 217, fig. 960 (1953). Stauroneis stefanssonii Ostr. in Rosenv. & Warm., Bot. Iceland 2 : 15, t. 2 fig. 18 (1918) (‘‘ Stefanssoni’’). For description and figure, see Hustedt (loc. cit.). From fresh and slightly brackish waters in northern Scandinavia and Iceland, and fossil from freshwater quaternary deposits in Sweden and Finland. Norway : Mouth of the Tana Elf (B.M. 13024 = Cleve & Moller, Diatoms n. 261, isotype). In the Synopsis of the Naviculoid Diatoms Cleve includes Stauroneis sagitta in the subdivision Schizostauron of the Naviculae microstigmaticae. The resemblance between the species to which this diatom is here referred and Capartogramma is discussed earlier in this paper (p. 53). It has been regarded as a distinct species by most authors, but Hustedt’s treatment of it as a variety of S. smithii seems justified. The division of the stauros on one or both sides in many specimens is the reason why Cleve referred it to Schizostauron, but this division of the stauros is comparable to that of a single tigillum sometimes found in Capartogramma and not to the possession of two diverging tigilla, which is the characteristic of that genus. Bull. B.M. (N.H.) Bot. 3, 2 PLATE 1 C Stereo-pairs of electron micrographs: A, Capartogramma karstenii (Zanon) Ross (x 2,500); B, C. crucicula (Grun., ex Cleve) Ross (x 5,500); C, C. amphoroides Ross (x 2,500). Phot. K. Little. Bull. B.M. (N.H.) Bot. 3, 2 PLATE 2 B C A. Capartogramma crucicula (Grun. ex Cleve) Ross; stereo-pair of electron micrographs. (x 13,000.) B,C. C. karstenii (Zanon) Ross ; electron micrographs. (B x 50,000; C x 35,000.) Phot. K. Little. ANGIOSPERMS OF THE CAMBRIDGE ANNOBON ISLAND EXPEDITION A. W. EXELL BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) - BOTANY , Vol. 3 No. 3 | LONDON : 1963 ANGIOSPERMS OF THE CAMBRIDGE ANNOBON ISLAND EXPEDITION } BY A. W. EXELL Pp. 93-118 ; Plates 3-12 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BOTANY Vol. 3 No. 3 LONDON : 1963 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), ¢stituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 3, No. 3 of the Botany series. © Trustees of the British Museum, 1963 PRINTED BY ORDER OF THE TRUSTEES. OF THE BRITISH MUSEUM Issued September, 1963 _ Price Fifteen Shillings ANGIOSPERMS OF THE CAMBRIDGE ANNOBON ISLAND EXPEDITION By A. W. EXELL THE following is an account of the more interesting Angiospermae collected on the Cambridge Annobon Island Expedition 1959. The two collectors were Fenella A. Melville (now Mrs. Wrigley) and T. C. Wrigley. I have included in the list all new records for Annobon and specimens of species already recorded for the island but meriting inclusion for one reason or another. It is particularly interesting to be able to record a number of gatherings of species described from Mildbraed’s 1g11 collection of which the types (formerly in the Berlin Herbarium) have been destroyed. These have largely been identified ex descriptione : if correctly named they provide (allowing for the very small size of the island) topotypes of the species concerned. Photographs of three of the Mildbraed types and seven of the new topotypes are included in this account as illustrations. Except for the Orchidaceae (which are at Kew) all the specimens collected on the Cambridge Expedition are represented in the British Museum Herbarium, and all specimens cited in this account are in the British Museum Herbarium unless otherwise indicated. In my statements of new records “the islands’’ means the three islands of Principe, S. Tomé and Annobon in the Gulf of Guinea. The four principal works which I have already published on the flora of these islands are: (1) Catalogue of the Vascular Plants of S. Tomé (with Principe and Annobon) (1944) ; (2) Supplement to the Catalogue of the Vascular Plants of S. Tomé (with Principe and Annobon) (1956) ; (3) Aditamento a Flora das Ithas de Séio Tomé e do Principe (Confer. Internacion. Afr. Oc., Sess. 6, 3 : 77-91) (1958), with A. Rozeira ; (4) Additions to the Flora of S. Tomé and Principe (Bull. Inst. Frang. Afr. Noire 21, Sér. A, n. 2) (1959). They are abbreviated in this account as follows : (1) Exell, Cat.; (2) Exell, Suppl.; (3) Exell, Aditam.; (4) Exell, Addit. The third of these works was published without my seeing a proof and is marred by typographical errors. I have taken the opportunity to bring the nomenclature up to date as regards a few species listed in these works and to correct various errors which have been pointed out to me, but I have not troubled to correct the fairly obvious orthographical mistakes in the Aditamento. Taxa not recorded for Annobon are placed within square brackets. I am indebted as usual to numerous colleagues at the British Museum (Natural History), the Royal Botanic Gardens, Kew, and other institutions ; and more particularly to J. E. Dandy (general arrangement), Miss D. Hillcoat (Leguminosae), F. N. Hepper (Rubiaceae), J. Lewis (Loganiaceae), W. T. Stearn (Solanaceae, Piperaceae, Loranthaceae), J. Léonard (Euphorbiaceae), V. S. Summerhayes (Orchidaceae) and W. D. Clayton (Gramineae). 96 ANGIOSPERMS OF ANNOBON ISLAND DILLENIACEAE TETRACERA ALNIFOLIA Willd.—Exell, Cat.: 100. ANNOBON: S.E. of Santa Cruz, 450 m., forest on ridge, scabrid climber and scrambler, 7 Aug. 1959, Wrigley 249. New record for Annobon. ANNONACEAE ANNONA MURICATA L.—Exell, Cat.: ror. ANNOBON : Near Ambo, 27 m., open woodland, 20 July 1959, Wrigley 63. New record for Annobon. Introduced. Soursop. ANNONA RETICULATA L.—Exell, Cat.: ror. ANNOBON : Near Ambo, 27 m., and near Crater Lake, open woodland, 16 July 1959, Wrigley 42. New record for Annobon. Introduced. Custard-apple. [CRUCIFERAE] [RorIPPA INDICA (L.) Hiern, Cat. Afr. Pl. Welw. 1: xxvi (1896).—Hara in Journ. Jap. Bot. 30 : 197 (1955). Sisymbrium indicum L., Syst. Nat., ed. 12, 2: 441 et Mant. Pl.: 93 (1767). Nasturtium sinapis (Burm. f.) O. E. Schulz.—Exell, Suppl.: 10 ; Aditam.: 81. Rorippa sinapis (Burm. f.) Ohwi & Hara.—Exell, Addit.: 450. S. TomE : Ribeira Peixe, Rozetva 236 (Herb. Porto Univ.). Diogo Vaz, Rozeira 869 (Herb. Porto Univ.). This introduced species has previously been recorded from S. Tomé under the names Nasturtium sinapis and Rorippa sinapis. It is widely distributed in south- east Asia from Japan and China to India and Malaysia. The name R. indica has been misapplied to another species, R. madagascariensis (DC.) Hara (tom. cit.: 198).] CAPPARACEAE CLEOME RUTIDOSPERMA DC., Prodr. I: 241 (1824). Cleome ciliata Schumach.—Exell, Cat.: 105 ; Suppl.: Io. ANNOBON: N.W. of the Island, 6 m., rocks and cassava plantations by sea, 11 July 1959, Melville 129. De Candolle’s name replaces that of Schumacher. The species is recorded for Principe, S. Tomé and Annobon and is widespread in tropical Africa southwards as far as Northern Rhodesia. PORTULACACEAE PORTULACA OLERACEA L.—Exell, Cat.: 110 ; Addit.: 450. ANNOBON : Ambo, by the expedition’s houses, 15 m., sandy place by house, 21 July 1959, Melville 168. New record for Annobon. ANGIOSPERMS OF ANNOBON ISLAND 97 TALINUM TRIANGULARE (Jacq.) Willd.—Exell, Cat.: 110 ; Suppl.: 11. ANNOBON:: N.W. of San Pedro, forest, 25 July 1959, Melville 194. New record for Annobon. Introduced. MALVACEAE SIDA RHOMBIFOLIA L.—Exell, Cat.: 115. ANNOBON : Near Capelle San Juan, 30 m., wayside in cultivated land, 10 Aug. 1959, Melville 258. New record for Annobon. GOSSYPIUM BARBADENSE L.—Exell, Cat.: 119. ANNOBON:: Near Ambo, 9 m., pathside, 11 July 1959, Wrigley 12. New record for Annobon. Introduced ; presumably an escape from cultivation. RUTACEAE CITRUS AURANTIFOLIA (Christm.) Swingle.—Exell, Cat.: 131. ANNOBON : Near Ambo, 24 m., waste ground, 12 July 1959, Wrigley 25. New record for Annobon. Introduced. Lime. OCHNACEAE OcHNA MEMBRANACEA Oliv., Fl. Trop. Afr. 1: 316 (1868).—Henriq. in Bol. Soc. Brot. 10 : 105 (1893). Diporochna quintasii Tiegh. in Ann. Sci. Nat., Sér. 8, Bot. 16 : 392 (1902). Ochna quintasii (Tiegh.) Exell, Cat.: 132. Ochna cf. gilgiana, Exell, Cat.: 132. ANNOBON : Between Capelle San Juan and San Pedro, 45 m., steep wood, cliffs by sea, 11 Aug. 1959, Melville 260. Recorded from Annobon by Mildbraed as O. cf. gilgiana. Also in S. Tomé and widespread in western tropical Africa, eastwards to Uganda and southwards to Angola. MELIACEAE TURRAEA GLOMERULIFLORA Harms.—Exell. Cat.: 136. (Plate 3.) ANNOBON : Ridge west of Crater Lake, 360 m., open woodland ; ‘ small sepals ; green petals ; white filament tube with 10 stamens’; 18 July 1959, Wrigley 52. Endemic to Annobon. Wrigley’s locality is quite near to that of the holotype, Mildbraed 6487 (Herb. Berlin, destroyed). A photograph of the holotype exists in the British Museum Herbarium and is reproduced in Plate 3. CELASTRACEAE Maytenus annobonensis (Loes. & Mildbr.) Exell, comb. nov. (Plate 4.) Gymnosporia annobonensis Loes. & Mildbr. in Notizbl. Bot. Gart. & Mus. Berl.-Dahl. 12: 30 (1934).—Exell, Cat.: 138. 98 ANGIOSPERMS OF ANNOBON ISLAND ANNOBON : Ridge west of Crater Lake, 360 m., open woodland ; “ flower white ; 3-merous fruit with I orange seed’; 18 July 1959, Wrigley 56. Endemic to Annobon. This material was collected in almost the same locality as the holotype, Mildbraed 6541 (Herb. Berlin, destroyed), and the only doubt about the identification lies in the fact that the original specimen had neither flowers nor fruit so that here “ex descriptione’’ has little meaning. Actually this new gathering by Wrigley confirms the supposition that the species belongs to the genus Maytenus (Gymnosporia). It differs from M. monodii Exell (endemic to S. Tomé) by its dense ferrugineous pubescence on the young stems, being nearer in this respect to M. welwitschiana Exell & Mendonca (Angola) but with larger flowers than in the latter species and with a glabrous (or only minutely pubescent) capsule. ANACARDIACEAE MANGIFERA INDICA L.—Exell, Cat.: 145. ANNOBON : Common tree, lower grassland and up to Crater Lake, regenerating very freely, 19 July 1959, Wrigley 6r. New record for Annobon. Introduced. CONNARACEAE AGELAEA OVALIS Schellenb.— Exell, Cat.: 148. (Plate 5.) ANNOBON : Ambo, go m.; dry valley ; “tree c. 30 ft.; fruit red and yellow with black and orange seed ; flowers off-white (pinkish) ’’; 2 Aug. 1959, Melville 232. Endemic to Annobon. Melville’s plant is almost certainly A. ovalis, of which the holotype, Mzldbraed 6679 (Herb. Berlin), has been destroyed. It is, however, difficult to believe that the plant is a tree and I suspect an error in the collector’s note. LEGUMINOSAE ARACHIS HYPOGAEA L.—Exell, Cat.: 155. ANNOBON : Cultivated on a small scale during the wet season, 18 Aug. 1959, Wrigley 297, fruit only. New record for Annobon. Introduced. DESMODIUM RAMOSISSIMUM G. Don.—Exell, Suppl.: 16. Desmodium mauritianum sensu Exell, Cat.: 158. ANNOBON : Path towards Pico do Fogo and Crater Lake, 90 m., pathside, 10 July 1959, Melville ro5. New record for Annobon. MUCUNA SLOANEI Fawc. & Rendle in Journ. of Bot. 55 : 36 (1917).—Hepper in Hutch.:& Dalziel, Fl. W: Trop. Afr.,-ed..2, 1: 561 (1958): Mucuna urens sensu Exell, Cat.: 160. ANNOBON : Near Ambo, 24 m., climbing over trees, 14 July 1959, Wrigley 3o. Recorded for S. Tomé and Annobon as M. urens ; widespread in tropical Africa and tropical America. Calabar Bean. ANGIOSPERMS OF ANNOBON ISLAND 99 VIGNA UNGUICULATA (L.) Walp., Repert. Bot. Syst. 1: 779 (1842).—Hepper in Hutch. & Dalziel, Fl. W. Trop. Afr., ed. 2, 1 : 569 (1958). Dolichos unguiculatus L., Sp. Pl. 2: 725 (1753). Clitoria alba G. Don, Gen. Syst. 2 : 215 (1832). Vigna alba (G. Don) Planch. ex Baker f.—Exell, Cat.: 161. ANNOBON: East path towards Crater Lake, 135 m., pathside and cassava plantation, running over rocks, 18 July 1959, Melville 162. New record for Annobon. The species occurs also in S. Tomé and is widespread in tropical and subtropical regions. CAJANUS CAJAN (L.) Millsp.—Exell, Cat.: 163. ANNOBON: Near Ambo, 30 m., open grassland, 23 July 1959, Wrigley 76. New record for Annobon. Introduced. ERIOSEMA GLOMERATUM (Guillem. & Perrott.) Hook. f. in Hook., Niger Fl.: 313 (1849). Rhynchosia glomerata Guillem. & Perrott. in Guillem., Perrott. & Rich., Tent. FI. Senegamb. I : 216 (1832). ANNOBON : Near Ambo, 30 m., pathside, 17 July 1959, Wrigley 46. New record of the genus Eviosema (DC.) Desv. for the islands. EF. glomeratum is widespread in tropical Africa. CAESALPINIA PULCHERRIMA (L.) Swartz.—Exell, Cat.: 165. ANNOBON : Mission, Ambo, 15 m., garden, 4 Aug. 1959, Melville 233. New record for Annobon. Cultivated. Barbados Pride. CASSIA OCCIDENTALIS L.—Exell, Cat.: 166. ANNOBON : Path between Capelle San Juan and San Pedro, 60 m., pathside, 2 Aug. 1959, Wrigley 230. New record for Annobon. CASSIA OBTUSIFOLIA L., Sp. Pl. : 377 (1753).—Brenan in Kew Bull. 13 : 248 (1958). Cassia tora sensu Exell, Cat.: 167. ANNOBON : Path between Capelle San Juan and San Pedro, 60 m., pathside, 2 Aug. 1959, Wrigley 230a. New record for Annobon. The species occurs also in S. Tomé and is widespread in the tropics. According to Brenan (tom. cit.: 248-252) C. tora L. is a different species which does not occur in West Africa. CRASSULACEAE KALANCHOE CRENATA (Andr.) Haw.—Exell, Cat.: 173 ; Suppl.: Ig. ANNOBON : Near shore near Ambo, 6 m., edge of dry river, 4 Aug. 1959, Wrigley 235. | New record for Annobon. 100 ANGIOSPERMS OF ANNOBON ISLAND RHIZOPHORACEAE CASSIPOUREA ANNOBONENSIS Mildbr. ex Alston.—Exell, Cat.: 174. (Plate 6.) ANNOBON : Pico do Fogo on S.W. side, 360 m., edge of forest at base of Pico ; “flowers 4-merous ; tree 30 ft. (branches well spaced for climbing)”; 25 July 1959, Melville rgo. Endemic to Annobon. The locality is near to that of the holotype, Mildbraed 6511 (Herb. Berlin, destroyed). A photograph of the holotype exists in the British Museum Herbarium and is reproduced in Plate 6. COMBRETACEAE TERMINALIA CATAPPA L.—Exell, Cat.: 175. ANNOBON: Near Ambo, 24 m., planted; ‘the nuts are eaten’; 18 Aug. 1959, Wrigley 296. New record for Annobon. Introduced. Indian Almond. MELASTOMATACEAE TRISTEMMA OREOTHAMNOS Mildbr.—Exell, Cat.: 178. (Plate 7.) ANNOBON : Pico Surcado, 420 m., forest ; “‘ flower pale pink’; 21 July 1959, Wrigley 65. Crater Lake, N.E. side, 360 m., forest near base of Pico do Fogo ; “corolla pink ; anthers yellow’; 25 July 1959, Melville 193. Endemic to Annobon. The holotype of this species, Mildbraed 6677 (Herb. Berlin), has been destroyed ; it was collected at a somewhat higher altitude on the neighbouring peak of Quioveo. PUNICACEAE PUNICA GRANATUM L., Sp. Pl. 1: 472 (1753). ANNOBON : Ambo, near the Mission, 15 m., cultivated, 2 Aug. 1959, Melville 219. New record of the genus Pumica L. for Annobon. P. granatum, the Pomegranate, is a widely cultivated species of Asiatic origin. CARICACEAE CARICA PAPAYA L.—Exell, Cat.: 183. ANNOBON : Crater Lake, 265 m., lakeside cultivation, 24 July 1959, Melville 182. New record for Annobon. Introduced. CUCURBITACEAE [PEPONIUM VOGELII (Hook. f.) Engler in Engler & Prantl, Nat. Pflanzenfam., Nachtr. zum 2-4 : 318 (1897).—Jeffrey in Kew Bull. 15 : 359 (1962). Peponia vogelii Hook. f. in Oliv., Fl. Trop. Afr. 2 : 526 (1871). Peponium bracteatum (Cogn.) Cogn.—Exell, Cat.: 184. Described from S. Tomé as Peponza bracteata Cogn., but this is now referred by Jeffrey (loc. cit.) to Pepbonium vogelii, a species widely distributed in tropical Africa and extending into Natal.] ANGIOSPERMS OF ANNOBON ISLAND IOI LAGENARIA BREVIFLORA (Benth.) Roberty in Bull. Inst. Franc. Afr. Noire 16, Sér. A: 795 (1954). Adenopus breviflorus Benth.—Exell, Cat.: 184. ANNOBON : Near Ambo, 24 m., pathside, 12 July 1959, Wrigley 19. Near Ambo, 30 m. (common in other villages), hedges, roofs, etc., climber, 9 Aug. 1959, Wrigley 254. New record for Annobon. Jeffrey (Kew Bull. 15 : 355 (1962)), in a new classification of the family Cucurbit- aceae, includes Adenopus Benth. within the genus Lagenarza Ser. LAGENARIA SICERARIA (Molina) Standl. in Publ. Field Mus. Nat. Hist., Bot. Ser. 3 : 435 (1930). Cucurbita lagenaria L., Sp. Pl. 2: 1010 (1753). Cucurbita siceraria Molina, Sagg. Stor. Nat. Chil.: 133, 355 (1782). ANNOBON : Ambo village, 9 m., cultivated ; “ fruit used as calabash”’ ; 20 Aug. 1959, Melville 302. New record for Annobon. Widespread and frequently cultivated in the tropics. Calabash, Bottle-gourd or White Pumpkin. LUFFA AEGYPTIACA Mill.— Exell, Cat.: 184. ANNOBON : Ambo, near the expedition’s houses, 24 m., climbing over top branches of Jatropha curcas, 12 July 1959, Melville 133. San Pedro, 15 m., village rubbish heap, 2 Aug. 1959, Wrigley 229. New record for Annobon. [DIPLOCYCLOs PALMATUS (L.) Jeffrey in Kew Bull. 15 : 352 (1962). Bryonia palmata L., Sp. Pl. 2: 1012 (1753). Bryonopsis laciniosa Naud. in Ann. Sci. Nat., Sér. 4, Bot. 12: 141 (1859) pro parte.— Exell, Cat.: 185. Jeffrey (tom. cit.: 352-354) has adjusted the nomenclature of this widely distri- buted palaeotropical species, which is recorded for S. Tomé. He adopts the generic name Dzplocyclos (Endl.) Post & Kuntze.] [ZEHNERIA GILLETII (De Wild.) Jeffrey in Kew Bull. 15 : 366 (1962). Melothria cordifolia Hook. f. in Oliv., Fl. Trop. Afr. 2 : 563 (1871) ; non Zehneria cordifolia Schweinf. ex Broun & Massey (1929). Melothria gilletit De Wild. in Ann. Mus. Congo, Bot. Sér. 5, 3 : 140, t. 13 figs. 4-6 (1907). Melothria capillacea sensu Exell, Cat.: 185 ; Suppl.: 22. Occurs in S. Tomé and on the mainland of western tropical Africa from Liberia to Gabon and Congo. This and the following species belong to the palaeotropical genus Zehneria Endl. which is separated by Jeffrey (tom. cit.: 343) from the exclusively New World genus Melothria L.| 102 ANGIOSPERMS OF ANNOBON ISLAND [ZEHNERIA SCABRA (L. f.) Sond. in Harv. & Sond., Fl. Cap. 2: 486 (1862).—Jeffrey in Kew Bull. 15 : 369 (1962). Bryonia scabra L. f., Suppl. Pl.: 423 (1781). Zehneria scrobiculata Hochst. ex A. Rich., Tent. Fl. Abyss. 1 : 287 (1848). Melothria minutifiora sensu Exell, Cat.: 185. Recorded from S. Tomé ; widely distributed in the Old World tropics, extending into South Africa. ] ARALIACEAE POLYSCIAS GUILFOYLEI (Bull) L. H. Bailey in Rhodora 18 : 153 (1916). Aralia guilfoylei Bull, Cat. 1873 : 4 (1873). Var. LACINIATA L. H. Bailey, loc. cit. (1916). ANNOBON : Capelle San Juan, 4:5 m., garden of Capelle, 14 Aug. 1959, Melville 281. New record for the islands. Introduced ; a cultivated plant of Polynesian origin. RUBIACEAE IxXORA COCCINEA L.—Exell, Suppl.: 24. ANNOBON: Ambo, near the Mission, 15 m., cultivated land, 2 Aug. 1959, Melville 278. New record for Annobon. Introduced. CoFFEA LIBERICA Bull ex Hiern.—G. Taylor in Exell, Cat.: 208. ANNOBON : Ridge west of Crater Lake, 360 m., open woodland, 18 July 1959, Wrigley 50. New record for Annobon. Introduced. BERTIERA ANNOBONENSIS G. Taylor.—G. Taylor in Exell, Cat.: 208. (Plate 8.) ANNOBON : Near highest point of Crater, 480 m., forest ; ‘“‘ undershrub, two specimens with white corolla, others fruit only ’’; 26 July 1959, Wrigley 93. Endemic to Annobon. Wrigley 93 is topotypical material of Mildbraed 6502 (Herb. Berlin), one of the destroyed syntypes. The other syntype, Mildbraed 6760, came from Santa Mina. UNCARIA AFRICANA G. Don, Gen. Syst. 3 : 471 (1834). ANNOBON : Pico Surcado and Monte Abecin, 570 m., forest climber, 11 Aug. 1959, Wrigley 304. New record of the genus Uncaria Schreb. for the islands. This is the typical variety of U. africana, distributed on the mainland of West Africa from Portuguese Guinea eastwards and extending to Sudan, Uganda and Tanganyika. [GEOPHILA NEURODICTYON (K. Schum.) Hepper in Kew Bull. 16 : 331 (1962). Psychotria neurodictyon K. Schum.—G. Taylor in Exell, Cat.: 213. Recorded from Principe. ANGIOSPERMS OF ANNOBON ISLAND 103 The generic name Geophila D. Don (non Bergeret) has been proposed for conserva- tion over Carinta W. F. Wight, the name adopted by G. Taylor (in Exell, Suppl.: 25) for the three species previously listed by him (in Exell, Cat.: 217) under Geophila ; cf. Taxon 9 : 88 (1960).] OLDENLANDIA HERBACEA (L.) Roxb., Hort. Bengal.: 11 (1814). Hedyotis herbacea L., Sp. Pl. 1 : 102 (1753). ANNOBON : N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Wrigley Fi; New record for the islands. Widely distributed in the Old World tropics. COMPOSITAE ADENOSTEMMA PERROTTETII DC.—Exell, Cat.: 222 ; Suppl.: 26. ANNOBON : South of Crater Lake, near higher point of Crater, 420 m., cultivated land, cassava plantations in the hills, 1 Aug. 1959, Melville 2z0. New record for Annobon. MIKANIA CORDATA (Burm. f.) B. L. Robinson.—Cannon in Exell, Suppl.: 27. Mikania scandens sensu Exell, Cat.: 222. Mikania sp., Cannon in Exell, Suppl.: 27. ANNOBON : Near Ambo, 30 m., scrub, climber, 12 July 1959, Wrigley 27. Previously recorded from Annobon as Mtkania sp. (DICHROCEPHALA INTEGRIFOLIA (L. f.) Kuntze, Revis. Gen. Pl. 1 : 333 (1891). Hippia integrifolia L. f., Suppl. Pl.: 389 (1781). Dichrocephala bicolor (Roth) Schlecht.—Exell, Cat.: 223 ; Suppl.: 31. The epithet integrifolia, though inapt, is the correct one for this widespread species which is recorded for S. Tomé. ] SYNEDRELLA NODIFLORA (L.) Gaertn.—Exell, Cat.: 226 ; Suppl.: 31. ANNOBON:: North shore of Crater Lake, 265 m., pathside, 24 July 1959, Wrigley 80. New record for Annobon. ELEUTHERANTHERA RUDERALIS (Swartz) Schultz Bip. in Bot. Zeit. 24 : 164 (1866). Melampodium ruderale Swartz, Fl. Ind. Occ. 3 : 1372 (1806). ANNOBON : Near Capelle San Juan, 30 m., wayside in cultivated land, 10 Aug. 1959, Melville 257. New record of the genus Eleutheranthera Poit. ex Bosc for the islands. E. ruderalis is a pantropical weed. CAMPANULACEAE CEPHALOSTIGMA PERROTTETII A. DC., Monogr. Campan.: 118 (1830). ANNOBON : Pico do Fogo, 360 m., among rocks, 24 July 1959, Wrigley go. New record of the genus Cephalostigma A. DC. for the islands. C. perrotteti is widespread in tropical Africa and tropical South America. 104 ANGIOSPERMS OF ANNOBON ISLAND APOCYNACEAE ALLAMANDA CATHARTICA L., Mant. Pl. Alt.: 214 (1771). ANNOBON : Ambo, near the Mission, 15 m., 2 Aug. 1959, Melville 220. New record of the genus Allamanda L. for the islands. A. cathartica, native of tropical America, is widely cultivated in the tropics. CATHARANTHUS ROSEUS (L.) G. Don.—Philipson in Exell, Cat.: 239. ANNOBON : Ambo, 27 m., waste ground, 19 July 1959, Wrigley 58. New record for Annobon. Introduced. PLUMERIA RUBRA L.—Philipson in Exell, Cat.: 241. ANNOBON : Near Ambo, 24 m., in front of church, planted, 18 Aug. 1959, Wrigley 293. New record for Annobon. Introduced. ASCLEPIADACEAE ASCLEPIAS CURASSAVICA L.—Exell, Cat.: 242 ; Suppl.: 34. ANNOBON : Ambo, 180 m. S.W. of church, 22 m., cleared ground, 10 July 1959, Melville ro2. New record for Annobon. Introduced. LOGANIACEAE (By J. Lewis) NUXIA CONGESTA R. Br. ex Fresen. in Flora 21 (2) : 606 (1838).—Bruce & Lewis in Five iiop: Air, Logan; 44, fig’ §.(7, 8) (1900): Nuxia congesta R. Br. in Salt, Voy. Abyss., App.: 63 (1814), nom. nud. Var. CONGESTA. (Plate 9.) Nuxia angolensis Gilg in Notizbl. K. Bot. Gart. & Mus. Berl. 1 : 74 (1895). Nuxia mannii Gilg in Engler, Bot. Jahrb. 30 : 376 (1901). Nuxia annobonensis Mildbr. in Wiss. Ergebn. Zweit. Deutsch. Z.-Afr.-Exped. 2: 163 (1922), nom. nud. Lachnopylis mannii (Gilg) Hutch. & Moss in Hutch. & Dalziel, Fl. W. Trop. Afr. 2: 20, fig. 185 (1931). Lachnopylis annobonensis Mildbr.—Philipson in Exell, Cat.: 244. Lachnopylis angolensis (Gilg) Philipson in Exell, Cat.: 245 (1944). ANNOBON : Pico do Fogo, 2 m. and less below top of Pico, rocky mountainside, 12 Aug. 1959, Melville 277. Widely distributed on the mountains of tropical Africa and extending into South Africa ; previously recorded for Annobon under the name Lachnopylis annobonensis, of which Melville 277 is an exact topotype. The holotype of L. annobonensis, Miuldbraed 6561 (Herb. Berlin), is destroyed, but there is an isotype in the British Museum Herbarium accompanied by a photograph of the holotype which is reproduced in Plate g. The opinions expressed by Bruce and Lewis (loc. cit.) and by Verdoorn (Bothalia 7 : 14 (1958)) concerning the great variability and wide range of this species are ANGIOSPERMS OF ANNOBON ISLAND 105 confirmed by more recent workers in Floras concerning West and South Africa about to be published, and I have now no hesitation in including N. angolensis and N. mannii as well as Lachnopylis annobonensis. The Annobon plant is very clearly the common form of the species which has been known as N. angolensis on the neighbouring mainland. On S. Tomé, however, the species is represented by a plant which may be distinguished varietally. [Var. thomensis (Philipson) J. Lewis, stat. nov. Lachnopylis thomensis Philipson in Exell, Cat.: 245 (1944).—Exell, Addit.: 462. Known from S. Tomé and Mount Kilimanjaro in Tanganyika. Specimens collected by Monod on the Pico in S. Tomé (Monod 11954, 11989) have unusually large elliptic leaves, about 15 cm. long and 6 cm. broad, bearing distinct traces of a white dendroid indumentum along the midrib and nerves beneath. However, they are exactly matched by a specimen collected by Haarer at a compar- able altitude on Mount Kilimanjaro (Herb. Kew) which has always been accepted as N.congesta. The occurrence of N. congesta at similar altitudes on other mountains makes subspecific rank unsuitable for these variants, but their moderate distinctness justifies recognition at the varietal level. | GENTIANACEAE CANSCORA DECUSSATA (Roxb.) J. A. & J. H. Schult., Mant. 3 : 229 (1827). Pladeva decussata Roxb., Hort. Bengal.: 10 (1814), nom. nud. Pladera decussata Roxb., Fl. Ind., ed. Carey, 1 : 418 (1820). ANNOBON: N.W. of the Island, 15 m., pathside, 11 July 1959, Melville 137. New record of the genus Canscora Lam. for the islands. C. decussata is widespread in the tropics of the Old World. EXACUM QUINQUENERVIUM Griseb., Gen. & Sp. Gentian.: 112 (1839). ANNOBON : N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Wrigley 9. New record of the genus Exacum L. for the islands. E. guinquenervium is widely distributed in tropical Africa, Madagascar and the Mascarene Islands. CONVOLVULACEAE IPOMOEA ALBA L., Sp. Pl. 1 : 161 (1753). Calonyction aculeatum (L.) House.—Exell, Cat.: 249. ANNOBON : South of Crater Lake, 450 m., opening in forest, climbing on plants, 1 Aug. 1959, Melville 207. New record for Annobon. Calonyction Choisy is now included in Ipomoea L., and C. aculeatum takes the name I. alba. IPOMOEA INVOLUCRATA Beauv., Fl. Oware & Benin 2: 52, t. 89 (1816). ANNOBON:: North of San Pedro, 420 m., clearings in forest, 21 Aug. 1959, Melville 305. New record for the islands. Widespread in tropical Africa. 106 ANGIOSPERMS OF ANNOBON ISLAND IPOMOEA MAURITIANA Jacq.—Exell, Addit.: 462. Ipomoea digitata sensu Exell, Cat.: 250. ANNOBON : Crater Lake, 265 m., lake edge, on Ficus and Elaeis, climber, 31 July 1959, Melville 206. New record for Annobon. IPpoMOEA NIL (L.) Roth, Catalect. Bot. 1 : 36 (1797). Convolvulus nil L., Sp. Pl., ed. 2, 1 : 219 (1762). ANNOBON : South of the Administrator’s house and just behind the expedition’s houses, 24 m., shady pathside, running over ground, 10 July 1959, Melville 103. Newrecord for theislands. Introduced ; a widespread tropical species of American origin. SOLANACEAE (By W. T. Stearn) LYCOPERSICON ESCULENTUM Mill.—Exell, Cat.: 252. Var. CERASIFORME (Dunal) Alef.—Exell, Cat.: 252. ANNOBON : San Pedro, 3 m., sea shore, cultivated, 11 Aug. 1959, Melville 262. New record for Annobon. Introduced. SOLANUM DASYPHYLLUM Schumach. in K. Danske Vid. Selsk. Naturvid. & Math. Afhandl. 3 : 146 (1828).—Bitter in Fedde, Repert. Sp. Nov., Beih. 16: 188 (1923). Solanum duplosinuatum Klotzsch.—Exell, Cat.: 253. ANNOBON : Between Pico do Fogo and San Pedro, 450 m., banana plantation, 26 July 1959, Wrigley 95. New record for Annobon ; recorded (with doubt) for Principe under the name S. duplosinuatum. Widespread in tropical Africa, extending into Natal. SOLANUM NODIFLORUM Jacq., Collect. 2 : 288 (1788). Solanum nigrum sensu Exell, Cat.: 253 ; Suppl.: 36. ANNOBON : Pico do Fogo, S.W. side, 390 m., bare soil between rocks, recently burnt, 25 July 1959, Melville 188. Recorded for Principe, S. Tomé and Annobon under the name S. nigrum. Wide- spread in the tropics, where it has been much confused with the closely allied S. nigrum L. PHYSALIS ANGULATA L.—Exell, Cat.: 254. ANNOBON: N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Wrigley 124. New record for Annobon. Probably introduced. PHYSALIS MICRANTHA Link, Enum. Pl. Hort. Bot. Berol. 1: 181 (1821).—O. E. Schulz in Urban, Symb. Antill. 6 : 147 (1909). ANGIOSPERMS OF ANNOBON ISLAND 107 Physalis minima sensu C. H. Wright in Dyer, Fl. Trop. Afr. 4 (2) : 247 (1906). ANNOBON : N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Melville 125. Pico do Fogo, 390 m., bare soil, 25 July 1959, Melville 195. New record for the islands. Widespread in tropical Africa and also recorded for India and tropical America. CAPSICUM ANNUUM L.—Exell, Cat.: 254. ANNOBON : Valley between Pico do Fogo and San Pedro, 450 m., banana planta- tion, 26 July 1959, Wrigley 94, 94a. Highest point of Crater, 420 m., cultivated land, 1 Aug. 1959, Melville 221. South of Santa Cruz, cultivated land, 19 Aug. 1959, Melville 314. New record for Annobon. Cultivated. DATURA CANDIDA (Pers.) Safford.—Exell, Suppl.: 36. ANNOBON : Ambo, in front of church, 15 m., garden, 22 Aug. 1959, Melville 312. New record for Annobon. Introduced ; widely spread as a cultivated plant in tropical and subtropical countries. NICOTIANA TABACUM L.—Exell, Cat.: 256. ANNOBON : Ambo, 27 m., waste ground ; “not (obviously) cultivated ; not used’; 19 July 1959, Wrigley 59. New record for Annobon. Introduced. SCROPHULARIACEAE LINDERNIA DIFFUSA (L.) Wettst.—Exell, Cat.: 256. Var. DIFFUSA. ANNOBON:: N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Wrigley IO. New record for Annobon. Var. PEDUNCULATA (Benth.) Skan.—Exell, Cat.: 256 ; Suppl.: 36. ANNOBON : Top of Pico Surcado, 11 July 1959, Wrigley roa. New record for Annobon. ALECTRA SESSILIFLORA (Vahl) Kuntze, Revis. Gen. Pl. 2 : 458 (1891). Gerardia sessiliflora Vahl, Symb. Bot. 3 : 79 (1794). Var. MONTICOLA (Engler) Melch. in Notizbl. Bot. Gart. & Mus. Berl.-Dahl. 15 : 126 (1940).—Hepper in Kew Bull. 14 : 406 (1960). Melasma indicum var. monticola Engler, Bot. Jahrb. 30 : 402 (1901). Alectra communis Hemsl.—Exell, Cat.: 257. ANNOBON : Pico do Fogo, S.W. side, 390 m., burnt area, bare ground, 25 July 1959, Melville 189. New record for Annobon. Recorded for Principe ; widespread in tropical Africa, Madagascar, Mauritius and south-eastern Asia from India to Formosa and the Philippines. 108 ANGIOSPERMS OF ANNOBON ISLAND ACANTHACEAE PHAULOPSIS MICRANTHA (Benth.) C. B. Clarke.—Exell, Cat.: 261. ANNOBON : N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Melville 120. New record for Annobon. DICLIPTERA VERTICILLATA (Forsk.) C. Christens. in Dansk Bot. Ark. 4 (3) : 11 (1922). Dianthera verticillata Forsk., Fl. Aegypt.-arab.: cClll, 9 (1775). Justicia umbellata Vahl, Enum. Pl. 1: 111 (1805). Dicliptera umbellata (Vahl) Juss.—Exell, Cat.: 263 ; Addit.: 463. ANNOBON : Near Ambo and near Crater Lake, near streams, etc., 18 July 1959, Wrigley 49. New record for Annobon. The species occurs in Principe and S. Tomé and is widespread in tropical Africa, extending into Arabia and India. [VERBENACEAE] [AVICENNIA GERMINANS (L.) L., Sp. Pl., ed. 3, 2 : 891 (1764).—Stearn in Kew Bull. 13: 34 (1958).—Exell, Aditam.: 86; Addit.: 464. Avicennia africana Beauv.—Exell, Cat.: 265. The earliest publication of the combination A. germinans was in 1764 as given above. The species is recorded from S. Tomé ; it occurs on the coasts of tropical America as well as West Africa. ] - LABIATAE LEONOTIS NEPETIFOLIA (L.) Ait. f—G. Taylor in Exell, Cat.: 266.—Exell, Suppl.: 38. ANNOBON : Ambo, by the expedition’s houses, 15 m., waste ground and common there ; ‘‘ natives dry and hang up to keep off mosquitoes ’’; 13 July 1959, Melville I4t. New record for Annobon. OcIMUM BASILIcUM L.—G. Taylor in Exell, Cat.: 266. ANNOBON : Ambo village, 15 m., pathside, 15 July 1959, Wrigley 4o. New record for Annobon. NYCTAGINACEAE BoOERHAVIA COCCINEA Mill., Gard. Dict., ed. 8, n. 4 (1768).—Meikle in Hutch. & Dalziel, Fl. W. Trop. Afr., ed. 2,1: 178, fig. 66 (1954). Boerhavia diffusa sensu Exell, Cat.: 268 pro parte. PRINCIPE : Without locality, Sept. 1853, Welwitsch 5391. S. Tomé : Without locality, Don. Rio do Ouro, 240 m., July 1885, Moller 590. Blu-blu, 130 m., 26 Sept. 1912, Watt 7090. Roca Rozema, 13 Oct. 1912, Watt 7350, 7357: ANNOBON : Without locality, 15 Feb. 1933, Exell 901. San Antonio, 60 m., village, between houses, 8 Aug. 1959, Melville 252. ANGIOSPERMS OF ANNOBON ISLAND 109 New records for Principe, S. Tomé and Annobon. Native of tropical America, now widespread in the tropics of both America and Africa. Since the publication of my Catalogue in 1944 Meikle (loc. cit.) has distinguished two species within the West African material referred to B. diffusa L. Re-examin- ation of the specimens in the British Museum Herbarium cited in the Catalogue shows that they all belong to B. coccinea as defined by Meikle. BOERHAVIA DIFFUSA L.—Exell, Cat.: 268 excl. parte. ANNOBON : Lower region, 0-150 m., 15 Feb. 1933, Exell 880. Towards Pico do Fogo, south of Ambo, go m., pathside in open semi-cultivated country, 10 July 1959, Melville 106. Recorded for Annobon by Mildbraed, but the identity of his plant is now uncertain. The record for Principe, and at least some of the records for S. Tomé, belong to B. coccinea as indicated above under that species. AMARANTHACEAE AMARANTHUS SPINOSUS L.—Exell, Cat.: 270. ANNOBON : San Pedro, 3 m., pathside in village, 21 July 1959, Melville 173. New record for Annobon. AMARANTHUS VIRIDIS L.—Exell, Cat.: 270. ANNOBON : N.W. corner of Crater Lake, 265 m., dry stream bed, 24 July 1959, Melville 179. New record for Annobon. ACHYRANTHES ASPERA L.—Exell, Cat.: 270. ANNOBON : Crater Lake, 270 m., forest paths, 25 July 1959, Melville 192. New record for Annobon. CYATHULA PROSTRATA (L.) Blume.—F xell, Cat.: 271. ANNOBON : Path to Pico Surcado (Monte Santa Mina), 510 m., dense vegetation in mist-forest, 21 July 1959, Melville 174. New record for Annobon. ALTERNANTHERA MARITIMA (Mart.) St.-Hil., Voy. Brésil 2: 437 (1833).—Keay in Hutch. & Dalziel, Fl. W. Trop. Afr., ed. 2, 1 : 154 (1954). Bucholzia maritima Mart., Nov. Gen. & Sp. Pl. 2: 50, t. 147 (1826). Telanthera maritima (Mart.) Moq.—Exell, Cat.: 271 ; Aditam.: 86. ANNOBON : Near Bird Island, N.W. coast, 6 m., sea cliff with wind-blown spray, 17 Aug. 1959, Wrigley 284. Recorded for Principe, S. Tomé and Annobon. Keay and other recent authors include Telanthera R. Br. under Alternanthera Forsk. ALTERNANTHERA SESSILIS (L.) R. Br. ex DC., Cat. Pl. Hort. Bot. Monspel.: 4, 77 (1813).—Exell, Cat.: 272 ; Suppl.: 30. 110 ANGIOSPERMS OF ANNOBON ISLAND ANNOBON : Dry stream east of Ambo, 30 m., between rocks, 1 Aug. 1959, Wrigley 205. Near Bird Island, N.W. coast, 9 m., sea cliffs, sometimes in wind-blown spray, 17 Aug. 1959, Melville 283. The earliest publication of the combination A. sessilis appears to be that given above. The species is recorded for Principe, S. Tomé and Annobon. BASELLACEAE BASELLA ALBA L.—Fxell, Cat.: 274. ANNOBON : San Pedro and Ambo, 15 m., roof of houses, vine rooted in the ground ; “ used as pot-herb’”’; 21 July 1959, Wrigley 67. New record for Annobon. Probably introduced. PIPERACEAE (By W. T. Stearn) PEPEROMIA HYGROPHILA Engler.—Exell, Cat.: 277. Peperomia annobonensis Mildbr.—Exell, Cat.: 276. ANNOBON: Pico Surcado, 510 m., mist-forest, very abundant epiphyte, 21 July 1959, Wrigley 7o. P. hygrophila was based by Engler on specimens collected at 2,500-2,600 m. altitude on Cameroons Mountain, evidently in a moist habitat as the epithet implies ; an isosyntype (Preuss 805) in the British Museum Herbarium represents a robust luxuriant individual with cuneate-based leaf-laminas 1-5-4 cm. broad and a flowering spike about 8 cm. long, these somewhat exceeding the dimensions given by Engler in his original description. Exell 156 from virgin forest on S. Tomé approaches Preuss’s plant in luxuriance, having laminas I-35 cm. broad, some cuneate, others almost rounded at the base, but with spikes 2-5-4:5 cm. long. Other specimens from S. Tomé bridge the gap between this and a small-leaved plant from the same island (Monod 11744) with laminas rounded or almost subcordate at the base and 0:5-1°4 cm. broad. In Monod 12228 the lamina varies from almost circular with rounded base to narrowly obovate with narrowly cuneate base. The spike in S. Tomé material varies in length from about 8 cm. to 2cm. I have examined the holotype of P. annobonensis (Mildbraed 6532) which is preserved at Berlin and which was collected on Annobon in dry Steganthus-Elaeis forest at 250-350 m. altitude. It differs from Wrigley’s mist-forest material in having mostly elliptic rather than mostly circular laminas, but both these Annobon plants come well within the range of variation of P. hygrophila as manifested on S. Tomé and Cameroons Mountain. LAURACEAE PERSEA AMERICANA Mill.—Exell, Cat.: 280. ANNOBON : S.W. of Crater Lake, 390 m., open forest, 5 Aug. 1959, Melville 239. New record for Annobon. Introduced. Avocado. LORANTHACEAE (By W. T. Stearn) VISCUM ENGLERI Tiegh. in Bull. Soc. Bot. France 43: 190 (1896).—Sprague in Dyer, Fl. Trop. Afr. 6 (1) : 405 (1911). ANGIOSPERMS OF ANNOBON ISLAND Il ANNOBON : Edge of forest near Pico do Fogo, 330 m., parasite on Fucus clarencensis, 24 July 1959, Wrigley 86. New record for the islands. Also in Angola and Tanganyika. The specimen, which has only unripe fruit, belongs to the small group of African species of Viscum with conspicuously flattened stems and the leaves reduced to scales. In the key to the tropical African species given by Sprague (tom. cit.: 394-395) it runs down to V. engleri, hitherto recorded only from Tanganyika ; and it comes within the range of variation of this species as indicated by specimens from the Usambara Mountains (the locus classicus) and the Sagara Mountains. Specimens of what is evidently the same species collected by Gossweiler (n. 9884, Herb. Kew) in the Cuanza Sul region of Angola suggest that V. englert may be widespread but sporadic in tropical Africa. EUPHORBIACEAE EUPHORBIA PULCHERRIMA Willd. ex Klotzsch in Allgem. Gartenzeit. 2 : 27 (1834). ANNOBON : Santa Cruz, 30 m. (also Ambo), planted by chapel, 8 Aug. 1959, Melville 269. New record for the islands. Introduced ; native of Mexico and Central America. Poinsettia. [CLEISTANTHUS LIBERICUS N. E. Br.—Exell, Cat.: 286. In the Catalogue I listed this species from S. Tomé, but with a query. Dr. J. Léonard considers (and I agree) that the S. Tomé plant was correctly identified as C. libericus by Hutchinson. ] THECACORIS ANNOBONAE Pax & Hoffm.—Exell, Cat.: 287. (Plate ro.) ANNOBON : Pico Surcado, 420 m., forest ; “‘shiny leaves ; very red wood ; racemes of flowers with 5 stamens”; 21 July 1959, Wrigley 66. Near highest point of Crater, 480 m., forest, 26 July 1959, Wrigley 92. Endemic to Annobon. This identification is probably correct but there is a slight element of doubt as the species was described from female inflorescences and fruit while Wrigley collected male inflorescences and fruit, so that his material can be compared only partly with the original description. JT. annobonae was described as having “ capsulae cocci pilosi’’; the fruits of Wrigley g2 are nearly glabrous but have vestiges of anindumentum. Although there is a suggestion that there may be another species of Thecacorts on the island (Exell, Cat.: 288) Wrigley’s specimens nevertheless in all probability represent the species (T. annobonae) collected by Mildbraed on the rim of the North Crater at c. 500 m. altitude and said to be very common. PHYLLANTHUS RETICULATUS Poir. in Encycl. Méth., Bot. 5 : 298 (1804).—G. L. Webster in Journ. Arnold Arb. 38 : 57, fig. 7 (1957). ANNOBON : East of Quioveo (Pico del Centro), edge of clearing in forest, straggler, 1 Aug. 1959, Melville 212. New record for the islands. Widespread in the Old World tropics ; introduced into the West Indies. 112 ANGIOSPERMS OF ANNOBON ISLAND JATROPHA MULTIFIDA L.—Exell, Cat.: 293 ; Suppl.: 42. ANNOBON : Ambo, near the Post Office, 12 m., pathside, cultivated patch, 16 July 1959, Melville 156. New record for Annobon. Introduced. CODIAEUM VARIEGATUM (L.) Blume.—Exell, Cat.: 296. ANNOBON : Due south of Crater Lake, 420 m., forest near path to Quioveo, 1 Aug. 1959, Melville 208. New record for Annobon. Introduced. Discoclaoxylon pubescens (Pax & Hoffm.) Exell, stat. nov. (Plate 11.) Discoclaoxylon occidentale var. pubescens Pax & Hoffm. in Engler, Bot. Jahrb. 58, Beibl. 130 : 39 (1923).—Exell, Cat.: 299. ANNOBON : Highest point of Crater, 480 m., forest ; “‘ large leaves, 2/5 phyllo- taxis ; catkin-like inflorescence ; 4-merous green flowers ’’; 26 July 1959, Wrigley 96. Endemic to Annobon. There is little doubt that Wrigley’s plant is identical with D. occidentale var. pubescens, described from rather insufficient female material consisting of three syntypes : Muldbraed 6492, 6555 and 6751 (Herb. Berlin, des- troyed). The male flowers in Wrigley’s material are considerably larger than those of D. occidentale (Muell. Arg.) Pax & Hoffm., measuring about 8 mm. in diameter when expanded. There are about five male flowers to each bract. They seem to expand one at a time and the pedicel is then 5-8 mm. long. There are 10-12 stamens. The leaves are similar to those of D. occidentale but are more hairy on the under surface, the petiole and the base of the midrib being tomentellous. I consider that these differences justify separate specific rank for the Annobon plant, and that D. occidentale is confined to S. Tomé. Discoclaoxylon (originally Claoxylon sect. Discoclaoxylon Muell. Arg.) consists of a group of West African species geographically separated from Claoxylon A. Juss. and sufficiently distinct to warrant the generic rank given them by Pax and Hoffmann (in Engler, Pflanzenr. 4 (147, 7) : 137 (1914)). Hutchinson and Dalziel (Fl. W. Trop. Afr. 1 : 301 (1928)) reunited Discoclaoxylon with Claoxylon and this treatment was followed by Keay (in Hutch. & Dalziel, op. cit., ed. 2, 1: 401 (1958)). The main differences between the two genera, according to Pax and Hoffmann (tom. cit.: 76, 100, 137), are : Claoxylon : extrastaminal disk none, but small glands usually present between the stamens ; stamens 10-200 or more (very rarely fewer than 15). Discoclaoxylon : extrastaminal disk urceolar, entire or lobulate ; stamens 6-12. Claoxylon does not occur on the continent of Africa but is distributed from Madagascar and the Mascarene Islands eastwards through southern Asia to Australia and Polynesia. The genus Discoclaoxylon now comprises four species : D. pedicellare (Muell. Arg.) Pax & Hoffm. (Fernando Po), D. occidentale (Muell. Arg.) Pax & Hoffm. (S: Tomé), D. pubescens (Pax & Hoffm.) Exell (Annobon) and D. hexandrum (Muell. Arg.) Pax ANGIOSPERMS OF ANNOBON ISLAND 113 & Hoffm. (Fernando Po, Cameroons and Congo). The last-named species, which has (3-) 6-8 stamens (the others have 10-12) is the only one so far found on the mainland of Africa. When male and female flowers and fruits of all four are known their relationship may become clear, but it is already apparent that the distribution is ‘“‘normal’’ with one West African species reaching Fernando Po, one endemic on that island, one on S. Tomé and a fourth on Annobon. The genus has not been recorded for Principe. It is one of the few “ island genera ’”’ with more species on the islands than on the mainland (cf. Calvoa in the Melastomataceae which has about half its known species on the islands). It seems desirable to typify the generic name Discoclaoxylon. Claoxylon sect. Discoclaoxylon Muell. Arg. (Flora 47 : 437 (1864)) was founded on three species, C. pedicellare Muell. Arg., C. occidentale Muell. Arg. and C. hexandrum Muell. Arg., without indication of type ; all three species were known only from male material. The position was unchanged when the section was later reviewed by its author (in DC., Prodr. 15 (2) : 779 (1866)). When Pax and Hoffmann raised the section to generic rank as Discoclaoxylon the third species (D. hexandrum) had become known from the mainland of West Africa and female material had become available, so that quite a large part of their generic description (relating to ovary, capsule and seeds) came from this species, female material being still unknown in the other two species. It is also quite possible that D. pedicellare and D. occidentale are now extinct and complete material may never become available. Taking these points into consideration it seems best to select D. hexandrum as type of Discoclaoxylon. ACALYPHA ANNOBONAE Pax & Hoffm.—Exell, Cat.: 299. (Plate 12.) ANNOBON : Ridge west of Crater Lake, 360 m., open woodland ; “‘ female and male flowers on same shrub ; low shrub 1-2 m.”’; 18 July 1959, Wrigley 55. Endemic to Annobon. Wrigley’s plant is practically a topotype. The holotype, Miuldbraed 6538 (Herb. Berlin), has been destroyed. RICINUS COMMUNIS L.—Exell, Cat.: 301. ANNOBON : Ambo village, 15 m., pathside, 15 July 1959, Wrigley 4r. New record for Annobon. Probably introduced. MORACEAE ARTOCARPUS COMMUNIS J. R. & G. Forst.—G. Taylor in Exell, Cat.: 304. ANNOBON: Near Ambo, 6 m., waste ground near sea, not seen to regenerate, 18 Aug. 1959, Wrigley 292. New record for Annobon. Introduced. Breadfruit. ARTOCARPUS HETEROPHYLLUS Lam. in Encycl. Méth., Bot. 3 : 209 (1789).—Jarrett in Journ. Arnold Arb. 40 : 334 (1959). Artocarpus integer sensu G. Taylor in Exell, Cat.: 304. ANNOBON:: Near highest point of Crater, 480 m. (also one sterile specimen on shore of Lake), forest, edge of clearing, 26 July 1959, Wrigley gr. 114 ANGIOSPERMS OF ANNOBON ISLAND New record for Annobon ; recorded for Principe and S. Tomé under the name A. integer. Introduced. Jack Fruit. The name A. integer (Thunb.) Merr. applies to a different species (cf. Jarrett, tom. cit.: 329). URTICACEAE URERA CAMEROONENSIS Wedd.—G. Taylor in Exell, Cat.: 312. ANNOBON : Common all over the Island in forest, forest scrambler, often forming lianes, 31 July 1959, Wrigley 223. New record for Annobon. ORCHIDACEAE (By V. S. Summerhayes) EPIPOGIUM ROSEUM (D. Don) Lindl. in Journ. Proc. Linn. Soc. Lond., Bot. I: 177 (1857). Limodorum roseum D. Don, Prodr. Fl. Nepal.: 30 (1825). ANNOBON : South of Crater Lake and east of Quioveo (Pico del Centro), 450 m., humus below trees with locally sparse fern flora on ground, 1 Aug. 1959, Melville 213 (Herb. Kew). New record of the genus Epipogium R. Br. for the islands. EE. roseum is widely but sporadically distributed in tropical Africa (Ghana, Cameroons, Fernando Po, Congo, Angola, Uganda), Indo-Malaysia, Australia and New Hebrides. PLATYLEPIS GLANDULOSA (Lindl.) Reichb. f—Dandy in Exell, Cat.: 319.—Exell, Suppl 47, ANNOBON : South of Santa Cruz (also Crater Lake, N.E. side), under forest, 24 July 1959, Melville 184 (Herb. Kew). New record for Annobon. The species extends from tropical Africa into Natal. CORYMBORKIS CORYMBOSA Thou., Hist. Pl. Orch., Tab. Esp. 1 (1822).—Summerh. in Kew Bull. 11 : 224 (1956). Corymborkis welwitschii (Reichb. f.) Kuntze.—Dandy in Exell, Cat.: 320. ANNOBON : Between Crater Lake and Monte Abecin, 450 m., forest, 19 Aug. 1959, Melville 299 (Herb. Kew). New record for Annobon. The species is recorded for S. Tomé and is generally distributed in tropical Africa, extending into Natal and eastern Cape Province ; it also occurs in Madagascar and Réunion. LIPARIS WELWITSCHII Reichb. f. in Flora 48 : 184 (1865). ANNOBON : Pico Surcado, 600 m., beneath canopy of ferns in forest, 11 Aug. 1959, Melville 266 (Herb. Kew). New record for the islands. Also in Angola. ANGIOSPERMS OF ANNOBON ISLAND 115 POLYSTACHYA RIDLEYI Rolfe.—Dandy in Exell, Cat.: 323. ANNOBON : South of Crater Lake, 510 m., epiphytic on Coffea trees in particular, 14 July 1959, Melville 154 (Herb. Kew). New record for Annobon. Hitherto believed to be endemic to S. Tomé. BULBOPHYLLUM MELANORRHACHIS (Reichb. f.) Reichb. f. ex De Wild., Pl. Bequaert. E 3) 93° (1921): Megaclinium melanorrhachis Reichb. f. in Gard. Chron., Ser. 2, 4 : 162 (1875). Bulbophyllum melanorrhachis Reichb. f., loc. cit. (1875), nom. syn. ANNOBON : Ridge west of Crater Lake, 510 m., epiphyte, 14 July 1959, Wrigley 31 (Herb. Kew). New record for the islands. Also on the mainland of West Africa from Guinea to Nigeria, Gabon and Congo. BOLUSIELLA TALBOTII (Rendle) Summerh. in Hutch. & Dalziel, Fl. W. Trop. Afr. 2 : 456 (1936). Angraecum talbotit Rendle, Cat. Pl. Talbot : 108, t. 15 figs. 6, 7 (1913). ANNOBON : Ridge west of Crater Lake, 510 m., epiphyte, 14 July 1959, Wrigley 33 (Herb. Kew). New record of the genus Bolusiella Schlechter for the islands. B. talbotii occurs on the mainland of West Africa from Sierra Leone to Nigeria. AMARYLLIDACEAE HYMENOCALLIs LITTORALIS (Jacq.) Salisb. in Trans. Hort. Soc. Lond. 1 : 338 (1812). Pancratium littorale Jacq., Select. Stirp. Amer. Hist.: 99, t. 179 fig. 94 (1763). ANNOBON : Mission, Ambo, 15 m., garden, 4 Aug. 1959, Melville 234. New record of the genus Hymenocallis Salisb. for the islands. H. littoralis is introduced into West Africa from tropical America. DIOSCOREACEAE DIOscOREA ALATA L.—Dandy in Exell, Cat.: 344.—Exell, Suppl.: 49. ANNOBON : Ambo, path to San Pedro, 24 m., cassava plantation, 22 Aug. 1959, Melville 308. New record for Annobon. Introduced. DIOSCOREA CAYENENSIS Lam.—Dandy in Exell, Cat.: 345.—Exell, Suppl.: 4g. ANNOBON : Ambo, south of the Governor’s house, near footpath, 10 July 1959, Wrigley 2. South of Santa Cruz, 420 m., openings for cultivation, common, scrambling, 7 Aug. 1959, Wrigley 247. New record for Annobon. DIoscOREA DUMETORUM (Kunth) Pax.—Dandy in Exell, Cat.: 345.—Exell, Suppl.: 49. 116 ANGIOSPERMS OF ANNOBON ISLAND ANNOBON : South of Ambo, towards Pico do Fogo, 135 m., climber, 10 July 1959, Melville rrr. North of Santa Cruz, on coast path, 45 m., opening in forest, climber, 8 Aug. 1959, Wrigley 253. New record for Annobon. LILIACEAE CORDYLINE FRUTICOSA (L.) A. Chev., Jard. Bot. Saigon : 66 (1919). Convallaria fruticosa L., Herb. Amboin.: 16 (1754). ANNOBON : Santa Cruz and near Quioveo, 420 m., forest near path, 17 July 1959, Wrigley 48. New record of the genus Cordyline Commers. ex Juss. for the islands. C. fruticosa is a native of south-east Asia, introduced into tropical Africa and America. SANSEVIERIA METALLICA Géréme & Labroy in Bull. Mus. Hist. Nat. Par. 9: 170, 173, fig. 2 (1903). ANNOBON : Ambo, near the Mission, 15 m., cultivated ground, 22 Aug. 1959, Wrigley 309. New record for Annobon. Native of tropical Africa ; introduced into other parts of the tropics. Species of Sansevieria (unnamed) have been introduced into S. Tomé (cf. Dandy in Exell, Cat.: 347) but none is reported from Principe. COMMELINACEAE [STANFIELDIELLA IMPERFORATA (C. B. Clarke) Brenan in Kew Bull. 14 : 284 (1960). Buforrestia imperforata C. B. Clarke.—Dandy in Exell, Cat.: 347. The genus Stanfieldiella Brenan has recently been separated from Buforrestia C. B. Clarke (cf. Brenan, tom. cit.: 280). S.imperforata var. imperforata is recorded from Principe and S. Tomé.] COMMELINA BENGHALENSIS L., Sp. Pl. 1: 41 (1753). ANNOBON : San Pedro, 15 m., streamside, 2 Aug. 1959, Wrigley 228. New record for the islands. Widely distributed in tropical and subtropical regions of the Old World. ARACEAE XANTHOSOMA SAGITTIFOLIUM (L.) Schott in Schott & Endl., Melet. Bot.: 19 (1832). Arum sagittifolium L., Sp. Pl. 2 : 966 (1753). ANNOBON : Above Crater Lake, 360-480 m., open woodland, probably planted, 16 July 1959, Wrigley 44. New record of the American genus Xanthosoma Schott for Annobon. X. sagittifolium is introduced into tropical Africa from tropical America where it is cultivated for its edible tubers. ANGIOSPERMS OF ANNOBON ISLAND L177 CYPERACEAE [HyPOLYTRUM. GRANDE (Uitt.) Koyama in Journ. Fac. Sci. Univ. Tokyo, Sect. 3, 8 : 68 (1961). Principina grandis Uitt.—Dandy in Exell, Cat.: 355. Koyama (loc. cit.); in his recent new arrangement of the family Cyperaceae, has included Principina Uitt. in Hypolytrum Rich. Principina was previously regarded as a monotypic genus endemic to Principe. | CYPERUS PEDUNCULATUS (R. Br.) Kern in Act. Bot. Neer] 7 : 798, fig. 4 (1958). Remirea maritima Aubl.; non Cyperus maritimus Poir.—Dandy in Exell, Cat.: 356.— Exell, Aditam.: 87. Remirea pedunculata R. Br., Prodr. Fl Nov. Holl.: 236 (1810), Recorded for Principe and Annobon, but not collected by the Cambridge Expedition. Remirea Aubl. is included in the large genus Cyperus L. by Kern (loc. cit.) and by Koyama (Journ. Fac. Sci. Univ. Tokyo, Sect. 3, 8 : 72 (1961)). GRAMINEAE (By W. D. Clayton) BAMBUSA VULGARIS Schrad. ex Wendl.—Dandy in Exell, Cat.: 365. ANNOBON : Crater Lake, south side, 265 m., lakeside vegetation, near clearing in forest, 12 Aug. 1959, Melville 276. New record for Annobon. Introduced. PANICUM BREVIFOLIUM L.—Dandy in Exell, Cat.: 370. ANNOBON : Pico do Fogo, S.W. side, 390 m., bare soil on rocky hillside, 25 July 1959, Melville 187. New record for Annobon. OPLISMENUS HIRTELLUS (L.) Beauv.—Dandy in Exell, Cat.: 373.—Exell, Suppl.: 51. ANNOBON : Path to Quioveo (Pico del Centro), just above north end of Crater Lake, 300 m., pathside in wood, 18 July 1959, Melville 163. New record for Annobon. PASPALUM COMMERSONII Lam.—Dandy in Exell, Cat.: 375. ANNOBON: N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Melville 122. New record for Annobon. PASPALUM CONJUGATUM Berg.—Dandy in Exell, Cat.: 375. ANNOBON: Near Ambo, 30 m., streamside, 12 July 1959, Wrigley 28. New record for Annobon. ANTHEPHORA CRISTATA (Doell) Hack. ex De Wild. & Dur.—Exell, Suppl.: 52. ANNOBON: N.E. coast near Ambo, 24 m., pathside in cassava plantation, 21 July 1959, Melville 171. New record for Annobon, 118 ANGIOSPERMS OF ANNOBON ISLAND Beckeropsis laxior W. D. Clayton, sp. nov.; affinis B. nubicae (Hochst.) Fig. & De Not. sed differt ramulis laxioribus subglabris, spiculis acuminatis, glumis superioribus longioribus, lemmatibus spinuloso-hipsidulis. (Holotype in Herb. Kew from Nigeria : Summit of Orosun, Idanre Hills, Keay FHI 22678.) Beckeropsis nubica sensu Dandy in Exell, Cat.: 378. : ANNOBON : S.W. side of Pico do Fogo, 390 m., bare soil after burning, 25 July 1959, Melville gr. New record for Annobon. Also in S. Tomé and on the mainland of West Africa from Ghana to Cameroons. PENNISETUM POLYSTACHION (L.) Schult——Dandy in Exell, Cat.: 378.—Exell, Aditam.: 89. ANNOBON : Ambo, by the expedition’s houses, 15 m., waste ground, common, 13 July 1959, Melville 139. New record for Annobon. SCHIZACHYRIUM BREVIFOLIUM (Swartz) Nees ex Buse in Pl. Junghuhn.: 359 (1854). Andropogon brevifolius Swartz, Nov. Gen. & Sp. Pl.: 26 (1788). ANNOBON:: N.W. of the Island, 15 m., cassava plantation, 11 July 1959, Melville 128. New record of the genus Schizachyrium Nees for the islands. S. brevifolium is widespread in the tropics. CYMBOPOGON CITRATUS (DC.) Stapf.—Dandy in Exell, Cat.: 380. ANNOBON : Ambo, just north of the expedition’s houses, fide Melville (no specimen collected). New record for Annobon. Introduced. CoIx LACRYMA-JOBI L.—Dandy in Exell, Cat.: 381. ANNOBON : Near Ambo, 24 m., streamside, 12 July 1959, Wrigley 21. New record for Annobon. Introduced. Bull. B.M. (N.H.) Bot. 3, 3 PLATE 3 Mecklenburg s 4040 nee GID Turraea glomerulifilora Harms (Mildbraed 6487, holotype) Bull. B.M. (N.H.) Bot. 3, 3 "ay Pease aunrlonenagd (Lets % rita. } Leelee, per. CAL E44 19.6 2. CAMBRIDGE Qa EXPEDITION TO ANNOBON ISLAND WEST TROPICAL AFRICA T. C. Watctey ano F. A. MELVILLE, 1959 ANNORGs Sidee west of Croter Lake, Open woodland, L200ft. Alte Flover white; Seaerous frudt ith 1 orange Maytenus annobonensis (Loes, & Mildbr.) Exell (Wrigley 56, topotype) ” 1, | rex Ey 1 ‘| PLATE 4 Bull. B.M. (N.H.) Bot. 3, 3 ae | (14 4 PLATE 5 CAMBRIDGE SamBMN EXPEDITION TO ANNOBON ISLAND WEST TROPICAL AFRICA T. (. Wreictey anp FL AL Menyrire, 1959 i } J (igelaca ovat; Sek elbus- : Der. 2orSacw 1962 AANOBON: Aabo, Dey val SOOM. Alt, Tree t 30ft. Fruit red and yellow With back onl oruye seed, Flowers off—wiite (:inkieh), F, ¥elvil 232, 2 Auge 195996 Agelaea ovalis Schellenb. (Melville 232, topotype) Bull. B.M. (N.H.) Bot. 3, 3 PLA LES Arb raat ; j i antl thee noi Friedrich Herzog zu Mecklenburg Zweite Expedition nach Central-Afrika 1910-1911 .. Coots prunes. dene bOntetts OAL de. He gee pect, ae owes. Aeh ee a ke o1 Annobon ; Be é Dtttoad han 2 lon- byrrnens a. /Fibetarelef Cu PbO a Cassipourea annobonensis Mildbr. ex Alston (Mildbraed 6511, holotype) Bull. B.M. (N.H.) Bot. 3, 3 PLATE 7 Tice CAMBRIDGE @0N@@088) EXPEDITION TO ANNOBON ISLAND WEST TROPICAL AFRICA T. C. Wrratey ann F. A. Menviiie, 1959 ANNOBON: Pieo Surgado, Forest ULS0ft, Alt. Flower pale pink, T.c. Trighley 65, 2k July 1959, Tristemma oreothamnos Mildbr. (Wrigley 65, topotype) Bull. B.M. (N.H.) Bot. 3, 3 PLATE 8 WERE ee ANNGBON: Nr. highest polst of crater. CAMBRIDGE SRBRGSRY EXPEDITION TO ANNOBON ISLAND Forest. 1600ft, Alt. WEST TROPICAL AFRICA Undershrub. 2 speciuens with white corollas; A others fruit oly. © Gh . A. Mr Ey 195 igl T. C. Warcrey anno F. A. Mrivinre, 1959 T.C. te 93. 26 July, 1959. Bertiera annobonensis G. Taylor (Wrigley 93, topotype) Bull. B.M. (N.H.) Bot. 3, 3 PLATE 9 Po Se * rare: te bntn hilt 2 ay ee te . Lda ye Souter iets en POST: ae Pare pte Tr tacth yy on Dr. J. Mildbraed. No? O% ; oo al eo: Nuxia congesta R. Br. ex Fresen. var. congesta (Mildbraed 6561, holotype of Lachnopylis annobonensis Mildbr.) Bull. B.M. (N.H.) Bot. 3, 3 PLATE 10 CAMBRIDGE QR EXPEDITION TO ANNOBON ISLAND WEST TROPICAL AFRICA T. C. Wrisctey ann F. A. Mervinne, 1959 Thecacoris annobonae Pax & Hoffm. (Wrigley 92, topotype) Bull. B.M. (N.H.) Bot. 3, 3 PLATE 11 ‘Diseeclaswylorn pubescens Pars 5 , tof, hee Der, Oe Seek i Fee AR BON, hichest poet of crater. CAMBRIDGE S2a¥BRS@8% EXPEDITION orest. ieoort, Alt. TO ANNOBON ISLAND Lurye leaves 2/ phyllotetigs, Catkin-like WEST TROPICAL AFRICA infior-scses, heacreus grees fruits, 7. trisley 96, a6 Taly 1959, T. C. Wrigiey ann F. A. Mrtvitee, 1959 Discoclaoxylon pubescens (Pax & Hoffm.) Exell (Wrigley 96, topotype) Bull. B.M. (N.H.) Bot. 3, 3 PLATE 12 ie Be a 1 A. OBON: Riige west of Crater Lake, Open woodland, 1200ft, alt, Q & ¢ flowers on same shrub. Low shrub T.o, “rigley 55, 18 July 1959, CAMBRIDGE SRRSBIR EXPEDITION TO ANNOBON: ISLAND WEST TROPICAL AFRICA T. C. Wriatey ano F. A. Metyinre, 1959 teal ha auvstnat fan vt Mifhin. per, Qto-fnta 19 6% Acalypha annobonae Pax & Hofim. (Wrigley 55, topotype) A REVISION OF THE GENUS _ PETRORHAGIA ( > & , KG: W-HAL) = July 21808; Ross 112 (BM; E; FI; K). Piano dei Favari, 30 June 1855, Huet du Pavillon (BM ; CGE; FI; K). Sorrentino, 1854, anon. (FI). Gragia, 1871, Lojacano (FI). Aetne, 1840, Tornabene (FI). Madonie, Todaro (EK). Monte S. Salvatore : Pizzo di Palermo et Pizzo Antenna, 15 and 22 July 1873, Strobl (K). Caltagirone, 4 June 1843, Herb. Bunbury (CGE). Prope Troina, 2 July 1855, Huet du Pavillon (BM ; CGE ; FI; K). Monte Scalone supra Polizzi, 22 July 1874, Strobl (K). Siracusa, 1841, Cassia (FI); Nov. 1929, Corrie & Chase H2287 (K). _Presso Augusto (Siracusa), Oct. 1911, Vaccari (FI). Caltanissetta, 14 June 1899, July rgor, Giovanni 308 (FI). This subspecies is difficult to separate from subsp. zlyrica. The populations in northern Greece and Macedonia cannot be satisfactorily placed under either subspecies, although here they are included in subsp. haynaldiana on the basis of the larger flowers and the stem often glabrous at the base. The populations in Italy and Sicily are similar and are also included here on the generally more robust habit, larger flowers, longer pedicels and the lower part of the stem sometimes glabrous. Tunica cretica, T. ochroleuca, T. compressa and T. armerioides, etc. of Italian authors are only minor variants of this subspecies. 1c. Petrorhagia illyrica subsp. angustifolia (Poir.) Ball & Heywood, comb. nov. Silene angustifolia Poir., Voy. Barb. 2 : 164 (1789). Gypsophila compressa Desf., Fl. Atlant. 1 : 343 (1798). Tunica compressa (Desf.) Fisch. & Meyer, Index Sem. Hort. Bot. Petrop. 4 : 50 (1837). Dianthella compressa (Desf.) Pomel, Mat. Fl. Atlant. : 9 (1860). Tunica davaeana Coss. in Bull. Soc. Bot. France 36 : 103 (1889).—Maire, Fl. Afr. Nord to : 284 (1963). Tunica angustifolia (Poir.) Briq., Prodr. Fl. Corse 1 : 544 (1910). Tunica scoparia Pampan. in Arch. Bot. Forli 12 : 25 (1936).—Maire, tom. cit. : 285 (1963). Tunica illyrica subsp. angustifolia (Poir.) Maire in Bull. Soc. Hist. Nat. Afr. Nord 30 : 265 (1939) ; Fl. Afr. Nord 10 : 283 (1963). Indumentum variable, stems usually densely pubescent throughout or glabrous in the lower part. Pedicels strongly compressed (usually more or less terete in all REVISION OF THE GENUS PETRORHAGIA 137 other subspecies). Calyx 5:5-8 mm. long ; costae I—veined, rarely 3-veined. Petals 8-11 mm. long. Morocco : Settat ad Guicer, 31 May 1913, Pitard 2699 (K). Settat, Bir Chafoi, 25 May 1913, Pitard 2697 (BM). Atlas med., Ras-el-Ma prope Azrou, 24 May 1926, Lindberg 4329 (W). Azrou, 30 June 1923, Jahandiez 628 (BM) ; 23 June 1926, Lindberg 7072 (K). High Atlas, Tadderte, 28 July 1955, Newbould 48 (BM). Anamer, Oed Arboa, 16 Aug. 1955, Newbould 136 (BM). Marrakesh-Ouarzazet road, 4 May 1937, Masters 46 (BM). Djebel Afougueur, 9 June 1875, Soc. Dauphinoise 1878, 1113b (K). Province Shedma, May 1871, J. Ball (BM). Ida Oubahil, 1875, Cosson (K). Shedma, Apr.-May 1871, Hooker (K). Siksona, May 1871, Hooker (K). Oujda, 21 May 1930, Faure (K). Sidi bou Ziane, 12 May 1913, Pitard 2698 (K). SPANISH Morocco : Atlas Rifain, valles du Ketama, 21 June 1934, Sennen 9293 (BM). ALGERIA : Constantine, 18 Aug. 1855, Choulette 22 (CGE; K). Mantourah, June 1869, E. G. Paris 308 (K). Oran, May 1889, Luzzet (BM). Mostaganem, 15 June 1853, Balansa 138 (K). Ain-Tindamine (Oran), 7 June 1921, Faure (BM). Plaine de Batna, June 1853, Balansa (K). Crezel vers le Nador de Ciuret, 27 June 1937, Faure (E). Ain-el-Ibil, Losobria (K). Tunisia : Djebel Zaghouan, 9 July 1854, Kralik (K). LisyA : Garian, 16 Nov. 1948, Johnson 54 (BM); 18 May 1952, Guichard KG/LIB/403 (BM) ; 24 May 1939, Sandwith 2758 (K). Suani Bin Adem, 10 Nov. 1952, Guichard KG/LIB/5 (BM). Benghazi, Grotta di Lete, 1 Apr. 1939, Sandwith e227 (K). A variable subspecies with a range of variation parallel to that found in subspecies allyrica and haynaldiana and in P. armerioides. It can be separated only with some difficulty from subsp. haynaldiana as the reliability of the pedicel character is uncertain. 1d. Petrorhagia illyrica subsp. taygetea (Boiss.) Ball & Heywood, stat. nov. (Fig. 3.) Tunica illyrica var. taygetea Boiss., Fl. Or. 1 : 521 (1867). Tunica taygetea (Boiss.) P. H. Davis in Not. R. Bot. Gard. Edin. 22: 165 (1957), excl. syn. T. cretica et Gypsophila cretica. Stem more or less densely glandular-pubescent throughout. Calyx 3-5-5 mm. long ; costae more or less 3-veined, the lateral veins often very weak. Petals 5°5-7°5 mm. long. GREECE: Peloponnisos: Androuvista, Montis Taygeti, June-July 1844, Heldreich (BM ; CGE; K>; isotypes) ; July 1870, Orphanides 593 (W-HAL). Megala Zonaria, 9 July 1899, Heldreich 1519 (E ; K ; W-HAL). Likoneri, 9 July 1902, Haldcsy 39 (W-HAL). This taxon has had a somewhat disturbed history. It was first distributed by Heldreich as Tunica cretica but Boissier described it as a new variety of T. alyrica. 138 REVISION OF THE GENUS PETRORHAGIA Halacsy, Conspectus Florae Graecae, again considered this plant to be identical with T. cretica, and appears to have been followed by later authors. Recently Davis showed that the name 7°. cretica had been misapplied by all authors since Fischer & Meyer, and that the Cretan plant required a new name. Unfortunately, Davis followed the majority of authors in regarding Taygetean plants as being identical with Cretan plants. We cannot agree with Heldreich, Halacsy or later authors that these two popula- tions are identical or even that they should be referred to the same species. If the venation of the calyx is ignored, the Cretan plant still differs from P. zllyrica and the Taygetean populations by a number of characters and appears to be one of the more distinct species in this critical group (for further discussion see under P. candica, p. 141). Here the Taygetean population is regarded as a subspecies of P. illyrica, showing some affinity with P. candica (cf. the smaller flowers, the feebly 3-veined calyx costae, and the longer pedicels) but clearly distinct from this latter species. Subsp. taygetea forms a fairly distinct unit within P. illyrica. It is most readily recognized by the venation of the calyx costae but similar venation occurs in plants of typical P. iallyrica in Attiki. The other characters are statistical rather than absolute differences, so these Taygetean plants are considered to represent a subspecies rather than a separate species. 2. Petrorhagia ochroleuca (Smith) Ball & Heywood, comb. nov. (Fig. 4.) Gypsophila ochroleuca Smith, Fl. Graec. Prodr. 1 : 281 (1809). Tunica ochroleuca (Smith) Fisch. & Meyer, Index Sem. Hort. Bot. Petrop. 4 : 49 (1837).— Boiss., Fl. Or. 1 : 520 (1867).—Hayek, Prodr. Fl. Penins. Balcan. 1 : 222 (1924). Stem up to 30 cm. high, usually much branched, glabrous except for the lowest internodes which are shortly and densely glandular-pubescent, sometimes the middle and upper parts of the stem sparsely glandular-pubescent (plants from Evvoia). Lower leaves 10-25 mm. long, linear, 3-veined. Inflorescence lax ; pedicels 4-14 mm. long. Calyx 4:5-5-5 mm. long, glabrous or rarely sparsely glandular ; costae 1-veined, sometimes with very feeble lateral veins ; teeth acute or acuminate. Petals 5~7 mm. long, linear-oblong, entire, pale-yellow, purplish at the base. Anthers white. Seeds 1-8-2 mm. long, 0-g-1:1 mm. broad, margin not thickened. Distribution : South-eastern Greece. GREECE: Attiki: Hymettus, Heldreich (BM; E; K ; W-HAL); Spruner (K) ; Haldcsy (E ; W-HAL). Hymettus, versus Pankrati, 13-14 June 1932, Rechinger fil. 1855 (W-RECH). Parnes, 18 June 1878, Heldreich (W-HAL) ; 18 July 1911, Tunta 1278 (W-HAL). Parnes, Karavola, 29 June 1930, Guiol 1271 (BM). Evvoia: Inter Psachna et Achmet Aga, 27 May 1955, Rechinger fil. 16509 (W). Limni, 28 May 1955, Rechinger fil. 16675 (W) (ad P. armerioidem trans.). P. ochroleuca is very similar to P. illyrica and is only doubtfully retained as a distinct species. It can be recognized by a combination of characters (indumentum, flower and anther colour, pedicel length), none of which on its own serves reliably to REVISION OF THE GENUS PETRORHAGIA 139 distinguish this species from P. illyrica. The material collected recently by Rechinger from Evvoia serves only to emphasize the doubtful status of P. ochroleuca. 3. Petrorhagia armerioides (Ser.) Ball & Heywood, comb. nov. Gypsophila armerioides Ser. in DC., Prodr. 1 : 353 (1824). Tunica sibthorpii Boiss., Diagn. Pl. Or. Nov. 2 (8) : 61 (1849), nom tllegit. ; Fl. Or. 1: 521 1867). Pi ares (Ser.) Halacsy, Consp. Fl. Graec. 1 : 194 (1900).—Hayek, Prodr. FI. Penins. Balcan. 1 : 222 (1924). Stem up to 30 cm. high, densely glandular-pubescent at the base and apex, usually glabrous in the middle, but sometimes glandular-pubescent throughout (plants from Greek mainland). Lower leaves 15-30 mm. long, linear or linear-oblong, 3-veined. Inflorescence fastigiate or lax with pedicels 1-5(—9) mm. long. Calyx 4-6(—6-5) mm. long, densely glandular-pubescent ; costae strongly 3-veined ; teeth triangular- mucronate or acuminate. Petals 6-8(-9) mm. long, oblong-spathulate, white, often purple-spotted at the base. Anthers purplish. Seeds 1-8-2-1 mm. long, I-I:3 mm. broad, margin not thickened. Distribution : South-eastern Greece, Aegean Islands, ? north-western Turkey. 3a. Petrorhagia armerioides var. armerioides. (Fig. 5.) Tunica argentii Meikle in Kew Bull. 9 : 106 (1954) pro parte, quoad pl. penins. Graec. Inflorescence dense, fastigiate ; pedicels usually 1-5 mm. long. GREECE : Sterea Ellas : Graecia, 1836, Zuccarint 592 (K). Ad ruinas Delphorum, 4 Aug. 1904, Maire 41 (W-HAL) (ad P. illyricam trans.)._ Prope Palaeo Kuntura, montis Pateras, June 1876, Heldreich (W-HAL). Circa Marathon, 1-4 July ro11, Tunta 1419 (W-HAL). Prope Athena, Reliquiae Orphanideae (E) (ad P. tllyricam trans.). Aegaleos prope Daphni, 23 June 1905, Bretzl (W-HAL). Parnassos, supra Arachova, 31 July 1852, Heldreich 5 (BM) (ad. P. allyricam trans.). Mt. Asine, Argolis, June 1931, Atchley 1103 (K). Prope Vraona, 1 June 1930, Gutol 1468 (BM). Euboea, Mt. Dirphys, 10-17 July 1880, Heldreich (W-HAL) (ad P. allyricam trans.). Kikladhes : Amorgos, 22-23 May 1808, Dorfler 8,169 (BM ; K ; W-HAL). Amorgos, ad portam Katapola, 30 June-6 July 1932, Rechinger fil. (W-RECH). Amorgos, Khozoviotissa, 4 Sept. 1939, Davis 935 (E; K). Naxos, 24-27 June 1889, Heldreich & Halécsy (W-HAL) ; Apr. 1898, Dérfler, Herb. Norm. 3806 (BM ; E ; W-HAL). Naxos, prope Apiranthos, 16 June 1898, Dérfler 148 (W-HAL). Naxos, m. Phanariotissa, 28 June 1932, Rechinger fil. 2259 (W-RECH). Tinos (Tenos), 20 May 1851, Heldreich (K). Tinos, Esomeria, 28 July 1go1, Heldreich (W-HAL). 30. Petrorhagia armerioides var. laxa (Bornm. & Rech.) Ball & Heywood, comb. nov. Tunica armerioides var. laxa Bornm. & Rech. in Beih. Bot. Centralbl. 54 (B) : 614 (1936). Tunica argentii Meikle in Kew Bull. 9 : 106 (1954), quoad typum, excl. pl. penins. Graec. Inflorescence lax ; pedicels 3-9 mm. long. GREECE: Samos: Monte Ambelos, 16-23 June 1932, Rechinger fil. 2098 140 REVISION OF THE GENUS PETRORHAGIA (W-RECH). Chios: Monte Plaka supra Karies, 12-14 May 1934, Rechinger fil. 5415 (W-RECH). Pelinaion, 12 July 1939, Platt 349 (K, holotype of Tunica argentit). Fics. 5-8. 5, Petrorhagia armeriotdes (Ser.) Ball & Heywood var. armerioides. 6, P.candica Ball & Heywood. 7, P. lycica (P. H. Davis) Ball & Heywood. 8, P. cretica (L.) Ball & Heywood. a, calyx ; b, petal. (All x5.) P. armervioides is quite distinct in the Aegean Islands, being distinguished by the 3-veined calyx costae, the indumentum, and the usually short pedicels and dense inflorescence. However, on the mainland of Greece and Evvoia intermediates between this species and P. ilyrica or P. ochroleuca, or between all three species, seem to occur fairly frequently. Nearly all the plants from these areas are to some extent intermediate as they are usually pubescent throughout the length of the stem. Plants with lax inflorescence and with feeble lateral veins on the calyx costae also occur occasionally (e.g. prope Athena, Reliqguiae Orphanideae). It seems desirable that these populations should be studied in considerable detail, but this must await further extensive collections particularly in Evvoia and Andros and other islands off REVISION OF THE GENUS PETRORHAGIA 141 the coast of Attiki. When such material is available it may be found necessary to regard P. armerioides and P. ochroleuca as subspecies of P. illyrica. The variety which occurs in Chios and Samos is rather unusual, although it appears to differ from typical Aegean plants only in the lax inflorescence. It should not be confused with the long-pedicelled plants occurring on the mainland of Greece and Evvoia which are intermediates between this species and P. ochroleuca or P. illyrnica. 4. Petrorhagia candica Ball & Heywood, sp. nov.; inflorescentia laxa pauciflora absque bracteis epicalycinis, calycis costis 3-venosis, petalis emarginatis retussive albis subtus roseo-venosis, seminibus nigris laevibus c. 1-5 mm. longis, facile distinguitur. (Fig. 6.) Gypsophila cretica Smith, Fl. Graec. Prodr. 1 : 280 (1809) pro parte, quoad descr., excl. syn..; Pi. Graec:.4 : 76, t. 384 (1823). Tumica cretica Fisch. & Meyer, Index Sem. Hort. Bot. Petrop. 4: 49 (1837) pro parte, quoad descr., excl. syn. L.—Boiss., Fl. Or. I : 520 (1867). Tunica taygetea P. H. Davis in Not. R. Bot. Gard. Edin. 22 : 165 (1957) pro parte, quoad syn. TI. creticam et Gypsophilam creticam. Stem up to 20 cm. high, simple or little branched, glabrous or sparsely glandular- pubescent. Lower leaves 10-20 mm. long, linear-subulate, 3-veined. Inflorescence lax, few (1-8)-flowered ; pedicels (4—)6-20 mm. long. Calyx (3—)3°5-4:5 mm. long, glabrous or sparsely glandular-pubescent ; costae 3-veined, the lateral veins some- times weak ; teeth acute or obtuse-mucronate. Petals 4:-5~7 mm. long, oblong- spathulate, retuse or emarginate, white with pink veins on the lower surface. Seeds (slightly immature) c. 1-5 mm. long. Distribution : Crete. CRETE : Crete, c. 1930, Guiol (BM; LIVU). Sitia, May 1846, Heldreich (BM, holo- type ; CGE; K); Szeber 204 (BM; K). Pyrgiotissa, Kryavrisis, 1 June 1904, Doérfler 384 (W-RECH). Hag. Vasilis, Preveli, 19 June 1904, Dérfler 632 (W-RECH). Hierapetra, montes Aphendi Kavusi, inter Schinokapsala et Orinon, 20 May 1942, Rechinger fil. 13183 (K). Sphakia, Levka Ori, inter Samaria et Potamos, 1942, Rechinger fil. 13802 (BM). As recently pointed out by Davis (1957), Saponaria cretica L. is the species described by Fischer and Meyer as Tunica pachygona and not the plant commonly called T. cretica. Unfortunately we cannot agree with Halacsy (Consp. Fl. Graec. 1 : 194 (1900)) or Davis that Tunica cretica auct. is identical with T. illyrica var. taygetea or that they can be included within the same species, so it is necessary to treat this Cretan plant as a new species. P. candica can be readily recognized by the emarginate or retuse petals, the 3-veined calyx costae, longer pedicels and relatively sparse indumentum. The cauline leaves tend to be shorter and there are more pairs per unit length than in other species of this section. This, combined with a lesser degree of branching, gives the plant a strict appearance not found elsewhere in the section. 142 REVISION OF THE GENUS PETRORHAGIA With some justification Davis minimizes the importance of 3-veined calyx costae, but his statement that ‘‘ botanists seem to have been hypnotised into seeing them [3-veined costae] in 7. cretica auct.” is not correct. Nor is his statement that Rechinger in Flora Aegaea keys out T. cretica auct. as having I-veined costae. The lateral veins are sometimes feeble, but never completely absent. All the material seen from Crete has proved to be P. candica while all “‘ Tunica cretica ’’’ from other areas is referable to other species. The only exception to this is Baldacci, Iter Creticum 107, which seems to be P. tllyrica subsp. tlyrica. The plants from this collection are all rather small and the three sheets seen are all very similar, so that there seems little likelihood of an error. Further investigation at this locality is desirable. A second collection by Baldacci, [ter Creticum Alterum 293, is probably wrongly labelled. 5. Petrorhagia lycica (P. H. Davis) Ball & Heywood, comb. nov. (Fig. 7 ; Plate 13 B.) Tunica lycica P. H. Davis in Not. R. Bot. Gard. Edin. 22 : 163 (1957). Stem 10-20 cm. high, branched in the upper part, densely glandular-pubescent. Lower leaves 10-15 mm. long, linear to linear-spathulate, 3-veined. Inflorescence lax, the pedicels 3-18 mm. long. Calyx 5-6-5 mm. long, glabrous or sparsely glandular-pubescent on the costae ; costae 3-veined ; teeth triangular, acuminate to acute-mucronate. Petals 7-9 mm. long, oblong-spathulate, white, purple-veined on the lower surface. Seeds 1-5 mm. long, 0-5 mm. broad, brown, minutely papillose. Distribution : Turkey (western Anatolia). TuRKEY : Mugla, Fethiye, Baba Dag above Akbel yayla, 30 July 1947, Davis 13675 (E, holotype ; K). Minara, 31 July 1947, Davis 13709 (BM ; E ; K). This recently described species does not seem to be very closely related to any of the other species in this section. Its seeds are more or less intermediate between the two types found in the genus, but appear to be most similar to those of P. candica. In other characters it resembles P. armerioides var. laxa and P. ochroleuca, but even so it seems to be a quite distinct species. For further discussion see Davis (1957). Subsect. b. Creticae Ball & Heywood Petrorhagia subsect. Creticae Ball & Heywood, subsect. nov. ; a subsectione [llyricis differt habitu annuo, seminibus majoribus 2-3-5 mm. longis. Type : P. cretica (L.) Ball & Heywood. Annual. Seeds 2—3:5 mm. long. 6. Petrorhagia cretica (L.) Ball & Heywood, comb. nov. (Fig. 8.) Saponaria cretica L., Sp. Pl., ed. 2, 1: 584 (1762). Gypsophila cretica (L.) Smith, Fl. Graec. Prodr. 1 : 280 (1809) quoad syn. Tunica cretica (L.) Fisch. & Meyer, Index Sem. Hort. Bot. Petrop. 4: 49 (1837) quoad syn. L. Tunica pachygona Fisch. & Meyer, tom. cit. : 50 (1837).—Boiss., Fl. Or. 1 : 522 (1867).— Hayek, Prodr. Fl. Penins. Balcan. 1 : 221 (1924). Gypsophila pachygona (Fisch. & Meyer) D. Dietr., Synops. Pl. 2 : 1543 (1840). Tunica brachypetala Jaub. & Spach, Ill. Pl. Or. 1: 11 (1842). REVISION OF THE GENUS PETRORHAGIA 143 Stem up to 40 cm. high, much branched, densely glandular-pubescent, viscid. Lower leaves 8-25 mm. long, oblong, subobtuse, more or less 3-veined. Inflorescence lax ; pedicels 4:5—20(—30) mm. long, glabrous or rarely glandular-pubescent. Calyx 6-10°5 mm. long, glabrous ; intercostal membranes 0-g-1:4 mm. broad. Petals included in the calyx or rarely shortly exserted, linear-spathulate, entire, white, sometimes reddish on the lower surface. Seeds 2-2-8 mm. long, I-3-I-9 mm. broad. Distribution : South-western Asia, Greece. GREECE : Thessalia : Kalampaka, 27 July 1896, Sintents 1223 (E ; K). TuRKEY : Olympi Bithyn., Pichler 135 (W-HAL). Yaila de Bozdagh (Tmolus occid.), 17 July 1854, Balansa ro8 (BM ; K). Vallis Meandri, June 1842, Boisster (BM ; K). Lydia, Magnesia, Mt. Sipylos, g-10 June 1906, Bornmiiller g122 (K). Elmalu, 13 June 1860, Bourgeau, Pl. Lycide 42 (E; K). Vil. Ankara, Beynam, 5 July 1947, Davis 1306 (BM; E; K). Vil. Ankara, Kibris gorge near Kayas, 6 July 1947, Davis 13141 (BM; E; K). Phrygia, Akscheher, 28 June 1899, Bornmiiller 4145 (E ; K). N.d’Ouchak, Phrygie vers, 26 July 1857, Balansa 1308 (BM ; K) ; Relig. Matlleanae 224 (BM; K ; W-HAL). Amasia, 24 June 1880, Bornmiiller r90 (BM ; K). Prov. Adana, district Feke, Gdksu gorge below Himmetli, g July 1952, Davis 19843 (K). Prov. Maras, district Goksun, Binboga dag above Yalak, 14 July 1952, Davis 19952 (K). Ad Angora Galatiae, 1892, Bornmiiller 3012 (BM ; E; K). Cappad. orient., Aucher-Eloy 638 (BM ; K). Sert. austr., Aucher- Eloy 4228 (BM). Armeria turc., Egu, Ekrekdere, 23 June 1890, Sintenis 2682 (K). Kurdistan 5 km. O. von Siirt, 12 June 1936, Frédin 30 (W). Mardin, 21 June 1888, Sintenis 1159 (K). Aintab, 16 June 1865, Haussknecht (BM). Inter Aintab et Marash, June 1907, Haradjian 1493 (W). Mt. Duluk Baba au N. d’Aintab, June 1907, Haradjian 1378 (K). SyRIA : Supra Damascum, mont. Dschebel Kasium, 13-15 May 1910, Bornmiiller 11459 (BM). Aleppo, Syria-Palestine, Montbret (K). Cyprus : Between Platres and Aphamis, 16 May 1941, Davis 3445 (E; K\). Circa Prodromo, in mt. Trodos, 17 June 1880, Sintenis & Rigo g60 (CGE ; K). Platres, 9g June, 18 June, 27 June 1938, Kennedy (K). Near Philani, 13 May 1908, Clement Reid (K). Parakklisha-Kellaki, Limasol, 24 May 1948, Kennedy 1615 (Kx). IRAQ : Gabal Hamrin, N.E. of Baghdad, 24 Nov. 1918 (seed, hort., Aug. 1920), W. E. Evans M/3 (E). Mam Zawita district (Lomana), 1-8 June 1955, Robertson 14456 (K). Jarmo, 1 May 1955, Helbaek 1286 (K). Jebel Khatchra near Balad Sinjar, 5 June 1934, Field & Lazar 646 (KK). Pl. Mesopot., Kurdistan et Mossul, 1841, Kotschy 331 (K). IRAN : Azna, 12 June 1937, Kéte 1378 (W). For discussion of the nomenclature of this species see Davis (1957). The type of Saponaria cretica is undoubtedly this species (see p. 141). 7. Petrorhagia arabica (Boiss.) Ball & Heywood, comb. nov. (Fig. 9.) Tunica arabica Boiss., Diagn. Pl. Or. Nov. 2 (8) : 62 (1849) ; Fl. Or. 1 : 523 (1867). Tunica pachygona var. hirtituba F. N. Williams in Journ. of Bot. 28 : 199 (1890). 144 REVISION OF THE GENUS PETRORHAGIA Stem up to 30 cm. high, much branched, glandular-pubescent, viscid. Lower leaves 15-30 mm. long, linear-oblong, subobtuse, more or less 3-veined. In- florescence lax, with pedicels 3-30 mm. long, densely glandular-pubescent. Calyx 6-5-9 mm. long, densely glandular-pubescent ; intercostal membranes 0-7—1-I mm. broad. Petals 9-12 mm. long, distinctly longer than the sepals, linear-spathulate, entire, white. Seeds 2-7—-3:5 mm. long, 1-6-2 mm. broad. [2 Fics. 9-12. 9, Petrorhagia arabica (Boiss.) Ball & Heywood. 10, P. hispidula (Boiss. & Heldr.) Ball & Heywood. 11, P. alpina (Habl.) Ball & Heywood subsp. alpina. 12, P. kennedyae (Jacks. & Turrill) Ball & Heywood. a, calyx ; b, petal. (All x5.) Distribution : Israel and Jordan. ISRAEL, JORDAN : Magdala, 1863-64, Lowne (BM ; CGE; E; K). Mt. Gilboa, top of pass from Bari to Faggira, 17 May 1942, Davis 4739 (E ; K). East of Jordan, 1873, J. A. Paine (K). Wad-el-Kat, 31 Mar. 1909, Meyers & Dinsmore 5311 (E ; K). South of Shahba, 28 June 1932, Dinsmore rrorz (K). N. of Nablus, I June 1935, Eig, Feinbrun & Leinkram 320 (E ; K). Inter Tiberiadem et Safed, 9 June 1881, Letourneux 328 (K). Shittim plain, 24 Apr. 1886, Post (K). REVISION OF THE GENUS PETRORHAGIA 145 This species generally appears to be distinct from P. cretica. It can be dis- tinguished by the usually much larger seeds, the pubescent calyx with compara- tively narrow intercostal membranes, and the petals always distinctly longer than the sepals. PP. cretica rarely has the calyx very slightly glandular at the base and more frequently has the petals slightly longer than the sepals. The frequency with which these apparent intermediates occur is uncertain, but if they should prove to be widespread in Syria, Israel and Jordan then P. arabica should perhaps be regarded as a subspecies of P. cretica. However, from the sparse material seen these two species appear to be distinct in this area, although there has undoubtedly been some confusion between them. 8. Petrorhagia hispidula (Boiss. & Heldr.) Ball & Heywood, comb. nov. (Fig.10.) Tunica hispidula Boiss. & Heldr. in Boiss., Diagn. Pl. Or. Nov. 2 (8) : 62 (1849).—Boiss., POP S 5231867). Stem up to 30 cm. high, much branched, glandular-pubescent, viscid. Lower leaves 10-25 mm. long, linear-oblong, linear or linear-subulate, more or less 3- veined. Inflorescence lax, with pedicels 4-15 mm. long, densely glandular-pubescent. Calyx 5-5-7 mm. long, pubescent or subglabrous ; intercostal membranes 0-4—0-7 mm. broad. Petals 5-5-8 mm. long, included in the calyx or slightly exserted, linear-spathulate, 4-lobed, white. Seeds 1-8—2:2 mm. long, 1-2-1-7 mm. broad. Distribution : Turkey (southern Anatolia). TuRKEY : Inter Adalia et Maela, July 1845, Heldreich (E ; K; W; W-HAL ; isotypes). Supra Adalia, July 1845, Heldreich (E). Near Gebiz, 22 July 1949, Davis 15476 (BM ; E; K). Sect. 2. PSEUDOGYPSOPHILA (A. Braun) Ball & Heywood Petrorhagia sect. Pseudogypsophila (A. Braun) Ball & Heywood, comb. nov. Tunica subgen. Leptopleura Jaub. & Spach, Ill. Pl. Or. 1: 11 (1842). Tunica sect. Pseudogypsophila A. Braun in Flora 26 : 384 (1843). Tunica sect. Leptopleura (Jaub. & Spach) Boiss., Fl. Or. 1 : 521 (1867). Type : Tunica stricta (Ledeb.) Fisch. & Meyer (=P. alpina (Habl.) Ball & Heywood). Annual. Leaves 1-5-veined. Epicalyx bracts absent. Petals not clawed. Seeds small, blackish-brown, reticulate-tuberculate, with thickened margin. 9. Petrorhagia alpina (Habl.) Ball & Heywood, comb. nov. Gypsophila alpina Habl., Neue Nord. Beitr. 4 : 57 (1783) ; in S. G. Gmel., Reise Russl. 4 : 178 (1784). Gypsophila stricta Ledeb., Ic. Pl. 1 : 4 (1829).—Bunge in Ledeb., Fl. Alt. 2 : 129 (1830). Tunica stricta (Ledeb.) Fisch. & Meyer, Index Sem. Hort. Bot. Petrop. 4: 50 (1837).— Boiss:, Bl, Or: 1..> .522 (1867). Dianthus recticaulis Fenzl in Ledeb., Fl. Ross. 1 : 287 (1842). Tunica alpina (Habl.) Bobrov, Bot. Zhurn. 43 : 1546 (1958). Stem up to 40 cm. high, erect, much branched, glabrous. Basal leaves rosulate, 146 REVISION OF THE GENUS PETRORHAGIA 5-30 mm. long, 1-4(-7) mm. broad, oblong- to linear-spathulate, acute or obtuse ; cauline leaves linear, acute, I-veined. Calyx 2-5-4:5(-5:5) mm. long, glabrous, green ; teeth shortly triangular-acute or obtuse-mucronate. Petals (3—-)3-5-6 mm. long, linear-oblong, entire, white. Seeds 0-7-1-2 mm. long, 0-4-0-7(-0-9) mm. broad, ovate-oblong. Distribution : Mountains of western and central Asia, eastwards to the western Himalaya ; southern Bulgaria. ga. Petrorhagia alpina subsp. alpina. (Fig. 11.) Leaves oblong-spathulate. Inflorescence strict ; pedicels 0-5—12(-20) mm. long, erect. TuRKEY : Nigde, Hasan Dag below Taspinar Yayla, 17 June 1952, Davis 19003 (BM; E; K). Rize, Vallée de Cimil, 16 July 1866, Balansa 1415 (BM; K). Kayseri, Bakir Dag, above Kisge, 2,000 m., 30 June 1952, Davis 19385 (BM ; E ; K). Adradices mont. Tech-Dagh, supra Erzeroum, July 1853, Huet (BM ; K). Giresun, Balaban daglari above Tamdere, 1,800 m., 6 Aug. 1952, Davis 20510 (BM; E; K) (ad subsp. olympicam vergens). Van, Catak, Kavussahap Dag, 2,900 m., 24 July 1954, Davis 23063 (BM; E; K). (Coruh, Kordevan dag near Kutul yayla, 28 June 1957, Davis 30237 (E). Kars, Yalnizcam Daglari above Yalnizgam, 2,100-2,300 m., 19 Aug. 1957, Davis 32501 (E). Erzincan, Kesis dag above Cimin, 2,300 m., 28 July 1957, Davis 31832B (E). U.S.S.R.: Armenia, Migri, inter et supra Vargavar et Lishkras, I,500-1,600 m., 29 June 1934, Karjagin (BM). Ahas Tauman, Iber. occid., Smirnoff (W-HAL). Azerbaidjan, Nakhichevanskaya, 17 June 1947, Grossheim, Ilinskaya & Kirpichnikov (BM). Kazakhstan, Semipalatin, Karkaraly, 1890, Korjinsky (BM). Altai, Ledebour (K). Dschany-bek, mont. Tarbagatai, mont. Kurtschum, 1840, Karelin & Kiriloff 136 (BM ; K). Prope Kolyvanskoje Osero, 12 May 1go1, Krylov (K). Tadzhikistan, Iovon, 10 June 1893, Komarov (K). Tadzhikistan, Kuli-Mohif, May 1893, Komarov (K). . IRAN : Mazanderan, in valle fluvii Calus, 9 June 1937, Rechinger fil. 913 (BM ; K). Tochtal, 3,300 m., July 1935, Lindsay 511 (BM). M. Demaurend, 2,100 m., 21 June 1843, Kotschy 354 (BM ; K). Alatau ad fl. Lepys, Kotschy 1296 (BM). PAKISTAN, KASHMIR: Kashmir, Falconer 237 (K). Huripur to Aliabad, Falconer 2984 (K). Shapujon, 2,100 m., 9 July 1876, Clarke 28596 (K). Aliabad, 2,700 m., 9 July 1876, Clarke 28677 (K). Nowgunje, 2,000 m., 19 July 1876, Clarke 29199 (K). Kunshoan, 2,250 m., Clarke 29368 (K). Liddar, S. Kashmir, 2,650 m., 5 July 1902, Drummond 14128 (K). Between Aru and Lidarwat, 6 July 1902, Drummond 14125 (K). Ushkra, 1,600 m., July 1891, McDruell (K). Kammi pass, above Shankangarh, 3,000 m., 16 July 1946, Stewart 22764 (K). Near Gulmarg, 2,400-2,700 m., 19 June 1893, Duthie 13016 (BM). Swat, Utrot, 2,400 m., Stewart & Rahman 25269 (BM). gb. Petrorhagia alpina subsp. olympica (Boiss.) Ball & Heywood, stat. nov. (Plate 13 C.) REVISION OF THE GENUS PETRORHAGIA 147 Tunica olympica Boiss., Diagn. Pl. Or. Nov. 2 (8) : 61 (1849) ; Fl. Or. 1: 522 (1867). Tunica filiformis Post in Bull. Herb. Boiss. 1 : 411 (1893). Leaves linear-spathulate. Inflorescence spreading ; pedicels (0:5—-)3-30 mm. long, patent. BuLGaRiA : In pinetis m. Pirin ad Suchodol, 14 July 1929, Stojanoff, Stefanoff & Georgieff (K). TURKEY : Olympus, Aug. 1842, Boissier (BM ; E; K ; isotypes) ; June-July 1862, Mill (BM ; K) ; July 1874, Pichler 135 (K) ; Bornmiiller 446 (E). Uludag, 1,800-2,180 m., 19 June 1956, Moore 7287 (E). M. Ida, prope Kareikas, 1893, Sintenis 599 (BM ; E ; K ; W-HAL). Kutahya, Gediz, Saphane dag, 6 Aug. 1950, Davis 18463 (BM ; E; K). Kutahya, Vili Muphlor, Sundras dag at Goékce ova, 27 July 1947, Davis 13613 (E). Vergers de l’Yaila de Bozdagh (Izmir), 28 July 1854, Balansa 105 (BM; K). Phrygia, Sultan dagh, Akscher, 25 June 1899, Bornmiiller 4147 (E ; K). Isparta, Sutculer (Isauria), between Tota bebi yayla and Daribiiykii, 30 July 1949, Davis 15882 (BM; E; K). Isparta, Sutculer, Dedegol dag, between Selkose and Oruz Gazi, yayla, 1,300 m., 1 Aug. 1949, Davis 15929 (E). Isparta, Sutculer, Dedegdl dag between Daribiiykii and Selkose, 30 July 1949, Davis 15866 (E; K). Wilajet Kastambuli (Kastamonu), Tossia in pinetis ad Lhakirla (Cakirlar), 16 June 1892, Sintenis 4651 (BM). Kure-Nahas (near Inebolu), 5 Sept. 1892, Sintenis 5150 (BM ; K). Amasia, m. Ak-dagh, 18 June 1889, Bornmiiller 184 (BM ; K) (ad subsp. alpinam vergens). Cappad. orient., Aucher-Eloy 556 (BM ; K) ; Montbret (K). Taurus, au nord du defile des Portes cilicieuses (Gulek Bogaz), June-July 1855, Balansa (EF). Anatolia orientalis, Sandz, Liggna-dagh, 4 July 1917, Schischkin (BM). Mt. Amanus, July 1801, Post (K, isotype of Tunica filiformis) ; 1906, Haradjian 381 (K) ; 800-1,000 m., Sept. 1913, Haradjian 4685 (W). Kusliji dagh, 1,700-2,200 m., Aug. 1908, Haradjian 2532 (K ; W). Bobrov (1958) has recently shown that Gypsophila alpina Habl. is identical with Tunica stricta (Ledeb.) Fisch. & Meyer. Hablizl’s epithet is therefore the earliest for the species. Boissier listed a number of characters by which Tunica olympica could be dis- tinguished from 7. stvicta (spreading panicle, longer slender pedicels, smaller obovate seeds). Our investigations have shown that these characters are not very satis- factory. The seeds of both taxa appear to be identical, while there is such a wide range in pedicel length, even on the same plant, that this can only be used on a statistical basis. The difference in the inflorescence, although reasonably satis- factory, is not always reliable, and there seems to be some intergradation in northern, central and eastern Turkey. We therefore treat T. olympica as a subspecies of P. alpina. 10. Petrorhagia kennedyae (Jacks. & Turrill) Ball & Heywood, comb. nov. (Fig. 12 ; Plate 13 p.) Tunica kennedyae Jacks. & Turrill in Bull. Misc. Inf. Kew 1938 : 462 (1938). Tunica cypria Rech. f. in Fedde, Repert. Sp. Nov. 47 : 163 (1939). 148 REVISION OF THE GENUS PETRORHAGIA Stem 5-30 cm. high, erect, usually branched from the base, shortly glandular- hispid at the base, glabrous or glabrescent in the middle, glandular and viscid at the apex. Basal leaves up to 12 mm. long and 4 mm. broad, oblong-elliptic, obtuse ; cauline leaves 5-15 mm. long, 0-5-2 mm. broad, 3-veined, linear, acute. Inflorescence lax, many-flowered ; pedicels 5-25 mm. long, spreading. Calyx 4°5-6:5 mm. long, glandular-viscid, the costae often reddish or purplish ; teeth broadly triangular-acute or obtuse-mucronate. Petals 6-8-5 mm. long, linear- spathulate, white, usually reddish on the lower surface, bifid. Seeds 1:3-1-6 mm. long, 0-6—0-8 mm. broad, oblong. Distribution : Cyprus. Cyprus : Troodos mountains, 17 July 1880, Sintenis & Rigo 764 (BM ; CGE ; K ; W-HAL) ; 20-25 June 1912, Haradjian 467 (K). Platres, 18 June 1938, Kennedy 1037 (K, holotype). Mesopotamus, 21 June 1939, Lindburg (K) ; on dry igneous slopes in Pinus forest, 1,200 m., 16 May 1941, Davis 3450 (E ; K). Krigos, Potamos, 20 June 1938, Kennedy 1038 (K). Stevros (Paphos), 1,200 m., 3 July 1940, Davis 1759 (E ; K). Rondhkias valley between Pano and Kykko, 450-600 m., 8 May 1941, Davis 3385 (E ; K). Above Alithenon, Merton 2408 (K). This species appears to have been first collected as long ago as 1880 by Sintenis and Rigo, but was not recognized until 1938 when Jackson & Turrill published a description ; soon afterwards Rechinger described it independently. The relationships of this species are somewhat uncertain. Jackson & Turrill placed it in Tunica sect. Pachypleura, primarily because of its superficial resemblance to P. cretica (Tunica pachygona). It resembles that species in its annual habit, glandular indumentum and comparatively large flowers, but differs in the one-veined calyx costae and the small, blackish-brown scutate seeds. These differ- ences are here considered to be of major taxonomic significance and indicate that this species should be placed in Sect. Pseudogypsophila together with P. alpina, which is also an annual species. However, P. kennedyae is quite distinct from P. alpina, and can be distinguished by a series of characters (e.g. indumentum, larger flowers, bifid petals, etc.). Sect. 3. PETRORHAGIA Petrorhagia sect. Petrorhagia ; sectio typica generis. Tunica sect. Gypsophiloides Griseb., Spicil. Fl. Rumel. 1 : 184 (1843), nom. illegit. Tunica sect. Pseudodianthus A. Braun in Flora 26 : 384 (1843), nom. illegit. Tunica sect. Eutunica Boiss., Fl. Or. 1 : 518 (1867), nom. illegit. Tunica sect. Tunicastrum F. N. Williams in Journ. of Bot. 28 : 194 (1890), nom. illegit. Type: P. saxifraga (L.) Link. Perennial. Leaves I-veined (rarely up to 5-veined). Flowers with or without epicalyx bracts. Petals not abruptly clawed. Seeds blackish-brown, tuberculate, with thickened margin. Subsect. a. Saxifragae Ball & Heywood Petrorhagia subsect. Saxifragae Ball & Heywood, subsect. nov.; subsectio typica REVISION OF THE GENUS PETRORHAGIA 149 sectionis Petrorhagiae, continens species caulibus valde ramosis, floribus solitariis vel fasciculatis, bracteis lanceolatis vel ovatis. Type: P. saxifraga (L.) Link. Stems much branched. Flowers solitary or fasciculate ; bracts and epicalyx bracts (when present) lanceolate or ovate, membranous, I-veined. 11. Petrorhagia phthiotica (Boiss. & Heldr.) Ball & Heywood, comb. nov. (Fig. 13 ; Plate 14 A.) Tunica phthiotica Boiss. & Heldr. in Boiss., Fl. Or., Suppl. : 82 (1888). Tunica ochroleuca var. phthiotica (Boiss. & Heldr.) Hayek, Prodr. Fl. Penins. Balcan. 1 : 223 (1924). Stems 3-20 cm. high, sparsely glandular-pubescent or subglabrous. Lower leaves 3-8 mm. long, I-3-veined. Inflorescence lax ; epicalyx absent. Calyx 3-5 mm. long, glabrous ; costae I-veined ; teeth broadly triangular to very obtuse, mucronate. Petals 4-6-5 mm. long, 0-5-1 mm. broad, oblong-spathulate, entire, pink. Seeds 1-2—1-6 mm. long, 0:55—0-75 mm. broad. [5 16 Fics. 13-16. 13, Petrorhagia phthiotica (Boiss. & Heldr.) Ball & Heywood. 14, P. fascicu- lata (Marg. & Reut.) Ball & Heywood. 15, P. saxifraga (L.) Link var. saxifraga. 16, P. graminea (Smith) Ball & Heywood. a, calyx ; b, petal ; c, epicalyx bract. (All x 5.) Distribution : South-eastern Greece (Oiti) ; very doubtful in north Peloponnisos. GREECE : Sterea Ellas: M. Oeta, 14 July 1879, ?Heldreich (W-HAL, ?isotype) ; 30 June 1935, Guiol 2366 (BM). Katafigon, Oeta, 11 June 1937, Balls & Gourlay 3244 (BM ; K). Peloponnisos : Kyllene at head of the gorge of Phlambouritza, June 1862, Mill (K) (? locality correct). 150 REVISION OF THE GENUS PETRORHAGIA The relationships of this species are somewhat uncertain, although it is perhaps nearest to P. fasciculata. However, P. phthiotica also shows some resemblance to P. alpina and it may represent a link between Sects. Petrorhagia and Pseudogypsophila. It was originally placed in his Tunica sect. Gypsophiloides by Boissier, but the very different seed type, in addition to a number of secondary characters, clearly dis- tinguishes it from the other members of that group, which we place under Petrorhagia sect. Pseudotunica. It is difficult to understand why Hayek regarded it as only a variety of Tunica ochroleuca, but this may have been due to the paucity of material, and the failure to recognize the importance of the seed characters. Apart from the seeds, P. phthiotica can easily be distinguished from P. ochroleuca by the smaller pink or purplish flowers, the broad triangular mucronate calyx teeth, and the different indumentum (cf. Davis, 1957). P. phthiotica differs from other members of Sect. Petrorhagia in a number of characters, the significance of which is somewhat uncertain. The distinctive calyx teeth, glabrous calyx, entire petals, complete absence of fasciculation of flowers, and the sparse patent-glandular hairs are the more obvious differences, and these may well prove to be more indicative of relationships than the perennial habit, the very narrow leaves, and the pink flowers. 12. Petrorhagia fasciculata (Marg. & Reut.) Ball & Heywood, comb. nov. (Fig. 14.) Gypsophila fasciculata Marg. & Reut. in Mém. Soc. Phys. & Hist. Nat. Genéve 8: 281 (1839).—Boiss., Fl. Or. 1 : 555 (1867). Tunica fasciculata (Marg. & Reut.) Boiss., Fl. Or., Suppl. : 82 (1888).—Hayek, Prodr. F1. Penins. Balcan. 1 : 223 (1924). Stems 5-30 cm. high, glandular-pubescent, sometimes glabrous above. Lower leaves 6-30 mm. long, 0-5—-1-8 mm. broad, usually 1-veined, rarely up to 5-veined. Inflorescence branched ; flowers usually fasciculate and bracteate, rarely solitary and then ebracteate. Calyx 2:5—5:5 mm. long, shortly or sparsely hirsute or glandu- lar ; costae I-veined ; teeth acute or subobtuse. Petals 3-5-6 mm. long, linear- spathulate, retuse, white or pale-yellow. Seeds 0-8-1 mm. long, 0-6—0-7 mm. broad. Distribution : Western Greece and Ionian Islands. GREECE : Sterea Ellas : Stratos, Aetolia, 1935, Guiol 2439 (BM). Arcania, prope Agrinion, 3 July 1893, Haldcsy (K ; W-HAL). Zacynthos, Mazztoiri (W-HAL). Hag. Ilias, Levkas, 30 June 1895, Baldacci 13 (BM ; K ; W-HAL). Kephallonia, Same (Lixourion), 1926, Bornmiiller 188 (BM ; K), rg0 (BM ; K). Cephallonia, mont. Phalaris, Aug. 1867, Heldreich (K). This species appears to be a reasonably distinct member of Sect. Petrorhagia and can be recognized by the white or pale-yellow, usually fasciculate flowers, the smaller calyx With more or less acute teeth and one-veined costae, and the patent, glandular hairs (sometimes even on the calyx). These last two characters are shared with P. phthiotica and this to some extent supports the view that P. phthiotica should be placed in this section. However, P. fasciculata has a number of important features in common with the other members of the section (e.g. the fasciculation REVISION OF THE GENUS PETRORHAGIA 151 of the flowers, the white hirtulose calyx indumentum, the more or less oblong calyx teeth, and the very narrow linear leaves) so that there can be no grounds for regarding it as separate. It should be noted that plants of P. fasciculata sometimes occur with all the flowers solitary, which are then ebracteate. 13. Petrorhagia saxifraga (L.) Link, Handb. Erkenn. Gewachse 2 : 235 (1831). Dianthus saxifragus L., Sp. Pl. 1 : 413 (1753). Gypsophila saxifraga (L.) L., Syst. Nat., ed. 10, 2 : 1028 (1759). Tunica saxifraga (L.) Scop., Fl. Carniol., ed. 2, I : 300 (1772). Dianthus filiformis Lam., Fl. Frang. 2 : 537 (1778), nom. illegit. Imperatia filiformis Moench, Meth. Pl. : 60 (1794), nom. tllegit. Gypsophila multicaulis Poir. in Encycl. Méth., Bot., Suppl. 2 : 875 (1812). Petrorhagia multicaulis (Poir.) Link, loc. cit. (1831). Gypsophila arenicola Dufour in Bull. Soc. Bot. France 7 : 240 (1860). Tunica erecta Jord. & Fourr., Brev. Pl. Nov. 1 : 10 (1866). Tunica bicolor Jord. & Fourr., loc. cit. (1866). Tunica xerophila Jord. & Fourr., loc. cit. (1866). Tunica ciliata Dulac, Fl. Dép. Haut.-Pyrén. : 260 (1867), nom. illegit. Tunica arenicola (Dufour) F. N. Williams in Journ. of Bot. 28 : 194 (1890). Kohlrauschia saxifraga (L.) Dandy in Watsonia 4 : 42 (1957). Stems 5-45 cm. high, glabrous, shortly papillose or scabrid-pubescent. Lower leaves 5-25 mm. long, 0-4-1°5(—2:5) mm. broad, I-veined. Inflorescence lax or fasciculate ; flowers usually with 4 epicalyx bracts. Calyx 3-6(-7) mm. long, glabrous or occasionally sparsely pubescent ; costae I-veined with 2 weak lateral veins ; teeth oblong, obtuse. Petals 45-10 mm. long, 1-2-3(-4) mm. broad, the limb obcordate, white or pink. Seeds o-g-1-6 mm. long, 0-6—-1-I mm. broad. Distribution : Central and southern Europe and south-western Asia. 13a. Petrorhagia saxifraga var. saxifraga. (Fig. 15 ; Plate 14 B.) Lower leaves 0:4-1:5 mm. broad. Flowers solitary. Calyx 3-6 mm. long. PorTuUGAL : Porto, 1848-50, Welwitsch 851 (BM; CGE). Carrica, margem direita do rio Douro a 2 km. de Barca d’Alva, 27 June 1955, Fernandes, Matos & Matos, Flora Lusitanica 5664 (BM). SPAIN : Valentia, montis Segaria, pr. Dianium, 8 May 1923, Gros, Flora Iberica 119 (BM ; K). Denia, Alicante, 12 May 1928, Ellman & Sandwith 1132 (K). FRANCE : Gard: West of river Rhone opposite Avignon, 13 June 1952, Souster 1249 (K). Vaucluse : Serignon, 1 June 1880, Delacour (K). Dréme : Romans-sur-Isére, June 1872, anon. (W-HAL). Rhdne: Silleuxbanne prés Lyon, 19 June 1859, Guichard (K ; W-HAL). Ain: Balan, 10 Sept. 1865, Miciol (BM ; W-HAL). Jura: Sept. 1857, Herb. E. A. Wilmott (K). Haute Savoie: La montagne de Mandallax prés d’Annecy, Billot 1431(2) (BM ; CGE). Savore : Mont. Lemene, prés de Chambéry, Billot 1431(r) (BM; CGE). JIsére: Grenoble, 6 July 1884, Magnier (W-HAL). Hautes-Alpes ; Briancon, Aug. 1870, Reverchon (K). Basses- Alpes : Le Roc, Castellare, 27 June 1950, Brenan 9781 (K). Var: St. Raphael, 152 REVISION OF THE GENUS PETRORHAGIA 3 June 1873 (FI). Alpes-Maritimes : Menton, Dec. 1866, Joad (E) ; Rechinger fil. 1507 (W-RECH) ; Nov. 1920, Lester-Garland (K). Corsica : Evisa, 13 June, 28 July 1884, Reverchon (BM ; E ; FI; K ; W-HAL). Pentes du Pigno, a Bastia, 2-13 June 1867, Mabille 21b (BM ; CGE; E; FI; K). Vizzavona, July 1907, Martelli (FI). Ragliano, 26 Sept. 1854, Reveliére 43 (BM). GERMANY: Munchen, Sfitzel 393 (BM; CGE; E; K). Prope Riesenfeld, 1 Aug. 1883, Woerlein (LIVU). SWITZERLAND : Geneva, June 1893, Ndgeli (CGE ; E) ; Botssier (K). Between Visp and Zermatt, Evans (E). Lausanne, Aug. 1879, Favrat & Barbey (K). Tarasp, unter Engadin, Aug. 1903, Crawford (E). Between Sierre and Vissoye, Aug. 1903, Lester-Garland (K). Austria : Tirol: Innsbruck, Sept. 1882, Evers (W-RECH). Botzen, 17 Aug. 1860, Babington (K). Ost Tivol: Lienz, June 1871, Gander (E; K; W-HAL). Karnten : (Klagenfurt, July 1872, Ehrenberger (W-HAL). Stevermark: Graz, July 1901, Hayek, Fl. Stir. 956 (BM ; E). Ober-Osterreich : Prope Styr, Dérfler, Fl. Austro-Hung. 3228 I (BM ; E;K ; W-HAL; W-RECH). Nieder-Osterreich : Marchfeld, 21 June 1910, Korb (W). Wien, Aug. 1876, Haldcsy (W-HAL). Burgenland : Podersdorf, 24 June 1923, Rechinger (W-RECH). CZECHOSLOVAKIA : Slovakia austr.-occid., prope Plavecky Ctvrtek, 20 June 1929, Domin & Krajina 35 (CGE ; K). Huncary : Fehér, Fehérvarcsurgé, 28 June 1923, Fl. Hung. 836 (BM; E; K). Tolna, ad stationem Tolnanemeti, 8 July 1922, Pillich (BM ; K). Szontagh, Odenburg (Sopron), 1861, Sempront (K). RoMANIA : Verciorova, 5 June 1887, Degen 20 (E; K; W-HAL). Baile Herculane, June 1907, Schneider 1215 (BM; K). Gura Vaii, June 1881, Grecescu (W-HAL). Kudijévar, 19 June 1875, Tauscher (E). ITaLy : Piemonte: Aosta, 16 June 1898, Vaccari (FI). Susa, 8 Aug. 1854, Parlatore (FI). Lombardia : Bormio, June 1870, Sommier (FI). Liguria: San Remo, Mar. 1880, Marchesetti (FI). Genova, Aug. 1856, Ardissone (FI). Venezia : Lungo il Bachiglione presso Padora, 3 June 1894, Fiori (FI). Venezia Carperica Vittorio, 6 June 1894, Pampanini (FI). Tridentina : S. Leonardo in Passiria, V. Valtina, 17 July 1949, Zenari (FI). Toscana: Firenze, Sept. 1867, Sept. 1875, Groves (FI; K). We d’Elbe, 27 Feb. 1871, Sommier (FI). Grossetto, M. Orace, 14 June 1918, Fiori (FI). Siena, Colle V. d’Elsa, 17 June 1918, Fiori (FI). Marche : Salendo al Vettore (Mti. Sibillini), 10 Aug. 1895, Sommier (FI). Ascoli, 3 July 1856, Parlatore (FI). Emilia: Riola, Rochetta, July 1883, Baldacci (FI). San Marino, 15 Oct. 1916, Pampanini 1267 (FI). Modena, 20 June 1882, Fiori (FI). Lazio: Tivoli, M. Rigali, Sept. 1901, Vaccari (FI). Terracina, 22 May 1893, Sommier (FI). Presso il Colosseo, Roma, 12 May 1886, Aiutz (FI) ; May 1836, Trevelyan (CGE). Abbruzzo: M. di Caramanica, 6 July 1872, Pedicino (FI). Majella, in valli Orfente, ad radic. montis Mucchia, 31 July 1874, Levier (FI). Campania ; Sorrento, June 1911, Guadagno (FI). Monte Alburno, 29 May 1951, REVISION OF THE GENUS PETRORHAGIA 153 Philippis & Moggi (FI). Capri, May 1843, Heldreich (K). Umbra: Mura di Assisi, 14 June 1935, Corradi (FI). Ierni, Alle Cascate, 26 June 1875, Marchesetti (FI). Puglia: Taranto, June 1919, Lacaita, Fl. Ital. 2434 (BM ; FI; K). Monte Nero presso S. Marco, 10 May 1893, Martelli (Fl). Basilicata: Mt. Sena delli Ciovoli, 4 Aug. 1929, Gavioli 14996 (FI). Vulture, July 1928, Biondi (FI). Calabria: Palizzi, 9 June 1877, Arcangeli (FI). Reggio, 15 May 1872, Anztz (FI). SARDINIA : Sadali, 24 July 1898, Martelli (FI). Terranora Pousania, 16 June 1899, Martelli (FI). Sicity : Etna, Sept. 1841, Parlatore (FI). Palermo, May 1905, Ross 515 (BM ; E ; FI). Madonie a Quacedda, June 1840, Parlatore (FI). Ins. Favigna, 5 May 1855, Huet (BM ; CGE; FI; K). Sorrentino, July 1845, Sorrentino (FI). Jucostavia : Slovenija: Near Bled, 1923, Leathes (BM). Radmannsdorf, Aug., Paulin, Fl. Carniolica 262 (BM). Hrvatska: Zagreb, 2 July 1879, Rossi (K). Pola, Capo Promontore, 6 June 1919, Vaccari (FI). Spalato (Split), Pichler, Fl. Austro-Hung. 3228 II (BM ; E; W-HAL ; W-RECH). Lesina, 18 May rg10, Keller (W-HAL). Dubrovnik, 26 June 1953, Dennis 52 (K). Srbya: Nischka Banja prope Nisch, 15 June 1955, Rechinger fil. 16064 (W). Vranja, 17 July 1887, Bornmiiller (E). Sveto-Petra, 26 Aug. 1921, Yermoloff (BM). Novi Pazar, 8 July 1956, Carpenter 342 (K). Bosna 1 Hercegovina: Nevesinje, 23 Aug. 1895, Callier m7t (K; W-HAL). Sarajevo, 22 June 1930, Gillrat-Smith 2662 (K). Mont. Porim, prope Mostar, July 1893, Vandas (K). Crna Gora: Lovcen, 9 July 1927, Korb (W). Mt. Jerinja, July 1906, Rohlena (BM). In colle Ljinbovic ad Podgorica, 11 Sept. 1902, Baldacci 212 (BM). Kostat, 14 July 1891, Baldacci (E ; W-HAL). Makedonija : Inter Debar et Mavrovi Hanovi, 12 June 1955, Rechinger fil. 15796 (W). Gjefgjeli, 1-2 June 1927, Rechinger fil. 1564 (W-RECH). Kaunferiji, June 1906, Adamovié (BM). ALBANIA : Pastrik, 31 Aug. 1916, Dérfler 343 (E; K). West of L. Ohridsko, 13 Aug. 1938, Hepburn 84 (K). Sarandé, 3 June 1933, Alston & Sandwith 1339 (BM ; K). M. Kusema, supra Vonitko, dist. Leskovik, 9 July 1896, Baldacci 215 (K ; W-HAL). Supra Sceperi, dist. Tepelené, 22 July 1894, Baldacci 55 (BM ; E). ButcarRiaA : Near Rila, 2 Aug. 1926, Turrill 1541b (K) ; 2 Aug. 1926, Stojanoff & Stefanoff (BM). Causovo, June 1908, Strzbrny (BM). GREECE : Ipiros: Arta, 5 July 1893, Haldcsy (W-HAL). Mt. Peristeri, 11 July 1893, Haldcsy (W-HAL). Metsovon, June-July 1929, Guiol 1467 (BM). Cephalonia, 19 Oct. 1834, Schimper & Wrest (K). Makedhonia: Prespa L., Io June 1932, Alston & Sandwith 250 (BM; K). Between Turica and Gumus Dere, south of Struma Plain, 5 June 1917, Turrill 381 (K). Peloponnisos: In m. Omplo pr. Patras, 7 June 1899, Heldreich (W-HAL) (ad P. gramineam trans.). Dhiakopton, 8 June 1899, Heldreich (W-HAL) (ad P. gramineam trans.). Between Dhiakopton and Megaspelion, 19 Oct. 1939, Davis roo3 (E ; K) (ad P. gramineam trans.). Panakhaikon Oros, Kokkinobryssi, 22 June 1949, Goulimy 50 (K). TurKEY : Bosphore, Aucher-Eloy 553 (BM; K). Tokat, Niksar-Karakus, 5 Sept. 1954, Davis 24898 (E). Trabzon, Macka, to July 1934, Balls 1626 (E ; 154 REVISION OF THE GENUS PETRORHAGIA K). Giresun, Tandere-Yavuzkenal, near Karinca, 13 Aug. 1952, Davis 20709 (BM; E; K). Rize, June 1866, René de Parquet (BM). Katahor, 7 July 1933, Balls 474 (E ; K). U.S.S.R.: Tiflis, Apr., Smirnoff (W-HAL) ; 29 May 1912, Holmberg 996 (W) ; June 1881, Brotherus 153 (BM). Iberia, Aug. 1824, Wilhelms (K). Armenia, Szovils 75 (K). Altaica, 1826, C. A. Meyer (BM). IRAN : Twenty-five miles west of Zorab, 14 June 1929, Cowan & Darlington (K). Mazanderan, Kudjur, inter Sanus et Kindj, 9-11 Aug. 1948, Rechinger fil. 6570 (BM ; K). Introduced and naturalized : SWEDEN : Skane, Paroecia Sirekopinge, 8 Aug. 1932, Samuelsson 769 (BM ; W). Sireképing, Klasinge, 19 July 1901, Henningson (BM). GREAT BriTAIN : Dorset : Abbotsbury, 5 Aug. 1925, Andrews (K). Pembroke : Tenby, 2 Sept. 1909, Barley (K). 13b. Petrorhagia saxifraga var. glomerata (Ten.) Ball & Heywood, comb. nov. Gypsophila rigida L., Sp. Pl. 1 : 408 (1753). Gypsophila saxifraga var. glomerata Ten., Fl. Nap. 4 : 202 (1830). Petrorhagia rigida (L.) Link, Handb. Erkenn. Gewachse 2 : 235 (1831). Tunica rigida (L.) Reichb., Ic. Fl. Germ. & Helv. 6: 42 (1844).—Boiss., Fl. Or. 1: 518 (1867). Very similar to var. saxifraga but flowers fasciculate. IraLy : Venezia: Bassano, 15 Oct. 1854, Ball (E). Lazio : Roma, Collossea, 30 Aug. 1856, anon. (FI). Avellino, July 1888, Milansz (FI). Campania : Monte Arasualo, 6 Oct. 1954, Agostini (FI). Inter Sorrento et Viso, 31 July 1874, Heldreich (BM ; W-HAL). Puglia: Modugno (Bari), 25 Sept. 1938, Capztaneo (FI). Calabria : Sila Fossiata (Longobucco), 29 July—3 Aug. 1918, Fort (FI). Sicity : Palermo, June 1904, Ross 516 (E ; FI) ; 1893, Gasparini (BM ; CGE). Etna, 1844, Tornabene (FI). Baida, 1848, Tineo (FI). Madonie, 1848, Tineo (FI). JucosLtavia: Hrvatska: Rijeka (Fiume), 19 Oct. 1863, A. M. Smith (FI). Makarska, Aug.—Oct. 1934, Gilliat-Smith 89 (K). ALBANIA : Durrés, 30 Aug. 1935, Alston & Sandwith 2740 (BM ; K). 13c. Petrorhagia saxifraga var. gasparrinii (Guss.) Ball & Heywood, comb. nov. Gypsophila gasparrinii Guss., Fl. Sicul. Synops. 1 : 474 (1843) (“ Gasparrini ”’). Tunica gasparrinii (Guss.) Nyman, Consp. Fl. Eur. : 100 (1878). Tunica saxifraga var. gasparrinii (Guss.) Fiori, Nuov. Fl. Anal. Ital. 1 : 505 (1924). Stems procumbent with numerous rosettes of leaves. Lower leaves I-2:5 mm. broad, oblong. Flowers solitary. Calyx 4-5-7 mm. long. Petals 7-10 mm. long. Sicity : Messina, June 1866, Sequenza 81 (FI) ; Nicotra 21 (W-HAL). Isnello, 1848, Tineo (FI). Madonie, 1866, Pasquale (F1). This very widespread species shows comparatively little variation throughout REVISION OF THE GENUS PETRORHAGIA 155 its range. There is a certain amount of intergradation with P. graminea in Peloponnisos and besides the typical variety there are the two varieties described above, which occur mainly in Italy and Sicily. Var. glomerata is characterized by the fasciculation of the flowers and this seems to occur sporadically throughout the range of the species. However, in Italy and Sicily it is much more frequent, and there tends to be a greater degree of fascicu- lation. Gypsophila rigida L. appears to be identical with this variety. Var. gasparrinu presents a different problem. This variety appears to be confined to Sicily, and perhaps southern Italy. If it were not for this very restricted distribu- tion var. gasparrinit would probably be regarded as no more than an abnormal growth form, but in the present circumstances a field investigation seems to be desirable to determine its true status. It is not at all clear whether var. glomerata or var. gasparrinii occur in discrete populations at all, but what evidence is available suggests that they may not. Numerous other varieties have been described, but it seems doubtful whether any of them deserve taxonomic recognition. 14. Petrorhagia graminea (Smith) Ball & Heywood, comb. nov. (Fig. 16.) Gypsophila graminea Smith, Fl. Graec. Prodr. 1 : 279 (1809). Tunica graminea (Smith) Boiss., Diagn. Pl. Or. Nov. 2 (8) : 60 (1849) ; Fl. Or. 1: 519 (1867). Caespitose, usually very woody at the base with dense rosettes of leaves ; flower- ing stems up to 40 cm. high, shortly and densely pubescent, sometimes glabrous above. Basal leaves 8-20 mm. long, 0-5-0o-9 mm. broad. Inflorescence usually lax, sometimes with a few flowers fastigiate ; epicalyx absent. Calyx 3°5-5:5 mm. long, densely pubescent, sometimes with the teeth subglabrous ; costae 1-veined with 2 obscure lateral veins ; teeth oblong, obtuse. Petals 5-10 mm. long, the limb 1-8—-2:5 mm. broad, obcordate, pink, reddish, or white becoming pink. Seeds I-I-1'5 mm. long, 0-8—0-9 mm. broad. Distribution : Southern Greece (Peloponnisos). GREECE : Peloponnisos : Messenia, Kalamata, Oct.—Dec. 1897, Heldretch 1410 (E; Kj; W-HAL); c. 1930, Guiol (BM). Elide, dans la forét de Danolada, 28 Oct. 1906, Maire & Petitmengin 2295 (W-HAL). Mt. Taygetos, Sept. 1929, Guiol 1494 (BM). This species, normally without epicalyx, appears to be very closely related to P. saxifraga. A number of collections from the Peloponnisos are intermediate between the two species, but it is not clear whether these represent occasional hybrid swarms or geographically intermediate populations. Only an extensive investigation in the area concerned is likely to provide a solution to this problem. Bossier placed this species in his Tunica sect. Gypsophiloides, but the evidence of intergradation with P. saxifraga quite clearly supports the transfer to Petrorhagia sect. Petrorhagia. The seed characters, the white-hirtulose, oblong obtuse calyx teeth, and the occasional fasciculation of the flowers also support this. 156 REVISION OF THE GENUS PETRORHAGIA 15. Petrorhagia rhiphaea (Pau & Font Quer) Ball & Heywood, comb. nov.. Tunica rhiphaea Pau & Font Quer in Font Quer, Iter Marocc. 1929: n. 147 (1930) ; in Cavanillesia 3 : 77 (1930). Stems 5-15 cm. high, glabrous in the lower part, glandular-pubescent at the apex. Basal leaves 4-8 mm. long, 0-4—0-7 mm. broad, linear with obtuse apex. Inflorescence lax ; epicalyx absent. Calyx 2:5-4:5 mm. long, glandular-pubescent ; costae feebly 3-veined ; teeth ovate to oblong, obtuse. Petals 4-7 mm. long, the limb 0-8-1 mm. broad, linear-oblong, retuse, pale-yellow with purple veins. Seeds (immature) 0-8-1-2 mm. long, 0-7—0-9 mm. broad. Distribution : Spanish Morocco. SPANISH Morocco : Inrupibus schistosis supra oppidum Mahzen dictum (Ktama), ad 1,350 m., 29 June 1929, Font Quer 147 (BM, isotype). This species is very similar to P. graminea and may eventually prove to be identical with it, or perhaps a subspecies. It is kept distinct here owing to the paucity of the material, and further collections are necessary before any definite conclusions can be reached. The main differences are the smaller flowers with pale-yellow, purple-veined (not pink, reddish or white becoming pink), much narrower petals, and the shorter leaves. Subsect. b. Thessalae Ball & Heywood Petrorhagia subsect. Thessalae Ball & Heywood, subsect. nov.; a subsectione Saxtfragis differt caulibus simplicibus vel sparse ramosis, floribus capitatis, bracteis late ovatis vel suborbicularibus. Type: P. thessala (Boiss.) Ball & Heywood. Stems simple or with few branches. Flowers capitate ; bracts broadly ovate or suborbicular, membranous and I-veined to brown-scarious and many-veined. 16. Petrorhagia thessala (Boiss.) Ball & Heywood, comb. nov. (Fig. 17.) Tunica thessala Boiss., Diagn. Pl. Or. Nov. 2 (8) : 63 (1849) ; Fl. Or. 1: 518 (1867). Gypsophila thessala (Boiss.) Nyman, Syll. Fl. Eur. : 238 (1854). Tunica orphanidesiana Clem. in Mem. R. Accad. Sci. Torino, Ser. 2, 16 : 17 (1855) ; Sertul. Or, 2 E7 (1855): Stems 10-35 cm. high, simple or slightly branched above, shortly papillose- pubescent at least at the base. Leaves up to 15 mm. long, linear. Heads up to 10-flowered ; outer bracts 6-many, the largest 6-10 mm. long, 2-5-4 mm. broad, about equalling the flowers, ovate, brown-scarious with white-membranous margin, rarely 5-7 mm. long, 1:5-2 mm. wide, ovate-lanceolate, and almost completely membranous except for the vein. Calyx 5-5-7 mm. long, glabrous or papillose with oblong, obtuse teeth. Petals 6-5-8 mm. long, linear-spathulate, entire, retuse or rarely crenate, white with pink or purple veins. Seeds 1-8-2-3 mm. long, I-3-1-7 mm. broad. REVISION OF THE GENUS PETRORHAGIA 157 Distribution : Greece ; doubtfully in southern Jugoslavia. GREECE : Séerea Ellas : Parnes, 24 Aug. 1911, Tunta 1279 (W-HAL) ; 28 June 1930, Guiol 1307 (BM) ; 17 July 1956, Rechinger fil. 18159 (W). Thessalia : Pelion, supra urbem Volo, 28 July 1893, Leonis, in Haldcsy, Iter Graec. Sec. (K ; W-HAL) Olympus, Aucher 537 (K, isotype). Olympi Thessaliae inter Hagios Elias et Litochoren, July 1857, Orphanides (BM ; CGE ; E ; K ; W-HAL). Hagios Elias prope Tyrnavo, 30 July 1885, Heldreich (W-HAL) ; 3 Aug. 1896, Sintenis 1106 (E; K; W-HAL). Milonna, 14 Aug. 1895, Formanek (W-HAL). Makedhonia : In monte Boz-dagh prope Serrae, ad monast. Prodromos, 15 July 1936, Rechinger fil. 10954 (W-RECH). Thraki: Near Oktchilar (Tochotai), Tedd 2041 (K). This species has a somewhat disjunct distribution, its main centre being Olimbos and neighbouring mountains, with apparently small populations on Parnis Oros, and scattered localities in Macedonia and Thrace. The plants from Tochotai (Oktchilar) differ somewhat from the rest of the species. The bracts of the inflorescence are smaller and almost completely hyaline. They approach P. macra and P. cyrenaica in this character, but otherwise resemble typical P. thessala. Records from Jugoslavia have not been confirmed, the specimens seen proving to be Petrorhagia saxifraga. However, the distribution in Greece is such that the species will probably be found to occur in both southern Jugoslavia and Bulgaria. ae eS a aes [7 18 Fics. 17,18. 17, Petrorhagia thessala (Boiss.) Ball & Heywood. 18, P. dianthoides (Smith) Ball & Heywood. a, calyx ; b, petal ; c, epicalyx bract or the largest bract from a head. (All x5.) 158 REVISION OF THE GENUS PETRORHAGIA 17. Petrorhagia cyrenaica (Durand & Barratte) Ball & Heywood, sp. nov. Tunica thessala var. cyrenaica Durand & Barratte, Fl. Lib. Prodr. : 33 (1910).—Maire, FI. Afr, Nord10 > 281 (1963). Stems 5-30 cm. high, glaucous, glabrous or papillose-scabrid at the base. Leaves up to 15 mm. long, linear to setaceous. Heads 1~-3-flowered ; outer bracts numerous, the largest 4-8 mm. long, 1:5-2:5 mm. broad, shorter than the flowers, ovate-lanceolate, white-membranous with brown-scarious veins. Calyx 4-5-6 mm. long, shortly and sparsely pubescent with oblong, obtuse teeth. Petals 5-5-8 mm. long, linear-spathulate, ?entire, white, pale-red on the lower surface. Seeds (immature) I-3-1-5 mm. long, 1:1 mm. broad. Distribution : Libya. LispyA. Zaouia Sidi-Mahdi entre Koubba et Lamloude, 13 May 1887, Taubert 559 (E, isotype). El Hania, 8 May 1934, Pampanini & Pichi-Sermolli 2577 (KK). Wadi Derna, 8 Apr. 1939, Sandwith 2487 (K). Very little material of this species has been seen, but it appears to be distinguish- able from P. thessala by its fewer-flowered heads with smaller flowers, and smaller almost entirely membranous bracts. Also, in the material seen, the calyx has always been sparsely pubescent, although a glabrous variant has been described by Pampanini (Arch. Bot. Forli 12: 24 (1936)). 18. Petrorhagia macra (Boiss. & Hausskn.) Ball & Heywood, comb. nov. Tunica macra Boiss. & Hausskn. in Boiss., Fl. Or., Suppl. : 81 (1888). Tunica gracilis F. N. Williams in Journ. of Bot. 28 : 196 (1890). Stems 25-40 cm. high, slightly branched above, pubescent at the base, glabrous elsewhere. Leaves up to 10 mm. long, oblong-linear to subulate. Heads 1-5- flowered ; outer bracts 4-many, the largest 3-5-5 mm. long, 0-7-1:3 mm. broad, triangular-lanceolate to ovate, acuminate, white-membranous with a brown-scarious vein. Calyx 5-5-8 mm. long, sparsely pubescent with more or less oblong-obtuse teeth. Petals 8-9 mm. long, linear-spathulate, entire, white. Seeds not known. Distribution : Iran (Kurdistan). IRAN: Monte Shahu, 1867, Haussknecht (K, isotype, also holotype of Tunica gracilis). Very similar to P. thessala and P. cyrenaica. Tunica gracilis was described from isotype material of T. macra. The specimen closely fits the description of T. macra, and there does not seem to be any reason why it should be regarded as a distinct species. 19. Petrorhagia dianthoides (Smith) Ball & Heywood, comb. nov. (Fig. 18.) Gypsophila dianthoides Smith, Fl. Graec. Prodr. 1 : 280 (1809). Tunica dianthoides (Smith) Fisch. & Meyer, Index Sem. Hort. Bot. Petrop. 4 : 48 (1837).— Boiss., Fl. Or. r : 518 (1867). Kohlrauschia sibthorpit Kunth, Fl. Berol. Fam. Nat. Dispos. 1 : 109 (1838), nom. illegit. Stems up to 40 cm. high, simple or slightly branched above, glabrous. Leaves up REVISION OF THE GENUS PETRORHAGIA 159 to 20 mm. long, linear to subulate. Heads 1-6(-8)-flowered ; outer bracts 4—many, the largest 4:5-10 mm. long, I-2:5 mm. broad, lanceolate to ovate-lanceolate, white- membranous except for the brown-scarious vein. Calyx 5~7 mm. long, pubescent, with triangular-lanceolate, more or less acute teeth. Petals 6-10 mm. long, oblong- spathulate, entire, white with red veins. Seeds (immature) I-1-3 mm. long, o-8—o-9 mm. broad. Distribution : Crete. CRETE : Mont. Sphacioticorum, July 1846, Heldreich (BM; CGE; E; K; W-HAL ; isotypes). Hag. Roumeli et Samaria, dist. Sphakia, 12-13 July 1899, Baldacci 21m (BM ; W-HAL) ; June 1932, Guwiol 2125 (BM). Aradhena, 24 July 1882, Spreitzenhofer (W-HAL) ; July 1932, Atchley 1694 (K). Paleocastis, Sitia, May 1852, Heldreich (E). Theriso gorge near Cnea, 22 July 1938, Davis 303 (BM ; E ; K). Imbros gorge, 25 May 1938, Davis 185 (E). This species appears to be one of the most distinct in the section, Sect. 4. DIANTHELLA (Boiss.) Ball & Heywood Petrorhagia sect. Dianthella (Boiss.) Ball & Heywood, comb. nov. Tunica sect. Dianthella Boiss., Fl. Or. 1 : 517 (1867). Type : Tunica pamphylica Boiss. & Balansa (= P. pamphylica (Boiss. & Balansa) Ball & Heywood). Annual. Leaves 3-veined. Flowers solitary or subcapitate ; bracts ovate, brown- or purplish-scarious, 3-many-veined. Calyx costae 3-7-veined. Petals not clawed. Seeds small, blackish-brown, reticulate-tuberculate, with thickened margin. 20. Petrorhagia pamphylica (Boiss. & Balansa) Ball & Heywood, comb. nov. (Fig. 19.) Tunica pamphylica Boiss. & Balansa in Boiss., Diagn. Pl. Or. Nov. 3 (6) : 27 (1859).— Boiss, Fly Ort = 517-(1867), Stems 10-35 cm. high, pseudodichotomously branched from the base, glabrous. Cauline leaves up to 16 mm. long, I mm. broad, linear, 3-5-veined. Flowers solitary ; epicalyx bracts 6-12, 3-4 mm. long, shorter than the calyx, oblong to ovate, acute, brown- or purplish-scarious with white-membranous margin and with 5-many thick veins. Calyx 5-7 mm. long ; costae very broad, 6~7-veined, the commisural membranes almost absent ; teeth o-8-1-1 mm. broad, lanceolate- acuminate. Petals 7-9 mm. long, the limb c. 0-5 mm. broad, linear-spathulate, obtuse, entire, pale-pink. Seeds 1-6-2 mm. long, 1-2-1-7 mm. broad. Distribution : Turkey (southern Anatolia). Turkey : Antalya, near Gebiz, stony hillside, 1,000 m., 22 July 1949, Davis 15478 (BM s Bue). 21. Petrorhagia peroninii (Boiss.) Ball & Heywood, comb. nov. (Fig. 20 ; Plate 14 C.) Tunica peroninii Boiss., Fl. Or., Suppl. : 81 (1888) (“ Peronini’’). 160 REVISION OF THE GENUS PETRORHAGIA Stems 15-30 cm. high, pseudodichotomously branched from the base, pruinose- papillose. Cauline leaves 8-20 mm. long, c. 1 mm. broad, linear, 3-veined. Flowers solitary or subcapitate ; bracts and epicalyx bracts 6—-many, up to 5-7-5 mm. long, exceeding the calyx, lanceolate, brown- or purplish-scarious with white-membranous margin, papillose, 3(—5)-veined. Calyx 5-7 mm. long ; costae 3-veined, with a narrow but distinct commissural membrane ; teeth 0-4-0-7 mm. broad, oblong- lanceolate. Petals 5:5-8-5 mm. long, the limb c. 0-5 mm. broad, linear-spathulate, obtuse, entire, white, sometimes purplish on the lower surface. Seeds I-I-1-:2 mm. long, 0-7-0-9 mm. broad. -- es Fania) 19 20 19, Petrorhagia pamphylica (Boiss. & Balansa) Ball & Heywood. 20, FIGs. I9, 20. a, calyx ; b, petal ; c, epicalyx bract. (All x5.) P. peroninii (Boiss.) Ball & Heywood. Distribution : Turkey (southern Anatolia). TuRKEY : Prov. Mersin, Dist. Anamur, montagne ouest d’Anamour, June 1872, Peronin 5 (BM; K> isotypes). Between Ferhenk and Anamur, rocky meta- morphic hills, 50 m., 19 Aug. 1949, Davis 16329 (BM; E; K). These two rare species, as already mentioned (p.126), appear in some respects to form a link between Petrorhagia and Dianthus. Sect. 5. KOonLRAUSCHIA (Kunth) Ball & Heywood Petrorhagia sect. Kohlrauschia (Kunth) Ball & Heywood, comb. nov. Kohlrauschia Kunth, Fl. Berol. Fam. Nat. Dispos. 1 : 108 (1838). Dianthus sect. Kohlrauschia (Kunth) Fenzl in Endl., Gen. Pl. : 971 (1840). Tunica sect. Kohlrauschia (Kunth) Griseb., Spicil. Fl. Rumel. 1 : 185 (1843). Dianthus subgen. Proliferastrum F. N. Williams in Journ. of Bot. 23 : 347 (1885). Type : Kohlrauschia prolifera (L.) Kunth (= P. prolifera (L.) Ball & Heywood). Annual. Leaves 3-veined. Flowers capitate ; bracts very broad, orbicular- ovate or suborbicular, brown-scarious, many-veined. Petals distinctly clawed. Seeds small, blackish-brown, reticulate to papillose, with thickened margin. REVISION OF THE GENUS PETRORHAGIA 161 22. Petrorhagia prolifera (L.) Ball & Heywood, comb. nov. (Plate 15 A.) Dianthus prolifer L., Sp. Pl. 1: 410 (1753). Dianthus diminutus L., op. cit., ed. 2, I : 587 (1762). Tunica prolifera (L.) Scop., Fl. Carniol., ed. 2, 1 : 299 (1772). Caryophyllus aridus Moench, Meth. Pl.: 59 (1794), nom. illegit. Kohlrauschia prolifera (L.) Kunth, Fl. Berol. Fam. Nat. Dispos. 1 : 10g (1838). Kohlrauschia diminuta (L.) Reichb., Ic. Fl. Germ. & Helv. 6: 43 (1844). Stems up to 50 cm. high, usually simple, glabrous or scabrid. Cauline leaves up to 40 mm. long and 2 mm. broad, linear-oblong, 3-veined ; sheath 1-2-5 mm. long, about as long as broad. Largest bracts of head 6-12 mm. long, 3~7(—8) mm. broad, usually obtuse. Calyx 10-13 mm. long ; costae 3—veined ; teeth oblong, obtuse. Petals 10-14 mm. long, the limb 2-3-5 mm. wide, obcordate, pink or purplish. Seeds I-3-I'9 mm. long, 0-8—1-I mm. broad, reticulate. Distribution : Central Europe, mountains of southern Europe, Caucasus, Turkey (northern Anatolia), mountains of western North Africa. SWEDEN : Paroecia Garde, Petard, Gotland, 30 June 1926, Samuelsson 797 (BM ; W). Prope Resmo, Oland, 15 July 1860, July 1864, Ahlberg (BM ; CGE; K). DEnMARK : Ins. Samsoe, 28 July 1887, Jensen (W-HAL). Ins. Bornholm, July 1848, Lange (CGE). Rorvig, Aug. 1866, Steenhiip (E). GERMANY : Sandhausen, 8 km. S. of Heidelberg Baden, 1 Aug. 1935, Stearn 70 (K). Bei Mannheim, 2 Aug. 1859, Schweinfurth (BM). Frankfurt, 24 June 1856, Bagge (BM). Bonn, 16 June 1851, Blackie (E). Schleswig-Holstein, June-July, Hansen 347 (BM). Calshalle, Kreuznach, 2 Sept. 1868, Tvimen (BM). S. von Aschaffenburg (Bavaria), July 1905, Rubner (E). Prope K1.-Kreutz, Brandenburg, Aug. 1922, Gort (K). PoLanp : Krappilk (Schlesien), 19 July 1899, Ziesché (E). CZECHOSLOVAKIA : Litomefice, 17 Sept. 1918, Korb (W). Pressburg, 7 Sept. 1913, Korb (W). Westfusse der Kahlberges (B.-Leipa), Aug. 1910, Petrak 330 (BM ; E). Prope Veveri Bytyska, Moravia occid., Aug. 1924, Sivjaev 26 (K). AustrRiA : Tirol: Boyen, Grabmayr (K). Nieder-Osterreich : Wiener-Neustadt, Kerner, Fl. Austro-Hung. 2509 (BM; K; W-HAL; W-RECH). Prope Wien, Aug. 1878, Halécsy (W-HAL). Baden, 18 July 1900, Krebs (E). Burgenland : Leithagebirge, Naglersberge, 21 June 1923, Korb (W). Am Eisenbahndamm zwischen Weiden und Neusiedler See, 21 Sept. 1924, Rechinger (W-RECH). Huncary : Ins. Csepel, 2 July 1913, Filarszky & Kiimmerle, Fl. Hung. Cent. IT, 90 (BM; E; K; W-RECH). Pest, June 1872, Richter (BM). Bei Magyarfalva, 24 June 1905, Vetter (W). RoMANIA : Baile Herculane, Caras Severin, 13 June 1922, Bujorean, Fl. Rom. 4or (BM ; E; K). Carmen Sylva, Constanta, 14 Aug. 1931, Enculescu, Fl. Rom. 4o1b (K). BULGARIA : Ladovo, 8 June 1898, Sivibrny (E). Dragoman, mont. Golem Cepan, 31 July 1930, Rechinger fil. (W-RECH). Djenden Tepe, bei Philipopel, 25 July 162 REVISION OF THE GENUS PETRORHAGIA 1891, Gheorghieff (W-HAL). Dolni Pasarel, 6 July 1957, Newman (K). Prope Varna, 16 July 1930, Rechinger fil. 1021 (W-RECH). Bei Tirnorvo, 1896, Urumoff 8r (W-HAL). Bei Lowtscha, 1895, Urumoff (W-HAL). Supra urbem Kolarov- grad, 9 July 1951, Zelezova, Pl. Bulgaricae Exs. 25 (K). Gorne Banja, 20 June 1921, Uromov (K). Plovdiv, 20 July 1930, Rechinger fil. 1165 (W-RECH). Near R. Topolnitza, 36 km. from Pazardzhik, 16 July 1950, Mason 27 (K). Prope Lakatnik, Stara Planina, 29 July 1930, Rechinger fil. 1760 (W-RECH). Deli Orman prope Russe (Rustschuk), 13 July 1930, Rechinger fil. 747 (W-RECH). JUGOSLAVIA : Makedonija: Mte. Ohrida, 1908, Dimonie (W-HAL). Veles, Ig June 1905, Adamovié (E). Bei RaduSe, 11 July 1937, Behy (W-RECH). Scardus geb., July 1876, Hertz (K). Gjefgjeli, 1-2 June 1927, Rechinger fil. 1577 (W-RECH). Bosna 1 Hercegovina: M. Bokievria prope Astrozai, 4 July 1926, Maly (K). Moscanica, 4 July 1933, Guilliat-Smith 3266 (K). Mostar, 8 July 1955, Hepper 870 (K). Srbija: Prope Vranja, 18 June 1900, Adamovié (E). S. of Streocha (Streoci), Prokletije, 18 July 1956, Carpenter 340 (K). Visoki Decani, prope Peé, 17-19 July 1933, Rechinger fil. & Scheffer 1070 (W-RECH). Fluv. Pecska Bistrica prope Pe¢, 3 July 1933, Rechinger fil. & Scheffer 121 (W-RECH). Dalmacya : Makarska, 1934, Jozipovié (K). Ragusam (Dubrovnik), June 1907, Adamovié (K). Napoda, Kriodsijc, 6 June 1906, Schneider (W). Popovici (Tepui Dol), 20 July 1926, Lorchuigg (W-RECH). Hrvatska : Vernde, Istria, 17 May 1875, Freyn (K). ALBANIA : Sarandé, 10 July 1932, Alston & Sandwith (BM). Mali Dajtit near Tirana, 9 Aug. 1938, Hepburn 26 (K). Near Richa, 4 July, Giuseppi (K). Selce, dist. Klementi, 2 Aug. 1901, Baldacci 192 (BM). GREECE : Sterea Ellas : Mavrolitharion to Mousonitsa (Oiti), 14 June 1937, Balls & Gourlay 3292 (BM; K). Ad Paliampela (Akarnania), 22 July 1889, Baldacci (K ; W-HAL). Ifivos: Janina, Sept.—Oct. 1895, Baldacci (BM); July 1931, Guiol (BM). Prope Dramissous, June-July 1929, Guiol (BM). Mt. Peristeri prope Syraka, 14 July 1893, Haldcsy (W-HAL) ; Aug. 1931, Atchley 992 (K). M. Murga (Tzoumerka), 6 July 1895, Baldacci 15 (K). Montes Pindus, prope Metsovon, 28 July 1956, Rechinger fil. 18298 (LIVU). Makedhonia: Lacus Dojran, 6 June 1936, Rechinger fil. 9147 (W-RECH). Boz-dagh prope Serrai, 15 July 1936, Rechinger fil. 11050 (W-RECH). Samarina to Perevali, 14 July 1937, Balls & Gourlay 3609 (E ; K). Near Thessaloniki, 19 Sept. 1917, Ramsbottom (BM). _ S. slopes of Krousia Oros, 30 June 1917, Turrill (BM). Thraki: Fluv. Nestos (Mesta) prope Toxotai, 12 June 1936, Rechinger fil. 9279 (W-RECH). Prope Alexandroupolis, 29 May 1934, Rechinger 6054 (W-RECH). Xanthi, 23 June 1931, Tedd 666 (K). Derecol valley near Xanthi, 4 Oct. 1929, Tedd 70 (K). SWITZERLAND : Between Sierre and Vissoie, Valais, 3 Aug. 1903, Lester-Garland (K). Prope Geneva, Boissier (K). Martigny, 27 Aug. 1858, Fraser (E). Lausanne, June 1879, Favrat & Barbey (K). Argau, Unteres Aaretal, 30 June 1954, Koch (W). Cressier, 10 June 1834, Neuchatch (E). ItaLy : Basilicata: Pignola, 8 June 1924, Gaviolo (FI). Gruppo del Pollino, REVISION OF THE GENUS PETRORHAGIA 163 19 July 1935, Gavioli (FI). Luconia, M. San Michele, 2 July 1932, Gavioli (FI). Campania : Monte Alburno, Da Corleto M. a C. Sierre, 12 Aug. 1936, Philippis (FI). Presso Caserta, June 1870, Tenace (FI). Lazio : Tivoli, Sul Monte Catillo, 12 June 1904, Vaccari (FI). Monte Albani, Nemi, 20 June 1875, Marchesetit (FI). Montero- tondo (Roma), 22 Oct. 1950, Steinberg (FI). Umbria : Stroncone, 22 Sept. 1893, Mari (FI). Pemgin, 30 Oct. 1947, Bettini (FI). Venezia: Mont. Ricco prope Monselice, May 1911, Béguinot 1657 (BM ; FI; K). Verona, 27 July 1888, Caruel (FI). Romagnano, Aug. 1888, Goivan (FI). Tirolo Italiano: S. Leonardo in Passiria, Alto Adige, 15 Aug. 1949, Zenari (FI). Cembra, 20 Sept. 1948, Whitehead (K). Emilia: M. Titano, San Marino, 25 June 1912 and 15 Oct. 1916, Pampanini (FI). Osservanza, presso Bologna, June 1900, Ceroni (FI). Spezzano, 19 July 1882, Fiori (FI). M. Chiarugola, Parma, 18 Aug. 1919, Mimo (FI). Marche: Mt. Fortino, Apennini Picenim, 1836, Marzialetti (FI). Mt. Catria, 23 July 1866, Parlatore (FI). Nevola e Mifr, 4 July 1934, Bettina (FI). Puglia : prope Gallipoli, June 1883, Groves (FI). Abbruzzo : Monte Morrone, 1891, Profeta (FI). Majella, 30 June 1905, Vaccari (FI). Prope Gagliano, Sirente, July 1876, Groves (FI), Lombardia: Malgrate, 1 Oct. 1893, Camperio (FI). Grumone, 3 Sept. 1864. Parlatore (¥F1). SicILy : Sagana, 12 Aug. 1836, Parlatore (FI). Etnae, Sept. 1844, Tornabene (FI). Maudenici, 4 June 1882, Borzi (FI). Corsica: Bastia, 28 Sept. 1880, Chubern (FI). Bastelica, 20 June 1878, Reverchon (K). FRANCE : Muerthe-et-Moselle : Nancy, 1871, Gamble (K). Jura: Serre, Sept. 1866, Blanche. Savoie: Prés Albens, 28 Sept. 1881, anon. (K). Hautes-Alpes : Gap, 3 July 1871, Reverchon (K). Briancon, 18 July 1871, Reverchon (K). Embruin, Sept. 1887, Gamble 19660 (K). Iséve : Pont de Beauvois, Sept. 1860, anon. (K). Grenoble, June 1840, Haden (BM). Puy de Dime: Puy de Montaudon, 16 Aug. 1927, Burtt-Davy 167/27 (K). Rhéne: Lyon, 1852, Jordan (K). Céte d’Or: Dijon, 1 Aug. 1894, Clarke 47629 (K). Somme: St. Valery-sur-Somme, 10 July 1959, Carter (K). Niévre : Nevers, Aug. 1840, J. Ball (E). Eure: Andelys, Aug. 1844, Mill (BM ; K). Orne: Alencon, 28 June 1886, Beaudoin (LIVU). Loire-Inférieure : Nantes, 1897, Blow (E). Seine-et-Oise: Point du Jour, prés Paris, 4 July 1875, Bonnet (K). Seine-et-Marne : Fontainbleau, Henslow (E). Haute-Garonne : Toulouse, 25 July 1952, Billot 2027 (BM ; LIVU). Hérault : W. of Montpellier, 14 Sept. 1953, Rawden 48 (K). Pyrénées-Orientales : Verneb, 15 July 1934, Wyatt (K). Hautes-Pyrénées : Gedre, July 1870, Bordére (K). Basses-Pyrénées : Laruns, 28 July 1922, Hutchinson, Matthews & Riley 290 (K). BELGIUM : Mariembourg, 23 June 1935, Mosserony (K). LuxemBourc : N. du Hobscheid, 18 June 1953, Reichling (BM). Eich, 23 Sept. 1925, Rundberg (W-RECH). Spain : Ur, Cerdagne, 21 Aug. 1956, Hooper (K). San Mateo (Castellén de la Plana), 1954, Rechinger fil. (W). 164 REVISION OF THE GENUS PETRORHAGIA Morocco: Fimelil, Atlas, 12 June 1926, Lindberg 3880 (K). Ras-el-Ma, prope Axron, 1926, Lindberg 4293 (W). U.S.S.R. : Crimea, 1855, Munro (K). Bei Sudak, 13 July 1896, Callier 378 (E ; W-HAL ; W-RECH). Lagodechi, fluv. Lagodech, Caucasus Kachetia, 17 July 1898, Tomin (E). TurKEY : Adana, Feke, Sencan Deresi between Giiriimze and Siiphandere, I July 1952, Davis 19603 (K). Adapazari, Arifiye, 1 July 1962, Davis 36281 (E). Introduced : GREAT BRITAIN : Sussex : Lewes, 1902, Davy (BM). Norfolk : near Northwold, 19 Sept. 1927, Little (BM). The typification and relationships of this species are discussed by Ball & Heywood (1962). 23. Petrorhagia nanteuilii (Burnat) Ball & Heywood, comb. nov. (Fig. 21 ; Plate 15 B.) Dianthus nanteuilii Burnat, Fl. Alp. Marit. 1 : 221 (1892). Tunica nanteuilii (Burnat) Giirke in K. Richt., Pl. Eur. 2 : 338 (1903). Tunica prolifera var. nanteuilit (Burnat) Briq., Prodr. Fl. Corse 1 : 569 (1910). Tunica prolifera subsp. nanteutlii (Burnat) Graebn. in Aschers. & Graebn., Synops. Mitteleur. El. 5 :(2) = :264: (t921): Kohlrauschia nanteuilii (Burnat) Ball & Heywood in Watsonia 5 : 115 (1962). Very similar to P. prolifera but the stems sometimes pubescent in the middle ; leaf sheaths 1-3-5 mm. long, sometimes up to twice as long as broad ; seeds tuberculate. Distribution : Western Europe and western North Africa. GREAT BRITAIN: Isle of Wight: Douvre Common, Aug. 1841, C. A. Wright (BM). Ryde, 6 July 1852, More (BM). Hampshire : Hayling Island, 18 Aug. 1922, Foggitt (BM). Portsea Island, 1890, Snope (LIVU). Sussex : Selsey, 1837, Forster (BM). E. of Pagham, 21 June 1914, Salmon (BM). Pagham, 25 June and 29 July 1901, Marshall (LIVU). Kent: Dungeness, July 1903, Chandler (BM). Hythe, Aug. 1903, Riddelsdell (BM). Norfolk: Stoke Ferry, July 1890, Aug. 1891, Cross (BM). CHANNEL ISLANDS: Jersey: Quenvais, 14 July 1901, Lester-Garland (LIVU). St. Ouens Bay, 14 June 1862, Briggs (LIVU). St. Brelade’s Bay, 8 June 1884, Hanbury (LIVU). Guernsey : C. 1858, Brewer (BM). FRANCE : Iséve : South of St. Marcellin, 12 Sept. 1953, Rawdon 20 (K). Coétes du Nord: St. Lunaire, 1883, Lester-Garland (K). Loire Inférieure : Pointe de Chemoulin, 1871, Genevier (BM ; CGE). Lot: Prés de Cahors, June 1855, Bullot 2027b (BM ; CGE ; LIVU). Guronde : Bordeaux, 1896-97, Neyraut (K ; W-HAL). Hautes-Pyrénées : Luz St. Sauveur, 18 Aug. 1932, Meinertzhagen (BM). Prope Cauterets, July 1845, Spruce (CGE ; K). Corsica : Prés Bastia, 17 June 1881, Chubern (FI). SARDINIA : Isola Maddalena, June 1893, Vaccari (FI). REVISION OF THE GENUS PETRORHAGIA 165 SPAIN : Versus Vallvidrena, prope Barcelone, June 1876, Tvemols (W-HAL). S. Juan de Alcaraz, 16 June 1860, Bourgeau 963 (CGE ; E; K). Bayona, Ponte- vedra, 20 June 1935, Schrafer (BM). Madrid, June 1919, Sennen 3662 (K). S. Guadarrama, 5 June 1933, Afchley 123 (K). Above Cercedilla, S. Guadarrama, 7 June 1926, Wilmott (BM). Alcantara Bridge, Toledo, 6 June 1926, Wilmott (BM). Despefiaperros, 2 June 1927, Wilmott & Lacaita (BM). Venta de Cardenas, 11 June 1952, Rotvainen (K). Slopes of Yelmo Grande, Sierra de Segura, 22 June 1955, Heywood 2802 (LIVU). El Campillo, Sierra de Segura, 21 June 1955, Heywood 2670 (LIVU). Barranco de la Cueva de los Cervales, S. Magina, 6 June 1925, Cuatrecasas 5311 (K). Minas de Beires, S. Nevada, 20 June 1926, Wilmott & Lofthouse (BM). Supra Guejar, S. Nevada, 22 Aug. 1844, Willkomm 136 (K). Prados del Aire, S. Nevada, 10 June 1926, Wilmott & Lofthouse (BM). Ronda, 7 May 1924, Ellman & Hubbard 231 (K). Gibraltar, 5 June 1913, Wolley-Dod 2113 (K). Guadacorte, 17 May 1912, Wolley-Dod 1060 (BM). Prope San Roque, 16 June 1871, J. Ball (K). Almendral (Cadiz), 18 May 1957, Taylor (K). 2| 22 Fics. 21, 22. 21, Petrorhagia nanteuilii (Burnat) Ball & Heywood. 22, P. glumacea (Bory & Chaub.) Ball & Heywood var. glumacea. a, calyx ; b, petal ; c, epicalyx bract or the largest bract from a head. (All x5.) PorTUGAL : Coimbra, 5 July 1887, Murray (BM). Estacao Velha, Coimbra, 3 June 1952, Matos (BM). Peneda de Saudade, Coimbra, June 1886, Moller (W-HAL). Near Lisbon, 1935, Scarlett 180 (K). Moita, June 1840, Welwitsch 222 (CGE ; K). Tapada da Ajuda, 5 June 1933, Atchley 256 (K) ; 3 May 1940, P. Silva 166 REVISION OF THE GENUS PETRORHAGIA (K). Villa Nova de Rainha, June 1846, Welwitsch (BM; K). Near Oporto, 16 July 1887, Murray (BM). Caldar de Gerez, June 1888, Tait (BM). Morocco: Tichka l’Agadal, 10 June 1936, Balls 2766 (BM; K). Amizmiz, 17 June 1936, Balls 2851 (BM ; K). MADEIRA: Madeira, Lindley (K) ; July 1862, Clarke (K). Funchal, 27 Mar. 1949, Sledge (BM). Bay E. of Pont Gorda, 27 Apr. 1924, Riley (K). CANARY IsLANDS: Tenerife: Teneriffae, Webb (K). Cuevas Negras, 27 Apr. 1845, Bourgeau 709 (K). Santa Ursula, 5 Mar. 1933, Asplund rog (K). Montana de la Horca Orotava, 26 May 1913, Sprague & Hutchinson 4o (K). Barranco del Carmen, 31 May and 1 June 1913, Sprague & Hutchinson 163 (K), 170 (K). La Palma: El Paso, 10 June 1913, Sprague & Hutchinson 303 (K). For full discussion of the separation of this species from P. prolifera see Ball & Heywood (1962). 24. Petrorhagia velutina (Guss.) Ball & Heywood, comb. nov. (Plate 15 Cc.) Dianthus velutinus Guss., Index Sem. Hort. Bocc. 1825 : 2 (1825). Tunica velutina (Guss.) Fisch. & Meyer, Index Sem. Hort. Petrop. 6 : 66 (1839). Gypsophila velutina (Guss.) D. Dietr., Synops. Pl. 2: 1542 (1840). Kohlrauschia velutina (Guss.) Reichb., Ic. Fl. Germ. & Helv. 4: 43 (1844). Stems up to 50 cm. high, usually simple, usually with crowded glandular hairs on the middle internodes, rarely almost glabrous. Cauline leaves up to 35 mm. long and 2mm. broad, linear-oblong, 3-veined ; sheath 3-7(-8) mm. long, at least twice as long as broad. Largest bracts of head 7-12 mm. long, 4:5-8-5 mm. broad, mucronate. Calyx 8-14 mm. long ; costae 3-veined ; teeth oblong, obtuse. Petals 11-16 mm. long, the limb 1-2-2-5 mm. broad, obcordate, sometimes bifid, pink or purplish. Seeds 1-1-3 mm. long, 0-7-0-8 mm. broad, strongly tuberculate or papillose. Distribution : Mediterranean region. PorTuGAL : Alemtejo, Redondo, May 1891, Moller 1086 (W-HAL). Prope Villa Nova, 1848-50, Welwitsch 1158 (BM). SPAIN : Vega de Granada, June 1864, Campo (BM). Sierra de Chiva, May 1844, Willkomm (BM). Barranco de Ohanes, Sierra Nevada, 21 June 1926, Wilmott & Lofthouse (BM). Near Venta de Saladillo (Gibraltar), 15 Apr. 1912, Wolley-Dod 651 (BM). Almendral (Cadiz), 18 May 1957, Taylor (K). FRANCE : Alpes-Maritimes: Prés Nice, 7 May 1861, Bourgeau (FI). Esteril, 7 May 1889, Bonafous & Vidal (K). Corsica : Pentes du Pigno, a Bastia, 20 June 1865, Mabille, Herb. Cors. 73 (BM ; CGE; FI; K). Bastelica, 20 June 1878, Reverchon (BM ; K ; W-HAL). Is. Sanguinare, 25 May 1848, Reguien (K). Ajaccio, Apr. 1899, Collett (K). Near Corte, Porter 176 (BM). Near Porto, 22 Apr. 1928, Edwards (BM). Travo, 29 Apr. 1930, Bates (BM). Iraty: Liguria: Pizzo, Apr.-May, Ricca (FI). Emilia: Bologna, Monti REVISION OF THE GENUS PETRORHAGIA 167 Sabbione, 11 June 1873, Marchesetti (FI). Presso Modena, 27 June 1808, Fiori (FI). Marche: Flora Picena, Orsini (FI). Toscana: Grosseto, 9 Apr. 1843, Parlatore (FI). Isola del Giglio, 24 Apr. 1895, Sommer (FI). Careggi, 13 May 1866, Bucci (FI). Pisa, 15 May 1870, Groves (FI; K) ; May 1856, Savi in Billot 2028 (BM). Firenze, 29 Apr. 1874, Leviery (CGE ; FI; K; W-HAL). S. Vincenzo, 2 May 1868, Parlatore (FI). Presso Buggiane, 1 Apr. 1933, Picht-Sermolli (FI). Colli di Terontola (Arezzo), 20 May 1916, Fiori (FI). Lazio: Lago d’Agnano, 1g Apr. 1841, Heldreicch (CGE; FI; K). Civita Castellana, 11 May 1847, Bentham (K). Roma, 17 May 1903, Vaccari (FI). Campania: Licola, 8 Apr. 1955, Agostini (FI). Monte Alburno, 29 May 1951, Philippis & Moggi (FI). Napoli, May 1889, Guadagno (FI). Mts. di Agerola, 15 May 1911, Guadagno (FI ; W-HAL). Between Amalfi and Salerno, June 1883, Lacaita (K). Lipari, Apr. 1902, Zodda (FI). Insula Caprearum, 8 May 1878, Heldreich (W-HAL). Puglia: Monte Calvo (Gargano), 27 May 1893, Martelli (FI). Mater Grazia, Leucarpide, Apr. 1877, Profeta (FI). Basilicata : Potenza, 13 June 1926, Gavioli (FI). Calabria : Stilo et Mongiana, 18 June 1877, Biondi (FI). Prope Pezzo, 3 May 1877, Biondi (FI). Rossano, 4 May 1872, Aint: (FI). Palene, 20 May 1877, Arcangeli (FI). Serra San Bruno, 30 Apr. 1884, Zwierlein (FI). SARDINIA : Capo Carlo, 12 Apr. 1894, Martelli (FI). Mte. Genargentu, 21-22 June 1897, Martelli (FI). Isola dei Cavoli, 12 Apr. 1894, Martelli (FI). Isola Maddalena, Apr. 1893, Vaccari (FI). Tempio, 23 May 1881, 3 June 1882, Reverchon 68 (E ; FI; K; W-HAL). Orune, May 1899, Martells (FI). Sassari, 7 May 1895, Martelli (FI). Cagliare, Mar. 1828, Thomas (K). SiciLy : Madonie, Castelvetrano, 1842, Parlatore (FI). Palermo (Panormitani), May, Todaro 437 (BM; CGE; FI; K; W-HAL). Chiani di la Penna, June, Mina 330 (FI). Piana del Greci, 21 May 1855, Huet du Pavillon (BM ; CGE ; FI; K). Messina, Canepo, 1865, Seguenza (FI). Pantelleria, Apr. 1890, Ross (FI). Castanea, May 1867, Mallandrino (FI). JUGOSLAVIA : Hrvatska: Pola (Pula), 7 June 1874, Freyn (FI). Promontore (Istria), 23 May 1898, Marchesetti (FI). Fasana, 10 June 1909, Korb (W). ALBANIA : Sarande, 3 June 1933, Alston & Sandwith 1372 (BM ; K). Levani, 23 Apr. 1918, Schneider (W). Ljusme, 1 May 1918, Schneider (W). Burearia : S. of Varna, 20 May 1925, Gilliat-Smith 1134 (K). GREECE : Thraki: Dedeagach (Alexandroupolis), 7 May 1891, Sintenis & Bornmiiller 75 (K ; W-HAL). Porto Lago, 17 Apr. 1933, Tedd 939 (K). Near Xanthi, 21 May 1929, Tedd 8 (K). Makedhonia : Cassandra, 4 May 1891, Abd-ur- Rahman Nadji (W-HAL). Simopetra, Athos, 13 Apr. 1934, Hill, Sandwith & Turrill 2337 (K). Karadagh inter Thessaloniki et Serrai, 8 June 1936, Rechinger fil. 9214 (W-RECH). Near Snevce, Apr.-May 1918, Cook (BM). Between Kalamaria and Mikra Bay, Mar. 1918, Ramsbottom (BM). Dora Tepe, E.N.E. of Doiran, 1918, Gooding (BM). Antartikon, 30 May 1932, Alston G& Sandwith 59 (BM; K). Vermion prope Naoussa, 30 May-1 June 1936, Rechinger fil. 8757 (W-RECH). Thessalia: Kalampaka, 20 May 1896, Sintenis 251 (W-HAL). Sterea Ellas : 168 REVISION OF THE GENUS PETRORHAGIA Korax prope Musinitza, 5 June 1899, Dérfler 283 (W-HAL). Agrapha, circa monasterium Korona, 20-28 June 1885, Heldreich (W-HAL). Porto Rafti, May 1929, Guiol (BM). Parnes, prope Driza, Apr. 1929, Guiol (BM). Hymettus, 25 Apr. 1878, 1884, Heldreich (E ; W-HAL). Evvoia: Mt. Ocha, 23 May 1955, Rechinger fil. 16291 (W). Peloponnisos : Messenia, May 1844, Heldreich (BM). Leontari, May 1844, Heldreich (BM). In ruinis Olympiae, Elis, 12 Apr. 1911, Haldcsy (W-HAL). Aegean Islands: Skyros, Apr. 1908, Tunta 663 (W-HAL). Hydra, 3-10 Apr. 1876, Heldretch (W-HAL). Melos, Mar. 1889, Heldreich & Haldcsy (BM ; K ; W-HAL). Tenos, Apr. 1889, Heldreich & Haldcsy (K ; W-HAL). Karpathos, Monte Lasto, 13 Mar. 1883, Pichler (BM ; K). Mytilini, Ajassos, 19 May 1934, Rechinger fil. 5638 (W-RECH). Chios, 14-15 Apr. 1939, Platt 126 (K). Samos, monte Kerki, 15 Apr. 1934, Rechinger fil. 4o40 (BM ; W-RECH). Kalimnos, 13 Apr. 1887, Major 676 (E). CRETE : Viano prope Christos, May 1g00, Leonis in Dérfler goto (BM; E; W-HAL). Hag. Vasilis, 9 May 1904, Dérfler goo (W-RECH). Lassithi, Krystal- lenia, 23 May 1914, Gandoger 2027 (K). Kissamos, Polyrrhenia, 6 Apr. 1915, Gandoger 8744 (K). Canee (Khania), 8 Apr. 1914, Gandoger 2792 (K). TURKEY : Messarburnum bei Konstantinopel, 23 May 1897, Nemetz (W-RECH). Smyrnae, 1827, Fleischer (E). Marmaris (Mugla), 24 Mar. 1956, Davis 25304 (K). Taktali dag (Antalya), 26 May 1906, Bornmiiller gog5 (K). Sencan Deresi between Girtimze and Siiphandere, Feke (Andana), 1 July 1952, Davis 19636 (K). Anamur (Mersin), 14 Apr. 1956, Davis 25954 (K). Erzerum, Zorab 137 (K). Montes Amanus, 1906, Haradjian 218 (W). Cyprus : Kyrenia, 20 Mar. 1956, Atherton 1194 (K). West of Orga, 30 Mar. 1956, Merton 2588 (K). Above Defetra, 27 Mar. 1952, Probyn 62 (K). Larnaca, 7-20 Apr. 1912, Haradjian 28 (K). Ayios Philon, near Rizokarpeso, 19 Feb. 1941, Davis 2257 (K). Platres, 18 June 1938, Kennedy r1gr (K). Ayia Grini (Morphan), 11 Mar. 1941, Davis 2538 (E). U.S.S.R.: Azerbaydzhan, Plantae Bakuenses, Ins. Swatoi, 20 May 1863, Bruhns (BM ; LIVU). SyRIA, ISRAEL, JORDAN, LEBANON : Kishnon R., 1863-64, Lowne (K). Garnala, 3 Apr. 1911, Meyers & Dinsmore 3561 (K). Above Ain Fit, 9 May 1911, Meyers & Dinsmore Gr561 (E ; K). Ramleh, 30 Mar. 1911, Dinsmore 1561 (E). Below Kulaia, 31 Mar. 1877, Post 228 (K). LisyA : Between Cyrene (Shahhat) and Apollonia (Marsa Susa), 4 Apr. 1939, Sandwith 2350 (K). ALGERIA : Bab-el-Oued, prés Alger, Apr. 1879, Allard, Soc. Dauph. 2388 (K). Birmadreis, Alger, 13 Apr. and 15 May 1851, Jarmin 129 (K). Mustapha sup., Algiers, 10 Apr. 1873, Joad (K). Oran, 1843, anon. (K). Introduced : AUSTRALIA : Western Australia : Cannington, Lower Cannington River, 12 Oct. 1910, Morrison 20012 (K). Victoria: Point Lonsdale, Oct.-Nov. 1912, Tilden 672 (K). Shepparton, 17 Sept. 1925, Britten (K). REVISION OF THE GENUS PETRORHAGIA 169 Hawall : Waimea, S. Kohala, 25 May 1938, Hosaka 2rro (K). UNION OF SouTH AFRICA: Cape of Good Hope: Tulbagh, District Wolseley, 4 Oct. 1955, Van Breda 555 (K). Caledon to Batriver, 9 Oct. 1955, Van Breda 564 (K). Jordans Bay, Stellenbosch, 4 Oct. 1947, Parker 4268 (K). 25. Petrorhagia glumacea (Bory & Chaub.) Ball & Heywood, comb. nov. Dianthus glumaceus Bory & Chaub. in Expéd. Sci. Morée, Sect. Sci. Phys. 3 (2): 340 (1832). Tunica glumacea (Bory & Chaub.) Boiss., Fl. Or. 1: 517 (1867). Kohlrauschia glumacea (Bory & Chaub.) Hayek, Prodr. Fl. Penins. Balcan. 1 : 224 (1924). Stems up to 50 cm. high, usually simple, glabrous or scabrid. Cauline leaves up to 40 mm. long and 2 mm. broad, linear-oblong, 3-veined ; sheath 1-2 mm. long, usually shorter than broad. Largest bracts of head 12-18 mm. long, 8-12 mm. broad, obtuse. Calyx (g-)10-13 mm. long; costae 3-veined ; teeth oblong, obtuse. Petals 12-18 mm. long, the limb 3-6 mm. broad, very variable in shape, pink or purplish. Seeds 1-5—2-2 mm. long, I-2-1-7 mm. broad, almost smooth to tuberculate. Distribution : Balkan peninsula. 25a. Petrorhagia glumacea var. glumacea. (Fig. 22 ; Plate 15 D.) Petal limb obdeltate, crenate to laciniate at the apex. Seeds 1-7-2:2 mm. long, I:4-1'7 mm. broad, almost smooth or slightly tuberculate. GREECE : Peloponnisos : Vastitza, 12-24 July 1855, Orphanides 462 (CGE ; K ; W-HAL). Messini, May 1844, Heldreich (BM ; CGE; E). Taygetos, Likorema, 28 June 1902, Haldcsy 51 (W-HAL). Taygetos, Megali Anastasova, 1 Sept. 1899, Heldreich (W-HAL). Prope Astros, Laconia, 6 May 1929, Guiol 556 (BM). Laconia, Sparti, 7 June 1958, Rechinger 19984 (LIVU ; W). Carytaena (Karytaina), 21 May 1929, Guiol 1053 (BM). Plain of Argos, Apr. 1855, Mill (CGE ; K). Mycenae (Mykene), 29 July 1888, Heider (W-HAL). Kyllene, prope Trikala, 30 June 1887, Heldreich (W-HAL). Kalavryta, 25 June 1899, Dérfler 412 (W-HAL) ; 21 June 1926, Bornmiiller 178 (BM). Lintye (Elis), 1 June 1899, Heldreich (W-HAL). Patras, 7 June 1893, Haldcsy (W-HAL). Nea Epidavros, May 1877, Tuckett (K). Dervenakia (Argolis), July 1930, Atchley 275 (K). Sterea Ellas: Attika, Pinatzt (LIVU). 250. Petrorhagia glumacea var. obcordata (Marg. & Reut.) Ball & Heywood, comb. nov. Dianthus obcordatus Marg. & Reut. in Mém. Soc. Phys. & Hist. Nat. Genéve 8 : 281 (1839). Kohlvauschia obcordata (Marg. & Reut.) Reichb., Ic. Fl. Germ. & Helv. 6: 43 (1844). Tunica glumacea var. obcordata (Marg. & Reut.) Boiss., Fl. Or. 1 : 517 (1867). Petal limb obcordate, usually entire. Seeds 1-5-1-8 mm. long, I-2-1°-5 mm. broad, tuberculate. GREECE : Peloponnisos : Mt. Chalkis ad sinum Patranum, 8 June 1893, Haldcsy (K ; W-HAL). Sterea Ellas : Mt. Parnassi prope Levadeia, 17 July 1888, Haldcsy (W-HAL). Mt. Parnassus above Delphi, 31 May 1927, Heard (BM). Gravia, 21 June 1937, Balls & Gourlay 3362 (BM; E; K). Kionae, supra Segritna, 170 REVISION OF THE GENUS PETRORHAGIA 11 July 1888, Haldcsy (W-HAL). Thessalia : Olympus, June-July 1929, Guiol (BM). Kalampaka, 20 May 1896, Sintenis 25z (K). Orman Magula, June 1885, Haussknecht (K). Agrapha, M. Pindi circa monast. Korona, 20-28 June 1885, Heldreich (E). Mt. Ossa, Tsagesi, 23 July 1936, Grebenchikoff (K). Ipiros : Metsovon, 19 July 1937, Balls & Gourlay 3649 (BM ; E; K). Mt. Mitcikeli, Sept. 1896, Baldacci (BM). Makedhonia: Kaream (Chalcidica), 1896, Dimitrijev(K). Stavros, July 1917, Turner (K). Kefallinia : Mt. Aenos, 13 June 1895, Grimburg (W-HAL). TuRKEY : Muratli, 20 June 1890, Degen (K). ALBANIA : Mavrova, Valona, 27 June 1894, Baldacci 234 (BM). Biskezmi, Ljnime, 24 Apr. 1918, Schneider (W). JuGcosLavia : Makedonya: Roszdan, 9 July 1893, Dérfler (W-HAL). Katlan- ovska Banja, near Skopje, seed Oct. 1953, Hort. Kew. 1954, Gore-Browne (K). Srbya: Rissovatz, Gosnza, 3 Oct. 1889, Sintenis (K ; W-HAL). Vranja, 19109, Lucas (BM). Bosnai Hercegovina : Uskoplje, 1 Apr. 1926, Korb (W). Crna Gora : Borkovici, fl. Piva, July 1909, Rohlena (K). Cattaro, 13 June 1867, Huter (BM ; CGE ; K); Puchler (BM; K; W-HAL). Hrvatska: Teodo, 4 June 1905, Schneider (W). Krivosije, bei Ledenice, 17 June 1926, Korb (W). Scaljari, Pichler, Fl. Austro-Hung. 2508 (K ; W-HAL ; W-RECH). Bei Combar nachst Casselnuovo, June 1876, Studniczka (W-HAL). Circa Spalatensi, 1843, Alexander (BM). Introduced : ItaLty : Trentino: Prope Povo et Roncegno, July 1899, Murr in Dérfler, Herb. Norm. 4orr (BM ; E). This is a somewhat variable species which has been confused with P. prolifera in the central and northern Balkan peninsula. All the Bulgarian specimens seen and many of those from Jugoslavia and northern Greece are undoubtedly P. prolifera. The status of the two varieties recognized is still not clear as it has not been possible to determine the extent of the correlation between the petal and seed characters. Most specimens are either in flower and without mature seeds or in fruit and without petals. Large almost smooth seeds occur only in Peloponnisos and southern Sterea Ellas while small tuberculate seeds occur in all areas except Peloponnisos. The petal type of var. glumacea is largely restricted to Peloponnisos, but intermediate types occur occasionally almost throughout the range of var. obcordata. A further difficulty is that in many specimens the petals are badly pressed owing to the bulkiness of the head so that it is difficult to see the precise shape and toothing of the petal limb. From this discussion it will be clear that there is the possibility that the two varieties recognized may eventually be treated as subspecies. However, the evidence available at the present is so inconclusive and unsatisfactory that it is impossible to justify raising them to this rank. Insufficiently known species Tunica syriaca Boiss., Diagn. Pl. Or. Nov. 2 (8) : 63 (1849) ; Fl. Or. 1 : 519 (1867).— Post, FI..Syr. Palest. & Sinai, ed. 2, 1 + 160 (1932). “ T. glaberrima nigrescenti-viridis perennis basi suffrutescens multicaulis, caulibus REVISION OF THE GENUS PETRORHAGIA 171 filiformibus rectis vel ascendentibus ad nodos superiores szepé geniculatis simplicibus unifloris rarius ramulo auctis nigricantibus asperulis, foliis parvis angusté lineari- triquetris crassiusculis obtusiusculis supra sulcatis subtus carinatis margine scabridis strictis internodio 3-4 plo brevioribus basi breviter connatis, bracteis 8-10 calyci subeequilongis lanceolatis acutissimis subpungentibus preter nervum rubrum crassum angusté albo-membranaceis, calyce oblongo campanulato brevissime puberulo tubo enervi albido dentibus tubo triplo brevioribus oblongis nigricantibus albo-marginatis, petalis angusté linearibus obtusis ungue nudo lamina pallidé rosea breviter exserta, capsula. . . “Hab. in rupibus Syri@ borealis poné urbem Antiochiam sitis ubi specimina pauca Junio 1846 legi. ‘“‘Caules tenuissimi 6-8 pollicares, folia 3-5 lin. longa, calyx cum bracteis 4 lin. longus. Species distinctissima propé T. dianthoidem collocanda.”’ The above is a transcription of Boissier’s protologue. Tristram (Survey of Western Palestine, Flora & Fauna : 240 (1884)) also records this species from northern Lebanon, but he does not give any further information. No material has been seen and no additional records have been traced. Even the most recent description (Post, 1932) is clearly based entirely on Boissier’s data. The description suggests that this species is nearest to Petrorhagia sect. Petrorhagia subsect. Thessalae, all the species of which are rare and local late-flowering perennials. Geographically it would form a link between the Balkan and North African species, and P. macra from Kurdistan. Excluded species Petrorhagia glomerata (Pall. ex Bieb.) Link, Handb. Erkenn. Gewichse 2 : 235 (1831) = Gypsophila glomerata Pall. ex Bieb. SPECIAL LITERATURE Batt, P. W., & HEywoop, V.H. (1962). The taxonomic separation of the cytological races of Kohlrauschia prolifera (L.) Kunth sensu lato. Watsonia 5 : 113-116. BarkoupaH, Y. I. (1962). A revision of Gypsophila, Bolanthus, Ankyropetalum and Phryna. Wentia 9 : 1-203. BécHER, T. W., LarRseEN, K., & Raun, K. (1953). Experimental and cytological studies on plant species. I. Kohlrauschia prolifera and Plantago coronopus. Hereditas 39 : 289-304. (1955). Experimental and cytological studies on plant species. II. Tvrifolium arvense and some other pauciennial herbs. Dan. Biol. Skr. 8 (No. 3). BoissiER, E. (1867). Flora Ovientalis. 1. Bale & Genéve. Danpy, J. E. (1957). Some new names in the British flora. Watsonia 4 : 41-48. DarRLINGTON, C. D., & WyLtE, A. P. (1955). Chromosome Atlas of Flowering Plants. London. Davis, P. H. (1957). New Turkish species of Tunica, Velezia and Potentilla. Not. R. Bot. Gard. Edin. 22 : 163-171. EHRENDORFER, F. (1958). Ein Variabilitaétszentrum als “ fossiler ’’ Hybrid-Komplex : Der ost-mediterrane Galium graecum 1..-G.canum Req.-Formenkreis. Eine Monographie. Osterr. Bot. Zeitschy. 105 : 229-279. 172 REVISION OF THE GENUS PETRORHAGIA Hrywoop, V. H. (1958). The presentation of taxonomic information. A short guide for contributors to Flora Europaea. Leicester. JANCHEN, E. (1963). Catalogus Florae Austriae. I. Teil : Pteridophyten und Anthophyten (Farne und Bliitenpflanzen). Erganzungsheft. Wien. Léve, A., & LévE, D. (1961). Chromosome numbers of Central and Northwest European plant species. Opera Botanica 5 : 1-581. Marre, R. (1963). Flore del’Afrique du Nord. 10. Paris. Wix.iams, F. N. (1890). Synopsis of the genus Tunica. Journ. of Bot. 28 : 193-199. Bull. B.M. (N.H.) Bot. 3, 4 PLATE 13 A B Cc D Seeds (x15). a a - 7 & 7 a ain - ‘ Fs = \ ec . A} ~ a iy -~ Cite ee ee a eee ee ae . eee. a MARINE ALGAE OF GOUGH ISLAND YVONNE M. CHAMBERLAIN BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BOTANY Vol. 3 No. 5 LONDON : 1965 fen a hee oa, MARINE ALGAE OF GOUGH ISLAND BY YVONNE M. CHAMBERLAIN Pp. 173-232 ; 80 Text-figures ; Plates 16-19 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BOTANY Vol. 3 No. 5 LONDON : 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper ts Vol. 3 No. 5 of the Botany serves. © Trustees of the British Museum (Natural History), 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued May, 1965 Price Thirty-two Shillings MARINE ALGAE OF GOUGH ISLAND By YVONNE M. CHAMBERLAIN (Mrs. BUTLER.) THE material reported on in this paper was collected by Dr. N. M. Wace during the stay of the Gough Island Scientific Survey on Gough Island from 13 November 1955 to 13 May 1956, and is now in the British Museum Herbarium. The collections were made on only a few days during this summer period and the species included here undoubtedly represent only a few of those present on the island. However, as Dr. Wace is an ecologist, he collected particularly those species which dominated the littoral and sub-littoral zones ; I think therefore that all the zone-forming species are included here together with the most abundant associated littoral species. Only two or three attempts at dredging were made and these were unfortunately fruitless ; none, therefore, of the smaller sub-littoral species, notably the Delesseriaceae, which undoubtedly grow round Gough Island, is included in the collections. ACKNOWLEDGEMENTS I am indebted to Dr. Wace for making the collections, which were well preserved and clearly labelled, also to him and Dr. M. W. Holdgate (the invertebrate zoologist of the expedition) for much helpful discussion on the littoral ecology of the island, and to Mr. J. J. van der Merwe who collected some more Blidingia and Enteromorpha when he stayed on Gough Island after the return of the Survey and thus helped to clarify some taxonomic and ecological problems. I worked on the material while I was a member of the staff of the British Museum (Natural History) and I am grateful to Mrs. F. L. Balfour-Browne, Mr. J. E. Dandy and Mr. R. Ross for general, editorial and nomenclatural help ; to Dr. W. T. Stearn who wrote the latin diagnoses of new species ; and to Mr. D. W. Cooper who made sections of the Corallinaceae. For general taxonomic guidance I am indebted to Dr. P. S. Dixon and to Mme. Marie Lemoine, with whom I was able to discuss the Corallinaceae when I visited the Muséum National d’Histoire Naturelle, Paris. Several of the species were sent to taxonomists expert in particular groups and their help is acknowledged in the appropriate places. Finally, I would like to thank the Directors of the following Institutions for permission to examine herbarium material: Royal Botanic Gardens, Kew ; Bolus Herbarium, Cape Town ; Herbarium of the University of California ; Rijksherbarium, Leyden ; Muséum National d’Histoire Naturelle, Paris ; and Dr. Egil Baardseth, who presented a duplicate collection of his Tristan da Cunha algae to the British Museum. MARINE ALGAE OF GOUGHIISLAND GEOGRAPHY OF GOUGH ISLAND Gough Island (Figs. 1, 2) is an outlier of the Tristan da Cunha group of islands situated near the southern end of the mid-Atlantic ridge. It lies 370 km. S.S.E. of Tristan da Cunha, more than 3,000 km. from South America to the west and Greenwich Tristan da Cunha : -— "Gough Prince Edward “Marion 60°E SS Sandwich “,Crozets \ S. Georgiaes i — + — = Ps Falklands § Orkneys ra Kerguelen New se. Paut .Amsterda #S. Shetlands \ . : rd \ raham Land .—:4 oe terra . »Masatierra del Fuego *Masafuera vu rv) c © 2 © > 5 UG > v ¢ g off 120°E 120°W +Macquarie sl 50° . Campbell »Aucklands \ Tasmamia od Antipodes™— "Snares Bounty: f 40° Chathame és er oN NEW ZEALAND :Kermadec 180° 150°W Fic. 1. The Southern Hemisphere. 2,400 km. from South Africa to the north-east. The island lies in the West Wind Drift and within the zone of sub-antarctic, cold-temperate water (Knox, 1960, p. 578) but may occasionally be influenced by the cold-temperate mixed water derived from the Brazil current. Gough Island lies between the Antarctic and Sub-tropical Convergences (Knox, 1960, p. 582) but is closer to the latter, which is here at, or somewhat to the south of, Tristan da Cunha. MARINE ALGAE OF GOUGH ISLAND 177 Church i] Rock 0 Point Dell Rocks Window Rock Slaughtered Seal Bay “o ( The Midshipman » Stand off Rock [000 yds_ Fic. 2. Sketch map of Gough Island to show localities ee marine algae were collected. LITTORAL ECOLOGY OF GOUGH ISLAND It is intended to publish a paper on the littoral ecology of Gough Island and only the main features will be summarized here. The entire island is very exposed in comparison with continental shores so that the terms sheltered, moderately exposed, etc., are relative only to conditions on the island. On the moderately exposed shore at Dell Rocks Beach the zonation is : Upper supra-littoral Lichina sp. Lower supra-littoral Porphyra tristanensis, Verrucaria sp. Upper littoral Blidingia minima in crannies and sheltered places. Inidaea laminarioides where more exposed. Enteromorpha bulbosa. Mid littoral Inidaea laminarioides gradually giving way to Rhodo- glossum revolutum. Lower littoral Inridaea laminarioides and Rhodoglossum revolutum succeeded by Durvillaea antarctica and Melobesteae. Sub-littoral The Durvillaea-Melobesieae zone extends down into the sub-littoral and below this Macrocystis pyrifera is dominant. 178 MARINE ALGAE OF GOUGH ISLAND Plate 16 A shows the shore at Isolda Rock which is somewhat more exposed than Dell Rocks Beach. Light-coloured Melobesteae can be seen on the left, a dark band of Iridaea laminariotdes in the centre and plants of Durvillaea antarctica on the right. On the more sheltered Midshipman Rock, the mid littoral is dominated by a band of Ralfsia sp., below which is a band of Melobesieae and Corallina officinalis. West Point Reefs is the most exposed locality ; here Ividaea laminarioides and Rhodo- glossum revolutum zones are absent and the mid littoral is dominated by a band of white Melobesteae succeeded by a band of Polysitphonia howei (probably) and other Rhodomelaceae. Macrocystis pyrifera is also absent on very exposed shores. Among the most common plants occurring in littoral rock pools throughout the island are: Corallina officinalis, Dermatolithon nodulosum, Inidaea undulosa, Scytostphon lomentaria and Codium fragile. The algal zonation on Gough Island is typical of the sub-antarctic pattern (Skottsberg, 1941 ; Knox, 1960, p. 601). It differs markedly from that of Tristan da Cunha, which has more in common with a South African shore. This corresponds with the fact that Gough Island is well south of the Sub-tropical Convergence while Tristan da Cunha is at its mean position. Very little is known about the east coast shore of South America at about the same latitude as Gough Island, but it is probable that further investigations will show that it has a zonation pattern similar to that of Gough Island. Further south, Tierra del Fuego and the Falkland Islands have a zonation pattern similar to that of Gough Island but with a greater number of typical antarctic species. LOCALITIES AND HABITATS, WITH SPECIES PRESENT (Entries in brackets relate to algae that have not been identified at species level and are not included in the Systematic List.) Dell Rocks Point. 10 Dec. 1955. Washed ashore at storm-tide mark. G.J.S.S. 1501. {Melobesieae.| Between the Admiral and the Commodore. 22 Dec. 1955. Between 18 and 55 m., and in gullies. G.J.S.S. 1502. Macrocystis pyrifera. Slaughtered Seal Bay. 22 Dec. 1955. Porphyra zone. G.I.S.S. 1503. Porphyra tristanensis. Slaughtered Seal Bay. 22 Dec. 1955. In pools from water level down to Io cm. G.I.S.S. 1504. Ulva lactuca, Corallina officinalis, Dermatolithon nodulosum. Slaughtered Seal Bay. 22 Dec. 1955. Below15cm.inrock pools. G.I.S.S. 1505- 1507. Codium fragile with epiphytic Feldmannia globifera, Giffordia granulosa and Sphacelaria furcigera, Gelidium regulare, Corallina officinalis, Dermatolithon nodulosum, Lithothamnium neglectum, Iridaea undulosa, Centroceras clavulatum. Beach south of Sophora Glen. 28 Dec. 1955. At tide mark on boulder beach. G.I.S.S. 1508. |Melobesieae, Gelidium sp.] Near Penguin Island. 2 Jan. 1956. At55m. G.I.S.S. 1509. Macrocystis pyrifera, [Melobesieae]. Outer Dell Rocks. 18 Feb. 1956. From mid-tide level to 2 m. below low-tide level. G.I.S.S. 1510. Durvillaea antarctica. MARINE ALGAE OF GOUGH ISLAND 179 Dell Rocks Beach. 19 Feb. 1956. Near low-tide mark on boulder beach. G.J.S.S. 511. Scytosiphon lomentaria. Window Rock. 18 Feb. 1956. Below low-water neaps on rocks amongst corallines. G.I.S.S. 1512. Cladophora radtosa. Midshipman Rock. 18 Feb. 1956. Porphyra zone, upper shore. G.I.S.S. 1573, 1514. Oscillatoria nigroviridis, Phormidium submembranaceum, Ulva lactuca, Porphyra tristanensis, Dermatolithon nodulosum, Iridaea laminarioides, Rhodo- glossum revolutum. Midshipman Rock. 18 Feb. 1956. Mid-littoral zone, on shiny olive-green deposit. G.I.S.S. 1515, 1517, 1518. [Ralfsia|, Corallina officinalis, Centroceras clavulatum, Microcladia alternata, |Delesseriaceae|, Herposiphonia paniculata, Polysiphonia howet. Midshipman Rock. 18 Feb. 1956. Lower littoral. GJ.S.S. 1516. Colpomenia sinuosa. Midshipman Rock. 18 Feb. 1956. Standing water in upper littoral. G.J.S.S. I519, 1520. Scytosiphon lomentaria, Centroceras clavulatum, Herposiphonia paniculata, | Rhodomelaceae}. Midshipman Rock. 18 Feb. 1956. Upper-littoral pools. G.J.S.S. 1527, 1522. Entophysalis conferta, E. deusta, Calothrix crustacea, Lyngbya confervoides, Ulva lactuca. Standoff Rock. 18 Feb. 1956. Mid-littoral coralline zone. G.J.S.S. 1524, 1525. Corallina officinalis, Rhodoglossum revolutum. Standoff Rock. 18 Feb. 1956. Above Corallina on less-exposed part of rock. G.I.S.S. 1526. Rhizoclonium ambiguum, R. riparium forma validum, Corallina officinalis, Iridaea laminarioides, Bostrychia mixta, Lophostphonia scopulorum. Standoff Rock. 18 Feb. 1956. Upper littoral, Porphyra zone. G.I.S.S. 1527. Entophysalis deusta, Calothrix crustacea, Plectonema calothrichoides, Bangia fuscopurpurea, Porphyra tristanensis, |Hildenbrandia}. Church Rock. 5 Mar. 1956. Vertical rock face. G.J.S.S. 1528-1535. Cladophora vadiosa, Codium adhaerens, Iridaea undulosa, Rhodoglossum revolutum, Centroceras clavulatum, |Nitophyllum sp.], Herposiphonia paniculata, Lophurella sp. West Point Reefs. 6 Mar. 1956. In and around bird-fouled spray-zone rock pools. G.I.S.S. 1536-1538, 1544. Entophysalis deusta, Calothrix crustacea, Plectonema calothrichoides, Enteromorpha bulbosa, Scytosiphon lomentaria, |Hildenbrandia? |. West Point Reefs. 6 Mar. 1956. In rock pools below spray zone. G.I.S.S. 1540, 1541. Ulva lactuca, Codium fragile, Corallina officinalis, Dermatolithon nodulosum, Rhodoglossum revolutum, Polysiphonia boergesenit. West Point Reefs. 6 Mar. 1956. Margins of pools below spray zone. G/.S.S. 1542, 1543. Ulva lactuca, Dermatolithon nodulosum. Isolda Rock. 6 Mar. 1956. Mid-littoral rock pools. G.I.S.S. 1545, 1546, 1550, 1551. Cladophora rupestris, Corallina officinalis, Dermatolithon nodulosum, Inridaea undulosa. Isolda Rock. 6 Mar. 1956. Lower-littoral rock pools. G.J.S.S. 1547, 1548. Cor- allina goughensis, Lithothamnium neglectum. 180 MARINE ALGAE OF GOUGH ISLAND Isolda Rock. 6 Mar. 1956. Margins of mid-littoral pools. G.J.S.S. 1549. Poly- siphonia howe. Isolda Rock. 6 Mar. 1956. Upper littoral on ridges between pools. G.J.S.S. 1552. Iridaea laminarioides, Rhodoglossum revolutum, Herposiphonia paniculata, Polysiphonia hower. Isolda Rock. 6 Mar. 1956. Upper-littoral pools. GJ.S.S. 1553. Corallina officinalis, Dermatolithon nodulosum, Lithothamnium neglectum, Rhodoglossum revolutum, Centroceras clavulatum. Window Pool. 12 Apr. 1956. In caves and wet crevices of spray-zone rocks. G.I.S.S. 1554. [Chlorophyta]. Window Pool. 12 Apr. 1956. In pools and sheltered channels of boulder beach. G.I.S.S.1555,1556. Blidingia minima, Inidaea undulosa, Rhodoglossum revolutum. Window Pool. 12 Apr. 1956. On walls of small shallow cave beside pool. G.I.S.S. 1557. Entophysalis conferta, Rhizoclonium ambiguum. Dell Rocks Beach. 12 Apr. 1956. In sheltered places between boulders and in shallow pools. G.I.S.S. 1558. Guiffordia mitchelliae. Dell Rocks Beach. 12 Apr. 1956. In pools and protected channels just above mid-tide level. G.I.S.S. 1559. Scytostphon lomentaria. Between Station and Dell Rocks. 1 Dec. 1960. High-water level. Van der Merwe 1,2, 4. Blidingia minima, Enteromorpha bulbosa. Capsize Sands. 1 Dec. 1960. Van der Merwe 3. Blidingia minima, Enteromorpha bulbosa. PHYTOGEOGRAPHY The elements of which the Gough Island marine algal flora is composed are : Cosmopolitan 19 species General southern hemisphere 4 species Southern cold-temperate 6 species North and south Atlantic 3 species Tristan da Cunha group endemics 6 species Endemic 2 species This list is necessarily incomplete and inaccurate for two reasons. The first is that, compared with, for example, the angiosperms, relatively little is known about the distribution of marine algae and, until fairly recently, authors have made little attempt to relate the flora of the area on which they were working to that of other areas. Secondly, there is at present a very wide variation of opinion among algal taxonomists as to what constitutes a species. In the past, a large number of new species have been described without regard to seasonal, ecological or geo- graphical variation, and also without reference to previous literature on or collections from the same area. On the whole, present-day phycologists regard species in the widest sense, as entities which are liable to great variation according to the external conditions mentioned above. I agree with this concept, and this may account for the large number of cosmopolitan and widely distributed species in the list. MARINE ALGAE OF GOUGH ISLAND 181 It is a well-known fact, however, that isolated islands usually have relatively high numbers of two groups of species in their flora, cosmopolitan species and endemic species, so that the list may be only an exaggeration of the true picture. Of the second group, the endemic species, two species so far known only from Gough Island are described, and as these are in the Corallinaceae, which is a very difficult and often ignored family, it is probable that they will later be found elsewhere. On the other hand, endemic species are more likely to be found among the smaller, less conspicuous species. As previously explained, the collector concentrated on the dominant species and many possible endemics have no doubt so far escaped notice. There were some more possibly new species in the collection but the material was too scanty, in my opinion, to merit their description as such. There are some rather surprising absences from Gough Island, most notably Hildenbrandia lecannelliert and Adenocystis utricularis. Sub-littoral members of the Delessertaseae and of Desmarestia are also missing from the collection, probably because of the difficulties of dredging. Two littoral members of the Delesseriaceae were collected but I was unable to identify them further than that they are probably species of Nitophyllum. The Hildenbrandia and Adenocystis may have been missed, but Dr. Holdgate, who has seen many southern cold-temperate shores, is fairly sure that they were not at any of the localities visited by the Survey. In conclusion, on present evidence it is not possible to make many generalizations about the Gough Island marine algal flora. It does however represent a typical sub- antarctic zonation pattern and the high number of endemic species (to the Tristan da Cunha group) is typical of isolated localities. SYSTEMATIC LIST CYANOPHYTA! CHAMAESIPHONACEAE ENTOPHYSALIS CONFERTA (Kiitz.) Drouet & Daily in Lloydia 11 : 79 (1948) ; in Butler Univ. Bot. Stud. 12 : 111, figs. 196-215 (1956). Palmella conferta Kiitz., Phyc. Germ.: 149 (1845). Midshipman Rock, in upper-littoral pool, 18 Feb. 1956, G.I.S.S. 1522. Window Pool, on walls of small shallow cave beside pool, 12 Apr. 1956, G.I.S.S. 1557. Cosmopolitan. This species occurred as an epiphyte on Calothrix crustacea and Rhizoclonium ambiguum. When Dr. Drouet discussed the identification of the Gough Island Cyanophyta with me, he commented that the only difference between this and the next species is that this is an epiphyte and E. deusta is a lithophyte. ENTOPHYSALIS DEUSTA (Menegh.) Drouet & Daily in Lloydia 11 : 79 (1948) ; in Butler Univ. Bot. Stud. 12 : 103, figs. 185-194, 247-250 (1956). Coccochloris deusta Menegh. in Atti 2 Riun. Sci. Ital. Torino 1840 : 173 (1841). 1Dr. Francis Drouet of the Academy of Natural Sciences of Philadelphia kindly identified the Cyanophyta. 182 MARINE ALGAE OF GOUGH ISLAND Midshipman Rock, in upper-littoral pool, 18 Feb. 1956, G.I.S.S. 1522. Standoff Rock, from upper littoral, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1527. West Point Reefs, encrusting layer on rock in and around spray-zone pools, 6 Mar. 1956, G.I.S.S. 1538, 1544. Cosmopolitan. The growth form of the specimens of this plant from Gough Island is the cushion type. It grows on rocks in the upper littoral zone where it forms a film, usually in association with other Cyanophyta. Calothrix crustacea is its most frequent associate. RIVULARIACEAE CALOTHRIX CRUSTACEA Born. & Flah. in Ann. Sci. Nat., Sér. 7, Bot. 3 : 359 (1886). Midshipman Rock, in upper-littoral pool, 18 Feb. 1956, G.J.S.S. 1522. Standoff Rock, from upper littoral, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1527. West Point Reefs, encrusting layer on rock in and around spray-zone pools, 6 Mar. 1956, GIS3S 7538; 1544: Cosmopolitan. C. crustacea is the most common blue-green alga collected on Gough Island. It grows in pools and on rocks in the upper littoral and spray zones of several localities, where it forms a film mixed with Entophysalis deusta and Lyngbya confervoides. It often bears epiphytes such as Plectonema calothrichoides and Entophysalis conferta. From the available information, there is no indication that the Cyanophyta are ever the dominants of a littoral zone on Gough Island, although they are certainly a noticeable feature of the upper littoral zone on the Midshipman Rock. Baardseth (I94I, p. 133) records Cyanophyta as forming a dense covering of rocks in the Porphyra zone on the exposed rocks of Tristan da Cunha and they frequently occur as the dominant or sub-dominant algal community at this shore level in other parts of the world. Skottsberg (1941) did not mention the Cyanophyta in his analysis of antarctic and sub-antarctic algal communities. OSCILLATORIACEAE LYNGBYA CONFERVOIDES Gom. in Ann. Sci. Nat., Sér. 7, Bot. 16 : 136, t. 3 figs. 5, 6 (1892). Midshipman Rock, in upper-littoral pool, 18 Feb. 1956, G.I.S.S. 1522. Cosmopolitan. Collected from the Midshipman Rock where it grew with Calothrix crustacea and Entophysalis deusta as a film on rock submerged in an upper-littoral rock pool. OSCILLATORIA NIGROVIRIDIS Gom. in Ann. Sci. Nat., Sér. 7, Bot. 16: 217, t. 6 fig. 20 (1892). Midshipman Rock, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1514. Cosmopolitan. This species was found growing with Phormidium submembranaceum on the Midshipman Rock. MARINE ALGAE OF GOUGH ISLAND 183 PHORMIDIUM SUBMEMBRANACEUM Gom. in Ann. Sci. Nat., Sér. 7, Bot. 16: 180, t. 5 fig. 13 (1892). Midshipman Rock, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1514. Cosmopolitan. Only a small amount of this alga was collected ; it came from the upper littoral zone on the Midshipman Rock where it was growing mixed with Oscillatoria nigroviridis forming a film on the rocks. PLECTONEMA CALOTHRICHOIDES Gom. in Bull. Soc. Bot. France 46 : 30, t. 1 figs. 6-10 (1899). Standoff Rock, from upper littoral, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1527. West Point Reefs, encrusting layer on rock in and around spray-zone pools, 6 Mar. 1956, G.I.S.S. 1544. Cosmopolitan. Growing as an epiphyte on Calothrix crustacea. CHLOROPHYTA ULVACEAE BLIDINGIA MINIMA (Nageli ex Kiitz.) Kylin in K. Fysiogr. Sallsk. Lund Forhandl. 17: 181 (1947).—Womersley in Austral. Journ. Mar. Freshw. Res. 7: 350 (1956). (Figs. 3-5.) Ulva intestinalis var. nana Sommerfelt, Suppl. Fl. Lapp.: 186 (1826). Enteromorpha minima Nageli ex Kiitz., Sp. Alg.: 482 (1849).—Kiitz., Tab. Phyc. 6: 16, t. 43 fig. 3 (1856).—Bliding in Bot. Notis. 1938: 84, figs. 1-6 (1938).—Sj6stedt in Svensk Bot. Tidskr. 33 : 18, fig. 1 (1939). Enteromorpha micrococca Kiitz., Tab. Phyc. 6 : 11, t. 30 fig. 2 (1856).—T. A. Stephenson in Ann. Natal Mus. 11 : 296 (1948). Enteromorpha gunniana J. G. Agardh in Act. Univ. Lund. 19 (2) : 122 (1883).—Hylm6 in K. Svensk. Vetenskapsakad. Handl, Ser.. 3, 17 (1) : 6 (1938).—Sjéstedt, tom. cit. : 54, fig. 11 (1939).—Skottsb. in K. Svensk. Vetenskapsakad. Handl., Ser. 3, 19 (4) : 75 (1941). —Levring in Act. Univ. Lund., Ny Foljd, Avd. 2, 56 (10) : 8 (1960). Enteromorpha nana (Sommerfelt) Sjéstedt, tom. cit.: 28, figs. 3-6 (1939).—V. J. Chapman in Journ. Linn. Soc. Lond., Bot. 55 : 412, figs. 57-63 (1956). Window Pool, abundant in pools and sheltered channels of boulder beach, 12 Apr. 1956, G.I.S.S. 1555. Between Station and Dell Rocks, high-water level, 1 Dec. 1960, Van der Merwe 2. Capsize Sands, 1 Dec. 1960, Van der Merwe 3. Cosmopolitan. This plant grew abundantly in the upper and mid littoral zones at Dell Rocks Beach and occurred both in sheltered rock crevices and in pools. It was also collected from Capsize Sands. It formed a low, bright-green turf which at times dominated the shore community. This species is found throughout the world in the upper littoral zone of the sea-shore and quite often dominates parts of this zone (cf. T. A. & A. Stephenson, 1954, p. 38). The Gough Island material consists of minute plants rarely more than 1 cm. high. Numerous fronds grow from a flattened basal part (Fig. 3) ; the fronds are tubular, 184 MARINE ALGAE OF GOUGH ISLAND ~ Imm Fics. 3-5. Blidingia minima (Nageli ex KUtz.) Kylin: 3, habit of plant ; 4, surface view of frond, = pyrenoid ; 5, transverse section of frond, os = outer surface, p = pyrenoid. the tubes being closed when young but open at the top when older. Some of the older plants are twisted while others are swollen at the top. The cells seen in surface view are irregular in shape and arrangement (Fig. 4) ; they have thick walls and a large single pyrenoid. The cells measure up to 7u in diameter but are mostly about 3-5u. In transverse section (Fig. 5) the cells are elongated and measure about 3-4 x 8-gu, and the whole section is about 15-18 deep. The inner wall is thick and may measure up to 8u deep. Numerous forms and varieties of Blidingia minima have been described, under the various names given in synonymy above, according to the size and shape of the fronds, the thickness of the cell walls and the nature of the basal part. The Gough Island material is most like the descriptions of Enteromorpha gunmniana (e.g. that of Sjdstedt), but Womersley, who has seen Agardh’s type material, does not think it sufficiently distinct to merit specific separation from the rest of the complex. Dr. Carl Bliding kindly confirmed the identification of this species. ENTEROMORPHA BULBOSA (Suhr) Montagne in Gaudich., Voy. Monde La Bonité, Bot. 1 : 3 (1846).—De Toni, Syll. Alg. 1 : 127 (1889). (Figs. 6-8.) Solenia bulbosa Suhr in Flora 22 : 72, t. 4 fig. 46 (1839). Enteromorpha africana Kiitz., Phyc. Gen. : 300 (1843) ; Tab. Phyc. 6: 14, t. 40 fig. 2 (1856). Enteromorpha hookeriana Kiitz., Sp. Alg.: 480 (1849) ; Tab. Phyc. 6: 13, t. 37 fig. 2 (1856). West Point Reefs, in bird-fouled spray-zone rock pools, 6 Mar. 1956, G.I.S.S. 1537. Between Station and Dell Rocks, high-water level, 1 Dec. 1960, Van der Merwe 1 ; same locality, seaweed turning white in sun, 1 Dec. 1960, Van der Merwe 4. Capsize Sands, 1 Dec. 1960, Van der Merwe 3. South America, Falkland Islands, South Africa, Crozet Islands, Kerguelen Island, Australia, New Zealand, Chatham Islands, Antarctica, ? Japan. MARINE ALGAE OF GOUGH ISLAND 185 Collections of E. bulbosa were made at the very exposed West Point Reefs where it was growing in penguin-fouled rock pools in the spray zone, at Dell Rocks Beach where it grew on littoral rocks, and at Capsize Sands. The green band recorded as being present on other parts of the island was probably dominated by this species and Blidingia minima. Fics. 6-8. Enteromorpha bulbosa (Suhr) Montagne : 6, part of mature plant ; 7, surface view of frond, p = pyrenoid ; 8, transverse section of frond, os = outer surface. The largest plants of this species collected from Gough Island (Fig. 6) are 6 cm. high and up to 5 mm. wide, but most of the plants are smaller than this. Both branched and unbranched plants were collected ; when branches are present they are long and very similar to the main axis, as is characteristic of this species, and not noticeably smaller than the main axis as in many other species of Enteromorpha. The thallus is tubular and compressed, the branches and axis are usually open at the top and often torn. The branches and main axis taper downwards to a terete base and some specimens still have the warty basal part by which they had been attached to the rocks. Many of the tubes are perforated (Fig. 6) by an amphipod ; this animal apparently lives in the tubes of FE. bulbosa and many specimens of it were found in situ. When seen in surface view (Fig. 7) the cells are irregularly arranged, are up to I5u in diameter, and have rounded corners. There is one fairly large pyrenoid per cell. In transverse section (Fig. 8) the cells are rectangular and measure about 25u. tall x 124 wide. The inner and outer membranes of the tube are fairly thick 186 MARINE ALGAE OF GOUGH ISLAND and the section measures about 30u deep altogether. The plants are grass-green in colour and tend to be darker at the base of the axis and branches than in the upper parts, a feature mentioned by von Suhr. The Gough Island plants agree well with von Suhr’s description, with Montagne’s material in the herbarium of the Muséum National d’Histoire Naturelle, Paris, and with Kiitzing’s figures of E. africana and E. hookeriana, both considered by De Toni to be E. bulbosa. E. bulbosa is a common plant of the sub-antarctic littoral ; according to Skottsberg (1941, p. 56) it forms a widespread community on shores with a loose, mobile substrate such as sand, and frequently occurs also in other associations such as littoral pools. £. bulbosa has been recorded from the antarctic (cf. Hariot, 1907, p. 3, and Gain, 1912, p. 26) and from warmer parts such as South Africa (Levring, 1938, p. 5), New Zealand (Chapman, 1956, p. 411) and Australia (Womersley, 1956, Pp. 351), but it is on sub-antarctic shores that it grows most prolifically. The identification of this species was confirmed by Dr. Bliding. DivatrActuca L., Sp: Pl2- 1103 (1753)e0 (Pigs. 6-12.) Slaughtered Seal Bay, in pools from water level down to I0 cm., 22 Dec. 1955, G.I.S.S. 1504. Midshipman Rock, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1514 ; same locality, dominant in upper-littoral pools, 18 Feb. 1956, G.J.S.S. 1527. West Point Reefs, pools below spray zone, 6 Mar. 1956, G.I.S.S. 1541 ; same locality, margins of pools below spray zone, 6 Mar. 1956, G.I.S.S. 1542, 1543. Cosmopolitan. This plant was very common on rocks and in pools in the upper littoral zone of the Gough Island shores. {RG EP ST ba ( Bi6 10u 10 il iz Fics. 9-12. UlvalactucaL. : 9, blade margin ; 10, surface view of frond ; 11, transverse section of frond ; 12, transverse section of stipe, semi-diagrammatic. The material collected from Gough Island varies considerably in size and shape. All the specimens have a somewhat fleshy stipe from which the blades expand ; the blades are sometimes entire but more often divided—sometimes right down to MARINE ALGAE OF GOUGH ISLAND 187 the holdfast. The blade margins (Fig. 9) vary from smooth to crenulate or fimbriate, all types of margin sometimes occurring on one plant. Although there is considerable variation in external appearance, the microscopic structure of the plants is remarkably constant. In the upper parts of the plants the cells are irregularly arranged, somewhat rounded and up to 20yu in diameter in surface view (Fig. 10). In section the blade is 60—70u thick (Fig. 11), the two rows of cells are separated by a hyphal layer and the cells are rather rounded and not more than one and a half times as high as they are wide. The stipe in section (Fig. 12) is much thicker than the blade and has a wide central region of hyphae. This material undoubtedly corresponds to what has been generally accepted as Ulva lactuca L.—an extremely common plant which grows throughout the world. Papenfuss (1960, pp. 303-305) has investigated Linnaeus’s material of U. lactuca and his concept of the genus Ulva. As a result of this investigation, Papenfuss concluded that the genus which Linnaeus really had in mind was that now known as Enteromorpha Link and proposed that the generic name Ulva should be conserved in the sense of Thuret and all subsequent authors ; this proposal has now been accepted, and I am here using the name in that sense. CLADOPHORACEAE RHIZOCLONIUM AMBIGUUM (Hook. & Harv.) Kiitz., Sp. Alg. : 387 (1849). (Figs. 13-15.) Conferva ambigua Hook. & Harv. in Hook., Lond. Journ. Bot. 4 : 295 (1845). Rhizoclonium hookeri Kiitz., op. cit. : 383 (1849) ; Tab. Phyc. 3 : 21, t. 67 fig. 3 (1853).— Stockmayer in Verh. Zool.-bot. Ges. Wien 40 : 584 (1890).—V. J. Chapman in Journ. Linn. Soc. Lond., Bot. 55 : 468, fig. 126 a (1956).—W. R. Taylor, Mar. Alg. E. Trop. & Subtrop. Coasts Amer. : 77, t. 2 fig. 5 (1960). Rhizoclonium africanum Kiitz., Tab. Phyc. 3 : 21, t. 67 fig. 2 (1853). Lychaete novae-zelandiae J. G. Agardh in Act. Univ. Lund. 14 (4) : 2 (1878). Rama novae-zelandiae (J. G. Agardh) V. J. Chapman in Trans. R. Soc. New Zeal. 80: 56, fig. 11 (1952). Standoff Rock, on less-exposed part of the rock above the Corallina zone, 18 Feb. 1956, G.I.S.S. 1526. Window Pool, on walls of small shallow cave beside pool, ¥2 Apr. 1950,G.1 58:5. 1557: Widely distributed in the Southern Hemisphere and the tropics as far north as Florida. ; recorded from A small quantity of this species was found in the Bostrychietum on Standoff Rock ; it occurred in greater quantity in a shallow cave beside Window Pool. The latter habitat compares with the grottos mentioned by Taylor. The plants consist of sparingly branched filaments (Figs. 13-15), the cells of which measure 60—goy wide and up to 150u long. The cell walls are thick and lamellated. Rhizoidal outgrowths are quite frequent and they usually terminate in a digitate hapteron which often attaches itself to other filaments of the plant. The cell contents do not show much structure, but prominent, dark, parietal bodies are present, which are waste products. A few sporangia are present but the nature of the spores cannot be determined in this preserved material. 188 MARINE ALGAE OF GOUGH ISLAND 1Ou i TT a a a) lO 15 Waa ala: Fics. 13-15. Rhizoclonium ambiguum (Hook. & Harv.) Kiitz. : 13, branched filament ; 14, filaments with rhizoid, y = rhizoid, h = hapteron, w = waste products ; 15, detail of branching. The Gough Island plants are very similar to the type material in the Kew Herbarium and the isotype material in the British Museum Herbarium. Stockmayer was of the opinion that Kiitzing’s Rhizoclonium hookeri and R. africanum were to be referred to the same species ; this view is now generally accepted. The type of R. hookeri is a specimen collected by J. D. Hooker on Kerguelen Island at the same time MARINE ALGAE OF GOUGH ISLAND 189 as he collected his Conferva ambigua but which he labelled C. linum. There is an iostype of R. hooker: in the British Museum Herbarium which in all ways exactly resembles R. ambiguum. This species is often called R. hookert by modern authors, but if, as most authors agree, that species and Rk. ambiguum are to be united, the latter name is correct because the publication of its basionym antedates that of R. hookert. The characteristic features of R. ambiguum are its sparse branching—most species of Rhizoclonium are unbranched—and the filament width, which ranges from 40u to 100u. It is usually found in marine habitats but is also recorded from brackish and fresh water (Taylor, loc. cit.). The type specimen was marine, although Kiitzing incorrectly stated that it came from fresh water. In 1952 Chapman erected a new genus, Rama, which differed from Rhizoclonium only in being sparingly branched ; to this genus he transferred Lychaete novae-zelandiae. Chapman said that Rama was similar to Spongomorpha. However, as the branching in Rama is dichotomous (Figs. 13, 15) as opposed to the characteristic lateral branching of Spongomorpha, they are not really very similar. In 1956 Chapman repeated his description of Rama novae-zelandiae and treated. Rhizoclonium hookeri, which he described as being unbranched and 47-70 wide, as a separate species, retaining it in the genus Riizoclonium. In Bornet & Thuret’s herbarium (in the Muséum National d’Histoire Naturalle, Paris) there is an isotype of Lychaete novae-zelandiae collected by Berggren at Warrington, New Zealand. This plant is sparingly branched and has rhizoids, the cell walls are thick and lamellated, and the filaments measure 60—80y. wide. It closely resembles the type material of R. ambiguum. In my view, Rama novae-zelandiae as well as Rhizoclonium hookert should be united with R. ambiguum. This species is characteristically branched, but this feature alone does not seem sufficient basis for generic separation and I have retained it in the genus Rhizoclonium. RHIZOCLONIUM RIPARIUM (Roth) Harv., Phyc. Brit.: t. 238 (1849). Forma VALIDUM (Hansg.) Foslie in Tromso Mus. Aarshefter 13 : 138 (1890).— Koster in Pubbl. Staz. Zool. Napoli 27 : 351 (1955). (Figs. 16-18.) Conferva tortuosa Dillw., Synops. Brit. Conferv.: t. 46 (1805)?. Rhizoclonium tortuosum (Dillw.) Kiitz., Phyc. Germ.: 205 (1845).—V. J. Chapman in Rhodora 41-: 20 (1939). Rhizoclonium riparium [subsp.?] b) validum Hansg. in Arch. Naturw. Landesdurchforsch. Bohm. 5 (6) : 79 (1886). Rhizoclonium riparium var. validum (Hansg.) De Toni, Syll. Alg. 1 : 279 (1889). Lola tortuosa (Dillw.) V. J. Chapman in Journ. Linn. Soc. Lond., Bot. 55 : 463 (1956). Standoff Rock, on less-exposed part of the rock above the Corallina zone, 18 Feb. 1956, G.I.S.S. 1526. Probably cosmopolitan. This alga was collected from the upper littoral zone of Standoff Rock, where it formed part of the Bostrychietum together with Bostrychia mixta, Rhizoclonium ambiguum and Spongomorpha sp. The plant consists of filaments (Figs. 16—18) 1 The publication date of L. W. Dillwyn’s A Synopsis of the British Confervae is cited by many authors as 1809. In fact it was published in fascicles issued from 1802 to 1809 and for nearly all species the precise date of publication is printed on the plate. 190 MARINE ALGAE OF GOUGH ISLAND which measure 30-37 wide with cells whose length is less than to two and a half times more than their width depending on how recently they have divided. The cell walls are lamellated and fairly thick. The filaments are unbranched but rhizoids are present which terminate in slightly digitate haptera (Figs. 16, 17). Cytological details are not discernible as the cell contents have shrunk. Growth takes place by intercalary cell division and all stages of this process can be seen in the material (Figs. 16, 18). 18 Liiiiititits 17 Ou Fics. 16-18. Rhizoclonium riparium forma validum (Hansg.) Foslie : 16, filament with rhizoid, m = newly divided cell, y = rhizoid ; 17, filament attached to R. ambiguum, h = hapteron ; 18, filament with dividing cell. No previous record of the occurrence of R. ripartum forma validum as a member of a Bostrychia mixta community has been found but R. riparium has often been recorded and some of the records probably refer to this forma. In the British Museum Herbarium there is a specimen of this plant from the spray zone of the shore at Knysna, South Africa, which has some Bostrychia mixta with it. This material was collected by Boodle and annotated by E. S. Barton but was not included by her in her lists of Cape algae. The spray zone is the usual habitat of this particular Bostrychietum (cf. Post, 1959, p. 498). Also in the British Museum Herbarium there is material of R. ripariwm forma validum from Kerguelen Island collected on the ‘‘Erebus”’ and ‘“ Terror’? Expedition and labelled Conferva tortuosa ® perreptans in J. D. Hooker’s handwriting. Joly’s record (1957, p. 59) of R. riparium from Brazil probably refers to forma validum since he records the filaments as being 35—40u wide. MARINE ALGAE OF GOUGH ISLAND IgI The name Rhizoclonium riparium forma validum is usually cited as having been published by Foslie in Wittrock & Nordstedt, Alg. Aq. Dulc. Exsicc. : n. 624 (1884), but the printed label of this number, which was included in Fasc. 14, reads : 624. Rhizoclonium riparium (Roth) Harv. Forma valida, crassit. cell. 26-36, altit. 4 minore—24 plo majore. ayer ar [Habitat]. M. Foslhe. The second line is clearly a comment on the characteristics of this particular specimen and not the last part of the name of a new taxon followed by a diagnosis. This view is confirmed by the list of new taxa in Fasc. 1-20 of the work (Wittrock & Nordstedt, Alg. Aq. Dulc. Exsicc. 21: 16 (1889)) where the entry under n. 624 has the same form as the printed label, whereas on the next page there appears : 944. Kh. hieroglyphicum (Ag.) Kiitz. f. calida (Kiitz.) Wittr. It may well be that, when communicating the specimen to Wittrock and Nordstedt, Foslie intended that it should be treated as a new forma, for in 1890 he treats the name Rhizoclonium riparium forma validum as having been published by them on n. 624 of their set of exsiccata, but this does not alter the fact that they did not do so. The epithet validum was first validly published for an infraspecific taxon in Rhizoclonium riparium by Hansgirg in his Prodromus der Algenflora von Bohmen, the taxon being based on the specimen issued by Wittrock & Nordstedt. Hansgirg nowhere states the rank to which he assigns the infraspecific taxa which he designates with roman letters, but it is clear from the way in which his work is set out that it is higher than variety, designated by Greek letters thus : “ var. 8) ’’, and they must therefore presumably be treated as subspecies. Hansgirg’s attribution of the name to Foslie is based on a mistake and its authorship cannot therefore be cited as “‘ Foslie ex Hansgirg’’; there is no evidence that Foslie ever regarded this taxon as having a rank higher than forma. De Toni reduced Hansgirg’s R. riparium b) validum to the rank of a variety in 1889. Although Foslie in 1890 makes no reference to Hansgirg or De Toni, he cites the same type and his name must be treated as an alteration of rank. This taxon is quite clearly delimited. It consists of plants of the Rhizoclonium type with unbranched filaments 28-45 wide but mostly in the 30-35y range. The cells are longer than wide unless they have divided very recently and usually have fairly thick walls, although wall thickness depends to some extent on external conditions. There has been much confusion as to what the taxon should be called and in the first place the difference between Conferva tortuosa Dillw. and C. tortuosa sensu J. G. Agardh (Alg. Medit. Adr.: 12 (1842), non Dillw.) had to be established. This was done in 1939 by Chapman, who included them in Rhizocloniwm tortuosum (Dillw.) Kitz. and Chaetomorpha capillaris (Kiitz.) Borg. respectively. He in- cluded R. riparium forma validum as a synonym of R. tortuosum. Subsequently, Chapman transferred R. tortuosum to Lola A. & G. Hamel. This genus, as described by A. & G. Hamel (1929, p. 1094), is distinguished from Rhizoclonium by the posses- sion of numerous parietal, not few axial, nuclei ; and from Chaetomorpha by having heterogamous as opposed to isogamous sexual reproduction. The Hamels also said 192 MARINE ALGAE OF GOUGH ISLAND that Lola was unbranched and only rarely had rhizoids. Chapman seems to have transferred R. tortuosum to Lola only because it was unbranched as he gives no details about either its cytology or reproduction. In view of the great variability of branching throughout the Cladophorales, I do not think this reason enough to transfer the plant to another genus. In 1931 (p. 25) G. Hamel included in the genus Lola the species Conferva implexa Harv. and quoted R. tortuosum Kitz. as a synonym of this. He said that it had 8-18 nuclei but made no comments on the reproduction. R. tortuosum is quite distinct from C. implexa and I do not think it is possible to take this statement as evidence of the cytological structure of R. tortuosum. In 1955 Koster published a very careful and exhaustive review of the genus Rhizoclonium in the Netherlands. In this she included R. riparium forma validum. Whilst agreeing with Chapman that this was the same taxon as R. tortuosum, she considered that its separation from R. riparium was untenable. In forma validum she included all plants, with or without rhizoids, with filaments 28-40u wide and in forma riparium plants, with or without rhizoids, with filaments 17-33 wide. Examination of Dillwyn material (not type) of Conferva tortuosa and Wittrock & Nordstedt’s n. 624 on which R. tortwosum forma validum is based, both in the British Museum Herbarium, confirmed that they agree with this concept. The Dillwyn specimen of C. tortuosa is without rhizoids and has filaments 35-40 wide, Wittrock & Nordstedt’s material has many rhizoids and the filaments are about 30u wide. I have not been able to make observations on the nuclei of any of the specimens I have examined as this was not found to be possible on pressed or formalin-preserved material. Sinha (1958) published the results of a cytological investigation on this and other species of the Cladophorales. Included in his results were the following chromosome counts : mitosis meiosis R. riparium (marine) . ‘ ZN 30 yak R. riparium (brackish and freshwater) : 230 R. riparium forma validum (brackish) 2n = .24 R. riparium forma validum (marine) : ai 24 R. tortuosum (marine) . ; , , Zn — 24 He then wrote: ‘ There is cytological support for Stockmayer’s view that R. riparium f. validum is a synonym of R. tortuwosum and that R. riparium is a distinct species whether found in marine, freshwater or brackish situations ” In view of all the evidence outlined above, I am in no doubt that Conferva tortuosa and R. riparium forma validum are synonymous and I am also in agreement with Koster in retaining this forma in the species R. riparium. CLADOPHORA RADIOSA (Suhr) De Toni, Syll. Alg. 1 : 354 (1889).—Papenf. in Journ. S. Afr. Bot. 9 : 79 (1943). (Figs. 19-21.) Conferva vadiosa Suhr in Flora 17 : 741 (1834).—Kiitz., Sp. Alg.: 390 (1849). Cladophora catenifera Kiitz., loc. cit. (1849) ; Tab. Phyc. 3: 24, t. 83 fig. 1 (1853).— E. S. Barton in Journ. of Bot. 34 : 193 (1896).—Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9 : 12, fig. 2 B (1941). MARINE ALGAE OF GOUGH ISLAND 193 Window Rock, frequent below low-water neaps on rocks amongst corallines, 18 Feb. 1956, G.J.S.S. 1512. Church Rock, dominant on vertical rock face, 5 Mar. 1956, G.I.S.S. 1529, 1535. South Africa, Tristan da Cunha, New Zealand, ? Pondicherry, ? Bermuda. C. radiosa grew very abundantly on Church Rock where it dominated the lower littoral zone of the rock and extended into the sub-littoral on the sheer southern face. It was also collected from the sub-littoral zone at Window Rock. Baardseth recorded it from littoral pools and from the sub-littoral zone on Tristan da Cunha. Imm 19 20 21 Fics. 19-21. Cladophora radiosa (Suhr) De Toni: 19, terminal branching ; 20, terminal branching of specimen from Cape of Good Hope labelled C. pellucida by Harvey ; 21, habit of Gough Island plant. 194 MARINE ALGAE OF GOUGH ISLAND The plant grows in dense, erect clumps up to 9 cm. high, which consist of numerous branched fronds. From the base of the fronds downward-growing rhizoids are produced which interweave and form a mat on the rock. The lowermost cell of the frond is very large, measuring up to 40 mm. long x 0:45 mm. wide with thick, lamellated walls up to 70y thick. The plants (Fig. 21) branch from about two-thirds of the way up ; the main branches are dichotomous and the secondary ones (Fig. 19) are dichotomous or trichotomous. The blunt apical cells measure from 700u. to I,300u in length and from 200y. to 330 in width, the shorter ones tending to be the widest. The upper cells measure up to 2,000n long x 200-330u wide and the walls are about 45u thick. There is a tendency for the cells to be somewhat swollen at the cross walls and more particularly at the dichotomies. An occasional swollen, barrel-like cell is present, but there are not enough of these to make them a characteristic feature of the plant. The nomenclature and taxonomy of this little-known species of Cladophora have become somewhat confused. Barton pointed out that Conferva radiosa and Cladophora catenifera are both the same taxon, and Papenfuss has confirmed this. Papentuss also attributes the combination Cladophora radiosa to Kiitzing. However, although Kiitzing in the Specres Algarum included this species under the general heading of Cladophora, he put it in Species Inquirendae at the end of the genus and retained the generic name Conferva for it. The first author to call the plant Cladophora vadiosa was therefore De Toni. The species is characterized by the very long basal cell, the rhizoidal base and the dichotomous to trichotomous nature of the branching. It bears some resem- blance to C. prolifera (Roth) Kiitz. but is much less wiry ; also that species, when pressed, is green whereas C. vadiosa is brown. The type specimen of C. catenifera is a specimen from the Cape of Good Hope collected by Harvey and sent to Kiitzing as C. pellucida. There are two isotypes of C. catenifera in the British Museum Herbarium (Fig. 20) with which the Gough Island material agrees quite well. The main difference lies in the rather longer upper cells of the Harvey material. The relative shortness of the cells of the Gough Island plant may be due to the extremely severe conditions of thelocality. Baardseth points out that the apparently barrel-like shape of the cells of C. catenifera as illustrated by Kiitzing was probably due to the preservation ; it was certainly not a noticeable feature of the Harvey material in the British Museum Herbarium after it had been soaked and gently warmed in a weak detergent solution. © Chapman (1956, p. 446) recorded this species from New Zealand and pointed out its similarity to C. feredayi Harv. Chapman gives no cell measurements and his drawing is not clear enough to enable comparison of the New Zealand material with material from other localities. CLADOPHORA RUPESTRIS (L.) Kiitz., Phyc. Gen. : 270 (1843). (Figs. 22-25.) Conferva rupestris L., Sp. Pl. 2 : 1167 (1753). Isolda Rock, frequent in mid-littoral rock pools, 6 Mar. 1956, G.I.S.S. 1545. Europe, North America, Brazil. MARINE ALGAE OF GOUGH ISLAND 195 Plants of this species grew in abundance attached to the base of Corallina plants in littoral rock pools on Isolda Rock. The plants (Fig. 22) are up to 5 cm. high and are fairly stiff; they are composed of long, straight, branching filaments. Branching (Figs. 23, 24) is mainly dichotomous but sometimes trichotomous, particularly in the lower parts of the plant ; there is thus no single main axis. The angle between the branches is always acute and the plants are not spreading. At the dichotomies, the basal cells of the branches coalesce (Fig. 24), a charac- teristic of this species. Growth is mainly intercalary. The apical cells are obtuse (Fig. 24) and measure 200-350 (-500)u long x 60-gou wide. Cells in the 4cm ZL l | | | ! 25 100u Figs. 22-25. Cladophorarupestris (L.) Kiitz. : 22, habit ofplant ; 23, terminal branching; 24, detail of branching and apex, a = apical cell ; 25, sporangia. 196 MARINE ALGAE OF GOUGH ISLAND upper branches measure 300-700u long x up to 200y wide and cells in the lower- most part of the plant measure up to 1,050u long x 150u wide. All the cells have the characteristically very thick, lamellated wall (Fig. 25) up to 45u thick the structure of which has been investigated by Frei & Preston (1961). The material was reproducing (Fig. 25), and spores escaping through a pore in the cell wall can be seen. In this preserved material it is not possible to determine the type of spore. Mr. C. van den Hoek, who kindly determined the material, said that had it come from Europe he would have had no hesitation in naming it C. rupestris but that it might be better to give it this name only tentatively in view of the geographical isolation of the locality from the main area of distribution. There have been various records of this species from the Southern Hemisphere, but the only reliable one is probably that of Joly (1957, p. 57) from Brazil. J. D. Hooker (1845-47, p. 495) and Dickie (1876, p. 47) recorded the species from Kerguelen, but the material on which these reports are based (in the Kew and British Museum Herbaria respectively) is all Sbongomorpha sp. There are several records of C. rupestris from South Africa but these have also been re-identified as other taxa: for example, Hohenacker, Meeresalg. 465 and Tyson, Alg. Exsicc. Austro-Afr. 13 have both been identified by Papenfuss as C. capensis. Hariot (1889, p. 18) recorded the species from Fuegia, but I was not able to find the specimen on which this record was based in the Herbarium of the Muséum National d’Histoire Naturelle, Paris. On present knowledge, therefore, it seems that the distribution of C. rupfestris in the southern hemisphere is restricted to the Atlantic Ocean. CODIACEAE! CODIUM FRAGILE (Suringar) Hariot in Miss. Sci. Cap Horn 5 : 32 (1889). Anthocodium fragile Suringar in Ann. Mus. Bot. Lugd.-Bat. 3 : 258 (1867) ; Alg. Jap. Mus. Bot. Lugd.-Bat.: 23, t. 8 (1870) (reimpr. in Natuurk. Verh. Holl. Maatsch. Wet., Ser. 3, 11:235-4..8 (1872)). Slaughtered Seal Bay, common below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1506. West Point Reefs, frequent in rock pools below spray zone, 6 Mar. 1956, G.I.S.S. 1540, 1541. Cosmopolitan. This species was common in littoral rock pools on Gough Island and was collected from West Point Reefs and Slaughtered Seal Bay. The material collected consists of large, robust plants up to 30 cm. tall with branches up to 2 cm. broad. The branching is irregular ; it is mostly dichotomous but sometimes numerous small branches are given off over a very small area or sections of branch present a warty appearance. The utricles are mucronate, and sporangia are present in G.J.S.S. 1506. Although the species C. fragile is so widespread, Silva has distinguished a number of subspecies each of which has a fairly restricted distribution. As he has not finished his work on the species from the South Atlantic, it is not possible to say which subspecies the Gough Island material represents. 1 Dr. P. C. Silva is at present working on Codium from the South Atlantic and is including the Gough Island material in his survey. MARINE ALGAE OF GOUGH ISLAND 197 CoDIUM ADHAERENS Agardh, Sp. Alg. 1 : 457 (1822).—O. C. Schmidt in Bibl. Bot. gi: 26, figs. 6, 7 (1923).—-Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9 : 15, fig. 5 D (1941). Agardhia adhaerens (Agardh) Cabrera in Phys. Sallsk. Lund Arsber. 1823 : 99 (1823). Church Rock, dominant over considerable areas of shaded vertical rock face, 5 Mar. 1956, G.I.S.S. 1532, 1535. Distribution not certain, probably chiefly Atlantic. This species was collected from Church Rock where it was growing on a shaded, vertical rock face. The specimens agree well with Baardseth’s description of the plants from Tristan da Cunha. PHAEOPHAYTA ECTOCARPACEAE FELDMANNIA GLOBIFERA (Kiitz.) Hamel in Bot. Notis. 1939 : 67 (1939).—Baards. in Res. Norw. Sci. Exped. Tristan da Cunha g : 19, fig. 8 c—D (1941).—Kuckuck in Helgol. Wiss. Meeresuntersuch. 6: 179, figs. 5-10 (1958). (Figs. 26-34.) Ectocarpus globifer Kiitz., Phyc. Gen.: 289 (1843).—Bornet in Bull. Soc. Bot. France 38 : 358, t. 7 figs. 6, 7 (1891).—Borgesen in K. Danske Vid. Selsk. Biol. Medd. 6 (2) : 48, figs. 25, 26 (1926). Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1506. Europe, Canary Islands, Tristan da Cunha. The plant was found growing in tufts on Codium fragile in rock pools at Slaughtered Seal Bay. The tufts are about 2 cm. high and have a rhizoidal base, the filaments of which weave in among the Codium utricles and anchor the tufts firmly. Kuckuck described two sorts of tuft in European material and the Gough Island plants re- semble his northern form in being rather compact and having a dense “ kernel ”’ of branched filaments where most of the reproductive bodies are found surrounded by an outer fringe of unbranched, lightly pigmented filaments. The erect filaments are branched, ribbon-like and 40-50u wide. These dimensions are similar to those given by Baardseth for the Tristan da Cunha material and by Bergesen for that from the Canary Islands but are rather smaller than in the Euro- pean specimens described by Kuckuck as being about 70 wide. The meristems are intercalary except in some young side branches (Fig. 26), which are still entirely meristematic. The meristems and branches occur in the lower two-thirds of the tufts and beyond this region the filaments extend into long, scantily pigmented hairs. Branching is irregular ; many branches arise from the growth of the stalk cell of old sporangia (Fig. 27). The branches are initially at right-angles to the filaments which bear them, but gradually the lower wall of the basal cell elongates more than the upper wall and the angle thus becomes acute. The plants were reproducing abundantly ; both unilocular and plurilocular sporangia are present. Both types of sporangium can be seen in a single tuft, but the two types are not found on the same filament, a condition illustrated by Kuckuck (loc. cit., figs. 9, 10). 198 MARINE ALGAE OF GOUGH ISLAND ee ET eal | i Es ea LS ed Fics. 26-34. Feldmannia globifera (Kiitz.) Hamel : 26, meristematic side branch ; 27, branch growing from basal cell of plurilocular sporangium ; 28, unilocular sporangium ; 29, unilocular sporangium with two-celled stalk ; 30, unilocular sporangium growing in the remains of an old one ; 31, unilocular sporangium with liberated spores ; 32, pluri- locular sporangia ; 33, spores being liberated from a plurilocular sporangium ; 34, germinating spores. Unilocular sporangia have only been recorded once previously : by Kuckuck on material from Rovigno (now Rovinj) on the Adriatic coast. In the Gough Island plants these sporangia (Figs. 28-31) are globose, and mature ones have a diameter of about 50u. With very few exceptions the sporangia have a one-celled stalk, but occasional two-celled stalks (Fig. 29) are present. Sometimes a new sporangium develops inside the ruptured wall of an old one (Fig. 30). Unilocular sporangia are usually borne laterally on the filaments but occasionally they develop terminally on young side branches. No unilocular sporangia containing ripe MARINE ALGAE OF GOUGH ISLAND 199 spores were found, and it was accordingly not possible to ascertain how they developed and were liberated. These unilocular sporangia differ somewhat from those described by Kuckuck as being sessile and 70-go0u high x 50-70u wide. Kuckuck also described a second type of unilocular sporangium. This type was stalked, globose and about 7oy in diameter, and was thus more like the Gough Island sporangia ; it differed however in being found characteristically on the same filament with plurilocular sporangia. The plurilocular sporangia (Figs. 32, 33) are nearly always stalked and lateral but occasional sessile or terminal ones are seen. They are solid-looking, ovoid bodies and mature ones measure up to 80u wide xX I00y long ; this is very similar in size and shape to those described by Kuckuck, Bornet and others. The spores differentiate in the sporangium and are liberated by the rupture of the apex of the sporangium (Fig. 33). Plurilocular sporangia also grow inside the ruptured wall of old sporangia. Many liberated spores are present in the material (Fig. 34) ; several had germinated and others had developed into the few-celled plantlet stage on the surface of the filaments. It is also common to find spores germinating while still inside the sporangia. These features were also commented on by Bornet. The Gough Island plants therefore agree closely with this species as it is known from other parts of the world. Dr. P. Kornmann, who kindly confirmed the identification, remarked on this close similarity. Baardseth’s record of this species from Tristan da Cunha was the first from the Southern Hemisphere but the species is well known from Europe and the Canary Islands. Good descriptions of the plant have been given by Bornet, Bgrgesen and Kuckuck. GIFFORDIA GRANULOSA (Smith) Hamel in Bot. Notis. 1939 : 67 (1939).—Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9g: 21, fig. 8 A (1941).—Kuckuck in Helgol. Wiss. Meeresuntersuch. 8 : 119, figs. 8,9 (1961). (Figs. 35, 36.) Conferva granulosa Smith in Sowerby, Engl. Bot. 33 : t. 2351 (1811). Ectocarpus granulosus (Smith) Agardh, Syst. Alg.: 163 (1824).—Sauvageau in Bull. Stat. Biol. Arcachon 30: 59, figs. 13, 14 (1933).—Lindauer, Chapman & Aiken in Nova Hedwigia 3 : 145, t. 82 fig. 4 (1961). Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1506. Widely distributed in warm and temperate waters. A small quantity of this plant was found growing as an epiphyte on Codium fragile in rock pools at Slaughtered Seal Bay ; it was intermingled with the much more abundant growth of Feldmannia globifera. The soft, feathery plants are up to 2°5 cm. tall and show the characteristic opposite branching (Fig. 35) of the species and the cortication of the lower part of the main axis by downward-growing rhizoids (Fig. 36). The main axis and branches measure 40-70u wide, which is rather narrow for this species (cf. up to 100u given by Kuckuck) and the material generally is in rather poor condition. The cells are barrel-shaped and in most cells the contents have disintegrated ; in some, however, the typical discoid chromato- phores can be clearly seen. The meristematic regions (Fig. 35) are intercalary 200 MARINE ALGAE OF GOUGH ISLAND 35 Fics. 35, 36. Giffordia granulosa (Smith) Hamel: 35, upper part of plant, m = meri- stematic region, s = sporangium ; 36, rhizoids growing from lower cells of plant. and in these parts the cells are wider than long. In the more mature parts, the cells are more elongated, especially in the branch apices where they are also very lightly pigmented. The apices are not prolonged into long hairs as they often are in this species, but are very like those illustrated by Sauvageau. Sessile sporangia are present on the upper sides of the side branches (Fig. 35). The young sporangia are quite clearly divided and plurilocular ; although divisions are not apparent in the mature sporangia, they are almost certainly all plurilocular. The apparent absence of cell walls is probably due to the generally poor state of preservation of the material. The sporangia are ovoid and mature ones measure 40-60u. wide x 60-80 long ; this is within the usual size range for this species. MARINE ALGAE OF GOUGH ISLAND 201 GIFFORDIA MITCHELLIAE (Harv.) Hamel in Bot. Notis. 1939 : 66 (1939).—Baards. in Res. Norw. Sci. Exped. Tristan da Cunha g : 21, fig. 7 E (1941).—Levring in Act. Univ. Lund., Ny Foljd, Avd. 2, 56 (10) : 14 (1960). (Figs. 37-41.) Ectocarpus mitchelliae Harv., Ner. Bor.-amer. 1 : 142, t. 12 fig. G (1851) (reimpr. in Smithson. Contrib. Knowl. 3 (4) : 142, t. 12 fig. G (1852)) (‘‘ Mitchellae ’’)—Beorgesen in Dansk Bot. Ark, 2 (2) : 6, figs. 3, 4 (1914) (reimpr. ut Alg. Dan. W. Ind.: 162, figs. 129, 130 (1914)) — Joly in Univ. Séo Paulo Fac. Fil. Ci. & Let. Bol. 217, Bot. 14 : 72, t. 5 figs. 7, 7a (195°7). Ectocarpus virescens Thur. ex Sauvageau in Journ. de Bot. 10 : 124, figs. 1-7 (1896).— Hamel, Phéophyc. France : 29, fig. 7 (1931). a7 4l 10 Fe eer oF i ear ee Fics. 37-41. Giffordia mitchelliae (Harv.) Hamel: 37, rhizoidal base of plant ; 38, branching, m = meristematic region ; 39, detail of cell contents, d = discoid chromato- phore, m = nucleus ; 40, plurilocular sporangium ; 41, dehisced plurilocular sporangium, m == meiospore. 202 MARINE ALGAE OF GOUGH ISLAND Dell Rocks Beach, occasional in sheltered places between boulders and in shallow pools, 12 Apr. 1956, G.I.S.S. 1558. Widely distributed in warm and temperate waters. This species was found growing on rocks in shallow pools and in sheltered places between boulders at Dell Rocks Beach. The plants are about 5 cm. tall and soft and feathery like the form from sheltered places on Tristan da Cunha reported by Baardseth. The base of the plant is rhizoidal (Fig. 37) and the upper part richly and irregularly branched (Fig. 38) ; the side branches are sometimes secund as shown in the figure, but this is not a pronounced feature of the plant. The main axes measure up to 40u. wide with cells up to twice as long as wide. The cells have discoid chromatophores (Fig. 39). Meristematic regions occur at the base of the side branches (Fig. 38) ; these branches usually taper somewhat towards the apex but the apical cells are always obtuse. Plurilocular sporangia are present (Figs. 40, 41) ; these are usually sessile and lateral on side branches but are occasionally found on the main axis or terminally on young side branches. The sporangia are elongated with rather a blunt apex ; they vary considerably in size, the largest measured being 1204 xX 30p. The loculi are about 6-7u high x 7—8y wide and thus correspond (like the ones described by Baardseth from Tristan da Cunha) to the meiospore loculi of Sauvageau (tom. rol ces UIE SPHACELARIACEAE SPHACELARIA FURCIGERA Kiitz., Tab. Phyc. 5 : 27, t. go fig. 2 (1855).—Reinke in Bibl. Bot. 23: 14, t. 4 figs. 5-13 (1891).—Skottsb. in Wiss. Ergebn. Schwed. Siidpol.-Exped., Bot. 4 (6) : 58 (1907).—Irvine in Trans. Bot. Soc. Edin. 37 : 28 (1956).—Levring in Act. Univ. Lund., Ny Foljd, Avd. 2, 56 (10) : 16 (1960). (Fig. 42.) Fic. 42. Sphacelaria furcigera Kiitz. : branching, propagule on left. MARINE ALGAE OF GOUGH ISLAND 203 Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1506. Widely distributed in warm, temperate and colder waters. This species was found growing as an epiphyte on Codium fragile in rock pools at Slaughtered Seal Bay. The plant is firmly anchored to the host by means of rhizoids which weave in among the Codium utricles and form knotted masses at intervals. Reinke shows the rhizoids of S. furcigera actually penetrating the tissue of the Sargassum on which it is growing, but the nature of the Codium thallus is such that the Sphacelaria cannot penetrate its tissues. The plants measure up to 12 mm. high ; the upper parts are sparingly branched (Fig. 42) and hairless; the filaments measure 20-40u wide, which accords well with the dimensions given by Reinke and Irvine ; the plants also bear a close resemblance to the figures of Kiitzing. Sporangia are absent but occasional bi-radiate propagules are present (Fig. 42), the appearance of which does not differ markedly from ordinary side branches. Some of the material was sent to Dr. Irvine who confirmed the identification and commented as follows : “I could find no well-developed propagules, though a few showed clear signs of developing two arms only. The young propagules appear to be second or even third generation from the base and there is no sign of hairs either on the propagules or elsewhere. The general dimensions, mode of branching, number of longitudinal divisions of segments (1-3) and form of propagules seem right for S. furcigera ; the lack of hairs is probably unimportant as in general I have found that their presence is very variable, apparently depending on the tem- perature, at least to some extent, i.e. the higher the temperature the more hairs. The only point that does worry me a little is the presence here and there of quite distinct little holdfasts like those of S. pennata ; in general though, the axes spring from dispersed rhizoids as is usual in this species. There are also very occasional secondary transverse divisions, but I do not really think that this is of much significance .”’ Skottsberg recorded the species from Tierra del Fuego and the Falkland Islands. PUNCTARIACEAE COLPOMENIA SINUOSA (Roth) Derb. & Solier in Suppl. Compt. Rend. Acad. Sci. Par. 1: II, t. 22 figs. 18-20 (1856).—De A. Saunders in Proc. Calif. Acad. Sci., Ser. 3, Bot. 1 : 164, t. 32 figs. 7, 8 (1898).—Setch & Gardn. in Univ. Calif. Publ. Bot. 8 : 539, t. 45 figs. 82-86 (1925). (Figs. 43-45; Plate 16 B.) Ulva sinuosa Roth, Catalect. Bot. 3 : 327, t. 12 (1806). Midshipman Rock, gelatinous form frequent in lower littoral, 18 Feb. 1956, GIS5. 1576. Probably cosmopolitan. This species was collected only once from Gough Island, where it grew quite abundantly on rocks of the lower littoral zone of the Midshipman Rock. The plants form brown, warty, gelatinous cushions which are up to I cm. thick (Plate 16 B). 204 MARINE ALGAE OF.GOUGH ISLAND In section (Figs. 43-45) the plant body consists of an outer layer, one or two cells deep, of more or less cubic cells up to about 12y in diameter, which are usually pigmented. Inside this cortex is the medulla, consisting of colourless, irregular cells which increase in size towards the centre. The middle of the thallus is hollow. Pigmented hairs are present ; these grow fairly deeply embedded in the thallus (Fig. 45) in cryptostomata and are about 7p in diameter. Plurilocular sporangia were found over much of the surface of the thallus ; the sporangia (Figs. 43, 44) arise from the outer cells of the cortex and consist of sixteen cells in two rows of eight. Mature sporangia measure up to 40u high x 8y wide. Paraphyses are present between some of the sporangia ; they are large and some- what clavate and measure about 50u X I5yu. The cortex and medulla together are about 160 thick. ; ; 2208 @)>ys Nu} i oe, lOul ee oy b F | i | L 44 43 Fics. 43-45. Colpomenia sinuosa (Roth) Derb. & Solier: 43, vertical section through thallus, p = paraphysis, ms = mature sporangium ; 44, vertical section through thallus, p = paraphysis, ys = young sporangium ; 45, vertical section through thallus, h = hair, c = cryptostoma. The type form of C. sinuosa is a balloon-like, hollow ball. Several varieties have however been described which closely resemble the tuberculate Gough Island plant. Saunders (tom. cit. : 164) described a new species, C. tuberculata, which was subse- quently reduced to C. sinuosa forma tuberculata by Setchell & Gardner (tom. cit. : 541). This forma is extremely like the Gough Island plant in external form but differs in microscopic detail. The cortex is about four cells deep and the sporangia, which measure up to 25 tall, have cells in a single row, not a double row. The MARINE ALGAE OF GOUGH ISLAND 205 Gough Island plant also strongly resembles Iyengaria stellata (Berg.) Borg. (1939, p. 92) externally, although it is smaller than that species (which includes Colpomenta capensis Levring (1938, p. 20)) in both macroscopic and microscopic detail. The generic status of [yengaria is rather dubious as it is separated from Colpomenta only by the absence of paraphyses among the plurilocular sporangia and an allegedly somewhat different growth process in the thallus. Re-investigation of [yengaria is needed ; meanwhile the Gough Island plant would seem to differ from it in the possession of sporangial paraphyses. In conclusion, I think that this alga comes within the generally accepted variation range of C. sinwosa, and while it does not externally resemble the type form of this species, it is yet not sufficiently like any of the described varieties to be referred to one of them. Tuberculate forms of C. sinuosa have been mentioned by various authors and of particular interest is that described and illustrated by Joly (1957, p. 84, t. 4 fig. 3 B) from Brazil, which appears to be very similar, macroscopically, to the Gough Island plant. Unfortunately, Joly does not give an account of the microscopic structure and further comparison is thus impossible. SCYTOSIPHON LOMENTARIA (Lyngb.) Endl., Gen. Pl., Suppl. 3 : 25 (1843).—Setch. & Gardn. in Univ. Calif. Publ. Bot. 8 : 531, t. 44 figs. 72, 74 (1925).—Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9: 31 (1941).—Silva in Univ. Calif. Publ. Bot. 25 : 258 (1952). Chorda lomentaria Lyngb., Tent. Hydrophyt. Dan.: 74, t. 18 fig. E (1819).: Dell Rocks Beach, near low-tide mark on boulder beach, 19 Feb. 1956, G.I.S.S. I51r ; same locality, frequent in pools and protected channels just above mid-tide level, 12 Apr. 1956, G.I.S.S. 1559. Midshipman Rock, in standing water in upper littoral, 18 Feb. 1956, G.J.S.S. 1519. West Point Reefs, in bird-fouled spray-zone rock pools, 6 Mar. 1956, G.I.S.S. 1536. Cosmopolitan. Of the four collections of this plant made on Gough Island, two (r5z9 and 1536) are of the form which Setchell & Gardner designate as cylindricus minor and the other two are of that which they call typicus. The plants of the latter come from the mid and lower shore while the others are from pools, often bird-fouled, on the upper shore. Baardseth recorded both forms from Tristan da Cunha. LESSONIACEAE MACROCYSTIS PYRIFERA (L.) Agardh, Sp. Alg. 1: 47 (1820).—Skottsb. in Wiss. Ergebn. Schwed. Siidpol.-Exped., Bot. 4 (6) : 80, figs. 91-170, t. 9 (1907).— Womersley in Univ. Calif. Publ. Bot. 27 : 118, tt. 1, 3, 4 (1954).—G. A. Knox in Proc. R. Soc. Lond., Ser. B, 152 : 588 (1960). (Figs. 46, 47.) Fucus pyriferus L., Mant. Pl. Alt.: 311 (1771). Between The Admiral and The Commodore, forming a belt around the island between 18 and 55 m., also in gullies, 22 Dec. 1955, G.J.S.S. 1502. Near Penguin Island, at 55 m., 2 Jan. 1956, G.I.S.S. 1509. 206 MARINE ALGAE OF GOUGH ISLAND 47 Fics. 46, 47. Macrocystis pyrifera (L.) Agardh: 46, holdfast, dh = dichotomous haptera ; 47, air bladder. Circum-subantarctic between latitudes 40° S. and 60° S. ; west coast of South America as far north as Peru ; west coast of North America. Two collections of M. pyrifera were brought back from Gough Island including an intact holdfast and an apical frond with the leaflets joined together terminally. The holdfast (Fig. 46) shows the dichotomous haptera typical of this species. The pear-shaped air bladders (Fig. 47) at the base of the leaflets are also a characteristic feature. Plants of this species formed a band round Gough Island from about 18 to about 55 m. immediately below the Durvillaea band ; it was absent only from the very exposed part of the western end of the island. DURVILLAEACEAE DURVILLAEA ANTARCTICA (Cham.) Hariot in Notarisia 7 : 1432 (1892).—Skottsb. in Wiss. Ergebn. Schwed. Siidpol.-Exped. 4 (6) : 140, fig. 172 (1907).—Naylor in Ann. of Bot., New Ser., 13 : 285, figs. 1-8, tt. 6, 7 (1949) ; in Trans. R. Soc. New Zeal. 80 : 277 (1953).—Lindauer, Chapman & Aiken in Nova Hedwigia 3 : 284, t. 65 (1961). Fucus antarcticus Cham. in Choris, Voy. Pitt. Monde, Travers. de Cronstadt Chili : 7, t. 7 1822). ie utilis Bory in D’Urville, Fl. Il. Malouines : 17, 22 (Dec. 18251) (reimpr. in Mém. Soc. Linn. Par. 4 : 589, 594 (Jan. 18261)), nom. invalid. 1 For these dates see under Jvidaea undulosa, p. 224 MARINE ALGAE OF GOUGH ISLAND 207 Durvillaea utilis Bory in Dict. Cl. Hist. Nat. 9 : 192, t. 49 (Feb. 18261), nom. illegit.; in Duperrey, Voy. Monde La Coquille, Bot., Crypt. : 65 (Dec. 1827), t. 1 (Jan. 1827), t. 2 fig. 1 (Mar. 1827).—Hook. & Harv. in Hook. f., Bot. Antarct. Voy. I : 176 (1845), 454 (1847).—Hariot in Miss. Sci. Cap Horn 5 : 53 (1889). Outer Dell Rocks, dominant from mid-tide level to 2 m. below low-tide level on immobile rocks, occasional in deep rock pools, 18 Feb. 1956, G.I.S.S. 15r0. Sub-antarctic Islands, South America, New Zealand. This species dominated the lower littoral and sub-littoral zones of the Gough Island shores. Two collections were made, one of young plants showing the shield- like attachment organ and divided, leathery frond ; the other a fragment of a mature frond showing the spongy, honeycomb texture. The young plants had no reproductive organs but the mature frond bore female conceptacles. The plants grew to an immense size with fronds up to 3 m. long and holdfasts at least 7-5 cm. thick. D. antarctica is a very characteristic species of the sub-antarctic sub-littoral ; its distribution is mapped by Knox (1960, p. 589). It does not grow as far north as Tristan da Cunha in the Atlantic. In the prefatory part of the Flore des Iles Malouines, D’ Urville gives an account of the general appearance of an alga and concludes: “‘ M. Chamisso I’avait trés-bien décrit sous le nom de Fucus antarcticus ; mais M. Bory y a reconnu le type d’un genre nouveau, et a jugé a propos de l’appeler Durvillaea utilis.’’ In the list of species, compiled, as far as the cryptogams were concerned, by Bory, the entry relating to this species reads: ‘27. Durvillaea utilis. Bory. Porro. Le Gentil (Voyage, tome II, pl. 3.) Fucus antarcticus. Chamisso (Voyage pitt. de Choris.) Laminaria porroidea. Lamour. (Dict. class. hist. nat. art. Géographie botanique).”’ “ Porra ’”’ of Le Gentil is a vernacular name and the botanical name provided for the species in his work is a “ pre-Linnaean”’ one: Fucus ramis ex tuberculo rotundo extentibus, foliis planis, profunde crenatis, pediculatis, pediculis uno versu dispositis (Le Gentil de la Galaisiére, Voy. Mers Inde 2: 215, t. 3 (1781)). The name Laminaria porroidea was never validly published, although Lamouroux (Dict. Sci. Nat. 25 : 189 (1822)) had published Laminaria porra, based on Le Gentil’s “ Porra’’. In his full account of Durvillaea in the Voyage . . . Coquille, Bory points out that D’Urville’s account was based primarily on a different alga, Laminaria buccinalis (L.) Lamour. (= Ecklonia maxima (Osb.) Papenf.), and that Le Gentil’s “‘ Porra ”’ also was not Durvillaea but probably Macrocystis. He states, however, that Lamouroux had imperfect specimens of Durvillaea from Valparaiso in his herbarium labelled Laminaria porroidea. The generic name Durvillaea, and hence the specific name D. utilis, is not validly published in the Flore des Iles Malouines. D’Urville’s remarks can scarcely be treated as a description. Even if they could, they would be a description of the species already described as Fucus antarcticus by Chamisso, and neither they nor Chamisso’s original description can be regarded as a descriptio generica-specifica, for the genus, though monotypic, is not founded on a new species. 1 For this date see under Ividaea undulosa, p. 224 208 MARINE ALGAE OF GOUGH ISLAND When the generic and specific names were validly published by Bory in the Dictionnaire Classique, he again cited Fucus antarcticus as a synonym of Durvillaea utilis, the epithet of which is consequently superfluous and illegitimate. In this, and throughout his works, Bory spelled the generic name Durvillaea, indicating that he intentionally latinized D’Urville’s name as “ Durvillaeus’’. The “ correction ”’ to ‘‘ Durvillea’’, the form usually adopted, cannot therefore be justified under Art. 73 of the International Code. RHODOPHYTA BANGIACEAE BANGIA FUSCOPURPUREA (Dillw.) Lyngb., Tent. Hydrophyt. Dan. : 83, t. 24 fig. c (1819). Conferva fuscopurpurea Dillw., Synops. Brit. Conferv.: t. 92 (1807). Standoff Rock, from upper littoral, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1527. Cosmopolitan. Only one collection of this species was made and that was from the supra-littoral zone of Standoff Rock where the Bangia was growing with Porphyra tristanensis, Hildenbrandia sp. and Verrucaria sp.—a very characteristic supra-littoral community. Although it was not recorded or collected from any other part of Gough Island, it is a rather inconspicuous plant when growing and probably occurs elsewhere on the island. PORPHYRA TRISTANENSIS Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9g: 36, figs. 14,15 F (1941). (Figs. 48-51.) Slaughtered Seal Bay, 22 Dec. 1955, G.I.S.S. 1503. Midshipman Rock, upper shore, 18 Feb. 1956, G.I.S.S. 1513. Standoff Rock, from upper littoral, 18 Feb. £056, GI'S.S..7527. Tristan da Cunha group of islands. P. tristanensis is the dominant plant of the Gough Island supra-littoral, where it forms a zone together with a lichen, Verrucaria sp. Some of the material was sent to Miss J. M. Graves who has studied South African Porphyra. I am grateful to her for drawing Figs. 48-50 and for the following description and comparison with P. capensis Kiitz. : ‘‘ The specimens correspond quite well with P. capensis as that species is found in moderately warm water (e.g. False Bay, Cape). The holdfast has the typically umbilicate form. In section [Fig. 48] the vegetative cells appear to have one chromatophore. The cells are not so markedly elongate as are those of typical P. capensis but I think they fall within the variation range. The protoplasts have unfortunately shrunken away from the cell walls which makes measurement of the cell dimensions a little difficult, but as far as I can make out, the vegetative cells are from 15-25 wide x 50-60 long as seen in transverse section. The ratio length/ breadth therefore varies from 2:1 to 4:1. The majority of the cells measured are about three times as high as broad as is the case in warm-water P. capensis. A MARINE ALGAE OF GOUGH ISLAND AD\Aa (BDO OA «@ D OBESSRHIET | eee is V Fics. 48-51. 50 Porphyra tristanensis Baards. : 48, transverse section through thallus, vegetative cells ; 49, surface view of thallus, c = carpospores ; 50, transverse section through thallus of G.IJ.S.S. 1513, ¢ = carpospores ; 51, transverse section through thallus of G.I.S.S. 1503, ¢ = carpospores. comparison of measurements of P. capensis and the Gough Island plants can be seen in the following table : Length of Width of Thickness vegetative cells in | vegetative cells in of transverse section | transverse section thallus Cold-water 100-135 | 12-26. 130-180 P. capensis | Warm-water 47-68 u. II-19p 70-80 P. capensis Gough Island 50-60. 15-25 up to gou Glas 1513 210 MARINE ALGAE OF GOUGH ISLAND “The plants are all carposporophytic. The number and arrangement of carpospores in the parent cell is similar to that in warm-water P. capensis. There is a total of sixteen carpospores per cell and these appear as two rows of four cells each [Fig. 50] in transverse section. The arrangement in surface view [Fig. 49] is somewhat more irregular than in P. capensis. The distribution of the fertile area on the thallus—an irregular marginal zone which is slightly broader at the distal end than at the sides—corresponds well with P. capensis. “ To sum up, I would say that the available material corresponds quite well with warm-water P. capensis.” I myself thoroughly investigated the plants in G.J.S.S. 1503 and found them to agree with Miss Graves’s description of G.J.S.S. 1513 as to the vegetative characters. In this material also, only carposporophytic plants are present and in these plants there are often up to 32 carpospores per parent cell (Fig. 51) ; this also comes within the variation range of P. capensis (see Isaac, 1957). The taxonomy of the genus Porphyra is extremely confused ; with such a simple structure, the number of characters is small and there seems to be considerable overlapping between species. Workers on European members of the genus are gradually elucidating the taxonomy by studying the appearance and development of living plants and it seems that only by doing this in conjunction with careful study of type material will it be possible to define specific limits within the genus. Unfortunately in this case only preserved material is available for study and all the plants are female. The plants are thus probably dioecious but it may be that it is a monoecious species in which the spermatia have not yet developed. I think there is little doubt, however, that this species is the same as that described by Baardseth from Tristan da Cunha and that this is probably the same as P. capensis from South Africa. Levring (1953, p. 464, and 1956, p. 410) published accounts of Porphyra in Australia and New Zealand and found that, of the two lithophytic species, P. umbilicalis J. G. Agardh was dioecious with usually 8 carpospores per parent cell while P. columbina Montagne was monoecious with 32 carpospores per parent cell. Other workers, however, consider that P. wmbilicalis can be either monoecious or dioecious and this probably also applies to P. columbina as it does to P. capensis. It seems, therefore, best to call the Gough Island plants P. tristanensis until the relationship between this and other southern cold-temperate species has been elucidated by the study of living material of them all. GELIDIACEAE GELIDIUM REGULARE Baards. in Res. Norw. Sci. Exped. Tristan da Cunha g : 52, figs. 23, 24 A (1941). (Fig. 52.) Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1507. Tristan da Cunha (Nightingale Island). G. regulare was found in littoral rock pools at Slaughtered Seal Bay ; it was growing with Corallina officinalis, Dermatolithon nodulosum, Lithothamnium lamellatum and Centroceras clavulatum. MARINE ALGAE OF GOUGH ISLAND 211 The plants (Fig. 52) measure up to 6 cm. tall. They show the distinct main axis and abundant, tripinnate branching described by Baardseth. Tetrasporangia are present. Baardseth pointed out that this species is very similar to the European species G. latifolium (Grev.) Born. & Thur., but until more is known of the distribution and variation range of that species it seems best to give the South Atlantic plants a different name. 5mm Q) ( WA 4, \ 9 y, fi X NUN HE Bs NY Sie = Ave y SE v4 f s) YZ, RS Wy ROT' ae Fic. 52. Gelidium regulare Baards. : habit of plant, ¢ = tetrasporangia. CORALLINACEAE CORALLINA OFFICINALIS L., Syst. Nat., ed. 10, 1 : 805 (1758). (Figs. 53, 54.) Slaughtered Seal Bay, in pools from water level down to 10 cm., 22 Dec. 1955, G.I.S.S. 1504 ; same locality, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1507. Midshipman Rock, dominant in mid-littoral zone on shiny olive-green deposit, 18 Feb. 1956, G.I.S.S. 1515, 1517, 1518. Standoff Rock, top of mid-littoral, 18 Feb. 1956, 212 MARINE ALGAE OF GOUGH ISLAND G.I.S.S. 1524, 1525, 1526. West Point Reefs, pools below spray zone,6 Mar. 1956, G.I.S.S: 1541. Isolda Rock, in shallow mid-littoral pools, 6 Mar. 1956, G.I.S.S. 1557; same locality, upper-littoral pools, 6 Mar. 1956, G.I.S.S. 1553. All Northern and Southern Hemisphere temperate seas, arctic, antarctic. C. officinalis was very common on the shores of Gough Island ; it grew in littoral rock pools, and in two places, Standoff Rock and the Midshipman Rock, it formed a zone on the rocks at mid-tide level. The plants (Fig. 53) are small, seldom exceeding 4 cm. in height. They have a very pronounced main axis bearing densely crowded, slender, pinnate branches which often branch again. The intergenicula of the main axis and first-order branches are flattened, the second-order branches are usually thread-like and terete. Reproductive conceptacles (Fig. 54) are abundant ; they are globose Imm Fics. 53, 54. Corallina officinalis L. : 53, habit of plant ; 54, conceptacles (c). with an apical pore and are occasionally antenniferous. In longitudinal section the cells of the intergenicula are seen to be in horizontal rows all of much the same height ; the cells measure about 60-7ou high x 7-gu wide and form lateral synapses. Each geniculum is composed of a single row of long cells with thick walls. The plants are frequently covered with epiphytes, Herpostphonia paniculata being the most common one. MARINE ALGAE OF GOUGH ISLAND 213 C. officinalis has frequently been recorded from southern cold and temperate waters, but in only one case (Skottsberg, 1923, p. 67) has the material been described and this material is obviously different from the Gough Island plants. The Gough Island plants differ from the type form of C. officinalis in their small size and very dense branching but I am of the opinion that they come within the variation range of this species. Dr. Levring, to whom I sent some material, said that it closely resembled his material from Chile and he thought that it should be included in C. officinalis (see Levring, 1960, p. 40). In the collections in the herbaria of the British Museum, Kew and the Muséum National d’Histoire Naturelle, Paris, there are a few specimens which much resemble the Gough Island material of C. officinalis and these specimens are variously identi- fied as C. officinalis, C. cuviert and C. chilensis. The distinctions between these three and other species, such as C. vancouvertensis, are far from clear and it seems best to refer the Gough Island plants to C. officinalis. Corallina goughensis Y. M. Chamberlain, sp. nov. (Figs. 55-60 ; Plate 16 Cc.) Planta usque ad 4 cm. alta, frondibus numerosis erectis e crusta basali ortis, per duos longitudinis trientes inferiores simplicibus intergeniculis teretibus, per trientem superiorem dichotome vel corymbose ramosis intergeniculis compressis. Inter- geniculorum cellulae dispositae in strata horizontalia extremam partem versus deorsum curvata ita corticem formantia ; cellulae medianae 35—70y longae, 6-8u latae, synapsibus lateralibus bene evolutis. Genicula unizonalia ; cellulae usque ad 250u longae. Conceptacula tetrasporica terminalia, in sinu inter ramulos duos posita, poris apicalibus. Tetrvasporae zonatim divisae, c. 160u longae, 60y latae. Isolda Rock, very abundant in lower-littoral rock pools, 6 Mar. 1956, G.I.S.S. 1547 (holotype). Not known elsewhere. Plant up to 4 cm. tall with numerous fronds springing from a basal crust. Lower two-thirds of plant unbranched with terete intergenicula ; upper third branched dichotomously or corymbosely with flattened intergenicula. Cells of intergenicula (Fig. 56) in horizontal rows curving out at the sides to form a cortex. Central cells 35-70u long x 6-8yu wide, lateral synapses (Fig. 56) very pronounced. Genicula unizonal (Figs. 57, 58), cells up to 250u long. Tetrasporic conceptacles (Fig. 59) present in the angles of the branches, pores apical. Tetraspores (Fig. 60) zonately divided, about 160y long x 60p wide. This plant was found growing in littoral rock pools at Isolda Rock where it was very abundant. Externally, the plant (Plate 16 c; Fig. 55) is characterized by its dichotomous-corymbose branching, internally by the very striking lateral synapses between the intergenicular cells (Fig. 56) which appear colourless in an otherwise stained section. It was very difficult to decide in which genus this plant should be included. In the most recent survey of the articulated Corallinaceae, Manza (1937, 1940) divides the species with only terminal conceptacles into the genera Corallina, Jania, 214 MARINE ALGAE OF GOUGH ISLAND 10u a a 1Ou, 56 intergenicular cell genicular cell 50u 60 Fics. 55-60. Corallina goughensis Y. M. Chamberlain : 55, habit of plant, ¢ = tetra- sporic conceptacle ; 56, vertical section of intergenicular cells, s1 = side view of lateral synapsis, ff = face view of lateral synapsis ; 57, vertical section of geniculus, ge = genicular cell, eg = extra-genicular portion, ig = intergenicular cell ; 58, detail of extra- genicular portion of genicular cell (diagrammatic), = pit connexion, c = granular contents, sg = deeply staining genicular cell wall ; 59, vertical section through frond g = geniculum, 7 = intergeniculum, ¢ = tetrasporic conceptacle ; 60, tetraspores. MARINE ALGAE OF GOUGH ISLAND 215 Arthrocardia and Duthiea!. With the exception of Duthiea, which is characterized by having lateral conceptacle pores, the genera are distinguished by their branching ; on this basis this plant should be classified as a Jania. However, in the Coral- linaceae the branching even in a single plant is liable to be extremely variable, and Suneson (1943, p. 48) has pointed out that there are even forms of Corallina officinalis which are dichotomous. Most authors distinguish Corallina from Jania by the length of the genicular cells, which are at least twice as long as the intergenicular cells in Corallina, and are about the same length as the intergenicular cells in Janda. On this basis I have classified this plant as a Corallina. The material is somewhat similar to Jania digitata Manza but the intergenicula are more flattened in C. goughensis and the upper branches are less regularly dichotomous than those of J. digitata. Dermatolithon nodulosum Y.M. Chamberlain, sp. nov. (Figs. 61-65 ; Plate 17.) Planta horizontaliter crustaceo-expansa, in statu vivo rosea, ad 2 cm. crassa, nodulos calcarios permultos confertim ferens. Crusta basalis ad 700w crassa, plerumque in vaginam circa nodulos producta, constata e strato supremo simplici cellulis epithallinis parvis planis, stratis medianis pluris horizontalibus cellulis perithallinis ad 50u longis et roy latis, strato infimo simplici cellulis hypothallinis obliquis ad 80y longis et Io latis. Noduli e perithallo bene evoluto cellulis ad 100 longis et 15y latis constati. Cellulae perithallinae et hypothallinae omnes synapsibus secundariis lateralibus apicem cellulae versus munitae. Conceptacula tetrasporangiifera in perithallo praesentia, plerumque supra superficiem crustae eminentia, desuper visa circularia, diametro interno ad 250u metienti, ad 150y alta. Tetrasporae zonatim divisae. Slaughtered Seal Bay, in pools from water level down to 10 cm., 22 Dec. 1955, G.I.S.S. 1504 ; same locality, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1507. Midshipman Rock, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1514. West Point Reefs, pools below spray zone, 6 Mar. 1956, G.I.S.S. 154r ; same locality, margins of pools below spray zone, 6 Mar. 1956, G.I.S.S. 1543. Isolda Rock, very abundant in shallow mid-littoral pools, 6 Mar. 1956, G.I.S.S. 1550, 1551 ; same locality, upper- littoral pools, 6 Mar. 1956, G.I.S.S. 1553 (holotype). Not known elsewhere. Plant (Plate 17 A) pink when living, up to 2 cm. thick, bearing nodules (Plate 17 B ; Fig. 61). Basal crust (Fig. 63) up to 700 thick, composed of a single layer of small flat epithallial cells, horizontal layers of perithallial cells measuring up to 50u long X roy wide and a single layer of oblique hypothallial cells measuring up to 80 long xX Iou wide. Nodules (Figs. 61, 62, 64) composed of a much-developed perithallus with cells roou long x I5u wide. All perithallial and hypothallial cells have secondary, lateral synapses which occur towards the top of the cell. Basal crust often extending to form a sleeve round the nodules (Fig. 62). Tetrasporic conceptacles (Figs. 62, 65) present in perithallus, usually projecting a little above the surface. Conceptacles circular in surface view, in section having a diameter up to 250u and being up to 150u. deep. Tetrasporangia zonately divided. } This name, Duthiea Manza (1937), is a later homonym of Duthiea Hack. (1895) (Gramineae). Fics. nodule showing direction of cell rows ; 62, vertical section through nodule showing position of conceptacles and sheath of basal crust cells (b), c = conceptacle ; 63, details of cells of basal crust in vertical section, e = epithallus, p = perithallus, h = hypothallus ; 64, detail of nodular cells in vertical and transverse section, sf = face view of secondary synapsis, ss = side view of secondary synapsis, p = primary pit connexion ; 65, trans- verse section of tetrasporic conceptacle. MARINE ALGAE OF GOUGH ISLAND lOul |imm level B level C level B 4p , i rt CT) Sip. 65 Dermatolithon nodulosum Y.M. Chamberlain : 61, vertical section through D. nodulosum was very abundant on the shores of Gough Island ; its conspicuous pink crusts grew on the margins of deep rock pools and over the rock surface of shallow pools ; occasionally it grew on the open rocks. The size and habit of this plant are unusual for a Dermatolithon, most species of which have a very thin thallus and are epiphytic or epizoic. Two other species, D. hapalidioides (Crouan) Foslie and D. papillosum (Zanard.) Foslie, have been recorded as growing on rocks in temperate waters although both are more commonly epiphytic or epizoic. D. nodulosum is similar to these species in microscopic. MARINE ALGAE OF GOUGH ISLAND 217 structure but is a very much larger plant and has coarser nodules. The other two species have so far only been recorded from the European and North African coasts (Hamel & Lemoine, 1952, pp. 61, 63). In young plants, only the basal crust is present. The epithallial and outer perithallial cells (Fig. 63) of this crust have homogeneous, rather dense contents, the inner perithallial cells become vacuolated and the hypothallial cells are also vacuolated. The nodules are composed almost entirely of horizontal rows of large perithallial cells (Figs. 61, 64), which are almost devoid of contents and form numerous secondary lateral synapses with the cells round them. What little remains of the cell contents usually collects round the primary pit connexions and secondary synapses. Only tetrasporangial conceptacles were found. These have one central pore (Plate 17 c; Fig. 65). The tetraspores are usually divided zonately into four but bispores are also present. The conceptacles do not cave in after the spores have been shed and many empty ones can be seen buried in the older parts of the perithallus (Plate 17 c). LITHOTHAMNIUM NEGLECTUM (Foslie) Foslie in Deutsch. Stidpol.-Exped. 1901-03 8 : 207 (1908) ; Contrib. Monogr. Lithothamnia, ed. Printz : 43, t. 9 figs. 1-4 (1929).—Lemoine in zme Exped. Antarct. Frang. 1908-10, Melobesiées : 14 (1913) ; apud Cotton in Journ. Linn. Soc. Lond., Bot. 43 : 196, t. 9 fig. 2, t. 10 figs. 3-6 (1915) ; in K. Svensk. Vetenskapsakad. Handl., Ny Foljd, 61 (4) ; 12 (1920)... (Big: 60:5 Plate 28.) Lithothamnium muelleri forma neglectum Foslie in K. Norske Vidensk. Selsk. Skr. 1899 (5) : 17 (1900). Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1507. Isolda Rock, frequent in lower-littoral rock pools, 6 Mar. 1956, G.I.S.S. 1548 ; same locality, upper-littoral pools, 6 Mar. 1956, G.I.S.S. 1553. Antarctic and sub-antarctic. Two collections (G.I.S.S. 1507, 1548) of this species come from rock pools at Slaughtered Seal Bay and Isolda Rock respectively. These plants are rather similar both macroscopically and microscopically. They have leaf-like pink thalli (Plate 18 A) with concentric markings. The thallus (Fig. 66) is up to about 400p. thick ; it has a single layer of epithallial cells, below which is the perithallus. The perithallial cells run vertically below the epithallus and measure up to 10u wide x 17p long ; they run horizontally in the middle of the thallus and curve downwards below to the hypothallus, the cells of which measure up to 50u long xX 7u wide. Tetrasporic conceptacles (Plate 18 c) occur in the perithallus ; these measure up to 300% wide xX 200u deep and open through numerous pores. The tetraspores are zonately divided and measure up to 145y long x 70 wide. The third collection (G.I.S.S. 1553) comes from Isolda Rock where, together with Lithophyllum sp., the plant formed a conspicuous band on the rocks in the upper mid-littoral zone. This plant (Plate 18 B) was white when growing, probably due to bleaching in its exposed position. A similar band was reported to occur in other 218 MARINE ALGAE OF GOUGH ISLAND 10u Fie I ae coe sa ep ee nF Sp (0 | 66 Fic. 66. Lithothamnium neglectum (Foslie) Foslie : vertical section through thallus of G.I.S.S. 1507, h = hypothallus, p = perithallus, e = epithallus. parts of the island but no other collection from it was made. This plant is smaller than the preceding ones both microscopically and macroscopically and it probably represents Foslie’s L. neglectum forma fragile (1908, p. 208). The two forms are well illustrated in Foslie’s photographs (1929, t. 9) of the species, G.I.S.S. 1553 resembling Foslie’s figs. 1-3 and G.J.S.S. 1507 and 1548 his fig. 4. GIGARTINACEAE IRIDAEA LAMINARIOIDES Bory in Duperrey, Voy. Monde La Coquille, Bot., Crypt. : 105 (18283), t. 11 fig. 1 A—-D (18271).—Kititz., Sp. Alg. : 726 (1849) ; Tab. Phyc. 17: 3, t.8 figs. c, d (1867).—Montagne in C. Gay, Hist. Fis. & Polit. Chile, Bot. 8 : 352 (1854).—Laing in Chilton, Subantarct. Is. New Zeal. 2 : 506 (1909).— Levring in Act. Univ. Lund., Ny Féljd, Avd. 2, 56 (10) : 56(1960). (Figs. 67-71.) Ividaea heterococca Kiitz., Tab. Phyc. 17 : 4, t. 11 (1867). Iridaea cordata sensu Cotton in Journ. Linn. Soc. Lond., Bot. 43 : 176 (1915) pro parte ; non J. G. Agardh. Ivridophycus laminarioides (Bory) Setch. & Gardn. in Proc. Nation. Acad. Sci. Wash. 22 : 470 (1936) ; in Univ. Calif. Publ. Bot. 19 : 206 (1937). 1 For the dates of the natural history parts of the Voyage ... Coquille see Sherborn & Woodward in Ann. & Mag. Nat. Hist., Ser. 7, 7 : 392 (1901). MARINE ALGAE OF GOUGH ISLAND 219 Ividophycus boryanum Setch. & Gardn. in Proc. Nation. Acad. Sci. Wash. 22 : 470 (1936), nom. nud. Ividophycus boryanum Setch. & Gardn. in Univ. Calif. Publ. Bot. 19 : 202 (1937). Iridaea boryana (Setch. & Gardn.) Skottsb. in K. Svensk. Vetenskapsakad. Handl., Ser. 3, 19 (4) : 82 (1941).—Levring in Ark. fér Bot. 31a (8) : 13 (1944) ; in Act. Univ. Lund., Ny Foljd, Avd. 2, 56 (10) : 55 (1960). Midshipman Rock, Porphyra zone, 18 Feb. 1956, G.I.S.S. 1514. Standoff Rock, dominant on less-exposed part of the rock above the Corallina zone, 18 Feb. 1956, G.I.S.S. 1526. Isolda Rock, occasional in upper littoral on ridges between pools, 6 Mar. 1956, G.I.S.S. 1552. Widespread in southern South America, Falkland Islands, Kerguelen, Crozet Islands, Auckland Island, ? North America. I. laminarioides was one of the most common littoral algae on Gough Island. On moderately sheltered boulder beaches it formed a conspicuous mid-littoral zone of bright orange-brown which was bounded above by a Blidingia zone and below often merged into a Rhodoglossum zone. It has been recorded as a dominant littoral plant in other regions, such as Chile (Levring, 1960) and Kerguelen and the Crozet Islands (Levring, 1944). The Gough Island plants (Figs. 67, 68) grow from communal, fairly thin basal crusts which spread over the rocks, each crust giving rise to numerous erect fronds. According to Dr. Holdgate, the smaller, orange-brown plants grew in the more exposed places while larger plants, which were usually more violet in colour, came from the sheltered channels and crannies between the boulders. The material collected measures up to 20 cm. long but is mostly under 9 cm. ; plants from more sheltered places undoubtedly grew longer than this, however. All the plants have a small, terete stipe above the basal crust; above the stipe is a strongly channelled apophysis from which develops the smooth-edged blade which typically has a cuneate base. Only one small piece of the material was found to be fertile ; this (Fig. 68) has tetrasporangia along one side of the edge of the blade. In transverse section (Fig. 69) the blade can be seen to consist of a central medulla of widely spaced, very thin and elongated cells ; towards the edge the cells become shorter and fatter and interweave more noticeably ; finally filaments of up to five, small, oblong cells are produced which lie at right-angles to the surface of the thallus and form a densely packed cortex. The tetrasporangia develop in the medulla ; the intercalary development of the tetrads of spores can be seen when they are young (Fig. 70) and ripe tetrads measure up to about 35y in diameter (Fig. 69). Various species of Ividaea occur in great abundance on rocky sub-antarctic shores. The fact that small samples of Jvidaea from this area have frequently been collected and often identified without reference to previous collections has resulted in a confused picture of the taxonomy of the genus. Setchell & Gardner in their 1937 paper made the first real attempt since Bory’s account in the Voyage... Coquille to sort out the taxonomy of the South American species, and their con- clusions have been accepted by most subsequent authors. Setchell & Gardner also changed the name of the genus to Ividophycus since Iridaea Bory is a later 220 MARINE ALGAE OF GOUGH ISLAND 10u 67 68 Fics. 67-71. Ividaea laminarioides Bory : 67, habit of plant ; 68, plant with tetra- sporangia, b = blade, ¢ = tetrasporangia, a = apophysis, s = stipe, d = basal disk ; 69, transverse section through blade with mature tetraspores, c = cortex, m = medulla, t = tetraspore ; 70, transverse section of blade with young tetraspores, m = medulla, t = tetraspore ; 71, transverse section of blade of Bory’s type specimen, c = cortex, m = medulla. homonym. Iridophycus has not been generally accepted, however, and Iridaea Bory has been conserved. There is an error in the entry referring to it in the list of nomina generica conservanda (Internation. Code Bot. Nomencl. Montreal 1959 : 205 (1961)) ; this reads: “‘ Ividaea Bory, Dict. Class. Hist. Nat.9g: 15. 1826”. On MARINE ALGAE OF GOUGH ISLAND 221 the page cited Bory used the spelling Jvzdea, on the next page Iridea twice and Iridaea twice, and on p. 192 of the same volume Jvidaea ; in his other publications Bory always used the spelling Ividaea. Among the species of Jvidaea described by Bory is I. laminartoides, the type material of which was collected at Concepcion, Chile, by D’Urville during the voyage of ‘‘ La Coquille’. Setchell & Gardner (1937, pp. 202, 206) considered that the type material—the smaller and larger plants of which Bory thought to be juvenile and adult stages of the same species—represented two separate species. They called the adult plants (Bory, 1827, t. 11 fig. 1 D, E) Iridophycus laminariordes and the so-called juvenile plants (fig. 1 A, B, c) I. boryanum. This last species they considered to be the same as Kiitzing’s Ividaea heterococca but did not give it that name because in their opinion the name was based on a monstrosity, the type material carrying a blue-green epiphyte mistaken by Kiitzing for reproductive organs of the Ividaea. Various authors (e.g. Skottsberg and Levring in his 1960 paper) have used Setchel] & Gardner’s classification of the genus but both Skottsberg and Levring pointed out that the separation of the two species is unsatisfactory since many intermediate forms occur. I had the opportunity of examining Bory’s type material of I. laminariovdes in the herbarium of the Muséum National d’ Histoire Naturelle, Paris, and I have also studied much other herbarium material of Jvidaea in the Paris, British Museum and Kew Herbaria. In my opinion, Bory’s type material does represent two species but I think that the plants on which Bory’s fig. I A, B, Cc and D were based are all the same species, D perhaps being from a more sheltered position on the shore, and that E (which has a ciliate margin) is a different species. As the fragment of E drawn by Bory represents all the material there is, I do not think it is possible to decide what species it is. J. laminariotdes is a very variable species, but whilst the blade and apophysis may sometimes have irregular outgrowths, they are never regularly ciliate like Bory’s fig. 1 E. The small form (Bory’s A, B and c), with which most of the Gough Island material agrees very closely, usually dries to a shade of brown if pressed when fresh while the larger form (D) is often more red or purple. As far as I could see when I examined the type material, the lower left-hand blade of D is in fact a separate plant and was never part of the main frond as shown by Bory ; the angle it makes with the main frond looks most unnatural in the drawing too. The distinctive features of J. laminarioides are the very short stipe, the pronounced apophysis which is channelled and has a smooth edge, and a blade with a cuneate base and a smooth edge ; the blade may or may not be divided. I also think that the fronds usually arise from a communal basal crust ; certainly those in the Gough Island material did so and much of the herbarium material I examined appeared to show this too ; unfortunately it was usually impossible to make sure about this since dried material of Jvidaea disintegrates very quickly when soaked out. This feature of the communal basal crust has been used by Setchell & Gardner (1937, p. 211) as one of the diagnostic characters of another of their new species, Inidophycus caespitipes. Skottsberg (1941, p. 82) commented on and published a drawing of this species and said how similar the erect fronds are to those of Ividaea 222 MARINE ALGAE OF GOUGH ISLAND boryana. Iridophycus caespitipes is probably only a form of Ividaea laminarioides and the more pronounced development of the basal crust could well be a response to the very exposed habitats from which it has been collected. While working on the Ividaea laminarioides group of species, I made sections of the blade of a considerable number of herbarium specimens. I found a wide range of variation in the tissues, the cells in some being much more densely packed than they were in others. In the filaments of the cortex, there was a variation both in the number of cells in the filaments and in the size and shape of the cells. Some of the variation may have been due to the method of pressing the specimens and I could not see that there was any significant correlation between the detailed microscopic structure of the blade and the external appearance of the plant. The observations of Setchell & Gardner as to the microscopic characteristics of their various species seem to have been based on a rather small sample in each case. The appearance of the cross-section of a blade from Bory’s type specimen of J. laminarioides (Fig. 71) was rather different in the cortical region from the Gough Island material (Fig. 69) but I do not think the difference is great enough to warrant specific separation of the two plants. The only plant of Ividophycus caespitipes that I have been able to examine is the fragment in the Kew Herbarium collected in the Falkland Islands by Mrs. Vallentin and sent to Setchell & Gardner by Cotton as Ividaea cordata (Setchell & Gardner, 1937, p. 212). The specimen is so incomplete that it gives no idea of the external characteristics of the species, but in cross-section the cortical cells of the blade are very considerably larger than those of all the other material which I sectioned. For this reason I have not included Ividophycus caespitipes as a synonym of Ividaea laminarioides. It is quite clear that, before the taxonomy of Ividaea can be satisfactorily worked out, much more work will have to be done with fresh material and field observations made by phycologists. Until this has been done, I think it is difficult to be sure whether the North American plants which are called I. laminarioides really belong to the same taxon as the South American ones, and also to sort out the relationship between J. laminarioides and other very similar species such as J. micans Bory and I. obovata Kiitz. IRIDAEA UNDULOSA Bory in D’Urville, Fl. Il. Malouines: 22 (Dec. 1825)! ; in Mém. Soc. Linn. Par. 4 : 594 (Jan. 1826).—Skottsb. in K. Svensk. Vetenskaps- akad. Handl., Ser. 3, 19 (4) : 85, figs. 4-7 (1941).—Levring in Act. Univ. Lund., Ny Foljd, Avd. 2, 56 (10) : 57 (1960). (Figs. 72, 73.) Ividaea crispata Bory in Dict. Cl. Hist. Nat. 9 : 16 (Feb. 1826).—Levring, loc. cit. (1960). Ividaea augustinae Bory in Duperrey, Voy. Monde La Coquille, Bot., Crypt.: 108 (1828), t. 12 (1827), nom. illegit.—Kiitz., Sp. Alg.: 726 (1849).—Montagne in C. Gay, Hist. Fis. & Polit. Chile, Bot. 8 : 353 (1854). Ividaea micans var. ciliolata Hook. & Harv. in Hook., Lond. Journ. Bot. 4 : 263 (1845). Ividaea cordata var. ciliolata (Hook. & Harv.) Hook. & Harv. in Hook. f., Bot. Antarct. Voy. I : 485 (1847). Ividaea ciliata Kiitz., Sp. Alg.: 726 (1849) ; Tab. Phyc. 17 : 4, t. 10 (1867).—Baards. in 1T am indebted to M. R. Deélepine for a photocopy of this paper. MARINE ALGAE OF GOUGH ISLAND 223 Res. Norw. Sci. Exped. Tristan da Cunha 9 : 76, fig. 41 E (1941).—Levring, loc. cit. (1960). Iridophycus undulosum (Bory) Setch. & Gardn. in Proc. Nation. Acad. Sci. Wash. 22 : 471 (1936) ; in Univ. Calif. Publ. Bot. 19 : 217, tt. 25, 26 (1937).—W. R. Taylor in Pap. Mich. Acad. Sci. Arts & Lett. 24 : 146 (1939). Iridophycus crispatum (Bory) Setch. & Gardn. in Proc. Nation. Acad. Sci. Wash. 22 : 471 (1936) ; in Univ. Calif. Publ. Bot. 19 : 221 (1937). Iridophycus ciliatum (Kiitz.) Setch. & Gardn. in Proc. Nation. Acad. Sci. Wash. 22 : 471 (1936) ; in Univ. Calif. Publ. Bot. 19 : 220 (1937). Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1505. Church Rock, abundant on vertical rock face, 5 Mar. 1956, G.I.S.S. 1528, 1535. Isolda Rock, frequent in mid-littoral rock pools, 6 Mar. 1956, G.I.S.S. 1546. Window Pool, in sheltered shallow pool, 12 Apr. 1956, G.I.S.S. 1556. Southern South America, Falkland Islands, Tristan da Cunha. Iridaea undulosa occurred quite frequently on the shores of Gough Island ; it was most often found in littoral rock pools but was also collected from the Cladophora vadiosa zone just above and below mid-tide level on the sheer southern face of og 9868 7 PRP i A Fics. 72,73. Ividaea undulosa Bory : 72, habit of plant, a = apophysis, e = emergences, b = blade, ¢ = tetrasporangia ; 73, transverse section of blade, c = cortex, m = medulla, t = tetrad of spores. 224 MARINE ALGAE OF GOUGH ISLAND Church Rock. The plants collected measure up to 17 cm. tall x 12 cm. wide (Fig. 72), are more or less ovate in shape and are of a deep, purple-red colour and a rather leathery texture when alive. The apophysis is flat or very slightly channelled and always bears obvious emergences ; it is unbranched but occasionally the emergences are large enough to look like young plants. The base of the blade is cuneate to reniform and the margin is, in most cases, ciliate ; in some younger plants itis smooth. Many of the plants are tetrasporic but no sexual reproductive organs were found. In section (Fig. 73) the blade can be seen to have the usual Ividaea structure ; the long, thin cells of the medulla are widely scattered and the cortex consists of three-celled filaments of fairly large cells. The very wide spacing of the cells may be more apparent than real ; the plants are extremely gelatinous and when sectioned the gelatinous substance immediately swells so much that it pushes the cells apart. Tetrads of spores develop in sporangia in the medulla in an inter- calary manner characteristic of this genus. Setchell & Gardner (1937, pp. 215-222) divided the South American species of Invidaea with ciliate margins and unbranched apophyses into three species, the specific delimitations depending upon whether the base of the blade was cuneate to reniform or cordate. Skottsberg (1941, pp. 85-88) considered that these inter- specific boundaries were untenable and that the three species should be treated as forming one taxon, a view which I share. The distribution seems to be limited to the Atlantic and Pacific coasts of South America and some South Atlantic Islands. The question of whether the name Iridaea undulosa was validly published in D’Urville’s Flore des Iles Malouines is not easy to decide. That work as a separate publication certainly antedates the ninth volume of the Dictionnaire Classique d’ Histoire Naturelle, and so also probably does the sixth part of the fourth volume of the Mémoires de la Société Linnéenne de Paris, in which it was also printed.! In the Flore des Iles Malouines there is a general but not a botanical description of Iridaea micans in the prefatory part by D’Urville, and in the list of species, for the cryptogamic part of which Bory was responsible, there is : 25. Iridaea micans. Ovato reniformis, undulosa, tenerrima, basi subcordata. Cap. (fide Gaud.). 26. a. I. undulosa. Laevis, fronde ovato-conica, basi obconica, crassiuscula, undulato-crispa. b. I. papillosa. Nowhere in this work is there an indication that Iv:daea is a new genus, unless this can be read into the statement : “‘L’éclat de ses couleurs lui a valu le nom d’Jridea micans”’. 1 Flore des Iles Malouines. Title-page dated 1825 ; not noticed in the Bibl. France ; received at weekly meeting of Paris Academy of Sciences on 9 January 1826 (Proc.-Verb. Acad. Sci. Par. 8 : 332 (1918)) ; therefore presumably December 1825. Mém. Soc. Linn. Par. 4 (6). Cover dated Jan. 1826; not noticed in Bibl. France but notices of other numbers suggest that it ran close to schedule ; not received by Paris Academy of Sciences ; date on cover (Jan. 1826) probably correct. Dict. Cl. Hist. Nat.9. Title-page dated Feb. 1826 ; noticed in Bibl. France 15 : 154 (25 Feb. 1826) ; received at weekly meeting of Paris Academy of Sciences on 20 March 1826 (Proc.-Verb. Acad. Sci. Par. 8 : 359 (1918)) ; therefore February 1826. MARINE ALGAE OF GOUGH ISLAND 225 , However, Bory’s article headed “ Iridée, Ividea’”’ in the ninth volume of the Dictionnaire Classique, which appeared only a few weeks later, begins “‘ Le genre ainsi designé par Stackhouse, ne pouvant étre conservé selon Lamouroux, nous adoptons ce nom pour un genre nouveau de Fucacées’’. The species he describes in this article include J. micans but not I. undulosa. In his fuller treatment of the genus in the Voyage . . . Coquille, he treats I. undulosa, and I. crispata, described in the Dictionnaire Classique, as synonyms of the new and _ superfluous name I. augustinae. Unless Ividaea micans and I. undulosa are regarded as having been published in the Flore des Iles Malouines as species of the genus Ividea Stackh., they are not validly published there. The spelling of the generic name in the part of that work for which Bory was responsible differs from Stackhouse’s, and we know from Bory’s subsequent publications that a different genus was intended. However, in deciding whether a name is validly published in a particular publication, matter published later cannot be taken into account. The decision on the place of valid publication of these two names therefore turns on the answer to the question : had neither vol. 9 of the Dictionnaire Classique nor the Voyage... Coquille been published, would J7vidaea in the Flore des Iles Malouines have been considered as a mis-spelling of Ividea Stackh. or as the name of a new genus for which no definition was provided? The definitions of the species in the Flore des Iles Malouines are inconsistent with membership of a genus defined thus : “ Jvidea. Fronde cartilaginea, tereti ; ramosissima, pinnata : ramis oppositis, supra decompositis, capillaribus . . . Fructificatio incognita .” (Stackhouse, Ner. Brit., ed. 2 : ix (1816)). They would nevertheless at the time have been regarded as belonging to the same group of algae as Ividea Stackh., and their names must therefore be regarded as having been validly published as names of new species of that genus. Their subsequent transfer to Ividaea Bory cannot be regarded as the making of new combinations, and hence there is the anomalous situation that these two correct names were validly published earlier than was the name of the genus which forms their first part. RHODOGLOSSUM REVOLUTUM Baards. in Res. Norw. Sci. Exped. Tristan da Cunha : 78, figs. 39 C, 40 E, 41 D (1941). (Figs. 74-77.) Midshipman Rock, Porphyra zone, 18 Feb. 1956, G.I.S.S. 15174. Standoff Rock, occasional in mid-littoral coralline zone, 18 Feb. 1956, G.IJ.S.S. 1524. Church Rock, occasional on vertical rock face, 5 Mar. 1956, G.I.S.S. 1534. West Point Reefs, pools below spray zone, 6 Mar. 1956, G.I.S.S. 1541. Isolda Rock, in upper littoral on ridges between pools, 6 Mar. 1956, G.I.S.S. 1552 ; same locality, upper-littoral pools, 6 Mar. 1956, G.I.S.S. 1553. Window Pool, abundant in sheltered shallow pool, 12 Apr. 1956, G.I.S.S. 1556. Tristan da Cunha (Inaccessible Island). R. revolutum was a common plant of the Gough Island littoral. It grew on rocks and in pools in the mid-littoral and in some cases, particularly on rather less exposed beaches such as Dell Rocks Beach, it formed a zone below the Ividaea laminarioides zone. The only previous record of this species is its original gathering from the 226 MARINE ALGAE OF GOUGH ISLAND littoral zone of Inaccessible Island. Baardseth gives no indication of its abundance there, but as the species is not mentioned in the ecological section of Baardseth’s work it can be assumed that it was not present in any great quantity and certainly did not dominate a zone as it did in places on Gough Island. It would seem therefore that this is a plant of very restricted distribution which attains its maximum development on Gough Island. The plants (Fig. 74) consist of a spreading basal crust from which arise erect fronds in all stages of development. The fronds are fleshy and of a dark reddish- brown colour ; the lower half of most plants is entire ; above this the plants usually branch dichotomously up to four or five times but occasional unbranched 10u Fics. 74-77. Rhodoglossum revolutum Baards.: 74, habit of plant ; 75, transverse section through blade, m = medulla, c = cortex ; 76, transverse section through blade with tetraspore mother cells (¢mc), c = cortex ; 77, mature tetrad of spores. plants are present. The margins of the fronds are entire and frequently inrolled. The fronds measure up to 6 or 7 cm. high and are extremely gelatinous, which causes them to shrink excessively when pressed. Baardseth describes the plants as being 2 or 3 cm. high, but this measurement must have been based on dried material ; the isotypes in the British Museum Herbarium measured up to 3 cm. on the herbarium sheet, but one specimen which was 2 cm. high when dry expanded to 3-5 cm. when soaked in water. The Gough Island plants measure up to I cm. across but are usually about 5 mm.; Baardseth recorded 4-8 mm. for the width. In section the fronds are flat except at the very bottom where they are terete. The frond is composed of a cortical region of filaments of small elongate cells at MARINE ALGAE OF GOUGH ISLAND 227 right-angles to the surface of the thallus and an inner medulla of long, thin branching cells which anastomose with one another (Fig. 75). Female and asexual reproductive organs are present. The carposporophytes develop on the concave surface of the upper parts of the thallus ; when mature they form a protuberance on the surface of the thallus and in section are seen to be surrounded by a pericarp, as shown in Baardseth’s drawing (his fig. 41 Dp). The tetrasporangia are also developed on the concave side of the upper thallus and can be seen as dark patches. They do not protrude from the thallus. The spore mother cells develop in accessory filaments (Fig. 76) from the lower cells of the cortex. When mature (Fig. 77), the tetrads of spores measure up to 35 in diameter. Baardseth commented at length on his reasons for including this species in the genus Rhodoglossum and the generally unsatisfactory distinctions between the genera Rhodoglossum, Gigartina and Inidaea. CERAMIACEAE CENTROCERAS CLAVULATUM (Agardh) Montagne in Explor. Sci. Algérie 1840-42, Bot. I : 140 (1846). Ceramium clavulatum Agardh in Kunth, Synops. Pl.: 2 (1822). Slaughtered Seal Bay, below 15 cm. in rock pools, 22 Dec. 1955, G.I.S.S. 1507. Midshipman Rock, mid-littoral, 18 Feb. 1956, G.J.S.S. 1518 ; same locality, occas- ional in standing water in upper littoral, 18 Feb. 1956, G.J.S.S. 1520. Church Rock, frequent on vertical rock face, 5 Mar. 1956, G.I.S.S. 1531, 1535. Isolda Rock, upper-littoral pools, 6 Mar. 1956, G.J.S.S. 1553. _ Widely distributed in warm and temperate waters. This plant was found frequently on Gough Island intertwined with other algae. MICROCLADIA ALTERNATA Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9g : 99, figs. 50 A, 5I C-E (1941). Midshipman Rock, mid-littoral, 18 Feb. 1956, G.I.S.S. 1578. Tristan da Cunha group of islands. A small quantity of this species was found growing as an epiphyte on Corallina officinalis at the Midshipman Rock. Female and tetrasporic plants are present and agree well with Baardseth’s description and material. RHODOMELACEAE BOSTRYCHIA MIXTA Hook. & Harv. in Hook., Lond. Journ. Bot. 4 : 270 (1845).— E. Post in Rev. Algol. 9 : 39 (1936). Standoff Rock, on less-exposed part of the rock above the Corallina zone, 18 Feb. 1950, G.I.S.S. 1526. Southern circumpolar temperate and cold waters, Japan, China. B. mixta is a very common plant of southern shores where it frequently dominates a community at or above high-tidemark, especially in places where there is a certain amount of freshwater seepage. On Gough Island it was found only at Standoff Rock, where it grew mixed with Rhizoclonium ambiguum, R. riparium forma validum 228 MARINE ALGAE OF GOUGH ISLAND and Spongomorpha sp. Dr. Erika Post kindly confirmed the identification of this plant. HERPOSIPHONIA PANICULATA Baards. in Res. Norw. Sci. Exped. Tristan da Cunha Q: 127, figs. 69 A, 71 A-C, 72 (1941). (Figs. 78, 79.) Midshipman Rock, in mid-littoral zone on shiny olive-green deposit, 18 Feb. 1956, G.I.S.S. 1515, 1517, 1518 ; same locality, in standing water in upper littoral, 18 Feb. 1956, G.I.S.S. 1520. Church Rock, vertical rock face, 5 Mar. 1956, G.I.S.S. 1535. Isolda Rock, in upper littoral on ridges between pools, 6 Mar. 1956, G.J.S.S. 1552. Tristan da Cunha group of islands. WOE WS SD WR te YY SR S CF RQ WIE WROTE Q Oe, 100u 19 78 Fics. 78,79. Herposiphonia paniculata Baards. : 78, ventral view of plant, y = rhizoid, s = short branch, / = long branch ; 79, tetrasporic branchlet. H. paniculata was a frequent epiphyte on Corallina officinalis. The plants (Fig. 78) measure up to 2:5 cm. high. The main axis consists of a central cell surrounded by 12-14 pericentral cells each about two to three times as long as wide. Each segment of the main axis gives off a branch ; on every fourth segment are alternating, ventral, long branches which repeat the structure of the main axis ; on the intervening segments are simple branches up to 27 segments long and Ioop. wide. This organization is characteristic of the genus. Tetrasporic (Fig. 79) and carposporophytic plants are present ; the mature cystocarps on the latter grow laterally on the simple branches and measure up to 350% wide X 500u long. The tetraspores are borne in branchlets of the compound branches. The Gough Island material agrees well with that described by Baardseth, who points out that this plant may later prove to be within the variation range of one of the other species of Herposiphonia with sexual organs occurring laterally on the branchlets. In habit, the plants closely resemble the North American species MARINE ALGAE OF GOUGH ISLAND 229 H. grandis Kylin and H. rigida N. L. Gardn., but they have consistently more pericentral cells than either of these species. LOPHURELLA sp. (Fig. 80.) Church Rock, occasional amongst Cladophora radiosa on vertical rock face, 5 Mar. 1956, G.I.S.S. 1533. It is unfortunately impossible to identify the species of this material with certainty as the plants are incomplete, the ultimate branches being absent. However, as the genus is predominantly sub-antarctic and antarctic, its presence on Gough Island is of interest. The wiry plants (Fig. 80) measure up to 5 cm. high and lcm 543 . RA 80 Fic. 80. Lophurella sp. : habit of plant. arise from a creeping, rhizomatous base. Tetraspores are present. The plants belong to one of two species, L. hookerana (J. G. Agardh) Falkenb. or L. patula (Hook. & Harv.) De Toni, but, without the terminal branchlets, it is impossible to say which. POLYSIPHONIA BOERGESENII Baards. in Res. Norw. Sci. Exped. Tristan da Cunha g: 116, figs. 64 c, 65 B—D (1941). West Point Reefs, pools below spray zone, 6 Mar. 1956, G.I.S.S. 1541. Tristan da Cunha group of islands. The Gough Island material of this species was found growing as an epiphyte on Codium fragile in rock pools at West Point Reefs. The material agrees exactly both with Baardseth’s description of the type and with isotype material in the British Museum Herbarium except in the size of the 230 MARINE ALGAE OF GOUGH ISLAND spermatangial branchlets. In the Gough Island material these are 140-180 long whereas Baardseth describes them as being about 75y long. However, when I measured the spermatangial branchlets in the isotype material, I found that these were of the same order of size as those in the Gough Island plants ; it seems therefore that Baardseth must have been measuring young branchlets. POLYSIPHONIA HOWEI Hollenb. apud W. R. Taylor in Allan Hancock Pacif. Exped. I2: 302, fig. 3 (1945).—Joly in Univ. Sao Paulo Fac. Fil. Ci. & Let. Bol. 217, Bot. 14: 164, t. 13 figs. 5, 5a (1957).—W. R. Taylor, Mar. Alg. E. Trop. & Subtrop. Coasts Amer. : 582 (1960). (Plate Ig A.) Polysiphonia sp. Baards. in Res. Norw. Sci. Exped. Tristan da Cunha 9 : 119 (1941). Midshipman Rock, mid-littoral, 18 Feb. 1956, G.I.S.S. 1517, 1518. Isolda Rock, abundant on margins of mid-littoral pools, 6 Mar. 1956, G.J.S.S. 1549 ; same locality, occasional in upper littoral on ridges between pools, 6 Mar. 1956, G.J.S.S. 1552. Atlantic and Caribbean from North Carolina to Brazil, Tristan da Cunha, Pacific from Panama to Colombia. P. howei was collected from Midshipman Rock and from Isolda Rock, where it grew in a low, brown turf on lower littoral rocks ; on Isolda Rock it actually domi- nated a zone below Iridaea laminarioides and above Durvillaea antarctica. A similar zone above the Durvillaea zone was reported on West Point Reefs and was probably dominated by P. howe1, but no collection was made. The plants from Isolda Rock (Plate 19 A) measure about 1 cm. high and have the habit of a Lophostphonia with well-developed prostrate shoots, which have rhizoids below and erect, sparingly branched filaments above. From the structure of the apices of the prostrate shoots it can be clearly seen, however, that this plant belongs to the genus Polysiphonia and not Lophostphonia ; the branches develop exo- genously on all sides of the shoot and the dorsiventral habit is due to the fact that the branches on the lower surface abort and rhizoids develop from the pericentral cells on the lower side. The rhizoids are long, thin-walled, cut off from the peri- central cell which produces them, and they end in a well-developed hapteron. The erect shoots are strongly recurved at their apices and the apices bear dichotomously branched trichoblasts. The filaments have 8-13, mostly 10-11, pericentral cells ; the segments of the basal filaments measure about 130% wide xX 150 long, those of the erect filaments 70-100n wide x 55-85yu long. The erect filaments branch rarely when sterile but more abundantly when tetrasporic. Tetraspores are pro- duced in the upper branches in somewhat spiral series ; there is one tetraspore per segment. This material is in close agreement with Hollenberg’s and Taylor’s descriptions of the species both in morphology and habitat. The plants from the Midshipman Rock are very similar but have somewhat longer segments ; those of the prostrate shoots are 100-1404 wide x 140-170y long. I do not think there is any doubt however that they belong to the same species. The plants recorded by Baardseth from the Tristan da Cunha group of islands as Polystphonia sp. are also P. howet. MARINE ALGAE OF GOUGH ISLAND 231 LOPHOSIPHONIA SCOPULORUM (Harv.) Womersley in Trans. R. Soc. S. Austral. 73 : 188 (1950).—Cribb in Univ. Queensl. Pap., Bot. 3: 138, t. I, t. 2 figs. 8-12 (1956). (Plate 19 B.) Polysiphonia scopulorum Harv. in Trans. R. Irish Acad. 22 : 540 (1855).—Segi in Journ. Fac. Fish. Pref. Univ. Mie 1 : 200, fig. 9, t. 3 fig. 7 (1951). Polysiphonia villum J. G. Agardh, Sp. Gen. Ord. Alg. 2 (3) : 941 (1863). Standoff Rock, on less-exposed part of the rock above the Corallina zone, 18 Feb. 1956, G.I.S.S. 1526. Widespread in the Pacific Ocean, Australia, New Zealand. This species was collected from Standoff Rock, where it formed part of a dense turf together with Polysiphonia sp. The plants (Plate 19 B) are rather flaccid and seldom measure more than 1-5 cm. high. They are composed of a prominent, prostrate basal filament which gives off rhizoids mainly from its lower surface and erect filaments mainly from its upper surface ; sometimes the positions are reversed. The rhizoids are long and end in frilly haptera ; their contents are a continuation of those of the parent pericentral cell. The filaments are ecorticate and have four pericentral cells throughout. The basal segments measure 140-250 long x c. 140u wide in the older parts and 70-gou long x c. 125y wide nearer the apex. The erect filaments have segments 70-120u long xX I1oy wide and these in their upper parts give off side branches with noticeably much shorter segments measuring 30-55u long xX 704 wide. Branching is quite frequent and in tetrasporic filaments it is abundant. Tetraspores are borne in the side branches and there is one per segment. Branching is endogenous ; no trichoblasts were seen. The Gough Island plants agree well with isotype material in the British Museum Herbarium and with Queensland material named by Dr. Cribb. SPECIAL LITERATURE BAARDSETH, E. (1941). The marine algae of Tristan da Cunha. Res. Norw. Sci. Exped. Tristan da Cunha 9 : 1-174. BORGESEN, F. (1939). Marine algae from the Iranian Gulf especially from the innermost part near Bushire and the Island Kharg. Dan. Sci. Invest. Ivan 1 : 47-141. CHAPMAN, V. J. (1956). The marine algae of New Zealand. Part I. Myxophyceae and Chlorophyceae. Journ. Linn. Soc. Lond., Bot. 55 : 333-501. DickiE, G. (1876). Contributions to the botany of H.M.S. ‘‘ Challenger”. XXIV. Marine algae collected by Mr. Moseley at the Island of Kerguelen. Journ. Linn. Soc. Lond., Bot. 15: 43-47- FREI, E., & PREston, R.D. (1961). Cell wall organization and wall growth in the filamentous green algae Cladophora and Chaetomorpha. I. The basic structure and its formation. Proc. R. Soc. Lond., Ser. B, 154 : 70-94. Gain, L. (1912). La flore algologique des régions antarctiques et subantarctiques. 2me. Expéd. Antarct. Frang., Bot. : 1-218. HameEL, A., & Hamet, G. (1929). Sur l’hétérogamie d’une Cladophoracée, Lola (nov. gen.) lubrica (Setch. et Gardn.). Compt. Rend. Acad. Sci. Par. 189 : 1094-1096. HAMEL, G. (1931). Chlorophycées des cétes frangaises. (Fin.) Rev. Algol. 6: 9-73. HAMEL, G., & LEMoOINE, P. (1952). Corallinacées de France et d’Afrique du Nord. Arch. Mus. Nation. Hist. Nat. Par., Sér. 7,1 : 15-136. Hariot, P. (1889). Algues. Miss. Sci. Cap Horn 5 : 3-109. (1907). Algues. Expéd. Antarct. Frang., Bot., Alg. : 1-9. 232 MARINE ALGAE OF GOUGH ISLAND Hooker, J. D. (1845-47). Botany of Fuegia, The Falklands, Kerguelen’s Land, etc. The Botany of the Antarctic Voyage. 1. Flora Antarctica : 209-574. Isaac, W. E. (1957). The distribution, ecology and taxonomy of Porphyra on South African coasts. Proc. Linn. Soc. Lond. 168 : 61-65. Jory, A. B. (1957). Contribuig¢ao ao conhecimento da flora ficol6gica marinha da Baia de Santos e Arredores. Univ. Sdo Paulo Fac. Fil. Ci. & Let. Bol. 217, Bot. 14 : 1-199. Knox, G. A. (1960). Littoral ecology and biogeography of the southern oceans. Proc. R. Soc. Lond., Ser. B, 152 : 577-624. LeEvrinG, T. (1938). Verzeichnis einiger Chlorophyceen und Phaeophyceen von Siidafrika. Act. Univ. Lund., Ny Foljd, Avd. 2, 34 (9) : I-25. (1953). The marine algae of Australia. I. Rhodophyta: Goniotrichales, Bangiales and Nemalionales. Ark. for Bot., Ser. 2, 2 : 457-530. (1956). Contributions to the marine algae of New Zealand. I. Rhodophyta : Gonio- trichales, Bangiales, Nemalionales and Bonnemaisoniales. Ark. for Bot., Ser.2,3: 407-432. (1960). Contributions to the marine algal flora of Chile. Act. Univ. Lund., Ny Foljd, Avd. 2, 56 (10) : 1-85. Manza, A. V. (1937). The genera of articulated corallines. Proc. Nation. Acad. Sci. Wash. 23 : 44-48. (1940). A revision of the genera of articulated corallines. Philipp. Journ. Sci. 71 : 239-310. Papenrfuss, G. F. (1960). On the genera of the Ulvales and the status of the order. Journ. Linn. Soc. Lond., Bot. 56 : 303-318. Post, E. (1959). Weitere Daten zur Verbreitung des Bostrychietum VII. Ayvch. Protistenk. 103 : 489-500. : StnHA, J. P. (1958). Chromosome numbers and life-cycles in members of Cladophorales. Brit. Phyc. Bull. 1 (6) : 24-27. SKOTTSBERG,C. (1923). Botanische Ergebnisse der Schwedischen Expedition nach Patagonien und dem Feuerlande 1907-09. IX. Marine algae 2. Rhodophyceae. K. Svensk. Vetensk- apsakad. Handl., Ny Foljd, 63 (8). (1941). Communities of marine algae in subantarctic and antarctic waters. K. Svensk. Vetenskapsakad. Handl., Ser. 3, 19 (4). STEPHENSON, T. A. & A. (1954). Life between tide-marks in North America. IIIa. Nova Scotia and Prince Edward Island : description of the region. Journ. Ecol. 42 : 14-70. SUNESON, S. (1943). The structure, life-history and taxonomy of the Swedish Corallinaceae. Act. Univ. Lund., Ny Foljd, Avd. 2, 39 (9) : 1-66. WomersLey, H. B.S. (1956). A critical survey of the marine algae of southern Australia. I. Chlorophyta. Austral. Journ. Mar. Freshw. Res. 7 : 343-383. Bull. B.M. (N.H.) Bot. 3, 5 PLATE 16 B Cc A. Theshore at Isolda Rock. B. Colpomenia sinuosa (Roth) Derb. & Solier, habit of plant (x2). C. Corallina goughensis Y. M. Chamberlain, habit of plant ( 2). Bull, B.M. (N.H.) Bot. 3, 5 PLATE 17 Dermatolithon nodulosum Y. M. Chamberlain. A. Habit of plant (x). B. Detail of nodules (x11). C. Transverse section through nodule with tetrasporic conceptacles. Bull. B.M. (N.H.) Bot. 3, 5 PLATE 18 Lithophyllum sp. Lithothamnion neglectum (Foslie) Foslie. A. Habit of G.I.S.S. 1548 (x $), B. Habit of G.I.S.S,. 1553 (x *). CC. Vertical section through G.I.S.S. 1507. oe i] Te | "= _ = owe : = i. - x a «tt a a ue a, a = or i) ; = : a a : 7 ae a) Xi : _ : > _ 7 _ - 7 7 : i ea | ; 7 - 7 : - : : 7 a | | : eas Bull. B.M. (N.H.) Bot. 3, 5 PLATE 19 | ‘A ) ~ B A. Polysiphonia howei Hollenb. Habit of plant (x10). B. Lophosiphonia scopulorum (Harv.) Womersley. Habit of plant (x 12). tf q THE CEYLON SPECIES OF ASPLENIUM W. A. SLEDGE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BOTANY Vol. 3 No. 6 LONDON: 1965 : af : Wise ere sre Apia eae at reeds Tee CEYLON SPECIES OF ASPLENIUM N.S OCT 1965 — x * é AL Wes BY W.. A. SLEDGE (University of Leeds) Xu/ J Ph. 233-277 ; 3 Text-figures ; Plate 20 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) BOTANY Vol. 3 No. 6 LONDON : 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), \y// 404 Ns Ne Asplenium disjunctum Sledge (holotype). a, whole plant; 0, frond tip with gemma; c, underside of pinnule ; d, rhizome scale ; é, spore. THE CEYLON SPECIES OF ASPLENIUM 275 differs in facies and in the position of the sori, which are always situated at the ends of the segments. The pinnules are narrower, less markedly cuneate and more sharply incised, the narrowness of the pinnules and their segments being accentuated by the tendency for their margins to be revolute in the dried specimens, a tendency which is not apparent in specimens of A. elmert. The latter species is known only from wet forests on high mountains in Luzon (Mt. Benguet) and Borneo (Mt. Kinabalu) at altitudes of 1,500-4,000 m. Both the stations for the Ceylon species are at low elevations on the eastern, dry side of the island. The name refers to the separate and distant stations, 50 miles (80 km.) apart, whence the two localized gatherings weremade. That from Mandagalain the Eastern Province is far removed from any other station whence I have seen collections. 24. Asplenium tenuifolium D. Don, Prodr. Fl. Nepal. : 8 (1825).—Bedd., Ferns S. Ind. : 44, t. 130 (1864) ; Handb. Ferns Brit. Ind. : 159 (1883). Rhizome oblique, paleaceous with narrow, finely pointed, dark-brown scales. Fronds tufted, tripinnate or quadripinnate, normally 30-40 x 6-12 cm., sometimes larger, lanceolate or oblong-lanceolate, the lowermost pinnae slightly reduced. Stipe and rhachis slender, grooved above, glabrous or with a few hairs. Lamina green, 10-30 cm. long, with 10-15 pairs of alternate, stalked, patent pinnae, the lower ones 5~7 cm. long, lanceolate, bipinnate to tripinnate, proliferous from scaly buds in the upper part, ultimate segments narrowed below and with acute to mucronate tips, apices of the frond and primary pinnae pinnatifid with linear segments ; glabrous above and beneath ;_ veins solitary, central, terminating well below the apices ; texture thinly herbaceous. Sori 1-2 to a segment; indusium pale, mem- branous, entire. Spores reniform to plano-convex, 30-36 xX 21I-27u, with a peri- spore forming a broad undulate wing and many anastomosing surface folds. CEYLON : Thwaites C.P. 3268 (BM; E; GH; K ; PDA—data cited on PDA sheet : Nuwara Eliya, Apr. 1856). Hakgala, Aug. 1881 (PDA). Nuwara Eliya, Robinson (K). Same locality, Ferguson (PDA). Same locality, Freeman 172 Cc (BM). Same locality, 1,950 m., rr Mar. 1954, Schmid 1492 (BM). Blackpool, near Nuwara Eliya, Freeman 170 a, 17% B (BM). Ramboda Pass, on rocks by stream in forest on the Maturata track, I,920 m., 17 Mar. 1954, Sledge 1306 (BM; K; US). 1899, Anderson (E). Ferguson (GH; US 815517). Gardner 1079 (K). Wall (E; GH). On rocks by streams about Nuwara Eliya. Ceylon, South India, North India from Nepal to Assam, west China (Yunnan), Thailand, Indochina, Formosa. This is a very distinct species on account of its finely dissected fronds. Hooker, Beddome and Clarke do not refer to the proliferous habit though many Indian specimens show small scaly buds, often dormant. In Ceylon specimens the fronds invariably bear buds, often two or three on each frond, which arise in the groove at or close to the junction of the pinna rhachis and first acroscopic pinnule. 276 THE CEYLON SPECIES OF ASPLENIUM 25. Asplenium decorum Kunze in Bot. Zeit. 6 : 176 (1848). Darea belangeri Bory in Bélanger, Voy. Ind.-or., Bot. 2 : 51 (1833). Asplenium belangeri (Bory) Kunze, loc. cit. (1848) ; non A. belangeri Bory (1833).— Bedd., Handb. Ferns Brit. Ind.: 161 (1883).—Hboltt., Fl. Malaya 2 : 434 (1954). Asplenium veitchianum T. Moore, Index Fil.: xlix (1857), 176 (1860). Asplenium prolongatum Hook., Sp. Fil. 3 : 209 (1860) pro parte, quoad specim. Zeyl.— Thw., Enum. Pl. Zeyl.: 384 (1864). Rhizome erect, scaly at the apex, scales ovate-acute, black with paler, brown edges. Fronds tufted, up to 30 cm. long. Stipes green, up to 10 cm. long, scaly below with narrow, dark, clathrate scales which are often jagged at the base, inter- mixed with and replaced higher on the stipe and on the under side of the rhachis and pinnae by scattered, pale-brown, appressed, irregularly laciniate to stellate- squamulose scales. Lamina pale-green, up to 20 cm. long, rarely more, 4-6 cm. broad, oblong-lanceolate to ovate in outline, apex acute or sometimes caudate, bipinnate ; midrib of rhachis raised above and bordered by narrow lateral wings ; pinnae 2-5 cm. long, linear-oblong cut down regularly into numerous (up to 12 pairs) linear, straight or somewhat falcate segments 1 mm. broad, the posterior margins decurrent to form a narrow wing to the pinna rhachis, the lowest acroscopic segment bifurcate, the rest simple and traversed by a single central vein; texture somewhat fleshy. Sori one to each segment on the anterior margin ; indusium whitish, not reflexed at maturity. Spores light-brown, plano-convex, 39-42 X 24-27u, with a perispore wing and surface folds. CEYLON : Thwaites C.P. 1348 (BM; E; GH; K; PDA). Kanneliya Forest Reserve, Southern Province, 150 m., 2 Apr. 1954, Sledge 1386 (BM). Beddome (BM). Ferguson (GH ;_ US 815511, 815512). Lowcountry forests, Hutchison (E). Wall (E + GL). 1837; Wight 132 (E). Epiphytic in wet forest in the south, especially (? exclusively) the Sinha Raja forest. Ceylon, Indochina, Malaya, Sumatra, Java, Borneo, Celebes, New Guinea, Philip- pines. Ceylon plants were referred by Hooker to A. prolongatum Hook. and were so named by Thwaites. They differ however in several respects. They lack the remarkably attenuated apical prolongation of the rhachis, they never bear a terminal bud, the shape of the frond is oblong-lanceolate to ovate, not linear-oblong, the pinnae are sparingly furfuraceous or stellate-squamulose beneath and their divisions are shorter but more numerous, and the basal scales are not so long or attenuated. Beddome’s illustration (Handb. Ferns Brit. Ind.: 162, fig. 79 (1883), as A. rutifolium) which represents typical A. prolongatum well illustrates the difference between this species and A. decorum. Sometimes the apex of the frond is drawn out or caudate, but it then bears simple, oblique and distant lobes and lacks the sudden transition from pinnae to long naked extensions bearing terminal bud swhich are so character- istic of A. prolongatum. The qualification in Hooker’s statement (Sec. Cent. Ferns: sub t. 42 (1861)) that “‘ Nearly every one exhibits the remarkable prolongation of the rachis, rooting at the apex ’”’ was evidently inserted to cover the few Ceylon specimens which were seen by him, as all the other gatherings he quoted (and DHE CEYLON SPECIES OF ASPLENIUM 277 many subsequent gatherings correctly referred to this species) are quite consistent in this character. Christensen evidently doubted that the Ceylon plant was identical with A. prolongatum as he excluded Ceylon from the distribution of that species as given in the third supplement to the Index Filicum. The African A. rutifolium (Berg.) Kunze, with which Ceylon plants were at one time identified, differs in its ovate or lanceolate pinnae, the basal acroscopic pinnules of which, and often the basiscopic ones also, are enlarged and pinnate. The epithet belangert which has been widely used for this species is not available as A. belangeri (Bory) Kunze is a later homonym of A. belangert Bory (= A. longissimum Blume). Moore pointed this out long ago (Index Fil: xlix (1857), 176 (1860)) and substituted the epithet vertchianum. But Kunze had already described A. decorum, based on a plant from Java (Zollinger 1260), which he considered specifically distinct from his A. belangert. The differences on which it was founded however are trivial and all recent workers have agreed in recognizing only one dareoid species in Malaysia, under which A. decorum has been cited as a synonym. This species requires further study. Whilst I am sure that the Ceylon population cannot be included under A. prolongatum Hook., I am by no means sure that it is identical with Malaysian A. decorum. Plants from Malaya and Indonesia are mostly considerably larger than those from Ceylon and have more regularly divided pinnae ; but the frond apex is the same in both and although Malaysian plants frequently produce buds they are not invariably proliferous as in A. prolongatum. A. decorum has often been regarded as a variety or form of A. ¢enerum differing only in its deeply and regularly pinnatifid pinnae. Although Holttum regards it as a distinct species, he considers it closely related to A. tenerum, with which it frequently grows in Malaya. In Ceylon, however, it is geographically separated from A. tenerum, which is confined to the mountainous region of the interior, A. decorum occurring only in the south of the island, and at lower elevations. Bull. B.M. (N.H.) Bot. 3, 6 PLATE 20 Srorsm anret dantbich efach obey oye bn ton Did he aan Pmwvnee + tome ‘ tee weg eee ne Aeesete: wed edency wana. See hanheceton. : Asplenium erectum Willd. (Bory 70, holotype.) INDEX TO VOLUME*3 The page numbers of the principal references and the new taxonomic names are printed in bold-face type, synonyms in italics. Names of infraspecific taxa are indexed only when they are new or are synonyms. Acalypha annobonae 113 Anthocodium fragile ; ; : - 196 ACANTHACEAE 108 APOCYNACEAE : ; : : aL LO! Achnanthes andicola 85 ARACEAE ‘ ; : ; : s- 116 baldjickii 81 Arachis hypogaea . : : : eos subsp. lorenziana 82 Aralia guilfoyler. : F : » 102 var. subquadrata 82 ARALIACEAE ; : ; : 5 LOZ baldjickii 82 Artocarpus communis. : , . 113 danica 76 heterophyllus. : s : . 113 var. maxima 72 integer 3 ‘ : : > E13 danica 2 Oe Arum sagittifolium ; : : . 116 fimbriata 72 ASCLEPIADACEAE . : : : 1 LO4 heteropsis 72 Asclepias curassavica. : 104 hungarica . 84 Asplenium, The Ceylon Species: of 233- 277 lorenziana . oi igtee Asplenium . : ; : : . 240 lyvata we 79 adiantoides : : : . 20%,.267 manifera 72 aethiopicum : ‘ ; : . 267 reichardtiana 70 affine : : : . 269 stvoemii G2 forma tenuisectum. ; : . 270 vaszaryi F 72 affine : : ; : : e270 Achnanthidium baldjickii 81 belangeri . ‘ : 3 : ) 270 hungaricum ‘ 84 brachyotus 3 : : : 252 neglectum 84 brasiliense : : : Z . 248 Achyranthes aspera 109 camptorhachis . : : : «248 Actinoneis lorenziana 82 caudatum . : : : . 258,259 Adenopus breviflorus 101 cheilosorum ; ‘ c : . 245 Adenostemma perrottetii 103 contiguum F ; : : . 259 Agelaea ovalis 98 coriaceum . : : ; , 255 Aghardia adhaerens 197 cristatum . : : : + 247 Alectra communis 107 cuneatum var. affine : : . ae], sessiliflora 107 decorum . ; : : : . 276 Allamanda cathartica 104 decrescens . : : : ; . 259 Alternanthera maritima . 109 disjunctum : : : : . 273 sessilis 109 elongatum . : ; : ‘ 253 AMARANTHACEAE 109 emarginatodentatum . : eas Amaranthus spinosus 109 ensiforme . : ; : : . 242 viridis : 109 erectum . : : ; ; . 248 AMARYLLIDACEAE . 115 evosodentatum ; 3 ‘ Sa eAT ANACARDIACEAE 98 erythrocaulon. , : ; 245 Andropogon brevifolius : 118 excisum 3 : F . . 2406 Angiosperms of the Cambridge Annobon falcatum . : ; ; . 261 Island Expedition 93-118 var. abbreviatuis : : : 3 204 Angraecum talbotii . 115 var. bipinnatum . : : . 262 Annobon Island Expedition, Angiosperms formosum . : : ; : . 243 of the Cambridge 93-118 furcatum . fi ; F ‘ = 267 Annona muricata > 96 gardneri. ‘ : é ; . 258 reticulata . 96 glaucophyllum . : é : . 265 ANNONACEAE 96 heterocarpum ‘ ; : ; . 2245 Anthephora cristata 117 hirtum : ‘ : ; a. 254 280 Asplenium—conid. inaequilaterale indicum Insigne laciniatum . laetum laserpititfolium longipes lunulatum . var. camptorhachis var. evectum var. sphenolobium var. tvapeziforme macraei forma sphenolobium macrophyllum minus multijugum mysovense . nidus nitidum nitidum normale obscurum . opacum pavonicum . pellucidum persicifolium var. laoham : planicaule . praemorsum var. latum prolongatum pyvamidatum vesectum serricula servviforme . spathulinum sphenolobium tenerum var. terminans teneyum tenuifolium trapeziforme unilaterale forma majus var. majus varians vettchiorum vulcanicum walkerae wightianum zenkeranum Avicennia africana germinans . Ball, P. W., and V. H. Heywood Bambusa vulgaris . : Bangia fuscopurpurea BANGIACEAE . Basella alba . INDEX BASELLACEAE : : . 2 2 EO 252 Beckeropsis Jaxior : ; ‘ . S18 264 nubica : : : : > “ris 265 Bertiera annobonensis : : ‘ . 102 264 Blidingia minima : : . 180,183 245 Boerhavia coccinea ; : ‘ . 108 267 diffusa ; : : , : - 109 256 diffusa : : : : ‘ . 108 248 Bolusiella talbotii . . 2 ; eet Ey 248 Bostrychia mixta ; : é 179,227 248 Bryonia palmata . : 2 : ~/ SIO 248 scabra : ‘ : : PLO? 252 Bryonopsis laciniosa ; IOL 248 Buchenavia and Ramatuella, N revision of 258 the genera . F , ; ; . I-46 243, 248 Buchenavia . : ‘ é ; : 4 243 acuminata 2 : : ; » «ao 262 callistachya : ; ; : ~ -20 242 capitata . : : : F see? 265 congesta . : ; ; : . 36 270 corrugata . : : : : : 19 243 discolor. : : ; F ee 247 fanshawei . 4 : : . eee 243 fluminensis : ; : : ese 243 gracilis : : ‘ ; : SY) 254 grandis. : ; ; : » oo 254 huberi : : ; ; : . 34 264 kleinii ; : , ‘ = : 14 267 longibracteata . : : : . 33 262 macahensis ; : ‘ : : 37 276 macrophylla F é : : ey 248 megalophylla. : ; : . 26 245 ochroprumna é : . : : 16 255 oxycarpa . : ‘ : : . 24 - 247 oxycarpa . : : : : - 2I 262, 205, 269 pallidovirens : : 5 : Sey 7/ 248 parvifolia . 2 : : : o&S 253 ptariensis . ‘ : : 3 a ee 253 pterocarpa : : ‘ : ee 258 pulcherrima . : : : oe 275 punctata . ; 3 ‘ : ae 3 252 reticulata . ; 3 j : ‘ 17 245 sericocarpa ; ‘ , : . 25 246 stellae c : ; . é <2 132 246 suaveolens , : : : . 21 272 tomentosa . : : Fi j 2 ld 276 vaupesana : : : : oe EZ 256 viridiflora . : ; : : . 28 256 Bucholzia maritima : : : . 109 255 Bucida angustifolia , : ‘ 3, a pk 254 capitata. : ‘ : oar iy 108 Buforrestia imperforata ‘ : : ee cde 108 Bulbophyllum melanorrhachis : - 215 Caesalpinia pulcherrima . : : a 119-172 Cajanus cajan : ; ‘ ‘ . 9 penile) Calonyction aculeatum ; A . 105 179, 208 Calothrix crustacea 179, 182 208 CAMPANULACEAE . P : : 4.” JEG 110 Canscora decussata ‘ ; ‘ . 105 INDEX Capartogramma, The diatom genus, and the identity of Schizostauron Capartogramma amphoroides crucicula . jeanit karstenii . rhombicum CAPPARACEAE Capsicum annuum . Carica papaya CARICACEAE . Caryophyllus avidus Cassia obtusifolia occidentalis tova Cassipourea annobouenss Catharanthus roseus CELASTRACEAE : Centroceras clavulatum . Cephalostigma perrottetii CERAMIACEAE : Ceramium clavulatum Ceylon species of Asplenium, The CHAMAESIPHONACEAE Chamberlain, Yvonne M. CHLOROPHYTA Chorda lomentaria . Citrus aurantifolia . Cladophora catenifera radiosa ‘ rupestris CLADOPHORACBAE . Clayton, W. D. Cleistanthus libericus Cleome ciliata rutidosperma Chitoria alba . Coccochloris deusta . Cocconeis beltmeyeri danica delicata : diaphana var. Bp dirupta flexella hungarica . CoDIACEAE Codiaeum variegatum Codium adhaerens . fragile é Coffea liberica Coix lacryma-jobi . Colpomenia sinuosa COMBRETACBAB : Commelina benghalensis COMMELINACEAE COMPOSITAE . Conferva ambigua fuscopurpurea granulosa 178, 179, 180, 227 ae 50 64 59 55 54 58 96 107 100 99 99 100 abd 103 227 227 233-277 181 173-232 180, 183 205 97 e192 179,192 179,194 187 117-118 111 96 2 As 179, 203 100 116 116 103 187 208 199 Conferva—contd. vadiosa vupestris tortuosa CONNARACEAE Convallaria fruticosa CONVOLVULACEAE . Convolvulus nil Corallina goughensis officinalis CORALLINACEAE Cordyline fruticosa Corymborkis corymbosa . welwitschit . CRASSULACEAE CRUCIFERAE . Cucurbita lagenaria sicevavia CUCURBITACEAB CYANOPHYTA ; Cyathula prostrata Cymbopogon citratus CYPERACEAE . ; Cyperus pedunculatus Darea belangeri Datura candida DELESSERIACEAE Dermatolithon nodulosum Desmodium mauritianum ramosissimum Dianthella compressa . Dianthera Bey iicillnts Dianthus sect. Kohlrauschia sect. Pseudotunica sect. Tunica ; subgen. Proliferastrum diminutus . filiformis glumaceus . nanteutlit obcordatus . prolifer vecticaulis . saxifvagus . velutinus Dichrocephala bicolor integrifolia Dicliptera umbellata verticillata DILLENIACEAE : Dimeregramma baldjickii Dioscorea alata cayenensis . dumetorum DIOSCOREACEAE Diplocyclos palmatus Diporochna quintasii 281 192 194 189 98 116 105 106 179, 213 178, 179, 180, 178, 179, 180 129, 211 282 INDEX Discoclaoxylon occidentale var. pubescens . 112 Gypsophila—contd. pubescens 112 glomerata . Dolichos unguiculatus : 99 gvaminea Durvillaea antarctica 178, 206 haynaldiana utilis : 206, 207 illyrica DURVILLAEACEAE . 206 multicaulis ochroleuca . EcTOCARPACEAE 197 pachygona Ectocarpus globifer . 197 vigida granulosus . 199 saxifraga mitchelliae . 201 var. glomerata . ViVESCENS 201 stricta Eleutheranthera ruderalis 103 thessala Enteromorpha africana a Od velutina bulbosa 179, 180, 184 gunniana 183 Hedyotis herbacea hookeriana . 184 Herposiphonia paniculata micrococca . 183 Heywood, V. H. minima 183 Hildenbrandia nana : 2 183 Hippia integrifolia . Entophysalis conferta 179, 180, 181 Hymenocallis littoralis deusta : 179, 181 Hypolytrum grande Epipogium roseum . 114 Eriosema glomeratum 99 Imperatia Euphorbia pulcherrima . 111 piliformis EUPHORBIACEAE LEY lpomoea alba Exacum quinquenervium 105 digitata Exell, A.W. ; 93-118 involucrata Exell, A. W., and C. A. Stace : 1-46 mauritiana nil 3 : Feldmannia globifera 178, 197 Ividaea augustinae . Fiedleria 129 boryana illyrica 133 ciliata Fucus antarcticus 206 cordata pyriferus 205 var. ciliolata crispata GELIDIACEAE Fe ZLO heterococca . Gelidium regulare . 178,210 laminarioides sp. 178 micans var. ciliolata GENTIANACEAE 105 undulosa Geophila neurodictyon 102 Iridophycus boryanum Gerardia sessiliflora o, A107 ciliatum : Giffordia granulosa 178, 199 crispatum . mitchelliae 180, 201 laminarioides GIGARTINACEAE 218 undulosum Gossypium barbadense ‘ 97. Txora coccinea Gough Island, Marine algae of 173-232 GRAMINEAE : : 117 Jatropha multifida Gymnosporia annobonensis 97 Justicia umbellata . Gypsophila sect. Petrorhagia 129 sect. Tunicastrum 132 Kalanchoe crenata subgen. Tunica 129 Kohlrauschia alpina 145 diminuta arenicola 151 glumacea armerioides 139 nanteutlit compressa . 136 obcordata cretica 141, 142 prolifera dianthoides 158 saxifraga fasciculata . 150 sibthor pit gasparrinii 154 velutina 171 155 134 133 151 138 142 154 151 154 145 156 166 ; 103 179, 180, 228 119-172 179 103 115 117 129 151 105 106 105 106 106 222 219 222 218 222 222 1 2io 179, 180, 218 eee "78, 179, 180, 222 219 223 223 218 223 102 112 108 99 129, 160 161 169 164 169 161 151 158 166 LABIATAE Lachnopylis angolensis annobonensis mannii thomensis Lagenaria breviflora siceraria LAURACEAE LEGUMINOSAE Leonotis nepetifolia LESSONIACEAE Lewis, J. LILIACEAE ; Limodorum voseum Lindernia diffusa Liparis welwitschii Lithothamnion muelleri forma neglectum d ae 179, 180, 217 neglectum . LOGANIACEAE Lola tortuosa . Lophosiphonia scopulorum Lophurella sp. LORANTHACEAE Luffa aegyptiaca ; Lychaete novae-zelandiae . Lycopersicon esculentum Lyngbya confervoides Macrocystis pyrifera MALVACEAE . Mangifera indica Marine algae of Gough Island . Maytenus annobonensis Megaclinium melanorrhachis Melampodium ruderale Melasma indicum var. monticola MELASTOMATACEAE MELIACEAE MELOBESIEAE Melothria capillacea cordifolia gilletia minutiflora : Microcladia alternata Microneis hungarica Mikania cordata scandens sp. MorRACEAE Mucuna sloanci urvens Nasturtium sinapis Navicula aqueductae bacilliformis caucasica crucicula jimbriata fusticulus INDEX 108 104 205 104-105 116 114 107 114 217 104 oO 179, 231 179, 229 110 101 187 . 106 179, 182 178,205 97 98 173-232 97 115 103 107 100 O77, 178 101 10L 101 Pe ey- 179,227 Navicula—contd. granum harstenit laevissima . laevissima . mutata , pseudopupula var. aqueductae pupula vaszaryt wittrockit Nicotiana tabacum Nitophyllum sp. Nuxia angolensis annobonensis congesta é var. thomensis mannit NYCTAGINACEAE Ochna cf. gilgiana membranacea quintasit OCHNACEAE . Ocimum basilicum . Odontidium baldjickit Oldenlandia herbacea Oplismenus hirtellus ORCHIDACEAE : Oscillatoria nigroviridis . OSCILLATORIACEAE Palmella conferta Pancratium littorale Panicum brevifolium Paspalum commersonii conjugatum Pennisetum polystachion Peperomia annobonensis . hygrophila Peponia vogelit Peponium bracteatum vogelii Persea americana Petrorhagia, A revision of the genus ‘ Petrorhagia sect. Dianthella_ sect. Kohlrauschia sect. Petrorhagia sect. Pseudogypsophila sect, Pseudotunica subsect. Creticae subsect. Illyricae subsect. Saxifragae subsect. Thessalae alpina subsp. olympica arabica armerioides var. laxa 179, 182 119-172 284 Petrorhagia—conld. candica cretica cyrenaica dianthoides fasciculata glomerata glumacea : var. obcordata graminea hispidula . illyrica : subsp. angustifolia subsp. haynaldiana subsp. taygetea kennedyae lycica macra multicaulis nanteuilii ochroleuca pamphylica peroninii phthiotica prolifera . rhiphaea . rigida saxifraga var. gasparrinii var. Slomerata thessala velutina PHAEOPHYTA Phaulopsis micrantha Phormidium submembranaceum Phyllanthus reticulatus . Physalis angulata . micrantha . minima PIPERACEAE . Pladera decussata Platylepis glandulosa Plectonema calothrichoides Pleurostauron sagitta Plumeria rubra Polyscias guilfoylei Polysiphonia boergesenii howei scopulorum Spies villum Polystachya ridleyi Porphyra tristanensis Portulaca oleracea . PORTULACACEAE Principina grandis . Psychotria neurodictyon PUNCTARIACEAE Punica granatum PUNICACEAE . 179, 179, 179, 179, 180, 178, 179,. INDEX 141 142 158 158 150 1 By fa 169 169 155 145 133 136 134 137 147 142 158 151 164 Ralfsia . Rama dobaciaelaniia Ramatuella argentea crispialata . var. obtusa latifolia maguirei obtusa virens , Remirea maritima . pedunculata 3 Rhaphoneis baldjickii fluminensis lorenziana . scutellum ; Rhizoclonium africanum ambiguum hookeri riparium subsp. ? validum var. validum tortuosum RHIZOPHORACEAE . ; Rhodoglossum revolutum RHODOMELACEAE RHODOPHYTA Rhynchosia glomerata Ricinus communis . RIVULARIACEAE Rorippa indica sinapis Ross, R. RUBIACEAE RUTACEAE Sansevieria metallica Saponaria cretica illyvica Schizachyrium brevifolium Schizostauron andicola brasiliense crucicula Simbriatum karstenit lindigianum ovatum veichardtianum vrhombicum tatricum verbanium SCROPHULARIACEAE Scytosiphon lomentaria . Sellaphora pupula . Sida rhombifolia Silene angustifolia . Sisymbrium indicum Sledge, W. A SOLANACEAE . 187 179, 180, 187 a ey) 179, 189 189 189 189 . 100 179, 180, 225 179, 227 208 99 113 182 96 107 : 96 233-277 106 Solanum dasyphyllum duplosinuatum nigyrum nodiflorum Solenia bulbosa Sphacelaria furcigera SPHACELARIACEAE . Stace, C. A. . Stanfieldiella imperforata Stauroneis andicola cornuta crucicula grunowtt harstenit lindigiana mervimacensts ovata veichardtiana rectangularis sagitta smithii stefanssonti tatrica verbania wittrockit Stearn, W. T. Summerhayes, V. S. Synedrella nodiflora Talinum triangulare Tarachia falcata furcata ; var. platyphylla nitida Telanthera mbalind Terminalia catappa hilariana macrophylla obovata oxycarpa punctata suaveolens . tomentosa vasivae Tetracera ainifolia., Thamnopteris nidus Thecacoris annobonae Trichomanes adiantoides aethiopicum Tristemma oreothamnos Tunica . ; sect. Dianthella . sect. Eutunica sect. Fiedleria . sect. Gypsophiloides sect. Kohlrauschia sect. Leptopleura sect. Pachypleura sect. Pleurotunica sect. Pseudodianthus INDEX 106 Tunica—contd. 106 sect. Pseudogypsophila 106 sect. Pseudosaponaria 106 sect. Tunicastrum 184 subgen. Imperatia 178, 202 subgen. Leptopleura 202 subgen. es I-46 alpina 116 angustifolia 84 arabica 72 avenicola 59 argentit 60 ayrmerioides 55 var. laxa 69 bicolor 60 brachypetala 69 ciliata 70 compressa . 87 cretica 92 cypria 92 davaeana 92 dianthoides 87 erecta 87 fasciculata . . 89 filiformis 106-1 07, IIO-III gasparrinit II4-115 glumacea 103 var. obcordata . ; gracilis 97 graminea 261 haynaldiana 267 hispidula 262 illyrica 265 subsp. angustifolia 109 subsp. haynaldiana . 100 var. haynaldiana 12 var. taygetea 32 kennedyae . 52 lycica 24 macra nf nanteutlit 25 ochroleuca . 19 var. phthiotica 21 olympica 96 orphanidesiana 242 pachygona ; 111 var. hirtituba . 261 pamphylica 267 peroninit 100 phthiotica 129 prolifera’ . 159 subsp. nanteuilii 148 var. nanteuilit . a ek29 rhiphaea 132, 148 rhodopea 160 vigida 145 saxifraga 132 var. gasparrinit 132 scoparia 148 sibthorpit 134, 134, Ve Vines BG. 133, 134, 285 145 132 148 129 145 132 145 136 143 I51 139 139 139 I51 142 151 136 142 147 136 158 I51 150 147 154 169 169 158 155 134 145 134 136 134 134 137 147 142 158 164 138 149 147 156 142 143 159 159 149 161 164 164 156 134 154 151 154 136 139 286 Tunica—contd. stricta syrviaca taygetea thessala var. cyrenaica . velutina xevophila Turraea glomeruliflora Ulva intestinalis var. nana lactuca sinuosa ULVACEAE 137, 141 178, 179, INDEX 145 170 156 158 166 151 97 183 186 203 183 Uncaria africana Urera cameroonensis URTICACEAE VERBENACEAE Vigna alba unguiculata Viscum engleri Xanthosoma sagittifolium Zehneria gilletii scabra scrobiculata do getetss pees bsp aereonns etetrtiiin Feely sees We vere PT ETE ce revere } + ae erepese Pr aesiat statebatelebesaae: pehebebe preees + Teh ahha atek etal ees bo ghee haict a1 idverbeerertrer’ cers yeas eat Hy shoe abee whe x tre rere weyte s seenneee! eeserestt tats stelatest ae ret re « ay thy. arte tf ~ye os bets +5 4 siz3! > 3et verere Eirsts shaese r +6 ree eeechet: a pesests + ahs pene! a Sphrpas sheds latetetel So iveede hs eeite te raster s ice . Tet Mpa eine +40 eee ett bomen beens Bs y deeet ah pebehehas: erenery, 7h; Sood ett sretetarst lever ers: Set etriest 4s poieaa os bat cbaboted ah pepertesitst SFeTe He t ante! tafeeiteastassca aigesyy sobs + é bo ted were eet =e athe : eaert sea shee ba bots ter ey ore simsetes a3 ener bash eee Creve peieces a + Hii > shit etiy Seatate settee. +h yee tbe athe attire WH preeeer es + i apres e' vtree ore i yeteos eerie Ste . 333 - 4 - St = = cs popes 3 Bret les es. ernie near’ ——. . es wopessrepoesee secs vena pokey aba ver oe batseapshee rest —rteee ee 2 . ee vaner meses eee eoece = Piste ets tone te tee ° “omnes eer et eset se eet kee ess nee fava tse se eds: Serer ees Se tm abe be rete ioe ess eeaeses basitt $1 irs apee—eneme 25 eh ears atr¢) sone: atta eter t poearies = what aeee ae. Seen los stegoecres ereresegesy ape Tents : tH eae poker nuntmen ve sereee = =Taese= ee & oteb-ecenee > teense wretnnepmnae pate ee oe Ter A eter ey ea poees petrsti pear FAT At se st see rae ors eee es Hees eek. eaee ehaeet - emia Serra eases * Saheeese Sra SETS, Setteces ate ena ee - - ry « stea-pate= Se Seren Fetsh at st atera iret ead _ 3 s5StSt sat < cree rere rere or aentaereret yee cache 35st25555. Serwubet oo 3825s Snece— ee acum sates he Peres ses ea see Pot erties estates es poe FE Cmte ee a <¢ = eeente SF St : rrssshsh serestes! 3 5 =s pers Sisteletets otrate eeeret erprew Fete res ete arietes ‘eter ates Vom Faeroe UA De imters ens sarees Fe Sectiiisetel tet yy akeeatetsien iotisegats { tosr ass? eis vhabhas wet