i ri _ 1 Rin tm hip thag Aan Par Aga te i eer Pte Rate, tpt atic Se Tass Par 2 eeF e ee wt=* - woe he Hire haga! wv * a > wire ” SPt oy fected sale . b a ekeeecat cee eR ate sateen tet = De rotates as : epee eed P ate tn "aims eto tanta avin eM ihecthehin ie Reanim hen aa Pi “ a oe pate: rere se me ~ See caipeaiend aan sie aie ae ie ea - ~~ 7 “ as ’ 2 eed av ¥ De® tient hs ms Pane Matias Parte el Sm nthnflng PEP Se sete og “et sP— Be” 7 oP a ee oe * ——) = arb § ren ah 8) a ae >" wie 02 the =) oh a oe Pa aici aps a Rok tea ina vats eis wens oad tene eka ® 6-6-$ ~ ate 463 een die en mS Ee eae TNO ST EE Gite ah Sete Sha Washoe SE a aicigcntt tn AHS f y - : 4 3 ws piensa 5 ul iis —— a ‘arya aN ae ™ ie tay mea : “ne 6 ZV f aa ro = kN Ta » hy aly i . ae Hina ot ay i wy x ROSE oo, be) ro | i | ‘i Il (| HE: CR. Gea WATSONIA JOURNAL OF THE POBANTCAL SOCIETY OF THE BRITISH ISLES Editor : M. C. F. PROCTOR, M.A., Ph.D. Volks PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES AT THE DEPARTMENT OF BOTANY, BRITISH MUSEUM (NATURAL History), LONDON 1961-63 DATES OF 2UBLICATION Part 1, pp. 1-46, 10 August 1961 Part 2, pp. 47-112, 27 December 1961 Part 3, pp. 113-176, 2 May 1962 Part 4, pp. 177-258, 23 November 1962 Part 5, pp. 259-328, 3 May 1963 Part 6, pp. 329-401, 30 November 1963 CONTENTS Part 1. SPARGANIUM IN BRITAIN. By C. D. K. Cook GERMINATION, SEEDLINGS AND THE FORMATION OF HAUSTORIA IN BUPHRASIA. By bE. F. Yeo . ie ie Ae a A ie a STUDIES ON WELSH | ORCHIDS. I. THE VARIATION OF DACTYLORCHIS PUR- PURELLA (T. & T. A. STEPH.) VERMEUL. IN NorTH WALES : STUDIES ON WELSH ORCHIDS. II. THE OCCURRENCE OF DACTYLORCHIS MAJALIS (REICHB.) VERMEUL. IN WALES. By R. H. Roberts : GERANIUM MICROPHYLLUM Hook. F. AS AN ADVENTIVE PLANT IN BRITAIN Byae. ©. Townsend . Mt A NEW VARIETY OF VALERIANELLA LOCUSTA ne) BETCKE. By 1D) E. Alien iParge2 INTERSPECIFIC RELATIONSHIPS AND INTRASPECIFIC VARIATION OF CHENOPODIUM ALBUM L. IN BRITAIN. I. THE TAXONOMIC DELIMITATION OF THE SPECIES. By M. J. Cole AN AUSTRALIASAN SPECIES OF CRASSULA INTRODUCED INTO ) BRITAIN, By J. R. Laundon NOTES ON RUMEX ACETOSA L. IN THE BRITISH ISLES. (BEITRAG | ZUR KENNTNIS VON Rumex No. XV). By K. H. Rechinger te Re a NOTE ON Dr. RECHINGER’S PAPER ON RUMEX ACETOSA. By J. E. Lousley . TRIFOLIUM OCCIDENTALE, A NEW SPECIES RELATED TO T. REPENS. By D. E. Coombe ot ee By, - es Se 5 SOME STUDIES IN CALYSTEGIA : COMPATIBILITY AND HYBRIDISATION IN C. SEPIUM AND C. SILVATICA. By Clive A. Stace TAXONOMY AND NOMENCLATURE IN SOME SPECIES OF THE GENUS ARUM I, Bye. 1: Prime BOOK REVIEWS Rant. 3: THE TAXONOMIC SEPARATION OF THE CYTOLOGICAL RACES OF KOHLRAUSCHIA PROLIFERA SENSU LATO. By W. Ball and V. H. Heywood ne INTERSPECIFIC RELATIONSHIPS AND INTRASPECIFIC VARIATION OF CHENOPODIUM ALBUM L. IN BRITAIN. II. THE CHROMOSOME NUMBERS OF C. ALBUM AND OTHER SPECIES. By M. J. Cole a STUDIES ON RANUNCULUS L. SUBGENUS BATRACHIUM (DC) TS Gray. It CHROMOSOME NUMBERS. By C. D. K. Cook 2 STUDIES IN THE BRITISH EPIPACTIS. V. EPIPACTIS LEPTOCHILA; WITH SOME NOTES ON E. DUNENSIS AND E. MUELLERI. By Donald P. Young STUDIES IN THE BRITISH EPIPACTIS. [V. SOME FURTHER NOTES ON E. PHYLLAN- THES. By Donald P. Young STUDIES IN THE BRITISH EPIPACTIS. VII. SEED DIMENSIONS AND ROOT DIAMETERS. By Donald P. Young. SOME NOTES ON GALEOPSIS LADANUM is AND iG ANGUSTIFOLIA EHR. EX Horrm. By C. C. Townsend We te CHROMOSOME NUMBER, MORPHOLOGY AND BREEDING BEHAVIOUR IN THE BRITISH SALICORNIAE. By D. H. Dalby . ae STUDIES IN THE BIOLOGY OF POA SUBCAERULEA ‘SM. By Dp: M. Barling je BOOK REVIEW iil enamels’ aA PAGE 1-10 11-22 23-36 37-42 43-44 45-46 A7T-58 59-63 64-66 67 68-87 88-105 106-109 110-111 113-116 117-122 123-126 127-135 136-139 140-142 143-149 150-162 163-173 174 Part 4. THE TAXONOMY OF POLYGONUM AVICULARE AND ITS ALLIES IN BRITAIN. By B. T. Styles NOTES ON BRITISH HIERACIA. II. THE SPECIES OF THE Once IsLANDS. By ie D. Sell and Cynly West) >: A STUDY OF VARIATION IN EUPHRASIA BY MEANS OF OUTDOOR CULTIVATION. By P. F. Yeo.. THE BRITISH FORMS OF TUBERARIA GUTTATA ie FOURREAU. By M. C. F. Proctor . LUZULA X BORRERI IN ENGLAND. By re E, Ebinger BOOK REVIEWS Parte: THE NOMENCLATURE OF ALCHEMILLA MINOR AUCT. BRIT. ae M. E. Bradshaw, P. D. Sell and S. M. Walters CIRCAEA IN THE BRITISH ISLES. By Peter HY Raven VARIATION IN SOME DIAGNOSTIC CHARACTERS OF THE SESSILE AND PEDUNCUL- LATE OAKS AND THEIR HYBRIDS IN SCOTLAND. By J. E. Cousens THE STATUS OF ORCHIS LATIFOLIA VAR. EBORENSIS GODFERY IN YORKSHIRE. By R. H. Roberts and O. L. Gilbert STUDIES IN RANUNCULUS (L.) SUBGENUS BATRACHIUM (DC,) ke GRAY. Il. GENERAL MORPHOLOGICAL CONSIDERATIONS IN THE TAXONOMY OF THE SUBGENUS. By C. D. K. Cook STUDIES ON ALCHEMILLA FILICAULIS BUS., SENSU LATO, AND A. MINIMA WALTERS. INTRODUCTION, AND I. MORPHOLOGICAL VARIATION IN A. FILICAULIS, SENSU LATO. By Margaret E. Bradshaw STUDIES ON ALCHEMILLA FILICAULIS BUS., SENSU LATO, AND A. MINIMA WALTERS. II. CYTOLOGY OF A. FILICAULIS, SENSU LATO. By Margaret E. Bradshaw BOOK REVIEW Part 6. ELEOCHARIS AUSTRIACA HAYEK, A SPECIES NEW TO THE BRITISH ISLES. By S. M. Walters VARIATION IN MELAMPYRUM PRATENSE L. By A. J. E. Smith FERTILE SEED PRODUCTION AND SELF-INCOMPATIBILITY OF HYPERICUM CALY- CINUM IN ENGLAND. By Sir Edward Salisbury THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS. By J. G. Done THE TAXONOMY OF POLYGONUM LAPATHIFOLIUM L., P. NODOSUM PERS. AND P. TOMENTOSUM SCHRANK. By J. Timson BOOK REVIEWS INDEX TO VOL. 5 1V PAGE 177-214 215-223 224-235 236-250 251-254 255-258 259-261 262-272 273-286 287-293 294-303 304-320 321-326 32H 329-335 336-367 368-376 377-385 386-395 396-398 399-401 ~ WATSONIA JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M, C. F. PROCTOR, M.A., Ph.D. Vol.5 JULY, 1961 Pt; 1 CONTENTS SPARGANIUM IN BRITAIN. By C. D. K. Cook re i a ag 1-10 GERMINATION, SEEDLINGS, AND THE FORMATION OF HAUSTORIA IN EUPHRASIA. By P. F. YEo ve aie He ss Ae oe os oe 2D STUDIES ON WELSH ORCHIDS. JI—THE VARIATION OF DACTYLORCHIS PURPURELLA (T. & T. A. STEPH.) VERMEUL. IN NORTH WALES. By R. H. ROBERTS .. be Me a a a aC oe 23-36 STUDIES ON WELSH ORCHIDS. II—THE OCCURRENCE OF DACTYLORCHIS MAJALIS (REICHB.) VERMEUL. IN WALES. By R. H. ROBERTS .... AD GERANIUM MICROPHYLLUM Hook. F. AS AN ADVENTIVE PLANT IN BRITAIN By C. C. TOWNSEND a ie Ue - i Ae .» 43-44 A NEw VARIETY OF VALERIANELLA LOCUSTA (L.) BETCKE. By D. E. ALLEN 45-46 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH Museum (NATURAL HIsToRY), LONDON PRICE FIFTEEN SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES Patroness: H.R.H. THE PRINCESS ROYAL OFFICERS . President : J. E: LOUSLEY Vice-Presidents : E. MILNE-REDHEAD, J. E. DANDY, Miss C. M. ROB and Dr. E. F. WARBURG Hon. General Secretary: J. G. DONY, 41 Somerset Avenue, Luton, Beds. . Hon. Treasurer: J. C. GARDINER, 301-307, Oxford Street, London, W.1. Hon. Editors: (Watsonia) M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT Hon. Meetings Secretary: H. J. M. BOWEN, Pomander House, Harwell, Berks. Hon. Field Secretary: P. C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Longfield, Kent. Hon. Junior Activities Secretary: P. F. HUNT, The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey. Applications for membership should be addressed to the Hon. Assistant Secretary, D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13, from whom copies of the Society’s Prospectus may be also obtained. WATSONIA Price to non-members, 15/— per part. To be obtained from E. B. BANGERTER, Botanical Society of the British Isles, c/o Dept. of Botany, British Museum, (Natural History), London, S.W.7. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary .Members, 15/- for Junior Members, and 10/— for Family Members, and the Society’s year runs from January 1st to December 31st. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13 SPARGANIUM IN BRITAIN By Cy Dyk Cook Institut fiir systematische Botanik der Universitat Miinchen The Sparganiaceae were monographed by Graebner (1900) in Das Pflanzenreich. About seven years later Professor Wladislaw Rothert, formerly at the University of Odessa, started to prepare a more comprehensive monograph. Unfortunately Rothert suffered the tragic fate of so inany elderly Polish and Russian scientists in the Russian Revolution and his work seems to have been lost. In 1911, however, he visited the herbaria at Kew and the British Museum and wrote notes on many of the herbarium sheets. I have used these notes as a guide to the taxonomy. SPARGANIUM L., Sp. P/., 971 (1753). Glabrous aquatic (occasionally semi-terrestrial) perennial herbs, reproducing vegeta- tively by long, thin, underground rhizomes. Stems simple or branched. Leaves linear, distichous, sheathing at the base, erect or floating. Flowers unisexual, crowded into separate globose capitula, the female capitula towards the base in each inflorescence; perianth of 3 — 6 radiate scales; male flowers of 3 — 8 stamens, the filaments sometimes partially united; female flowers of one, occasionally two, rarely three, fused carpels with a single style persisting in fruit, and as many stigmas as carpels. Fruit drupaceous with a dry, spongy exocarp, and a hard endocarp; seed albuminous, with a large embryo. Pollination mainly by wind. There are about fifteen species in the North Temperate regions, extending from sub- arctic Scandinavia and North America to the Mediterranean and Mississippi Basin; the distributional belt stretches across North America, Europe, and Asia to Japan. Two species occur in South Australia and New Zealand. S. erectum has been collected at 12,000ft. in Tibet. The morphology and biology of the genus are dealt with by Gliick (1911, 1924), Kirchner, Loew & Schroter (1908), Muenscher (1944), and Steinbauer & Neal (1948). Most of the species have been investigated cytologically; so far, they have all been found to have the diploid chromosome number of 2n = 30. (Wulff 1938, Hagerup 1941, Lohammer 1942, Love & Love 1948, 1956). In view of this uniformity, the cytology was not investigated further by the writer. The genus was widely distributed in the Northern Hemisphere in late Tertiary times. The fossil forms are very much like the plants of to-day except in the larger number of loculi. S. multiloculare, described by Reid from Bembridge, showed two to five loculi,while Dr. Hartz recorded an inter-glacial form of S. erectum with two to four loculi. Only S. erectum to-day frequently shows plurilocular ovaries; the highest number seen is three. Ascherson & Graebner, Synopsis der mitteleuropdischen Flora 1, 279, (1897), proposed three sections :— I. ERECTA : Large plants with distinctly keeled or triquetrous leaves, the floating ones (when present) keeled towards the apex. Style and stigma long, filiform. II. NATANTIA : Floating leaves rounded or flat on the back, never keeled, the midrib usually obsolete towards the apex. Erect leaves obtusely triquetrous or keeled towards the base. Ill. MINIMA: Small plants. Leaves all flat, not keeled. Stigma ovoid, often sessile. - Their section Natantia is ill-defined, and floating forms of Erecta or erect forms of Minima can easily be incorporated into it. Watsonia 5 (1), 1961. SMITHSONIAN 43 ae 2 Cc. D.*KeCOOK Holmberg (Bot. Notis. (1922), 203-209) suggested that the genus be subdivided into two subgenera. He based his classification on a single perianth character :— I. MELANOSPARGANIUM : Tepala crassiora firmiora fusco-atra. Il. XANTHOSPARGANIUM : Tepala tenuia dilute colorata. This classification is a great improvement on that of Ascherson & Graebner, as it splits the large erect forms from the smaller floating ones. In addition to the diagnostic characters cited by Holmberg, the seed of Melanosparganium (which must now be called Sparganium) has six to ten longitudinal ridges, whereas in Xanthosparganium the seed is smooth. Hybrids between the subgenera have not been observed by the author. KEY TO BRITISH SPARGANIA la. Perianth segments black-tipped; inflorescence branched, male capitula on lateral branches; seed with 6-10 longitudinal ridges : subgen. SPARGANIUM. 1. S. erectum 2a. Fruit with distinct shoulder, upper part dark brown to black. 3a. Fruit large, (S—) 6-8 (-10) mm. long (excluding style), (3—) 4-6 (-7) mm. wide (at shoulder); upper part of fruit flattened. subsp. erectum 3b. Fruit smaller, 6-7 (-8) mm. long, 2°5—4°5 mm. wide; upper part of fruit domed, wrinkled below style. subsp. microcarpum 2b. Fruit with indistinct shoulder, spherical to ellipsoidal, upper and lower parts uniform, shiny, light brown. 4a. Fruit ellipsoid, 7-9 mm. long, 2-35 mm. wide. subsp. neglectum 4b. Fruit + spherical, 5-8 mm. long, 4-7 mm. wide. subsp. oocarpum 1b. Perianth segments translucent, not black-tipped; inflorescence unbranched, all male capitula on the main axis; seeds smooth: subgen. XANTHOSPARGANIUM. 5a. Male capitula remote, usually more than three; cauline leaves keeled but not inflated at the base, triangular in cross-section. 2. S. emersum 5b. Male capitula approximated, usually less than three; cauline leaves not keeled at ihe base, flat in cross-section. 6a. Leaf-like bract subtending lowest female capitulum 10-60 cm. long, at least twice as long as the whole inflorescence; male capitula usually two, elongated. 3. S. angustifolium 6b. Leaf-like bract subtending lowest female capitulum 1-5 (—8) cm. long, barely exceeding the inflorescence; male capitula usually solitary. 4. S. minimum Sub-genus SPARGANIUM (Melanosparganium Holmberg, 1922). Perianth segments thick, with dark brown to black tips. Seeds with 6 — 10 longitudinal ridges. SPARGANIUM ERECTUM L., Sp. P/., 971 (1753). S. ramosum Huds., Fl. Angl., 2, 401 (1778). The following description covers all the segregates discussed later : Plant large, erect, semi-terrestrial, (30 —) 50 — 150 (~ 200) cm. tall. Leaves triangular in cross-section, usually erect (may be floating in young plants or in deep water); apex broadly rounded, truncate or apiculate. Inflorescence branched (rarely simple) with male capitula borne above the female capitula on lateral branches. On banks of ponds and slow flowing rivers, in ditches and on ungrazed marshland. Within S. erectum L. there are about five forms with distinct fruit-shapes. In Europe, Asia and North America these fruit-shapes have been recognised and given taxonomic ranks varying from species to variety. In Britain four forms have been found, but so far only S. neglectum Beeby has received any attention. During the past two years a critical search for characters correlated with fruit-shape has been undertaken, but with little success. The only difference found was a slight variation in geographical range. Ecological differences have been reported, but these could not be confirmed. In fact, the opposite seemed to be the case ! Along the Trent and Mersey canal south of Derby, for example, all four forms were found growing within Watsonia 5 (1), 1961. SPARGANIUM IN BRITAIN 3 a quarter of a mile of each other, and all under what appeared to be similar conditions. Despite their sympatric distribution, intermediate fruit-shapes were not found. Fig. 1. Fruits of the subspecies of Sparganium erectum L., X c.7. (a) subsp. erectum, (b) subsp. microcarpum (Neuman) Hylander, (c) subsp. neglectum (Beeby) Schinz & Thell., (d) subsp. oocarpum (Celak.) C. D. K. Cook. I am adopting the classification used by NHylander (1953) in his Nordisk Karlvaxtfiora. \n this work subspecific rank is assigned to three of the forms which I have recognised as occurring in Britain; the fourth (S. oocarpum Celak) is here given subspecific rank in conformity with Hylander’s treatment. S. ERECTUM L., Sp. P/., 971 (1753) subsp. ERECTUM. S. ramosum Huds. emend. Beeby, J. Bot. Lond., 23, 26, 193 (1885); S. ramosum subsp. polyedrum Asch. & Graeb., Syn. mittel- europ. FI., 1, 283 (1897); S. erectum subsp. polyedrum Schinz & Thell., F/. Schweiz. ed. 3, 2, 14 (1914); S. ramosum var. polyedrum Holmberg, Skan. FI., 1, 79 (1922); S. polyedrum Juz. in Komarov, F/. U.R.S.S., 1, 219 (1934). Fruit cuneate-obpyramidal, (5 —-) 6-8 (— 10) mm. long, (3 -) 4-6 (—~7) mm. wide, with a distinct shoulder between upper and lower part; lower part pyramidal, (4 -) 5-7 (— 8) mm. long, light brown in colour; upper part flattened, dark brown to black in colour; in cross-section sharply 3 — 5-angled. Style less than 2 mm. long, persistent. Ovary usually bilocular, but occasionally uni- to trilocular (Fig. 1a). Watsonia 5 (1), 1961. 4 C.D. “KECOOK From herbarium material and descriptions it appears that the North American S. eurycarpum Engelmann is the same as S. erectum subsp. polyedrum (Asch. & Graeb.) Schinz. & Thell. from Europe and Asia. If this can be established then the following should be included in the synonymy: S. eurycarpum Engelmann, Gray. Man. Bot. 5, 48 (1864); S. californicum Greene, Bull. Calif. Acad., 3, 11 (1884). South Sweden and Finland to South Mediterranean, extending eastwards to Central Siberia. In Britain occurs south of the Wash only. S. eurycarpum in North America extends from coast to coast, the southern limit being in Florida; its northern limits are unknown. S. ERECTUM subsp. MICROCARPUM (Neuman) Hylander, Nordisk Kdrlvdxtflora, 1, 83 (1953). S. ramosum forma microcarpum Neuman, Krok. Hartm. Handbk., 12, 112 (1889); S. micro- carpum Celak., Osterr. Bot. Zeitschr., 66, 281 (1896); S. ramosum var. microcarpum Asch. & Graeb., Syn. mitteleurop. Fl., 1, 281 (1897). Type, collected from central Czechoslovakia by Neuman, ? in Prague (7.v.). Fruit obpyramidal with a rounded apex, 6 — 7(—8) mm. long, 2:5—4-5 mm. wide, with a distinct shoulder between the upper and lower parts; lower part pyramidal, 4- 6 mm. long, light brown with a slight constriction below the shoulder. The upper part domed with longitudinal ridges below the style, dark brown in colour. Below the shoulder fruit 3 —5 angled in cross-section. Style persistent, less than 2mm. long. Ovary usually uni- locular, rarely bilocular (Fig. 1b). Extends from the Arctic circle southwards to north Africa and east to Siberia. Occurs throughout the British Isles. S. ERECTUM subsp. NEGLECTUM (Beeby) Schinz & Thell. F/. Schweiz., ed 3, 2, 14 (1914). SS. neglectum Beeby, J. Bot. Lond., 23, 26, 193 (1885); S. ramosum subsp. neglectum Neuman, Krok. Hartm. Handbk., 12, 112 (1889); S. erectum var. neglectum Fiori & Paoletti, F/. Anal. Ttal., 1, 146 (1896-98). Type collected at Albury ponds near Guildford, Surrey; lectotype in Herb. Kew ! Isotype in Herb. Edin. ! Fruit ellipsoidal, 7-9 mm. long, 2-3-5 mm. wide, shoulder between upper and lower parts indistinct. Upper and lower parts conical, uniform light brown, shiny. Lower part 4-6 mm. long. In cross-section fruit barely angled. Style persistent, usually more than 2mm. long. Ovary unilocular. (Fig. Ic.). From South Sweden to North Africa, extending eastwards to the Caucasus. In Britain most common south of the Wash, but northern limit in Westmorland. S. erectum subsp. oocarpum (Celak,) comb, nov. S.neglectum var. oocarpum Celak., Osterr. Bot. Zeitschr., 21, 5 (1896); S. oocarpum Ostenf. & Hansen, Bot. Tidsskr., 21, 5 (1897). Described from specimens collected from Bohemia ? now in Herb. Berlin n.yv., Fruit ovoid to spherical, 5-8 mm. long, 4-7 mm. wide, uniform light brown in colour, shoulder between upper and lower parts indistinct, upper part hemispherical, lower part broadly pyramidal, 2-5-5 mm. long, almost circular in cross-section. Style persistent, usually less than 2 mm. long. Ovary unilocular, rarely bilocular (Fig. 1d). Distribution imperfectly known in Europe; specimens have been seen from Turkey and North Africa. Not found north of the Wash in Britain. Very poor fertility is shown in all the specimens that have been examined. This may be due to hybrid origin, although this has not been tested; on morphological grounds, the suggested parents are subsp. erectum and subsp. neglectum. Sub-genus XANTHOSPARGANIUM Holmberg 1922. Perianth scales thin, lightly coloured. Seeds without longitudinal ridges. SPARGANIUM EMERSUM Rehman, Verh. Nat. Ver. Briinn, 10, 80 (1871). S. erectum L., Sp. P1., 971, p.p. (saltem quoad [ partim?}) (1753) ; emend. Wahlenberg, FI. Suec., 2, 582 (1826); Watsonia 5 (1), 1961. SPARGANIUM IN BRITAIN 5 S. simplex Huds. sensu Curtis, F/. Lond., 5, 66 (c. 1788); S. multipedunculatum Morong., Bull. Torr. Bot. Club, 15, 79 (1888); S. glehnii Meinsh., Mel. Biol. Acad. St. Petersb., 13, 388-390 (1893); S. simile Meinsh., loc. cit. (1893). S. splendens Meinsh., loc. cit (1893). S. subvaginatum Meinsh., loc. cit. (1893); S. diversifolium var. acaule Fernald & Eames, Rhodora, 9, 88 (1907); S. chlorocarpum Rydberg, N. Amer. Fl., 17, 8 (1909). Type collected in the River Rov at Bar in the Ukraine, syntype in Herb. Kew ! Erect or floating aquatic, 20-60 cm. tall. Leaves triangular in cross-section, sheathing at the base but not inflated. Inflorescence simple, unbranched; male capitula 3-10, distinct, remote and all borne on the main axis; anthers when mature 6-8 times as long as broad; female capitula 3-6 with lower ones frequently stalked. Fruit pedunculate, ellipsoidal, frequently with a constriction around the centre; peduncle 2-3 mm. long; style persistent, 3-4 mm. long. \ Fig. 2. Flowering shoots and fruiting capitula (x c. 4) of (a) Sparganium angustifolium Michx., and (b) S. minimum Wallr. Watsonia 5 (1), 1961. 6 C. D. K. COOK In shallow rivers and canals and by the edge of ponds and lakes, on wet mud or in water to a depth of 100 cms. Throughout North Temperate regions from the Arctic Circle south to lat. 40°N. Occurs throughout the British Isles, but not found on very high ground. S. ANGUSTIFOLIUM Michx., F/. Bor. Am., 2, 189 (1803). S. natans L., Sp. Pl., 971, p.p. (1753); S. simplex Huds., Fl. Angl., 2, 401, p.p. (1778); S. alpinum D. Don ex G. Don, in Loud. Hort. Brit., 375, nomen nudum (1830); S. affine Schnizlein, Typhac., 27 (1845); S. boreale Laestadius ex Beurl, Oefeurs. Vet. Akad. Faerhaud., 9, 192 (1852); S. vaginatum Larsson, Fi. Verml., 259 (1859); S. borderi Focke, Bremen Abh., 5, 409 (1877). Described from specimens collected from Lake Mistassini in Canada. Type specimen ? at Herb. Mus. d’Hist. Nat. Paris n.y. Floating (rarely erect) aquatic. Leaves flat in cross-section, sheathing and inflated at the base. Inflorescence simple, unbranched; male capitula 1—2 (—3), clustered together, appearing as an elongated capitulum on the main axis; anthers 3-4 times as long as broad; female capitula 2-4, lower capitula stalked; leaf-like bract subtending lowest female capitulum 10-60 cm. long and at least twice as long as the whole inflorescence. Fruit shortly pedunculate, ellipsoidal and light brown in colour (Fig. 2a). In highland or northern acidic peat lochs, in water from 10 to 150cm. deep. An arctic-alpine species found in suitable habitats in Europe from Iceland to the Pyrenees and Alps, extending eastwards across Asia to Japan. Distribution limits for North America unknown. In Britain found in north Scotland, Pennines, north Wales and north- ern and western Ireland. There is also a single location near Beaulieu, Hampshire. S. MINIMUM Wallr., Erster Beitrag Fl. Hercyn, 2, 297 (1840). S. natans L., Sp. Pl., 971, nom. ambig. (1753); S. simplex Huds., Fl. Angl., 2, 401 nom. illegit., quoad. var. (1778); S. natans var. minimum Hartm., Handb. Skand. FI., 43 (1820); S. gramineum Wallr., loc. cit., ined. in syn. (1840); S. minimum Fries, Summa. Veg. Skand., 2, 560 (1849); Bot. Notiser, 154 (1849); S. rostratum Larsson, Fl. Verml., 260 (1859); S. septentrionale Meinsh., Bull. Soc. Nat. Moscow, N.S., 3, 174 (1889); S. ratis Meinsh., /oc. cit. (1889); S. flaccidum Meinsh., Mel. Biol. Acad. St. Petersb., 13, 393—4 (1893); S. perpusillum Meinsh.., loc. cit. (1893). The original description was based on plants from the southern Harz Mountains in Germany. No specimens have been found in Wallroth’s herbarium at Halle. Floating (very rarely erect) aquatic. Leaves flat, usually translucent, barely inflated at the base. Inflorescence simple, unbranched; male capitula one (rarely two), if two then close together and appearing as one; female capitula (1—) 2-3, usually sessile; leaf-like bract subtending lowest female capitulum 1-5 (—8) cm. long, barely exceeding the whole inflor- escence. Fruit sessile, obovoid with very short persistent style, uniform, shiny light brown in colour. (Fig. 2b). In lochs, pools or ditches with a rich organic substratum, in 10-50cm. of water. Temperate and arctic regions of North America, Europe and Asia. Found throughout Britain in suitable habitats. NOMENCLATURE The nomenclature of the British Spargania is difficult to elucidate, mainly because the earlier taxonomists chose to diagnose the species on characters that do not remain constant under all ecological conditions. Linnaeus in his Species Plantarum, 971 (1753), recognised two species of Sparganium : 1. S. erectum (foliis erectis triquetris). 2. S. natans (foliis decumbentibus planis). Hudson (Flora Anglica, ed. 2 (1778), 2, 401) was apparently not satisfied with Linnaeus’s classification, so he dropped the Linnaean species and proposed two new ones : 1. S. ramosum (foliis ensiformibus triquetris, caule ramoso). Watsonia 5 (1), 1961. SPARGANIUM IN BRITAIN i 2. S. simplex (foliis ensiformibus planis, caule simplice). The original description of S. erectum L. covers the plant later described as S. ramosum by Hudson. S. erectum L. is cited by Hudson as a synonym of S. ramosum Huds., so that the latter is an illegitimate superfluous name. Within his S. simplex Hudson described a variety 8 natans (foliis decumbentibus planis) ; as this is the same diagnosis as that of S. natans L., the varietal name must be based on the Linnaean species, so that the name S. simplex Huds. is illegitimate also. Curtis in Flora Londinensis (c. 1788), 5, 66, produced the earliest known full des- criptions and illustrations of the two Hudson species. The Curtis description of S. simplex Hudson is the one accepted today; unfortunately it differs from the S. simplex diagnosed by Hudson. S. simplex Huds. was originally diagnosed as having flat leaves, while Curtis illustrated and described it with triquetrous leaves. In the Linnaean herbarium there are three specimens of Sparganium: one of S. erectum L. and two of S. natans L.; of these, the specimen of S. erectum L. and one of the sheets of S. natans L. bear these names in Linnaeus’ handwriting. The specimen of S. erectum L. was found to be S. simplex Hudson sensu Curtis. The specimen in Clifford’s herbarium based on the Hortus Cliffortianus (referred to in synonymy by Linnaeus) was examined in the British Museum, and was also found to be S. simplex Hudson sensu Curtis. Linnaeus in Species Plantarum cited under S. erectum L. a variety B (non ramosum) with a reference to Bauhin’s Theatre, (1620, p. 231) and Pinax, (1623, p. 15). Under Bauhin’s list of synonyms there is a reference to S. alternum in L’Obel’s Historia Plantarum (1570), 41. The illustration of S. a/ternum shows a plant that is undoubtedly S. simplex Hudson sensu Curtis. On the other hand, under S. erectum (var. «), Linnaeus cites S. ramosum in Bauhin’s Theatre and Pinax. By referring back to the illustrations in L’Obel’s Historia Plantarum, it can be seen that var. « refers to S. ramosum Hudson. It appears justifiable to assume that Linnaeus recognised the differences between the plants later described as S. ramosum Hudson and S. simplex Hudson sensu Curtis, but did not think them worthy of specific rank. It is quite likely that Linnaeus chose to represent only var. 8 in his herbarium as var. « is a large and difficult plant to mount on a small herbarium sheet. Today it is recognised that there are two species within S. erectum L. From the preceding argument, S. ramosum Hudson should bear the name S. erectum L. For the plant which has been called S. simplex (following Curtis’s misinterpretation of Hudson’s name), the earliest available legitimate name must be used. This appears to be S. emersum Rehman, Verh. Nat. Ver. Briinn, 10, 80, (1871). The sheet of S. natans L. bearing this name in Linnaeus’s handwriting has on it a well-preserved specimen of S. minimum Wallr. while the other sheet bears a poor specimen of S. emersum Rehman. Hylander (1945) rejected S. natans as a nomen ambiguum. This decision was fully justified and the earliest available legitimate name appears to be S. minimum Wallr. In 1903 Fernald examined type material of S. angustifolium Michx. and declared it to be the same as S. affine Schnizl. Apparently this had been noticed before as Morong (1888) remarked ‘“‘ Engelmann has seen Michaux’s specimen at Paris and it is the same as S. affine,’ but he then proceeded to drop the name S. angustifolium and use the later name S. affine. S. angustifolium is the earlier name and must of course be used in place of S. affine. BREEDING BEHAVIOUR Sparganium is strongly protogynous and in S. erectum the lower female capitula have frequently passed their fertile stage before the first male capitulum matures. This insures cross-fertilization for some flowers at least, but the lower capitula rarely set much seed. No self-incompatibility or agamospermous mechanisms have been found. Watsonia 5 (1), 1961. 8 Cc. D. K. COOK Many hybrids have been reported, but the only hybrids I have seen (including examina- tion of herbarium material) are between S. emersum and S. angustifolium. These two species are fairly widely separated ecologically, the former being a species of eutrophic and the latter a species of oligotrophic waters, but they occasionally grow close together in parts of western Scotland. The hybrids appear to be fully fertile and many plants have been found that showed introgression. At Stoer in Sutherland and on the Isle of Raasay the appearance of the hybrids tends towards S. emersum while in Galloway introgression appears to be towards S. angustifolium. This allopatric introgression appears to be similar to the type found in Typha by Fassett & Calhoun (1952). EVALUATION OF THE TAXONOMIC CHARACTERS The object of much of this study was to ascertain those characters which are of most value in separating the species. Many of the characters that have been used by previous authors have proved inconstant or readily modified by environment. The charac- ters are annotated below and their value assessed. The key to the species is based on those characters that have proved most constant. a. Leaf Habit. The difference between erect and floating leaves is relative and is controlled largely by the habitat. S. erectum usually has erect leaves but in deep or flowing water they may be floating. S. emersum has erect and floating leaves, the floating leaves being found on deep-water forms and on forms exposed to wave action or water currents. S. angust- ifolium has floating leaves, but occasionally erect ones are found in very sheltered habitats. S. minimum is very rarely found with erect leaves. Cross-section. S. erectum and S. emersum show a triangular cross-section. This is a good character when examining specimens in the field, but cannot easily be seen on her- barium specimens. S. angustifolium and S. minimum both have a flat cross-section. This character breaks down in S. emersum in the north of Scotland, possibly owing to hybridiza- tion. Apex. In S. erectum the apex of the leaf has been used as a character to separate the various subspecies. Subsp. neglectum has been reported as having a truncate apex and subsp. erectum an apiculate one. These are not good characters and forms of subsp. neglectum have been seen with apiculate apices and forms of subsp. erectum with truncate apices. Cross Veins. S. erectum has been described with obscure cross-veins and longitudinal veins appearing as pellucid lines without a dark border, while S. emersum is said to have distinct cross-veins and longitudinal veins with a dark green border. I can find no justifica- tion for these statements after examining living and dead material of both species. Sheath. In S. erectum and S. emersum the leaf bases sheathe the stem. In S. angust- ifolium the leaf bases are inflated. This is a reliable taxonomic character for studies in the field especially when there are no flowers, but it is difficult to see in herbarium specimens. In S. minimum the leaf base is flat and barely sheathes the stem. b. Stem Colour. The base of the stem has been reported to be pink in S. erectum subsp. neglec- tum, but this is an unreliable character and is occasionally found in all subspecies of S. erectum especially when growing in stagnant water. c. Inflorescence Branching. The inflorescence is branched only in S. erectum; in all the other species it is simple. Leaf-like bract subtending the lowest female capitulum. In S. erectum and S. emersum this bract is triangular in cross-section and about one to one-and-a-half times as long as the inflorescence. In S. angustifolium and S. minimum it is flat in cross-section. In'S. Watsonia 5 (1), 1961. SPARGANIUM IN BRITAIN 9 angustifolium it is usually between 10 and 60cm long, the absolute length being directly proportional to the depth of water and the length of the peduncle supporting the lowest female capitulum. In S. minimum the bract is between | and 5 (—8) cm long; again, the absolute length is controlled by the habitat conditions. On this character S. angustifolium and S. minimum can easily be separated. In S. angustifolium the bract is usually more than twice as long as the whole inflorescence, but in S. minimum it barely exceeds it. The taxo- nomic value of this character does not appear to have been previously recognised. d. Flower Perianth segments. In S. erectum the perianth segments are thick with a black tip. In the other species they are thin with a translucent tip. This is a good character but is sometimes a little difficult to see on very old herbarium specimens. Beeby (1885) described the perianth segments of the female flowers of subspecies neg/ectum as being linear with a broad spathulate apex, while those of subspecies erectum were described as mostly ligulate, more membranous and scarcely or not enlarged at the apex. No distinctions can be made between these two types as intermediates appear to be more common than the extremes. In all the subspecies of S. erectum the perianth segments are very variable and determinations can not be made using this character. Anther. In S. erectum and S. emersum the anthers are six to eight times as long as broad, and in S. angustifolium and S. natans they are only three to four times as long as broad. This is a good character, but it breaks down in S. emersum in the north of Scotland. Carpel. In S. erectum the carpel is sessile and frequently plurilocular. Subsp. erectum usually has bilocular ovaries, whereas subspp. microcarpum and oocarpum usually have unilocular ovaries though bilocular ones are occasionally seen. S. emersum and S. angusti- folium have shortly pedunculate carpels with long styles which persist in fruit. S. minimum has sessile carpels with short styles that do not persist in fruit; these characters are constant. Seed. The seed has 6-10 longitudinal ridges in S. erectum but is smooth in the other species. This character does not appear to have been previously noted. ACKNOWLEDGMENTS This study was started at Edinburgh University under the direction of Dr. P. H. Davis. I would like to thank him, Dr. S. M. Walters and Prof. Dr. H. Merxmiiller for valuable discussions. I would also like to thank Mr. J. E. Dandy for sparing so much time on the nomenclatural problems. REFERENCES CLAPHAM, A. R., TuTIN, T. G. & WARBURG, E. F. (1952). Flora of the British Isles. Cambridge. ASCHERSON, P. F. A. & GRAEBNER, P. (1897). Synopsis der mitteleuropdischen Flora, 1. Leipzig. BAUHIN, K. (1620). Prodromus Theatri Botanici. Frankfurt-am-Main. BAUHIN, K. (1623). Pinax Theatri Botanici. Basel. Curtis, W. (c. 1788). Flora Londinensis, 5. London. FAsseTT, N. C. & CALHOUN, B. (1952). Introgression between Typha latifolia and T. angustifolia. Evolution, 6, 367. FERNALD, M. L. (1922). Notes on Sparganium. Rhodora, 24, 26. GLuck, H. (1911, 1924). Biologische und morphologische Untersuchungen iiber Wasser- und Sumpfgewtchse, 3, 547-558; 4, 1-5. Jena. GRAEBNER, P. (1900). Sparganiaceae. Pflanzenreich, 2. GRONTVED, J. (1945). Spargania in Denmark. Bot. Tidsskr. 50, 215. Hacerup, O. (1941). Nordisk Kromosom-Tal. I. Bor. Tidsskr. 45, 385. HOLMBERG, (1922). Anteckningar till nya Skandinaviska Floran, Il Bor. Notis., 203-9. Hupson, W. (1778). Flora Anglica, 2, Ed. 2. London. HYLANDeER, N. (1945). Nomenklatorische und systematische Studien iiber nordische Gefisspflanzen. Stockholm. HyYLAnper, N. (1953). Nordisk Kérlvéxtflora. Stockholm. KIRCHNER, O. VON, Loew, E. & SCHROTER, C. (1908). Lebensgeschichte der Bliitenpflanzen Mitteleuropas. Stuttgart. LINNAEUS, C. (1737). Hortus Cliffortianus. Amsterdam. Watsonia 5 (1), 1961. 10 C.D. Ky COOK LINNAEUS, C. (1753). Species Plantarum. Stockholm. L’OseL, M. DE (1570). Historia Plantarum. Antwerp. LOHAMMER in LOvE A. & Léve, D. (1941). Chromosome numbers of Skandinavian plant species. Bot. Notis., 27. Love, A & LOveE, D. (1948). Chromosome Numbers of Northern Plant Species. Reykjavik. Love, A. & L6vE, D. (1956). Cytotaxonomical conspectus of Icelandic Flora. Acta Hort. Gotob., 20, 65. Morona, T. (1888). Studies in the Typhaceae. II, Sparganium. Bull. Torrey Bot. Cl.,15. 73-81. MUENSCHER, W. C. (1944). Aquatic Plants of the U.S.A. Ithaca. STEINBAUER, G. P. & NEAL, (1948). U.S. Spargania. Papers Mich. Acad. Sci., 34, 33. Wu rr, H. D. (1938). Chromosomenstudien an der schleswigholsteinischen Angiospermen-Flora. Ber. deutsch. Bot. Ges., 56, 247-254. Watsonia 5 (1), 1961. GERMINATION, SEEDLINGS, AND THE FORMATION OF HAUSTORIA IN EUPHRASIA By Ae Jes NEO. University Botanic Garden, Cambridge PREVIOUS WORK ON GERMINATION All the investigators quoted below on the subject of germination have found that germination is independent of the presence of roots of other plants. Koch (1891) sowed seed of Euphrasia officinalis L. (an aggregate species; the species most likely to have been employed are E. stricta Host, and E. rostkoviana Hayne) in pots in the autumn, and obtained free germination the following spring. It is not stated whether the pots were in the open or not. Wettstein (1896) described some germination experiments in his monograph on Euphrasia. He sowed seed in troughs in the open (in some of which seed of other plants had been sown previously, in others of which the soil was root-free) and on damp blotting paper in the windows of a room at 4-10°C. The species sown in the open were : rostkoviana Mayne from three different localities hirtella Jord. pumila Kerner salisburgensis Funck minima Jacq. (two varieties). By Gy By By Only E. rostkoviana and E. hirtella were sown on blotting paper. Seed of all these species, collected in 1893 and sown on 10 October 1893, germinated in March 1894, germination indoors being simultaneous with, or up to eleven days earlier than, that outdoors. Seed of E. minima and E. rostkoviana, sown under the same conditions on 25 January 1894, germinated five to twelve days later than the corresponding October sowings. Seed of E. rostkoviana from 1893, sown outdoors on | April 1894, failed to germinate; sown indoors on blotting paper on 3 April 1894, 3 seeds out of 70 germinated on 15 April 1894. Seed of E. rostkoviana collected in 1892 and sown on 10 October 1893 failed to germinate. The highest percentage germination recorded by Wettstein was 70 : 42 out of 60 seeds of E. rostkoviana sown in October and germinating the following year (Wettstein, 1897). Wettstein concluded that the time of germination was independent of the time of sowing; and that, if seed does not germinate during the spring following its formation, it loses its viability about the time when it would normally be germinating. He later stated, however, (1898) that the pots in which no germination had taken place were left only until the following October. However, Heinricher (1898a) demonstrated a different state of affairs in regard to germination time. Using E. stricta seed of the 1895 crop sown in pots in 1896 on 27 February, 28 March, 16 April, 22 May and 23 June, he showed that the two earliest sowings gave quite good germination in March and April respectively, the next two poor germination in May and June respectively, and the last gave no germination in 1896 but good germination at the beginning of March 1897. In addition, a thick sowing of 1894 seed on 27 February 1896 gave | seedling in March 1896, and a sowing of 1895 seed on 24 January 1897 gave 36 seedlings by 3 March 1897. He thus demonstrated a restriction of germination to the spring, and the retention of viability into the second spring after seed-formation, if germina- tion cannot take place in the first. It was not stated where the pots were kept. Other results for E. stricta (Heinricher, 1898a) give some precise times for germination; thus, of some seed sown in a pot in a warm house on 21 January 1897, 1 seedling germinated i Watsonia 5 (1), 1961. 12 P. FEEO 14 days later; sowings in pots on 27 February 1896 had started germinating 19 days later; of many seeds sown in a pot on 28 March 1896, 4 had germinated after 19 days, 20 after 35 days, and 36 after 48 days. Further results showed that seed of E. rostkoviana can, like that of E. stricta, germinate in the second spring after its formation. E. salisburgensis and E. minima, sown in open ground in late October or in November, had germinated by 15 April 1897. A pot of E. minima seed, sown at the same time and kept in the window of the Botanical Institute, Innsbruck, had produced 2 seedlings by 21 Feb- ruary, and 27 by 7 April 1897 (Heinricher, 1898b). In Heinricher’s experiments the time between sowing and germination was two to four weeks, unless the sowing took place in the autumn or early winter, in which case germination was delayed until February or March; germination was confined to the period February to June. Nichols (1934) included EF. americana Wettst. (= E. brevipila Burn. & Gremli*) in an investigation of the conditions required for germination of a number of American plants. He found that seeds of this species, collected from the coast of Maine, would not - germinate when kept in a greenhouse, but those exposed to winter conditions for 71 days germinated 14 to 21 days after being brought into the greenhouse. Neidhardt (1947) sowed seed of E. rostkoviana (collected in the years 1934 and 1936-38) in the winter of 1938-39, and in the spring of 1939. Samples of 100 seeds were sown on damp blotting paper in petri-dishes. They were subjected to various conditions of light and temperature. Indoor temperatures up to 30°C and outdoor temperatures ranging from — 5°C to +5°C were used. Varying alternations of high and low temperatures were employed also. Only the seed of the 1938 crop germinated, and about 20% germinated in all samples except in one that was kept at a constant temperature of 30°C, in which 10% germinated. A test using one seed per dish gave germination at the same time as the others. Germination was said to take five to seven days; this is very quick, but germination would be visible sooner in seeds on blotting paper than in seeds in the soil; however, it seems extraordinarily quick for the sample outdoors in temperatures ranging from — 5°C to + 5°C, and may in fact have referred to germination times obtained only at the higher of the temperatures used. Seed samples sown for cultivation in March, with 1, 2, 5, or 100 seeds per pot of soil, germinated in 15-20 days, with percentages up to 40. It will be noted that Neidhardt failed to get seed to germinate except in the season after it was formed. Some of my own observations on germination in Euphrasia now follow. Loss OF VIABILITY IN STORAGE Seed removed from herbarium specimens of Euphrasia collected in 1951 and sown in the winter of 1951-2 showed very poor germination, and IJ attributed this to desiccation in storage. Ten seeds of each of eight different samples were sown on damp filter-paper on 22 November and kept in the laboratory; only one seed germinated, on 25 February 1952. Some of the seeds were attacked by mould, but a number were still unaffected on 2 April. Samples of from 50 to 150 seeds, representing several species, were sown on 1 April. About half were kept in the greenhouse, and the rest outside. No germination could be detected on 22 April or 6 May 1952. Seeds of a sample of E. pseudokerneri Pugsl. were sown on three separate dates and their germination gives evidence of declining via- bility : seed was sown in each of ten pots on 6 November and germination took place in January, February and March, 59 seedlings appearing altogether; 90 seeds were sown on 19 February in sand, and placed under fluorescent lighting in a laboratory; one seedling appeared on 20 March and no more had germinated by 22 April; a third sowing, of 81 *P, D. Sell and P. F. Yeo, unpublished. Watsonia 5 (1), 1961. GERMINATION IN EUPHRASIA 13 seeds, on | April, had produced no seedlings by 6 May, the pot having been kept in a cold frame. All these seeds had been stored in paper envelopes in a very well heated laboratory. All seed collected in 1952 was extracted from the plants as soon as possible and kept in a refrigerator at about 8°C in plastic bags sealed with cellulose tape, and this resulted in much better germination. In later years storage in a refrigerator was not possible, but an effort was made to keep the seeds out of very warm places in the summer and well heated rooms in winter, and at least some seeds have germinated in the majority of samples sown. The approximate time of ripening is known for some seeds produced by artificial self- and cross-pollination; those ripening earlier spend longer in storage, and Table | shows a great decline in viability in storage in 1957 and a small one in 1958. The germination figures are the combined totals for the two years after seed production. TABLE | Germination of seeds collected at different times Percentage Proportion of samples Date collected No. sown germination showing germination 1957 | (sown 6 Dec. 1957) | Late July, early August 251 372 Z Out Oba Mainly September 82 15-0 4outof 4 Mainly October 208 26:9 | 6 out of 7 1958 (sown 16 Dec. 1958) Late August, September 550 | 9-3 9 out of 14 Mainly October 416 | 12-0 | 9 out of 9 In January 1955 two samples of 1952 seed and one of 1953 seed were sown; there was no germination in 1955. PERIOD OF GERMINATION The following history of the seeds of E. pseudokerneri already mentioned (collected at Holywell Mound, S. Lincs.) gives some indication of the germination behaviour. Group A (2 pots) | Group B (8 pots) 11 Oct. 1951: Seed collected Ii Oct 195i Scedicollected 6 Nov. 1951 : Seed sown in pots and 6 Nov. 1951: Seed sown in pots and placed in greenhouse placed in greenhouse 26 Nov. 1951 : Pots put out of doors 28 Jan. 1952: 1 seedling appeared , 1-14 Feb. 1952: 9 seedlings appeared 4-12 Feb. 1952: Pots brought into green- house 15-29 Feb. 1952: 32 seedlings appeared 1-14 Mar. 1952: 17 seedlings appeared After three weeks in the greenhouse it appeared that the seeds were unable to germinate, and eight of the pots were therefore put outside in the expectation that frost might break the dormancy. In fact, dormancy ended in the pots in the greenhouse twelve weeks after sowing. (The seeds were probably exposed to frosts before being collected on 11 October, as there were ground frosts of 29°F, 29°F, and 26°F on 9, 10 and 11 October respectively at Cottesmore, 4 miles to the south-west). I then guessed that the seeds outside would also have become germinable and were now being retarded only by the low temper- ature outside; I therefore brought them into the greenhouse, and my supposition was con- Watsonia 5 (1), 1961. 14 P. REYEO firmed, for seedlings began appearing after ten days, and the last had appeared at the end of 28 days. The same effect can be seen in the figures in Tables 2 and 4 (p. 16). TABLE 2 Germination in 1954 of seed of E. occidentalis Wettst., collected at Perranporth, 1953 (25 seeds were sown in each pot on 9 November 1953) No. of seedlings produced Pot indoors Pot outdoors 21 January l 2 11 February 2 2 15 February 6 | 2 (brought indoors) 18 February 14 2 22 February 19 | 3) 25 February 20 iL tt 1 March | 20 | 17 4 March 20 19 A great many samples of Euphrasia seeds have been sown in the course of my work on the genus, the majority in the November or December after they had ripened. In these, germination in the open has taken place from January to April, the peak of germination being in March at Leicester in 1953, and in March or April according to season at Cambridge in 1954-60. In my garden at Cambridge the first seedlings of E. pseudokerneri have appeared at the end of January in 1959 and 1960. In nature I have found seedlings of E. anglica Pugsl. in Charnwood Forest, Leicestershire, on 28 February, 1953, and of E. pseudokerneri at Box Hill, Surrey, in March, 1953. If sowing takes place later than December, germination is likely to start late and continue until a later date. Thus, some samples sown on 17 February 1954 began to germinate in late March or early April, and continued germinating until late April or, in one case, mid-May. The latest sowing to have produced seedlings the same year was made on 14 April 1955 (E. rostkoviana from Friuli, Italy); two seedlings came up in the latter half of May. An exceptional case was that of a sample of seed of E. stricta (from Poland), sown on 15 March 1956. The beginning of germination was overlooked, and on 26 September the pot contained, in addition to a plant of Trifolium repens L. and three plants of Euphrasia with ripe seed, three apparently newly-germinated seedlings of Euphrasia. On 28 May 1952, 63 seeds of FE. pseudokerneri, collected in the greenhouse in the preceding fortnight (an unusual time of year for seed-setting), were sown and kept in the greenhouse for the rest of the summer, but there was no germination. EXPOSURE TO WINTER COLD AFTER SOWING Of the samples of wild-collected seed sown in December 1952, the total number sown inside was 23, of which germination occurred in 12; a total of 52 samples was sown outside and germination occurred in 47. Some counted samples of seeds produced in 1952 by enforced self-pollination were sown in pots on 20 December 1952 and kept in the heated greenhouse. In one sample (of E. occidentalis) germination took place the following spring. In the rest there was no germina- tion, but the pots were kept in the open from May 1953 until February 1954, when they were brought into a cool greenhouse, and the layer of moss on the surface of the pots was broken up. The germination is shown in Table 3. Seeds produced by artificial hybrid- isation in 1952 were treated in the same way; these seeds produced four seedlings in 1953 and about 68 in 1954. In one pot, two seedlings appeared in 1953 and three more in 1954. Watsonia 5 (1), 1961. GERMINATION IN EUPHRASIA Ils) TABLE 3 Germination of 1952 seed in 1954 Euphrasia species | No. of seeds sown No. of seeds germinating | | nemorosa | 40 | 5 pseudokerneri | 35 28 pseudokerneri | 13 I anglica 5) 6 It appears from these observations that it is better, for practical purposes, to put seed out of doors in the winter for germination. I have, however, recorded a few instances of good germination of seeds kept in a heated greenhouse from the time of sowing. Thus the sample of seed produced by self-pollination from E. occidentalis from Barton-on-Sea, S. Hants, v.c. 11, treated in the same way as the samples in Table 3, yielded 15 seedlings from 47 seeds in 1953 (the pot was not kept until 1954); and a large sample of seed of an atypical form of E. nemorosa (Pers.) H. Mart. from Roborough Down, S. Devon, v.c. 3, collected in 1952, germinated profusely, without exposure to winter cold, in February and March 1953. GERMINATION IN SECOND AND LATER SPRINGS AFTER SOWING In the preceding section examples were given in which germination was completely postponed until the second spring after the seed was sown; in addition one example was mentioned in which germination took place in both the first and the second spring after sowing. Both kinds of behaviour are recorded in Table 4 (p. 16). Seed pots of all the sowings made in the winter of 1957-1958 were kept and three samples showed total post- ponement of germination until 1959; they were sown respectively in December 1957 (only one seed germinated), February 1958, and March 1958. Germination in both first and second springs after sowing has occurred in a number of samples sown late (January to April), but it also occurs in early-sown samples, for six of the eight samples sown on 6 December 1957 produced seedlings in 1958 and 1959, and in 1960 germination has taken place in six samples which produced seedlings in 1959 after sowing in December 1958. I have also recorded germination in the third and fourth springs after sowing. Thus a large sample of seed of E. nemorosa was sown on 17 February 1954, and probably about 90 seeds germinated the same year, about 200 in 1955, and one in 1956. A sample of seed of E. micrantha Reichb., from a population that had. been maintained in cultivation since 1953, was sown on 14 December 1956 and produced 28 seedlings in 1957, 9 in 1958, 14 in 1959 and 2 in 1960. A sample of E. nemorosa, sown 6 December 1957, gave 29 seedlings in 1958, 5 in 1959, and 2 in 1960, and a sample of E. brevipila, sown 19 March 1958, ger- minated first in 1959 to produce 21 seedlings, and had produced 2 more by 17 March 1960. A GERMINATION TRIAL This trial was intended to compare the germination of seed samples sown at different _ dates, and of samples kept in the greenhouse from the time of sowing, or put in cold frames for varying periods before being brought into the greenhouse, or kept throughout in cold frames. Owing to a failure of heating in the greenhouse, all samples were exposed to freezing temperatures for a few days, and so no comparison was possible between samples exposed to cold and samples never chilled. Some of the pots were kept for a second year, having been kept out of doors during the second winter. The results for E. nemorosa are shown in Table 4. This table shows the occurrence of high percentages of germination, an apparent decline in viability in storage, an apparent decline in viability with time spent out of doors in the first winter (for which I cannot suggest an explanation), a delay in germination in pots still out of doors when germination is be- ginning inside, germination in two successive years, and postponement of germination in Watsonia 5 (1), 1961. 16 P. “Rap xaeO TABLE 4 Seed germination of Euphrasia nemorosa from Friday Street, Surrey, 26 Sept. 1953 (25 seeds per pot) Date sown 9 November 1953 | 8 December 1953 | 13 January 1954 1s rebruaryoen Pot number 1 Sans) Anes | 1 2 3 4 1 | 2. | 3 | 4 | 1 | 2, | 3 PS tenn ie oy Rees Weareallaene | ete ae Time outside (weeks)| 0} 43 | 9 | 153) All | Oo} 59s) All 10 | 44 All | All | 0 | 24 All No. germinating : March 1954 ieee 2 Wie | 6) | April 1954 1 7 | 5 he April 1955 Sli) tie eS A (ee hlre | | o| 9 Total 22 | eal [a ew 7\ -7|.2 a | ne | | the later sowings. A similar set of sowings was made with E. occidentalis from Perranporth on the coast of Cornwall. It was thought possible that, although E. nemorosa would probably require frost treatment before germinating, E. occidentalis from an almost frost- free locality would need little or no cold-treatment. This question could not be tested owing to the freezing of the greenhouse. The results, some of which are shown in Table 2, were similar to those given by E. nemorosa. The following differences were found : the three pots of the November and December sowings that were kept a second year produced only one seedling and I was in doubt whether even that was a Euphrasia; and the main period of germination was about three weeks earlier, both in the greenhouse and out of doors while a few seedlings came up a month earlier still, in a mild spell in mid-January. VARIATION IN GERMINATION I have sown few counted samples of seeds apart from those produced in cross-pollina- tion or enforced self-pollination. The variation in germination seen in Table 3 seems to be typical of these and also of wild seed, although some variation may be due to different periods of storage and different times of collecting. However, there is a possibility that the testas of Euphrasia seeds are easily damaged and that the embryos are killed when the testas are damaged, with reduction in percentage viability as a consequence. Seeds of controlled pollination were usually collected with a curved needle after moistening, and I found that the ridges on the testa were easily broken in this process. The same needle has also been used for extracting seeds from herbarium material. A new trial of germination has now been started, using counted seeds taken from mature specimens by shaking, so as to avoid any unnatural damage to the seeds. The seed pots will be kept until the first spring when no seedlings come up. DISCUSSION OF GERMINATION BEHAVIOUR My observations which indicate failure of seed to withstand dry storage conditions are consistent with the results of Wettstein and Neidhardt; Heinricher, however, obtained germination in seed that had been stored for more than a year. Loss of viability in storage was found by Vallance (1952) in Rhinanthus, and he overcame it by starting his germination experiments within a few weeks of the ripening of the seed.- The germination of Euphrasia seed is normally restricted to the spring, and the species seem to have a rigidly fixed life-cycle, which is partly governed by the restricted germination period. All the Eurasian and North American species are annuals, flowering in the summer and dying in the autumn. A seasonally fixed life-cycle is probably the rule in annuals. It seems clear that one of the factors concerned in controlling germination is the need. for exposure to winter cold. However, there must be others, because late sowing, with its reduced exposure to the winter, sometimes leads to postponed germination and sometimes Watsonia 5 (1), 1961. GERMINATION IN EUPHRASIA 7! does not, while some seeds do not in any case germinate in their first spring after sowing. Among the results quoted from previous workers on Euphrasia, only those of Nichols appear to provide positive evidence of better germination resulting from outdoor winter conditions, while the results of Neidhardt are complete enough to suggest positively an absence of any such effect, except for a reduction in germination at the very high temperature of 30°C. However, Heinricher (1909, p. 285) made observations on Rhinanthus which suggested that the duration and severity of winter cold had an effect on germination. The controlled experiments of Vallance (1952) on Rhinanthus showed that seeds kept moist at 20°C failed to germinate, even after a year of such treatment. Germination required a period of several weeks of chilling, after which it took place more rapidly at higher temperatures. It seems probable that Euphrasia has similar requirements, though if so the rather frequent germination under greenhouse conditions would have to be accounted for. However, the maximum temperature which will give effective chilling has not been determined, and the question of whether chilling can be effective if it shows a diurnal rhythm has not been examined. There is a possibility that night frosts in late summer might have provided a sufficient chilling effect in some seed samples that have been used in experiments. It is noteworthy that Vallance obtained germination throughout the year, though no seeds germinated less than two months after ripening. It appears from my results that a spread of germination over a period of years is normal, though the proportion of second-spring germination is doubtless increased by late sowing. Previous workers who obtained germination over two or three years only did so in rather late-sown samples. However, in 1909 Heinricher (1909, p. 284) appeared to regard a spread of germination over two years as normal in Euphrasia, Odontites and Rhinanthus, and mentioned an occurrence of third-spring germination in Rhinanthus. The spread of germination over a number of years is probably due to hereditary differences in the seeds, maintained by selection. This spread of germination is clearly of advantage in the event of an unfavourable season, such as a very dry summer in the lowlands, or an unusually cool and short one in the mountains or in the Arctic. THE SEEDLINGS OF EUPHRASIA AND THE FORMATION OF HAUSTORIA Parasitism in the Rhinanthoideae was discovered by Decaisne (1847). He found that the roots of Alectorolophus (Rhinanthus), Melampyrum and Odontites had haustoria by which they were attached to the roots of ‘“‘ Gramineae, of shrubs and even of trees.’ (Melampyrum is a parasite of shrubs and trees). His observations were supported by those of Henslow (1847) who found Euphrasia officinalis* and E. odontites (Odontites verna (Bell.) Dum.) bearing haustoria by which their roots were attached to the roots of Gramineae. Henslow also found isolated Euphrasia plants, which he supposed had destroyed their hosts. Koch (1891) found that in Euphrasia growing parasitically on other species the haustoria only seized living roots, and they only parasitized the youngest or thinnest host roots. In Melampyrum, as opposed to Euphrasia, the haustoria attach themselves to dead organic matter. Saprophytism occurs in Euphrasia and Rhinanthus only after the end of the parasitic activity of the haustoria, when the affected host roots have died. Haustoria were observed also by Wettstein (1896), being numerous on plants of E. rostkoviana Hayne grown in - troughs in the open, in which grass seed had been sown the previous year and in which they grew well and flowered. Heinricher (1898b) grew E. salisburgensis Funck, E. rost- koviana and E. minima Jord. out of doors, and found well-grown Euphrasia plants with haustorial attachments to a variety of other plants. Wettstein reported that the identification of host plants in the field was difficult because the Euphrasia roots are very thin and because when the Euphrasias are flowering, the host roots are dying, and nourishment continues saprophytically. However, he recorded the formation of haustoria by several species of Euphrasia on a variety of host-plants in the wild (1896, 1897), as also did Crosby-Browne (1950) for E. salisburgensis. Parasitism of Euphrasia plants by others of their kind was observed by Koch (1891). Euphrasia seeds were germinated in pots and where they were very crowded some individuals *Formerly all British Euphrasiae were referred to E. officinalis L., which is now treated as a nomen ambiguum. Watsonia 5 (1), 1961. 18 P. FPIYEO erew faster than the rest and were found to be attached to their neighbours by haustoria. Wettstein (1897) also found Euphrasia plants attached to one another by haustoria in pots of E. rostkoviana which had been sown with no host-plant, although in one a weed was being parasitised by the Euphrasia. Heinricher (1898a) reported that thickly-sown seedlings of Odontites odontites (O. verna) with no host-plant had haustorial initials at the contact points of their root systems on 29 April, germination having started on 12 April. The genus Odontites is closely related to Euphrasia. The anatomy of the root of Euphrasia has been described by Neidhardt (1947). Wettstein (1896) gave anatomical drawings of haustoria, and described the root system as consisting of extremely fine roots, with few root hairs, which occurred singly near root tips and on haustoria, and densely in a ring at the base of the hypocotyl, where they play a part in the initial anchorage of the seedling. He reported that his observations on the anatomy of the haustoria of E. rostkoviana agreed fully with those of Koch on “E. officin- alis.”’ Koch (1891) investigated in detail the anatomy of the haustorium. Characteristic of Euphrasia are the small size and slight growth in length of the haustoria, the simple tracheal strand (formed from one row of cells), and the clasping structures round the edge of the contact surface of the haustoria. Root hairs are more or less isolated. The attacked host root becomes internally disorganized, and the attack becomes saprophytic. About this stage, the haustorium becomes filled with food reserves and its tissue becomes loose, with intercellular spaces. The haustorium penetrates the cortex and comes into contact with the vascular tissue of the host root. Neidhardt (1947) stated that the roots of seedlings growing alone penetrate deeply into the soil, while seedlings growing with host-plants have comparatively short taproots which bear abundant branches. He also recorded that, whereas lateral roots arise endogenously, growth of haustoria begins in the outermost layer of the cortex. I have not consulted the anatomical works by Hovelacque and by Spoerri, nor the physiological work by Kostytschew, that are cited by Neidhardt. ‘ An attempt was made by Fraysse to find out what materials were obtained from the host by the parasite (Fraysse, 1906). Fraysse (/.c., p. 99) observed haustoria of E. officinalis in fields in Savoie in September attached to the roots of various Gramineae, Compositae and Leguminosae, and to the rhizoids of mosses, etc. He found autoparasitism (and parasitism) commonly in a relative of Euphrasia, Odontites rubra var. serotina (O. verna subsp. serotina (Dum.) E. F. Warb.) (Fraysse, 1906 p. 89). Regarding penetration by haustoria, Fraysse found that in Trifolium and Taraxacum they usually penetrated the cortex but reached the host xylem only exceptionally (p. 99), the parasite in this case being E. officinalis. In Monocotyledons the haustoria always reached the centre of the root, and sometimes divided the entire stele. Fraysse tested the haustoria and adjacent regions of host and parasite for starch (p. 92). In this investigation the parasite was Odontites. The roots were collected in August; some from plants which had not yet started to flower, and others from freely-flowering specimens. In the roots of Ranunculus repens and Trifolium repens, testing with iodine showed a disappearance of starch from the region around the haustoria. From the distribution of reducing sugar in the haustorium, as indicated by the Fehling reaction, Fraysse concluded that glucose was passing from host to parasite. The roots of the grasses collected were poor in glucose as well as in starch, and on them the haustoria of Odontites contained little carbohydrate; consequently, Fraysse concluded that they obtain from grasses only water and whatever solutes it contains. Haustoria of Odontites on Leontodon autumnalis behaved as on Ranunculus and Trifolium. | In the haustoria of Euphrasia Fraysse found the same physiological mechanism as in Odontites. He states that the haustoria were constituted only for obtaining carbohydrate, but mentions the presence of abundant protein granules in Euphrasia haustoria on Trifolium and Taraxacum (p. 99 seq.). My own observations support many of those of previous workers, and slightly extend them. The appearance of the seedlings of Euphrasia is shown in Fig. 1 (p.19). Photographs of Watsonia 5 (1), 1961. GERMINATION IN EUPHRASIA 19 sets of seedlings of seven species, including the seedlings shown in Fig. 1 b—d, were taken on 1 or 2 April 1953. The youngest seedlings are just free of the testa, or have the cotyledons still within it; the oldest show the first pair of leaves. The degree of root-branching in plants of apparently similar age appears to vary from population to population. At the base of the hypocotyl a tuft of root-hairs develops; they give the impression of being effective in anchorage, and are probably necessary to ensure that the cotyledons come above ground. Root initials Fig. 1. (a) Seedling of Euphrasia occidentalis, grown in the greenhouse, drawn on 3 February 1953, four weeks after the cotyledons had appeared above ground. (b)—(d) seedlings of E. confusa, drawn from photo- graph taken at beginning of April 1953 (Root hairs not shown, except for tuft at base of hypocotyl). All approx. x 2 If they were absent the seed and cotyledons would give anchorage, and the hypocotyl would probably never appear above ground. Root-hairs are sparse elsewhere on the root system of the seedling; they are usually absent from much of it, and occur in groups here and there (Fig.1 a). They also seem to develop around haustoria. It was noted that on five seedlings of E. occidentalis sand grains and organic debris were found firmly attached wherever root hairs were present. The root system often appears to branch more in pieces of peat in the potting compost than elsewhere. Lateral branches are swollen at their base, and the initial of a lateral branch looks similar to a detached haustorium (Fig. 1). No haustoria could be seen by direct observation in any of the sets of seedlings photographed; seedlings which were left for a much longer time in the pot in which they had germinated formed haustoria, however. Two small pots, one containing seedlings of E. micrantha and the other E. confusa Pugsl., were investigated on 20 June 1953. Neither pot contained any plants other than Euphrasia and in both Watsonia 5 (1), 1961. 20 P. FraeyEO haustorial connections between Euphrasia roots were seen, as well as haustoria which had become disconnected when unearthed. In the £. micrantha sample, roots connected by a haustorium were traced back to two different plants (Fig. 2.). In this sample a haustorium was also found connecting two branches of the same root system; another Fig. 2. Two pieces of Euphrasia root, each belonging to a different plant, one connected to the other by two haustoria. (E. micrantha. 20 June 1953). : connection of this kind was found on 13 August 1955 in the roots of an isolated plant of E. nemorosa growing in a pot with no host plant. Connections between Euphrasia roots were found in 1954 in two cultivated samples of E. nemorosa in which the plants also had haustorial connections with Medicago lupulina L., and in 1955 when the Euphrasia was also parasitic on Hieracium pilosella L. Haustoria attached to cultivated M. /upulina were also seen in 1954 on E. anglica Pugsl., several hybrids and, in abundance, on E£. salisburgensis (from Austria) and E. pseudokerneri Pugsl. Haustoria on host roots were found also on E. micrantha. The intended host was M. lupulina, but young plants of Ranunculus sp. and Lamium Purpureum were also in the pot at the time of examination. Haustoria were found connecting E. nemorosa to Hieracium pilosella and E. scottica Wettst. to M. lupulina (in 1955 and 1956 respectively). These observations were made when the plants were well-grown and were about to be pressed. The soil ball was knocked out of the pot and the haustoria were seen in the mass of roots at its periphery. Two poorly-developed plants of E. pseudokerneri examined in May 1952 were found to have haustoria. One had flowered and fruited precociously in a pot with Pelargonium x hortorum Bailey; when dug up the Euphrasia had detached haustoria and one haustorium attached to a dead piece of Pelargonium root. The other Euphrasia was dug up when nearly dead; it was growing in the presence of Plantago lanceolata and Prunella vulgaris and it had a few haustoria where its root was closely applied to a dead root of one of these plants. Very vigorous plants of E. nemorosa were grown in the garden in embedded ‘ whale- hide ’ pots (which rot fairly quickly in the soil) with Plantago lanceolata as host. The roots of two host plants, together with their parasites, were pulled up and preserved in alcohol. The roots of each host, with its parasite, formed a dense mass in which there were numerous haustoria, but in which it was difficult to make out connections. However, each host was seen to be connected to its parasite by haustoria. In both cases haustoria were seen on large Plantago roots six or seven times as thick as the Euphrasia roots bearing the haustoria, and in one case also on relatively thin host roots. I have not been able to trace the roots parasitised by wild Euphrasia plants back to their origins, except for those of a plant of Trifolium repens parasitised by E. nemorosa This was found by Mr. E. K. Horwood growing isolated in cinders which were very easily washed out of the roots. I have been able to find haustoria attached to the roots of unknown host species in young E. occidentalis and E. pseudokerneri dug up in early May and having four pairs of leaves visible. Some of the haustoria of these plants and of the plants cultivated with Plantago lanceolata in the garden appeared to be terminal instead of lateral as they usually are. Euphrasia seedlings which are not established on host-plants show very slow growth Watsonia 5 (1), 1961. GERMINATION IN EUPHRASIA ah above ground, and the leaves they produce are small. However, the root system grows quite actively, and produces an appreciable length of very fine roots. For cultivation it is usually necessary to transplant the seedlings from the pots in which they germinate into pots provided with host plants. The earlier the seedlings are planted out the less they are damaged in the process, and it is therefore necessary to plant them out when they are extremely small. In practice it seems better not to plant them out as soon as they are visible above ground, but to do so two to four weeks after they appear. For example, suitable seedlings for planting out would be those shown in Fig. | (a), (c) and (d), while the plant in Fig. | (b) would have too little root to make it desirable to transplant it. DISCUSSION OF HAUSTORIA Some small differences between my observations and those of previous workers may be noted. Haustorial attachments to live host roots can be found, and are in fact the rule in my experience, in Euphrasia plants that are well into their flowering period, whereas Wettstein stated that the host roots were dying when the Euphrasia plants were in flower. It seems possible that the saprophytic phase of the haustorium described by other workers exists alongside the parasitic phase, and that new haustoria continue to be formed until the Euphrasia plant begins to die. Kinds of haustorial attachments that have apparently not been recorded before are those between two Euphrasia roots formed even when a host- plant is present, and those between two roots belonging to the same Euphrasia plant. It seems, therefore, that Euphrasia roots have no immunity to attack by their own kind, and that the necessary stimulus to haustorium formation is merely contact with another root. This is consistent with the wide host range which Euphrasia species show (see Crosby- Browne, 1950). The occurrence of terminal haustoria does not seem to have been reported before. SUMMARY 1. Previous work is summarised. 2. Seed of Euphrasia may lose its viability in a few months in dry storage conditions, and has rarely been germinated when stored for a year and never when stored for two years or more. 3. Germination is normally confined to the months January to May. 4. Some samples give good germination without exposure to winter cold. Others fail to germinate, or germinate less well, without it. 5. Germination may be entirely or partially delayed until the second spring after seed- sowing; complete postponement is more common with late sowing (January to April). Partial postponement is probably the rule, and germination may occur in the third and even the fourth spring. 6. Germination behaviour and the percentage germination have been found to be very variable. 7. Euphrasia roots are very slender and are almost devoid of root-hairs except at the base of the hypocotyl. Above-ground growth is slow before establishment on the host. Young seedlings can be transplanted. 8. Haustoria sometimes appear to be terminal on the Euphrasia root; usually, however they are lateral. 9. Haustoria may attack roots of Euphrasia belonging to another plant or even to their own plant. 10. Various species of Euphrasia were found to have haustoria attached to the roots of a number of host-plants. ACKNOWLEDGMENTS I am indebted to Professor T. G. Tutin and the University of Leicester for valuable help at the commencement of this work and to the Meteorological Office for supplying information about frosts at Cottesmore. Watsonia 5 (1), 1961. 2? P. FO WEO REFERENCES CrosBy-Browng, A. J. (1950). The root parasitism of Euphrasia salisburgensis Funck. Watsonia, 1, 354-355. DEcAISNE, M. J. (1847). Sur le parasitisme des Rhinanthacées. Ann. Sci. Nat., Sér. 3, 8, 5-9. Fraysse, A. (1906). Contribution a la biologie des plantes phanérogames parasites. Montpellier. HEINRICHER, E. (1898a). Die Griinen Halbschmarotzer, I. Jahrb. Wiss. Bot., 31, 77-124. HEINRICHER, E. (1898b). Die Griinen Halbschmarotzer, II. Jarhb. Wiss. Bot., 32, 389-452. HEINRICHER, E. (1909). Die Griinen Halbschmarotzer, V. Jahrb. Wiss. Bot., 46, 273-376. HensLow, J. S. (1847). Parasites. Gard. Chron. for 1847, 605. Kocu, L. (1891). Zur Entwicklungsgeschichte der Rhinantheen (II, Euphrasia officinalis L.). Jahrb. Wiss. Bot., 22, 1-34. NEIDHARDT, G. (1947). Euphrasia rostkoviana Hayne, Der Augentrost. Die Pharmazie, 3 Beih., 1 Ergan- zungsbd. NicHo.s, G. E. (1934). The Influence of Exposure to Winter Temperatures upon Seed Germination in Various Native American Plants. Ecology, 15, 364-373. VALLANCE, K. B. (1952). The Germination of the Seeds of Rhinanthus Crista-galli. Ann. Bot., new series, 16, 409-420. WETTSTEIN, R. VON (1896). Monographie der Gattung Euphrasia. Leipzig. WETTSTEIN, R. (1897). Zur Kenntniss der Ernahrungsverhaltnisse von Euphrasia Arten. Oesterr. Bot. Zeitschr., 47, 319-324. WETTSTEIN, R. (1898). Bemerkungen zur Abhandlung E. Heinricher’s “‘ Die Griinen Halbschmarotzer I, .. .” Jahrb. Wiss. Bot., 31, 197-206. Watsonia 5 (1), 1961. STUDIES ON WELSH ORCHIDS I. THE VARIATION OF DACTYLORCHIS PURPURELLA (T. & T. A. STEPH.) VERMEUL. IN NORTH WALES By R. H. ROBERTS ABSTRACT Among ten spatially isolated populations of D. purpurella in North Wales statistically significant differences have been found in most of the phenotypic characters. Highly significant differences exist between populations only short distances apart. The phenotypic variation follows a reticulate pattern and, as far as these results are concerned, no useful taxonomic subdivisions appear to be possible. Two groups of correlated characters have been found : (a) among three vegetative characters : the number of sheathing leaves, the number of non-sheathing leaves and leaf width ; (6) among three floral dimensions : labellum width, labellum length and spur thickness. INTRODUCTION The marsh orchid now known as Dactylorchis purpurella was first described by T. & T. A. Stephenson (1920) from plants found near Aberystwyth, Cardiganshire, v.c. 46. The variability of the species was apparent from the outset, for the Stephensons recognized two forms : dwarf plants with flowers of a vivid reddish-purple colour and a small, almost entire, irregularly diamond-shaped labellum (‘form A’); and taller plants with flowers of a rich, dark purple, and a rather larger, more rounded, shallowly trilobed labellum (‘form B’). It is clear that while they felt that “form A’ was quite distinct from other marsh orchids, they were not so confident about ‘ form B.’ In 1921 they visited the Isle of Arran and in several stations saw plants of ‘form B’ growing with what they accepted as * Orchis pulchella’ (= O. praetermissa var. pulchella Druce). The latter differed from their ‘form B’ only in having unspotted leaves and being somewhat larger. Without declaring the complete identity of these forms the Stephensons (1922) nevertheless realised their very close similarity and suggested that the var. pulchella would be better separated from O. praetermissa Druce. Eight years later T. Stephenson (1930) still referred to Orchis purpurella ‘ form B’ and the var. pulchella as distinct forms, although he emphasized the slightness of the differences which separated them. This variety was finally transferred to O. purpurella by H. W. Pugsley (1935), but it was clear that even as a variety it had little to distinguish it from the “type ”’ (Hall, 1937). Further varieties of O. purpurella subsequently described by T. Stephenson (1937) are var. maculosa, in which the leaves are spotted all over and not merely on either the apical or basal half as normally found; and var. crassifolia, in which the plants are much bigger, more leafy and have much larger labella. Variations in other directions were also evident : in the population on which the description of var. maculosa was based some of the plants had flowers of the dark crimson-purple typical of the species, but many had rose-coloured flowers; while, in the Orkneys, populations apparently indistinguishable in other respects from D. purpurella were found to contain a proportion of heavily leaf-spotted plants (1930). As Pugsley (1935) pointed out, there was little to separate these from D. majalis (Rchb.) Vermln. except their later flowering. In North Wales D. purpurella is frequently found in suitable habitats. Even within 23 Watsonia 5 (1), 1961. 24 R. H. ROBERTS this restricted geographical area it shows a wide range of variation both within populations and between one population and another. The populations vary from those composed almost entirely of small, distinctive plants with short spikes of small, purple flowers (ap- proaching Stephensons’ form A), through some in which the majority of individuals are larger, leafier and with broader, more rounded, obscurely trilobed labella (presumably Stephensons’ form B), to populations which, in their vegetative features, appear to approach the var. crassifolia Steph. Many of the Anglesey populations are very nearly and sometimes completely lacking in leaf-marking. As these are usually composed of rather large, leafy plants, they have occasionally been recorded as D. praetermissa (Druce) Vermeul., a species which has not been found so far in v.c. 52. In order to obtain some impression of the pattern of variation among these populations a biometric study of ten colonies was made in the flowering seasons in 1957, 1958 and 1959. The results are presented below. METHODS The morphological characters studied are those which have been used in previous work on the dactylorchids (Heslop-Harrison, 1948, et seqg.), namely : (a) stature; (6) number of leaves per plant; (c) leaf dimensions; (d) leaf marking; (e) labellum dimensions; (f) spur dimensions, and in addition, (g) the number of non-sheathing, bract-like leaves has been included. To avoid the unnecessary destruction which mass collecting would entail, measure- ments of all the vegetative characters were made in the field; flower colour, shape of spike and leaf marking were also noted for each plant in the sample. One homologous flower was collected from each plant and kept in a stoppered jar to prevent wilting. As soon as possible afterwards labella and spurs from these flowers were mounted by pasting on thin card. They were then placed under moderate pressure to dry and measurements made within a week to avoid any errors due to shrinking. THE POPULATIONS The localities of the populations shown in Fig. | are: Pl. Damp, base-rich meadow at Brithdir, near Dolgellau, Merioneth, v.c. 48. P2. Dune slack at Mochras, Merioneth, v.c. 48. P3. Marshy tract fringing the Artro estuary at Mochras, Merioneth. This colony is separated from P2 by over 365 yards, the intervening ground being occupied partly by a sandy ridge and partly by salt marsh. The colonies P2 and P3 have therefore been treated as distinct populations. P4. Dune slack separated from P3 by only 95 yards of sand dunes. This colony had many plants with unusually large spikes of paler flowers, with rather larger, more trilobed labella. Because of these differences it was treated as a separate population for biometric purposes. It is not suggested that it is isolated from P3 as a breeding unit. PS. Damp meadow land along the River Cegin, near Bangor, Caernarvonshire, v.c. 49. Dactylorchis fuchsii and several hybrids of this species with D. purpurella are the only other orchids here. P6. Cors Erddreiniog, 3 mls. N.N.E. of Llangefni, Anglesey, v.c. 52, is an area of fen where the most abundant marsh orchid is D. traunsteineri (Saut.) Vermeul. (Lacey & Roberts, 1958). D. purpurella occurs sparsely in the small tracts of damp grassland and sedge-meadow around the main fen area. P7. Marshy bottom of a disused limestone quarry, on the edge of Malltraeth Marsh, near Llangristiolus, Anglesey, v.c. 52. The orchids grow among grasses and sedges, with a few plants of the hybrid D. fuchsii I CO CO CO mH ok -_ NNwWWNNN Colony 66 (C. sepium x C. silvatica) w sw 22 4 22 4 24 5) 19 3 23 4 25 5 Mean 4 wie lossy 1-69 1-85 1-73 1-92 2:08 im) TABLE 3. MM mie 1-40 1:30 1-40 1-50 1:30 1-44 mie 0-50 0-50 0-73 0-69 0-83 0-61 mie 1-54 1:38 1:46 jicshil 1-43 1:36 mle 1-3 1-42 1:36 1-46 1-42 1:70 mie 0:70 0-70 1-00 0-80 0-78 0:62 mie 1-00 1-08 1-08 1-18 1-08 1-08 IN CALYSTEGIA Data compiled from 6 of the 70 small colonies. i) DADADUNUDAUNS i=) SS Li) NNNYANH LY ie) =x i) APRA HA HL RS oS Es es i GOMES Gsats WwWAWWAUNG 95 Ts 15 Ts 24 1855 ZT Ts 10 Ts 17 Watsonia 5 (2), 1961. 96 CLIVE A. STACE first described. The Ts values are plotted in Fig. 6 : here each individual value represents a colony rather than a flower. As might well be expected, the graph in Fig. 6 shows con- siderable similarity to that in Fig. 5. Two main peaks, representing the two species, are present, together with numerous intermediates of varying character. It seems likely that the intermediates shown in Figs. 5 and 6 are hybrids, C. sepium x C. silvatica, although stronger evidence for this assumption will be brought forward below. 3. POLLINATION AND GERMINATION STUDIES Having obtained an accurate method of separating the two species and the putative hybrid, I carried out a series of pollination experiments on these three taxa. Advantage was taken of the regular flower opening sequence in Calystegia, a flower being in bud one day, in flower the next, and withered the next. Large buds, ready to open the following day, were enclosed in pollen-proof bags made of grease-proof paper (polythene and cellophane being rejected to alleviate condensation). Later a quicker method was employed, whereby the tip of the corolla was tied with knitting wool to prevent it from opening, the corolla itself acting as the pollen-proof bag. With intended cross-pollinations, as opposed to self- pollinations, emasculation was usually effected just prior to enclosure in the pollen-proof bag. This entailed baring the inner parts of the flower and excising the anthers, the filaments being left intact to prevent unnecessary damage. Emasculation was not always carried out, however, as Calystegia was known to be self-incompatible, both from the literature and from previous personal experience. However, where it was not effected, self-pollination was prevented wherever possible by supporting the flower in its bag with the apex uppermost (the stigma being distal to the anthers). Anther dehiscence occurs almost as soon as corolla opening, and is neither introrse nor extrorse, but lateral (although Warnstorf (Knuth, 1909) says it is extrorse). When the flowers had opened on the second day pollina- tion was effected. Pollen from the chosen source was smothered on to the inside and outside of the two stigma lobes with the aid of a pencil point, after which the bag or wool was replaced. Each colony used as a pollen source was designated a colour, and a ribbon of that colour was tied on to the pedicel of all flowers pollinated from that source. Wherever possible reciprocal pollinations were effected, and pollen from as many sources as possible was used for the various flowers of each colony used as a female parent. With three taxa, there are twelve possibilities of pollination: three selfings, three intraspecific crosses and six interspecific crosses. In addition self-pollinations were effected by pollination from one flower on to another known certainly to be on the same plant. All fifteen of these possibilities were carried out, and the results are briefly summarised below. Selfings Attempts Successes 1. sepium flowers selfed 26 0 2. sepium flowers of the same plant crossed 5) 0 3. Hybrid flowers selfed 9 0 4. Hybrid flowers of the same plant crossed 10 1 5. silvatica flowers selfed 12 0 6. silvatica flowers of the same plant crossed 8 0 Intraspecific Crosses 7. sepium onto sepium 28 20 8. Hybrid onto hybrid 12 8 9. silvatica onto silvatica 14 13 Interspecific Crosses (Hybridisations) 10. Hybrid onto sepium 10 6 11. silvatica onto sepium 12 8 12. sepium onto hybrid 8 4 13. silvatica onto hybrid 8 3 14. sepium onto silvatica 21 11 15. Hybrid onto silvatica 8 2 Watsonia 5 (2), 1961. COMPATIBILITY AND HYBRIDISATION IN CALYSTEGIA 97 As may be seen from the above results, all three taxa are self-incompatible. Although one of the seventy self-pollinations surprisingly proved fertile, it is possible that this was the result of contamination by compatible pollen, especially as it involved transference of pollen from one flower to another rather than from the anthers to stigma of the same flower. On the other hand it could indicate that Calystegia is incompletely self-incompatible. All intraspecific and interspecific pollinations, in all nine of the possible combinations, proved successful to some extent, though very variously so. Although figures for inter- specific pollinations are mostly the lower of the two sets, little significance should be attached to this as in many cases different colonies were involved, and the figures represent the sums of three seasons’ work. The figures show that the two species are highly interfertile and that the hybrid is fertile in crosses with other hybrids and both parents, on both male and female sides. These results are considered extremely strong evidence that the intermediates between the two species, C. sepium and C. silvatica, are in fact hybrids. The biological nature of some populations was later investigated, the results fully confirming those of Walters & Martin (1958). Small colonies, isomorphic to the eye, proved to be single clones since all combinations of pollinations proved unsuccessful. Larger colonies, nearly always polymorphic to the eye to some degree, were composed of two or more clones, since cross-pollination in certain directions produced good fertile seed. Some fairly large (isomorphic) colonies, however, proved to be single clones. Seed-set in Calystegia has been observed to be good in the area of study during the years 1958-60. Populations of either species and of all types of intermediates commonly bear very good quantities of seed. As Baker (1957) points out, small populations of a single clone are less likely to produce seed than are large multiclonal populations, and Baker observed that in the London area the small colonies were frequently sterile or sparsely fertile, whilst larger colonies not far distant had good seed-set. For some reason this situation appears not to exist in the present area of study, since almost all colonies produced good seed-set. In some of the larger colonies, which superficially appeared to have a better seed-set than the smaller colonies, exact counts showed that the total number of seeds produced, expressed as a percentage of the total number of seeds possible, was no larger than on many of the small uniclonal plants, which in extreme cases merely consisted of a single branch with no more than ten flowers. There are ideally and at the most four seeds per capsule in Calystegia. Hegi (1927) states ‘ seeds 4, rarely 3,’ but this certainly appears to be an opti- mistic report. According to the number of ovules fertilized there may be any number of seeds from nil to four, capsules with one or two seeds being quite frequent, but those with nil, three or four more so. The number of seeds per capsule did not vary from species to species, nor from hybrid to species, and as far as could be ascertained the hybrid and its back-crosses appeared to be equally as fertile as the two parental species. I have found one capsule of C. sepium with five good seeds. The unsuccessful self-pollinations soon made themselves quite evident as the whole flower structure, bracteoles and distal end of the pedicel had turned brown and had shrivelled by early in September. Observation on the pollen tubes of pollen on the stigma of the same flower was therefore made. Flowers were self-pollinated in the usual way, and the style and stigma removed twelve, twenty-four, thirty-six and forty-eight hours after pollination. The styles were boiled in cotton blue in lactophenol for about three minutes, when the dye stained all parts and the lactophenol entered and softened the tissues. The styles, then a tangled, limp mass, were removed and washed in lactophenol and each was squashed flat on a slide under a coverslip and ‘ tapped out.’ Under the microscope the pollen grains and tubes appeared a dark intense blue, and the stylar and stigmatic tissue a pale blue. In all cases the pollen tubes, if present, were short and contorted, and had never penetrated more than a fractional distance into the stigma. In cases where compatible pollen had been deposited on the stigma, pollen tubes were seen to be long and straight, entering the stylar tissue. Quite obviously, assuming incompatibility to be of the normal type (Lewis, 1954), the inhibitory reaction, or lack of stimulatory reaction, (as the case may be) occurs in the stigma. Excision of the stigma did not allow self-pollination to result in ferti- lization, however, but the results of this experiment must be regarded as inconclusive as Watsonia 5 (2), 1961. 98 CLIVE A: STACE yet. It is now generally considered that stigmatic inhibition indicates a sporophytic in- compatibility system. Peters (Hegi, 1927) found that the pollen of C. sepium is very sensitive to damp, and that in foggy conditions 50% of the grains burst. He could not get the grains to germinate away from the stigmata, glucose, sucrose and even stigma extracts not providing the essential medium. These observations were confirmed with both species on several occasions : strong glucose solutions caused plasmolysis; very weak solutions caused bursting (as did water); and intermediate solutions had no effect at all. On all plants examined the percentage of grains which appeared to be fertile (full of contents, spherical and staining black with iodine) was over 95%. Using seeds collected from natural pollinations, attempts were made to find the best time to germinate the seeds. Three types of seed were used : white soft seeds not yet really ripe, although in brown capsules; hard, dark brown, fully ripe seeds; and similar seeds chipped with a scalpel. After about three days all the white soft seeds had germinated; the hard unchipped seeds had not yet imbibed water; and the chipped seeds provided mixed results, some germinating freely, others not, but in all cases the hard testa in some way impeding the unfolding of the cotyledons. In Calystegia germination is epigeal, two thin green ovate cotyledons being raised on a fairly long hypocotyl. In all seeds examined an embryo was present, and if there was no germination it was prevented either by the hard testa or by attacks from mould. Exactly similar results were obtained with seeds produced by experimental intraspecific and interspecific pollinations. Thus the presence of ripe seed may be taken to indicate a successful pollination. In artificial pollinations rather more fruits had one or two seeds than was the case in natural pollinations. In natural conditions the seeds probably fall to the ground, taking all winter for their testa to be fully water-saturated and thus made pliable. Special adaptations for seed dis- persal appear to be more or less absent, since the capsules have no explosive device, and the four sutures are not lines of dehiscence. Irregular pole-to-pole cracks appear in the capsule and many seeds drop out by reason of shaking caused by air currents. However, many seeds remain trapped in their old capsules for at least a year, and ultimately they probably rot away with the capsule wall. However, in 1946 Toole & Brown reported that after 39 years over half of the seeds of C. sepium were still viable (Salisbury, 1961). 4. DISTRIBUTION AND DESCRIPTION OF THE HYBRID From the previous sections it seems highly probable that hybrids between Calystegia sepium and C. silvatica occur in this country, and these hybrids are fertile, forming F, hybrids as well as backcrosses with either parent. Twelve such colonies were found in the neighbourhood of Tunbridge Wells, Kent, v.c. 16. One British record only has apparently been published — namely ‘ in the vicinity of Cambridge,’ v.c. 29 (Walters & Martin, 1958), but it seems probable that the hybrid does or will occur wherever the two species are in close proximity. Dr. W. T. Stearn informs me that intermediates occur with both parents by the Thames at Kew (Surrey, v.c. 17). In the Kew herbarium specimens are present from : Alderney, v.c. S (Jackson, 1932); Surrey, v.c. 17 (Clarke, 1901; Summerhayes & Milne- Redhead, 1932; Turrill, 1956); Berkshire, v.c. 22 (Elliot, 1945); Huntingdonshire, v.c. 31 (Sandwith & Gilbert, 1958): E. Gloucestershire, v.c. 33 (Riddelsdell, 1934); Glamorganshire, v.c.41 (Wade, 1947); and Merionethshire, v.c. 48 (Milne-Redhead, 1948). Thus 1901 appears to be the first British record. The earliest material of the hybrid seen by the author (and which also pre-dates any published record) is a specimen labelled ‘Anglia, Herb. Forsyth ’ in the handwriting of Hooker (senior) at the Kew herbarium. It is on the same sheet as a specimen of C. sepium collected in France by Hooker in 1819, and it is also labelled C. sepium. The hybrid is most probably of the same period as the other specimen, and of garden origin. In order to obtain an idea of the foreign distribution of the hybrid, the ranges of the two parents are clearly of some significance. C. sepium, in one form or another, appears Watsonia 5 (2), 1961. COMPATIBILITY AND HYBRIDISATION IN CALYSTEGIA 99 to be pan-North-Temperate, so hybrid distribution will be limited only by that of C. silvatica. C. silvatica is a native of south Europe from S. Spain and Portugal (?) eastwards to the Caspian Sea. Although neither Lousley (1948) nor Tutin (1952) mention north Africa in its area of distribution, it is obviously native there. It is mentioned in all the thorough floras of north-west Africa (Battandier, 1890; Sauvage & Vindt, 1954; Murbeck, 1905; and others), and specimens from Morocco and Algeria are present in the British Museum dating from the first half of the nineteenth century. Its range in this continent appears to be throughout most of Morocco (including Tangier), Algeria and Tunisia. It is also found in Malta (Borg, 1927), but it has not been recorded from the Balearics. These older literature records, however, should probably not be used without further confirmation. C. silvatica is now naturalised in various more northerly areas of Europe (Scholz, 1960), as it is in Britain. A specimen in the British Museum herbarium from Portugal (1954) may be a native or naturalised example. In investigating the nomenclature of the C. sepium-silvatica complex a variety of names is encountered. Although several authors have attempted to split off species from C. silvatica as here understood, this does not seem to be possible. The two usual synonyms are C. sylvestris and C. inflata. Pomel (1876) described a species which he named C. physoides from North Africa. Pau (1924) described a variety of C. sepium (var. tangerina) which is clearly a synonym of C. physoides, and was treated as such by subsequent authors (e.g. Sauvage & Vindt, 1954). Moreover, examination of Pomel’s description of C. physoides brings one to the conclusion that this taxon cannot be separated from C. silvatica, at least specifically. This conclusion was also reached by Battandier (1890), who gave C. physoides as a synonym of C. silvatica Griseb. The present author, then, considers that there is but one species in the C. si/vatica group, which, as Dandy (1958) concluded, is to be referred to as C. silvatica (Kit.) Griseb. Even more names are to be associated with the C. sepium group, although most of these have been treated as varieties rather than as separate species. Pomel (1876), however, described C. obtusa from N.W. Africa, which should probably be included with C. sepium. If all the names involved in the C. sepium-silvatica complex which are referable to either C. sepium (L.) R. Br. or to C. silvatica (Kit.) Griseb. are discarded, three names (involving two taxa) remain. The earliest of these is Convolvulus lucanus, named by Tenore (1826) from South Italy. Don (1837) transferred it to its present genus as Calystegia lucana, but Fiori (1926) reduced it to a variety of Convolvulus sepium, as he did C. silvatica. Choisy (1845) called it Calystegia sepium var. tubata, but it is clear that he refers to the same plant, as he gives the synonym, and the only specimen he cites is one sent to him by Tenore himself. Examination of the descriptions of Convolvulus lucanus given by Tenore (1826; 1824-29; and 1831, Sy//.), and of the excellent plate given in his Flora Napolitana, shows clearly that his plant is Calystegia sepium x silvatica. The plate depicts a plant exactly intermediate between the two parents, and this fact is actually mentioned in his third cited work. The third name is Calystegia barbara, described by Pomel (1874) from North Africa, and reduced to C. sepium var. barbara by Jahandiez & Maire (1934). Pomel’s description is fairly lengthy, and it appears that the plant is also a hybrid between C. sepium and C. silvatica, although perhaps closer to the former. Battandier (1890) reached the same conclusion, as he merely states under C. barbara ‘intermediate between the two preceding.’ The hybrid should thus be referred to as Calystegia x lucana (Tenore) G. Don. Thus, although Brummitt & Heywood (1960) say that they find little evidence of hybrids where C. silvatica is native (or elsewhere), it is not surprising to find that hybrids are localised both in the literature and as herbarium specimens, under a great variety of names. These localities are perhaps best mentioned geographically rather than chronologically :— NORTH AFRICA — the first record for this continent is that of Pomel (1874), who gives several localities in Algeria, which are repeated, with one addition, by Battandier (1890). A good specimen from Algeria (Gandoger, 1879) is also in the British Museum herbarium, labelled as C. sagittata. Sauvage & Vindt (1954) record the hybrid from Morocco, giving one locality only (Beni Mellal, Jahandiez, 1925), which was previously Watsonia 5 (2), 1961. 100 CLIVE A. STACE recorded by Jahandiez & Maire (1934) but without a date. A specimen with these exact details of collector, date and locality is in the British Museum herbarium, named C. sepium var. barbara. The specimen may be a hybrid between C. sepium and C. silvatica, but as previously implied it is much closer to C. sepium than to the other parent. Nomenclaturally however, it is still a synonym of C. x Jucana. Ball (1878) also recorded the hybrid from Morocco, stating ‘ specimina nostra intermedia sunt.’ A further specimen of C. x Jucana from Morocco (Trethewy, 1933) is in the Kew herbarium. RUSSIA —A single specimen from Lenkoran on the Caspian Sea in S.W. Russia is in the Kew herbarium (Hohenacker, 1838). This is probably the specimen Lousley (1948) cited as C. sylvestris when he gave the distribution as‘... . east to the Caspian Sea at Lenkoran.’ GREECE - A specimen from Greek Macedonia is in the Kew herbarium (Russell, 1918). ALBANIA-A specimen from Albania is in the British Museum herbarium (Baldacci, 1898). ITALY —- Tenore (1826) described the hybrid from Italy, the type localities being Auletta and Lauria in the region Lucania in south Italy. A number of other localities are given in his five cited works, all at about the same latitude as Rome or further south. A specimen is in the Kew herbarium which was sent to J. Gay by Tenore in 1830. It is labelled by Tenore, ‘ Macchia Matthei e Auletta.’ Which of the two localities it was collected from is not clear. Auletta (one of the type localities) is in south Italy, in Lucania, whilst Macchia Matthei is close to Rome. It is possibly some (hitherto unrecognised) sort of type specimen of Convolvulus lucanus. An additional possibility is that it is the holotype of Choisy’s C. sepium var. tubata, since Choisy (1845) said that the only specimen he had seen was one sent from Italy by Tenore. There is a second specimen at Kew sent by Tenore to Gay, which Tenore also labelled Convolvulus lucanus. This specimen is clearly C. silvatica, however, so it is possible that Tenore did not have a very clear idea of his * Convolvulus lucanus’. This specimen was sent to Gay in 1827 and is labelled ‘Nella Basilicata Nel Gargano etc.’ by Tenore. It is most unlikely that this was the specimen referred to by Choisy (1845), as he had a good idea of C. silvatica. Ball (1878) recorded C. Jucana from Italy as he states ‘ formae intermediae praesertim in Italia occurrunt.’ SPAIN — A specimen from El Cobre, Gibraltar (Wolley-Dod, 1912) labelled C. sepium is in the British Museum herbarium. This is the same locality as is given in Wolley-Dod (1914) under ‘ C. sepium var. sylvestris (?)’. Some of the descriptions of the hybrid taxon are fairly lengthy, and Tenore’s plate is most satisfactory, but measurements have never been included. A description of the hybrid is thus given here, from British material the author has seen in the field (see also Fig. 3). CALYSTEGIA X LUCANA (Tenore) G. Don (pro sp.) (= C. sepium (L.) R. Br. x C. silvatica (Kit.) Griseb.); Convolvulus lucanus Tenore, Fl. Neap. Prod. App., 5, 9, (1826); Convolvulus sepium var. lucanus Fiori, Nuoy. Fl. Anal. It., 2, 296 (1926); Calystegia lucana G. Don, Gen. Syst., 4, 296, (1837); Calystegia sepium var. tubata Choisy in DC., Prod. Syst. Nat. 9, 433, (1845); Calystegia barbara Pomel, Nouv. Mat. Fl. Atlant., 1, 83, (1874); Calvstegia sepium var. barbara Jahandiez & Maire, Cat. Pl. Maroc, 3, 591 (1934). A fertile hybrid between C. sepium and C. silvatica, capable of crossing with other similar plants and with either parent. It is intermediate between its two parents in all characters, and may be distinguished from them by : pedicels 30-100 mm. long; corolla white (always?), 41-62 mm. long; stamens 20-21 mm. long; style and stigma 20-23 mm. long: bracteoles broadly ovate, 14-25 mm. wide when flattened out, acute, obtuse or mucronate at apex, weakly cordate at base, slightly to strongly inflated, midrib very promi- nent especially at base, edges overlapping at each side and partially obscuring the calyx. Ratio of midrib-to-midrib (m) to edge-to-edge (e) measurement of bracteoles in natural condition 0-85-1-45. Watsonia 5 (2), 1961. COMPATIBILITY AND HYBRIDISATION IN CALYSTEGIA 101 Known from Britain, Morocco, Algeria, S.W. Russia, Greece, Albania, Italy and S. Spain. Since not all the previous literature giving details of measurements of C. sepium and C. silvatica flowers is satisfactory, this opportunity is taken of compiling a list of the most useful taxonomic measurements of the three taxa (Table 4). The figures given in this table represent the normal ranges found in the 72 colonies studied, only very abnormal measure- ments being omitted, and they have been found to hold good for all the British herbarium - material seen. Great caution should be exercised in applying these figures to foreign material, however. With regard to some characters, the literature proves somewhat controversial, due probably in part to vastly different areas and times of observation. For example, Stearn (1951) gives the stamen length of C. sepium as 15 mm., and of C. silvatica 25mm. In addition to the information given in Table 4, pollen grain sizes were also investigated. Hegi (1927) gives the diameter of C. sepium pollen as 76-84 », and Erdtman (1952) as 75-80 «. Diameters of grains were first measured in colony 71, when it was seen that they afforded good separation of the three taxa : the diameters encountered were 67-77 p, 77-86 » and 82-91 uw for C. sepium, C. x lucana and C. silvatica respectively. However, it was later found that this distinction is not constant, for one colony of C. sepium possessed grains of 83-88 pu. Literature regarding seed and capsule shape and size is especially diverse. The accounts of Pospichal (1899) and Hegi (1927) are very similar, the latter probably adapted from the former. Neither account refers to any substantial difference in either capsule or seed shape, and the only measurement given is ‘ about 5 mm.’ for the seed length in C. sepium TABLE 4. Characteristic measurements of the 3 taxa. C. sepium C. X lucana C. silvatica Length of corolla 36-55 mm. 41-62 mm. 58-82 mm. Length of pedicel 24-80 mm. 30-100 mm. 70-140 (—200) mm. Length of stamens 17-18 mm. 20-21 mm. 23-26 mm. Length of style + stigma 16-19 mm. 20-23 mm. 24-28 mm. Width of bracteoles 10-18 mm. 14-25 mm. 21-45 mm. Bracteole ratio, w/e 1:00-1:50 (—1:75) 1-40-2-40 (1:75—) 2:00—3°25 Bracteole ratio, m/e 0-40-1-10 0-85-1-45 1-15—2-20 Length of capsule 9-10 mm. — 8:5-10 mm. (mean = 9-15 mm.) (mean = 9:60 mm.) Width of capsule 8-5-10 mm. — 9-11 mm. (mean = 9:40 mm.) (mean = 9-85 mm.) Length of capsule beak 1-5—2:0 mm. ~- 2:5—4-0 mm. Length of seeds 4-5-5-0 mm. ~~ 4-5-—6:0 mm. (Hegi). Although a number of authors give minor differences between the fruits of the two species, none give the striking differences described by Tutin (1952). With regard to the capsule Tutin states that in C. sepium it is subglobose and 7-8 mm. long, whilst in C. silvatica it is ovoid, acute, and about 12 mm. long. Furthermore Tutin states that in C. sepium the seeds are 4—5 mm. long and more or less round but wrinkled, whilst in C. si/vatica they are 6-7 mm. long, triangular-ovoid and not wrinkled. In neither case do the data coincide with those collected by the present author (Table 4). Examination of many colonies produced no constant differences in the seeds of the two species, either in shape, size or texture. With regard to the capsule, no constant size difference was found. The capsules of the two species showed a rather minor difference in most cases, the beak being larger and stouter in C. silvatica, and less abruptly delimited from the rest of the capsule. The overall shapes are the same in both species, however, never approaching a condition which Watsonia § (2), 1961. 102 CLIVE A. STACE could be described as acute, but being broadly ovoid to subglobose, and very obtuse to truncate at the apex. As pointed out in Section 2, it is important to use only four-seeded capsules for these considerations. Apart from the fact that a capsule with fewer than four seeds is less rounded, the seeds are quite different in shape, since the two flat faces of the seeds, which are present where the seed adjoins its two neighbours in the capsule, are absent when there are no neighbouring seeds. In extreme cases the solitary seed is quite round in section. Seeds of C. silvatica, however, do tend to have more conspicuously flattened faces than those of C. sepium, and the hilum tends to be more sunken. An additional precaution is that only absolutely ripe seeds should be used, as the apparently mature yet soft and still hydrated seeds of both species often found in brown capsules are considerably larger than in the truly mature (hard, dry) state. Although Pospichal (1899) and Hegi (1927) give the seed colour as black, Tutin (1952) gives it as dark brown. Observation showed that seed colour varies from light brown to black, sometimes on the same plant. Professor Tutin has kindly informed me that his original observations on fruit characters were carried out with small samples, and that subsequent sampling showed the characters, as described above, to be inconstant. All the known colonies of C. x lucana were separated by shorter distances from pure colonies of the parent species than from other hybrid colonies, and were thus relatively unlikely to be pollinated by the same taxon (although eight such pollinations were success- fully performed artificially). Fruit found on hybrid colonies varied in form from and between that of the two species, and measurements are thus omitted from Table 4, being of little or no significance. 5. DISCUSSION Since the hybrid is highly fertile and capable of forming backcrosses with either parent, as well as forming F, hybrid generations, intermediates of every possible degree may be expected to occur. Fig. 6 shows that the hybrid characters in the area investigated already overlap those of C. sepium, but not those of C. silvatica. If individual flowers are included it is found that the only 7s values possible (4-40) that are not occupied by at least one flower are 37, 39 and 40. 3 Because of this overlap it is difficult to designate a range of structure in terms of Ts values to the hybrid, but the author considers this range to be 13-21, whilst those of C. sepium and C. silvatica are 6-12 and 23-31 (-35) respectively. In colony 71 (Fig. 5) it is interesting to note that the average 7s value for the hybrid element (17-5) is extremely close to the halfway point (18) between the mean 7s values for C. sepium (9) and C. silvatica (27). The mean Ts values for C. sepium and C. silvatica colonies in colonies 1-70 (Fig. 6) are exactly the same as in colony 71, namely 9 and 27 respectively. The mean value for the hybrid colonies is lower, however, being 16. This indicates that the majority of hybrid colonies encountered resulted from crosses involving smaller than average parents, since from the data in Fig. 5 it seems that hybrids are almost exactly intermediate between their two parents. The hybrid colonies in the area of study (12 out of 70 colonies) represent some 17% of the total. The name of the hybrid, C. x Jucana, should theoretically be applied to all plants possessing a genotype derived from two species, whether or not these plants are morphologically distinguishable from either parent. Undoubtedly, from this situation, originally hybrid populations may merge into either parent by continued backcrossing to that parent. The results of this, hybrid swarms, provide interesting genetical investigations into introgression and its effects. Anderson (1949) quite correctly criticises workers who, having proved that hybrids exist in the wild, do not continue their investigations further along the lines indicated above. Unfortunately, due to lack of time and facilities, genetical investigations could not be pursued in the present situation. Introgression may well have a bearing on pink-flowered forms of C. sepium Watsonia 5 (2), 1961. COMPATIBILITY AND HYBRIDISATION IN CALYSTEGIA 103 and C. silvatica, since Brummitt & Heywood (1960) state that the former species may have pink corollas, but that C. silvatica is never predominantly pink-flowered. If hybridisation occurs, and back-crossing to C. silvatica follows, pink-flower genes will undoubtedly leak into the latter species. Brummitt & Heywood (1960) refer all predominantly pink- flowered plants with inflated bracteoles to their C. pulchra (which Scholz (1960) reduces to a variety of C. sylvestris). I have so far reached no personal conclusions as to the taxo- nomic status of pink-flowered Calystegiae. The present work may at first seem in part open to the same criticism that Pritchard (1960) describes for Gentianella. This is that we have (possibly in error) assigned equal importance to each of the four characters used in the hybrid index, since each was trans- ferred to a scale of 1-10. Pritchard, however, was attempting to separate a series of popula- tions, previously recognised as a single group, into two subspecies. The present author suggests that in the present situation this criticism (if it existed) would assume far less importance, since here we are attempting to find a numerical method of expressing the difference between two well-defined taxa, a difference at once obvious to the eye. When we find that method, and have proved it to be successful with a large number of colonies (only plants appearing intermediate to the eye actually falling between the two species ranges), it is relatively much less important than in the case of Gentianella that some of the characters may be given slightly too much or too little significance. In the case of Gentianella there is no good yardstick informing the worker that he is on the right lines. In addition, the fact that many workers have used the hybrid index with great success (cf. Anderson, 1949) would suggest that its method can be safely applied here. As has been found with many other self-incompatible plants, the majority of out- crossings prove successful. Assuming the genetical basis of incompatibility in Calystegia to be of the normal S-allele type, results would tend to suggest that there are a very large number of S-alleles involved in this system, since the chances of finding a similar S-allele even in fairly close proximity (outside the same clone) are apparently small. It seems likely, therefore, that C. silvatica has been introduced into this country on numerous occasions. As far as the experimental taxonomist is concerned, there is little doubt as to the category that the C. sepium-silvatica complex falls into. Using the nomenclature of Turesson, as modified by Clausen, Keck and Hiesey (Heslop-Harrison, 1953), this complex would comprise a single ecospecies if we assume that hybridisation involves no loss of fertility. The number of results so far obtained, however, are not sufficient to be certain of this, and it may be that some slight loss in fertility is incurred, especially after several generations. In this case the complex would fall within a single coenospecies. Not until Tutin (1959) relegated C. silvatica to C. sepium subsp. silvatica (Kit.) Tutin* had the two taxa been regarded as anything but ‘ good’ species by modern British tax- onomists. Brummitt & Heywood (1960) keep the two species apart, mainly on the evidence that they do not hybridise. Although, in fact, they do readily hybridise, and natural hybrids are very widespread, many pairs or more of taxa are known which hybridise freely and yet are retained as species nomenclaturally. One of the best examples is the genus Geum, where whole groups of species comprise a single coenospecies, and several pairs a single ecospecies (Gajewski, 1957). As Gajewski points out, the evolution of an interspecific sterility barrier may not occur apace with morphological differentiation and there is no reason to assume that the two are directly related. A quick glance through the species of Calystegia in the British Museum herbarium shows that several of them are quite close to C. sepium and C. silvatica in appearance, and from the diverse remarks made on the Sheets it appears that considerable difficulty has been encountered in the past in naming many specimens. At least two or three species are closer to C. sepium than is C. silvatica. It is thus clear that only after a detailed cytotaxonomic investigation, on the lines of Gajewski’s work on Geum, will it be possible to state with any certainty the relation of C. sepium and C. silvatica to the other species and to each other. Until this is done it is not possible to assess the importance of interfertility between C. sepium and C. silvatica as *I have since found that this combination was first made by Braun-Blanquet & Maire, Fl. Maroc, 217 (1924). Watsonia 5 (2), 1961. 104 CLIVE A. STACE taxonomic evidence. If several other groups of species in the genus are found to be capable of hybridisation amongst themselves and/or with the C. sepium/silvatica group, then all the species are best treated as taxonomically distinct, since introduction of the sub-species concept here would result in very few species with numerous subspecies —a structure which Gajewski rejects. It might equally be found, on the other hand, that the only two species hybridising in the genus are C. sepium and C. silvatica, when introduction of the subspecies concept (cf. Tutin, 1959) would be advantageous, especially since C. silvatica is severely limited geographically when compared to C. sepium. The author suggests, however, that until any such monographic work is undertaken the two taxa, C. sepium and C. silvatica, are best maintained as distinct species, producing the fertile hybrid C. x lucana. I do not suggest that conclusive proof of the hybrid origin of the numerous inter- mediates found in the wild has been produced, but I consider that the circumstantial evidence outlined above is sufficient to assume this. The results are published as they stand, as there seems to be little chance that I shall be able to pursue the subject further for some time. 6. SUMMARY 1. Of the several characters which have been used in the past to separate C. sepium and C. silvatica, some are considered inconstant, their use being limited to extreme or ‘typical’ plants. It would seem that the best diagnostic characters are the sizes of some parts and the shape of the bracteoles, so that plants without flowers may not be referable to their correct taxon in all cases. 2. By means of a hybrid index of four variables C. sepium and C. silvatica may be easily separated, and any intermediates (hybrids) which might exist recognised. Of seventy small colonies examined twelve (or 17°%) proved to be of hybrid derivation. Two poly- morphic colonies were also investigated, one being solely a mixture of the two species, the other a mixture of the two species with the hybrid between them. 3. In the wild, all three taxa are found to be commonly fertile from natural pollination, and pollen from them is seen to be full and spherical, although germination in glucose solution did not occur. 4. Pollination experiments showed that all three taxa are (probably 100%) self- incompatible. Pollen tube studies showed that the contorted tubes resulting from germina- tion on incompatible styles entered the stigma for a short distance, but that they failed to reach anywhere near the style. 5. Further pollination experiments showed that the three taxa are freely interfertile in all six possible combinations, and that small populations are usually single clones, whilst large (polymorphic) ones are multiclonal. 6. All seeds obtained (either from natural pollinations or from artificial intraspecific or interspecific crosses) proved easy to germinate when in the sub-mature stage. After the seeds become completely mature and desiccated germination may be effected by chipping followed by a long period of soaking, but naturally the seeds appear to enter a dormant stage. Thus the presence of full seeds always indicates fertility. 7. Literature research showed that the hybrid has been long known in North Africa and in Italy, as C. Jucana and C. barbara. The hybrid should be called C. x J/ucana (Tenore) G. Don. A list of the most important diagnostic measurements separating the three taxa is given. C. x Jucana is known from several areas of England and Wales and also from the Channel Isles, Spain, Italy, Albania, Greece, S.W. Russia and N.W. Africa. 8. Difficulties in assigning limits to a fertile hybrid are pointed out, and the possibilities of introgression and its possible effects on pink-flower characters are commented upon. 9. The present situation is compared to that in Geum, and reasons put forward for maintaining the two species distinct, at least for the present. A specimen of the hybrid from Colony 71 (Ts = 16), on which many of the pollination experiments were performed, has been deposited in the British Museum herbarium (BM). Watsonia 5 (2), 1961. COMPATIBILITY AND HYBRIDISATION IN CALYSTEGIA 105 ACKNOWLEDGMENTS I am grateful to Mr. J. Carpenter for interest and help in the earlier work, to Mr. T. Miiller for translation of several passages from the German, and to Mr. J. E. Dandy and Dr. W. T. Stearn for valuable advice on nomenclature. I should also like to express my thanks to Dr. S. M. Walters for kindly reading through the typescript and suggesting some improvements, and to Dr. V. H. Heywood for similar help. REFERENCES ANDERSON, E. (1949). Introgressive Hybridisation. London & New York. Baker, H. G. (1957). Plant Notes; Calystegia. Proc. B.S.B.I., 2, 241. BALL, J. (1877-78). Spicilegirem Florae Maroccanae, 4 (1878). J. Linn. Soc., 16. BATTANDIER, J. A. (1890). Convolvulaceae in Battandier, J. A. & Trabut, L. (1888-95). Flore de I’ Algérie, 4. Bora, J. (1927). Descriptive Flora of the Maltese Islands. BrIQUET, J. (1910— ) Prodrome de la Flore Corse, 3 (2) (1955). BRUMMITT, R. K. & HEywoop, V. H. (1960). Pink Flowered Calystegiae of the Calystegia sepium complex in the British Isles. Proc. B.S.B.I., 3, 384. Cuoisy, J. D. (1845). Convolvulaceae, in De Candolle, A. Prodromus Systematis Naturalis Regni Vegetabilis, 9. CLARKE, W. A. (1897). First Records of British Flowering Plants. Danpy, J. E. (1958). List of British Vascular Plants. London. Don, G. (1831-37). A General System of Gardening and Botany, 4, (1837). ERDTMAN, G. (1952). Pollen Morphology and Plant Taxonomy. 1-Angiosperms. Stockholm & Waltham, Mass. Fiori, A. (1923-49). Nuova Flora Analitica D’Italia, 2 (2). (1926). GAJEWSKI, W. (1957). A Cytogenetic Study on the Genus Geum. Monographiae Botanicae, Warsaw. 4. HARRISON, J. HeEsLop- (1953). New Concepts in Flowering Plant Taxonomy. Hea, G. (1906-31). J/lustrierte Flora von Mittel-Europa, 5 (3) (1927). HUTCHINSON, J. (1945). Common Wild Flowers. Harmondsworth. JAHANDIEZ, E. & Maire, R. (1931-34). Catalogue des Plantes du Maroc, 3 (1934) KNUTH, P. (1906-9). Handbook of Flower Pollination, 3 (1909) tr. J. R. Ainsworth Davis. Lewis, D. (1954). Comparative Incompatibility in Angiosperms and Fungi, Adv. Gen. 6, 235. Lous.Ley, J. E. (1948). Calystegia sylvestris,in Plant Notes. Rept. Bot. Soc. & Exch. Cl., 13, 265. MurBeECck, S. (1905). Contributions . . . de la Flore du Nord-Ouest del Afrique et... de la Tunisie. Sér. 2. Pau, C. (1924). Plantas del Norte de Yebala. Mem. Real Soc. Espan. Hist. Nat., 12, 364. PomeEL, A. (1874-76). Nouveaux Matériaux pour la Flore Atlantique, 1 & 2. PosPICHAL, E. (1897-99). Flora des Oesterreichischen Kiistenlandes, 2 (1899). PRAEGER, R. Li. (1934). The Botanist in Ireland, Dublin. PRITCHARD, N. M. (1960). Gentianella in Britain. II. Watsonia, 4, 218. SALISBURY, E. J. (1961). Weeds and Aliens. London. SAUVAGE, C. & VINDT, J. (1952— ). Flore du Maroc, 2 (1954). SCHOLZ, H. (1960). Notiz iiber Calystegia sylvestris, Willdenowia, 2, 398. STEARN, W. T. (1951). Calystegia sepium & C. sylvestris. J. R. Hort. Soc., 76, 114. TENORE, M. (1826). Ad Florae Neapolitanae Prodromum App. 5. TENORE, M. (1827). Viaggio in Alcuni Luoghi della Basiliata e della Calabria Citeriore. TENORE, M. (1811-36). Flora Napolitana, 3 (1824-29). TENORE, M. (1831). Succinta Relazione del Viaggio fatto in Abruzzo ed in alcune parti dello Stato Pontificio. (Reprint from Atti dell’ Accademia Pontaniana, I, 147). TENORE, M. (1831). Sylloge Plantarum Vascularium Florae Neapolitanae. TuTIn, T. G. (1952). Convolvulaceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Cambridge. TuTIn, T. G. (1959). Convolvulaceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Excursion Flora of the British Isles. Cambridge. WALTERS, S. M. & Martin, A. (1958). Self-incompatibility and Hybridisation in Calystegia. Exhibition Meeting 1957. Proc. B.S.B.I., 3, 104. WALTERS, S. M. & Wess, D. A. (1956). Calystegia, in Plant Notes. Proc. B.S.B.I., 2, 22. Warkins, A. E. (1925). Genetic and Cytological Studies in Wheat. II. Journ. Gen., 15, 323 Wo LLeEy-Dop, A. H. (1914). A Flora of Gibraltar and the Neighbourhood, Journ. Bot., 52 (suppl.). Watsonia 5 (2), 1961. TAXONOMY AND NOMENCLATURE IN SOME SPECIES OF THE GENUS ARUM L. By C. T. PRIME 1. THE RELATIVE STATUS OF ARUM ITALICUM AND ARUM NEGLECTUM Recent studies of the distribution of Arum neglectum (Townsend) Ridley and Arum italicum Miller (Prime, Buckle and Lovis, 1955, 1960) have made it difficult to sustain the two as separate species. The differences between the two have been discussed by Ridley (1938) but plants intermediate in every respect can be found in large populations on the Continent, and in the extreme west of England. There is a ratio cline, the italicum form being commoner in the Mediterranean area, and the neg/ectum form in south Britain. Particularly attractive forms of Arum italicum with dark green leaves, diverging basal lobes and a well-marbled white venation have been cultivated in this country and are occasionally found as escapes. This fact complicates the interpretation of the distribution in the British Isles, but the main facts of the cline are clear enough, and it seems best to treat the two as subspecies. The first real attempt to classify the common Aroids was made by Tournefort (1719). He describes three genera, Arum, Dracunculus and Arisarum. Linnaeus’s copy of this book, now in possession of the Linnean Society, contains his names against the species listed, and he included them all as belonging to the single genus Arum. Tournefort gives the synonymy for Arum maculatum and amongst the names may be noted ‘Arum venis albis C. B. Pinax’ and ‘Arum yenis albis, Italicum, maximum, H. R. Par.’ (See Bauhin, C. (1623) and Joncquet, D. (1665)). Against all the synonyms, including the two above, Linnaeus has written only one name, Arum maculatum, so it is clear that he considered all these to refer to the one species. Miller (1759) describes Arum maculatum and Arum italicum separately. The descrip- tion of Arum italicum reads: “ Arum acaule foliis hastatis acutis petiolis longissimis spatha erecta. Arum without Stalk, pointed Spear-shaped leaves, with very long Petals, and a large upright Spathe. This is the Arum venis albis, Italicum, maximum. H. R. Par. Largest Italian Arum with white Veins.”’ Again, Linnaeus in his own annotated copy of the Dictionary ignores this species, so he evidently regarded its description by Miller as unwarranted splitting. Miller goes on to say that the plant “‘ grows naturally in Italy, Spain and Portugal from whence I have received the Seeds. The Leaves of this Sort rise a foot and half high, are very large, running out to a Point; these are finely veined with white, interspersed with black Spots, which together with the fine shining green of their Surface, make a pretty Variety. The Flowers grow near a Foot high, and have very long upright Spathas, which are of a pale green, inclining to white; these appear at the End of April, or Beginning of May, and the Seeds are ripe in August.” It seems therefore quite clear that Miller intended his description to apply to the more common Mediterranean plant with marked white veins and not to the plant found in this country. Townsend (1883) first described the latter as A. italicum var. neglectum. He makes the following observation “‘ The Isle of Wight plant more frequently has dark coloured spots on the leaves, which are rarely seen on the continental A. italicum. The basal lobes are less triangular in outline, and are less spreading than those of the usual continental plant; the apex of the leaf is more rounded. The Isle of Wight form is in my experience rare on the Continent, but I have it from Cannes, together with the usual form from 106 Watsonia 5 (2), 1961. TAXONOMY AND NOMENCLATURE IN ARUM 107 Bordighera, and both retain their own characters. A. italicum appears earlier but flowers later than A. maculatum.” He cites a specimen in the British Museum Herbarium collected by Trimen. Ridley (1938) gave a fuller description, and raised the variety to specific rank. Though Townsend’s diagnosis is rather slight, and not quite characteristic of A. neglectum as understood today, the name is validly published. As a subspecies of A. italicum it becomes Arum italicum Mill. subsp. neglectum (Townsend) Prime, comb. nov. (A. italicum var. neglectum Townsend, Fl. Hants., (1883); A. neglectum Ridley, J. Bot., Lond., 78: 144 (1938). Holotype: Ventnor, October 1866, Trimen (BM)). The principal differences between the two subspecies are : Subsp. italicum Subsp. neglectum 1. Leaf-shape relatively narrow. Length/ Leaf-shape relatively broad. Length/ width 1-7. width 1-5. 2. Leaf-apex pointed, almost acuminate. Leaf-apex rounded at side, acute. 3. Basal lobes of leaf relatively narrow, Basal lobes relatively broad, less divergent, diverging, not overlapping. Angle sometimes overlapping. Angle between between mid-line of lobe and petiole mid-line of lobe and petrole about 45°. about 60°. | 4. White venation very marked. White venation absent or inconspicuous. 5. Leaves dark green. Leaves lighter green. 6. No. of seeds per fruit 2-4. No. of seeds per fruit 1-2. 7. Fruits rather more turgid. Fruits less turgid, slightly more oblong in shape. oe The last difference is a reflection of the different number of seeds in the fruit. The Distinction between Arum italicum and Arum maculatum. Arum maculatum, it may be emphasised, is a very different plant from Arum italicum. A. italicum is larger in all its parts; it is winter-green, the leaves being fully expanded by early November. The petioles are longer in proportion to the blade length and the laminae stand more erect. The spathes are relatively larger than the spadix when compared with A. maculatum. The spathes droop earlier, and the reproductive organs occupy a relatively greater space, for the ring of ovaries is about twice as long as broad. In A. maculatum this ring is about one and a half times as long as broad. The rudimentary male flowers do not taper very directly into the filaments, and the spadix is yellow with a fairly sharply contracted base. The fruits are larger and contain two to five seeds, which are also larger than those of A. maculatum. The fruits of A. italicum are also slower to turn to the bright red colour of autumn, and are usually hardly as bright in colour. 2. THE SUBSPECIES OF ARUM MACULATUM The Danish plant with 28 chromosomes, an account of which has already been given (Prime, 1955) is a distinct form of Arum maculatum and should be treated as a separate subspecies. This raises the problem of the correct typification of A. maculatum L. It would of course be convenient if it could be demonstrated that the name A. maculatum subsp. maculatum should be applied to the more widely distributed and commoner plant with 2n = 56 chromosomes. As Linnaeus himself said (Critica botanica, p. 246 in W. T. Stearn, 1957), “If an accepted genus has to be split up into several, according to the Law of Nature and of Science, then the name which formerly belonged to the whole should be kept to denote the best known and officinal plant.” It seems most likely that the name A. maculatum was first given to this plant, but although at the moment plants with 28 chromosomes are known only with certainty from Denmark, they may occur elsewhere. Schmucker (1925) gives 2n = c.32 for some German plants, probably in error for 27 = 28, and it seems not unreasonable to believe that plants Watsonia 5 (2), 1961. 108 Ca PRIME with 2n = 28 may occur in Germany and even further afield. In Denmark, Arum maculatum (2n = 56) has been widely planted, but the plant with 28 chromosomes is confined to the southern part of the country. It is common on the islands of Lolland and Falster; frequent on Bornholm, Mon, and the south and west part of Zeeland. In east Jutland, where it is only found in rich soils and woods towards the sea, its existence is probably due to the fact that the summer is relatively warm and long. The name A. maculatum was first used for the species by Tabernaemontanus (1590) and Bauhin (1623), and this name was retained by Linnaeus in Species Plantarum (1753). The specimen in the Linnean herbarium (No. 1079-8) is most likely of the subspecies with 2n = 56, but the nature and condition of the specimen make it impossible to decide with certainty. Another early reference to a specimen is in the Hortus Cliffortianus (1737). The phrase name here is the same as in the Species Plantarum, 1.e. ““Arum acaule, foliis hastatis integerrimus, spadice clavato,’ Habit and distribution are given as “ Crescit ad sepes in umbrosis per Germanium, Hollandiam, Angliam, Galliam, Italiam.” The species described in the Hortus Cliffortianus are essentially those growing at that time in Clifford’s garden, which was at Hartekamp in Holland, and they are also the plants repiesented in his her- barium. This herbarium is in the British Museum, and has been consulted, but the specimen labelled in a later unknown hand is not Arum maculatum at all but a quite different unidenti- fied Aroid. The only illustrated work cited by Linnaeus is by l’Obel (1591), but the illustration is inadequate for a definite identification to be made. It is thus not possible at the present time to select a satisfactory nomenclatural type for Arum maculatum. However, in the absence of conclusive evidence demonstrating which of the two subspecies is in fact the type subspecies of A. maculatum, it nevertheless remains more probable that Linnaeus originally described the plant with 2” = 56 chromosomes, and therefore the name A. maculatum subsp. maculatum should be applied to this plant. Accordingly, the plant with 2n = 28 chromosomes is described below as a new subspecies. Arum maculatum subsp. danicum Prime, subsp. nov. Surculus maximus cormi in medio situs; folia immaculata et pro longitudine latiora quam in subsp. maculato, minus hastata; spatha ad imam partem collata, brevior quam in subsp. maculato; spadix cylindrica, non latior basim versus. Chromosomata 2n = 28. Holotypus : Grénnese Skov, nr. Frederiksvoerk, North Zeeland, T. W. Bocher, 11 May 1961, in Herb. Ke@benhavn (C). Isotypus in Herb. Mus. Brit. (BM). Distributio. Solum in Dania meridionali-orientali. Main shoot of corm centrally situated; leaves unspotted, relatively broader than in subsp. maculatum, less hastate; ratio of spathe length to the basal part less than in subsp. maculatum; spadix uniformly cylindrical, not tapering from the base. Chromosome number 2n = 28. Distribution (so far as is known) confined to Denmark, mainly in woods on the islands and in south-east Jutland (Kgie, 1939). I should like to acknowledge with many thanks the very great help I have received in preparing this paper from Mr. J. F. M. Cannon and Dr. W. T. Stearn of the British Museum (Natural History), and from Dr. J. D. Lovis of Leeds University. REFERENCES BAUHIN, C. (1623). Theatri Botanici. Basle. JONCQUET, D. (1665). Hortus Regius Parisiensis. Paris. Kgie, A. and Kare, M. (1939). Udbredelsen af Geraniaceae, Araceae, Lemnaceae og Droseraceae i Danmark. Bot. Tidskr., 45, 73-100. LINNAEUS, C. (1737). Hortus Cliffortianus. Amsterdam. LINNAEUS, C. (1753). Species Plantarum. Stockholm. L’OBEL, M. DE (1591). Stirpium Icones Antwerp. Watsonia 5 (2), 1961. TAXONOMY AND NOMENCLATURE IN ARUM 109 MILLER, P. (1759). Gardener’s Dictionary, Ed. 7. London. Prime, C. T. (1955). Problems of speciation in the British species of Arum. Species Studies in the British Flora (ed. J. R. Lousley), p. 195, B.S.B.I., London. PRIME, C. T., BUCKLE, O. and Lovis, J. D. (1955). The distribution and ecology of Arum neglectum in Southern England. Part I. Sussex, Hants. and Dorset. Proc. B.S.B.J. 1, 287. Prime, C. T., BucKLE, O. and Lovis, J. D. (1960). The distribution and ecology of Arum neglectum in Southern England and Wales. Part II. Dorset, Devon, Cornwall, Isles of Scilly, Channel Islands and Wales. Proc. B.S.B.I., 4, 26. Rip ey, H. N. (1938). Arum neglectum (Towns.) Ridley, J. Bot. 76, 144. SCHMUCKER, TH. (1925). Beitrage zur Biologie und Physiologie von Arum maculatum. Flora 18, 19, 460. STEARN, W. T. (1957). Linnaeus’s Species Plantarum. Vol. I with introduction by W. T. Stearn. Ray Society, London. TABERNAEMONTANUS, J. T. (1590). Eicones Plantarum. Frankfurt. TOURNEFORT, J. P. (1719). Institutiones Rei Herbariae. Paris. TOWNSEND, F. (1883). Flora of Hampshire. London. Watsonia 5 (2), 1961. BOOK REVIEWS New Records for Illinois Vascular Plants. Winterringer, G. S. and Evers, R. A. (Eds.). Pp. 135. Illinois State Museum, Scientific Papers Series, Vol. XI. 1960. The State of Illinois is comparable in area with England and Wales, being about 400 miles long and 200 miles wide. It is divided into 102 counties and these, the basic unit of plant recording, are thus almost equivalent in area to the British vice-county. This work is essentially a topographical botany and first supplement rolled into one, and is the sort of work H. C. Watson would have liked to have produced if the printing processes of today had been available 100 years ago. Instead of a printed list of counties, there is a map for each species (sometimes two rare species share one map) showing by a symbol, conventionally placed in the centre, those counties from which a species has been reliably recorded. In all there are 1,375 maps dealing with over 1,500 species; the number of vascular plants in Illinois is of the same order as the number found in England and Wales. Vascular Plants of Illinois by G. N. Jones and G. D. Fuller was published in 1954, and the present work is proof of the stimulus which it gave to the study of the Illinois Flora. Within 6 years 70 species have been added, an increase of about 5%, and details of these finds are listed. Many of them are European species, introduced into the United States, which are still spreading, and a British botanist visiting Illinois will feel more at home in the future now that he can expect to find Butomus umbellatus, Calamagrostis epigejos, Puccinellia distans, Epipactis helleborine, Cerastium atrovirens, Rosa micrantha, Trifolium fragiferum, Myosotis sylvatica, Verbascum virgatum, Carduus acanthoides, Hypochaeris radicata and Senecio viscosus. All of these have been recently recorded for the first time in the State. The number of new county records made in the same period is phenomenal. In some cases the number of county records has doubled in 6 years and this suggests that botanical recording in this part of the United States has now reached the position in which we found ourselves in Britain over 100 years ago. Further evidence for this is obtained from an assessment of the number of species which are known from every county in Illinois—the total is only about 20, whereas in England and Wales with a similar area and number of counties the total is at least ten times as great. It seems that, if new county records are your special pleasure, you stand a better change of success in Montgomery and Cumberland, Illinois, than in v.cs 47 and 70, Great Britain. Considering the level of knowledge of the flora which has been achieved, then the use of maps to show species distribution on a county basis is understandable. However, the value of the work is less great than it might be for a foreign botanist whose main interest in the flora is likely to be the distribution of species he knows elsewhere in relation to climate, geology, and topography, for no maps of these features are included. It would have increased the cost of production very little for such maps, on the same scale, to have been added. Then we might have understood why Geum rivale is confined to the north-east of the State and Arabidopsis thaliana to the south and what combination of factors determines that in distribution Silene vulgaris and Myosurus minimus are almost mutually exclusive. The omission is particularly frustrating for the British botanist for Illinois is in the part of North America which has a climate most similar to our own to judge from maps of Thornthwaite’s climatic regions. The greatest difficulty which presents itself to a British plant geographer is status. One assumes that the same difficulty applies in the United States : yet no attempt is made in the work to distinguish between native and non-native distributions : another hazard which confronts the would-be interpreter and some- what reduces the reliance which can be placed upon the maps. F. H. PERRING. A California Flora. Philip A. Munz (in collaboration with David D. Keck). Pp. 1681, with 2 maps, 134 text figures and colour frontispiece. California University Press, Berkeley and Los Angeles; and Cam- bridge University Press, 1959. Price 86s. 6d. It is tempting for a European botanist to consider California in relation to Spain, which is of similar size, lies nearly in the same latitude, and has a similar huge total (nearly 6,000 species) of vascular plants, but the physical features and geological history of the two countries are so different that, in the end, it is perhaps best for him to confine any comparison to the plants of the drier, desert areas in California and the more arid Spanish provinces. Some British botanists will be interested in California from a different motive : many of our aliens or garden plants, for instance in the Compositaec, Polemoniaceae, Boraginaceae and Rhamnaceae (Ceanothus), are natives of this State while, in reverse, several hundreds of European and North African species are completely naturalised there. Others, plant geographers, will perhaps concentrate on the extension south (or the failure to extend south) from the Rocky Mountains into the Californian 110 Watsonia 5 (2), 1961. BOOK REVIEWS 111 sierras of familiar genera and even species of northern, sometimes alpine or subalpine, distribution; while tree-lovers will be most interested in the great Redwood and Douglas Fir forests of the moist coastal region. Professor Munz’s excellent new flora of California, with its brief but lucid introductory sections, can be warmly recommended to European taxonomists, ecologists and plant geographers. It discards the 6 Merriam Life Zones recognized by the late W. L. Jepson, whose large-scale Flora of California, begun in 1909, is due to be completed. Instead, the author has 5 Biotic Provinces, 11 major Vegetation Types, and 29 Plant Communities, which are fully classified, and throughout the text the native species are assigned to one or more of these, with a description of habitat, altitudinal range and distribution within the State. The body of the work follows customary lines, with indented keys, and text figures illustrating characters of orders, families or subdivisions of families, not of lower ranks. Chromosome numbers (with author citations) are given when known, and references to revisions follow generic descriptions. The order of the families is unconventional and was prepared by Professor Munz’s collaborator, Dr. David D. Keck. There are two small maps, for counties and main topographical features; European readers could have wished for a larger, coloured, folding map with place-names. Thin paper, like that of Fernald’s edition of Gray’s *“ Manual’, was inevitable or the book would have needed two volumes. The author does not give a summary of the composition of the flora, but a statistical report on his book was published in Leaflets of Western Botany, vol. ix, No. 8, pp. 117-123 (November 1960), by Gladys L. Smith and Anita M. Noldeke, and this was followed by a summary of the endemism prepared by A. M. Noldeke and J. T. Howell (1.c., pp. 124-127). These writers provide a total of 5,675 native species, with 1,414 (or 29-02%) endemic species and 29 endemic genera. Some idea of the extent of evolution within Californian genera may be gathered from the large representation of Carex (144 species), Astragalus (93), Phacelia (87), Lupinus (82), Mimulus (77), Eriogonum (76), Cryptantha (59), Trifolium (49) and Arctostaphylos (43). The high endemism is explained, as Professor Munz points out, partly by the nature of the region, so broken up into separate mountain ranges and low-lying valleys, and partly by changes in climate and migrations both north and south and from altitude to altitude. With such a field of exploration before them, and such a nice book to work with, British botanists might soon give up their exhausting climbs in the Scottish Highlands, and even skip the more rewarding limestone cliffs of Asiatic Turkey. N. Y. SANDWITH. Watsonia 5 (2), 1961. INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the taxonomy or distribution of British vascular plants or Charophytes are invited from both members of the Society and others. They should be typewritten, with wide margins, double spaced, on one side of the paper only; contributors are recommended to keep a carbon copy of their typescripts. The form adopted in recent parts of Watsonia should be used for layout, headings, citations and references. Contributors are urged to avoid very complicated hierarchies of headings and sub-headings, and to check carefully the consistency of those that they use. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British flowering plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (N.H.) and B.S.B.I., 1958), and may then be cited without authorities. Otherwise, authors of names must be cited, at least on the first occasion where they appear in the text. Except for citations of the place of publication of plant names, full references should be listed in alphabetical order of authors names at the end of the paper. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals, 3rd ed. (1952). References to herbaria should include the abbreviations given in British Herbaria (B.S.B.1., 1957) and Index Herbariorum 3rd ed. (1.A.P.T., 1956). Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new taxa, and the main conclusions. Line-drawings should be boldly drawn in Indian ink on Bristol board or similar smooth white card, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. Graphs can be reproduced from originals on graph paper with faint blue ruling (not grey or any other colour), but all lines to appear on the finished block must be inserted in Indian ink. Lettering on line-drawings and graphs should be inserted lightly in pencil, and will be finished in uniform style. If an illustration includes plant-names or place-names, it is advisable to type these clearly on a separate sheet of paper. Photo- graphs can only be accepted where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable for the necessary reduction. It should be remem- bered that the finest detail on the originals may be lost even on the best half-tone blocks. If in doubt about the citation of names or references, or the presentation of illustrations or tabular matter, contri- butors are advised to consult the editor before submitting their typescripts. Authors will normally receive both galley-proofs and page-proofs. Particular care should be given to the thorough checking of references and tables. Alterations in page-proof should be avoided as far as possible, and any words or phrases deleted replaced by others of equal length. Twenty-five separates are given free to authors of papers. Further copies may be obtained at the printer’s current price, and must be ordered when the proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to the Editor, Dr. M. C. F. Proctor, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. IRISH NATURALISTS’ JOURNAL A Magazine of Natural History Published Every Quarter by the I.N.J. Committee. Edited by Miss M. P. H. KERTLAND, M.Sc., with the assistance of Sectional Editors. Annual Subscription, 10/- post free. Single Parts, 3/6. All communications to be addressed to:— The Editor, Department of Botany, Queen’s University, Belfast. PUBLISHED JULY 1958 LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANpy for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf. of a sub-committee of the Society appointed for this purpose. Dr. George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and will without doubt become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8 vo., 176+xvi pages. Bound in cloth. Price : 10/-. Interleaved Copies 15/-. Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. A BIBLIOGRAPHICAL INDEX OF THE BRITISH FLORA COMPILED BY N. DOUGLAS SIMPSON, ™.A., F.Ls. The purpose of this work is to provide references to sources of information whereby flowering plants, vascular cryptogams and Charophytes found in Britain may be identified, their history traced and their geographical range determined. In addition, information is provided on plantlore, local names, poisonous plants and weeds. The work has been in preparation for nearly twenty years and contains over 65,000 entries, including references to books, articles and manuscripts relating to the flora of the British Isles from the fifteenth century to the present time. Demy 4to, bound, 448 pages in double column. Limited Edition of 750 copies. PRICE £3 I5s. (postage extra) Obtainable from the compiler at “‘ Maesbury,’ 3 Cavendish Road, Bournemouth, Hants. The Salisbury Press Ltd., Salisbury—283898 “ __ WATSONIA JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 5 APRIL, 1962 Pt 3 CONTENTS THE TAXONOMIC SEPARATION OF THE CYTOLOGICAL RACES OF KOHLRAUSCHIA PROLIFERA SENSU LATO. By P. W. BALL and V. H. HEywoop INTERSPECIFIC RELATIONSHIPS AND INTRASPECIFIC VARIATION OF CHENOPODIUM ALBUM L. IN BRITAIN. II. THE CHROMOSOME NUMBERS OF C. ALBUM AND OTHER SPECIES. By M. J. COLE STUDIES ON RANUNCULUS L. SUBGENUS BATRACHIUM (DC.) A. Gray. I. CHROMOSOME NUMBERS. By C. D. K. Cook STUDIES IN THE BRITISH EPIPACTIS. V. EPIPACTIS LEPTOCHILA; WITH SOME NOTES ON E. DUNENSIS AND E. MUELLERI. By DONALD P. YOUNG STUDIES IN THE BRITISH EPIPACTIS. VI. SOME FURTHER NOTES ON E. PHYLLANTHES. By DONALD P. YOUNG .. STUDIES IN THE BRITISH EpipActis. VII. SEED DIMENSIONS AND ROOT DIAMETERS. By DONALD P. YOUNG SOME NOTES ON GALEOPSIS LADANUM L. AND G. ANGUSTIFOLIA EHRH. EX Horrm. By C. C. TOWNSEND CHROMOSOME NUMBER, MORPHOLOGY, AND BREEDING BEHAVIOUR IN THE BRITISH SALICORNIAE. By D. H. DALBY.. STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. By D. M. BARLING Book REVIEW. . 113-116 117-122 123-126 127-135 136-139 140-142 143-149 150-162 163-55 174 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH Museum (NATURAL HisTory), LONDON PRICE FIFTEEN SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES Patroness: H.R.H. THE PRINCESS ROYAL OFFICERS President: J. E. LOUSLEY Vice-Presidents : E. MILNE-REDHEAD, J. E. DANDY, Miss C. M. ROB and Dr. E. F. WARBURG Hon. General Secretary: J. G. DONY, 41 Somerset Avenue, Luton, Beds. Hon. Treasurer: J. C. GARDINER, Thrift House, 12&14 Wigmore Street, London, W.1. Hon. Editors: (Watsonia) M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT Hon. Meetings Secretary: H. J. M. BOWEN, Pomander House, Harwell, Berks. Hon. Field Secretary: P. C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Longfield, Kent. Hon. Junior Activities Secretary: P. F. HUNT, The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey. Applications for membership should be addressed to the Hon. Assistant Secretary, D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13, from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 15/— per part. To be obtained from E. B. BANGERTER, Botanical Society of the British Isles, c/o Dept. of Botany, British Museum, (Natural History), London, S.W.7. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/- for Junior Members, and 10/- for Family Members, and the Society’s year runs from January Ist to December 31st. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS Ail enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13 THE TAXONOMIC SEPARATION OF THE CYTOLOGICAL RACES OF KOHLRAUSCHIA PROLIFERA (L.) KUNTH SENSU LATO By P. W. BALL and V. H. HEYwoop University of Liverpool INTRODUCTION It has been realised for some considerable time that Koh/rauschia prolifera (L.) Kunth contains both diploid and tetraploid plants (Blackburn, 1933, Bocher et al., 1953, 1955). The cytotypes appear to have different distributions, the more widespread diploid being replaced in W.S.W. Europe by the tetraploid. Because of the close morphological similarity between the cytotypes, and the difficulty of separating them, K. prolifera has come to be widely referred to in the literature as an example of the phenomenon of semi-cryptic polyploidy (cf. Heywood, 1958, Larsen, 1960). In the course of a revision of the whole genus, we have had occasion to make a detailed study of the K. prolifera group, as a result of which it has become apparent that not only can the diploid and tetraploid races be separated fairly satisfactorily on the basis of their seed-coat morphology, but the differences had been recognised precisely many years previously. Although it has usually been assumed that the tetraploids have arisen from the diploids by autoploidy, the morphological evidence suggests that another diploid species, K. velutina (Guss.) Reichb., is involved and that the tetraploid has been formed as a result of allopoly- ploidy between it and diploid K. prolifera. TYPIFICATION OF K. PROLIFERA This species was based on Dianthus prolifer L., Sp. Pl., 410 (1753). Typification has proved difficult since there is no positive evidence that the specimen of Dianthus prolifer in the Linnean Herbarium has a claim to being selected as a lectotype. The sheet does not bear the Species Plantarum number normally found on those of species described in the first edition. If, however, the specimen is accepted as authentic there is no reason to suppose that it is other than diploid Kohlrauschia prolifera, although in the absence of seeds it cannot be identified with certainty. The synonyms and distribution given by Linnaeus could apply to either the diploid or tetraploid species. This kind of situation arises occasionally with the typification of species in polyploid groups where the various species may be separated from one another by micro-characters which are not available for examination in the type material. In such cases the wisest course is to accept a typifica- tion which does not upset the traditional interpretation of the species, unless there is strong evidence to the contrary. Accordingly the name K. prolifera (L.) Kunth is accepted here for the diploid species which occurs in most countries of Europe and to a small extent in N.W. Africa and Anatolia. THE IDENTITY OF DIANTHUS NANTEUILII BURNAT During the investigation of the genus it was necessary to account for Dianthus nanteuilii Burnat which was published in Flore des Alpes Maritimes, 1: 221-222 (1892) with a full description based on 38 specimens from various localities at Cannes and Agay. The leaf sheaths are described by Burnat as ‘ aussi larges que longues ’ and the seeds are described in considerable detail : ‘absolutely intermediate between the two preceding species [D. prolifer and D. velutinus] in dimensions and shape : more convex dorsally than in lulls) Watsonia 5 (3), 1962. SMITHSONIAN pin ge INSTITUTION ~~?" ° rT. oP > 114 P. W. BALL and V. H. HEYWOOD D. velutinus, and less hollow-concave than in the latter, the outer surface striate-tubercu- lous, with tubercles similar to those of D. velutinus, but much closer together and less projecting.’ Burnat considered the species (apart from the rare presence of glandular hairs) inter- mediate between D. prolifer and D. velutinus, but Nanteuil thought that a hybrid origin was unlikely in view of the commonness of the new form in the vicinity of Cannes while only one specimen of D. velutinus was seen. D. prolifer did not on the other hand appear to be less abundant although it flowered later. He gave the following times of maturity of the first seeds : velutinus 20th May intermediate Ist June prolifer 25th June Time of flowering needs further investigation both in the field and in cultivation, but D. velutinus does appear in the light of later observations to be the earliest of the three to flower (cf. Bécher et al., 1953). Further examination of large numbers of individuals by Nanteuil failed to reveal any transitional forms. This is confirmed by our own comparative studies on a wide range of material from diverse provenances. It is interesting to note that, although Burnat appears to have been the first to publish a description of this species, there are two older sheets in the Kew Herbarium from the Pyrenees which have manuscript names and extensive and accurate descriptions appended. KOHLRAUSCHIA VELUTINA (GUSS.) REICHB. The third taxon involved in this complex, K. velutina, is characterized by its long leaf sheaths (at least twice as long as broad) and by its strongly tuberculate seeds. The epithet velutina refers to the dense glandular-tomentose indumentum which is usually found on some of the internodes in the middle part of the stem. This character is not, however, constant and forms with glabrous stems occur which have in the past been confused with K. prolifera, especially in Italy, Sicily and Sardinia. K. velutina is diploid (Boécher et a/. 1953, 1955) and is regarded by some writers, such as Briquet (1910), as only subspecifically distinct from K. prolifera. It is widely distributed in the Mediterranean region from Portugal and Spain to Turkey and Palestine. The karyo- type has, however, been shown to be quite distinctive (Bocher et al., loc. cit.) as discussed below. TAXONOMIC COMPARISONS The accompanying table sets out the differences between the three units. From this it will be seen that several characters can be employed to separate them, with K. nanteuilii occupying an intermediate position. Of these characters, however, the only constant and reliable ones are those of the seed testa. TABLE | K. prolifera K. nanteuilii K. velutina Internodes Glabrous Glabrous to Densely glandular- (middle part of stem) tomentose tomentose, rarely glabrous Leaf sheaths | Broader than long to | 1-2 times as long as | At least twice as long | about as long as broad > broad as broad Width of Petal limb 2-3-5 mm | 2-3 mm 1-2-2-5 (— 3) mm Inner bracts Obtuse | Obtuse or mucronate Mucronate Seed size 1-3-1-9 mm 1-2-1-9 mm 1-1-3 (— 1-4) mm Testa pattern Reticulate Tuberculate Tuberculate to papillose (Plate 6a) (Plate 6b) | (Plate 6c) Watsonia 5 (3), 1962. LANNE) (top) K. prolifera (Kobenhayn 3325), b. (middle) K. nanteuilii (K@benhavn 3248) c. (bottom) K. velutina (Kebenhayn 3303) CYTOLOGICAL RACES OF KOHLRAUSCHIA PROLIFERA 115 The details of the seed testa are shown in Plate 5. The characteristic pattern of each species is easily appreciated by using a low-power dissecting microscope, although with a little practice an ordinary | @20 oe Kk Scege nS oS we & © ‘ee efe® 7 Pa ed ® A . og” @ @8 6 PE @ bad G0 & g ® 8 © e2@ e Seeed @ be Be ed ®@ 1 setsges P tamegier e8 Fig. 1. Chromosomes from the genus Chenopodium. a. C. viride L. (2n = 18) Oslo, Norway; b. C. ficifolium Sm. (2n = 18) Southampton; c. C. murale L. (2n = 18) Southampton; d. C. album L. (2n = 54) Southampton; e. C. reticulatum Aell. (2n = 54) Southampton; f. C. urbicum L. (2n = 36) Holbury, Hants; g. C. bonus-henricus L. (2n = 36) Southampton; h. C. berlandieri Mog. subsp. zschackei, (J. Murr) Zob. (2n = 36) Southampton; i. C. polyspermum L. (2n = 18) Southampton; j. C. opulifolium Schrad. ex Koch & Ziz (n = 27) Southampton; k. C. variabile Aell. (x = 18) Romsey, Hants; 1. C. album L. (n = 27) Saskatchewan, Canada; m. C. reticulatum Aell. (n = 27) Southampton. a-i, root tip mitoses x c. 1000. j-m, PMC metaphases x c. 1350. All drawings are either by camera lucida or are tracings from photographs. Watsonia 5 (3), 1962. 120 M. J. COLE TABLE 2 Chromosome counts of C. album L. (incl. C. reticulatum Aell.) and the localities of the foreign material Locality Seed marking R.T./P.M.C. 2n Saskatchewan, Canada A P.M.C. 54 Ottawa, Canada A R.T. 54 Minnesota, U.S.A. | A Rel 54 Carenac, France A R.T. 54 Neuchatel, Switzerland A | R.T. | 54 Coburg, Germany A | R.T. 54 Bremen, Germany A R.T. 54 Copenhagen, Denmark A R.T. | 54 Pavia, Italy A R.T. 54 Christchurch, New Zealand A Ree 54 Rouen, France | R R.T. 54 Vienne, France | R R.T. 54 Dandenong, Australia | R Ran, 54 13 Counts from America, Europe and Australasia; all hexaploid, 2n = 54. TABLE 3 Chromosome counts of Chenopodium spp. excl. C. album L. Localities of British material Species Locality VCs R.T/P.M.C. 2n C. ficifolium Sm. | Southampton, Hants. 11 Ro 18 C. murale L. | Southampton, Hants. 1 R.T. 18 C. polyspermum L. _ Romsey, Hants. 11 R.T. 18 C. viride L. (C. suecicum J. Mutr) | Enfield, Middx. PAI R.T. 18 C. viride L. (C. suecicum J. Murr) | Newark, Notts. | 56 R.T. 18 C. viride L. (C. suecicum J. Murr) | Dumfries, Dumfries. We, P.M.C. 18 C. bonus-henricus L. | Southampton, Hants. , 11 Rede 36 C. rubrum. L. | Southampton, Hants. 11 R.T. 36 +C. opulifolium Schrad. | Southampton, Hants. 11 | P.M.@c& Rete a4: *C. berlandieri Mog. ssp. zschackei (J. Murr) Zob.| Southampton, Hants. 11 R.T. 36 *C, urbicum L. Holbury, Hants. 11 Reis 36 *C. variabile Aell. Romsey, Hants. 11 |P.M.C.&R.T.| 36 *New determinations. First count on British material contradicting a previous Continental record (see text). Other numbers are first counts on British material confirming previous Continental records. (ii) Foreign material. Seed sent from colleagues abroad enabled counts to be made from other continents. The results (see Table 2) are in agreement with those from Britain : only the hexaploid count was recorded and this from plants with both smooth and reticulate seed coats. (b) Species of Chenopodium other than C. album Seed was available from ten additional species and their chromosome numbers were determined (Table 3 and Fig. 1). Of these six are first records for this country and confirm previous counts from abroad, three (C. urbicum L., C. variabile Aell., and C. berlandieri Mog. subsp. zschackei (J. Murr) Zob.) are new determinations and one(C. opulifolium Schrad. ex Koch & Ziz 2n = 54) differs from the previous record from Germany (Wulff, 1936) Of 2n — 36; Watsonia 5 (3), 1962. CHROMOSOME NUMBERS OF CHENOPODIUM 121 4. DISCUSSION The results of this investigation suggest with reasonable certainty that C. album L. (incl. C. reticulatum Acll.) exists, at least in this country, and probably elsewhere, only as a hexaploid (2n = 54). Attempts however should be made to reconcile this result with the previous contradictory reports in the literature, since it was Kjellmark (1934) who alone obtained the hexaploid count. One possible explanation put forward by Aellen & Just (1943) to account for these divergent counts was that of polysomaty. This phenomenon has been reported extensively from the Chenopodiaceae and in particular from Chenopodium, (Wulff, 1936; Lorz, 1937; Maude, 1940; Witte, 1947). However, hexaploid cells are unlikely to have arisen as a consequence of simple polysomaty (only tetraploid and octoploid cells would be expected) and it can be assumed that any hexaploid counts reported are probably indicative of a true diploid number and are not of polysomatic origin. A much more feasible explanation is that these discrepancies are errors arising from the taxonomic misidentification of the original material used, mistakes which are easily made in this critical genus. For example, out of five samples of seed sent as C. album L. from four independent sources in Scandinavia, four were suspected on seed characters to be C. viride L. Two chromosome numbers were recorded, a diploid (2n = 18) from the material suspected to be C. viride and a hexaploid (2n = 54) from the single specimen of C. album. It is suggested that similar confusion between these two species might easily explain the previously recorded reports of 2n = 18 for C. album of Winge (1917) and Love & Love (1944) from Denmark and Sweden respectively : their material was probably C. viride L. (C. suecicum J. Murr) which is notably abundant in Scandinavia. From this country also Maude (1940) reports a diploid number of 2” = 18 for C. album which may similarly refer to C. viride, known to occur in Merton Park. Certainly material of C. album from that locality collected personally in 1954 gave the hexaploid count (see Table 1). The record of 2n = 36 from the U.S.A. attributed to C. album (Cooper, 1935; Witte, 1947) may also be a consequence of misidentification, this time with C. berlandieri Moa. subsp. zschackei (J. Murr) Zob. (2n = 36). There is ample evidence to show that this last species has been long confused with C. album in the U.S.A. (cf. Wahl, 1952) and data from the Kew herbarium, for example, show that, out of 23 specimens from North America initially determined as C. album, Aellen redetermines 18 of them as C. berlandieri subsp. zschackei. On the basis of these observations it is suggested that taxonomic misidentification is largely responsible for the previous reports of varying chromosome numbers for C. album, and whilst this conclusion must remain tentative until further counts have been made, the evidence is sufficiently strong to suggest that in Britain, at least, C. album L. (incl. C. reti- culatum Aell.) exists only as the hexaploid 2n = 54. REFERENCES AELLEN, P. & Just, TH. (1943). Key and synopsis of the American species of Chenopodium. Amer. Midl. Nat., 30, 47. BHARGAVA, H. (1936). The life-history of Chenopodium album. L. Proc. Indian Acad. Sci., B. 4, 179. Cote, M. J. (1957). Variation and interspecific relationships of Chenopodium album L. in Britain. Ph.D. Thesis, University of Southampton. Cote, M. J. (1951). Interspecific relationships and intraspecific variation in C. album L. in Britain. I. The taxonomic delimitation of the species. Watsonia, 5, 47. Concer, A. & FAIRCHILD, L. (1953). A quickfreeze method for making smear slides permanent. Stain. Téch., 28, 6. Cooper, G. C. (1935) Microsporogenesis in Chenopodiaceae. Bot. Gaz., 97, 169. Watsonia 5 (3), 1962. 2 M. J. COLE KAWATANI, T. & OHNO, T. (1950). Chromosome numbers of the genus Chenopodium. Jap. J. Bot., 25, 177. KJELLMARK, S. (1934). Einige neue Chromosomenzahlen in der Familie Chenopodiaceae. Bot. Notis., 1-2, 136. Lorz, A. P. (1937). Cytological investigations on five Chenopodiaceous genera with special emphasis on chromosome morphology and somatic doubling in Spinacia. Cytologia, 8, 241. Love, A. & D. (1944). Cytotaxonomical studies on Boreal plants. Arch. fiir Bot., 31, 1. MaupeE, P. F. (1940). Chromosome numbers in some British plants. New Phytol., 39, 17. Omara, J. C. (1948). Acetic acid methods for Chromosome studies at prophase and metaphase in meri- stems. Stain Tech., 23, 201. WAHL, H. (1952). A preliminary study of the genus Chenopodium. Bartonia, 27, 1. WINGE, O. (1917). The Chromosomes : their number and general importance. C.R. Trav. Labor. Carlsberg, 13131: Witte, M. B. (1947). A comparative cytological study of three species of the Chenopodiaceae. Bull. Torrey Bot. Cl., 74, 6, 443. Wu rr, W. D. (1936). Die Polysomatie der Chenopodiaceen. Planta, 26, 275. Watsonia 5 (3), 1962. STUDIES ON RANUNCULUS L. SUBGENUS BATRACHIUM (DC.) A. GRAY I. CHROMOSOME NUMBERS By Co DP K:. Cook Botany Department, University of Liverpool It has been thought worth while to publish a list of chromosome numbers even though work is still in progress. It is hoped to publish a taxonomic account later but in order to avoid undue confusion the references to the original descriptions are cited. A complete set of herbarium specimens of all the plants counted is deposited in the herbarium of the Botany School, Cambridge, England. An incomplete set is in the Botanische Staatssamm- lung, Miinchen, Germany. The specimens in Miinchen are suffixed by M. The collection of microscopical preparations will, for the time being, be kept by myself. All counts are cited as somatic numbers. When localities alone are cited it means that these are the origins of material that I have counted. R. HEDERACEUS L., Sp. Pl., 556 (1753). 2n = 16. Langlet (1927), Bocher (1938). The Lizard, Cornwall, M3; Sellings, Kent; an unknown locality in Portugal; Eifel, Wirftal, Priim, Germany, M. An artificial auto- tetraploid (2n = 32) has been induced from the Lizard populationfusing colchicine treat- ments. There is no evidence that this tetraploid occurs in nature. R. OMIOPHYLLUS Ten., Fl. Nap., 4: 338 (? 1830). (R. lenormandii F. W. Schultz). 2n = 16. 6km N. of Polizzi Generosa, Madonie Mts., Sicily, M. 2n = 32. Larter (1932). Two Bridges, Dartmoor, Devon, M. R. TRIPARTITUS DC., Icon. Pl. Gall. Rar. 1, 15, tab. 49 (1808). (R. lutarius (Revel) Bouvet). 2n = 48. (Fig. la). The Lizard, Cornwall, M; Arne, Dorset. The count 2n = 32, Cook (1959), was an error. R. OLOLEUCOS Lloyd, Fl. Loire-Inf., 3 (1844). 2n = 16. (Fig. 1b). Maarne, 20 km S.E. of Utrecht, Netherlands, M. R. BAUDOTII Godr., Mem. Soc. Roy. Nancy 1839, 21, fig. 4 (1840). 2n = 32. Bocher (1938), Christiansen in Sgrensen (1955). Brading, Isle of Wight, M; Zicksee, St. Andrat-Frauenkirchen, S. E. Austria, M. R. PELTATUS Schrank, Baier. Fl., 2: 103 (1789). 2n = 32. Between Erlangen and Dechsendorf, Niirnberg, Germany, M. Herbarium specimens of plants collected in Denmark and counted by K. Larsen have been seen in the herbarium of the Universitetets Botaniske Museum, Copenhagen, Denmark. The Danish localities are: Hestehaven (18.V.1952); Bastrup (13.V.1953); Plejelt, N. of Fredensborg (17.V.1953); Vejenbred (25.V.1953); Baesbakke, Fyns Hoved (7.VI.1953). i. AQUATILIS L., Sp. Pl. 556 (1753). 2n = 48. Christiansen in Sgrensen (1955). Castle Donington, Leicestershire, M3; Stretham, Cambridgeshire, M; Tuddenham, Suffolk; Pocking, Starnberg, S. Germany, M; Erling, Herrsching, S. Germany, M. R. aquatilis and R. peltatus are frequently confused. The following counts of 2n = 32 could refer to either. Larter (1932), Bocher (1938), Svaélov in Ehrenberg (1945), Delay (1947), Turala in Skalinska (1959). 123 Watsonia 5 (3), 1962. 124 C.D: Kock R. TRICHOPHYLLUS Chaix in Vill., Hist. Pl. Dauph. 1, 335 (1786). 2n = 16. Langlet (1927). 2n = 32. Reese (1957). Wicken Fen, Cambridgeshire, M; Woodwalton Fen, Hunting- donshire; Bretby, Derbyshire; the river Wiirm, Obermenzing, Miinchen, S. Germany, M; Ascholding, S. Germany, M; Oberstdorf, S. Germany, M. (d) Fig. 1. Root-tip mitoses in Ranunculus spp., x c.3509. (a) R. tripartitus, Lizard, Cornwall, 2n = 48. (b) R. ololeucos, Maarne, Holland, 2n = 16. (c) R. rionii, Prater, Wien, Austria, 2n = 16. (d) R. sphaero- spermus, between Qurna and Madina, Iraq, 2” = 16. Watsonia 5 (3), 1962. CHROMOSOME NUMBERS OF RANUNCULUS SUBGENUS BATRACHIUM _ 125 R. TRICHOPHYLLUS subsp. LUTULENTUS (Perrier & Songeon) Vierhapper, Abh. Zool.-Bot. Ges. Wien, 16: 109 (1935); including R. confervoides Fr. 2n = 32. Mattick in Tischler (1950), A. & D. Léve (1956), Jorgensen, Sorensen & Westergaard (1958). Funtensee, Berchtesgadener Alpen, S. Germany, M; Geissalpsee, Allgauer Alpen, S. Germany, M. R. RIONI Lagger, Flora, 31 : 49 (1848). 2n = 16. (Fig. Ic). Pond near the ‘ Lusthaus’, Prater, Wien, Austria, M (Dupla- Exsiccata ex hb. Mus. Wien, Ch. Cook & A. Patzak, 24.V1.1960); Himberg, S.E. Austria, M; Madgil, Basra, Iraq. The original material of the Iraq collection is in the Naturhis- torisches Museum, Wien, Austria, (K. H. Rechinger fil., No. 15808, 16.11.1957). R. SPHAEROSPERMUS Boiss. & Blanche in Boiss., Diag. ser. 2, 3 (5): 6 (1856); not R. aquatilis subsp. sphaerospermus sensu Clapham in Clapham, Tutin & Warburg (1952) which belongs in the “ pseudofluitans’”’ group. 2n = 16. (Fig. Id). Between Qurna and Madina, Iraq; the original material of this collection is in the Naturhistorisches Museum, Wien, Austria, (K. H. Rechinger fil., No. 8477, 17.111.1957). R. CIRCINATUS Sibth., Fl. Oxon., 175 (1794). 2n = 16. Scheerer (1939), Turala in Skalinska (1959). Cheddar, Somerset; Hemington, Leicestershire; Wicken Fen, Cambridgeshire; Oberlaus, Glonn, S. Germany; Nymphen- burg, Miinchen, S. Germany, M. R. FLUITANS Lam., Fl. Fr., 3: 184 (1778). 2n = 16. Quorndon, Derbyshire; Donaueschingen, Wirttemberg, Germany, M3. 2n = 24. Schleissheimer Kanal, Dachau, S. Germany, M, (sterile). 2n = 32. Temple Bridge, Suffolk, M; Wicken Fen, Cambridgeshire; Whatstandwell, Derbyshire. THE ‘ PSEUDOFLUITANS ” COMPLEX. This complex contains many topo- and ecodemes some of which are genetically distinct. They are more or less confined to flowing water and range from types that look like R. fluitans to types that look like R. peltatus. As the whole group is imperfectly known no attempt has been made to name the various genodemes. Some are simple hybrids involving R. fluitans and R. trichophyllus, R. peltatus, R. aquatilis and, perhaps, R. circinatus. There is evidence that some are allopolyploids from these hybrids and some morpholo- gically similar types have almost certainly arisen in different localities. 2n = 24. The River Wiirm, Obermenzing, Miinchen, Germany, M, (sterile). 2n = 32. River Elz, Prechtal, Pfalz, Germany, M, (sterile); Sigmaringen, Wurttem- berg, Germany, M, (fertile). 2n = 40. Peakirk, Northamptonshire; Monsal Dale, Derbyshire, (both sterile). 2n = 48. Greywell, Hampshire; Puddletown, Dorset; Foulmere, Cambridgeshire; Oberlaus, Glonn, S. Germany, M, (all fertile). REFERENCES BocHeErR, T. W. (1938). Cytological studies in the genus Ranunculus. Dansk Bot. Archiv, 9, 1-33. Bocuer, T. W. & LarsEN, K (1950). Chromosome numbers of some arctic and boreal flowering plants. Medd. om Gronl., 147, No. 6, 1-32. CLAPHAM, A. R., TuTin, T. G. & WARBURG, E. F. (1952). Flora of the British Isles. Cambridge. Cook, C. D. K. (1959). The Ranunculus tripartitus complex. Proc. Bot. Soc. Brit. Isles, 3 (3), 326. Deay, C. (1947). Recherches sur la structure des noyaux quiescents chez les phanérogames. Rev. Cyrol. & Cytophysiol. Vég., 9, 169-223, 10, 103-229. EHRENBERG, L. (1945). Kromosomtalen hos nagra K4rlvaxter. Botaniska Notiser, 436. JORGENSEN, C. A., SORENSEN, T. & WESTERGAARD, M. (1958). The flowering plants of Greenland. A taxo- nomical and cytological survey. Biol. Skrifter Dan. Selsk., 9, No. 4, 1-172. LANGLET, O. (1927). Beitrage zur Zytologie der Ranunculazeen. Svensk Bot. Tidskr., 21, 1-17. Watsonia 5 (3), 1962. 126 C. D. K. COOK LARTER, L. N. H. (1932). Chromosome variations and behaviour in Ranunculus. J. Genet., 26, 255-285. Love, A & D. (1956). Cytotaxonomical conspectus of the Icelandic Flora. Acta Hort. Gotob., 20, 65-295. REESE, G. (1957). Uber die Polyploidiespektren in der nordsaharischen Wiistenflora. Flora, 144, 598-634. SCHEERER, H. (1939). Chromosomenzahlen aus der schleswig-holsteinischen Flora. Planta, 29, 637-648. SORENSEN, T. (1955). Hybriden Ranunculus baudotii x R. radians. Bot. Tidsskr., 52 113-124. TISCHLER, G. (1950). Die Chromosomenzahlen der Gefdsspflanzen Mitteleuropas. The Hague. TurALA, K., in SKALINSKA, M. ef al. (1959). Further studies in chromosome numbers of Polish Angiosperms (Dicotyledons). Acta Soc. Bot. Pol., 28, 491-493. Watsonia 5 (3), 1962. STUDIES IN THE BRITISH EPIPACTIS V. EPIPACTIS LEPTOCHILA; WITH SOME NOTES ON £. DUNENSIS AND £. MUELLERI By DONALD P. YOUNG ABSTRACT Records for Epipactis leptochila (Godf.) Godf., sensu stricto, are enumerated. In England it is confined to calcareous areas in the south, where it is frequent on the Chilterns and Cotswolds. It also occurs in France, Germany, Denmark and Switzerland. The plant is strongly calcicolous, and its usual habitat is in beech- woods with an open ground-flora and in heavy shade. Some variation in floral morphology occurs in this species. Godfery’s type, in particular, has the column and lip much more elongated than in the bulk of specimens. A freak plant from Germany had a column like that of E. muelleri, implying that such a column- form can arise by mutation. An achlorophyllose plant has also been seen. The equation of E media sensu Bab. non Fr. with E. leptochila was based on false premises. E. cleistogama C. Thomas may be a form or state of the last species. E. dunensis (TY. & T. A. Steph.) Godf. is only known from five British vice-counties; Contunental records cannot be substantiated. E. muelleri Godf. is known from France, Switzerland, the Benelux countries and Germany. It is a plant of lightly shaded habitats, and is closely related to E. dunensis. As originally conceived (Godfery 1921b; Stephenson & Stephenson 1921b), Epipactis leptochila (Godf.) Godf. included all the then known self-fertilised forms except E. muelleri Godf., which was distinguished by a different fertilisation mechanism. The separation from it of E. dunensis (Godfery 1926) and later of E. vectensis (Brooke & Rose 1940; now called E. phyllanthes G.E.Sm.) has left the residual E. leptochila as a much better-defined entity. Now that the somewhat confused records for these species have been sorted out, it is possible to reappraise the distribution, ecology, and morphological variation of E. leptochila, sensu stricto. DISTRIBUTION In Britain, E. /eptochila is frequent on the escarpments of the Chilterns and Cotswolds. Elsewhere it is uncommon, in scattered localities along the chalk range from Wiltshire to Kent; on the Carboniferous Limestone of the Wye Valley and Cheddar; and in one or two localities in Devon. On the Continent, it is in rather widely separated localities from Denmark to the Swiss Jura; its distribution is very incompletely known, and its eastern limit is uncertain (Fig. 1). I have seen the colonies marked ! in situ. British Records V.c. 3. S. DEVON : Dunsford and Cornwood (Martin & Fraser 1939) (TOR; the specimen from the first locality is now too worm-eaten to confirm, but the second is correct). 6. N. SOMERSET : ash-whitebeam scrub, Cheddar Gorge, 1957, J. T. H. Knight (K). 8. S. WILTS.: beech-wood on chalk, Winterslow, 1957, A. Roseweir! (K). 11. S. HANTS.: beech-woods on chalk, Hursley, 1954 (K, herb. Young), and W. Tytherley, 1956, A. Roseweir! (K). These are the first two authentic records for Hants; all previous ones are referable to E. phyllanthes. 127 Watsonia 5 (3), 1962. 128 13. 15: 17. 20. 22 D3: 24. 33: DONALD P. YOUNG Fig. 1. Known distribution of Epipactis leptochila. W. SUSSEX: wood on chalk, Treyford, 1960 Miss D. W. Fawdry, conf. V. S. Summerhayes who had found a doubtful specimen at the same spot 12 years previously. Earlier records refer to E. phyllanthes. E. KENT : ash-hazel coppices on chalk, Kingston! (Brooke & Rose 1940) and Womenswold, 1955, B. J. Brooke, R. Gorer & F. Rose. SURREY : beech-woods, W. Horsley (type locality), and E. Horsley, F. Rose; beech-yew wood, Woldingham, long known and still extant; Kingswood Valley, 1924 (Salmon 1931), not seen recently. All are on chalk. HERTS.: beech-wood on chalk, Tring, 1943, H. W. Pugsley (BM), 1953! (K, herb. Young). BERKS.: frequent on chalk between Streatley and Pangbourne. Bisham, 1925, C. B. Tahourdin (SLBI). OXFORD : frequent on the chalk. BUCKS.: frequent on chalk above the Thames between Hambleden and Marlow, and for about 5 miles northwards; thence in scattered localities as far as Chequers. E. GLOS.: frequent on the Cotswold escarpment and valleys from Stroud to Cheltenham, thence apparently less frequent as far as the Worcs. border, and eastwards as far as Guiting (cf. Riddelsdell, Hedley, & Price 1948). Watsonia 5 (3), 1962. EPIPACTIS LEPTOCHILA 129 34. W. GLOS.: frequent on the Cotswolds from Stroud to Wotton-under-Edge. Madgett (Wye Valley), 1933, J. E. Lousley (herb. Lousley). 35. MONMOUTH : between Tintern and Wyndcliff, 1920, C. E. Salmon (BM). Records for the following vice-counties are either erroneous or based on insufficient evidence, or else refer to segregates now made separate species: V.c. 9. Three records are given in the Dorset Flora (Good 1948). Of these, one (Badbury) is not supported by specimens, and the other two prove to be of E. helieborine (specimens in herb. Good). V.c. 10, 12: Published records all refer to E. phyllanthes. V.c. 14: Wolley-Dod’s tentative record from Saxonbury Hill (cf. Wolley-Dod 1937) has never been confirmed, and it appears that E. purpurata was mistaken for it. The locality is an oak-wood on sandstone, and seems now to be an unlikely habitat for E. leptochila. V.c. 30: An erroneous record has been withdrawn (Dony 1953). V.c. 40: The record by Godfery (1919) from Bomere Pool was based on a single admittedly poor specimen, said to be of ‘the only Epipactis in sight’ there. This specimen is not now extant, nor has the record ever been confirmed. Independent searches of the woods around Bomere Pool by Miss E. P. A. Jones and myself brought to light only E. helleborine, in a somewhat small-leaved form that might simulate E. leptochila if in poor condition. A specimen of E. helleborine collected there by Leighton in 1835 is in herb. BM. These woods are on sandstone, and vary from very dry to marshy, but seem unlikely to support E. leptochila. V.c. 51: The record refers to E. phyllanthes. V.c. 52, 59, and 60: The records refer to E. dunensis. V.c. 66 in the Comital Flora was a misprint for 60, but has been repeated by copyists. European Records FRANCE _ Seine-Mme.: beech-forests on chalk in two places, Forét d’Eu, 1959, B.S.B.I. field meeting! (herb. Young). GERMANY Lr. Saxony : on calcareous soils in the neighbourhood of Stadtoldendorf and Hildesheim (Krésche 1929); beech-forests on Jurassic Limestone, S.E. of Hildesheim and Osterwald W. of Elze! on chalk, Sieben Berge E. of Alfeld! on Muschelkalk, Holzberg S. of Stadtolden- dorf! (herb. Young). Wirttemberg : beech-forests on Jurassic Limestone around Urach, 1953, F. Rose! (Herb. Young). Sachsen-Anhalt : “in sylvis mont. calc. umbrosis,” Alte Stolberg, 1885, Vocke (UPS). DENMARK On chalk, Mons Klint and near Hesnes (Falster) (Young 1953)... SWITZERLAND Berne : fir-beech forests on Jurassic Limestone near Tavannes (Young & Renz 1958) (herb. Young, herb. Renz) AUSTRIA Ost-Tirol : mixed woods on limestone, Matrei, 1961, C. D. Sayers (K). The record from Bomere Pool raises a point of synonymy. Bomere Pool is the /ocus classicus for Epipactis media, sensu Babington, and on this basis Godfery (1919; 1933, p. 75) regarded E. media Bab. non Fr. as a synonym of E. leptochila, at least in part. Since E. leptochila does not in fact grow there, and LE. helleborine certainly does, this synonymy cannot be upheld. The question of what Babington intended by E. media has been discussed by Stephenson and Stephenson (1921a), and there is little to add to their remarks, except to note Babington’s revealing statement (1852), ‘ The true £. Jatifolia is a much less frequent plant in this country than my E. media, which is often mistaken for it by British botanists.’ ECOLOGY The most striking feature that emerges is that the plant is strongly calcicolous —- much more strictly so than any other European Epipactis. Every station for which geological Watsonia 5 (3), 1962. 130 DONALD P. YOUNG data is available is on calcareous rock, and this comprises the great bulk of records. It is also noticeable that it occurs particularly on steep slopes where the calcareous rock is free from overlying drift. Conversely, no station has been recorded as on neutral or acid soils (the Devonshire records ought to be reinvestigated with this in mind). Tree cover is usually beech, occasionally admixed with yew or fir. At Cheddar it is under heavily-shading scrub, and in Kent it is in ash-hazel coppices. The species usually affects heavy shade, and soon disappears if the trees are cleared. It is normally associated with a low and very open ground-flora. In a typical locality on the Cotswolds, the principal associated species were Fragaria vesca, Sanicula, Viola riviniana, Bromus ramosus, Hedera, Hieracium sp., Epipactis helleborine, Cephalanthera damasonium, Neottia nidus-avis, and Pyrola minor. In this country, E. leptochila can be found in good quantity in most stations, and hundreds of plants may occur in a square km. of woodland. The reverse seems to be the case in Germany, where it is very unusual to find more than half-a-dozen plants together. This seems to be the consequence of the rather different nature of central European beech forests, where beech regenerates freely. They are subjected to regular attention of foresters engaged in trimming, thinning, and felling. This leads to a cycle of disturbance of the shade cover, alternating with the growth of a shrub layer of beech (and ash) seedlings, Rubus idaeus, etc. The places most productive of Epipactis spp. are forests of young trees which have not reached the reproductive stage, and which shelter a ground flora which is quite open or almost absent. In suitable places in the Schwabische Alb around Urach, E. leptochila is associated with E. helleborine, E. atrorubens, Cardamine bulbifera, Lathyrus vernus, Cephalanthera damasonium, C. rubra, and Epipogium aphyllum. At the other end of its range, at Mons Klint in Denmark, E. Jeptochila is again associated with E. atrorubens, but in Britain their ecological requirements have diverged and the two are confined to quite different areas. Cephalanthera rubra is, in all its known stations in this country, in areas where E. leptochila is frequent. Epipogium is also associated with E. leptochila in the Chilterns, and Cardamine bulbifera is frequent in the same area, but the association is not invariable because neither of these is strictly calcicole. E. leptochila is one of the earliest species in the genus to flower. In a normal English summer it starts in mid-July and continues for 3-4 weeks, thus being about 3 weeks ahead of E. helleborine. On the Continent it is later, commencing in early August, still ahead of E. helleborine but later than E. atrorubens. VARIATION Like other species of the genus, E. leptochila shows considerable and sometimes puzzling variability. Variation in size and number of flowers is dependent on the age and state of nourishment of the plant, and has no other significance. There is a moderate range of leaf-size, from 6:0 x 2:5 to 10-0 x 4:5 cm. for the largest leaf of mature plants; the length/breadth ratio is fairly constant, but narrower leaves (to 8-5 x 1:5 cm.) do occur Fig. 2. Labellum and column of Epipactis leptochila (v.c. 33 : Cranham Woods). xX 5. Watsonia 5 (3), 1962. EPIPACTIS LEPTOCHILA 131 occasionally. The floral architecture shows some variability, mainly in the breadth of the labellum and the shape of the column. Perhaps the only definable characters of the species (apart from absence of rostellum) are that the epichile is acute and at least as long as broad, and that the column has a well-marked clinandrium. Characteristic, but not invariable, features are that the epichile is usually very noticeably longer than broad, acuminate, and prolonged at the base into broad wings decurrent on the hypochile, giving the epichile a sagittate shape; and that the prominence on the column that bears the anther is prolonged forward, sometimes so much so as to form a short broad filament (Fig. 2). It is unfortunate that Godfery’s type, represented by a population at West Horsley, is an unusual and extreme variant having a long narrow epichile and a longly stipitate anther, well shown in his illustrations (Godfery 1920, pl. 553 fig. 1, and 1933, pl. K fig. 1). E. leptochila var. cordata Brooke (1950), which perhaps represents the other extreme, is described as having a cordate and broadly acuminate epichile. Unfortunately the colony on which this was based has disappeared, and no specimens or illustrations survive. An elaborate catalogue of varia- tions has been given by Krosche (1930, 1932, 1936). The value of this compilation is more than doubtful, since it includes, uncritically, both normal and teratological variants, and is overshadowed by attempts to force them into a taxonomic framework. The West Horsley population is the only one known to me where the distinct characters of the population are obvious, but genotypic differentiation between populations is to be expected and could probably be demonstrated biometrically. Two abnormal plants seen in recent years deserve mention. The first was from a scattered population in the Sieben Berge (Germany) showing rather considerable variation in floral architecture. In one plant the column, in every flower, had no clinandrium, so that the pollinia overhung the stigma exactly as in E. muelleri. This variant does not appear anywhere in Krésche’s papers, but on his arrangement it would fall within E. muelleri — where, from the shape of the labellum, tepal size, and other characters, it certainly did not belong. Assuming that the absence of clinandrium was genotypic — and there is no reason for supposing otherwise — then the occurrence of this form amongst an otherwise normal population implies that it had arisen by mutation. Hence the similar column-shape of E. muelleri could conceivably have arisen from a single mutation, rather than by gradual evolution. The other noteworthy freak was a single achlorophyllose plant found in Hampshire (v.c. 11) by Mr. A. Roseweir in 1954, amongst a large and otherwise normal population. This plant reappeared yearly up to 1957. Despite its deficiency in chlorophyll, and also despite being picked (by persons unknown) on two occasions before reaching anthesis, it kept all the vigour of a normal plant. Its cells were probably not completely without chlorophyll, since the stems and flowers were lemon-yellow and the leaves a pale greenish- yellow. [Coloured photographs of this were taken by Mr. D. E. Kimmins, and a copy has been deposited in the British Museum (Natural History)]. Achlorophyllose plants have been reported in other species of Epipactis and in Cephalanthera. The phenomenon has recently been discussed by Burgeff (1959), who regards it as demonstrating that the plants can live entirely saprophytically with the aid of the symbiotic fungus. No other explanation appears possible, although objections have been raised in the past to this idea; not the least of these is that the roots are often well below the humus layer which supplies saprophytic nourishment. Burgeff suggests that nourishment can be transmitted for some distance through the fungal hyphae; it would be desirable to have more direct evidence on this. The hypochile of E. leptochila usually contains nectar, which is sweetish to the taste but does not seem to attract insects. In other autogamous species the hypochile is always quite dry inside. E. CLEISTOGAMA C. THOMAS This taxon has been distinguished from E. /eptochila by the different fertilisation mechanism : instead of the pollinia falling bodily on to the stigma in the opened flower, Watsonia 5 (3), 1962. 132 DONALD P. YOUNG in E. cleistogama the flowers do not open, and friable pollen is scattered in the bud stage (Thomas 1948). On gross morphology there seems to be no significant difference between the two; the distinctive features ascribed to E. cleistogama, such as robust habit, are within the range of E. leptochila. The plant has, moreover, apparently vanished from its former stations (where E. /eptochila still occurs). If it reappears, it would be desirable to watch marked plants for several seasons to see whether they are constantly cleistogamous. For the present, it seems best to regard it as a form or state of E. leptochila. According to Krodsche (1929, 1932, 1936), all German forms of E. /eptochila shed friable pollen in the bud, but complete cleistogamy has not been noted in German examples. E. DUNENSIS (T. & T. A. STEPH.) GODF. The known distribution of E. dunensis in Britain is as follows : V.c. 52. ANGLESEY : Newborough Warren, well-known. 59. §. LANCS.: Hall Rd. to Southport, well-known. 60. W. LANCS.: Lytham, 1873, E. F. Linton (CGE)—probably other records exist, but plant not seen recently. 68. CHEVIOTLAND : Holy I., 1958, A. J. Smith! (E). 69b. N. LANCS.: Roanhead-Sandscale dunes, 1952, G. Wilson (K). The following records for E. dunensis on the Continent have been published, but in no case do any specimens exist : FRANCE : Coutainville (Manche) (Meslin 1928). BELGIUM : Nieuport-Bains and (?) le Coq-sur-Mer (Godfery 1933, p. 78). GERMANY : coast opposite Usedom (Pomerania) (Godfery 1933, p. 78). I have visited the alleged localities in France and Belgium, and have found no E. dunensis nor ground apparently suitable for it. At Coutainville and also a few miles north of it, there is a little of a dwarf dunal form of E. helleborine. Meslin’s description and figure could well refer to this. The plant he depicts has internodes shorter than the leaves, and a very flexuous stem, which point to E. helleborine rather than to E. dunensis. He speaks of the rostellum being evanescent in the freshly-opened flower; in exposed situations this does in fact occur with E. helleborine. In E. dunensis the rostellum disappears in the early bud stage. (A further diagnostic difference is the usual dull purple suffusion of the flowers of E. helleborine, whereas E. dunensis has pale green tepals and the lip marked with Fig. 3. Dissections of flower of Epipactis muelleri (France : Thorenc). The items are from different flowers, and show minor differences of shape and size. Watsonia 5 (3), 1962. EPIPACTIS LEPTOCHILA 133 rose-pink). Several specimens from coastal localities in Belgium in the herbarium of the Brussels botanic garden (BR), labelled E. dunensis, are either E. helleborine or E. palustris. The Belgian coastal dunes have suffered much in recent years from building and desiccation. Nannfeldt (1946, p. 5 footnote) similarly doubts the Pomeranian record, after examination of much herbarium material from that area. The evidence for the occurrence of E. dunensis on the Continent is thus unsatisfactory, and we may claim it as a British endemic. However, suitable habitats for it do exist on the French and Dutch coasts, and it would be worth searching for there. It is curious that the common dunal species of the Channel and North Sea coasts of the Continent is E. helleborine, whereas in Britain this is extremely rare in dunes — Kenfig is the only such station certainly known to me. E. MUELLERI GODF. This species is not British, but it is appropriate to summarise present knowledge of it in relation to our own species. Detailed descriptions have been given by Godfery (1921a), Zimmermann (1922, as Parapactis epipactoides), and Krodsche (1934), and recently an excellent one, illustrated, by Reichling (1955). Some diagrams of the dissected flower are given here (Fig. 3). Throughout its range it varies but little. Characteristic recognition features are the slender habit and narrow undulate leaves. The column is constantly as described by Godfery and other authors, with virtually no clinandrium, i.e. the anther is attached almost directly above the stigma and the pollinia overhang the latter. The anther has, usually, a hooked empty tip which may be clearly visible in soaked-out herbarium specimens. The labellum only varies to the extent that it may be cordate or rhomboid, but it is always short, broader than long, and obtuse. The flowers are smaller than those of E. leptochila. E. muelleri is very closely related to E. dunensis, which it resembles in vegetative parts and in flower size, structure (except the column), and colour. The roots are thicker (2:24 + 054mm., against 1-48 + 0-28), but this could be the result of different habit Fig. 4. Known distribution of Epipactis muelleri (black circles and shading) and of E. dunensis (open circles) Watsonia 5 (3), 1962. 134 DONALD P. YOUNG (see Part VI, p. 140). The ecological requirements of the two species are somewhat similar, except that E. dunensis is entirely maritime and EF. muvelleri entirely inland. The autecology of E. muelleri has been discussed by Reichling (1955). It is a thermo- phile, affecting open woods and forest borders and clearings, often amongst grass. It is calcicole, but not so strongly so as E. leptochila. The ecology of the latter, which favours cool, heavily shaded situations, is in strong contrast. The authenticated distribution of E. muelleri is as follows, although it must be very incompletely known (see also Fig. 4). FRANCE Pas-de-Calais : Desvres, 1959, B.S.B.J. field meeting! Somme : Cambron, 1959, idem! Seine-Mme.: near Hodeng, 1959, idem! (L, herb. Young); near Bazinval, 1960, M. de Blangermont & J. Liger (Herb. Young). Hte. Sadne : N. of Champlitte, 1959, B. J. Brooke & R. Gorer (Herb. Young). Alpes-Mme.: Thorenc! (K, herb. Young; locus typicus). Pyr. Or.: Bourgmadame, 1926, Sennen (Pl. d’Esp., 5883) (LD). SWITZERLAND Valais : near Vallorbe, 1955, C. Sipkes (L, herb. Young). BELGIUM District Calcaire Moisan, in several places (Young 1958). LUXEMBURG Frequent! (Reichling 1955). HOLLAND Limburg: near Maastricht, 1945, H. W. E. Croockewit (herb. Vermeulen). GERMANY Rheinland : Driburg (Miller 1868, p. 7; the classical description) : Echternachterbriick, 1908, J. Groves (BM); near Miunstereifel, 1927-8, H. Hoppner (Orchid. exsicc. VI, 137 & 76) (LD). Lr. Saxony : Hildesheim region and near Stadtoldenorf (Kr6ésche 1934). Wiirttemberg : Tuttlingen, Wirmlingen (Zimmermann 1922). Thtiringen : near Sondershausen, 1885, F. Heinmann (UPS). Other records are given by Godfery, Zimmermann, Sod, etc., but without seeing specimens or precise descriptions I defer acceptance of them. Further search is needed to determine its distribution more precisely. ACKNOWLEDGMENTS My thanks are offered to Messrs. B. J. Brooke, E. P. Bury, A. J. Smith and especially to Prof. L. Reichling, Mr. A. Roseweir, and Mr. C. Sipkes for their kindness in taking me to see Epipactis colonies; to Dr. F. Rose and Mr. V. S. Summerhayes for helpful dis- cussions; and to several herbarium authorities for allowing me facilities to examine specimens. REFERENCES BABINGTON, C. C. (1852). The new Epipactis. Gard. Chron., 693. BrookE, B. J. (1950). Wild Orchids of Britain. London, p. 122. BROOKE, B. J. & Rose, F. (1940). A new species of British Epipactis. J. Bot., Lond., 78, 81-89. BurGEFF, H. (1959), in C. L. Withner, The Orchids. New York, pp. 388-389. Dony, J. G. (1953). Flora of Bedfordshire. Luton, p. 405. GopFERY, M. J. (1919). ‘ Epipactis media (Fries!)’ Bab. J. Bot., Lond., 57, 80-83. Goprery, M. J. (1920). Epipactis viridiflora Reich. J. Bot., Lond., 58, 33-37 and pl. 553. GopreryY, M. J. (1921a). A new European Epipactis. J. Bot., Lond., 59, 101-106. GoprerY, M. J. (1921b). Epipactis leptochila Godf. J. Bot., Lond., 59, 146-147. GoprFeERY, M. J. (1926). Epipactis dunensis Godf. J. Bot., Lond., 64, 65-68. GoprerY, M. J. (1933). Monograph and Iconograph of the native British Orchidaceae. Oxford. Goop, R. D’O. (1948). A geographical Handbook of the Dorset Flora. Dorchester, p. 217. Watsonia 5 (3), 1962. EPIPACTIS LEPTOCHILA 135 KRrOscHE, E. (1929). Nochmals Epipactis viridifloraauct. (em.) f. auctiflora Krésche. Repert. nov. Spec. Regn. veg., 26, 88-92. KroscuHeE, E. (1930). Beobachtungen an der Gesamtart Epipactis latifolia im braunschweiger Weserlande und bei Hildesheim. Repert. nov. Spec. Regn. veg. 27, 368-379. KroscHe, E. (1932). Erganzungen zu den ‘Beobachtungen an der Gesamtart Epipactis latifolia All.’ Repert. nov. Spec. Regn. veg. 30, 239-245. KroscuHeE, E. (1934). Epipactis latifolia All. B. Muelleri (Godf.), Repert. nov. Spec. Regn. veg. 35, 102. Kroscue, E. (1936). Gliederungstabelle der Epipactis latifolia All. (sensu lat.). Repert. nov. Spec. Regn. veg. 40, 360-362. MartTIN, W. K. & FRASER, G. T. (1939). Flora of Devon. Arbroath, p. 596. MESLIN, R. (1928). Epipactis dunensis Godf. on the French coast. J. Bot., Lond., 66, 217-218. MULLER, H. (1868). Beobachtungen an westfalischen Orchideen. Verh. Naturhist. Ver. preuss. Rheinlande u. Westphalens, 25, 1-62. NANNEELDT, J. A. (1946). Tre f6r Norden nya Epipactis-arter, E. persica Hausskn., E. leptochila (Godf.) Godf., och E. purpurata Sm. Bot. Not., 1946, 1-28. REICHLING, L. (1955). Les Epipactis de la flore luxemburgeoise. Arch. Inst. G. D. Luxemb., Sect. Sci. Nat., (2) 22, 123-145. RIDDELSDELL, H. J., HEDLEY, G. W. & Price, W. R. (1948). Flora of Gloucestershire. Arbroath. pp. 447-448. SaLmon, C. E. (1931). Flora of Surrey. London. p. 599. STEPHENSON, T. & STEPHENSON, T. A. (1921a). Epipactis latifolia in Britain. J. Bot., Lond., 59, 33-39. STEPHENSON, T. & STEPHENSON, T. A. (1921b). Epipactis viridiflora. J. Bot., Lond., 205. Tuomas, C. (1948), in Riddelsdell, Hedley & Price (1948), pp. 612-613 and pl. 42-43. WoLLEY-Dop, A. H. (1937). Flora of Sussex. Hastings. pp. 424, 560. Youna, D. P. (1953). Autogamous Epipactis in Scandinavia. Bot. Not., 1953, 253-270. Youna, D. P. (1958). Le genre Epipactis en Belgique. Bull. Jard. Bot. Etat, Brux., 28, 123-127. Youn, D. P. & RENZ, J. (1958). Epipactis leptochila Godf.—its occurrence in Switzerland and its relation- ship to other Epipactis species. Bauhinia, 1, 151-156. ZIMMERMANN, W. (1922). Parapactis noy. gen., eine tibersehene Orchidaceengattung. Repert. nov Spec. Regn. veg. 18, 283-287. Watsonia 5 (3), 1962. STUDIES IN THE BRITISH EPIPACTIS VI. SOME FURTHER NOTES ON E. PHYLLANTHES By DONALD P. YOUNG ABSTRACT Self-fertilisation of E. phyllanthes occurs by germination of poilen in the clinandrium. The somatic chromosome number is 36. Plants answering to var. pendula have now been found in several places in the south of England, and there is no material dividing-line between this form and var. vectensis. A list of new records is given. The species also occurs in Ireland, and in France and Denmark, and its distribution appears to be Atlantic in pattern. One natural habitat for it is in wet willow-holts on the alluvium of chalk streams. The following notes are supplementary to Part IV of this series (Young 1952). FERTILISATION In Part IV the speculation was advanced that E. phyllanthes might be apomictic, as the process of fertilisation was not visible. Emasculation experiments showed that, if the pollinia were removed before their membrane had ruptured, no seed was set. If, however, the pollinia had become friable (which occurred long before the perianth had opened), it was not possible to remove them completely, and eventually a limited number of seeds matured in the capsule. Later, Dr. O. Hagerup kindly made sections of the column of an unexpanded flower, and demonstrated that the pollen was germinating in the clinandrium, on the upper edge of the stigma. From there pollen-tubes were passing down the stylar canal into the ovary, and embryos were already formed. Although union of gametes was not actually observed, he had no doubts that self-fertilisation was occurring in this case (Fig. 1). Fig. 1. Longitudinal section of column from unopened flower of Epipactis phyllanthes (v.c. 24 : Dorney), showing germination of pollen; x 35. (Preparation and drawing by O. Hagerup). Dr. Hagerup also made a chromosome count on Danish E. phyllanthes, and found 2n = 36*. Most other European species normally have 2n = 40. *This was first reported by Hagerup (1947) under the name E. leptochila, which has been quoted by other authors. To the best of my knowledge, the chromosomes of E. leptochila sensu stricto have not yet been counted. 136 Watsonia 5 (3), 1962. FURTHER NOTES ON EPIPACTIS PHYLLANTHES 137 VARIATION Plants which would come under var. pendula, i.e. with hypochile equalling the epichile, have now been found at E. Tisted (v.c. 12), Thakeham (v.c. 13), Wilstone (v.c. 20), and Harpsden (v.c. 23). These are within the area in which var. vectensis (hypochile shorter than epichile) and varieties with a degenerate lip are general. Other plants, e.g. from Thriplow (v.c. 29), possess a broad cordate epichile like var. pendula, but a smaller hypo- chile. There is thus no material dividing-line, either geographical or morphological, between var. pendula and var. vectensis. The names are retained here for convenience of description. Lip variation in colonies recently discovered has been on the same lines as previously described. As observed in Part IV, a few colonies contain more than one lip-form, but in the majority the lip-shape is constant throughout the population. On the other hand, variation between colonies is very great. The distribution of forms with a degenerate lip- shape is centred on the southern counties, but within this area the different forms seem to be distributed quite at random. In parts of Hampshire and Wiltshire, particularly, colonies are thickly scattered within a few miles of one another, yet scarcely two are identical in flower-form. Their evident isolation suggests that very little long-distance dispersal by seed takes place. Probably most of these small populations have been reduced to a single biotype. They may be the relics of a very variable population extending over the southern counties when they were mainly forest-covered. Unfortunately, the lip-shape in colonies outside Britain has not been studied very fully, on account of scanty and poor material. However, in all specimens where it could be discerned, the lip was perfectly differentiated. DISTRIBUTION E. phyllanthes has been found on the Continent and also in Ireland. Its distribution appears to be Atlantic in character, running from the Pyrenees up the west coast of France, through England and Ireland, and reappearing in Denmark (Fig. 2). It should be looked Fig. 2. Known distribution of Epipactis phyllanthes. Watsonia 5 (3), 1962. 138 DONALD P. YOUNG for in northern France, where it is unrecorded but likely to occur. In Parts II and IV, reference was made to an allied Scandinavian plant tentatively included under E. persica (Sod) Hausskn. ex Nannf. Further investigations by the writer (Young 1953) showed that this was best regarded as a separate species, E. confusa D. P. Young. This is a more slender plant, with narrower leaves, smaller and less pendent ovaries, smaller seeds, thinner and less copious roots, and 40 somatic chromosomes. It is related to E. persica (sensu stricto), which is an Asiatic species extending from Turkey to Afghanistan. The ranges of E. phyllanthes and EF. confusa just overlap in Denmark. FRANCE DENMARK European Records for E. phyllanthes Maine-et-Loire : recorded by Jovet (1957); 1 have not seen the specimen. Vendée : pine forest on sand-dunes, Forét d’Olonne, 1880, C. Pontarlier (LD, UPS), 1953! (herb. Young) (lip small, perfect; var. pendula). Charente-Mme.: pine forest, Forét de Foulerot near St. Georges d’Oléron, 1891, N. Reau (P); Forét d’Arvert, 1889, A. Guillon (P). Basses-Pyr.: Bayonne, 1836, Grenier (P); pine plantations on banks of R. Adour, Boucau and Anglet, 1883, Blanchet (MANCH, P) (lip perfect, rather narrow). Hte.-Garonne : wood at mouth of the Ariége, Toulouse, 1850, Timbal-Lagrave (P). Pyr. Or.: Olette, 1853, Loret (P). Jutland, Falster (var. pendula; Young 1953). New British Records The following are supplementary to the list given in Part IV. Some have already been published in “‘ Plant Records,” but references are only given to where substantial further information can be found. Records are recent unless a date is quoted. The lip-form is indicated as follows : (a) var. phyllanthes, (b) var. degenera, (c) vat. vectensis, (d) var. pendula. I have seen the colonies marked ! in situ. VEC): 8. ite 12: 13: 16. Le 20. S. SOM.: roadside near Taunton, 1891, C. Bailey (MANCH). S. WILTS.: beech plantations, Boyton (c; epichile broad, reflexed), B. M. Stratton (K, herb. Young); edge of beech-wood, Erlestoke (c), W.O. Cobbett!; under beeches, Harnham (6b), D.E. Coombe!; beech-belts, Berwick St. James (c), Steeple Langford (a), and Durnford (c), A. Roseweir; beech-belt, Winterbourne (c, epichile only just longer than hypochile), J. Hemsley & A. Roseweir; plantation, Stapleford, 1876, E. S. Marshall (K); beech-yew belt, Stapleford (c), A. Roseweir; beech-belt, Coombe Bissett, Miss H. Lamb (K). S. HANTS.: willow-holts, Nursling (c) (several hundred plants over a wide area; K, herb. Young), near Swaythling (5) (K), and Bossington (6) (K), and plantation on river mud, Eling (c) (Herb. Young), P. Bowman!; plantation on alluvium of Itchen, Twyford (6, differentiation almost perfect), A. Roseweir!; canal bank amongst Petasites, Otterbourne (6), Mrs. J. Goater!; beech-wood, Bishop’s Waltham (5) (K), and beech-belts, Little Somborne (6), W. Tytherley (6), and Broughton (5) (K), A. Roseweir. N. HANTS.: beech-yew belt, Crawley (c) (K), A. Roseweir!; some miles from the previously- known Crawley locality, where P. Bowman found var. vectensis occurs as well as var. degenera!; beech-woods, E. Tisted (d) and Cranbourne, beech-clump, Hurstbourne Priors, and beech-belt, Overton (5), A. Roseweir; in adjoining beech-belts, Nether Wallop and Abbott’s Ann (a) (K), Miss D. Stevens!; hedgebank, Conford (a), Mrs. E. Briggs & Mrs. Missen!. W. SUSSEX : copse beside lake, Arundel (56), Miss B. M. C. Morgan & Miss D. Powell (herb. Young); under oaks on sandy soil, W. Chiltington (a), Miss D. A. P. Long & Mrs. Thatcher; shady roadside, on chalky grit (over sand), Thakeham (d), J. T. H. Knight. W. KENT : plantation belt, Eynsford (6), Mrs. Denton! (K, S, UPS, herb. Young). SURREY : Mr. Spicer’s park near the railroad, Esher, n.d., H. C. Watson (K); this would be on the bank of the R. Mole. HERTS.: amongst grass in orchard, Chorley Wood, R. F. Turney! (K); edge of wood on bank of stream, Wilstone (d), R. I. Sworder (herb. Young). Watsonia 5 (3), 1962. FURTHER NOTES ON EPIPACTIS PHYLLANTHES 139 23. OXON.: plantation beside tributary of Thames, Crowmarsh (c), 1925, R. R. Hutchinson (CYN), 1956! (K); beech-wood, Harpsden (d), G. J. Munday! (BR, K, herb. Young); beech-wood, Bix, D. J. Tennant; beech-wood, Rotherfield Greys (b), V. .S. Summerhayes (K). 24. BUCKS.: beech-wood, Chalfont St. Giles, R. F. Turney! (K); beech-wood, Great Marlow, F. Rose. 29. CAMBS.: mixed wood on bank of stream, Thriplow (c; epichile broad, cordate), Mrs. G. Crompton (CGE, herb. Young). 48. MERION.: sand-dunes, amongst Salix repens, Morfa Dyffryn (d) (Benoit 1959; K, herb. Young). 51. FLINT.: now known to be in a number of places around Rhyd-y-Mwyn and Pant-y-Mwyn in Mold R.D., under willows, beech, and ash (d). 55. LEICS.: ‘ Leicestershire,’ n.d., R. M. Norman in herb. Syme (BM). 60. W. LANCS.: sandhills between St. Anne’s-on-Sea and Fairhaven (d), 1890 and 1902, and waste ground, St. Anne’s, 1901, C. Bailey (MANCH). Giege Sane, YORK = S. Cave (). (Young 1955; K): H20. WICKLOW : dunes, Brittas Bay (?c) (Webb 1953, Sipkes 1954; TCD, herb. Young). H33. FERMANAGH : damp mossy woods on limestone, Lurgan River glen, 1948, R. D. Meikle & J. McK. Moon (K). H40. DERRY : beech-wood on sandy soil, Castledawson (d), D. McClintock & J. McK. Moon (herb. Young). ECOLOGY It has become clear that one natural habitat for E. phyllanthes is in willow-holts on the calcareous alluvium of chalk streams. There are numerous such habitats along the rivers Test and Itchen, even on the tidal reaches. The plant also grows in equivalent habitats such as plantations on stream and river banks. It seems to have been overlooked in such places until recently; their herbaceous flora is usually coarse and botanically uninviting, but E. helleborine sometimes accompanies E. phyllanthes in them. These orchids are quite amenable to periodical inundation, but they avoid hollows where the ground is permanently waterlogged, keeping to ridges and banks where the drainage is better. Most of the recently reported woodland stations have again been in small copses, tree-belts and the like. When the plant has been found in extensive woods it has usually been near the margins. This behaviour has already been remarked in Part IV. E. phyllanthes clearly prefers comparatively light shading, in contrast to E. leptochila and E. purpurata which seek heavy shade in the interior of woods. E. helleborine is intermediate in this respect, and the Continental E. muelleri (see Part V) requires only very light shade. ACKNOWLEDGEMENTS I am indebted to the late Dr. O. Hagerup for his help and advice, and for permission to use his drawing; also to my many correspondents who have sent me information and specimens, particularly Messrs. P. Bowman and A. Roseweir, and to Mr. V. S. Summerhayes for making information about new records available to me. [ am also grateful for the facilities afforded me at the various herbaria visited. REFERENCES BENOIT, P. M. (1959). Epipactis phyllanthes in Merioneth. Nature in Wales, 5, 808-813. HaGeErRup, O. (1947). The spontaneous formation of haploid, polyploid, and aneuploid embryos in some Orchids. K. Dansk. Vidensk. Selsk., Biol. Medd., 20, No. 9. JoveT, P. (1957). Notes et remarques floristiques et taxonomiques. Bull. Soc. Bot. France, 104, 87-99. SrpKes, C. (1954). Epipactis phyllanthes G.E.Sm. in County Wicklow. Irish Nat. J., 11, 113-114 & plate. Wess, D. A. (1953). Epipactis phyllanthes G.E.Sm.: an orchid new to Ireland. Jrish Nat. J., 11, 90-91. Youna, D. P. (1952). Studies in the British Epipactis. IV. A revision of the pAyllanthes-vectensis-pendula group. Watsonia, 2, 259-276. Youne, D. P. (1953). Autogamous Epipactis in Scandinavia. Bot. Not., 1953, 253-270. Youna, D. P. (1955). Epipactis phyllanthes in Yorkshire. Naturalist, 65. Watsonia 5 (3), 1962. STUDIES IN THE BRITISH EPIPACTIS Vil. SEED DIMENSIONS AND ROOT DIAMETERS By DONALD P. YOUNG ABSTRACT Seed-dimensions and root-diameters are characters of some taxonomic significance in Epipactis species. Preliminary data on these measurements is given in the tables. Epipactis are notoriously variable plants, and it is not surprising that their quantitative characters show a wide amplitude and much overlapping between species. The data which follows indicates that the dimensions of the roots and seeds are more absolute characters and of taxonomic significance. It is not the result of a serious biometric study, but simply arose from examination of material that happened to be available. It is presented as an indication of where properly designed studies might be profitable. SEED DIMENSIONS The size of seeds is well known to be almost independent of the size and vigour of the parent plant. Dymes (1921, 1923) has shown that seed morphology is sometimes a useful taxonomic character in the Orchidaceae. Table 1 gives the dimensions of the seeds of all species of Epipactis, Section Epipactis, of Europe and the Mediterranean region, including five non-British species. Samples were mounted in Canada balsam, and measurements were made by means of a micrometer eyepiece on ten seeds taken at random (but rejecting twisted or undeveloped ones). On the basis of size and shape of the seeds, the species fall mainly into two groups : (i) E. purpurata, E. microphylla, E. leptochila, and the three glabrous-stemmed species E. phyllanthes, E. confusa, and E. persica, with large seeds 1-15-1-3 mm. long, and about 4-5 times as long as broad (5-2 in the first species); (11) E. helleborine, E. atrorubens, E. muelleri, and E. dunensis, with smaller seeds, 0-9-1-0 mm. long, length/breadth ratio about TABLE 1 Seed dimensions of Epipactis species. Length, mm. Breadth, mm. Species | Number of | No. of seeds | — Length/breadth | collections | measured | Mean | s.d.¥ | Mean | s.d.f | CO A eee BE. purpurata | 3 30 1:29" | 0-14 | 025-7) o-085 5-2 *persica | 3 27 1:29 O18 77) 70:29 0-04 4:5 phyllanthes 1/9 170 1-28 O:120'7) 50-28 0-045 4-5 *microphylla 1 10 1:16 | 006 | 026 | 0-04 4:5 *confusa 3 30 Feat) 0-15 0-27 0-035 | 4-3 leptochila 2 20 1-15 0-04 0-27 0:03 | 4-3 *condensata 2 | 18 1-19 0-12 ~ | 631 0-04 | 3-9 atrorubens 3: Nl E30 1:04 | 017 | 030 | 0035 | 3-5 helleborine Mt Ritdeese! 0 0:97 | 009 0:27 | 0-025 3-4 dunensis | 5 | 44 0-95 0-17 | 0-29 | 0-04 | 3-3 *muelleri | 1 | 10 0-91 0-12 | 0-27 0-025 3-4 * Non-British + Standard deviation 140 Watsonia 5 (3), 1962. SEED DIMENSIONS AND ROOT DIAMETERS OF EPIPACTIS 141 3:4. The shorter seeds tend to a bluntly clavate shape, whereas the longer ones are more fusiform, with the areolation of the testa elongated. EF. condensata Boiss. (E. troodi Lindb. f.) is somewhat anomalous, with very large seeds 1:2 mm. long but with a length/breadth ratio only 3-9. This species, from the Levant, is quite different in appearance from any of the other species mentioned, and need not be further discussed here. The standard deviation of the pooled length measurements for each species was in most cases less than -++ 10%, emphasising the absolute nature of the character. However, marked differences were sometimes found between different colonies of the same species, particularly with E. atrorubens. At present, too little data is available to say whether this is significant. Some influence of habitat and also of season is to be expected. The almost identical dimensions of the seeds from E. dunensis and E. muelleri should be noted. This reinforces the suggestion (see Part V) that these two are very closely related. ROOT DIAMETER The roots of Epipactis plants are emitted from the rhizome and the base of the stem and are, strictly speaking, all adventitious. They are cylindrical with no taper except at the extreme tip, and they are of practically uniform thickness regardless of their length or position, or of the age or size of the plant. The thickness is characteristic of the species : the thin wiry roots of EF. dunensis, for instance, can be distinguished at a glance from the thick fleshy ones of E. phyllanthes. Some data on the root diameters of all the European species (except E. palustris) is given in Table 2. Two or three pieces of root about 1 cm. long were cross-sectioned, and a random selection of sections was mounted in Canada balsam. The diameter of each section was measured in two perpendicular directions. The standard deviations, as can be seen from the table, are mostly around + 20%. The average diameters range from 1-5 to 3-1 mm. over the various species, so that the root diameter may provide a character of useful significance. There is one apparent ecological effect here, that plants (E. he/leborine and E. phyllanthes) TABLE 2 Root diameters in Epipactis spp. Number of Root diameter, mm. Species es Stations Plants | Measurements; Mean | s.d. E. purpurata | 2 3 | 42 Pans: 9) ose phyllanthes (inland) | 1 7 | 84 294° 1) = 0:37 phyllanthes (dunal) | 2 3 | 32 226 Fon 0:30 *microphylla 1 1 2 2:61 | 0-42 *confusa 4 4 | 48 2254 Si 10:40 *muelleri 5) 6 74 es 32-24 0-54 heileborine (inland) 3 | 3 | 50 e298 all aiOr48 helleborine (dunal) | 4 | 4 | 48 ee 188/55 2029 leptochila | 6 | 6 | 39 ele tle 0:34 atrorubens | 3 3 48 Ol 10:28 | | | | dunensis | 2 | 4 14 14g | 0.28 * Non-British. growing in dune sand have roots noticeably thinner than those from woodland habitats. These are separated in Table 2. Measurements from more stations will be needed before a satisfactory test of statistical significance can be applied (for E. phyllanthes the available data gave P < 0-05 as between dune and inland colonies), but the existence of such a Watsonia 5 (3), 1962. 142 DONALD P. YOUNG difference would not be surprising. If granted, it brings an interesting corollary: the difference in root-diameters between E. dunensis and E. muelleri, which at first sight seems to separate them widely, might be explained (at least in part) as the result of their different habitats. The specific root-sizes do not fall naturally into groups. Neglecting dune plants, they might be arbitrarily classified as (i) species with thick roots (2:5-3-1_mm.), E. confusa, E. microphylla, E. phyllanthes, E. purpurata; (ii) those with thinner roots (1-9-2-:2 mm.), E. atrorubens, E. leptochila, E. helleborine, E. muelleri. Perhaps E. dunensis could be regarded as a third class, with very thin roots. REFERENCES Dynes, T. A. (1921). Notes on the seeds of the British Dactylorchids. Rep. Bot. Soc. Exch. Cl., 6, 432-440. Dynes, T. A. (1923). Proc. Linn. Soc., 135, 48-49; cf. also Godfery, M. J. (1923). The seeds of the Marsh Orchids. Orchid Rey., 31, 266-268. Watsonia 5 (3), 1962. SOME NOTES ON GALEOPSIS LADANUM L. AND G. ANGUSTIFOLIA EHRH. EX HOFFM. By C. C. TOWNSEND The Herbarium, Royal Botanic Gardens, Kew ABSTRACT Galeopsis angustifolia Ehrh. ex Hoffm. is formally typified. The characters normally used to separate the two species are reviewed, and that of leaf shape is shown to have only limited application. Attention is drawn to the structure of the hairs of the calyx in these two plants, which seems to be diagnostic as far as tested. The status of the varieties listed in the second edition of Druce’s British Plant List is investigated; only one is believed to be worthy of recognition. There appears to be no reliable evidence of the hybridisa- tion of the two species. PREAMBLE My interest in the British forms of Galeopsis ladanum and G. angustifolia arose when, as an amateur botanist, I found myself unable satisfactorily to place many gatherings on the basis of the account given by Warburg (1952). In conversation with other botanists, I found that I was not alone in this, and have found subsequently in some of the larger herbaria sheets annotated with such remarks as ‘seems to have the upper indument of angustifolia and the leaves of ladanum.’ It was hoped that an investigation into the varieties listed by Druce (1928) within the British flora would shed some light on these anomalous forms, but rather did this confuse the issue further. The present notes are written with two objects in view; firstly, to clear up the differences between these two species, and secondly to assess the value of the varieties listed by Druce. HISTORICAL The foundation of most modern work on Galeopsis has been the monumental mono- graph of Briquet (1893). The influence of this, with the inevitable changes of status and additions of infra-specific entities of varying worth, may be seen in all the Continental Floras. These range in their taxonomic concepts from the confusing treatment of Rouy & Foucaud (1909) to the ultra-conservative account by Fiori (1925-9); the most recent truly critical paper is that of Henrard (1919). TYPIFICATION OF THE SPECIES The typification of G. ladanum L. presents no difficulty. As has been widely noted, an excellent specimen exists in the Linnean Herbarium at Burlington House. But to the best of my knowledge G. angustifolia has never been formally typified. The name G. angustifolia first appeared on exsiccatae which were distributed by Ehrhart in 1792 (No. 137, mis-cited by Briquet, 1893, as No. 132). He published no description, but the name was validated by Hoffmann (1804), who published a short but adequate diagnosis and cited Ehrhart’s exsiccata number as the basis of the name. Williams (1910) suggested that in fact Hoffmann published angustifolia as a variety of G. ladanum and that Persoon (1807) was the first to accord it specific rank. However, it would seem that the ‘unnumbered species’ listed by Hoffmann (and other older botanists) were intended to take specific rank, though perhaps as species concerning whose status the writer was in some doubt. Mr. J. E. Dandy advises me (in Jitt., 7 Feb. 1961) that such was his view in compiling the List of British Vascular Plants. 143 Watsonia 5 (3), 1962. 144 C. C. TOWNSEND Williams makes the following remarks on this exsiccata : ‘On p. 246 of the monograph, var. (or subsp.) angustifolia is mentioned as having first been described by Ehrhart, P/. exsicc. no. 132 (1792) as a species. This merely refers to a series of specimens (and poor ones) from Ehrhart’s garden, bearing Linnean or other names, several of them obviously cultivated examples, but without either descriptions or other references. There seem to have been 160 of them, divided into ten bundies. In Kew Herbarium Library there is a catalogue of the whole series of names, bound up with the first two bundies. But the name of Galeopsis angustifolia is not in the list at all. No. 132, which is the plant cited in the monograph, is Rumex pulcher; while No. 137, which is the plant cited by Hoffmann, is Artemisia sieversiana.’ It seemed to me that since the date of circulation of the exsiccata referred to by Williams differed from that cited by Briquet, and the numbers given by Briquet and Hoffmann did not correspond to specimens of G. angustifolia in the list which he mentions, it was very probable that there were other sheets distributed by Ehrhart of which Williams was unaware. This view has been vindicated by the discovery of at least two sheets of Ehrhart’s G. angustifolia No. 137 still extant. These are at Halle (HAL) and Goettingen (GOET), and by courtesy of Profs. Meusel and Firbas I have been able to examine both. Both agree well with Hoffmann’s description, but that from Halle is the more homogeneous, and I therefore formally designate this as the lectotype of Galeopsis angustifolia Ehrh. ex Hoffm. (Plate 7a). It has appressed hairs to the calyx and agrees well with specimens of G. ladanum subsp. angustifolia var. kerneri Brig. from Briquet’s herbarium at Geneva (G) which I have had on loan through the kindness of Prof. Baehni. DISTINCTION BETWEEN THE TWO SPECIES The difficulties which have been experienced in distinguishing between G. angustifolia and G. ladanum, at any rate in British and the Continental material which I have seen, are the result of undue stress on the value of leaf characters. Indeed, Warburg stresses these in italics as the chief means of separating these perfectly distinct species. Yet it has long been realised that G. angustifolia frequently has broad-leaved forms. Briquet’s variety odontota is described as having the leaf superficies with a length : breadth ratio of 3:1, which is quite as broad as in many gatherings of G. Jadanum which I have seen. Rouy & Foucaud described a var. Jatifolia of ‘ Race calcarea’ with ‘ feuilles ovales-lancéolées ’ and Henrard’s G. /adanum ssp. angustifolia var. calcarea subvar. platyphylla (Plate 7b) is undoubted G. angustifolia with a superficies of 4-0 x 1:8 cm. In fact, after examining a large number of gatherings of both species, one is made aware that leaf shape in this species pair is as variable as in, for example, the genus Mentha; and as Graham (1954) has done good service in recommending the quashing of all varietal epithets based on leaf shape in Mentha aquatica L., so ‘ varieties’ of, particularly, G. angustifolia based on leaf shape are scientifically worthless, as a complete range from broadly ovate to narrowly linear may be found, though lanceolate is probably the commonest shape. Likewise I have seen leaf serration quite as pronounced in G. angustifolia as in any G. ladanum material which has been examined. However, it may be said that G. Jadanum does not seem to produce forms with linear or linear-lanceolate leaves, and to this extent leaf-shape would appear a reliable character; the real pitfall is in the broad-leaved G. angustifolia forms. Thus we are left with the other commonly used separating character, that of calyx indument. After testing this on a great number of specimens I am satisfied that, correctly understood, it affords a means of accurate identification. Unfortunately the macroscopic difference is of a nature which is not readily expressed in words, and unquestionably the best means of apprehending it is to ‘ get one’s eye in’ by examining a quantity of authentic material. As is pointed out by Warburg the calyx of G. ladanum appears green, while that of G. angustifolia appears whitish or canescent. At first sight this appears due to the fact that the hairs on the calyx of G. ladanum are patent, thus rendering the tube visible; but if fresh or fairly recent herbarium material is viewed under a lens it is at once apparent that the hairs of G. ladanum have a peculiar transparent and glistening appearance while those of G. Watsonia 5 (3), 1962. GALEOPSIS LADANUM AND G. ANGUSTIFOLIA 145 angustifolia appear a dull harsh white. To ascertain the reason for this, the hairs were examined under the high power of the compound microscope (x 400). It was found that the white colour of the hairs of G. angustifolia was due to the crowded, coarse papillae with which the hairs are furnished. In G. /Jadanum, on the other hand, the hairs are for the most part almost devoid of papillae and quite transparent—so much so that the inner walls of the cells of the hair are visible (Fig. 1). All the hairs in G. Jadanum are not always as smooth as figured, but in material I have examined papillae of sufficient prominence to show any irregularity in the profile of the hair are very rare; it is difficult to determine whether such unevenness in the hair surface as does occur is in fact caused by a very low papilla or a very shallow punctum, and the inner cell walls remain visible. In old herbarium material the glistening appearance under the lens is lost due to the shrinkage of the cells, but is restored somewhat by boiling out. I cannot find any previous notice of this character, which is diagnostic as far as I have been able to test it. It may also be noted that when the indument of the calyx in G. angustifolia is so sparse as to render any appreciable area of the tube visible, the hairs are usually short and tightly appressed, glands are few, and there is often development of anthocyanin pigmentation which tints the tube purplish or brownish. In fresh material the glistening appearance of the glands in G. ladanum is very striking. Fig. 1. Calyx hairs, x 400, of (a) Galeopsis ladanum L. (b) Galeopsis angustifolia Ehrh. ex Hoffm. Hybrids between the two species have been reported from the Continent, but I have had no difficulty in assigning all the material which I have seen to either one species or the other. G. Jadanum is clearly a plant of considerable rarity in this country, occurring in waste places, docks, railway tracks, vegetable gardens and other abodes of the casual alien, which is clearly its status in Britain. It is thus hardly to be expected that, in any event, hybrids would be of any but the rarest occurrence with us. According to Briquet, the hybrid recorded by Haussknecht (1884) is in fact typical Jadanum. Briquet himself cites some intermediates between ssp. intermedia and ssp. angustifolia ; but since these are for the most part said to approach ssp. intermedia in leaf shape probably broad-leaved angustifolia forms are involved. ON THE VARIETIES LISTED IN DRUCE’S “ BRITISH PLANT LIST,” ED. 2 (1928) In this list Druce notes only one species, 578/4 G. ladanum L., with eight varieties. These will be dealt with seriatim :— b. intermedia (Vill.). G. ladanum L. was divided by Briquet into two subspecies, subsp. intermedia (Vill.) and subsp. angustifolia. It is as clear as can be ascertained in the absence of an authentic specimen in Villars’ herbarium at Grenoble that Villars’ G. intermedia was, Watsonia 5 (3), 1962. 146 C. C. TOWNSEND as has been long accepted, a form of G. ladanum L. sensu stricto. Villars’ second and better description (1778), including the phrase ‘toute la plante est visqueuse,’ considered in con- junction with the plate, which shows a broad-leaved plant with patent glandular hairs on the calyx, can hardly apply to G. angustifolia. Villars appears to have regarded his plant as an intermediate between G. ladanum and G. tetrahit on the grounds of flower size—not a constant character in G. Jadanum, or in most species of the genus. c. angustifolia (Ehrh.). This becomes the second species, G. angustifolia. d. Jatifolia (Hoffm.). With his original description Hoffmann (1804) cites t. 884 of Sowerby & Smith’s English Botany. This plate is not only clearly G. /Jadanum, but in the accompany- ing description it is mentioned that the plant illustrated agrees well with the Linnaean specimen. Thus G. /atifolia Hoffm., like G. intermedia Vill., falls into the synonymy of G. ladanum L. sensu Stricto. e. kerneri Brig. As has been pointed out above, specimens of var. kerneri from Briquet’s herbarium match the lectotype of G. angustifolia. Thus, assuming that two or more varieties are recognised within G. angustifolia, ‘var. kerneri’ falls into the synonymy of var. angustifolia. f. campestris Timb. This variety is placed by Briquet in the synonymy of his var. orophila (an earlier MS. name from an exsiccata distributed by Timbal-Lagrave which he takes up), which is distinguished from ‘var. kerneri’ by the length of the calyx teeth (5 mm. long as compared with 2-3 mm. for the latter variety). That the two names do in fact apply to the same form is clear not only from Timbal-Lagrave’s description (1885) : ‘lobes du calice aussi longs que le tube . . . . terminés par une épine blanche et longue,’ but also from examination of the exsiccata distributed through the Société Dauphinoise on which this was based compared with specimens of var. orophila from Briquet’s herbarium. Since the epithet campestris is the earlier at varietal level, this is the correct one. I do not consider, however, that the variety is worth maintaining. The length of the calyx teeth in G. angustifolia varies tremendously, quite independently of other characters. I have seen quite a lot of British material that matches well * var. orophila ’ from Briquet’s herbarium — e.g., a gathering by E. F. Linton from Diddington Wood, Hunts. (v.c. 31), 29 July 1879, in Cambridge (CGE); and also others which provide a good match for the type of G. ladanum subsp. angustifolia var. calcarea subvar. longidentata Henrard. These apparently represent the greatest length reached by the calyx teeth in Great Britain. At the other end of the scale are specimens agreeing with the type of G. ladanum subsp. an- gustifolia var. calcarea subvar. microdus Henrard — e.g., a gathering from ‘shingle bank between Southampton and Netley,’ S. Hants. (v.c. 11), Sept. 1846, from Herb. Babington in Cambridge (CGE). On the Continent greater variation even than this occurs, from var. spinosa Benth., with calyx teeth said to be over 8 mm. long, to var. inermis (Posp.) Fiori with muticous calyx teeth. I have seen no material agreeing with the original descriptions of these, but they are clearly angustifolia forms. g. odontota Briq. This variety is based on leaf shape, Briquet keying it out from “ var. kerneri’ thus :— Leaf serration always of irregular, scarcely perceptible or almost non-existent, triangular teeth less than 1 mm. deep. ‘var. kerneri.’ Leaf serration of stronger, more regular teeth, triangular, 1-1-5 mm. deep. var. odontota (& var. berteti Briq.). Var. odontota, which Briquet himself cites from Miller’s Dale, is also said to have broad leaves of which the bases are rounded and shortly attenuate into the petiole. As has been stated above, no useful purpose will be served by maintaining varieties of this species based on leaf shape, and var. odontota thus falls into the synonymy of var. angustifolia. Watsonia 5 (3), 1962. ‘plelusyH ByjAyddinjd “reaqns “bug (yuoyss) DaiDI0I “ea “biig (WYOH x9 “YIYy”) vyofijsnsuv ‘dss “q wnuvpy sisdoajvy jo adA\ojoH (q) yesess ea tora fe fey : “WOH xo (a Oe ee — ‘YlYA vyof{Ijsnsuv sisdoajvy Jo adk\o1aqT_ (ke) Osacen! ; : Sap : : SESS SS [ee a5 greg mqomndor apdoor9 : A S aS ere . passonueny - 4 | 2 DEER OLY . ee oe - | oe — ay oe ¥ : : oe ae : of DIGRESS AGH RES OZGUQRHETS QRS 23RD. . - NSLIVISNY SHOSINVLOS eSrr ON AMIVQIN BIA bss Po a a - = Le S i cu re len > 3 ors SISNSIVH SILVLISYSAINN WhINVeuaH BNLYOSN BSR = d/ ALVId VL ALVId GALEOPSIS LADANUM AND G. ANGUSTIFOLIA 147 h. monticola Lannes. In some areas, particularly on the chalk in eastern and south- eastern England, there occurs a form of G. angustifolia which in its extreme state appears very distinct. It is a rather short plant, branched almost from the base, canescent, and with the calyx densely villous with patent, long white hairs. This form was first described without ambiguity by Schoénheit (1832), as a species, G. calcarea, attention being particu- larly called to the ‘ calycis patenti-villosuli.” It is the form often called * var. canescens ’ by referees of British exchange clubs, but its correct place under calcarea was noted by Salmon (1931). Briquet reduced Schénheit’s species to a variety of G. Jadanum ssp. angustifolia. However, since according to the International Code a name has priority only within its own rank, consideration must be given to two varietal names of earlier date which Briquet cites in synonymy here. One of these is var. monticola Lannes, the name used by Druce. This, however, is a omen nudum which appeared on a herbarium label, Magnier’s Flora selecta No. 940. Neither on the sheet nor in the Scrinia which gave notes on Magnier’s plants does any description appear. The only note which the label on the specimen bears is ‘ an G. monticola Jord ined.?’ It is undoubtedly G. calcarea, but the epithet monticola must be discarded as never validly published. The name G. angustifola var. arenaria Gren. Godr. was indicated as the correct one for the form under consideration by Brenan (1956). But what does not appear clear is why Briquet should have placed this name within the synonymy of calcarea at all. The original description is very brief: ‘Dents du calice plus courtes; plante blanchatre et parfois glanduleuse dans le haut.’ There is no mention of patent hairs on the calyx, and in fact Godron cites G. canescens Schultes, which Briquet maintains as a distinct variety, as a synonym. The only two specimens of G. angustifolia var. arenaria Godr., named by the author, which exist in his herbarium, were kindly sent to me on loan by the Director of the Ecole Nationale Supérieure Agronomique de Nancy. Both bear collection dates later than 1850, the date of publication of Godron’s varietal epithet. These two specimens represent diverse elements. Under Briquet’s classification, interpreted in the light of specimens from his herbarium at Geneva, one specimen, which has the calyces almost or quite eglandular and totally appressed-hairy, would have been called *‘ G. ladanum ssp. angustifolia var. kerneri’ (i.e., G. angustifolia var. angustifolia); the other, which has the calyces considerably glandular and with some patent hairs, would have been called *G. ladanum ssp. angustifolia var. calcarea.’ Indications are, therefore, that Godron would have applied his epithet to any reasonably canescent form of G. angustifolia irrespective of other characters. What is clear is that whatever was in Godron’s mind, he was in fact reducing Galeopsis canescens Schultes to the rank of variety. The present writer’s views on G. canescens are given later in this paper, and these views thus gain some additional support from the fact that Godron appears to have interpreted Schultes’ species in the same manner. Thus G. angustifolia var. arenaria Godr. and G. canescens Schultes can now only be regarded as merely representing any + canescent forms of Galeopsis angustifolia. The correct name for the form of G. angustifolia with the patent-villous calyces remains G. angustifolia var. caicarea (Schonh.) C. E. Salmon, based on the beautifully circumscribed G. calcarea Schonheit. In addition to the patent calyx hairs, var. calcarea is usually con- siderably glandular around the inflorescence. Intermediates between var. calcarea and var. angustifolia certainly occur not infre- quently, and it is a matter of opinion as to whether the former should be maintained; in particular, forms with long appressed hairs on the calyx tube, and patent hairs in the region of the teeth and sinuses between them, are clear intermediates. Nevertheless, the appearance of the variety is so distinct in the field, and it is so predominant over large areas of arable chalkland, that it seems worthy of recognition. It may be noted that the hairs of the calyx tend to spread even further on drying and storing. i. canescens (Schultes) (not Schultz, as in the British Plant List). This is certainly the most elusive of the varieties of G. angustifolia. In the synoptic table which forms chapter XXXI of the Monograph, Briquet separates it from vars. odontota and kerneri, which it is said to resemble in the appressed indumentum of the calyx, in the following manner :— Watsonia 5 (3), 1962. 148 C. C. TOWNSEND Plant with branches greenish, pilose or glabrous, but not with a hoary indumentum. vars. odontota, kerneri (&c.). Plant with branches densely and closely hairy, the stem, leaves, bracteoles and calyces likewise being thickly canescent. var. canescens. The difficulty in resolving precisely what Schultes’s plant was is due principally to the fact that no authentic specimen appears to be extant at Munich, Berlin, or other Continental institutions. Fortunately, however, his description is a good and full one. The plant pictured by Briquet’s description on p. 254 of his monograph is one with so thick an indumentum that the whole plant appears ashy-white (though, as often, the specimens in his herbarium so named do not always bear this out). Likewise, the leaves are depicted as having their margins more or less revolute; and in all but one of Briquet’s specimens (a specimen from the Somme, ex herb. Schmidely, coll. E. Gonse — this is densely white-canescent) the plants have linear, entire or slightly stnuose leaves. In his original description Schultes (1809) clearly states ‘caule....subtomentoso .... foliis /anceolatis subtomentosis subserratis.’ Also, in Oesterreichs Flora, Schultes gives to his G. canescens the vernacular name ‘ Graugriiner Hohlzahn’ and refers to the stems and leaves as ‘ etwas filzig.” For such a plant as Briquet describes, * graugrtiner ’ would hardly be appropriate; ‘ weisgraulich’ or ‘ weisgrau’ seem more applicable. It appears to me that G. canescens was merely one of the more canescent forms of G. angustifolia var. angustifolia, and that Godron was correct in using the epithet arenaria to cover the same range. Briquet’s plant is said to occur in dry or stony places by the sea, or by lakes. Under such xerophytic conditions many species take on an unusually tomentose appearance (as may be seen in many plants as they occur, for example, at Braunton Burrows); similarly, leaf margins may become revolute. Briquet’s form is probably of this kind; but it must be confessed that in this country G. angustifolia grows on maritime shingle (as for example at Wolferton, Norfolk, at Dungeness and the Crumbles, Eastbourne) without developing such tomentum, and the variety may prove worth keeping up. If so, the name var. /ittoralis de Vicq & de Brutelette (1864) should be used, as the description of these authors (‘ velue blanchatre ’) is far more applicable to Briquet’s plant than that of Schultes. Mr. Brenan informs me (in Jitt., 1 March 1957) that ‘... I have seen what appears to be canescens as a sand dune plant in southern France, and it looks very different from anything I have seen in Britain.’ I have seen no British material which could be placed here with any confidence. The confusion which has existed in the past between canescens and calcarea may well be due to Hegetschweiler (1840), who describes his G. canescens as ‘ die Kelch... mit... etwas abstehenden Harchen besetzt.’ G. canescens Heg. is, of course, an illegitimate later homonym. It is thus proposed that Galeopsis ladanum L. agg. in Great Britain be listed thus : Galeopsis ladanum L. Galeopsis angustifolia Ehrh. ex Hoffm. Var. angustifolia Var. calcarea (Schénh.) C. E. Salmon. ACKNOWLEDGMENTS Acknowledgments are due to the authorities of the various institutions mentioned in the course of the paper, and to those at Paris (P) and Leiden (L) for the loan of certain exsiccatae and type material ; to the authorities of the British Museum and Kew for facilities given prior to my taking up my present post in the latter Herbarium ; to Dr. S. M. Walters for similar facilities freely given during my residence in Cambridge, and to Mr. P. D. Sell for personal help in the Cambridge Herbarium ; to Mr. A. A. Bullock and Mr. J. E. Dandy Watsonia 5 (3), 1962. GALEOPSIS LADANUM AND G. ANGUSTIFOLIA 149 for advice on points of nomenclature ; to Dr. H. Heine for help in certain German interpretations ; to Miss M. Grierson for the preparation of Fig. 1 and to all who have kindly lent specimens for examination. REFERENCES BRENAN, J. P. M. (1956). Notes on the flora of Oxfordshire and Berkshire, II. Proc. B.S.B.1. 2, 111. BriQueT, J. (1893). Monographie du genre Galeopsis. Brussells. Danby, J. E. (1958). List of British Vascular Plants. London. Druce, G. C. (1928). British Plant List, Ed. 2, 93. Fiori, A. (1925-9). Nuova Flora analitica d’Italia, 2, 418 ff. Florence. GRAHAM, R. A. (1954). Mint Notes V. Watsonia 3, 109. Gopron, D. A. (1850), in Grenier, J. C. M. & Godron, D. A. Flore de France, 2, 684. Paris. HAUSSKNECHT, C. (1892). Mitt. thiiring. bot. Ver. 8. HEGETSCHWEILER, J. (1840). Die Flora der Schweiz, 570. Zurich. HENRARD, J. TH. (1919). Galeopsis : Een systematisch floristische-studie. Nederl. Kruidk. Arch., 1918, 158 ff. HOFFMANN, G. F. (1804). Deutschlands Flora, Ed. 2, 2, 8. Erlangen. PERSOON, C. H. (1807). Synopsis Plantarum, 2, 122. Paris. Rouy, G. & Foucaup, J. (1909). Flore de France, 11, 279 ff. Asniéres & Rochefort. SALMON, C. E. (1931). Flora of Surrey, 530. SCHONHEIT, P. (1832). Phytographische Bemerkungen, in Flora, 15, 593-4. SCHULTES, J. A. (1809). Observationes Botanicae, 108. Innsbruck. ScHULTES, J. A. (1814). O0cesterreichs Flora, Ed. 2, 2, 151. Vienna. SOWERBY, J. & SMITH, J. E. (1801). English Botany, 13, 884. London. TIMBAL—LAGRAVE, E. P. M. (1885). Bull. Soc. Dauphin., Etud. biol. 12, 575. DE VicQ, E. & DE BRUTELETTE, B. (1864). Catalogue raisonné des plantes vasculaires de la Somme, 487. Abbeville. VILLARS, D. (1779). Prospectus de Il’ Histoire des plantes du Dauphiné, 21. Grenoble. VILLARS, D. (1787). Histoire des plantes du Dauphiné, 2, 387. Grenoble. WARBURG, E. F. (1952), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Cambridge. WILLIAMS, F. N. (1910). Prodromus Florae Britannicae Pt. 6, 396 ff. Brentford. Watsonia 5 (3), 1962. CHROMOSOME NUMBER, MORPHOLOGY AND BREEDING BEHAVIOUR IN THE BRITISH SALICORNIAE By D. H. DALBy Botany Department, Imperial College, London ABSTRACT Chromosome counts on British Salicornia confirm the presence of two series, diploids with 2n = 18, and tetraploids with 2x = 36. These two series differ in size of pollen, stomatal guard cells and seeds; and also in general morphology. An attempt is made to correlate these numbers with the taxa generally recognised as being present in the British flora. The plants are self-compatible, and appear usually to be self-pollinated, though there is limited evidence for occasional anemophily and possible out-breeding. Infra-specific variation may be attributed to the presence of numerous virtually pure-breeding races perpetuated by autogamy. CONTENTS PAGE INTRODUCTION .. as wai a sie Ae ae ae - i 150 MORPHOLOGY .. ae ae a =e oe ry a Meaimes |5\() NAMES USED .. a ae ie ate “if hes Be aS op let CHROMOSOME NUMBER .. ae or as ig ars ae ts ae 151 MORPHOLOGICAL DIFFERENCES BETWEEN DIPLOIDS AND TETRAPLOIDS .. nee ae Ee 153 BREEDING BEHAVIOUR .. ae xt ae a he i Sa ie 158 ACKNOWLEDGMENTS Ms ye a rae Me = =~ a ee 161 REFERENCES Bes a ae ae 5 ae ae ee a > 161 INTRODUCTION The genus Salicornia has for many years been a source of trouble to systematists, and we may echo the words of Baxter (1839) : ‘ Botanists of the highest authority differ in opinion respecting the specific distinctions of the British Salicorniae.’ This is especially unfortunate in view of the important role played by these plants in the ecology of saltmarshes and similar habitats. In order to achieve a satisfactory understanding of this group, one must take into account various aspects of the plants’ general biology such as breeding behaviour and phenotypic plasticity, which until recently have been neglected or altogether ignored. The taxonomic problems presented by this genus cannot be solved by the study of herbarium material alone. In the present paper attention is directed mainly to the relationship between mor- phology and chromosome number, together with a discussion of breeding behaviour. The significance of the variation patterns seen in field colonies, and the particular problems of growth habit and taxonomy, will be elaborated elsewhere. MORPHOLOGY Salicornia L., a genus of succulent halophytes belonging to the Chenopodiaceae, is represented in Britain by one perennial and several annual species, limited in distribution to coastal saltmarshes. Free leaves are absent, and the fleshy internodes (“ segments ”) have apparently been derived from the fusion of the decurrent leaf bases with the outer cortical tissues of the stem (de Fraine, 1913). Branching is regularly decussate, and in large plants 150 Watsonia 5 (3), 1962. BRITISH SALICORNIAE 1B) may be of the fourth order. When mature every branch ends in an inflorescence, referred to in this paper as a spike, though perhaps a more correct description would be a * spike of three-flowered cymules.’ Branches which are purely vegetative at maturity are confined to the perennial species. Fig. 1. Longitudinal section through Salicornia flower. ad—anther, dehisced, au—anther, unopened, o-ovule, p—perianth, s—stigma, v—vascular strand. The flowers are much reduced morphologically (Fig. 1), the perianth consisting of three minute flaps and being immersed in the cortical tissues so that its upper surface is flush with that of the stem. Normally each flower contains two stamens, but sometimes only one is present. The solitary ovules (and also the mature seeds) lie transversely in the ovary, parallel to the main axis of the stem, and each with its micropylar end towards the funicle (Fig. 1). NAMES USED There is no general agreement as to the number or status of the taxa occurring on our coasts. Illustrative of this variation in personal judgement are the differing treatments of the British species advanced by Tutin (1952,1959) and by Ball & Tutin (1959). Extreme views on the species concept in the genus are those of Wilmott (1939, and unpublished), and KG6nig (1960). For reference purposes some taxonomic system must be followed, so the names given by Dandy (1958) are being used in this paper as they conform more closely with the views of the present writer. These are : Salicornia perennis Mill., Gard. Dict. ed. 8, no. 2 (1768). S. dolichostachya Moss, New Phytol., 11, 409 (1912). Seuropaca ., Sp. Pl., 3 (1753). S. ramosissima Woods, in Henfrey, Bot. Gaz. 3, 29. S. pusilla Woods, in Henfrey, Bot. Gaz. 3, 30 (1851). Making allowance for the synonymy given by Dandy, these species are interpreted in the sense of Tutin (1952), where references may be found to the appropriate figures in Moss (1914). Ball & Tutin’s new species, being published later, are not included by Dandy; _ reference will be made to these in a subsequent paper. CHROMOSOME NUMBER Of the published chromosome counts on European Salicornia known to me, those which can be assigned to particular segregates within the genus are included in Table 1. Watsonia 5 (3), 1962. S74 D. H. DALBY TABLE 1. Chromosome counts for European Salicornia species. Ludwig | Hambler| Ball & Dalby 1950 1954 Tutin 1959 | unpublished Wulff Maude | K6nig} Castro & 1936, 1937] 1939 | 1939 | Fontes 1946 . perennis — 18 — 18 — 18 — 18 S S. pusilla — — — — — 16? 18 18 S. ramosissima — — 18, 36 —— — 18 18 18, 36 S. europaea 38 — -— 18 18, 36 36 18, 36 36 S. dolichostachya — — 36 — — 36 36 36 Some of the counts given above were recorded for taxa not recognised by Dandy (1958). Such taxa have been grouped as far as possible to conform with the species listed by Dandy. This table shows clearly that there are two chromosome series present, one diploid with 2m = 18, and the other tetraploid with 2n = 36. No triploid hybrids have so far been recorded. There is considerable agreement as to the chromosome numbers for each taxon, apart from S. ramosissima and S. europaea, both of which appear to contain diploid and tetraploid races. It is clear, from other evidence, that these names are being used to cover a range of forms whose precise taxonomic status is open to debate. With regard to the aneuploid numbers which have been published, Hambler (1954) was uncertain of his own count of 2m = 16 for S. pusilla, and counts by other authors for this very distinct species give uniformly 2n = 18. Wulff’s (1936, 1937) figure of 2n = 38 needs some comment. He writes (1937) ‘ Ich fand in zahlreichen Platten eindeutig 2n — 38 Chromosomen, wodurch S. herbacea L. als polyploide, abgeleitete Form gekennzeichnet ist.’ His conclusion as to polyploidy is certainly correct, but it is strange that he did not find any plates with 2n = 36. No plate that I have examined has shown an aneuploid number, all examples where exact counts were possible proving to be euploid 18 or 36. In the present study, chromosome counts were made from metaphase plates of mitotic divisions of root-tips,obtained from plants in the field and from seedlings in culture. roan Suk a> TU ERS en ~ WS . ZN Ne ? Clac,) a “Ere PIT ye pute, L 39,b¢ 17 ek i, MEME es J we Re BA i 372 ON Ws D K L E Py ANF We ING A, a —9 0 lO p M G Saar Fig. 2. Metaphase plates from root-tip sections. Diploids (2n = 18) A—C, S. ramosissima, Hayling Island, Hants (Nos. 68, 55, 69); D, E, S. ramosissima, Shingle Street, Suffolk (Nos. 12, 11); F, S. ramosissima, Flatford, Suffolk (s.n.); G, S. pusilla, Hayling Island, Hants (No. 52). Tetraploids (2n = 36). H,S. europaea, Blakeney Point, Norfolk (s.n.); I, S. europaea, Colne Point, Essex (No. 35); J, S. dolichostachya, Hayling Island, Hants (No. 71); K, S. europaea, Hayling Island, Hants (No. 54); L, S. europaea, Flatford, Suffolk (No. 19); M, S. europaea, Shingle Street, Suffolk (No. 14). Watsonia 5 (3), 1962. BRITISH SALICORNIAE 153 Langlet’s modification of Navashin’s fixative was used, this fluid also being used for storage until required. The stain used was gentian violet, which was usually taken up well by the chromosomes, and differentiation yielded very satisfactory plates. Some counts were made from acetic orcein squashes of fresh root-tips. Some examples of metaphase plates are given in Fig. 2. After fixation the chromosomes are small and range from about 0-6 u to 1-8 w in length, with very little in the way of dis- tinctive morphology to enable the recognition of homologous chromosomes or particular karyotypes. Nevertheless, diploid plates show sufficient variation in chromosome shape and size to suggest genetic heterogeneity within the scope of the species as recognised in the present study. In two plates at least from plant No. 52 (S. pusilla), one chromosome (arrowed in Fig. 2 G) was seen to bear a satellite. MORPHOLOGICAL DIFFERENCES BETWEEN DIPLOIDS AND TETRAPLOIDS It was thought that diploid and tetraploid plants within the genus might well show systematic differences in size or form of their parts, and that this could be used as a primary taxonomic division in the annual species. A range of specimens was therefore examined, including plants from all the species recognised by Dandy (1958). As mentioned above, no definite aneuploid counts were made, and in a few polyploids where clumping made precise counting difficult, they were regarded as having the expected number of 2n = 36. (a) Pollen grains Pollen grains were taken from mature anthers, and mounted in acetocarmine, their diameters being drawn with a camera lucida at a magnification of x 800. The grains are almost spherical, and their volumes were assumed to be the volumes of spheres with similar diameters. It should be noted that measurements of volume are of greater value than say diameters or lengths, though slightly more difficult to obtain. 20 = I2 pA 70 7 nn 19 14 = 13 i POLLEN GRAIN VOLUME IN 103 CU. O Fig. 3. Pollen grain size for 25 plants of known chromosome number. Note. In Figs. 3 to 5, dots and asterisks mark the mean values for diploids and tetraploids respectively. The shorter bars above and below mark distances of -- 2 standard error of the mean (non-overlap indicates p = < 0-005 for samples of similar size). The longer bars mark distances of + 2 standard deviation, as an estimate of the range. This range will be exceeded by about 1 in 20 in a normal sample. Watsonia 5 (3), 1962. 154 D. H. DALBY The data for pollen size are given in Fig. 3, where it is seen that in general the pollen grains are larger in the tetraploids than in the diploids, the change-over being at about 10:5 x 10% cu. » volume (27-2 » diameter). (6) Stomatal guard cells The stomata are rather variable in size, even on a single plant, and vary considerably in frequency, being most numerous towards the tops of the fleshy internodes. They are arranged with their long axes at right angles to the line of the stem, a feature often found in xeromorphs and succulents. Strips of epidermis were removed from living plants, and placed immediately in absolute alcohol. The length of each stoma was drawn with a camera lucida at a magnification of x 800, and the volume calculated as that of a sphere with diameter equal to the stoma length. These volumes will always bear a constant relationship to the actual stomatal volumes, provided the stomata are of constant shape. This was assumed to be so. Bb O 35 30 25 20 STOMATAL EQUIVALENT VOLUME IN 10% CU. Fig. 4. Stomatal size for 44 plants of known chromosome number. See also note to Fig. 3. The stomatal sizes are shown in Fig. 4, where, as might be expected, they are generally larger in the tetraploids than in the diploids, the change-over point between the two groups being at approximately 13 x 10% cu. » for equivalent volume, and 29-17 uw for length. There is however considerable overlap, part of which seems to be due to certain diploid plants having shown unusual vigour in their vegetative growth. These plants, referred to later in this paper, were selected in the field because of their noticeably robust growth and it was expected that they would prove to be polyploids related to, for example, the ‘ Typ. Nr. 3’ of K6nig (1939). These plants proved in reality always to be diploids. In this con- nection, it is interesting that out of four plants whose pollen size was compared with that of their F, progeny, the only one to show a significant change in size was one of these robust *‘ pseudopolyploids.’ This was plant No. 11, and the significantly smaller F, pollen seemed correlated with poorer vegetative growth. Clearly pollen size is more closely correlated with chromosome number than is stomatal size in Salicornia, but in neither case is the relationship sufficiently precise as to be used alone as a reliable indicator of chromosome number without actually making a count. Watsonia 5 (3), 1962. BRITISH SALICORNIAE 155 (c) Seed size Seeds were taken from the central flowers of cymes on heads of plants of known chromosome number. The seeds from lateral flowers would have been equally satisfactory for this purpose, but seeds from central and lateral flowers should not be mixed, as within each cymule the central flower always has a larger seed than has either of the laterals. 2.8 Ov Oo 2.6 KS) hb ne) Le) Je O SEED LENGTH IN MMS a ® 1.0 HSSESS Fig. 5. Seed size for 18 plants of known chromosome number. See also note to Fig. 3. Seed lengths are shown in Fig. 5, where it is seen that tetraploids have larger seeds than do the diploids, although there is considerable overlapping. The change-over size is at about 1-6 mm. length. Plants 28 and 9 were robust ‘pseudopolyploids,’ and once again the increase in size of a particular structure seems to follow vigorous vegetative growth in the rest of the plant. The largest tetraploid sample comes from plant 39, a natural mutant of S. europaea, in which the cymules are one-flowered. Here, in the absence of competition from the lateral flowers, the central seeds grew to an abnormally large size. (d) Ratio of numbers of fertile to sterile segments The ratio of numbers of fertile to sterile segments shows a break with very little overlap at about 55,* the data being given in Fig. 6. The ‘ pseudopolyploid ’ plants referred to above are labelled ‘A’ in Figs. 6 and 7. They are all regarded as belonging to S. ramosissima, eK 40k kK CK ek oA eA eA @A @A @A © 6 @ ee fo) 0 86 68 68:9 moe weleaeio oe iar alee th ae oe a or 10 20 30 40 60 80 100 200 400 FERTILE : STERILE SEGMENT NUMBER RATIO Fig. 6. Ratio of number of fertile to sterile segments for 36 plants of known chromosome number. Rings— diploids (S. pusilla), dots-diploids (other than SS. pusilla), A-robust forms, asterisks—tetraploids. Note. The horizontal scale is plotted logarithmically. * Tn this paper, all ratios are expressed as single figures, the second (always 100) being omitted for brevity. Watsonia 5 (3), 1962. 156 D. H. DALBY ratio = 20 | atio=50 ratio = |OO 5 10 15 20 25 30 fertile segment number Fig. 7. Change in ratio of numbers of fertile to sterile segments in F, progeny in S. ramosissima. Open circles—parent plants; solid circles-mean values for F, progeny in culture. Lines connecting circles contin- uous if change is significant (p = < 0-005), broken if not significant. Plants 9, 11, 12 and 65 were robust ‘A’ forms, the remainder being of normal size. some evidence for this being given in Fig. 7, which shows the change in the segment number ratio between several parent plants and their progeny in cultivation. Clearly the ‘A’ plants lose their high values for the segment number ratio, and assume a constant one of between 40 and 50, typical of normal S. ramosissima in the field. One may reasonably assume that in normal circumstances of growth, there is little or no overlap between diploids and tetraploids in their segment number ratio. OA SEGMENT SHAPE EN AQ 1 RATIO: 2 OO B i i= git cor! sel SEGMENT SHAPE INDEX : | A+C ees Ase se 2B - 100 Fig. 8. Derivation of segment shape ratio and index. Means for 2 or 3 segments (central by distance along terminal spike of main axis) used in calculations, except for progeny and field colony samples where only one central segment was measured. (e) Segment shape index A segment shape index (see Fig. 8) also reveals differences between the two groups. The break is seen in Fig. 9 to be at about 65. Very few diploids overlap the tetraploid range, and those that do are in fact all ‘A’ plants. Watsonia 5 (3), 1962. BRITISH SALICORNIAE il 5)7/ xe Ok ke FOUR & A A @, 2% A % r) eee © © 45 50 55 60 65 70 75 80 85 90 95 SEGMENT SHAPE INDEX Fig. 9. Segment shape index for 39 plants of known chromosome number. For segment index see Fig. 8. Rings—diploids (S. pusilla), dots—diploids (other than SS. pusilla), A-robust forms, asterisks—tetraploids. (f) Anthocyanin Anthocyanin occurs widely in the European species of Salicornia, and its varied dis- tribution and intensity are responsible for their often brilliant autumnal coloration. The presence or absence of anthocyanin in 43 plants investigated is shown in Table 2. TABLE 2 Diploids Tetraploids Anthocyanin present 20 “. Anthocyanin absent 4 15 The correlation between chromosome number and presence or absence of anthocyanin is highly significant, though not absolute (testing for 77, p = 0-001). Anthocyanin is almost restricted to the diploids, though it does occur rarely in tetraploids. Strictly however, it is the ability to produce anthocyanin in favourable circumstances that should be measured, rather than its actual appearance. Too much shading prevents the red pigment being de- veloped, a feature sometimes encountered in prostrate forms which are green beneath yet bright red above. (g) General conclusions The results discussed above show that the two levels of chromosome number are quite distinct morphologically for several characters which can be measured with some precision. Other differences may also be noticed (some are listed by Ball & Tutin (1959) and K6nig (1960)), and it is gratifying to observe the general agreement reached here. One striking difference however, is seen in Ball & Tutin’s data for pollen size. They regard a pollen diameter of about 29-5 u as marking the change from diploid to tetraploid, whereas the present writer would put it at 27-2 ». Approximately half the tetraploids plotted in Fig. 3 come within Ball & Tutin’s diploid range for mean grain size. Their data for stoma and seed size are however almost identical with those given in this paper. The assigning of chromosome numbers to the different species of Salicornia is a difficult problem, and will not meet with the same general agreement as has been reached over the correlation between gross morphology and chromosome number. This is because of the instability of the systematic treatments of the genus referred to at the start of this paper. Thus although S. pusilla is definitely a diploid, and S. dolichostachya is definitely a tetraploid, both S. ramosissima and S. europaea appear to contain diploid and tetraploid forms. It is from this S. ramosissima—europaea complex that several new species have recently been separated, and doubtless others could too with equal justification. Watsonia 5 (3), 1962. 158 D. H. DALBY It seems to me that it is in this S. ramosissima—europaea complex that we should seek a basic Salicornia-type from which the other forms have been derived. This is wholly opposed to Salisbury’s (1940) views, where he believed S. dolichostachya to be a basic type, allied to S. perennis. In the first instance, S. dolichostachya is polyploid, and so cannot be regarded as ancestral to the diploid species. Secondly it is very close morphologically to some forms of S. europaea, and differs quite markedly from S. perennis. These differences are in fact so great as to have led Moss (1948) into transferring the latter species from Salicornia to Arthrocnemum. BREEDING BEHAVIOUR Genera showing taxonomic difficulties often possess reproductive peculiarities, and as little is known of the reproductive methods in these plants, the remainder of this paper is devoted to the pollination and fertilisation of Salicornia. (a) Pollen fertility The fertility of pollen from a number of plants was determined by mounting fresh grains in acetocarmine; those which took up the stain deeply being considered fertile, whilst those which remained colourless or pale, and which were usually mis-shapen or shrunken, were regarded as being sterile. Data for pollen sterility are given in Table 3. TABLE 3 Percentage sterility of Salicornia pollen. 2n = 18 (counted) | 2n = 18 (inferred from morphology) Shingle Street 10-5 | Colne Point S:5= 8-5 | Hayling Island 15-0 6:0 | 9:5 55) 6:8 3-0 Colne Point ils) | 2n = 36 (counted) | 2n = 36 (inferred from morphology) Blakeney 5:0 | St. Mary’s Bay 100-0** 2:0 1-0 Shingle Street 5:0 Dovercourt 10:33*> PDS) 7:5 0-0 | 5:0 Colne Point 82-0 | 3-0 Hayling Island PIS) | 2:0 (165) | * — 100 grains examined. ** _ 83 grains examined. *** _ ¢.150 grains examined. At least 200 grains examined from each plant unless otherwise stated. From these figures it is seen that normal plants show a fertility of between 90% and 100%, whether they are diploid or tetraploid. From this one may suppose that the tetra- ploids are, in general, allo- rather than autoploids. The markedly reduced fertility of some specimens indicates some genetic unbalance, possibly resulting from the hybridisation of related, but not identical, biotypes. Pollen tubes were seen in abundance on gynoecia dissected out and mounted whole, although efforts to germinate pollen artificially met with no success. Fresh stained dissec- Watsonia 5 (3), 1962. BRITISH SALICORNIAE 159 tions have revealed pollen tubes within the stylar tissues, and sections have shown traces of what are probably pollen tubes at the entrance to the micropyle. Such serial sections of flowers cut at and soon after pollination have cast little light on the actual process of the presumed sexual fusion. The embryo sac appears to have eight nuclei, and the embryo develops normally from the micropylar end. An endosperm is formed, though scanty in amount, as is correctly recorded by Volkens (1893), though overlooked by Moss (1914) and Tutin (1952). (6) Pollination Knuth (1909) quotes Schulz as saying that the flowers of Salicornia are feebly proto- gynous, but possess persistent stigmas so that in consequence of the proximity of the anthers automatic self-pollination is easily possible. Moss (1912) says ‘ The species of Salicornia are wind-pollinated; and hybrids are often abundant when allied species grow together.’ Ball & Tutin (1959) describe a species S. obscura with ‘ usually cleistogamous flowers.’ My own observations suggest that it is very likely that many of the annual forms are self-pollinated in nature, as ripe dehiscing anthers may be seen in contact with presum- ably receptive stigmas, and their pollen spilling on to the stigmatic papillae (see Fig. 1). Usually it seems that Salicornia is weakly protogynous, and sometimes it may be markedly so (as for example S. perennis). Often, however, stigmas and anthers appear simultaneously, and even when protogynous, pollen from earlier flowers on one part of a plant may be transferred to stigmas on later flowers on another part of the same plant. The possibility of wind-pollination sometimes happening must therefore be considered, though without an experiment using distinctive marker-genes, it may not be possible to prove. The experiment described below was carried out at Blakeney Point, Norfolk, to help in assessing the possibility of anemophily in Salicornia. Blakeney Point was selected because there is, on the ‘ Pel/vetia Marsh,’ a great expanse of almost pure Salicornia which comes to an abrupt stop on the northern edge of the marsh against sand-dunes, where the vegetation is almost solely Ammophila arenaria. On the southern side of the marsh there is about 4 to ? mile of bare mud or water, and then more saltmarsh with much Salicornia. A day was selected when the wind was blowing steadily from the south, and Salicornia on the marsh was shedding its pollen. Microscope slides coated thinly with gelatin contain- I5 @ H » S, ¥ sand dunes | shingle : salt Roe ' marsh - G mney 8 F Seed horizontal scale iY ee D GiB A (metres)5O ION ee ee ain: cr 79," Square cover glass fe) fe) number of chenopod pollen grains pe Or O 1200 Fig. 10. Wind dispersion of pollen along transect from saltmarsh to sand dunes, Blakeney, Norfolk. The lines marking the sampling points A to H are not drawn to scale vertically. Watsonia 5 (3), 1962. 160 D. H. DALBY ing a little acetocarmine were exposed vertically at about one foot above the ground at intervals along a transect running from saltmarsh to dune-crest (see Fig. 10). After 12 hours, the slides were taken indoors, coverslips placed in position and the slides warmed gently to melt the gelatin. The catch for total Chenopodiaceae-pollen is shown below each sampling point in Fig. 10. The catch at D is rather larger than might have been expected, probably because the freshening wind raised grains which had come to rest on the shingle (the slide was covered with wind-blown sand). In contrast, the catch at E is smaller than to be expected, as the wind was deflected clear of this sampling site on rising from the blow-out. This experiment certainly proves that pollen of Chenopodiaceae is carried by the wind, and knowing that Salicornia was by far the most numerous genus at anthesis on the marsh, we may well assume that the majority of this wind-blown pollen did in fact come from that genus. It should be mentioned that related genera in the Chenopodiaceae have very similar pollen grains, differing only very slightly in size and sculpturing of the exine. (c) Self-compatibility The compatibility relationships of several plants were examined by considering specimens gathered from different saltmarshes before the flowers had opened. The flowering branches were washed carefully to remove any grains that might be adhering, and the selected spikes were enclosed in small polythene bags tied loosely over cotton-wool plugs. Other branches were left as controls. When the control branches had set seed, the ‘bagged ’ branches were dissected, and the number of seeds set counted. In all instances counts were restricted to the central flowers of the cymules. The results are set out in Table 4. TABLE 4. Effect of ‘ bagging’ on setting of seed in Salicornia. | = = al Control Experimental Maximum potential | Number Maximum potential Number seed number | actually set seed number actually set S. ramosissima | Hayling Island | a — | — 21 7) b — | — | 22 18 Shingle Street | | | Cc 5 | 2 | 12 a d 13 8 | 18 | 10 e 16 11 | 26 | 12 S. europaea | Pagham | | f 22 15 ! 26 | 15 g 33 31 | 28 23 h PI 17 | 30 | 2 S. dolichostachya | | Blakeney | | i 10 | 10 23 18 28 13 It will be seen that ‘ bagging ’ has caused a slight but significant reduction in the quantity of seed set (17 = 5-93; 0-025 > p > 0-01). Assuming reproduction to be sexual, it is clear from the substantial quantity of seeds set that the plants used in the experiment were self-compatible. A further conclusion is that the physical effects of the ‘ bagging’ (eli- Watsonia 5 (3), 1962. BRITISH SALICORNIAE 16] mination of air currents near the anthers) have reduced the chances of pollination taking place. Thus though these plants are self-compatible, we may assume that in nature some pollen transfer takes place from flower to flower, and possibly from plant to plant. (d) Apomixis There still remains the possibility that some species of Salicornia may be apomictic, which might explain their taxonomically critical nature. Plants of S. europaea from Pagham, Sussex, were used in a simple experiment in which the stigmas were removed from 17 protogynous flowers by making a cut parallel to the seg- ment surface. 16 of the experimental flowers failed to set seed during a period when uncut control flowers did so. The single seed set may have been due to pollination having taken place before the experiment began. This evidence, though scanty, is against autonomous apomixis in S. europaea, but would not of course indicate the absence of pseudogamous apomixis. The evidence in favour of sexual reproduction is, however, strong, though not yet conclusive. (e) General conclusions The main significance of the work described above lies in its bearing on specific variation limits in the genus. Hybrids may sometimes occur, although as far as the writer knows there have not yet been any proven records of hybridity in this genus. Hybridisation could certainly be a physical possibility, as Salicornia pollen may be carried in considerable quan- tities for short distances by the wind. The reduced pollen fertility sometimes encountered is evidence for crossing, possibly between related, but not identical, biotypes. Reproduction is considered to be sexual, whilst the preponderance of self-fertilisation would be expected to produce more or less pure-breeding lines which could be regarded taxonomically as microspecies. Segregates from the occasional crosses between different lines could then act as the sources of new lines. This pattern of variation is wholly in keeping with the views expressed by Stebbins (1958) in considering the advantages to an annual species following self-fertilisation. ACKNOWLEDGMENTS My thanks are due to Professor J. Heslop-Harrison, who supervised my work at University College, London from 1951 to 1954, and to Professor W. H. Pearsall. Numerous others have helped me in various ways — in particular Mr. F. J. Bingley and Dr. F. Rose, to whom I am especially grateful. The work was carried out during the tenure of a post-graduate research studentship awarded by the Nature Conservancy, and certain extra field expenses were covered by a grant from the Central Research Fund of the University of London. The contents of this paper formed part of a thesis accepted for the degree of Ph.D. by the University of London. REFERENCES BALL, P. W. & TuTin, T. G. (1959). Notes on annual species of Salicornia in Britain. Watsonia, 4, 193-205. BAXTER, W. (1839). British Phaenogamous Botany. Oxford. Castro, D. DE & Fontes, F. C. (1946). Primeiro contacto citologico com a flora haldfila dos Salgados de Sacavém. Broteria, Cienc. nat., 15, 38. Danpy, J. E. (1958). List of British Vascular Plants. London. FRAINE, E. DE (1913). The anatomy of the genus Salicornia. J. Linn. Soc. (Bot.), 41, 317. HAMBLER, D. J. (1954). Chromosome numbers in British Salicornia. Nature, Lond., 173, 547. Knut, P. E. O. W. (1909). Handbook of Flower Pollination, vol. 3, tr. J. R. Ainsworth Davis. Oxford. Konic, D. (1939). Die Chromosomenverhiltnisse der deutschen Salicornien. Planta, 29, 361. KO6nic, D. (1960). Beitraége zur Kenntnis der deutschen Salicornien. Mitt. flor—soziol. Arb. Gemeinsch., N.F. 8, 5-58. Lupwic, W. (1950). Der Queller (Salicornia europaea) in der Wetterau. Natur u. Volk, 80, 176. Watsonia 5 (3), 1962. 162 D. H. DALBY MAUvDE, P. F. (1939). The Merton Catalogue. A list of the chromosome numerals of species of British flower- ing plants. New Phytol., 38, 1. Moss, C. E. (1912). The International Phytogeographical Excursion in the British Isles. XII. Remarks on the characters and nomenclature of some critical plants noticed on the excursion. New Phytol., 11, 398. Moss. C, E. (1914). The Cambridge British Flora, 2. Cambridge. Moss, C. E. (1948), in Barker, W. F. ef al., Plantae novae africanae, ser. 28. J. S. Afr. Bot., 14, 40. SALISBURY, E. J. (1940), in Huxley, J., The New Systematics. Oxford. STEBBINS, G. L. (1958). Self fertilization and population variability in the higher plants. Amer. Nat., 91, 337. TuTIN, T. G. (1952), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Cambridge. Tutin, T. G. (1959), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Excursion Flora of the British Isles. Cambridge. VOLKENS, G. (1893), in Engler, A. & Prantl, K., Die natiirlichen Pflanzenfamilien, 3, abt. 1a, 36. WitmotTT, A. J. (1939), in Martin, W. K. & Fraser, G. T., Flora of Devon. Arbroath. Wu rr, H. D. (1936). Die Polysomatie der Chenopodiaceen. Planta, 26, 275. WuLetF, H. D. (1937). Karyologische Untersuchungen an der Halophytenflora Schleswig-Holsteins. Jb. wiss. Bot., 84, 812. Watsonia 5 (3), 1962. STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. By D. M. BARLING Royal Agricultural College, Cirencester ABSTRACT Poa subcaerulea has a widespread distribution in Great Britain, being common in the hill country of Scotland and Wales, and in sand dunes along the British coast. It is essentially a plant of moist habitats. There is considerable variation within populations but all show taxonomic characters of great constancy. P. subcaerulea panicles are usually small, with a low number of spikelets, a tendency to two branches at the lowest node and glumes with three nerves. The rhizomes spread extensively and normally have one tiller per tillered node. Leaves are broad and may be hairy. Pollen diameter is from 30-6—37:°5 » and fertility is high. Seeds are long. Polyembryony has been found in all samples collected. Chromosome numbers varied between 2n = 54 and 2n = 109. Autopolyploidy is widespread and chromosome number varies within populations. Apomixis may be the main method of seed production, and together with vegetative reproduction by rhizomes gives considerable cloning. Taxonomically P. subcaerulea is quite distinct from other species of the P. pratensis group occurring in the British Isles, but it is essential to consider vegetative as well as floral characteristics in classification. Transplant studies have shown the great constancy of these taxonomic characters. INTRODUCTION The critical nature of Poa pratensis L. sensu lato is now well appreciated and this, together with its economic importance, has attracted much study to this group which is usually classified either into distinct ecological form complexes (e.g. Akerberg, 1942), or subspecies (e.g. Hiitonen 1933; Hylander 1941, 1953, 1955; Tutin 1952, 1957), or divided into several separate species (e. g. Lindman 1926, Roshevitz 1934; Hultén 1950; Hubbard 1954; Melderis 1955; Dandy 1958). The concept of the last-mentioned authorities is followed in this paper. Lindman (1926), who was the first to be particularly interested in this group, divided it into five species : P. pratensis L. sensu stricto, P. angustifolia L., P. alpigena (Fr.) Lindm., P. subcaerulea Sm. and P. irrigata Lindm. Hylander (1941, 1953, 1955) has treated these entities as subspecies of P. pratensis, and has included P. subcaerulea in subsp. irrigata (Lindm.) Lindb. f. On the specific level the epithet subcaerulea published by Smith in 1802 is much earlier than irrigata given by Lindman in 1905. More recently Clausen & Hiesey (1958) have added P. arctica R. Br. to this group, as it forms intermediates with P. alpigena in Sweden (Nygren, 1950). Three of these species, P. pratensis L. sensu stricto (P. pratensis L. subsp. pratensis), P. angustifolia L. (P. pratensis subsp. angustifolia (L.) Lindb. f.) and P. subcaerulea Sm. (P. irrigata Lindm., P. pratensis subsp. subcaerulea (Sm.) Hiitonen. P. pratensis subsp. irrigata (Lindm.) Lindb. f., P. pratensis subsp. alpigena auct. flor. brit.) are present in Britain. Both the other species, the low-arctic P. alpigena (Fr.) Lindm. (P. pratensis subsp. alpigena (Fr.) Hiitonen) and arctic-alpine P. arctica R. Br. have a circumpolar distribution, but neither occurs in the British Isles. Melderis has reviewed the taxonomic interest of P. pratensis, sensu lato, and with regard to P. subcaerulea has pointed out that no work has been carried out on this species in Britain, though Léve (1952) has investigated Icelandic populations and Akerberg a limited number of Swedish plants. Since then data on Welsh populations have been recorded (Barling 1957, 1959). 163 Watsonia 5 (3), 1962. 164 D. M. BARLING It would appear from Akerberg that P. subcaerulea is itself taxonomically very variable, for he describes a shore and a forest form, which hybridise with other forms of P. pratensis. Indeed the reticulate relationship of the various forms of P. pratensis, sensu lato, is well known and a similar pattern appears to exist between many Poae (Gustaffson 1947, Stebbins 1950; Nygren, Melderis, Clausen & Hiesey). The variety of taxonomic opinions on P. subcaerulea, together with the knowledge of hybridisation with other species of the P. pratensis group, not to mention the ecological specialisation of species, makes this group a necessary study for a fuller understanding of the complex relationships within the Poae. The present paper is an account of more extensive studies on populations and transplants of British origin. MATERIAL AND METHODS Populations of P. subcaerulea from centres distributed through most of its latitudinal range in the British Isles have been studied as follows :— v.c. 3, Slapton Sands and Dawlish Warren (sea level); v.c. 34, Hazel Hill and Five Acres (800 ft.); v.c. 41, Cwmdare : Lluestai Llwydion (900 ft.), Tir-Evan-Bach-Traws (600 ft.), Pwll Rhys (600 ft.), Tir Morris (750 ft.), Cefn-y-Gyngon (1,100 ft.), Windber (950 ft.), Bwlifa (950 ft.); v.c. 41, Pen Rhys (800 ft.), Craig-y-Llyn (1,900 ft.), Limeslade (sea level), Jersey Marine (Sea level); v.c. 42, Pontpren (600 ft.), Penderyn (600 ft.); v.c. 44, Pendine (sea level); v.c. 45, Maenclochog (800 ft.), Rosebush (800 ft.), Wolf’s Castle (300 ft.); v.c. 46, Ffair-Fach (1,000 ft.); v.c. 60 St. Annes (sea level), Bispham (sea level); v.c. 69, Grange-over-Sands, (sea level), Windermere (800 ft.), Kirkstone Pass (1,500 ft.), Knock Fell (2,200 ft.); v.c. 70, Dunmail Raise (2,000 ft.); v.c. 77, Bailliesmuir (300 ft.); v.c. 90, Montrose (sea level); v.c. 92, Balmoral (800 ft.); v.c. 98, Glenbeg (800 ft.). The transplants were grown in 8 in. pots immersed in a washed gravel bed situated at Cirencester (Lat. 51° 43’ N., Long. 1° 57’ W.), at an elevation of 440 ft. The soil used was a calcareous loam derived from the Great Oolite. DISTRIBUTION AND ECOLOGY P. subcaerulea was described by Smith (1802) from specimens gathered by Rev. H. Davies in Anglesey. In his English Flora, Smith (1824) mentioned that this species occurs in mountainous situations in Wales, Anglesey, Westmorland, Cumberland and Scotland. According to Hubbard, P. subcaerulea has a widespread distribution in the British Isles, though it is more common in the north than in the south. It is found in a variety of habi- tats : moist mountain grasslands, stream sides, dune slacks and sea shores, and the last environments have resulted in its distribution along the south coast of England. Outside the British Isles, this species has been recorded from Fennoscandia, Denmark, the northern part of the European U.S.S.R. and Iceland (cf. Hylander 1953). It seems that this species is common in N.W. Europe, from at least 50°N. to slightly within the Arctic Circle, overlapping over much of this territory with other species of the P. pratensis group. An excellent map showing the distribution of P. subcaerulea in N.W. Europe is given by Hultén. It is probably at its southern limits in the British Isles and is replaced to the south by the P. pratensis complex. This may account for its ecological prevalence in the hill or mountain areas as opposed to lowlands of the British Isles. Its distribution in the British Isles is incompletely known. Habitats, particularly in South Wales, are as follows :— 1. Permanent grasslands. The plant was found in enclosed and open hill grasslands, particularly where Agrostis tenuis or fine-leaved fescues (Festuca rubra and F. ovina) were dominant. In these pastures the plants are not very obvious, their long rhizomes and single tillers giving a very diffuse appearance. It is most noticeable in flower, or where conditions may lead to local concentration as on pathways. It has been recorded in tussocky swards of Molinia caerulea, usually as isolated plants with very long culms and rather poor vegetative development, and also amongst swards where Juncus effusus is prominent. Watsonia 5 (3), 1962. 165 STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. 00-0 + 00-1 v0-:0 + 70-1 70:0 + FIT 00-0 + 00:1 70-0 + 90-1 00-0 + 00-1 00:0 + 00-1 00-0 + 00-1 00:0 + 00-1 00-0 + 00-1 apou pasaipi /S42]]LL, bL:9 + 09-6LT ETS F OE-SL C681 + 8h-h87 S71 + SPSL 10:9 + 06-771 C8-€ + S6:78 66-11 + €v-861 OI-L F 00-9€1 SO-€I + O8-SST 09-ST + LL-LEI TI-O1 + 9v-001 60:8 + $9-LOI LI-L + Ob-8¥1 69-6 + 09-971 @8:1 + €0-LE 66:1 + OF-LZ CEGam Vere, 9V-7 + PI-9€ €8-1 + T€-672 69:7 F 69-6€ 61-7 + 69-06 OL‘ F EL-9V L€-% + 0-€€ 6S:% + C8-LZ 17-7 F S187 Sv:T + LI-€7 0¢-F * OF-0S O7:T + SE-1E “WU wjnd fo Jyslayy ajaiuvd dad sjajayids O1-0 F SI-Z ZL-0 + 06-1 60:0 + 07-7 80:0 + TE-Z 90-0 + SI-7 | 77-0 + 09-2 Jib 0) = CLG vI-0 + €9-7 | 01-0 + OI-2 | €1l-0 + 81-7 7-0 + LY-7 00-0 + 00-2 v1-0 + OE-€ ZL-0 + 09-2 apou samo] SayoUuDlg 6I-T + $8-Sbp O£€-1 + LO-ZE 80-7 + 9P-TS 00:7 + PI-Lv 19-T + CL-8€ L6:0 + 07-€P 68-1 F ZL-€¢ 69-T F OI-7S 19-7 + O0@:SP 8L-b + V7-8E Lv-t + Ov-0P 6L:1 + SE-LZ 97:7 F L9-7¢ €6:1 F 9P-9€ “UNU Yyjsua] °40] {uy Dansavaqns “q{ Ul suorje;ndod pyim wWosj eyep [esso;oydiopy *][ AIAV,], OSIOA PLOY puvlsseIH eunp poxI puvyjsseiIHn puvjsselH do} YO ounp poxI puvjsseiH puvyjssein PUL]SSeIS OSIOA PLOY Joyoys deoys diy aseA\ ouNnp PpoxI 9104S IDUIQUHT [elowyeg soquolH (solioyst{ Av) SSOIUO/] SSeq SUO01SyIDY UAT[SATOH weydsig s,ouuy 3S uvog jo \soloy ‘SOIOY OAL | uAIOpUuog | uA] ]-A-3IeIZ) uoBUAD-4-UJOD JOQPUIMA UdTIVAMA YSI|MeEg spurs uojdeyis uL3110 Watsonia 5 (3), 1962. 166 D. M. BARLING 2. Road-verge grasslands. On the hill areas the fertilising effects of road dust has led to the formation of improved swards a few yards wide in what is often rough grassland dominated by Nardus stricta or Molinia, and on these areas P. subcaerulea is frequently found. 3. Sheep shelters. Where localised shelter or ““ camping ”’ 1s taken by sheep, as under walls, banks or stunted hawthorn, then the heavy trampling and faeces concentration result in localised modification of the flora. On these small areas bare ground is obvious, and P. subcaerulea is frequently found, often in well developed mats, together with such species as Agrostis tenuis, Poa annua and Stellaria media. These modifications are often isolated amongst large acreages of rough grazings. 4. Waste places and walls. Plants have been found growing on walls at a variety of sites in north Glamorgan and south Breconshire. P. subcaerulea has been repeatedly found on tips of colliery waste, where the loose weathered shale allows the spread and develop- ment of long rhizomes and provides conditions similar to the sand dune environment. It is also found in swards developed on the stable areas of colliery tips where it may be accompanied by Agrostis tenuis and Festuca ovina. 5. Sand dunes. P. subcaerulea is frequently found on sand dunes, particularly on fixed dunes and slacks. In the present study plants have been collected from Dawlish Warren, Jersey Marine, St. Anne’s and Montrose. Plants are sometimes found in the shifting sand and also in the turf of nearby golf courses. It is also common on walls, cliff top grassland, and gravelly shores in these areas. At Dawlish it is found on fixed dunes and also in slacks among such species as Spiranthes spiralis. 6. Other habitats in which the plant has been found include the ballast of railway tracks, roadside gutters, garden pathways and hedgerows. PLANT CHARACTERS Wild populations were selected from a wide range of geographical and ecological environments to provide representative data (Table 1). Plants follow a regular pattern with short inflorescences and a low spikelet number. There was usually one tiller per tillered node, and rhizomes were difficult to excavate in dense swards, and much rotting of rhizomes was found. However, long rhizomes easily excavated were characteristic of sand dunes, and the somewhat similar waste-tip areas. Culms were short in grazed grasslands but where they were sheltered by unpalatable herbage, such as tufts of Juncus effusus, they might be very tall as at Cwmdare where heights of 510, 505, 460 and 425 mm. were found. The lengthening of fertile tillers is due to internode elongation which increases to a maximum in the node below the panicle. Two branches at the lowest node is common but 1-5 branches may be found in wild or transplant panicles. Data for transplants is found in Table 2. The number of panicle nodes seldom exceeds twelve (Barling 1959) and the glumes are both three-nerved. TABLE 2 Inflorescence data in transplants of P. subcaerulea a ee ee Onein Culm height No. basal NON nih cle Panicle length | mm. branches mm. Lluestai Ilwydion grassland 350-18 + 13-83 2:41 + 0-13 41-13 + 2:43 60-93 + 1-92 Craig-y-Llyn 408-60 + 17-08 2°42 + 0:14 43-72 + 3-89 70-17 + 3-42 Cefn-y-Gyngon 319-29 + 8-08 2°76 + 0:08 42-96 + 2:01 57:09 + 1-80 Knock Fell 316:57 + 10-39 2°48 + 0:08 40:73 + 2:84 | 62°87 + 2:98 Rhizomes are produced as early as February and actively during March, April and May. Internode length measured has varied from 15-33 + 0:59 mm. to 22:85 + 2:-45mm., which is usually shorter than that found in Poa angustifolia, and the number of tillers per tillered node is also less (Table 3). Watsonia 5 (3), 1962. STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. 167 TABLE 3 Tillers/Tillered Node. Origin No. tillers/tillered node 1 2 3 4 5 6 7 8 OF sO Mos 12s lise 1A P. subcaerulea Windber a. waste tip 116 12 8 - = a = va at = = ue is es b. path 122 20 4 ~ = - = = = ee a = as a Jersey Marine 70 4 = = a =a = wh 2 = = 3 a = Slapton Sands 356 5) = = = a ut Es - a “i ss an ai Dawlish | 360 19/ 1 _ = = = E = a » “s is x P. angustifolia Cirencester - ~ 3) 14 7 8 11 i 9 8 2 3 2 2 P. pratensis Cirencester 26 ”)\| 19 4 1 1 _ == = = u = = ee Steadings 4 8 14 14 6 2 2 = = = = a sas es Roots are actively produced from the end of February to early May. Leaves are produced from March onwards and the number of green leaves found at Windber was 3-9 + 0-3 in August, none in December and 1-48 + 0-2 in March. Hairiness of the blade is frequently found and is a ready means of finding biotypes. Leaf blade length increases then decreases on fertile tillers. Plants of P. subcaerulea that are kept in a heated greenhouse in autumn remain vegeta- tive the next season, but if kept in the open and then placed inside during January they head in the greenhouse. Heading is usually accelerated by the higher temperatures. There is quite obviously a response to late autumn low temperature and daylength similar to that reported for P. pratensis by Peterson & Loomis (1949). Inflorescence primordia have been found in late December and January. FLORAL BIOLOGY AND BREEDING MECHANISM Anthesis in P. subcaerulea commences about two thirds of the way up the panicle at about the seventh or eighth node. Other spikelets flower above this node, and then at lower nodes, flowering at the lowest node last of all. The basal floret of a spikelet opens first and the top floret often does not open at all. Details of anthesis have been described elsewhere (Barling 1959) and it follows, essentially, that of typical cross-pollinating species. Pollen grain size was measured as the greatest diameter, and for ten transplants had TABLE 4 Seed production in enclosed and open pollinated panicles of P. subcaerulea. Average seeds/panicle Seeds set/spikelet Origin No. selfings | enclosed open enclosed open Lluestai Ilwydion 1 3 68 66 PPA 2:6 Lluestai Ilwydion 3 3 47 60 | 1-2 3-0 Tir-Evan-Bach-Traws 1 | 5 66 70 | 2°6 2°8 Tir Morris 1 | 3 49 68 | ilP7/ 2:4 Windber 1 3 96 95 | 3-8 3-9 Bwilfa Farm 3 372 391 | 39 4-0 Penderyn 5 5 124 165 | PT) BPS) Watsonia 5 (3), 1962. 168 D. M. BARLING a mean of 34-66 » with a range of 26-45 wu, whilst 4 wild plants had a mean of 33-38 p and a range of 27-5-40y. Pollen fertility, determined by staining, was a mean of 89% in the transplants and 81-75% in the wild material. The seed production of open-pollinated and selfed panicles was determined for a number of plants (Table 4). The yield of many of the selfed panicles compares favourably with that found under open pollination of transplants or field material. SEED, GERMINATION AND POLYEMBRYONY, AND SEEDLINGS Seed length was longer than found in P. angustifolia, and hand threshed samples from six wild populations gave an average mean of 3-30 + 0-045, mm. and from six transplant populations 3-25 -+ 0-044. mm. Seed samples were collected from a wide range of sixteen centres in 1958 and were germinated on moist filter paper in petri dishes at room temperature and full illumination; twelve transplant samples were similarly sown. The wild samples totalled 1,485 seeds which gave 1,055 seeds with one seedling, 110 seeds with two, and 4 seeds with three seedlings. Polyembryony was present in all samples and amounted to 7:7%, with germination equal to 78:°8%. The transplant seeds totalled 1,174 with 758 single, 114 twin and 11 triplet seedlings, having 10-7% polyembryony and 75-2% germination. Selfed seed was obtained from various transplants and generally germinated well. However, some samples contained much light seed and a comparison with more normal- sized seed was made (Table 5). TABLE 5 Germination data for selfed seed samples of P. subcaerulea. Penderyn (thin) Slapton (full) Date Seedlings|seed [of Poly- | % Germ- | Seedlings/seed % Poly- |\% Germ- 1 2. 3 4 |embryony| ination 1 2 3 embryony | ination 10 Sept. 1958 (Seed sown) | | (Seed sown) | 25;Sept. 1958 9 6 - - 6 ke als (a3 10 1 11 84 29 Sept. 1958 13 q ~ — | a VemeZO lew 9 2 11 84 3 Oct. 1958 22 9 — — | 9 ee | 13 9 2 11 84 6 Oct. 1958 67 13 - =m 13 | 80 73 9 2 11 84 9 Oct. 1958 640 Ni V2 23 | 87 73 9 2 11 84 | | Fifty seeds of each known transplant of varied geographical origin were sown half an inch deep in John Innes Seed Compost, together with some from transplants of P. angustifolia. It was found that the first leaves of P. subcaerulea and P. angustifolia were very similar but that the second leaf was wider and generally shorter in P. subcaerulea. Width differences were quite clear to the eye in the second leaf stage (Table 6). The leaf blade colour of P. subcaerulea was of a darker or bluer green. Also in P. subcaerulea the first tiller was found exclusively in the axil of the first leaf whereas in P. angustifolia it was found in the second leaf in many cases. In P. subcaerulea the first tiller either pene- trated the sheath of the first leaf giving typical extra-vaginal tillering, or emerged sideways or took a very small angle pushing the sheath and blade towards the soil, whilst the second tiller was normally found in the axil of the second leaf and was intra-vaginal. In P. angusti- folia the tillering was exclusively intra-vaginal. This early extra-vaginal tillering was the most obvious of all the features recorded and revealed the characteristic tendency to a creeping as opposed to a bunched habit. Although the measurements of leaf blade width offered some hope of quantitative dis- crimination it was felt that for rapid work and single seedling identification it was unreliable in itself, and should be backed by colour features if discernible, as well as tiller behaviour. Watsonia 5 (3), 1962. STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. 169 TABLE 6 Seedling data in P. subcaerulea and some P. angustifolia | Ist leaf 2nd leaf Origin Width Length Width Length (mm.) (mm.) (mm.) (mm.) P. subcaerulea Cefn-y-Gyngon 0-82 36°25 1-44 46-00 Craig-y-Llyn 0-55 22:40 1-17 39-60 Knock Fell 0-68 20-14 1-02 36°40 Pontpren 0-76 29-40 1-38 44-50 Penderyn 0:60 22-00 1-26 34-20 Lluestai Ilwydion 0:66 36°20 | 1-13 32-00 P. angustifolia I. | Deer Park 1 0:55 18-83 0-89 43-67 Deer Park 2 0°55 23-38 1-00 46-00 Deer Park 3 0-51 23-83 0-89 46°67 CYTOLOGY In the present work mitosis has been studied in root tips and meiosis in pollen mother cells. The root tips were stained in aceto-orcein or Feulgen and then squashed and irrigated in the stain if necessary when the counts were being made. Metaphase plates invariably contained large numbers of small, densely arranged chromosomes. In some cases there was difficulty in determining the exact number, but counts were made of three cells for each plant. PMC smears were stained with aceto-orcein. Mitosis counts have given high numbers and it is possible to get widely different chromosome numbers within relatively small areas. Thus at Cefn-y-Gyngon, plants from different sheep shelters gave the different numbers of 70 and 98. Other numbers determined were Dawlish Warren 1, 2n = 92; Dawlish Warren 2, 2n = 84; Slapton Sands 1, 2n = 54; Cwmdare, 2n = 65; Lluestai Ilwydion grasslands 1, 2n = 72; Lluestai Ilwydion tip, 2n = 90; Tir-Evan-Bach-Traws, 2n = 105; Bwllfa Farm, 2n = 56; Limeslade, 2n = 92; Penderyn 8, 2n = 83; St. Anne’s Airport, 2n = 109; Grange over Sands, 2n = 98; Knock Fell 1, 2n = 84. At meiosis univalents are commonly seen at first division when they usually lag at anaphase, they may divide and some are excluded from the interphase nuclei. Multivalents are common at first metaphase and segregation of chromosomes is clearly uneven in some mother cells. At second division lagging is again common and exclusion has been seen in some cells in the tetrad stage. The pattern is identical with that illustrated for north American P. pratensis by Nielsen (1945, Figs. 22-46). The number of univalents for various transplants was: Slapton Sands 2:90 -- 0:18, Lluestai Ilwydion grassland 2-70 -+ 0-20, Penderyn 8, 1-85 + 0-16, Jersey Marine 1, 3:20 + 0-27, Knock Fell 2-80 + 0-20. In Icelandic P. subcaerulea, Love found chromosome numbers of 2” = 82-147. Within an area of 10 metres diameter he found plants with 2n = 82, 84, 91, 105, 111, 112, 113 and 119 chromosomes and provides the most detailed evidence of variation within populations. In addition, Akerberg has given numbers of 2n = 87 (for a form of P. pratensis which comes close to P. subcaerulea), 90, 90-95, for Swedish material. TRANSPLANTS Transplants have been much quoted already in the data on plant characters and it remains to mention the major features of genetic and taxonomic value derived from their study at Cirencester. An obvious factor was the genetic variation within and between populations. It would be difficult to show that the morphological variation revealed has Watsonia 5 (3), 1962. 170 D. M. BARLING any adaptive value, but it shows that apomixis where present is not a barrier to variation. The transplants have shown that many characteristics seen in the wild such as short culms, small spikelets and small panicles, are usually environmental effects. Thus greater values for the features measured have been recorded in transplants (Tables 1 & 2), and many of the extremes seen in the field are not seen under cultivation. When studied together the transplants from widely scattered localities show no great differences in gross morph- ology. It was not possible to detect differences according to altitude. Three-nerved acuminate glumes have been constant. Branching at the lowest panicle node has been variable within clones with a tendency to a greater number of branches than found in the wild. The number of spikelets, panicle length and number of panicle nodes all usually show much lower values than in P. angustifolia, but in some clones there are approaches in quantity to this species. These inherited tendencies are found within populations. The vegetative character of great constancy is the single tillering per node which has been recorded in all transplants. Hairiness of the leaf blade is a ready means of detecting population variability and it is inherited in clones. Hairs at the collar of the leaf blade are also constant in all transplants. Fig. 1 compares characters in P. angustifolia and P. subcaerulea transplants. — 3 ay ee 85 = 0] VA 8S 0 ae) HO Qs oD 3 wa 8S 3 : A an cab of XN 8 a0 — 50: | Tillers| Node 0 ; 30 | Panicle Length 0 in cm. — le 7" so — —~ lw a 100 Branches at 50 Lowest Panicle Node 0 In g2 tars) 4008S 30 Spikelets | [Panicle 0 i » Seed Length [ 0 in mm. i Dey Suey EW ERS EN Zh aL = = 0 50 Pollen Le geeed oper Me Diane i nee a Se OB ae Fig 1. Taxonomic characters in P. subcaerulea (solid black) and P. angustifolia (outline) Watsonia 5 (3), 1962. STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. 171 DISCUSSION One of the obvious conclusions that can be made from these studies is that P. subcaerulea forms a morphological and ecological complex of widespread distribution in the British Isles, possessing characteristics of great similarity to related forms. All the morphological features are important in the taxonomy of the group and are seen to rest on vegetative, inflorescence and seed characters. The outstanding vegetative feature is the production of single tillers on spreading rhizomes, which contrasts markedly with P. angustifolia (Barling 1959). Leaf blades may be hairy, a ready means of seeing variability in populations. The inflorescence is characteristically short, with a low number of spikelets and a tendency towards two branches at the lowest node. The number of nodes is also low and all these characters contrast with the values found in P. angustifolia (Barling 1959). Whilst these may be the general and focal features of taxonomy, the expected variability about these features is easily found. Panicles with spikelet numbers and branches similar to P. angustifolia may be seen though they are combined with the usual vegetative features of P. subcaerulea. Whilst these characters are more typical of related complexes, the decision to include such plants in P. subcaerulea is based on their presence within field populations of the group, their three-nerved acuminate glumes, and their conspicuous vegetative behaviour. The taxonomic difficulty is due to the retention of a basic type of inflorescence, long rhizomes, and hairy seed bases throughout the species of the P. pratensis group, in spite of the tremendous variability in ecological and physiological directions. The cytology of the group shows high polyploidy and there is a wide range of numbers and variation within the same habitat, features found in other populations of P. subcaerulea (Love, Akerberg) and P. angustifolia (Barling, Akerberg). If one considers the chromosome numbers of complexes in relation to their typical habitats, then the higher polyploids are of the wetter and more extreme climates. Such ecological variation within races has been reported by Winge (1940) and Jones (1958), amongst others. It is likely that increasing polyploidy in P. pratensis, sensu lato, has allowed some of its forms to survive in areas with more adverse conditions. However, the situation is far from clear as wide ranges of chromo- some numbers have been reported from within each of the distinct taxa (see Melderis), and also from within populations (Léve). Thus it would be necessary to consider the average polyploidy of British populations on an altitudinal basis, and also to compare these with data from more northerly areas (Léve). Hagerup (1931) proposed that since the majority of polyploids have a better adaptability than diploids then an increase in polyploids could be expected in the flora of areas of more extreme climate. In this work no discussion of the degree of polyploidy, only its incidence, was undertaken. Love & Love (1958) have discussed this aspect and whilst supporting Hagerup’s hypothesis, point out that the degree of polyploidy is only connected with distribution in exceptional circumstances. It appears that the apomictic Poae may present suitable material for studying the degree of polyploidy in relationship to environmental extremes, such as latitudinal and altitudinal distribution might involve. Thus the range of chromosome numbers reported by Love for Icelandic material is greater than that found in the more limited work of the present study. In spite of this high polyploidy there is little disturbance at meiosis, and the occurrence of univalents and multivalents at metaphase of the first division appears to be insufficient to affect pollen fertility. There is good seed production when inflorescences are selfed and this, together with the constancy of polyembryony in seed from all sources, is probably indicative of apomixis. Akerberg has shown sexual and apomictic forms in Swedish material and Love apomixis in Icelandic plants. The effects of this on the breeding system are similar to those discussed for P. angustifolia (Barling 1959), and the opportunity may now be taken of discussing this feature in relationship to P. pratensis as a whole. Léve has reported a range of chromosome numbers from 27 = 82-119 in a sample area of ten metres diameter. A similar situation, on a less detailed basis, is found in the present work. There is no difference in visible taxonomic characters associated with this range, though certain features such as pollen diameter, height, etc., do show correlation Watsonia 5 (3), 1962. 172 D. M. BARLING in Icelandic material. It seems that these taxonomic characters depend upon genes that are not lost by loss of chromosomes, and this is undoubtedly due to autopolyploidy of a high degree bringing about much repetition of basic hereditary material (multivalents at meiosis support this point). The same appears to hold for other forms of P. pratensis, sensu lato, and this autopolyploidy and its association with apomixis have been discussed in a thorough fashion by Darlington (1956). There is obviously much ecological and physiological differentiation between these forms of P. pratensis which is of a genetic nature and this may have occurred at a much earlier period than the inception of the present levels of polyploidy, much nearer to the presumably diploid and sexual ancestors. The present differences in average polyploidy that seem to occur between P. angustifolia and P. subcaerulea are not the only reasons for their distinctiveness. If they have chromosomes in common they also have complex genetic differences, as their ecology shows. P. subcaerulea is distributed widely in Britain in two major habitats of contrasting altitude, namely hill and mountain grasslands and dune slacks and shores. Both are essen- tially moist environments. Large gaps are nevertheless present in its distribution and these are probably the result of unfavourable habitat, though the under-recording that would arise from its critical nature may complicate the issue. The species is common in Icelandic grasslands and north west Europe, and is undoubtedly at the southern edge of its range, in much the same way as Lolium perenne is at its northern edge in the British Isles. This might account for its greater tendency, in the dense competitive grassland environment, to occur in hill areas, habitats of more extreme climate, and so behave as P. pratensis, sensu stricto, does in southern Europe and north Africa (Tutin) when it is at its southernmost points. The uncommonness of the form in the south of England (Hubbard), and its more or less restricted occurrence in the specialised maritime habitats, lend further support, and fits with many of Salisbury’s (1932) generalisations. Clausen & Hiesey suggest a latitudinal zonation of the forms of P. pratensis, with P. arctica, P. alpigena, P. subcaerulea and P. pratensis replacing one another from north to south. Whilst this is broadly true — for example P. arctica and P. alpigena are not known in Great Britain —-it is not completely so since in many areas they are sympatric. Much overlap occurs in Sweden (Akerberg) and certainly on a broad basis in the British Isles between P. angustifolia., P. subcaerulea and P. pratensis. This tendency to geographical, coupled with ecological, specialisation in the poly- morphic P. pratensis allows widespread colonisation and the development and maintenance of distinct physiological forms simultaneously. But the fact that the group is basically pastoral introduces a complexity of distribution since the variety of agricultural practices will lead to a modification of ecological barriers, and P. pratensis, sensu stricto, the main agriculturally useful form, will migrate to areas which, under more natural conditions, might be more suitable for others. Similarly, if there is a tendency for various pasture types to merge, there may be some areas in which the forms may be brought into very close contact, and even pollinating distance. In conclusion it appears that in Britain P. subcaerulea is a species with a tendency to inhabit damp environments, hill grasslands and fixed dunes and shores. It is relatively easy to distinguish the typical panicles from other forms of P. pratensis, sensu lato but variation is such that panicle details may be similar to P. pratensis or P. angustifolia and so it is necessary to examine glume and vegetative characters. However, work in re- lationship to P. alpigena is much needed. A fuller and more detailed account of the dis- tribution of P. subcaerulea in Great Britain would be of value in assessing the complex relationships of the other species of the P. pratensis group, and could indicate to what degree apomixis has permitted a fragmentation of distribution to small but suitable habitats. There is, even at this stage, much to suggest that P. subcaerulea and P. angustifolia are more or less extreme forms of P. pratensis, sensu lato, whether considered ecologically, cytologically or taxonomically. Watsonia 5 (3), 1962. STUDIES IN THE BIOLOGY OF POA SUBCAERULEA SM. 173 ACKNOWLEDGMENTS It is a great pleasure to record my thanks to Dr. A. Melderis for his interest and suggestions, as well as his help in the final stages of the work. Thanks are also due to Principal F. H. Garner for facilities, and to Mr. J. Stevens for aid from time to time. REFERENCES AKERBERG, E. (1942). Cytogenetic studies in Poa pratensis and its hybrid with Poa alpina. Hereditas, 28, 1-126. BARLING, D. M. (1957). Poa pratensis subspecies subcaerulea in N. Glamorgan and S. Brecon. Nature in Wales, 3, 429-433. BARLING, D. M. (1959), Biological Studies in Welsh Poa subcaerulea. Nature in Wales, 5, 774-780 BARLING, D. M. (1959). Biological studies in Poa angustifolia. Watsonia, 4, 147-168. CLAUSEN, J. (1952). New Poa hybrids, Carn. Inst. Wash. Y.B., 51, 111-117. CLAUSEN, J. and Higsey, W. M. (1958). Experimental studies on the nature of species. IV. Genetic structure of ecological races. Carn. Inst. Wash., Pub. 615. Washington, D.C. Danpy, J. E. (1958). List of British Vascular Plants. London. DARLINGTON, C. D. (1956). Chromosome Botany. London. GUSTAFSSON, A. (1947). Apomixis in higher plants III. Acta Univ. Lund., 43, 183-370. Hacerup, O. (1931). Uber Polyploidie in Bezichung zu Klima, Okologie und Phylogenie. Hereditas, 16, 19-40. HurTonen, I. (1933). Suomen Kasvio. Helsinki. HusBBarpb, C. E. (1954). Grasses, Penguin Books, London. HUuLtTeEN, E. (1950). Atlas of the Distribution of Vascular Plants in North-west Europe. Stockholm. HyLAnper, N. (1941). Foérteckning Over Skandinaviens vdxter. I. Kdrlvdéxter. Lund. Hy Lanper, N. (1953). Nordisk kdrlvéxtflora, 1. Stockholm. HyLAnper, N. (1955). Férteckning Over Nordens vdxter. I. Kdrlxvdéxter. Lund. Jones, K. (1958). Cytotaxonomic studies in Holcus mollis L. 1. The chromosome complex of Holcus mollis L. New Phyt., 57, 191-210. LinpMAN, C. A. M. (1905). Poa irrigata, en ny nordisk art af pratensis-typen. Botaniska Notiser 1905, 73-90. LINDMAN, C. A. M. (1926). in Holmberg, O.R., Skandinaviens Flora. 1 (2), 202-205. Stockholm. Love, A. (1952). Preparatory studies for breeding Icelandic Poa irrigata. Hereditas, 38, 11-32. Love, A and Love, D. (1957). Arctic polyploidy. Proc. Genet. Soc. Canada, 2, 23-27. MELperRIs, A. (1955). Species problems in recent Scandinavian works on grasses, in Species studies in the British Flora. B.S.B.I. conference report, No. 3, 140-159. London. NIELSEN, E. L. (1945). Cytology and breeding behaviour of selected plants of Poa pratensis. Bot. Gaz., 106, 357-382. NyoGren, A. (1950). Cytological and embryological studies in arctic Poae. Symb. Bot. Upsaliensis, 10, 1-64. PeTERSON, M. L. and Loomis, W. E. (1949). Effects of photoperiod and temperature on growth and flowering of Kentucky Blue Grass. Plant Physiol., 24, 31-43. RosHEViTZ, R. (1934). in Komarov, V.L., Flora U.S.S.R., 2, 388-390. Leningrad. SALISBURY, E. J. (1932). The East Anglian Flora. A study in comparative geography. Norwich. SmitH, J. E. (1802). English Botany, 14, 1004. London. SmitH, J. E. (1824). The English Flora 1, 126. London. STEBBINS, G. L. (1950). Variation and evolution in plants. New York. Tutin, T. G. (1952). Gramineae, in Clapham, A. R., Tutin, T. G. and Warburg, E. F., Flora of the British Isles. Cambridge. WincE, O. (1940). Taxonomic and evolutionary studies in Erophila based on cytogenetic investigations. Compt. Rend. Lab. Carlsberg, Ser. Phys., 23, 17-39. Watsonia 5 (3), 1962. BOOK REVIEW Flora of New Zealand. Vol. 1: Indigenous Tracheophyta: Psilopsida, Lycopsida, Filicopsida, Gymnos- permae, Dicotyledones. H. H. Allan. Pp. liv. -+ 1095, with 4 maps in colour on end papers and 40 figures in text. Government Printer, Wellington, 1961. Price £5 5s. Od. (New Zealand currency). The flora of New Zealand is remarkable for the high proportion of endemics, complex hybrid popula- tions, the large number of attractive species contributed to horticulture, and the impressive spread of intro- ductions. The publication of the late Dr. H. H. Allan’s new Flora of a country with a climate so similar to our Own is an event of considerable interest to British botanists. The previous most recent work is the second edition of Cheeseman’s Manual issued in 1925. The first volume of the new Flora describes 1457 species, whereas of the same groups Cheeseman described 1363. But statistics alone cannot convey the measure in the advance in knowledge of New Zealand plants in the last 35 years, or of the higher technical standards used in its presentation. Cheeseman’s book stood high amongst the most useful Floras of the world; Allan’s entirely new work is a great advance on Cheeseman and should take its place in every botanical library. Most of the features of the best Floras are included. There are dichotomous keys to families, genera and species — combined with clear warnings in the more difficult genera about their limitations. Original descriptions are cited for all species and varieties, and latin descriptions of new taxa collected together at the end of the volume. Characters are set out clearly, and measurements and numbers cover extremes as well as normal specimens. A most commendable feature is citation of type localities and the location of type specimens. Many of the latter are in Britain, and thus this feature will be especially valuable to workers in this country. The task of critically reexamining type material must have involved an immense amount of work, but it is probably in this that Dr. Allan’s work makes its biggest advance on Cheeseman’s; many misinterpretations have been corrected, while the very full critical notes added in small type to so many genera and species will pave the way for further advances. As an example of the value of this approach Allan’s account of Crassula (Tillaea) helmsii is appropriate, as this species was discussed at length in the last part of Watsonia. Allan gives the type locality as near Greymouth, and the type as collected by R. Helms in the herbarium of the Dominion Museum, Wellington, and refers to a paper by Court stating that this type has been compared with Victorian specimens of Crassula recurva (Hook. f.) Ostenf. and that no specific differences between the two were found. Full citations like this will save future workers much time, en- courage sound nomenclature, and facilitate future research. The present volume includes only species accepted as indigenous though this must considerably reduce its practical use as ‘“‘ a means of identifying plants,”’ which, as is pointed out in the preface is the “‘ primary purpose of a regional Flora.”’ Visitors and resident botanists will find many plants which are not mentioned in this book. Even aliens like the blackberries and sweet briar which occur in great quantity are excluded. If a Flora is to serve adequately its purpose as a means of identification it should include all established species to be found at the time of publication. To limit the scope to those regarded as indigenous is to ignore the changes which have taken place during the past two centuries, and to involve users in failing to find names for many of the plants they find, or, even worse, making erroneous determinations simply because they were unaware that they were dealing with an introduced species. This treatment is the more surprising as Dr. Allan wrote A Handbook of the Naturalized Flora of New Zealand published in 1940, and was responsible for other valuable publications on aliens. Even if it is intended to include description of introductions in the second volume — and Cheeseman listed them in an appendix — this will be less useful than including them in systematic order. Dr. Allan died in 1957 leaving a manuscript which included all the indigenous vascular plants except the monocotyledons. We are indebted to his colleague Miss Lucy B. Moore for filling in gaps, making final checks, seeing the volume through the press and adding a useful section of supplementary notes includ- ing abstracts of very recent papers. It is to be regretted that there is no reference in the preface or on the dust-cover to any arrangements for completing the work. Monocotyledons form such a relatively small proportion of the New Zealand Tracheophyta that the labour of preparing Volume 2 should be very much less than for the volume now in our hands. The selection of type and standard of printing reflect great credit on the Government Printer, but it is unfortunate that such thin paper was chosen. This is too flimsy for satisfactory use in library or herbarium, and unpractical for consultation in the field. Portability is extremely desirable for a Flora but the fact has to be faced that a paper has yet to be invented which will give 1150 pages of print easy to handle in a volume less than one inch thick. From the heavy paper of Cheeseman, the Government Printer has gone to the other extreme in company with other publishers of modern floras — it should not be difficult to find a paper of intermediate weight which would be more serviceable. Dr. Allan and Miss Moore have given us a first class new Flora which will be valued by workers in many countries. J. E. LOUSLEY 174 Watsonia 5 (3), 1962. ERRATA 175 p. 65. Line 6 from bottom of page, for Mullett read Mullet. p. 66. Line 1, for Mullett read Mullet; line 3, for Glosh read Glash; line 4, for Mollet read Mullet Watsonia 5 (3), 1962. a : Seca Rcrips set abe ig cat « see 4 Ca - \ pled =| ‘ , - y ® r AS vt < - ‘ } = \ : < f Z ’ cd ‘ - ‘ fx a, teat om 4 ' me od - e a es c i mn 7 fe we r. at VO a 5 ee ee i mre — re he sir aie ana > is aad pes eye cp a i a , Flora of the British Isles A. R. CLAPHAM, T. G. TUTIN & E. F. WARBURG The first edition of the standard British Flora appeared in 1952, and has been reprinted several times. This new edition, largely rewritten and completely reset, embodies the results of research in the last eight years. Most of the larger keys have been modified for easier use, and accounts of introduced plants have been extensively altered. As the Excursion Flora (1959) now satisfies the need for a pocket edition, the page size has been increased. SECOND EDITION. 70s. net from all bookshops CAMBRIDGE UNIVERSITY PRESS IRISH NATURALISTS’ JOURNAE A Magazine of Natural History Published Every Quarter by the I.N.J. Committee. Edited by Miss M. P. H. KERTLAND, M.Sc., with the assistance of Sectional Editors. Annual Subscription, 10/- post free. Single Parts, 3/6. All communications to be addressed to :— The Editor, Department of Botany, Queen’s University, Belfast. INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the taxonomy or distribution of British vascular plants or Charophytes are invited from both members of the Society and others. They should be typewritten, with wide margins, double spaced, on one side of the paper only; contributors are recommended to keep a carbon copy of their typescripts. The form adopted in recent parts of Watsonia should be used for layout, headings, citations and references. Contributors are urged to avoid very complicated hierarchies of headings and sub-headings, and to check carefully the consistency of those that they use. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British flowering plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (N.H.) and B.S.B.I., 1958), and may then be cited without authorities. Otherwise, authors of names must be cited, at least on the first occasion where they appear in the text. Except for citations of the place of publication of plant names, full references should be listed in alphabetical order of authors names at the end of the paper. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals, 3rd ed. (1952). References to herbaria should include the abbreviations given in British Herbaria (B.S.B.1., 1958) and Index Herbariorum 3rd ed. (I.A.P.T., 1956). Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new taxa, and the main conclusions. Line-drawings should be boldly drawn in Indian ink on Bristol board or similar smooth white card, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. Graphs can be reproduced from originals on graph paper with faint blue ruling (not grey or any other colour), but all lines to appear on the finished block must be inserted in Indian ink. Lettering on line-drawings and graphs should be inserted lightly in pencil, and will be finished in uniform style. If an illustration includes plant-names or place-names, it is advisable to type these clearly on a separate sheet of paper. Photo- graphs can only be accepted where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable for the necessary reduction. It should be remem- bered that the finest detail on the originals may be lost even on the best half-tone blocks. If in doubt about the citation of names or references, or the presentation of illustrations or tabular matter, contri- butors are advised to consult the editor before submitting their typescripts. Authors will normally receive both galley-proofs and page-proofs. Particular care should be given to the thorough checking of references and tables. Alterations in page-proof should be avoided as far as possible, and any words or phrases deleted replaced by others of equal length. Twenty-five separates are given free to authors of papers. Further copies may be obtained at the printer’s current price, and must be ordered when the proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to the Editor, Dr. M. C. F. Proctor, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. A BIBLIOGRAPHICAL INDEX OF THE BRITISH FLORA COMPILED BY N. DOUGLAS SIMPSON, »™.A., F.LS. The purpose of this work is to provide references to sources of information whereby flowering plants, vascular cryptogams and Charophytes found in Britain may be identified, their history traced and their geographical range determined. In addition, information is provided on plantlore, local names, poisonous plants and weeds. The work has been in preparation for nearly twenty years and contains over 65,000 entries, including references to books, articles and manuscripts relating to the flora of the British Isles from the fifteenth century to the present time. Demy 4to, bound, 448 pages in double column. Limited Edition of 750 copies. PRICE £3 I5s. (postage extra) Obtainable from the compiler at ‘‘ Maesbury,’ 3 Cavendish Road, Bournemouth, Hants. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANDy for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf. of a sub-committee of the Society appointed for this purpose. Dr. George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and will without doubt become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8 vo., 176+-xvi pages. Bound in cloth. Price : 10/- (postage extra). Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. A DARWIN CENTENARY Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the very successful Conference arranged by the Botanical Society of the British Isles in 1959. The contributors include Lady Barlow, G. D. H. Bell, Prof. T. W. Bocher, Prof. W. H. Camp, Prof. N. Hylander, Prof. I. H. Manton, Prof. T. G. Tutin and Prof. D. H. Valentine. Demo 8 vo., 140 pages, with 7 half-tone plates and 1\ maps and figures in the text. Bound in buckram. PRICE £1 (postage extra) Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7 The Salisbury Press Ltd., Salisbury —31724 W3 4 e- = =6EMhWWAT SONIA JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 5 OCTOBER, 1962 Pt. 4 CONTENTS THE TAXONOMY OF POLYGONUM AVICULARE AND ITS ALLIES IN BRITAIN. By B. T. STYLES... Ke Re bie ae at ws oc Ab 177-214 NOTES ON BRITISH HIERACIA. II. THE SPECIES OF THE ORKNEY ISLANDS. By P. D. SELL and CYRIL WEST ae Hs oe se bs a 215-223 A STUDY OF VARIATION IN EUPHRASIA BY MEANS OF OUTDOOR CULTIVATION. By P. F. YEO ee ee ee ee eo ee eo ee ee 224-235 THE BRITISH FORMS OF TUBERARIA GUTTATA (L.) FOURREAU. By M. C. F. PROCTOR 236-250 LUZULA X BORRERI IN ENGLAND. By JOHN E. EBINGER un < ae 251-254 BOOK REVIEWS is ae - oa Ss oe ve a a 255-258 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH MuszuM (NATURAL History), LONDON PRICE FIFTEEN SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES Patroness: H.R.H. THE PRINCESS ROYAL OFFICERS President : J. EK. LOUSLEY Vice-Presidents : J. E. DANDY, Miss C. M. ROB, Dr. E. F. WARBURG and Dr. S.M. WALTERS Hon. General Secretary: J. G. DONY, 9 Stanton Road, Luton, Beds. Hon. Treasurer: J. C. GARDINER, Thrift House, 12&14 Wigmore Street, London, W.1. Hon. Editors: (Watsonia) M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT Hon. Meetings Secretary: H. J. M. BOWEN, Pomander House, Harwell, Berks. Hon. Field Secretary: P.C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Longfield, Kent. Hon. Junior Activities Secretary: P. F. HUNT, The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey. —=— Applications for membership should be addressed to the Hon. Assistant Secretary, D. H. KENT, 75 Adelaide Road, West Haling, London, W.13, from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 15/— per part. To be obtained from E. B. BANGERTER, Botanical Society of the British Isies, c/o Dept. of Botany, British Museum, (Natural History), London, S.W.7. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/- for Junior Members, and 10/— for Family Members, and the Society’s year runs from January Ast to December 3ist. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13 THE TAXONOMY OF POLYGONUM AVICULARE AND ITS ALLIES IN BRITAIN By B. 2. STYLES Commonwealth Forestry Institute, Oxford+ ABSTRACT The taxonomic history of Polygonum species in the section Polygonum (Avicularia) relating to the British flora is outlined. The merits of the various morphological characters used by authors to separate taxa in previous treatments is discussed. Evidence is produced, supported by results of biometrical, ecological and cytological studies, that the British and W. European forms of the weedy Polygonum aviculare, sensu lato comprise four species and a reclassification of these is proposed. Problems in the related British species in the group are discussed and clarified. A key to the British species is included. CONTENTS 1. HISTORICAL INTRODUCTION a, ts a: las Aes ot Lee 2. ANALYSIS OF TAXONOMIC CHARACTERS ey ny hy ss re 182 3. BIOMETRICAL STUDIES . 43 Me re i! we a 186 4. CYTOLOGY AND BREEDING SYSTEMS .. Ae “ i he, ae 200 5. ‘TAXONOMIC REVISION AND DISCUSSION ie ag: ait Se on 203 ACKNOWLEDGMENTS REFERENCES 1. HISTORICAL INTRODUCTION As the nomenclatural problem encountered in the taxonomy of P. aviculare, sensu lato is complicated and confused, a more detailed historical taxonomic sketch than usual of the species concerned is necessary for a full understanding of the problem as it exists today. This however includes only the barest detail relevant to the species described. Linnaeus (1753) includes P. aviculare in a division of the genus : **Polygonum foliis indivisis, floribus octandris.’ As the other sections are given names by him, e.g. Persicaria, Helxine, Bistorta, it seems clear that he intended Po/ygonum as the sectional name for this group. In it he includes the following species: P. aviculare, P. erectum, P. articulatum and P. divaricatum. Of these P. aviculare is the best known, the commonest and also one which stands first in the group. It is therefore proposed to make P. aviculare the type species of the genus and Polygonum should be regarded as the name of the section according to the Jnternational Code of Botanical Nomenclature (1956), replacing later names. The section was named Centinode by De Candolle (1815) with a brief description which was later amplified by Meisner (1826) who called it Avicularia. None of the characters given separate this section sharply from others in the genus. However, there should be little difficulty in recognizing species belonging to it; the axillary flowers alone are a sufficient guide. There is a tendency for certain of the species, e.g. P. arenarium Waldst. & Kit. and P. ramosissimum Michx., to have the flowers aggregated at the upper nodes only of the leafless branches, forming long spike-like racemes. The fruit characters of these species are however similar to those of the more typical species of this section. +Formerly of the Botany School, Oxford LT Watsonia 5 (4), 1962. 178 B. ff. (STYLES Polygonum aviculare, sensu lato Linnaeus (1753) used the binomial P. aviculare and gives the following very brief description : ‘Polygonum floribus octandris trigynis axillaribus, foliis lanceolatis, caule procumbente herbaceo.’ This is an extremely broad description of the plant and would include all the species of the section. The closely related P. maritimum appears in the group headed ‘ Persicaria pistillo bifidis,’ but the diagnosis is : ‘ Polygonum floribus pentandris trigynis axillaribus, foliis lanceolatis, caule stipulis obtecto fruticoso.’ Linnaeus also gives 4 varieties to P. aviculare, which are impossible to relate to any species which have been subsequently described in the aggregate. The first “ species’ to be separated from P. aviculare L. was P. littorale. The original place of publication of this binomial is often wrongly cited. H. F. Link (in Schrader 1800) states in a letter written while on a journey to Portugal that he has seen a variety of P. aviculare which is similar to P. maritimum, and which has thick juicy leaves and the branches covered wholly with sheaths. This note of Link’s does not however contain any name but some authors have referred to it as being the place where P. /ittorale was first published. The first publication of the epithet /ittorale is in fact by Persoon (1805) when he gives *P. littoraley with the following description, *foliis crassis succulentis, caule minus ramoso.’ He cites Link’s original plant in Schrader. Link (1821) later described P. /ittorale with: Caule procumbente suffruticoso, foliis oblongis acutis venosis carnosis, vaginis ciliatis internodiis multo brevioribus. Hab. in Europae maritimis. Caulis suffrutescens angulatus. Foliis petiolo brevissimo lamina 10 ‘‘5”’ acuta. Cor. alba. Vulgo pro P. maritimo colitur at hujus vaginae internodiis aequilia. A. P. aviculari differt foliis latioribus magis carnosis, caule suffruticoso.’ From this one learns that the main differences between P. /ittorale and P. aviculare are slight, as it is said to differ in only the broader fleshy leaves and the woody or shrubby stem. P. maritimum differs in having ochreae as long as or longer than the internodes. Since this time the plant has sometimes been given the status of a species, sometimes as subspecies or as a variety. Koch (1837) reduces it to that of a variety of P. aviculare whereas Meisner (1826) gives it the status of full species. It is difficult to visualize the kind of plant which Persoon first described and its origin is not quoted in the diagnosis. Many authors of later works have included it under P. maritimum. Lindman (1896) contends that Persoon may have in fact described a form of P. maritimum. Babington (1922), Lindman (1912), Moss (1914) and Rechinger (1958) follow Koch in retaining the plant as a maritime variety of P. aviculare. Tutin (1952) retains it as a ‘ microspecies ’ in the Polygonum aviculare aggregate distinguished by characters of the fruit. It is interesting to note that in the last five years A. & D. Love (1956 a, & b), Chrtek (1956), Scholz (1959) and Pauvels (1959) do not mention this plant, and its status as a species may be taken as doubtful (see later). I have made exhaustive attempts to trace the type specimen of P. /ittorale on the Continent. Unfortunately, Persoon’s herbarium is apparently lost and almost all Links’ specimens at the Berlin-Dahlem Herbarium, including the Polygonaceae, were destroyed during the last war. A specimen named by Koch P. aviculare var. littorale has been seen from the Rijksherbarium, Leiden (L), which fits P. aviculare, sensu stricto. The specimen exhibited the characteristic fleshiness of the leaves and stems associated with plants growing under halophytic conditions. It seems necessary, therefore, to treat P. littorale as a nomen dubium. tSpecies given by Persoon with an asterisk appear to him to be uncertain or obscure Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 179 An investigation of British maritime forms in P. aviculare sensu lato has shown that no plants answering to the description of P. /ittorale Pers. exist in this country. A number of herbarium specimens labelled under this name exhibited great morphological differences and were a mixture of either P. aviculare, sensu stricto, or P. arenastrum. Living plants of the group collected by myself had rather fleshy, glaucous stems and leaves, but these phenotypic differences did not persist in cultivation. Persoon (1805) also named plants collected by Thiébaud in the P. aviculare group as *P. monspeliense with the following description : *P. erecta, foliis ellipticis crenulatis, caule erecto stipulis scariosis incisis. Hab. Monspelii; Thiébaud. Distincta species videtur.’ Many later authors in floras have cited Thi¢baud as the authority for the name but this is incorrect. It is the earliest name for the taxon P. aviculare, sensu stricto (P. hetero- phyllum Lindm.) if the Linnean name is rejected for any reason (see p. 205). I have obtained a photograph from the De Candolle Herbarium at Geneva (G—DC) of P. monspeliense collected by Thiébaud, and itis typical of plants which I today call P. aviculare, sensu stricto. This name will be considered iater. Besser (1822) described a species that he collected in * Volhynia and Podolia’ (now the Ukraine, Russia), which he called Polygonum neglectum. He gives the following des- cription : ‘P. floribus 7 octandris, trigynis, subspicatis, foliis lanceolatis, ochreis laceris, internodiis brevioribus, caulibus diffusis patentibus ramosis mihi.’ There have been several different interpretations of this description and it has been the centre of much confusion. Although the plant has never been listed as a member of the British flora, a discussion of its taxonomy is relevant from the point of view of nomencla- ture. Besser does not describe the fruit of the plant he collected and later authors describe it differently. Ledebour (1850) gives P. neglectum as a synonym of P. arenarium Waldst. & Kit., whilst Meisner (1856) includes it pro parte under this species and pro parte under his own P. aviculare var. angustissimum. Boreau (1857), on the other hand, suggests that it might be an earlier name for Jordan’s P. rurivagum. Several of the more recent treatments of the variation in this group have included P. neglectum as a full species. Notable examples of these are Komarov (1936), Lembke (1948), A. & D. Love (1956) and Scholz (1959). Whilst A. & D. Léve record the plant as being common in Canada, the other three authors are all concerned with its occurrence in eastern Europe. This species will be discussed later. The Russian botanists Meyer & Bunge (in Ledebour (1824)) first described P. oxy- spermum from the island of Osel in the Baltic. Their description is : *Foliis lineari-lanceolatis subeveniis, floralibus diminutis, ochreis laceris internodiis brevioribus, caule ramoso, ramis erecto-patulis, corollae laciniis apice concavis patulo-conniventibus, semine acuminato corollam superante.’ P. oxyspermum is kept as a species by Meisner in De Candolle (1856), and is placed in the herbaceous section with P. aviculare, whereas the very closely related P. raii is placed in the suffruticose section with P. maritimum. The name appears in several of the older east-European Floras. A plant was named P. acadiense by Fernald (1914) as an endemic in N. America but Ostenfeld (in Fernald (1916)) states that after examining Danish and Scandinavian material labelled as P. raii he discovered that many of the plants referred to this were in fact P. acadiense. He found it to be widely distributed along the coasts of the Baltic and gives records for Denmark, Sweden, Gulf of Riga and arctic Norway. True P. raii is according to him confined to the channel coasts (England and France) and neigh- bouring areas. Samuelsson (1931) identified P. acadiense as conspecific with P. oxyspermum after examining the type of Meyer & Bunge in Leningrad. P. oxyspermum was recorded in Great Britain during the last century and will be discussed later. Watsonia 5 (4), 1962. 180 By Gi. SIMEES A ‘species’ of a most dubious nature, P. robertii was described by Loiseleur-Deslong- champs (1827). Specimens were gathered by a M. Robert from a locality near Toulon, France and sent to Loiseleur for determination. The description is : ‘Planta caulibus herbaceis procumbentibus diffusis, foliis ovato-lanceolatis virentibus, stipulis mem- branaceis acutis, floribus 2-3 axillaribus pedunculatis.’ No indication is given of the character of the fruit, but from Rouy (1910) one gathers that the main differences between this plant and P. aviculare are to be found here. In P. robertii the fruits are small and shining, and the plant is said to be a biennial or perennial. This species is of interest to British botanists in that the name has been frequently quoted as a synonym of P. raii Bab. and has been taken by some to be the correct name for this species. C. E. Britton, however, commenting on a specimen gathered from the Killiney Sands, Co. Dublin (v.c. H21) and labelled P. raii Bab. 1920, mentions : * The two names are not synonymous. P. robertii 1s a plant of the Mediterranean area apparently not reaching the British shores, and distinguished by its biennial or perennial root. The achenes are much smaller than in P. raii and polished almost as much as P. maritimum, Babington (1836) described P. raii* as a new species from plants sent to him from Marazion Bay, Cornwall, as follows : ‘ Caule procumbente herbaceo, ochreis 2-partitis ovatis demum laceris venis paucis distantibus simplici- bus, floribus axillaribus, caryopside laevissima perianthio longiore.’ He was considerably confused about the existence of P. robertii and mentions in a letter to W. J. Hooker ‘ P. robertii appears to be a plant that is quite unknown to all botanists except its author and I am quite unable to state whether it is the same as my [Ps (ROMs No authentic material of P. robertii could be traced in France. Specimens in the Herbarium, Laboratoire de Phanérogamie, Paris (P) labelled as such were not P. raii but maritime forms of P. aviculare, sensu lato. However, Reynier (1905) states ‘ Mutel, ayant examiné dans Vherbier de Robert la plante authentique, la décrit variation a tiges herbacées moins dures du P. maritimum, Lindman (1904) mentions that a specimen of this ‘species’ in the State Museum, Stockholm, labelled P. robertii Loiseleur, Toulon, * belongs decidedly to P. aviculare s.s.’ I feel sure that M. Robert’s plant was merely a form of P. aviculare with rather more shining fruits than normal. As no authentic specimens of this species are extant this name must be treated as a nomen dubium. Boreau (1857), incorporating the researches of Jordan, threw the olen! of nomen- clature in the P. aviculare aggregate into even greater confusion by adding five new names and elevating other varietal names to specific status. The four binomials of Jordan (previously unpublished) are P. agrestinum, P. humifusum, P. microspermum and P. ruriva- gum, whilst Boreau himself described P. arenastrum as a new species as well as P. polycnemiforme (Lecog & Lamotte) Bor. based on P. aviculare var. polycnemiforme Lecog & Lamotte, and P. denudatum (Desv.) Bor. based on P. aviculare var. denudatum Desv. Two of the binomials given by Boreau are relevant to the work on the British plants. These are P. rurivagum Jord. and P. arenastrum Bor. The relevant parts of his description are given here for purposes of typification. * P. rurivagum Jord. ex Bor. Tige de 1 a5 dec. gréle, flexueuse, dressée .... a nervures saillantes, gaines brun-rougeatres. .. . fruit brun mat, un peu luisant, finement chagriné, trigone, a faces excavees, ovoide, aigu au sommet . . . Champs sablonneux apres la moisson.’ ‘ P. arenastrum Bor. Tiges trés rameuses a rameaux longs étalées, sur le sol, tous trés garnis de feuilles rapprochees, petites, ovales ou ovales oblongues, gaines courtes. . .. fruit brun chatdain, assez luisant, court, terminé en pointe. Sables, graviers etc.’ *P. rayi or raji are mistaken orthographies adopted by many Continental, Scandinavian and American botanists, but no employed by Babington in his original diagnosis Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 18] These two names have been taken up as being the correct ones for taxa in the British flora. Lange (1880) after a botanical expedition to Greenland described a new variety P. aviculare var. boreale which was later elevated by Small (1895) to specific rank. Lange’s Original description is as follows :—- ‘Foliis petiolatis, 4-1 pollicaribus, elliptico-obovatis, obtusiusculis; glomerulis 2—5 floris, floribus pedicellatis, pedicello fere perigonii longitudine; ochreis latis, acutis v. obtusiusculis, apice fissis.” North America, Siberia, Iceland, Faeroes. Druce in this country was much puzzled by populations which he encountered from the Shetland Islands, subsequently called by him (1922) P. aviculare var. grandiflora, and this led me to an investigation of the plants of this group occurring in these islands. A number of populations have been obtained from the Shetlands and most of these on investigation have been found to form a distinct taxon. The plants agree with Lange’s type collection in the petiolate leaves and in the large fruit and persistent perianth. They have also been compared with gatherings from Iceland, Greenland and Northern Scandinavia. It is regarded here as a distinct species, P. boreale, new to the British Isles. An effort to produce a more satisfactory taxonomic treatment for the species occurring in Europe was attempted by Lindman (1904, 1912). He studied the plants over a greater area of their range than any previous workers and made the group the object of special investigation to which many more recent authors have referred. He divides the group into three species, P. calcatum Lindm., P. heterophyllum Lindm. and P. aeguale Lindm., with a number of varieties. He first separated P. calcatum (1904) and later (1912) published the other names in dividing up the forms remaining under P. aviculare. It is regrettable, however, that the differences between P. calcatum and P. aequale are not clear from his description. Lindman states furthermore, that P. calcatum is as common in Scandinavia as P. aviculare, (including P. aequale). The relevant features of Lindman’s description of P. calcatum are : ‘ Annual, prostrate, leaves tending to be small elliptical, ovate or obovate. . . Peianth united to the middle, compressed, tubular; segments pale green edged with white. Stamens 5. Nut small 2-23 mm long, broadly ovate-lanceolate, gradually narrowing at the tip, compressed trigonous, the sides convex, smooth. Those of P. aequale are : ‘Annual, stem erect or prostrate, leaves of main stem and flowering branches equal or differing in size, all commonly the same shape. .. Perianth segments equalling the length of the fused base or twice as long. .. Persistent perianth as long as the fruit or a little shorter. Fruit trigonous c. 2-3 mm long, ovate or lanceolate, black.’ Lindman’s other species, P. heterophyllum, is a more satisfactory entity, and is distin- guished from the others by the heterophyllous nature of the leaves and by the fruit which is larger, generally up to 3 mm, with three dull, concave sides. Lindman gives five varieties of the latter with the name rurivagum as one of them. Hybrids between all species of the group are said to occur. It is presumed that Lindman used new binomials for these taxa on the grounds that the Linnean Polygonum aviculare is a nomen ambiguum, but nowhere in the literature does he state this. Moss (1914) follows Lindman very closely in his extensive treatment of the group in Britain but raises P. rurivagum to full specific rank. Tutin (1952) in our most recent Flora treats the Linnean P. aviculare as an ‘aggregate species’ and includes the following as segregates, P. aviculare L., sensu stricto (= P. hetero- phyllum Lindm.), P. littorale Link, P. rurivagum Jord., P. aequale Lindm., P. calcatum Lindm. His account is based on that of Moss (1914) but less detail of the morphology and distribution of the plants in Britain is given. The four latest studies by specialists on the taxonomy of this difficult group have all appeared since this present work was begun in 1956. The conclusions of these workers differ to some extent among themselves and from my own. There is, however, some measure of agreement over such issues as the number of taxa into which the variation pattern of Watsonia 5 (4), 1962. 182 B. T. STYLES P. aviculare in the wide sense can be divided if allowance is made for the areas in which these taxonomists have worked. There is, however, still great confusion and difficulty over the question of the correct binomials which are to be used in naming these taxa. Every effort has been made to discuss the more difficult problems with them. These dis- cussions have helped in reaching the conclusions included in this present treatment, which I believe to be the correct ones. A. & D. Love (1956) have worked on Canadian plants and base their reasons for the delimitation of taxa mainly on cytological evidence. They give little detail of the morphology of the species they include and this has made exact comparisons with them and the British taxa difficult in most cases. These authors retain P. heterophyllum Lindm., P. aviculare L. (= P. aequale Lindm.) and P. rurivagum Jord. (!). They include also P. neglectum Bess. and state that it ‘has been introduced, and is common in Canada.’ P. calcatum Lindm. is reduced to a subspecies of P. aviculare L. P. boreale is treated as a subsp. boreale of P. heterophyllum. No reason is given by A. & D. Love for equating P. aequale Lindm. with P. aviculare L. The interest of the work by A. & D. Léve lies in the fact that they are the first workers in this field to attempt a complete correlation of cytological observations with morpholo- gical differences in these plants. Their work, however, contains several counts of chromo- some numbers which are not consistent with those made on the same taxa in this country. A discussion of these differences will appear later. The treatments of Chrtek (1956) and Scholz (1958-59) may be considered jointly since they have both worked in eastern Europe. Although the rank of the taxa they pro- pose is considerably different and a new binomial is introduced their results are similar. Chrtek proposes four groups at the level of subspecies. These are subsp. monspeliense Pers., subsp. rectum Chrtek, subsp. calcatum (Lindm.) Thell. (new to Czechoslovakia) and subsp. aviculare (= aequale Lindm.). The name subsp. rectum is a new one to the already long list of names in the aggregate, and Scholz (1959) elevates it to specific rank in his treatment* and keeps the taxa above as full species together with P. neglectum. The value of Chrtek and Scholz’s work is difficult to assess since they each seem to have worked in a vacuum and have referred to existing literature rather than specimens on which to base their conclusions. Types have been entirely neglected. Pauvels (1959) has been able to establish the existence of two separable taxa in P. aviculare, sensu lato in the Belgian flora. He contends that plants gathered from different localities in Belgium fall easily into two groups, P. aviculare L., sensu stricto, and P. aequale Lindm. which are both extremely variable. 2. ANALYSIS OF TAXONOMIC CHARACTERS (a) Vegetative Characters Habit. The habit and size of the plant have been frequently used in the past as diagnostic for a particular species, (Boreau 1857, Ascherson & Graebner 1913, Chrtek 1956). Habit varies with the age of the plant and is a most variable feature depending on environmental conditions, and | consider it to be of little taxonomic significance. Ochreae. The hyaline outgrowths at the nodes which characterize the family Polygonaceae are of only limited use as taxonomic criteria in the P. aviculare aggregate. The ochreae of P. rurivagum are however longer than those of the other species recognized. Ochrea-length and the type and number of veins are, on the other hand, highly signi- ficant characters in the separation of P. raii and P. maritimum; those of P. maritimum are frequently as long as the upper internodes, and silvery in colour, with 8-12 branched *He drops this name later (1960) in a key to the species in this group. Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 183 veins. Those of P. raii are shorter; the mean length of a population from Hayling Island, S. Hants (v.c. 11), was 0-6 cm, with up to a maximum of 6 unbranched veins. Leaf size and shape: Heterophylly The leaves are all stmple and entire from 0-3 cm to 5-5 cm in length, and 0-1 cm to 1-8 cm wide, varying in shape from very narrowly lanceolate, to oblong-lanceolate and ovate-spathulate. The petiole is evident only in P. boreale where it is longer than, and protrudes beyond, the enveloping ochrea. Stress has been laid on leaf-characters by previous workers, and they are frequently the only ones mentioned in their keys, (e.g. Boreau 1857). Because of the caducous nature of the foliage, leaf characters must be used with caution, and in this work they have always been related to other more stable features such as fruit- size and length of the fruiting perianth. Two characters were used for biometrical data : 1. A linear measurement of the length and breadth of the largest stem leaf. The length was measured from the point of insertion of the leaf, including the petiole if present, to the apex. The width was taken at the broadest part of the leaf. 2. The index of heterophylly for each plant in any given sample was obtained by dividing the length of the main stem leaf obtained in | by the length of the lowest leaf of the branch subtended by this stem leaf. Fig. 1. Camera lucida drawings of fruiting perianths of Polygonum spp., x 8. (a) P. rurivagum, Devil’s Dyke, Cambs. (b) P. arenastrum, Chedworth, Glos. (c) P. aviculare, Brill, Bucks. (d) P. boreale, nr. Scalloway, Shetland Is. Watsonia 5 (4), 1962. 184 B.-T.- SEVEES (b) Flower and Fruit Characters It is very probable that the small size of the flowers and fruits and their insignificant position in the axils of the leaves is a reason for their having been neglected as taxonomic criteria by many earlier authors. The very striking differences between the fruits of P. maritimum and P. aviculare are not mentioned by Linnaeus. The flowers of the Polygonaceae are monochlamydeous. In the species in this section there are 5 (rarely 4 or 6) perianth segments which vary in length according to the species. The perianth segments have a green midrib and a red, pink or greenish-white petaloid border. The width of this border varies, being greater in the two littoral species P. raii and P. maritimum, and in P. boreale, than in the other species. Colour-differences between the species have been given as diagnostic by many authors in the past, but the present study has revealed that there is no correlation between colour and the different species. Fig. 2. Camera lucida drawings of fruiting perianths of Polygonum spp., X 8. (a) P. raii, Hayling I., S. Hants., (b) P. oxyspermum, Irevik, Gotland, Sweden. (c) P. maritimum, Herm, Channel Is. The perianth is persistent and encloses the fruit forming an enveloping cover. It is dispersed with it when mature. The length of the fruiting perianth segments shows a considerable amount of variation. In some species the perianth is divided almost com- pletely to the base of the flower so that there is practically no tube, whereas in others it is less deeply divided. The tube is then longer, sometimes three quarters of the total perianth length. The length of the persistent perianth, the length of its fused portion and the length of the perianth segments have been found to be diagnostic. (See Fig. | & 2). The following Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 185 parameters were measured using a low-power binocular microscope with a x 10 eye- piece containing a calibrated micrometer scale and a x 3-4 objective. Length of the persistent perianth. 4. Length of the perianth segments, obtained by subtracting the length of the joined portion of perianth from the total perianth length. Total perianth length Perianth segment length 5. The ratio Stamen number Lindman (1912) states that P. heterophyllum has generally eight stamens, five in an outer whorl and three in an inner, whereas P. aequale has five, a whorl of three and another of two. Three typical popuiations were analysed in detail for this character. There is wide variation in the number of stamens in the flowers of any one plant in all three species (Table 1). Stamen number seems to vary with the age of the flower in the inflorescence. Also flowers developed late in the year were found on the whole to have a lower number than earlier formed ones. ABER “1: Percentage frequency of stamen number in flowers of 3 different species. Number of stamens Species | 4 5 6 i 8 91 eKoralsiNo: No. of | | of flowers | plants P. aviculare, sensu stricto 1-26 1-95 7:36 27:8 59-6 1-95 | 163 | SH P. arenastrum | 2:82 42-4 8:02 26 a5574 0 148 | 48 P. rurivagum iO) Dorin ilo, 20:6.) oo.0) *, 1:03 | 194 | 50 | | Fruit The fruit develops from a tricarpellary ovary to form a nut which is usually three-sided. It may be examined by removing the persistent perianth with a sharp scalpel. Absolute length of the fruit has been used by many botanists to segregate species in the past, as in P. microspermum for very small fruited specimens, and P. heterophyllum (= P. aviculare L., sensu Stricto) with large fruit. On the whole the longer fruits tend to be broader, so that fruit-size offers a supplementary character to that of fruit-length. I therefore measured the following parameters using the method mentioned above. 6. Fruit length, from the base of the apex of the fruit, but not including the trifid style which is sometimes found persisting. 7. Fruit size (fruit length x breadth). The length of the fruit as measured above multiplied by its greatest width when the fruit is lying on its broadest face. Perhaps one of the most interesting features of the fruit is the strong correlation between size and shape as seen in cross-section. (See Fig. 3, a-e). It is at once noticeable that the longer fruits of heterophyllous plants almost always have all three sides of the fruit concave, and that one of them is broader than the other two. The more homophyllous plants tend, on the other hand, to have smaller fruit of a different shape. In this case it exhibits in cross-section two convex sides with the third side concave. This third side is narrower than the other two. The narrowness of the concave side, recognisable in the Watsonia 5 (4), 1962. 186 B. T. STYLES fruit of certain plants, was the reason for Lindman (1904) separating P. calcatum as a species from the Linnean P. aviculare complex. It is difficult to obtain biometric data on variation of the width of this concave side. As there is variation in the narrow side from ‘ narrow — very narrow — ridge like’ in populations referable to P. arenastrum it is considered that P. calcatum Lindm. as a taxonomic entity does not exist. ae rs A Fig. 3. Camera lucida drawings of cross sections of fruits of Polygonum spp., X 8. (a) P. arenastrum, Chedworth, Glos. (b) P. arenastrum, Newborough Warren, Anglesey. (c) P. rurivagum, Babraham, Cambs. (d) P. boreale, Shetland Is. (e) P. aviculare, Caernarvon, N. Wales. The fruits of P. aviculare, sensu stricto, are always lustreless and dull with rows of punctate dots. The sides of the fruit are concave. Occasionally four-sided fruits are found. Populations of P. arenastrum have fruits which vary considerably from completely shining to shining only on the edges. P. rurivaguin has fruits which are as long as those of P. aviculare, sensu stricto, but narrower; they also are shiny on their edges. The surfaces of fruits of the two maritime species, P. raii and P. maritimum, are very distinctly shining and highly polished. This, together with their larger size, makes them distinguishable from plants of P. aviculare, sensu lato. A feature which has caused some difficulty in the past is the fact that the fruits of plants in P. aviculare sensu lato may be totally included within the persistent perianth or may project very slightly from it. This has, however, during the course of the work, been found to be variable in any population of either P. arenastrum or P. aviculare, sensu stricto. The fruits of P. rurivagum are always exserted, a point noted by Moss (1914). Those of any single plant are usually all of one type, but even on a single plant there may be variation, since fruits formed late in the year are nearly always slightly exserted. This will be further considered. The fruits of P. raii are always markedly exserted from the persistent perianth whereas those of P. maritimum are enclosed, or only slightly exserted. 3. BIOMETRICAL STUDIES As a basis for my biometrical studies, I decided to take as my primary groups for preliminary investigation those given by Tutin (1952), who has followed in the main the treatment by Moss (1914), in order to see if they presented differences enough to warrant their delimitation as taxa in this country. The groups are P. heterophyllum Lindm. (now Watsonia 5 (4), 1962. PLATE 8 Gc d Fruits of Polygonum spp., < 2:5. (a) P. rurivagum, Babraham, Cambs. (b) P. arenastrum, Chedworth, Glos. (c) P. arenastrum, Newborough Warren, Anglesey. (d) P. aviculare, Brill, Bucks. PLATE 9 c d Fruits of Polygonum spp., 2:5. (a) P. boreale, Easterhoull, Shetland Is. (b) P. raii, Hayling I., S. Hants. (c) P. oxyspermum, Laeso, Denmark. (d) P. maritimum, Figueira da Foz, Portugal. TAXONOMY OF POLYGONUM AVICULARE 187 P. aviculare, sensu stricto), P. aequale Lindm. (now P. arenastrum), P. rurivagum Jord., P. calcatum Lindm. and P. littorale Pers. P. microspermum was also included as this binomial is given by Dandy (1958). I examined in detail a very large number of specimens of the group from British and European herbaria, to ascertain the variation of the plant from many localities over its entire range in this country and western Europe. From this study it was apparent that there is a very great variation in general morpho- logy and habit of the plant according to the type of habitat and degree of sheltering, trampl- ing and crowding to which the plants are exposed, and many characters used by previous authors appeared to be very unreliable, or showed such variability from one plant to another as to be taxonomically useless. Certain features of the fruit and persistent perianth which lend themselves easily to biometric treatment were found to be more constant over a larger number of specimens. Measurement made on non-random samples produced evidence that at least three groups could be defined on these characters. The variation patterns, for instance of linear measurements of fruit, perianth-segment- and perianth- length, fruit size and certain ratios, showed marked discontinuities between the three groups. These corresponded to Tutin’s P. heterophyllum, P. aequale and P. rurivagum. A cursory examination also revealed that the larger-fruited P. heterophyllum tended to possess leaves of a larger size, whereas P. aequale which is small-fruited (by comparison) has shorter, narrower leaves. Thus a possibility existed of correlating the very variable vegetative characters with those of the fruit which are less variable (see scatter diagrams, Figs. 8-10). I have gathered a large number of random samples from throughout the range of the plant in Great Britain and from as many ecologically different habitats as possible. All Species in the group are gregarious and found in patches (often alone on bare ground), or intermixed with other weeds. Sampling was done by gathering plants from a transect cutting through the centre of a patch and from around the edges so that a true representative collection was made of all forms. Composite data for all populations of the four taxa proposed is presented. In the herbarium the samples were analysed for the quantitative parameters already discussed. Their arithmetical means (%), standard deviations (c) and ranges (after Heslop-Harrison 1952) were obtained. Data for a number of samples are presented in Tables 2-6. The same measurements were made on all plants*. Certain other qualitative characters, such as surface-texture and shape of the fruit, which it was not found possible to treat biometrically, have already been discussed. Analysis of Individual Characters An idea of the differences between populations of the same group can be gained from a study of Tables 2-6. In these tables the correct names for the groups have been applied and will be used in the discussion here of their variability. Fruit Characters Sample data for the linear measurements of the fruits, as seen in the tables, show that the means for the different populations in the same group agree remarkably well in this dimension. i.e. = 3-88 for the Chedworth population, 3-85 for that of Clevedon and 3-79 for that at Brill (Table 2). In this research where I have used large samples, the significance of the difference between the means was determined by calculating the standard error of the difference. If the difference between the two means is greater than twice the standard error then the means may be taken to be significantly different (P = 0-05). The fruit length of P. arenastrumis significantly smaller and that of P. boreale (Table 5, Shetland plants) significantly larger than P. aviculare, sensu stricto. The fruit of P. rurivagum is as long as that of P. aviculare and the range of variation is almost completely overlapping. The means for the *The measurements in the tables are given in arbitrary units, where one arbitrary unit =0°793 mm. Watsonia 5 (4), 1962. STYLES Berar 188 “IojouIVied yors 105 sjueld ;eNprAIpul Jo sjusuIOINseow jo 1aquinu =A | OJIIAJS NSUAS SAADINIIAD “q JO Sojdures WOpuUeI INOF JOJ vJep [RONsNeIg “7 AIAVL .<2]DAO}11}], WUNUOBA]OT SW x 08-c-OV: | Ss za £0-C | Oc-£-09: I Se CxG | 06:c-0S-1 WL KORE O83 x05 eel OG (Wo) YIsuUsT Jeo] youRIg “6 ¢0-c-€7¢'l peat 1S-] Qe = Col 28S Alloa ep | eI 14531 (opiGeN oI Iv-] £9-0 Aj[Aydosajoy JO xopuy °*g 86:0-SS:0 cyO SL-0 (oO) = PEED) ES) | $6:0°8$5:0 97:0 8-0 Sisal) W/E) = Gsit0) (Wid) YIpPRaig JesT */ | | . O¢-7-06') 91% — 60°€ | OFS209°0 3 OC-Ge sVS-5 | OV CS ey eV G Vane OS Vase = LG. eaoCre (Wid) YyBUZ] Jeo “9 CONS 0C-0 areal Wacol aa een Vall | (VSL Ea || LG. Ome | Ok fee | iC:0} 26e-1 Ysus] “Sas/yISUE] “Jog “¢ O02 Geo l-0 FOr | SCAPUI(G= = Oi Eats OGV=SL:-G 29-0 SS:¢ | OF V0S:G “Gl-0 =s09:¢ YISUs] JUSUBIg “p OF SaO Gavan a7. 9-6: | 08:S-00-7 67:0 O6L:7 O8:S=0l- SEV 0 = 96-7 | SenOG7Se0678= S250) — 10:5 ysus] YURIIOd jUDISISIOg “€ OF-GI=CS- LS 8V2b =. 89°6 | VAP Ne Bho — ESA OVEO)II | 09-€1-0S-6 98-1 Iv-Ol 09-ZI-€6:9 8H $9-6 | oZIS OI °C | | 09-7-OV-£ WEA) VERS O70 -6S:0) 96:5 | Sy-y-Or-e 65:0 SBE 09-7-O£-£ TOO). fekolote yisug] wnt "| asuv. o a ABUDA o x IBUDA o ao IBUDA 9 xv | ADJOUDAD ov = N CoN 0s =N Sy = Nt | "SYONg “]] 14g yUOAADUIID ) yJeSdIWUos SUOpaday “SOIDQ “Yllompay)D | "pajZys osIMIBUJO ssofUN “WIL ¢6/-9 = jWUN oUO ‘s}TUN ATeIIGIe UI S}USWOINSeo| Watsonia 5 (4), 1962. [89 OF POLYGONUM AVICULARE TAXONOMY SoalevO 160 ¢8-I-$9:0 «IVT 09:020G-0 > 55-0 EASED LONG CEG0S 1) 81-0 Or-c-0£:1 OF-0 0c-7-08-C 7S-0 S91a0CV — SE-T OS-€-0S:c Lv-0 asUuDdA 9 vy —N x JISAOUOG “MOsLLOG SGSOS-0= Teale 728-0 enol meye 0) = GI] OS S003) = 1é0- Ol SZ212002) 26-0 8G0n01-0— 06:0" 65-0 6S:0eGl-0= =9£-0 09 -1s59:05 Vel 360.) WIS Ho S721 OOPGAUS AE = O10) = (yea OL-cs09 1. SE-0 WLECAVEIE —- 0) Size Sc:c Or-I 80-0 VRE, GSW horG 00-7-0S:c 19:0 0c-9-O1-€ Sale ESV, ES:S=O0S°S 1G EAMES (OEY = Lore OE=07:c — 98:0 asUudA fo) a ISUDA o OS = N She == INL (auvT]T Mppyovne) psof{fxoO xassng ‘sousog OSI ¢3-05 By] LoASOGIs “Lé-0 8S:0-0£:0 ~=1E-0 On:Gesc = aalizal GlicGaHE Sol aS \!40) cy-c-Sv-l 60:0 SO-E=S8°C, 69:0 WEES GE ASI OV-€-OL:C 95:0 asUuDdd fo) OS —N Ol-T le-] IS (sspd-dg) p41ofxO "Po}e}S SSIMIOYIO ssayUN “WU ¢6/-Q = WUN oUO ‘s}lUN ATeIIGIe Ul S}USWIOINSES|A ‘1OG wndsjspuadD “q JO Sajdwes 10; eyep eonsners "¢ ITEVE O[DAOIJ 1], WNUOSAJOG SY (uid) YISUI] Jeo, YOuRIg AyjAydorajay JO xopuy (Wd) YIpRAIG jeaT (us) YyysUs] JeoT YIsus] “Sos/YISUS] “dg USus, JUOWISAG yjsua] UURIIod 1Ud}SISIAg ozIS WNIA yysuo] JINI4 41dJOUDAD 6 Watsonia 5 (4), 1962. STYLES Box. 190 eG=exerlk teva Ov-c-OV-I 872-0 SvV:0-07:0 9-0 OG.0209°C IE Se es} 1E0) 00-€-OL:T CIO OS-r-00-€ IS:0 Oc-11-80:9 67-1 Oc-b-0C-€ 84-0 a3UuDA o 0S = N 6S:1 v6:1 1¢-0 IT-€ 6:1 DEC JEEPS 6¢°8 ENE IS ‘squUiD) “‘wpyviqog 60:7F1-1 Ih €S:CaveEl< €¢:0 Iv-0+1:0 4-0 Norenor|! (Werte creel Agal — te60) Oc-€-79-1 60:0 L8-y-Oc:€ cS-0 OlGlsVS:9 = Zeb OS-y-OV-€ 69-0 asuvd o So IN: EE 06:I-0L:0 8S:l LS-I evV-c-SE-1 Sc-0 8-0 SEU I) SIK0) v8-C OV: Fa0LS OS:C e931 Es kee|l lO) 8P-C Ov-€-06:1 O1-0 €6-¢ 0c-S-09:'€ 19-0 6S°8 OCNsS9-9 56ee1 06:¢ O8-b-Or-€ IL-0 o asUuvdA o Co ==ENi ‘sydog “WUDYyIA YY xassny “psofvas $0: 89-1 SS Iss "‘poye}s SsIMIOYJO ssopuUN “WU ¢6/-0 = YUN 9UO ‘s}iUN ATeIIQIe UT s}yUstUOINSEaP] ‘piog¢ winsoalins ‘gq JO sojdures Wopurls so1y} Joy veep [eonsnRS “p AIaVE (Wid) YSUS] Jeo] YOuvIg Ay[Aydos9}0y JO xopuy (wo) YIpReiq jeoT (wd) Yus] JeoT YyysUg] “Sos/YISUST “3g yisusy yUoWUIsAS yysus] YjuLIIod jUS}sISIOg OZIS INIA yysuo, WnI4 4dJOUDAD Watsonia 5 (4), 1962. 19] TAXONOMY OF POLYGONUM AVICULARE (SECAGE NI e332 LEER VE AW) vO-L-€8:0 SS-0 SUS SSKE (ANS LEVIES Cl-0 OF a0 es CIO Ol-lacv-7 ~~ CS:0 EO e9 E-Oree 08-1 Ov-S-00-7 18:0 asUuDdd o Se IN IS BULISANT YINOS VSsSDA LSCAR A Sil OF:CROC TT cV:0 $9-I-SL:0 S-0 OS-7-OE-C = -€ Livelasall GVO) 09-7-0c-€ CIO O1-:9-09:7 cS-0 O@-CI-08-8 ET 08-7-08-€ SZ-0 asUuDdd o Ov = N f4ONnG sAdjsv7z IL-1 16-1 vI-T 8C-e oon 88-€ GES 08-01 CIV IS O8:cs0S:1 VET Mealy Se.0 OL:1-08:0 S:0 00:Ss02:C = C628 WAST ENO 08-r-0c:€ CIO 0€:9-0S-7 L6-0 S6-rI-€6:8 LVI 09-7-OL:€ €L:0 asuDdA 0 0S = N ADMO]]DIS 16:1 v8-1 HE |S) CS-€ ce: S6-€ ces 6V- IT VI-v IS OL:c-06:T cr 06-I-cc-Il I£€-0 OS:I-06:0 _‘IS-0 OV-y-OL:C COON SvV-I-Sc-Il 11-0 09:S-OE-€ S10 O€-L-08:7 S-0 00-81-00-01 vS-I OL:S-OI'b £8-0 asUuvd 0 8 = N 204 vAING IS "poie}s osIMIOYJO ssoyuN “WWW ¢6/-0 = WuN suo ‘s}IUN AIeIIQGIe UI s}USWOMsKo/ “Spueys] PULT}OYS 94} WoI, ajva10g ‘g JO sojdwes wWopueI INoJ JOJ eIep eonsne}sS "¢ ATaVE, (wid) YSUs] Jeo, Yourig Ay[Aydos9}0y JO xopuy (wid) Uypeoiq jeoT (wid) Y\sUS] JeoT uySug] ‘Sos/YISUST “Jog YyIsus] JUSWISEC sus] YJUeLIod yUAd)sSISIOg ozIS yINIA Yyysug] WNIZ AdJOWUDID Watsonia 5 (4), 1962. T. STYLES B. 192 Gio Te SV 80 ee e0 Cars) Slee OS eer Ineo One 01. 002) €6:7-0S-1 Trl 60-72 | (wd) YISus] jee] yourlg “6 Slice wo S608 ven | CV:Cs0G:15 SCO EL | iors coe ier Lec Oe oT | AyAydosojoy JO xopuy °g Z9-1-78-0 5-0 SIT | OF-0-91-0 99:0 12-0 CS OgeIrO es Se 182-0 SU ec0r 79:0 c8:0 | (ud) = YIpeolg JeoT “1 OL7-09'% CEE CHE v8-€-80:% IE 81-2 | 761-€6-:0 Le? EET 99-b-1E-T €9-€ OF-€ | (Cite) preicey) gee) Sy-I-ST-1 TIO SEI | 98-I-SE-1 81:0 19-1 | Cleo S10) 081 19-I-87-1 07-0 €€-1 yIsus] “Bos/ISus] “log °¢ 97S-0S-E SIO SOF | OC-E-FL I O10 66-72 | CET-HET 60:0 8L'1 977-097 91-0 LPE YySus] WSUIsog “y SS-9-$9-y 15-0 5-6 P8-b-97-E 6-0 00-4 | €6€-19-'T €S0 €E€-€ 08-S-€0- 8r0 96-P Ysa] yyURLiod JUcIsISIOg “€ CySI-1S-6 =9PS-T SLT OL-II-@r-9 9E-T 6-8 | 9L:9-$8-7 PET OFS 8L:ZI-00-8 OS-1 90-01 ozIS NIA “7 O@-S-S6-€ PLO SEr OcteceG CON ene | Orecrc e067 OS-b-OF-€ 79-0 98-€ ysus] UNI] OBUDA fo) 6 | IBUD.E ia) a | OSUDA fa) a aSUuDA o iG 4IJOWUDIDG a]DI1OG ‘d UNGDALINA ‘d LUAAJSDUIAD * OJIIAJS NSUIS ‘IAD]NIIAD ‘d "‘PoyeIS ISIMIOYJO SsopUN ‘WU ¢6/-0 = HUN dUO ‘s}IUN AIeIIIGIe UI S]UaUeINsKa (sojduies [[e) aj02109 ‘g “wnspvalind *g “UINA|SDUAAD “gq ‘AADINIIAD “gq “CX¥} INOJ OY) JOJ VEp yeonsteIs aIsOduWIOD ‘9 ATEV], Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 193 aggregate data can be seen in Table 6. There is a significant difference between the means 3-86 for P. aviculare, sensu stricto and 2:92 for P. arenastrum. Fruits from plants that I gathered in Brill (Table 2) and grew the next year in the experimental garden showed a range in length from 3-50 to 4:50, and plants from fruits collected on the Oxford by-pass gave offspring with the same range in fruit-length as the parent plants. This character therefore appears to be very stable under different conditions. 40 30 frequency % N S co) fruit-length (mm) Fig. 4. Variation in fruit-length in (a) P. arenastrum, (b) P. rurivagum and (c) P. aviculare. Fruit-length and -breadth have been multiplied together to give numbers which ex- press size more adequately than either linear dimensions. The fruits of P. boreale are both broader and longer than those of P. aviculare; the product is therefore greater. Although there is some degree of overlap the two have significantly different means. Variation in fruit size between P. arenastrum and P. aviculare is almost discontinuous. frequency (°%) perianth-length (mm) Fig. 5. Variation in perianth-length of (a) P. arenastrum, (b) P. rurivagum and (c) P. aviculare. Perianth Characters The variation of perianth-length in the different groups follows the same pattern as fruit length. From the tables it is seen that perianths of P. aviculare, sensu stricto are longer (% = 4-96) than those of P. arenastrum (% = 3-33) and P. rurivagum (% = 4:00). The Watsonia 5 (4), 1962. frequency (°%) 194 B. T. STYLES distributions of the last two are almost coincident (Fig. 5). Perianth-length in P. boreale (x = 5°40) is the longest in the group and differs significantly from that of P. aviculare although there is again some overlap in the variation Fig 6. The range of this character in the group is great. The length of the tube in relation to the total length of the perianth is very interesting and significant in the taxonomy of the group. This relationship has been expressed as a ratio. In P. arenastrum, which has a comparatively short perianth and long fused portion, up to half its length, with consequently shorter perianth segments, this ratio approaches and exceeds, in some cases, 2 in the fruiting stage. P. aviculare, on the other hand, has a longer perianth and short tube with correspondingly longer perianth segments. In this species the ratio is nearer unity. Study of the results shows that P. aviculare, sensu stricto has a mean value of 1-33, whereas in P. arenastrum itis as high as 1:80. Variation in this ratio is great, however (see Fig. 7, and Tables 2-4). The population of P. arenastrum from Berrow showed variation from 1:50 to 2:25, and another from Oxford (By-pass) varied from 1:57 to 2:12. In P. aviculare the Chedworth population varied from 1-24 to 1-61, and that from Brill between 1-34 and 1-62. From the histogram of the variation of this ratio it is clear that two well defined groups may be demarcated. P. rurivagum has, as already stated, a perianth length approaching that of P. arenastrum (x = 4-00 and 3-33 for aggregate data) but has.the short tube of P. aviculare. In respect of this ratio, therefore, P. rurivagum is intermediate and the mean value from the composite data is 1-61. The length of the perianth segments is greatest in flowers of P. boreale (% = 4-05); 30 20 10 perianth-length (mm) Fig. 6. Variation in perianth-length of (a) P. aviculare and (b) P. boreale. this is significantly different from the mean (3-47) for P. aviculare, sensu stricto (see Tables 2, 6). The shortest perianth segments occur in P. arenastrum with % = 1-78. Lindman (1904) mentions that P. calcatum has the perianth fused up to half its length whereas P. aequale has the perianth segments as long as the fused portion or twice as long (see Table 3). Therefore the ratio mentioned above should always be approximately 2 for P. calcatum whereas that for P. aequale would be more variable. In no population have I found this to be constantly the case and as already mentioned this ratio varies con- siderably, from 1:50 to 2:20, and there is variation from one flower to another on the same plant. Vegetative Characters The four species may be identified on leaf size and shape, although as previously pointed out these characters must be used with caution. Populations of P. aviculare sensu lato are extremely variable (Styles 1960); from the tables it is seen that plants of P. rurivagum Watsonia 5 (4), 1962. 6:0 SS SEES SO TAXONOMY OF POLYGONUM AVICULARE 195 60 50 40 30 frequency (%) 20 werner enw www wm Oe ese www = =| we ew @& = oo ol — i=) SL J ’ ' ' ' po Hae —— od porte neee ‘ i 1 2 1-4 16 1°8 2°0 ratio perianth-length/segment-length Fig. 7. Variation in the ratio perianth-length: segment-length in (a) P. aviculare, (b) P. rurivagum and (c) P. arenastrum. are separable from those of P. aviculare on leaf breadth, (% = 0°27, 0-82), although on this character alone they would be indistinguishable from P. arenastrum (% = 0-28). The leaves of P. rurivagum are, however, longer, % = 2:78 to % = 1-35 in P. arenastrum, and there would be no difficulty in recognizing either in the field. From the Index of Hetero- phylly for the four species it is seen in the tables that P. arenastrum is more homophyllous than the others, although this feature varies widely from population to population owing to factors already mentioned (p. 183). TABLE 7. Parameters of population of plants from Arthur’s Seat, Edinburgh named P. calcatum (N = 25) Measurements in arbitrary units, unless otherwise stated. Parameter ay o range Fruit length 2-98 0:58 2:40-3:40 Fruit size (l. x b.) 6:22 1-34 4:65-7:80 Persistent-perianth length So5\7/ 0:63 2:90-4:20 Segment length 1-96 0-12 1-70—2-60 Per. length/Seg. length 1-71 0:13 1:55-2:05 Stem leaf length (cm) 1-63 2:10 1:25-2:58 Leaf breadth (cm) 0:33 0-32 0:22-0:47 Index of heterophylly 1-36 0:21 1:20-1:44 Branch-leaf length (cm) 1-24 1:38 1:05-1:48 Watsonia 5 (4), 1962. 196 B. 1. SPMEES In Table 7 are presented data for a population from Arthur’s Seat, Edinburgh (v.c. 83). Mr. B. L. Burtt tells me that they were gathered from a spot which is ‘ traditionally ’ known as a station for P. calcatum. The appearance of the plants in the first instance gave me no reason to suspect that they were any different from plants in other populations named P. arenastrum. This proved to be the case when I examined certain of the characters quantitatively. None of the significant characters mentioned by Lindman could be seen only in these plants and not in others. The same is true of samples of Lindman’s own specimens obtained from Stockholm. Some British authors have taken P. calcatum to have fruits which are two-sided (bi-convex) instead of normally 3-sided. During my biometrical analysis of many popula- tions of P. arenastrum | discovered that a certain percentage of fruits of this species were always bi-convex, the highest being for a sample of plants collected from the sand-dunes at Berrow, Somerset with 7-:72°% (see Table 8). These are not constantly found on single plants, and when I raised plants from 10 of such fruits they yielded normal trigonous fruit. This atypical fruit shape is almost certainly due to their development from ovaries with two styles instead of three. TABLE 8. Percentage of bi-convex fruits in populations of P. arenastrum | | Sample | No. | % of population Jackdaw Lane, Oxford v.c. 23 5 | 2-70 Berrow v.c. 6 | 7 | UP Te. Bognor Regis v.c. 14 | 12 | 6:60 Oxford by-pass v.c. 23 | 8 2-20 In 1958 I obtained from the Jardin Botanique de I’Etat, Brussels (BR), six authentic speci- mens of Jordan’s P. microspermum. | have analysed these biometrically in order to ascertain whether they are in fact different morphologically from small plants in populations I had named P. arenastrum (see Table 9). TABLE 9. Statistical data from six herbarium specimens named by Jordan P. microspermum Measurements in arbitrary units, unless otherwise stated. Parameter x G range | | Fruit length | 2-94 | 0:35 | 2:70-3:40 Fruit size (1. x b.) | 5:94 | 1-38 | 4-48-6-40 Persistent-perianth length Bal2 0-51 1-80-3-70 Segment length | 1:70 | 0-12 | 1-60-1-90 Per. length/Seg. length 1:68 | 0-18 | 1-50—1-85 Stem-leaf length (cm) 1-02 1-81 | 0:94-1:35 Leaf breadth (cm) | 0-15 0:25 0:10-1:28 Index of heterophylly | 1-10 0-25 1-00—-1-25 Branch-leaf length (cm) | 0:92 i727) 0:72-1:24 Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE Oy Even though this sample is very small, from the data for fruit and perianth characters there is no evidence that these plants form a distinct taxon, or that the plants are any different from ones which fall in the normal range of P. arenastrum. As already indicated the fruits of the four species in the P. aviculare aggregate differ somewhat in shape and this is, in some respects, a more useful discrimination than size. Unfortunately, however, no method of presenting this feature in quantitative terms could be devised. | 1-4 ) 6 2 e r) e r) 1:0 es @ @,4 @ ® ak ee e & 8 © @ e006? SZ e e e & ~ ® @ 2 ® ® r 6) 6 4 ee e © 1 é : cD) = X MHOOGS e so (6) “4 PRO OS re) x 09900 x 009 O SOOM x Sx 22068 *x*xO x OO O x ba x fo} 7) x © ° segment-length (mm) Fig. 8. Scatter diagram to show the relationship between leaf-breadth and segment-length in P. aviculare (solid dots), P. arenastrum (crosses) and P. rurivagum (open circles). From Fig. 8 separation of the three species P. aviculare, P. arenastrum and P. rurivagum is possible using leaf breadth and perianth-segment length. Fig. 9 deals with the separation of P. arenastrum and P. aviculare, sensu stricto on leaf-length and the size of the fruit. P. boreale may be distinguished from P. aviculare in general by the larger size of its floral and fruiting parts, although the separation is not complete on these characters alone (Fig. 10). Watsonia 5 (4), 1962. 198 B. T. STYLES leaf-length (cm) 3 4 ) 6 i 8 9 10 fruit-size (mm?) Fig. 9. Scatter diagram to show the relationship between fruit size and leaf length in P. aviculare (dots) and P. arenastrum (crosses). perianth-length/segment-length 3:5 4°5 aye) perianth-length (mm) Fig. 10. Scatter diagram to show the relationship between perianth-length and the ratio perianth-length : segment-length in P. aviculare (crosses) and P. boreale (dots). Watsonia 5 (4), 1962. 199 TAXONOMY OF POLYGONUM AVICULARE yseuusqd OV (soueIy pue | jesnyio0g osye) | 0c | | (APMION OSTP) OY | OV | Or Or | | Foes ee 09 | 09 | Ob | Ov | Ob 09 09 | 09 09 ‘Or 09 ‘Or =—s 09 “Ov -_~ Q — aS QD — 8b ean ws b = aoe re SEES SoS SWS ee ~~ N 3 _ S J = = — sjaanog OV 08 09 OV 09 pueysoy) 09 0c OV 09 09 “Or ‘OZ (q puv 19667) (puv]2.] % vpvuvy) ano] ‘qd puny (epeues pur (dupsunff) (S¢61) pajod (DIADUIPUDIS,) (861 ‘Cr6I) OV 09 ‘Or uosssapuy (DIADUIPUDIG) WNjIa] dou wmniwusadsSAxo | d ‘d ‘d ‘d ‘d ‘d ‘d d ‘d | (awo0Yy) LUN LADUL | 0Z | IDA * winjv2]v2 ajo1oqjyy * | ajDa10g WNBDALANA (ajonbav *q) wndAjspua.io | (winyj€ydosajey “d) Or | | OJIIJS NSUuas ‘asojNIIAD Or Or | OID] nsuas ‘aADjNIIAD aN | S | AaB q Se S| e & is =e 5 | saivads Bo S88 Se ® SS Si | e SS = has} "(UZ ‘sIoquIOU ploydip |e) wnuos{jog UONIES OY} UI SPIOIEI SWOSOWOIYD ‘Ol AIAV], Watsonia 5 (4), 1962. 200 B.-T., SPMEES 4. CYTOLOGY AND BREEDING SYSTEMS Chromosome Numbers Results of previous cytological investigations on Polygonum aviculare and its allies are summarized in Table 10. Jaretzky (1928) confined his work to six species which he apparently obtained from botanic gardens rather than natural habitats. As he gives no details of the morphology of the plants he studied the value of his observations is very limited. The tetraploid number for his species P. monspeliense Pers. is surprising, since the type of this plant is easily referred to P. aviculare, sensu stricto, and might have been expected to be hexaploid. The tetraploid number is given for a form of P. aviculare, to which, he states, he is unable to give a name as he is not conversant with the very many varieties of this species. A. & D. Love (1942a) working in Sweden record the tetraploid number for plants labelled * P. aviculare’ with a remark that the specimens studied seemed to be nearer P. heterophyllum Lindm. In a later publication (1956), however, they state that the identi- fication of the voucher specimens was erroneous since they belong to P. neglectum Bess. As this species has never been mentioned in any Scandinavian Flora its existence here is considered most doubtful. Andersson (in A. & D. Léve 1942b) published the numbers 2n = 40, 60 for plants in Polygonum aviculare, sensu lato and later in A. & D. Léve (1948) gave the former number for plants identified as P. aequale and the latter for plants approaching P. heterophyllum. Both Polya (1948) and Pauvels (1959) have published results consistent with these findings, and the latter also discovered P. raii new to Belgium with a chromosome number ol 2n — 40: The most important work published on the cytogenetic and taxonomic variability within and between species of Polygonum is that by A. & D. Love (1956a) on the Icelandic flora and (1956b) on plants in eastern North America. As will be seen from the table several of their chromosome counts do not agree with my own or those of other workers in this field. Perhaps the most interesting is the fact that they record the diploid number of 2n = 20 for P. rurivagum with the other species in the aggregate forming a polyploid series with an octoploid of 2n = 80. The littoral species P. raii is also given as hexaploid with 2n = 60, as is P. boreale from Iceiand. In both these reports the nomenclature is confused and will be discussed later. Very recently Scholz (1960) has published chremosome’ counts substantiating my own records. Cytological investigations on the British taxa abana in P. aviculare, sensu lato were carried out on germinating seedlings, fruits having been gathered from naturally occurring populations, widely distributed ecologically and geographically. Plants of the littoral species P. raii, P. maritimum and P. oxyspermum have been examined from only a limited number of populations. This was due mainly to the fact that, in the case of the first two, the species are rare or very rare, and in the last because authentic material from stations abroad was not forthcoming, although the greatest possible effort was made to obtain viable seed from several north-European and Scandinavian countries. Great difficulties were experienced in obtaining a sufficient germination, and a method following that published by Justice (1941) was finally used to raise the germination rate to about 60%. A method of obtaining metaphase chromosome counts was developed from squash techniques following Darlington & La Cour (1947). Translucent root-tips were treated in 1% colchicine or p-dichlorobenzene for 12 to 24 hours and chromosomes stained using the Feulgen technique, with 6-8 minutes hydrolysis. The chromosomes are extremely small with a slight variation in size from 0-8-1-2 uw. Lists of the British taxa with their chromosome numbers are given in Tables 11 and 12. The diploid number 2” = 20, given by A. & D. Love for P. rurivagum, was not found in any British populations of this species which all proved to be hexaploid with 2n = 60. This agrees with the results of Scholz (1960). As the British plants agree very well morpho- Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 201 TABLE Mil: Chromosome records of species recognized in P. aviculare, sensu lato in Britain. Species | Locality ie, Chromosome number 2n P. aviculare, sensu stricto Foreshore shingle, Caernarvon 49 60 Farmyard, University Farm, Wytham 2 60 Waste ground, building site, Brill 24 60 Entrance to gateway, | Chedworth, Glos. 33 60 P. arenastrum Between flagstones, northern by-pass, Oxford 23 40 Sand dune, path to sea, | Berrow, Somerset 6 | 40 Rough cart track, Bodmin, Cornwall 2 40 P. rurivagum Barley stubble, Gog Magog Hill, Cambs. 29 60 Chalky cornfield, Seaford, Sussex 14 | 60 Bean field, | University Farm, Wytham 2D 60 P. boreale | Shetland Islands | 112 40 TABLE 12 Chromosome records for British and European plants of P. maritimum, P. raii and P. oxyspermum Chromosome Species Locality | Origin number 2n P. maritimum | Sea shore, Herm Channel Isles | 20 Sea shore, Figueira da Foz Portugal 20 Banyuls, Pyrenees S. France 20 P. raii | Shingle, West Town, Hayling Island, | Hants. (v.c. 11) | England 40 Sandy beach, Maenporth, Cornwall | Wick 2) England 40 Sands, Tromsoe Island Norway | 40 P. oxyspermum Leudrig Strait, Laeso, Copenhagen Denmark | 40 Watsonia 5 (4), 1962. 202 B. T. SS@YEES logically with specimens of the same name from Jordan’s herbarium at Lyon it can only be supposed that the Canadian plant is something other than P. rurivagum Jord. and perhaps indigenous there. Plants of P. arenastrum and P. aviculare from this country were consistently found to be tetraploid and hexaploid respectively. Specimens from populations of varying morphology of the two species were studied and counts were made on as many root tips as possible per population. Piants of P. boreale obtained from the Shetland Isles through the courtesy of Mr. W. Scott were consistently found to be tetraploid with 2n = 40. This observation does not agree with that of A. & D. Love. The identification of their plants is however suspect. Among weedy species of Polygonum, therefore, there do not appear to be any intraspecific chromosome races. Diploid (2n = 20) and tetraploid (2n = 40) numbers were found in P. maritimum and P. raii respectively in both British and European populations whilst a single population of P. oxyspermum from Denmark was also tetraploid. The number for P. maritimum thus agrees with the observations of Jaretzky on plants from the Mediterranean. The number for P. raii and P. oxyspermum differ from the 2n = 60 and 80 respectively given by A. & D. Léve. These authors mention that eastern Canadian plants of P. raii come nearer the race named subsp. norvegicum Sam. which is stated to be more northern, having biue-green foliage, shorter internodes, flowers which are white rather than pink or greenish white, and somewhat broader achenes. They suggest that there may be two races of differing chromosome numbers: a British-west-European— Channel race which is tetraploid and a northern Scandinavian and east Canadian race which is hexaploid. My own counts on Norwegian plants do not, however, support this view. It is possible that the Canadian populations do differ from the European ones in chromosome number; the same may be true of P. oxyspermum, but my efforts to obtain viable fruits of these species from several Canadian botanical institutions were unsuccessful. Both Andersson and A. & D. Léve record 2” = 50 for supposed hybrids between P. aviculare, sensu stricto and P. arenastrum but I have never met with a plant I suspected to be a hybrid either from its morphology or its chromosomes. Breeding Systems P. aviculare agg. and the other species of this section are inbreeding and the flowers appear to be always seif-fertilized. This mechanism leads to difficulties in their taxonomy. On many occasions populations have been carefully observed in the field, but no insect visitors have been recorded. The flowers are scentless and their inconspicuous size, colour, and position on the plant show no adaptations to out-breeding. Frequently in the course of emasculation experiments unopened buds were found to contain two or three stamens already dehisced, and many flowers especially at the ends of the older branches open when still enclosed in the ochreae. They then behave as though they were cleistogamous. My own attempts to obtain hybrids between the various species in the aggregate were not conclusive, due mainly to technical difficulties. Such results as were obtained suggested that hybridity is extremely rare among species of this group. A. and D. Léve (1956b) have pointed out that a low degree of apomixis cannot be ruled out. I could not verify this suggestion but I am of the opinion that it does not occur in this group. It is in any case extremely rare among annual weeds (Warburg 1960). The condition whereby the fruits formed in late autumn become swollen and much exserted from the persistent perianth does not so far as I am able to tell have anything to do with apomixis. The embryos and endosperm of these fruits appear to be almost twice as large as normal and they are not as A. & D. Love state ‘inflated and mostly empty.’ Samples of these swollen fruits from several populations were grown in the experimental garden at Oxford but the plants raised from them were normal in every way. I can offer no explana- tion for this abnormal development, and I know of no cases of similar developments among other genera. Watsonia 5 (4), 1962. eS a ee aw aes TAXONOMY OF POLYGONUM AVICULARE 203 The possible origin of these species in the P. aviculare aggregate presents an interesting problem. Many plants are known to have spread from their original areas through the induction of polyploidy (Poa spp., Paeonia spp.). It is interesting to note that the mediter- ranean species P. maritimum never behaves as a weed and only survives in a natural (littoral) habitat. It is thought that weedy species have been derived through hybridization and allopolyploidy of this diploid with one or more unknown species. The place of origin of the weedy species of Polygonum is, however, uncertain. It is interesting to note that certain North American diploid species (e.g. P. achoreum Blake, P. buxiforme Small and P. pro- lificum (Small) Robins) are also confined to maritime zones such as sand-dunes and salt- marshes. 5. TAXONOMIC REVISION AND DISCUSSION From the morphological and cytological evidence presented it is clear that the differences between the groups which can be delimited as species within P. aviculare, sensu lato are rel- atively clear-cut and well defined. This is perhaps surprising, considering the chaos and un- certainty which exist in many floras regarding the number of taxa which can be recognized in this group. It is now proposed to discuss very briefly the chief causes of difficulty in the taxonomy of P. aviculare, and the reasons for calling the groups species. Finally an account of the taxonomy of the group as it now stands after this revision is presented. The synonomy is complete except for some varietal names, which, however, have not been used in accounts of Polygonum in the British flora. I have found the fruit and persistent perianth the most useful and stable criteria in the group and they are the most reliable for classification. There is, however, wide variation in size even on a single plant, and the extent of this can only be ascertained by extensive measurements. The fruits are inconspicuous and concealed in the more evident vegetative parts of the plant. Old plants of P. aviculare, sensu lato are especially difficult to identify since they are frequently reduced to bare stems by the dropping off of the leaves. All species therefore tend to have the same facies and the fruits are then the only guide. Care must also be exercised with specimens gathered in the late autumn because of the enlargement of the fruit, a condition mentioned previously. Maritime forms of P. aviculare in this condition have frequently been called P. raii. A. & D. Léve mention that the recognition of the tetraploid P. arenastrum and hexaploid P. aviculare as species is “ beyond discussion.’ I am in full agreement with these authors on this point. Not only are they morphologically distinct on a number of reliable taxonomic characters but the numerical chromosome differences result in an effective sterility barrier. Hybrids between them appear to be extremely rare in nature. The two species have also different ecological niches, though some intermingling does occur. P. rurivagum has the same chromosome number as P. aviculare but my hybridization experiments have shown that there is probably a sterility barrier. Morphologically the plant differs from the other taxa in a number of characters. The extreme narrowness of the leaves, the longer ochreae and short perianth serve to distinguish it from typical forms of P. aviculare, sensu stricto. Both occur in cultivated fields but remain distinct. P. rurivagum seems, also, to be confined to basic soils in this country and west Europe and has a limited distribution. Plants of P. boreale are among the most robust I have found in the aggregate. This is Surprising since it is the most northerly of all the species. P. boreale is separable from all the others by the large size of the fruit and perianth, and is recognized as a species mainly because it differs in having the very evidently petiolate leaves and the tetraploid chromosome number. P. arenastrum is perhaps the most distinct morphologically of all the species in the aggregate. P. aviculare and P. rurivagum are probably closely related. P. boreale, though tetraploid, exhibits many of the features of P. aviculare. P. maritimum and P. raii are less variable than the weedy species in the group; this is possibly due to the fact that they are Watsonia 5 (4), 1962. 204 B. T. STYLES always confined to littoral habitats. The differences between them are well-marked. Hybrids of these have not, so far as I can trace, ever been recorded. [ee eR AVICULARE IZ Spee el 362-311 753): . centinodium Lam., Fl. Francaise 3: 237-8 (1778). nom. illegit. . erectum Roth., Beitr. Bot. 2: 131 (1783) nom. illegit. . monspeliense Pers., Synop. Plant. 439 (1805). . aviculare var. vulgare Desv., Obs. Pl. Angers 98 (1818). . aviculare var. littorale Koch, Syn. Fl. Germ. Helv. 618 (1837). . aviculare var. polycnemiforme Lecoqg & Lamotte, Catal. Rais. Pl. 322 (1848). . provinciale K. Koch, Linnaea 22: 204 (1849). . aviculare B erectum Ledeb., Fl. Ross. 3: 532 (1850). . aviculare 5 vegetum Ledeb., Fl. Ross. 3: 532 (1850). . aviculare var. angustissimum Meisn. in DC. Prodr. 14: 98 (1856). . denudatum (Desv.) Bor., Fl. Centre France, Ed. 3, 2: 559 (1857). . agrestinum Jord. ex Bor., Fl. Centre France, Eds35 221559 (1857): . polycnemiforme (Lecoqg & Lamotte) Bor., Fl. Centre France, Ed. 3, 2: 559 (1857). . humifusum Jord. ex Bor., Fl. Centre France, Ed. 3, 2: 559 (1857) nom. illegit. . heterophyllum Lindm., Svensk Bot. Tidskr. 6: 690-1 (1912) nom. illegit. . aviculare subsp. heterophyllum (Lindm.) Aschers. & Graebn., Synops. Mitteleur : 848 (1913). . aviculare subsp. rectum Chrtek, Preslia 28: 365 (1956). . aviculare subsp. monspeliense (Pers.) Chrtek, Preslia 28: 365 (1956). P. rectum (Chrtek) Scholz, Ber. Deutsch. Bot. Ges. 72: 65-66 (1959). [P. littorale auct.]. Annual. Plant robust, erect, prostrate or spreading, heterophyllous when young; stems up to 2m long. Leaves lanceolate to ovate-lanceolate, 2-5-5 cm x 1-5 cm, subacute, those of the main stems 2-3 times as long as those of the flowering branches: petioles short (up to 2 mm) or almost absent, included in the ochreae. Ochreae silvery when young, c. 6mm, becoming lacerate with age. Inflorescence 3—6-flowered. Perianth-segments united at the base only, distinctly veined and overlapping in fruit, pink, purple or white at the edges. Stamens (5)-8-(9). Fruit 2:5-3-5 mm, half as broad, punctate, dull brown, with three -+ equal concave sides, (rarely 4-sided), included or sometimes slightly project- ing from the persistent perianth. Flowering the whole season, from July to November. le nef ae) BG elacp as) ms} melt ae) ae) ome! ch splash lac) lash las) Ino Ss} ns} toe) Holotype in the Linnean Herbarium (LINN), London. ; ARENASTRUM Bor. Pl. "Centre France id 37:22) 559 (lisa): . microspermum Jord. ex Bor., Fl. Centre France, Ed. 3, 2: 560 (1857). . calcatum Lindm., Bot. Not. 139 (1904). . aequale Lindm., Svensk Bot. Tidskr. 6: 692 (1912). P. aviculare subsp. aequale (Lindm.) Aschers. & Graebn., Synops. Mitteleur. FI. 4: 848 (1913). P. aviculare subsp. calcatum (Lindm.) Thell., Ber. Schweiz. Bot. Ges. 22: 120 (1913). P. x lindmanii Druce Rep. Bot. Soc. Exch. Cl. 8: 873 (1928). [P. aviculare subsp. aviculare sensu Chrtek, Preslia 28: 366. (1956)]. [P. aviculare auct. (Komarov 1936, Lembke 1948, A. & D. Live 1956a & b) non L.]. Ps oe Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 205 Annual. Plant usually 5-50 cm, forming a dense prostrate mat or sometimes erect. Leaves elliptic or elliptic-lanceolate, up to 2cm long and 0:5 cm wide, those of the main stem and flowering branches -+ equal in size, subsessile. In old specimens the leaves are often more or less crowded towards the ends of the branches. Ochreae short, c. 4mm. Inflorescence 2-3-flowered. Flowers smaller than in P. aviculare; stamens (4)—5-(8). Segments of the persistent perianth united for up to half their length, greenish-white or pink. Fruit 1-5-2-5 mm, dull, but sometimes shining on the edges, with two convex and one narrowly concave sides, (rarely with two sides concave and one convex), or sometimes biconvex, brown to black, not or only a little exserted from the persistent perianth. Flowers July to November. Holotype not traced (but see later). The taxonomy of these two species is best considered together both from the point of view of nomenclature and their geographical distribution. The nomenclatural problems of the two species are most involved. In splitting the Linnean P. aviculare, Lindman (1912), according to the International Code of Botanical Nomenclature, should have kept the latter binomial for one of his segregated species. Instead, however, he published two new names, P. heierophyllum and P. aequale, giving P. aviculare L. pro parte under both. In trying to equate P. aviculare with one of these names in order to conform with the accepted rules, some authors (Komarov 1936, Lembke 1948, Chrtek 1956 and A. & D. Léve 1956) have given the name P. aviculare to P. aequale, whereas others (Moss 1914, Tutin 1952, Dandy 1958 and Rechinger (in Hegi 1958)) have given it to P. heterophyllum. Those authors who take P. aequale Lindm. to be P. aviculare L., retain P. heterophyllum as the name for the other species. This name, however, is illegitimate since Lindman himself includes under it subsp. rurivagum based on the earlier P. rurivagum Jordan ex Boreau. A. & D. Léve and Chrtek give no reasons for retaining P. aviculare L. as the correct name for P. arenastrum. In the Linnean herbarium under P. aviculare there are 5 specimens but only one bears the name P. aviculare in Linnaeus’ handwriting. It is a large leaved heterophyllous plant with fruit c. 3mm and a persistent perianth with segments divided almost to the base. This is P. heterophyllum Lindm. and the Linnean name must therefore stand for it. The specimen falls well within the range of variation of British and west-European plants that I call P. aviculare, sensu stricto. Since the phrase-name in Hortus Cliffortianus 1s not cited by Linnaeus in Species Plantarum, this specimen may be regarded as the type. It seems better to retain this well-known name since it has been used correctly by many British and European botanists. P. monspeliense is the correct name if P. aviculare is rejected, since P. centinodium is illegitimate. The name P. aequale Lindm. cannot be retained because earlier names are available. I have obtained specimens of P. aviculare, sensu lato, from Boreau’s herbarium at Angers, but unfortunately there is no specimen here which can be taken as the type of his P. arenastrum. Plants however, collected and labeiled as this (in what is taken to be Boreau’s handwriting) at a date later than 1857 fall very well into the range of variation exhibited by British plants. There thus seems no doubt of Boreau’s intentions as all specimens labelled P. arenastrum in his herbarium belong to the same species. I propose therefore to use this name, P. arenastrum Bor., for the taxon usually called P. aequale Lindm. P. arenastrum was frequently used by botanists in this country prior to the publication of the name P. aequale by Lindman either as a specific name or as a variety of P. aviculare. Although the names P. arenastrum and P. microspermum were published at the same time by Boreau (1857) I retain the former in that it refers to plants with broader leaves which are the commoner form. The two names have not been combined previously so far as I am able to discover. Scholz (1959) rejects the name P. aviculare L. on the grounds that it is a nomen ambiguum and replaces P. heterophyllum by P. monspeliense. As already mentioned, the latter is, in fact, the correct name for this species if P. aviculare is not used. As I have Watsonia 5 (4), 1962. 206 By ‘Tt. (SEYES been able to typify the Linnean species, there is no reason why the name should not be retained. ; I do not at the present time propose to recognize any infraspecific categories within the two taxa. As is well known in plants where a high degree of inbreeding is prevalent, homogeneous populations occur which differ only slightly from other populations. This has often led to the description of numerous meaningless microspecies in Polygonum as in other common weed species. Both P. aviculare and P. arenastrum are very polymorphic and are markedly affected by habitat conditions. They also show great genetic plasticity in all their morphological characters. Until more controlled breeding experiments can be carried out it is perhaps best to consider them as two variable groups with the status of species in the ‘Linnean’ sense. Ecotypic forms of P. arenastrum are recognizable and these show a certain degree of constancy in habitats of the same type. These showed some morphological differences when their offspring were raised in the experimental garden. I have found that P. calcatum Lindm. cannot be maintained as a distinct taxon in this country at least. The differences between it and P. aequale are not clearly stated by Lindman (1904, 1912) and those attributed to it by Moss (1914) and other authors have not been found sufficiently constant to warrant its maintenance as a species or even subspecies (see p. 196 above and Table 7). Professor H. Nannfeldt of Uppsala (in /itt.) states that in Sweden also many botanists agree that this plant is not a distinct species. It is also suggested that P. microspermum Jord. is not a distinct taxon. Measurements of a small number of plants named by Jordan as this have been obtained from several European herbaria. These plants were analysed but were found to be inseparable from forms of P. arenastrum. In this country narrow-leaved plants with small fruits have been given this name. Small-fruited forms are biometrically inseparable from the overall variation pattern of this character in P. arenastrum (see Table 9). A study of the type of Besser’s P. neglectum from Kiev (KW) shows that it is a different form from any met with in west Europe. The specimen is a poor one but agrees fairly well with the figure given by Lembke (1948) from Polish specimens. It has proved impossible to obtain an idea of the morphological range of this species owing to its being very poorly represented in European herbaria, and hopelessly confused in the literature. It is probable that it has a more easterly distribution and could have been introduced into Canada. Geographical Distribution British Isles Both species are very common as weeds throughout the whole of this country. Druce (1932) records P. aviculare, sensu lato, for every vice-county in Great Britain, mentioning that the common plant is P. heterophyllum Lindm. I have seen specimens of both species in British herbaria from every vice-county; I have also received many records besides from the BSBI Distribution Maps Scheme. It is clear that the two species have different ecological preferences. The estimated percentages in Table 13 must be regarded as only a very rough guide since not all habitats fall clearly into any of the categories listed. These results were obtained from data compiled during my field work on approximately 200.habitats in Great Britain. The two species are often found growing together forming mixed populations. P. aviculare occurs most commonly in cultivated areas, cornfields, etc. (87%), whereas P. arenastrum tends to be found at the edges of fields, in bare patches, and on paths and roadsides (82%). It is on the whole a smaller plant, and most frequently prostrate, which seems to account for the ecological differences. P. arenastrum appears to be able to tolerate a greater amount of treading and trampling than P. aviculare. The former has been observed on footpaths which were trampled on daily by human beings and farm animals, but the plants were not eliminated. In maritime areas, I have noticed too that the two species occur in about the same frequency, but again P. arenastrum survives in places where trampling is most severe. Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 207 TABLE 13. Ecological preferences of P. aviculare and P. arenastrum. Habitat P. aviculare | P. arenastrum ——— Ty a a ec Ta a ar na a ry | Cornfields 66% | 10% Other cultivated fields | e.g. potato fields, Brassica crops, legumes Zw, | 34% | | Tracks and roadsides | (pavements, metalled streets, footpaths) ne | B64 Waste ground (farmyards, building sites, etc.) | A, | ie, Maritime habitats Ou o% Other habitats (gardens, etc.) | fs | PEs On the whole the latter species thrives in drier situations than P. aviculare, as on gravel drives, roadsides, pavements and tracks. Foreign Distribution P. aviculare agg. has a world-wide distribution except for the Antarctic, and it is presumed that the two species P. aviculare and P. arenastrum have similar geographical distributions though with some ecological differences. I have seen herbarium specimens of both from most European countries and there are records for many parts of Asia. P. plebejum is, however, said to replace P. aviculare in some parts of India, Ceylon and parts of Africa. P. aviculare and P. arenastrum are now well established in N. America, but detailed information on their occurrence in S. America and Australasia is lacking. 3. P. RURIVAGUM Jord. ex Bor., Fl. Centre de la France, Ed. 3, 2: 560 (1857). P. aviculare var. rurivagum (Jord. ex Bor.) Gentil, Inv. pl. Sarthe, 218 (1892). P. heterophyllum subsp. rurivagum (Jord. ex Bor.) Lindm., Svensk Bot. Tidskr. 6: 691 @912). Annual. Plant very slender, flexuous, suberect, seldom more than 30cm, hetero- phyllous. Leaves linear-lanceolate to linear, acute, 1-5~3-5 cm long and 0-5—1-4 mm wide. Ochreae brownish-red below, c. 10 mm long, longer than in related species. Inflorescence 1-2-flowered, seldom more. Perianth segments narrow, shorter than P. aviculare, reddish, not overlapping. Fruit 2-5-3-5mm x 1-5-2 mm, exserted from the persistent perianth, with 3 concave sides, scarcely shining. Fruit as long as, but narrower than that of P. aviculare. Flowers from August to November. Holotype not seen (but see below). There are no Jordan specimens of P. rurivagum in the Boreau herbarium, but there are others, the collector of which is unknown (the handwriting is evidently not that of Jordan). I have, however, seen three specimens of the latter from his own collections at Lyon gathered by himself in 1860. This is three years after the publication of Boreau’s Flore du Centre de la France (Ed. 3). As the specimens in the two herbaria are the same plant I propose to keep the name P. rurivagum for this species. Geographical Distribution British Isles This species is not well known, and is probably overlooked in Great Britain. Its distribution needs further investigation. I experienced considerable difficulty in tracing populations of this plant and have found very few records of it. Watsonia 5 (4), 1962. 208 B. T. STYLES P. rurivagum occurs most frequently in cornfields and cultivated places, and so far as is known, only on chalky or other light soils. All my records are from the south and south-west of England. It is easily recognizable from P. aviculare, with which it sometimes grows, by its slender and flexuous habit, very narrow acute leaves and long reddish brown ochreae. I have gathered material myself from vice-countries 8, 13, 14, 22 and 29, and further localities have been provided by the BSBI Plant Distribution Maps Scheme or herbarium specimens from v.c. 7, 9, 11, 16, 17, 18, 24, 30, 31, 54 and 58. There are specimens from all of these except 54 and 58 where the determinations must remain doubtful. Moss (1914) gives the plant’s further distribution as Leicestershire, Durham, Dum- barton and Perthshire. These records again need confirmation. Foreign Distribution The species was first recorded from France by Jordan. I have seen many French specimens and there are duplicates of these in the herbarium at Munich (M). It will pro- bably prove much more common with further searching and will no doubt be found to occur in most western European countries. As stated above, A. & D. Love record P. rurivagum from Canada with the diploid chromosome number (2” = 20). The achenes of this plant are stated to be small, from 1-1-5 (rarely up to 2 mm) long, whilst British plants of this species have fruits which are as long as in P. aviculare but narrower. I have sent British specimens of this species to Professor A. Love for examination and he agrees that they differ from the Canadian plants to which he gives this name. 4. P. BOREALE (Lge.) Small, Bull. Torr. Bot. Club 21: 479 (1894). P. aviculare var. boreale Lge., Medd. om Groenl. 3: 105 (1880). [P. islandicum Meisn. ex Small, Mem. Dept. Bot. Columbia Coll. 1: 108, et tab. 41. (1895) nom. illegit.]. P. heterophyllum var boreale (Lge.) Lindm., Svensk Bot. Tidskr. 6: 691 (1912). P. heterophyllum subsp. boreale (Lge.) A. & D. Love, Acta Hort. Gotoburgensis 20: 216 (1956). Annual. Plant up to 100 cm, erect or suberect, simple or sparingly branched, markedly heterophyllous. Stem leaves oblong-obovate to spathulate, the lower blades 3-5 cm long and 0:5-1-8cm broad; petioles 4-8 mm, projecting from the ochreae. Ochreae 5-8 mm, silvery or brownish, only slightly lacerate. Flowers large, c. 4-5 mm. Petaloid margins of perianth segments broader than those of P. aviculare L., pinkish or white. Segments inclined to open in fruit, longer than in the other species, c. 2:5—3-0 mm. Fruit triquetrous, punctate, dull, light or dark brown, 3:5-4-5 mm x c. 2:5 mm broad, with 3 broadly concave sides (sometimes 4), included, or with the apex just protruding from the lax perianth segments. Flowers June to October. This species has the largest fruit yet observed in plants of the Polygonum aviculare aggregate, and has flowers approaching P. raii and P. maritimum in size. Holotype, Botanical Museum, Copenhagen, Denmark (C) Legit J. Lange, Greenland, 1880. Geographical Distribution British Isles This species is new to the British Isles, although Druce (1912) used the name P. aviculare var. boreale for plants gathered at Loch Leven, Fife (v.c. 85). There was, how- ever, at the time great confusion between this variety and the plant then known as P. littorale. It is at present only known from the Shetlands and Orkney Islands. Druce (1922) records a forma grandiflora [sic] of P. aviculare var. littorale from Uyea Sound, Shetland, and it was from a study of this specimen and others in the Druce Herbarium, Oxford, that I decided to investigate the Shetland populations further. Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 209 Foreign Distribution I have seen harbarium specimens of this species from Iceland, Faroes, Greenland and Northern Scandinavia. A. & D. Léve’s records for northern Canada need to be checked but may well be correct, since the description of the Canadian plants studied agrees with European and Greenland plants but differs from their earlier description of plants from Iceland. Small (1894), after raising Lange’s variety boreale of P. aviculare to full specific rank, published the name P. islandicum in 1895. It is apparent that after having published the first of the binomials he received a sheet of P. boreale from De Candolle’s herbarium at Geneva collected from Iceland in 1866 by M. E. Jardin labelled by Meisner as P. islandicum The latter author, however, never published this name and it is therefore not valid. The name P. boreale (Lge.) Small must be retained for this species. Further work is needed to establish the extent of the distribution of P. boreale in northern Scotland. 5. P. RAW Bab., Trans. Linn. Soc. Lond. 17: 458-9 (1836). P. dubium Deakin, Florig. Brit. 2: 576 et tab. 656 (1845), nom. illegit. P. maritimum f raii (Bab.) Lloyd, Fl. Ouest France, ed. 2: 430 (1868). [P. littorale sensu Gren. & Godr., Fl. de France 3: 51-2 (1855), non Link]. P. raii subsp. norvegicum Sam., Acta Hort Bergiani 11: 72 (1931). [P. robertii auct.]. Annual. Plant prostrate, glabrous, sometimes glaucous and slightly fleshy, 10-100 cm. Stems more or less woody at the base. Leaves 1-3°5 cm, elliptic-lanceolate to linear- lanceolate, usually flat, though the young ones sometimes appear revolute. Ochreae shorter than upper internodes, laciniate, c. 5mm, hyaline and silvery in the upper part with 3-6 unbranched veins. Inflorescence 2-6-flowered. Flowers large, pink or white. Perianth segments c. 3mm, with broad petaloid margins. Stamens 8. Fruit 5-6 mm long x 3-3-5mm broad, light brown, flattened, smooth, shining, much exceeding the persistent perianth. Flowers from June to September. Holotype in the University Herbarium, Cambridge (CGE). Legit Borrer. Sands between Marazion and Penzance, Cornwall, 1836. Geographical Distribution British Isles P. raii is an interesting Atlantic species which formerly occurred fairly frequently on sands and shingle in maritime areas just above the high tide mark around the coasts of Britain (see Fig. 11). Associated species growing with P. raii at Freshwater Bay, Pembrokeshire (v.c. 45) were Euphorbia paralias, E. portlandica, Salsola kali, Cakile maritima, Ammophila arenaria and Chenopodium spp. From records received by the BSBI Distribution Maps Scheme there is evidence that it is decreasing in frequency in this country. I have searched for it, without success, in many of the localities in Wales and in the west and south of England where it has been recorded in the literature as common. The plant was very local in two areas (Hayling Island, S. Hants (v.c. 11), and Maenporth, Cornwall (v.c. 1)). It was formerly common locally along the coasts of south England, Wales, west Scotland and the Hebrides and on the east from Durham to Angus. It has been recorded from nearly all the coastal counties in Ireland. Druce gives the following vice-county records of the species in the British Isles. S, 1-6, 9-11, 13-16, 18, 34, 37, 41, 44-46, 48-52, 53, 54, 58-60, 66-76, 82, 83, 85, 87, 90, 97-99, 100-104, 110 H, 1-3, 5-6, 9, 12, 15-16, 20-22, 27, 31, 34, 35, 38-40, Watsonia 5 (4), 1962. 210 B. T. STYLES I have added the plant new to v.c. 106, where it was recorded by Marshall at Ballintore, E. Ross in 1910, but named by him P. aviculare var. littorale Link (BM). I have seen speci- mens from all the above vice-counties except 34, 37, 53 and 54. The last two records are suspect. Vice-county 37 is inland (Worcs.), but the occurrence of P. raii is possible since the river Severn is tidal along its lower reaches. Foreign Distribution Moss (1914) gives the extra-British distribution of P. raii as southern Scandinavia, Denmark, Germany, Belgium, France, northern Russia, Spain, Italy and the west coast of North America. Specimens are extant from the following; north-west France, Belgium, north Germany (Heligoland) (Scholz 1957), Norway and Denmark. Dr. Bengt Petter- sson tells me (in /itt.) that P. raii does not occur in Sweden where it is replaced by P. oxyspermum. Specimens tentatively named R. raii from Portugal by Dr. A. R. Pinto da Silva were P. maritimum (atypical). I have been unable to trace specimens or records for Spain and Italy, and, as the species appears to have an exclusively Atlantic distribution, I regard them as having been based on misidentified material. Fig. 11. Distribution in Europe of P. maritimum (solid line), P. raii (dashes’ and P. oxyspermumi (dots). P. raii is recorded from northern Russia by Komarov (1936). The name P. raii is given in certain east-European Floras, Savalescu (1952), Stoyanoff & Stefanoff (1933) for plants collected around the Black Sea. I have seen specimens named P. raii from this area. They are morphologically very close to west-European plants to which I give this name and I am investigating them further. Chrtek (1960) has described two new species in the group from this area. One of these also seems to be very close to P. raii. The north American records present a problem. Many recent Floras and Manuals of selected areas of this continent, e.g. Fernald (1950), Gleason (1952) and Roland (1947), give accounts of it but it appears that it may be a different cytological race. I have seen specimens in the Kew herbarium (K) which are undoubtedly P. raii. I can find no basis for retaining Samuelsson’s subspecies norvegicum (1931) of P. raii. The differences attributed to it are small. It differs from the type in having more flowers in the inflorescence and narrower non-overlapping perianth segments with a broad white edge. All of these characters are most variable and are not of great taxonomic significance within the section. Both pink and white flowers occur in plants of P. raii from Great Britain. Plants from Norway have been called exclusively subspecies norvegicum. Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE 211 Many British, Continental and American botanists during the latter half of the last century named plants belonging under P. raii Bab. as P. robertii Lois. The confusion arose in this country in the first instance because no British botanists had seen authentic material of the latter, while Meisner (in De Candolle 1856) states that material sent to him by Loise- leur-Deslongchamps was a mixture of P. raii, P. littorale and P. aviculare and that the name ought to be completely abolished. Loiseleur’s species was described from maritime sands of the Mediteranean, and since P. raii, according to Rouy (1910), does not occur south of the shores of the English Channel, it is hardly probable that the plant collected by Robert on the sands near Toulon is identical with the northern one. Furthermore, Rouy maintains P. robertii is a very distinct plant of the Mediterranean sands with shining achenes 2-3 mm long. Under these circumstances it is correct to keep the name P. raii for Babington’s plant, and no change in the binomial is proposed. 6. P. OXYSPERMUM Mey. & Bge., in Ledebour, Ind. Sem. Hort. Acad. Dorpat. Suppl. 2: 5 (1824). P. acadiense Fernald, Rhodora 16: 188 (1914). After examining a number of specimens of P. oxyspermum from Denmark, Sweden and Scotland, I feel that the status of this plant as a full species is doubtful. Many of the differences given by Samuelsson are based on characters which show great variability. Although I was able to detect the colour differences of the achenes and perianth on plants collected in 1957, I found this impossible in older specimens. The narrowness of the fruit is an important feature, but I was unable to separate P. raii from it on this character alone. Although P. oxyspermum has a Baltic distribution (neglecting Canadian records) there is considerable overlap of this species with P. raii in southern Scandinavia and Denmark (Fig. 12). Before any taxonomic conclusion can be made on the distinctness of these two species more information is required on their behaviour in these areas of overlap. There is a chance of hybridization since the plants occupy similar types of habitat, the flowers are larger and more attractive than those of P. aviculare, sensu lato and the European populations have the same chromosome number. I think that P. oxyspermum and P. raii may be better treated as geographical sub- species of one species, but until more information is forthcoming on the facts mentioned above the question must remain unanswered. The two are being grown under ex- perimental conditions and I hope to publish a fuller account of these two species later. There are several specimens in Herb. Mus. Brit. which have been labelled by Samuelsson as P. oxyspermum, although the original collectors identified them as P. raii. The localities are: Gosfort, Haddington 1850 (v.c. 82); Mornfar, Forfar, 1845 (v.c. 90); Carnoustie, Angus, 1845 (v.c. 90); Musselburgh, W. Lothian, 1842 (v.c. 83). There is also another, originally named as P. aviculare var. littorale, collected from a salt marsh, Warrenly, Yorks., v.c. 62, which is also called ? P. oxyspermum by Samuelsson. I have studied these specimens, but at the moment feel uncertain about their correct determination. They are all old and the colour and narrowness of the fruit is not evident. The Carnoustie plant (the same form was also obtained from Dundee), did not have red borders to the flowers, which is given as diagnostic for the species by Samuelsson (1931), and I could not see that the perianth segments were any narrower than in specimens of P. raii. There are no very evident differences in leaf morphology and internode length, which are very variable characters in the group. 7. P. MARITIMUM L., Sp. Pl. 361 (1753). P. maritimum var. confusum Rouy, Fl. France 12: 110 (1910). A glabrous, glaucous, prostrate or erect shrubby (in south Europe) perennial, 10-50cm; root stock woody and stout at the base. Leaves 0-5—2:5 cm, elliptic-lanceolate, greyish- green, margins usually revolute. Ochreae very conspicuous and silvery-white above, usually Watsonia 5 (4), 1962. 212 B. T. STYLES longer than the upper internodes which they envelop, 2-lobed at first but eventually be- coming lacerate, with 6-12 strong branched veins. Inflorescence 1—4-flowered. Flowers twice as large as those of P. aviculare. Perianth segments c. 2-2-5 mm, with pink or white broad petaloid margins. Stamens 8. Fruit 445mm long x c. 2.5mm broad, ovoid, acute, reddish chestnut-brown, smooth, shining, as long as or slightly exceeding the persistent perianth. Flowers July to September. Holotype in Linnean Herbarium (LINN), London. Geographical Distribution The British Isles This species is now thought to be extinct on the mainland of Great Britain, since no plants have been found since about 1939 (coll. C. West nr. Mullion, Cornwall, v.c. 1) in spite of the increased activities of the members of the BSBI engaged in the Maps Scheme. The well known colony of approximately 30 plants was seen on Herm, Channel Islands in 1961. They were growing in bare unstable coarse shell-sand just above the high tide mark with Beta vulgaris subsp. maritima, Carex arenaria and Brachythecium albicans as associate species. Although never very abundant in any one locality in this country, I have seen herbarium specimens from the following vice-counties in the south of England : S. (Guernsey, Herm) 1 (including Scilly Isles), 2-5, 6, 9, 11. Druce also gives the following v.c. records in brackets : 13, 28, 46, 49, 75. All are suspect and no specimens are available for their confirmation. P. maritimum is a Mediterranean species which reaches its northernmost limits in Great Britain. As in many other plants at the edge of their geographical range, the British speci- mens are not. quite typical and differ in certain characters from the main body of the Continental material. The Herm plants are more herbaceous than those obtained from Sacavem, Portugal and Banyuls, Pyr.-or., France and have shorter ochreae with less conspicuous veins. The British populations on Herm (fide Mrs. F. Le Sueur) behave as perennials but other specimens seen from Braunton Burrows suggest from their herbaceous appearance that the plant might have behaved as an annual on the mainland. Jt is quite possible that fruits of P. maritimum could be carried by ocean currents and thereby washed up on to the shore by tides. I have no doubt that the plant could become established again in the south and west of England given the right conditions. Foreign Distribution P. maritimum occurs frequently on sandy shores and shingle in the following European areas from which I have seen specimens : western (Manche) and southern France, Iberian Peninsula, Macaronesia, Mediterranean Europe, Corsica; also in Asia Minor and north Africa. Moss (1914) gives the following further records, Cape Colony, South Africa (rare), North America (Massachusetts to Florida) and South America. I have seen no specimens nor can I trace records from the first and last continents mentioned, and consider the occurrence of the plant here most doubtful. The North American records are, according to Fernald (1913), all based on wrongly identified material. Linnaeus (1753) includes American plants with his frutescent Mediterranean species P. maritimum, saying ‘Habitat Monspelii, in Italia, Virginia, [perennial].’ Several American authors, Pursh (1814), Torrey (1843) and Small (1895), have included P. maritimum in their Floras and Manuals but it is now fairly certain that the American plant referred to by Linnaeus is in fact P. glaucum Nutt. This species is an American endemic and differs from P. maritimum in that it is always annual and herbaceous, with shorter ochreae, 7-10 mm, and in the fruits, which are distinctly smaller, 3-4 mm long and 1-6-2:2 mm wide. It is very probable that Linnaeus included his American plants of this species under P. maritimum. There is still much confusion in American floras over the maritime species in this section and experimental work on their taxonomy is urgently needed. Watsonia 5 (4), 1962. TAXONOMY OF POLYGONUM AVICULARE PAN 3) ARTIFICIAL KEY TO THE BRITISH SPECIES* 1. Fruit dull, striate, enclosed by or slightly longer than the persistent perianth. 3 Fruit smooth, distinctly shining, as long as or longer than the persistent perianth. 2 2. Ochreae with 4-6 unbranched veins, shorter than the upper internodes; fruit 5-6 mm long, exceeding persistent perianth. 5, ACH Ochreae with 8-12 branched veins, as long as upper internodes; fruit 4-4-5 mm long, enclosed by or slightly projecting from persistent perianth (very rare). 7. maritimum 3. Branch leaves much smaller than stem leaves; persistent perianth divided almost to base; fruit trigonous, with 3 concave sides. 4 Branch and stem leaves + equal; persistent perianth divided for half its length; fruit with 2 sides convex, one concave. 2. arenastrum 4. Stem leaves narrow, linear-lanceolate, 1-4 mm broad; perianth segments and fruit narrow. 3. rurivagum Stem leaves 5-18 mm broad; perianth segments overlapping; fruit broad. 5) 5. Stem leaves ovate-lanceolate, subsessile or with petioles c. 2mm included in the ochreae; fruit 2:-5—3-5 mm long. 1. aviculare Stem leaves obovate-spathulate; petioles 4-8 mm, projecting from ochreae; fruit 3-5—4-5 mm long (Shetland). 4. boreale ACKNOWLEDGMENTS My grateful thanks are due to Dr. E. F. Warburg for his help and guidance during the progress of this work and for his careful criticism of the manuscript. Also to Dr. K. R. Lewis for help with the cytological problems involved, and to numerous correspondents who have lent or given me specimens. REFERENCES ANDERSON, E. (1949). JIntrogressive Hybridization. New York. ANDERSSON, E. (1942) in A. & D. Love (1942a). ANDERSSON, E. (1948) in A. & D. Love (1948). ASCHERSON, P. & GRAEBNER, P. (1913). Synopsis der Mitteleuropdischen Flora, 4. Leipzig. BABINGTON, C. C. (1836). On several new or incompletely und2rstood British and European plants. Trans. Linn. Soc. Lond., 17, 451-464. Besser, W. (1822). Enumeratio plantarum hucusque in Volhynia, Podolia, Gub. Kiioviensi, Bessarabia Cis- Tyraica et circa Odessam collectarum, etc. Vilna. BoreEAu, A. (1857). Flore du Centre de la France, 2, Ed. 3. Paris CHRTEK, J. (1956). Proménlivost drihu Polygonum aviculare L. v C.S.R. Preslia, 28, 362-368. CHRTEK, J. (1960). Neue Arten der Gattung Polygonum L. s. str. Preslia, 32, 366-368. Danby, J. E. (1958). List of British Vascular Plants. London. DARLINGTON, C. D. & LA Cour, L. F. -(1947). The Handling of Chromosomes. Ed. 2. London. De CANDOLLE, A. -P. & LAMARCK, J. B. (1815). Flore Francaise, 3. Ed. 3. Paris. Druce, G. C. (1922).- Flora Zetlandica. Arbroath. Druce, G. C. (1932). The Comital Flora of the British Isles. Arbroath. FERNALD, M. L. (1913). Some N. American relatives of Polygonum maritimum. Rhodora, 15, 68-72. FERNALD, M. L. (1914). A new maritime Polygonum from Nova Scotia. Rhodora, 16, 187-189. FERNALD, M. L. (1952). Gray's Manual of Botany. Ed. 7. New York. GLEASON, H. A. (1952). The New Britton and Brown, Illustrated Flora of the Northeastern United States and Adjacent Canada, 2, Lancaster, Pa. HESLOP-HARRISON, J. (1952). Statistical methods in plant taxonomy. Yaxon, 1, 53-59, 73-78. International Code of Botanical Nomenclature, 1956. Utrecht. JARETZKY, R. (1928). Histologische und Karyologische Studien an Polygonaceen. Jahrb. wiss. Botan. 69, 357-499. Justice, O. L. (1941). A study of dormancy in seeds of Polygonum. Mem. Cornell Agric. Expt. Stat. 235-243. Kocu, W. D. Jj. (1837). Synopsis Florae Germanicae et Helveticae etc., 2, Ed. 1. Frankfurt a/M. Komarovy, V. L. (1936). Flora U.R.S.S., 5, Leningrad—Moscow. *Identifiable with certainty only if ripe fruit and stem leaves at least at 2nd or 3rd node are present. + see also 6. oxysperum Watsonia 5 (4), 1962. 214 BY To (SRYEES LANGE, J. (1880). Conspectus Florae Groenlandicae. Medd. om Grénl., 3, 1-214. LEDEBOUR, C. F. (1850). Flora Rossica, 3, Stuttgart. LeEMBKE, J. (1948). Etude sur les menues espéces appartenant a l’espéce collective Polygonum aviculare L. Acta Soc. Bot. Polon. 19, 201-207. LINDMAN, C. A. M. (1896). Polygonum aviculare L. f. littoralis (Link) i Skandinavien. Bot. Not. 75-80. LINDMAN, C. A. M. (1904). Polygonum calcatum nov. spec. inter Avicularia. Bot. Not. 139-147. LINDMAN, C. A. M. (1912). Wie ist die Kollectivart Polygonum aviculare za spalten? Svensk Bot. Tidskr. 6, 673-696. LInK, H. F. (1821). Enumeratio Plantarum Horti Regii Botanici Berolinensis altera, 1, Berlin. LINNAEUS, C. (1753). Species Plantarum. Stockholm. LOISELEUR—DESLONGSCHAMBPS, M. J. L. (1827). Nouvelle notice a ajouter a la Flore de France. Mém. Soc. Linn. Paris, 6, 1-40. Love, A. & D. (1942a). Chromosome numbers of Scandinavian Plant Species. Bot. Not. 19-59. Love, A. & D. (1942b). Cytotaxonomic studies on boreal plants I. Some observations on Swedish & Icelandic plants. Kgl. Fysiograf. Sallskap. Lund Férh, 12, 1-19. Love, A. & D. (1948). Chromosome numbers of Northern Plant Species. Iceland Univ. Inst. App. Sci. Dept. Agric. Rep. B, 3, 1-131. Love, A. & D. (1956a). Cytotaxonomic conspectus of the Icelandic Flora. Acta Horti Gotoburgensis 29, 65-290. Love, A. & D. (1956b). Chromosomes and taxonomy of eastern N. American Polygonum. Canad. Journ. Bot. 34, 501-521. MEISNER, C. F. (1826). Monographiae Generis Polygoni Prodromus. Geneva. MEISNER, C. F. (1856). Polygonaceae in A.-P. De Candolle, Prodromus Systematis Naturalis 14. Paris. Meyer, C. A. & Bunce, D. A. (1824), in Ledebour, C. F. Indicis Seminum Horti Academici Dorpatensis. Supplementum 2. Dorpat. Moss, C. E. (1914). The Cambridge British Flora 2, Cambridge. OSTENFELD, C. H. (1916), in Fernald, M. L. On the occurrence of Polygonum acadiense in Denmark. Bot. Tidsskr. 34, 253-255. PAUVELS, L. (1959). Etudes critiques sur quelques Polygonum de Belgique. Bull. Soc. Roy. Bot. Belg. 91, 291-297. PERSOON, C. H. (1805). Synopsis Plantarum etc. 1, Paris. Potya, L. (1948). Chromosome numbers of certain alkali plants. Arch. Biol. Hung. Ser. II. 18, 145-148. PursuH, F. (1814). Florae Americae Septentrionalis 1, London. RECHINGER, K. H. (1958). Hegi’s Illustrierte Flora von Mitteleuropa 3, Ed. 2. Miinchen. REYNIER, A. (1905). Notes sur les plantes distribuées. Bull. Herb. Boiss. 5, 980-983. ROLAND, A. G. (1947). The Flora of Nova Scotia. Proc. Nova Scotian Inst. Sc. 21, 3, 95-642. Rouy, G. (1910). Flore de France 16. Paris. SAMUELSSON, G. (1931). Polygonum oxyspermum Mey. et Bge. und P. raii Bab. ssp. norvegicum Sam. n. ssp. Acta Horti Bergiani 1i, 67-80. SAVALESCU, T. (1952). Flora Republicii Populare Romine 1, Bucharest. ScHoLz, H. (1957). Uber ein Vorkommen des Strandknéterichs Polygonum raii in Deutschland. Ber. Dtsch. Bot. Ges. 70, 484-488. SCHOLZ, H. (1958-9). Die Systematik des europaischen Polygonum aviculare L. I, Ber. Dtsch. Bot. Ges. 71, 427-434; II ibid. 72, 63-72. SCHOLZ, H. (1960). Bestimmungsschliissel fir die Sammelart Polygonum aviculare L. Verh. Bot. Prov. Brand. 98-100, 180-182. SCHRADER, J. (1800). Journal fiir die Botanik 1, 1. Gottingen. SMALL, J. K. (1894). Some notes on the rarer species of Polygonum, Bull. Torr. Bot. Club 21, 476-482. SMALL, J. K. (1895). Monograph of the N. American species of the genus Polygonum. Mem. Dept. Bot. Columbia College 1, 1-183. STOYANOFF, N. & STEFANOFF, B. (1933). Flora of Bulgaria, Ed. 2. Sofia. StyLes, B. T. (1960). The Taxonomy of the British Knotgrasses, in The Biology of Weeds. British Ecological Society Symposium No. 1, 48-54. TorreEY, J. (1843). Flora of New York State 2. New York. TuTIin, T. G. (1952), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Ed. 1. Cambridge. WARBURG, E. F. (1960). Some taxonomic problems in weedy species, in The Biology of weeds. British Ecological Society Symposium No. 1, 43-47. ; Watsonia 5 (4), 1962. NOTES ON BRITISH HIERACIA. I. THE SPECIES OF THE ORKNEY ISLANDS By P. D. SELL and CyrRIL WEST Botany School, Cambridge During the period 1920 to 1933 more than a dozen new species and varieties of Hieracium from the Orkneys were described by H. Dahlstedt, based on specimens sent to him for identification by the late H. H. Johnston. Many of these have been ‘ lumped ’ by H. W. Pugsley (1948). We have made a special study of the hawkweeds of this area in an attempt to discover how far there was justification for Dahlstedt’s new species and for their treatment by Pugsley. All the specimens on which these new species were founded are in the collections made by H. H. Johnston and J. Sinclair. We are grateful to Mr. Sinclair for allowing us to examine the specimens in his private herbarium, and to the Regius Keeper of the Royal Botanic Garden, Edinburgh, for the loan of Johnston’s material. A few specimens from other collections have also been examined. As a result of a careful examination of all the available material we have found that we are in general agreement with Pugsiey that many of Dahlstedt’s species can be matched with species already described. It is well known that certain species of Hieracium are extremely variable and not easily distinguishable from one another, especially when poorly grown or late-flowering. Johnston often sent Dahlstedt few specimens of inadequate, late-flowering material. This and the fact that Dahlstedt was not well acquainted with the British Hieracia may account for the large number of names he applied to Johnston’s plants. We recognise in all twelve species of Hieracium from the Orkney Isles. Ten of these species belong to the Subgenus Hieracium, namely H. anglicum Fries and H. iricum Fries (Sect. Cerinthoidea Fries); H. argenteum Fries, H. orimeles Hanb. ex W. R. Linton and H. scoticum Hanb. (Sect. Oreadea Fries); H. sarcophylloides Dahlst. (Sect. Suboreadea Pugsley); H. euprepes Hanb. and H. caledonicum Hanb. (Sect. Vulgata Fries); and Hi. latobrigorum (Zahn) Roffey and H. maritimum (Hanb.) Hanb. (Sect. Fo/iosa Fries). The other two species, H. pilosella L. and H. aurantiacum L. sensu Stricto, belong to the subgenus Pilosella Tausch. Of these twelve species H. aurantiacum and H. pilosella present no difficulty. H. aurantiacum is a garden escape of central European origin, easily recognised by its reddish-orange ligules. H. pilosella is widespread in the British Isles and continental Europe but is absent from the Shetlands, Faeroes and Iceland. Its stoloniferous, scapi- gerous habit makes it a plant easy to identify. H. latobrigorum and H. maritimum differ from the rest by their very leafy stems. H.. maritimum shows little variation in form, and elsewhere occurs only in a few places in N. Scotland, the Shetlands (as H. obesifolium Pugsl.), Ireland and Scandinavia. H. latobrigorum, on the other hand, is a very variable plant to which several names have been applied. It is widely distributed in northern Britain, north-eastern Ireland and central Europe. H. anglicum and H. iricum can usually be separated from the remainder by their semi-amplexicaul cauline leaves, very pilose-tipped ligules and furcate inflorescence of large heads. They may however be difficult to distinguish from each other, and their differences can only be fully appreciated by examination of a series of authentic herbarium sheets. H. iricum normally has more numerous stem leaves, and is more shaggy in appearance, but plants of H. anglicum occur (var. amplexicaule Backh.) that superficially 215 Watsonia 5 (4), 1962. 216 P. D. SELL andy CYRIL” WES® resemble it. In fact, among the specimens we examined there were several of H. anglicum which had been labelled H. iricum. H. anglicum is not uncommon in northern Britain and Ireland, but is extremely rare in Wales. It occurs in the Faeroes (as H. perampliforme Dahlst.) and in Iceland (as the variety vestmannaense Omang). It is absent from the Shet- land Isles and the Continent. JH. iricum occurs in Scotland, northern England, Ireland and the Faeroes (as H. peramplum Dahlst.), but is not recorded from the Shetlands, Iceland or continental Europe. The remaining six species are variable and are not easy to distinguish from one another, especially when poorly grown or late flowering. H. argenteum Fries is most readily distinguishable by the characteristic leaden colour of its leaves. H. pseudomicrodon Dahlst., which was founded on exsiccatae from Hoy, is, as Pugsley (1948) rightly pointed out, identical with this species. H. argenteum is widely spread throughout Scotland, northern England, and Scandinavia, and is of local occurrence in Ireland, Wales and Iceland (as H. microdon Dahlst.). It has not been found in the Shet- lands or in the Faeroes. H. orimeles Hanb. ex W. R. Linton, which is abundant in the Shetlands (as H. beeb- yanum Pugsl.) and local in western Scotland, northern England and Wales, is known only from a single locality in the Orkneys. This species closely resembles H. argenteum, both having narrowly ovate to lanceolate basal leaves, 1-3 usually spreading cauline leaves, and narrow acute phyllaries. It can however be distinguished from H. argenteum by the colour and clothing of its leaves. H. sarcophylloides Dahlst. which was originally described from the Faeroes, is identical with H. dasypodum Dahlst. from the Orkneys. H. sarcophylloides is the older name. It is a variable species, which in its typical form can be separated from the other Orcadian species by its rather broadly ovate, truncate-based leaves. The type material of H. sinclairii Dahlst. matches H. sarcophylloides so closely that it is not easy to understand why Pugsley claimed, even with some reservation, that it was closely allied to H. oistophyllum Pugs. H. pycnodon var. acutidens Dahlst. is based on specimens of H. sarcophylloides with leaf bases which are less truncate than usual. The Scandinavian H. pycnodon Dahlst. may also be referable here, but the material we have seen is insufficient to enable us to arrive at a definite opinion. H. sarcophylloides also occurs in Sutherland, Argyll and the Faeroes. Although H. euprepes is so variable that it is not easy to pick out any definite distin- guishing characters, the general facies of a well-grown plant with its elliptical, dark green leaves, tall wiry stem with 1-3 rather narrow stem-leaves, and rather condensed inflorescence of narrow, often geminate heads, is characteristic. We agree with Pugsley (1948) in placing a number of Dahlstedt’s species under H. orcadense, but we consider that species to be indistinguishable from H. euprepes. The type specimen of H. sagittatum subsp. sagittatum var. abrasum Dahlst., which was considered by Pugsley to be a variety of H. oistophyllum, bears littie resemblance to that species. It is without doubt H. euprepes. H. euprepes is not uncommon in Scotland and Wales, and has a local distribution in north- east Ireland and the Faeroes (as H. ardisodon Dahlst.). It has not been recorded from the Shetlands, Iceland or the Continent. In Wales, northern England, northern Ireland, and the greater part of Scotland, H. caledonicum can be recognised by its thick, ovate, caesious leaves and broad, truncate- tipped phyllaries, but in the extreme north of Scotland, the Orkneys and the Faeroes (as H. subrubicundum Dahlst., H. perintegrum Dahlst., H. cordifrons Dahlst., H. leniscotum Omang, H. subcordifrons (Zahn) Omang) it shows many modifications of form. A variety (var. austroislandicum Omang) of H. perintegrum Dahlst. has been described from Iceland, but we have seen no specimens of it. H. caledonicum is not known from the Shetlands or the Continent. H. scoticum can usually be distinguished by its numerous (2-7) dark green stem-leaves and long deflexed hairs on the base of the stem. It is of local occurrence in Wales, northern England, Scotland, north-west Ireland and the Faeroes (as H. scoticiforme a but is not recorded from Iceland, the Shetlands or the Continent. Watsonia 5 (4), 1962. TauLe 1. The Distribution in Europe of the species of Hieracium which occur in the Orkneys ee SS ee Mainland Northern Southern Species 9 half. of half of Wales Treland Shetland Isles Faeroes Iceland Scandinavia Continental Scotland England England [ Europe ae + ct R H. i Dahls Feo 5 — + + —_ as H. perampliforme Dahlst. vestmannaense _— — H, anglicum Onune ae 1H irieum ae + — = =f — as H. peramplum Dahlst. = — = + HH, argenteum ap aP = + + = — as H. microdon Dahlst. + = fe H. scoticum ar +R —_ +R + — as H. scoticiforme Dahlst. — — — H, orimeles AP +R = +4 a a == = ie Si H, sarcophylloides +R = _ — — — + = 2 as H. pycnodon Dahlst. = + piennrenes + = = + + = as H. ardisodon Dahlst. = » zs sr as H. subrubicundum Dahlst. 2 as H. perintegrum H. perintegrum BEB var, Th + + = + + — H. cordifrons Dahist. austroislandicum — = HH Realedouicun H, leniscotum Omang Omang H. subcordifrons (Zahn) Omang H. latobrigorum ar + = = + = oa = = + R H. maritimunt +R = = = +R as H. obesifolium Pugsl. = =, +R eS H. pilosella® t + + a t = = = + re | ig alte i Hi aurantiacum, sensu lato} | +N +N +N +N | EN = & pe | 4N e ++ = present absent R = rare N = naturalised *H. pilosella also occurs in Asia Minor and Siberia. tH. aurantiacum, sensu lato, occurs as a naturalised plant in North America, img 23 = i S Sel) se eS, : ; is f i s) . 1 . a - Cah < q ; ; = t { ne Nay ‘ : 3 an iy 7 a ‘ " ie t. Weep han ¥ : 7 ‘ ' Beane a R ba b, D ‘ js i p de = ont 7 salle ~ Paar si eey yt Ru bahrain ta iy # 7 t ¢ cit : ft 4 kK i ; - By ] i . | me 4 ; i + tei } ] t : y ‘ - - = ; ig at : ; i & — Alesehes : a0 2 } 5 é « *, his 5 ; . x . 4 é “4s = ‘4 HIERACIA OF THE ORKNEY ISLANDS JA) In the Orkney Isles the Hieracia are mainly distributed along the south coast of the Mainland and on the island of Hoy, usually occurring on rocks, banks of streams or low sea cliffs, where there is a high humidity and good drainage, and where they are protected from the depredations of grazing animals, particularly sheep. If the distribution of the Orkney species of the Subgenus Hieracium is considered, it will be seen from Table | that all the species occur on the Scottish mainland while many are to be found in Ireland, Wales and the north of England. They are completely absent from central and southern England. Perhaps the most striking feature is that the only species represented in the Shetlands are H. orimeles and H. maritimum. When considering the European distribution of the group we have attempted to assign to what we consider to be their proper place certain specific and varietal names created by Dahlstedt and Omang from material gathered in the Faeroes and in Iceland. The Orcadian species are well represented in the Faeroes whereas only two or possibly three have been recorded from Iceland. They are very poorly represented in Continental Europe, only H. argenteum, H. maritimum and possibly H. sarcophylloides occurring in Scandinavia, and only H. /atobrigorum in central Europe. Not a single species of the Sect. A/pestria has been found in the Orkneys, in spite of the fact that numerous species of this section are common in the Shetlands, Faeroes, Iceland and Scandinavia. H. sparsifolium Lindeb. (H. stictophyllum Dahlst. ex W. R. Linton) of the Section Tridentata Fries is common in the Shetlands (as H. pseudo- protractum Pugsl. non Noig.), the Faeroes (as H. faeroense Dahlst. and H. epileucoides Dahlst.), [Iceland and Scandinavia, but is absent from the Orkneys. The following key may help visitors to the Orkneys to identify the species of Hieracium which are known to occur there. Identifications should always be confirmed by comparison with authentic specimens. Early to mid-July is probably the time to see most species at their best, but H. pilosella and H. aurantiacum usually flower about a month earlier while H. latobrigorum and H. maritimum may not flower until late July or early August. After the key the full synonymy of each species and a list of all the Orkney specimens we have examined are given. The National Grid References are only given as a guide to the locality and may not denote the exact spot where the plant was found. ARTIFICIAL KEY TO THE SPECIES OF HIERACIUM RECORDED IN THE ORKNEYS 1. Ligules orange-red H. aurantiacum Ligules yellow (sometimes with red stripe on back) 2 2. Plant stoloniferous; ligules usually with red stripe on back H. pilosella Plant not stoloniferous; ligules concolorous 3 3. Cauline leaves usually more than 12 4 Cauline leaves usually less than 7 5 4. Leaves mostly ovate to elliptic-lanceolate; phyllaries with many microglands, a number of very small glands and a few simple hairs ; Hi. latobrigorum Leaves narrowly oblong; phyliaries with very few, very smal! glands, and/or an occasional simple hair, or glabrous H. maritimum 5. Ligules strongly pilose-tipped; cauline leaves semi-amplexicaul 6 Ligules glabrous; cauline leaves not amplexicaul i 6. Large cauline leaves rarely more than 2, hairs of stem long but not shaggy H. anglicum Large cauline leaves usually more than 3; hairs of stem long and shaggy HA. iricum 7. Phyllaries narrow and + acute 8 Phyllaries broader and + obtuse 10 8. Leaves yellowish-green, rather broadly ovate, often with a truncate base, cauline usually 0, rarely | H. sarcophylloides Leaves rather glaucous (leaden) or light green, narrowly ovate or lanceolate, + cuneate-based, cauline 1-3 9 9. Leaves rather glaucous (leaden), almost glabrous above H. argenteum Leaves light green, densely short-setose above H. orimeles Watsonia 5 (4), 1962. 218 PD. SEEL and CYRIL WEST 10. Leaves rather dark green, + elliptic, base cuneate or attenuate, cauline 0-3; phyllaries c. 10 mm. long, linear-lanceolate H. euprepes Leaves caesious, + ovate or elliptic, base rounded or cuneate, cauline 0-3; phyllaries 10-12 mm. long, broad linear-lanceolate, truncate-tipped H. caledonicum Leaves rather dark green, + ovate or occasionally elliptic, base cuneate, cauline 2-7; phyllaries 11-12 mm. long, linear-lanceolate Hi. scoticum Subgenus HIERACIUM Section CERINTHOIDEA Fries HIERACIUM ANGLICUM Fries (1848) 93; Fortescue (1882) 370; W. R. Linton (1887) 154; E. S. Marshall (1901) 269; Spence (1914) 43; Johnston (1914) 209; Pugsley (1948) 68. [H. iricum sensu Spence (1914) 41 pro parte.] | H. anglicum var. cerinthiforme Backh., Bennett (1915) 55. H. mougeotii subsp. anglicum var. cerinthiforme (Hanb.) Zahn, Johnston (1928) 12. H. mougeotii subsp. anglicum forma acutifolium (Backh.) Zahn, Johnston (1928) 12. H. patens Dahlst. in Johnston (1934) 473 (Lectotype: west side of Nowt Bield, Ward Hill, Hoy, 10 July 1933, H. H. Johnston No. 4742 (E)). Southwest side of Berriedale, Hoy, 30/1901, 8 August 1927, H. H. Johnston nos. 3651 and 3655A (E). Near the Kame, Hoy, 30/1904, 9 August 1886, H. H. Johnston (E). Ward Hill, Rousay, 12 August 1880, H. H. Johnston (E); 28 August 1883, H. H. Johnston (E); 10 August 1886, W. R. Linton (BM); 23 July 1907, M. Spence (E); 30 August 1916, H. H. Johnston no. 426 (E); 26 July 1927, J. Sinclair no. 436 (herb. Sinclair); 31 July 1928, J. Sinclair no. 591 (herb. Sinclair & E). Valley Burn, Hoy, 12 July 1928, J. Sinclair no. 571 (herb. Sinclair) and H. H. Johnston nos. 4006 and 4008 (E). Burn of Segal, Hoy, 30/2002, 11 July 1933, H. H. Johnston no. 4749 (E). Lenders Fea, Skecking Gill, Hoy, 30/2004, 25 July 1925, J. Sinclair no. 155 (herb. Sinclair), H. H. Johnston no. 3120 (BM). Braebusier Burn, Hoy, 30/2105, 7 July 1928, J. Sinclair no. 568 (herb. Sinclair) and H. H. Johnston no. 3990 (E). Glen of Gair (Greor), Ward Hill, Hoy, 30/2203, 20 August 1885, H. H. Johnston (BM). West side of Sel Wick, Hoy, 30/2205, 7 July 1928, J. Sinclair no. 566 (herb. Sinclair), H. H. Johnston no. 3987 (E). Trowie Glen, Hoy, 30/2300, 28 July 1925, H. H. Johnston no. 3138 (E, BM); 5 August 1927, H. H. Johnston no. 3644B (E). Nowt Bield, 30/2301, 5 August 1927, H. H. Johnston no. 3646 (E); 2 August 1932, H. H. Johnston no. 4340 (E); 10 July 1933, J. Sinclair no. 954 (herb. Sinclair) and H. H. Johnston nos. 4740, 4741 and 4742 (E); 1953, C. West (CGE). Junction of the burn of the Nowt Bield and the Trowie Glen, Hoy, 30/2300, 16 July 1928, H. H. Johnston no. 4046 (E). Burn of Quoys, Hoy, 30/2402, 1 August 1927, H. H. Johnston no. 3616 (herb. Sinclair) and no. 3616A (E); 5 August 1927, J. Sinclair no. 457 (herb. Sinclair); 6 July 1928, J. Sinclair no. 637 (herb. Sinclair). Dwarfie Hamars, July 1953, C. West (CGE). Red Glen, between Burandie and Dwarfie Hamars, Hoy, 30/2500, 10 August. 1886, H. H. Johnston (E). Bring Head, Hoy, 30/2701, 15 August 1937, J. Sinclair no. 481 (herb. Sinclair). Scapa, August 1875, J. Boswell-Syme (BM). Westerow, Rackwick, Hoy, 39/2098, 14 July 1928, J. Sinclair nos. 578 and 639 (herb. Sinclair), H. H. Johnston nos. 4025 and 4025A (E). South Burn, Hoy, 39/2099, 14 July 1928, H. H. Johnston no. 4039 (E); 20 July 1932, H. H. Johnston no. 4515 (E). Pegal Bay, 39/3097, August 1880, J. B. Fortescue (CGE, BM); 22 August 1894, H. H. Johnston (E); 16 July 1912, H. H. Johnston (E); 3 August 1927, J. Sinclair no. 453 (herb. Sinclair & E); 7 July 1933, H. H. Johnston nos. 4727 and 4728 (E). HIERACIUM IRICUM Fries (1848) 60; Spence (1914) 41 pro parte; Pugsley (1948) 80. Burn of the Nowt Bield, Hoy, 30/2300, 2 August 1932, H. H. Johnston no. 4540A (E). Dwarfie Hamars, Hoy, 30/2400, 10 August 1886, H. H. Johnston (E, BM). Burn of Quoys, Hoy, 30/2402, 3 July 1926, J. Sinclair no. 328 (herb. Sinclair); 1 August 1927, H. H. Johnston no. 3616 (E); 6 July 1928, J. Sinclair no. 634 (herb. Sinclair); 1953, C. West (CGE). Section OREADEA Fries HIERACIUM ARGENTEUM Fries (1848) 99; Pugsley (1948) 92. [| H. rivale var. dasythrix sensu Johnston (1920) 32.] H. pseudomicrodon Dahlst. in Johnstone (1929a) 5 (Lectotype: Grassy rocky crags at Watsonia 5 (4), 1962. HIERACIA OF THE ORKNEY ISLANDS 219 burnside in a ravine on the southwest side of Berrie Dale, below a waterfall, Hoy, alt. 280 ft., 8 July 1927, H. H. Johnston, no. 3657 (E)); Johnston & Sinclair (1930) 24. [H. saxifragum sensu Dahlst. in Johnston (1933a) 3; sensu Johnston (1933b) 23, sensu Johnston (1934) 472.] Southeast side of Enegars, Kame, Hoy, 30/1904, 19 August 1932, J. Sinclair no. 770 (herb. Sinclair & E); 8 July 1933, H. H. Johnston no. 4731 (E, BM). Berrie Dale, Hoy, 30/2001, 7 September 1914, H. H. Johnston no. 362 (E); 8 July 1927, H. H. Johnston no. 3657 (FE); 8 August 1927, J. Sinclair no. 471 (herb. Sinclair); 26 August 1928, H. H. Johnston and J. Sinclair (herb. Sinclair & E); 1 August 1932, J. Sinclair no. 760 (herb. Sinclair & E). HIERACIUM SCOTICUM Hanb. (1888) 206; Spence (1914) 42 and 130; Johnston (1914) 220; Johnston (1925) 6; Pugsley (1948) 98. [H. gothicum sensu Johnston (1882) 370.] [H. rubicundum sensu Johnston (1914) 210 (quoad loc. Malsetter).| H. scoticum var. submaculatum Dahlst. in Johnston (1929b) 5 and (1930) 24 (Lectotype: Heathery rocky ravine at burnside, Roonie (? = Runcie) Gill, Hoy, 210 ft. alt., 15 August 1929, H. H. Johnston no. 4241 (E)). Berrie Dale, 30/2001, 22 August 1929, H. H. Johnston no. 4280 (herb. Sinclair & E). Burn of Sowa Dee, Sandwick, 30/2314, 26 August 1912, H. H. Johnston (E); 15 July 1913, H. H. Johnston (E); 23 July 1925, H. H. Johnston no. 3107 (E, CGE); 30 June 1928, J. Sinclair no. 645 (herb. Sinclair); 1954, N. D. Simpson no. 54/153 (herb. Simpson). Hangarback, Gyre, Orphir, 30/3403, 18 August 1885, H. H. Johnston (E). North Hill, Westray, 30/4048, 15 July 1883, H. H. Johnston (E); 28 August 1913, H. H. Johnston (E). Roonie (? = Runcie) Gill, Hoy, 39/2299, 15 August 1929, H. H. Johnston no. 4241 (E). Malsetter, Waas, Hoy, 39/2789, 11 August 1913, H. H. Johnston (E). North-east side of Aith Hope, South Walls, Hoy, 39/2989, 21 July 1913, H. H. Johnston (£); 25 July 1913, H. H. Johnston (E); 4 August 1913, H. H. Johnston (E); 9 August 1928, J. Sinclair no. 597 (herb. Sinclair & E); 2 August 1929, J. Sinclair no. 663 (herb. Sinclair & E); 20 August 1929, H. H. Johnston no. 4259 (E). West side of Fara, South Isles, 39/3195, 30 July 1924, H. H. Johnston no. 2669 (E). HIERACIUM ORIMELES Hanb. ex W. R. Linton (1901) 106; Pugsley (1948) 102. Waulkmill Bay, 30/3806, July 1953, C. West (CGE). Sect. SUBOREADEA Pugsley HIERACIUM SARCOPHYLLOIDES Dahlst. in Warming (1903) 629 (Holotype: Osterd, Kodlen ad Eide at 400 m., 17 July 1895, H. G. Simmons in Herb. Copenhagen). [H. pallidum sensu Johnston (1882) 370.] [H. silvaticum var. tricolor sensu Spence (1914) 131; sensu Johnston (1914) 211.] [H. caesium sensu Spence (1914) 43.] [H. murorum sensu Spence (1914) 43.] [H. britannicum sensu Spence (1914) 42; sensu Johnston (1925) 6.] [H. sarcophyllum var. expallidiforme sensu Bennett (1915) 55; sensu Johnston (1920) 32.] [H. rubicundum var. boswellii sensu Johnston (1920) 31.] H. dasypodum Dahlst. in Johnston (1929a) 5 (Lectotype: Heathery crags on hillside between Grut Fea and the Glen of Button, Hoy, 12 July 1928, H. H. Johnston no. 4012 (E)); Johnston (1929b) 3; Johnston & Sinclair (1930) 24; Pugsley (1948) 115. H. pycnodon var. acutidens Dahist. in Johnston (1929a) 4 (Lectotype: Grassy, rocky banks of burn, White Glen, Ward Hill, Hoy, 16 August 1928, J. Sinclair no. 602 (E)); Johnston (1929b) 3; Sinclair (1930) 25. H. sinclairii Dahlst. in Johnston (1932) 5 (Lectotype: Moist mossy soil on sandstone cliffs, Dwarfie Hamars, Hoy, 24 July 1931, J. Sinclair no. 706 (E)). Between Grut Fea and the Glen of Button, Hoy, 30/1901, 12 July 1928, J. Sinclair no. 575 (herb. Sinclair) and H. H. Johnston no. 4012 (E, BM). South of the meadow of the Kame, Hoy, at 750 ft., 39/1904, 11 September 1914, H. H. Johnston no. 373 (E) and Enegars, south-east of Kame, 880 ft., 30/1904, Watsonia 5 (4), 1962. 220 P. D. SELL and CYRIL WEST 8 July 1933, H. H. Johnston no. 4735 (E). Ward Hill, Hoy, 30/2301, 31 July 1928, J. Sinclair no. 590 (herb. Sinclair & E); 16 August 1928, J. Sinclair no. 602 (herb. Sinclair & E); 22 August 1929, H. H. Johnston no. 4274 (E); 14 August 1938, J. Sinclair no. 1234 (herb. Sinclair); July 1953, C. West (CGE). Dwarfie Hamars, 30/2400, 28 August 1883, H. H. Johnston (E); 14 July 1900, E. S. Marshall no. 2328 (BM) and as Linton Set no. 178 (BM, CGE); 22 July 1912, H. H. Johnston (E); 24 July 1925, J. Sinclair no. 147 (herb. Sinclair); 26 July 1929, J. Sinclair no. 661 (herb. Sinclair & BM); 24 July 1931, J. Sinclair no. 716 (E, BM) and no. 715 (herb. Sinclair & E); 2 August 1932, H. H. Johnston nos. 4534 and 4535 (E); July 1953, C. West (CGE). Bring Head, Hoy, 30/2702, 4 July 1877, H. H. Johnston (E); 7 July 1933, H. H. Johnston no. 4726 (E) and J. Sinclair no. 952 (herb. Sinclair). Sect. VULGATA Fries HIERACIUM EUPREPES Hanb. (1892) 206 (Lectotype: Craig Dulyn, Caernarvon, 28 July 1891, F. J. Hanbury (BM)); Pugsley (1948) 166. [H. vulgatum sensu Fortescue (1882) 370, sensu Johnston (1882) 370; sensu Spence (1914) 42. | H. orcadense W. R. Linton (1893) 196; Johnston in Cryer (1913) 264; Johnston (1914) 211; Spence (1914) 41 and 131; Pugsley (1948) 168. | H. orarium var. fulvum sensu Marshall in Marshall & White (1901) 638.] | H. ? caesium sensu Spence (1914) 43.] [H. buglossoides sensu Spence (1914) 130; sensu Johnston (1914) 211.] |H. schmidtii var. crinigerum sensu Spence (1914) 41; sensu Johnston (1925) 6.] [H. ? orarium sensu Bennett (1915) 55.] [H. silvaticum var. subtenue sensu Johnston (1926) 300.] [H. holopleurum sensu Dahlst. in Johnston (1928) 11; sensu Johnston (1929b) 3; sensu Johnston (1929c) 619.] H. sagittatum subsp. sagittatum var. abrasum Dahlst. in Johnston (1929a) 6 (Lectotype: Heathery, sandy banks of Valley Burn, Hoy, 50 ft. alt., 12 July 1928, H. H. Johnston no. 4007 (E)); Johnston & Sinclair (1930) 35. H. subalpestrifrons Dahlst. in Johnston (1929a) 7 (Lectotype: Grassy, heathery banks of the Burn of Quoys, Hoy, 6 July 1928, J. Sinclair no. 565 in herb. Sinclair); Johnston (1929b) 4; Johnston and Sinclair (1930) 25. H. paraliaeforme Dahlst. in Johnston (1932) 7 (Lectotype: Moist mossy soil on sandstone cliffs, Dwarfie Hamars, Hoy, 24 July 1931, J. Sinclair no. 717 in herb. Sinclair). H. kalsoense subsp. burnense Druce & Zahn in Druce (1932 for 1931) 559 (Holotype: South Burn near Rackwick, Hoy, August 1931, G. C. Druce (OXF)). H. sagittaticeps Dahlst. in Johnston (1932) 4 (Holotype: Heathery banks of the Burn of Segal, Hoy, 8 August 1931, H. H. Johnston no. 4425 (E)). H. subexpallescens Dahlst. in Johnston (1933 a) 3 (Lectotype: Heathery freestone sea- cliffs, Enegras, Craig Gate, southeast of Rackwick, Hoy, 1 August 1932, H. H. Johnston no. 4532 (E)); Johnston (1933 b) 23. H. oistophyllum var. abrasum (Dahlst.) Pugsl. (1948) 159. North Geo, between Bre Brough and The Sow, 30/1802, 13 August 1929, J. Sinclair no. 671 (herb. Sinclair). Grut Fea, Hoy, 30/1901, 7 August 1938, J. Sinclair no. 1233 (herb. Sinclair). Enegars, south- east of the Meadow of the Kame, Hoy, 30/1904, 8 July 1933, J. Sinclair (herb. Sinclair) and H. H. Johnston no. 4734 (BM). South-west side of Berriedale, Hoy, 30/2001, 26 August 1928, J. Sinclair no. 699 (herb. Sinclair). Burn of Segal, Hoy, 30/200!, 8 August 1931, H. H. Johnston no. 4425 (KE); 1 August 1932, J. Sinclair no. 761 (herb. Sinclair) and H. H. Johnston no. 4533 (BM). Trowie Glen, Hoy, 30/2300, 5 August 1927, J. Sinclair no. 462 (herb. Sinclair). West side of Nowt Bield, Ward Hill, 30/2301, 8 July 1929, J. Sinclair no. 660 (herb. Sinclair). Dwarfie Hamars, 30/2400, 14 July 1905, E. S. Marshall and W. A. Shoolbred as Linton Set no. 142 (BM, CGE); 1921, G. C. Druce (BM); 24 July 1931, J. Sinclair no. 717 (herb. Sinclair & BM); 2 August 1932, H. H. Johnston no. 4536(E). Burn of Quoys, Hoy, 30/2402, 31 July 1894, F. J. Hanbury (BM); 6 July 1928, J. Sinclair no. 565 (herb. Sinclair). West side of Waulkmill Bay, Orphir, Mainland, 30/3706, 15 August 1881, H. H. Johnston (E); 17 July 1900, E. S. Marshall no. 2331 (CGE); 20 July 1912, H. H. Johnston (BM, CGE); 9 August 1933, J. Sinclair no. 959 (herb. Sinclair); July 1953, C. West (CGE); 1954, N. D. Simpson no. 54/084 (herb. Simpson). Enegras, Craig Gate, south-east of Rackwick, 39/1998, Watsonia 5 (4), 1962. HIERACIA OF THE ORKNEY ISLANDS 221 1 August 1932, J. Sinclair no. 759 (herb. Sinclair) and H. H. Johnston no. 4532 (BM, E). South Burn, Hoy, 39/2099, 12 June 1929, H. H. Johnston no. 715 (BM); 23 August 1927, J. Sinclair no. 486 (herb. Sinclair) ; 14 July 1928, H. H. Johnston no. 4038 (BM). Roonie (? = Runcie) Gill, 39/2299, 26 July 1929, J. Sinclair no. 662 (herb. Sinclair); 15 August 1929, H. H. Johnston nos. 4241A (E) and 4245 (BM); 21 August 1938, J. Sinclair no. 1235 (herb. Sinclair). Valley Burn, Hoy, 12 July 1928, J. Sinclair no. 572 (herb. Sinclair) and H. H. Johnston no. 4007 (BM, E). HIERACIUM CALEDONICUM Hanb. (1889) 75; E. F. & W. R. Linton (1893) 178; W. R. Linton (1905) 26; Johnston (1913) 263; Spence (1914) 130; Johnston (1914) 210; Pugsley (1948) 106. H. rubicundum Hanb. (1892) 208; E. F. & W. R. Linton (1893) 178; Johnston in Cryer (1913) 262; Johnston (1914) 210 pro maximo parte; Spence (1914) 130. H. boswellii E. F. Linton (1893) 178. | H. pulchellum sensu Spence (1914) 43.] [H. ? murorum sensu Spence (1914) 43.| H. rubicundiforme (Zahn) Roffey (1925) 26 no. 1028. Grut Fea, Hoy, 30/1901, 8 August 1927, H. H. Johnston no. 2659 (herb. Sinclair). South-east of Enegars, Kame, Hoy, 880 ft. 30/1904, 19 August 1932, J. Sinclair no. 769 (herb. Sinclair). Berriedale, Hoy, 30/2001, 7 September 1914, H. H. Johnston no. 360 (E); 8 August 1927, J. Sinclair no. 470 (herb. Sinclair) and H. H. Johnston no. 3655 (E). West side of Waulkmill Bay, Orphir, 30/3706, 19 and 20 July 1912, H. H. Johnston (E, CGE); 3 August 1924, R. J Burdon (CGE); July 1953, C. West (CGE); 1954, N. D. Simpson nos. 54/085, 54/088 and 54/339 (herb. Simpson). Ward Hill, Rousay, 30/3829, 30 August i916, H. H. Johnston no. 422 (E); 23 July 1917, M. Spence (E). Fitty Hill, Westray, 30/4244; 3 August 1934,.H. H. Johnston no. 4948 (herb. Sinclair & E). Sea-cliff, Scapa, St. Ola, Mainland, 30/4308, 5 July 1912, H. H. Johnston (E, CGE); 1953, C. West (CGE). Scapa Pier, 30/4408, 18 July 1894, F. J. Hanbury (CGE); 17 July 1900, E. S. Marshall and W. A. Shoolbred as Linton Set no. i131 (CGE). Sect. FOLIOSA Fries HIERACIUM LATOBRIGORUM (Zahn) Roffey (1925) 29 no. 1230; Pugsley (1948) 272. | H. strictum sensu Johnston (1882) 370; sensu Fortescue (1882) 370; sensu Bennett (1915) 55. ] H. strictum var. amplidentatum Hanb. pro parte (1894) 232 (quoad loc. Pegal Burn); sensu Spence (1914) 42. [H. corymbosum vat. salicifolium sensu Spence (1914) 42; sensu Johnston (1920) 43 and (1925) 6: [H. auratum sensu Spence (1914) 42; sensu Johnston (1925) 7.] H. inuloides subsp. striatum var. pseudauratum Zahn (1921) 901; Johnston (1927) 411. | H. inuloides subsp. strictum var. amplidentatum sensu Johnston (1929a) 8.] | H. crocatum var. normale sensu Druce (1932 for 1931) 561.] H. aestivum subsp. hemitrachys Druce & Zahn in Druce (1932 for 1931) 561 (Holotype: Pegal Bay, Hoy, August 1931, G. C. Druce (OXF)). H. johnstonii Dahlst. in Johnston (1932) 8 (Lectotype: north side of Pegal Head, 13 August 1931, H. H. Johnston no. 4461 (K)). H. polyphyllum Dahlst. in Johnston (1933a) 2 & 5 (Lectotype: Grassy sea cliffs between Lee Craig and Stanger Head, Hoxa Sound, Flotta, 15 August 1932, H. H. Johnston no. 4560 (E)); Johnston (1933b) 24; non Willd., Enum. hort. Berol., Suppl., 54 (1813). H. subpolyphyllum Pugsl. (1946) 347 and (1948) 279 (Lectotype: that of H. polyphyllum Dahlst. non Willd.). Hobbister Rocks, Orphir, 30/3805, August 1873, J. B. Syme (BM, E) and 1874 and 1879 (BM); August _ 1880, J. B. T. Fortescue (BM); July 1953, C. West (CGE); 1954, N. D. Simpson no. 54/098 (herb. Simpson). Pegal Bay, North Walls, Hoy, 39/3098, August 1880, J. B. T. Fortescue (CGE, OXF); 22 August 1894, H. H. Johnston (E, CGE, OXF); 9 August 1926, H. H. Johnston, no. 3484 (EK); 20 August 1929, J. Sinclair no. 678 (herb. Sinclair); 13 August 1931, H. H. Johnston no. 4461 (E) and J. Sinclair no. 724 (herb. Sinclair) ; August 1931, G. C. Druce (OXF). Between Lee Craig and Stanger Head, Hoxa Sound, Flotta, 39/3793, 26 September 1924, H. H. Johnston no. 2847 (E); 15 August 1932, J. Sinclair no. 763 (herb. Sinclair) and H. H. Johnston no. 4560 (E, BM). : Watsonia 5 (4), 1962. apap P. D. SELL and CYRIL WEST HIERACIUM MARITIMUM Hanb. (1904) 267; Pugsley (1948) 289. [H. inuloides subsp. strictum forma angustifolium sensu Dahlst. in Johnston (1929a) 8; sensu Johnston & Sinclair (1930) 26; sensu Johnston (1934) 472.] Aith Hope, South Walls, Hoy, 39/2989, 9 August 1928, J. Sinclair no. 596 (herb. Sinclair); 20 August 1929, H. H. Johnston no. 4258 (EZ, CGE). Subgen. PILOSELLA Tausch Sect. PILOSELLINA Fries HIERACIUM PILOSELLA L. (1753) 800; Fortescue (1882) 370; Spence (1914) 41; Pugsley (1948) 314. Bu Farm, Hoy, 30/2304, 5 July 1926, J. Sinclair no. 330 (herb. Sinclair). Swanbister, 30/3405, July 1849, J. T. Syme (BM). St. Andrews, Rousay in Spence (1914). Not uncommon on Mainland and Hoy according to J. T. Syme in Fortescue (1882). var. TRICHOSCAPUM (Naeg. & Peter) Pugsl. (1948) 315. Near Maeshow, Mainland, 30/3112, August 1880, J. B. Syme (BM). Sect. COLLININA Naeg. & Peter HIERACIUM AURANTIACUM L. (1753) 801, sensu stricto; Johnston (1927) 415. Top of a stone wall, 90ft. alt., near the garden of Holland House, Papa Westray, 30/4851, 22 June 1926, H. H. Johnston no. 3330 (E). REFERENCES BENNETT, A. (1915). Notes on the Flora of the Orkney Isles. Trans. Bot. Soc. Edinb., 27. DaHLsTeDT, H. (1903). The Hieracia from the Faeroes, in Warming, E., Botany of the Faeroes, 2. Copenhagen. Druce, G. C. & ZAHN, K. H. (1932). Rep. Bot. (Soc.) Exch. Cl., 9. ForTESCUE, W. I. (1882). A New List of the Flowering Plants and Ferns of Orkney. Scot. Nat. 47. Fries, E. (1848). Symbolae ad Historiam Hieraciorum. Uppsala. HAnsury, F. J. (1888). Notes on some Hieracia new to Britain. J. Bot., Lond., 26. Hansury, F. J. (1889). Further notes on Hieracia new to Britain. J. Bot., Lond., 27. HAnNBuRY, F. J. (1892). J. Bot., Lond., 30. Hansury, F. J. (1894). Notes on British Hieracia. J. Bot., Lond., 32. Hanpsury, F. J. (1904). Hieracium in Babington, C. C. Manual of British Botany, ed. 9. London. JoHNSTON, H. H. (1913) in Cryer, J., Report of the Distributer for 1912 in Rep. Bot. (Soc.) Exch. Cl. 3. Jounston, H. H. (1914). Additions to the Flora of Orkney, and Notes on some rare or interesting Orkney Plants. Trans. Bot. Soc. Edinb., 26, 207-217 & 217-226. JounstTon, H. H. (1920). Additions to the Flora of Orkney, and Observations on “‘ Notes on the Flora of the Orkney Isles by A. Bennett.’’ Trans. Bot. Soc. Edinb., 28, 23-42 & 43-46. JOHNSTON, H. H. (1925). Additions to the Flora of Orkney. (This is a corrected reprinting of a paper published in Ann. Scot. Nat. Hist. 173-183 (1895). We have quoted pages of reprint). JOHNSTON, H. H. (1926). Additions to the Flora of Orkney. Trans. Bot. Soc. Edinb., 29. JOHNSTON, H. H. (1927). Trans. Bot. Soc., Edinb., 29. JOHNSTON, H. H. (1928). Additions to the Flora of Orkney. Edinburgh. JOHNSTON, H. H. (1929 a & b). Additions to the Flora of Orkney, Edinburgh. JOHNSTON, H. H. (1929c). Plant Notes in Rep. Bot. (Soc.) Exch. Cl. for 1928., 8. JouHNsTon, H. H. (1930) with or without Sinclair, J., for 1929, Plant notes in Rep. Bot. (Soc.) Exch. Cl., 9. JOHNSTON, H. H. (1932). Additions to the Flora of Orkney. Edinburgh. JOHNSTON, H. H. (1933a). Additions to the Flora of Orkney. Edinburgh. JOHNSTON, H. H. (1933b). Plant Notes. Rep. Bot. (Soc.) Exch. Cl. 10. JoHNSTON, H. H. (1934). Plant Notes. Rep. Bot. (Soc.) Exch. Cl. 10. LINNAEUS, C. (1753). Species Plantarum, ed. 1. Stockholm. LINTON, E. F. & W. R. (1893). British Hawkweeds in J. Bot., Lond., 31. Watsonia 5 (4), 1962. HIERACIA OF THE ORKNEY ISLANDS 223 LINTON, W. R. (1887). Rep. Bot. (Soc.) Exch. Cl. 1. LINTON, W. R. (1901). British Hieracia. J. Bot., Lond., 39. Linton, W. R. (1905). An Account of the British Hieracia. London. MARSHALL, E. S. (1901). Plants of N. Scotland, 1900. J. Bot., Lond., 39. MarsHALL, E. S. & Wuite, J. W. (1901). Report of the Distributer in Rep. Bot. (Soc.) Exch. Cl. 1. Pucs.ey, H. W. (1946). Hieracium in Clapham, A. R., Check List of British Vascular Plants. J. Ecol., 31. PuGsLey, H. W. (1948). Prodromus of the British Hieracia. J. Linn. Soc., Lond., 54. Rorrey, J. (1925). Hieracium in London Catalogue of British Plants, ed. 11, London. Sincxair, J. (1930). Plant Notes. Rep. Bot. (Soc.) Exch. Cl. 9. SPENCE, M. (1914). Flora Orcadensis. Kirkwall. ZAHN, K. H. (1921). Hieracium in Engler, A., Pflanzenreich, IV. 280. Watsonia 5 (4), 1962. A STUDY OF VARIATION EUPHRASIA BY MEANS OF OUTDOOR CULTIVATION Byte) EO University Botanic Garden, Cambridge* ABSTRACT Samples of Euphrasia from populations at six localities on chalk or limestone in the Midlands and south-east of England were cultivated in the garden in order to compare them without interference from environmental effects. The variation between populations in the wild and in cultivation is described. It was concluded that five samples were more or less normal EF. nemorosa, differing only slightiy from one another; also that one of them was mixed with, and apparently hybridising with, E. stricta (not previously known in Britain, and possibly an alien introduction). One population appeared to representa slightly divergent ecotype of E. nemorosa and another an extreme ecotype of the same species. A hybrid between the latter and E. pseudokerneri appeared among the plants cultivated. Four more samples were grown a year later and it was concluded that these comprised three normal forms of E. nemorosa and a form of E. confusa. Two forms of E. anglica showed fairly marked differences although growing only three-quarters of a mile apart. The differences may have been caused by hybridisation at one of the localities. The cultures provided instructive information on interspecific and intraspecific variation in a taxonomic group in which the most closely related species are very similar to one another and in which variation within species is considerable. INTRODUCTION Owing to the fact that the species of Euphrasia are annual and semi-parasitic plants they show great variations in vigour. These may be seen within populations, but overall differences of vigour occur between populations as well. Ignorance of the effects of this variation on morphology suggested that the sampling of wild populations for morphological study would be unreliable. A large-scale garden trial cultivation was therefore undertaken to investigate variation in E, nemorosa. A small trial of E. anglica was carried out at the same time, and further populations of E. nemorosa were later com- pared by the same method. The species of Euphrasia are variable, ill-defined, and subject to hybridisation, though each has several characters by which it normally differs from its nearest relatives (Yeo, 1955). In spite of the difficulty of recognizing even species on the basis of small and re- latively inconstant characters, &. nemorosa was subdivided by Pugsley (1930) into five varieties, and Warburg (1952) stated that three of them were distinct ecologically and might prove to be worthy of treatment as subspecies. It was thought that the cultures might help in the understanding of some of these varieties, especially of var. calcarea Pugsl., since most of the samples of this species were collected from chalk or limestone. COMPARISON OF SEVEN EUPHRASIA SAMPLES GROWN IN THE GARDEN Description of experiment The ‘ whalehide pot’ method of cultivation was used in this experiment, which was carried out at Leicester in 1953. Each bituminized paper pot was planted in March with one seedling of Euphrasia and one of Plantago lanceolata ; the pots were then embedded in the garden; Euphrasia and host plants that died during the few days occupied by this process were replaced at once. The aim of the experiment was to compare the offspring *Most of the work here described was done during the tenure of a Research Scholarship at the then University College of Leicester. 224 Watsonia 5 (4), 1962. VARIATION IN EUPHRASIA pipe) of wild populations from six localities when grown in the same environment. Six samples were grown in a set of randomised blocks, an arrangement which permits an analysis of variance to be made on the statistical data. There were six blocks, each consisting of six rows of fourteen Euphrasia plants each. Each sample of Euphrasia was represented by one row in each block. The individual plants were assigned to the blocks at random, and the order of the rows in each block was also chosen at random. A seventh sample, of which few seedlings were availabie, was grown in a row of ten plants near the main experiment. On four occasions from mid-May to mid-June a record was made of the survival of the Euphrasias, signs of establishment on the host, and signs of disorder in the plants. Leaves and flowers were taken from the plants and mounted on cellulose tape on glass. From each plant two leaves were taken, one subtending the last-but-one normally-developed branch or pair of branches (called ‘ leaf 2”), and one from the fourth node above the upper- most branch or pair of branches (called ‘ bract 4’). One flower from each plant was mounted; the aim in collecting these was to obtain mature full-sized flowers which had not begun to wrinkle or shrink. Most of the flowers chosen were therefore ones which had not been shed but which could be pulled off easily. The mounting was begun on 29 July 1953. Between 27 August and 7 September all the plants were pressed, except for some which were pressed early because they were wilting. The height of the pressed plants was measured, and various measurements were made on the mounted leaves and flowers. For this purpose, the lantern-plate cover-glasses on which they were mounted were put into a projector and measurements were made on their enlarged images. In addition, photographs were taken of individual plants in the garden, and photographic contact-prints were made of the mounted leaves and flowers. The Euphrasia seed for this experiment was extracted from many different plants of large or fairly large gatherings of herbarium specimens. The samples wili be referred to by their serial numbers. The six samples used in the main part of this experiment were : E151A Juniper Top, Box Hill, Surrey, v.c. 17; flinty, rabbit-grazed turf on chalk; E166 Watlington Hill, Oxon., v.c. 23; chalky field; E167A near Medmenham, Bucks., v.c. 24; chalky field (with E167B, see below); E210 Waltham Quarry, Waltham on the Wolds, Leics., v.c. 55; grassiand on oolite; E211 near Croxton Kerrial, on Leics.-Lincs. border, v.c. 53/55; grassy track on oolite, about 6 miles from E210; E215 Bedford Purlieus, Northants., v.c. 32; woodland ride on oolite, about 19 miles from E211 and about 17 miles from E210. The additional group of 10 plants grown nearby, numbered E167B, was from the same locality as E167A and was growing mixed with it; all samples, with the probable exception of E167A, were considered to be Euphrasia nemorosa. At Box Hill the popula- tion of E151A was in contact with and apparently hybridising with E. pseudokerneri. Plants thought to be hybrids were excluded from the gathering. Effect of replacement of dead seedlings Many seedlings of Euphrasia, and some of the host, were replaced from 22 to 26 April, about four weeks after the initial potting-up and early replacement of the Euphrasias and hosts. Survival during May and June was slightly less good where replacements had been made, but the percentage of survivors probably established on the host was approximately the same whether replacements had taken place or not. Variation between populations in survival and establishment Table | shows that there was considerable variation in survival up to 31 May between the different samples. This variation was greater than that between the blocks, each of which contained all the samples. The table shows a similar result for establishment on the Watsonia 5 (4), 1962. 226 PR. By YEO host. These figures show that the populations differed from one another in characters affecting their establishment on the host provided and their survival in the garden at Leicester. TABLE 1. Survival and establishment of Euphrasia in 1953 in the garden | % Of Euphrasias alive on 22-26 April | % of surviving Euphrasias probably that survived until 31 May | established on kest on 31 May Sample E1SIA 50 44 E166 73 83 E167A 81 84 E210 93 | 85 E211 92 94 E215 82 67 Block 1 83 74 2 85 75 3 87 qs 4 a7 81 5 75 | 0 6 74 | 81 Deaths of plants later in the season, after they had become established on the host, were mostly due to fraying of the stem base, which in turn was probably the result of attacks of damping-off fungi upon the young seedlings. Many plants which did not die wilted readily in dry weather from this cause. However, the sample E151A, which suffered the heaviest mortality before establishment, showed no ill effects from stem-fraying, apparently because of its dwarf habit. Variation between populations in habit Variation in habit is illustrated by photographs of some of the living plants, and by height measurements. Some of the leaf characters will also be considered here. The plants grew with great luxuriance, and as a result all samples were much more bushy than their wild parents had been. Sample E151A (Plate 10a) was dwarf, with short internodes, few branches, and flowering from a relatively low node. The leaves were very thick and readily developed anthocyanin; their green colour was pale and they were not shiny. The areas between the veins were flat on the upper surface of the leaves and very slightly concave beneath; the veins appeared on the upper surface as narrow grooves. In all these characters E151A was different from all the other samples, except E167A, which resembled it in its development of anthocyanin. Sample E166 (Plate 10b) was characterised in habit by its small leaves which left the branches more exposed than in other samples. The remaining samples were all very similar in habit, but El167A (Plate 10c) was distinguishable by its leaf-shape, which will be described later, and by the considerable development of anthocyanin in its leaves. The other three samples, E210, E211 and E215, could not be distinguished from one another in the garden by habit and foliage. It can be seen from Plate 10d, that the leaf surface was similar to that of E166 and E167A. The measurements of plant height, together with numerical data obtained for ten other characters (Table 2), have been subjected to an analysis of variance. This work was kindly carried out by Dr. D. A. Wilkins at the Scottish Plant Breeding Station. It was found that, in all eleven characters, differences existed among the populations signi- ficant at the 0°1% level of probability. Watsonia 5 (4), 1962. VARIATION IN EUPHRASIA paps) | For the character of plant height (character 1), Tables 2 and 3 show that EI51A was significantly shorter than all other samples, and that sample E167A was significantly shorter than the two tallest samples. Table 2 shows that E151A was in fact only about half the height of E210 and E211. Variation between populations in leaves The leaves of the populations are illustrated in Fig. 1 a-f. Differences between samples E166, E210, E211 and E215 are not very noticeable in the silhouettes, but EI151A and E167A are conspicuously distinct. E151A (Fig. 1a) has disproportionately few teeth for the size of its leaves and the teeth are necessarily relatively large (Table 2, character 5). The same applies to E167A (Fig. Ic) in lesser degree, but this sample has the longest teeth of all. The leaves of E167A are also more elongated and have the teeth directed more towards the apex than in other samples. Table 3 shows which of the measured or counted differences are Statistically significant. It is noteworthy that there are no significant foliar or habit differences between E210, E211 and E215, which were the samples that could not be distinguished by these characters when they were being grown. A comparison of the figures for the length and breadth of the leaves (Table 2) indicates a variation between the samples in leaf shape. TABLE 2. Averages of measurements made on Euphrasia plants (Measurements in cm for character 1, in mm for characters 2-11; greatest and least averages for each character are in bold type) Population E151A E166 E167A E210 |) E211 E215 Habit character | 1. Heightofplantafterpressing) 12-7 | 22:7 bss) 25-4 | 25:0 22°6 Foliar characters | | | 2. Length of ‘ bract 4’ | | (see p. 225) 9:40 Bae 10:2) | 13:2 12-2 11:9 10-7 3. Breadth of ‘ bract 4’ Wee inp Old | L169 130 | 12:0 11-7 4. Greatest no. of teeth on a | | | side of ‘ bract 4’ 0:53 .| 0:91 |) 0:73 110°) 27.097 | 1-02 5. Length of distal side of a | | tooth on the widest part | of ‘ bract 4’ 190 | 1-56 2:09 1:97 1-74 1-91 Floral characters | | | 6. Length of upper lip of | | corolla plus tube 731 |). 656.,-|- +.4-79 7:56 | 7:60 7-93 7. Length of lower lip of | | | corolla plus tube Sa19e 760 | 9-04 8-74 8°67 9-17 8. Length of mid-lobe of | | lower lip of corolla 2:90 | 2:13 | 2:76 2:70 2:67 2:71 9. Depth of emargination of | | mid-lobe of lower lip of | | corolla 059 060 081 0-79 | 0-90 0-99 10. Breadth of base of mid- | | lobe of lower lip of | | | corolla Ray 12) 13) 1:5 1:7 16s 4 1:6 11. Greatest breadth of mid- | | lobe of lower lip of | | corolla 2:27 1:97) | 2:37 2-71 2:73 3:07 Watsonia 5 (4), 1962. 228 PP eyYyEO TABLE 3. Characters in which pairs of Euvphrasia samples showed a statistically significant difference (P= 01%) (Characters numbered as in Table 2) Characters in which pairs of samples differed significantly Pairs of samples Foliar Floral E151A and E166 1 E151A and E167A | i E151A and E210 I | 1 S oN —_ — os Nn ft SN oI = — \O CO oS — — E151A and E211 E151A and E215 E166 and E167A E166 and E210 E166 and E211 E166 and E215 | 5 E167A and E210 | 1 4 E167A and E211 1 4 2 & » ) mE DNADDADOD INN NN — O- “ — E167A and E215 E210 and E211 E210 and E215 E211 and E215 —I © XO © — = ~ 7 . — —a Leaves of two untypical plants, plant | of EISIA (counting from left to right) and plant 2 of E167A, have been included in Fig. 1. These will be discussed later. The number of teeth on a side of * bract 4’ attained for each population the following maxima :— E1SIA 4 teeth (attained by about ? of plants) E166 7 teeth (attained by about 4 of plants) E167A 6 teeth (attained by about of plants) E210 9 teeth (attained by about + of plants) E211 8 teeth (attained by about 4 of plants) E215 8 teeth (attained by about 4 of plants) When the two untypical plants were omitted, each population produced its maximum tooth-number in each of the six blocks of the field trial. It seems, therefore, that there was a characteristic maximum tooth-number for each population, which was not attained by all individuals. It is interesting to note that E210 had a greater maximum tooth-number than E211 and E215, which it closely resembled (p. 226). Variation between populations in flowers This variation is also shown by Tables 2 and 3, and some flowers are illustrated in Fig. 2 a-f, p. 230. It is clear that samples E151A and E166 were the most distinctive in flower shape : EISIA by its fan-shaped lower lip with the lobes shallowly emarginate and weakly dilated in relation to their length, and E166 by its small size and narrowly emarginate lobes, the lobules on either side of the emargination tending to turn in. E210, E211 and E215, with no significant habit or foliar differences, were also found to differ in their flowers. The tables show that E210 has a less deeply emarginate mid-lobe than the other two, E215 has a longer lower lip and tube than E211, and E215 has a wider (more dilated) mid-lobe than either E210 or E211. Watsonia 5 (4), 1962. VARIATION IN EUPHRASIA 229 bad ¢ €@ # Fig. 1. Leaf silhouettes of Euphrasia, x 4. a-g,‘° bract4’;h andi, ‘ bract 6.’ a, E151A; b, E166; ewblo7jA} d. E2106" E211 E215: ¢, El6G7B2 h, El68s 1, E219: Variation within populations Dr. Wilkins calculated the standard deviations of each of the six populations for one foliar and one floral character. These are given in Table 4; they show that E210 and E211 were the most variable populations. The wild populations at the localities of E210 and E211 were numerically larger than the other four populations, and they were certainly sampled over a larger area than the others, except possibly E167A, which also formed an extensive population although it was not extremely abundant. Population E151A was very uniform. However, it included one untypical plant that stood out sharply from the rest. Compared with the other plants of this population, it will be seen from Fig. la that this plant had 5 teeth on a side of ‘ bract 4’ instead of 3 or 4. In addition the corolla lobes were broader, the leaf surface not so flat, the habit more branched, and the flowers more abundant. This untypical plant was evidently a hybrid between EI51A and E. pseudokerneri which grew at the same locality, and it seems likely that it was an F1 hybrid, in view of the effort made to exclude hybrids from the original collection of EISIA. Watsonia 5 (4), 1962. 230 Pr: Ee VEO Fig. 2. Corollas of Euphrasia, upper and lower lips separated. a, E166; b, E151A; c, E167A; d, E210; es, B21ils GE 21550 2B 219: ch E168: Watsonia 5 (4), 1962. VARIATION IN EUPHRASIA 231 TABLE 4. Standard deviations for Euphrasia populations (Unit of measurement = 1 mm) Standard deviation for | Standard deviation for Population | length of * bract 4° greatest width of mid-lobe of lower corolla-lip EISIA 0:97 2117) E166 1-83 18] E167A 1-62 374 E210 2°34 | 390 E211 1-96 | 501 E215 | 1-62 326 The other plant with an untypical leaf shape belonged to population E167A. The leaf of this plant was, however, similar in shape to those of the populations E166, E210, E211 and E215; it was the only ‘ bract 4” produced by E167A with more than six teeth on one side. Leaves of this shape also occurred in plants 1, 2, 3, 4 and possibly 6 of E167B (Fig. 1g), counting from left to right). The leaves of the only other two survivors out of the original ten plants of E167B were like those of E167A , though the leaf of plant 5 is perhaps intermediate, as were some of the leaves of E167A. The occurrence of intermediate leaf- shapes suggests that the two forms were hybridising. Comparison of wild and cultivated plants The wild plants of EISI1A were extremely dwarf and compact, but freely branched. The leaves were few-toothed, and fleshy-looking but shiny. The plants were seeding freely when collected on 24 July 1952, which is an early date for E. nemorosa to be in such an advanced condition. E166 was taller and had longer internodes than E151A, but was dwarfer and had smaller leaves and flowers than the other populations. E167A had a distinctive leaf-shape, but was otherwise rather similar in habit to the remaining three populations. The latter were all collected on the same day, and appeared to differ slightly from one another, E211 having stout, dense flowering spikes, and E215 particularly long internodes, compared with E210. Population E151A, which was the most distinct in the field, was also the most distinct in cultivation. Its few leaf-teeth, dwarf habit, and early flowering were shown to be heredi- tary, although the internodes were not quite as short in cultivation as in nature. The leaves became even more fleshy in cultivation, and developed a flat, non-shiny surface. In cultivation, population E166 was little different from E210, E211 and E215, except in its small leaves and flowers; it looked about as distinct as it did in the wild. EI167A in the main retained its characteristic leaf-shape in cultivation. The slight differences which E210, E211 and E215 showed in the wild disappeared in cultivation, where they could not easily be distinguished by eye. E215 was in fact the shortest of these three in cultivation, in spite of having had the longest internodes in the wild. On the other hand small differences between these samples were detected statistically. CULTIVATION OF FURTHER SAMPLES IN 1954 In 1954 an attempt was made to repeat the 1953 experiment on a smaller scale, using another five Euphrasia populations. The plants were grown at Cambridge by the method used at Leicester, but conditions were apparently less favourable, and very few plants survived. However, a fairly good idea of the characters of four of the populations was obtained. These were believed to be E. nemorosa and were as follows : Watsonia 5 (4), 1962. 232 Po. PY EO E417 between Warslow and Elkstones, Staffs., v.c. 39; long grass and stony patches by roadside on acid moorland but accompanied by some more or less calcicolous plants; E421 Friday Street, Surrey, v.c. 17; sandy field; E429 West Harling Heath, W. Norfolk, v.c. 28; calcareous sandy soil; E430 Devil’s Dyke, Cambs., v.c. 29; on part of chalk dyke levelled in 1943. In general appearance the cultivated plants of three populations (E417, E421 and E430), although distinguishable by eye, were much like E210, E211 and E215 grown in 1953. E417 showed the closest resemblance to the 1953 samples just mentioned, being very vigorous and rather large-leaved. E430 and E421 were rather smaller plants with somethat narrower leaves, E421 differing from E430 in that its branches diverged for a greater proportion of their length and became vertical only at the tips. The fourth popula- tion, E429, had smaller leaves and thinner stems than the others, and it was not so tall as E417. Most of the E429 plants were quite luxuriant, however, for they produced a great profusion of branches; these were usually widely spreading and somewhat flexuous. The flowers were largest in E417 and smallest in E430, and the flowers of E429 differed from those of the other samples in being lilac instead of white. The flower-shape was different in each sample. Wild populations of the 1954 culture The differences between the 1954 samples in cultivation were mostly similar to differences seen in the wild plants from the same localities. The distinctive features of the cultivated E429 were all evident in the wild except for the low spreading habit. COMPARISON OF SAMPLES OF E. ANGLICA GROWN IN THE GARDEN Two samples of EF. anglica were grown in the garden at Leicester in 1953. There were ten ‘ whalehide’ pots for each sample, and each pot was planted with two Euphrasia seedlings and a plant of Luzula campestris. The pots of each sample were arranged in a row, the two rows being side by side. The seed of the two samples (E168 and E219) was collected from two localities three-quarters of a mile apart in Charnwood Forest, Leicestershire. The two populations differed in habit in cultivation, E168 having shorter internodes than E219. The average height after pressing of the 13 surviving plants of E168 was 72cm (range : 3°3-9:3 cm), and that of the 16 survivors of E219 was 11-1 cm (range: 44-16°5 cm). ‘ Bract 6,’ defined in the same way as ‘bract 4’ (p. 225), was mounted for each plant (Fig. th, i). In E168 the bracts did not attain so large a size as in E219, and they also had fewer teeth, the greatest number of teeth on a side of ‘ bract 6’ averaging 5-0 in E168 (range: 4-6) and 6-2 in E219 (range: 5-7). The flowers of E168 were smaller, but more uniform in size, than those of E219, and they also differed from them in shape (Fig. 2g, h, p. 230). Some measurements were made on the flowers. The depth of emargination of the mid-lobe of the lower lip averaged 0-39 mm in Ei68 and 0-67 mm in E219. In E168 the width of the mid-lobe averaged 1-29 mm at the base and 1-63 mm at the widest point, while in E219 it averaged 1-49 mm at the base and 2-23 mm at the widest point. TAXONOMY The three most distinct forms of supposed EF. nemorosa that were cultivated were E151A, E167A and E429. In order to look for further populations of the type of E1S1A I spent two days in 1954 at Box Hill and Mickleham Downs. The area from which EI51A had been collected was revisited and the same form collected again. Several other gatherings of E. nemorosa from both grassy fields and dry chalky slopes were made, but none was at all like E151A. They appeared to be fairly normal forms of E. nemorosa, although Watsonia 5 (4), 1962. PLATE 10 © d Euphrasia plants growing in the garden; the plants were luxuriant and each photograph shows only a portion of a plant. a, E151A; b, E166; c, El167A; d, E211. VARIATION IN EUPHRASIA p28) those from the poorest habitats were the most dwarf. The form represented by EI151A appears to deserve taxonomic recognition on morphological grounds, but it does not seem to be worth while to accord it such recognition when only one colony is known. In view of its similarity to E. nemorosa in leaf-outline and flower-size, it seems best to regard it as an extreme ecotype of that species. Regarding the plants from Medmenham (E167A and B), I consider that two species were present, E. stricta Lehm., represented by nearly all the cultivated plants of E167A, and EF. nemorosa, represented by one cultivated plant of E167A and four or five of E167B. Other samples of £. stricta, obtained from foreign sources, that I cultivated, had long, very acute leaf-teeth, similar to those of E167A, and they were usually rather few in number. The upper cauline or lower floral leaves were usually rather narrow, and some had a rather rounded base, but others were truncate. Usually E. stricta has rather large lilac flowers, and few short erect branches. These two characters were lacking in E167A. However, they are not always present in E. stricta, and Professor W. Rothmaler, to whom I sent plants of the same type as E167A, collected in 1954 from the same locality, considered that they were E. stricta forma parviflora Sag. After the publication of Wettstein’s (1896) monograph, E. stricta was widely reported from Britain. In Pugsley’s (1930) revision, however, E. stricta was not accepted as occurring in Britain; plants previously identified as E. stricta were referred by him to EF. nemorosa, E. confusa and E. pseudokerneri. | find these identifications acceptable, but I have encountered a few populations of E. nemorosa in which there were resemblances to E. stricta, and it may be that £. stricta was formerly in Britain, or has been introduced from time to time, but has in most places been unable to maintain itself as a distinct entity. Probably £. stricta was introduced at Medmenham, possibly at the same time as another alien that was growing with it, namely Prunella laciniata. A description of E. stricta follows. EUPHRASIA STRICTA, Wolf ex J. F. Lehmann, Primae linae Florae Herbipolensis, 43 (1809), emend. Host, Flora Austriaca, 2, 185 (1831). Stem erect, usually with 0-4 pairs of erect branches, cauline and few to many of the lower floral internodes 14-3 times as long as the leaves, upper floral shorter than the leaves; flowering commencing at the 7th to 12th node. Leaves glabrous or nearly so, usually strongly pigmented with anthocyanin; cauline ovate or narrowly ovate, obtuse or acute, up to c. 14mm long, with up to 5 pairs of obtuse, acute or aristate teeth; floral leaves relatively broader, ovate or rhombic, acute, more or less rounded or cuneate at base, with 4-6 acute or aristate teeth, teeth of upper cauline and floral leaves directed towards the apex of the leaves. Calyx 45-6 mm long, with long slender aristate teeth. Corolla white or lilac with the usual markings, (5-5—) 8-11 mm long, measured from base of tube to apex of upper lip. Capsule rounded, truncate or retuse at apex, 4-6 mm long, almost always distinctly shorter than the calyx-teeth. E. stricta is widespread in Europe as a plant of dry grassland. It is best distinguished from E. nemorosa and E. pseudokerneri, its nearest allies in Britain, by its habit and foliage. The National Grid Reference of the locality at Medmenham, Bucks., v. c. 24, is 41/8186. The plant was collected from a chalky pasture and a rough field by a wood on 12 August 1952 (Yeo, no. E167A, E169A) and i9 August 1953 (Yeo, no. E401A, E402A). The small leaves, short internodes, ascending main stem, and abundant, spreading and often flexuous, branches of E429 are characters of EF. confusa. Wild plants from the Breckland locality of E429 differed from the cultivated E429 in being more erect and nemorosa-like in habit. Excursions to the Breckland, made since 1954, have shown that plants which are identical with E. confusa from the West of England occur there; in addition, there are many populations which, although similar to E. confusa, show some resemblance to £. nemorosa. Most of these populations are probably best referred to E. confusa, a species which was first identified as occurring in the Breckland by Pugsley in 1939 (specimen in CGE). Watsonia 5 (4), 1962. 234 P. F. YEO The experiments described here have emphasized the differences between these three distinctive forms. The hereditary character of the distinctive habit of E151A has been confirmed and a number of lesser differences which together have a considerable effect on the appearance of the cultivated plants have been brought out; the distinctive leaf- shape of E167A was easier to appreciate in the luxuriant cultivated plants; the small leaf-size of E429 was still evident in the luxuriant plants, while habit differences from £. nemorosa were increased. The remaining samples of E. nemorosa were all slightly different from each other in cultiva- tion but the differences were not always the same as the ones they showed in the wild; the experiment thus revealed examples of phenotypic and genotypic variation between populations of generally similar plants. This general similarity was evident among samples collected over a fairly wide geographical area. The most distinct population among these was the small-leaved, small-flowered E166 from Watlington Hill. Its characters are pro- bably a response to the dryness of the habitat, and the form can thus be regarded as an ecotype of FE. nemorosa, though a much less extreme one than EISIA. The differences shown by the two populations of E. anglica that were grown are quite considerable in view of their separation in the wild by only three-quarters of a mile. How- ever, £. nemorosa was growing with E219, but not with E168. CONCLUSIONS Cultivation showed that all samples compared differed genetically. This indicates some degree of reproductive isolation, which in most cases must be due to the spatial separation of colonies. Where the differences are slight they can tentatively be attributed to chance; where they are rather larger and have been classed as ecotypic they can be attributed to response through selection to special habitat conditions. The differences, of a similar magnitude to the ecotypic ones, between the two populations of the diploid species FE. anglica which were compared may well be due to introgressive hybridisation, for the tetraploid FE. nemorosa was growing with E219, but not with E168. There is evidence (Yeo 1956) that introgression takes place from tetraploid species into diploid species of Euphrasia, and it is, therefore, possible that this had occurred at the locality of E219. Another population of E. anglica, growing some miles away from E168 and also unmixed with any other species, was cultivated in the greenhouse with additional plants of E168 and, although found to differ from it in flower-shape, leaf-shape and habit, the differences were never greater, and were mostly smaller, than those between E168 and E219. The differences between sample E167A and the samples of EL. nemorosa compared with it are considered to indicate that E167A is E. stricta, while the differences between E429 and samples referred to E. nemorosa are considered to indicate that E429 is E. confusa. There were indications that hybridisation was taking place between E. nemorosa and E. stricta, two rather similar species with the same chromosome number which would be expected to hybridize freely, judging by the behaviour of other Euphrasia species. The observations on variation within E. nemorosa can do no more than provide a useful initial contribution to the material required for the evaluation of the varieties of this species. Cultivation was considered necessary in order to eliminate the effects of phenotypic variation in vigour before making morphological comparisons. However, the results of comparing the very well-grown plants of rather uniform vigour suggest that at least leaf- shape and flower-shape can be reasonably satisfactorily compared in wild material on the basis of selections of the more vigorous individuals. ACKNOWLEDGMENTS I wish gratefully to acknowledge help from Professor T. G. Tutin, under whose supervision most of this work was carried out, and who has suggested improvements to Watsonia 5 (4), 1962. VARIATION IN EUPHRASIA 239 the manuscript of this paper, from the University of Leicester, which provided a Research Scholarship and travelling allowances, from Dr. E. F. Warburg, who showed me the Euphrasia populations at Medmenham and Watlington, from Dr. D. A. Wilkins, who did the statistical work, and from the late Professor W. Rothmaler, who helped with the taxonomy and nomenclature of E. stricta. REFERENCES Pucstey, H. W. (1930). A Revision of the British Euphrasiae. J. Linn. Soc. Lond., Bot., 48, 467-544. WARBURG, E. F. (1952), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Cambridge. WETTSTEIN, R. VON (1896). Monographie der Gattung Euphrasia. Leipzig. YEO, P. F. (1955). The Species Concept in Euphrasia. In J. E. Lousley (ed.), Species Studies in the British Flora. Arbroath. Yeo, P. F. (1956). Hybridisation Between Diploid and Tetraploid Species of Euphrasia. Watsonia, 3, 253-269. Watsonia 5 (4), 1962. THE BRITISH FORMS OF TUBERARIA GUITTATA (L.) FOURREAU By M. C. F. PRocTorR Department of Botany, University of Exeter ABSTRACT An account is given of the variation in British Tuberaria guttata, and of its previous taxonomic treat- ment : the taxonomic value of various characters is examined. Most of the Welsh and Irish plants includ- ing the type population of Helianthemum breweri Planch. differ from 7. guttata as it occurs in the Channel Isles and northern France in their shorter stature, the more common presence of bracts and other characters. All the characters said to distinguish breweri intergrade continuously with those of typical guttata in both herbarium and cultivated material, and are only loosely correlated. The more compact Welsh and Irish plants appear to be comparable with plants in similar exposed coastal habitats in north-west France. It is concluded that 7. guttata shows ecotypic differentiation in relation to exposure on the Atlantic coast of Europe, and that the populations combining short diffuse habit and numerous bracts may be of polytopic origin. It is suggested that they should not be given formal taxonomic recognition. 1. INTRODUCTION Like a number of other widespread Mediterranean species, Tuberaria guttata extends northwards up the west coast of Europe to a northern limit in the British Isles. Up to the north coast of France its distribution is more or less continuous, and the Channel Islands lie on the northern fringe of this essentially continuous area. But north of the English Channel its range is disjunct, and it occurs only in widely separated colonies on the coasts of north Wales and western Ireland. Plants from the best known of these colonies, on Holyhead Mountain in Anglesey, were described by Planchon in 1844 as a new species, Helianthemum breweri, and, as a variety or subspecies, breweri has become firmly established in the British literature. Authors have differed considerably in the value they have attached to the various characters said to separate it from typical 7. guttata, and in the British populations that they have referred to breweri and to guttata. Most seem to have assumed that it was deserving of taxonomic recognition, and that all British plants could be assigned satisfactorily either to breweri or to guttata. The purpose of this paper is to show that none of the delimitations of breweri proposed hitherto is completely satisfactory, and that the variation pattern in 7. guttata in Britain and neighbouring parts of the Continent is more complicated, and taxonomically intractable, than has usually been assumed. 2. THE HISTORY OF HELIANTHEMUM BREWERI PLANCH. The Holyhead colony of Tuberaria guttata was discovered by Samuel Brewer in 1726 (see Hyde 1930). Dillenius wrote to Brewer on 31 May 1727 ‘I desire him (i.e. the Rev. Mr. Green) to look after the plant you sent from Holyhead .... It is a Cistus and seems to be new.’ Brewer received further specimens of the plant from Mr. Green on 5 August 1727. Daillenius had evidently intended to describe the plant himself, but, in the event, it was first mentioned in print over half a century later by Hudson (1778), as ‘ Cistus guttatus .... Habitat in pratis arenosis, in monte Llech ddt prope Holyhead in insula Mona.’ Curtis (1775-98) figured under the same name a bracteate and spotted-petalled plant, but did not say whether it was wild or cultivated, and gave the Isle of Man as a locality, evidently mistranslating ‘Insula Mona.’ In this he was followed by Sowerby & Smith (1790-1814, t. 544), who figure a rather stiffly formal little plant with numerous bracts and say that ‘ Having no hopes of obtaining wild specimens in a state fit for drawing, we have been 236 Watsonia 5 (4), 1962. BRITISH FORMS OF TUBERARIA GUTTATA yp obliged, like Mr. Curtis, to have recourse to a garden one, communicated by the Rev. Mr. Watts, F.L.S. of Ashill, Norfolk ...’° Smith (1800) gave the localities ‘In Jersey, Sherard. In the Isle of Man, Mr. Brewer. Huds.’ and his description notes *. . . saepius ebracteati, interdum bracteis lanceolatis, solitariis, ad basim pedicellorum.’ Later (1825), quoting the Rev. H. Davies, he correctly gave the locality as Anglesey. Supposedly Manx plants were distributed by Dickson (Hortus Siccus Britannicus), 1793-9, but the specimens are tall slender plants quite unlike any other British examples of T. guttata. As far as I know, there is no other evidence that they actually came from the Isle of Man, and they are probably garden specimens to which the name of the (supposed) best-known British locality was attached. Planchon knew T. guttata well in the south of France, and saw specimens of the Holy- head plant in the herbarium of W. Wilson of Warrington. He says (Planchon 1844) ‘ the straggling mode of growth, with short, numerous and dichotomous bunches of flowers, the existence of bracteas even to the upper flowers, and of pedicels, which at first bend down, but rise up when the fruit is ripe and stand at an angle of forty-five degrees to the axis of the cluster, such were the characters, which at first glance, suggested the idea that the plant was specifically distinct from H. guttatum.’ Planchon knew that cultivation made no difference to “the main point,’ the presence of bracts. He gives a terse diagnosis, followed by a longer Latin description: HELIANTHEMUM BREWERI, Planchon Helianthemum (e sectione Tuberaria); annuum a basi ramosum subdiffusum viscidulo-hispidulum, pedicellis bracteatis, defloratis subdeflexis, fructiferis erecto-patentibus, petalis angustis immaculatis, staminibus 8-12, seminibus quam in H. guttato majoribus. Citus guttatus, Huds. Fl. Angl., p. 232, ex local. citat. Cistus guttatus, Smith Fl. Brit. t. 2, p. 573, (pro parte et quoad plant. Brewerian, non Engl. Bot.). Has. Anglesey, Mr. Brewer, (Hudson), Holyhead Mountain, Rev. Mr. Williams, and also near Amlwch, Rev. H. Davies, W. Wilson, Esq. Herba a basi ramosa, vel, axi abbreviato pluricaulis, viscidulo-hispida. Caules 4-6, subdiffusi, semel bis-trichotomi, in racemos laxos, bracteatos desinuentes. Folia pauca, opposita, sessilia ,inferiora 4—6, rosulata, late obovata, semipollicaria, obtusissima; caulina stipulata, lineari oblonga, obtusiuscula, omnia pilis stellatis simplicibus intermixtis, utrinque hispidula. Bracteae conspicuae, lineares, non raro falcatae, inferiores 3-4 lin. longae, hinc ad summos flores gradatim minores, nunquam deficientes. Pedicelli graciles, 6-8 lin. longi, floriferi subdefiexi, deflorati diu sic persistentes, tandem, maturo semine, erecto- patentes. Calyx ut in H. guttato. Sepala conspicue nigropunctata; trium majorum alterum latus in prae- floratione tectum, petaloidea-membranaceum. Petala in flore unico observata, angusta, lutea, immaculata. Stamina definita (8-12) pistillo vix longiora. Ovarium et fructus vix H. guttati, sed semina pauciora et evidenter majara. Planchon emphasised the short diffuse habit of growth and the presence of bracts as the most important characters of his new species. He had seen an imperfect specimen of the Jersey plant, and considered it ‘ merely a hispid variety of H. guttatum, The Annals and Magazine of Natural History for the same year contains an account of a meeting of the Botanical Society of London, at which W. R. Crotch exhibited specimens of Planchon’s new species, and a note by C. C. B[abington] reporting ‘true H. guttatum,’ collected at Three Castles Head by Miss H. Townsend. H. breweri appeared in the second edition of Babington’s Manual of British Botany (1847) as a separate species, the presence of bracts being given as a diagnostic character. Bentham (1858) says rather non-committally*... the Anglesea specimens are rather stunted, with the leaves broader than usual, and have been published as a species under the name of H. Breweri. Hooker and Arnott (1860) give breweri as a variety under H. guttatum. In the seventh edition of the Manual (1874), Babington reduced breweri to a variety of guttatum, and altered the description to ‘ Racemes with or without bracts on the same plant.’ Syme (1864) and Hooker (1870) rank the plant as a subspecies of H. guttatum. Inishbofin, W. Galway, was added to the list of localities of var. breweri Watsonia 5 (4), 1962. 238 Mi Co Ey PROCTOR by A. G. More (1876), and appeared in the ninth edition of Babington’s Manual (1884); according to Praeger (1901) ‘ The var. Breweri accompanies the type at Three Castles Head and Inishbofin, and alone grows on Inishark.’ (see below for detailed list of localities). Planchon’s conclusions seem to have been generally accepted by British botanists, who contrasted breweri with the Channel Islands plant : indeed tall ebracteate plants from the /ocus classicus of breweri were named guttata by J. G. Baker (Druce 1890, 1891). But Griffith (1895) says ‘I have sown seeds of the supposed H. guttatum and seeds of undoubted H. Breweri from the same place in rich soil in my garden, with the result that they all turned out to be the same - viz. H. Breweri.’ In this conclusion he is strongly supported by Druce (1902) who considered that *... the characters given to H. breweri both by its original describer and subsequent writers will need amending; as those of bracteate inflorescence, diffuse growth, and unspotted petals given by Planchon are found to be inconstant.... Still, the Anglesey plant has a different facies from the H. guttatum of Jersey, the leaves being broader and more obovate, and of a more coriaceous texture. . and I think it to be a good geographical race.’ Druce had earlier (1893) named the ebracteate plants from Holyhead * f. ebracteata.’ Grosser (1903), regarding T. guttata on a continental scale, united populations on the west coast of Europe from Wales and Ireland to Portugal in var. breweri. The main dia- gnostic characters of his variety are low, diffuse habit, and the pubescence of the leaves, which are described as hispid and densely stellate-tomentose on both surfaces : he makes no mention of bracts. The distribution of the variety is given as ‘ Strandgebiete von West-England und Nord-West-Frankreich entlang der westfranzésischen Kuste bis in das maritime Nord- und Mittel-Portugal (Sierra de Cintra).’ He cites Welwitsch It. Lusit. (1840) nos. 433, 1317, 1526 and (1851) no. 49, and includes H. guttatum var. maritimum Lloyd and H. litorale Rouy and Fouc. as synonyms. As he included England among the localities for his var. genuina, it seems that he accepted the opinion of British botanists that at least the Channel Islands material belonged to typical T. guttata. Warburg (1952) accepts Grosser’s statement of the Continental distribution of 7. guttata var. breweri, and his description and account of its British distribution follow accepted British opinion. In my account of 7. guttata for the Biological Flora (Proctor 1960), I suggested that the name breweri should be limited to the Welsh and Irish populations; though recognising that the taxon so defined was not completely satisfactory. At my sugges- tion, Dr. Warburg adopted this delimitation of subsp. breweri in the second edition of the Flora (1962). 3. THE VARIATION IN BRITISH T. GUTTATA (a) Distribution I believe the following list of the British localities of T. guttata is substantially complete: I have cited specimens and literature records only in the case of less well-known localities, or where the record is otherwise interesting. V.c. S. CHANNEL ISLANDS. Jersey: West Mount, St. Helier; Noirmont; St. Brelade’s Bay; Beauport; La Moye; Portelet; Corbiére; Petit Port; headland north of Petit Port; L’Etac (Lester-Garland 1903); above Pinnacle Rock, between L’Etacq and Grosnez (F. Le Sueur, March 1960); ‘ observed in Jersey by Mr. Sherard on the west-side of the Island near Grosnez Castle.’ (Ray 1690 p. 238); Grosnez (Lester-Garland 1903). Alderney: south coast, on cliff top west of Val du Sud. (Except where specific references are given, specimens from all localities are in either K, BM or CGE). 49. CAERNS. Mainland opposite Bardsey I. (1938, W. Hughes D’Aeth), 1939 (and det.) Miss B. M. Morgan (Rept. Bot. Soc. Exch. Cl. for 1943-4, 705 (1946)). 52. ANGLESEY. Holyhead Mountain, S. Brewer, 1726, and many more recent collectors; near Porthdafarch, C. C. Townsend, 8 July 1952 (herb. Townsend); ‘Lady Verney tells me she has found this [H. breweri] in another locality about three miles from the Jocus classicus in Anglesey.’ (Druce 1919) (Perhaps the same locality as the last); between Clegyr Mawr and the Watsonia 5 (4), 1962. BRITISH FORMS OF TUBERARIA GUTTATA 239 sea, Llanrhyddlad (Griffith 1895, as H. breweri); Gader [Trwyn y Gader = Carmel Head] in the parish of Llanvair in Cornwey [Llanfairynghornwy] at the north end of the Island facing the Skerries about half a mile from the sea upon the rocks where the soil is not deep, among the Scilla bifolia and Sedum rubens: it flowers in June. Mr. Williams [undated] (BM);..... Amlwch, i.e. N.W. of the Village among Heath, before the discovery of the great copper mine, plentifully, H. Davies [undated] (BM). H3. W. CORK. Three Castles Head (Babington, 1844; Moore & More, 1866, as H. guttatum); ibid., type and var. breweri (Colgan & Scully 1898; Praeger 1901); Calf Islands (Praeger 1934). H16. W. GALWAY. Inishbofin (type and var. breweri), Inishark (var. breweri) (Colgan & Scully 1898; Praeger 1901). H27. W. MAYO. Inishturk (Praeger 1934). (6) Cultivation experiments My own cultivation experiments were begun in 1953, the most extensive being carried out in 1956, 1959 and 1961 (see Figs. 3-6). It was found most convenient to grow plants singly in four-inch pots (1956) or in rows in boxes (1959, 1961). The seeds are hard-coated, and it often proved difficult to germinate enough seeds from all the populations at nearly enough the same time for the seedlings to be comparable. Plants grown from wild-collected seedlings were several weeks more forward than those grown from seed germinated in the garden, and all plants took much longer to flower and fruit than in the wild. Indeed in Cambridge it was often difficult to induce the Three Castles Head plants to flower at all. T. guttata is a plastic species, and the different conditions of different years produced very different-looking plants, so that comparisons cannot be made between populations grown in different seasons. However, the relative differences between populations in any one season were not greatly affected. From comparison of several years’ results I am satis- fied that the main conclusions set out below are generally true, though further experiments would quite probably produce results differing in detail from mine. In spite of some diff- erence in the age of the seedlings, plants from any one population were generally rather uniform in size and other characters, and I found no evidence for the coexistence of two taxa in any British population. In a number of cases seed was saved from cultivated plants for later experiments. This was made easier by the fact that the plants grown were apparently all normally self- pollinated: as additional precautions, plants kept for saving seed were isolated as far as possible, or seed was used which had set when other populations were not in flower. (c) General observations on the taxonomic characters Of the characters emphasised by Planchon, the ‘narrow unspotted petals’ were based on observation of only a single, aberrant, flower. Seed size has received little atten- tion. Seed lengths for a number of samples of 7. guttata are given in Table 1. There is an evident tendency for the northern and western populations to have somewhat larger seeds, but there is too much local variation, both in north and south Europe, for this to be a useful diagnostic character. The presence of bracts was recognised to be inconstant by Babington and Druce, and the size and number of bracts produced depends at least in part on environmental conditions. Thus on the one hand it is possible to find ebracteate individuals in popula- tions of generally bracteate ‘ breweri,’ while on the other hand plants of 7. guttata from Jersey grown in Exeter produced numerous bracts. The bracts are of two kinds. Broad leafy bracts are sometimes seen in the lower part of the inflorescence, and it is not always easy to differentiate between these and the uppermost leaves. Higher up the inflorescence the bracts resemble the narrow outer sepals in shape and texture. These small bracts may become almost constant in the upper parts of inflorescences which are wholly ebracteate below (e.g. from about the fifth flower in La Moye plants, and the third in Holyhead Watsonia 5 (4), 1962. 240 Mi: °C. PAEROCTOR TABLE 1. Seed length in some populations of Tuberaria guttata. | Number of seeds Mean length | Standard Provenance | measured (mm) deviation Messina, Sicily 5 584 "042 Monte Nuovo, Naples 6 534 "045 Appenines between Florence and Bologna 5 699 ‘066 Sintra (Estagao Agronomica Nacional)* 10 59 "044 Sintra (I. A. Williams)* 10 -688 ‘060 Welwitsch It. Lusit. 433 5 501 037 Toulon 5 613 054 Montpellier 4 613 046 Lyon, a la Pape 5 598 "042 Mouilleron-en-Pareds, Vendée 5 662 053 Croisic, Loire-inf. 5 -560 | 035 Baie des Trépassés, Finistere* | 10 636 030 Corbiere, Jersey* | 10 621 038 L’Etacq, Jersey* 10 | “608 041 Ffynnon Fair, Caerns.* | 10 | -739 | ‘070 Holyhead, Anglesey* | 10 | -702 | -073 Clegyr Mawr, Anglesey* | 10 “675 048 Inishbofin (Middle Quarter)* | 10 | ‘703 045 fe a For further details of populations marked with an asterisk, see appendix. The other measurements are from herbarium specimens in CGE. Measurements of herbarium specimens of British 7. guttata. This table is based mainly on material in BM, TABLE 2. | with some Welsh and Irish material from CGE, EXR, K and herb. C. C. Townsend. | *These are the numbers of specimens whose height was measured : in several cases it was not possible to score all of these for bract number and stellate pubescence. Several of the small sets of data are based on one or a few gatherings, and in these cases the figures given are not suitable for a significance test of the difference in height between populations, since season-to-season differences contribute variation which may not be adequately reflected in the standard deviations. | Height (cm) | % with | SSS | % pedicels stellate | Specimens | | | % diffuse with pubescence | Provenance _examined* mean | s.d. | habit bracts on upper | | | | | | leaf | | | | surface | Channel Isles | | | | | Jersey (S. Coast) \tpi2dev yer ALOAT rele: 4:64 25 slit SOP ae 9 100 Jersey (W. Coast) AQ OO AOS A alae 3e 57 Bye 3 96 Alderney | 2 15-0.) Be sulplinebeeedataal 0 0 100 | Wales | | | | | Opposite Bardsey Island 5 | 32 0-75 | 0 — 100 | Holyhead nes 59 sees ltaian3-49 | aul 26 [eee 60 94 | Porthdafarch | 9 O22 | 2:35 | 0 | 62 89 | * Gader ’ | 12 MES Inhalt e y223G0e aul 8 | 21 100 | Clegyr Mawr | Sah tialt QekpeD. | 2-49 O01 4° Aeo0 100 Ireland | | | | d Three Casiles Head | I our O8)a nn en2-04 Feats 31 | 19 23 | Inishbofin 14 | Sei SOnr RAE CRLGT Dae! Ba 24 | 29 | 79 : | | Watsonia 5 (4), 1962. BRITISH FORMS OF TUBERARIA GUTTATA 241 plants, among material cultivated in 1961), so there is a general tendency for the proportion of bracts to flowers to rise as the season advances (see Table 3). There is a negative correla- tion, but not a close one, between plant height and bract number (r = — -24 in cultivated material in September 1961 (not statistically significant, P > 0-3): if the La Moye plants are ignored as possibly aberrant, r = — -75; P 0-05). TABLE 3. Height and bract frequency of plants of 7. guttata cultivated in 1961. The upper measurement in each row was made on 8 Aug. 1961, the lower on 8 Sept. 1961. | | Height (cm) | Population | Number of plants ery a. | % pedicels with measured | mean | s.d. bracts Brittany | | Baie des Trépassés 4 | qs | 2:0 | 0 | 3 | 12-3 | 0-47 0 | Channel Islands | La Moye | 2 | 15-0 1:0 | 743) D, 16°5 15 47 Corbiére | 9 | 106 Wed 22 7 8 | iat | 0-93 13 L’Etacq | 9 | 41 37 wt 8 | 10-6 0-70 36 Wales | | Holyhead 6 | 3°8 | 0:69 = z | 9-0 | 1:0 63 Clegyr Mawr 8 | 5:9 | 3c | 0 1 | 9-7 | i167) 13 (Clegyr Mawr*) 7 | 14-9 | 0:84 19 Ffynnon Fair 6 | 5-2 | 0-69 ee 5 | 10-0 1:6 32 Ireland | | Inishbofin 10 | 3-0 | = == | 6 | 6:3 | 0-94 53 Druce considered relative leaf-width a more useful and more constant character of Holyhead breweri than any of the foregoing. However, if the logarithm of leaf-width is plotted against the logarithm of leaf-length, the points for all populations fall near to a straight line. This suggests that the relation of leaf-breadth to leaf-length is allometric, and that the essential difference between the ‘ breweri’ populations and typical guttata is the obvious one of size. Environmental factors evidently affect the relation between leaf- breadth and -length, and the leaves of plants in cultivation are regularly wider than those of plants in wild habitats. There remain the characters of stature and habit. The size of the plant depends greatly on environment, so that although stature differences may be striking in the field, cultivation experiments are necessary to assess their significance. In cultivation, different *Wild-collected seedlings, measured 8 Aug. 1961. Watsonia 5 (4), 1962. 242 M. C. F. PROCTOR 15 + + e+ rae a 10 + Oo+ os die a> 21 cis ane S e ts 2 0 00 0@ eet oy + ‘“ @@ e = eo ee + + fa] a & 5 e ab O ® € S @® +O OO = os 25 10 20 30 40 50 60 70 Leaf-length (mm) Fig. 1. Log leaf-breadth plotted against log leaf-length for lower cauline leaves of 68 herbarium specimens of T. guttata. Crosses, Channel Isles material; solid dots, Wales; open circles, Ireland. 40 i + 30 Gy ate oo + ‘ =f Q 20 4% =) “et? E sgt. . = oD + eI ++ 2 CORT tt = 10 e+ e+e a OL (D) me] + +0 ++ G + ae eA 10 20 BOM 4DuSo 00 Leaf-length (mm) Fig. 2. Log leaf-breadth plotted against log leaf-length for basal rosettes of T. guttata in cultivation. The graph is based on two measurements, taken on 8 and 20 June 1961, of the longest rosette-leaf of each of 62 plants. The measurements are taken from the same material as Figs. 5 and 6. Symbols as in Fig. 1. populations matured at different rates, southern plants in general developing more rapidly and flowering earlier than those from further north. Hence it was possible for plants to be similar in size and habit in the rosette stage and at maturity, but strikingly different at any one moment in the intervening period (see figs. 3-5). These differences in rate of maturation are obviously important taxonomically and physiologically : so far as possible I have avoided their incidental effects on other characters by making comparative Watsonia 5 (4), 1962. BRITISH FORMS OF TUBERARIA GUTTATA 243 measurements on mature plants. Diffuse habit of growth is not as useful a character as it appears at first sight. Simple individuals can be found in all populations of 7. guttata, especially in dry seasons. All the populations I have grown become more diffuse in cultiva- tion, and, given sufficient water, more diffuse as the season progresses, and the differences observed in the field are probably very largely a direct reflection of habitat conditions, especially water-supply and competition. The inflorescence does not appear to me to provide any clear characters apart from length which might separate breweri from typical guttata: the pedicels in all the British populations appear to behave in the same way as Planchon describes for H. breweri. It is interesting, as Druce (1919) points out, that cultivation tends to accentuate several of the features usually considered characteristic of breweri; diffuse habit, broad leaves, and the occurrence of bracts. Without cultivated plants of typical 7. guttata for comparison this undoubtedly presented Planchon and Griffith with an even more striking picture of the difference between Holyhead breweri and Channel Islands 7. guttata than that which really exists. (d) Comparison of British populations The Channel Islands plants do not differ in any taxonomically significant way from the plants of neighbouring parts of the French coast, and they are the standard by which other populations have generally been judged by British botanists. Tall sparingly branched plants from Jersey can be matched closely in habit, leaf shape, and inflorescence with plants from as far afield as Spain and Calabria; but taken as a whole the Channel Islands plants are shorter and more branched than south European populations. Typically they range from about 5—15 cm in height, with 3 or 4 pairs of opposite cauline leaves above the basal rosette. Even vigorous individuals are often sparingly branched; but plants showing diffuse branching from the base are also common. The leaves are pilose or hispid on both surfaces, and stellate-hairy below; and almost invariably show a greater or lesser amount of stellate tomentum above. They are usually said to be ebracteate, but bracts occasionally occur in wild plants, often several on one individual (see Table 2). Their scarcity in wild material probably reflects the fact that flowering is over much earlier in Jersey than in the more northern and western localities. There are usually one or more alternate, lanceolate, stipulate leaves between the opposite-decussate lower cauline leaves and the first flowers. In cultivation, plants from Jersey show substantial variation in size and in the number of bracts produced. Plants from the Jocus classicus of breweri at Holyhead are shorter (usually less than 10 cm), and vigorous examples are normally diffusely branched. However, small impove- rished plants are usually simple, and in fact diffuse branching was found to be substantially less frequent in herbarium material from Holyhead than in that from the Channel Islands. There are most commonly 3 pairs of opposite-decussate cauline leaves, and the alternate, stipulate, upper cauline leaves are often absent. As Druce observes, the leaves are dark green, rather coriaceous in texture, and relatively wider than those of Jersey plants (but see pp. 241-2). Stellate hairs are usually present on the upper surface of the leaf, forming a sparse felt, at least near the apex. Bracts are usually present, subtending some or all of the pedicels, but by no means constantly so. The Holyhead plants have a rather distinctive appearance, with their short rigid spreading inflorescences, and their compact dark green vegetative growth. They remain shorter than the Channel Islands plants in cultivation. Their habitat, in the bare patches between wind-cut heathers on the steep slopes above the cliffs facing South Stack, is extremely exposed. The Caernarvonshire locality opposite Bardsey Island is equally exposed : specimens I have seen from here are similar in stature to Holyhead plants (2-4 cm), but simple and more slender in habit. In cultivation they differ in their untidier and more vigorous vege- tative growth, fewer bracts, and in the less coriaceous texture and paler colour of the leaves: they become almost as tall as the most compact of the Jersey plants (l’Etacq). Watsonia 5 (4), 1962. 144 M. C. F. PROCTOR Plants from the Clegyr Mawr colony are similar in stature, but with laxer inflorescences (mean distance between successive pedicels in 1961 experiment 8-66 mm, against 7-20 mm in material from Ffynnon Fair; difference significant at P < 0-05), and the presence of bracts is much less constant. Some herbarium specimens can scarcely be distinguished from the largest plants collected at Holyhead, while others do not differ in any significant respect from small Channel Islands plants, which plants from this population resemble in cultivation. The Clegyr Mawr population grows among low wind-cut heather on a slight rocky knoll in the shallow valley leading from Clegyr Mawr to the sea, and is probably considerably less exposed than either the Holyhead or Lleyn populations. The exact location of Williams’s population *‘ Gader.... facing the Skerries....’ is not certain. His specimens resemble the Clegyr Mawr plants, and it is perhaps possible that the locality is the same. His description suggests a position near Carmel Head (c. Nat. Grid Ref. SH 296924), where there is certainly suitable-looking ground, but if 7. guttata occurred here | have been unable to refind it. In any case the habitat must have been less exposed than at Holyhead. Irish plants from Three Castles Head, Inishbofin, and Inishturk are all diffuse in habit when well-grown, though they are extremely plastic and, as in Wales, starved specimens are often simple. Bracts are often present, though less regularly than in the Holyhead plants: they appear to be significantly more frequent on Inishbofin than at Three Castles Head. I have not seen herbarium material from the Calf Islands, but there seems to be little variation in size and habit in Irish 7. guttata from the remaining localities beyond that due to plasticity (though there is a surprisingly tall, simple, specimen from Three Fig. 3. 1956 cultivation experiment; silhouettes of representative plants, x }. a, Jersey; b, Baie des Trépassés; c, Clegyr Mawr, Anglesey; d, Holyhead, Anglesey; e, Three Castles Head, Co. Cork. Watsonia 5 (4), 1962. BRITISH FORMS OF TUBERARIA GUTTATA 245 Castles Head in Babington’s herbarium (CGE)). Plants grown from seed from Three Castles Head and Inishbofin uniformly produced very dense much-branched leafy rosettes, with many short inflorescences late in the summer. TABLE 4. Ranking of cultivated material of six populations of 7. guttata by seven characters (1961 experiment). | Bract Leaf Stem Number | Order of | Seed Height | frequency colour | pubescence | of branches flowering length Population (tall- ( few- (dark- (sparse- (many- | (early- (small- short) many) light) dense) few) late) large) Corbiére 1 1 1 1 2 1 2 L’Etacq 2 4 2, 5 6 3 1 Ffynnon Fair 3) 3 6 6 4 6 6 Clegir Mawr 4 | Deine 4 3 5 4 | 3 Holyhead 5 | 6 | 3 2 1 2 4 Inishbofin 6, ae ee 4 5 ee 5 i Fig. 4. 1959 cultivation experiment; silhouettes of representative plants, x }. a, Sintra, Portugal, 20 July; b, ibid., 27 Aug.; c, Jersey, 20 July; d, ibid., 28 July; e, ibid., 4 Aug.; f, Holyhead, 20 July; g, ibid., 4 Aug.; h, Ffynnon Fair, 28 July; i, ibid., 27 Aug.; j, Middle Quarter Hill, Inishbofin, 28 July; k, ibid., 18 Aug.; 1, ibid., 27 Aug.; m, West end of Inishbofin, 18 Aug.; n, ibid, 27 Aug.; The dates given are when plants were pressed. Watsonia 5 (4), 1962. M. C. F. PROCTOR 246 Q “pigi OF -P99S Pd}da][OO-pfiIMm WOIJ) “SNYW g “IMeY] IABID “puvjolay pue soyeAA WOT [elem : ‘pe °t x deg 8 “pigi ‘I s-3ny g “uyoqysiuy ‘ JO1END [PPI 9S 1dag 8 “pig ‘Y -SNY 8 “ez uouudyy “y ¢ dag g “pig? ‘a S-3ny g “IMI IABOIO ‘p {(s3uT] 6 q 6 Go "ydasg BNY § ‘pesysoH JUSUULIBAX9 UOTJCAIND [OS] “9 “SI ‘a Sydag g BT “q :3dag 8 oy} pue Aueyig woul [ert ¢ ‘deg g pig) ‘J :-3ny 8 “AOSIO£ ‘pigi “p 3ny g ‘Aosiog¢ ‘aigIqiog fo , 6 ‘bov1q «I 9 3ny g ‘Aosiog ‘oko “Aueyiiig ‘sossedoiy, sop oreg ‘e -% x s}eUl : JUSUTIIAdKS UONANIND [96] “S “BI 6 Spuvysy] jouueyD Watsonia 5 (4), 1962. BRITISH FORMS OF TUBERARIA GUTTATA 247 TABLE 5. Spearman rank correlation coefficients calculated from the data of Table 4. Only in the cases of the correla- tions between flowering date and leaf colour and stem pubescence, and between seed length and leaf colour is p great enough to establish a significant correlation between two characters with P < -05. How- ever, in all the remaining cases the departure from unity demonstrates the absence of close correlation. Flowering Seed Bracts Colour Pubescence | Branching date length Height oral ‘60 09 = 43 60 uf Bracts -26 ee :20 = up 2115) ee Bison Colour ry ri 66 14 94 ss, “89 Pubescence ai | 2 66 “83 ne °3)7/ Branching o | | 43 — :37 Flowering date cc fh Oey | . 71 In size and habit the populations form a graded series from the tallest plants from Jersey (La Moye) to the most compact Irish plants from Inishbofin. The relation between height and some other characters is illustrated in Tables 4 and 5. Of these characters, only bract number and leaf colour are markedly correlated with height, and even here the correlation is not statistically significant at the 5% level. In general, each character is correlated with not more than four others: two characters (e.g. bract frequency and leaf colour) may both show correlation with a third character (e.g. height), without themselves being closely correlated. The whole body of data suggests a network of loose correlations in several unrelated directions, in which no single direction of variation predominates strongly over the others. It is noteworthy that in Table 5 the variation associated with leaf colour and flowering date appears to be at least as important as that associated with the more familiar and obvious characters of height, and frequency of bracts. (e) T. guttata in north-west France Plants on the exposed coasts of Brittany and the Bay of Biscay are characteristically short in stature and diffuse in habit, usually with rather denser stellate and simple pubes- cence on the upper surface of the leaves than in the British populations. Some (including the only sample I have had in cultivation) closely resemble the Channel Islands plants and are more or less ebracteate; but populations exist which appear to be very comparable with the populations at Holyhead and in the west of Ireland : low-growing much-branched plants, with short bracteate inflorescences, e.g. Rochers mar. de Pornichet 4 Chemaulin (Loire-inf.), July 1874, Herb. E. Gadeceau (BM); St. Brevin (Loire-inf.), 23 June 1856, J. Lloyd (BM); Croisic (Loire-inf.), 23 June 1880 J. Lloyd (BM); ibid., June 1851, Drs. Thuret & Broceaux {?], Herb. Gaston Genevier (CGE). These variable west coast plants were described by Lloyd (1876) as H. guttatum var. maritimum. maritimum. H. alyssoides Pesn. cat., 59, non Vent. Plante de 5-10 cent. rameuse, étalée, hérissée, blanchatre; rac. epaisse, dure, mais [annuelle]. Rochers marit. Rouy & Foucaud (1895) call the plant Helianthemum litorale, citing Lloyd’s name as a synonym, and they give stature and presence of stellate pubescence on the upper surface of the leaves as diagnostic characters. The habitat is given as “‘ Rochers maritimes des bords de la Manche et de |’Océan.”’ Neither Lloyd nor Rouy & Foucaud mention bracts. The specimens mentioned above, collected by Lloyd and named var. maritimum, are bracteate, but the name evidently included taller and more or less ebracteate plants Watsonia 5 (4), 1962. 248 M. C. F. PROCTOR as well, e.g. Falaise de Carteret, (Manche), 31 May 1888, Herb. L. Corbiére (K), which is very like material from the more exposed localities in Jersey. What is particularly significant is Lloyd’s comment after his description of var. maritimum: ‘ Revient graduelle- ment au type, 2 mésure que l’on s’éloigne de la mer.’ It appears, then, that in north-west France we have the common pattern of a wide- spread species passing gradually into a dwarf maritime ecotype on exposed rocky cliffs; a correlation between the ‘ breweri’ characters and exposure is also apparent in Britain, and it seems reasonable to suppose that the disjunct British populations represent scattered fragments of a similar pattern. A similar pattern may also exist on exposed coasts in N. Spain and Portugal: the plants cited by Grosser provide little evidence of it, but two very short compact flowering specimens (c. 1-5-2 cm high) in the Kew Herbarium (Adjuda Hill, Lisbon, May 1935, S. C. Atchley Fl. Portugal 401; Foothiils facing the Atlantic between Estoril and Cintra, May 1935, S.C. Atchley, Fl. Portugal 402), look much more likely to be comparable with British ° breweri.’ They differ from any British or French material in their very dense grey stellate tomentum with only sparse simple bristles. The common feature running through the extreme west coast forms of 7. guttata is the short much- branched growth habit, often associated with bracteate inflorescences: in fact the characters emphasised by Planchon as distinguishing breweri. This may be due to a tendency for other, non-adaptive, characters to show developmental correlations with the characters responding to selection for short stature, and perhaps for long vegetative period; and if this is so, the “ breweri’ characters are likely to be of polytopic origin in the various popula- tions. The differing indumentum of British and French (and Portuguese) plants, and the variation in frequency of bracts and other characters in the British populations, are consistent with this possibility. It would be interesting to know something of the morphogenesis of the leaf and shoot pattern in 7. guttata. 4. DISCUSSION The question arises whether these diverse populations can be distributed into useful taxa. The two extreme, and simple, possibilities are either to limit the name breweri to the Holyhead population, or to follow Grosser and adopt a wide delimitation of breweri to include a range of Atlantic coast plants, of varying stature, with or without bracts. The first possibility would logically require similar separate taxonomic recognition of the Caernarvonshire, Irish and north-west French breweri-like forms, and although there are zoological precedents for a treatment of this kind in dealing with isolated small populations within the British Isles, I feel that this would be of little scientific value in the case of a generally variable species like 7. guttata. On the other hand, Grosser’s delimitation of T. guttata var. breweri provides no satisfactory line of distinction from typical guttata. Stellate pubescence on the upper surface of the leaves is commoner near the Atlantic coast, but it is found in a proportion of plants throughout much of the range of the species (and in the British Isles it is least constant at Three Castles Head !). Grosser’s delimitation results in the abandonment of virtually all the characters of Planchon’s original H. breweri. Willkomm?’s figure of T. variabilis var. cinerea (‘ad specimen Welwitschianum delineatam ’) is certainly rather like the Welsh and Irish plants, and has a few bracts; but specimens at Kew and Cambridge show Welwitsch It. Lusit. (1840) No. 433 (which Willkomm and Grosser cite) to be a rather tall plant (11-23 cm), having little in common with the Holyhead plant. The Channel Islands plants can scarcely be excluded from Grosser’s variety, though by implication he seems to include them in typical gutfata; and they were specifically excluded from the original delimitation of breweri and included in typical guttata by Planchon. Any intermediate treatment breaks down over one or both of two difficulties: the difficulty of providing usable and taxonomically meaningful lines of demarcation, and the fact that most of the possible groupings of the populations obviously lack genetic coherence. The extreme populations are certainly strikingly different from typical T. guttata, but on Watsonia 5 (A), 1962. BRITISH FORMS OF TUBERARLA GUTTATA 249 analysis their characters reduce to the slight, and obviously adaptive, one of stature, in which they intergrade with the typical form of the species. Other characters, which might be expected to be taxonomically more reliable, do not correlate at all closely with this, and do not appear to show any major taxonomically significant pattern of their own. There is a strong argument for simply recognising the existence of a pattern of ecotypic differentiation in north-west European T. guttfata in relation to exposure, and not attempting any formal taxonomic treatment of it. On the present evidence I am forced to the conclusion that no orthodox taxonomic treatment of the variation in these 7. guttata populations is practicable or of any real value. ACKNOWLEDGMENTS The work on which this paper is based was begun during the tenure of a research studentship from the Nature Conservancy. I would like to thank all those who helped by providing seed, living plants, herbarium material, or in other ways; and the botanic garden staffs at Cambridge and Exeter and Miss M. A. Turner for their care and patience with experimental material. REFERENCES BABINGTON, C. C. (1843). Manual of British Botany. London. Ed. 2, 1847; ed. 3, 1851; ed. 4, 1856; ed. 5, 1862; ed. 6, 1867; ed. 7, 1874; ed. 8, 1881; ed. 9 (edited H. & J. Groves), 1884. B[ABINGTON], C. C. (1844) [ Note on Helianthemum guttatum|. Ann. Mag. Nat. Hist., 14, 514. BENTHAM, G. (1858). Handbook of the British Flora. London. CoLGANn, N. & SCULLY, R. W. (1898). Contributions towards a Cybele Hibernica. London. Curtis, W. (1775-98). Flora Londinensis, 5. London. Druce, G. C. (1890). Helianthemum guttatum in Anglesea. J. Bot. Lond., 28, 315. Druce, G. C. (1891). Rept. Bot. (Soc.) Exch. Cl. (1890), 1, 283. 7 WRwUGE, Gz, (1893). Rept. Bot. (Soc.) Exch. Cl. (1892), 1, 355. Druce, G. C. (1902). Helianthemum breweri Planch. J. Bot. Lond., 40, 231. Druce, G. C. (1920). Helianthemum breweri Planchon, in Plant Notes etc. for 1919. Rept. Bot. (Soc.) Exch. Cl. (1919), 5, 549. GRIFFITH, J. E. (1895). The Flora of Anglesey and Caernarvonshire. Bangor. Grosser, W. (1903). Cistaceae in Engler, A., Pflanzenreich, IV, 193. Hooker, J. D. (1870). The Student’s Flora of tie British Islands. London. Hooker, W. J. & ArnotT, G. A. W. (1860). The British Flora. Ed. 8. London. Hupson, W. (1778). Flora Anglica. Ed. 2. London. Hype, H. A. (1931). Samuel Brewer’s diary: a chapter in the history of botanical exploration in North Wales. Rept. Bot. (Soc.) Exch. Cl. (1930), 9, supplement, 1-30. LESTER—GARLAND, L. V. (1903). A Flora of the Island of Jersey. London. LLOYD, J. (1876). Flore de Quest de France. Nantes. Moore, D. & Mors, A. G. (1866). Contributions towards a Cybele Hibernica. London. Morg, A. G. (1876). Report on the flora of Inish-bofin, Galway. Proc. Roy. Irish Acad., 2 (Science), 553-578. PLANCHON, J. E. (1844). Description of a new British species of Helianthemum. London Journal of Botany, 3, 617. PrAgGer, R. Li. (1901). Irish Topographical Botany. Proc. Roy. Irish Acad., 3rd series, 7. PRAEGER, R. LL. (1934). The Botanist in Ireland. Dublin. Proctor, M. C. F. (1960). Biological Flora of the British Isles. Tuberaria guttata (L.) Fourr. J. Ecol., 48, 243-253. Ray, J. (1690). Synopsis Methodica Stirpium Britannicarum. London. Rouy, G. & Foucaup, J. (1895). Flore de France, 2. Asniéres & Rochefort. SmiTH, J. E. (1800). Flora Britannica, 2. London. Smitu, J. E. (1825). The English Flora, 3. London. SOwersy, J. & SmiTu, J. E. (1790-1814). English Botany. London. SyME, J. T. BOSWELL (1864). Sowerby’s English Botany. Ed. 3. London. Wareura, E. F. (1952) in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Ed. 1. Cambridge. Ed. 2, 1962. WILLKoMM, H. M. (1856). Jcones et descriptiones plantarum Europae austrooccidentalis, 2. Leipzig. Watsonia 5 (4), 1962. 250 M. °C) FY PROCTOR APPENDIX : MATERIAL OF T. GUTTATA USED IN CULTIVATION EXPERIMENTS Jersey. Seed from cultivated plants, probably originally from near Corbiére. D. McClintock. Near Corbiére, Jersey, 22 March 1960, Mrs. F. Le Sueur. Above Pinnacle Rock, between L’Etacq and Grosnez (on granite), Jersey, 14 March 1960, Mrs. F. Le Sueur. Exposed slope above cliffs facing Bardsey Island, near Ffynnon Fair, Aberdaron, Lleyn Peninsula, Caerns., April, 1958. Holyhead Mountain, opposite South Stack, 31 Aug. 1954, D. E. Coombe (and later collections by M.C.F.P.). Bare patches between wind-cut Calluna and Erica cinerea, on low rocky knoll between Clegyr Mawr farm and the sea, near Llanrhyddlad, Anglesey, July 1955. Seedlings collected 1 April 1961. Summit of Three Castles Head, W. Cork, on open rocky ground (Devonian slates). Patches of more closed vegetation (from which 7. guttata is absent) with Calluna, Erica cinerea, Cladonia spp., 14 Aug. 1952, S. M. Walters. Sloping bank, edge of pool below Three Castles Head, with Calluna, Erica and into marsh with Drosera, Sphagnum, Potamogeton polygonifolius etc., 14 Aug. 1952, S. M. Walters. Bare patches in Calluna heath near summit of Middle Quarter Hill, Inishbofin, W. Galway, 14 Aug. 1958. Rock outcrop in pasture near west end of Inishbofin, facing Inishark, 14 Aug. 1958. Dry short turf on cliffs above Baie des Trépassés on the Pointe du Van, near the Pointe du Raz, Finisiére, France, July 1952, C. D. Pigott. Sintra, Mercés, Estremadura, Portugal (from Estagao Agronomica Nacional, Sacavém). Roadside, Pe. da Sena, towurds Sintra, Portugal, 12 Oct., 1957, I. A. Williams. Watsonia 5 (4), 1962. LUZULA x BORRERI IN ENGLAND By JOHN E. EBINGER Roanoke College, Salem, Virginia ABSTRACT The hybrid Luzula x borreri was studied using the procedures for determining introgressive hybridisa- tion. Most of the hybrids were sterile and intermediate in their characteristics between the parents, L. pilosa and L. forsteri. A few plants, however, were found that showed characteristics similar to L. pilosa and occasionally a few seeds were produced by the hybrids. This similarity between some of the hybrids and L. pilosa as well as the rare production of seeds indicates that occasionally there is some gene flow between these two species. Luzula pilosa and L. forsteri are woodland species which belong to the subgenus Pterodes. Both are wide-ranging taxa that occur in Europe, with L. pilosa having a more northern distribution. The range of L. forsteri, however, overlaps that of its more northern relative and through most of this region hybridisation has been reported. The hybrid was first described under the name L. Borreri by Babington (1851) who inserted it doubtfully as a species. He further mentioned that only two seeds were found on the specimen. This taxon was later considered by Syme (1872) to be a variety of L. pilosa. However, Buchenau (1890) was the first to recognise it as a hybrid. Generally this hybrid is completely sterile. Richards (1952) mentions that Luzula x borreri is not uncommon with the parents, and that it resembles L. pilosa, although it is often taller. According to him the capsules of the hybrids are much shorter than the perianth and always sterile. In recent hybridisation experiments, Nordenskicld (1957) indicates that L. forsteri (2n = 24) and L. pilosa (2n = 66) can be crossed readily. In all cases the hybrids were completely sterile. It was also indicated that in the F, hybrid the 33 small chromosomes of L. pi/osa sometimes pair with the 12 larger chromo- somes of L. forsteri. Although the hybrid is usually sterile, occasionally a few seeds develop. Therefore there is a chance for the transfer of genes between parents. To determine if there is gene flow between these two species, the procedure described by Anderson (1949) for determin- ing introgressive hybridization was used. The present study is primarily confined to collections of these taxa and their hybrid from England. MATERIALS AND METHODS This study was based on herbarium material consisting of more than 100 collections from England. Also, over 200 European specimens were examined to sample the total variability of the two taxa. Fruiting specimens were used for all measurements and the results were plotted on a scatter diagram to show the variability of both species and their hybrid. Material used in this study was obtained from the Gray Herbarium (GH), the Royal Botanic Gardens, Kew (K) and the United States National Museum (US). DISCUSSION Luzula pilosa and L. forsteri are morphologically distinct from one another. L. forsteri has a short projecting mucro extending from the callose tip of the leaves; the basal leaves are usually less than 3mm wide; the cauline leaves are usually more than 4 cm long; the pedicels of the fruiting inflorescence are erect to slightly spreading; and the caruncle of the seeds is erect and does not exceed 0-6 mm in length. In contrast, Za Watsonia 5 (4), 1962. 252 JOHN E. EBINGER L. pilosa \acks a projecting mucro on the callose tip of the leaf, the basal leaves are over 4 mm wide, the cauline leaves are less than 3 cm long, the pedicels of the fruiting inflorescences are strongly spreading to reflexed; and the caruncle of the seeds is strongly curved and over 0-6 mm long. In most of the above traits the hybrid is intermediate between the two species. The projecting mucro on the callose leaf-tip may or may not be present and the cauline leaf length is usually between 3 and 5cm. The curvature of the pedicels is also variable, but they are usually strongly spreading as in L. pilosa. In contrast, the basal leaf width is more than 4mm, similar to that of L. pilosa. The distinctive difference, however, is that the seeds of the hybrid are generally small and abortive. Measurements of the length of the cauline leaves and the width of the basal leaves were made on all available specimens of these two species and their hybrids from England. Observations were also made on the presence or absence of a projection on the callose tip of the leaf, the curvature of the pedicels, and the presence or absence of seeds. These results were then plotted on a pictorialised scatter diagram (Fig. 1) in which the abscissa represents the width of the basal leaves and the ordinate the length of the cauline leaves. A line radiating from the top left of the circle indicates that there is a projecting mucro on the callose tip of the leaf, and a line radiating to the right indicates that the pedicels of the inflorescence are erect, or nearly so. Plants with abortive seeds are indicated by solid circles while plants with normal seeds are represented by open circles. 4 5 7 $ PEDICELS ERECT s MUCRONATE TIP PRESENT = ( ¥ y @ SEEDS ABORTIVE : 9p) Bal PORE ES QO es YE YYYBY es a | Se od coe mn | Vane yey ae Bb & ¥ ‘eae ede = b z b 7 34 anrohvo « a 6° © Oo oole = 4 2 4 222 6 900 00 00 OF 0 fe) se) | : I 2 3 4 5 6 7 8 BASAL LEAF WIDTH (mm) Fig. 1. Pictorialized scatter diagram showing the relationship of cauline-leaf length to the basal-leaf width in Luzula pilosa, Luzula forsteri and their hybrid. The bar radiating to the left of the circle indicates that there is a projecting mucro on the callose tip of the leaf, while a line radiating to the right of the circle indicates that the pedicels of the inflorescence are erect to slightly spreading. An open circle indicates that seeds are present while a solid circle indicates that the seeds are abortive. Watsonia 5 (4), 1962. LUZULA xX BORRERI IN ENGLAND 253 As can be seen from the diagram (Fig. 1) three fairly definite groups are present. The group at the upper left represents Luzula forsteri, the group at the lower right re- presents L. pilosa and the group lying between these two species, with solid circles, represents the hybrid. The hybrid apparently resembles L. pi/osa more than L. forsteri, and the similar leaf measurements as well as the other traits indicate that occasionally the hybrid back-crosses with L. pilosa. Five specimens with abortive seeds have the same leaf measurements and the spreading pedicels of L. pilosa |J. P. M. Brenan 6248; C. E. Britton (16 May 1909); R. W. Butcher (17 May 1925); Milne-Redhead 6135; B. M. Watkins 1170]. Of these five, three also lack a projecting mucro on the callose tip of the leaves. In addition, on two of the hybrids a few seeds were found | A. Bromfield (12 Aug. 1850); £. C. Wallace (27 May 1933)]. In both instances there were no more than four seeds, and no more than two to any one capsule. It is not known whether these seeds are viable, and it was not possible to determine this during the present investigation as only herbarium material was used. However, these seeds were as large as those of L. pilosa and the embryos were well developed. These plants with a few seeds are also represented by solid circles on the diagram. Furthermore, all of the plants with abortive seed, including these two specimens on which a few seeds were found (all specimens represented by solid circles), are considered to be L. x borreri. Though the seeds of the hybrid are generally abortive, the anthers are well developed. They are very long (usually 1-5 mm), three times as long as the short filaments. By contrast, the anthers of Luzula pilosa are about 1-0 mm long, half the length of the filaments while the anthers of L. forsteri are slightly shorter. Though the anther lengths in both species are variable, in no instance are they as long as in the hybrid. Further- more, the pollen of the hybrid is similar to that of the two species though slightly larger. CONCLUSIONS It has generally been considered that the hybrid between Luzula pilosa and L. forsteri is completely sterile, but present information indicates that occasionally the hybrid sets a few seeds, and may back-cross with one of the parents. Most of the hybrids are intermediate between the two parents and have abortive seeds. However, five specimens were found that, though they usually had abortive seeds, had the same leaf measurements and the spreading pedicels of L. pilosa and three of these also lack a projecting mucro on the callose leaf-tip. This similarity between some of the hybrids and L. pilosa as well as the rare formation of a few seeds on the hybrid indicates that occasionally there is some gene flow between these two species. REFERENCES ANDERSON, E. (1949). Jntrogressive Hybridization. New York. BABINGTON, C. C. (1851). Manual of British Botany, Ed. 3. London. BUCHENAU, F. (1890). Monographia Juncacearum. Bot. Jahrb., 12, 1-495. NORDENSKIOLD, H. (1957). Hybridization experiments in the genus Luzula UI. The subgenus Prerodes. Bot. Not., 110, 1-16. RICHARDS, P. W. (1952) in Clapham, A. R., Tutin, T. G. and Warburg, E. F. Flora of the British Isles, Ed. 1. Cambridge. SyMgE, J. T. B. (1872). Report of the Botanical Exchange Club for the year 1872. J. Bot., Lond., 10, 237-247. REPRESENTATIVE SPECIMENS LUZULA FORSTERI (Smith) DC., Syn. Pl. Fl. Gall. 150 (1806). ENGLAND : A. Bennett 1002 (US); C. R. Billups 1549 (K); G. C. Druce (June, 1914) (GH); J. S. Gamble 18768 (K); A. G. Gregor (26 May 1926) (K); J. D. Grose (31 May 1939) (K); A. B. Jackson (May 1937) (ix); Jackson & Green 1656 (K); J. E. Little 2 (K); B. T. Lowne (19 May 1906) (K); A. G. More (May 1855) (K); C. E. Moss (1913) (K); W. B. Turrill (2 May 1926) (K); V. S. Summerhayes 1936 (K); C. Waterfall (14 May 1919) (K); D. P. Young 3992 (K). Watsonia 5 (4), 1962. 254 JOHN E. EBINGER LUZULA X BORRERI Bromf. ex Bab., Man. Br. Bot. Ed. 3., 333, (1851). ENGLAND : J. P. M. Brenan 6248 (K); A. Bromfield (12 Aug. 1850) (US); C. E. Britton (16 May 1909) (K), (23 May 1909) (K); (16 May 1926) (K); R. W. Butcher (17 May 1925) (K); C. B. Clarke (25 June 1891) (K); J. S. Gamble 26063 (K), 28500 (K); P. Hall 2299 (K); J. E. Lousley (18 June 1939) (K); Marshall & Sandwith (29 May 1918) (K); Milne-Redhead 6135 (K); E. C. Wallace (27 May 1933) (K); B. M. Watkins 1170 (US); R. Yindlay (April 1927) (K). LUZULA PILOSA (L.) Willd., Enum. Hort. Berol. 393 (1809). ENGLAND : C. C. Babington (2 June 1852) (K); R. A. Blakelock (7 May 1939) (K); C. E. Britton (21 May 1893) (K); H. E. Bunker 141 (i); B. Burtt 234 (KK); J. Cunnack 13536 (K); P. E. Goodliff (30 March 1957) (K); C. E. Hubbard 12667 (K); J. H. Lewis (19 July 1877) (GH); F. K. Makins 1910 (K); A. Meebold (Sune 1924) (K); Milne-Redhead 6136 (K); H. K. A. Shaw (18 May 1930) (K); T. A. Sprague (1 May 1932) (K, US); V. S. Summerhayes 2159 (K); D. P. Young 3993 (K). Watsonia 5 (4), 1962. BOOK REVIEWS Plant Communities of the Scottish Highlands. Donald N. McVean and Derek A. Ratcliffe. Pp. xiii + 445, with 36 figures and 34 maps in text, 26 plates, and two folding maps in end-pocket. Mono- graphs of the Nature Conservancy, No. 1. H.M.S.O., London, 1962. Price £3 17s. 6d. net. The tradition of descriptive ecology which culminated in Sir Arthur Tansley’s The British Islands and their Vegetation has not been prominent in Britain since the War, and for that reason alone the appearance of this book would be a notable event, and one to be welcomed. It sets out to give a monographic account of the vegetation types of the area north of the ‘ Highland Line’ (but excluding the largely agricultural country to the east, and coastal habitats) - an area which is extensive even by Continental standards. The work which this survey records grew out of Prof. M. E. D. Poore’s phytosociological work in Breadalbane, and the authors’ methods and concepts closely follow his. After a brief introductory chapter the greater part of the text is occupied by description and discussion of the communities recognised — 77 of them, not counting facies and a number of vegetation types which are not formally defined and named. These are grouped partly in life-form categories, partly ecologically: forest and scrub, dwarf-shrub heath, grass-heaths, herb and fern meadows, moss heaths, ombrogenous mires or bogs, soligenous mires, and springs and flushes. Chapter 10 tentatively relates the vegetation units described to the phytosociological classification used by the Scandinavian ecologists, and briefly compares Scottish and Scandinavian mountain vegetation. There follows a thoughtful chapter on climate and vegeta- tion, a chapter on soils, and a brief chapter on plant-geographical factors. Floristic analyses of the communities are set out in 60 tables, which take up 170 pages of the book — about as much as the text. The tabies are a tribute to the care and thoroughness of the authors’ field-work, and, quite apart from their primary purpose, will be a mine of information to anyone interested in the ecological or geographical distribution of vascular plants, bryophytes or macrolichens in the Highlands. The distribution of a number of selected communities is shown in a series of maps, which illustrate some of the main trends of vegetational variation; and a series of vegetation charts indicate the occurrence and altitudinal range of the main communities in a series of representative areas. The book is completed by twelve pages of soil and mire-water analyses, lists of calcicole, calcifuge and indifferent species, a summary of the distribution and soil preferences of the communities described, 26 photographs of representative pieces of vegetation (all at least adequate and some magnificent), an excellent index, and two folding maps showing the present and reconstructed past distribution of oak, pine and birch woodland in Scotland. Two points which emerge forcibly from the authors’ work are the large number of recognisable vegeta- tion types to be found even on acid mountains, and the richness in species of many of these. Thus even the species-poor Rhacomitrium heath of exposed summits averages 16 species (vascular plants, bryophytes, and macrolichens) in the 11 samples listed; not many of the communities average less than 20 species in a sample (usually 4 m2), and a number of the richer ones average between 40 and 50. In general, the character of the vegetation types emerges clearly from the descriptions and tables, and many of them will be recognised by those who have searched for plants in the Highlands. There is certainly a strong phytosociological ele- ment in the intuitive ‘ feel’ that one acquires for the habitat of a plant, and it is interesting to compare ones field experience with the summary of constant and dominant species and indications of fidelity in the des- cribed communities in tables 1 and 2, and with the floristic tables and the comments in the text. Thus Loiseleuria, Salix lapponum, Athyrium alpestre, Dryas and Carex saxatilis are all constant and more or less faithful species of particular noda, and many of the rarer species are confined, or nearly so, to particular communities. The authors have attempted so much, and in the main succeeded so well, that any criticism virtually amounts to an attempt to define the standard at which a study of this sort should aim. First there is the question of comprehensiveness. The authors have specifically excluded coastal vegetation, and they have not dealt with rock-face communities, which they consider are best dealt with in terms of the autecology of the species concerned, or with aquatic communities — some of which are certainly interesting and charac- teristic in the Highlands. Then there are the epiphytic bryophyte communities of woodland and scrub, and this brings one back to the mosaics and bryophyte micro-communities within some of the vegetation types already recognised. A line must be drawn somewhere. However, at least some rock-crevice communi- ties lend themselves to phytosociological treatment in other areas, and it would have been interesting to ee) Watsonia 5 (4), 1962. 256 BOOK REVIEWS have some account of the main types found in the Highlands; is there any regularity in the associates of Woodsia alpina, or Asplenium septentrionaie ? The number of noda recognised is large, and in some cases noda seem to be separated on rather tenuous grounds. However, the authors have undertaken a difficult and unusual task in providing a set of vegetation units to cover an area embracing as much climatic and edaphic variation as the Highlands — a problem which is not faced in quite the same way by the various regional compilations published on the Continent. As they point out, the choice is between a smaller number of more homogeneous units of local distribution, individually well-characterised, but difficult to separate from their equivalents in neighbouring areas, or between a smaller number of widely distributed but variable units which are difficult to characterise. This is a general problem which has no completely satisfactory solution, and compromise is necessary. With the vegetation analyseG into a number of noda, the reader is then faced with the problem of synthesis. Tine authors have taken a great deal of trouble to meet his needs in the maps and vegetation charts, with their explanatory text. Nevertheless, the reader’s task wouid probably have been made easier by a summary chapter outlining the main types of pattern shown in the charts, and the principal trends illustrated in the maps. This need is partly met in the chapters on climate and soils; but much of the relevant information has to be extracted piecemeal from the descrip- tions of the individual plant communities. One hopes that we shall be given the opportunity to study in more detail the evidence on which the two woodland maps are based. Finally, a minor point; if Latin names are to be used for plant communities, surely the usual forms “‘Agrostido-Festucetum’ and ‘Vaccinietum’ are etymological!y preferable to the forms used here ? It is rather refreshing to find that the authors have not added to the tiresome and sterile controversy which tends to surround discussion of phytosociological work, but have let their results stand as their own justification. Phytosociology has generally been discussed in terms of ecology in Britain, but it has as much or more to offer to those whose interests are in floristics or phytogeography. Systematic and thorough phytosociological study obviously has much to add to our knowledge of British vegetation. The authors have managed a nice balance between the coherent formal (but sometimes over-rigid) framework of Con- tinental phytosociology, and commonsense empirical (but sometimes vague) approach of pre-1939 British ecology. It would be a pity if descriptive work in Britain were now to be tied too closely to the ideas that have served some of the major Continental schools for the past few decades, or if it were to divert too much effort from related subjects. On the other hand, there is much scope for descriptive work on vegetation in Britain; Continental phytosociology represents one substantial body of experience on which to draw, and there are now encouraging signs both in Britain and on the Continent of convergence between tradi- tional descriptive methods and recent applications of multivariate techniques. The book is well printed on good paper, and the layout of the text and illustrations is clear and pleasing; misprints are few. The problem of printing extensive floristic tables has been well handled, and in spite of the small type-size the tables are admirably clear. Plant Communities of the Scottish Highlands is undoub- tediy the most important post-war contribution to descriptive ecology in Britain, and it is fitting that it should be the first of the Monographs of the Nature Conservancy. It sets a high standard. M. C. FY Procror Flora of the British Isles. Second Edition. A. R. Clapham, T. G. Tutin and E. F. Warburg. Pp. xlviii + 1269, with 87 figures in the text. Cambridge University Press, 1962. Price £3 10s. Od. Publication of the new Flora of the British Isles in 1952 has proved to be the most important event for a generation in the study of British plants, and the second edition was eagerly awaited. The large number of corrections and additions which have been made is to a considerable extent due to the stimulus to new work prompted by the first edition. The book has been largely re-written, and completely re-set in a new format, and, excellent as it was before, it is greatly improved. The new format is an outstanding improvement. The first edition was produced as a compromise between a work to be used in the field and one suitable for a library, and was far from ideal for either purpose. Printed on thin paper, more suitable for a Bible to be read at leisure than for a Flora to be con- sulted in haste under difficult conditions, many ccepies of the old ‘ CTW’ became soiled and tattered shortly after purchase. The shorter and more portable Excursion Flora is now available for use in the field, and the Flora of the British Isles is printed on good paper with a larger page size entirely suitable for the herbarium and library. Although the second edition contains about 320 pages less than the first, the contents are more, better displayed, and easier to consult. Watsonia 5 (4), 1962. BOOK REVIEWS 257 Some genera or paris of genera have been completely re-written in line with recent work. Polypodium, Salicornia, Oxalis, Ulmus (reduced from 7 species to 3), Gentianella, Rhinanthus (reduced from 8 species to 3), and Sparganium are examples. The new account of Rubus is more suitable for most users of the book, leaving specialists to consult W. C. R. Watson’s Rubi of the British Isles for detailed accounts of the species. The revision of Orchidaceae, and especially Dactylorchis, is outstanding, but surely Platanthera bifolia is characteristically found on more acid soils than P. chlorantha and the description of habitats on p. 1033 requires correction. Many readers will be pleased to find Polygala oxyptera sunk in synonomy, and Cochlearia alpina, Daucus gummifer, and Epilobium lamyi reduced to subspecies. In contrast to the first edition, there are few names which will be unfamiliar to most British botanists, though some of the scientific names introduced then are not retained. As the authors acknowledge in their preface, the work of J. E. Dandy is a step towards nomenclatural stability, and the names used are in general those of the List of British Vascular Plants. In the Flora, unlike that work, Scirpus is split up into Tricho- phorum, Isolepis, Eleogiton, Holoschoenus and Schoenoplectus. Changes based on careful study and tending towards stability are necessary but frequent alterations are difficult to justify. If the fern known as Dryo- pteris spinulosa is to be called D. carthusiana, the recent change to D. lanceolatocristata would appear to have been too hasty. Fortunately similar examples are few. At the time the first edition appeared the spelling of certain specific epithets with a small initial letter was highly controversial, and these were indicated by repeating the name with an initial capital. The new practice is now so generally accepted by British botanists (though still regretted by a few) that the cases where initial capitals are appropriate are no longer indicated. Another change in the second edition is that illustrations are no longer cited for each species. The justifica- tion for this is the production of the companion volumes of Flora of the British Isles Ilh:strations, of which two volumes have appeared, but these do not cover, and will not cover, all the species described, and not all the illustrations are as good as those formerly cited. It is to be hoped that the authors will consider re- storing this very useful feature in the next edition. The preface is dated as long ago as November 1958 and in bringing the book up to date the authors were handicapped by the length of time the pages were in the press. This is particularly evident with the paragraphs dealing with distribution, since although they were able to incorporate the major additions revealed by the work of the BSBI Distribution Maps Scheme, they had not got the advantage of the Atlas to check the overall pattern or the detail. The distribution of plants found in the Channel Islands is especially defective — for example, Ranunculus paludosus is not * Found only in dry places near St. Aubyns,’ Silene conoidea is not ‘ established in Jersey ’ (and never has been), and Spergularia bocconii is far more plentiful and widespread in Jersey and Guernsey than in any of the counties listed, though its occurrence is recognised by an ‘S.’ A protest too must be lodged against the use of ‘ Scilly Is.’ to refer to the archipelago officially known as the ‘ Isles of Scilly.” For a shortened form ‘ Scilly ’ is adequate, more acceptable to the residents, and runs no risk of being tautologous. The treatment of alien plants has been immensely improved but even so it remains by far the weakest feature of the Flora. Many aliens in the first edition which are now seldom seen have been dropped, but some of those in the Cruciferae and Compositae still retained would be more appropriate in a Victorian Flora than in one published in 1962. Similarly there are statements on distribution, which are fifty or more years out of date, such as that Herniaria hirsuta is ‘ Naturalized on sandy ground at Christchurch, Hants.’, and that Reseda phyteuma is ‘ Well established in Surrey.’ In general the statements about occurrences are too vague to be useful and the means of introduction are seldom mentioned. Later workers will have great difficulty in checking claims of persistence where no locality where the species is established is mentioned. For example, the statement that Arabis caucasica is * occasionally naturalized’ is vague. A reference to its spread since 1890 on limestone rocks in Derbyshire would indicate that in at least one district it is thoroughly established. For aliens for which special studies have been published there can be little excuse for some of the statements given. Thus Lacey’s paper on Galinsoga which appeared in 1957 must have been available when the entries that G. parviflora is ‘ Well established .... in London and elsewhere in S. England’ and G. ciliata ‘ A casual recently recognized in S. England and S. Wales’ were checked for the new edition and left unaltered. They are very much more widespread than this and still spreading, as the entries in Plant Records show. The selection of alien species for inclusion is always a difficult matter since every botanist’s experience differs, but there can be little doubt about the lack of merit of some of the plants included, and the claims of some of those left out. Inula britannica, Filago arvensis, and Crepis nicaensis are examples of those included in large type. Centaurea diluta, Erigeron bonariensis and Salvia reflexa are commoner plants which the reader is much more likely to find, while Juncus pallidus has current claims for inclusion as established. To include Erodium botrys, and leave out several equally common species of the same genus, and to describe ten alien species of Oxalis and only two of Amaranthus are clearly anomalous. It seems strange that the annual Ambrosia artemisiifolia, which seldom persists for long, is given space while the Watsonia 5 (4), 1962. 258 BOOK REVIEWS perennial A. psilostachya, which has been at St. Annes for 75 years, is not mentioned. Woody aliens, on the other hand, receive very much better treatment, and this is the best account of them which has yet appeared. This new edition of ‘CTW’ is a first-class work which can hold its own by comparison with any over- seas Flora of similar size. In general the revision is thorough and brought as nearly up to date as the time- table for printing allowed. Very few errors escaped notice in the proof reading (though “‘ the Isle of Dogs, Kent’ on page 227 should not have been missed). This is a work which no British botanist can afford to be without. J. E. LoUSLEY Watsonia 5 (4), 1962. us Hie uh He wee VIG vn ban © Ke Lat ¢ Cy: ny A BIBLIOGRAPHICAL INDEX OF THE BRITISH FLORA COMPILED BY N. DOUGLAS SIMPSON, ™.aA., F.Ls. The purpose of this work is to provide references to sources of information whereby flowering plants, vascular cryptogams and Charophytes found in Britain may be identified, their history traced and their geographical range determined. In addition, information is provided on plantlore, local names, poisonous plants and weeds. ) The work has been in preparation for nearly twenty years and contains over 65,000 entries, including references to books, articles and manuscripts relating to the flora of the British Isles from the fifteenth century to the present time. Demy Ato, bound, 448 pages in double column. Limited Edition of 750 copies. PRICE £3 15s. (postage extra) Obtainable from the compiler at “‘ Maesbury,’ 3 Cavendish Road, Bournemouth, Hants. IRISH NATURALISTS’ JOURNAL A Magazine of Natural History Published Every Quarter by the I.N.J. Committee. Edited by Miss M. P. H. KERTLAND, M.Sc., with the assistance of Sectional Editors. Annual Subscription, 10/- post free. Single Parts, 3/6. All communications to be addressed to:— The Editor, Department of Botany, Queen’s University, Belfast. — ee ee ee INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the taxonomy or distribution of British vascular plants or Charophytes are invited from both members of the Society and others. They should be typewritten, with wide margins, double spaced, on one side of the paper only; contributors are recommended to keep a carbon copy of their typescripts. The form adopted in recent parts of Watsonia should be used for layout, headings, citations and references. Con- tributors are urged to avoid very complicated hierarchies of headings and sub-headings, and to check carefully the consistency of those that they use. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British vascular plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (N.H.) and B.S.B.L, 1958), and may then be cited without authorities. Otherwise, authors of names must be cited, at least on the first occasion where they appear in the text. Except for citations of the place of publica- tion of plant names, full references should be listed in alphabetical order of authors names at the end of the paper. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals, ed. 3 (London, 1952). References to herbaria should include the abbreviations given in British Herbaria (B.S.B.1., 1957) and Index Herbariorum ed. 3 (1.A.P.T., 1956). Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new taxa, and the main conclusions. Line drawings should be boldly drawn in Indian ink on Bristol board or similar smooth white card, firm smooth-surface white cartridge paper of good quality, or suitable draughtsman’s tracing materials, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. No half-tone shading or wash should be introduced into line-drawings; all shading should be put in the form of separate dots or lines, and where large areas of uniform shading are required (e.g. in maps and diagrams) this should preferably be done by means of suitable ‘ mechanical tints.’ Graphs can be reproduced from originals on graph paper with faint blue ruling (not grey or any other colour), but all lines to appear on the finished _ block must be inserted in Indian ink. Lettering on line-drawings and graphs should be _ inserted lightly in pencil, and will be finished in uniform style. If an illustration includes plant-names or place-names, it is advisable to type these clearly on a separate sheet of _ paper. Photographs can only be accepted where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable _ for the necessary reduction. It should be remembered that the finest detail on the originals may be lost even on the best half-tone blocks. If in doubt about the citation of names or references, or the presentation of illustrations or tabular matter, contributors are advised ‘ to consult the editor before submitting their typescripts. Authors will normally receive both galley-proofs and page-proofs. Particular care should be given to the thorough checking of references and tables. Alterations in page- proof should be avoided as far as possible, and any words or phrases deleted replaced by others of equal length. Twenty-five separates are given free to authors of papers. Further copies may be - obtained at the Society’s current price; orders should be sent to the editor when the page- _ proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to the Editor, Dr. M. C. F. Proctor, Hatherly Biological Laboratories, of Wales Road, Exeter. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANpy for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf. of a sub-committee of the Society appointed for this purpose. Sir George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and have become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8 vo., 176+xvi pages. Bound in cloth. Price : 10/- (postage extra). Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. A DARWIN CENTENARY Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the very successful Conference arranged by the Botanical Society of the British Isles in 1959. The contributors include Lady Barlow, G. D. H. Bell, Prof. T. W. Bacher, Prof. W. H. Camp, Prof. N. Hylander, Prof. I. H. Manton, Prof. T. G. Tutin and Prof. D. H. Valentine. Demo 8 vo., 140 pages, with 7 half-tone plates and 11 maps and figures in the text. Bound in buckram. PRICE £1 (postage extra) Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7 SS sss The Salisbury Press Ltd., Salisbury—34256 [1 641 WATSONIA. JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 5 APRIL, 1963 Pt 5 | CONTENTS THE NOMENCLATURE OF ALCHEMILLA MINOR AUCT. BRIT. By M. E. BRADSHAW, P. D. SELL and S. M. WALTERS ae Me oe re a a 259-261 4 CIRCAEA IN THE BRITISH ISLES. By PETER H. RAVEN .. oe a 262-272 VARIATION IN SOME DIAGNOSTIC CHARACTERS OF THE SESSILE AND PEDUNCULATE OAKS AND THEIR HYBRIDS IN SCOTLAND. By J. E. COUSENS .. ee 273-286 THE STATUS OF ORCHIS LATIFOLIA VAR. EBORENSIS GODFERY IN YORKSHIRE. By R. H. Roperts and O. L. GILBERT ie a - fe 287-293 STUDIES IN RANUNCULUS SUBGENUS BATRACHIUM (DC.) A. GRAY. IJ. GENERAL MORPHOLOGICAL CONSIDERATIONS IN THE TAXONOMY OF THE SUBGENUS. By C. D. K. Cook a oe we Ae 294-303 STUDIES ON ALCHEMILLA FILICAULIS BUS., SENSU LATO, AND A. MINIMA WALTERS. INTRODUCTION, AND I. MORPHOLOGICAL VARIATION IN A. FILICAULIS, SENSU LATO. By MARGARET E. BRADSHAW ss ay 304-320 STUDIES ON ALCHEMILLA FILICAULIS BUS., SENSU LATO, and A. MINIMA WALTERS. II. CyTOLOGY OF A. FILICAULIS, SENSU LATO. By MARGARET E. BRADSHAW .. ey ne oy. oe ip ae: re \. 321-326 BOOK REVIEW ae 5 i ise A ae f; 4f Ne S27 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH MUSEUM (NATURAL History), LONDON PRICE FIFTEEN SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES Patroness: H.R.H. THE PRINCESS ROYAL OFFICERS President : J. E. LOUSLEY Vice-Presidents : J. E. DANDY, Miss C. M. ROB, Dr. E. F. WARBURG and Dr. S.M. WALTERS Hon. General Secretary: J. G. DONY, 9 Stanton Road, Luton, Beds. Hon. Treasurer: J.C. GARDINER, Thrift House, 12&14 Wigmore Street, London, W.1. Hon. Editers: (Watsonia) M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT Hon. Meetings Secretary: H. J. M. BOWEN, Pomander House, Harwell, Berks. Hon. Field Secretary: P.C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Longfield, Kent. Hon. Junior Activities Secretary: P. F. HUNT, The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey. ae Applications for membership should be addressed to the Hon. General Secretary, Dr. J. G. DONY, 9 Stanton Road, Luton, Beds., from whom copies of the Society’s Prospectus may aiso be obtained. WATSONIA Price to non-members, 15/— per part. To be obtained from E. B. BANGERTER, Botanica! Society of the British Isles, c/o Dept. of Botany, British Museum, (Natural History), London, S.W.7. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/- for Junior Members, and 10/— for Family Members, and the Society’s year runs from January Ist to December 31st. All subscriptions should be paid te the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13 THE NOMENCLATURE OF ALCHEMILLA MINOR AUCT. BRIT. By M. E. BRADSHAW, P. D. SELL and S. M. WALTERS This paper attempts to elucidate the various names applied to the taxon described in Clapham, Tutin & Warburg (1952) as Alchemilla minor Huds. sec. Wilmott, and to find its correct name. The problem has been tackled before by Wilmott (1939), Rothmaler (1941 and 1944), Samuelsson (1943) and Hylander (1945). Our conclusions disagree with those of Rothmaler and Wilmott and support those of Samuelsson and Hylander. We are very grateful to Mr. J. E. Dandy for all the time and trouble he has taken to answer numerous problems of nomenclature and typification. The first name to be considered is 4. hybrida (L.) L. (1756) (A. alpina var. hybrida L. (1753)). The diagnosis given for the basionym reads * A/chemilla alpina pubescens minor. Tournef. inst. 508. Pluk. phyt. 240 f. 1. The name A. alpina pubescens minor “ H.R.P.” was actually first published by Plukenet (1692) under his plate 240, fig. 2 (not 1 as quoted by Linnaeus), and then in his text (1696). It was later published by Tournefort (1719) who also attributed it to ““H.R.P.” Plukenet’s figure (reproduced in Fig. 1) was based on a specimen in his herbarium now on fol. 22, vol. 95, of Herb. Sloane at the British Museum (Plate lla). This specimen, which is incomplete (lacking basal leaves), originated in the Botanic Gardens at Paris (‘ Hort. Reg. Paris.’), from which it would have been sent to Plukenet by Tournefort. In the Tournefort herbarium at Paris (P) is a better specimen labelled ‘ A/chemilla alpina, pubescens minor hort reg Par’ (Plate 12a) which is definitely identifiable with A. /apeyrousii Buser (1893). There being nothing to the contrary, it may safely be assumed that Plukenet’s specimen is a duplicate of Tournefort’s and is therefore also referable to A. /apeyrousii. There are no specimens in Linnaeus’ own Fig. 1. Plukenet’s figure of ‘A/chemilla alpina pubescens minor.’ 259 Watsonia 5 (5), 1963. 260 M. E. BRADSHAW, P. D. SELL and S. M. WALTERS herbarium, but there are two sheets in his Hortus Cliffortianus at the British Museum. The Hortus Cliffortianus is not cited in the original publication of the name A. alpina var. hybrida, but the specimens were certainly seen by Linnaeus before 1753. Both are labelled with the Tournefort descriptive name ‘ A/chemilla alpina, pubescens minor’ and both are in our opinion A. /apeyrousii Buser (though Rothmaler (1962) refers one to the species we are here considering). There is thus little doubt which taxon Linnaeus meant by A. hybrida. The next name that has been applied is A. minor Huds. (1762). Under this species is cited exactly the same synonym as under A. alpina var. hybrida L., 1.e. * Alchemilla alpina pubescens minor. Tourn. 508. Pluk. Ph. t. 240 f. 12 Consequently the publication of A. hybrida as a species in 1756 makes Hudson’s name a superfluous nomen illegitimum. As to what British plant Hudson had in mind it is difficult to tell, as none of his specimens has been traced, but the taxon we are discussing has never been found in Westmorland and Hudson’s plant is more likely to have been A. vestita (Buser) Raunk. Next is A. pubescens Lam. (1791) which is also a superfluous nomen illegitimum, as is A. montana Willd. (1809). Both make either direct or indirect reference to A. hybrida (L.) L. These superfluous names probably arose because the authors concerned did not know that Linnaeus had raised his variety hybrida to specific rank. This brings us to the name A. g/laucescens Wallr. (1840), which is the correct specific name of the taxon under discussion. This name first appears on page 134 of Linnaea in 1840. Here Wallroth says that, owing to the heterogeneous nature of the specimens labelled A. montana in the Willdenow herbarium, he will from now on call the plant he had described as A. montana in 1815 A. glaucescens. Thus the type description of A. glaucescens is that published as A. montana in 1815. Later in the year 1840 Wallroth published in the same work (p. 549) a new variety of glaucescens, var. alpestris. From his notes below this description it is clear that he names his type variety a/pina, and the only original specimen which we have as yet been able to trace is one of Willdenow’s specimens of A. montana in the Berlin herbarium (B) Plate (11b), annotated by Wallroth as A. glaucescens var. alpina. This specimen is referable to the taxon here discussed and it 1s proposed to select it as a lectotype until other material can be found with a better claim. As all the names here discussed apply to one or the other of two taxa it is proposed to set out their synonymy and give comparative descriptions and distributions. ALCHEMILLA GLAUCESCENS Wallr. (1840) 134 (Lectotype : Willdenow specimen in Herb. Berlin annotated by Wallroth); Samuelsson (1943) 31; Hylander (1945) 205; Clapham, Tutin & Warburg (1960), 10, fig. 588. | A. montana sensu Wallr. (1815) 26 (quoad descript. exclud. syn.).| A. glaucescens var. alpina Wallr. (1840) 549. [A. pubescens sensu Salmon (1928) 345; sensu Druce (1928) 36, no. 190: 1.] [A. hybrida sensu Butcher & Strudwick (1930) 131, plate 144; sensu Rothmaler (1962)] [ A. minor sensu Wilmott (1939) 249; sensu Walters (1949) 9, 16; sensu Warburg in Clapham, Tutin & Warburg (1952) 506; sensu Ross-Craig (1956) 9, plate 1; sensu Dandy (1958) 55, no 22071377 0) Plant small, up to 20 cm tall, grey-green, silvery. Stem fairly stout, with strong lateral branches giving a dichotomous appearance, yellow-green, often reddening in the summer, densely hairy with spreading or slightly ascending hairs. Radical leaves small, 3-5 cm broad, usually round in outline, basal sinus usually closed, basal lobes often overlapping; lobes 7 or incompletely 9, short, }-} radius of leaf, broad, semicircular, often overlapping at the base; teeth large, subobtuse, 6-7 on each side of lobe; upper surface usually densely hairy, lower very densely hairy with long, silky, silvery hairs; petioles densely hairy with long spreading or slightly ascending hairs. Cauline leaves small, reniform, incompletely 5-lobed, the lobes broad and rounded, densely hairy on both surfaces and with few large teeth. Flowers in dense clusters appearing silvery; pedicels, calyx and outer surface of the sepals densely hairy. Watsonia 5 (5), 1963. “TEM SUadsaonv]s BI[NWU2yo]p JO 9AK}O1DIT “(ESLT) “I ppligdy “eA DUId|D D]]IWaYyrP JO UowIdeds odA J, HERBARIUM MUSEL BRITANNICI : Vy] * Te ZL rs 3 bp Pe oe bp POPES Ae oe BIg +6 Pipe (i ERIE GEIS a Lies : : ‘ [MH ee moor pa of oe. ‘ f (A 7 77 | . d / _. ig ‘SS 6s age Loe t i % P: tHe 5: iy f i : 4 : \ § 5 \ \ 8 Be 3 Bit elk) oe : , joe ie NSWI9NdS AdAL ‘el [ ALVId OL] aALvid “WUNTIeGIOH 1OJOUINO |], UI , JOUI Suaasagqnd DUId|D DyNUaYyap , Jo UdUIDEd¢ ‘josng usnosdady] oyjiuayrp JO 9d410}09T yi ¥ Wot bai PES WE SIT b d ig iy DEBE. BOK. fag h 6 hp RAGUSA (seh $6: ip SBOE AISEIGIS B14 DASE ITAL SIS Us LAGE fy iy i BAR SS YS aK BOUSB UA IHEEY ty PMY AS G4 i i Y Gi yp Y GZ Uy, f $ aapgnngenaroccans y ¥ y ¥ y y Y, y y j Y SEQ ‘QZI aLWId | "eT. ALVId THE NOMENCLATURE OF ALCHEMILLA MINOR AUCT. BRIT. 261 Distribution : Austria, Belgium, British Isles (N. England, Scotland and W. Ireland), Bulgaria, Corsica, Czechoslovakia, Denmark, Finland, France (but not Pyrenees), Germany, Switzerland, Hungary, Italy, Jugoslavia, Netherlands, Norway, Poland, Roumania, Russia and Sweden. ALCHEMILLA HYBRIDA (L.) L. (1756) 49. A. alpina var. hybrida L. (1753) 123 (Holotype : Plukenet specimen on fol. 22 of vol. 95 in herb. Sloane (BM)). A. minor Huds. (1762) 59 (superfluous nomen illegitimum for A. hybrida (L.) L.). A. pubescens Lam. (1791) 347 (superfluous nomen illegitimum for A. hybrida (.) L.). A. montana Willd. (1809) 170 (superfluous nomen illegitimum for A. hybrida (L.) L. A. lapeyrousii Buser (1893) app. 2, 18 (Lectotype : Sommet du Puy-de-Dome, France, 27/9/1892, J. Héribaud in herb. Boissier (Herb. Geneve (G) Plate 12b). A. anglica Rothm. (1937) 167 (superfluous nomen illegitimum for A. minor Huds.). Differs from A. glaucescens Wallr.: plant larger, up to 25cm tall, generally less hairy in all its parts, especially the flowers; stem not bifurcating nor reddening in summer, and the branches less strong; radical leaves reniform, flat, sinus open, lobes longer, c. 4 radius of leaf, more triangular, teeth smaller, more acute and more connivent; pedicels with few hairs or glabrous; and calyx and sepals not densely hairy. Distribution : France (Pyrenees, Cévennes and Auvergne); Spain (Pyrenees). REFERENCES Buser, R. (1893). Notes sur les plantes distribuées, et diagnoses des espéces nouvelles ou peu connues. Bull. Herb. Boiss., 1 app. 2, 18. BUTCHER, R. W. & STRUDWICK, F. E. (1930). Further Illustrations of British Plants. Ashford. CLAPHAM, A. R., TuTIN, T. G. & WaArBurRG, E. F. (1952). Flora of the British Isles. Ed. 2 (1962). Cambridge. CLAPHAM, A. R., TuTIn, T. G. & WaArburG, E. F. (1960). Flora of the British Isles. Illustrations, part Il, Rosaceae-Polemoniaceae. Drawings by S. J. Roles. Cambridge. Danpy, J. E. (1958). List of British Vascular Plants. London. Druce, G. C. (1928). British Plant List, ed. 2. Arbroath. Hupson, W. (1762). Flora Anglica. London. HyYLANpeER, N. (1945). Nomenklatorische und Systematische Studien tiber Nordische Gefasspflanzen. Uppsala Universitets Arsskrift, 7, Uppsala. LAMARCK, J. B. P. DE (1791). Tableau encyclopédique. Paris. LINNAEUus, C. (1753). Species Plantarum. Stockholm. LINNAEUS, C. (1756). Amoenitates Academicae, 3. PLUKENET, L. (1692). Phytographia tab. 240 fig. 2. London; and (1696) Almagestum Botanicum, 18. Ross-Craic, S. (1956). Drawings of British Plants, 9. London. ROTHMALER, W. (1937). Systematische Vorarbeiten zu einer Monographie der Gattung A/lchemilla (L.) Scop. VII. Aufteilung der Gattung und Nomenklatur. Feddes Repert. 42, 164. ROTHMALER, W. (1941). Jbid. IX. Uber Alchemilla-Arten aus Osteuropa und Asien, 50, 245. ROTHMALER, W. (1944). Zur Nomenklatur der Europdischen A/chemilla-Arten. Svensk Bot. Tidskr., 38, 102. ROTHMALER, W. (1962). Systematische Vorarbeiten zu einer Monographie der Gattung Alchemilla. X. Feddes Repert. 66, 194-234. SALMON, C. E. (1928). Alchemilla pubescens Lam. as a British Plant. J. Bot., Lond., 66, 345. SAMUELSSON, G. (1943). Die Nomenklatur der nordeuropdischen Alchemillen. Svensk Bot. Tidskr., 37, 27. TOURNEFORT, J. P. DE (1719). Jnstitutiones Rei Herbariae, ed 3. 1, 508. Paris. WALLROTH, F. W. (1815). Annus botanicus, 26. Halle. WALLROTH, F. W. (1840). Exonion zu Hampe’s Prodromus Florae Hercyniae. Linnaea, 14, 134 and 549. WALTERS, S. M. (1949). Alchemilla vulgaris L. aggr. in Britain. Watsonia, 1, 6. WILLDENOW, K. L. (1809). Enumeratio plantarum horti regii botanici Berolinensis. Berlin. Witmort, A. J. (1939). Nomenclature of Two British Alchemillas. J. Bot., Lond., 77, 249, Watsonia 5 (5), 1963. CIRCAEA IN THE BRITISH ISLES By PETER H. RAVEN Division of Systematic Biology, Stanford University, California, U.S.A. Although the three European taxa of Circaea — C. alpina L., C. x intermedia Ehrh. and C. lutetiana L. — have in general been well understood on the Continent, the distinc- tion between C. alpina and C. x intermedia has been a persistent source of confusion and subject of discussion in the British Isles (Baker 1951, p. 306). The result of this confusion has been the classification of a number of individuals of C. intermedia as C. alpina and the consequent obscuring of some remarkable patterns of distribution. It is the aim of this paper to re-examine the British distribution of C. a/pina and its relation to the hybrid origin and spread of C. x intermedia. CHARACTERS DISTINGUISHING THE THREE ENTITIES Although the three taxa of Circaea considered here (Figs. 1-3) are usually quite distinct, there has in the past been some confusion among them, particularly, as noted above, in Britain. Early in the course of the present study, it was discovered that the most useful character separating C. alpina from the others was the nature of its inflorescence. In C. alpina the rachis of the inflorescence does not elongate until after the open flowers have dropped; consequently, all of the open flowers are clustered at the stem apex (Fig. la). In C. /utetiana on the other hand the rachis elongates before the flowers open, and the open flowers are well spaced (Fig. 1g). In C. x intermedia, the spacing of the flowers is similar to that of C. /utetiana, but the inflorescence is somewhat more contracted (Fig. Id). This character is well shown by Ross-Craig (1958, plates 35-37). When its importance was realised, it was found to be correlated with a number of other features, some stressed in earlier descriptions. Many of them are enumerated in the following table on p. 263. Comment is necessary on some of these characters. Bracteoles are found with a very low frequency in C. /utetiana, but I have not seen them in British or Irish specimens. The disc, which is shown in C. /utetiana in Fig. lh, is a prominent, elevated, nectar-secreting ring at the apex of the hypanthium. The flowers of this species are frequently visited by small flies, and, as the anthers are held away from the stigma, self-pollination is probably a relatively exceptional event, although Mr. P. M. Benoit informs me that the plants of this species he tested were self-compatible. In fine weather the anthers of C. /utetiana begin to shed pollen an hour or two after the flowers open. In C. alpina, on the other hand, no disc is present, and consequently no nectar, and the anthers often deposit pollen on their own stigma before the buds open, which doubtless causes a high degree of self- pollination. In fact, it is common for the later flowers of C. alpina not to open; instead, the buds become large and pale and eventually produce fruit cleistogamously. In C. x intermedia the disc is usually present but low and obscure. Evidence of the hybrid nature of C. x intermedia is found in its complete morphological intermediacy between the two parents, as shown by the preceding table. In addition it is absolutely sterile. The anthers of C. x intermedia often fall undehisced, and I have never seen a single well-filled, morphologically normal, pollen grain in a plant of this taxon. Likewise, plants of C. intermedia fail to produce mature fruit. On the other hand, the plants of C. alpina and C. lutetiana generally set full complements of seed, and those that I examined, with the exception noted below, were consistent in having pollen more than 90 per cent fertile. Watsonia 5 (5), 1963. CIRCAEA IN THE BRITISH ISLES 263 Rhizomes Overwintering portion Stolons from lower axils Stem height Stem pubescence Petioles Leaf base Leaf apex Leaf margin Leaf pubescence Inflorescence elongation Bracteoles Pedicels Hypanthium Disc Sepals Petals Petal base Filaments Pollen Fruits Fruit hairs lutetiana thick entire rhizome absent 15-60 cm dense, appressed below | hairy all round truncate or slightly cordate acuminate mostly remotely denticulate strigulose along veins and margins before flowers fall absent densely glandular-pubescent 1-1-2 mm long, about equal to ovary in length conspicuous, dark, 0-2-0:4 mm high pale green; densely glandular-pubescent 2-4 xX 2:2-5 mm, deeply notched truncate to rounded 2:5—5:5 mm long fertile 3=4) << 2-25 mim, bilocular 0:7-1:1 mm long, dense < intermedia alpina intermediate most of rhizome present 10-45 cm more sparse hairy above, | subglabrous below shallowly cordate abruptly acuminate | dentate subglabrous before flowers fall present, setaceous sparsely glandular-pubescent 0:5—1-:2 mm long, shorter than ovary shallow, rarely to | 0-2 mm high | whitish; sparsely glandular-pubescent 1-8-4 x 2-3-5 mm, deeply notched rounded to cuneate 2-5 mm long sterile | Up to 25 << 1-2 mim, falling immature 0:5—0:6 mm long, dense slender, forming terminal tubers in autumn only tubers present 5-30 cm absent below glabrous cordate subacuminate or acute deeply dentate glabrous after flowers fall present, setaceous glabrous 0-1-0-2 mm long, much shorter than ovary absent white; glabrous 0-6-1:4 x 0-4-0-9 mm, often shallowly notched cuneate 1-1-5 mm long fertile 2 4 AOD DE Gk is 4‘ DOoe Ch onoonex bD SSCOOSCESSH OSL i SP SES 24 Es aims DES SOSSOVOSEDS 9 8 OBS FON ES O Ges "ON ON-Y: o> ze he Fig. 3. Distribution of Circaea lutetiana L. in the British Isles. Watsonia 5 (5), 1963. CIRCAEA IN THE BRITISH ISLES 267 0 1 2 3 4 | Fig. 4 Distribution of Circaea intermedia Ehrh. in the British Isles. Solid dots, 1930 onwards; open circles, before 1930; crosses, introductions. with scattered localities in Wales and a single station in mainland Scotland (Westerness). Since the restriction of range proposed here for C. a/pina is so striking, I will amplify my sources in greater detail. All of the British material labelled C. alpina or C. x intermedia in the following institutions has been examined : Department of Botany, The University, Aberdeen (ABD): Department of Botany, The University, Birmingham (BIRM); Botany School, University of Cambridge (CGE); Department of Botany, National Museum of Wales (NMW): National Museum of Ireland (DBN); School of Botany, Trinity College, Dublin (TCD); The Royal Botanic Garden, Edinburgh (E); Department of Botany, University of Glasgow (GL); Royal Botanic Gardens, Kew (K), including herb. Watson; Department of Botany, British Museum (Natural History) (BM); The Manchester Museum, The University (MANCH); Druce Herbarium, Department of Botany, University of Oxford (OXF): Botanical Department, The University, St. Andrews (STA). In addition, material from the private herbaria of J. E. Lousley, R. H. Roberts, N. Y. Sandwith and E. C. Wallace has been examined by kind permission. The resulting information has been used in drawing up the map of the distribution of C. alpina (Fig. 8). As now understood, the species is confined to only 10 vice-counties, all in Great Britain and Arran, instead of the 45 and 11 Irish vice-counties mentioned by Watsonia 5 (5), 1963. 268 PETER H. RAVEN atk pe 0 Fig. 5. Distribution of Circaea alpina L. in the British Isles. Solid dots, 1930 onwards; open circles, before 1930. Note the isolated Westerness record in grid square 27/07. Clapham (1952, p. 614). The following list includes all British material of C. alpina that has been examined : V.c. 41, Glamorgan: near a waterfall, west cliffs of the Rhondda Valley, 31 July 1890, Ley (BIRM; BM). V.c. 42, Brecon: Craig-y-Cilau, 23 July 1952, Guile (NMW). V.c. 48, Merioneth: Cliffs, Llaethnant, Pennant Dyfi, Llanymawddwy, 16 July 1955, Condry & Richards (NMW; herb. Sandwith); same, 6 July 1961, Raven & Condry 16294 (BM). V.c. 49, Caernarvon: wooded bank opposite the Dolbadarn Inn (shores of Llyn Peris), 3 July 1832, - (CGE); Cobdew Bridge to Capel Curig Lakes, Coed Bryn Engan, 22 July 1948, E. Roberts 354 (BM); woods above Bryn Engan, Capel Curig area, R. H. Roberts (herb. R. H. Roberts). V.c. 60, W. Lancs.: gravel margin of forest in oak wood between Wray Castle and Perry, Windermere, 10 Aug. 1920, Adamson (BM); gravel stream bank in oak wood, west side of Windermere, 5 Aug. 1920, Adamson (BM). V.c. 65, N. W. Yorks.: above Rawthey Bridge near Sedbergh, 1938, Sledge (BM; TCD; herb. Lousley; herb. Wallace). V.c. 69, Westm.: Glencoyne Dale, Ullswater, 30 July 1937, Lousley (BM; K; NMW;; herb. Lousley; herb. Wallace); Glencoyne Wood, Ullswater, in large patches on stabilised scree in a good many places throughout the wood, 29 Aug. 1955, Hervey H1828/1955 (K); same, 3 July 1961, Raven 16222 (BM); same, Watsonia 5 (5), 1963. CIRCAEA IN THE BRITISH ISLES 269 24 July 1937, Carter (BM; K); Ullswater, 12 June 1893, Clarke 47531 (K); Stybarrow Crag, Ullswater, 19 Aug. 1880, Slater (CGE); Dungeon Ghyll, Aug. 1870, Baker (BM); abundant in Rough Crag, Riggindale, 30 July 1910, Ley (BIRM); rocks of Branstree, above Mardale Green, Aug. 1857, Watson (K-herb. Watson: TCD); Mardale, 1866, Watson (CGE); Swindale, Sept. 1856, Watson (K,herb. Watson). V.c. 70, Cumb.: steep shady bank above Glencoyne Farm, Glencoyne Dale, Ullswater, 1937, Lousley (herb. Lousley); Thirlmere, June 1895, Tennant (CGE): west edge of Thirlmere, 29 Aug. 1881, Ley (BIRM); damp rocks, Fisher Ghyll between Armboth Fell and Thirlmere, July 1904, Waterfall (K, MANCH); Armboth Fell, July 1904, Mason (OXF); Lodore Wood, 20 July 1884, Parsons (K); Lodore Waterfall, Derwentwater, 1806, Turner (K); Derwentwater, 28 June 1842, Wright (BM); Troutdale, July 1904, Waterfall (K); in a moist place in Johnny’s Wood, near Rosthwaite, Borrowdale, 8 Aug. 1946, Whellan (herb. Lousley; herb. Sandwith); Borrowdale, 31 July 1863, Hind (TCD); Borrowdale, Glaramara, at bottom of wet gully on north face in very deep shade under large boulder, 1200 feet, 14 June 1953, Cannon & Herbert 2193 (BM); Glaramara, amongst dry overhanging boulders, 1300 feet, 7 July 1951, Park 215 (BM); rocks above Seathwaite, 2000 feet, in wet Sphagnum, Aug. 1879,-(OXF); mountains, Keswick, July 1860, Ward (MANCH) ; plentiful about Stonethwaite, near Keswick, 13 July 1947, Gerrans 73 (BM); Grange Fell, 7 July 1891, Wolley-Dod (BM). V.c. 97, Westerness: steep slope of ravine, Coine, 30 June 1957, McCallum Webster 1012 (K). V.c. 100, Arran, among bracken, Benlester Glen, 300 feet, 3 Aug. 1937, Mackechnie (BM; herb. Wallace); same, 1000 feet, 8 Aug. 1937, Mackechnie (herb. Wallace); same, 1000 feet, 10 Aug. 1933, under Preris (herb. Lousley; herb. Wallace); woodland, Monamore Glen, 300 feet, 23 July 1937, Mackechnie (BM; herb. Lousley herb. Wallace); Glen Cloy, Aug. 1872, Craig-Christie 502 (ABD); Lamlash, July 18990, Wilkie (GL); same, Aug. 1928, Mackechnie (herb. Wallace); Arran, 29 Aug. 1888, Thompson (BM). Not assigned to vice-county: ‘ North Wales,’ Weiss (MANCH). One of the more interesting facts that emerges from these maps and the listing of specimens is that C. alpina does not have a particularly northern or montane distribution in the British Isles. Rather, while conforming rather closely to the sort of distribution expected of a member of the “‘ Continental Northern ”’ pattern as outlined by Matthews (1955), it occupies a highly fragmented range along the western side of Great Britain. Its extreme rarity on the mainland of Scotland and apparent absence from the Hebrides and from Ireland were entirely unexpected, the more so since C. x intermedia is common over much of this area. Of equal interest is the occurrence of C. a/pina in the Rhondda Valley at a locality which is south of the present range of C. x intermedia; it would be of importance to study this site ecologically if it could be re-located. A few examples may be given of other species which have ranges in Britain similar to that of C. alpina (kindly suggested by Dr. F. H. Perring), but it should be noted that none of them is of the same geographical group when total range is considered. These are Carum verticillatum (W. Europe), Vicia orobus (W. Europe), Centaurium littorale (coasts of W. Europe and inland from Austria to S. Russia) and Rhynchosinapis monensis (endemic). ORIGIN AND SPREAD OF CIRCAEA X INTERMEDIA On the continent of Europe, C. x intermedia occurs mostly in or near the zone of contact between C. alpina and C. lutetiana. In Scandinavia, for example, C. alpina is widespread, C. /utetiana is present only in the south, and C. intermedia is nearly re- stricted to the area where its two parents overlap (Hultén 1950, maps 1293-1295). As can be seen from an inspection of Figs. 3, 4 and 5, however, the situation is quite different in Britain. Circaea alpina doubtless had a wider distribution in the British Isles in cooler times nearer to the last glacial period, but today it is restricted to relatively few localities. On the other hand, C. /utetiana is a woodland species that must have mostly expanded its area since the last glacial period. The morphological intermediacy and complete sterility of C. x intermedia, taken together with the lack of other possible parents, point to the fact that it must be regarded as being of hybrid origin between C. alpina and C. /utetiana. Despite this, C. alpina is now absent from many areas where C. x intermedia is common, such as Ireland, the Hebrides and much of mainland Scotland. Moreover, C. /utetiana also is absent from some of these areas, particularly northern Scotland and Orkney. Watsonia 5 (5), 1963. 270 PETER H. RAVEN These apparent discrepancies may be explained in two ways. Either C. x intermedia has achieved its present wide distribution in the British Isles by means of its remarkably vigorous vegetative reproduction, or its distribution reflects the one-time area of C. alpina where it came into contact with C. /utetiana, or both factors may be important. Correlated with these considerations is the question as to whether C.xintermedia in Britain and Ireland originated from a single hybridisation, or whether it is produced with fair frequency whenever its parents come into contact. The latter question may be considered first. Arguing for the fairly frequent production of the hybrid is the sort of situation on which I made observations in Glencoyne Wood, Ullswater, Westmorland, on the 3 July 1961. Here all three species grew in a moist shaded area of woodland dominated by bracken and species of Carex and Juncus. Circaea alpina occurred in two small patches scarcely 3m across, whereas C. /utetiana was more widespread both on this flat and in the neigh- bouring woods. The individuals of C. intermedia formed a band approximately 2 m across, ringing the small patches of C. alpina, and they were in turn surrounded by the abundant C. /utetiana. It is extremely probable that these hybrids were produced in situ. Furthermore, C. x intermedia, as has been remarked by many writers of floras, is very variable in habit, pubescence, flower size, leaf margin and other characteristics. This sort of variability would be most unexpected if all the individuals concerned had been derived by vegetative reproduction following a single hybridisation. Finally, the range of C. x intermedia, from Britain to the Caucasus, argues against such an hypothesis. On the other hand, it would be a serious mistake to underestimate the capacity of C. x inter- media for vegetative spread, particularly when it 1s remembered that these plants are often serious garden weeds where they occur. Vegetative reproduction seems to have been important in the spread of certain other well-known plants that occur in gardens, such as Aegopodium podagraria, species of Calystegia and Tussilago farfara; Circaea x intermedia has obviously attained part of its present range in the same way. A particularly suggestive example of this, pointed out to me by Professor Webb, is its local occurrence in Co. Wicklow, Ireland, well south of its main area of distribution; here it is exclusively a garden weed and very probably introduced. Also in Merioneth, for example, it is mainly restricted to disturbed areas and man-made habitats, particularly in gardens and along roadsides, and is probably extending its range in connection with cultivation (Benoit & Richards 1961). Thus, in summary, it does not appear likely that all of the present distribution of C. x intermedia in Britain and Ireland can be explained by its admittedly vigorous vegeta- tive reproduction. A more probable hypothesis is that C. alpina was more widespread nearer the last glacial maximum. As it contracted from this wider range to the few stations it occupies at the present time, it was often in contact with C. /utetiana. Spreading as it must have done with the increase in wooded areas, C. /utetiana may eventually have come to occupy an area even more extensive than at present during the warm, dry period approx- imately 5,000 years ago (Matthews 1955). At the height of this warm, dry period, C. alpina may have been even more restricted in range than it is today, judging from the western, oceanic nature of its present range in Britain. As their ranges shifted, hybridisation between C. alpina and C. lutetiana probably occurred fairly frequently, producing numerous biotypes of C. x intermedia with intermediate morphology and ecological requirements. From these centres, the sterile C. x intermedia spread by means of its efficient vegetative reproduction and, in some instances, as a ruderal aided by man. It is important to note that the capacity for vegetative spread was already present in one of the parents, C. /utetiana, as gardeners will attest, and that it was this pre-determined capacity that allowed the sterile C. x intermedia to become established and spread; it is evidently even more vigorous in its spread than C. /utetiana. Progressive changes in climate finally led to the apparent extinction of C. alpina in Ireland and over much of Scotland and allowed its survival only in those few, particularly favourable localities where it persists to the present day. Its apparent absence from Ireland is particularly interesting in view of its present western oceanic distribution in Britain, which suggests that conditions in at least some Watsonia 5 (5), 1963. CIRCAEA IN THE BRITISH ISLES 271 portions of Ireland may be suitable for it at the present time. It is possible that the shift to a cooler, more oceanic climate in the past 5,000 years has caused the restriction of C. /utetiana in the north and perhaps even allowed C. a/pina to spread slightly in the Lake District and on Arran. Meanwhile, C. x intermedia, apparently well suited to the climatic conditions prevalent in northern Ireland, Scotland, northwest England, and portions of Wales, has flourished in those areas and become locally the commonest or even the only representative of the genus present in some of them, despite its complete dependence on vegetative propagation. Thus Circaea x intermedia presents a truly remarkable example of a sterile hybrid better suited to a particular set of conditions than either parent, and consequently replacing them in certain areas. Two similar examples from the British flora that may be cited are Nuphar x intermedia, the apparently highly sterile diploid hybrid between JN. /utea and N. pumila (Heslop-Harrison 1953) and Rorippa ~_ Sterilis (R. nasturtium-aquaticum x R. microphylla), a sterile triploid which occurs locally in the absence of both parents in much the same areas as Circaea < intermedia. Such patterns of variability are characterized by the ability of hybridisation to produce a large number of relatively uniform genotypes suited to a particular environment correlated with an apparent loss in evolutionary flexibility. Similar patterns occur frequently in the angiosperms but they are usually a product of the spread of fertile allopolyploids rather than of sterile, vegetatively propagated hybrids. ACKNOWLEDGMENTS The research leading to this paper was completed during the tenure of a United States National Science Foundation post-doctoral fellowship spent in London in 1960-61. Iam grateful to Mr. J. E. Dandy, Keeper, Department of Botany, British Museum (Natural History), and to Sir George Taylor, Director, Royal Botanic Gardens, Kew, for the facilities generously extended to me during my stay. I would also like to acknowledge specially the many helpful comments of Mr. Peter M. Benoit, Barmouth, whose perceptive observa- tions on the flora of Merioneth are well known. Thanks are also due Mr. D. E. Allen, Rondon. Dr nH. G. Baker, University of California, Berkeley; Mr. A. ©. Chater, Aberystwyth; Mr. W. M. Condry, Machynlleth; Dr. F. H. Perring, Cambridge; Mr. Evan Roberts, Capel Curig; Mr. R. H. Roberts, Bangor; Mr. N. Y. Sandwith, Kew; and Prof. D. A. Webb, Dublin, who aided this study in various ways; to those in charge of the herbaria listed above, for permission to study the material entrusted to their care; and to Dr. C. R. Metcalfe and Dr. K. Jones of the Jodrell Laboratory, Kew, for affording me the facilities to make the cytological studies reported herein. Finally, I would like to thank Miss Gretel W. Dalby, who prepared the illustrations of the plants, and to acknowledge the kind permission of the Distribution Maps Scheme, Botanical Society of the British Isles, for the maps of the three species of Circaea to be reprinted here. REFERENCES BAKER, H. G. (1951). Hybridization and natural gene-flow between higher plants. Biol. Rev. 26, 302-337. Benoit, P. M. & RicHArDs, M. (1961). Circaea, in A contribution to a flora of Merioneth. Nature in Wales, Us Se BuTcHeErR, R. W. (1961). A New Illustrated British Flora, Part I. London. CLAPHAM, A. R. (1952). Onagraceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F. Flora of the British Isles, Ed. 1, 596-614. Cambridge. Ed. 2 (1962). CLAPHAM, A. R., TuTin, T. G. & WARBURG E. F. (1960). Flora of the British Isles. Ulustrations, part II. Drawings by S. J. Roles. Cambridge. Fitcu, W. H. & SmitH, W. G. (1924). Illustrations of the British Flora. Ashford. Watsonia 5 (5), 1963. Pie, PETER H. RAVEN HESLOP-HARRISON, Y. 1953). Nuphar intermedia Ledeb., a presumed relict hybrid, in Britain. Watsonia 3, 7-25. HULTEN, E. (1950). Atlas of the Distribution of Vascular Plants in NW. Europe. Stockholm. Love, A. & D. (1961). Chromosome numbers of central and northwest European plant species. Opera Botanica, Lund 5, 1-581. MattuHews, J. R. (1955). Origin and Distribution of the British Flora. London. Ross-Craic, S. (1958). Drawings of British Plants, Part 11. Droseraceae — Ficoideae. London. Upp.itnc, A. (1929). Die Chromosomenzahlen von drei Circaea-Arten. Hereditas 12, 294-296. Watsonia 5 (5), 1963. VARIATION OF SOME DIAGNOSTIC CHARACTERS OF THE SESSILE AND PEDUNCULATE OAKS AND THEIR HYBRIDS IN SCOTLAND By oly 2) COUSENS Forestry Department, University of Edinburgh ABSTRACT A collection of Scottish oaks including nearly 1,000 fertile specimens has been analysed in respect of selected diagnostic characters. Pictorialized scatter-diagrams confirm that only two taxa are involved and that considerably more than half the fertile material cannot be diagnosed with confidence as belonging to either. The variation patterns within population samples are so exactly what would be expected from widespread and massive introgression that the objections to this interpretation are called into question. All the robur-dominated woods appear to be quite strongly introgressed but there are a few petraea woods which are relatively homogeneous and show no evidence of introgression. The latter are used to demonstrate the variation that may be expected from reasonably good petraea in Scotland. INTRODUCTION Q. robur L. and Q. petraea (Matt.) Liebl. have been described in varying detail many times since they were first distinguished. Jones (1959) has recently summarised the more important British and Continental contributions to our knowledge of the two species. The list below of some of the more reliable diagnostic characters is extracted from his account. Character robur petraea Leaves of spring shoots Shape obovate | ovate of the crown Lobing deep, irregular, 3-5(-6) | shallow, regular, 5—6(—8) pairs pairs Petiole short 2-3-7 mm long 13-25 mm Leaf base cordate with strong cordate to cuneate, weak Acorns Abaxial surface Colour Longitudinal stripes auricles glabrous, occasional simple hairs pale fawn olive green on fresh mature acorns auricles always some stellate hairs uniform dark brown absent Fruiting peduncle Total length 2-9 cm 0-3(—4) cm Thickness slender stout Pubescence glabrous some pubescence Buds Size small | large Apex obtuse acute oe Sona Two of these characters are reputedly qualitative : the striations on the maturing acorn of robur and the abaxial stellate pubescence on the leaf of petraea. The first is of little practical use in the field unless collecting is confined to the critical periods of good Watsonia 5 (5), 1963. PAS 274 JE. ‘COUSENS fruiting years. The second is disputed : Jones follows Moss (1914) in accepting abaxial stellate hairs as a petraea character, but several Continental workers have described aff. robur forms as having them (e.g. Weimarck 1947). Of the remaining quantitative characters only petiole length has ranges for robur and petraea which do not overlap. Scottish collec- tions however, do not conform to the specifications above; they show overlap in this as well. Confident identifications of oaks in Scotland can therefore be made only when the individual exhibits most of the diagnostic characters in the extreme forms associated with one of the two species. A majority of individuals have to be classified as intermediate either because intermediate values for diagnostic characters predominate or because extreme robur and petraea characters occur on the same tree. Many Scottish woods have a very high proportion of such intermediate forms and for many years it was commonly assumed that they were of hybrid origin (e.g. Burtt Davy 1933, Tansley 1940). Opinion changed as the results of work by Pyatnitski and Dengler (refs. in Jones 1959) became known; they had demonstrated experimentally that the two species when crossed set on the average only 2°% fertile seed. Jones expresses the new views when he defines the species more broadly than before and claims that hybrid forms, even in woods where both species are present, are unlikely to exceed 5%. While I agree that Jones’ descriptions of the two species are a great improvement on former ones, they do not seem to clarify the position in Scotland (cf. attempts to apply earlier descriptions in Scotland by Greville (1841)). The field botanist has neglected the oaks. The reasons are probably the obvious ones, the difficulty of collecting comparable material from mature trees and the long history of planting of these economically important species. If the field botanist has to accept the possibility of planting he can no longer place much ecological significance on the presence or absence of one or other species in a particular site. Anderson (unpublished) has studied the history of oak planting in Scotland in great detail. There appear to be old planting records in all the areas where oak woodland occurs today and there are many records in what now seem the most unlikely places by reason of their remoteness or the roughness of the terrain. Tansley (1939) evades the problems of planting by classifying the older oak woodlands as semi-natural and saying that a plantation on an oak site would eventually become indistinguishable from a natural wood; but he makes no stipulations about the provenance of the planting stock used to form the plantation. There is considerable evidence, discussed by Jones (ibid. pp. 175 and 215), to show that robur was consistently preferred for planting irrespective of the species originally occupying the site. A majority of the older oakwoods remaining in Scotland include many stems which have patently developed from coppice stools. Some areas, as in the Trossachs, are known to have been under intensive coppice management (or coppice with standards) for long periods. Between 1600 A.D. and 1900 A.D. coppice management spread throughout Scotland, at first for rural produce and later for tan-bark production and charcoal for smelting. The woodlands least intensively managed are likely to have been those least accessible from the sea and the main land-routes; the north-shore oakwoods of some of the inland lochs of the north-west may come in this category. The youngest coppice woods date from the First World War but many still obviously coppice woods are more than twice that age. Generally few standards remain though large stumps often show that they were present. On the better sites stools have usually been singled but very few stems show conspicuously good form. Woods derived from coppice have an unnaturally high stocking of oak; other woody species (except hazel) have presumably been eliminated deliberately. High oak stocking could conceivably have been attained with natural regenera- tion alone, but it seems more likely that planting was frequently necessary. OBJECTS The original objective, which led to the work described in this paper, was the location of oakwoods of indigenous Scottish stock. The immediate objective was a practical defini- Watsonia 5 (5), 1963. VARIATION OF OAKS IN SCOTLAND ZV) tion of robur and petraea in terms of suitable diagnostic characters. As will be seen, the work carried out to this end has gone a long way towards the next objective — to discover whether there were distinctive variation patterns within oak populations which could be used to assess their status. COLLECTING METHODS AND CHOICE OF DIAGNOSTIC CHARACTERS Variation on the tree is considerable and if comparisons between trees are to be meaningful an attempt must be made to minimise variation from this source. A preliminary study was therefore made of variation patterns on the tree and through the growing season with the aim of determining the most reliable diagnostic characters and the best general collecting procedure. While the results of this study suggest that there may be significant differences in the variation patterns of robur and petraea it is the resemblances that are more important at this stage before the species have been defined. The pattern of leaf variation on the sun-shoot, for example, is remarkably consistent. There is a gradation from small basal leaves which are rather irregular, comparatively broad with deeper lobes, more ovate and have a more cordate base, through larger mid-shoot leaves to the apical leaves which are also smaller than average but are more regular, narrower with shallower lobes, more obovate and have a less cordate (or even cuneate) base. Mid-shoot leaves are most re- presentative and also show the best auricle development. On horizontal shoots leaves arising from the under side are larger. A trial was made using two mid-shoot leaves for biometric measurements. A comparison of average figures for all leaves with those obtained from pairs of mid-shoot leaves can be made in Columns A and B of Table 1 below. TABLE | Variation of leaf-length and petiole-length with position on the tree | | Leaf-length _ Petiole-length | Petiole % Tree No. Part sampled (mm) | (mm) | | | one Bees: A B | A B 1. petraea _ Lower crown—W 80-2 SPOR ala O24 10:3 11-7 2 (isolated) | Lower crown—N | 82:1 91-9 9-2 OZ -IHilew flea | Lower crown—E ——:76°3 88:4 Be) 12a es 13-0 13°8 _ Lower crown—S 74-7 86-7 12:3 14 Ae 6:5 16:6 | | | 2. Hybrid _ Lower crown—W 90:4 a Ofelia) a an6:3 eS GRO) 7-0 (robur affinity) | Lower crown—N 89-4 100:0 6:7 TP 75 71 | Lower crown—E | 97:0 1063 8-0 oS | Bx 8:0 Lower crown—S | 96:3 104-4 |. 6-9 icles) 12 6°8 3. robur Lower crown—N = 72°8 86:2 3-9 49 | 5:4 527] (isolated) Lower crown—S 86:3 102-9 4-6 G25) 5:3 6:0 | | | 4. robur _ Top of crown | UES | 2:6 | 3°3 (NW ofa Upper crown—SSE | Tica | NG | 3-6 large gap) Lower crown—SSE_ 85:0 | 2A 2:8 Inside crown | | (shade shoots) LOLSen | 3:51 71 3-4 _ Epicormic shoots | 99:38 | 3:0) | 3-0 | | | | *A — Averages for all leaves on 5 representative terminal shoots (sun-shoots except as indicated). B — Averages for two mid-shoot leaves from each of the shoots (ten leaves), Watsonia 5 (5), 1963. 276 J. E. COUSENS These data show that there is marked variation with position on the tree; the collecting position should thus be standardised. The south-east aspect was eventually chosen because it appeared that the difference in petiole-length for robur and petraea would then be greatest. The figures also suggest a relationship between leaf dimensions and insolation or exposure. For a study of variation through the growing season 12 trees were selected in Dalkeith Old Wood. This wood includes both robur and petraea with many intermediate types (see Table 3 below). Collections were made at the same point (SE of the accessible part of the crown) at increasing intervals from the time the buds began to swell. The pertinent observations from these data were : (a) abaxial stellate pubescence remained substantially the same from the time the leaf was fully expanded till it fell in autumn even though scored from different leaves at each collection, (b) buds of all specimens were small and obtuse at first - only towards autumn were robur and petraea types readily distinguishable, (c) flowering peduncles reached their maximum length about mid-June but continued to increase in diameter for several weeks, (d) after July the peduncle beyond the most distal developing acorn shrivelled and sometimes fell off; it follows that total length of fruiting peduncle must be an unreliable character for comparative purposes, (e) during the same period the peduncle became more woody at the base and de- creased in diameter roughly to the same extent as peduncles collected in July (after drying). (f) peduncles of all specimens were densely pubescent at first; aff. robur forms were practically glabrous by mid-June; on aff. petraea forms the pubescence persisted at least on those parts protected from rubbing, (g) peduncle length, particularly on aff. robur forms, could show considerable variation on the tree and sometimes on the shoot. The period available each year for collecting ran from mid-June till early October. It was desirable therefore to select characters which would be comparable throughout this period. Leaf characters were suitable but buds and the acorn with its fruiting peduncle were not. In view of the importance attached to peduncle characters in diagnosis alternative biometric measurements were investigated. Length to the first flower locus proved very satisfactory; it does not change as the acorns develop and the flower bract persists even if the fruit does not develop; using this character the proportion of nominally fertile specimens is increased and better separation of typical robur and petraea is obtained. As both peduncle pubescence and peduncle diameter seemed promising diagnostic characters, collecting was planned to start not earlier than mid-July. The following characters were eventually selected for recording: leaf length, petiole length, auricle type, abaxial stellate pubescence, lobe number, lobe depth, length of peduncle (total and to the first bract), peduncle diameter and peduncle pubescence. Field collections were made with a pole-cutter reaching to twenty feet. A small branch was cut and a fertile (if present) terminal sun-shoot selected as typical of such shoots on it. Two representative mid-shoot leaves were detached and pressed and dried separately. After measurement shoots and leaves were mounted and retained for reference. Collection presented no problems in the open or scrub woods which predominate in Scotland. In woods with a closed high canopy samples had to be taken from the nearest gap at paced intervals or along roads and rides running roughly east-west. In a few woods sampling had to be done along the southern margin after checking that these marginal trees were similar to those inside the wood. In practice it was some time before the importance of taking a satisfactory population sample was appreciated. Many of the earlier collections were rather small and some were selective. Collections have now been made in 127 different woods well distributed through- out Scotland (details appear in the Forestry Commission Research Reports for 1961 and Watsonia 5 (5), 1963. VARIATION OF OAKS IN SCOTLAND 277 1962) and 55 of them include sufficient fertile specimens, systematically collected, for complete analysis as population samples on the lines described below. ANALYSIS Of the characters recorded as continuous variables only petiole (as petiole °% of leaf length) and peduncle length (to first bract) provided reasonable separation of typical examples of robur and petraea; leaf length, lobe number and lobe depth all provided minimal separation. Six characters were thus used, the two above (the primary characters) and four secondary characters — auricle type, stellate pubescence (abaxial), peduncle pubescence and peduncle diameter. The latter were first studied in some detail to deter- mine which parts of their ranges were diagnostic of one or other species. Arbitrary limits were set and then each character was considered in turn, but only for those specimens for which all the other characters were positively diagnostic of robur or petraea. Each reappraisal was followed if necessary by reclassification and the process repeated until no further adjustment of limits was required. The definitions arrived at may be seen in Table 4. A specimen showing all four secondary characters within the diagnostic range for one species belongs to what is hereafter called the Theoretical Species Type (or TST). A specimen showing one character in the indeterminate range lies within the apparent normal range of the species but may not belong to it if the difference is actually due to hybridisation. The four characters each with a robur, indeterminate and petraea range provide 81 (3) possible combinations which can be classified according to their differences from either TST. TABLE 2 Classification of Secondary Character Combinations Secondary character combination classes Character I I Ill IV Vv VI VII Vill 1X Combinations | Theoretical Degrees of difference Theoretical | petraea 1 2 3 + robur ' combination 4 a 5) 1 combination Possible | | 4 10 16 19 16 10 4 1 Actual 1 4 10 13 18 14 10 4 The number of combinations which can appear in a population sample is determined by the variability of the population sampled and the sample size. However, quite a small sample (15-20 specimens) is likely to show the range of combination classes present in the population. A general picture of the variability in Scottish oakwoods is conveniently presented in terms of the range and frequency of these combination classes. If obvious plantations are excluded a majority of the woods sampled fall into category A below; aff. petraea forms predominate, while aff. robur forms (Classes VIII & IX) are absent. Intermediate forms (Classes IHI-VII) dominated 15 woods (category B), two of them plantations. Only 8 woods fell in category C (robur woods) and 4 of these were obvious plantations; collectively they included a much higher proportion of intermediate classes than the petraea woods. The mixed woods (category D) totalled 7; they were distinctive in having both aff. robur and aff. petraea forms but without a complete series of linking intermediate classes. One of them (Knockman Wood, Galloway) was a rather poor plantation of robur Watsonia 5 (5), 1963. 278 7. 'E. COUSENS TABLE 3 Representative population samples classified on the range and frequency of their component secondary character combination classes | | | 1 OES: Combination classes — no. of specimens General | Wood | Grid a category | sampled | Square fae | II Jil IV vi Vile oh Vili NAS SIX | eee A. petraea-| Glentrool | NX | 20 6 | 26 dominated | Elgin NJ io a4 3 p) | 9 woods | Spean Bridge | NN 5 6 4 3 Puts | Loch Sunart | NM 4 2D 3 2 2 13 | Ptarmigan 1 ININ 8 8 8 1 D 2: 1 30 B. | Kirkwood NY | per MOE es Bp 15 Dominated Loch Katrine =. NN | p) 3 2 3 4 1 1 16 by inter- | Dalkeith | ee NGI ee ot I 3 9 6 12 9 7 2 50 mediates | | C. robur- | Erchless NH 2 1 2, 1 3 3 2, dominated | Gourdie iNOFy 1 3 3 1 2 10 woods _ Fearnoch | NM 2 3 1 6 D. Mixed | Aikieside | NT 9 2 1 12 woods Tore of Troup; NJ 2 1 2 3 1 2 2 13 | Springkell PRG dia eae: i 5) 5 2 4 20 Totals re 6 31 24 28 18 34 23 D2. 14 | 250 Other population samples | 122103 58 S)// 66 37 58 49 21 yt All population samples | 178 134 82 85 84 7h 81 71 BBE oval affinity bounded by a wall beyond which were a few remnant perraea; the few good petraea in the plantation may well have been natural seedlings. Another (Aikieside, Berwickshire) consisted mainly of good petraea derived from coppice, with scattered aff. robur trees which were probably planted in gaps at the same time as a neighbouring robur plantation was established. The next step was the introduction of the primary characters, used as co-ordinates of scatter diagrams, to discover whether the TSTs themselves were discretely differentiated entities. Fig. 1 shows that they are not; the distribution of values for each is strongly directional with a marked trend towards the concentration centre of values for the other TST and slight overlap. Classes II and VIII showed the same trend more strongly and the overlap was greater. There is in fact a fairly regular gradation right through from Class I (TST petraea) to Class 1X (TST robur) when judged from the mean values for each class (see Fig. 1). The combination classes are, of course, artificial aggregates of individuals drawn from many rather different types of population. It will only be possible to formulate ‘normal’ limits of variation for either species in Scotland if populations can be found which exhibit them. The final test therefore lies in analysis of variation within populations. For this purpose pictorialized scatter-diagrams were used; they allow simultaneous appraisal of all the variables and the population samples do not need to be large or identical in size. The secondary characters were represented by the symbols shown in Table 4 below. Watsonia 5 (5), 1963. VARIATION OF OAKS IN SCOTLAND 279 fe ®@ Individual values for petraea TST Individual values for robur TST Mean values for Combination Classes O/ /O Petiole Peduncle length to first bract (mm) Fig. 1. Scatter diagram of combined petiole % and peduncle length values for all specimens classified as Theoretical Species Types (TSTs). Continuous lines enclose Concentration Centres (CC) for petraea (p) and robur (r). Broken lines outline the limits of variation of each TST. Figures 2-7 show examples covering the range of woodland types. As might be expected from the analyses above, the greater the number of secondary character combinations in the sample, the wider is the scatter. The important observation here is that this principle applies also to the scatter of TST values. Jf the TSTs from the more heterogeneous samples had been excluded in preparing Fig. I then they would have appeared as discretely differentiated entities. The most homogeneous samples from pefraea-dominated woods all include some individuals differing from the TST in one degree (Class If combinations) and all four possible variants occur. There are no equivalent samples from robur-dominated woods. DISCUSSION These findings cannot be reconciled with the view Jones (1959) has expressed about the current status of robur and petraea in Britain. If they are acceptable taxa then these data show that the proportion of hybrids in Scotland is at least 50°%% and probably 75%. Alternatives which have not been seriously countenanced for a long time are (a) that the two taxa are not discrete entities and (b) that a third taxon, intermediate between robur and petraea, 1s involved. In spite of intensive collecting in reputed robur woods and plantations during 1961 the amount of robur material is only just adequate to demonstrate a concentration centre for combined petiole and peduncle values of the robur TST. Never- theless it is fairly clear that two, and only two, taxa are involved in these data. Introgressive hybridization (or simply ‘ introgression’) was defined by Anderson (1949) as a natural process of gene exchange arising from repeated back-crossing between hybrids and the parent species. ‘Introgression’ is used here in the same sense, though the process is unnatural to the extent that it has been unintentionally accelerated by Man’s activities. The variation described in Scottish oak populations is so exactly what would be Watsonia 5 (5), 1963. 280 J. E. COUSENS TABLE 4 The diagnostic ranges for secondary characters with symbols* Character Diagnosis | Symbol 1. ABAXIAL STELLATE PUBESCENCE | Both types abundant petraca | @ Recorded separately for the small pro- | : | a ee ; P y é P | One type absent or indeterminate | @ strate trichomes on the lamina and | | : 2 ly (One. Or both sparse | large erect trichomes beside the midrib. | | Both types absent robur | ee 2. AURICLE TYPE _ Lobes weak (or nil); petraea | ce, Based on (a) the development of the lamina not sharply | basal lobes which in the extreme robur reflexed form overlap the petiole and (b) the : | P P ; >) Medium lobes not indeterminate | oy sharpness of the reflexion of the | ws reaching the petiole; | lami yhere it joins th tiol , lamina where it joins the petiole iemninny Sharply eefiewcd | Lobes well developed robur | S. reaching the petiole on at least one leaf; lamina sharply reflexed 3. PEDUNCLE DIAMETER Stout petraea | 7) Measured by gauge (cut to | mm, 2mm diam. or over | 1:5 mm and 2mm) at the narrowest : | s Intermediate indeterminate | a point below first bract | | = c. 1-5 mm | Slender robur | — 1 mm diam. or less 4. PEDUNCLE PUBESCENCE Very pubescent or petraea ct Caducous pubescence lost by about pubescent at least on mid-June. Persistent pubescence tend- protected parts of ing to rub off on all except very short | peduncle eduncles : a Trace of pubescence indeterminate _ “Ny Glabrous robur Theoretical Species Combination petraea TST XH Theoretical Species Combination iee AGsat b expected of two mutually introgressing species that conflicting evidence should be critically examined. The fertility barrier between the species demonstrated by Pyatnitski and Dengler is the only serious problem. This may be either not as important or not as generally applicable as workers in this field have assumed. Experimental crosses need to be made in other parts of the range of the two species and, possibly more important still, the success of back-crosses needs investigation. * For a more detailed description of these characters with illustrations of auricle types see Cousens (1962) Watsonia 5 (5), 1963. VARIATION OF OAKS IN SCOTLAND 281 AO , Z Cg: a abd € Oe, ae -& we Sa Fig. 2, GLENTROOL Fig. 3. EDGERSTONE Mature oak woodland. Reputed remnant of old Jedforest. or 20° 50 40.5 ZORU UNO RAO SON 204 Fig. 4. PTARMIGAN Fig. 5. KIRKWOOD Coppice regrowth, Policy wood probably IS Boer. c. 120 years old. planted c. 150 years ago. ei oe Ot OF PHS O & ere 2 Xn Oe & -@ Ke) = Oo e O- oO- oan) x & (e) 2 | & eo o= ow 6 Se H o Vice es 20 30. «40 = 50 20 30 40 50 20 Figs 65 “GOURDIE Fig. 7. KNOCKMAN WOOD Policy wood. Walled plantation on old 15 sessile oak site. ~~ 10+ i=) 2 oa ae g b O- es ns SS 8= 0 10 20 30 40 50 10 20 30 40 50 60 Peduncle length to first bract (mm) Peduncle length to first bract (mm) Figs. 2-7. Scatter diagrams for six oak populations. Concentration Centres for petraea (from Fig. 10) and robur (from Fig. 1) are shown by dotted lines. As Stebbins (1950) has pointed out oaks are long-lived and form more or less closed communities : interspecific hybrids becoming established in such communities will tend to be isolated among many individuals of the parent species and back-crosses rather than Watsonia 5 (5), 1963. 282 J. E. COUSENS interhyorid crosses will be the rule: furthermore one hybrid is potentially capable of producing an enormous number of back-cross progeny during its lifetime. The interspecific fertility barrier may be relatively unimportant if the hybrids are fully fertile with either parent. Numerous examples of introgession in American oakwoods have already been cited (see Tucker & Muller 1958, Cottam, Tucker & Drobnik 1959, Tucker 1960 and Tucker 1961 for some recent examples). The historical facts point to widespread and extensive planting activity with a strong preference for robur planting stock. Q. robur must have been introduced in this way into many areas where petraea dominated the natural woodland and the chances of the cross robur 2° x petraea 3 appearing must have been very high. Collection of acorns from these robur nuclei would have given the hybrids very favourable chances of becoming established as crop trees in new plantings, particularly if a proportion of them exhibited hybrid vigour (refs. in Dengler 1941). Reasons will be put forward below for believing that the Fl hybrid bears a strong superficial resemblance to robur; collection of acorns from such hybrids may frequently have been made in the mistaken belief that they were good robur. Petiole of 30 40 50 60 70 80 90 Peduncle length to first bract (mm) Fig. 8. Specimens with abnormally long peduncles. The only suggestion of hybrid vigour in these data lies in eleven values for peduncle length to first bract beyond the range of the robur TST. They include 10 different secondary character combinations and every class from II to VIII is represented (Fig. 8). The numbers of specimens recorded in the intermediate combination classes (III to VII) happen to be of similar magnitude (see Table 3). It seemed that analysis of the frequency with which robur, petraea and indeterminate characters appeared among these putative TABLE 5 Frequency (%)' of sccoudaly characters in intermediate combinations | | | robur | 25 | 29 | Auricle | Abaxial | Peduncle | Peduncle ; | | | or, Character type pubescence | diamete) | pubescence RAG | | petraea 23 | 40 | 19 | 48 Indeterminate 52 | 31 | 28 | 36 | 53 | 16 Watsonia 5 (5), 1963. VARIATION OF OAKS IN SCOTLAND 283 0/ /O Petiole 10 20 30 40 50 Peduncle length to first bract (mm) Fig. 9. All values for @~< ?anF.1. hybrid hybrids might yield significant results. Only auricle type shows a maximum in the inde- terminate range, the expectation if the character is under polygene control. The other three characters appear to exhibit partial dominance for petraea pubescence of leaf and peduncle and robur slenderness of peduncle. If these conclusions are valid the interspecific hybrid will consist mainly of forms with medium auricles, well-developed abaxial pubescence and slender pubescent peduncles. This combination is in fact the most frequent one in these combination classes and provides the only example of an intermediate combination dominat- ing a population sample (viz. Kirkwood — see also Table 3 and Fig. 5). Figure 9 is a scatter-diagram for all specimens showing this combination : it will be seen that they show extremely wide scatter in the zone between the concentration centres for the TSTs. This combination is also found in a form described as Q. robur subsp. puberula (Lasch) Weim. by Weimarck (1947) who noted several ‘ aberrant forms’ —a f. brevipedunculata and a f. petiolaris (suggesting that petiole-peduncle values for the sub species would also show very wide scatter) and a f. /ongipedunculata (for which an equivalent did not occur in these data though it might correspond to the forms tentatively suggested above as due to hybrid vigour). QO. robur subsp. puberula is reported from Norway (Risdal 1955) and is said to become more and more frequent eastwards through Scandinavia; it is also reported as fully fertile (Weimarck 1947). It is interesting to note that Hoeg (1929) described the most easily recognisable interspecific hybrid as resembling robur but with abaxial stellate pubescence on the leaves. On the question of hybrid fertility the data offer a little evidence through a comparison of fertile and non-fertile specimens in certain collections. All are from petraea-dominated woods (category A of Table 3). The above collections were all made in 1961 when fertile material, particularly of aff. robur forms, was less plentiful than in the unusually good fruiting years 1959 and 1960. The proportion of possible hybrids in the collections rises from 42% (22/52) to 57% (49/86) if infertile material is included. Consideration of fertile material alone will lead to an underestimate of hybridity. It is strange that none of the samples from plantations proved to consist of good robur forms, for many of them must have been established with stock raised from the many large consignments of acorns known to have been imported from England and the Continent. Could it be that all importations of acorns have included a high proportion of introgressed stock ? Or was the late Professor Anderson correct in his contention that robur is not native in Scotland and that hybrids thrive better here than good robur ? Maxwell (1900) Watsonia 5 (5), 1963. 284 , J. E. COUSENS TABLE 6 Percentage fertility by combination classes of auricle type and abaxial pubescence Ptarmigan Ross Wood | Loch Ard Combined | | Total Fertile | Total Fertile | Total Fertile Total Bak ae | D. mS Geracae | alo 16) WS ile ah 5 37 : 81 i ecient cm 1 difference Le 2 Once 3) i | 8 5 23 52 2 difference ent 2 6 2 ce 2 1 16 44 3 difference Mecita Pia Nah es oe a 1 8 37 4 difference ee ye oe | 1 ok 1 eS ee 2 Totals | 50 30 19 Oo La 12a E86 S20 alee claimed that both petraea and hybrids were much superior to robur for planting in Scotland. Only one observation can be offered in this context, namely, that several old petraea woods are growing on very heavy clays where robur might have been expected if it were native. The wider implications of the introgression hypothesis extend into the realms of pure speculation. Widespread planting of oaks in the past would have accelerated any introgres- sion that was in progress, but has there been enough time (perhaps nine centuries) for planting to be held completely responsible ? One of Tucker’s examples of introgression (Cottam, Tucker & Drobnik 1959) is traced back to the Miocene era. In Scotland intro- gression could not date back much before the Climatic Optimum some 5,000 years ago. It is possible that robur reached Scotland naturally in the more favourable climate of those times. Petiole %, 10 20 30 Peduncle length to first bract (mm) Fig. 10. Scatter diagram for the four most homogeneous petraea samples combined. Watsonia 5 (5), 1963. VARIATION CONCLUSIONS OF OAKS IN SCOTLAND 285 Q. robur and Q. petraea seem likely to be good taxa but they are very hard to define in Scotland, mainly because there are no obvious natural populations in which to assess the normal ranges of variation of the species. Striking variation patterns are found in Scottish oakwoods but practically all the variation can be completely accounted for by postulating (a) limited interspecific hybridization and (b) extensive introgression initiated or accelerated by Man’s planting activities. TABLE 7 Variation recorded in the most homogeneous samples from petraea woods in Scotland? Elevation No. of specimens Leaf-length*+ (cm) Petiole-length* (mm) Petiole %* Pairs of lobes* Lobe depth %-* No. with deficient abaxial pubescence Auricle type Peduncle* total length (mm) Peduncle* length to first bract (mm) No. with peduncle* diameter intermediate No. with only trace of peduncle pubescence * a rather early collection t For illustrations showing the range of leaf variation in these samples see Cousens (1962) Watsonia 5 (5), 1963. mean | nil weak medium | | | ranges )| mean range mean Glentrool SW Galloway Ce 25Oit (75 m) 26 8:5—15-2 11-4 7:5—32:0 14:2 7T-0-21:5 12°6 Aikieside+ SE Coastal en 200 Tt (60 m) 11 7:5—14:7 10:6 8-0-19-0 12°6 7:0-16:0 Des + mean for two mid-shoot leaves Cheviot We Borders c. 600 ft (180 m) 6 7-4-11-3 9:6 tignoring the one obviously non-petraea individual Menstrie Ochils c. 500 ft (150 m) 6 8-4-10:°4 9-9 9:0-14-0 ts 9:7-13°8 11:8 4-0-7:0 5-4 34-60 | E. Central | | | | Combined Totals < longest peduncle on specimens 286 J. E. COUSENS Six homotypic characters have been tentatively defined. All are independent variables judging from the high proportion (93 %) of the possible character combinations occurring. For each one normal variation in certain directions is indistinguishable from that due to introgression and all six characters appear to be needed to define the taxa. In terms of these characters there are a few woods for which the population samples are reasonably homogeneous and without an introgressive trend in their scatter-diagrams (Fig. 10). They are all petraea woods; variation in them is summarized in Table 7. They provide the best material available in these data on which to base a description of petraea in Scotland. There seems little prospect of getting similar material in Scotland for robur. ACKNOWLEDGMENTS This work was carried out with the aid of a grant from the Forestry Commission whose officers in Scotland also provided information about the location of oak stands. My colleague Mr. D. C. Malcolm made some of the collections in 1961 and assisted with others. Their help is gratefully acknowledged. REFERENCES ANDERSON, E. (1949). Jntrogressive Hybridization. New York. ANDERSON, M. L. (1961). A History of Scottish Forestry (in MS). Burttr—Davy, V. (1933). The British Oaks, Quart. J. For. 27, 4. Cottam, W. P., TUCKER, J. M. & DRoBNIK, R. (1959). Some clues to the Great Basin postpluvial climates provided by oak distributions. Ecology. 40, 361-377. Cousens, J. E. (1962). Notes on the status of the sessile and pedunculate oaks in Scotland and their identification. Scot. For. 16, (3). DENGLER, A. (1941). Bericht tiber Kreuzungversuche zwischen Trauben-und Stieleiche (Q. sessiliflora Smith u. Q. pedunculata Ehrh.). Mitt. H-Goéring-Akad. Dt. Forstwiss. 1. Hora, E. (1929). Om mellemformerne mellem Q. robur L. og Q. sessiliflora Mart. Bot. Tidsskr. 40. JoNEs, E. W. (1959). Biological Flora of the British Isles. Quercus L. J. Ecol. 47. MAXWELL, H. (1900). Correspondence. Gdnrs.’ Chron. No. 724 p. 341. Moss, C. E. (1914-20). The Cambridge British Flora. Cambridge. RisDAL, M. (1955). On the two Oak species native in Norway. Norwegian For. Res. Inst. Report No. 46. STEBBINS, G. L. (1950). Variation and Evolution in Plants. New York. TANSLEY, A. G. (1939). The British Isles and their Vegetation. Cambridge. TANSLEY, A. G. (1940). Forestry 14 (1). TUCKER, J. M. & MULLER, C. H. (1958). A revaluation of the derivation of Q. margharetta from Q. gambellii. Evolution 12, 1-17. Tucker, J. M. (1960). QO. dunnii and Q. chrysolepis in Arizona. Brittonia 12. Tucker, J. M. (1961). Studies in the Q. undulata complex. I. A preliminary statement. Amer. J. Bot. 48, 202-208. WEIMARCK, H. (1947). De nordiska ekarna (with English summary). Bot. Notis. 1947. Forestry Commission Research Reports 1961 and 1962. H.M.S.O. Watsonia 5 (5), 1963. THE STATUS OF ORCHIS LATIFOLIA VAR. EBORENSIS GODFERY IN YORKSHIRE By RK. oH. Rorerrs and ©. L. GILBert ABSTRACT The Yorkshire dactylorchids formerly known as Orchis traunsteinerioides var. eborensis are shown to be referable to Dact)lorchis traunsteineri, the separate populations showing extreme divergence in certain characters such as stature, leaf length, and the size and shape of the labellum. This is, apparently, the result of a high degree of genetic isolation, combined in some instances with the effect of the extremely small size of the populations. INTRODUCTION In his classic work on British Orchids, Godfery (1933) described plants from some Yorkshire and Durham localities as a new variety under the name ‘ Orchis latifolia var. eborensis. According to him their salient features were dwarf stature (-_ 12 cm), very narrow, usually spotted leaves and a short, few-flowered spike. This new variety was also said to be early flowering, the only other orchid in flower at the same time being Orchis mascula. Three years later Pugsley (1936) described two Irish marsh orchid forms as new sub- species of the Continental Orchis majalis Rchb. One of these, based on plants from Co. Wicklow, he named subsp. Traunsteinerioides, thereby recognizing its close resemblance to the Continental Orchis traunsteineri Saut. Despite this, however, he decided that these plants were more closely allied to Orchis majalis Rchb., under which the new subspecies was therefore placed. The following year Pugsley (1939) examined a colony of Godfery’s var. eborensis near Hellifield, Yorkshire, and decided to place this variety under his subsp. Traunsteine- rioides, from which it was distinguished by its dwarf habit, and rather smaller flowers with a less distinctly deltoid labellum. Pugsley also recognized that in this latter feature these plants showed an approach to typical O. majalis. Following the elevation by Wilmott (1938) of O. majalis subsp. occidentalis to the rank of full species as Orchis occidentalis, Pugsley (1940) raised the subsp. Traunsteine- rioides also to species rank under the name Orchis traunsteinerioides. Subsequently, further colonies of it were discovered both in Ireland and in England (Pugsley 1946), and a biometric study of four of them by J. Heslop-Harrison (1953) showed beyond any doubt that all belong to the Continental species Dactylorchis traunsteineri (Saut.) VermlIn. The var. eborensis, however, was not studied by Heslop-Harrison, who suggested that further field work on the Yorkshire and Durham plants was desirable. Since the publication of Heslop-Harrison’s paper, further finds of D. traunsteineri have been made in Wales, Ireland and eastern England (Lacey 1955; Heslop-Harrison 1956; Lacey & Roberts 1958; Roberts 1960; Bellamy and Rose, undated; Heslop- Harrison, in /itt.). Subsequent studies of the Welsh populations have shown that in two of the Anglesey colonies a proportion of the plants have labella which are not sub-deltoid or obcordate, and have their lateral lobes angled and notched. In addition, the majority of the plants in these two colonies are very small. In both these features they strongly recalled the description of the var. eborensis. 287 Watsonia 5 (5), 1963. 288 R. H. ROBERTS and O. L. GILBERT Consequently in 1960 a search was made for the var. eborensis near Hellifield, and, with the help of information supplied by Miss C. M. Rob, at Rievaulx. The orchids were found at both places : at Hellifield a thorough search revealed about 22 plants in flower; at Rievaulx, where Miss Rob had informed us that the colony was a very small one, only 9 flowering individuals were found. In order that a comparison of these plants might be made with D. traunsteineri elsewhere, and particularly with those in Anglesey, a biometric study of the two colonies was made in 1961. All measurements were made in the same manner as in previous studies of the dactylorchids, those of vegetative parts being made in the field. All differences between means were tested for significance by means of ‘tf’ tests. HABITAT The close association of D. traunsteineri with areas in which Schoenus nigricans is locally dominant has been frequently remarked. The plants of the Hellifield colony are almost entirely restricted to the vicinity of the small patches of Schoenus which occur at a few places in the area. At the Rievaulx locality the orchids occupy a slightly wetter de- TABLE | Species associated with Dactylorchis traunsteineri in Yorkshire and Anglesey 1, Hellifield; 2, Rievaulx; 3, Pentraeth. 1 2, 3 | 2 3 Selaginella selaginoides if D. incarnata Oo Equisetum fluviatile l 0 D. purpurella O E. palustre fe r vi Crepis paludosa r r Trollius europaeus r Taraxacum officinale Ranunculus acris O O O Juncus subnodulosus 0 la R. flammula @) J. articulatus 0 r O Viola palustris O Eleocharis uniglumis i Polygala vulgaris O O Eriophorum latifolium 0 Hypericum tetrapterum r if E. angustifolium i) f Linum catharticum o-f— o-f o-f Schoenus nigricans ld 1 Id Genista anglica O Carex disticha o O Lotus pedunculatus O O C. diandra la Filipendula ulmaria ] 0 C. nigra a O a Potentilla erecta fi 1 0 C. elata O Parnassia palustris if la C. dioica la l Hydrocotyle vulgaris if if Csiflacca ia O ir Angelica sylvestris O r 0 C. panicea of if uf Primula farinosa te if C. pulicaris r oO Anagallis tenella } ) C. hostiana O f-a Ajuga reptans } r C. lepidocarpa IE O ie Plantago lanceolata 0 0 Molinia caerulea a a a Galium uliginosum 0 Briza media a 0 Valeriana dioica a 0 Festuca rubra i a Succisa pratensis ie O if F. ovina a Cirsium palustre O O O Centaurea nigra r O Serratula tinctoria if Fissidens adianthoides O O oO Leontodon hispidus O Thuidium tamariscinum yf a Pedicularis palustris if if Campylium stellatum if f ie Pinguicula vulgaris if if if Drepanocladus revolvens if. O yi Prunella vulgaris O of oO Scorpidium scorpioides o Listera ovata r Acrocladium giganteum l Ophrys insectifera i Acrocladium cuspidatum a a a Epipactis palustris O Uf ie Ctenidium molluscum a a a Dactylorchis fuchsil if r O Watsonia 5 (5), 1963. ORCHIS LATIFOLIA VAR. EBORENSIS GODFERY 289 pression in a damp meadow. Here again there are extensive Schoenus flushes. At both localities the community in which the orchids occur is very similar (Table 1), the soil being strongly flushed with calcareous ground water, and the habitat kept in a fairly open condition as a result of water-logging and trampling by cattle — conditions which are also met with in the Anglesey locality at Pentraeth. RESULTS The sample data for vegetative characters in the Hellifield and Rievaulx colonies are given in Table 2, together with those from one of the Anglesey colonies with features recalling the description of var. eborensis. The sample numbers are small and may seem inadequate for a biometric comparison. It is, perhaps, relevant therefore to point out that they represent all that could be found in the course of one season. The colonies, particularly that at Rievaulx, are very small ones, so that even the small samples we have been able to examine embrace a high proportion of the total populations. Stature Basing his description mainly on plants from near Helmsley, Yorkshire, Godfery (1933) gave the stature of var. eborensis as + 12cm. However, five specimens from the Helmsley locality, collected by T. J. Foggitt (1455, 1456 (BM)), have statures from 19-5 cm to 25:5cm. For the Hellifield plants Pugsley (1936) gave the range 15-25 cm, but the four plants gathered by him from this locality (BM) vary in stature from 16-0 to 31-5 cm. TABLE 2 Sample data for vegetative characters | | | | Stature | Total number | Leaf length | Leaf width | Incidence of Locality N (cm) | of leaves | (cm) (cm) | leaf marking | | | (%) | | mean s.e.m. | mean s.e.m.| mean s.e.m. | mean $.e.m. | Hellifield Pe Ise 606) | 49 011 | 92 023 | 12 0.03 | 36 Rievaulx 9 pee Si Or tee 27 OS ies O77, O54 1-0 0:08 || 67 Fentracth | 50 | 145 060 | 37 008 | 84 022 | 1:0 0-03 30 The data obtained from Hellifield and Rievaulx (Table 2) show that the mean stature in the two populations differs considerably, that of the Rievaulx plants being extremely small. Four specimens from the Rievaulx locality, gathered by Miss C. M. Rob in 1948 (K), show a similar range of stature. It can be seen that the mean stature of the Anglesey colony of D. traunsteineri quoted here is practically intermediate between those of the two Yorkshire colonies. Leaf dimensions Leaf-length, like stature, is a very variable character in D. traunsteineri and cannot be used to separate it from other species of marsh orchids. It is worth noting, however, that while the mean leaf-length of the Rievaulx colony is smaller than that of any other for which data are available, that of the Hellifield plants falls within the range of values found elsewhere in the British Isles for D. traunsteineri. Leaf-width, on the other hand, has been shown to be one of the most important characters for discriminating D. traunsteineri from other tetraploid marsh orchid species (Heslop-Harrison 1953). Both the mean and range of leaf-width in the Hellifield and Watsonia 5 (5), 1963. 290 R. H. ROBERTS, and“@: 1. GILBERT Rievaulx colonies are in close agreement with those of the Anglesey colony (Tables 2 and 3) and with all the Britannic populations for which data are available (Heslop-Harrison 1953, Lacey and Roberts 1958). Moreover, examination of herbarium material of var. eborensis from other Yorkshire localities and of D. traunsteineri from several Continental ones, including the type locality, has shown a remarkable uniformity in this character among all of them (Table 3). Leaf-number It was shown by Heslop-Harrison that the vegetative characters of leaf-number and leaf-width, taken together, completely distinguish the colonies of D. traunsteineri from those of the other tetraploid marsh orchids; later studies have fully confirmed this. It can be seen that in both the mean and range of leaf-number the Hellifield colony agrees very closely with that at Pentraeth, and indeed with all the other colonies previously studied. In the Rievaulx sample the mean number of leaves per plant is even smaller, but the range falls within that of the Pentraeth colony. TABLE 3 Ranges of variation of characters in populations of D. traunsteineri (The numbers in brackets were found only in single instances) Total number No. of Leaf width No. of flowers Locality of leaves non-sheathing (ci) in leaves inflorescence Hellifield 3-5 (6) 0-1 (2) 0-8—1-5 5-14 Rievaulx 2-4 0-1 0-6-1-3 5-10 other Yorks. localities taken together* 3-5 0°6—1-3 Pentraeth, Anglesey 2-5 (6) 0-1 (2) 0-7-1-5 2-14 Continental localities* 3-5 0:5-1:5 * Data from herbarium specimens Again, examination of herbarium specimens of var. eborensis from other localities in Yorkshire shows that the range of leaf-number in these also corresponds closely both with that in the Pentraeth colony and in all the herbarium material from Continental stations (Table 3). Leaf marking and shape The incidence of leaf marking in D. traunsteineri shows considerable variation from one locality to another and is completely lacking in some populations (Heslop-Harrison 1953, Lacey and Roberts 1958). In the Hellifield colony 36% of the plants have leaf marking either in the form of a few small dots or transverse bars in the apical half of the leaves; in the Rievaulx colony 67% of the plants had leaf marking consisting of a combination of narrow bars and dots. Both the form and incidence of leaf marking in these two popula- tions agree well with those found elsewhere in colonies of D. traunsteineri (Heslop-Harrison 1953). The leaves at both localities are narrowly lanceolate, in many cases linear-lanceolate, the lower leaves widest usually just below the middle, with an acute, slightly hooded tip. Watsonia 5 (5), 1963. PLATE 13 Cc Random samples of flower dissections from the three populations: (a) labella of 16 plants and spurs of 15 plants from the Hellifield colony; (b) labella and spurs of 9 plants from the Rievaulx colony; (c) labella ' of 60 plants from the Pentraeth, Anglesey, colony. All x 1. ORCHIS LATIFOLIA VAR. EBORENSIS GODFERY Zoi Floral characters According to Pugsley (1939) the plants of var. eborensis were distinguished from O. traunsteinerioides by their rather smaller flowers in which the labella were less distinctly deltoid. A study of the Welsh colonies has shown that there is considerably greater variation in labellum and spur dimensions between separate populations of the species than was apparent from Heslop-Harrison’s original study. Mean labellum-shape also varies from one colony to another. The sample data (Table 4) show that in both labellum and spur dimensions the differences between the Hellifield and Pentraeth populations are very small and statistically not significant. This is an unexpected and striking result. TABLE 4 Sample data for labellum and spur dimensions | Labellum length | Labellum width | Spur length | Spur width Locality N | (cm) (cm) | (cm) | (cm) | | | can mean 5.e.m. mean §.@.™m. mean S.e.m. | mean S.e.m. Hellifield 16 0-80 0-022 | 0-96 0-025 0-85 0-018 | 0-34 0-008 Rievaulx 9 la0;66 0-028 0-73 0-046 0-81 0:033 | 0:30 0-010 Pentraeth 40 0:79 0-014 1-02 0-018 0-83 0-020 | 0:35 0-010 On the other hand the fact emerges that the Rievaulx population differs from the others in both labellum-length and labellum-width by large and statistically significant amounts. It also differs significantly from them in spur-width, but the small differences in spur-length fail to be significant. The position with regard to labellum shape is rather different. By assuming that the Hellifield plants represented ‘the normal condition of var. eborensis’ Pugsley (1939) erroneously concluded that the labellum shapes found there are also those prevailing in the other colonies of var. eborensis. The present study has shown that this is not so. While it is true that, on the whole, the labella in all of the Yorkshire colonies are less deeply tri-lobed than in most other localities, these Yorkshire populations are evidently not homogeneous with respect to labellum shape. The Hellifield plants appear to deviate more from the typical labellum shape of D. traunsteineri than those at Rievaulx, where the labella, though small, clearly display the sub-deltoid shape more characteristic of this species (Plate 13). As far as may be inferred from herbarium specimens the situation among the other Yorkshire colonies seems to be much the same : some, like the Helmsley plants, showing greater, and others, like those from Masham and Carperby, less deviation from the more typical labellum shapes (Fig. 1). Fig. 1. Labellum shapes of ‘var. eborensis’ from other Yorkshire localities: (a) Masham; (b) Helmsley; (c) Carperby. All drawn from herbarium specimens. The colour of the flowers in var. eborensis was described by Godfery as a dull red-violet. There is clearly some justification for this description, for in a large proportion of the plants of the two colonies examined flower colour is rather darker than in the Anglesey Watsonia 5 (5), 1963. 292 R. H. ROBERTS and O. L. GILBERT and Caernarvonshire plants. But there is also a considerable overlap in this, a few indivi- duals in all of the Welsh colonies having dull, deep red-purple flowers exactly matching the deeper shades in var. eborensis. Another characteristic of var. eborensis mentioned by Godfery (1933) is that the lateral lobes of the labellum are reflexed, and this is very noticeable in these two populations. It is, of course, a character normally found in D. traunsteineri. Inflorescence The occurrence in D. traunsteineri of some plants with very few flowers in the spike has frequently been emphasized (Pugsley 1946, Heslop-Harrison 1953). In this there is a remarkably close agreement between the two Yorkshire colonies and the Anglesey one : ineach of them the number of flowers in the inflorescence is consistently low (Table 3), the mode in all three lying from 9 to 11. Flowering-time In the British Isles D. traunsteineri is one of the earliest flowering marsh orchids, commencing in mid-May and extending into the middle of June (Heslop-Harrison 1953, Roberts 1961). In a particularly early season a few of the Anglesey plants have been in flower as early as 7 May. As Godfery had observed, var. eborensis is also early flowering. In 1961 both the Hellifield and Rievaulx plants started flowering in the second half of May and reached a peak in the first week in June. DISCUSSION AND CONCLUSIONS The results show that there are comparatively large and statistically significant differences between the Hellifield and Rievaulx colonies in both vegetative and floral dimensions. The mean expression of labellum shape also differs to a marked degree in the two populations. On the other hand, between the Hellifield colony and that of undisputed D. trauns- feineri at Pentraeth there is a remarkably close agreement in most of the morphological characters, the chief exceptions being stature and the overall expression of labellum shape. The characters which, according to Pugsley, separated the var. eborensis from his O. traunsteinerioides are not uniformly represented among the Yorkshire colonies. Evi- dently the latter have diverged to a greater or lesser extent both from one another as well as from the other populations of D. traunsteineri in the British Isles. Despite this the results show conclusively that all these Yorkshire populations should be referred to D. trauns- teineri. In numbers of individuals these populations are very small: Pugsley estimated the Hellifield colony at just over fifty plants in 1937; in 1961 a careful search revealed about a half of this number. At Rievaulx only nine plants were found and this seems to be con- sistent with the numbers occurring in other seasons (Miss C. M. Rob, in /itt.). Because of their restricted habitat requirements these populations are geographically isolated from one another, as well as from all other populations of the species in the British Isles. Between many of them gene exchange either does not take place, or does so at a very low rate. Under these conditions it is not surprising that some degree of morphological divergence has arisen among the Britannic populations, particularly so among those which, like these in Yorkshire, consist of very small numbers of individuals. In these the range of biotypes is necessarily restricted and the operation of genetic drift has probably been of far greater significance. As has been shown, these population differences are most apparent in stature and in the size and shape of the labellum; and, it will be recalled, it is the diversity in these particular Watsonia 5 (5), 1963. ORCHIS LATIFOLIA VAR. EBORENSIS GODFERY 293 characters which in the past has been the main obstacle to the correct identification of these northern colonies. The recognition of these Yorkshire plants as D. traunsteineri considerably broadens our appreciation of the range of variation to be found within this species in the British Isles. This, however, is not unexpected. In Holland, as Vermeulen (1949) has shown, the species shows considerable variation in both stature and shape of labellum; in the Baltic states it displays even greater variation, the regional variants in some instances being sufficiently well-marked to have become the basis of several subspecies (Vermeulen 1947). LOCALITIES The known localities of D. traunsteineri in Yorkshire and the exsiccatae from them are listed below. V.c. 62, North-east York. Beckdale, near Helmsley, 1905 and 1922, T. J. Foggitt (BM); Ashberry, near Rievaulx, 1937, T. J. Foggitt (BM); ibid., 1948, Miss C. M. Rob (K). V.c. 64, Mid-west York. Near Hellifield, 1937, H.W.P. & W.A.S. (BM). V.c. 65, North-west York. Carperby, Wensleydale, 1885, F. Arnold Lees (BM); Masham, 1893, A. B. Sampson (K); Tanfield, 1906, 7. J. Foggitt (BM). The var. eborensis was also stated by Godfery to occur in Durham, but there appear to be no herbarium specimens of it from that vice-county. Nevertheless, further explora- tion may well prove the occurrence of D. traunsteineri not only there but in other fen areas in the north of England and in Scotland. ACKNOWLEDGMENTS We wish to thank Professor J. Heslop-Harrison for kindly reading this note in manu- script and for his valuable suggestions; Miss C. M. Rob for details of the Rievaulx locality; and the authorities of the herbaria at the British Museum (Natural History) and at the Royal Botanic Gardens, Kew, for permission to consult their collections. Thanks are also due to Mr. H. T. Davies, Senior Technician at the Department of Botany, University College, Bangor, for taking the photographs of floral parts. REFERENCES BELLAMY, Davip J. & Rose, FRANCIS (undated). The Waveney-Ouse Valley Fens of the Suffolk-Norfolk Border. Trans. Suffolk Nat. Soc. 11, 367-385. GoprFeryY, M. J. (1933). Monograph and Iconograph of the Native British Orchidaceae. Cambridge. HeEsLop-HArrIson, J. (1953). Studies in Orchis L. Ul. Orchis traunsteineri Saut. in the British Isles. Watsonia 2, 371-391. HEsLop-HARRISON, J. (1956). Dactylorchis traunsteineri Saut. in Co. Antrim. Irish Nat. J. 12, 56-57. Lacey, W. S. (1955). Orchis traunsteineri Saut. in Wales. Proc. Bot. Soc. Brit. Is. 1, 297-300. Lacey, W. S. & Roserts, R. H. (1958). Further notes on Dactylorchis traunsteineri (Saut.) Vermeul. in Wales. Proc. Bot. Soc. Brit. Is. 3, 22-27. Pusey, H. W. (1936). New British Marsh Orchids. Proc. Linn. Soc. Lond., 148, 121-125. Pucs.ey, H. W. (1939). Recent work on Dactylorchids. J. Bot., Lond. 77, 50-56. Puastey, H. W. (1940). Further notes on British dactylorchids. J. Bot., Lond. 78, 177-181. PuGs.ey, H. W. (1946). Orchis traunsteinerioides Pugsl. in Britain. Naturalist, Lond. No. 819, 47. Roserts, R. H. (1960). The Wicklow Marsh Orchid in Anglesey. Nature in Wales 6, 14-17. Roperts, R. H. (1961). Studies on Welsh Orchids Il. The occurrence of Dactylorchis majalis (Reichb.) Vermeul. in Wales. Watsonia 5, 37-42. VERMEULEN, P. (1947). Studies on Dactylorchids. Utrecht. VERMEULEN, P. (1949). Varieties and forms of Dutch orchids. Ned. kruidk. Arch. 56, 225-242. Witmort, A. J. (1938). Orchis occidentalis (Pugsley) Wilmott, in Campbell, M.S., Further botanising in the Outer Hebrides. Rep. Bot. Soc. Exch. Cl. 11, 551. Waisonia 5 (5), 1963. STUDIES IN RANUNCULUS SUBGENUS BATRACHIUM (DC.) A. GRAY Il. GENERAL MORPHOLOGICAL CONSIDERATIONS IN THE TAXONOMY OF THE SUBGENUS By GC. DK Coow Hartley Botanical Laboratories, The University, Liverpool ABSTRACT An account is given of the characters that are taxonomically important in separating subgenus Batrachium from Ranunculus. The morphology of the stipules and achenes is discussed in some detail as they have been largely overlooked. Although of no direct taxonomic importance, embryology, teratology and chromosome size are described. INTRODUCTION Batrachium was first described by De Candolle (1818) as a section of Ranunculus. S. F. Gray (1821) gave it generic status and A. Gray (1886) designated it a subgenus of Ranunculus. Some recent authors such as Kreczetovicz in Komarov (1937), Janchen (1958), Rostrup (1958) and Love (1961) have chosen to give Batrachium generic rank, while most other workers include it within Ranunculus; the infrageneric rank is frequently not designated, but Ascherson & Graebner (1935), Benson (1948) and Clapham (1952) have recognised it as a subgenus. The full description of De Candolle’s section Batrachium is as follows : Pericarpia ovata mutica rugulis transversis striata. Flores albi ungue flavo foveola ad basin nectarifera exsquamulata. Plantae aquaticae; folia glaberrima, emersa dentato-sublobata, immersa capillaceo-multifida; radices fibrosae; pedunculi uniflori oppositifolii. Clapham adds that the species are annual or perennial and that they have stipules. In most Floras it is a combination of these characters that is normally used to delimit this group. There are, however, some exceptional species which will be discussed later. The batrachian group contains about eighteen morphologically recognisable species, found in climatically temperate regions throughout the Northern Hemisphere and in limited areas of South America, South Africa, Australia, Tasmania and New Zealand. Atlantic Europe is the region containing the greatest species density (Fig. la) and the greatest morphological diversity (Fig. 1b). The distributions shown on Figs. la and 1b are semi-diagrammatic and the species boundaries are not as smooth as illustrated. It is hoped to publish detailed distribution maps of each species later. Except in South America the species that are heterophyllous or entirely laminate-leaved are always found in areas also occupied by species that develop only capillary leaves. Most of the species are annuals or short-lived perennials but R. fluitans Lam., R. pseudofluitans (Syme) Baker & Foggitt ex Newbould and R. circinatus Sibth. are long- lived perennials. Some species such as R. tripartitus DC., R. pe/tatus Schrank, R. aquatilis L. and R. trichophyllus Chaix may be annuals or perennials. They are often found in temporarily aquatic habitats: if the habitat dries out in summer they behave as annuals, but if the water remains throughout the year they may perennate. If cultivated in an aquatic habitat these species usually die in the third year, but this varies in different races of each species. For example, two morphologically distinguishable races of R. peltatus. one from Portugal and the other from southern France, are obligate annuals in cultivation in England. Watsonia 5 (5), 1963. 294 MORPHOLOGY OF RANUNCULUS SUBGENUS BATRACHIUM Pie) Le ee 1 species (a) 2 species 4444 4 species -==-- 6 species uuu Q species uuu olants with only laminate leaves plants with only capillary leaves ---=-- plants with both laminate and capillary leaves (b) ‘ Of Ca a ify 19 1 4 \ ee Fig. 1. Distribution of (a) species density, and (6) morphological diversity in Ranunculus subgen. Batrachium. Watsonia 5 (5), 1963. 296 Cc. D. KY COOK R. hederaceus L. and R. omiophyllus Ten. (R. lenormandii F. Schultz) are primarily terrestrial plants of wet places but all the other species are aquatic, completely or partly submerged in water. All the aquatic species cultivated (R. tripartitus, R. ololeucos Lloyd, R. baudotii Godr., R. peltatus, R. aquatilis, R. trichophyllus, R. rionii Lagger, R. sphaerosphermus Boiss. & Blanche, R. pseudofluitans, R. fluitans and R. circinatus) can be grown as terrestrial plants during the summer in water-logged soil. DISCUSSION OF CHARACTERS Hairiness Almost all descriptions of Batrachium state that all the species are glabrous. This is not the case. All the specimens that I have seen bear unicellular hairs on the stem, leaves and stipules. Even the submerged capillary leaves have hairs surrounding the hydathodes. Heterophylly At all times of the year, in or out of water, R. hederaceus and R. omiophyllus produce only laminate leaves, while R. trichophyllus, R. rionii, R. sphaerospermus, R. circinatus and some races of R. aquatilis, R. pseudofluitans and R. fluitans bear only capillary leaves. The other species, R. /obbii (Hiern) A. Gray, R. tripartitus, R. ololeucos, R. baudotii, R. peltatus and some races of R. aquatilis, R. pseudofluitans and R. fluitans show heterophylly with capillary and laminate leaves. This type of heterophylly is not confined to Batrachium but is found in some other aquatic species of Ranunculus such as R. polyphyllus Waldst. & Kit. ex Willd., R. gmelinii DC. and R. flabellaris Raf. In most aquatic plants that have both capillary and laminate leaves the change from one type to the other is gradual with the formation of sequential intermediates and is largely controlled by the presence or absence of water. In R. /obbii, R. tripartitus, R. ololeucos, R. baudotii, R. peltatus and R. aquatilis the change from one type of leaf to the other type is abrupt. In R. pe/tatus and R. aquatilis the initiation of the laminate leaf must take place underwater and if the apex is lifted above the water a capillary leaf of the terres- trial type develops. Intermediate types of leaves are occasionally found but they are not sequential intermediates but more of a mixture of the two types. The presence of these intermediate leaves can be correlated with pollen sterility, and they are found in high polyploids and hybrids. I have only just started studies on heterophylly but it appears that, in the presence of water, the form of the leaf is primarily controlled by temperature and day-length. Stipules In taxonomic accounts the Ranunculaceae are usually described as being exstipulate, but occasionally, for example Lawrence (1951), an exception is made in the case of Thalictrum, and Clapham (1952) describes both Thalictrum and Batrachiumas stipulate. Hiern (1871), Freyn (1890), Gliick (1919, 1924), Salisbury (1934) and Drew (1936) in monographic accounts have all described stipules in Batrachium and have used them as taxonomic characters within the subgenus. Morphologists, however, have been more reluctant to call them stipules and Goffart (1901) called them auricules, while Weberling (1956) called them Scheidenlappen or lobes of the leaf-sheath. To define stipules it is necessary to look at the development of the leaf. The first visible sign of leaf formation is a swelling below and to the side of the apical meristem. This swelling is usually called the leaf primordium and it may remain small and localised on one side of the stem apex or it may grow, occasionally forming a ring around the apex. The leaf proper (petiole and lamina) develops from a localised area of this primordium so the whole primordium contributes cells to the axis and the leaf proper. Eichler (1861) Watsonia 5 (5), 1963. a MORPHOLOGY OF RANUNCULUS SUBGENUS BATRACHIUM 297 proposed two terms : Unterblatt and Oberblatt. The Oberblatt is the area in the primor- dium that gives rise to the leaf proper and the Unterblatt is that which is left. He then defines stipules as being a product of the Unterb/att. Sinnott & Bailey (1914) carried out more detailed anatomical investigations of leaf development and found themselves in agreement with Eichler; they suggested ‘ leaf-base of the primordial leaf’ as a translation of Unterblatt. Colomb (1887) defined the stipule as ‘an appendage inserted on the stem at the base of the leaf all the bundles of which are derived entirely from corresponding foliar bundles.’ Sinnott & Bailey stated that this definition would fit Eichler’s but they found that it was often hard to apply and could not be adopted as stipules occasionally lack bundles. Sinnott & Bailey pointed out that leaf-sheaths, stipules, ligules and some other append- ages at the base of the petiole arise from the Unterb/att and are morphologically homologous. Nobody has doubted that the leaf-sheaths found in the Ranunculaceae originate from the Unterblatt. It has been noted that these leaf-sheaths are occasionally lobed, so the problem is one of terminology. When should the lobe on a leaf-sheath be called a stipule ? Weberling suggested that if the lobes show proleptic growth, that is, if they anticipate growth of the leaf proper (Oberb/att), then they should be called stipules. The lobes in Batrachium do anticipate growth of the leaf proper and therefore should be called stipules. It is perhaps worth mentioning that Tro//ius and Caltha also show well-developed stipules but the stipule ruptures as the next leaf develops. Petals Comparative studies on the floral parts of the Ranunculaceae (Eichler 1878, Prantl 1888) have shown that the so-called petals of Ranunculus are strictly speaking homologous to nectaries and are occasionally called Nektarbldtter (usually translated into English as honey-leaves). It is proposed to call the Nektarb/ldtter petals as this usage is generally accepted in Ranunculus and does not cause confusion. The petals in Batrachium are usually described as being white with a yellow claw. This is true for the majority of species but in races of R. ololeucos they are entirely white and in R. flavidus (Hand.-Mazz.) C. D. K. Cook* they are uniformly pale yellow. This species seems to be allied to R. aquatilis and is confined to the Himalayas. The petals have a matt surface (Parkin 1928) and do not contain starch. This appears to be a good distinction between subgenus Batrachium and section Hecatonia of subgenus Ranunculus which contains the most batrachian-like species. This character is not of much practical use as it requires microscopical examination of living petals. The nectary pit is highly reduced (Benson 1940). It may be lunate, circular or pyriform. Nectary characters are useful taxonomically within the subgenus but cannot be used to separate Batrachium from the rest of Ranunculus. Achenes The presence of transverse ridges on the achene is perhaps the most important diag- nostic character of Batrachium, but it must be used with caution. R. rivularis Banks & Sol., R. undosus Melville, R. papulentus Melville and R. inundatus R. Br. ex DC., semi-aquatic species from Australasia, have transverse ridges on the achene. The ridges, however, are somewhat irregular and ill-defined. Anatomically the pericarp of R. rivularis consists of an inner layer of regular, almost symmetrical sclereids five to eight cells deep and an outer layer of loose parenchyma five to eight cells deep. The pattern is the same in longitu- dinal and transverse sections. The pericarp of Batrachium has a much more complicated structure and consists of three well defined layers. The innermost consists of a single layer of fibres forming a continuous sheath arranged tangentially around the achene at right angles to the longitu- *Ranunculus flavidus (Hand.-Mazz.) C. D. K. Cook, comb. nov. Batrachium flavidum Hand.-Mazz., Acta Hort. Gotob. 13, 168 (1939) Watsonia 5 (5), 1963. 298 C. D. K. COOK dinal axis. The fibres are as much as twenty times as long as broad and have thick walls with numerous simple pits. The next layer consists of a layer of elongated sclereids, three to four cells deep and from five to ten times as long as broad with thick walls and many simple pits. The sclereids lie at right angles to the fibres, parallel to the longitudinal axis. Each cell lies directly on top of the one below. The ends are slightly swollen and tend to turn away from the longitudinal axis. As these cells are in simple ranks the upturned ends form quite large ridges between 0-4 and 0-1 mm apart. The ridges lie at right angles to the longitudinal axis and are found on the lateral walls of the achene. The outer layer consists of between four and eight layers of loose parenchymatous cells which wither and die as the achene matures. The outermost layer may or may not bear unicellular hair cells. Fig. 2a shows a median transverse section and Fig. 2b a median longitudinal section. The ends of the sclereids swell and turn outwards before fertilization and, to a large extent, subsequent development is not dependent on fertilization. Normally, the walls of the sclereids start to lignify just as the endosperm starts laying down cell walls. Lignin is first laid down at the ends of each cell while the middle is still elongating. The fibres of the inner layer do not become lignified until the embryo is almost mature. Fig. 2. Sections of the lateral wall of achenes of Ranunculus spp., c. half-way from base to apex. (a) R. tripartitus 175/57, T.8.; (b) ibid., L.S.; (c) R. sceleratus, Bacs-Bodrog, nr. Zombor, Hungary, J. Proden May 1912 (achenes ridged), L.S.; (d) R. sceleratus (achenes non-ridged), L.S. The important property of these ridges on the batrachian achene is that they are areas of weakness. When the achenes are roughly handled they break along the ridges. During germination the pericarp ruptures along the ad- and abaxial crests which are areas of upturned sclereids. Although there is no direct evidence it is tempting to regard the ridges as possible passages of water through the periderm when the seeds are wetted after being dried. Watsonia 5 (5), 1963. MORPHOLOGY OF RANUNCULUS SUBGENUS BATRACHIUM 299 Some races of R. sceleratus L. have transverse ridges on the lateral walls of the achene that look similar to those found in Batrachium. Plants with ridged achenes are sympatric with non-ridged ones. Anatomically the periderm of R. sceleratus is similar to that found in Batrachium (Fig. 2c). The ends of the sclereids, however, are not markedly swollen and it appears that the ridges are largely formed by a collapse of the central part of the cell. The unridged race has a layer of symmetrical sclerenchyma cells, three deep, between the sclereids and the outer parenchyma (Fig. 2d). When the sclerenchyma is present the sclereids do not collapse and no ridges are formed. Lonay (1901, 1907) examined the structure of the achenes of many species of Ranunculus and it would appear that these elongated, ranked, sclereids may well be confined to Batrachium and R. sceleratus. Chute (1930) stated that the achene of Batrachium represented the most reduced type of any known. She said ‘so complete is this reduction that the ventrals, as distinct bundles, have completely disappeared; hence the ovule trace appears to come from the dorsals.’ The vascular anatomy of so few species of Ranunculus has been studied that it is not known if it is possible to use the lack of ventral bundles as a diagnostic character. Embryology Each achene contains a single anatropous ovule. The nucellus is monosporic, 8-nucleate, of the Polygonum-type with early fusion of the endosperm nuclei. Most pollen tubes reach the nucellus by travelling down the funicle, none have been seen ascending the micropyle. After fertilization the endosperm nucleus divides rapidly and cell walls are not laid down until the embryo has reached a sixteen- or thirty-two-celled stage. The embryo of R. tripartitus has the same development as the Myosurus variation of the Onagrad type (terminology after Johansen 1950). A limited number of embryological stages have been seen in R. hederaceus, R. omiophyllus and R. aquatilis; they all appear similar to R. tripartitus. No self-incompatibility, pseudogamy or agamospermous processes have been discovered. The synergid cells disappear before or just after fertilization; their disappearance does not appear to be correlated with pollination. The antipodal cells do not disappear until late in embryo development. They become enlarged and show considerable activity and are often borne on a short well-differentiated stalk. Enlargement of the antipodal cells has been seen in several species within the Ranunculaceae, Fumariaceae and Papaveraceae (Coulter 1898, Osterwalder 1898, Grafl 1941, Tschermak-Woess 1956). Not only do the antipodal cells become enlarged, they show giant chromosomes which continually divide forming highly polyploid cells; Hasitschka-Jenshke (1959) estimated cells up to 64-ploid for Eranthis hyemalis (L.) Salisb. An interesting feature that has not, I believe, been noted before is that the activity and growth of these cells is increased if pollination is withheld. Conversely, if pollination is carried out as soon as the nucellus is mature, there is very little development of the antipodal cells. In Batrachium the antipodal cells have never been seen to divide. Teratological variants Although teratological variants can hardly be used in taxonomy, it is, perhaps, worth mentioning the commonest types found. Occasionally one sepal is partly petaloid; some- times it is laterally asymmetrical with one half petaloid and sometimes it is the distal end that is petaloid. Petaloid sepals are without nectary pits. Intermediates between petals and stamens are more common; one extreme is an almost perfect stamen with a flattened filament and the other extreme is an almost perfect petal bearing one or two functional anther loculi. In the latter case the anther may occur laterally or centrally on the petal Watsonia § (5), 1963. 300 CysDA Ki iCOOKk but if it is central it usually replaces the nectary pit; sometimes this type of anther is so reduced that it looks like a pollen-producing nectary. Supernumerary petals (more than five) are quite common in R. fluitans, R. pseudo- fluitans, R. peltatus and R. aquatilis. They appear to be formed at the expense of stamens and in one population of R. pseudofluitans in the River Wye in Derbyshire plants have up to fifteen petals and no stamens. The plants examined in this population were sterile pentaploids. This ‘ double-flowered °’ character has been seen to breed true in cultivation in plants of R. pe/tatus collected from Slayley in Derbyshire. ——— ——— = Fig. 3. Chromosomes of Ranunculus spp. from root-tip squashes, < 2500; Feulgen staining, all except (d) pre-treated for 3 hrs with «-bromonaphthalene. (a) R. ficaria, under Hippophae rhamnoides by pond, Botanic Garden, Munich, Germany, 2” = 32; (b) R. sceleratus, from collection of aquatic plants, Botanic Garden, Munich, 2n = 32; (c) and (d) R. trichophyllus, collected wild from pool by R. Stillach, S of Moserbriicke, 1:5 km SW of Oberstdorf, S Germany (specimens deposited in M). Watsonia 5 (5), 1963. MORPHOLOGY OF RANUNCULUS SUBGENUS BATRACHIUM 30] Chromosome size It has been suggested that Batrachium could be delimited from the rest of Ranunculus on chromosome size, but no really significant difference has been found. Larter (1932) reported that the chromosomes of R. pe/tatus were less than one sixth the bulk of those of R. acris L. and considerably smaller than those of the other species illustrated in his paper, (R. constantinopolitanus (DC.) @Urv., R. trilobus Desf., R. serbicus Vis., R. parviflorus L., R. arvensis L., R. repens L., R. ficaria L., R. chius DC., R. nelsonii A. Gray, R. sprunerianus Boiss., R. oxyspermus Bieb., and R. ophioglossifolius Vill.). From the work of Skalinska et al. (1959) it can be seen that the chromosomes of R. circinatus are approximately the same size as those of R. flammula but smaller than those of R. montanus and much larger than those of /sopyrum thalictroides L. and Thalictrum aquilegiifolium L. Langlet (1927) doubted that Batrachium was a distinct genus on the basis of chromo- some size and number and in his later work (1932) on the relationships within the Ranunculaceae he included Batrachium within the same category as Ranunculus. Other workers on cytological variation in the Ranunculaceae (Coonen 1939, Gregory 1941, Kurita 1958) have also included Batrachium within Ranunculus. It is difficult to compare chromosome sizes from data given by different publications as there are either variations in size within a single species or inaccuracies in the scales quoted. Fig. 3a shows chromosome preparations from root-tip squashes, with Feulgen staining (12 min. hydrolysis in normal hydrochloric acid at 60°C), all prepared within a week and all drawn with the same microscope and camera lucida. Except for Fig. 3b, the roots were given a three hour pre-treatment in saturated «-bromo-naphthalene solu- tion. This pre-treatment is necessary as the chromosomes of Ranunculus are rather long and somewhat sticky at metaphase and it is seldom that one gets a late prophase prepara- tion with a good chromosome spread such as that shown in Fig. 3d. It can be seen that the chromosomes of R. ficaria illustrated are larger than those of R. trichophyllus while those of R. sceleratus are smaller. In Batrachium there is a certain amount of variation in chromosome size; none has chromosomes as large as those of R. ficaria but in R. tripartitus they are as small as those of R. sce/eratus. In many genera there are species with different chromosome sizes; Anemone, also within the Ranunculaceae, is a good example (Langlet 1932, fig. 2; Gregory 1941, plate 2.) CONCLUSION In the introduction it was pointed out that the status of Batrachium has varied from section to genus. The group shows most affinities to Ranunculus and it cannot be confused with any other genera in the Ranunculaceae. However, it can only be separated from Ranunculus by the combination of the minimum of two characters (transversiey-ridged achenes and matt-surfaced petals). Sectional status is inappropriate so the choice lies between genus and subgenus, but it is felt that the level of differentiation is insufficient to merit generic status so Batrachium is given subgeneric rank. A suggested description of Ranunculus subgenus Batrachium is as follows : Aquatic or semi-terrestrial annuals or perennials. Leaves mostly cauline, laminate or finely dissected; stipules membranous, partly fused to the petiole. Sepals usually five, not petaloid, caducous. Petals five or more, with matt surface; nectary-pit lunate, circular or pyriform. Achenes not strongly compressed, with regular transverse ridges from 0-4 to 1 mm apart on the lateral walls. The lectotype of subgenus Batrachium is Ranunculus hederaceus, sheet number 74 (Savage 1945) in the Linnaean Herbarium, Linnean Society, London (LINN), selected by Bodom and Brown, 1913, 7/f. Fi. N.U.S. ed. 2; 2, 115. ACKNOWLEDGMENTS Financial support during the major part of these studies was provided by the Watsonia 5 (5), 1963. 302 C.D. K, GOOK Nature Conservancy and the D.S.I.R. to whom grateful acknowledgments are made. I would also like to thank Dr. Focko Weberling for discussion on stipules and Dr. V. H. Heywood for help with the manuscript. REFERENCES ASCHERSON, P. & GRAEBNER, P. (19?5). Synopsis der mitteleuropdischen Flora 5 (3), 74-98. BENSON, L. (1940). The North American subdivisions of Ranunculus. Amer. J. Bot. 27, no. 9, 799. BENSON, L. (1948). Ranunculus in North America. Amer. Mid. Nat. 40, 1-261; and Suppl., (1954) Joc. cit. 52, 328-369. CuHuTeE, H. (1930). Achene morphology. Amer. J. Bot. 17, 709. CLAPHAM, A. R. (1952). In Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles, Cambridge. CoLoms, M. G. (1887). Recherches sur les stipules. Ann. Sci. Nat. sér. 7, Bot. 6, 1-76. COOoNEN, L. P. (1939). The chromosomes of the Ranunculaceae. Amer. J. Bot. 26, 49-58. CouLter, J. M. (1898). Ranunculus embryology. Bot. Gaz. 25, 73-88. DE CANDOLLE, A. P. (1818). Regni Vegetabilis Systema Naturale 1, 232-233. Drew, W. B. (1936). The North American representatives of Ranunculus section Batrachium. Rhodora 38, 1-47. EIicHLeR, A. W. (1861). Zur Entwicklungsgeschichte des Blattes mit besonderer Beriicksichtigung der Nebenblattbildungen. Inaugural Dissertation, Marburg. EICHLER, A. W. (1878). Bliithendiagramme 2, 160. Leipzig. FRreyYN, J. (1890). Zur Kenntniss einiger Arten der Gattung Ranunculus. Bot. Centralol. 40, 1-23. GLUckK, H. (1919). Blatt- und bliitenmorphologische Studien. Jena. GLUcK, H. (1924). Biologische und morphologische Untersuchungen tiber Wasser- und Sumpfgewachse 4. Jena. GOFFART, J. (1901). Recherches sur l’anatomie des feuilles dans les renonculacées. Arch. Inst. Bot. Univ. Liége 3 (2) 1-190. GratL, [. (1941). Uber das Wachstum der Antipodenkerne von Caltha palustris. Chromosoma 2, 1-11. Gray, A. (1886). Revision of Ranunculus in North America. Proc. Amer. Acad. Arts and Sci. 21, 363-378. Gray, S. F. (1821). A Natural Arrangement of British Plants 2, 720. GreGorY, W. C. (1941). Phylogenetic and cytological studies in the Ranunculaceae. Trans. Amer. Phil. Soc. N.S. 31, 443-521. HASITSCHKA-JENSCHKE, G. (1959). Vergleichende karyologische Untersuchungen an Antipoden. Chromo- soma 10, 229-267. HIERN, W. P. (1871). On the forms and distribution over the world of the Batrachium section of Ranunculus. J. Bot., Lond. 9, 43-49; 65-69; 97-107. JANCHEN, E. (1958). Catalogus Florae Austriae 2, 195-197. Wien. JOHANSEN, D. A. (1950). Plant Embryology. Waltham, Mass. Komarovy, L. (1937). Flora U.R.S.S. 7, 335-350. Leningrad. Kurita, M. (1958). Chromosome studies on Ranunculaceae. VIII. Karyotype and Phylogeny. Rep. Biol. Inst. Ehime Univ. No. 5, 1-14. LANGLET, O. (1927). Beitrage zur Zytologie der Ranunculazeen. Svensk Bot. Tidskr. 21, 1-17. LANGLET, O. (1932). Uber Chromosomenverhiltnisse und Systematik der Ranunculaceae. Svensk Bot. Tidskr. 26, 381-400. LarTER, L. N. H. (1932). Chromosome variation and behaviour in Ranunculus. J. Genet. 26, 255-285. LAWRENCE, G. M. G. (1951). Taxonomy of Vascular Plants. New York. Lonay, H. (1901). Contribution a l’anatomie des renonculées, structure des péricarpes et des spermodermes. Arch. Inst. Bot. Univ. Liége 3, 1-164; & Suppl. (1907) loc. cit. 4, 1-33. Love, A. & D. (1961). Chromosome numbers of Central and Northwest European plant species. Opera Botanica, Lund, 5, 162. OSTERWALDER, A. (1898). Beitrage zur Embryologie von Aconitum napellus. Flora 85, 254-292. PARKIN, J. (1928). The glossy petal of Ranunculus. Ann. Bot., Lond. 42, 739-755. PRANTL, K. (1888). Beitrage zur Morphologie und Systematik der Ranunculaceen. Bot. Jb. 9, 225-273. RostrupP, E. (1958). Den Danske Flora. Copenhagen. SALISBURY, E. J. (1934). On the morphology, ecology and distribution of Ranunculus hederaceus and R. lenormandii. J. Bot., Lond. 72, 185-196. SAVAGE, S. (1945). A Catalogue of the Linnaean Herbarium. London. Watsonia 5 (5), 1963. MORPHOLOGY OF RANUNCULUS SUBGENUS BATRACHIUM 303 SKALINSKA, M. ef a/. (1959). Further studies in chromosome numbers of Polish Angiosperms (Dicotyledons). Acta Soc. Bot. Pol. 28, 491-493. SINNOTT, E. W. & BaILey, I. W. (1914). Investigations on the phylogeny of the angiosperms. 3: Nodal anatomy and the morphology of stipules. Amer. J. Bot. 1, 441-453. TSCHERMAK-WoOESS, E. (1956). Notizen tiber die Riesenkerne und ‘ Riesenchromosomen’ in den Antipoden von Aconitum. Chromosoma 8, 114-134. WEBERLING, F. (1956). Morphologische und entwicklungsgeschichtliche Untersuchungen tber die Ausbildung des Unterblattes bei dikotyleden Gewdchsen. Beitr. z. Biol. d. Pflanzen 32, 27-105. Watsonia 5 (5), 1963. STUDIES ON ALCHEMILLA FILICAULIS BUS., SENSU LATO, AND AY MINIMA WALTERS. INTRODUCTION, AND I. MORPHOLOGICAL VARIATION IN A. FILICAULIS., SENSU LATO By MARGARET E. BRADSHAW Department of Extra-Mural Studies, University of Durham ABSTRACT Alchemilla filicaulis Bus., A. vestita (Bus.) Raunk. and A. minima Walters form a critical group within the A. vulgaris L. aggregate. The results of a biosystematic study of variation in morphology, habit and cytology are presented in a series of papers. These indicate that 4. vestita should be treated as a subspecies of A. filicaulis but A. minima should be maintained as a species. In paper I morphology and habit are considered. Density of the indumentum of the flower-stem and leaves is genetically determined but greater variation occurs in the more hairy plants, subsp. vestita, than in the less hairy subsp. fi/icaulis. Plants of intermediate hairiness may be genotypes or environmentally- induced variants. Colour of the stipules and leaf shape are genetically determined but can be modified by environmental factors. Lowland, meadow and rock-ledge plants are tall with few shoots; a dwarf much branched genodeme occurs in close-grazed upland grassland. Such ecogenodemes are analogous to the ecotypes of sexual species and important in the evolution of apomicts. INTRODUCTION In recent years a number of papers have been published on the taxonomy and distribu- tion of the British species of the apomictic A/chemilla vulgaris L. aggregate by Walters (1949, 1952), Bradshaw (1957), Bradshaw & Walters (1961) and Bradshaw (1963). In Britain there are twelve named taxa and a critical group temporarily referred to as the Lawers *° acutidens’ (see Clapham, Tutin, & Warburg, 1962). Several species of the Alchemilla vulgaris aggregate were shown to be obligate apomicts by Strasburger (1905) and Murbeck (1901). All the British taxa are thought to be obligate apomicts. Walters’ paper (1949) illuminated some of the difficulties associated with the group formed by A. filicaulis Bus., A. vestita (Bus.) Raunk. and A. minima Walters. Of the first two, Walters wrote *“ The morphological difference is very slight — the only “ good ” character appears to be the hairiness of the inflorescence — nevertheless, the two can with rare exceptions be satisfactorily separated in the field, and exhibit somewhat different geographical ranges.’ Exactly how ‘* good’ needed to be ascertained, as field collections made between 1951 and 1954 contained several plants of intermediate hairiness not readily assignable to either taxon. Furthermore dwarf plants of these two taxa from closely grazed habitats are very similar to A. minima. In the field, the latter proved not so easy to separate from the former as suggested by Walters (1949), so that further study of the dwarf populations was needed to ascertain the proportions of genotypic and environmentally induced dwarfs. The present work is based on a biosystematic study of these taxa with particular reference to variation in morphology, habit and cytology. As a result of these investiga- tions certain changes of status are recommended. The new designations and reasons for the changes are briefly given here in order that these names can be used throughout the papers in which fuller explanations will appear. 304 Watsonia 5 (5), 1963. VARIATION IN ALCHEMILLA FILICAULIS 305 Satisfactory separation of A. filicaulis and A. vestita by morphological characteristics is not possible because of intermediates. The two taxa do exhibit somewhat different ecological and geographical distributions; and it is accordingly proposed that they be ranked as subspecies of A. filicaulis Bus., viz : (1) subsp. filicaulis (2) subsp. vestita (Bus.) M. E. Bradshaw stat. nov. (A. filicaulis Bus. forma vestita Bus., Bull. Herb. Boiss. 1, app. 2, 23 (1893). A. minima Walters is retained at specific rank.* IP MORPHOLOGICAL VARIATION IN, A. FILICAULIS BUS., SENSU LATO A. filicaulis, sensu stricto, was first described by Buser (1893). From it he separated ‘Une f. vestita. As his description is rather long and not easily accessible the following shortened translation is given : Plant small to medium sized, rhizome medium thick. Lower part of stem and petioles usually with spreading hairs; upper part of stem and pedicels glabrous. Leaves variously clothed with spreading hairs often only on the veins below and folds above. Earliest petioles and leaves often + glabrous, later petioles with spreading hairs. Base of the petioles and stipules usually wine-red. Leaves medium (3-9 cm wide, 2-5-8 cm long), + reniform in outline, with wide basal sinus, somewhat blue-green, paler below, lobes 7 or incompletely 9, round or + triangular, toothed all round; teeth (9—) 11-13 (-15), sub- acute, somewhat connivent, all + equal except the small terminal tooth. Stems (5—) 20-30 (-40) cm, ascending, slender. Inflorescence medium, flowers large, 3-4 mm, clusters not very dense. Of the forma vestita Buser continued ‘(la plante entiere plus ou moins velue) a été constaté en outre dans les Cevennes, en Normandie, au Jura (Reculet). Dans ces localités la forme semble se présenter comme une race indépendante, a l’exclusion du type, tandis qau’au Saleve on ne la rencontre qu’en individus égrenés et rares parmi la forme normale fréquente.’ Raunkiaer (1906) raised this forma to specific rank, no doubt impressed by its frequency in Denmark. 1. VARIATION IN HAIRINESS (a) The flowering stem It has already been indicated that the hair cover on plants of A. filicaulis, sensu lato, varies from dense to slightly hairy. A measure of this hairiness is most conveniently made on the flowering stem, where the density of the indumentum usually decreases from the base upwards. This change, however, may be slight, in which case the plant will be either densely hairy throughout, or with all but the basal internodes glabrous. The hairiness of the flower-stems was measured by direct visual comparison of the density of the indu- mentum on the internodes with a set of six standards. This method was of sufficient accuracy, and easier to apply than an analysis based on the number of hairs per unit length of stem, which would have wasted too much time obtaining a degree of exactitude which was not necessary. (i) Method The flower-stems were classified in the following way. Standards were set up to provide six classes of degree of hairiness of the stem in Fig. 1. * In his latest paper, Rothmaler (1962) treats these three taxa as varieties of A. filicaulis. Watsonia 5 (5), 1963. 306 MARGARET E. BRADSHAW | Fig. 1. Classes of hair density on internodes of the flower-stem. Drawing based on a photograph (x 4). Internodes II, IV, VI and VIII were thus scored for hairiness according to this scale, other internodes being scored where necessary. Since some flower-stems have more internodes than others comparison was made over the lower six internodes only. The grade of hairiness of the flower-stem corresponds to the decrease in hairiness of the inter- nodes of the flower-stem, and since internode II is usually densely hairy, it was assumed to be class 5 in all cases. Therefore, the grade for each flower-stem was calculated as the difference between the class of the VIth internode and class 5. In Fig. 2, in grade 5 the Vith internode is glabrous, the difference between this class (0) and the most hairy state (5) is 5; in grade 0, all internodes are class 5, so that there is no difference between the class of internode VI and the most hairy state (5). This gives seven grades of hairiness of the flower-stem (0-6). It should be noted that in grades 0-4 the internodes above the sixth (VI) may be hairy; in grade 6, internode V as well as VI is glabrous, internodes below the fifth (V) may be glabrous; also, in some cases the most hairy internode may be less hairy than class 5. 1 2 3 + 5) 6 Grade of hairiness of the flower-stem Hair density on internodes Internodes Fig. 2. Diagrammatic representation of the seven grades of hairiness of the flawer-stems. Watsonia 5 (5), 1963. VARIATION IN ALCHEMILLA FILICAULIS 307 (ii) Variation in the grade of hairiness of the flower-stem of individual plants (i.e. variation within a plant) Preliminary studies showed that a slight variation may occur in the grade of hairiness of two or more flower-stems on the same plant. For example, the grade of hairiness may be 0 and 1, or 3, 4 and 3, in both plants the variation is one grade; those plants with 0, 1, 2 or 1, 1, 3 show a variation of 2 grades, etc. Analyses of 353 plants are shown in Fig. 3. In the whole sample, 83% showed either no variation or a difference of only one grade between the two or three flower-stems which were scored. When the more glabrous (Grade 6) and more hairy (Grade 0-3) plants were plotted separately, it was clear,that most variation occurred in the more hairy plants. 100 100+ @ 90 90 80 80 70+ @ 70 60 Percentage of sample Bb Percentage of sample ol r * ch ea ae aCe fees aC 0 2 3 Number of grades of hairiness Fig. 3. Variation in grade of hairiness of two or more flower-stems on the same plant. Crosses, A. filicaulis, sensu lato (353 plants); dots, grades 0-3 (163 plants); diamonds, grade 6 (72 plants). (iii) Variation from year to year Number of grades of hairiness Fig. 4. Variation in grade of hairiness of the flower-stems on the same plant from year to year. Crosses, A. filicaulis, sensu lato (102 plants); dots, grades 0-3 (82 plants) ; diamonds, grade 6 (8 plants). This analysis was made on 102 plants grown in the garden for three (in some cases two) years. In most cases it is based on the mean of three flower-stems per plant in the final year, and one flower-stem per plant in the other year(s). (Fig. 4). Whilst 74°% show variation of no more than | grade, 26% have greater variation; this is somewhat higher than was found within the plants and no doubt is in part due to the analysis being based on a single flower-stem from each plant in the earlier years, thus giving an over-estimation of the difference. In both of these analyses the more hairy plants, grades 0-3, show greater variability than the less hairy grade 6. Watsonia 5 (5), 1963. 308 MARGARET E. BRADSHAW From the above observations on plants grown under similar conditions in the garden, it is clear that there is some fluctuation in the density of the indumentum between the several flower-stems of each plant and on the same plant from year to year. The less hairy plants are least variable, but even in the more variable hairy plants the oscillation is over only a portion of the total range of hairiness. Thus this method of analysis does provide a means of discriminating between the various degrees of hairiness of the flower-stem found in A. filicaulis, sensu lato. (iv) Variation within families Offspring of plants of different hair density were raised from seed. As far as possible, 20 plants of each family were grown in boxes of John Innes Compost No. | containing 10 plants each. Three mature flower-stems were collected from each plant and scored for hairiness by the method described. Results are given below (Table 1). TABLE | Variation in hairiness of the flower-stems within families. (grade mean, usually of three flower-stems per plant) Hairiness grade Code Frequency of each grade of hairiness number | | | in each family | Parent Family | 0 1 2 3 4 5 6 4/38/371 0 | 0-65 | Tue 13 4/28/326 2 0-8 | 4 18 1 4/11/1116 | 2 1-1 3 11 6 4/9/101 | 2 1-2 2) soe 4/3/45 2 1:4 1 10 f 1 4/30/318 3 ilS9/ 1 ‘| 10 2 4/39/269 | 1 1-85 5 13 y) H | 1 2-0 5 Seats 4/30/316 | 3 | 2-25 | Dae atl 7 4/39/272 | 5 | 4:8 4 ile 4/8/84 | 6 | 5°8 4 16— 4/15/216 | 6 5:9 2 37 4/16/207 6 | 6:0 | 10 Sc. 8 | 6 6:0 13 4/10/109 | 6 6:0 | 5 52/1 | | 1-0 | Sa id 4 In line with the earlier observations, least variation occurred in the more glabrous families with means of 5-8, 5-9 and 4-8 (families 4/8/84; 4/15/216 and 4/39/272). In addi- tion three small families of 5, 10 and 13 plants respectively and means of grade 6 showed no variation. Other nearly uniform families are at the other end of the scale with means of 0-5 and 0-65 (families 4/11 and 4/38/371). The families with means between 0-8 and 2-25 were much more variable, even containing plants of three or four grades of hairiness. One would have liked more examples of families with the grade mean of 0, 4 or 5. The range of variation found in all these families is of the same order of magnitude as that found within individual members of the families: in no case does it exceed the maximum range found in individuals. Watsonia 5 (5), 1963. a a a meena canteen tame nr VARIATION IN ALCHEMILLA FILICAULIS 309 — (vy) Variation in hair character of parents and offspring In the same families, comparison was made between the hairiness of the parent (the mean over a number of years) and the variation found in the families (as above). As before the greatest similarity was found in the more glabrous families (4/8/84; 4/15/216 and 4/39/272). The one family of a grade 0 parent showed a shift of the mean towards grade |, but other families of parental grade 0-3 are those with a wide range in the offspring. This spreads over 2—4 grades usually ranging about the mean of the parent, but rarely is the mean of the offspring the same as that of the parent. Care must be taken when drawing conclusions from these breeding experiments because the most hairy (0) and intermediate grades (4-5) are represented by only one family each. It is clear that determination of the hairiness of the flower-stem in A. fi/icaulis, sensu lato is largely genetical. In the least hairy plants this is within quite narrow limits. Family 4/39/272 is of particular interest; this is a plant of the dubious intermediate grade. Here the hair character is maintained in the offspring with little variation and shows very well that the intermediate character is not improbably genetically determined. Results of breeding from grades (0—) 1-3 suggest that these grades should be grouped together; the method of analysis reveals an amount of variation which may be genotypic but is more probably environmental in origin. This is supported by the evidence of wide variation within single plants and the fluctuation above or below the grade of the parent. At this stage it is convenient to correlate the grades used in classification of the flower- stem with the taxonomic treatment recommended at the beginning of this paper. Grades 0-3 may be regarded as subsp. vestita and grade 6 as subsp. filicaulis; grades 4-5 are intermediate in character. (vi) Variation in plants grown under field and garden conditions The preceding observations have been made on plants grown under garden conditions. Investigations were also made on the hairiness of plants as collected in the field and after a period of cultivation in the garden for one or more years. These 30 plants gave 57% showing no variation and 33% with variation of | grade; the remaining 10% varied 2 or 3 grades; this suggests that on the whole variation with change of growing conditions is within the range found in individual plants. Certain rather sparsely hairy plants do require special mention; some of these with glabrous or nearly glabrous flower stems in the field produced more normal, grade 6 flower stems after cultivation. These were all plants of high altitudes and will be discussed fully after variation in hairiness of the leaf has been considered. (b) Other parts of the flower-stem Variation in hair density is not limited to the main branches of the flower-stem only. On the pedicels (ultimate inflorescence branches) hairiness is of the same nature as the rest of the branches, being absent in the more glabrous grades. Hairiness of the urceole shows a similar series. Urceoles in the more hairy grades are usually also densely hairy. In other grades the density decreases, but rarely are all the urceoles on the flower-stem without hairs, though in grade 6 there may be no more than one or two hairs on a few urceoles only. (c) The leaves Variation occurs in the hairiness of the petiole and of the upper and lower leaf surfaces. No detailed analysis was made on the hair density of the petiole but a series from densely hairy in the grade 0-3 plants to sparsely hairy or almost glabrous in grade 6 does occur. More detailed work was done on the lamina of the leaf. Mature summer leaves were compared. Only in the most densely hairy plants are the hairs apparently evenly distributed over the surfaces of the leaf; in less hairy leaves decrease of hair cover is localised so that Watsonia 5 (5), 1963. 310 MARGARET E. BRADSHAW hairs persist on the folds and ends of the lobes of the upper surface and the veins and basal lobes of the lower. Such a distribution of hairs cannot be measured quantitatively, so a series of standards had to be used. These were chosen for upper and lower surfaces of the leaf as in Fig. 5. as Fig. 5. Classes of hairiness of upper and Jower leaf surfaces. In classes 4 and 5 hairs are shown on a portion of the leaf only. Preliminary investigations of the hairiness of both leaf surfaces of plants of different grades of hairiness of the flower-stem showed a close correlation between the lower surface and the flower-stem grade (Table 2). TABLE 2 Correlation of flower-stem hairiness. Mean values for class of leaf hairiness of several plants with flower-stems with the same grade of hairiness. Leaf Flower-stem |—— i grade Lower surface | Upper surface 0 4-4 4-4 1 | 3-0 3-4 2 | DT) PETS 3 Be] 3°3 4 t75 PTS 5 cS 2-7, 6 1-1 gi] Watsonia 5 (5), 1963. VARIATION IN ALCHEMILLA FILICAULIS Sit Further comparisons between flower-stem and leaf hairiness (lower surface) were made on cultivated plants. Observations on three leaves per plant showed very little variation within the plant, even in the hairy ones. Means of these supported the preliminary findings of correlation between the hairiness of the flower-stem and the leaf, e.g. in family 4/38/371 of flower-stem hairiness 0-1, leaf class was consistently 4; i.e. maximum flower- stem hairiness with near maximum leaf hairiness. In Sc. 8 and 4/15/216, both of grade 6 flower-stem, leaf hairiness is of minimum density. Results of scoring 218 garden and field plants assessed on one flower-stem and one leaf are given in Table 3. These results support those from the families at both ends of the series and, as with the hairiness of the flower-stem, greater variability was found in the intermediate grades. TABLE 3 Correlation of flower-stem hairiness and hairiness of the lower surface of the leaf. Grade of hairiness of the flower-stem 0 1 2 3 4 5 6 oy I 5 41 Ses 2 3 3 2 9 16 = Se 3 I 9 14 12 8 12 6 = See 12 16 10 5 5 + Sa 5 6 13 4 2 gx s oe | Total 19 38 31 22 15 30 63 Certain almost glabrous plants can now be considered. These formed a very small proportion (16) of all the plants collected in the field. In most cases the lower surfaces of the leaves were hairy only on the nerves; the flower-stems were glabrous, including all or most of the urceoles. All were very dwarfed when collected and of high montane habitats, including very wet flushes. After cultivation for 2-3 years they increased in size and all but one developed the hairs typical of the more glabrous variants of A. filicaulis, sensu stricto. In plant 6/28/201 the petioles and leaves became hairy as in A. filicaulis, sensu Stricto, but the flower-stem remained glabrous. This was the only plant in which the absence of hairs appeared to be genetically determined. Details of variation in five plants are given in Table 4. 2. VARIATION IN LEAF SHAPE From the beginning of these investigations it was apparent that there was considerable variation in the shape of the leaf, especially in that of the lobes; but this was difficult to measure and too variable to assess qualitatively. Only when leaves of families were available did the genotypic nature of the variation become clear. Three mature leaves were collected from samples of twenty plants of different families. Though the general shape of the leaf in A. filicaulis, sensu lato, is reniform, the angle of the basal sinus does vary a little; of greater significance is the variation in the length of the lobes, most obviously in the median lobe. This may be broad and shallow or long and narrow or even with a truncate end. Leaves of five representative plants from four samples of twenty raised from different parent plants are shown in Fig. 6. The relative length of the lobe decreases in the order 4/9/101, 4/11/116, 2/1, 4/30/316. The great similarity within each family clearly shows the genotypic nature of this character. This range is shown in plants of the grade 0-3 hairiness range; from general observations it is known Watsonia 5 (5), 1963. 312 MARGARET E. BRADSHAW TABLE 4 Hairiness of five dwarf montane plants when collected and in cultivation Leaf hairiness | Flower-stem hairiness Code no. Date Petiole | Lamina Internodes | Urceoles lower upper eo IVAW ae 6/22/172.| 1956 f | class 1 class 1, distal 2/3 nerve | class 2 | Class 00) Om OO 'few on most | 1957 | class 2-3* | class 1, distal 2/3 nerve class 2 Classsa aml ee Ona) few 1958 g) — — —— (1 sGlass 7 Oo4. a SiaaO, 0 — 6/28/201 | 1956 f | class 1 | class 1, distal 2/3 nerve | class2 | class 0 O° most hairy | 1957 g| class 2-3* | class 1, distal 2/3 nerve | class 2 | Class. 0" 0. 0? aifeweentmias: | 1958 g — | — | — | “sclasse 0) 22 (0) “Os s0080)g | | class 0 0. Or - | | | | class 0 0: 230 | | | | | | 6/28/202 | 1956f| class1 | class 1, distal 1/2 nerve) 1/few | class 0 0 0 | few on most | 1957 g| class 2-4* | class 1, distal 2/3 nerve} class2 | class . 1 0 0 | very few 1958 g} class4 | class 1, distal 2/3 nerve} — | class . 3 2 On, | | | class 3 3 1 ae | | class 3 0 0 | 6/28/203 | 1956f| O hairs | class 1, distal 1/4 nerve 1/few | class. --"0' 4) sOetaaO 0 1957 g| class 3—5* | class 1, distal 3/4 merve| class2 | class . 4 1 0°. O | fewonsome 1958 g| hairy | class 1, distal 3/4 nerve — |. Gass 4 A eat = | | class 4 2. les Og 6/26/181 | 1956f| class 2 | class 1, distal 3/4 nerve ; class 3 Class: “12 MOGey AG | 20 1957 g| class 5 | class 2, distal 3/4 nerve | class 3 Classi Ale 2a 0 | few on most * = 3 leaves te Rte Wee te Wate te ete HA He He VE, ete HeHe Hy we Ve de fe He fp —sfielGs ss 2).——seardent 4M 16 49.10! Fig. 6. Variation in leaf-lobe shape between families. Three leaves from each of five plants in each family. Watsonia 5 (5), 1963. VARIATION IN ALCHEMILLA FILICAULIS Sie that a range of leaf shapes is also found in plants of the other hair groups. It would appear that this variation is quite independent of the variation in hairiness. A preliminary analysis of the variation in the shape of the leaf-lobe has been made. Measurements were made of the median lobe of a well grown leaf; though the ratio length/(4 greatest breadth) makes no allowance for the degree of curvature of the sides of the lobes, it does differentiate between the long and narrow, and short and broad lobes. Analysis of c. 100 herbarium specimens showed that there is considerable variation in lobe shape; the values obtained were between 0:58 and 3:26 with a mean of 1:36. Although the sample was not random, it showed there is considerable variation in the lobe size throughout Britain, without local concentrations of any one size. TABLE 5 Variation in shape of the median lobe of the leaf in several populations. ) } Population | No. of plants | Length of lobe code no. | 4 breadth. mean Seas 4/39 6 Oni. 0-11 mixed 4/44-52 | 9 0-9 0-04 4/9-11 10 1:03 0:07 4/22 7 ile] 0:16 4/3 9 licl 0-20 4/8 9 ile 0:21 5/48 | 10 1-18 0-06 5/45 | 12 12 0-13 Values for some populations of uneven size from different habitats but grown under the same garden conditions for one or two years are given in Table 5. The range of variation within the populations is usually wide (e.g. 0-7—1-7), but some populations are less variable and centre about different means: e.g. the mixed collection with 0-9 + 0-04 and 5/48 with 1:18 + 0-06. Whilst the first two samples (Table 5) are from long rough grass at 1300 ft and c. 350 ft the last two are from high altitude pasture at c. 2000 ft. Further investigation may show that leaf-lobe shape is associated with habitats. 3. VARIATION IN STIPULE COLOUR In A. filicaulis, sensu lato, the base of the petiole is usually wine-red due to antho- cyanin; the colour may extend along the stipules and usually occurs on the rhizome. Occasionally concolourous forms without the pigment are found. Cultivation and breeding have shown that the presence or absence of anthocyanin is genetically determined. All the offspring of three concoloured plants, (60, 20, 20 plants respectively) were without the pigment. Offspring of pigmented plants were coloured but the intensity varied considerably. This variation is found in both subspecies in A. filicaulis, sensu lato. It was noticeable in wild material that the deeply coloured plants were most frequently found in long grass, whilst plants of open or grazed habitats were more variable and usually less deeply coloured. Experiments have supported these observations. Samples of seedlings from several plants were grown under the following conditions : (a) in grass which was allowed to grow tall (meadow). (b) in grass which was kept short with a lawn mower. (c) open soil in the garden. (d) boxes (see page 308). (e) as (d), these plants were kept short (2:5 cm above soil level) by cutting throughout the summer. Results are given in Table 6. Watsonia 5 (5), 1963. 314 MARGARET E. BRADSHAW TABLE 6 Variation in colour of the base of the petiole Code number Garden plot Boxes subsp. vestita a (20 pl.) DAQOipl)) 1) ne eU pI) d 40pk), | eCopt 4/3/45 | 3 Y=, 2 De Bk 3 oe 4/12/173 Fetes lS") petal oy alle ee DQ, 2D | 4/39/270 | Lest Q0piyae 2 4/30/318 | | Des 2—,2 4/9/101 | | fee | 2 subsp. filicaulis | | | 4/15/216 | 3 LZ a | 26 | 24° Gopal) | | | a tall grass d open soil pl. = plant 2 = intermediate b short grass e short grass 1 = no wine colour 3 = deep wine-red c open soil This shows that the wide variation in the intensity of the wine colour is correlated with the conditions in which the plants are grown. In the shade produced by closely associated plants (a and d) the bases of the plants are deep wine-red (3). Under simulated grazed conditions (b and e) the colouring is slight - maybe only a very little on the rhizome or stipules, rarely is it completely absent. Widely spaced plants grown in open soil (c) were intermediate in colour. It seems probable that the quantity of pigment present is correlated with the amount of light which reaches the base of the plants. Concolourous plants are readily recognised in rough grass, meadow and woodland habitats; they are difficult to determine in open or grazed habitats, since genetically deter- mined coloured plants may be so lacking in colour that only careful examination, preferably followed by cultivation, will separate the concoloured genotypes. They are very difficult to determine in herbarium material. Only a few concoloured plants have been confirmed during these studies, and then occurrence is sporadic. Map reference No. Nr. Malham, Yorks. several plants 34/68 Scarborough, Yorks. 1 plant 54/08 Upper Teesdale | plant 35/83 S. Scottish Uplands 3 plants 36/0612 Central Highlands 1 plant 27/6544 W. Pennines 2 plants 35/7131 Only around Malham is this form known to be at all frequent. Further cultivation experiments on grazed plants will undoubtedly show that the concoloured plant is more widespread. 4. MORPHOLOGICAL DELIMITATION OF SUBSP. FILICAULIS, SUBSP. VESTITA AND THE INTERMEDIATE The subsp. filicaulis and subsp. vestita have been separated by the single character of hair density, in particular that of the inflorescence. This hair density has been shown in the preceding observations and cultivation experiments to be genetically determined; but it is subject to variation which is most marked in the more hairy plants. On the basis of these data it is now possible to define more exactly the limitations in the morphological characters of the two taxa and their intermediate. This, of course, is based on the hair character only, since there is no evidence that the observed variation in leaf shape is in any way correlated with the variation in hair density. Watsonia 5 (5), 1963. VARIATION IN ALCHEMILLA FILICAULIS 315 Subsp. filicaulis usually has flower-stems with hairiness grade 6 and the lower surface of the leaf class | or 2. Thirty plants in cultivation and 92% (of 203 plants) of the offspring of type 6 parents had this grade of hairiness of the flower-stem, i.e. with internode V glabrous. The remaining 8% of the offspring of type 6 parents had some hairs (class 1) on internode V. It is not yet known if the hairiness of these plants is environmentally induced or genetically determined. Internode IV is often glabrous or with a low hairiness value (class | or 2). Usually internodes II and HI are densely hairy and so the abrupt change between III and IV is very noticeable; at first, it was thought that this was charac- teristic of subsp. filicaulis but this is not so, as these internodes may be less hairy. No attempt has been made to subdivide the less hairy forms within subsp. fi/icaulis as in most cases these appear to be environmentally-induced states (p. 311). The leaves are very uniform in hairiness of the lower surface, which is hairy only on the nerves or also on the lowest lobes; a very small proportion are thinly hairy all over. From the preceding data over 90% of the plants referable to subsp. fi/icaulis are separable from other plants of A. filicaulis, sensu lato, on the basis of flower-stem hairiness supported by leaf hairiness. Subsp. vestita forms the other end of the hairiness scale. All observations have shown a large amount of variation of an environmentally determined nature but it is not possible to say whether the wide variation seen in the families is mainly the same or partly genotypic. It seems best to include in subsp. vestita all those plants with flower-stems of hairiness grades 0-3; this includes plants densely hairy throughout, and those moderately hairy (class 2) on internode VI and maybe glabrous in the upper branches. Though it is apparent in the breeding experiment that the most hairy plants are a little less variable than the others, there is no reason why these plants should be considered to be anything more than equivalent to the most glabrous members of subsp. filicaulis. As in the flower-stem the leaves show a greater range of variability than in subsp. fi/icaulis; in 110 plants hairiness of the lower surface of the leaf was : class 5 in 25 plants class 4 in 43 plants class 3 in 36 plants class 2 in 6 plants Thus 95 % were within the more hairy half of the scale (classes 3-5) class only 5% were of the most glabrous classes (1 and 2). The intermediates between the lower limit of hairiness of subsp. vestita and the upper limit of subsp. filicaulis are those plants with flower-stems of hairiness grade 4 and 5. These are not so frequent as the other genodemes. In cultivation the grade of hairiness of the flower stem was maintained over several years with slight variations between grades 4 and 6. It was only possible to raise offspring of one plant of grade 5; all these were of grade 4 and 5. Although so near to subsp. fi/icaulis, no plant had a mean of grade 6 though single flower-stems on some plants were this grade. (The leaves are also intermediate in the class of hairiness of the lower surface). 5. VARIATION IN THE HABIT OF A. FILICAULIS, SENSU LATO Considerable variation in the size of plants of A. filicaulis, sensu lato, was very obvious in field samples collected from grazed pasture, meadow and wood over a wide altitudinal range. Walters’ recognition (1949) of A. minima as a dwarf species opened up the question of whether in fact these other assumed dwarf variants of more ‘normal’ species were environmental dwarfs or not. A preliminary experiment to test this was set up. Habitat samples were obtained mainly from close-grazed pasture, but also from long grass and woodland, from nine sites between 450 ft (138 m) in East Durham and 2,500 ft (723 m) on the Pennine escarpment Watsonia 5 (5), 1963. 316 MARGARET E. BRADSHAW near Cross Fell to the west, all about latitude 54° 40’N; and from 2,000 ft (610 m) on Ingleborough where A. minima grows. As far as possible equivalent-sized pieces of un- branched rhizome of ten plants from each habitat were planted in boxes of John Innes Compost No. | in March, 1957. The plants were grown under similar conditions until August and then scored for the following characters : 1. number of rhizome branches. 2. number of flower-stems. 3. length of the petioles of three well-developed leaves. 4. breadth of the lamina of three well-developed leaves. 5. length of the three longest flower-stems. In 3, 4 and 5 the mean value of three measurements per plant was used. Results are shown as graphs in Figs. 7 and 8. = — — Now -« ° SUS ==: —_— pn — i) —~@— f LL oe 8 oo + St i ERhg:| « aclaiay eles 33 Sia ey + aS 4 4 3 2: il Oi a OS es Se LE SAS SR ee pee 1000 ft. 2000 ft. 3000 ft. Altitude 305 m 610 m 915m Fig. 7. Variation in breadth of leaf and length of petiole of population samples from various altitudes after cultivation. Leaf measurements, plain crosses; petiole measurements, solid dots (pasture) and open circles (meadow). (Standard deviation is given as a vertical line above and below the mean). The sample from Ingleborough differed greatly from the others in the high number of flower stems (15-7) and short petioles, which are characteristics of A. minima. It was surprising that all the plants in this sample were of this kind; and, since no A. filicaulis, sensu lato, plants were present, this sample will be considered no further here, but will be included in a later paper on A. minima. In the other samples, although the dwarf plants attained a greater size in cultivation than in the field, a correlation between size of plant (as measured by petiole length and leaf size) and altitude was revealed. Fig. 7 shows the steady decrease in size found in the pasture plants as the altitude increases; the absence of pasture samples between 1,350 ft (411 m) and 2,000 ft (610 m) emphasises the difference between the lowland and montane populations. Additional correlation exists between the numbers of shoots produced per plant and altitude (Fig. 8). This is positive, the plants of high altitudes producing more shoots than those of low altitudes; variation within samples was greater here but significant differences were found between high and low samples. Watsonia 5 (5), 1963. . i VARIATION IN ALCHEMILLA FILICAULIS 317 | Number of shoots per plant 0 phatase eave octal ee 1000 ft. 2000 ft. 3000 ft. Altitude 305 m 610 m 915 m Fig. 8. Variation in the number of shoots produced per plant after cultivation of population samples from various altitudes; solid dots, pasture, open circles, meadow. (Standard deviation is given as a vertical line above and below the mean). The two samples from meadows show the same correlation with altitude, and somewhat surprisingly fit with little distortion into the curve of the pasture plants. A single woodland sample was not included in the graphs because of the greater difference in the nature of the habitat. The number of shoots and breadth of the leaves are very similar to those of samples from the other habitats at 950 ft (290 m); the petioles are slightly but not signi- ficantly longer. The number of the flower-stems per plant has not been considered any further because there did not appear to be any obvious correlation with any of the other factors. The combined effect of these variables is that the montane plants form shorter, more dense clumps, whilst the more lowland plants are taller, less branched and hence less compact. Two of these pasture samples from 2,500 ft (723 m) and 950 ft (290 m) have been grown in a garden plot for several years. The low, dense growth habit has been maintained by seven of the ten plants from 2,500 ft (723 m); the other three became taller and of lowland habit. This was revealed in the first year in the plot and has been maintained. Subsequent inspection of the data used in Figs. 7 and 8 did, in fact, show that these plants had leaf measurements slightly greater than the means of the samples. All plants in the 950 ft (290 m) sample maintained the lowland type of growth. Seed progeny samples have been grown, and were analysed in a similar way in their second year. Table 7 shows that the characteristics of the parents are maintained in the families. Families 1, 2 and 3 from the highest altitudes show the characteristic low dense habit. Those from 2,000 ft (610 m) and 1,350 ft (411 m) have maintained the same low growth for four years but had produced fewer shoots per plant than families 1—3 in two years. The number of shoots was not recorded in later years. Although family 6 is from 1,800 ft (549 m) it is of more lowland form and taller in cultivation. The lowest pasture family (7) from 1,200 ft (366 m) is of lowland stature. Four families from a lower pasture and other habitats maintained the lowland form of their parents. These results confirm that the habit characters are genetically determined. Watsonia 5 (5), 1963. 318 MARGARET E. BRADSHAW TABLE 7 Analysis of the number of shoots per plant and leaf size in several families | | | | | Height Code Ales No. of Shoot | Petiole length | Leaf breadth | after IND Oe GER) | Habitat | plants no. | (cm) | (cm) 4 years _*Parent | Offspring | * Parent. Offspring | | “1957 Jl 1959 O57 ee 959 | 5/47/267 | 2,400 | pasture | 20 | 98+0-41; 3:1 7440-28; 3:5 3:9 + 0:29 |, ‘shore 1 2 | 5/46/261 | 2,400] pasture | 10 | 102+40-76; 3:5 | 554040) 3-6 | 3-9 +0-15| short 3 | 5/6/46 | 2,000) pasture | 20 | 1234034) 28 | 684048) 3-8 | 4:0+0-07| short 4 | 5/45/247 | 2,000} pasture | 10 | 3:540-45/ 1:6 | 6740-34) 2:7 | 5-4+0-43| short 5 | 5/1/13 | 1,350! pasture | 20 | 5:-4+0-30} 2:3 | 69+0-32| 3-4 | 4-5 +0-21| short 6 | 5/4/29 | 1,800| pasture | 20 | 484034; 5:8 | 754033) 49 | 5-4 + 0-21 | medium 7 | 4/33e 1,200 | pasture | 20 | 6:3+0:53| 10-0 |11:1+0:50; 5-4 | 61+0-12| tall 8 | 4/15/216| 950| grass | 20 | 3:9+0-32| 11-5 |10-3+0-34| 7:3 | 6740-40] tall 9 | 4/39/272 | 1,300; meadow} 10 | 3:740-42}] 10-0 |10:1+0-34] 7:2 | 7440-13] tall 10 | 4/30/318 |; 950} meadow} 20 | 4:7 + 0-44 not available | tall 11 | 4/8/89 950! wood | 20 | 4940-22) 12:2 |11:0+0-88; 7-4 | 68+0-56| tall *The measurements were made in different years, therefore comparative values between the parents and between the families should be considered, not the direct parent-family value. This variation in size and habit is closely connected with environmental factors; as is to be expected the 4. filicaulis, sensu lato, plants are small in closely grazed habitats; so far, in all lowland samples, this character appears to be only environmentally induced, but at the higher altitudes probably the majority of the dwarf A. filicaulis, sensu lato plants of the close grazed montane habitats (above 1,500 ft, 458 m) are the dwarf, much-branched variant. The cline in size and habit shown in Figs. 7 and 8 and Table 7 is most obviously correlated with altitude and hence, the associated factors of climate, temperature, exposure and their effects on the soil. In the pasture habitats the grazing pressure is very important. All the montane habitats are very much more closely grazed by sheep than the lowland pastures, and may have been subject to this close grazing for a long time, probably several centuries. It is difficult to assess the relative importance of the biotic and altitudinal factors in the evolution of this variant. The low-growing, much-branched variant is so obviously well adapted for survival! under high grazing pressure that it appears to be axiomatic that grazing pressure has been of great importance. This is confirmed by field observations on rock-ledge plants protected from grazing which are similar in size and habit to the more lowland plants. It seems reasonable to assume that altitudinal factors also have been of some importance in the development of the cline, since variation is continuous over the whole altitudinal range of samples from the long grass habitats as well as those from the grazed habitats. The gradual variations found in these two characters, size and shoot number, form clines similar to those known in sexual species; for instance the well known example of Achillea lanulosa, investigated by Clausen, Keck & Hiesey (1948), which shows a similar correlation between plant height and altitude, forming several ecotypes in isolated localities between c. 4,000 ft (1,220 m) and 10,000 ft (3,050 m); or Plantago maritima where Gregor (1939) recognised ecoclines related to ecological gradients. In A. /anulosa the ecotypes are geographically and morphologically isolated but in P. maritima where variation is continuous, separation is difficult, and each end of the cline may be regarded as an ecotype. In Alchemilla filicaulis, sensu lato. the evidence so far obtained suggests a parallel with the ecoclines found in P. maritima, with the possibility of a separation into a dwarf montane and a tall ecotypic variant or ecogenodeme. Watsonia 5 (5), 1963. VARIATION IN ALCHEMILLA FILICAULIS 319 The existence of genodemes within the species in Alchemilla vulgaris agg., similar to the ecotypes in sexual species, was first demonstrated by Turesson (1943). The initial work, based on eleven species, was qualitative and showed variation in morphological and physiological characters in several species, from widely spaced localities in Sweden; those within A. filicaulis differed in habit, time of flowering and mildew resistance as follows: | | Type | Habitat Flowering | Mildew Lapland ' small low tussock earliest | very severely damaged Scania | compact well formed tussock latest | severely damaged Uppland | Ks | intermediate is Smaland loose spreading tussock | ip | less attacked Gotenburg large, loose spreading | late | slight Photographs of two types were given. Turesson regarded this species as the most variable of those which he studied. Later work on seed progenies of some species has confirmed the genotypic nature of some of these characters (Turesson 1956). In the earlier work Turesson introduced the term ‘agamotype’ for the different genodemes, which he regarded as equivalent to the ecotypes of the sexual species. This work revealed the genetic heterogeneity of species, which had been regarded as uniform and monotypic apomicts. Recently, Grun (1955) has shown that morphological variation also occurs in the obligate apomict Poa nervosa. ACKNOWLEDGMENTS I am pleased to record my thanks to Professor D. H. Valentine and Dr. S. Max Walters for encouragement and their interest in the work presented in this and the subsequent papers in this series. I wish to thank the Nature Conservancy and Research Fund Committee of the Durham Colleges for financial assistance while this work was being done, Professor Valentine for providing working space in the Department of Botany, and the Laboratory staff for technical assistance. REFERENCES BRADSHAW, M. E. (1957) in Lousley, J. E., (ed.) Progress in the study of the British Flora, 106-108. London. BRADSHAW, M. E. & WALTERS, S. M. (1961). A Russian Alchemilla in South Scotland. Watsonia 4, 281-282. BRADSHAW, M. E. (1963). The distribution and status of five species of the Al/chemilla vulgaris L. aggregate in Upper Teesdale. /. Ecol. 50, 681-706. BRADSHAW, M. E., SELL, P. D. and WALTERS, S. M. (1963). The nomenclature of Alchemilla minor auct. brit. Watsonia 5, 259-261. Busser, R. (1892). Alchemilles nouvelles frangaises. Bull. Herb. Boiss. 51, Appendix 2, 18-35. CLAPHAM, A. R., TuTIN, T. G. & WaArRBuRG, E. F. (1962). Flora of the British Isles, Ed. 2, 394-401. Cambridge. CLAUSEN, J., KECK, D. & HigsEy, W. (1948). Experimental studies on the nature of species. IJ. Carnegie Inst. Washington, Publ. No. 581. Grecor, J. W. (1939). Experimental taxonomy, IV. New Phytol. 38, 293-322. GRun, P. (1955). Cytogenetic Studies in Poa. III. Variation within Poa nervosa, an obligate apomict. Amer. J. Bot. 42, 778-784. MurRBECK, S. (1901). Parthogenetische Embryobildung in der Gattung Alchemilla. Lunds Univ. Arssk. 36, Afd. 2, Nr. 7, 1-41. RAUNKIAER, C. (1906). Dansk Ekskursions Flora. Kobenhavn. ROTHMALER, W. (1962). Systematische Vorarbeiten zu einer Monografie der Gattung A/chemilia X. Feddes Repert. 66, 194-234. STRASBURGER, E. (1905). Die Apogamie der Eualchemillen und allgemeine Gesichtspunkte, die sich aus ihr ergeben. Jahrb. wiss. Bot. 41, 88-164. Watsonia 5 (5), 1963. 320 MARGARET E. BRADSHAW TURESSON, G. (1934). Variation in the apomictic microspecies of Alchemilla vulgaris L. Bot. Notiser, 413-427. TURESSON, G. (1956). Variation in the apomictic micrsopecies of Alchemilla vulgaris L. Il, Bot. Notiser 109, 400-404. WALTERS, S. M. (1949). Alchemilla vulgaris L. agg. in Britain, Watsonia 1, 6-18. WALTERS, S. M. (1952). Alchemilla subcrenata Buser in Britain, Watsonia 2, 277-8. APPENDIX—LOCALITIES Key to code numbers, a/b/c, where, a = year b= population and locality, and c = number of the individual plant. Code No. Year Locality O.S. Nat. Grid. ref. 2/1/ 1952 Craven 34/720780 4/3/ 1954 Weardale 35/964392 4/8/ 1954 Teesdale 35/922274 4/9/ 1954 Teesdale 35/921273 4/10/ 1954 Teesdale 35/861303 4/11 1954 Craven 34/910633 4/11/ 1954 Teesdale 35/921273 4/12/ 1954 Teesdale 35/930258 4/15/ 1954 Teesdale 35/922274 4/16/ 1954 Teesdale 35/926270 4/22/ 1954 Teesdale 35/866313 4/28/ 1954 Weardale 35/875388 4/30/ 1954 Weardale 35/900378 4/38/ 1954 Bishop Auckland 45/214294 4/39/ 1954 Teesdale 35/865309 4/45/ 1954 E. Durham 45/338384 5/1/ 1955 Teesdale 35/855284 5/4/ 1955 Teesdale 35/833273 5/6/ 1955 Teesdale 35/837264 5/44 and 45/ 1955 Mickle Fell 35/815247 5/46-52/ 1955 Knock Fell 35/7130 6/22/ 1956 Ben Lawers 27/658444 6/26/ 1956 Ben Lawers 27/635405 6/28/ 1956 Ben Lawers 27/6443 H 1946 Scarborough 54/08 Sc. 8 Borghamn, Sweden Watsonia 5 (5), 1963. STUDIES ON ALCHEMILLA FILICAULIS BUSER, SENSU LATO AND A. MINIMA WALTERS. Il. CYTOLOGY OF A. PITCAUIETS, SENSU “LATO By MARGARET E. BRADSHAW Department of Extra-Mural Studies, University of Durham. ABSTRACT All chromosome counts of segregates in the Alchemilla vulgaris L. aggregate are listed. Cytological preparations of root-tips, grown in water-culture, and pollen-mother-cells of A. filicaulis, sensu lato, were examined with a phase-contrast microscope. It was difficult to obtain exact counts; in both subspecies 2n = c. 101-110, and one count of a dwarf montane plant (see p. 323) gave 2n = c. 105 + 2. P.M.C.’s stained more easily, but the small and variable number of probable bivalents made interpretation difficult. One plant of A. filicaulis had 2n = c. 150. INTRODUCTION Cytological studies of the segregates in the A/chemilla vulgaris L. aggregate have proved very difficult. Although the chromosomes are small it was difficult to spread the large number sufficiently well to obtain accurate counts. Love & Léve (1961) list the counts obtained so far within the A. vulgaris aggregate. All are high polyploids and apomicts. The particular type of embryology will be described in the third paper of this series. Turesson’s counts (1957) of the A. vulgaris L. agg. appear to be the most accurate. but he was not able to give absolutely exact figures, each value being within range of the figures given. However, the accuracy is sufficient to show that some species have a number of 2n = cc. 105-106; these are: A. cymatophylla Juz., A. sarmatica Juz. and A. xanthochlora Rothm. Most of the others probably have several numbers between 2m = 100 and 110; these are: A. glaucescens Wallr. (A. minor Huds.), A. acutiloba Opiz, A. swbcrenata Bus., A. gracilis Opiz., A. subglobosa C. G. West, A. vestita (Bus.) Raunk., A. filicaulis Bus.., A. monticola Opiz., A. glabra Neyg., A. obtusa Bus., A. wichurae (Bus.) Stef., A. murbeckiana Bus. and A. glomerulans Bus. The lower values of 27 = 93 and 96 are undoubtedly counts of insufficiently spread plates, the latter an obvious multiple of 8 which has been considered to be the basic number. A higher range of numbers is found in A. oxyodonta Bus. (2n = c. 165-168), A. os- karssonii Love (= A. glacialis Oskarss.) (2n = c. 144) and A. transpolaris Juz. (= A. borealis Semis) (27 — c. 130-152). In addition to the chromosome numbers given for A. wichurae, A. murbeckiana, A. oxyodonta and A. glomerulans, the value 2n = c. 64 is reported by Love & Love (1956) and for the last by Bocher (1938) also. Differences of this magnitude are unlikely to be errors in counting, so that if the material has been identified correctly they imply the exis- tence of intraspecific chromosome races. There may be some significance in the fact that 321 Watsonia 5 (5), 1963. S22 MARGARET E. BRADSHAW Turesson’s material was of Scandinavian origin whilst Bocher’s was from Greenland and Love & Léve’s was from Iceland; though the material of Jorgensen et a/., was also from Greenland. Obviously some rechecking is needed here. CYTOLOGY OF A. FILICAULIS, SENSU LATO Method Preparations of root-tips and pollen-mother-cells were made, and both had their special difficulties. Cells of the root-tips seemed to divide rarely; the chromosomes were difficult to spread and stained poorly. The pollen-mother-cells were easier to handle and stained better, but were more difficult to interpret than the root-tip preparations. The best results were obtained from roots grown by a simple water-culture method. All except the young roots were stripped from the rhizome of a healthy shoot before it was suspended in a beaker of water. When new adventitious roots had grown, after a few weeks, several millimetres of the yellow-tipped roots were cut off and pre-treated with 0-002 M 8- hydroxyquinoline. The tubes were immersed in running water (c.15°C) for 4 hours. Root-tips were fixed in acetic-alcohol, placed in a domestic refrigerator (c. 4°C) for 3-6 days to increase the spread of the chromosomes (Walker 1956), and then stored in a deep-freeze refrigerator. It proved difficult to get well spread, deeply stained chromosomes; best results were obtained with aceto-carmine stain using a modification of the combined staining and maceration technique of Proctor (1955). The root-tip was heated in about | cc of an aceto-carmine —- N HCl mixture in the proportions 9:1 in a test-tube in a boiling water-bath for 5-7 min. Iron was introduced by standing a cleaned needle in the test-tube or earlier as a few drops of ferric acetate in the fixative. The hydrochloric acid aided maceration and hence helped to in- crease the spread of the chromosomes but had the disadvantage of reducing the stain uptake by the chromatin material. However, better definition was obtained by the use of a phase- contrast microscope. For meiosis flower buds were fixed when the inflorescences were still very compact, ‘preferably in April, though flowering continues until September or even later. The flower- buds were separated and fixed in acetic-alcohol and chloroform (1 : 3 : 1) for at least 24 hr, then stored in 70% alcohol in a deep-freeze. Meiosis takes place when the flower buds are about 2 mm long. The four anthers were dissected out into a drop of aceto-carmine, split, and as much of the wall as possible removed. At the correct stage for metaphase plates the pollen-mother-cells float out as separate round cells. These were stained in the usual way. Only a small drop of aceto-carmine must be used or the PMC will float to the edge of the cover-slip and be lost; more stain can be added from the edge of the cover-slip if air gaps remain. Suitable preparations were made permanent in Euparal. All the chromosome numbers were high (2n = c. 100). To obtain accurate counts, an enlarged photographic print was used as a base on which to ink-in the chromosomes as each was observed under the micro- scope. In this way, no chromosome is missed or duplicated, and any uncertainty about the exact number is due to inability to interpret the preparation. This method is described by Manton (1950). Results The high number of chromosomes, and the difficulty of getting them well spread, made it almost impossible to obtain exact counts; somatic counts from the root-tips are given below. Watsonia 5 (5), 1963. CYTOLOGY OF ALCHEMILLA FILICAULIS 323 (a) Mitotic counts Subspecies Code No. Locality and grid ref. | 2n vestita | 4/9/101 Teesdale SSIs |. LOS 222 | | 105+2 | 104+2 vestita | 4/24/300a Teesdale 35/861296 | 104+1 | h WO See 106 + 1 (Plate 14a) 101 +1 | 106 + 1 vestita 4/39/269a | Teesdale 35/865309 104 +2 | | 10441 | 105 + 1 | 104 + 1 vestita 4/30/318b | Weardale 35/900378 106 exact (Fig. la) vestita 5/45/250a Mickle Fell 35/815247 105 +2 (montane) EA filicaulis 3(60) n =| ~— Meal nan 27/5838 =| 15243 hy | Tarmachan a | Uses 2 (Fig. 1d) The probable margin of error is given in each case. The count for 4/30/318b is thought to be exact. (b) Meiotic counts Preparation of pollen-mother-cell squashes was relatively easier than the root-tips and stain differentiation was better, but the interpretation was much more difficult. In all cases there were many univalents, and it was not easy to determine which were bivalents or even if multivalents were present; no certain cases of the latter were found. Certain closely associated pairs of chromosomes have been queried as probable bivalents; whether they are interpreted as such or regarded as univalents makes no difference to the total chromosome number. All interpretations are made in the light of somatic counts obtained from the root-tips. In general, when the larger chromatin bodies were regarded as bivalents and the smaller as univalents, the total chromosome number obtained was near the somatic value of the species; thus this interpretation seems justified and does give some idea of the number of bivalents which usually occur. It is essential to know the somatic value when studying meiotic figures as other interpretations could easily be made; because of this all chromosome numbers are approximate only. Subspecies Code No. Locality and grid ref. 2n vestita 4/2/31 Weardale 35/963392 109? 4/45/414b E. Durham 35/338384 c. 108 filicaulis 4/3/55a Weardale 35/964392 c. 103 €. 1035 filicaulis 3(60)b Meall nan Tarmachan 27/5838 c. 150 Watsonia 5 (5), 1963. a ce 324 MARGARET E. BRADSHAW Each plate will be considered separately. In the explanatory diagrams, the univalents are outlined, the bivalents and closely associated chromosomes are blacked in. subsp. vestita 4/2/31 (Fig. 1c). This was very well squashed and is probably a late anaphase, though it could be regarded as a metaphase IJ. All the chromatin bodies are much the same size and if interpreted as single chromosomes give 27 = 109 or 110; which is near the number obtained for the somatic counts in the species and the same as that given for A. vestita by Turesson. ; : Ro ae . ° (=) $e es @ fete AS on oa 9% 4 2 ¢ Oiere 40 iNy C50 ono OF e@ e.° e 28 oO Oo Cras “2 e @ 0°C~e BF O 2 e e as see Fy @ C4 ® is eee Rte ous ES spose S& 50° &I.° 2@e PY hd 2 2 Q Oa @ Oo 3 of % e % On as € Mood % CBS. © efee ® : eeneee Oe Hh’ OO s e* e Og: e@ 2 . a 2% O CS) O oe, 2 wre oO . 200 a ,.° . panes @) 08? ) ones O % Ae % - ‘3 *e 6° : reas : @ 2 @ @, Se eo? % 6 Cc 02S b &e $ Ge oe 8 © @ a S @ es Ry oe 3 Sutene @ Se g Crs °°. a = = > 8 OY ° og offs so Sa ° x a CEQ) ese e% 3 ceteris ?b + 0° ? b> ao 28 ° aetece (eas @ Oggi: Sees we cegen oe? Wee GOSS see erh %, O° S S ete*s te ¢ o = } Se ge 4 2 aie ae ° geese af Fig. 1. Chromosomes of Alchemilla filicaulis, sensu lato. (a) subsp. vestita, Weardale (RT, 2” = 106, x 770}: (b) subsp. filicaulis, Weardale (PMC, 2n = 103, x 530); (c) subsp. vestita, Weardale (PMC, 2n = 109 or 110, x 1070}; (d) subsp. filicaulis, Meall nan Tarmachan (RT, 2” = 150 + 2, x 760); (e) subsp. filicaulis, Meall nan Tarmachan (PMC, 2” = c. 150, x 1000). Doubtful bivalents are marked ?b. For further explanation see text. Qi 08: subsp. filicaulis. ‘The only somatic count of British material is of somewhat atypical material 3(60) which is considered separately below. Turesson has published values 2” = 101 to 110 for Scandinavian plants. Interpretation of the meiotic figures of British material has been based on the assumption that this has approximately the same range of numbers. subsp. filicaulis 4/3/55a (Plate 14c). A fairly straight forward metaphase I with 20 II and 631 giving 2n = 103. Same plant (Fig. 1b) also a metaphase I plate with 2n = 103 : 18 II and 671. Although the number of bivalents varies the total number of chromosomes is 2n = 103; it is of particular interest to know if the somatic count is the same, but a countable preparation was not obtained. subsp. vestita 4/45/414b (Plate 14b). This is a metaphase plate with 21 I and 66 I giving Watsonia 5 (5), 1963. PLATE 14 Chromosome plates and explanatory diagrams of Alchemilla filicaulis, sensu lato. (a) subsp. vestita, Teesdale (RT, 2n = 106 + 1, x 1340); (b) subsp. vestita, E. Durham (PMC, 2” =c. 108, x 1310); (c) subsp. filicaulis, Weardale (PMC, 2n = c. 103, x 1020). Doubtful bivalents are marked ?b. For further explanation see text. CYTOLOGY OF ALCHEMILLA FILICAULIS S25 subsp. filicaulis 3(60)b (Fig. le). The presence of the nucleolus and indication of an outer membrane suggest that this is at diakinesis or maybe metaphase I. The diffuse nature of the chromatin bodies makes their interpretation as univalents or bivalents difficult and hence the total number uncertain. Two clusters A and B probably consist of a total of 6 or 7 chromosomes. Of the other bodies, if the larger are interpreted as bivalents, the probable values are 20 II and 1171 giving a total of 159; if all are considered as univalents the total is 144, one above and the other below the somatic value of 2n = c. 150. (c) Cytology of subsp. vestita. The mean values obtained for lowland plants of subsp. vestita were : Somatic counts 2n = 104, 105, 105, 106. Meiotic counts 27 = c. 108, c. 109. In spite of the closeness of the somatic numbers and the difficulty of getting exact counts, it is very probable that several chromosome numbers do occur in the species. Turesson (1957) obtained a range of numbers between 101 and 110 for several species of A. vulgaris agg. He gave 2n = 110 for the only two plants of ssp. vestita which he examined. A count was obtained for only one of the dwarf montane plants; this gave 27 — 105, which is within the range of the lowland plants. (d) Cytology of subsp. filicaulis The only count obtained of a British lowland plant was from a pollen-mother-cell, 2n = c. 103. Turesson (1957) found a range of numbers : 2n = LOZ iOS OA OS elOG Oi e108; 109" 110 No. of plants 2) 6 1 7 5 I 3 4 2 This overlaps the range found in subsp. ves‘ita. Most interesting and surprising is the value 2n = 150-152 obtained for a subsp. filicaulis plant, 3(60), which represents a Scottish mountain population. These plants have the usual subsp. filicaulis hair character and wine-red base, but the leaves have very short broad lobes and sharply pointed, somewhat connivent teeth, rather similar to A. wichurae; the leaves tend also to be more shiny and crisp than in other subsp. fiJicaulis plants. They could easily be recognised in the field from amongst plants of A. vulgaris agg. Whether or not this plant should be classified as subsp. fi/icaulis will be considered later. The chromosome numbers obtained by Turesson (1957) and myself for subsp. vestita and subsp. filicaulis agree well and show no major differences in the cytology of the two taxa except in the case already mentioned. Both have a range of numbers which are similar to those found in several distinct morphological species in A. vulgaris agg. (Turesson 1957). The difference in hair density which separates subsp. filicaulis from subsp. vestita is not correlated with any constant difference in the chromosome numbers of these taxa. So far no counts are available for the genodeme of intermediate hairiness. (e) Cytology of the ecogenodemes A chromosome count of one example of the dwarf montane ecogenodeme (Bradshaw 1963, p. ) is clearly an insufficient basis for any conclusions on the correlation between chromosome numbers and ‘ ecotypic’ variation. If it is representative of the dwarf ecogeno- deme then there would appear to be no major cytological differences between the ecogeno- demes. Further data on the same kind of variation are provided by Turesson (1956). He worked with six species (but not A. filicaulis), and also found statistically significant differences in stem and petiole length and leaf size between genodemes. Later (1957) he published chromosome counts of many species but it is not clear if the numbers are those of the plants used in the earlier work or not. Turesson did not give the chromosome numbers of his morphological genodemes individually ; neither is it known if the plants used in his experiment Watsonia 5 (5), 1963. 326 MARGARET E. BRADSHAW all came from the same or different types of habitat. Turesson concluded ‘ Our Alchemillas apparently do not adapt themselves to different habitats by any change in chromosome numbers. Difference in habitat is implied by the wide latitudinal range of the original samples. Only by obtaining exact counts of several plants of each genodeme will it be possible to determine if this kind of variation is correlated with differences in the chromosome numbers. REFERENCES Bocuer, T. (1938). Biological distributional types in the flora of Greenland. Medd. Gronland 106, (2), 1-339. BrapsHAw, M. E. (1963). Studies on Alchemilla filicaulis Bus., sensu lato, and A. minima Walters. Intro- duction and I. Morphological variation in A. filicaulis Buser, sensu lato. Watsonia 5, 304-320. JORGENSEN, C. A., SORENSEN, T. & WESTERGAARD, M. (1958). Dansk. Vid. Selsk. Biol. Skr. 9, 1-172. Love, A. & D. (1956). Cytotaxonomical conspectus of the Iceland flora. Acta Horti Gotob. 20, 65-291. Love, A. & D. (1961). Chromosome numbers of central and northwest European plant species. Opera bot. 5, 206-209. MANTON, I. (1950). Problems of Cytology and Evolution in the Pteridophyta. Cambridge. Proctor, M. C. F. (1955). Some chromosome counts in the European Cistaceae. Watsonia 3, 154-159. TurEsson, G. (1956). Variation in the apomictic microspecies of A/chemilla vulgaris L. Ul. Bot. Notiser 109, (4), 400-404. TURESSON, G. (1957). Variation in the apomictic microspecies of Alchemilla vulgaris L. 1. Bot. Notiser 110, (4), 413-422. WALKER, T. G. (1956). Cytotaxonomic studies in the tropical fern genus Pteris. Ph.D. Thesis, University of Beeds: Watsonia 5 (5), 1963. BOOK REVIEW A Handbook to Plants in Victoria. Vol. 1. Ferns, Conifers and Monocotyledons. J. H. Willis. Melbourne University Press, 1962. Price 45s. Od. net. With the steadily increasing interest of Australians in their native flora the provision of up-to-date Floras covering different parts of the country is becoming an urgent necessity. Floras of most of the States of the Commonwealth have been published in the past but the great majority of them are now out of date and out of print. This is true of Victoria, Ewart’s Flora having appeared over 30 years ago. There is there- fore a great need for a new Flora embodying the research carried out during these intervening years and correcting the errors in the previous account. The work here reviewed is designed to fill this gap. It will be published in two volumes, the first one of which deals with the Ferns, Gymnosperms and Monocotyledons. The Dicotyledons will occupy a second larger volume. The present work is intended to be more of a pocket handbook for use in the field than Ewart’s account. The author, who is a member of the staff at the Melbourne Herbarium, has an extensive knowledge of Vic- torian plants, both in the field and in the herbarium. He has also been able to examine many important historical specimens in this country while acting as liaison officer at the Royal Botanic Gardens, Kew. The Flora is in the form of a key, separate descriptions of the species being furnished only for species of genera which are monotypic in the State. In addition to synonymy, references to illustrations are included with each species and the distribution within Victoria and other parts of Australia is given. Many British and European plants have been introduced into Australia, a large proportion of which have become quite naturalised, sometimes locally even replacing the native flora. The English botanist will encounter many old friends in the pages of this Flora; he will also find species which he has seen as introductions in this country but which are native in Australia, for example, Juncus pallidus. Altogether Mr. Willis has provided a very useful book for the field worker in Victoria and adjacent parts of the neighbouring States, and its publication will no doubt stimulate further work on Victorian plants. It can be recommended to any plant lover visiting this State and wishing to name the plants which he encounters. The appearance of volume 2 will be looked forward to with interest. V. S. SUMMERHAYES 327 Watsonia 5 (5), 1963 British Isles A. R. CLAPHAM, T. G. TUTIN & E. F. WARBURG The first edition of the standard British Flora appeared in 1952, and has been reprinted several times. This new edition, largely rewritten and completely reset, embodies the results of research Flora of ‘the | in the last eight years. Most of the larger keys have been modified | for easier use, and accounts of introduced plants have been | extensively altered. As the Excursion Flora (1959) now satisfies the need for a pocket edition, the page size has been increased. SECOND EDITION. 70s. net from all bookshops CAMBRIDGE” UNIVERSE, Vise Ress IRISH NATURALISTS’ JOUKINE A Magazine of Natural History Published Every Quarter by the I.N.J. Committee. Edited by Miss M. P. H. KERTLAND, M.Sc., with the assistance of Sectional Editors. Annual Subscription, 10/- post free. Single Parts, 3/6. | All communications to be addressed to :— The Editor, Department of Botany, Queen’s University, Belfast. INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the taxonomy or distribution of British vascular plants or Charophytes are invited from both members of the Society and others. They should be typewritten, with wide margins, double spaced, on one side of the paper only; contributors are recommended to keep a carbon copy of their typescripts. The form adopted in recent parts of Watsonia should be used for layout, headings, citations and references. Con- tributors are urged to avoid very complicated hierarchies of headings and sub-headings, and to check carefully the consistency of those that they use. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British vascular plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (N.H.) and B.S.B.I., 1958), and may then be cited without authorities. Otherwise, authors of names must be cited, at least on the first occasion where they appear in the text. Except for citations of the place of publica- tion of plant names, full references should be listed in alphabetical order of authors names at the end of the paper. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals, ed. 3 (London, 1952). References to herbaria should include the abbreviations given in British Herbaria (B.S.B.1., 1957) and Index Herbariorum ed. 3 (1.A.P.T., 1956). Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new taxa, and the main conclusions. Line drawings should be boldly drawn in Indian ink on Bristol board or similar smooth white card, firm smooth-surface white cartridge paper of good quality, or suitable draughtsman’s tracing materials, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. No half-tone shading or wash should be introduced into line-drawings; all shading should be put in the form of separate dots or lines, and where large areas of uniform shading are required (e.g. in maps and diagrams) this should preferably be done by means of suitable “ mechanical tints.’ Graphs can be reproduced from originals on graph paper with faint b/ve ruling (not grey or any other colour), but all lines to appear on the finished block must be inserted in Indian ink. Lettering on line-drawings and graphs should be inserted lightly in pencil, and will be finished in uniform style. If an illustration includes plant-names or place-names, it is advisable to type these clearly on a separate sheet of paper. Photographs can only be accepted where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable for the necessary reduction. It should be remembered that the finest detail on the originals may be lost even on the best half-tone blocks. If in doubt about the citation of names or references, or the presentation of illustrations or tabular matter, contributors are advised to consult the editor before submitting their typescripts. Authors will normally receive both galley-proofs and page-proofs. Particular care should be given to the thorough checking of references and tables. Alterations in page- proof should be avoided as far as possible, and any words or phrases deleted replaced by others of equal length. Twenty-five separates are given free to authors of papers. Further copies may be obtained at the Society’s current price; orders should be sent to the editor when the page- proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to the Editor, Dr. M. C. F. Proctor, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANpy for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf. of a sub-committee of the Society appointed for this purpose. Sir George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and have become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. — Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8 vo., 176+xvi pages. Bound in cloth. Price : 10/- (postage extra). Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. A DARWIN CENTENARY Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the very successful Conference arranged by the Botanical Society of the British Isles in 1959. The contributors include Lady Barlow, G. D. H. Bell, Prof. T. W. Bocher, Prof. W. H. Camp, Prof. N. Hylander, Prof. I. H. Manton, Prof. T. G. Tutin and Prof. D. H. Valentine. Derro 8 vo., 140 pages, with 7 half-tone plates and 11 maps and figures in the text. Bound in buckram, PRICE £l (postage extra) Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7 The Salisbury Press Ltd., Salisbury—37325 $00.5¢7 “= WATSONIA By [Selon 4 JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 5 NOVEMBER, 1963 Pt. 6 CONTENTS ELEOCHARIS AUSTRIACA HAYEK, A SPECIES NEW TO THE BRITISH ISLES, By S. M. WALTERS _... a oe i oe ae se ea 329=335 VARIATION IN MELAMPYRUM PRATENSE L., By A. J. E. SMITH .. af -. 336-367 FERTILE SEED PRODUCTION AND SELF-INCOMPATIBILITY OF HYPERICUM CALY- CINUM IN ENGLAND, By SIR EDWARD SALISBURY ae Ne .. 368-376 THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS, By J.G. Dony 377-385 THE TAXONOMY OF POLYGONUM LAPATHIFOLIUM L., P. NODOSUM PERS. AND P. TOMENTOSUM SCHRANK, By J. TIMSON .. ye ie ee .. 386-395 BOOK REVIEWS a ae st ff SN S$ “e a .. 396-398 t/ ees ee AD WL .. 399-401 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH MUSEUM (NATURAL HIsTOoRY), LONDON PRICE FIFTEEN SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES Patroness: H.R.H. THE PRINCESS ROYAL OFFICERS President : J. EK. LOUSLEY Vice-Presidents : J. E. DANDY, Miss C. M. ROB, Dr. E. F. WARBURG and Dr. S.M. WALTERS Hon. General Secretary: J. G. DONY, 9 Stanton Road, Luton, Beds. Hon. Treasurer: J. C. GARDINER, Thrift House, 12&14 Wigmore Street, London, W.1. Hon. Editors : (Watsonia) M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT Hon. Meetings Secretary: H. J. M. BOWEN, Pomander House, Harwell, Berks. Hon. Field Secretary: P.C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Longfield, Kent. Hon. Junior Activities Secretary: P. F. HUNT, The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey. Applications for membership should be addressed to the Hon. General Secretary, Dr. J. G. DONY, 9 Stanton Road, Luton, Beds., from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 15/— per part. To be obtained from E. B. BANGERTER, Botanica! Society of the British Isles, c/o Dept. of Botany, British Museum, (Natural History), London, S.W.7. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/- for Junior Members, and 10/- for Family Members, and the Society’s year runs from January Ast to December 31st. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed te D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13 ELEOCHARIS AUSTRIACA HAYEK, A SPECIES NEW TO THE BRITISH ISLES By S. M. WALTERS Botany School, University of Cambridge ABSTRACT Eleocharis austriaca Hayek, discovered in 1947 in Yorkshire by N. Y. Sandwith, but not identified until 1960, is described and illustrated. It is an easily recognised species of the E. palustris group, and is now known to be widespread mainly in the mountains of Continental Europe from the Pyrenees to the Urals, occurring also in Norway. Six British localities are now known, in vice-counties 64, 67 and 70. Some information is given on the habitat in three of these localities, all of which are by upland rivers. The group of species to which the common Eleocharis palustris belongs presents some difficulties of identification, and the critical taxonomy of the European representatives of the E. palustris group needs further elucidation. It is not, therefore, surprising that one of these species has only recently been detected as a member of the British flora. This is E. austriaca Hayek, originally described from Central Europe, but now known to be quite widespread in the European mountains. Material of Eleocharis austriaca was first collected in Britain by Mr. N. Y. Sandwith in July 1947 ‘in a marshy ox-bow of the R. Wharfe below Buckden,’ Upper Wharfedale, v.c. 64. Sandwith was obviously impressed by the differences between this plant and normal £. palustris, and submitted part of the material to me in 1948. At that time my own work on the European species allied to E. palustris was only just beginning, and I did not know E. austriaca; it was not until 1953 that I was able to publish a short paper on the species in question (Walters 1953). In 1960 I refound the sheet of Sandwith’s original collection, and realised what the plant was. In the following season, Sandwith and others revisited the locality, and found the plant still growing there in some quantity. The detailed account of the plant, and the illustrations accompanying this description, are based upon fresh material from the original locality. I am most grateful to Mr. Sandwith for the material, and for notes on the habitat (Table 1a). Very recently, during the preparation of this paper, Dr. G. A. Swan has found four new localities for Eleocharis austriaca in Northumberland (v.c. 67) and one in Cumberland (v.c. 70), and was good enough to show me one of these in the North Tyne Valley. Dr. Swan and Dr. M. E. Bradshaw have kindly supplied further information about the other North Tyne locality by the Kielder Burn, which they visited on the same day. (See Appendix and Table | for details of these localities.) In habit, Eleocharis austriaca resembles the other British Palustres Q.e. E. palustris and E. uniglumis) in the possession of a vigorous rhizome system, from which arise leafless stems terminating in simple flowering spikes (Fig. 1A). The spikes (Fig. 1B) are charac- teristically compact and bluntly conical, however, contrasting with the typically longer, more cylindrical spikes of E. palustris or E. uniglumis. (The difference between £. austriaca and the rather rarer subsp. microcarpa of E. palustris, in respect of spike-size and -shape, and glume-size, is rather less than that between it and E. palustris subsp. vulgaris. As, however, subsp. microcarpa is not known in N. England, recognition of £. austriaca by general appearance is relatively easy.)* * Strandhede (1960) considers that subsp. microcarpa, not subsp. vulgaris, must be the type subspecies, i.e., E. palustris subsp. palustris, according to the current International Code of Botanical Nomenclature. To avoid confusion, I have here used the names as originally published (Walters 1949). 329 Watsonia 5 (6), 1963. ELEOCHARIS AUSTRIACA HAYEK 330 Fig. 1 Eleocharis austriaca. F. Tip of bristle x 40 D. Fruit «12 E. Glume x 6 ection of stem x 8 ’ C G. Single Flower B. Fruiting Spike x 3 A. Habit x 1 H.S «x 14 Watsonia 5 (6), 1963. S. M. WALTERS 331 Two other characters which were rather obvious inthe North Tyne population, but which may not apply generally, were the dark brown colour of the spikes, and the extremely high proportion of fertile stems borne rather close together on a relatively short rhizome. When the plant is dug up, another character is obvious — the absence or very poor development of the reddish-purple colour so characteristic of the base of the stem in both E. palustris and E. uniglumis. This colour character seems reliable on both British and Continental material, and can be used to some extent on dried specimens. Field recognition of Eleocharis austriaca is therefore relatively easy — at least in Britain where, so far as we know, a fourth species, EL. mamillata, does not occur. Detailed characters of the stem anatomy, and particularly of the ripe fruit, however, provide the precise qualitative differences by which the species can be certainly recognised. The differences in stem anatomy can be described in several ways, but can perhaps be most easily shown by the epidermal pattern of strips of tissue (readily prepared by scraping off lightly the internal tissues from a strip of stem; Fig. 2A). There is a very obvious difference between FE. austriaca and E. palustris (Fig. 2B) which can be described as follows. The vascular bundles, with which are associated the stomatal rows, are widely- spaced in E. austriaca, and much more closely packed in E. palustris. Between the stomatal rows can be seen the rows of elongated fibre-cells; in E. austriaca there are up to 5 rows of ordinary epidermal cells between these fibre-rows, whereas in E. palustris only 1-3 rows are found between the fibre-rows. These differences are clearly seen in Figs. 2A and B. (E. uniglumis resembles E. palustris in these characters, whilst the Scandinavian E. mamillata is closely similar to E. austriaca in all vegetative characters). These anatomical differences are responsible for a difference readily observable in fresh material, namely that E. austriaca stems (Fig. 1H) are weak and readily cracked, in contrast to the more pliant stems of E. palustris. In fresh material, and even more in dried specimens, it is relatively easy to count the number of vascular bundles; stems of E. austriaca have 10-16 main bundles, which show as ridges in dried material, whilst those of E. palustris have 20 or more, which are far less conspicuous in the dried state. The ripe fruit of E. austriaca differs from that of E. palustris in two main features. Firstly, the perianth-bristles are rather well-developed, always exceeding the fruit, and they are often 5 in number, by the bifurcation of the abaxial bristle (Figs. 1C and D, 2C). In contrast, the bristles of E. palustris are not infrequently relatively short and always 4 in number (Fig. 2B). In practice, this means that if three or four ripe fruits of the two species are examined, at least one of the austriaca fruits will be found to have either a divided bristle or a pair in the abaxial position. Secondly, the style-base is typically long and narrow, with only a small constriction at the junction with the nut. This contrasts with the more broadly triangular style-base, typical of E. palustris. There is, however, a good deal of variation in style-base shape, and the character cannot be used reliably on single fruits. Its value is greater in distinguishing FE. austriaca from E. mamillata, which has the low broad style-base indicated by the name. The differences between EF. austriaca and E. mamillata in style-base and in other characters are discussed and illustrated by Walters (1953) and Strandhede (1962). The European distribution of E. austriaca is as yet very incompletely known, but it seems to occur mainly by mountain lakes and streams from the Pyrenees to the Urals; in Scandinavia, where E. mamillata is widespread, E. austriaca seems to be rare (Strandhede 1962). Its occurrence in N. England therefore fits well with the distribution of species such as Primula farinosa, a widespread European mountain plant not occurring in the Arctic. A complete list of localities is given in Appendix B. There is a general similarity between all the British habitats of E. austriaca. All the habitats are relatively impermanent ones by rivers, and the plant seems to form vigorous stands where there is rather open alluvial gravel or sand. The Wharfedale locality was re-visited by Mr. N. Y. Sandwith on 9 July, 1961, when the species-list (a) in Table 1 was compiled. The plant is growing in an old ox-bow of Watsonia 5 (6), 1963. 332 ELEOCHARIS AUSTRIACA HAYEK A. Eleocharis austriaca: Epidermis of stem x 100 B. Eleocharis palustris subsp. microcarpa: Epidermis of stem Xx 150 and (inset) fruit x 12 (Material from Aldenham Reservoir, Herts, v.c. 20, coll. M. A. Grierson, 1961 (K)). C. Fruits of E. austriaca, showing range of bristle-number and -form, x 12 Watsonia 5 (6), 1963. S. M. WALTERS 323 the River Wharfe, rooted in silty mud. A sample of the mud gave on analysis 33-1 °% of organic matter, and a pH of 6°8. TABLE 1. Localities of E. austriaca—lists of associated species | @ 6) © @ 6) (© Achillea millefolium =r Leontodon autumnalis | de A. ptarmica +. Linum catharticum ae Agrostis stolonifera ST se ai Lotus corniculatus “| Alnus glutinosa a= ae Lysimachia nemorum alle Angelica sylvestris + Mentha aquatica ao Anthoxanthum odoratum + Myosotis caespitosa + Bellis perennis + M. scorpioides -- Callitriche sp. + Parnassia palustris ae Caltha palustris + Pedicularis palustris + Cardamine fiexuosa — Phalaris arundinacea a C. pratensis Se Plantago lanceolata a Carex demissa = P. major + Centaurea nigra - | Poa trivialis aie Cynosurus cristatus -- Potamogeton natans + Deschampsia cespitosa + P. polygonifolius a= Eleocharis palustris + Potentilla anserina + Epilobium palustre + Prunella vulgaris + 4- E. nerterioides ++ Ranunculus flammula 4+ + a Equisetum arvense a -+- R. repens + = a E. fluviatile -- Rhinanthus minor sens /at. + Euphrasia brevipila + Rumex obtusifolius a E. nemorosa + Sagina procumbens +- Festuca arundinacea iaie + Salix sp.* + Filipendula ulmaria + ++ S. triandra 4. Galium palustre + ++ + S. viminalis +t. Galium uliginosum + Sparganium erectum a Glyceria fluitans + Succisa pratensis + Holcus lanatus oa Trifolium pratense + Juncus acutiflorus -— oo T. repens a J. articulatus - =" Triglochin palustris + J. bufonius 4. Tussilago farfara + + J. effusus ++ + 5= Veronica beccabunga 4. + + Lathyrus pratensis + V. chamaedrys 4+" (a) Original locality in Wharfedale, 9 July, 1961 (c) Kielder Burn, 23 Aug. 1962 (6) North Tyne near Otterstone Lee, 23 Aug. 1962 * Seedlings. The locality in the North Tyne Valley near Otterstone Lee, discovered by Dr. G. A. Swan, was shown by him to Dr. M. E. Bradshaw and myself on 23 August, 1962. It is very similar to the Wharfedale habitat; the plant is growing in a silted backwater by the side of the river, in a community containing six species common to both localities (see Table 1b). A sample of the mud from the rhizomes of the plant here showed an organic content of 6% and a pH of 7:1. Dr. Swan and Dr. Bradshaw visited the Kielder Burn locality on the same day, and furnished a description and species list (Table Ic). Here Eleocharis austriaca grows in a gravelly and silty backwater by the stream in an open community with much bare gravel. The soil sample gave a very low organic content @-7 7) and a pH of 7:0. | The significant difference in organic content between the Wharfedale locality and the other two emphasises the only obvious difference, namely that the Northumberland habitats were less stable and more obviously subject to flooding. One could, however, reasonably assume that there had been an earlier seral stage in the Wharfedale habitat, before the ox-bow was cut off from the parent river at its upstream end, when the community would have been much more open. Detailed information on Continental habitats of -£. austriaca for comparison with these British ones is not yet available, but Strandhede (1962) indicates that the Norwegian Watsonia 5 (6), 1963. 334 ELEOCHARIS AUSTRIACA HAYEK localities by upland rivers, at any rate, are very similar, and also points out that E. austriaca differs in its ecological requirements from the closely similar E. mamillata, which is charac- teristic of peaty ground, often of acid reaction. Some Continental habitats are by upland lakes (e.g. the locus classicus of E. benedicta, cf. Appendix); it would be interesting to have further information about such localities. I am greatly indebted to Miss M. A. Grierson, of the Herbarium, Royal Botanic Gardens, Kew, for the line-drawings reproduced in Figs. 1 and 2, and to Mr. P. Freeman, of the Botany School, Cambridge, for the soil analyses. REFERENCES Ho vs, J. (1960). Nové nebo méné zndmé rostliny kvéteny CSSR. Preslia 32, 423-5. KrecetTovic, V. I. & Zoz, I. G. (4940), in Bordzilovsky & Lavrenko (eds.) Flora R.S.S. Ucr. 2, 418-437. PopLecH, D. (1960). Die Arten der Eleocharis palustris-Gruppe in Bayern. Ber. Bay. Bot. Ges. 33, 105. STRANDHEDE, S.-O. (1960). A note on Scirpus palustris L. Bot. Not. 113, 161-171. STRANDHEDE, S.-O. (1962). Eleocharis Palustres in Scandinavia and Finland. Bot. Not. 114, 417-434. WALTERS, S. M. (1949). Biological Flora of the British Isles. Eleocharis R. Br. J. Ecol, 37, 192-206. WALTERS, S. ME Coe Eleocharis mamillata Lindb. fil. and allied species. Ber. Schweiz. Bot. Ges. 63, WALTERS, S. M. (1959). Heleocharis austriaca Hayek and Gilyceria declinata Bréb., two plants new to the flora of Poland. Fragm. Flor. et Geobot. (Krakow) 5, 239-244. ZINSERLING, G. (1935). in V. L. Komaroy (ed.) Flora U.R.S.S. 3, 75-6 and 581. APPENDIX A. Nomenclature and Synonymy Eleocharis austriaca Hayek, Flora Stiriaca Exsiccata, Lieferung 19/20 (1910). E. benedicta Beauverd, Bull. Soc. Bot. Genéve, sér. IT (2) 13, 245 (1921). E. leptostylopodiata G. Zinserling in Komarovy (ed.), Flora U.R.S.S. 3 (Addenda), 581 (1935). B. List of Localities (i) Britain V.c. 64, Mid-West Yorks: marshy ox-bow of the R. Wharfe below Buckden (Grid ref. 34/945760), N.Y. Sandwith, July 1947 (ref. no. 3185); July 1961 (ref. no. 5906). V.c. 67, S. Northumberland: by R. Rede near Byrness (Grid ref. 36/770022) 22 July 1962; by R. North Tyne near Otterstone Lee (Grid ref. 35/670879) 6 Aug. 1962 (see details on p. 333); by Kielder Burn (Grid ref. 35/651957) 6 Aug. 1962 (see details on p. 333); in silted-up pond by R. North Tyne near Plashetts (Grid ref. 35/664897) 29 Sept. 1962. V.c. 70, Cumberland: by R. Irthing near Gowk Banks (Grid ref. 35/682737) 16 Sept. 1962. (Localities in V.c.’s 67 and 70 were discovered by G. A. and M. Swan). (ii) Continental Europe AUSTRIA Steiermark: ‘ Stiria media in stagno exsiccato ad pagum St. Peter prope urbem Graz, 370 m. Juni 1906, K. Fritsch,’ lectotype in (H); cf. Walters (1953) p. 284). Tirol: ‘ in paludibus prope Lienz (?) Pichler, Juni 1865 ° (mixed sheet with E. palustris ) (BM). CZECHOSLOVAKIA Beskydy and Tatra (Holub, 1960). FRANCE Haute-Savoie: Plage du lac Bénit, 1580 m, massif des Vergys, 15 Aug. 1922, G. Beauverd, Ch. Duffour Exsicc. 4277 (E. benedicta Beauv., type locality) (G); Servoz, Lac Vert, 3 Sept 1934, M. E. Edmonds No. 9 (K); Chalets de Méri, 1910, de Palézieux (G). Jura: Bords du lac des Rousses 17 July 1869, Herb. Manceau (P). Hautes-Pyrénées: Vailée supérieure du lac de Gaube, 31 Aug. 1836, E. Doassans (as E. uniglumis) (K); Cautarets, le Cayan—alt. 1630 m, fin aott, Mouillard, Ch. Duffour Exsicc. 1104 (P). GERMANY Several localities are given for Bavaria by Podlech (1960) (M). Watsonia 5 (6), 1963. S. M. WALTERS 335 ITALY Prov. Belluno: Shore of Lago Alleghe, Caprile, 29 Aug. 1862, Herb. Churchill (K). JUGOSLAVIA Flora Bosnae: ‘ ad Koran prope Pale,’ (near Sarajevo), c. 860 m, 21 July 1929, leg. Maly (K). NoRWAY Three localities (2 in Saltdal, Norrland, and one in Sdr Tréndelag) are given by Strandhede (1962). POLAND Three localities in the Tatra (W. Carpathians) (KRA) (Walters 1959). SWITZERLAND Apparently widespread; add to the 27 Swiss localities listed in Walters (1953) (mostly in ZT): Fribourg: Commune de Charmey, rive du lac de Montsalvans, 803 m, 19 July 1849, H. Brunner, Soc. Franc. Ech. Pl. Vasc. Exsicc, B. de Retz, fasc. 3, No. 819 (K), Engadine: Prés Bevers 7. 1867. L. Favrat & W. Barbey (K). WES:S.R: The distribution given by Zinserling in Komarov (1935) for Heleocharis leptostylopodiata is European Russia to Middle Volga region, Ural Mountains and adjacent territories, Siberia (Kusnetskij Alatau), Amur region and the Caucasus. The disjunct nature of the distribution is stressed. The type material (LE) is from ‘ Baschkiria, Canton Zalaia,’ and is in excellent fruiting condition. There can be no doubt as to the identity of Zinserling’s species with EF. austriaca. In their account of Heleocharis in Flora R.S.S. Ucr. (1940), Kreéetovié & Zoz do not give the species from the Ukraine. It does, however, grow within the present boundaries of the Ukraine, as I have seen the following specimen: Orelac, near Sniatyn, on R. Prut (Podolia, W. Ukraine) 17 Sept. 1880. A. Sleidifiski (KRA). [The Herbaria in which material is to be seen are indicated by the standard abbreviations as given in ed. 4 of Index Herbariorum (Regnum Vegetabile 15, 1959). I am grateful to the authorities in these in- stitutions for facilities to consult and borrow material.] Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. By A. J. E. SMITH Department of Botany, University College, Swansea* ABSTRACT Melampyrum pratense is morphologically extremely variable in Britain and this variability has been found to occur at three levels. There is geographical variation brought about by climatic selection from a basic pattern. Superimposed upon this geographical variation is ecological variation caused by selection in different habitats. The third level of variation is brought about by the random isolation of particular genotypes in individual populations leading to minor but detectable differences between populations in similar ecological habitats. It was found that there are two morphologically separable groups within M. pratense which differ also ecologically in Great Britain and these have been given subspecific rank: subsp. commutatum is restricted to base-rich habitats in southern England; the remaining British populations belong to subsp. pratense. INTRODUCTION Melampyrum pratense is an annual hemiparasitic member of the subfamily Rhinan- thoideae of the Scrophulariaceae. The plant is found throughout most of Great Britain and Ireland, and in Britain usually occurs as small isolated populations of up to about two hundred plants. In this account only British plants are dealt with, as insufficient material from Ireland was seen. PARASITOLOGY Melampyrum pratense, in common with other members of the Rhinanthoideae, para- sitises host plants by means of root haustoria. This parasitic habit appears first to have been noted by Decaisne (1847) and the nature of the parasitic attachment has been studied in considerable detail by subsequent authors. Little is known of the physiology of the parasitism of the Rhinanthoideae but from field observations of M. pratense it is possible to suggest what some of these physiological requirements might be. In Britain, the seeds of M. pratense germinate about December and develop an extensive root system, but little vegetative growth takes place beyond the production of cotyledons until March. If the roots become attached to a suitable host during this period rapid growth commences towards the end of March and flowering begins within six to eight weeks. If no contact with a host is made the cotyledons expand and two or three pairs of very diminutive stem leaves are produced, after which growth ceases. During May and June these undeveloped plants gradually wither away. Plants of M. pratense do not appear to be able to parasitise one another, unlike species of Euphrasia (veo. 1961): Plants of M. pratense are able to thrive on soils which are exceedingly low in nutrients. Growth of other vascular plants in such habitats is extremely slow in comparison with M. pratense which reaches its full size in a matter of a few weeks. It seems that this can only be by the acquisition of nutrients from the host plants and it is evident that, in nutrient-deficient habitats at least, the species obtains a considerable proportion of its nutrients from the host. It has been suggested by J. L. Harper (personal communication) that the requirement is for growth substances and that these are responsible for the rapid burst of growth men- tioned above. It is difficult to conceive of this occurring without the acquisition of nutrients * Now at Department of Botany, University College of North Wales, Bangor. 336 Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 337 from a host, particularly on poor soils, so that whilst it may be possible that growth sub- stances are necessary, these are not the sole requirement. The ability to parasitise also appears to be important in connection with water relations. During a dry period in the summer of 1959 a perfectly healthy population of M. pratense was seen in a wood on Keston Common, Kent, in which other herbaceous plants were dead or severely wilted. In this locality Melampyrum was presumably growing on the roots of tree species. A similar situation was noted in an experimental plot at Oxford when, after a period of drought during which weeds were dead or dying, young plants of M. pratense parasitic on Betula and Calluna were growing vigorously. M. pratense is a shallow rooted species but is evidently able to obtain a fully adequate water supply from the host species and is thus independent of soil water. Host plants It is frequently assumed that M..pratense parasitises herbaceous plants but all the evidence suggests that the hosts are woody plants. Although it is rare for populations to occur without any associated herbaceous species, M. pratense is never found away from trees or shrubs and, in parts of northern Britain where large woody plants are absent, it is associated with Calluna, Erica, Myrica or combinations of the three. It is extremely difficult in the field to trace root connections and it has not been possible to determine whether M. pratense parasitises herbaceous plants, but from the fact that it never occurs solely with these plants, this is unlikely. In several populations examined, connections have been traced between M. pratense roots and those of Quercus robur, Corylus avellana and Betula pubescens. In cultivation experiments the plant has been successfully grown on Betula pubescens, Calluna vulgaris and Rubus fruticosus, but not on herbaceous plants. From field observations it is clear that Fagus sylvatica, Sorbus aucu- paria, Quercus petraea, Vaccinium myrtillus, V. oxycoccus, Erica tetralix and Myrica gale are also host plants. Whilst it is evident that M. pratense can parasitise a number of woody species—it is not clear whether particular populations are host-specific. Bog populations in northern Scotland do not spread to neighbouring birch woods where the plant might be expected to grow. In the New Forest, Hampshire, it is always associated with trees, although in certain other parts of the country it grows on Calluna, Erica or Myrica, all three of which occur in the New Forest. ECOLOGY M. pratense shows considerable ecological diversity and populations from different types of habitats tend to be morphologically different. The three main types of habitat in Britain are (i) calcareous scrub and woodland, (11) acid woodlands and (ii) bogs. These habitats are subdivisible as follows: (i) A. chalk and limestone habitats. (ii) B. acid woodland in the New Forest, Hampshire. C. other acid woodland habitats in southern England (neogenic region). D. acid woodland habitats in south-west England (palaeogenic region) E. acid woodland habitats in northern England F. acid woodland habitats in Scotland (iii) G. bogs in northern Scotland H. bogs in northern England The most commonly associated species in each habitat are listed in Table 1. Those of types E and F are so similar that they are listed together. The acid woodland habitats were grouped into four types because of geographical differences and likewise the two bog types. Neogenic and palaeogenic regions of southern England were considered separately as it was thought that plants from these areas of different origin might differ morphologically. Watsonia 5 (6), 1963. 338 A. J. E. SMITH TABLE 1. Commonly associated species from the different ecological types > bd OQ wo) to Re Ps a) an Ecological type Acer campestre Anthoxanthum odoratum Betula spp. Calluna vulgaris Corylus avellana Crataegus monogyna Deschampsia flexuosa Drosera spp. Erica tetralix Eriophorum spp. Hedera helix Holcus lanatus Ilex aquifolium Mpyrica gale Molina caerulea Narthecium ossifragum Potentilla erecta Potentilla palustris Pteridium aquilinum Quercus robur Quercus petraea Rubus fruticosus Solidago virgaurea Teucrium scorodonia Thelycrania sanguinea Trichophorum caespitosum Vaccinium myrtillus Vaccinium oxycoccus Viburnum lantana sepals demas eb eed La bara | cbse PM Seles le el ara tt ele feces se | | ae PS Sesto Rae tole ty se | | Pe seer eet | PSE sibs db El arse | bee TA IST se qeaese |) feaesear Past se ||| eed se Se ses Tra it a arsrrsersal fl sea ter | In southern England, populations occupy two very different habitats; chalky or lime- stone souls with a pH of 6-8 or more, and acidic soils with pH values of up to about 5:5. On the chalk of Kent, Surrey, Hampshire and Oxfordshire, M. pratense usually occurs on the edge of scrub and chalk grassland or as a hedgerow plant. Populations of M. pratense are also found in similar habitats on the oolitic limestone of the Cotswolds and the carboniferous limestone of the Wye Valley and associated species are much the same. On acid soils in southern England M. pratense usually occurs at the edges of rides, paths or roads through woods, or in clearings, and in common with populations from calcareous habitats only thrives in light shade or in the open. Where shading becomes heavy the number and vigour of the plants is reduced. Populations from woodland in the New Forest are quite characteristic in appearance and differ very much from other southern forms but are sometimes indistinguishable from plants from Scottish bogs. They differ from these, however, in always being associated with trees. As will be mentioned later, the New Forest populations of M. pratense form a disjunct feature in an otherwise more or less continuous picture of variation in southern England. In south-west England and Wales M. pratense grows in habitats very similar to the acidic woodlands of southern England but populations also occur on shallow soils and on steep and often rocky banks. This type of habitat is the one most commonly found in northern England and parts of Scotland and plants from such habitats are morphologically similar. As M. pratense is an annual it is clear that in these places where the soil is so shallow and nutrient-deficient the ability to thrive is closely correlated with the parasitic habit. All the habitats so far described are usually well drained, but there are two types of habitat where this is not so. These are blanket bogs in northern Scotland and bogs or mosses in northern England. In Ross, Sutherland and Caithness M. pratense was found growing in considerable quantity in boggy flushes and waterlogged peat. Watsonia 5 (6), 1963. 4 , ) | VARIATION IN MELAMPYRUM PRATENSE L. 339 In the only two known bog habitats visited in England, at Wybunbury Moss, Cheshire, and Brigham Moss, Cumberland, the associated species differ somewhat from the Scottish bogs and the plants of M. pratense are rooted in Sphagnum rather than in peat. From an investigation of M. pratense in the field the following points were apparent: (i) Plants from ecologically different habitats are frequently morphologically distinct. (ii) Plants from habitats with a low nutrient status are generally smaller than those from richer soils. (iii) The host plants from various habitats are not necessarily the same. From these points two questions arise. First, whether morphological differences are phenotypic or genotypic, and secondly, whether the apparently different morphological races are host-specific. A series of cultivation experiments were carried out in an attempt to answer these questions. CULTIVATION EXPERIMENTS Although up to 50% of the seeds sown (with or without a host) germinated, only about 0:-5% reached maturity and because of this the value of the results obtained from the cultivation experiments was much reduced. Samples, each of twenty seeds all from one population, were sown in pots with forty- three of the most commonly associated species, but only two plants matured, one on Betula pubescens and one on Rubus fruticosus, and it was not possible to come to any con- clusion concerning host range. In another experiment seeds were collected from seventy-five wild populations as follows :— Acid woods te a. we 50) samples Scottish bogs... mn “2 ae Oesamiples Calcareous habitats ae es eo samples Wybunbury Moss, Cheshire .. Saesample Twenty seeds were sown from each sample under the following conditions: (i) Acid ground (pH 4:5) with Betula pubescens as host in a forestry nursery bed at Bagley Wood, Berkshire. (ii) As (i) but with Calluna vulgaris as host. (111) Calcareous ground (pH 8-8-2) with Betula pubescens as host at the Botany School, Oxford. Of the seeds sown on Betula pubescens one or more plants from five samples reached maturity as follows: Four plants of sample 90 from an acid woodland from which Betula pubescens was absent. Although the soil in Bagley Wood was much richer than that from which the seed had been collected, morphologically the plants were indistinguishable from the wild population. One plant from sample 80 and two from sample 84, both from Scottish bogs. Again Betula pubescens was not present in the original locality, and soil conditions were very different. The plants were morphologically indistinguishable from those of the populations from which the seeds were collected, apart from the plant from sample 80 which was shorter. Two plants from sample 100 and one from 103, both from calcareous habitats. The plants from sample 100 were very stunted and died as soon as flowering started, but the plant from sample 103 was normal although growing in acid soil. Only three of the seeds sown on Calluna reached maturity. Two of these came from and were indistinguishable from the sample collected from Wybunbury Moss, Cheshire, despite the very different habitat. The other plant was from sample 107 and grew perfectly satisfactorily on Calluna vulgaris which was absent from the original calcareous habitat. It is evident from these experiments that whilst seeds of M. pratense germinate freely, Watsonia 5 (6), 1963. 340 A. J. E. SMITH even in the absence of a host, establishment on a host appears difficult. In a wild population examined near Oxford in May, a number of undeveloped seedlings like those described previously were found, although not in the great proportions (about 99%) found in cul- tivation. It is evident that the plant is difficult to grow successfully but the reasons for this are not apparent as plants that did grow had come from a variety of very different habitats. Although the cultivation experiments were not successful as a whole, certain con- clusions can be drawn from them in combination with field observations. M. pratense does not necessarily appear to be host-specific beyond being confined to woody species. That there is a lack of specificity for any particular woody species, in some populations at least, is shown by the ability to grow on Betula or on Calluna when not previously associated with these species. However, as mentioned earlier, M. pratense does not appear to be able to spread from one habitat with certain associated woody species to a neighbouring habitat in which different species are present. Plants grown in cultivation under very different conditions from those appertaining in the wild did not differ morphologically (with the exception of height in sample 80 and the stunted plants of sample 100) from the populations from which they originally came. It would seem, therefore, that for these populations at least, population differences are genotypic rather than phenotypic. MORPHOLOGICAL CHARACTERS A large number of morphological characters have been used by previous authors for the discrimination of various subspecific taxa. Attempts to name forms using descrip- tions of Ronniger (1914), Beauverd (1916), Soo (1928), Britton (1943) and Jasiewicz (1958) frequently failed, either because of the inadequacy of the descriptions, or because the plants in question did not fit any prescribed taxon. Many of the taxa are based on different numbers of parts or ranges of size and when a large number of specimens are examined it is often found that variation in the particular characters is continuous. Some of the subspecies and varieties are based on minor morphological variants and two or more can often be found in the same population. The most recent account dealing with British forms is that of Britton (1943) in which some eleven subspecies, seventeen varieties and ten forms of M. pratense are described. From examination of herbarium material it was evident that morphological variation in M. pratense is to some extent correlated with geographical distribution and, in order to investigate this further, 140 random samples of 25 plants each were collected from four main areas of Britain (southern and northern England, and central and northern Scotland). These areas were selected as being those in which the species is most abundant. On attempting to group the populations objectively by eye it was found that there were several morphological groups, linked by intermediates which had an ecological basis. It was also clear that the ecological groups corresponded with those described earlier. Similar results were obtained using scatter diagrams. As it was not possible to separate all the samples into discrete groups, either by eye or by the use of diagrammatic methods, it was decided to classify populations on the basis of geographical distribution and habitat in order to analyse the variability biometrically. The characters selected for examination are listed below and were selected either because they had been regarded as important by previous authors or because they were considered to be amenable to statistical analysis. Length of hypocotyl. Height. Number of sterile nodes and total number of nodes Number of pairs of cauline and intercalary leaves. Branching habit. Ae ea eee Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 341 6. Number of sterile and fertile branches. 7. Presence or absence of cotyledons at maturity. 8. Shape and size of cauline leaves. 9. Number of pairs of entire and toothed bracts. 10. Degree of toothing of bracts. 11. Calyx length. 12. Corolla length and colour. 13. Anther and anther-appendage length. 14. Pollen size. For conciseness the ecological types are referred as A, B, C, etc., as described on page 337. Type H has been omitted as there were only two samples of this type. STATISTICAL ANALYSIS OF MORPHOLOGICAL CHARACTERS A biometric analysis of certain of the listed characters was carried out to obtain comparative information on the morphology and also on the degree of genetic variability in the different ecological types and to see if any of the morphological differences were of taxonomic significance. Twelve selected populations were also compared. Seven of these populations were from bogs in Sutherland and five from chalk habitats in Kent. These were selected in order to compare two groups of populations differing widely ecologically and geographically from the point of view of morphology and genetical variability. Morphological differences and genetic variability in plant populations can be investi- gated by means of analyses of variance of means and standard deviations. The method used for M. pratense was based on that described by Day & Fisher (1937) for the analysis of populations of Plantago maritima. In the analysis of genetic variability it is necessary to make an allowance for the wide diversity of some of the mean values (Day & Fisher 1937), e.g. mean heights vary from about 6 to 37 cm. Differences in means might be expected to be accompanied by similar differences in the standard deviations. If this were so, analysis of variance of the standard deviations would supply little further information than the analysis of the means. It is possible to compare variabilities by dividing the standard deviation by the mean and obtaining a ‘ coefficient of variation.’ This method would be satisfactory if an increase in the mean carried with it a proportional increase in the standard deviation. In biological material this is not always so (Day & Fisher 1937). To eliminate the effect of widely differing means when comparing variabilities, an analysis of covariance of the means and the standard deviations of the populations must be carried out. An analysis of covariance of means and standard deviations was carried out on characters 2, 8, 11, 12 and 13 and an analysis of variance on 3, 4, 9 and 10. Where differences among types or populations are significant, the least significant difference is estimated using the Student-Newman-Keuls’ test (Steel & Torrie 1960). This method is used because the estimation of the least significant difference using the usual t-test is suitable only for independent comparisons involving two means. When more than two means are compared the comparisons are non-independent and the f-test is not a truly valid criterion. The Student-Newman-Keuls’ procedure takes into account the number of means. As the number of populations from each ecological type was not equal (numbers varying from six to twenty), the least significant difference was estimated using the following formula: | q a(p,n) s / 4 [: +) 1 J where q « is a value read from a table (Table A8, Steele & Torrie 1960) and Watsonia 5 (6), 1963. 342 A. J. E. SMITH = number of means number of degrees of freedom of the error variance = s/f, or 7 levelvon significance =" ‘error variance t,t; = number of observations in means for desired comparison. Tables 2, 3, 4 and 5 show the significant differences between means and variabilities of the ecological types and selected populations. Ao Brg | TAXONOMIC VALUE OF MORPHOLOGICAL CHARACTERS (i) Length of hypocotyl Hypocotyl length was not treated statistically as from field observations it was clear that length varies very much with density of surrounding vegetation. Length also varies with the depth of burial of the seed and this is important in view of the seeds being dis- persed by ants. Although hypocotyl length has been used by Britton (1943) as a taxonomic character it cannot be regarded as having any value. (ii) Height All measurements of height refer to the length of the main axis from the cotyledonary node to the stem apex. Total height is unsatisfactory because of the variable length of the hypocotyl. Height decreases markedly northwards (Fig. 1), but as can be seen from N 000 (Approx. 58° 59’ N) N. Scotland 900 800 © S C. Scotland S 700 Sy) Re = 60 S 0 = 6 500 N. England & : 400 = 300 2 00 S. England (Approx. 50° 45’ N) 100 Height (cm) Fig. 1. Geographical scatter diagram of height showing diminution of height of populations northwards. Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 343 Fig. 2, variation is almost continuous so that there are no clear groupings based on height alone, although in combination with characters such as leaf shape, ecological types A and G form two groups only partially merging with other populations. Height has been widely used in the delimitation of subspecies and varieties but its unsuitability is seen from the description of the subsp. vulgatum (Pers.) Ronn. in which height is said to vary from 10 to 50cm. This covers almost the whole range of heights of all other forms of M. pratense. (111) and (iv) Number of nodes and sterile nodes on the main axis and pairs of cauline and intercalary leaves These characters are taken together because the sum of pairs of cauline and inter- calary leaves is equal to the number of sterile nodes. Considerable importance has been attached by previous authors to the node at which the inflorescence commences (i.e. sterile 40 35 Height (cm) IN S 4 8 12 16 20 Leaf Index (length/width) Fig. 2. Height plotted against leaf index. Crosses, populations from calcareous habitats; solid dots, all other populations. node number plus one) and the relative numbers of cauline and intercalary leaves. Cauline leaves are those leaves with axillary branches, and intercalary leaves the leaves without axillary branches between the cauline leaves and bracts. It is not always possible to distinguish between intercalary leaves and bracts as the lowest flower buds, in the axils of untoothed bracts, sometimes abort and fall off, especially if the plant is growing in deep shade. Further, in old plants dehisced fruits are also lost so that the number of intercalary leaves is not a satisfactory character. It is of interest to note, however, that populations of type A and type G do not differ significantly with Watsonia 5 (6), 1963. 344 A. J. E. SMITH regard to intercalary leaf number but differ from all other populations. As type A is found in southern England and type G in northern Scotland, this upsets the south-north geograph- ical cline in this character (see Fig. 3.). N 000 (Approx. 58° 50’ N) N. Scotland 900 800 = S C. Scotland 700 = : 600 > 2 500 N. England : 400 z, 300 200 (Approx. 50° ae S. England Pairs of intercalary leaves Fig. 3. Geographical scatter diagram of number of pairs of intercalary leaves showing similarity of northern bog populations (open circles) and populations from calcareous habitats (crosses). Britton (1943) separates six of his subspecies into two groups on the basis of cauline leaf number and the position of the inflorescence. Four subspecies are defined as having one pair of cauline leaves and the inflorescence commencing at nodes 2-3-(4). The other two are said to have 1-3 pairs of cauline leaves and the inflorescence commencing at nodes 4-5. Identification of forms so described is frequently impossible as there is considerable variation in the number of sterile nodes, cauline and intercalary leaves both within and between populations. There appears to be no correlation between number of parts and habitat in Britton’s taxa and hence these taxa appear to be of little significance. (v) Branching habit The angle of branching varies from horizontal to almost vertical and from cultivation experiments appears to be genetically controlled. The character was not analysed for two Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 345 reasons. Firstly, in pressed specimens the branches were distorted. Secondly populations where the main axis was prostrate had branches which were likewise horizontal but rose vertically at the tips and these were difficult to compare with other populations. The character might have proved of interest if analysed in the field. (vi) Number of sterile and fertile branches Although this character has been used taxonomically it is of little use, as the propor- tion of sterile to fertile branches decreases with age and unless all populations are sampled at the same time comparisons are not valid. N 1000 (Approx. 58° 50’ N) 900 N. Scotland 800 C. Scotland 700 600 500 N. England 400 Northings of Grid Reference 300 200 (Approx. 50° 45’ N) 100 S. England 4 8 12 16 20 Leaf Index (length/width) Fig. 4. Geographical scatter diagram of leaf index (length/width) showing narrowing of leaves northwards. Solid dots, populations from acid habitats; crosses, populations from calcareous habitats. (vii) Presence or absence of cotyledons at maturity This character is again affected by age, although in populations of small plants the cotyledons are frequently still present when flowering commences. With large plants the stems become woody at the base and the cotyledons are lost. As with the number of sterile and fertile branches, all populations would have to be sampled at the same time and allow- ance would have to be made for the effect of climatic differences on the rate of growth due to geographical distribution and altitude. (vill) Cauline leaves Leaf shape is one of the most conspicuously variable characters of M. pratense and a number of infraspecific taxa have been largely based on leaf shape. One of these, subsp. commutatum (Tausch) C. E. Britton (1943), is defined as having lanceolate to ovate leaves, all other forms having narrower leaves. From field observations, examination of herbarium Watsonia 5 (6), 1963. 346 A. J. E. SMITH specimens and statistical analysis, this subspecies, corresponding with ecological type A, appears to be more distinct than any others described. Leaf shape is closely correlated with geographical distribution and height, as can be seen from Figs. 4 and 2. A measurement of leaf shape is obtained from the length/breadth ratio of one of the uppermost pair of cauline leaves of each plant. Length is total length of the leaf and width is measured across the widest portion. In Fig. 5, where leaf length is plotted against leaf width, it is clear that there are two patterns of variation, one made by plants from calcareous habitats and the other from the remaining populations. Plants from calcareous habitats, with the exception of some populations from Kent, are quite distinct with regard to leaf shape from all other populations, the leaves being ovate to ovate-lanceolate instead of lanceolate to linear-lanceolate. Despite some of the Kentish populations inter-grading in leaf shape with populations on non-calcareous soils in the same area, type A populations are significantly different from all other populations in leaf shape. (ix) Number of entire and toothed bracts This character does not appear to be of much use taxonomically, as the number of entire bracts is correlated with height, short plants having more entire bracts than tall plants. Leaf width (mm.) 20 30 £40 50: 5 760 70 ~=80 Leaf length (mm.) Fig. 5. Leaf width plotted against leaf length showing difference between plants from calcareous habitats (crosses) and other populations (solid dots). (x) Degree of toothing of bracts The number of teeth per bract varies very greatly from population to population. The various British taxa are described as having few to many teeth per bract, with teeth small and blunt to long and acuminate. Usually the number and size of bracts and the extent of toothing is correlated. If the bracts are few in number they tend to be small with Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 347 blunt teeth and, if numerous, are larger with long acute teeth. Because of the great variability in toothing of bracts, both within and between populations, and because the nature of the teeth depends upon the number of bracts, the toothing of bracts is not a good character. (xi) Calyx length The calyx of M. pratense is four-lobed, with the two upper lobes frequently longer than the lower. The calyx becomes enlarged in fruit and all measurements were made on the oldest flower of a stem or branch. The length of the calyx tube is relatively constant throughout the country but total calyx length, measured from the base of the calyx to the apex of the longest tooth, was found to be very variable. The mean calyx length of populations examined varies from about 3:5 mm to 7:8 mm, and there is a higher proportion of populations with long calyces in the north of Britain than in the south. ° » bo Ww Aunmereferefiin ” r fo ° 9 ent bien eel brent ere @ Se) ~ iS) a Q ™s o ° ° 5t} 17 400KM —— T T y T T W.LONG. 9° e if °° sy at r o° E.LONG. Fig. 6. Distribution map of the golden-yellow flowered form of M. pratense (var. hians Druce). The dotted line roughly divides the western palaeogenic regions of Britain from the south-eastern neogenic regions. (xii) Corolla length and colour Variation in corolla colour is a conspicuous character and there are three distinct colour forms which do not merge. The first type has a whitish or pale-yellow corolla tube with a pale-yellow mouth. This is widespread throughout Britain. The second type of flower is uniformly golden-yellow, sometimes verging on orange. This occurs as pure populations and apart from an unconfirmed record from Sussex appears to be restricted to palaeogenic regions of Britain (Fig. 6). It also occurs in Ireland but there are insufficient Watsonia 5 (6), 1963. 348 A. J. E. SMITH data to gain an adequate idea of its distribution. It appears to be very rare or absent on the Continent. The golden-yellow form was first described by Druce (1884) as var. hians and was raised later to subspecific rank but, apart from flower colour and geographical distribution, there is nothing to distinguish it from other forms of M. pratense. The third colour form which is limited to northern Britain has a pale-yellow or whitish corolla tipped with crimson and usually occurs mixed with plants with pale corollas. It is found in Scandinavia and in mountainous areas of other parts of Europe. It does not normally occur in pure populations in Britain and does not show any particular geogra- phical distribution other than tending to be northern and montane. In the first important monograph on the genus Melampyrum (Beauverd 1916) great stress 1s laid on the colour of the corolla after anthesis in MW. pratense and the two subspecies are based on this. In subsp. eu-pratense Beauv. the corolla is said to become purple to pale red after anthesis and in subsp. vulgatum Pers. the corolla becomes blackish. Britton (1943) follows Beauverd and, although he recognises more subspecies, these are in two groups corresponding to the original two subspecies of Beauverd. It is not clear either from Beauverd or from Britton whether anthesis means (correctly) opening of the flower or (incorrectly) shedding of pollen. In many of the populations examined, the flowers became slightly to markedly pinkish as they began to wither. The anthers usually burst just after the flowers open but there is no change in colour until the flowers begin to wither. No blackish corollas were seen other than decaying ones that had not fallen off. From vegetative characters a number of the populations could be allocated to varieties of either of Beauverd’s subspecies when flower colour was ignored. If, however, flower colour was taken into consideration it was impossible to name many of the populations simply because none had blackish corollas. It is impossible to identify pressed specimens using either Beauverd’s or Britton’s descrip- tions as corolla colour is lost on drying and identification can only be made by comparison with specimens named by Beauverd or Britton. It is considered that corolla colour either N 000 N. Scotland (Approx. 58° 50’ N) oe 0° de ie o B00' ky 3" Ny tecsterorm ie 800 S ol. eu toate ee C. Scotland iY) ~ 700 Ss Re BS 600 Oo SS ee S & @@ 3 @ Zs 500 . “ N. England Sy SS 5 400 = ° 300 200 e® 3g $8 AY OO or Approx. 50° 45’ N ee. (App a oe & Ped ce S. England 12 13. 14 Hey MG Corolla length (mm.) Fig. 7. Geographical scatter diagram of corolla length. Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 349 after anthesis or pollen shedding is a useless character and Beauverd’s two subspecies are worthless and that Britton’s identifications are based on invalid criteria. Corolla length is variable, population means varying from 11-1 to 16-1 mm and individual corollas from 11 to 18 mm, and there are literature records of corollas up to 20 mm. There is a higher proportion of populations with short corollas in the north of Britain than in the south (see Fig. 7). There is no correlation between corolla and calyx length. (xill) Anther and anther-appendage length Until recently anther length has not been used as a taxonomic character, although Beauverd (1916) used the relative length of the basal hairs of the anthers and the length of the anther appendages. The uselessness of this is seen in that different ‘taxa’ dis- tinguished on this character sometimes occur in the same population and even on one plant. Whilst attempting to find a character to separate badly pressed specimens of M. sylvaticum from depauperate forms of M. pratense it was found that there was great vari- ation in the anther length of M. pratense and that anther length was of no use in sep- arating the two species. Anther length (length of loculus plus appendage) varies in Britain from about 2 to 3-5 mm. Populations with anthers 2:5 to 3-5 mm are widespread throughout the country but those with anthers less than 2-5 mm long are limited to northern and montane areas (see Fig. 8). There is no correlation between anther, corolla and calyx lengths. Obser- vations on Continental material revealed that average anther length decreases northwards and Jasiewicz (1958) remarks that plants with small anthers are montane in Poland. N 000 (Approx. 58° 50° N) eee! she N. Scotland 900 : 800 8 S C. Scotland S 700 ge : 600 © 3 500 N. England oy = aS S 400 = 300 200 (Approx. 50° 45’ N) S. England 100 Anther length (mm.) Fig. 8. Geographical scatter diagram of anther length. Watsonia 5 (6), 1963. 350 A. J. E: SMITH Jasiewicz (1958) divides M. pratense into two groups, the first mainly lowland with anther loculi (2:2—) 2-4 — 2-7 (-3-2) mm long and appendages (0-:2-) 0-3 —0-5 (0:6) mm long and the second mainly montane with anther loculi (1-6—) 1-8 — 2:2 (2:7) mm long and appendages (0-5-) 0-6—0-8 (-1-0) long. It is clear that the proportions between loculus length and appendage length are different in the two groups and are of importance in separating them. Measurements of loculus length and appendage length of the anterior pair of anthers in flowers from each population sampled were carried out and it was found that appendage length was directly proportional to loculus length and that there were not two groups. The same was found of Continental material throughout the range of the species. Hence it seems unlikely that there are two types of anthers, as suggested by Jasiewicz, unless they are restricted only to Poland. Insufficient Polish material was seen to come to any con- clusion about this. (xiv) Pollen For each sample, twenty-five pollen grains were measured from each of three plants, making seventy-five pollen grains per sample. In most of the samples the pollen was 95-100% good, but about 2% of the plants examined had up to 50% of the pollen grains empty. Pollen varied in diameter (18—-) 19-22 (-24)u. In some populations pollen grains were elliptical, measuring approximately 18 x 21y, but in most the pollen grains were spherical. In a few populations there were occasional larger pollen grains, up to 27, these presumably being unreduced. Whilst the average diameter in populations was mostly 20 to 21u, two samples had mean diameters of 23 and 23-6 and one had 18-5u. It does not seem that variation in pollen size can be put down to differences in chromosome number as in all the samples on which chromosome counts were made, including the population with small pollen grains, the diploid chromosome number was eighteen. Pollen of M. sylvaticum varies from 21-27 in diameter and, hence, pollen size cannot be regarded as a reliable character for separating the two species. GEOGRAPHICAL VARIATION From Fig. 1 it is clear that height diminishes northwards and from Fig. 4 it can be seen that there is a decrease in leaf width in relation to leaf length northwards. In southern Britain there is also a slight decrease in height from east to west. There is an increase in calyx length northwards, brought about by the occurrence of some populations with longer calyces than those of the south. Similarly there is a slight decrease in corolla and anther lengths northwards, but the cline in these two characters is not smooth, as popu- lations from central Scotland have shorter corollas and anthers than populations elsewhere in Britain. This is seen in Figs. 7 and 8. VARIATION IN CONTINENTAL M. PRATENSE As British material of M. pratense showed distinct geographical trends in certain characters it was thought that it might be informative to study the geographical variation of M. pratense on the Continent. To obtain a comparison with British material, British and Continental specimens were divided into the following groups which are designated by the prefixed letter or figures in the geographical scatter diagrams (Figs 9 to 11.). i S. England (v.c. 1-35, 51° N) ii N. England (v.c. 60-70, 54° 15’ N) iii Central Scotland (v.c. 88-89, 56° 20’ N) iv N. Scotland (v.c. 105-108, 58° N) Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 351 NN N Norway (68° 30’ N) P Poland (52° N) SN S Norway (61° 30’ N) € Czechoslovakia (50° N) NS N Sweden (62° N) Ba Bavaria (49° N) SSS Sweden (58° N) T Thuringia (51° N) F Finland (62° N) LA Lower Austria (48° 15’ N) D Denmark (56° N) G Geneva (46° 15’ N) H Holland (53° N) I Italy (northern) (45° N) B Belgium (51° N) J Jugoslavia (45° N) Only areas from which more than twenty-five specimens were seen have been used in plotting the scatter diagrams. The countries or districts used do not represent the dis- tribution of the plant but merely the areas in which it has most frequently been collected, or from which a sufficient quantity of material was seen. The herbarium at Geneva contains a collection of M. pratense from neighbouring French and Swiss provinces and hence the name of the city has been used to designate a geographical area. Height Whilst height decreases northwards the diminution is much more rapid in Britain than in north-west Europe. Plants from Scotland are markedly smaller than those from similar latitudes in Norway and Sweden and, even in the extreme north of Norway, the plants hardly approach the small stature of those of northern Scotland (see Fig. 9). 65° | 60° 55° Degrees Latitude North 50° | 45° } io. 18 a oS 3 Height (cm.) Fig. 9. Geographical scatter diagram of height showing differences between British material (crosses) and Continental material (solid dots). For interpretation of lettering see text. Watsonia 5 (6), 1963. 352 A. 3. ESM Leaf index The slight increase in leaf index from latitudes 45° N to 70° N on the Continent contrasts markedly with pronounced increase in that of British material, as can be seen from Fig. 10. Number of branches The number of pairs of branches of British plants does not differ noticeably from that of northern European plants although from the different trends in height this is unexpected. Calyx length In Britain there is a northward increase in calyx length and the trend appears more extreme than on the Continent. Degrees Latitude North 5 10 15 Leaf Index (length/width) Fig. 10. Geographical scatter diagram of leaf index (length/width) showing differences between British material (crosses) and Continental material (solid dots). For interpretation of lettering see text. Corolla length Whilst there is a slight decrease in corolla length northwards in British plants, the variation in Continental material does not reveal any noticeable northerly trend. Watsonia § (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. B50 Anther length Except in central Scotland the anthers of British plants tend to be longer than those of Continental plants from similar latitudes and in Europe as a whole there is a steady de- crease in length from south to north (Fig. 11). @iin - 65° @F @6NS @5SN SS 1607 S < r. < @ss = ><) = Yaw 8 —] @bd S ose S So Kw = @xH ap @T @B Xe) 50° 4 @ Ba @.a eG A oe el ef 2:50 22S 3-00 Anther length (mm) Fig. 11. Geographical scatter diagram of anther length. Crosses, British material; solid dots, Continental material. For interpretation of lettering see text. General comments It might be expected that plants of M. pratense from west and north Scandinavia would resemble plants from northern Scotland, but this is not so. Northern Scandinavian plants are taller, have markedly broader leaves, slightly longer corollas and the same number of pairs of branches (or cauline leaves.) This gives the plants from these two areas a strikingly different appearance. These differences may be due to climatic and day-length differences in the two areas. Woodland plants from other parts of Scandinavia resemble more closely British plants though few are as small as those of many populations found in some parts of Britain. Plants collected from limestone areas in various parts of Europe from Gotland southwards closely resemble the forms from chalk and limestone in Britain. All plants from Belgium and Holland are very similar to plants of south and east acid-woodlands in Britain. Watsonia 5 (6), 1963. 354 A. J. E. SMITH RESULTS OF STATISTICAL ANALYSES In Tables 2 and 4 the means of particular characters for the populations of each ecological type are arranged in ascending order of magnitude. In Table 3 the means of variabilities (obtained by taking the logarithm of the standard deviations) are likewise arranged. In Table 5 again the means of variabilities (obtained by squaring the standard TABLE 2 Significance of differences between means (in brackets) of ecological types A to G. Any two means not underscored by the same line are significantly different. Any two means underscored by the same line are not significantly different using the SNK test. Character and level of significance Ecological type and mean Height (cm) A G D E F B G yo (27-8) (22-4) (16:5) (12-8) (11-8) (11-2) (10-4) I% A C 1D) E F B G No. sterile nodes A GS G D B E iF Sf 7) (4-67) (4-53) (3-79) (3-73) (3-63) (3-59) Ios A C G D B E F Total No. nodes A (C G D B E BE Sarandelio7 (11-6) (10-1) (8-18) (8-14) (8:01) (7-89) (6:91) No. intercalary leaves A G C B D E li Sypand 1% (1:79) Ghe7/7)) (1:24) (1:18) (1:06) (0-99) (0-93) Leaf shape (1/b) A Cc D B E F G Syjeande1o% ©:53) (8-29) (9-44) (0-29) (1-11) (2-95) 5Gs-c7) No. of entire bracts A (G F D E G B Of (1-26) @=73) (2-23) (2:28) (2:36) (2-56) 2-73) 1% A C F E G B No. of teeth per bract A CO} E D B G F S/orandeliy, (6°52) (5-48) (3-41) (3:25) (2°81) 2-420) (1°85) Calyx length (mm) G D A E F B G 3% (il) (5:2) 6:25) S645) ene) (5-7) (6°35) 1 EI D A E F B G Corolla length (mm) D A € B E G F a and 1% (14-8) (14-4) (14:1) (14-1) (13-6) (13-6) (12:8) Anther length (mm) B D GC A G E F 5% (3-25) (3-17) (3-16) (3-08) (2:88) (2-70) (2:42) 9% B D C A G E F deviations) are treated similarly. Although the variabilities in Tables 3 and 5 were obtained by different methods used in the original research, they are essentially similar since the log. of the square of the standard deviation (variance) is equal to twice the logarithm of the standard deviation. These variabilities represent the degree of genetic variability. The significance of differences between the means and variabilities have been analysed statistically as described on p. 341 at the 5% and 1% levels. Means or variabilities that are Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 355 TABLE 3 Significance of differences between variabilities of ecological types A to G Character and level Type and variability mean of significance Height Not significant 3% and 1% Leaf shape F (CG A ES B D G 5% (1238) (1244) (1260) (1263) (1296) (1336) (1361) es Not significant Calyx length Not significant 5% and 1% Corolla length F E A B é D G Syoand 1% (625) (701) (708) (710) (712) (783) (821) Anther length Syqcand 1, Not significant not significantly different are underscored by the same line. Where differences are sig- nificant the means or variabilities are not underscored by the same line. This provides a pictorial method of comparing ecological types in any one of the characters analysed. From Table 2 the following points can be seen: Type A is significantly different from all other types in five characters at the 5% level and four characters at 1%. Types C and F are significantly different from all other types in two characters at the 5% and 1% levels. Types B and G are significantly different from all other types in one character at the a, level: The morphological differences between type 4 and other types are greater than those between any of the other types. In view of the marked ecological difference as well, type A forms a distinct unit as will be mentioned later. It is evident that the various ecological types differ from one another to a greater or lesser extent. It is also clear, however, that the characters in which the differences occur are not necessarily the same for any two types. From Table 4 it can be seen that there are significant differences between populations from Scottish bogs (type G, samples 70-80) with respect to the number of nodes and calyx and corolla length, and similarly with populations from calcareous habitats (type 4A, samples 101-109) in Kent. This suggests that, whilst there is selection for particular morpho- logical forms in particular habitats, there are distinct genetical differences between closely sited populations in any one ecological type. Tables 3 and 5 both indicate that there is relatively little difference in genetical variability between individual populations and between ecological types. The significance of this will be discussed later. CYTOLOGY Root-tips of pot-grown seedlings from fifteen samples were stained with feulgen and squashed in aceto-carmine. All had a diploid chromosome number of 18 and all except four had the chromosome morphology illustrated in Fig. 12a. In four samples there were additional pairs of satellites as illustrated in Figs. 12b to 12e. The constrictions forming the satellites mark the position of nucleolar organisers and it has been suggested (K. Lewis, personal communication) that where chromosomes are Watsonia 5 (6), 1963. JJ. ES SMITE 356 601 cOl We 8L LOI 901 cl, OL 801 10T fit EL OL SL 08 Zot et (ps1) (+t (bl) O-rD (7D G&D (PD Grd) ©rp “ED GED OD ED 7) GID %S 601 cOl IL SL LOI 901 ZL OL SOI 101 LL EL OL Ej 08 YyIsua] DjJO1OD ST (L6-) (86-b) (SI-S) (8h-S) (19-5) (96:5) (66:5) (L0-9) (Or-9) (8r-9) (OS-9) (95-9) (S69) (46-9) (8-L) %S 80l COL 601 901 LOL EL GIL 9L OL SL 101 08 IL Gs ee (uu) yp Sua] xdJOD (t->) (pb) (0-5) (€-L) 1-8) (-8) @ZD G-stD SI) (SI) (0-91) (1-91) (91) (LI) (78D YT Puo %oS 801 601 €Ol LOI 101 901 SL EL SL CL OL jit Ve 08 OL adoys fo0aT €Ol LOI 801 SL 101 601 ZL 901 €L 08 jijh QL IL SL OL SI (9-71) (8-ID (G-ID (8-01) (-01) (6) (8-8) (9-8) (6-8) (¢-8) (8) (8) @L GL ©-2L) oss €Ol LOI 80l SL 101 601 (ofl 901 EL 08 LL 9L IL Gl OL sapou fo ‘ON €-[ 601 801 LOL 901 101 SL GL vl 08 EL Ue QL SL IL Zot (€-1€) (S-6Z) (-8z) (€-Sz) (-€Z) (€%) CrI) OL) WOLD (€-01) COD (6) ©-6) (6) (ZL) % €OI 601 SOI LOL 901 IOL SL TL PL 08 EL LL 9L CL IL (Ud) 14810] uDau pub uoljvjndog JIUDIYIUSIS JO 12h] PUD 4dJIDADY D (pV 9d4q “judy UT S}eIQeY SNOoIvITeD WO 60] 0} [OT ‘dod ‘Hy od4q ‘s8o0q Ysi}}09g UIsy}1OU WI Og 01 OL ‘Sdod) suoljeindod ¢] JO Suva USAM}9Q BOUdIOYIP JO SOUBOYIUSIG “p ATAV], Watsonia 5 (6), 1963. 357) VARIATION IN MELAMPYRUM PRATENSE L. JUBOYIUSIS JON % (€-16) (6-SS) (1-97) (2-Sbh) (-rr) (8:7) (Ip) (O-LE) (7-7E) (0-7E) (-1E) (1%) SI) C-SI ED %S CL SL EL LOI cOl 601 IL 9L jf 08 101 OL 901 801 GE YIZUI] VJJO1OD (GED (Gop OOD 186) “G6), “Ce6) (98) (78) (08) (IL) (99) (8h) (€bh) ©-ET) (7-6) %oT Puw YS LL IL 80l cOl 601 EL 101 SOL cL OL LOI gL OL 08 TL yisua] XAJDD JUBOYIUSIS JON] a adoys {vaT JUVSYIUSIS JON sapou fo “ON JUBOHTUSIS JON MI (0-77) (€-91) (L-€1) 7D Il) (ID @ID (ID Or) (8-01) (01) 6) WL) 9-9) G&D oS 80 OL il SL LOL tL OL CL 901 08 EL ee 601 0) 101 JYSIOET PIUDIYIUGIS ubau Apigoliva puv uoljvojndog fO Jada] puvo sajavsvyy) (p IIQeL ul se) Suonelndod useajyYy JO SarjIqei4seA UsdMJOq DOUdIOYIP Jo souKsyIusIG “CS ATV], Watsonia 5 (6), 1963. 358 A. J. E. SMITH 4 ee ? aN ” P) el —> 86" gs iy 4 i. Zt & tt wo —» 4 "SAF, : 32 2. @ w tA iy a Sst S$ — ¢ a QS es ca hoon Ee 0 BR a wh os b ~ 8. as ¥ \ sh. “> an i\] \ e =e & r ®& cmos a ~ | ra Me » So eo? a i” ey es ee \ oe, te ¢& Nils Sass oe ste co] 4 pt} Se 7] wee ar So 4 & Q “% 5 ii soe d lOp Fig. 12. Camera lucida drawings of root-tip chromosomes of M. pratense; (a), normal chromosome com- plement (sample 96); (b)—(e), abnormal chromosome complements with additional satellites and/or non-homologous chromosomes (samples 99, 26, 100 and 97 respectively). non-homologous this might be due to the suppression of some of the nucleolar organisers by those remaining. The chromosomes are very heterochromatic and, because of staining in the resting nuclei, it is not possible to count the number of nucleoli. Staining of the root tip chromosomes with feulgen and aceto-carmine was very satis- factory, but great difficulty was had in staining meiotic chromosomes in pollen-mother- cells. Aceto-carmine and aceto-orceine produced poor results with fresh material although the former stain was satisfactory with material fixed in 1:3 acetic alcohol. Best results were obtained by warming fresh anthers for about five minutes in a mixture of five drops each of aceto-orceine and aceto-lacmoid and three drops of normal hydrochloric acid. Examination of fresh material was desirable as in fixed material the chromosomes become very woolly in appearance. It proved difficult to find dividing pollen-mother-cells but sucessful preparations were made from material from ten populations. In all these the results were identical and the chromosomes always showed the same pattern of behaviour. At metaphase eight bivalents became arranged around the ninth on the metaphase plate as in Fig. 13a. At anaphase the eight bivalents appear to move to the wall of the pollen-mother-cell and fuse with each other at the ends to form a ring (Fig. 13b). That they fuse is suggested by the fact that when heavily squashed the dispersed chromosomes remain joined by bridges (Fig. 13d). The central bivalent becomes attached at its ends to the inside of the ring. In fresh material, however treated, the pollen-mother-cells always become orientated so that there appears to be only a single ring of chromosomes at ana- phase. In fixed material orientation of the pollen-mother-cells is at random and it is clear in such preparations that there are two rings. At telophase (Fig. 13d) the chromosomes become separated but still have the same arrangement as at metaphase and this is retained until the later stages of the second division. 7 In M. sylvaticum there is the same phenomenon at anaphase but metaphase was not seen and the chromosomes were too woolly to count in the preparations made. Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 359 ‘ axl | oge rh Ue “ Fig. 13. (a)—(c), stages in meiosis of pollen mother cells of M. pratense ( (a), metaphase; (b) anaphase; (c), telophase); (d) heavily squashed cells showing apparent bridges between chromosomes at anaphase. The chromosomes of M. pratense are relatively large when compared with those of related genera such as Rhinanthus and Euphrasia and it may be that their peculiar behaviour is only a mechanical effect dictated by the lack of space for random orientation in the pollen-mother-cells. The bivalents are not sufficiently different morphologically to determine whether the same chromosomes always occupy the same position at anaphase. Resting nuclei of both root-tips and pollen-mother-cells are very heterochromatic and the same situation is reported in M. cristatum (Hambler 1954). Melampyrum differs in this respect from Rhinanthus, the species of which have varying numbers of heteochromatic B chromosomes (Hambler 1954) which form stainable blobs in resting nuclei. DISCUSSION It is evident from the foregoing account that there are three types of variation in M. pratense, (i) geographical variation, (ii) ecological variation and (iii) variation from population to population. (1) Geographical variation In Britain there is a marked south-north cline in such characters as height and leaf shape. These clines can be explained on the basis of natural selection acting through climate. Such clines have been shown in Alnus glutinosa (McVean 1953). With calyx, corolla and anther lengths the pattern differs. There is not a gradual change in size, but there is a change in the proportion of populations with floral parts of particular dimensions. It is difficult to conceive of a reason for this. If, for example, short anthers and long calyces are of selective advantage in the north then it might be expected that they would have been selected for in all populations in such areas. In northern Scandinavia there is a higher proportion of such populations than in Scotland and it may be that short anthers and long calyces (which are not necessarily selected for together in any one population) have a slight advantage under more extreme climatic conditions. This being so, selection for them would increase in a northerly direction and they might even be selected against in the south. In Britain this particularly applies to anther lengths, there being no populations with short anthers in the south. Similarly with calyx length there are very few specimens with long calyces from southern Europe. Watsonia 5 (6), 1963. 360 A.J) SE. SSMITE Corolla colour is also a puzzling feature. Taking first the golden-yellow form, this occurs in pure populations in palaeogenic regions of Britain and with the possible exception of a record from Wurttemburg is apparently absent elsewhere except in Ireland. That it occurs in pure populations indicates that it is of selective advantage and when a mutation to this colour form occurs it spreads through the whole population. If it was a normal mutant it would be expected to occur throughout the whole geographical range. It is possible that the selective advantage of golden-yellow flowers is connected with a particular pollinator in certain areas. The plants with crimson-tipped corollas are northern and montane and in Britain they mainly occur mixed with the normal form. The situation on the Continent is unknown. As will be mentioned later M. pratense is usually self-pollinated so that these mixed popula- tions are unlikely to be instances of balanced polymorphism, and the fact that pure popula- tions of the crimson-flowered form are rare suggests that it is of no particular selective advantage. As with the golden-yellow form its absence from a considerable part of the species’ range is inexplicable. (11) Ecological variation Whilst some characters show an almost continuous geographical cline, others, for a considerable part of their distribution, show a similar cline, but there are certain anomalies which appear to have an ecological basis. The most conspicuous irregularity is in the cline of intercalary leaf number and is caused by the northern bog populations (see Fig. 3) which have a larger number of intercalary leaves than any other populations except those of type A. Evidently plants with numerous intercalary leaves have been selected for in Scottish blanket bogs. Plants from calcareous habitats might also be said to cause discontinuities in the leaf- shape cline (Fig. 4) though, as they all occur in southern Britain where leaves are broader than elsewhere, this is not obvious. Leaf shape of the calcicolous forms is significantly different from all other forms, but otherwise the northerly change in leaf shape is more or less continuous. It seems that the broad-leaf character has been strongly selected for in nearly all basic habitats, both in Britain and on the Continent. Whether or not broad leaves are of selective advantage in calcareous habitats is not clear, but they may be an external manifestation of a physiological difference. It was mentioned earlier that some chalk populations in Kent were indistinguishable from populations occuring in acid pockets in the same area. There are two possible explantions for this. The first is that there may originally have been two types of population, one calcifuge and narrow-leaved and the other calcicole and broad-leaved. In Kent conditions on the chalk, where there are acid pockets of soil, would provide adjacent habitats for the two types. Since they would be close together hybridisation and introgression would be pos- sible. None of the populations from chalk soils in Kent have such broad leaves as those from calcareous soils in Hampshire, Somerset and Gloucestershire where the populations appear to be more isolated. This suggests that the narrower leaves of some Kent populations may be due to introgression. The second possibility is that originally there was only one taxon composed of narrow- leaved plants and that there was selection for broad-leaved forms where populations developed on calcareous soils. This selection would presumably be relatively slow and it might be that in Kent populations from acid soils might still be spreading to calcareous soils, giving forms with a variety of leaf shapes. From each of the ecological types mentioned (i.e. A to H) plants of characteristic appearance have been collected (Fig. 14-19), although with the partial exception of type A, these various types intergrade to form a continuous series of morphological forms, i.e. they are a series of ill-defined ecotypes. This ecotypic differentiation has been super- Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 361 imposed upon continuous geographical variation so that discontinuities in clines of certain morphological characters have been produced. One very marked exception to the pattern of variation in southern Britain, other than the calcareous plants, is the form found in the New Forest. The reason for this exception can be dealt with in considering variation in M. pratense in Britain as a whole. It is possible that at the end of the last Ice Age the form of M. pratense found in southern Britain was like the form now found in northern Scottish bogs. As the ice retreated this form would have migrated northwards. As climatic and vegetation changes occurred it is pos- sible that, beginning in the south, this bog form became more variable because of less rigorous climatic selection and diversification of habitats. In any one type of habitat particular gene combinations were probably selected for, resulting in the phenotypes characteristic of the particular habitats. The anomalous New Forest populations (type B) could be relicts of an earlier phase in the post-glacial period, as they much more resemble northern plants, especially those from blanket bogs, than other forms. This does, however, raise the question why this form of M. pratense has persisted in the New Forest whilst elsewhere in southern England there appears to have been selection for different and more vigorous forms. Furthermore this suggestion implies the existence of wet acid peat in southern Britain in the late-glacial and early Post-glacial periods. All the evidence suggests that the only boggy areas were base-rich shallow valley mires and fens, totally different types of habitats from the wet acid peat of northern Scotland and the dry acid podsols of the New Forest. Fig. 14. Plant of type A (subsp. commutatum) from an oolitic limestone bank, Wotton-under-Edge, Gloucestershire (grid ref. ST 954940), x 4. There is an alternative and more likely explanation which will account for geographical variation of the species in Britain and also for the existence of the New Forest populations. The New Forest plants and those from northern Scottish bogs differ markedly from Continental plants, but other forms of M. pratense in Britain, expecially those from southern England, resemble closely those forms from similar latitudes in western Europe. Watsonia 5 (6), 1963. 362 A. J. E. SMITH It is suggested that M. pratense migrated from southern Britain northwards after the last Ice Age, either as a periglacial survival or as animmigrant. As the migrant forms spread northwards there was a diminution in size, possibly because of a more rigorous climate, and a diversification of forms resulting from the colonisation of new habitats. Fig. 15. Plants of type B (New Forest form) from Warwick Slade, near Lyndhurst, Hampshire (grid ref. SU 268065), xi. Cf, Fig. 18. If, as is suggested by McVean & Ratcliffe (1962), northern Scotland was until recent times wooded, suitable habitats for M. pratense would have been present. When the woods were destroyed M. pratense could have survived on the subsequent blanket bog as a woodland relict. This may also account for the absence of the species from raised bogs in England and Wales, since these may never have been wooded (e.g. Tregaron Bog, Borth Bog). Plants on nutrient deficient soils, such as the acid peats of northern Scotland and the podsols of the New Forest, tend to be small and the similarity between New Forest plants of M. pratense and those from Scottish blanket bog may be entirely accidental. Status of ecological variants of M. pratense From Table 2 it can be seen that in types E, F and G for certain (and not necessarily the same) morphological features any one of the types is significantly different from the Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 363 others, indicating that there has been selection for a particular expression of a morphological character in a particular habitat. In none of these types, however, has selection proceeded sufficiently far to delimit well-defined ecotypes. From a taxonomic point of view all the types from northern Britain (i.e. E, F and G) can be regarded as a single taxon. Type H which was not dealt with statistically, can, from its morphology, be regarded as falling within this taxon. In southern Britain the situation is different. Again referring to Table 2 it is evident that, ignoring type B, types C and D are indistinguishable except in the degree of toothing of the bracts and can be included in one taxonomic unit. Type A differs considerably from all other types, in five characters at 5% level and in four at the 1% level. The situation here seems to be either that ecological selection has pro- ceeded further than in other ecological types, or, as was suggested earlier, there were two distinct types of plant to start with. In Europe plants of M. pratense from calcareous habitats are usually distinct although in a few areas, e.g. Pas de Calais (F. Rose, personal communication), plants with narrower leaves like those from Kent are found in such habitats. 3 Fig. 16. Plants of type C from Fence Wood, near Newbury, Berkshire (grid ref. SU 511717), x ¢. Even when the intermediate Kentish populations are included, the calcicolous forms are sufficiently well-defined to be regarded as a taxon distinct from that containing types Cand D. Types C and D differ significantly from all of types E, F and G only in the number of nodes and toothing of bracts. Although types C and D tend to be larger in number and size of parts than the northern types there is a completely intergrading series, so that type C cannot be considered as distinct from the northern forms. There are a few New Forest populations tending towards the plants of type C and it is considered best to regard Watsonia 5 (6), 1963. 364 ASC ES) SMITE the New Forest plants as a form of type C. Thus there are two taxa, one consisting only of type A and the other of types B to H. For convenience, nomenclature being dealt with later, these will be referred to as subsp. commutatum (plants from calcareous habitats) and subsp. pratense (plants from acid habitats). Plants like those of type C are widespread throughout the geographical range of M. pratense. On the Continent there is a multiplicity of variant forms but from examination of herbarium specimens it seems that many of the Continental forms are of restricted distribution and that selection has produced different forms in similar types of habitat. The exact status of many of these forms is not clear as insufficient material has been seen, but it is suspected that the situation is as in Britain, i.e. that these forms are ill-defined. With the exception of type 4 plants, all material seen from the Continent comes within the range of subsp. pratense. Plants of subsp. commutatum are widespread being found in many of the areas where there is base-rich rock. It is clear that subsp. commutatum and subsp. pratense are eco- logically and to a considerable extent morphologically distinct and hence the most suitable taxonomic status for them is that of subspecies. Fig. 17. Plants resembling types E and F from Struan, Perthshire (grid ref. NN 806655), x 4. (111) Variation at the population level That there is variation between individual populations is evident from Table 4, but before considering this it is necessary to deal with the genetics of the species. In Britain M. pratense usually occurs in small isolated populations. According to Knuth (1906-9), the plant is pollinated by bumble bees and failing this, it selfs. No obser- vations were made of insect visitors, but in several samples when pollen grains were being Watsonsa 5 (6), 1963. | } | } | f | ; 1 | VARIATION IN MELAMPYRUM PRATENSE L. 365 Fig. 19. Extreme form of plants of type H from Wybunbury Moss, Cheshire (grid ref. SJ 695502), x 1. measured it was found that the anthers dehisced before the flowers opened. During the course of cultivation experiments at Bagley Wood, where M. pratense does not occur naturally, a solitary plant flowered about three weeks before any others and set seed, again suggesting that there is no self-incompatibility mechanism. Further, in Britain, many of the popula- tions of M. pratense are so isolated as to form small inbreeding units and such populations could be expected to be more or less homozygous. Watsonia 5 (6), 1963. 366 A. J. E. SMITH If the populations are homozygous it would be expected that the level of genetic variability both from population to population and from ecological type to ecological type would be relatively uniform. That this is so is indicated in Tables 3 and 5. It was suggested earlier that M. pratense spread across Britain after the last Ice Age and it is probable that populations became isolated, either by fragmentation of large populations or the distribution of solitary or small numbers of seeds. Random assortments of genes would become fixed in individual isolated populations. This would account for the many minor but noticeable differences between individual populations in similar ecological habitats. From the point of view of general inter-population variability M. pratense is morpho- logically much more diverse in southern Britain than in the north. This may be due to greater ecological diversity or to the less extreme climatic conditions of the south or a combination of both. At present M. pratense is decreasing in wooded parts of Britain because of felling of some woods and the overgrowth of others. In northern Scotland it appears to be limited to waterlogged areas that do not get burnt in the periodical firing of heather that is carried out in connection with grazing. TAXONOMY OF M. PRATENSE In past investigations of M. pratense solitary plants or small numbers of plants have been collected in different parts of Britain and the Continent. Because of geographical, ecological and population differentiation these plants have appeared to be distinct from one another and this has led to the description of many subspecies, varieties, subvarieties, forms and subforms, which do not have any validity when a large number of plants collected continuously over a geographical region are examined. It has, however, been found that plants fall into two morphological groups that are sufficiently distinct to be considered as subspecies i.e. the subsp. commutatum and subsp. pratense mentioned earlier. The specimen of M. pratense in the Linnean herbarium comes within the range of those plants which have for convenience been called subsp. pratense and which should there- fore be called M. pratense L. subsp. pratense. The plants with golden-yellow flowers belong to subsp. pratense. This form was first recognised as var. hians by Druce (1884) and later regarded as a subspecies by Beauverd (1914). The taxon is probably best regarded as a variety, being characterised by flower colour and a distinct geographical distribution. The correct name is therefore M. pratense L. subsp. pratense var. hians Druce. The plants with crimson-tipped corollas cannot be regarded as having any taxonomic status since they do not occur in pure populations and have no distinct distribution, but can for convenience be called f. purpureum. Plants resembling subsp. commutatum were distributed as M. commutatum Tausch by Tausch in 1832 but with no description. The original specimens distributed by Tausch have not been seen but from later-named specimens and descriptions it is clear that Tausch’s plant was the one that for convenience has been called subsp. commutatum. Tausch’s plant was described by Kerner in 1870 who said it should be regarded as a variety of M. pratense, and the correct citation would be M. pratense var. commutatum Tausch ex Kerner. The plant was raised to a subspecies by Britton in 1943 and, as this appears to be the correct taxonomic status, the valid name of the plant is M. pratense L. subsp. commutatum (Tausch ex Kerner) C. E. Britton. Descriptions of the subspecies M. PRATENSE L. subsp. PRATENSE: Plants (3-)9—25(-60) cm high (excluding hypocoty]). Uppermost cauline leaves (10—)20-80(-110) mm long, (1—-)2—10(-20) mm wide, leaf length/ Watsonia 5 (6), 1963. VARIATION IN MELAMPYRUM PRATENSE L. 367 breadth ratio (S—)7-15(-20). Calyx 3.5-7(-12) mm long. Corolla tube whitish or pale yellow, mouth pale yellow, usually darker than the tube; or mouth crimson-streaked; or whole corolla golden yellow (var. hians Druce); (11—)12-14(-18) mm long. Anthers (2.0-) 2.4—3.2(—3.8) mm. Plants of acid soils in woodlands, hedgerows and bogs, widely distributed but rare in East Anglia, the Midlands and southern Scotland. M. PRATENSE L. subsp. COMMUTATUM (Tausch ex Kerner)C. E. Britton: Plants (20—-)25-45(—60) cm high (excluding hypocotyl). Uppermost cauline leaves (30—)40—-70(-100) mm long, (4—)8—20(—27) mm wide, leaf length/breadth ratio (2-)3-8(-12). Calyx (3.5—-)4.0-5.5(-6.0) mm long, corolla 13-15.5 mm, anthers 2.5-3.5 mm. Restricted to calcareous or base-rich habitats, usually a scrub or hedgerow plant, common in south-east England, rare in the south-west. Also recorded from Oxfordshire, Somerset and Lancashire but probably extinct in these counties. ACKNOWLEDGMENTS I wish to express my thanks to Dr. E. F. Warburg for his help and guidance during the progress of this work, to the curators of the many herbaria from which material has been examined, and to the numerous individuals who lent specimens and provided information. The work was carried out at the Botany School Oxford, during the tenure of a post- graduate research studentship awarded by the Department of Scientific and Industrial Research. REFERENCES BEAUVERD, G. (1916). Monographie du genre Melampyrum. Mém. Soc. Phys. Genéve 38, 219-653. BriTTON, C. E. (1943). The genus Melampyrum in Britain. Trans. Bot. Soc. Edinb. 33, 357-379. Day, B. & FisHer, R. A. (1937). The comparison of variability in populations having unequal means. Ann, Eugen. Lond. 7, 333-348. DECAISNE, (1847). Sur le parasitisme des Rhinanthacées. Ann. Sci. nat., sér III 8, 5. Druce, G. C. (1884). A new variety of cow-wheat: Melampyrum pratense var. hians. Naturalist, Lond. 10, 35-36. HAMBLER, D. J. (1954). Cytology of the Scrophulariaceae and Orobanchaceae. Nature, Lond. 174, 838. JAsIEWIcZ, A. (1958). Polskie gatunki rodzaju Melampyrum L. Fragm. Flor. Geobot. 4, 17-120. Kerner, A. (1870). Ueber einige Arten der Gattung Melampyrum. Ost. bot. 20, 265-273. KwnutTH, P. (1906-9). Handbook of Flower Pollination. McVEAN, D. N. (1953). Regional variation of A/nus glutinosa (L.) Gaertn. in Britain. Watsonia 3, 26-32. McVEANn, D.N. & Ratcuirre, D. A. (1962). Plant Communities of the Scottish Highlands. London. RONNIGER, K. (1914). Sched. ad fl. exsicc. austro-hung., nos. 3697-3700. Soo, R. von (1927-8). Systematische Monographie der Gattung Melampyrum, I-Il. Repert. nov. Spec. Regn. veg. 23, 159-176, 385-395; 24, 127-193. STEEL, R. G. D. & Torrir, J. H. (1960). Principles and Procedures of Statistics. Yeo, P. F. (1961). Germination, seedlings and the formation of haustoria in Euphrasia. Watsonia 5, 11-22. Watsonia 5 (6), 1963. FERTILE SEED PRODUCTION AND SELF-INCOMPATIBILITY OF HYPERICUM CALYCINUM IN ENGLAND By E. J. SALISBURY ABSTRACT Hypericum calycinum is shown to produce, in England, a small number of viable seeds which, in well- grown capsules, approximates to a constant proportion. Ovaries contain an average of some 800 or more ovules which, following self-pollination, exhibit a wide range of development. Well developed seeds yield a far higher percentage germination if sown when immature and it is suggested that self-incompatibility mani- fests itself not only through the first male gamete by degrees of failure in embryo development but also, through the second male gamete by inefficient endospermic nutrition during the attainment of the resting stage in this exalbuminous species. Germination, which extends over a period of months, is of the inter- mittent type, whether at normal or higher temperatures, and appears to be photoperiodically indifferent. Seedlings have not, however, been observed near colonies and possible reasons for this are discussed. The Rose-of-Sharon (Hypericum calycinum) is found growing wild in south-eastern Europe and was introduced into this country in 1676 by Sir George Wheeler who had collected the plants from the neighbourhood of Constantinople (Bot. Mag. t. 146 (1796) ). It has thus been with us not far short of three centuries and is more or less naturalised in many localities throughout Great Britain and Ireland, especially in the southern counties of England (Perring & Walters 1962, p. 57). More than sixty years ago, however, Hanbury & Marshall (1899) commented upon the fact that, although then well established in a few places in Kent, the Rose-of-Sharon apparently never spread in this country ‘ except by root extension’. So too, that excellent field botanist, James White, stated that ‘ this species is believed not to ripen seed anywhere in this country but to spread and maintain itself by root-extension only (White 1912). Most gardeners and field botanists would probably subscribe to this generalisation although the phrase ‘ root-extension ” is liable to mislead, since the vegetative spread of H. calycinum is mainly, if not exclusively, by means of axillary underground stems. The distinction is significant since in other members of this genus, for example H. perforatum and H. pulchrum (see Salisbury 1942, Fig. 32), it is actually the roots which are responsible for the vegetative extension, since adventitious shoots arise from them. The production of such root-shoots can be greatly stimulated in H. perforatum by cutting down the erect stems but I have observed no such response in H. calycinum. Vegetative extension takes place slowly. The stoloniferous shoots (Fig. 3) grow obliquely upwards from the underground portion of the previous year’s leafy shoots. The number arising from each is usually two to four, commonly paired and involving from one to three nodes. Each stolon bears some 8 to 15 pairs of small brown scale leaves below the soil and develops normal foliage leaves on reaching the light. The internodes do not usually attain a length of more than 1-2 cm so that the radial extension in a season rarely exceeds 15 to 20cm. The very large areas, as along the cutting between Leatherhead and Dorking, that are often occupied exclusively by the Rose-of-Sharon are visual evidence both of the efficacy of its stoloniferous spread and its stability. A similar generalisation concerning the absence of reproduction by seed in Britain has been made with respect to the two Periwinkles (Vinca major and Vinca minor) likewise characterised by conspicuous vegetative vigour. Both these species I recorded as having fruited successfully in this country in 1961 (see Salisbury 1961). In the warm late autumn of 1961, examination of capsules of AH. calycinum revealed a small number of, apparently, fully developed seeds and this discovery, as also my experience with the species 368 Watsonia 5 (6), 1963. SELF-INCOMPATIBILITY OF HYPERICUM CALYCINUM IN ENGLAND — 369 DAILY GERMIEATIONS 19-4 fain 60 ons 2) ’ her +. WS SS BS Se Se ey i < . 4 Pies) , €7\PpAYs —> Oo WITH MO GERMmmMs- 1% 4 ’ ot hoe ee Fury — Aue—Sepr AIth iQe . | OVER PERIOD OF 293 DAYS e Fig. 1. Intermittent germination of 300 apparently fertile seeds of H. calycinum, with (inset) details of the fruit, seeds and seedlings. Magnifications as shown by millimetre scales. of Vinca, led me to look carefully at diverse colonies of H. calycinum, both in 1961 and subsequently, from which it would appear that some seed production, though very sparse, is not uncommon, at least in southern England. Normally the Rose-of-Sharon is probably self-incompatible and, again as with the Periwinkles, this would tend to be accentuated by the large colonies formed by the vegetative spread of a single individual, since the attentions of pollinating insects might be confined to the flowers on only one plant, even if another individual were in the neighbourhood. Moreover it is quite likely that the entire population of Hypericum calycinum in this country may have been derived from a single stock. However that may be, the five styles that surmount the ovary are relatively long (see Fig. 1 B) so that perhaps only exceptionally favourable weather conditions when pollination occurs, especially with regard to tempera- ture, can enable the pollen-tubes from the plant’s own pollen to traverse this distance and Watsonia 5 (6), 1963. 370 E. J. SALISBURY grow sufficiently rapidly to effect fertilisation. Eight large capsules obtained in October 1961, and eight more in November 1962, were dissected carefully and the numbers of apparently fertile seeds, abortive seeds and ovules alike counted with the following results. Fully-developed apparently Percentage Obviously Total number of ‘ fertile’ seeds ‘fertile’ seeds abortive seeds seeds in capsule 11 1:8 581 592 18 29 597 615 2D 3-4 629 651 7 1-0 647 654 13 2:0 647 660 37 52 678 715 36 4-7 721 757 24 302 767 791 18 2? 828 846 14 1-6 840 854 15 185) 966 981 2] Dal 964 991 19 1:8 1040 1059 11 1-0 1071 1082 3) Sse) 1058 1093 172 1-0 ISS 1167 Average number of Average % Average number of Average total number fertile seeds 20 [s. d. 2:4 abortive seeds of seeds in capsule 9-6: s.e.m. = 2:4] 824 844 [s.d. 261: s.e.m. = 65] Those sixteen capsules were deliberately selected because of their exceptional size and might therefore be expected to contain the highest proportion of fertile seeds and thus provide an indication of maximum fertility. In these the observed range was from seven to thirty-seven with an average of twenty. A random sample, of a number of other 1961 capsules, was also examined, in which the apparently fertile seeds were alone counted, and these provided an average of only 6-2 per capsule. The apparently fertile seeds were 1-5 to 2 mm in length and from 0:7 to 0-9 mm in width. Almost cylindrical in shape, with rounded ends and a slight ridge along one side, the symmetry of the seeds, as well as their size, usually distinguishes the viable from the abortive ones. The colour is dark brown and the average weight of a seed is 0-:00058 gm (the largest from 0-0006 to 0-0007 gm). This is about five or six times the weight of the seeds of the herbaceous species (see Salisbury 1942); they are, nevertheless, sufficiently small and light for gusts of wind to disperse them readily (Fig. 1, C & F). Subsequently, three hundred of the ‘ apparently fertile ’ seeds, harvested in late autumn, were sown on | May following. The outcome ts presented in the accompanying histogram (Fig. 1) from which it will be seen that germination began within three weeks and was then intermittent through May and June, with a total of only 19% in the first forty days followed by a pause of seventy-one days and then a further surge of germinations bringing the total to 33% in about 54 months with only two more (0-6 %) on 15 December, after 205 days. It should be emphasised that, in order to obtain a sufficiency for a representative test the seeds were collected from various plants at different times. More seeds were collected in 1962. Of these some were harvested from a number of capsules, produced by a single colony and all obtained and sown on the same day, 4 November. There was a total of 183 apparently good seeds but whereas 106 were deep-brown and presumably mature, the remaining 77 ranged in colour from cream to pale-brown and were clearly immature. Both batches were sown side-by-side in the same seed-pan and placed where the tempera- tures during the daytime attained about 20° C and the minimum night temperatures were about 5° C. It was anticipated that the immature seeds might yield a low germination but, in fact, their percentage was the highest observed for the species, namely 92%, and the much lower percentage germination of the riper seeds from the same capsules, namely 62:2 %, Watsonia 5 (6), 1963. SELF-INCOMPATIBILITY OF HYPERICUM CALYCINUM IN ENGLAND Si7/I would strongly suggest inadequate nutrition by the endosperm during the later stages of maturation, a point to which I shall revert. The pattern of germination, in both batches, was again intermittent but, as might be expected from the temperatures being appreciably higher than those in which the 1961 seeds were germinated, it was far less protracted (Fig. 2) occupying about ninety-nine days instead of two-hundred and five. A point that may be of significance is that although the number of viable seeds was almost the same in both the 1962 groups, germinations of the immature seeds took place on only twenty-three separate days compared with germinations on thirty-three days for the mature seeds. When considering self-incompatibility attention has, hitherto, been almost exclusively focussed upon the normal embryo which, in endospermic seeds, has achieved its resting stage before the transfer of food from the endosperm. In an exalbuminous seed such as that of H. calycinum, the resting stage is not attained till after the endospermic nutrition has been transferred. “Wa fe TUT TVITT TTT TTT TUTTI - Dec. N Fee. %e = | 41 HYPERICUM , 40|= CALYCINUM 106 BROWN MATURE SEEDS. 5 = HARVESTED AND SOWN NOVEMBER 4th. © = © 5 = | % ol TOTAL PERCENTAGE GERMINATION @A°A% 2, = BQ « Z 3 be Ale | : is : 5 eed ] E : 412 | | E bag f 3 |= é = » 9 > 43 12 77 CREAM-COLOURED OR PALE-BROWN, IMMATURE, SEEDS FROM SAME CAPSULES;SOWN NOV.4th. 10 TOTAL PERCENTAGE GERMINATION Q2Q% 5 9 : a Sic. : a 7 ra 3 6 ‘ ee 6 5 & ; | P y 4 & . f Tl 4 ® + i a0 A 4 3 La =] { : i ) 5 ia Dec. tents J Bc terns tt at 4 = TAN. Fig. 2. Germination of fully formed mature and immature seeds from the same capsules sown immediately after harvesting, showing much higher percentage germination of latter. Watsonia 5 (6), 1963. 372 E. J. SALISBURY Having regard to the fact that partially developed embryos were found in some of the seeds regarded as abortive and which a green coloration suggested were present in an appreciable proportion, 1,800 of the better-developed of these were sown (at the same time as the ‘good’ seeds and under like conditions), to check the validity of one’s assessment of their abortive character. Of these seeds which, from their size and shape, were deemed to be abortive, two germinated after 68 days and ten more between the 83rd and 116th day after sowing. The error, in judging the abortive seeds visually, would then appear to be almost negligible and, since those tested consisted of the most promising only, the error of assessment was certainly much fess than 6 in 1000. It is, moreover, not unlikely that this small fraction proved viable only because they were sown immediately and should be considered in the 704 SEEDS GOWN 2331 704 SEEDS SOW 3. : BW s SS, ) sete, DAYS OF CEARMINATIONS 19624 Averace pon isa ravits 13°7 Move 8-12 srevs 22) = = ce i. NY m = 3 = Lad = S Os oF Loe = = be vo oc tat a. NUMBER ¢ OF'GOOD" SEEDS PER FRUIT Fig. 3. Upper diagram, shows daily germination of 704 freshly-ripened and apparently good seeds over period of 85 days. Inset, drawing of stolon half natural size. | Lower diagram shows number and frequency of ‘ good’ seeds in 132 fruits. light of the enhanced percentage germination of immature seeds. However, the noteworthy feature that these germination tests collectively reveal is that there is obviously no abrupt transition from the viable to the non-viable but the distinction between them is one of degree rather than of kind. Most of the fruits examined were from Sussex plants growing at approximately 49° 48’ N. Fruit collected near Wendover (c. 51° 45’ N) contained some apparently fertile seeds and it would be of interest to know how far northwards the ability Watsonia 5 (6), 1963. SELF-INCOMPATIBILITY OF HYPERICUM CALYCINUM IN ENGLAND | 373 to produce fertile seeds can extend’. It should be noted that there is no obvious negative correlation between seed production and the formation of stolons (cf. Fig. 3), and that the emphasis here on vegetative spread, rather than seed development, is not, as in so many species, a concomitant of polyploidy since the Rose-of-Sharon is a diploid species with twenty chromosomes (Darlington & Wylie 1955 p.114). From various localities 132 fruits were collected in October 1962, before they were quite mature, but beginning to change colour. As the seeds at this stage are not yet fully pigmented, one is able not merely to distinguish between the possibly viable seeds and the apparently abortive ones but also in some degree to assess, in many of the latter, the amount of development before abortion. Even the collapsed and withered ovules were still dis- tinguishable, somewhat resembling microscopic autumn leaves. The well-developed symmetrical seeds, that germination tests indicate as including almost all the only viable ones, ranged in number from one to thirty-seven in a capsule and the frequencies of occurrence are represented in the accompanying graph (Fig. 4). The arithmetic mean was 13-8, compared with 6-2 in 1961, but it will be seen that the mode is between 8 and 12 (the average for the capsules represented in Fig. 4 was 18 but these were not a random sample, having been selected for range of size). OF i | SHARON a [ca] bs CORRELATION 5g OG ees | pene eer ee iw 4 ' &2 FRUITS fa pesuden ae, Fertile. and “re Tetal eee ges i ¢ eee : ome 2 "fertile! on a Q &©& = 40 a = BS AVERAGE = 350-600 ee: *7| @ cot i puiybaind pais | ' : | now moeth as d-o Wd '0 060 mao entry Oo Se ee sso- = 700s ! AiR DOOR ee yas oh OU TOTAL NUMBERS[OF'FERTILS’ AND ABORTIVE SEEDS AND OVULES] IN CAPSULES OF H.CALYCINUM Fig. 4. Each spot represents one capsule; its position in the vertical direction corresponds to the number of apparently fertile seeds which it contained and its horizontal position to the total content of ovules whether fertilised or abortive. O—Capsules from ‘Highdown’ It may, I think, be assumed that the usual percentage of those that would germinate is unlikely to exceed the average for the 1187 actually tested, namely 48%. If so, and if the average total of ovules in the fruit be taken as about 800, this would imply an effective development of not more than about eight per thousand of the ovules present. This may represent the normal incidence of compatibility and effective fertilisation in this species when self-pollinated. One can thus envisage that, in a normally self-incompatible species, the triple fusion which produces the endosperm and involves the second male nucleus may similarly exhibit partial or complete failure with correspondingly variable effect upon the development of the embryo. Precocious planting of seeds with ineffective endosperms, might however provide substitutional nutrition and so account for the enhanced percentage germination of the immature seeds of a capsule. 1From information provided by Professor Pichi Sermolli, it would seem that in Northen Italy, at Genoa, seed production of this species is as sparse and unreliable as with us. Watsonia 5 (6), 1963. 374 E. J. SALISBURY Another collection of 704 mature seeds was obtained, about three weeks later in 1962, from several widely separated colonies, growing in various situations and soils. These also were sown at once, under the same conditions as those just considered. This com- posite sample yielded only 48-4% seedlings, a reduction perhaps associated with the seeds having ripened in the less favourable temperatures of mid-November. These 704 seeds, having been culled from capsules produced in distinctive environments, the intermittence, as might have been anticipated, is obscured but is nevertheless manifest as repeated surges in the germinations (cf. Fig. 3). Elsewhere evidence has been furnished (Salisbury 1962 p. 394 et seq.) to indicate that the potentiality for intermittence is genetically determined but that its manifestation is probably dependent upon developmental conditions. If this be accepted we should not expect the intermittent germination of seeds from distinctive individuals and habitats to exhibit contemporaneity. Since the 1961 seeds were sown in May and all the 1962 seeds late in autumn, it would appear that the difference in photoperiodic environment (short nights for the one: long nights for the others) had little, if any, effect. No replication was possible since all available seeds have been utilised for the experiments here described. The lower percentage germination of the 300 seeds of 1961 (a decrease of 15%, as compared with that of the 704 seeds of 1962) is to be attributed not merely to seasonal differences but to the the former having been kept for several months whereas the latter were sown immediately. The decrease of germination with increasing age is a familiar feature accompanying the passage of years, or even months, but here, as we have seen, this decline appears to go back to the period within the capsule and a gradual diminution in the proportion of viable seeds would seem to be indicated from the time of their formation onwards. Although the number of ovules to be fertilised in an ovary is so large and, as mentioned later, may be over 1400, it seems unlikely that inadequate pollination plays any part in bring- ing about the poor seed production, since there is a multitude of stamens (350 or more in a flower) each of which perhaps produces some 900, or more, pollen-grains; a ratio of pollen-grains to ovules of the order of 300 tol! Moreover the greater the number of ovules present in an ovary the more would appear to be fertilised (cf. Fig. 4), a relationship that could scarcely obtain were quantity, and not quality, the limiting condition. Investigators, using a diversity of types, have provided evidence that the incompatibility between individuals of a species depends on a series of alternative genetic characteristics. These determine the biochemical relations between the germinating pollen-grain and the tissues of the pollinated plant, a subtle disparity between them being necessary for effective growth and fertilisation. External conditions, as Professor D. Lewis demonstrated with respect to temperature (Lewis 1942), can modify their interaction. It is scarcely surprising therefore that some pollen tubes do, in fact, reach the egg cell and effect fertilisation. Even for the markedly self-incompatible sweet cherries, Crane & Lawrence (1938) obtained about 0-1 % fertility when selfed. Thus, the occurrence of a small proportion of viable seeds in self-pollinated H. calycinum is not unexpected. Why, in the race against time to the ovules, some pollen-tubes achieve success may be an outcome of the shuffling of the genes that takes place during the formation of the sexual cells. Theoretical considerations lead us to recognise that the nuclei of both pollen-grains and egg cells, though derived from the same parent, may yet have individual differences that could influence their effective union. So we may postulate that successful self-fertilisation could depend upon the chance mating of such male and female nuclei as exhibit compatible constitutions. If this hypothesis be correct then with an increased number of matings the number of successes should augment although the proportion of these might be expected to remain roughly constant. Both these conditions are, in fact, apparently fulfilled. Owing to the fortunate circumstances that the successes and failures in the fruit can alike be determined with a fair degree of accuracy, it is possible to check these features although such determinations, owing to their laborious and time-consuming character, could only be made upon a limited number of fruits, actually fifty-one, selected for the Watsonia 5 (6), 1963. SELF-INCOMPATIBILITY OF HYPERICUM CALYCINUM IN ENGLAND | 375 range of size they presented. The data for fifty of these are displayed in Fig. 4. By grouping these capsules, according to increments of about 200 in their ovular content, the rise in the average numbers of ‘ fertile seeds ’, accompanying the total increase, is made evident. The examination included some large capsules from a colony, established for more than half-a-century, growing on an exceptionally sheltered, south-aspect slope in Sir Frederick Stern’s garden at ‘ Highdown’. One of these (not included in Fig. 4) was quite outstanding, with no less than 58 apparently fertile seeds out of a total content of 1458; a proportion still barely achieving 4%, of which not more than a third would probably germinate. If all the fifty-one fruits in which the contents were determined be grouped, we find that the proportion of apparently fertile seeds is strikingly similar, in all but the lowest category of contents, namely 370 to 600, 1:-4%; 600 to 800, 2:6%; 800 to 1,000, 2:2% and 1000, to 1,458, 2:3%. The facts then would appear to correspond with our hypothetical presentation, the more so that the discrepancy in the class-interval representing the smallest capsules could be due to nutritional stresses inoperative for the larger fruits. A very noteworthy feature of the incompatibility in this species is that various stages of abortion appear to have occurred subsequent to nuclear fusion. The green colour of many of the abortive seeds is indicative of the presence of embryos arrested in their develop- ment at different stages. The pollen tubes of incompatibles may often grow in the style to different degrees and any which reach the egg-cell may effect fertilisation. The genetic barriers thus act by affecting the temporal aspects of growth and perhaps this extends further, beyond the fertilisation of the egg, to affect also mating of the second male nucleus with the fused polar nuclei to provide the endospermic nourishment for the develop- ing embryo. If in H. calycinum this second fusion is similarly retarded it might explain the arrest of the embryo at various stages. The much higher % germination of immature seeds is in harmony with such an interpretation. It would not appear likely that nutrition materially influences the number of ‘ good’ seeds produced since, although the shoots of the Rose-of-Sharon are normally single- flowered, I have found paired fruits not infrequently and very rarely a shoot bearing three fruits. On three-fruited and two-fruited shoots alike, the capsules did not exhibit any diminution below the average in numbers of ‘ good’ seeds. Although flowering may begin in July, I have rarely found ripe fruit before the end of October and frequently not till late November. This retarded maturation may be the result of the different climatic conditions that prevail here, compared with those of S.E. Europe. Dehiscence takes place from the apex towards the base into five segments (Fig. 1, D & G) surrounding a central, conical column formed by the placentas. Frequently, however, the capsules do not appear to split open sufficiently widely for the larger seeds to escape readily. Though sometimes erect, the fruits are usually inclined so that, when the segments spread at the apex, the contained seeds should be shaken out by the gusts of wind. Two circumstances appear to militate against the successful operation of this ‘ censer’ device. Firstly, owing to the high proportion of seeds which abort, the capsule is largely empty and as it dries and contracts may do so without creating the tensions that, with a well-filled capsule, would more readily ensure the effective rupture and gaping apart of the segments towards the top of the fruit. To test this aspect, ten capsules were selected which had pronounced dehiscence slits and these were each in turn held by the stalk and violently shaken and jarred, to simulate wind movement in an exaggerated form. The result was that 54% of the total of fully-developed seeds escaped. However the 46% that were still retained included, not unnaturally, the larger ones, and if only 22% of the total were actually viable might well account for all of these. The second circumstance is that since the fruits ripen so late in the year the chance of a thorough sun-baking in these latitudes is meagre. Capsules, which had dehisced but in which the slits were too narrow to permit the escape of any but the abortive seeds, were placed in a warm room to simulate sun-drying. Only after several days was it that the segments separated sufficiently to permit the ready escape of the fully-developed seeds. This would seem to indicate that lack of adequate insolation probably impairs seed-dispersal Watsonia 5 (6), 1963. 376 E. J. SALISBURY under the climatic conditions that normally prevail here in late autumn. Obviously, with decay of the capsules, all seeds would eventually be released but, during the intervening period, they would have been subjected to great fluctuations of temperature and probably also to recurrent changes from dryness to saturation, conditions known to be the reverse of those conducive to the retention of viability. Towards the end of February a random collection of 62 fruits was obtained from an exposed habitat and their contents examined. All the capsules were dehisced and had been subjected to severe winds that had attained gale force cn several occasions. Nevertheless although just under one third contained no seeds, 43 of the capsules contained apparently good seeds the number of which ranged from 1 to 32, the average being 4:2 (s.d. = 6), so that even under these extreme conditions, so favourable to dispersal, an average of only about eight ‘good’ seeds would be liberated from a capsule and this would imply an average seedling potential for the escaping seeds of not more than four. In an actual test of 210 of these retained ‘good’ seeds germination attained 37-6°%% and was completed in 13 days, which suggests the leaching out or inactivation of an inhibitor. Ineffective dispersal would then seem to be an important contributing factor, but not a complete explanation, for the apparent absence of seedlings. From the foregoing considerations it is not unlikely that an appreciable proportion of the potentially viable seeds may not reach the ground until after their germinative capacity has become impaired. Further, the scanty seeds and even scantier seedlings might perhaps be largely, or even entirely, accounted for by predators. It will be obvious from the data presented in Figs. 1 and 3 that germination takes place in a discontinuous manner which implies that the number of seedlings at any one time could, at best, only be small and so might readily be overlooked. However, I have in fact searched the fringes of colonies, in various localities and on various soils, for seedlings, so far without success. Furthermore, there does not seem to be that degree of discontinuity in occurrence, near the borders of colonies, that might be expected if the species had spread from seedlings. Although we now know, with certainty, that some viable seeds are produced by the species in southern England, it still remains to be ascertained whether, in the open, seedlings do actually develop and, if so, when. Are conditions normally adverse to germination or, if not, what are the circumstances responsible for the non-survival of seedlings? Only further observations can resolve these questions. REFERENCES Crane, M. B. & LAwreNnce, W. J. C. (1938). The Genetics of Garden Plants. Ed. 2. London. DARLINGTON, C. D. & Wy ig, A. P. (1955). Chromosome Atias of Flowering Plants. London. HaAnsury, F. J. & MARSHALL, E. S. (1899). Flora of Kent. London. Lewis, D. (1943). The physiology of incompatibility in plants. I. The effect of temperature. Proc. Roy. Soc. 131, 13-26. PERRING, F. H. & WALTERS, S. M. (1962). Atlas of the British Flora. London. SALISBURY, E. J. (1942). The Reproductive Capacity of Plants. London. SALISBURY, E. J. (1961). The fruiting of the periwinkles (Vinca major and V. minor). J. R. Hort. Soc. 86, 489-492. SALISBURY, E. J. (1962). Masters Memorial Lectures, 1962. The biology of garden weeds. PartI. J. R, Hort. Soc. 87, 338-350, 390-404. WuiteE, J. (1912). The Bristol Flora. Bristol, Watsonia 5 (6), 1963, — THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS By J. G. Dony The success of the Distribution Maps Scheme of the Botanical Society of the British Isles has caused a number of workers on local floras to adopt divisions of 10-kilometre grid squares of the National Grid as a basis for their recording. In 1954 the Birmingham Natural History and Philosophical Society started to use a one-kilometre grid for collecting records for a revision of the flora of Warwickshire. This was, however, slightly amended at a later stage. Two years later, in 1956, E. S. Edees began to use a two-kilometre grid (i.e. four square kilometres) for a survey of the Staffordshire flora. In 1955 the author started work on a revision of the flora of Hertfordshire and in 1957 decided to collect records as Edees had done on the basis of a two-kilometre square, referred to below as a tetrad. Few records were made on this basis until 1959 and the results summarised here are mainly those of four years’ field work. Fig. 1. Hertfordshire. The boundary of Watsonian vice-county 20, Hertford, is shown with a dotted line where it differs from that of the present (1963) administrative county. The prefixing numbers of the 10-kilometre grid squares are omitted but they are given in Table 1. The area of survey (see Fig. 1) is the present administrative county of Hertfordshire with those parts of the Watsonian vice-county 20, Hertford, which now lie in neighbouring administrative counties. This has a total area of about 1,631 square kilometres and is very irregular in shape. It includes only four complete 10-kilometre grid squares and parts 8/7 Watsonia 5 (6), 1963. 378 JG, DONY of twenty-five others, making a total of 336 complete tetrads and parts of 131. The object of this paper is to try to determine the average number of plant species that could be expected to be found in a tetrad or in any other given area in Hertfordshire. After eight years’ intensive field work and the contribution of records from fellow workers the total number of species to be found in the county is reasonably well known and it is unlikely that in the next few years many more will be found other than aliens and casuals. Almost complete records have also been obtained for the full 10-kilometre grid squares and for those the greater part of which is included in the county. From these records it is possible to determine the number of species now known to occur in parts of the county larger than a 10-kilometre grid square. The species considered for the purpose of this exercise are those given in the List of British Vascular Plants (Dandy 1958). Aliens, casuals and garden escapes are not included unless they are in the List, and are neither microspecies nor hybrids. Records made before TABLE |. Summary of records of species made in Hertfordshire (1955-1962). 10-kilometre | Approximate Total species Total | Total Species grid square | area in Herts. recorded | area | recorded | (sq. km.) | | (sq. km.) | 52/24 | 10 225 19 (A) 282 52/34 | 9 | 230 53/13 | 12 | 421 | 52/23 85 544 220 (B) 651 52/33 | 98 | 521 52/43 | 25 415 52/02 | 6 303 52/12 87 | 499 52/22 | 100 | 604 364 (C) 712 52/382 | 100 588 | 52/42 68 | 579 52/52 | 3 | 117 42/81 | 12 265 | | 42/91 | 37 560 52/01 | 84 | 488 52/11 | 87 | 528 493 (D) 834 52/21 100 | 640 52/31 99 | 688 52/41 | 73 | 579 42/90 | 25 451 52/00 96 597 | 52/10 100 594 | 377 (BE) 758 52/20 84 561 | 52/30 | ip, | 591 51/09 | 66 | 647 51/19 62 | 554 158 (F) 724 51/29 30 | 461 A+B+C (52/2434, 52/03-13, 52/02-42) 603 (G) 798 E+F (42/90-30, 51/09-29) 635 (H) 836 D (see above) 493 (D) 834 GD AACS | 899 G+H 1138 (J) 908 D+H | 1038 (K) 911 Total Ls ei8634 | 956 | The totals of species given above are not totals of the columns from which they are derived, as a species recorded from more than one ten-kilometre grid square is counted only once and so on for the larger areas. A very common species included in the grand total may well have been recorded in all the twenty-seven ten-kilometre grid squares from which the totals have been made. Watsonia 5 (6), 1963. THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS _ 379 1955, when the survey began, are also excluded. The results of the work done until the end of 1963 are summarised in Table 1. These figures enable one to show by means of a simple graph (Fig. 2) the average number of species that one might expect to find in any given part of the county. This emphasises the fact, seen less readily from the figures themselves, that some parts of the county are floristically richer than others. If the whole county were as productive as these the total number of species found in the county would probably be greater. On the other hand, if all the county were as unrewarding as other portions are, the total number of species would be fewer. The gradient of the line drawn to show the probable average of number of species in a given area is, however, too great in the region of the smaller areas to be of any use in determining the average number of species one would expect to find in an area as small as a tetrad. 900 800 700 600 500 Species 400 300 200 100 200 400 600 800 1000 1200 1400 1600 Area (sq. km.) Fig. 2. Distribution of records of species by area in Hertfordshire. A different result is obtained if a logarithmic scale is used for both the area and the number of species. The curve now approximates to a straight line. Indeed it can in the areas being considered here be assumed to be a straight line. This is shown in Fig. 3 for records from a considerable area of the county, i.e. from 60 square kilometres to the whole of the area. The line drawn shows the expectation of records based on the assump- tion that as the survey proceeds very few additional species will be found in the whole county or in the larger parts (I, J and K) which are each approximately two-thirds of the county. It is certain, however, that more species already known to appear elsewhere in the county are awaiting record in the smaller areas such as a 10-kilometre square. The present average of species recorded from the complete or almost complete 10-kilometre grid squares (52/33, 52/22, 52/32, 52/21, 52/31, 52/00, 52/10) is 608 and the linear relationship suggests that it should be 623. If additions are made at the rate they have been hitherto in the survey it will probably reach this figure by the time the task is completed. If this method is to be useful for the purpose in hand it is necessary to consider records from both larger and smaller areas. The species considered are those in Dandy’s list, that is 2,137 for the whole of the British Isles (area 310,600 square kilometres) and this Waisonia 5 (6), 1963. Log no. of species 380 Ja 1G, DONY 900 600 750 1000 1250 1500 Log area (sq. km.) Fig. 3. The expectation of records of species from the larger areas in Hertfordshire. The records shown in Fig. 2 are now shown on logarithmic scales for both area and the number of species recorded for the res- pective areas. The line drawn to show the average expectation of records is based on the assumption that few additional records will be made for the whole of the county or for the areas shown as I, J and K, each representing about two-thirds of the county. Additional records may, however, be expected from the smaller areas. gives amaximum reference point. During 1962 close studies were made of 106 small habitats having floras representative of that of the county as a whole. Each study was made in a measured area of five yards radius (65-7 square metres) which was visited twice in the year and the frequencies of the various species noted. The number of species recorded for the studies varied from 82 to 13 with a median of 45. Samples taken from small areas will show a great range in the number of species present as they will include both rich and comparatively barren plant communities. The full results are given in Table 2. Also in TABLE 2. 82, 82,79, 78, 77, 77, 16, 716, 72, 65, 64, 63, 62) 62, 61, 60, 59, 59, 57, 57. 57, 56,05) son Seomes 52, 52, 52, 51, 50, 50, 50, 49, 49, 49, 49, 49, 48, 48, 47, 47, 47, 46, 46, 46, 46, 46, 46, 46, 45, 45, 45, 44,44, 44, 43, 43, 42, 41, 40, 40, 39, 39, 38, 38, 38, 38, 37, 37, 36, 36, 36, 36, 36, 36, 34, 34, 33, 33, 32, 32, 32,32, 32, 31,,31,30; 26,26; 26, 24) 24 235 23722) 21520) 18, lie alan salon 1962 a town garden with an area, including the house, of 297 square metres, yielded 84 species of wild plants and it is probable that urban sites generally have floras comparable with this. These figures allow an extension of the line of average expectation (Fig. 4). The change in gradient at 100 square kilometres arises from the fact that Hertfordshire being a lowland county is more floristically rich and varied than the British Isles as a whole. The results show that the average expectation of species in a tetrad is 335 and in a square kilometre 275. It also allows an estimate to be made, with less accuracy, of the maximum and minimum expectation. For a 10-kilometre grid square these are 708 and 500 and the present records for these areas in Hertfordshire range from 688 to 521. The maxima and minima to be expected from a tetrad are 446 and 224 and from a square kilometre 402 and 155. Watsonia 5 (6), 1963 THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS _ 381 log. species SS <—— Hopkins 1955 ten kilometre grid square Herttordshire | square Area of 106 kilometre | | : ! | Eaat enema ee - rn | 10 100 1,000 | 10 100 10GO 10900 100,000 I 10 100 1CO0- 1Q000 100,000 sq.cm. sq metres sq. km log. area Fig. 4. The average expectation of records of species in Hertfordshire. The dotted line at the lower end of the line of average expectation shows the number of species estimated by Hopkins to occur in various areas in one plant community which he studied. This was not in Hertfordshire and was the community, among the twelve he examined, with the largest number of species. The line drawn to show the average expectation of species is based on two assumptions: the first that the median of the number of species recorded for the 106 studies made of small areas is the average expectation of the number of species that would be found in Hertfordshire in areas of the same size as the areas of the studies: the second that the total number of species to be found in the whole of Hertfordshire is known as this consists of a large number of kinds of habitats each type of which has been studied. The lines of probable maximum and minimum expectation are based on the assumption that the greatest and smallest numbers of species recorded for the 106 studies represent the range of floristic diversity of areas throughout Hertfordshire comparable - with the areas of these studies. It is further assumed that this range will diminish as the area is increased to a point somewhat larger than the county. Records between the average and maximum expectation for tetrads have been obtained as the result of long and thorough search in the field in widely separated parts of the county. The method I have used for enumerating the tetrads is shown in Fig. 5. 31.A. 395* species were recorded by Mr A. G. Brown and Mr H. Williams of the John Innes Horti- cultural Institution who spent many junch hours accounting for all the plants that could be found on the Bayfcrdbury Estate. : 31.R. 376 species were recorded largely as the result of a concentrated effort by a number of members of the Wild Flower Society on a one-day field meeting. 32.V. 342 species were found by Dr and Mrs Lloyd-Evans during a number of visits in one season. Large totals have also been made in some marginal tetrads: *The number of species given for tetrads from this point in the paper includes some casuals, garden escapes, etc. not included in Dandy’s list as I have no easy means of excluding them. They are few and do not affect materially the main conclusions reached Watsonia 5 (6), 1963. 382 J. G. DONY 23.B. (area 3-5 square kilometres, expected average 320). 312 species recorded by Mr Harry and Miss Doris Meyer in a semi-urban area they know intimately. 09.R. (area 3-75 square kilometres, expected average 322). 309 species recorded by Mr G. Day who has listed all the plants he could find in Moor Park. 09.J. (area 2-7 square kilometres, expected average 317). 320 species recorded. Fig. 5. The enumeration of tetrads within a 10-kilometre square. Each 10-kilometre grid square (e.g. 52/31) is divided into 25 tetrads which are lettered from A to Z (omitting O to avoid confusion with Q) following the sequence used in the numbering of the 10-kilometre squares themselves in the National Grid. The prefixing numbers, i.e. 42, 51, 52, are omitted as there is no duplication of the second numbers in an area as small as a county. The tetrads thus follow a sequence of 31.A., 31.B., 31.C., etc., beginning at the bottom left- hand corner of the 10-kilometre square. The average expectation of the number of species to be found in a tetrad may be used as a basis for determining the thoroughness with which an area has been studied. The total number of records made for the 336 full tetrads in Hertfordshire, by the end of 1962, was 72,594, an average of 216 per tetrad. As the expected average is 335 this gives a 61% cover. It would not be possible except by wearisome calculations to determine the cover for marginal tetrads but it is probably at least as high. A summary of the records is given in Table 3. The peculiar distribution of records in Table 3 arises from targets of 150 and 200 records having been set during the field work. When 200 species had been listed from a tetrad and the rate of recording had slowed down there was a great temptation to move on TABLE 3. Summary of records of species made in individual tetrads in Hertfordshire (1955-1962). Full Tetrads 395, 376, 342, 307, 297, 295, 294, 293, 292, 290, 289, 286, 281, 280, 278, 276, 274(2), 271, 269(2), 268, 265, 264, 262, 256(2), 255(2), 254, 253, 252, 248(2), 247, 245, 242(4), 241(2), 240(2), 239(2), 238, 237(2), 236(2), 235(2), 234, 233, 232(3), 231, 230(3), 229(4), 228(4), 227(3), 226(3), 225, 224(4), 223(2), 222(6), 221(8), 220(3), 219(4), 218(4), 217(6), 216(3), 215(3), 214(7), 213(4), 212(15), 2119), 210(13), 209(8), 208(2), 207(13), 206(14), 205(21), 204(15), 203(18), 202(16), 201(29), 193, 185, 183(2), 176, 175, 172, 170, 167, 165, 162, 160(3), 159, 158, 157, 155(@2), 154, 153(@2), 152(2), 151(2). Total: 72,594. Marginal Tetrads 320, 312, 309, 289, 283, 280, 265, 254, 252(2), 237, 235, 234, 233); 230; 229) 226),223,220n7i9) 218, 217(4), 216(2), 215, 214(3), 213(4), 212, 211, 209(2), 207(6), 206(4), 205(6), 204(4), 203(10), 202(4), 201(6), 196, 187, 185(2), 172(2), 175, 172, 171, 168, 167, 166, 164, 157(2), 154(2), 153(2), 152, 143, 133; 128, 127, 123, 119(2), 117, 115, 114, 105, 102, 97, 91, 80; 67, 55, 54, 33: 27, 5. MotaleZ2eoiie and begin work in virgin country. As 200 species is 899% of the probable minimum expectation it has proved to be an exceedingly well-chosen target. In some cases it has been very difficult to find this number. It has, however, resulted in the more intensive Watsonia 5 (6), 1963. THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS — 383 search of the less rewarding country to the neglect of the better areas. After four years of field work and with records made for complete or almost complete tetrads ranging from 143 to 395 there is no species recorded for all. Thirteen are awaiting record from fewer than five tetrads and another twenty-four from fewer than twenty. The records made so far indicate that between seventy and one hundred species may be in all tetrads. It is obvious that a great deal of work could still be done. It must be left to each worker using tetrads as a basis for recording to decide for himself when a satisfactory cover has been made. I have been able to make some very useful maps with a cover of 61° and to pursue the study much further would occupy a long period of time. The concept of an average expectation has its limitations. Ifa cricketer has a batting average of sixty it does not follow that he scored this number of runs in his last innings or will score it in the next, neither does it mean that if one’s expectation of life is established as eleven years one will indeed live that time. If the average expectation of species from a tetrad is 335 it does not help the worker in the field to know what to expect from any particular tetrad. I recognise in Hertfordshire twenty different types of habitats and the various tetrads may contain one or more of these. A tetrad with only one or two different habitat types would with effort probably yield only 225 species but one with as many as ten habitat types could well give 400.* The relationship between species and area in very small areas can be solved only by the very close study of plant communities. This i have not done but it has attracted much attention from ecologists (Arrhenius 1921, 1922, Greig-Smith 1957, Gleason 1922, 1925, Hopkins 1955, 1957, Williams 1943, 1944, 1947, etc.). One of the most thorough studies of this kind was the investigation of twelve very different communities by Hopkins (1955). The number of species he recorded was surprisingly small especially as he included bryo- phytes and lichens as well as vascular plants. On the other hand his studies were limited to one short period and there were ‘additional species’ in most of his communities which 1 square mile British Isles x/2 Log no. of species. Ge na ee 1 10 100 1,000 10,000 100,000 Log area (sq. miles) Fig. 6. Watson’s Hypothesis. It is obvious from Cybele Britannica that Surrey was the county he had in mind. *This was an experience well known to workers for the Distribution Maps Scheme. One lady returned exhausted after a mapping meeting with the information, amusing to the rest, that she had had a ‘big square’. She had only meant that her square contained a large number of plant communities that needed to be visited before the day’s work could be said to be done. Watsonia 5 (6), 1963. 384 J. G. DONY did not enter into his samples. I have chosen one of Hopkins’ communities with species closely comparable with the lower ranges of my line of expectation and using a log area— log species scale it fits my conclusions perfectly. (Fig. 4.) Much has been written on the species-area relationship but H. C. Watson was as near the truth as anyone has been or is likely to be when he observed as early as 1835 that ‘on the average a single county appears to contain nearly one half the total number of species in Britain; and it would, perhaps, not be a very erroneous guess to say that a single mile contains half the species of a county’. Watson is making in simple language a relationship which, by the use of a logarithmic scale, may be made a linear one (Fig. 6). The matter appears to have rested there until ecologists became interested in the build-up of species in plant communities. It was claimed by Arrhenius (1921) that ‘the number of species increases continuously as the area increases’ and that this increase could be expressed by the simple formula y ( X i when y = the number of species and x = the area con- Sie NES taining these species. This gives a linear relationship when log species and log area scales are used as has been demonstrated in this paper.* The relationship demonstrated by Arrhenius was challenged by Gleason (1922) who suggested that a linear relationship could only be arived at by plotting the number of Species against the logarithm of the area. Ecologists have tended ever since to accept this and much work has been done to establish the exact nature of the species—log area curve and the limits of its linear distance. It must be stressed that the problem of the ecologists is different from mine and their attentions have been limited to uniform plant Annobon Principé Sao Tomé Fernando Po Log no. of species. 100 10 100 1,000 Log area (sq. km.) Fig. 7. A comparison of Island Floras, based on the list of ‘significant species’ given in Exell’s Flora of Sao Tome. *The above formula can be expressed more simply as y = ax" where y is the number of species, a is a constant and x is the area. It would appear from the method demonstrated here that the average number of species included in Dandy’s list that could be found in any given area of the British Isles can be expressed as Vii Si. x0°1738 given x as the area in square kilometres or yY = 21.8 x0:1738 given x as the area in square metres. 3 : This would give an average expectation of species for a ten-kilometre grid square of 527 which would agree with most botanists’ experience in the field. : ; R ; ; The corresponding relationship for Hertfordshire and probably most lowland English counties could be expressed as y = 275 x0-1772 given x as the area in square kilometres or y = 21.8 x0-1772 given x as the area in square metres. Watsonia 5 (6), 1963. THE EXPECTATION OF PLANT RECORDS FROM PRESCRIBED AREAS — 385 communities and they have always assumed that the situation changes when the sample areas extend into other communities. The method outlined here needs to be tested further by making comparisons of the number of species known to occur in the various parts of the British Isles, e.g. England and Wales, Scotland, Ireland and the Isle of Man and also for the separate islands of the Channel Islands group. I give (Fig. 7) its application to some closely grouped islands off the African coast taken from Exell’s Flora of Sio Tomé. If the method were confirmed by further data it would be possible to calculate the degree of cover obtained for the Distribution Maps Scheme. I wish to thank Dr C. B. Williams for suggesting to me the use of the logarithmic scale and helping me to apply it to the Hertfordshire problem; Dr M. C. F. Proctor for drawing my attention to the work done by ecologists with a similar problem; Dr B. Hopkins for allowing me to use the data quoted in his paper; Mr E. S. Edees for helping me to confirm my conclusions by providing me with details of his work in Staffordshire; and my many co-workers who have carried out the wearisome task of providing records from such small areas. Amongst the latter I thank especially Mr and Mrs P. C. Hall who made themselves responsible for collecting records from about 20°% of the county in a part which it would have been difficult for me to visit frequently. REFERENCES ARRHENIUS, O. (1921). Species and area. J. Ecol., 9, 95-99. ARRHENIUS, O. (1922). A new method for the analysis of plant communities. J. Ecol., 10, 185-199. Danpy, J. E. (1958). List of British vascular plants. London. GREIG-SMITH, P. (1957). Quantitative Plant Ecology. London. GLEASON, H. A. (1922). On the relation between species and area. Ecology, 3, 158-162. GLEASON, H. A. (1925). Species and area. Ecology, 6, 66-74. Hopkins, B. (1955). The species-area relations of plant communities. J. Ecol., 43, 409-426. HopkKIns, B. (1957). The concept of minimal area. J. Ecol., 45, 441-449. PERRING, F. H. & WALTERS, S. M. (1962). Atlas of the British Flora. London. EXELL, A. W. & WILD, H. (1961). A statistical analysis of a sample of the Flora Zambesiaca. Kirkia, 2, 108-130. EXELL, A. W. (1944). Catalogue of the Vascular Plants of Sao Tomé. London. Watson, H. C. (1835). Remarks on the geographical distribution of British Plants. London. Watson, H. C. (1847). Cybele Britannica. London. WILLIAMS, C. B. (1943). Area and number of species. Nature, 152, 264-5. WILLIAMS, C. B. (1944). Some applications of the logarithmic series and the index of diversity to ecological problems. J. Ecol., 32, 1-44. WILLIAMS, C. B. (1947). The logarithmic series and the comparison of island floras. Proc. Linn. Soc., 158, 104-8. Watsonia 5 (6), 1963. THE TAXONOMY OF POLYGONUM LAPATHIFOLIUM L., P. NODOSUM PERS., AND P. TOMENTOSUM SCHRANK By J. TIMSON Botany Department, University of Leicester.* ABSTRACT An account is given of the variation in Polygonum lapathifolium L., and of its previous taxonomic treatment. The taxonomic value of various characters is examined. Herbarium, field, and cultivation studies are described. It is suggested that P. nodosum Pers, and P. tomentosum Schrank are in fact merely parts of the range of variation found within P. /apathifolium L. and that they should no longer be regarded as distinct species. 1. HISTORICAL INTRODUCTION Polygonum lapathifolium L. is a species which exhibits considerable intraspecific variation and is frequently regarded as ‘plastic’ or ‘critical’. There are two main sources of confusion: (a) From at least the time of Linnaeus many botanists have regarded P. lapathifolium as a pale form of P. persicaria. Chromosome counts have shown that P. lapathifolium has 2n = 22; while P. persicaria has 2n = 44, and there are constant morphological differences between the species. Identification, however, has often still been made on the basis of perianth colour, and herbarium material of the white form of P. persicaria is often incorrectly labelled. P. persicaria may easily be distinguished from P. lapathifolium since the former has eglandular perianths, peduncles and leaves. (b) Two other names, P. nodosum Pers. (1805) and P. tomentosum Schrank (1789) have been described and published as species. Unfortunately, the characters given in the literature for the delimitation of these from P. Japathifolium vary from author to author and sometimes they are regarded as synonyms or subspecies, ete. Britton (1933) states that the only reliable character is the fruits of P. nodosum which are ‘smaller on the whole, less roundish and with a more ovate outline’ than those of P. lapathifolium. n fact the size of the fruits, which is discussed later, is largely dependent on the environment of the parent plant. Britton also states that the glands of the peduncles and perianths are less numerous in P. nodosum. There is certainly some variation in the density of glands in different populations but there is no agreement in the literature with regard to this character. Tutin (1952) uses ‘peduncles sparsely glandular’ against ‘peduncles densely glandular’ to separate P. lapathifolium from P. nodosum respectively. Tutin also states that P. lapathifolium has ‘leaves with sunk pellucid glands beneath’ while P. nodosum has ‘leaves with golden glands beneath’. Moss (1914) and Davey (1909) regard the presence of red spots on the stem as characteristic of P. nodosum, and Persoon in his original descrip- tion gives the stem as spoited. P. tomentosum is a name rarely used in Britain but commonly used by authors on the Continent. Schrank in his original description states that the leaves are tomentose beneath, and it is interesting to note that he does not give a description for P. lapathifolium. Plants which correspond to his description of P. tomentosum are not uncommon in Britain where they are usually regarded as P. lapathifolium or P. nodosum. Schuster (1906) erected an elaborate hierarchy of taxa which he believed to be an approach to a natural system of classification for P. lapathifolium. The first division in his key uses the leaf glands character to divide subsp. verumwhich has pellucid glands, from subsp. punctatum which has yellow glands. Each of these subspecies is further divided *Now at Dept. of Science, Hatfield College of Technology, Hatfield, Herts. This paper is largely taken from a thesis accepted for the degree of Ph.D. in the University of Leicester. 386 Watsonia 5 (6), 1963. THE TAXONOMY OF POLYGONUM LAPATHIFOLIUM L. 387 into varieties and some of the latter are subdivided into several forms. Most of the charac- ters used for the varietal and form divisions are phenotypically variable, e.g. degree of branching, length of internodes, etc. Some of the forms seem rather improbable; in particular f. natans Schroter (Veg. d. Bodensees 2, 1902), which is described as inhabiting lakes as a free-floating form with internodes 15 cm long, and with many adventitious roots on the greatly swollen nodes. Danser (1921) carried out cultivation experiments and found that the variation due to the environment was less than had been previously thought. He lists the following characters as being affected by the environment: plant height and habit, the abundance of branching, the swelling of the nodes and the splitting of the ochreae, the size of the leaves and whether they are tomentose or not. He regards the following characters as constant: the shape of the leaves, the colour of the perianth and other organs, the presence or absence of red spots on the stem, and the shape and the presence or absence of the leaf blotches. On the basis of this work Danser erects a classification in which the shape of the leaves and of the leaf blotches is important. The data presented here agree with Danser’s conclusions with the following exceptions: 1. The tomentose leaves were only found. during the cultivation experiments on plants raised from the fruits of tomentose plants. 2. The leaf blotches vary even on a single plant and there is even more variation within a population which is otherwise uniform. Leaves corresponding to two of Danser’s taxa have been found more than once ona single plant. 2. ‘TAXONOMIC CHARACTERS STUDIED (a) Vegetative characters Habit The habit and size of the plant have been used by various authors to delimit taxa below the species level, e.g. Fernald (1950). Habit varies with the immediate environment of the plant, and fruits from prostrate plants wiil produce erect plants in suitable habitats. There appears to be only one taxon in which habit is important. This is subsp. danubiale Kern. ( = brittingeri Opiz) which von Koch (1935) showed retained a prostrate habit in cultivation. This plant does not appear to occur in Britain. With living material the habit has been scored as E = erect or D = decumbent. Stems The stem is clearly divided into nodes and internodes and the amount of swelling of the nodes has been used as a diagnostic character. Danser (1921) showed that this swelling was largely the result of moist conditions. Anatomical examination showed that the nodes contained many starch grains and appear to function as storage organs. It is not surprising, therefore, that larger plants growing in favourable habitats with a fairly high water content should have swollen nodes. The stems are sometimes glabrous but are more usually very slightly pubescent; they are never very pubescent. Stem pubescence 1s, therefore, not suitable for biometrical treatment and has been scored as P = present or A = absent. The anthocyanin pigment of the stems, which appears as the stem matures, may bein the form of a diffuse, unlocalised, light red colouration, or in the form of discrete, more intensely red spots. The former occurs mainly on the south-facing sides of the stems and is not formed on the ventral side of prostrate stems. It is clearly correlated with the environ- ment and probably with light incidence. The red spots are found on the entire stem surface including the ventral part of prostrate stems although the colour is darker on the parts with the highest light incidence. Both characters were scored as ‘present’ or ‘absent’. Leaves : The leaves of P. lapathifolium are lanceolate to linear-lanceolate and vary in size with Watsonia 5 (6), 1963. 388 J. TIMSON the environment. The length of the leaves was measured in centimetres from the apex to the point of fusion with the petiole. The maximum breadth was also measured in centimetres. An index was then obtained by dividing the length by the breadth. The presence or absence of a leaf blotch, and of hairs on both surfaces was recorded. Where the tomentose condition was found this was recorded as T. Yellow glands on the lower surface were recorded as P if present; the A, recording absence of yellow glands, indicates that the glands were in fact pellucid. Ochreae and ochreolae The fused membranous stipules known as ochreae which are characteristic of the family Polygonaceae are of little taxonomic value in section Persicaria in which P. lapathi- folium is placed. The cilia of the ochreae are often said to be longer in P. persicaria than in P. lapathifolium, and some authors regard the latter as eciliate. This is rarely if ever quite true as it is usually possible to find at least a few short cilia. The difference in length is an indifferent character for separating the species. The ochreolae or floral ochreae are found in the inflorescence and like the ochreae may be ciliate or eciliate, glabrous or pubescent, and glandular or eglandular. Peduncles The peduncles of P. lapathifolium are glandular with stalked glands. Those of P. persicaria are usually regarded as eglandular. Schotsman (1950), however, records the presence of ‘sometimes a few stalked glands’ on the peduncle of P. persicaria. It is not impossible that this plant was an anthocyanin-containing specimen of P. lapathifolium. The peduncles are also occasionally pubescent. (6) Flower and fruit characters Flowers The flowers of P. lapathifolium ate hermaphrodite, and the perianth segments, which are free, are not divided into sepals and petals. They are not large, seldom reaching 4-0 mm in length and they remain attached to the fruit when it is shed. The perianth segments of P. lapathifolium are glandular and those of P. persicaria are usually regarded as being eglan- dular. References to microscopic perianth glands on P. persicaria have been found on herbarium sheets in the British Museum and Schotsman (1950) records ‘pluricellular yellowish glands on some flowers’ of this species. I have not seen these; they are in any case of a different order of magnitude from the perianth glands in P. lapathifolium which can be clearly seen with a hand lens. There are usually 4 perianth segments (occasionally 5) in P. lapathifolium and 6 (5) stamens with 2 (3) styles. The styles of P. /apathifolium, unlike those of P. persicaria, are separate to below the middle. The perianth glands were scored as present or absent, and the number of perianth segments was recorded on fresh material. The colour of the perianth was scored as Pk = pink (i.e. with anthocyanin) or W = white (i.e. without anthocyanin). The number of stamens and styles was also recorded. Fruit The ‘seeds’ of P. lapathifolium are in fact nuts which fall from the parent plant with the dead perianth still attached. The length of the perianth relative to the fruits has been used by some authors with but little success. Moss (1914) records all species in section Persicaria as having nuts ‘as long as the persistent perianth’ except for P. nodosum which had nuts ‘scarcely as long as the perianth’. Styles (1962) found this relationship of use in section Polygonum but it appears to be of little value in section Persicaria. The length and maximum breadth of the nuts (without perianth) were measured in millimetres to the nearest 0-1 mm, each set of measurements being on a sample of mature nuts usually fifty or more in number. An index was obtained by dividing the length by the breadth. The arithmetic mean, standard deviation and the range of these measurements were calculated. Watsonia 5 (6), 1963. THE TAXONOMY OF POLYGONUM LAPATHIFOLIUM L. 389 TABLE 1 Data from herbarium material. yi pe 5) 4 ) 6 7 8 9 10 Leaf length 13-3 12:3 13-4 10-0 13-7 6°8 6:5 14-3 10-2 7:3 Leaf breadth 3-2 3-8 3-9 25 4-1 1-3 1-3 4°] 2°3 2°4 Leaf index 4:2 ae 3°4 4-0 3°4 5:2 30 4-5 4:4 3-0 Leaf blotch |p iB PE A A P E IP P P Leaf hairs above 1B id P |B iP [e A Ee |p RP Leaf hairs beneath P iE IP IP P z EB iE P A Leaf yellow glands P P 1p A A P P \e P A Ochrea cilia A A A A A A 1p P IP Je Ochrea hairs IE 1B IE A A P P 1b A lp Perianth colour Pk Pk Pk uy W W W Inks W Pk Perianth glands 12 iE P P 1p P P P P 1B Peduncle glands | Res MA P P I P P 1F BE 12 Peduncle hairs P IE A A A A A A A A Ochreola cilia A A A A |B A iP ie P P Ochreola hairs A A A A A A P |e A A Stem hairs Ee P P P P iP A 1p A 1p Stem red spots A A Pe A A A P 1p A A Origin : 1. European Russia, 1908 (BM, No. 95/6/8/2848). 6. Surrey, 1959 B. A. Kneller Herb. Univ. Leicester). 2. Tepelen, Albania, 1935 (BM, No. 95/6/8/2420). 7. Scotland, 1959 U. K. Duncan (Herb. Univ. Leicester). 3. Athens, Greece, undated (BM, No. 95/6/8/789). 8. Tlemcen, Algeria, 1929 (MANCH). 4. Valais, Switzerland, 1895 R. Murry 9. Osses, France, 1961 £. K. Horwood (BM, No. 95/6/8). (Herb. Univ. Leicester). 5. Barcelonia, Spain, 1923 F. Jenner 10. Thraki, Greece, 1961 P. W. Ball (BM, No. 95/6/82/4857). (Herb. Univ. Leicester). The nuts of P. Japathifolium are usually ovate, biconcave to planoconcave, brown and shiny; a few trigonous nuts also occur. (c) Results Herbarium data More than 500 herbarium specimens have been seen and about 150 (well-preserved specimens selected to cover a wide geographical area) have been scored. Some typical scoring data are presented in Table 1. It will be seen that the length and breadth of the leaves varies considerably but that there is much less variation in the leaf index. The leaf blotch is usually present and has been found in most cases but it is not always easy to see on preserved material and is, therefore, an uncertain character to score on herbarium material. The perianth colour usually preserves well enough for scoring purposes but in the case of specimen No. 4 it was no longer possible to see it, perhaps because of the age of this specimen. Data from field material About 40 natural populations were scored and the data from some typical populations are given in Table 2.* In the case of measurements, the entries in the table are means. Where there was variation in quantitative (scored) characters, as in numbers of flower parts, within a population, the less common type is given in brackets. Although the leaf blotch was present in all the populations the amount of pigment present in each leaf varied widely within each population. It appeared that the plants growing in less exposed situ- ations had less pigment. *Small populations are usually quite uniform and this becomes evident quite rapidly. With larger populations, e.g. No. 1 in Table 2, about 100 specimens selected at random were scored. Watsonia 5 (6), 1963. 390 J. TIMSON TABLE 2 Data from natural populations Habit Stem pigment Stem red spots Stem hairs Leaf length Leaf breadth Leaf index Leaf hairs above Leaf hairs beneath Leaf yellow glands Leaf blotch Ochrea cilia Ochrea hairs Peduncle glands Peduncle hairs Perianth glands Perianth colour Perianth segments 4 (5) Stamens 6 Styles 2 (3) Ochreola cilia P Ochreola hairs P A Ochreola glands P P — Hoody yy yheNyydiT — m NO XO on Oe) co fetacha-lacha-la-la-ha- hee koala halal es Wav — oun TP UVUVUVUS HA UYYY ppp ry ar = ion ha Cee ee ea eS oie rg NAAR Fg NAAR rg re The habitats: 1. Borough-on-the-hill v.c. 55. 19 July, 1959. A large population of c. 2-3,000 plants in wheat and barley fields. 2. Skeffington, v. c. 55. 31 July, 1960. Hollow by roadside containing water. 3. Billesdon, v.c. 55. 18 Sept., 1960. Damp ground by side of road. 4. Burton Overy, v.c. 55. 24 July, 1961. Roadside, bare patch of ground with few other plants. TABLE 3 Data from cultivated plants. Ll bo Uy oN H OV Habit E(D) (D) Stem pigment iP Stem red spots A Stem hairs A(P) Leaf length Leaf breadth Leaf index Leaf hairs above Leaf hairs beneath Leaf yellow glands Leaf blotch Ochrea cilia Ochrea hairs Peduncle glands Peduncle hairs Perianth glands Perianth colour Perianth segments Stamens tyles Ochreola cilia Ochreola hairs Ochreola glands bey g Ae Urg oO it Ow wWwn~n WwW oOV~] NWN m= \O 00 Won —ON DODO OOH ORE AM OD > hy A B00 yD Py io PPrPr oops UP UUUUUY @ — PAW Aon ISIN TP U an SUP UUUUUH > Uo qn SUP UUUUUA URS Un & SU nu sUPVUVUTA VES UD ee one Origin of nuts: 1. Cambridge Botanic Garden; collected Ryston, Norfolk. 2. Botanical Garden, Lund, Sweden, as P. lapathifolium subsp. nodosum. 3. as 1. Collected West Dereham, Norfolk, as P. nodosum. Grown 1959. 4. as 3. Grown 1960. 5. Habitat 1 in Table 2. Plant with no anthocyanin in perianth. 6. Hortus Botanicus Bergianus, Stockholm, as P. lapathifolium subsp. nodosum. Watsonia 5 (6), 1963. PLATE 15a PLATE 15b erage ae} (a) Fruits of Polygonum lapathifolium and (b) fruits of ‘ Te OS py Ayeyimuiea yey Ay tA NA ANT Cfae POPaabAa aaa Tey OAs 04 Aa Oy Ayer 0s | wa msn \—— = BS Sy pti0s 05 64.04 As A504 Fon = Ss BOAR a ANN faa= ACMA, w = ON a et 04 Ay 0 A AY Fram A Ay Aa By A Ay Ay = PON) 61/0) 04070104104 wm > moe ag) = Rate ONE aa AA AA AA = NA CNP SD fy a Et fy fy ly yy nas A. Ay Aa Be 4 punoy JOquINyy SATBY WU01S SIIBY V[OITYIO BI[ID VTOOTYIO siey BoIYIO el[lo VIIYIO yyeousg sirey Jeary dA0QP SIIeY Jeary Yoo[q Jeo] XOpUI JeoT Yypeolq Jes] yisug] Jeo7q INO[OS YVWURIIOg OSOJUSUIO} “JeoT SpULIS MOT[OA JeoT s1ods pol WI9}S Ke) SIa}OVILVYO [VOW oul 9 Adv L Watsonia 5 (6), 1963. 394 J. TIMSON to above as the Critical Characters, but there are other less discrete variations which may also be important. Only controlled cultivation on a large scale would reveal the limits of genetical variation. Since all the pure lines will continue to exist and will not blend or segregate because of the autogamous breeding system, it is clear that P. lapathifolium will continue to consist of a number of distinct pure lines—at least twelve and probably more— and to name all these lines would only add to the existing confusion. It would probably be possible to give names to many combinations of characters so that a Rubus-like situation ona smaller scale would result. Fora taxonomist taking a special interest in P. /apathifolium the 16 combinations of the Critical Characters could form the basis of a classification but for less specialised use, e.g. in floras, I suggest that the species P. lapathifolium should not be subdivided. 6. TAXONOMIC REVISION P. LAPATHIFOLIUM L. Spec. Pl. ed 1; 360 (1753). Principal synonyms: P. tomentosum Schrank, Baiersche Flora (1789) P. nodosum Pers., Synopsis 1 (1805). P. brittingeri Opiz, Natural. 8: 74 (1824). P. danubiale Kern., Oesterr. Botan. Zeitschr. 25: 255 (1875). P. petecticale Druce, Fl. Bucks. 287 (1926); Rep. Bot. (Soc.) Exch. Cl., (1926). Annual. Plant erect, decumbent or nearly prostrate. Stem green often becoming red from the base in the autumn, slightly pubescent, with or without discrete red spots. Leaves 3:7—15-1 x 0:7—4-6 cm, 3:0—5:3 times as long as broad, lanceolate, usually with dark central blotch, usually pubescent on both surfaces, with or without tomentose hairs beneath, glandular beneath with either pellucid or golden-yellow glands. Ochreae not or very shortly ciliate, pubescent, occasionally glandular. Peduncle glandular, with or without a few short hairs. Inflorescences dense-flowered, one or many, terminal. Perianth glandular, with or without pink anthocyanin, with 4 (5) segments. Stamens (5) 6; styles 2 (3). Ochreolae not or shortly ciliate, glabrous or slightly pubescent, sometimes glandular. Fruits biconcave, rarely trigonous, c. 2:5 x 2:25 mm, brown to dark brown, shiny. Holotype. The specimen in the Linnean Herbarium CLINN) London is clearly not P. lapathifolium and does not appear to be a herb at all but to be a branch of a shrub. Fortun- ately Linnaeus refers in Species Plantarum to the specimen in Hortus Cliffortianus, and this specimen, which is in the British Museum Herbarium (BM) (No. 42/2), agrees with Linnaeus’s description and with the generally accepted concept of P. /apathifolium. Geographical Distribution P. lapathifolium is a widespread weed in Britain and it is probably present in all vice-counties. Simmonds (1945) suggests that it may be more common in the south than in the north of Britain but there is no evidence that this is more than a reflection of the more intensive agriculture in the south. It is found as a weed throughout Europe to 65° N in Iceland (Moss 1914), 68° N in Lapland (Lindman 1926), 70° N in Norway (Blytt 1906), and in Russia to 70° 20’ N (Komarov 1936). It is also recorded from north India (Hooker 1890), South Africa (Moss), North, Central and South America and the West Indies (Small 1913). It is regarded as native in Europe and Asia and as a naturalised alien in America and Australasia. Specimens very similar to the British and European material have been seen in the British Museum Herbarium from the United States, Canada, Japan, China, and Tibet (‘marshy places near Lhasa’). It is a weed of arable land, disturbed ground, and river banks and appears to be relatively intolerant of competition. It probably, therefore, does not extend far beyond the limits of cultivation. ACKNOWLEDGEMENTS My grateful thanks are due to Professor T. G. Tutin, under whose supervision this work was carried out, for his advice and encouragement. Watsonia 5 (6), 1963. THE TAXONOMY OF POLYGONUM LAPATHIFOLIUM L. SB REFERENCES BiyTT, A. (1906). Haandbog i Norges Flora. Oslo. BriTTon, C. E. (1933.) British Polygona, Section Persicaria. J. Bot., Lond. 71, 90-98. Danser, B. H. (1921). Contributions a la systématique du Polygonum lapathifolium. Rec. Tray. Bot. Neerl. 18, 125-210. Davey, F. H. (1909). Flora of Cornwall, Penryn. FERNALD, M. L. (1940). Gray’s Manual of Botany. 8th. Ed. New York. Hooker, J. D. (1890). Flora of British India 5. London. Kocnu, von W. (1935). Stipa joannis Celak. und Polygonum brittingeri Opiz fur die Schweiz neue Bluten- pflanze. Ber. der Schweiz. Bot. Ges. 44, 337. Komaroy, V. L. (1936). Flora URSS 5. Leningrad. LINDMAN, C. A. M. (1926). Svensk Fanerogamflora. 2nd Ed. Stockholm. LINNAEUS, C. (1753). Species Plantarum. Stockholm. Moss, C. E. (1914). Cambridge British Flora 2. Cambridge. Persoon, C. H. (1805). Synopsis Plantarum 1. Paris. SCHOTSMAN, H. D. (1950). Debouw der klieren van enige Polygonum soorten en bastaarden. Over. uit. het. Ned. Kruid. Arch. 57, 262-275. SCHRANK, F. von P. (1789). Baiersche Flora. Munchen. ScHuSTER, J. (1906). Versuch einer natiirlichen Systematik des Polygonum lapathifolium L. Mitt. Bay. Bot. Ges. 50, 74-78. Simmonps, N. W. (1945). Polygonum L. J. Ecol. 33, 117-143. SMALL, J. K. (1913). Flora of the South-Eastern United States. 2nd Ed. New York. STYLES, B. T. (1962). The taxonomy of Polygonum aviculare and its allies in Britain. Watsonia 5, 177-214. TuTin, T. G., (1952). Polygonum, in Clapham, A. R., Tutin, T. G. & Warburg, E. F. Flora of the British Isles. Cambridge. Watsonia 5 (6), 1963. BOOK REVIEWS Atlas of the British Flora. Edited by F. H. Perring and S. M. Walters. Pp. xxiv + 432. Botanical Society of the British Isles and Thomas Nelson and Sons Ltd., London and Edinburgh, 1962. Price £5 5s. Od. The proposal for mapping the distribution of British plants made by Professor A. R. Clapham at the B.S.B.I. Conference in 1950 has been most faithfully and ably implemented under the editorship of Drs. Perring and Walters. Between 1954 and 1960 over 1,500,000 field records, contributed by some 1,500 amateur and professional botanists, as well as a vast accumulation of data from literature and herbaria, were sorted and, by ultra-modern mechanical methods, tabulated on to maps. The superlative volume which was the end product of this activity was made possible only by the support of the Nuffield Foundation and the Nature Conservancy, who generously financed the scheme from its inauguration. The Azlas contains over 1,600 maps, preceded by a 15-page introduction covering the history of the mapping of plant distribution, an account of the origin and history of the B.S.B.I. Maps Scheme and details of the production of the Atlas. The maps are arranged, in most cases, four to a page, and on them are depicted the distributions of about 1,700 species of flowering plants and vascular cryptograms which are found in Great Britain and Ireland. The basic recording unit is the 10 x 10-km. square of the Ordnance Survey National Grid, and each map contains approximately 3,500 of these. It is indeed an achievement that all but 7 of the 3,500 squares were visited by botanists during the course of the survey and floristic lists compiled from each of them. A brief study of some of the maps quickly reveals a number of interesting facts — the truly southern distribution of such species as Euonymus europaeus, Tamus communis, Rosa arvensis, Viburnum lantana and Euphorbia amygdaloides. The typical northern distribution as shown by Polygonum viviparum, Betula nana and Juncus trifidus. The western distribution of Lotus hispidus and Rubia peregrina, and the eastern distribu- ~ tion of Chenopodium botryodes, Trifolium ochroleucon, Verbascum pulverulentum and Sonchus palustris. The scarcity in Scotland, and in Ireland, of such common lowland English weeds as Aethusa cynapium, Euphorbia peplus, Convolyulus arvensis, Tragopogon pratensis subsp. minor and Hordeum murinum. The paucity in western England of common southern plants like Silaum silaus, Bryonia dioica, Lysimachia nummularia, Primula veris, Galium verum and Carex hirta, and the very restricted distribution of Oenanthe crocata in East Anglia. A more detailed study of the maps shows that the rarer species have been very well documented, but that, in southern England at least, many common species have been under-recorded. The 10 x 10-km. square is undoubtedly a much more adequate recording unit than the vice-county, though, unfortunately, the placing of a single dot within a square still does not indicate the frequency of a species within that square. Thus, for Clematis vitalba, the dot in 51/28 represents isolated occurrences about Willesden, Kingsbury, Highgate and Hampstead, where the plant is an introduction, while the dot in 51/09 represents the species as abundant and native. Many other instances come readily to mind. It is, however, very difficult to see how this problem may be overcome without introducing a mass of symbols indicating different degrees of frequency or rarity. The selection of species to be mapped must have been a complex one for the eminent committee appointed for the task, and so far as native species are concerned there appears to be little to criticise. The choice of alien species is less satisfactory, and too many well-established species have been excluded — to name but a few — Sisymbrium loeselii, Rosa rugosa, Astrantia major (perhaps native in Salop?), Falcaria vulgaris, Erica lusitanica, Trachystemon orientalis, Campanula medium, Cicerbita macrophylla, Lagarosiphon majus and Lilium pyrenaicum. Was it lack of data, or space, that compelled their exclusion? If not, I feel that they have as great a claim to inclusion as Fagopyrum esculentum, Rorippa austriaca, Coronilla varia, Laburnum anagyroides, Cerastium tomentosum, Sedum spurium, Tilia x europaea and Phalaris canariensis — the latter, in particular, does little more than plot the distribution of pet-birds and rubbish-tips. Hybrids are mostly excluded, though a map is given of the distribution of Apium nodiflorum x repens. I should like to have seen maps of Eguisetum x litorale, Geum x intermedium and Senecio x ostenfeldii but it is likely that these will appear with others in the critical supplement to the Atlas now in the course of preparation. Subspecies are also omitted with the exception of those of Arenaria norvegica, Armeria maritima and Asparagus officinalis. It would have been useful if others had been given, particularly Fumaria muralis 396 Watsonia § (6), 1963. BOOK REVIEWS 397 subsp. boraei, which has a very different distribution from F. muralis subsp. muralis. We are, however, promised a further selection of subspecies in the critical supplement. Locking through the maps a number of random points come to mind — is not Clematis vitalba native on the Magnesian Limestone of Yorks. and Durham? How many of the dots on the map of Ranunculus aquatilis refer to R. peltatus which is best regarded as a distinct species? The ‘introductions’ shown for Aconitum anglicum may well refer to hortal forms of A. napellus. Are many of the old records of Cheno- podium urbicum valid — or were they merely forms of C. rubrum? Herniaria glabra is shown as a casual in 51/18, yet it has persisted as an introduction in the ‘square’ for at least 20 years. Silene nutans could be found in its Surrey (51/17) station well after 1930. May not Euphorbia lathyrus be native in some woods in southern England — particularly in Surrey and Sussex? Anthriscus caucalis is often only casual in some of its inland localities. The method of using different symbols on a single map to show pre- and post-1930 records is a very good one that could have been usefully extended to many other diminishing species, e.g. Berberis vulgaris, Erophila verna, Viola palustris, Geranium lucidum, Saxifraga tridactylites, various species of Potamogeton and Orchis simia. It is helpful also in showing the spread of a number of adventive species, e.g. Impatiens capensis, Heracleum mantegazzianum, Buddleja davidii and Veronica filiformis, since 1930. Curiously, Impatiens glandulifera is not included under this method. Yet another useful idea is the system of showing inland records on railway ballast of certain coastal species. e.g. Cochlearia danica, Cerastium atrovirens and Corrigiola litoralis. A major feature of the work is the inclusion of a series of transparent overlay maps which nestle neatly in a pocket at the end of the volume. These show rivers, vice-counties, altitude, location of chalk and limestone, average temperatures, humidity and rainfall. It is by the combined use of the overlays and the maps that the correlations between plant distribution, climate, altitude and soil are revealed and a veritable mine of information made readily available. This is the vital purpose for which the Af/as was created, and for which contributors and editors alike may take the greatest credit. An appendix lists vice-county records excluded from the maps, usually because of the absence of localized records. Lastly, there is a short bibliography, which in my cpinion is quite inadequate for a work of this impor- tance; and it is an important work — one of the most significant books published on the flora of our islands. It will remain for ever a monument to the enthusiasts who created it. D. H. KENT Native Wild Plants of Eastern Canada and the adjacent northeastern United States. F. H. Montgomery. Pp. xxxvi + 193, with 24 small coloured plates and 298 line drawings in the text. Toronto: Ryerson Press. (Agents: Bailey Bros. and Swinfen Ltd.), 1963. Price £2 2s. California Desert Wild Flowers. Phillip A. Munz. Pp. 122, with 96 colour photographs and 172 line drawings in the text, 2 maps. Berkeley and Los Angeles: University of California Press (Agents: Cambridge University Press), 1963. Price £1 4s. Here are two more little flower-books, both written con amore by professors for amateurs in opposite corners of the North American Continent. Professor Montgomery’s book, illustrated with his own drawings, will be appreciated by those who swear by our Collins ‘Pocket Guide to Wild Flowers’ but, as it deals with only about 400 native and mostly herbaceous wild flowers of eastern Canada, and excludes the grasses, sedges and ferns, it is perhaps more for the beginner on the spot than for the visitor from Britain. Prof. Munz’s Californian desert book will have a wider appeal because it deals with a single, and most remarkable, vegetation type. I suppose that less than half a dozen of the strange and beautiful plants shown and described here will ever be found in Britain, even in gardens, but very many of us will visit southern California and, with vivid memories of Walt Disney’s film or the Sherman Hoyt Cactus House at Kew, will want to brave the heat and the rattlesnakes. Prof. Munz tells us in less than three pages most of what we need to know about plant life in the Mojave and Colorado Deserts. His selected shrubs and herbs — with such fascinating common names — are placed in five sections, the first devoted to ferns and cone-bearers while the remaining four are based on flower-colour. N. Y. SANDWITH Skanes Flora. Henning Weimarck. Bokférlaget Corona AB, Lund, 1963. Price 45 Swedish Kronor Skane, the southernmost province of Sweden, has a rich and interesting flora. Professor Weimarck’s new account of it (in Swedish) will be welcomed by botanists interested in phytogeography and taxo- nomy as well as by those who live in or visit the region covered. Watsonia 5 (6), 1963. 398 BOOK REVIEWS To the English reader this book contains some features which will be unfamiliar in local floras. Most important of these is that it really is a flora in the fullest sense of the word. It contains clear but concise descriptions of families, genera and species, as well as keys for identification, so that it forms a complete field guide to the vascular plants of Skane. Chromosome numbers are also given, those determined from native material being distinguished by asterisks; these provide impressive evidence of the activity of Swedish cytologists. Ecological and distributional data, including the presence or absence of species in the neighbouring provinces and in Denmark, are also given, and there is a good glossary. The book embodies the results of many years of botanical exploration carried out by Professor and Fru Weimarck, at first by bicycle and later by car, with the assistance of a number of others. It is excellently printed and strongly bound, and provided with a useful map. It can be strongly recommended to anyone interested in Swedish plants and is a necessity for botanists visiting the beautiful and still largely unspoiled province of Skane. T. G. TUTIN Watsonia 5 (6), 1963. INDEX New taxa, names and combinations are given in heavy type Alchemilla filicaulis, chromosome numbers, 323 Alchemilla filicaulis subsp. vestita, 305 Alchemilla filicaulis, variation in, 305 glaucescens, 260 hybrida, 261 minima, 304 vestita, 304 Allen, D. E., A new variety of Valerianella locusta (L.) Betcke, 45 An Australasian species of Crassula introduced into Britain, by J. R. Laundon, 59 Arum italicum, 106 italicum subsp. neglectum, 107 maculatum, 107 maculatum subsp. danicum, 108 neglectum, 106 Atlas of the British Flora, ed. F. H. Perring & S. M. Walters (Review by D. H. Kent), 396 Ball, P. W. & Heywood, V. H., The taxonomic separation of the cytological races of Kohl- rauschia prolifera sensu lato, 113 Barling, D. M., Studies on the biology of Poa subcaerulea Sm., 163 Batrachium, general considerations in taxonomy, 294 chromosome numbers, 123 Beitrag zur Kenntnis von Rumex No. XV, by K. H. Rechinger, 64 Bradshaw, M. E., Sell, P. D. & Walters, S. M., The nomenclature of Alchemilla minor auct. brit., 259 Bradshaw, Margaret E., Studies on Alchemilla filicaulis, Bus., sensu lato, and A. minima Walters. Introduction, and I. Morphological variation in A. filicaulis, sensu lato, 304 Ii. Cytology of A. jilicaulis, sensu lato, 324 Breeding behaviour, of Salicornia, 158 mechanism, of Poa subcaerulea, 167 British forms of Tuberaria guttata (L.) Fourreau, by M. C. F. Proctor, 236 Californian Desert Wild Flowers, by Philip A. Munz (Review by N. Y. Sandwith), 397 Californian Flora, by Philip A. Munz (Review by N. Y. Sandwith), 110 Calystegia x lucana, 100 sepium, 88 silyatica, 88 Chenopodium, chromosome numbers, 118, 120 Chenopodium album, 47 reticulatum, 49 viride, 48 Chromosome number, Chenopodium album, 118 berlanderi, 120 bonus-henricus, 120 ficifolium, 120 murale, 120 opulifolium, 120 polyspermum, 120 reticulatum, 118 rubrum, 120 urbicum, 120 variabile, 120 viride, 120 Circaea alpina, 265 intermedia, 265 lutetiana, 265 Watsonia 5 (6), 1963. 399 Chromosome number, Dactylorchis majalis, 38 Melampyrum pratense, 335 Poa subcaerulea, 169 Polygonum arenastrum, 201 aviculare, 201 boreale, 201 maritimum, 201 oxyspermum, 201 raii, 201 rurivagum, 201 Ranunculus aquatilis, 123 baudotii, 123 circinatus, 125 fluitans, 125 hederaceus, 123 ololeucos, 123 omiophyllus, 123 peltatus, 123 pseudofluitans, 125 rionii, 125 sphaerospermus, 125 trichophyllus, 124 subsp. Jutulentus, 125 tripartitus, 123 Salicornia dolichostachya, 152 europaea, 152 perennis, 152 pusilla, 152 ramosissima, 152 Trifolium occidentale, 81 Chromosome number, morphology and breeding behaviour in the British Salicorniae, by D. H. Dalby, 150 Circaea, chromosomes, 265 Circaea in the British Isles, by Peter H. Raven, 262 Cole, M. J., Interspecific relationships and intra- specific variation of Chenopodium album L. in Britain. I. The taxonomic delimitation of the species, 47 II. The chromosome numbers of C. album L. and other species, 117 Compatibility, in Calystegia, 96 in Hypericum calycinum, 368 in Salicornia, 160 in Trifolium occidentale, 81 in Trifolium repens, 81 Cook, C. D. K., Sparganium in Britain, 1 Studies on Ranunculus L. sub- genus Batrachium (DC.) A. Gray. I. Chromosome numbers, 123 Il. General morphological considerations in the taxonomy of the subgenus, 294 Coombe, D. E., Trifolium occidentale, a new species related to T. repens L., 68 Cousens, J. E., Variation of ‘some diagnostic characters of ‘the sessile and pedunculate oaks and their hybrids in Scotland, 273 Crassula helmsii, 59 Cynosurus cristatus—Betonica officinalis nodum, 77 Cytology, of Melampyrum pratense, 355 of Poa subcaerulea, 169 Dactylorchis majalis, chromosome number, 38 majalis subsp. cambrensis, 41 purpurella, variation in North Wales, 23 traunsteineri, in Yorkshire, 287 400 INDEX Dalby, D. H., Chromosome number, morphology and breeding behaviour in the British Salicorniae, 150 Dianthus nanteuillii, 113 Dony, J. G., The expectation of plant records from prescribed areas, 377 Ebinger, John E., Luzula x borreriin England, 251 Eleocharis austriaca Hayek, a species new to the British Isles, by S. M. Walters, 329 Eleocharis leptostylepodiata, 334 Epipactis, seed dimensions and root diameters, 140 Epipactis cleistogama, 131 dunensis, 132 leptochila, 127 muelleri, 133 Phyllanthes, 136 Euphrasia, 11 Euphrasia, variation in, 224 Euphrasia anglica, 232 confusa, 233 nemorosa, 224 Siricta, 233 Expectation of plant records from prescribed areas, by J. G. Dony, 377 Fertile seed production and self-incompatibility of Hypericum calycinum in England, by E. J. Salisbury, 368 Flora of the. British Isles, by A. R. Clapham, T. G. Tutin & E. F. Warburg, ed. 2 (Review by J. E. Lousley), 256 Flora of New Zeaiand Vol. 1, by H. H. Allen (Review by J. E. Lousley), 174 Floral biology, of Poa subcaerulea, 167 Galeopsis angustifolia, 143 var. calcarea, 147, 148 ladanum, 143 Geranium microphyllum Hook. f. as an adventive plant in Britain, by C. C. Townsend, 43 Germination, of Poa subcaerulea, 168 Germination, seedlings and the formation of haustoria in Euphrasia, by P. F. Yeo, 11 Gilbert, O. L., see Roberts, R. H. Handbook to Plants in Victoria, Vol. 1, by J. H. Willis (Review by V. S. Summerhayes), 327 Haustoria, in Euphrasia, 11 Heleocharis leptostylopodiata, 334 Helianthemum breweri, 236 Heywood, V. H., see Ball, P. W. Hieracia, of the Orkney Islands, 215 Hieracium anglicum, 218 argenteum, 218 aurantiacum, 222 caledonicum, 221 euprepes, 220 iricum, 218 latobrigorum, 221 maritimum, 222 orimeles, 219 pilosella, 222 sarcophylloides, 219 scoticum, 219 Hypericum calycinum, 368 Interspecific relationships and intraspecific varia- tion of Chenopodium album L. in Britain. I. The taxonomic delimitation of the species, by M. J. Cole, 47 Il. The chromosome numbers of C. album L. and other species, 117 Introgression, in Scottish oaks, 279 Watsonia 5 (6), 1963. Key to the Hieracia of the Orkney Islands, 217 Polygonum ayiculare and related species, 213 British Spargania, 2 Kohlrauschia nanteuillii, 115 prolifera, 113 velutina, 114 Laundon, J. R., An Australasian species of Crassula introduced into Britain, 59 Lousley, J. E., Notes on Rumex acetosa L., 67 et x borreri in England, by John E. Ebinger, Luzula forsteri, 251 pilosa, 251 Melampyrum pratense, chromosome number, 355 variation, 336 Melampyrum pratense subsp. commutatum, 366 Morphology, of Salicornia, 150 Native Wild Plants of Eastern Canada and the adjacent northeastern United States by F. H. Montgomery (Review by N. Y. Sandwith), 397 New Records for Illinois Vascular Plants, by G. S. Winterringer & R. A. Evans (Review by F. H. Perring), 110 New variety of Valerianella locusta (L.) Betcke, by D. E. Allen, 45 Nomenclature of Alchemilla minor auct. brit., by M.E. Bradshaw, P. D. Sell &S. M. Walters, 259 Notes on British Hieracia. II. The species of the Orkney Islands, by P. D. Sell & Cyril West, 215 Notes on Rumex acetosa L. in the British Isles (Beitrag zur Kenntnis von Rumex No. XY), by K. H. Rechinger, 64 Notes on Rumex acetosa L., by J. E. Lousley, 67 Orchis latifolia var. eborensis, 287 Orkney Islands, Hieracia, 215 Plant Communities of the Scottish Highlands, by Donald N. McVean & Derek A. Ratcliffe (Review by M. C. F. Proctor), 255 Plant records, expectation from prescribed areas, 377 Poa angustifolia, 170 Poa subcaerulea, 163 cytology, 169 Pollination, of Salicornia, 159 Polyembryony, in Poa subcaerulea, 168 Polygonum arenastrum, 204 aviculare, 204 aviculare and allies, chromosome numbers, 200 boreale, 208 lapathifolivm, maritimum, 211 nodosum, 386 oxyspermum, 211 raii, 209 rurivagum, 207 tomentosum, 386 Prime, C. T., Taxonomy and nomenclature in some species of the genus Arum L., 106 Proctor, M. C. F., The British forms of Tuberaria guttata (L.) Fourreau, 236 Quercus petraea, in Scotland, 273 robur, in Scotland, 273 INDEX 40] Ranunculus, chromosome numbers, 123 Ranunculus flavidus, 297 Ranunculus subgen. Batrachium, 294 Raven, Peter H., Circaea in the British Isles, 262 Rechinger, K. H., Notes on Rumex acetosa in the British Isles (Beitrag zur Kenntnis von Rumex No. XV), 64 Roberts, R. H., Studies on Welsh orchids. I. The variation of Dactylorchis purpurella (T. & T. A. Steph.) Vermeul. in north Wales, 23 II. The occurrence of Dactylorchis majalis (Reichb.) Vermeul. in Wales, 37 Roberts, R. H. & Gilbert, O. L., The status of Orchis latifolia var. eborensis in Yorkshire, 287 Root diameters, of Epipactis, 141 Rumex acetosa, 64 Rumex hibernicus, 65 Salicornia, 150 chromosome numbers, 151 morphology of diploids and tetraploias, 153 Salisbury, E. J., Fertile seed production and self- incompatibility of Hypericum calycinum in England, 368 Seed dimensions, of Epipactis, 140 Seed production, of Hypericum calycinum, 368 Seedlings, of Euphrasia, 11 Sell, P. D., see Bradshaw, M. E. Sell, P. D. & West, Cyril, Notes on British Hieracia. II. The species of the Orkney Islands, 215 Skanes Flora, by Henning Weimarck (Review by T. G. Tutin), 397 Smith, A. J. E., Variation in Melampyrum pratense i, 3386 Some notes on Galeopsis ladanum L. and G. angustifolia Ehrh. ex Hoffm., by C. C. Townsend, 143 Some studies in Calystegia: compatibility and hybridisation in C. sepium and C. silvatica, by Clive A. Stace, 88 Sparganium in Britain, by C. D. K. Cook, 1 Sparganium angustifolium, 4 emersum, 4 erectum, 2 subsp. erectum, 3 microcarpum, 4 neglectum, 4 erectum subsp. oocarpum, 4 minimum, 4 Species—area relations, 377 Stace, Clive A., Some studies in Calystegia: compatibility and hybridisation in C. sepium and C. silvatica, 88 Status of Orchis latifolia var. eborensis Godfery in Yorkshire, by R. H. Roberts & O. L. Gilbert, 287 Studies on Alchemilla filicaulis Bus., sensu lato, and A. minima Walters. Introduction, and I. Morphological variation in A. filicaulis, sensu lato, by Margaret E. Bradshaw, 304 Ii. Cytology of A. filicaulis, sensu lato, 321 Studies on the biology of Poa subcaerulea Sm., by D. M. Barling, 163 Studies in the British Epipactis. V. Epipactis leptochila; with some notes on E. dunensis and E. muelleri, by Donaid P. Young, 127. VI. Some further notes on E. phyllanthes, 136. VII. Seed dimensions and root diameters, 140 Studies on Ranunculus L. subgenus Batrachium (DC.) A. Gray. I. Chromosome numbers, by C]Dake Cook, 123 Watsonia 5 (6), 1963. Il. General morphologicalcon- siderations in the taxonomy of the subgenus, 294 Studies on Welsh Orchids. I. The variation of Dactylorchis purpurella (T. & T. A. Steph.) Vermeul. in north Wales, by R. H. Roberts, 23 II. The occurrence of Dactylorchis majalis (Reichb.) Vermeul. in Wales, 37 Study of variation in Euphrasia by means of outdoor cultivation, by P. F. Yeo, 224 Styles, B. T., The taxonomy of Polygonum aviculare and its allies in Britain, 177 Taxonomy and nomenclature in some species of the genus Arum L., by C. T. Prime, 106 Taxonomy of Polygonum aviculare and its allies in Britain, by B. T. Styles, 177 Taxonomy of Polygonum lapathifolium L. P. nodo- sum Pers. and P. tomentosum Schrank, by J. Timson, 386 Taxonomic separation of the cytological races of Kohlrauschia prolifera, sensu lato, by P. W. Ball & V. H. Heywood, 113 Timson, J., The taxonomy of Polygonum lapathi- folium L., P. nodosum Pers. and P. tomentosum Schrank, 386 Townsend, C. C., Geranium microphyllum Hook. as an adventive plant in Britain, 43 Some notes on Galeopsis ladanum L. and G. angustifolia Ehrh. ex Hofim., 143 Trifolium biasolettii, 83 Trifolium occidentale, 70 Trifolium occidentale, chromosome number, 81 Trifolium occidentale, a new species related to T. repens L., by D. E. Coombe, 68 Liga occidentale—Bupleurum baldense nodum, Trifolium occidentale—Herniaria ciliolata—Cata- podium marinum nodum, 76 Trifolium occidentale—Scilla autumnalis—Jasione montana nodum, 76 Trifolium repens, 68 Tuberaria guttata, British forms of, 236 Valerianella locusta var. dunensis, 45 Variation, in Alchemilla filicaulis, 305 Chenopodium album, 47 Dactylorchis purpurella, 23 Euphrasia, 224 Variation in Melampyrum pratense L., by A. J. E. Smith, 336 Variation, of Polygonum aviculare, 188 Tuberaria guttaia, 238 Variation of some diagnostic characters of the sessile and pedunculate oaks and their hybrids in Scotland, by J. E. Cousens, 273 Walters, S. M., Eleocharis austriaca Hayek, a species new to the British Isles, 329 see Bradshaw, M. E. West, Cyril, see Sell, P. D. Yeo, P. F., Germination, seedlings and the formation of haustoria in Eup/irasia, 11 A study of variation in Euphrasia by means of outdoor cultivation, 224 Young, Donald P., Studies in the British Epipactis. V. Epipactis leptochila, with some notes on E. dunensis and E. muelleri, 127 VI. Some further notes on E. phyllanthes, 136 VII. Seed dimensions and root diameters, 140 pe phe neEy Suds A BIBLIOGRAPHICAL INDEX OF THE BRITISH FLORA COMPILED BY N. DOUGLAS SIMPSON, »™.a., F.Ls. The purpose of this work is to provide references to sources of information whereby flowering plants, vascular cryptogams and Charophytes found in Britain may be identified, their history traced and their geographical range determined. In addition, information is provided on plantlore, local names, poisonous plants and weeds. The work has been in preparation for nearly twenty years and contains over 65,000 entries, including references to books, articles and manuscripts relating to the flora of the British Isles from the fifteenth century to the present time. Demy 4to, bound, 448 pages in double column. Limited Edition of 750 copies. PRICE £3 I5s. (postage extra) Obtainable from the compiler at ‘‘ Maesbury,”’ 3 Cavendish Road, Bournemouth, Hants. ROYAL IRISH ACADEMY Robert Lioyd Praeger Fund for Field Natural History Grants, not normally exceeding £40 in any one year, will be awarded for field work relevant to the Natural History of Ireland. Application forms, which should be returned before February 15, may be obtained from The Secretary, Royal Irish Academy, 19, Dawson Street, Dublin, 2, who will also be glad to give further information. IRISH NATURALISTS’ JOURNAL A Magazine of Natural History Published Every Quarter by the I.N.J. Committee. Edited by Miss M. P. H. KERTLAND, M.Sc., with the assistance of Sectional! Editors. Annual Subscription, 19/- post free. Single Parts, 3/6. All communications to be addressed to :— The Editor, Department of Botany, Queen’s University, Belfast. INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the taxonomy or distribution of British vascular plants or Charophytes are invited from both members of the Society and others. They should be typewritten, with wide margins, double spaced, on one side of the paper only; contributors are recommended to keep a carbon copy of their typescripts. The form adopted in recent parts of Watsonia should be used for layout, headings, citations and references. Con- tributors are urged to avoid very complicated hierarchies of headings and sub-headings, and to check carefully the consistency of those that they use. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British vascular plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (N.H.) and B.S.B.I., 1958), and may then be cited without authorities. Otherwise, authors of names must be cited, at least on the first occasion where they appear in the text. Except for citations of the place of publica- tion of plant names, full references should be listed in alphabetical order of authors names at the end of the paper. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals, ed. 3 (London, 1952). References to herbaria should include the abbreviations given in British Herbaria (B.S.B.1., 1957) and Index Herbariorum ed. 3 (1.A.P.T., 1956). Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new taxa, and the main conclusions. Line drawings should be boldly drawn in Indian ink on Bristol board or similar smooth white card, firm smooth-surface white cartridge paper of good quality, or suitable draughtsman’s tracing materials, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. No half-tone shading or wash should be introduced into line-drawings; all shading should be put in the form of separate dots or lines, and where large areas of uniform shading are required (e.g. in maps and diagrams) this should preferably be done by means of suitable ‘ mechanical tints.” Graphs can be reproduced from originals on graph paper with faint blue ruling (not grey or any other colour), but all lines to appear on the finished block must be inserted in Indian ink. Lettering on line-drawings and graphs should be inserted lightly in pencil, and will be finished in uniform style. If an illustration includes plant-names or place-names, it is advisable to type these clearly on a separate sheet of paper. Photographs can only be accepted where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable for the necessary reduction. It should be remembered that the finest detail on the originals may be lost even on the best half-tone blocks. If in doubt about the citation of names or references, or the presentation of illustrations or tabular matter, contributors are advised to consult the editor before submitting their typescripts. Authors will normally receive both galley-proofs and page-proofs. Particular care should be given to the thorough checking of references and tables. Alterations in page- proof should be avoided as far as possible, and any words or phrases deleted replaced by others of equal length. Twenty-five separates are given free to authors of papers. Further copies may be obtained at the Society’s current price; orders should be sent to the editor when the page- proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to the Editor, Dr. M. C. F. Proctor, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANpy for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on _ behalf of a sub-committee of the Society appointed for this purpose. Sir George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and have become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8 vo., 176+xvi pages. Bound in cloth. Price : 10/- ( postage extra). Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. A DARWIN CENTENARY Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the very successful Conference arranged by the Botanical Society of the British Isles in 1959. The contributors include Lady Barlow, G. D. H. Bell, Prof. T. W. Bocher, Prof. W. H. Camp, Prof. N. Hylander, Prof. I. H. Manton, Prof. T. G. Tutin and Prof. D. H. Valentine. Demo 8 vo., 140 pages, with 7 half-tone plates and 11 maps and figures in the text. Bound in buckram. PRICE £1 (postage extra) Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7 The Salisbury Press Ltd., Salisbury—40478 © ¥% WSS Se thy ne A aay K ea Ps 7 vn Ca 0 en ANE Wy, ny \ oa shy Ns, ait ptt Lg tee ei “ite ETH By a ect wane sath? ist ee Se eae! o 3243-2 =e ies = prs? tt = ll Lars A, ie’ aA Soe tor a as ~ ray = fil see as ed & aes Ny be =a aia, bays % oe LESS a P eS Se a =) =} ae eS He Be ES e LAd! ane i) rf Vea sy y ——_ see, i | : i ii sh | | | iP to Cre Aalto se ; Ee aat ty On y ‘ ; Ai a8 ya y rs Fala cya ye Us Mal : : 4 ay el hak Th (fig “a ” ‘ ‘ 4 se Fy 0 ata } bord y Ot “ } , a Sha a " , F My : B i * “ ‘ uron phic ty tt Sie y an i, eS : a 5 sal a Wy 4 P Ml Peas oe ett wand " ire . th “ f ites i ; arn Fe : 5 Pear ae