I >3 •5^ k OF THL U N I V LRS ITY or ILLINOIS 581.942 E>65b yib’A- BIOLOGY JIOLOGY. Return this book on or before the Latest Date stamped below. University of Illinois Library FEB~Q-Q-mie y L161— H41 Digitized by the Internet Archive in 2018 with funding from BHL-SIL-FEDLINK https://archive.org/details/reportofconferen4195bota B.S.B.I. Conference Reports, Number Four SPECIES STUDIES IN THE BRITISH FLORA SPECIES STUDIES IN THE BRITISH FLORA BEING THE REPORT OF THE CONFERENCE UNDER THE TITLE OF THE SPECIES CONCEPT IN ITS RELATION TO THE BRITISH FLORA HELD IN 1954 BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES EDITED BY J. E. LOUSLEY 1955 Published by THE BOTANICAL SOCIETY OF THE BRITISH ISLES, c/o DEPARTMENT OF BOTANY. BRITISH MUSEUM (NATURAL HISTORY). CROMWELL ROAD. LONDON, S.W.7 Sold by the Hotanical Society of the British Isles, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. Printed by T. Bunct-e & Co. JiTD., Market Place, Arbroath. MADE IN GREAT BRITAIN, VI 'feldc' S'^/. sZ--, Z/ / CONTENTS ^0. ^ Page Editorial Note ... ... ... ... ... ... ... 7 Programme of the Conference ... ... ... ... 8 List of Members and Guests who attended the Conference 1 1 Introductory Remarks. Prof. T. G. Tutin ... ... ... 15 X The Experimental Approach to the Species Problem. Prof. S. C. Harland ... ... ... ... ... 16 Species Problems in Plants with Reduced Floral Structure. Prof. T. G. Tutin ... ... ... ... ... 21 The Stability of some Specific Characters as shown by Fossil Records. Dr. H. Hamshaw Thomas .. . ... 27 Variability within Species. Dr. J. H. Burnett Morphological Characters in the Discrimination of Species and Hybrids. Dr. R. Melville Problems Associated with the Species Concept in Mycology. Prof. Alan Burges ■ The Analysis of Variation within the Genus Fucus. Dr. E. Burrows and Dr. S. M. Lodge The Species Concept amongst Bryologists. Dr. E. W. Jones 32 55 65 83 86 The Importance of Ferns to an understanding of the British Flora. Prof. I. Manton ... ... ... 90 The Problem of Asplenium trichomanes. J. D. Lovis ... 99 The Two Sub-species of Asplenium adiantum-nigrum in Britain. Miss M. G. Shivas ... 104 The Dryopteris spinulosa complex in Europe. Dr. S. Walker ... Caltha in the British Flora. Dr. G. Panigrahi C Stellaria nemorum and the Species Concept. Dr. Ch. H. Andreas ... -N G Clinal Variation in Flower Size in Lotus corniculatus. Miss B. A. Poulter 105 107 111 115 Page The Section Eu-Callitriche in the Netherlands. Dr. H. D. ScHOTSMAN . ... ... 116 Ecotypical Variation in Adoxa moschatellina. Dr. H. G. Baker ... . . 118 Variation in Centaurium in West Lancashire. Miss W. M. T. O’Connor ... ... ... ... 119 Problems of Hybridization and Species Limits in some Erica Species. Peter A. Gay ... ... ... 126 The Species Concept in Euphrasia. P. F. Yeo ... ... 128 Some Variation in Salicornia and its Significance. D. H. Dalby ... ... ... ... ... ... 134 Problems of Speciation in the British Species of Arum. Dr. C. T. Prime ... ... ... ... ... 136 Species Problems in Recent Scandinavian Works on Grasses. Dr. A. Melderis ... ... ... ... 140 The Conflict of Categories. Prof. J. Heslop-Harrison ... 160 The Species Concept and Experimental Taxonomy. J. S. L. Gilmour ... ... ... ... ... 173 The Future of Synthetic Taxonomy. Dr. W. B. Turrill 177 Concluding Remarks. Dr. R. W. Butcher .. . ... ... 182 Tlie Distribution-Maps Scheme ... ... ... ... 183 Index 184 The theme chosen for the fourth Conference arranged by the Botanical Society of the British Isles is one of considerable importance to everyone interested in contemporary work on the British flora. The “Species Concept” as it concerns individual workers is usually viewed from a more or less narrow angle, and the purpose of the Conference was to bring together problems and evidence from as wide a field of current research as possible. The papers read proved of outstanding interest and import¬ ance. They include relevant studies of fossil botany, fungi, algae, bryophytes, ferns and flowering plants and of cytology, bio¬ metrics, experimental taxonomy and ecology. They cover recent and new work, some of which has not been previously published, and collectively they provide a broad approach to the subject not available elsewhere. The papers printed in this report have been arranged so that those covering general subjects appear first, followed by accounts dealing with groups of cryptogams and phanerogams, and concluding with contributions which cover suggestions for future developments. The sequence in which they were given at the Conference is set out in the Programme printed overleaf. Scientific names used in the papers are those selected by the speakers and it has not been thought advisable to attempt to standardise them. The arrangements for the Conference were in the hands of the Meetings Committee of the Society. Much of the work fell on Dr. J. G. Dony, Honorary Meetings Secretary, and to his energy and enthusiasm, together with that of the other officers and members who assisted, the success of the Conference was due. I am grateful to J. P. M. Brenan, J. E. Dandy, D. H. Kent, N. Y. Sandwith, H. K. Airy Shaw, and E. F. Warburg, members of the Society’s Publications Committee, for reading the proof and making many valuable suggestions, and to J. G. Dony for preparing the index. J. E. Lousley. 8 CONFERENCE PROGRAMME 1954 THE SPECIES CONCEPT in its relation to the British Flora FRIDAY, April 9th First Session 10.00 a.m. Registration 10.15 A welcome to the Conference The President: The Rev. Canon C. E. RAVEN 10.30 The Experimental Approach to the Species Problem Prof. S. C. HARLAND 11.15 The Importance of Ferns for the understanding of the British Flora Prof. IRENE MANTON 12.00 Interval for Lunch Second Session 1.45 p.m. The Vryopferis spimdosa complex in Europe Dr. S. Walker 2.00 The stability of some specific characters as shown by the fossil records Dr. H. HAMSHAW THOMAS 2.45 The problem of Aspleniuni- tricliomanes Mr. J. D. Lovis 3.00 The Species Concept among Bryologists Dr. E. W. JONES 3.45 The analysis of variation within the genus F'ucus Dr. E. Burrows and Dr. S. M. Lodge 4.00 Interval for Tea Third Session 4.45 Problems associated with the Species Concept in Mycology Prof. A. BURGES 5.30 Problems of speoiation in the British species of Arum Dr. 0. T. Prime 5.45 The conflict of categories Dr. J. HESLOP-HARRISON 6. .30 The Relationship of the two subspecies of Aspleni'iim adiantnm-niqrum in Britain Miss M. G. Shtvas 6.45 Interval for Dinner 9 The Hall will be open from 7 p.m. when the exhibits illustrating papers read on both days can be inspected 8.15 OPEN MEETING TO WHICH MEMBERS AND GUESTS NOT ATTENDING THE CONFERENCE ARE ALSO INVITED The Distribution Maps Scheme Prof. A. R. CLAPHAM The part Field Botanists may play in the Maps Scheme Dr. S. M. WALTERS There will also he an exhibit of the equipment for recording and mapping to he used in the Scheme SATURDAY, April 10th First Session 10.15a.m. The Species Concept and Experimental Taxonomy Mr. J. S. L. GILMOUR 11.0 11.40 12.00 Species Problems in plants with reduced floral structnres Prof. T. G. TIITTN Species Problems in recent. Scandinavian works on Grasses Dr. A. MELDERIS Variation in Snlicornia and its significance Mr. D. H. Dalby 12.15 p.m Morphological characters in the discrimination of species and hybrids Dr. R. MELV1LT>E 1.00 Interval for Lunch 2.80 2.45 3.00 3.45 4.00 4.15 5.00 Second Session Problems of hybridization and species limits in some Erira species Mr. P. A. G.ay The Species (’oncept in Euphrasia Mr. P. F. Yeo Variability within species Dr. J. H. lUTRNETT Cnlfhn in tlie British Flora Mr. G. Panto RAET Variation in Centaurium Miss W. T. M. O’ Connor The future of Synthetic Taxonomy ‘ Dr. W. B. TURRTLT. Closing Remarks by the President 10 SUNDAY, April 11th FIELD MEETING TO BOX HILL jointly with British Bryological Society Leader: Mr. E. C. Wallace The programme printed above is as circulated before the Conference. It was carried out as stated except that the President was unable to be present and Professor T. G. Tutin and Dr. R. W. Butcher, Vice-Presidents, opened and closed the proceedings in his place. Four additional exhibits were included as follows: — St ell aria nemorvm L. arid tire Species Concept Dr. Ch. H. Andreas Ecotypical Variation in Adoxa moschafellinn L. Dr. H. G. Baker C'linal Variation in Flower Size in Lotvs corniculatus L. Miss B. A. POULTER The Section Fyu-Callitriche in the Netherlands INlisS H. D. SCHOTSMAN 11 LIST OF MEMBERS AND GUESTS WHO ATTENDED THE CONFERENCE, April 9 and 10, 1954 (The following list does not include those who only attended the open meeting on the eA^ening of the first day.) F. W. Adams G M. J. Christmas G A. D. G. AgneAv Prof. A. R. Clapham G Mrs. Alderson G T. H. Clifford D. E. Allen M. J. Cole K. L. Alvin T. G. Collett G Dr. Ch. H. Andreas Miss A. Conolly Miss J, AndreAvs J. A. Crabbe G. M. Ash A. C. Crundwell Miss D. E. Ashhurst D. H. Dalby A. G. Bailey G Miss G. W. Dalby G Miss A. M. Baird G Miss I. Daniels Dr. H. G. Baker G Miss E. Davenport G P. W. Ball D. Davidson E. B. Bangerter Dr. E. W. Davies Miss F. M. Barton Dr. P. H. Davis G Miss A. A. Baylis G Miss M. de Vos Miss D. Baylis Miss 0. R. Dewey Miss K. Benson-Evans G Miss C. I. Dickinson Miss D. Bexon G Miss J. Dilnot Dr. K. B. Blackburn Dr. J. G. Dony Miss N. M. Blaikley T. R. Eagles G R. A. Blakelock Rev. E. A. Elliot M. Borrill G J. Farrand B. N. Bowden G S. Feinsilber Dr. H. J. M. Bowen R. S. R. Fitter Miss M. E. BradshaAv H. J. Fletcher Dr. 0. E. Brett G L. L. Forman D. W. Brett Miss H. Franks G. M. Brown G E. J. Friend G J. R. Lang Brown P. A. Gay 0. Buckle Mrs. P. A. Gay Miss W. F. Buckle Mrs. A. N. Gibby G Miss N. T. Burhidgc J. S. L. Gilmour Prof. A. Burges Mrs. S. Gilmour D. H. Burnett D. R. Glendinning Dr. J. H. Burnett Miss C. M. Goodman Dr. E. Burrows G Miss B. J. Golding Dr. R. W. Butcher K. M. Good way Miss D. A. Cadbury Miss V. Gordon Mrs. C. M. A. Cadell R. A. Graham J. F. M. Cannon G S. W. Greene Mrs. J. F. M. Cannon G Dr. J. W. Gregor B. V. Cave G Miss M. Gregory Miss Y. Chamberlain Miss M. E. Griffiths G S. K. Chaudhuvi P. C. Hall Mrs. P. C. Hall D. J. Hn in bier (t Miss M. P. Hancock P. 1). Hanson D. J. Harberd G Mrs. S. C. TTarland R. Hailey Dr. -I. G. Hawkes G ]\I iss A. R. Haygartb-Jackson J. H. Hemsley F. N. Hepper Dr. J. Heslop-Harrison G Mrs. J. Heslop-Harrison G Miss M. M. Hindmarsb Miss O. Holbek Dr. M. G. Hughes G Dr. A. T. Hnnziker G A. A. Idle Miss E. M. C. Tsherwood G P. W. James G Miss F. M. Jarrett A. C. Jermy Dr. E. W. Jones G Miss M. Kefallino D. H. Kent G C. C. King G Miss E. E. King G R. Knowles Miss R, G. B. Laidlaw Dr. J. M. Lambert G Miss J. Laptain W. N. Lawfield (; Mrs. W. N. Lawfield Miss S. M. Littleboy Miss C. E. Longfield J. E. Lonsley G .1 . Lovell J. D. Lovis G Mrs. J. D. Lovis Dr. A. G. Lyon Prof. T. Manton G A. R. H. Martin D. McClintock G Miss C. Macdonald G Miss C. M. Medd R. D. Meikle Dr. A. Melderis Dr. R. Melville G Miss D. Meyer H. Meyer Miss M. E. Mil ward R. Minor G D. M. Moore Miss B. ]\r. C. Morgan E. Nelmes G P. J. Newbonld P. M. Newey P. R. Norman G G. N. Oakeshott Miss W. Al. T. O’Connor J. Onnsted J. R. G. Packer Miss P. A. Padmore G G. Panigrahi G K. Parry G. J. Paxman R. AI. Payne F. Perring C. D. Pigott Ali ss B. A. Poulter Dr. C. T. Prime N. At. Pritchard Aliss J. P. Pugh R. C. Readett G 1). A. Recaldin 1). A. Reid B. W. Ribbons Aliss 0. At. Rob N. K. B. Robson J. Grant Roger I. H. Rorison Dr. F. Rose Dr. E. M. Rosser Mrs. B. H. S. Russell J. S. Ryland R. E. Sandell N. Y. Sandwith G Atrs. V. 0. Sankey G J. Sansome J. A. Sargent Mrs. N. Saunders G Aliss H. D. Schotsman G R. C. Seeley G R. 0. Sharpies G J. D. Shepherd G Miss M. G. Shivas Aliss P. M. Smith G Miss S. G. Smith Dr. E. Smithson J. E. S. vSonster l:} ^liss ('. J . Si)urgiii (i W. T. Steam E. \j. Swaim (i ^Jiss A. M. Swiiiiiey Cl Miss J. Taylor (1 R. Teasdale (> Dr. H. Hamsliaw Thomas C R. G. Thomas G Mrs. Thornton G Miss K. Toiisny G Miss J. Turnham Dr. AV. R. Tiirrill Prof. T. G. Tutin Prof. D. H. Valentine Dr. J. G. Vanghan N. M. Wace G Miss J. Wakefield G Dr. S. AValker E. C. Wallace Dr. S. M. Walters P. J. Wanstall Dr. E. E. Warhnrg G l». S. Ward G Dr. Ik d. Watson Mrs. W. Boyd AVatt Miss M. MeCallnm W(d)stor Mrs. B. AVelch G R. P. AA^eston A. W. Westriip Miss D. M. AVethered G E. White Miss AI. Ai. AAliitijig B. A. Whitton G D. A. AVilkins Aliss AI. A. AA^illiams Al. H. AVilliamson Aliss R. AVitton Aliss A. E. Wood P. J. AA'ood G S. R. J. AAh)odell -1. E. Woodhead P. E. Yeo Dr. D. P. Young INTltOD V CTOH Y' IIK JM AKKS INTRODUCTORY REMARKS The President of the Society, The Rev. Canon C. E. Raven, was unable to attend the Conference owing to bereavement, and at very short notice. Professor T. G. Tutin, a Vice-President, opened the proceedings with the following remarks: — I am sure we all greatly regret the absence of our President, and that it would be your wish to offer him our sincere sympathy in his recent loss. We shall miss not only his opening remarks but also the experience and skill which made his chairmanship of the last Conference so memorable. It is now my pleasant duty to welcome on your behalf our guests and our speakers for today. I feel sure that, with the happy combination of youth and enthusiasm, and experience and enthusiasm, that we see in this hall, our Conference is bound to be a great success. We are, I take it, here to widen our knowledge of the nature of species, and not necessarily to attempt an improved definition of these important but somewhat nebulous taxa. In fact it seems doubtful whether the time-honoured definition “a species is What a competent taxonomist thinks is a species” can be bettered, except by defining a competent taxonomist. We shall certainly learn more of the kind of criteria used for the recognition of species in different groups of plants and go away, when the time comes, with a clearer idea of the importance of variation. Though as a Society we are primarily interested in what we are pleased to call the higher plants, we are fortunate in having an opportunity today of hearing about the species problems of those who might be termed lower botanists. I use the word in no derogatory sense : their problems are, I think, of the same basic kind as ours and their difficulties often greater. It will be a matter of great interest to us to see how the idea of the species among bryologists and mycologists compares with our own. In addition to hearing about the variation which occurs within a species at any given time we shall also learn something of the permanence of certain characters in geological time. This may save us from the taxonomic nightmares we might have if we went away thinking that species are perpetually in a state of flux. It has become increasingly apparent during this century that for an understanding of the nature of species we must turn to the cytogeneticists and the users of experimental grounds. It is equally clear that classical taxonomy will never lose its value and that it will continue, at the very least, to provide a universal filing system for all kinds of botanical knowledge. We are getting off on the right foot this morning with contributions from two of our most eminent cytogeneticists* and I am sure you are look- ■ ing forward as eagerly as I am to hearing what they have to say. *Se6 Conference Programme on page 8. I() STKCIKS S'I'l DIKS IN 'INK I{|{IT1SII KI-OIIA THE EXPERIMENTAL APPROACH TO THE SPECIES PROBLEM 8. C. Haiiland (University of Manchester) (Professor Harlaiid was unable to attend to read liis paper owing to illness. It was read in liis absence by Miss Angela It. Haygartli- J ackson.) When this subject was first suggested to me for presentation I looked at it something like this: Hasn’t this subject been dis¬ cussed over and over again ? Hasn’t everything that can be said about it already been said? But on further consideration I thought that it might be possible to illustrate by a few examples what I conceive to be some of the most fruitful lines of tliought which have been opened up by the use of the experimental method as applied to the species problem. There are about a quarter of a million species of angiosperms in existence. They have to be classified and described in such a way that any plant brought in from the wild, or grown from a packet of seed obtained from a Botanic garden, can be identi¬ fied. It can be pigeon-holed, so to speak. This task has been one of great difficulty and great magnitude. On the whole it has been accomplished with considerable skill and precision. Errors have of course been made. Sometimes, as in the genus Gossypium, species have been allocated to it which, by the use of the experi¬ mental method, clearly belong elsewhere; and sometimes species which have been assigned to other genera have had to be incor¬ porated in the genus. Two examples may be given. The species formerly known as Thurheria thespesioides is now known as Gossypium thurberi and the species formerly known as Gossypium kirkii is now known as Gossypioides kirkii. I have mentioned these two examples because they illustrate how the experimental method works in practice and what are its tools. First let me reiterate a statement which I have frequently made to my students : that the most important thing about a plant is its chemistry, and the most important biological tools at our disposal are those which reveal in the plant kingdom fundamental biochemical differences and relationships. Orthodox taxonomy does not aim to do this. Its contribution to bio¬ chemistry is — to use a phase commonly used elsewhere — purely coincidental. The experimental method has four main weapons in its armoury. These are, first grafting, second crossing, third cytology and fourth genetics. There are other considerations, such as those arising from ecology, and there is the general information provided by biotic relationships, which are outside the scope of TtlK EXrElUAlKN'J Al. Al’l'itOAfli TO THE bJ'ECTES I'llOELEM. 17 this paper but are nevertheless important. Let me take up these methods one by one and discuss them in the light of the species problem. Grafting If two species can be grafted, it means that they possess in common some factors of a biochemical nature — they are bio¬ chemically related. Information on grafting relationships can be of quite extraordinary value, but for the most part is lacking. Our information, scanty as it is, has been derived principally from the horticulturist. We know, for example, that the pear, quince, apple and hawthorn are all capable of being intergrafted. At least, the quince can be used as a stock for both apple and pear, and hawthorn for apple. Similarly lilac can be grafted on privet; and tobacco, tomato and petunia on the potato. Plants in the same genus can probably almost always be grafted. Plants of the same family can often be grafted, though there are many exceptions. Plants belonging to different families can probably never be grafted. I am aware that cases have been recorded in the literature of alleged successful grafts of members of different families. But these require confirmation. To return to the Gossypium case previously mentioned. Here we had a species, Gossypium kirkii, which could not be grafted on any other known member of the genus. Morphologically the main difference from other cottons was in the possession of a ribbed or winged stem, a character not regarded as important by the taxonomist. On its grafting relationships it was clearly not a Gossypium, that is, it was biochemically different. Now if two species will not graft, they will also not hybridize, and I am not aware of any exception to this rule. Gossypium kirkii fol¬ lowed this rule. Later it was found by Skovsted that it possessed 12 pairs of chromosomes instead of 13 pairs and consequently it was put in a new genus, along with an endemic Madagascar species — Gossypioides hrevilanatum. This case admirably demonstrates what the experimental method can do. A geologist who encounters a territory for the first time makes a rapid survey of the principal features. He can map the main formations and say a good deal about their geological relationships. But even in this country the detailed and precise study of small areas still goes on. And so it is with taxonomy. Crossing and Genetical Relationships What I have said about grafting applies also to plant relation¬ ships as revealed by the results of inter-specific, or even inter¬ generic, hybridization. If two species will graft, they are in some way biochemically related. If they will both graft and cross, the relationship is presumably closer. Let me refer again to Gossypium. The plant known as Arizona Wild Cotton was ]8 SI’KCIKS STL’DIKS IX TilK l{iaTISll Fl.OUA formerly called Thurheria thespesioides. We found first that it would graft on all the species of cotton in our very large collec¬ tion. It would also hybridize with several species, and cytological examination showed that it possessed 13 pairs of chromosomes. One cross with the wild Peruvian cotton, Gossypium raimondil, which also had 13 pairs of chromosomes, proved to be highly fertile both in the first and second generations, and on these grounds it became necessary to change the taxonomic status of this species and it is now known as Gossypium thurberi. Similarly and on the same criteria Erioxylum aridum became Gossypium aridum. Another case may be mentioned. In our studies of the cyto¬ genetics of the genus Senecio, we have made observations on the status of two species, Senecio gallicus and Senecio squalidus. The former was obtained from the Jardin des Plantes in Paris, and the latter is, of course, now widely distributed in this country. Both these species proved to be self-incompatible, a condition by no means infrequent in Compositae, but they hybridized with ease, and both possess 10 pairs of chromosomes. Obviously the nomenclature should be revised and one of these names eliminated. Further preliminary results concerning our experiments with the genus Senecio may be mentioned. At first we were concerned only with the possibility of using groundsel (Senecio vulgaris) as a sort of plant Drosophila. But for this purpose it proved unsuitable. First it is mainly self-fertilised and the progenies of single plants taken from the wild rarely segregate. Second, the great variation exhibited seems to depend on genes leading to minute and unanalysable differences in morphology and physio¬ logy. Third, the life history was not so short as we had been led to expect from casual observation. The discovery of a male sterile form known as “strap”, from its narrow strap-like leaves, led to other lines of work in which the possibilities of interspecific hybridization are being explored. It is stated in the literature that hybrids between S. vulgaris, with 20 pairs of chromosomes, ^nd S. squalidus, with 10 pairs of chromosomes, occur in nature. Using the male sterile S. vulgaris as a female we ultimately got a hybrid, which has been carried on by cuttings. It is, as a rule, completely sterile though a few seeds were obtained by open pollination. Treatment of the hybrid with colchicine induced a certain amount of fertility, and we have raised a few seedlings which may be fertile hexaploids and constitute what is virtually a new species. The artificial produc¬ tion of a new species by making sterile hybrids fertile is now a recognised procedure, not only in experimental taxonomy, but also in plant breeding. Another hybrid has been made between Senecio vulgaris and Senecio inaequidens. Both these species have 20 pairs of chromo¬ somes but present wide differences in morphology and physiology-. TllK KXI'EKlMEM’Ai. AlTitOACH TO THE SPECIES PROBLEM. 19 S. vulgaris is annual and self-fertilising: S. inaequidens is peren¬ nial and self-incompatible. Analysis of the crosses using genetical and cytological methods should tell us a great deal about their relationships. Perhaps the best example of the kind of result which may be expected when a whole genus is studied using the techniques of cytology and experimental taxonomy, is provided by the monu¬ mental and admirable work of Babcock on the genus Crepis. VVe are working on species inter-relationships in the genus Fragaria. Without going into details it may be mentioned that a polyploid series exists in this genus. The European wild straw¬ berry, Fragaria vesca, has 7 pairs of chromosomes, while the the two species Fragaria chiloensis and Fragaria virginiana have 28 pairs and are octoploids. The cultivated garden strawberry has resulted from crosses between these two octoploid species, and may be called an octoploid cultigen. Even casual inspection of the diploid and the octoploid must lead to the conclusion that they are members of the same genus and closely related. It is, however, very difficult to get hybrids between diploid and octo¬ ploid and when they occur they are pentaploids and almost sterile. However, by first doubling the chromosome number of the diploid and making it a tetraploid, crossing with the octo¬ ploid becomes easy and large numbers of relatively fertile hexaploids have been obtained. It now becomes possible to conduct a genetical examination of these synthetic types. I mention the Fragaria case because ease of crossability be¬ tween species of undisputed taxonomic status varies greatly. At one extreme is the ease with which maize, Zea mays, will cross with the Teosinte — a member of another genus, and at the other is the recent hybrid between tomato and a wild Solarium made by Rick. Here it was only possible to make the hybrid by employing a whole series of the most refined techniques. The crossed fruits had to be prevented from abscission by the use of hormones. The tiny embryo had to be extracted at an early stage and grown in an artificial medium. The resulting plant proved to be sterile, but when colchicine treatment was used it was converted into what may be called a completely new species. It is convenient at this point to summarize briefly what are the real aims of the experimental method. First, the direct aim is not that of taxonomic revision. Intensive study of a genus and other closely related genera by means of tools of greater precision — the microscope and the breeding plot, must, however, inevitably lead to taxonomic revision. As I have mentioned, species will often have their taxonomic status changed, though as I emphasized on a previous occasion, what is really surprising is the fact that so few changes are found to be necessary. The aim of the experi¬ mental method goes further than that of the relatively naive one of telling the taxonomist where he went wrong. It seeks to formulate in a very precise way the relationships between species STKCIKS STl DIKS IN TIIK I{1UT1M1 KLOUA ‘JO in terms of genes, cliroinosomes, and cytoplasm, and ultimately in terms of evolutionary processes. The taxonomist is interested in orderly and logical classifica¬ tion. The experimentalist is necessarily also interested in classifi¬ cation, but he would like his own criteria brought in to make taxonomy more precise. He is also concerned with species build¬ ing and species evolution. He wants to know what happens to genes when species are isolated for a few million years. It is suspected that genes do not remain constant throughout long periods of time — that they evolve into new multiple allelomorphic systems with different combinatory properties. But only by in¬ tensive studies of the genetical architecture of carefully chosen single genera, can light be thrown on these profoundly interesting questions. Prof, T. Gr. Tutin congratulated Miss Haygarth-J ackson on the very lucid way in which she had read Prof. Harland’s paper, and expressed the gratitude of the meeting to her. Mr. J. OuNSTED asked what methods were used in colchicine treat¬ ment. Miss Haygarth-J ACKSON replied that many methods have been used. One which was used at Manchester was to grow seedlings and to add a drop of colchicine to their cotyledons. Another was to take seedlings which had just germinated and to soak the radicles in colchicine. Different strengths were used and the period of immersion varied — for example, the radicles might be soaked in a 1% colchicine solution for 12 hours. Dr. H. G. Baker enquired whether any work had been done on Senecio squalidus collected from the areas where the species is believed to be native. Miss Haygarth-J ACKSON replied that all the material they had used at Manchester was collected in this country where the species is, of course, an alien. It was quite likely that the aggressive colonising Senecio squalidus in Britain had changed its nature, and they were anxious to obtain plants from the Mediterranean for study and com¬ parison. SPECIES PROBLEMS IX PLAXTS WITH REDT'CEH FLORAL STRCCTURE 21 SPECIES PROBLEMS IN PLANTS WITH REDUCED FLORAL STRUCTURE T. G. Tutin (University College Leicester). It is well known that the great majority of flowering plants can be recognised by their vegetative characters at least as easily •as by their reproductive ones. Any experienced field botanist identifies the majority of species he meets by their general ‘look’, that is by a combination of vegetative and floral characters, and he is not usually greatly disturbed if the plant has no flower or fruit. In spite of this the characters given in books for the identifica¬ tion of plants and, consequently, those usually recommended to the attention of students, are based mainly on features such as the shape, size and colour of the floral parts, including the fruits and seeds. This is, of course, contrary to the normal practice of botanists in the day to day identification of plants, and has rather the same effect on the speed with which the beginner recognises a species as working to rule has on the operation of a railway. There seem to be two main reasons for this devotion to floral characters and the relegation of vegetative ones to a very secondary position in botanical writing. The first is the perfectly valid one that reproductive features provide the most convenient and most satisfactory means of making a classification and, further, their validity has been tested and their respectability established by at least two centuries of use. The second seems to be very largely a matter of convenience reinforced by the limitations of the written word as a means of communication. The vegetative parts of a plant show far more phenotypic variation than the floral parts, and it is obviously both more troublesome to describe a range of variation and less easy to compare patterns of variation than to compare features in which the variation is sufficiently slight to be neglected. Further, to the casual observer the leaves of, for example, sycamore and plane may appear sufficiently alike to be confused with one another. To the practised eye the colour, texture and exact shape of the lobes are, taken together, sufficiently different to make con¬ fusion impossible. There are probably several other features as well which go to make up the general ‘look’ of the leaves but, even neglecting these, how are we to write down in clear terms the difference between the leaves of these two species so that the reader cannot possibly mistake one for the other? The ecologist, who cannot afford to wait for a plant to flower or to neglect one that is sterile, makes regular use of non-floral characters with, it is to be hoped, great success and reliability. 22 SPECIES STUDIES IN THE BRITISH FLORA The taxonomist has been forced to do the same in distinguishing species in which the flowers are devoid of perianth or incon¬ veniently small and apparently lacking in diagnostic characters. Salix and Ulmus may be instanced as genera in which leaf charac¬ ters are freely used for the delimitation of species and the recogni¬ tion of hybrids. In Alchemilla (Walters, 1949) also, leaf shape and the distribution of hairs are the main characters used for the separation of the different apomicts. In this case the fact that it is apomicts that are being identified makes the problem easier since the range of variation is small compared with that found in sexual species. In the Gramineae the floral structures are so reduced and condensed that the spikelet, a whole branch of the inflorescence, has to be used for descriptive purposes, in place of the individual flower. The taxonomist’s problems are further increased by the apparently very uniform type of leaf and the great number of species. Non-floral characters have been used in this family for at least 50 years, together with floral characters, for the definition of major groups such as tribes and genera and, to some extent, e.g., by Armstrong (1917), for the recognition of species. The degree of difficulty in discriminating species appears to vary considerably between different genera in this family, as in most others. In some, e.g., Melica and Deschampsia, characters of the spikelet and inflorescence seem adequate, but in Agropyron section Elytrigia and in Puccinellia, among others, floral features are of more limited application. In Agropyron leaf characters have for long been used. For instance Babington pointed out that A. repens has “. . . . ribs on upperside not much raised nor nearly hiding the intermediate surface of the leaf”, while A. pungens has “Ribs on upperside of leaf so broad and so elevated as nearly to hide the intermediate part of the leaf”. Characters such as these seem to provide the easiest and safest means of distinguishing the species in flower and the only means when sterile. It is interesting to note, however, that Babington supplies most of his information about the leaf structure in the notes after the diagnosis proper, as though it were not quite respectable to pay attention to such trivialities. In contrast to this is the recent treatment of the genus in Flora N eerlandica, where not only species but also hybrids between them are clearly distinguished by drawings of transverse sections of the leaves. Plants which appear from their leaf structure to be hybrids are found to have a high percentage of abortive pollen, a fact which. confirms the value and reliability of these vegetative features. It will be noticed that, largely on the strength of the leaf structure, Jansen and Wachter have raised Agropyron repens var. maritimum to specific rank. This plant occurs on the east coast of England and possibly elsewhere and is to be found in herbaria sometimes under A. repens and sometimes under A. . . — r- . * •- '’" '-■lS^^' '^m ■ ■ ' t' . X ’ -' - iW - . {«' / i* V ^ f' * '^ ^*‘? .V “r-' V'.!^'C^, ’ > » 4- •■•• ■ *>'‘.!!i.iit‘5’’:.-;' V:,..’ ■. .' '.■rSwi kr‘- ./ ■:.:^ i-* ^ - • ^ 4>. N el * « ip X • - . ■ V.,; ' ^c; V ’^r r* Plate I. FiR. 3. Fig. 4. Lower epidermis of uppermost culm leaves of Puccinellia spp. Fig. 1. P. mari- tima from Blakeney; Fig. 2. ‘P. maritima var. hibernica’ from Strangforcl Lough; Fig. 3. P. distans from Heacham; Fig. 4. P. pseudodistans from Bowers ^Nlarsh, S. Essex. .411 X c. 500. SPECIES PROBLEMS IX PLANTS M^ITH REDUCED FLORAL STRUCTURE pungens x repens. It can be distinguished from the former by the thick ribs and completely smooth upper epidermis, and from the latter by the absence of abortive pollen grains, the completely smooth epidermis and the laxer spike with smaller spikelets. In Puccinellia non-floral characters of a rather different kind have been extensively used by Sprensen in his recent revision of the numerous Greenland species of the genus. He points out that “the rather featureless appearance of the spikelets of Puc¬ cinellia may render difficult an immediate recognition of the individual species” and goes on to say : “The shape of the panicle is of little value because the branches in certain species may be ascending or reflexed, according to the vigour and the develop¬ mental stage.” He has, however, used the structure and size of the epidermal cells and stomata, apparently with great success. The range of variation to be taken into consideration is reduced by comparing the epidermis of the uppermost culm leaves at maturity, and seems to be little, if at all, greater than that encountered in floral parts. In P. maritima, for example, some of the cells of the upper epidermis bear conspicuous blunt projections while the lower epidermis is smooth or nearly so. In P. distans on the other hand the lower epidermis has rows of pointed conical projections. In the light of this work it seemed of interest to look at the leaf epidermis of some British Puccinellia species. There are known to be in this country two entities in this genus about which there has been a good deal of uncertainty : P. pseudodistans, and the Irish plant which has been variously called P. festuciformis Praeger (non Hayne) and P. maritima var. hihernica Druce. The latter has a distinct appearance which attracted the attention of skilled field botanists such as Praeger and Druce, but is rather lacking in floral characters to distinguish it clearly from P. mari¬ tima. If, however, preparations of leaf epidermis of the two are compared abundant differences are evident. Plate I, fig. 1, represents the lower epidermis of P. maritima. Tlie guard cells here are 34-38y long and the stomata deeply sunk. The sur¬ rounding cells overlap the guard cells leaving a roughly crucifonn opening which is about 17/^ long and 24/^ wide. Plate I, fig. 2, shows a corresponding piece of the lower epidermis of what we can call, for the time being, P. maritima var. hihernica. The guard cells here are 54-56y long, about IJ times the length of those in P. maritima. The stomata themselves are only slightly below the surrounding epidermal cells, which leave an oval to sometimes almost circular pore giving access to the stoma. This pore is commonly 25-30y long by 17-20/x wide. There are, of course, other differences in the shape and size of epidermal cells, the sinuosity of their walls, their differentiation into long and short cells, and the proportion of stomata to other cells. This, com¬ bined with the difference in general appearance which has struck competent observers who have seen the two plants in the field. 24 SPECIES STTTDTES IN THE BRITISH FLORA seems to me to justify the recognition of two distinct species. Whether the Irish plant is the same as the southern European plant with which it has been identified or whether it will prove 'x) be an endemic remains to be seen. It is interesting to note n this connection that a variety of different chromosome num¬ bers have been recorded for P. maritima. From Sweden the number 2n = 56 is recorded, c. 60 from Portugal, 63 from England and 70 from Germany. Unless there are several errors among these counts there is a strong suggestion that apomixis occurs, at least in some strains, and this would help to explain the occur¬ rence of ‘small’ and perhaps rather local species. These might well differ in general, but indefinable, appearance and in charac¬ ters such as epidermal structure, without showing any major differences in reproductive parts. In comparing the epidermis of P. distans and P. pseudodistans we find differences at least as great. Plate I, fig. 3, shows the lower epidermis from the uppermost culm leaf of P. distans. The guard cells here are 30-35/x long, just as they are in P. pseudo¬ distans, but the pore is somewhat cruciform and considerably broader than long. The relatively short and broad epidermal cells are a striking feature of this species. In P. pseudodistans (Plate I, fig 4) the epidermal cells are long and narrow and interspersed with short cells each of which bears a blunt forward-directed protuberance. The epidermal cell immediately below the stoma, though a normal long cell, bears a similar protuberance which makes a flap overlapping the guard cells. The pore is oval and about twice as long as broad. The upper epidermis of these two species also shows interest¬ ing differences. The stomata of P. distans have a flap formed by a short narrow protuberance which overlaps the base of the guard cells obliquely; the corresponding structure in P. pseudo¬ distans is longer and lies symmetrically over the guard cells. Once again a variety of chromosome numbers is recorded for P. distans, but this time they form a polyploid series; 2n=14, 28, 42. It is of course possible that one of these numbers does in fact refer to P. pseudodistans, and, though on the face of it there is no evidence of apomixis, it is quite possible that the hexaploid, or even the tetraploid, may prove partially or completely agamo- spermous. There are a number of other problems among British grasses where epidermal and other non-floral characters may be of great use. For example the discrimination of Koeleria species always seems to be a difficult task, though we are fortunate in having only one common one. The question of the identity of the plant known as K. albescens might perhaps be solved by the use of these criteria, which may also prove helpful to anyone interested in the intricacies of the varied and puzzling forms to be met with on continental holidays. SPECIES PUOBEEMS IN PLANTS WITH PEPFCEP FLORAL STRCCTURE 25 The genus Salicornia is once more attracting the attention of a number of workers in this country. It presents problems to the taxonomist resembling in some ways those met with in the Gramineae though, owing to the great reduction of the leaves, species seem to be more difficult to delimit than in any grass genus. Most, if not all, experienced field botanists who have studied Salicornia at all, have been able to recognise subjectively far more species than they could define on paper. Obvious vegetative characters such as colour, branching, habit and the length, shape and number of internodes in the fertile branches seem often to provide the only means of recognis¬ ing the entities in this genus. It is to be presumed that all these featmes show a considerable amount of variation, but the extent of it needs to be studied in cultivation. Fortunately it is possible to grow Salicornia successfully in pots of ordinary garden soil. The plants develop satis¬ factorily if the soil is watered with a 3% salt solution at the time the seed is sown. There is some evidence that germination is better without salt, but the seedlings do not seem to grow pro¬ perly without it. It might well be profitable to examine carefully the epidermal and other anatomical characters of Salicornia species in the hope of getting help from them. Miss de Fraine published in 1913 the results of an investigation of the distribution of spirally thickened cells, but doubt has been cast on the value of this as a specific character and the whole question of anatomical features is in need of re-investigation on a wider basis. Finally, I should like to stray for a moment beyond the strict limits set by the title of this contribution and say a word or two about groups in which the floral structure is not reduced so much as remarkably simple and uniform. The genus Allium will illustrate the point. Here the flowers consist esssentially of 2 whorls of petaloid perianth segments, 2 whorls of stamens and a trilocular ovary with nearly always 2 ovules in each cell. Fortunately, the leaves are fairly varied and so are the underground parts and certain floral characters such as colour, so the recognition of species is usually not too difficult. It seems likely, however, that this large genus is not in fact a natural group but includes several divergent groups placed together because of their smell, their umbellate inflorescence and their flowers which are, after all, of the basic monocotyledonous pattern and not necessarily a guide to affinities. Non-floral structures suggest that this is so. For instance, different species show epigeal or hypogeal germination, some have rhizomes and no bulbs and among the others there is a great diversity of bulb structure. These are surely features which should be taken into account at the generic level as well as at the species level. 26 SPECIES STUDIES IN THE BRITISH FLORA REFERENCES. Armstrong, S. F., 1917, British grasses and their employment in agricul¬ ture. Cambridge. Babington, C. C., 1922, Manual of British Botany, 10th edition. London. De Fraine, E., 191-3, The anatomy of the genus Saiicornia, J. Linn. Soc., 41, 317-348. Jansen, P., 1951, Flora Neerlandica, 1, 2. Amsterdam. Sorensen, Thorvald, 1953, A revision of the Greenland species of Puccinellia Pari., Meddelelser om (honland 136, 1-180. Walters, S. M., 1949, Alchemilla vulgaris L. agg. in Britain. Watsonia, 1, 6-18. Dr. J. Heslop-Harrison suggested that in using vegetative charac¬ ters difficulties arose mainly when we had to rely on size. In such cases characters could only be found by statistical methods. Whenever it proved possible to find ‘‘either or” characters we should be able to distinguish units. Mr. J. D. Lovis said, concerning Professor Tutin’s examination of epidermal structure in Puccinellia pseudodistans, he would be most interested to know from where his material came; from Britain, Holland or the Mediterranean. Mr. Lovis went on to say that he had found that plants resulting from seeds of P. pseudodistans from Pitsea, Essex, given to him by Mr. Lousley, had a chromosome number of 2n = 2S, the same number as P. fasciculata. Regarding the three levels of poly¬ ploidy reported for P. distans, so far he had only been able to recover the hexaploid (2n = 42) from Britain. Experimental garden trials of P. pseudodistans and related species had been started at Leeds but he was not prepared to make any comment on these at the present time. Prof. Tutin replied that his material of P. pseudodistans also came from Mr. Lousley. STABILITY OF SOME SPECIFIC CHARACTERS AS SHOWN BY FOSSIL RECORDS 27 THE STABILITY OF SOME SPECIFIC CHARACTERS AS SHOWN BY FOSSIL RECORDS H. Hamshaw Thomas (Botany School, Cambridge) The concept of a species as a group of individual plants, dis¬ tinguishable from other groups in their genetical composition and usually recognisable by morphological characters, involves a time factor. According to our ideas of evolution the ancestors of forms, now recognisable as a distinct species, were part of an assemblage of plants which in the course of time also gave rise to types now regarded as forming other species, and probably to other forms which have since become extinct. Relatively small changes in genetic constitution seem to have resulted in differentiation on the species level, while greater or more frequent changes have given forms which we recognise as distinct genera. It would be of great interest to know something of the time taken in the differentiation of both species and genera. But although the quantity of plant material that has come down to us from the past is not inconsiderable, it is mainly in the form of fragments, usually isolated from each other. We can seldom reconstruct the original appearance of the whole plant of any extinct vascular type. From the study of the remains of pieces of stems, of leaves, seeds or fruits and pollen grains, we are, how¬ ever, often able to obtain reliable information about the main morphological features of many types related to plants living today. This enables us to identify genera, or to recognise generic differences, but we are seldom able to state the more important specific characters of any extinct form taken as a whole. The conditions necessary for the preservation of plants as fossils are of rare occurrence, and consequently we have little chance of seeing the way in which populations of organisms change during the passage of time. Palaeontologists are much more favourably placed when they study the evolutionary changes in vertebrate animals. But though the study of fossilised plants does not give much help in reconstructing the history of species, it does provide a very useful background to our studies of living forms. The last years have seen very great additions to our knowledge of the past history of the vascular plants, especially of the gymno- sperms and the ferns. The remains of ancient flowering plants have been studied in several parts of the world with results of considerable importance; this work is not very widely known though it carries considerable implications for those interested in angiosperm taxonomy. 28 SPFCTES STI'DTES IN THE BRITISH FLORA Most ot the botanists of my generation were taught to think that the evolutionary development of the angiosperms has been a very rapid process. This belief, which is still widely held, had a noticeable influence on the study of the group. It was based on two ideas, both without factual support, but resting on negative evidence: — (a) The flowering plants are thought to have orig¬ inated in the Cretaceous period. If this were true the widely different types now living must all have evolved in the course of about 100 million years, (b) That the flowering plants sprang from gymnospermous ancestors with groups of fertile leaves arranged in strobili ; the diverse forms of flowers seen today were thought to have evolved from these stroboli. Tlie fact that very few remains of plants looking like angio¬ sperms have yet been found in rocks older than those of the Cretaceous period does not prove that angiosperms did not exist at an earlier period. On the other hand, when we first find dicotyledonous remains in the Lower Cretaceous strata in Eng¬ land (Stopes, M., Phil. Trans. Boy. Soc. London, B203, 75, 1912) they are so diverse in anatomy that they can scarcely be regarded as having recently sprung from a common ancestor. It seems to me more probable that the early flowering plants were in existence for some fifty million years, or more, before they spread into the vegetation and areas in which they had some chance of being preserv^ed. The second idea is quite without support from historical evidence. If it were true, it would follow that very extensive evolutionary changes have taken place in the floral parts of all modem types. But this theory originated in the idealistic concepts of floral structure current in the early days of the 19th century, and there is no justification for applying these concepts to the problems of plant evolution. The difference between an idealistic morphology and an evolutionary mor¬ phology, to which the writer drew attention many years ago, has recently been ably demonstrated by Dr. Agnes Arber in her book on the Natural Philosophy of Plant Form. When one comes to examine all the fossil evidence which has been rapidly accumulating in recent years a very different picture is presented. Angiosperm evolution seems to have been as slow as in other groups of land plants; many types appear to have persisted over very long periods of time with comparatively little change in their morphological characters. The material available for study is very extensive. From rocks dating back some 90 millions of years we have the remains of the leaves of dicotyledonous plants, often found in abundance, whose shape and venation can be clearly seen. In some examples the cuticles of the leaves are also preser\"ed, and after suitable chemical treatment the details of the stomata and epidermal cells can be studied under the microscope. Many of these forms were collected and studied during the 19th century and it was realised that they could be closely matched with living forms. Early in the present century there was a strong feeling that the identifi¬ cations made by the earlier workers in this field could not be substantiated, but more recent investigations have shown that they were probably correct in many instances. This has come through the systematic study of the fruits and seeds of Tertiary plants. Clement Reid and his wife, who recently died, became interested in the many remains of fruits and seeds which could be found as fossils in the Tertiary and Quaternary beds of Britain. They got together a large herbarium of the seeds and fruits of living plants and found that it was possible to identify a con¬ siderable number of the fossil forms by matching them with the seeds and fruits of recent species. This work has been ably continued by Miss Chandler and other workers in England and on the Continent. It has led to the identification of a large number of the genera in the successive floras of the Tertiary period from the period of the London Clay onwards. In addition to the evidence from the leaves, the seeds and the fruits, we now have a considerable body of evidence derived from the study of the pollen which is preserved in layers of peat or fresh water clays. This helps to complete the picture of the earlier flowering plants. Well preserved remains of flowers are seldom found, but some remarkably perfect flowers, easily recognisable as derived from species of Quercus and Cinnamomum, were discovered some time ago preserved in amber, the fossilised resin of a tree, of Oligocene age. The remains of inflorescences and flowers have also been found at different places in the United States of America, but the details of their floral structure can seldom be made out. In many parts of the world pieces of petrified dicotyledonous wood can be found. Their structural features are often beauti¬ fully preserved, but in the past their identification has been a matter of great difficulty owing to the absence of collections of woods from living trees and shrubs with which they could be compared. Now that properly arranged and tabulated collections of the wood of almost all of the known trees of the world have been formed in England and America, we can look forward to the discovery of many new facts about the affinities of the trees of Cretaceous and Tertiary times. Taking the evidence as a whole, it would seem that the majority of the genera known to have existed in Lower Tertiary times are still living somewhere in the world today. In as far as comparison is possible the modern forms do not differ widely from their early representatives. From an examination of the species included in a genus of living plants, there seems some justification for the view that the leaves of a dicotyledon are the organs most likely to show varia¬ tion with the passage of time. If, then, the flowering plants have been evolving rapidly during the past one hundred million years, we should find great differences between the form and structure of modern leaves and that of the earlier forms of the same genus. .'50 Hl’KClKjj JSTUDlESj IN THE IJiUTitell EEOllA But in fact the changes in form and venation since tlie commence¬ ment of the Tertiary period are usually small, and many of the types whose cuticles have been studied show that in the Eocene period their stomata and epidermal cells very closely resembled those of the living species. The beautiful British material from Hampshire and the Isle of Wight has not recently received the critical study it deserves, but some years ago the late Miss Baji- dulska {Journ. Linn. Soc., Botany, 1924-28) showed that a number of leaves found in the Eocene beds at Bournemouth closely resembled living species of seve-ral genera of the Lauraceae in form, venation and cuticular structure. In recent years the critical examination of Tertiary leaves from different horizons has been actively pursued in Western North America by Prof. Chaney and his associates. They have collected considerable numbers of forms from different localities, and have mainly worked with the remains of leaves. The majority of these could be identified with modern genera, and some with species, now living in Mexico, Central America, China and the basin of the Mississippi. So much information of this kind has now been pub¬ lished, often accompanied by photographs of fossil and modern leaves side by side, that it is not possible to hold any longer the vdew that the leaves of dicotyledonous trees have varied rapidly with the passage of the years. It must be noticed, however, that the greater part of the evidence for the stability of characters comes from the study of trees or shrubs. Unfortunately we have little evidence from the remains of herbaceous plants, since these are seldom preserved as fossils. It may well be that evolutionary changes in herbs take place more quickly since their life cycle is completed so much more rapidly. We have one valuable piece of evidence which seems to support this view. The seeds of Stratiotes aloides have been collected from various localities in beds laid down in the Quaternary period, and comparable seeds have been found at a number of other horizons as far back as the Upper Eocene. Ail these forms were carefully studied and compared by Miss Chandler {Quart. Journ. Geol. Soc., 79, 117, 1923) who found that they showed a series of changes in form and structure. The changes suggested a main line of evolution resulting in the present-day form, together with one or two subsidiary lines. Unfortunately we have little or no evidence of the vegetative parts of the plants from which these seeds were derived. The fossil records of the ferns and the gymnosperms are similar to that of the angiosperms. Many forms seem to persist through very long periods of time with little change, wUile new forms appear at intervals. Wlien examined without morpho¬ logical preconceptions the new characters which appear are seen to be gradual modifications of ancestral characters rather than radical changes in structure. Through changes in climate and geographical conditions old types have died out and further changes have been favoured, but in general the evidence shows that plants of all groups possess great morphological stability. All this suggests to me that the flowering plants, as a group, must be very much older than has been usually thought, and that the reason why they scarcely appear in the geological record before Cretaceous times is that they originated, grew and developed on dry ground in areas where they could not be preserved as fossils, or in places from which the deposits in which they were buried have been removed by denudation. Prof. D. H. Valentine enquired whether pollen grains referable to Angiosperms had been found in deposits of Jurassic age. Dr. Thomas said that such pollen grains had been found in Jurassic deposits but there were only three or four records, and much more work was required. Dr. R. Melville asked how far back the records of Stratiotes extended and whether the lecturer could tell us of any sequences of leaf-form in Dicotyledons which went back any length of time. Dr. Thomas replied that the earliest records of Stratiotes were from the top of the Eocene. Nothing of value had been done on leaf sequence. The Americans have material suitable for work of this kind but have been interested in other aspects. Miss C. Longfield said that a dragonfly, an Aeschna, was entirely dependent on Stratiotes and a fossil dragonfly wing from the Eocene was almost identical. It was known that the Odonata date back to the Permian and an interesting link between plants and dragonflies was suggested. .SI’KCIES , STUDIES IN 'I'llE ItlllTiSH El.OKA ;^2 VARIABILITY WITHIN SPECIES J. H. Burnett (The University of Liverpool). Variation is an ubiquitous phenomenon in all organisms. It has been the especial concern of taxonomists to assess the relative importance given to variants in relation to classification. At present this is a particularly difficult activity to undertake for a number of reasons. The study of variation involves two inter¬ dependent and simultaneous investigations. On the one hand, it is necessary to determine which properties of an organism vary and in what ways ; on the other, to determine what are the causes of variation. An ancillary process, important for taxonomy, is the development of adequate techniques to describe the kinds and causes of variation. The application of the results of these investigations to taxonomy is difficult for three reasons. Variants differ in many ways. They are of different kinds, magnitudes, and stabilities. They may occur with different frequencies, or possess different distributions in assemblages of related individuals. The introduc¬ tion of new techniques has increased enormously the range of features which may be studied, and each new feature, so intro¬ duced, may vary in one or several of the ways described. Initially there is, therefore, the difficulty of appreciating and containing the great amount of information available. A second difficulty arises from progress m the study of the causes of variation. There has been a tendency for the description of unique, restricted patterns of variation to be replaced by the description of universal pro¬ cesses. This tendency can be demonstrated by a selection of statements on the relationship of variation to taxonomic practice over the last three hundred years. For instance, in 1674 John Ray wrote: — “Having observed that most herbarists, mistaking many accidents for notes of specific distinction, which indeed are not, have un¬ necessarily multiplied beings, contrary to that well known philosophic precept: I think it may not be unuseful, in order to the determining of the number of species more certainly and agreeably to nature, to enumerate such accidents and then give my reasons why I judge them not sufficient to infer a specific difference.” After this passage followed an enumeration of different kinds of variants which were regarded as unworthy of specific rank, e.g., floral abnormalities, changes consequent upon transplanta¬ tion of mountain forms to lowlands, and other conditions induced artificially. Canon Raven has remarked : “It is one of the out- VARIAIJILITY WITHIN SI’ECIKS 3:^ standing qualities of Ray’s achievement that he has refused to base specific differences on other than structural qualities.” This is true, but Ray also considered it necessary that forms should breed true within these limits : ''Distincta propagatio ex semine” Ray recognized that variation arose both from inherent, internal causes and from the influence of the external environment. The study of variation embraced both types but only that due to the former cause was of taxonomic significance. Two hundred years later Darwin was concerned with the same problem as Ray. He attempted to assess the significance of inherent variation in his theory of natural selection, thus : “I have called this principle, by which each slightest variation, if useful, is preserved, by the term of natural selection.” (Origin, p. 32). He described the application of this principle to taxonomic method in the last chapter of the Origin of Species : — “When the views advanced by me in this volume or by Mr. Wallace in the Linnean Journal or when analogous views on the origin of species are generally admitted, we can dimly foresee that there will be a considerable revolution in natural history. Systematists will be able to pursue their labours as at present, but they will not be incessantly haunted by the shadowy doubt whether this or that form be in essence a species. This, I feel sure, and I speak after experience, will be no slight relief. The endless disputes whether or not some fifty species of British Brambles are true species will cease. Systematists will have only to decide (not that this will be easy) whether any form be suffi¬ ciently constant and distinct from other forms, to be capable of definition and, if definable, whether the differences be sufficiently important to deserve a specific name .... It is quite possible that forms now generally acknowledged to be merely varieties may hereafter be thought worthy of specific names, as with the prim¬ rose and cowslip; and in this case scientific and popular language will come into accordance. In short, we shall have to treat species in the same manner as those naturalists treat genera, who admit that genera are merely artificial combinations made for con¬ venience. This may not be a cheering prospect, but at least we shall be freed from the vain search for the undiscovered and un- discoverable essence of the term species.” Read to-day, parts of this extract reveal a streak of the wildest optimism in Darwin, and most biologists will welcome the golden age of taxonomy, which he predicts, with no slight relief! Both Darwin and Ray were concerned with the delineation of species. Because Darwin attempted to account for the processes involved in the origin and maintenance of species he was led to regard species as relative rather than absolute units. The study of genetics in the twentieth century has resulted in an even greater emphasis on the nature of the processes in¬ volved in the development and maintenance of species.. For example, fifteen years ago it was possible to read in The New Systematics that — .‘M SI’KC'IKS STUDIKS IN TIIK UKITISII KLOKA “The only valid principles are those that we can derive, not from fixed classes but from changing processes. To do this we must go beyond the species to find out what it is made of. We must proceed (by collaboration) to examine its chromosome structure and system of reproduction in relation to its range of variation and ecological character. From them we can determine what is the genetic species of Ray, the unit of reproduction, a unit which cannot be used for summary diagnosis, but which can be used for discovering and relating the processes of variation and the principles of evolution.” (Darlington, 1940). Studies of the kind described in this extract evidently lead to different activities from those commonly engaged in by taxonomists! Divergent procedures with identical materials are likely to lead to confusion and controversy. It is not inevitable that they will lead to different conclusions. The third difficulty stems from the inadequate development of suitable techniques for describing variation, especially its frequency and distribution. This difficulty is a component of the first distinguished here, that derived from the great increase in the data on variation. No doubt this will be overcome, in time, by application and ingenuity. MUTATION 1 . Chromosomal a. Numerical b. Structural 2. Genic I BREEDING SYSTEM 1. Meiosis- Segregation 2. Fertilisation-Recombination 3. Inbreeding and Outbreeding GENOTYPE 7" Internal E N V I R O N M E N T External PHENOTYPE Pig. 1. The relationship of Genotype, Phenotype, the Environment and the processes which cause genetic diversity. VAlUAlJiLlTi' WITHIM 35 It is not certain that the second source of difficulty can be so readily resolved, for it arises from the application of different methodologies. As Ray realized, the appearance and properties of organisms are developed through the interplay of their inherit¬ ance and the environment. From fertilization, when the inherited potentialities come together, there is a constant interplay between them and with the environment outside them, until death. The nett result of this “fruitful struggle” is the phenotype. The origin and manifestation of variation of the phenotype are shown diagrammatically (Fig 1). Now, in the first instance, taxonomists are concerned to study phenotypes. Their final assessment of the taxonomic status oi an entity will be conditioned by the range of information they obtain concerning its phenotypic variation, not excluding any information available on the causes of this variation, genetic or otherwise. But those workers who adopt the principles, “derived from changing processes”, will be concerned, primarily, to classify the kinds of genetic diversity involved in the development of the entity. It is relevant, therefore, to consider the manifestations of phenotypes m relation both to genetic diversity and permanence. Such a consideration makes it possible to judge the kinds of result to be obtained by the study either of “fixed classes” or “changing processes”. Accordingly, a number of examples are set out to illustrate the ways in which different kinds of genetic diversity are manifested phenotypically. Save for hybridization, diversity which arises from different kinds of breeding systems has been omitted, since it has been reviewed recently (Baker, 1953; Gustafsson, 1946-47; Stebbins, 1950). The examples given in the tables are selected to illustrate: — 1. The magnitude of the phenotypic expression of the variation of the genotype. Variants detectable without the use of special techniques are classified as “Distinct”. Those whose detection requires special techniques, or those where the variation tends to be continuous, are distinguished as “Semi-cryptic” and those not readily distinguishable are classified as “Cryptic”. 2. The frequency or distribution of the variants, in space and / or time, whenever known. 3. The taxonomic status, if any, given to each variant. SI’K(;iK« STI’DIKS IN THK MRITISH FLORA U) (a) NUMERICAIr— 1>0LYPL01DY TABLE 1 PHENOTYPIC ; CHROMOSOME TAXONOMIC EXPRESSION PLANT NOS. DISTRIBUTION STATUS i lianunculus ficaria | 1 1 1 2n: 16 and Geogi-aphical and 32 i ecological ’ ; Subspp. u Eleocharis palustris 2n : 16 and 38 Geographical and ecological Subspp. M H Til P Monotiopa hypophegea M. hypopithys 2n: 16 2n : 48 ? ? Sp Sp. 1 o (xdlium jmlustre . 2ti: 24, 48iEcological and 96' t iVars. or subspp. H P s Q , 1 1 1 M Xfl 1 1 i 1 i 1 Valeriana officinalis 2n: (14), 28 Geographical and|Spp., snb- 1 and 56i ecological [ spp. and : I vars. i 1 ' i 1 M ' 1 1 i ^ Salix caprea \2n: 38, 76 1 P PS 1 1 I VAKIABn.ITV AVTTHIX SPF^CTKS CHROMOSOMAL VARIATION COMMENTS Distinguishable by several morphological characters which seem to be correlated with the cytological condition (Marsden- Jones and Turrill, 1952), They may possibly occupy different ecological niches (Turrill, 1948) and the tetraploid appears to be the predominant form in Sweden and (?) Denmark (Perje, 1952). Characters by which these forms may be distinguished have been given but the distinction is not always easy or possible to make. The two forms differ both in their ecological tolerances and in their geographical ranges (Walters, 1949). These plants are readily distinguished morphologically and have been found, recently, to differ cytologically (Hagerup, 1944 : Love & Love, 1944), It is assumed that the correlation between morphology and cytology is constant and presumably it is on this assumption that they are given specific rank in Clapham, Tutin & Warburg (1952). They are not known to show ecological or regional differences in distribution. The cytology of this species w’as described by Hancock (1942), Avho also described different ecological preferences for each cytotype. Earlier, the diploid and the octoploid had been recognized on morphological grounds as var. hinceolatxun and var. witheringii or var. angiistifoliwn , respectively. In its range of morphological expression the tetraploid overlaps both the others. Therefore, precise discrimination is not always possible. The cytological work on British and Polish material (Skalihska, 1947, 1950, 1951) and the taxonomic work on British and European plants (Walther, 1949; Sprague, 1952) demonstrate that in some parts of the range of this aggregate there is a correlation between morphological and cytological features. This does not always obtain in the British Isles and Western Europe. Wilkinson (1944) was unable to distinguish cytologically distinct material by morphological characters. The distribution of the two forms is unknown. 38 SPECIES STUDIES IN THE RRITISH FLORA TABLE 1 (b) NUMERICAL— DYSPLOIDY DISTINCT Viola riviniana 2n : 40, 46 and 47 Sporadic Forma ! SEMI-CRYPTIO 1 Erophila duplex 2n : 30-40 (?) Regional Forma or Spp. u 1 1 M H Iris pseudacorus 2n: 24, 32- , ? — >• 34 u (c) STRUCTURAL PHENOTYPIC TAXONOMIC EXPRESSION PLANT • DISTRIBUTION STATUS Trillium hamtscJi aticum 1 j Regional O M 1 i Da tura sfra m o niu m Geograpliieal H i ! PS o Paris qua.drifalia ? — VARIABILITY WITHIN SI’ECIES CHROMOSOMAL VARIATION (Continued) Oq Certain forms are distinguished by the possession of adventitious shoots. They are sporadic in occurrence and, so far, this morphological feature has always been found to he correlated with a dysploid number of chromosomes (Valentine, 1949). This aggregate group of cytologically heterogeneous plants shows con¬ siderable similarity although distinct forms occur which may have characteristic distributions. In general, there is not a high correlation between morphology, cytology and/or distribution (Winge, 1940). These chromosome numbers have been recorded but there is no evidence that morphological or distributional features are associated with the dysploidic condition (Ehrenberg, 1945). COMMENTS By studying the chromosomes of plants chilled before examination seg¬ ments which showed differential staining were detected. The frequency of these differential chromosome types was not the same in different populations (Haga & Kurabayshi, 1948). “Prime- types” occur with different pairs of chromosomes which have reciprocal translocations. These are only detectable by breeding and cytological experiments. The prime-types show different geographical distributions. In eastern U.S.A. several types occur whereas P.Ts 2 and 3 occur almost exclusively in Peru (Blakeslee, Bergner & Avery, 1937) . Populations from the Austrian Tirol possessed inversions in every chromosome arm. Morphologically these plants, and the species throughout its range, were very constant. How widespread this cytological condition is in P. quadrifoUa is unknown (Geitler, 1937 and 1938) . 40 SI*FX'IES STUDIES IN THE BRITISH FLORA TABLE 2 (a) AFFECTING SEVERAL CHARACTERS PHENOTYPIC TAXONOMIC EXPRESSION PLANT DISTRIBUTION STATUS Kumex ohtusifolius Geographical Subspp. H O H cc c Almis glutinosa Ecological Vars. Veronica anagallls- Sporadic Vars. U M aquatica e- 1^ f!? o s 1/} Laminin jyiirpiirenm Geographical & seasonal • — Deschainpsia cespi- Regional & tosa ecological o . H o Fomes pinicola Geographical & ( ?) ecological — VARTA15TLITY WITHIN SPECIES 41 GENIC VARIATION COMMENTS This species is wide ranging and four well marked subspecies, which differ both in their morphology and geographical ranges, have been described. Three of these have been recorded for the British Isles (Lousley, 1938). The habit, size and shape of leaf and catkin size are variable in this species. Certain combinations of these features show some degree of correlation and have been described as distinct varieties (Moss, 1914). But this variability can be described more precisely as clinal and can be correlated with a S.E.-N.W. climatic gradient in Britain (McVean, 1953). _ _ Small forms of this species with thicker leaves than are usual and densely glandular-hairy inflorescences have been described as distinct varieties under a multitude of different names. In Britain the majority of these forms are adaptive responses to environmental conditions but occasionally some forms are found wliich are genetically fixed in this habit. Miintzing (1932) described winter-annual and summer-annual forms of this species which are virtually genetically isolated and morphologically indistinguishable. The frequency of winter /summer annuals has been shown to vary in a more or less clinal manner from Scandinavia south¬ wards to Asia Minor (Bernstrom, 1953). This species is cytologically and morphologically uniform in N. America but physiologically (e.g. in vigour of growth, height, number of flower¬ ing stems, fruiting and survival after frost injury) very diverse. There is some correlation between certain of these physiological features and on this basis five groups can be experimentally distinguished (Lawrence, 1945). This fungus is widely distributed in N. America and Europe. It is rare in this country. There is some morphological variation in the fruit body on certain host trees. Erom their breeding behaviour isolates can be divided into three partly interfertile groups which are quite in¬ dependent of the morphologically variant groups. Two of the breeding groups in N. America show very little interfertility although individuals from each may grow side by side. One of the N. American groups is almost completely unable to cross with European forms, the other is (piite interfertile (Mounce & Macrae, 1938). SPECIES STUDIES IN THE BRITISH FLORA TABLE 2 (b) AFFECTING SINGLE CHARACTERS PHENOTYPIC TAXONOMIC EXPRESSION PLANT DISTRIBUTION STATUS O »-( H Painin sfiptirus Geographical OQ Q i 1 ! Plan faq o ari f 1 iiui 1 i ♦ Regional Var. I I SEMI-CRYPTIC 7) i q i fa L’.s p nrp urea ; Sporadic or (?) regional li'ici AMLS cam m ini is Ecological u H Oi Trifolium repens Ecological & — rH geographical O 1 1 VARIABILITY WITHIN SPECIES 43 GENIC VARIATION (Continued) COMMENTS This fungus is probably distributed throughout the world. All the N. American forms examined possess a phosphorescent mycelium; forms in Europe, including those in Britain, do not. The difference is due to a single pair of allelomorphic genes which apparently have different geographical distributions (Macrae, 1942). A single dominant gene determines the presence of small, deep wine-red spots on the leaves of this species. The frequency of occurrence of plants with spotted leaves varies in different populations and does not appear to be correlated with any other feature in this country. However, when populations from areas between E. America and Europe are compared, the average percentage frequency of plants with spotted leaves in the various sub-regions tends to follow a geographical sequence or topocline (Gregor, 1939). Two forms of this species occur in Britain. In one, the plant is uniformly pubescent, in the other, the leaves are less hairy and the stem is glabrous at the base but becomes hairy in the region of the inflor¬ escence. These conditions are due to a pair of allelomorphs, the latter condition being dominant to the former. Its distribution appears to be sporadic although the recessive condition is apparently more frequent (Saunders, 1918). Leaves with or without a waxy bloom are determined by a pair of allelomorphs lB:h. In Peru, coastal populations contained 0T5% plants with bloom and this percentage increased in populations at successively higher altitudes, viz. 585 ft. — 22-2% ; 1950 ft. — 30T % ; 2762 ft. — 50-0% ; 7764 ft. — 100-0%. There is evidently a correlation between ecological and climatic features and the presence or absence of a waxy bloom on the leaves i.e. variation is distributed clinally (Harland, 1947). The presence or absence of cyanogenetic glucosides in this species is determined by the allelomorphs Ac:ac. The frequency distribution of this gene pair has been studied in wild populations in the British Isles and in Europe as far as E. Russia and the Near East. There was a general tendency for the frequency of Ac to decrease in the direction Mediterranean Basin to N.E. Europe, and the distribution of the gene was apparently correlated with the January isotherms (Daday, 1954). 44 SI’KCIFS STrDIRS IX THK HRITISH FLORA TABLE 3 HYBRID COMBINATION. Veronica anaqallis-aquafica x F, catenata PHENOTA'PIC EXPRESSION o sc h-l H CO S Sffirrio (iquotirus x S. jarohaed os o M H Ph >> P3 o M H di O Snxifraqa hirsiita X .S', spaihvlaris {Symphytum officinale VAHIAIJILITY WITHIN SI’KCIKS VARIATION ARISING FROM HYBRIDISATION COMMENTS Hybrids between these species are quite sterile and are readily dis¬ tinguished by morphological features such as greater size, an increased number of flowers and irregular development of floral bracts. Hybrids between these species occur frequently in Ireland but are less common in this country. They are not fertile with either parent and a characteristic sight is a hybrid swarm in the area between a population of S. aquatu us on wet peat and one of 8. jacohaea on the well drained part (Praeger, 1951). Hybrid swarms are developed between these species in Ireland. In S.W. Ireland both parental and hybrid forms occur but in County Mayo only S. spafhularis and hybrids now occur (Webb, 1950). A number of variants are known with different flower colours. It has been suggested that this is due to a history of past hybridization (Turrill, 1948). These examples provide evidence that a particular kind of genetic process may show different kinds of phenotypic expres¬ sion in different species. Thus, for instance, the phenotypes of polyploids and their diploid progenitors may be distinguished in some cases by their morphology, ecology and distribution or, in the extreme condition, they may only be distinguished on cyto- logical examination. In general then, the possible array of phenotypes associated with a particular kind of genetic situation form a complete series in respect of their readily observable characters, a spectrum of manifestations. Most situations are more complex than those represented in the tables, which were selected, in part, for their simplicity. In¬ creased complexity may arise from several causes. There may be more than one kind of genetic mechanism giving rise to different kinds of variant. These processes may affect different characters in different or correlated ways. In addition, the kinds and causes of variability may be different in different parts of a species range. Some of the examples given illustrate these points. Erophila duplex, cited as an example of dysploidy, is a complex amongst a group of forms comprising a polyploid series (Winge, 1940). Superimposed upon the morphological variation correlated with dysploidy in Viola riviniana there is a wide range of morpho¬ logical expression associated with different ecological tolerances. These reach their extremes in subspecies nemorosa and subspecies minor (Valentine, 1941). The physiological variation of Fomes pinicola is independent of a particular kind of morphological development of the fungus found amongst some of the indivi¬ duals when they occur on coniferous trees. These morphological variants may belong to either of the physiologically dissimilar and partially inter-fertile races A and B. An even more complex situation in another thallophyte, Fucus, is described by Burrows & Lodge (see p. 83). The form of these plants is apparently readily modified by different environmental conditions but some of this variation is probably due also to hybridization. Moreover, different ecological behaviour is shown by some species in different parts of their geographical range, e.g. F. infiatus and F. vesiculosus. In the former species there is also a continuous gradation in the Faeroes between two forms which are separable both in morphology and ecological tolerance in the British Isles and Scandinavia. Variation also occurs in N. American species of Fucus but at present it cannot be correlated with that found in N.W. Europe because critical comparisons have not been made. Different patterns of behaviour are shown in different regions by flowering plants. Valeriana officinalis is an excellent example. In the British Isles there is a range of morphologically inter¬ grading forms superimposed upon a genetic background of tw'o polyploid variants. This condition also occurs on the western mainland of Europe but in Poland there are three morphologically and cytologically distinct forms. These forms overlap both phenotypically and cytologically with some of the forms which VAMIAMIJ.ITV WITHIN Sl’iaTKS occur in N.W. Europe (Skalinska, 1951). A more subtle form of regional differentiation is that which occurs in Lamium pur- pureum. This species is known to occur as both a summer-annual and a winter-annual form (Miintzing, 1932). More recently it has been shown that the proportion of summer-annuals increases in the more north-westerly populations while in Portugal, France, Switzerland, Italy, Syria and Central Europe, wdnter-annual forms predominate. There is also an increase in the degree of development of the winter-annual habit from N.W. to S.E., i.e. the most southerly summer-annuals approach the winter-annuals in habit and behaviour more than those from more northerly latitudes (Bernstrom, 1953). There is little doubt that the study of the variability of British species outside the United Kingdom has not been pro¬ secuted as vigorously as it might have been. A less parochial attitude in this matter is desirable. This is especially so in respect of the status, frequency and distribution of “minor” variants which are often dismissed as “not worthy of distinction by name” (Wilmott, 1949). A recent study of Trifolium repens illustrates the value of such an investigation. Twenty years ago, in a discussion on the taxonomic treatment of biological races, Turrill (1931) mentioned that some individuals of T. repens possess a cyanogenetic glucoside, others do not. The condition is due to a pair of allelomorphic genes. The study of the distri¬ bution of these two forms in wild populations was begun by the late R. D. Williams and has been completed and published by Daday (1954). The frequency of the occurrence of the glucoside- determining gene in populations is correlated with the January isotherms. In general, its frequency decreases from Asia Minor and W. Central Russia to N.W. Europe. There is evidence that similar trends occur in morphological characters e.g. length of the corolla tube, but this requires further study. It may be noted that this kind of pattern of variation can only be ade¬ quately represented as a gene-frequency map. The striking relationship between this “cryptic” character and an environ¬ mental factor could never have been detected by a study tJ British material alone. In quite a different way the data presented by Manton (see pp. 90-97) on the causes of variation in British, European and N. American ferns has contributed valuable information. Not only has it provided data on geo¬ graphical distribution and possible migrations but it has provided direct evidence for the immediate phylogeny of certain species. Such an unequivocal demonstration of natural relation¬ ships is obviously of value to taxonomists. It is convenient at this point to consider briefly phenotypic variation against a constant genetic background. Such varia¬ bility is shown by most plants and has been studied for a longer period than other forms of variation. Some species and races vary in this way more than others. Dr. F. H. Whitehead, of Oxford, has studied a particularly interesting case, Cerastiurn tetrandi'um , and has generously permitted me to quote some of his unpublished results. He has shown that, in response to different light intensities, different individuals from cloned material have shown variation which transcends that found in nature as a result of either genetic or environmental causes. The features studied included general habit, and relative xeromorphy of the leaves as measured by their hairiness and stomatal fre¬ quency per unit area. Thus Cerastium tetrandrum evidently possesses an extraordinary range of phenotypic plasticity which appears to be adaptive in relation to the environment. How widespread such behaviour may be is unknown. It is still true that “our knowledge of the plasticity of even common plants is verj^ meagre and even facts already known are sometimes ignored” (Turrill, 1938). There is some evidence that most of the modifications which are apparently adaptive to environmental conditions are due to variation and selection of the genotype. This adaptive response may be manifested in at least two ways. Either, distinct geno¬ types may be found in different environments, e.g., in Achillea borealis and A. lanulosa (Clausen, Keck and Hiesey, 1948 : Hiesey, 1953), or, the same genotypes may occur with different frequencies in different environments, e.g., Plantago maritima (Gregor, 1946). Doubtless intermediate conditions between these extremes will be found. Another very frequent condition of some interest is the occurrence of phenotypically similar individuals whose appearance is determined in different ways. In some the growth form is inherent and invariable, in others it is an adaptive response to the environmental conditions. An example of this situation is known in Veronica anagallis-aquatica. Individuals in exposed, well- drained situations often have short, stout stems, a small leaf area and highly glandular-pubescent inflorescences. Usually such individuals become taller, develop larger leaves and show a reduc¬ tion of pubescence when transplanted to a medium loam soil in protected pot cultures. However, a plant found on the Kenfig dunes and sent to me by Mr. B. L. Burtt has now maintained its original characters for two generations of pot culture. Its reduced, pubescent habit is genotypically conditioned. It may be sug¬ gested that careful observation of such cases might provide evidence for, or against, the “Baldwin effect” recently described by Simpson (1953). This is a condition simulating the inheritance of acquired characters. Whether or not this will be the case, it is certain that more precise studies of the nature and extent of phenotypic plasticity are necessary. Preliminary investigations using the transplant technique will give useful results, and can be carried out by amateur and professional alike, but more precise experimentation is also desirable. This rapid survey of the causes and manifestations of varia¬ tion in plants reveals one feature with great clarity. This is, that variation is rarely manifested in a discontinuous manner. Even VAIUAIUMTV WITHIN fcjl'ECJiiS 49 when a particular genetic process gives rise to discrete classes of individuals, e.g., diploid v. tetraploid, allelomorph A v. allelomorph B, their phenotypic expression may or may not reflect this distinc¬ tion. Moreover, in many cases where the alternative genetic determinants and their phenotypes are distinguishable, e.g., presence v. absence of glucoside in Trifolium repens, the frequency distribution of the characters varies continuously over the range of the species. This feature, the continuous manifestation of varia¬ tion, is one of the main reasons why variability, especially at the infra-specific level, is difficult to describe with precision. It is now desirable to assess the taxonomic use of knowledge of the causes of variation and the kinds and patterns of its pheno¬ typic manifestations. The great attraction of an approach which classifies variation by processes rather than by their products is that the former are relatively few in number and distinct from each other. The gejnetical behaviour consequent upon them is now quite well known and the results of any process can be predicted with some certainty. In this way the indescribable phenotypic chaos can be avoided and the material reduced to comparative orderliness. For example, most of the British species of Taraxacum are probably triploid or tetraploid and apomictic. Some non-British forms are regularly amphimictic, some occasionally amphimictic and many more apomictic. Different degrees of polyploidy also occur. This information effectively summarises the genetic situa¬ tion. It does not assist the identification or the ordering of the many hundreds of distinct morphological forms that exist. Turrill (1937) has written: “From the standpoint of genetic isolation the biotypes are species; from the standpoint of character-combina¬ tions, morphological differences and similarities, and ecology, the bio types, within any one section and within any one distributional area at least, are varieties of one or a few species”. The genetic information demonstrates how the baby has been delivered but it leaves the taxonomist still holding the baby ! This is not always the case. The work on polyploidy in ferns, described by Manton, has provided evidence of species relationship and so has assisted classification directly. More striking instances are the grouping of the numerous forms within Viola section Melanium into poly¬ ploid groups (Clausen, 1931) or the separation of Veronica and Hehe on cytological as well as morphological grounds (Frankel, 1941). In a similar manner observations on the correlation be¬ tween gene frequencies and particular environmental factors, e.g., in Ricinus communis (vide Table 2), demonstrate the importance of such factors for both the past and the future of the species. But they do not greatly affect the taxonomic treatment of the species. This does not mean that such studies are irrelevant for taxonomy. One of the most desirable and difficult accomplish¬ ments which a taxonomist can acquire is to so develop his judg¬ ment or intuition that he can determine the “taxonomic value” of any particular character. In this matter the study of the causes 50 SI’KCIKS STLDIK.S IN THE inU'I'lSH KE()|{ A of variation may be of inestimable, objective value. If it is neces¬ sary to select some character for taxonomic weighting, it is desir¬ able that it should either be important for the structure and future of the species, or that it should be correlated with such a charac¬ ter. This kind of information is available from studies on the processes involved in speciation. The apparent conflict, or rather, lack of common objective, between these two methods of studying variation is to be resolved by a consideration of their respective aims. This is a matter to which attention has been drawn by Gilmour (see pp. 173-176). Here it is sufficient to compare two statements on the aims of taxonomy. The first is due to Huxley (1940). “Fundamentally, the problem of systematics, regarded as a branch of general biology, is that of detecting evolution at work”. The second is adopted from Mayr (1942). “The field of activity of the systema- tist can perhaps be subdivided under three headings: identifica¬ tion (analytical stage) . . . : classification (synthetic stage) . . . : study of species formation and the factors of evolution”. For Mayr the “basic task of the systematist” is identification. To which Huxley might be expected to retort : “But identification of what?” So far as variability within species is concerned, it seems that little can be gained by accepting either attitude to the exclusion of the other. At the infra-specific level, variation is to be observed in its most protean guise. For the constant flux and reflux of genotype and environment either has not yet resulted in the stabilisation of new specific patterns or else they are maintained in a dynamic equilibrium. The more stabilisation sets in, the more readily the situation becomes amenable to description in terms of fixed classes, i.e. to taxonomic treatment. All other situa¬ tion must be described, as precisely as possible, in terms both of the processes involved and their consequences and manifestations. The interest of these situations is more evolutionary than taxonomic. It would be a matter of surprise if a foetus could be described by the same terminology and judged by the same yard¬ stick as would be employed to describe it when it had developed into a man in the prime of life. So it is within the species. Just as there is some connection and correlation between adult and foetus, so there is between species and infra-specific entities. Tlie first of each of these are assessed according to their achievements and status, the second are of interest for their development and potentialities. It is some encouragement, perhaps, to realize that the problem of describing natural phenomena with accuracy is of long stand¬ ing. Mrs. Arber has discussed Plato’s views on scientific method and has remarked : “few would dispute that . . . Plato had seized the essential fact that any scientific system of explanation has a certain static finality, and hence must be imperfectly compatible with the unceasing flux of Nature”. (Arber, 1954). VAUIAJHLITV WITHIN Sl'EUlES 51 It is our privilege and our pleasure to improve taxonomy so that it will become as nearly compatible with nature as ever human ingenuity can devise. Summary The study of variation by taxonomists is difficult at present for three reasons: — 1. The numbers and kinds of characters which vary are large and increase with every application of new obser¬ vational techniques. 2. There is an apparent conflict between those who seek to describe and define classes of variants and those who wish to classify variation in terms of the processes which cause it. 3. Techniques of description are not yet sufficiently com¬ prehensive or precise to summarize the different mani¬ festations of variation adequately. Examples are given, in tabular form, of variations due to different genetic processes. The phenotypic manifestation, frequency or distribution and taxonomic status of each variant is given. It is evident that particular genetic processes may be mani¬ fested in different ways in different species. In widespread species the kinds of genetic diversity as well as the phenotypes may be different in different regions. Attention is drawn to the desirability of studying variation in British species outside the United Kingdom as well as within it. Variation due to pheno¬ typic plasticity is briefly discussed. Adaptive variation may be due either to phenotypic plasticity or genetic diversity. It is observed that variation is rarely discontinuous in its expression. This is one of the reasons which make it difficult to classify. The relative value of the study of the kinds, causes and de¬ scriptions of variation for taxonomy are discussed. I acknowledge critical reading of manuscripts by Professor A. Burges during the preparation of this paper. References Arber, a., 1954, The mind and the eye. Cambridge. Baker, H. G., 1953, Race formation and reproductive methods in flowering plants, 8.E.B. Symposia, 7, 114-145. Bernstrom, P., 1953, Cytogenetic intraspecific studies in Lamium II, Hereditas, 39, 381-437. Blakeslee, a. F., Bergner, A. D. & Avery, A. G., 1937, Geographical distribution of chromosomal prime types in Datura stra¬ monium, Cytologia, Fujii Jubilee Vol., 1070-1093. Clapham, a. R., Tutin, T. G. & Warburg, E. F., 1952, Flora of the British Isles. Cambridge. Clausen, J., 1931, Cytogenetic and taxonomic investigations on Melanium Violets, Hereditas, 15, 219-308. UKIVtKSAV Of ILUmOk*' UBRA«V 52 SI’IOCIKS STUDIES IN TllK 15KIT1SH FUOilA Clausen, J., Keck, D. D. & Hiesey, W. M., 1948, Experimental studies on the nature of species III — Environmental responses of climatic races of Achillea, Carnegie Inst. Washington , Pnbl. No. 581. Daday, H., 1954, Gene frequencies in wild populations of Trifolium repens I — Distribution by geography. Heredity, 8, 61-78. Darlington, C. D., 1940, Taxonomic species and genetic systems, in The New Systematics, ed. J. S. Huxley, London. Darwin, C., 1859, The Origin of Species, ed. 1. London. Ehrenberg, L., 1945, Kromosomtalen hos nagra kiiilvaxtcr, Bot. Notiser, 1945, 430-437. Erankel, O., 1941, Cytology and taxonomy of Veronica, Hebe and Pygmaea, Nature, 147, 117. Geitler, L., 1937, Cytogenetische Untersuchungen an naturlichen Popu- lationen von Paris quadrifolia, Zeitschr. Indukt. Ahstanun. & Vererhungsl., 73, 182-197. - , 1938, Weitere cytogenetische Untersuchungen an naturlichen Populationen von Paris quadrifolia, op. cit., 75, 161-190. Gregor, J. W., 1939, Experimental taxonomy IV — Population differen¬ tiation in North American and European sea plantains allied to Plantago maritima L., New Phyt., 38, 293-322. - , 1946, Ecotypic differentiation, New Phyt., 45, 254-270. Gustafsson, a., 1946-47, Apomixis in Higher Plants 1-3, Lunds Univ. Arsskr., (N.f.), 42/3; 43/2; 43/12. Haga, T, & Kurabayshi, M,, 1948, Chromosomal variation in Trillium kamtschaticum, Jap. J. Genet., _23, 12-13. Summarized in, Kihara, H., 1950, Japanese papers on genetics, 1942- 1948. English abstracts — Heredity, 4, 391-408. Hagerup, 0., 1944, Notes on some Boreal polyploids, Hereditas, 30, 152-160. Hancock, B. L., 1942, Cytological and ecological notes on some species of Galium L. emend. Scop., New Phyt., 41, 70-78. Harland, S. C., 1947, An alteration of gene frequency in Ricinus com¬ munis L. due to climatic conditions. Heredity, 1, 121-125. Hiesey, W. M., 1953, Comparative growth between and within climatic races of Achillea under controlled conditions. Evolution, 7, 297-316. Huxley, J. S., 1940, Introductory: Towards the New Systematics, in The New Systematics, ed. J. Huxley, London. J.,AWRENCE, W. E., 1945, Some ecotypic relations of Deschampsia caes- pitosa, Amer. J . Bot., 32, 298-314. Lousley, j. E. 1938, Notes on British Rumices, I., Hep. Bot. Soc. <£• E.C., 12, 118-157. Love, A. & Love, D., 1944, Cytotaxonomical studies on Boreal plants III — Some new chromosome numbers of Scandinavian plants, Arhiv f. Bot., 31A, 12. Macrae, ll., 1942, Interfertility studies and inheritance of luminosity in Panus stipticus, Canad. J. Research, 20, C, 411-434. McVean, D. N., 1953, Regional variation of Alnus glutinosa (L.) Gaertn. in Britain, Watsonia, 3, 6-32. VARIABILITY WITHIN SPECIES 5.^ Marsden- Jones, E. M. & Titrrill, W. B., 1952, Studies on Ranun¬ culus ficaria, J. Genetics, 50, 522-534. Mayr, E., 1942, Systeinatics and the origin of species. New York. Moss, C. E., 1914, Cambridge Briti.-. (). IJlmus glabra x plotii. Leaf spectium of a i-epresentative short shoot of a nothomorph srowing near HolsteacL Leicestershire (Melville no. 37.8.3). Fiff. 7. IJimus glabra x pJoUi. Leaf spectruin of a sliort shoot with leaf shapes tending towards U. plotii, from the same tree as Fig. 5 (Melville 37.83) Fig. 8. Ulmus X elegantissima Horwood {U. glabra x plotii). Leaf spectrum of a short shoot with shapes closely resembling those of U. plotii. Tree from the type area near Launde, Leicestershire (Melville no. 36.261). MOlU’HOJ.OdlCAL CUAllACTlIKS OF Sl’EC'IFS AND IIVDIUUS Gi. but the apical cusp is not so well developed and the basal lobe does not overlap the petiole, dhere are other more subtle differ¬ ences also. Notice the relatively wide spacing of the lateral nerves of the distal leaf. The variation between short shoots is greater than in the species. The leaves of a second short shoot from the same tree are shown in Fig. 7. Here the distal and sub- distal leaves resemble more closely those of U . plotii, though the apical cusp is more strongly developed and the spacing of the lateral nerves is intermediate. The leaves towards the base of the shoot resemble the corresponding leaves of U. glabra more closely so that in this shoot there is a change in the balance of growth factors from one parent to another — from U. glabra to U. plotii. On looking back at the first shoot (Fig. 6) it can be seen that a change in the balance between the parents has taken place there also. It is in the same direction, but has not gone so far. There is a fluctuating balance between the two parents in the hybrid which changes from shoot to shoot in the adult and also changes during the ontogenetic sequence. The phenomenon of fluctuating balance between alternative parental characters is widespread in hybrids. In segregates and back crosses the mean balance point may verge toward one or other of the parents. For example, another segregate from U. glabra x plotii (Fig. 8), closely matching Horwood’s type of U. X elegantissima, approaches to U. plotii in its leaf shape. The spectrum of the short shoot illustrated suggests, even, some of the leaves of the proliferating shoots of U. plotii and the nerva- ture is very like that species. Nevertheless, the habit, brandling and bud characters were closest to U. glabra. Some other shoots had leaf spectra close to that shown in Fig. 7, but with the sequence finishing at the subdistal leaf. Such was the maximum expression of U. glabra in the leaves of this individual. As another example of a hybrid segregate closely resembling in fohage one of its parents, I illustrate a roadside tree (Fig. 9) growing near Hitchin in Hertfordshire. Any novice gathering this could be pardoned for identifying it as a form of U. carpini- folia. The habit was erect and moderately spreading, but with rather long slender pendulous branchlets. The latter and the dark green pigmentation of the leaves were probably inherited from U. plotii. That U. plotii contributed to its inheritance was showm by the occurrence on the epicormic shoots of leaves almost identical in shape and nervation with those of U. plotii. One is shown at the left of the figure; insect damage prevented my obtaining a complete leaf spectrum of the epicormic shoots. The number of lateral nerves in the distal and subdistal leaves of the normal short shoots is larger than was to be expected. Other studies of hybrid leaves indicate that the number is usually inter¬ mediate between those of the parents. In this tree, there are generally one or two extra pairs above that usual for XJ. carpini- jolia and four or five more than in U. plotii. For this reason, I ()2 M'Kl IKS S'l'KDJKS i.N TIIK ItIUTiSII K1>()MA Fig. y. Uirnus carpinifuiia x piolii {X(jlabra). Li:af spectrum ot a represen¬ tative short shoot from a tiee growing near Hitchin, Hertfordshire, with, at left, a leaf from an epicormic shoot having leaf shape and nervation similar to U. plotii (Melville no. 51.5). Fig. lu. Uirnus carpinifoliu x piolii. Leaf spectrum of a representative slioj-t shoot from a tree (Melville no. 39.79) growing at Norton Bridge, Staf¬ fordshire. Fig. 11. Uirnus cut pinijoiiu x piolii. Leaf .spectrum fioni a short shoot with symmetrical leaves, same tree as Fig. 9 (Melville no. 39-79). MOlll'iiOI.OGiC'AJi CilAKACTKKS OF SL’JiClKJS AND ilVBllJ 1)S (3-3 suspect that U. glahra contributed to its inheritance. Two neighbouring trees showed definite evidence of this triple parent¬ age. Apart from its parentage, the tree is interesting as an example of the change of balance with the developmental phase. Early in the ontogenetic sequence U. plotii is subdominant, but U. carpinifolia is virtually dominant in the adult. As a final example I will take another individual U. carpini- jolia X plotii, a tree approaching U. carpinifolia in habit, growing near Norton Bridge m Staffordshire. The leaf spectrum of a characteristic short shoot is depicted in Fig. 10. These leaves approach nearest to U. carpinifolia in shape, but the lateral nerves are rather too widely spaced, and the third leaf from the right in its upper half closely resembles the comparable portion of the corresponding leaf in U. plotii. In addition to the asymmetrical¬ leaved short shoots, smaller numbers of others occur with sym¬ metrical leaves. Both kinds are found in the adult phase, inter¬ mingled on the normal branches. Comparison of the leaves of the parent species suggests that the four upper leaves of this shoot (Fig. 11) consist of two short sides of an asymmetrical leaf juxtaposed. Notice that the petioles are longer than in ordinary leaves, as would be expected if a basal lobe had been suppressed. A greater petiole length in such leaves has been confirmed by numerous measurements. The distal and subdistal leaves are built on the plan of TJ. plotii short sides. At the third leaf, the balance swings over to U. carpinifolia and the basal leaf is best explained as consisting of two long sides of the corresponding leaf of U. carpinifolia. There is evidently a complex interaction here between the parent species, a fuller explanation of which must await another occasion Perhaps sufficient has been said to indicate both the com¬ plexity and diversity of leaf form that occurs in elm hybrids. The variation is greater than is found in individuals of species. More¬ over it contains discordant elements. The variation in species is more circumscribed and harmonious. The two kinds of varia¬ tion can be called concordant and discordant. Critical study of the discordant variation in the leaf spectra of the hybrids makes it possible to determine the parentage. Dr. H. G. Baker said that it was evident that there had been exten¬ sive hybridisation in TJlrnus and it seemed that the hybrids were extremely fertile. He asked whether Dr. Melville recognised subspecific or ecotypic variation in the group. Dr. Melville replied that eight species were recognised in the new Flora, and there was evidence of another in East Anglia which had been almost hybridised out of existence. Ulmus glahra and U . procera have distinct varieties, but he did not recognise any subspecies. There was a tremendous number of hybrid individuals and there were indications that they are usually fertile. (34 Sl'Kt'lKS STUDIES IN THE lUllTiSll EDOllA Prof. D. H. Valentine asked whether all or some of the hybrids were of the Fj generation, as one would expect the F, plants to show con¬ cordant rather than discordant variation. Dr. Melville said that studies of garden hybrids had revealed interesting examples of discordant variation as, for example, in Bihes odoratuni x B. sangulneuin, and it did not necessarily follow that concordant variation would be found in Fi plants. Mr. S. Walker enquired whether any cytological or genetical work had been done on Ulmus. Dr. Melville, in reply, pointed out that breeding experiments were hardly practical with Uhnus as it would take more than a lifetime to obtain useful results. It was therefore necessary to rely mainly on morphological characters. l>llOJ{l>KMS ASSOCIATKI) WiTJl TIIK SrEClKS CONCEI'T IN MYCOLOGY 05 PROBLEMS ASSOCIATED WITH THE SPECIES CONCEPT IN MYCOLOGY Alan Burges (University of Liverpool). iSysteinatic and Taxonomic Botany has developed during the last two to three hundred years a fairly satisfactory system of delimiting species, despite all the controversies concerning the (Species Concept and the value of the Type Method. However, it cannot be emphasised too strongly that this success has been primarily in the study of higher plants. It has been a great loss to systematic botany that, particularly in academic circles, it was for so long regarded as synonymous with the classification of the angiosperms. “To know one’s plants” was to be able to identify the seed plants and ferns. To look at Bryophytes, Lichens or Fungi was indecent. It was something that the mycologists, or such people, did, but it wasn’t systematic Botany! It is not surprising therefore that the problems con¬ fronting the mycologists and workers in comparable groups were seldom included in discussions on systematics. When attempts are made to apply the general concepts and methods current in angiosperm systematics to the fungi, many difficulties arise. From classical times onwards fungi have been regarded generally as “a treacherous and mutable tribe”. Their capacity to vary is notorious and is largely responsible for the difficulties associated with their classification. Some of the variability is associated with the growth form of fungi and the shortness of their genera¬ tion time, but a large part comes from their possession of evolu¬ tionary and nuclear mechanisms not found in other groups. The difficulty of preparing good herbarium specimens, especially in the Agarics, is a further hindrance. We may represent the gradual evolution of species diagram- matically, as in Fig. 12. This is, of course, open to many objections, but even if we complicate and refine the picture by considering convergent evolution and reticulate systems or intermittent periods of speciation, this would not affect tlie present argument. The line A may be regarded as an instantaneous observation.* Of the various species observed, A^, A2, A3, A^, etc. are all good species, distinct from each other, although A2 might be regarded as very “plastic”. A. and A^ on the other hand would be difficult, the so-called “critical” species. If we were to observe Aj, there would probably be a considerable change between X and Y, but this would be a gradual change. *J believe Professor A. R. Claphain of Sheffield has used a similar idea to this in his “time slice’’. sTUDiKS IN Till-; ijuiit.sii floka GC JJiagiaiii showing- the i-elatioii between the constancy of species and tJie rale of evolutu n. (For full explanation, see text.) The situation given by our observation A probably represents fairly well the situation in any angiosperm systematics. The slow change from X to Y is not detectable in the short period (300 years) of systematic botany. In comparison with the life of an angiosperm species the life of the systematist is brief. If now we observe not just at instant A but continuously from A to B, the position becomes A^astly more complicated. Many species show continued and irreversible variation and the proportion of “difficult” to “good” species rises alarmingly. The important thing about the time-scale is not the time itself but the rate of change of the species. If this is high, then the period A to B may well fall within the lifetime of an observer. The rate of the change is governed by many factors, but it would seem that the generation time of the individuals in the species, and the nature of the evolutionary mechanisms, are two of the more important. It is only when progeny are tested in the environment that selection can act. Thus, although a bacterium with a generation time of 20 minutes and a tree with a generation time of 20 years may both have the same mutation rate per generation, the rates of evolutionary change are potenti¬ ally vastly different. Most fungi have a short generation time, particularly if vege¬ tative reproduction is considered, and the unfortunate mycologist is often faced with a situation resembling the internal from A to B. Most of our knowledge of rapid change in fungi comes from laboratory studies of organisms in culture. Rapid change is. I'ltOliLiiMS A«SO(JiATliI) WITH THE SL’ECiES CONCKL’T JN MiCOliOUV (37 however, known also in the field. Waterhouse (1952) has shown that in the rusts new physiological races can appear and rapidly displace the older races, particularly where the new race is able to attack a previously immune host. Race 34 of Fuccinia graminis tritici first appeared in West Australia in 1926, it spread rapidly through the wheat growing areas of Australia and New Zealand and by 1929 had completely supplanted the six previously occur¬ ring races. A similar situation arose in 1941, when a new race, r. 126, a close relative of r. 34, in turn replaced r. 34, its possible parent. A third major change occurred in 1948. A single speci¬ men from Queensland of a wheat variety “Yalta” previously resistant, was found to be heavily infected by a new race, r. 222. In 1949 it was identified in 5 out of the 122 field collections, 4 from Queensland and 1 from New South Wales. In 1950 the new race formed 512 out of the 704 isolations and was widespread in Queensland, New South Wales, Victoria and Tasmania. The new races apparently arise as mutants which are rapidly propagated by uredospores, successive generations following each other at 10-14 day intervals. The old races are at a disadvantage owing to host resistance and the new races, free from serious competition, increase rapidly and soon become epidemic. It is difficult to compare such changes directl3^ with those in flowering plants. A comparison between a generation of time of 20 mins, and 20 years gives a factor of approximately 500,000, but it is highly improbable that a bacterium or a yeast would maintain such a high rate of division for long, and it might be necessary to take a figure, of something like 10,000. Thus, on these grounds alone, one might anticipate seeing changes in a yeast or a bacterium during an experimental study, comparable to those occurring in a tree species during the course of some 10,000 years. In addition, in the fungi, the selective pressure can be very high, particularly where a new mutant attacks a pre¬ viously resistant host. Such high selective pressures lead to rapid changes in the populations and are probably rare in flowering plants. One might perhaps regard the rapid success of Spartina townsendii in colonising a previously “resistant” habitat as a comparable example. When the general evolution of fungal species is considered, it is difficult to assess the importance of rapid selection due to a short generation time. It may well be that it has played only a small part in the general picture but under certain circum¬ stances, as in the case of the wheat rusts, or in organisms like yeasts or bacteria, it provides a means by which changes in the population may occur with surprising speed. Sources of Variation in Fungi. Fungi have long been noted for their variability. This varia¬ tion may take many forms. There is first the variation found in any particular character, such as spore size, and graphs of the distribution of size about a mean are familiar to most workers in 08 SI’KCIKS STUJ)IKS IN THK BRITISH FJ-ORA systematic mycology. The degree of dispersion about the mean for different characters may differ considerably as in Fig. 13a, character A being less variable than B. Then there is the varia¬ tion brought about by the change in position of the mean for a character or group of characters as in Fig. 13b. This change may be a temporary one due to a change in environment, or a permanent one due to some form of mutation. This is, of course, the situation in all organisms. Fig. 13a. A comparison between the freciuency diagram of a character — (A) which is fairly constant, and (B) which is more variable. Fig. I3b. The shirt of frequency curve for a given character or species. This may be temporary due to transplanting to a new environment or permanent due to mutation. In delimiting a species we would naturally prefer to base a diagnosis on characters which vary least. The work of Vuillemin (1910) and of Mason (1933) has shown that different types of spores have different origins. Measurements show that conidia which are associated with a nuclear division (Conidia vera), or spores such as ascospores, vary far less in their measurements than do spores such as phragmospores which are essentially detached fragments of hyphae. The ease of specific or generic discrimina¬ tion is often closely associated with the spore type available. This is particularly true in the Fungi Imperfecti. Phenotypic Plasticity. The organism as we see it (phenotype) is the outcome of the interaction between the genetic make-up of the organism and its environment. Fungi are as a whole extremely tolerant of big changes in their environment and phenotypic plasticity is often very marked. Many fungi will thrive over a pH range from 3-9 and survive a multitude of changes in carbon and nitrogen sources. Under such conditions, colour, growth-form, and even spore size are very susceptible to change. The change may even be of the order which has frequently been given generic status. The genera Gloeosporium and Colletotrichum are closely allied. PROBLEMS ASSOCIATED WITH THE SPECIES CONCEPT IN MYCOLOGY 69 The latter has a zone of setae surrounding the acervnlus but there are no setae in Gloeosporium. Pure culture experiments have shown that one fungus, usually known as Collet otrichum gloeos- porioides, which is widespread in the tropics, is capable of infect¬ ing 40 different host species. The resulting infections had been attributed to 25 different species in two genera. On mango, if the fruit were infected, no setae were formed and the fungus was a typical Gloeosporium ; on stems and leaves setae occurred and the fungus was regarded as a Collet otrichum (Bessey, 1950, p. 552). Usually the variations produced by transferring the fungi to different media are edaphic only, and correspond to variations seen in transplant experiments. Return to the original substrate restores the organism to its original form. Mutations. Mutational changes in fungi are well known. The intensive studies of Neurospora include many dealing with gene mutations and the subsequent morphological changes which may occur. Striking changes due to modification of the cytoplasm are also known, from the work of Ephrussi (1952) and his colleagues. When large numbers of colonies of yeast were grown, it was found that some of them had a very much reduced growth rate and an altered physiology. On normal media the proportion of new types produced was small, but the addition of acriflavine to the medium gave almost 100% conversion to the new form. It was suggested that during the normal process of budding, occa¬ sional cells were formed in which the cytoplasm going into the bud lacked some of the cytoplasmic components and the deficient cells thus formed gave rise to the new type colonies. Addition of acriflavine inhibited the reproduction of the cytoplasmic com¬ ponent without preventing cell division, thus almost all the new cells produced lacked the component. Genic or cytoplasmic mutations are represented by a permanent shift of the mean in figure 13b. The rapidity with which a fungus may undergo a permanent change is at least partly dependent on the ease with which a mutant can be perpetuated. A higher plant has many millions of nuclei so that we can reasonably expect that in any particular individual there are a large number of mutated genes but, unless these mutants occur in a germ cell or in tissue which will give rise to germ cells, the mutants will be lost. Since the proportion of germ to vegetative cells is very low, only a minute fraction of the mutants has any hope of survival. Occasionally a vegetative mutant may arise, but its chance of propagation under natural conditions is small. In a fungus the situation is very different. A culture in a petri dish will, like the higher plant, have many nuclei and numerous mutated genes but, unlike the higher plant, these mutations, unless they are deleterious, have a good chance of being perpetuated and the hyphal tip containing the mutant gene may rapidly develop a “sector” different from the parent. 70 SPECIES STUDIES IN THE ERITISH FLORA Heterokaryosis. In the higher organisms there is a fixed relation between nucleus and cell and a fixed alternation between haploid and diploid. In many fungi this is not so. In the Phycomycetes and Ascomycetes, cell-wall formation is not usually associated with nuclear division. The compartments into which the hyphae are divided do not correspond to cells in higher organisms. On the whole, the number of nuclei in a compartment is usually about the same. For instance, one often finds five or six nuclei per “cell” but elsewhere in the same hypha one may find as few as one or as many as ten. In most organisms, the plant is derived from a single nucleus, either haploid or diploid, so that any single gene is present in only one form or at the most two, as in a heterozygous diploid. In fungi, fusions between hyphae from several different spores are common and interchange of nuclei occurs. This can lead to the formation of a mycelium with hyphae containing nuclei of several different genetic constitutions (heterokaryosis) and the resulting mycelium is the outcome of the interaction between the nuclei with different gene complements. During growth of the organisms, hyphae may be formed which have a different grouping of the nuclear types, a difference either in kind or in proportion, and a new variant is born. There appear to be at least two types of heterokaryosis. The best known is common in the Ascomycetes and Fungi Imperfecti, which have multinucleate compartments in their hyphae. The nuclei may be of several genetic kinds and come together as the result of fusion between hyphae derived from different spores. In nature, selection pressure probably keeps the heterokaryons fairly stable, but, in the laboratory, removal of competition and stringent selection leads to a burst of variation and presents the mycologist with a multitude of different strains. A less familiar form occurs in fungi like Fusarium where Buxton (1954) has shown that although the tip of the hypha is multinucleate and presumably heterokaryotic, cross- wall formation produces uninucleate com¬ partments. Thus cells are produced containing single nuclei, but adjacent cells may have nuclei with different gene complements. Lateral outgrowths from these cells will give initially hyphae of a single genetic type. Local areas may, therefore, be wholly or pre¬ dominantly of one genetic origin, probably corresponding to the patch mutants of Miller (1946). Secondary anastomoses restore the heterokaryotic condition at the growing tips. Baper and Antonio (1954) have described a form of hetero¬ karyosis in Basidiomycetes where a mycelium which is haploid as regards mating type still carries two types of nuclei physio¬ logically distinct. Whether such a condition will be found in nature remains to be seen. Recently Pontecorvo et ol. (1953) have emphasized the im¬ portance of another source of variation in fungi. In forms such as Aspergillus niger, where no sexual reproductive mechanism is known, it has been found that among the normal haploid nuclei PROBLEMS ASSOCIATED WITH THE SPECIES CONCEPT IN MYCOLOGY 71 there is a small proportion of diploids. If the original mycelium was heterokaryotic, some of these diploids can be shown to be heterozygous for characters derived from the different genetic types forming the heterokaryons. In an artificially produced heterokaryon, Pontecorvo found that 3 in 10^ conidia contained heterozygous diploid nuclei. Subsequent growth of conidia gave colonies which showed mitotic segregation and recombination of the characters and a small proportion of the nuclei in these, 1 in 10"^, that is 3 in 10’^ of the original nuclei, became haploid (Ponte- corv^o, 1953b). The rate of occurrence of the heterozygous diploids is of the order which would be expected for a normal gene mutation. The additional factor of the slow rate of haploidization means that the nett effect of the whole process may be small. The ability to achieve new combinations of groups of genes by mitotic segrega¬ tion confers on the organism some of the benefits of sexual repro¬ duction. The variability and the adaptive capacity of such a mechanism would presumably be intermediate between gene mutation and normal sexual reproduction. Growth in Culture. A great many of our difficulties in specific discrimination in the fungi are of our own making. By the isolation and continued growdh of organisms on media which are often very different from their natural substrates, we induce a multitude of variants and impose entirely new selective conditions. Tlie dangers associated with the culture technique were well appreciated by some of the older mycologists, and it was indeed a source of continual friction between the older and younger mycologists at one time. The epilogue to volume 2 of Grove’s British Stem- and Leaf-Fungi (1937) is a relic of what was once a widespread dissension. He describes the plight of systematic mycologists. “Some have plunged into the thorny thickets of synonymy, floundering amid the multitudinous meticulosities of Nomenclatorialism, from which they rarely emerge unscathed; others wandered discon¬ solate over arid deserts of Petripatellism, plodding dully along a path, meandering and redeless, which could of necessity lead nowhither; still others filled up their space with frills, such as a boring discussion of the dietetical predilections of the patient for concocted foods.” He describes ‘Nomenclatorialism’ as “an intricate esoteric art which strives to affix to every living creature (plant or animal) a definite unchangeable Latin label in strict accordance with the very latest views about Scientific Nomen¬ clature. The object of the art was to reach finality; but it has not attained that end, nor can it, so long as the multiplicity of nature is rivalled by the variety of men’s minds. A naturalist should take heed lest he become too nomenclatorialistically minded”. He describes ‘Petripatellism’ “as the state of mind of a mvcologist who studies his fungus in a laboratory, on agar-slants or Petri-dishes without paying equal regard to what the fungus 72 SPECIES STUDIES IN THE BRITISH FLORA can do out of doors in the wide and untrammelled field”. Unfor¬ tunately, few people seem to have read or at least paid much attention to this epilogue. It is true, of course, that Grove was equally at fault when he dismissed so readily and unsympathetic¬ ally the results* obtained from pure culture work. The breeding programmes associated with modern genetical studies have brought home very clearly to geneticists the way in which intensive breeding and culture of an organism, such as Drosophila or Neurospora, has produced a host of laboratory strains differing in their morphology and physiology from each other and from the original form from which they came, and for which it was necessary to find a special designation — the “wild type”. As yet, in mycology, there has been little recognition of the need to define a “wild type” as distinct from the tamed and modified laboratory organisms, and the work of Miller in the Fusaria is a most valuable contribution to this aspect of the study of systematics. In the maintenance and identification of culture collections it is of importance that the “wild type” be cultivated and not the laboratory mutant. Applicability of Angio sperm Taxonomic Concepts to Mycology. With the development of experimental taxonomy of Angio- sperms the concepts of taxa at specific and subspecific levels underwent numerous changes. Although there are still manv matters in dispute a number of terms and ideas have met with general acceptance. It is of interest to see how some of these concepts can be applied to systematic mycology. Linneons and Jordanons. Close on a century ago Jordan showed that Draha (Erophila) verna of Linnaeus could be analysed into a large number of dis¬ tinct strains. Since then it has been found that a large propor¬ tion of the species studied are similarly complex. In some cases, experimental work has shown that the smaller units are all inter- fertile but ecological or other factors tend to give some degree of isolation to groups of characters. In mycology, it has also been found that many “Linneons” can be split into smaller and distinctive units, for example Corti- cium coronilla H & L., which Biggs (1937) has shown contains a large number of strains differing in cultural characteristics, spore and basidium size, etc. For normal purposes, there is little to be gained by giving such strains special status and to rank them as species would merely produce confusion. Aggregates and Coenospecies. Where a species seems to be particularly plastic, or where there is a group of very closely allied species differing from each other by relatively few characters, it is convenient to use some term to indicate that a broad view of the species is being taken. PROBLEMS ASSOCIATED WITH THE SPECIES CONCEPT IN MYCOLOGY 7:^ The term coenospecies, proposed by Turesson (1922), was de¬ signed for some groups of closely allied species and would at times be applicable to fungi. Subsequent attempts to redefine the term have not been happy and its use has tended to decrease. Recently the more vague term aggregate species has gained popularity, perhaps because of its fewer implications. In the new British Flora there are a number of examples of the convenient use of the aggregate species. One may cite Alchemilla vulgans agg. under which eleven critical species are described. For many purposes the aggregate species is adequate, but for critical work identification to the finer level is necessary. In the difficult genera Aspergillus and Penicillium, the idea of an aggregate species has proved extremely useful as Thom (1952) has pointed out. The early monographs by Thom and Church on the Aspergilli in 1926 and the Penicillia in 1930 were landmarks in systematic mycology, and emphasised very clearly the special problems associated with such difficult groups of fungi. In recent years, both these monographs have been replaced by new ones from the same authors, the Aspergilli in 1945 and the Penicillia in 1949. The contrast between the old and the new treatments makes a most interesting study. In the work on Aspergillus, the difficulties of identifying the innumerable strains, often differing from one another by minute morphological charac¬ ters, or by a single chemical property, has meant the abandon¬ ment of the narrower specific limit found in the earlier works. The various strains are now placed into groups, and each group contains a number of series. Within the series specific rank is still accorded to certain well marked forms, but it is clear that there is no difficulty in finding a complete range of forms linking a number of the so-called species which had previously been re¬ garded as distinct. For most purposes identification of an organ¬ ism to its series is adequate. It is difficult to make a direct comparison between the units in Aspergilli and species in higher plants, but it is probably justifiable to regard the series of Thom and Raper as approaching very closely to the aggregate species as used by Clapham, Tutin and Warburg in the new Flora of the British Isles, and the species of Thom and Raper as corresponding to the critical species. The intensive study of Penicillia and Aspergilli, stimulated by biochemistry and industrial chemistry, has been paralleled in the genus Fusarium because of its interest to agriculture and plant pathology. The widespread occurrence of forms of Fusarium attacking the underground parts of economic plants, and the occurrences of apparently saprophytic forms in the soil, has made it desirable to have some methods of species discrimination within the group. For many years, the works of Reinking, Sher- bakoff, and Wollenweber have formed the basis for identification. The most detailed treatment of Wollenweber and Reinking (1935) listed some 65 species, with 78 varieties, grouped into 16 sections. Despite this careful monograph, the identification of a particular 74 SPECIES STUDIES IN THE BRITISH ELORA isolate was the task for a specialist and involved a great deal of cultural work and examination. The most recent discussions of the problem by Hansen and Snyder (1940) and by Miller (1946), have both proposed drastic simplification of the older works. As a broad generalisation, it may be said that the recent workers re¬ gard the sections of Wollenweber as equivalent to species, and the majority of the older species as varieties of a few well marked pleomorphic species. Again, one is tempted to equate the new concept of the fungal species with the aggregate species of the angiosperm systematist. Any detailed consideration of the problem though must await the results of genetical evidence re¬ garding the infertility of the various units. Infra-specific Units. In mycology, as in all other branches of botany, there is a great variation in the use of infra-specific units. The most detailed studies have come from workers in the rusts where the economic importance of being able to recognise individual strains has led to intensive investigation of minor taxonomic units. In Puccinia graminis it has long been realised that there are groups of strains which attack predominantly wheat or barley or oats, etc. These strains have been grouped as sub-species and the trinomial nomenclature favoured by zoologists adopted. Thus the strains on wheat and on oats are grouped as P. graminis tritici or P. graminis avenae respectively. Genetical work has shown that these groups are somewhat interfertile but have each become adapted to their own ecological niches. They are in effect eco- species. The subspecies are themselves capable of being further sub¬ divided into physiological races. These differ by very minor morphological characters and are distinguished on their ability to attack different members of a range of test varieties of the host. The individual races are given a number. Genetical analysis has shown that the differences between many of the races are very small and may be due to a small number of gene differences. Some of the races, e.g., r.34, are known to be hetero¬ zygous, the common segregates from r.S4 being r.ll and r.56. (Waterhouse, 1952). Natural hybrids. References to the natural occurrence of hybrids between different species of fungi are usually vague. Perhaps the best example is that of Allomyces javanicus which Emerson and Wil¬ son (1954) have shown is a natural hybrid with a complex and unstable chromosome complement. Reports of hybrids in the higher fungi are often based on little evidence, difficult forms being attributed to hybridisation solely on the grounds that the specimen is intermediate between the concepts a mycologist has of the two presumptive parents. The tendency to class certain forms of Mycena or Hygrophorus as PROBLEMS ASSOCIATED WITH THE SPECIES CONCEPT IN MYCOLOGY hybrids may simply be an expression of the difficulty of placing the large numbers of apomictic races in these genera. Careful analyses of presumed hybrid populations in the Basidiomycetes are rare. Parker- Rhodes (1950), as the result of the application of his method of tetrasporic analysis, considers that natural hybridisation occurs between Psilocyhe hullacea and P. coprophila on Skokholm Island giving rise to a population of intermediates which he has described as P. scocholmica. Accounts of the artificial production of hybrids are not uncommon par¬ ticularly in the Phycomycetes, for example, that of Raper (1950) who cross-mated a number of species of Achlya. In some, e.g. A. amhisexualis and A. americana, he obtained complete compati¬ bility. In others, the incompatibility was either partial or com¬ plete. In the higher groups artificial hybrids are rare. Roma- gnesi (1948) has reported the experimental production of a hybrid between Agrocyhe [Pholiota) praecox and A. sphalenomorpha. There seems little doubt that under laboratory conditions there is no difficulty in obtaining hybrids in some genera. The apparent scarcity of natural hybrids might well be due to the difficulty of recognising a hybrid as such when it is found. Apomixis. Development of a gamete or a gametangium without sexual fusion is common and widespread in the fungi, and mycology courses are full of examples of “progressive loss of sex”. Many species possess no distinct morphological differentiation of sex organs yet still retain the essentials of sex, the fusion of nuclei • followed by meiosis. True apomixis, however, occurs in all the groups. In the Phycomycetes, Cutler (1942) has shown that Sporodinia grandis, although forming apparently good gametangia which fuse, does not follow this with nuclear fusion but develops a zygote which subsequently germinates to give a sporangium containing the original haploid nuclei. There are numerous examples of the development of a fruit body in the Ascomycetes without nuclear fusion. In many of these well developed sex organs still occur, as, for example, in Ascobolus equinus. In the Basidiomycetes it has been shown by Kiihner (1938) that in the genus Mycena apomixis is common. It is interesting that the species of this genus are generally regarded as not diffi¬ cult provided care is taken to observe the fine details of the shape of the basidia and cystidia and the spore size. Work with angiosperms has shown that apomixis is often associated with the occurrence of a large number of forms, each reasonably distinct and constant, but separated from other forms by very small differences in morphological characters. This would seem to be the situation in the agarics such as Mycena and Hygro- phorus. In the Ascomycetes it is possible that the apparent absence of the fine splits usually associated with apomixis may be due to the occurrence of heterokaryosis, which, with its intro- 76 si»p:cies studies ix the British flora duction of different gene complements, confers a plasticity unknown in the gene isolation of the angiosperm apomict. Cytotaxonomy. Few detailed cytotaxonomical studies of fungi are available. It is clear, however, that both allopolyploidy and aneuploidy have played a part in the evolution of fungal species and genera. Figure 14 summarises most of the available chromosome numbers Fig. 14. The distribution of chromosome numbers in the different groups of fungi for fungi. No attempt has been made to assess the relative reliability of the counts and all records seen have been included except for the yeasts and the smuts. In both of these groups two chromosomes are frequently reported. The text figures, however, are so unconvincing that it seems best to disregard them for the present. IMIOMLKMS ASSOCIATKl) WITH THK SIMiCIKS CONCEPT IN MVCOJ.OCV i I Figure 14 suggests very strongly that the basic chromosoinii number in the fungi is n==4. In the Phycomycetes, 2 and 3-ploid series are common and higher polyploids occur. Most of the high counts are in the genus Alloinyces but high numbers also occur in the few available counts for Myxomycetes. In contrast, the Basidiomycetes show" little evidence of polyploidy. The best and only detailed cytotaxonomic study at present available is that of Emerson and Wilson (1954) in the genus Allomyces. The chromosome numbers of the different species and strains are given in Table 1. This shows clearly that poly¬ ploidy occurs both within a genus and within what is normally regarded as a single morphological species. Table 1. Chromosome numbers in Allomyces from Emerson and Wilson (1954). Species. Chromosomes. A. arhuscula. 8 Brazil, Portugal, India, Fiji, Philippines, U.S.A., Australia. 16 Trinidad. 24? Belgian Congo. 22-26 Queensland. 32 Illinois. A. macrogynus. 14 India. 28 Burma. 50 + Philippines. A. javanicus. , 13, 14, 14 or 15, 16 or 17, 19, 21. This species is regarded as a natural hybrid. A. arhuscula x A. macrogynus. 20, 22, 27, 36, 42, 42 + , 44. An artificial hybrid. Recent investigations of McGinnis (1953 and 1954) suggest that in the rusts the basic chromosome number is 3, as, for instance, in Puccinia coronata. P. graminis, with n= 6, is regarded as a polyploid and the clear evidence of secondary association of the chromosomes suggests that it is of hybrid origin. Difficulties Peculiar to Certain Groups. In the Phycomycetes, systematic discrimination does not seem to have presented any outstanding difficulties. Certain groups, e.g. the Mucorales by Zycha (1935), have been well mono¬ graphed and most workers find it possible to place with reason¬ able confidence the majority of the strains they isolate. This is also true of the aquatic Phycomycetes which have been mono¬ graphed by Sparrow (1943). The Ascomycetes, again, as a group, do not seem to present any major difficulties except for a few- orders. Some of these, how*- ever, are particularly troublesome, like the yeasts. It is inevitable that in unicellular organisms the number of available morpho¬ logical characters is small and it is necessary to fall back on physiology and biochemical reactions. This is notoriously un¬ satisfactory owing to the rate at which characters are lost or “adaptations” occur, Romagnesi (1948) has given an excellent summary of the problems and methods of systematics in the Basidiomycetes. Many of the problems in this group arise from the difficulty of preparing good herbarium specimens. This has meant that tradition has played a very large part in the concept of the different species. Much of the accumulated knowledge of Agarics has been passed on by word of mouth and traditions have been built up in the different countries regarding the local concept of what constitutes a particular species. Probably in no other group IS it necessary so often to qualify a specific name by a comment such as ^^sensu Lange, non Bresadola” etc. In recent years there has been an attempt to correlate the traditional concepts from different countries and to encourage the collection of herbarium material. This is particularly important in association with species-lists of plants occurring in any district, as there is hardly an example of a check-list supported by reference specimens. In most genera of the Agarics the species appear to be reason¬ ably well separated, and the number of presumed naturally occur¬ ring hybrids is very small. In certain groups, as e.g. Tricholowxi, it seems that the older concept was too narrow, and that it is necessary to lump some of the earlier species, as, for example, in the T. terreum Fr. — T. scalpturatum Fr. group, where forms such as T. argyraceum (Bull.) Fr. and T. chrysites (Jungh.) Gill., mark distinct points in a wide range of colour and scaliness of the pileus or again to place the different colour forms and varia¬ tions such as T. brevipes (Bull.) Fr., T. excissum Fr., T. humile (Pers.) Fr., T. oreinum Fr., T. paedidum Fr., T. patulum Fr., T. phaeopodium (Bull.) Quel., T. subpidverulentum (Pers.) Fr. etc., all under T. melaleucum (Pers.) Fr. Colour has, of course, played a great part in the recognition of species in the Agarics, and while on the whole it is an extremely valuable character, it does at times appear to depend on very small genetic differences. From time to time, attempts have been made to apply chemi¬ cal tests to the identification of fungi and of lichens. This has often aroused most violent partisanship, and the sensible view that the chemical tests simply add one or two more characters to the already recognised morphological characters is surprisingly seldom taken. In a group which is already difficult and in which characters are few or very variable, one would imagine that any additional evidence would be welcome. The opposition there¬ fore to the introduction by Schaeffer and Moller (1938) of chemi¬ cal tests to aid specific discrimination in the notoriously difficult genus Psalliota seems inexphcable. There is no essential difference between recognising a chemical in the pileus because it is brightly coloured and can be seen by the eye, and perceiv¬ ing its presence by means of a chemical test. The problems associated with the classification of the Fungi Imperfect! are largely due to the absence of the perfect stage. Many higher plants have vegetative forms of reproduction but most of the species still retain their normal sexual mechanisms as well. In the fungi it seems that large groups of organisms spread primarily by vegetative spores and seldom, if at all, reproduce sexually. Since the systems of classification are based primarily on sexual reproductive mechanisms and spores, the mycologist is left with a host of organisms producing no apparent sexual stage. These he has dumped in a group called the Fungi Imperfect!. It was at one time assumed, tacitly at least, that each fungus would eventually reveal a sexual stage and could then be re¬ habilitated and placed into its proper systematic position. It seems, certainly at the moment, that this is a vain hope and since taxonomy is mainly a matter of convenience and expediency, names have been given to these imperfects. The systematic arrangement has been frankly artificial on the form and grouping of the vegetative reproductive units. Suppose we were forced to exclude flowers from the classifica¬ tion of the angiosj)erms and use vegetative reproduction only. We might form a genus containing all the species with hulhs — this would not be a bad group, in fact surprisingly close to the Liliaceae : a genus based on corms, however, would be poor — while one based on tubers would bring together Dahlia, Solarium, Orchis, etc. The difficulties of classification in the Fungi Imper- fecti are of this order. It is true, though, that they are usually difficulties of generic rather than specific level. The Fungi Imperfecti present a further taxonomic problem since several fungi which are distinct in their perfect stage may have very similar imperfect forms, e.g. Neurospora spp. and Monilia sitophila. If we had only the Monilia stages, we would almost certainly consider that we had only a single species — and a fairly constant one at that. It is not at all improbable that some of our variable species of the Imperfecti are aggregates of the imperfect stages of several different fungi. The treatments accorded the different genera have varied greatly with the author concerned. In some, as in Fusarium, the modern tendency is to “lump”. In others, like Alternaria, a large number of species is still accepted. General Discussion. It is clear that it is possible to define a group or groups of organisms either in terms of their morphological pattern, or in terms of their genetic behaviour elucidated by breeding experi¬ ments or cytological investigations. Often the two definitions coincide and the new systematics confirms the old, but where differences occur some compromise is usually sought. In the higher plants most of the lack of agreement is due to polyploidy 80 SI'KCIKS STUDIES IN THE 15RITIS11 FI.ORA unaccoiiipanied by marked morphological changes, or by apomixis. The former seldom causes practical difficulties in systematics. Apomixis, on the other hand, is only too well known in the genus Ruhus, with its some 4000 morphological units in North Western Europe alone, or in Hieracium, which probably has twice as many. Some special convention will have to be devised to handle such situations. In the fungi the impact of apomixis on systematics has been slight. It is clear that in the genus Mycena a mixture of mictic and apomictic forms is common, so that one might expect a chaos comparable to that found in Ruhus. Perhaps because of the difficulty of getting good herbarium specimens for comparison of collections from many localities the careful studies of slight differences of morphology have not been made, and a Pandora’s box of microspecies awaits opening. On the other hand, the efficiency of the spore dispersal mechanisms may mean that in¬ dividual apomicts can potentially spread over very wide areas and competition has eliminated all but the most successful forms. The phenomenon of heterokaryosis confers a potential plasticity far in excess of that seen in a normal haploid or diploid organism. It would seem reasonable, therefore, to have much wider species concepts in genera where heterokaryosis occurs. The use of aggregate species in Penicillium and Aspergillus and the recent lumping in Fusarium, although based on older systematic methods, can be strongly supported on the genetical grounds that heterokaryosis occurs in these genera. When dealing with pathogenic forms, the assumption that a species is distinct because it is on a different host, even though there may be accompanying morphological differences, is far from justified. It is essential that transplant experiments to a more familiar host be attempted to eliminate the possibility that the so-called different species is not just an ecad. It seems reasonable to conclude that the problems of species discrimination in the fungi are * essentially the same as in other organisms, although they may be accentuated by the rate of speciation or by great plasticity. One cannot dismiss the fungi from the general schemes nor can one ignore the lessons learnt from fungi in considering the species concept in higher organisms. I am most grateful to my colleagues, particularly Dr. J. Burnett, for helpful discussions during the preparation of the above paper. Literature Cited. Bessev, E. a., 1950, Morphology and Taxonomy of Fungi. Blakistou Co., Philadelphia. Biggs, R., 1937, The species concept in Chrticiuni coionilla, Mycologio , 29, 686-705. rilOlil.KMS ASSOCIATED WITH THE Sl'ECIES CONCEI’T IN MA'COHOOV 8.1 Buxton, E. W., 1954, Heterokaryosis and variability in Fusarium oxy- sporiiim f. gladioli (Snyder and Hansen), J. Gen. Micru- hiol., 10, 71-84. CuTLEii, V. M., 1942, Nuclear behaviour in the Mucorales. II. The Rhizopus, Phycomyces and Sporodinia patterns. Bull. Torr. Bat. Club, 69, 592-616. Emerson, R. & Wilson, C., 1954, Interspecific hybrids and the cyto¬ genetics and cytotaxonomy of Eualloinyces, Mycologia, 46, (in press). Ephrussi, B., 1952, Nucleo-cytoplasmic relations in micro-organisms, 1-127. Oxf. Univ. Press. Grove, W. B., 1937, British Stetn and Leaf Fungi, 2. Cambridge Univ. Press. Kuhner, R., 1938, Le genre Mycena, Encyclopedie Mycologique, 10, Paris. McGinnis, R. C., 1953, Cytological studies of chromosomes of Rust Fungi. I. The mitotic chromosomes of Puccinia graminis, Canad. J. Bot., 31, 522-526. - , 1954, II. The mitotic chromosomes of Puccinia coronata, Canad. J. Bot., 32, 213-214. Mason, E. W., 1933, Annotated account of fungi received at the Im¬ perial Mycological Institute, List II, Fascicle 3. Miller, J. J., 1946, Cultural and Taxonomic studies on certain Fusaria, I and II, Canad. J. Res., C., 24, 188-223. Barker-Rhodes, a. F., 1950, The basidiomycetes of Skokholm Island. IV. A case of hj'bridization in Psilocybe (Deconica), 'New Fhyt., 49, 335-343. PoNTEcoRVo, G., 1953, Diploids and mitotic recombinations, p. 18, in Adaptation in Micro-organisms — 3rd Symposium of the Society for General Microbiology. PoNTEcoRVo, G., Roper, J. A. & Forbes, E., 1953, Genetic Recombina¬ tion without sexual reproduction in Aspergillus niger, J . Gen. Microbiol., 8, 198-210. Rarer, J. R., 1950, Sexual hormones in Adilya. VII. The hormonal mechanism in homothallic species, Bot. Gaz., 112, 1-24. Rarer, J. R. & San Antonio, J. P., 1954, Heterokaryotic mutagenesis in hymenornycetes. I. Heterokaryosis in Schizophylliyn commune, Amer. J. Bot., 41, 69-86. Romagnesi, H., 1948, Les problemes et les methodes de la systematique des champignons superieurs, Bull. Soc. Mycol. France, 64, 53-100. Schaffer, J., 1938, Beitrag ziir Psalliota-Forschung, Ann. Myc., 36, 64-85. Sparrow, F. K., 1943, Aquatic Phycomycetes. Univ. Mich. Press. Snyder, W. C. & Hansen, 1940, The species concept in Fusarium, Amer. J. Bot., 27, 64-67. Thom, C., 1952, Molds, Mutants and Monographers, Mycologia, 44, 61-85. Turesson, G., 1922, The genotypical response of the plant species to the habitat, Hereditas, 3, 211-350. 82 SI’KCIKS STUDIKS IN THIO imiTlSH FI.OHA VuiLLEMiN, P., 1910, Materiaiix pour uiiG' classification ratioiicllo dcs Fungi Iinperfecti, Compt. rend., 150, 882-884. Waterhouse, W. L., 1952, Australian rust studies, IX. Pliysiologic Race Determinations and Surveys of cereal rusts, Froc. Linn. Soc. N.S.W., 77, 209-258. WoLLENWEBER, H. W. & Reinking, O. A., 1935, Die Fusarien, Hire Beschreihung , Schadwirkung and Bekdmpfutig . Berlin. Zycha, H., 1935, Mucorineae, Kryptogamenfiora der Mark Branden¬ burg, 6a. Leipzig. TEIK ANALYSIS Ob' VARIATION WITHIN THb) (JKNUS FUCUS THE ANALYSIS OF VARIATION WITHIN THE GENUS FUCUS E. Burrows and S. M. Lodge (University of Liverpool). The exhibit under this title illustrated for the Algae a species problem of a type frequently found among flowering plants. Within the British Isles there are to be found five taxonomic species of Fucus. Of these F. spiralis L., F. vesiculosus L. and F. serratus L. are common on rocky places all round the coast. F. ceranoides L. is also widely distributed in its habitat in the mouths of fresh-water streams. The fifth species, F. infiatus L. has its centre of distribution much further north and has only a very limited distribution in the British Isles. The status of these species and their ecological relationships are still far from clear. Of F. ceranoides little can be said at present but it presents an interesting problem. It is readily distinguished by its dark colour, delicate papery texture, fine thread-like stipe and very pointed receptacles. These are, however, all characters influenced by habitat conditions. At different tide levels F. ceranoides may show, in addition, characters of any of the other three species, F. spiralis, F. vesiculosus and F. serratus, and it is possible that it includes estuarine forms of all three. F. spiralis, F. vesiculosus and F. serratus are distributed in relation to tide levels in such a way that, when all three are pre¬ sent on the shore in abundance, they are limited to fairly well defined zones with F. spiralis in the upper part, F. vesiculosus in the mid-tide region and F. serratus in the lower part and extend¬ ing well below L.W.S.T. Under such conditions the species appear to be separated by clear cut characters. In situations where the shore is more sparsely covered by vegetation or where clearance experiments have been carried out, the zones become less distinct, the species extending their ranges up and down the shore, and individuals are then often more difficult to refer to one or other taxonomic species. F. serratus is usually distinct, but F. spiralis and F. vesiculosus have often little to separate them except the hermaphrodite condition of the former and the dioecious condi¬ tion of the latter. It is to be emphasised that, for intertidal species, a difference of only two or three vertical feet in tidal posi¬ tion may mean a considerable difference in ecological conditions m terms of exposure to a drying atmosphere, changing light intensity and temperature. These three species are interfertile and hybrid sporelings have been raised in culture. There is a good deal of evidence to show that hybrid sporelings are also formed in nature, but that com¬ petition with the parent species, under the conditions prevailing 81 Sl'KCIKS STUDIES IN THE IHUTISH El.OKA in a well developed shore zonation, prevents their establishment. Suspected hybrid swarm populations have been found in a num¬ ber of places under either naturally or experimentally disturbed conditions. The results of an analysis of a suspected hybrid swarm population between F. spiralis and F. vesiculosus was shown in the exhibit. Anderson’s hybrid index method was used for the analysis and the characters included were those normally employed by taxonomists in separating the species. The analysis showed about 20 % of F. spiralis, a little less of F. vesiculosus and a large range of intermediates between them. It is possible, how¬ ever, that this range of intermediates is merely giving an indication of the degree of plasticity of the species and does not, in fact, represent a hybrid swarm. On the information available at present it is difficult to decide between these two interpreta¬ tions. The exhibit was particularly concerned with the variation shown by F. vesiculosus and F. in flatus. All the species of Fucus are variable but these two are particularly so. In both cases the form of the plant varies with the degree of exposure to the effects of open sea breakers. The effect of rough water is to reduce the growth rate by rendering photosynthesis difficult so that narrow’ dwarfed plants are produced and, in the case of F. vesiculosus these lack the typical air vesicles. For F. vesiculosus, an attempt has been made to analyse this variation on a population basis and, for this purpose, an adaptation of Wilmott’s grid method for recording variation within a critical group has been used. Several populations growing under different degrees of exposure to open sea breakers have been analysed and defined by this method, and the results show variation on an ecocline from sheltered to exposed conditions. The extremes of this range have long ago been defined as varieties of F. vesiculosus. The sheltered water form was distinguished as F. vesiculosus var. vadorum Aresch. and that from rough water as F. vesiculosus var. evesi- culosus Cotton. How far this variation is accounted for by plasticity of phenotype or what degree of isolation the extremes have is not known, but the problem is now being tackled by culture methods. The plants can be grown under standard culture conditions to a length of 5 or 6 cms. and attempts are being made to plant them out on selected shores to finish their growth. F. infiatus, which differs from F. vesiculosus in a number of characters and especially in being hermaphrodite instead of dioecious, varies in form very strikingly with degree of exposure. Here again the extremes of the range have been distinguished, for sheltered water as F. inflatus i. edentatus (de la Pyl.) Rosenv. and for rough water as F. inflatus L. f. distichus (L.) B0rgesen. This species is common further north in the Faeroes, Iceland and Norway and there the extremes are linked by intermediate forms. In Britain only the extreme forms are found and these are restricted both ecologically and geographically: F. inflatus THE ANAT.YSIS OF VARIATION WITHIN THE GENUS FUCUS 85 f. edentatus is found only in Lerwick and Scalloway harbours in Shetland and in North Haven in Fair Isle; all of these are functioning harbours. F. inflatus f. distichus is confined to exposed coasts on the Atlantic facing shores of the British Isles. It occurs as far south as Kilkee in the west of Ireland but its extreme southern limit is not known. One feature which is very striking is the similarity in appear¬ ance between F. vesiculosus var. evesiculosus and F. infiatus f. distichus. The two species are very different and easy to dis¬ tinguish at one end of their variation range, but almost identical in appearance under rough water conditions. The two can be separated on the sexual characters in that the former is dioecious and the latter hermaphrodite. The exliibit did little more really than illustrate the species problem for Fucus. A start has been made on its solution by the development of suitable culture techniques and an attempt to sort out the variants of the taxonomic species on a population basis. Experimental work in the field has begun to give informa¬ tion on the ecological relationships of the components of the genus, but of the real status of the components still very little is known. Reference. WiLMOTT, A. J., 1950, A new method for the identification and study of critical groups, Proc. JAnn. Soc., 162, 83-98. 86 SPECIES STUDIES IN THE IHIITISH FLORA THE SPECIES CONCEPT AMONGST BRYOLOGISTS E. W. Jones (Imperial Forestry Institute, Oxford). Bryologists have tended to apply to bryophytes the concepts of species formed in the study of flowering plants; they have not been great innovators of new concepts. The amateur status of a high proportion of our leading bryologists is doubtless one important reason for this situation; the late H. N. Dixon, for example, was a school-master with a classical education, Richard Spruce a mathematical school-master, and Max Fleischer, whose work on the mosses of Java forms the basis of our modern classi¬ fication of mosses, was an artist by profession. Men such as these had neither the training nor the resources to allow them to develop new methods, even if such were needed. Moreover, most bryologists have served their apprenticeship in the study of flowering plants and they are at a disadvantage compared with the student of flowering plants because mosses and liverworts are relatively difficult to cultivate, particularly from spores. Consequently the criteria used for defining species have changed and multiplied but the fundamental concepts have altered but little. The earlier bryologists attached great importance to the gross morphological features of the vegetative and (from analogy with flowering plants) more especially the reproductive organs. \Vlien, at the close of the 18th century, Hedwig introduced the use of finer details, he was criticised, e.g. by Menzies (1797) who wrote ‘No generic or specific characters ought ever to be adopted that cannot easily and distinctly be seen by the assistance of a single- lens magnifier such as botanists commonly carry in their pockets’ (quoted by Steere, 1947). Bruch and Schimper figured many anatomical details in their great Bryologia Europaea (1836-55), and Limpricht described in detail the anatomical structure of all species in his volumes on European mosses in Rabenhorst’s Kryptog amen flora (1890-1904). The size and character of the cells of the leaf are now universally recognised as giving specific characters of great value, and the structure of the stem, midrib etc. as seen in cross section is often characteristic, especially in mosses. The search for additional characters of systematic importance continues, and increasing attention is being paid to the nature of the cell-contents. Thus the oil-bodies, which are present in most hepatics, often present valuable specific characters, but unfortunately in many species they disappear soon after death in specimens preserved by the usual methods. In some groups also chromosomes have been studied. Lorbeer found that European female plants otherwise indistinguishable from tlie THE SPECIES CONCEPT AMONGST BRYOLOGISTS 87 American Sphaerocarpus texanus Austin differed from the American plants of this species in some details of the X-chromo- some, and he therefore separated the European plant as a different species under the name of S. europaeus — a course which has not met with the approval of so excellent a hepaticologist as K. Muller (1951); this affords an excellent illustration of the difficulty of knowing how far to proceed in splitting up into specific units and of the necessity for taking into account what is practicable when doing so. Many bryophytes are extremely plastic, and one of the greatest practical difficulties of the bryologist lies in distinguish¬ ing between true genotypes and modifications which are the direct result of the influence of environment. Many of these latter have been described in the past as species or varieties, and indeed some bryologists, even in relatively recent years, have not always excluded the environmentally-produced modification from their concept of the species ; thus in 1916 when Douin (a mathematician) had by cultivation produced a Cephaloziella completely different in appearance from the plant which he originally gathered, he described it as a new species because, as he said with a mathematician’s logic, he would have described it as a new species had he found it growing wild (quoted by K. Muller, 1947, p. 15); similarly Ingham in 1908 could speak of Drepanocladus wilsoni (Schp.) as being a derivative of Z). lyco- podioides (Brid.), and of D. sendtneri (Schp.) as originating from various species. Thus critical work has, in the course of time, generally reduced the number of forms admitted as species. During the past twenty-five years, however, some critical studies combined with experimental work have revealed hitherto un¬ suspected genotypical variants, and there has been a tendency to describe all such genotypes as ‘species’. This practice, which may be compared with that followed by the students of Salices or Rubi, leaves no place for ‘varieties’, and in my opinion it goes too far. There is perhaps no reason why the concept of the species as applied to bryophytes should be any different from that applied to flowering plants unless it lies in the different relationship between sporophytes and gametophyte. In most ferns both gametoph;^e and sporophyte are independent, though the former, in the form of the prothallus, is reduced and plays a brief part in the life cycle. In flowering plants the gametophyte is no longer independent and is represented only by a short phase in the development of the egg-cell and the pollen-grain. In bryophytes, on the contrary, it is the gametophyte which is the independent plant, and the sporophyte, though sometimes com¬ plex in structure, remains attached to and dependent upon the gametophyte and has become the biological equivalent of a seed- vessel. T^e principal effect of this difference on the practice of systematy is to reduce greatly the part which is ascribed to 88 SPECIES STUDIES IN THE BRITISH FLORA hybridisation as a source of variation. When an archegonium is fertilised by an antherozoid of another species, a hybrid capsule is produced, which of course remains attached to the unaltered parent plant. Such hybrid capsules have been frequently pro¬ duced experimentally; they are generally sterile. They have also been recorded as occurring naturally, when they have been recognised by the occurrence together of both parents and the presence of abnormal capsules, variable and intermediate in character, or sterile, and usually mixed with normal capsules. A few such plants with hybrid capsules have been described as species. Hybrid capsules may well be much more frequent than we imagine, for they would obviously be very easily overlooked. When they produce fertile spores it is probable that in reduction division the genes of either parent will tend to segregate, but they might well produce some offspring of intermediate character. There would be no clear evidence in the field, however, to distin¬ guish between these intermediate gametophytes of hybrid origin and genotypic variants of other kinds; it may well be therefore that hybridisation contributes far more than we suspect to varia¬ bility in some groups of species. Experimentally it is possible to induce portions of a hybrid capsule to produce a protonema aposporically, and from this protonema a hybrid gametophyte can arise ; it is not known whether any process akin to this ever occurs in nature. While processes homologous with apomictic reproduction of flowering plants are not known among bryophytes, the same perpetuation of every genotypic variation is brought about by profuse vegetative reproduction, which may be effected not only by specialised gemmae but also by broken leaves, fragments of rhizoid, protonema etc. Polyploidy seems to be relatively uncommon among bryo¬ phytes; haploid and diploid gametophytes are the most usual, though a few triploid and tetraploid species are known. References . Ingham, W., 1908, Notes on Harpidia, Bevue Bryologique, 35, 8. Muller, K., 1947, Morphologische Pntersiicliiingen zur Anfklarnng einiger europaischer Lebermoose, Beitr. zur Kryptogamen- flora der Schweiz, 10, Heft 2. - , 1951, Die Lebermoose Europas, in Rabenhorst’s Kryptogamen- flora von Deutschland, Oesterreich und der Schweiz, 6, Ed. 3, Leipzig, Steere, W. C., 1947, A consideration of the concept of genus in Mnsci, The Bryologist, 50, 247-258. In reply to Mr. J. D. Lovis, Dr. Jones said that he did not wish to give an impression that spore characters should not be used in Bryo¬ phytes — on the contrary, he was quite willing to make use of them if they proved to be of value. THE SPECIES CONCEPT AMONGST BRYOLOGISTS 89 Dr. H. G. Baker remarked on the infrequency of polyploids amongst Bryophytes. He said he had been impressed by the high chromosome numbers given in Vaarama’s papers on Finnish material, and these numbers do not appear to be arrangeable on a basic number. Dr Jones replied to the effect that cytology provides only one set of characters out of many, and should not be over emphasised. Dr. R. W. Butcher enquired whether the protonema offered charac¬ ters of value. Dr. Jones said that there were small differences but generally speaking the characters of the protonema are common to those of the family rather than of species. Nevertheless it was possible with sufficient experience to recognise the protonema of some common species in the field. 90 SPECIES STUDIES IN THE BRITISH FLORA THE IMPORTANCE OF FERNS TO AN UNDERSTANDING OF THE BRITISH FLORA I. Manton (University of Leeds). My subject has been chosen under pressure from my Depart¬ ment in Leeds who have begged me to give a general picture of the scheme of work into which the separate problems being developed by students and colleagues can be fitted in proper perspective. I was at first reluctant to do this owing to the extent of unpublished data, mostly compiled by other people, which it would be necessary to quote. It is, however, undoubtedly a subject relevant to the topic of this meeting and, on reflection, it became clear to me that by an audience of this kind the dis¬ cussion of unpublished work would be unlikely to be abused. I shall not, however, discuss the species concept as such, since whatever concept one uses it is an inescapable fact that behind the morphologically recognisable taxon there is always hetero¬ geneity of some kind, except perhaps in the special case of obligate apomicts with no mutation rate, for which the species is, in fact, a clone. Elsewhere, as has been shown with dazzling clarity by the papers of Clausen, Keck and Hiesey in the Experimental Studies in the Nature of Species (1940-48), one must recognise the existence at least of physiological heterogeneity on which the main pressure of natural selection of biotypes operates, regardless of whether there is or is not a correlation between physiology and visible structural variation. Similarly, at the cytological level, one must recognise the existence of populations of more or less similar plants with different chromosome numbers or attributes, irrespective of the names by which one chooses to designate them. I do not therefore propose to discuss names, but rather to show the types of conclusions which can be drawn if the analysis of populations by cytological means is carried out on a large enough scale. The question of scale is, however, all-important and an alternative title for my talk might have been to the effect that you cannot understand either individual British species or the British flora as a whole if you confine your attention to this country. The next few minutes will explain this statement. At the outset I must reluctantly remind my audience of the book I published in 1950 under the title of Problems of Cytology and Evolution in the Pteridophyta. It will be necessary to quote some essential facts from it although I will not recapitulate more than is indispensable to the understanding of the work which has been carried out since. Table 1 contains a list of the complexes which cytology had revealed among pre\dously recog¬ nised British fern species. THE IMPORTANCE OF FERNS TO AN UNDERSTANDING OF THE BRITISH FLORA 9i Table 1. Sexual "species" with several forms — Europe. Dryopteris fdix-mas agg. 2n 4n D. dilatata 2n 4n D. villarsii 2n 4n Asplenium trichomanes 2n 4n A. adiantum-nigrum 2n 4n Polypodium vulgare 2n 4n 6n Cystopteris fragilis 4n Qn Enumerating the facts for each more precisely, we have : — 1. The D. filix-mas complex with two forms. The diploid, subsequently named Dryopteris ahhreviata, is characteristically a plant of screes and rocks in mountains. The tetraploid D. filix-mas proper is a lowland plant, probably more abundant in individuals than the diploid in the country as a whole. 2. The D. dilatata complex, likewise with two forms. The tetraploid is abundant and widespread but the diploid has so far only been traced to one plant brought from Ben Lawers by Mr. A. H. G. Alston, though it will probably turn up elsewhere. This subject is being studied by Mr. Stanley Walker. 3. D. horreri (D. paleacea) is an obligate apomict (therefore not in the list) existing in two forms, diploid and triploid, which when hybridized with D. fdix-mas (tetraploid) give the occasional single plants of tetraploid and pentaploid which have been found in Britain. 4. D. villarsii is a tetraploid virtually confined to the lime¬ stone of the northern Pennines. It is the sole type in Britain but it is listed here because all the plants reputed to be D. villarsii so far examined from Switzerland and the French Alps have been diploid. There is therefore a problem as to its nature which takes one at once outside the British Flora. This problem is being studied by Mr. G. Panigrahi. 5. Asplenium adiantum-nigrum exists in two forms of which the common one is tetraploid. A diploid has, however, been collected in Ireland by Miss M. G. Shivas, who went there to look for it and she is exploring the situation further. 6. A. trichomanes. The diploid of this turned up originally at Kew in a plant sent in from Wales. Diploid populations have since been located not only in Wales but also in the Lake District by Mr. J. D. Lovis, who is studying the problem in detail and who hopes to extend the investigation to Scotland and Ireland. Diploid populations are not very extensive, and over most of the country only the tetraploid occurs. 7. Cystopteris fragilis contains three spore forms and two chromosome numbers. The commonest type is hexaploid with large spiny spores. A tetraploid with smaller spiny spores has been found locally in Scotland and the Lake District and a tetraploid without spines on the spores (the var. or species dicJcieana of floras) was originally known wild near Aberdeen but has perhaps now been exterminated by collectors though, 92 SPECIES STUDIES IN THE BRITISH FLORA fortunately for science, stocks have been kept alive in the gardens of amateur fern enthusiasts. 8. Poly podium vulgare also exists in three forms which can hybridize with each other. The commonest is tetraploid. The next commonest, especially in the western parts of the country, is the hexaploid. The diploid is local, being known only from special calcareous localities such as South Devon, Galway Bay and the Cheddar Gorge. The existence of this kind of cytological variation has, of course, been known for a long time and in many countries other than Britain. Various comments have been made upon it from time to time in the literature and one in particular is perhaps important to recall, namely the correlation between percentage of polyploids and latitude, which was first suggested by Tischler, Hagerup and others, and elaborated in greater detail by Love and Love, whose table was quoted on p. 282 of Manton, 1950. These last authors worked with Monocotvledons and Dicotvledons on statistics drawn from a range of latitudes in Northern Europe ( Spit zber gen and Iceland to Denmark and Schleswig Holstein). Bather similar fisrures were obtained in ferns by us when com¬ paring Britain with Madeira, a flora which is, however, perhaps too small to be whollv significant. The interpretation of these statistics is the doubtful question and though some correlation with recent geological history involving glaciation is a possibility to be borne in mind, the correlation, even if it exists, is not a simnle one. This was clearlv shown bv our first considerable excursion awav from the British flora when we examined the tropical flora of Ceylon. Some of the results of this were discussed at the Oxford Svmposium of the Society for Experimental Biologv in 1952 and others will be mentioned here. Table 2 shows a list of the heterogeneous “species” which we found in a mere five weeks of random collecting in Ceylon. Table 2. Sexual ‘'species' with several forms — Ceylon. Gleichenia linearis 2n 4n Thelypteris hrunnea — 4:n 6n T. ifaccida 2n 4n Cheilanthes farinosa 2n 4:n Levtochilus lanceolatus 2n 4r? H'ljpolerns punctata 2n 4n Asplenium lunulatum 2n 4n Pteris ensiformis 2n 4n Cyclosorus parasiticus 2n 4n Antronhyum plantagineum — 4n 6n Athyrium macrocarpum — 4n 6n A . solenopteris — 4n 6n Asplenium affine — — — 8n 12n TiJK IM I'OKTANCK OF FKllNS TO AN U M) KUSTAN DINC! OF TllK lUUTlSil FJ.ORA Uii Tins list can only contain a small proportion of the variants actually present on the island because multiple cytotypes can¬ not be detected unless a species has been examined from at least two places and many of our species were only successfully examined once. In actual fact, as one knows from European experience, multiple types may escape detection in spite of very numerous random samples if one type is much more abundant than another, so that it is obvious that cytological heterogeneity^ must be extremely well represented in the Ceylon flora. For polyploidy as a whole figures quoted in 1952 showed that not less than 60 per cent, of the Ceylon fern flora examined was polyploid, as opposed to 53 per cenh in Britain and 42 per cent, in Madeira. Moreover the grade of polyploidy often encountered was very much higher than in Europe, a fact which was discussed in 1952 in a way which need not be repeated here, except to the extent necessary to point the moral that polyploidy as such can¬ not possibly be either a simple climatic response to latitude or a direct adaptation to the Ice Age as has sometimes been suggested in all seriousness by students of European floras. What then can we infer? And what further action can we take? There are two lines of enquiry which can be applied to our British (or any other) cytotypes which can give important further insight if they are applied on a sufficient scale. These are, firstly, the study of geographical distribution over as large a part of the earth’s surface as possible and, secondly, the study of genetical affinity as revealed by chromosome pairing in hybrids. The first, geographical distribution, is the more difficult to work out with any completeness since the facts must necessarily be based in the first instance on new collecting. They can never be ascertained accurately from the literature owing to widespread confusion of the cytotypes both with each other and with other taxa. Thus on two occasions when floristic lists citing Dryopteris filix-mas in Madeira and Ceylon respectively were followed up, we encountered apogamous cytotypes, diploid in Madeira and triploid in Ceylon, having a general resemblance to D. horreri and not to D. filix-mas at all. On the other hand, herbarium studies can yield reliable results if it has been possible at the outset to correlate mor¬ phological criteria with the cytology. An example of a favourable case in which a herbarium has been so used is in Mr. Panigrahi’s work on the tropical Cyclosorus parasiticus complex quoted in Table 2. C. parasiticus sensu lato is a well known taxonomic tangle of numerous and common species spread all over the oriental tropics and extending as far west as Madeira. Many collections . from Malaya, Ceylon, Africa and Madeira had proved to be tetra- ploid, but one diploid specimen was detected in Ceylon. This proved to correspond, on herbarium characters, to a much less common species, C. repandulus, previously only known from six gatherings scattered in New Guinea, Malaya, and Madagascar but not previously known from Ceylon itself. The facts at once sug- ‘J1 Sl’KClKS STUIJIKS IX THIO JmiTlSH FJ/IKA gest an analogy with diploid D. dUatata in Britain, and the gene- tical behaviour has indeed proved to be similar. The weakness of herbarium studies is that, especially in remote regions, insufficient opportunity may be afforded of checking the determinations by direct cytological analysis. For this reason we have preferred, in dealing with our own flora, to compile our geographical information directly from the cytological records, even though this, at the moment, precludes the possibility of drawing maps in which the distributional areas detected are delimited by accurate lines. In view of this limitation we may list the available data for the non-British ranges of the various cytotypes previously enumerated as follows : — 1. D. fdix-mas. The tetraploid is widespread in Europe. We have no records for America and most of the extra-European records which we have followed up (e.g., Madeira and Ceylon) have been attributable to other species. The diploid (D. ahhreviata) has been found in the Auvergne, but personal com¬ munication from Dr. A. Love reports that this is the only type known to him in Iceland, a fact which may be highly significant. 2. D. dilatata. The tetraploid is abundant in Britain and extends to the edge of Scandinavia where we found it on Bjorko in the outer archipelago near Stockholm. Over most of the Scandinavian peninsula, however, only the diploid is recorded and we have also a diploid from Greenland. In Switzerland diploids occur at a high altitude (cf. Man ton, 1950) and they are widespread in North America (cf. Manton and Walker, 1953). 3 D. horreri. The diploid in Madeira and the triploid of the same or a related species in Ceylon mentioned above are the only non-European records known to us. Otherwise the facts, as in Manton (1950), indicate that triploids are the commonest type in Europe, though they may not be uniform genetically owing to the possibility of frequent re-synthesis from the diploid by crosses with the sexual species. 4. D. villarsii. Nothing to add to the statement on p. 91. 5. Asplenium adiantum-nigrum as studied by Miss Shivas is diploid in Portugal, in the Mediterranean basin, in Madeira and in parts of Ireland. It is tetraploid in its one American locality and also tetraploid in Kenya. 6. Asplenium trichomanes has been studied in detail both bj’’ Mr. Lovis and by Dr. D. Meyer in Germany. Diploids are abundant in north-eastern Germany and in Norway. They are reported from Canada by Britton (1954) and have been detected by Mr. Lovis in recent collections from the Himalayas and Australia. Tetraploids are present both in Europe and America and perhaps elsewhere. 7. Cystopteris fragilis. The tetraploid with smooth spores has been found in Northern Europe and in Greenland. The tetra¬ ploid with spiny spores is the dominant type in Northern Scan¬ dinavia, in east and west Canada, and one plant has been sent THK IM I’ORTANCE OF FKIINS TO AN UN DFKSTAN OlNCi OF THE lUUTlSIl FEOKA IJo to US from Australia. The hexaploid occurs in southern Scandinavia and across Europe as far as Madeira. It has not yet been found in America. 8. Polypodium vulgare. Most parts of Europe are dominated by the tetraploid, though the hexaploid is abundant along the Atlantic seaboard from Portugal to Denmark and prob¬ ably S. Norway. It also occurs in the Alps. The diploid has a Mediterranean distribution closely resembling that of Asplenium adiantum-nigrum, though it penetrates northward in favoured localities such as the Rhone Valley. In America diploids occur along both the eastern and western seaboards, and there is a tetraploid of probably separate and local origin on each side of the continent (cf. Manton and Shivas, 1953; Manton, 1951). Focussing attention for the moment on Europe and summing up these various results, there are two important generalisations which arise from them. On the one hand, it is clear that in every case it is the higher numbered cytotype which is the commoner one in the particular latitudes which cross the British Isles. Secondly, in every case the lower numbered cytotype has a re¬ stricted, or even discontinuous, distribution within Great Britain, while outside it, at least as regards Europe, they all fall into one of two alternative distribution patterns, viz. ; two {Polypodium and Asplenium adiantum-nigrum) show Lusitanian affinities, hav¬ ing a continuous range for their diploids in the Mediterranean region; the remainder, in so far as the facts are known, i.e. cer¬ tainly for Dryopteris dilatata and Cystopteris fragilis and probably for D. filix-mas and Asplenium trichomanes , have boreal affinities with a continuous range for the low numbered cytotypes in the latitude of northern Scandinavia. Before interpreting these facts we need to add the evidence concerning phyletic relationship between the cytotypes which is given by chromosome pairing. For this it is convenient to quote a list (Table 3) reproduced in Manton, 1953, which sums up the facts not only for the cytotypes under discussion but also for a number of well marked pairs of generally accepted species with the 2n : 4n cytology. Table 3. Chromosome pairing in triploid ferns (not apogamous). s denotes a synthesized hybrid, the others are wild. Many trivalents: s Osmunda regalis autotriploid n pairs -f n univalents : Asplenium trichomanes (2n) x septentrionale Sl'EClKS STUDIES IN TDE BIUTISJI ELOKA uo xl. adult ermum x viride s A. adiantum-nigrum 2n x 4n. Polystichum lonchitis x aculeatuni s P. setiferum x aculeatum Woodsia ilvensis x alpina s Dryoptens filix-mas x abbreviata s D. dilatata 2n x 4n 3n univalents: Polypodium vulgare 2n x 4n The uniformity of pattern is impressive. We have only one case [Polypodium) in which our local diploid does not show signs of close genetical affinity with the tetraploid hybridising with it. All the others (and Polypodium also when crossed with some American diploids) are manifestly related to each other in a manner suggesting tliat the tetraploids are allopolyploid, in each case with the genome of the diploid used as one ancestral type. This is not a result which would necessarily be repeated in every small area which one cared to examine, for we have already reason to think that, both in Ceylon and in eastern North America (unpublished work of Mr. S. Walker on the D. spinulosa complex), hybrids showing complete failure of chromosome pair¬ ing are relatively far commoner than here. It is therefore one of the local characteristics of the west European flora. The interpretation, for Britain, would appear to be as follows. Our flora seems fairly recently to have been enriched by a large number of tetraploid forms well adjusted to our present climate and which are still closely akin to diploid components in our flora. In almost every case, however, we have only been able to trace one diploid ancestral type within our area and moreover these diploids, when closely examined, show distributional characteristics suggesting relict status in some kind of relation to recent glaciation. The Lusitanian group may perhaps be the oldest, since they are southern biotypes which could have sur¬ vived the glacial periods, or part of these, in the Mediterranean area, but which cannot now spread very effectively away from this area, since their physiological powers of adjustment are now too limited. The boreal group, on the other hand, could have entered Britain more directly in relation to the colder climatic periods, since when they seem to have persisted in mountains. All three elements, however (the tetraploids and the boreal and Lusitanian diploids) have quite definitely entered this country from outside; they have not been formed here. They have pre¬ sumably come to us via the European mainland, but it is doubt¬ ful whether any of them were formed even there, for we have clear evidence in several cases, e.g. Polypodium, D. dilatata, Cystopteris, Asplenium trichomanes, of a circumpolar range THE iMl’OIlTANCE Ob’ EEIHNS TO AN UXDEltyTANDlNL: OE THE EiU'iTSH ELOKA Ui aiiiung tile ancestral types wliich, within Britain itself, seem to be relicts. As a small island on the edge of a large land mass such a history is not unexpected, but the study of the continental main¬ land immediately adjacent to us will not carry us much further. Continental Europe, like Britain, reflects the marks of recent gross changes of chmate, but not many components of our present flora have been originated there (hexaploid Polypodium is the best known case where this has happened). It is possible that, with further search, a few more parental types will come to hght, but, as things stand at present, the rule, that one only has been found of the necessary two parents of almost all the allopolyploid immigrant species detected, apphes as much to Europe as to Britain. If it were only a single case, this would be of no con¬ sequence and one could postulate extinction of the second parent in explanation. Wholesale extinction of one parent only in species after species is, however, more difficult to accept and a more plausible explanation is that they must be sought for elsewhere. If we look at the map of the world even Europe is a very small part of the continental land masses of the northern hemi¬ sphere. Of far greater geographical extent are North America and Asia, and both these are relevant to us. North America, however, is intrinsically unlikely to solve om whole problem owing to its position at the opposite end to ourselves of the land mass joined by “Beringia” (cf. Hulten, 1937). Asia is intrinsic¬ ally likely to be far more important, both as the possible cradle for our tetraploid immigrants and also as the place which may still, perhaps, harbour their missing diploid ancestors. Further progress in our understanding of the British Flora therefore requires rather urgently some botanical penetration of the Iron Curtain. This perhaps exemplifles as well as anything could do the thesis with which my talk began. REFERENCES. Hulten, E., 1937, Outline of the history of the arctic and boreal biota. Stockholm. Manton, I., 1950, Problems of cytology and evolution in the Pterido- phyta. Cambridge. — — , 1951, The cytology of Polypodium in North America, Nature, 167, 37. - , 1953, The cytological evolution of the fern flora of Ceylon, in Symposia of the Society for Experimental Biology, 7 {Evolution). jM ANTON, I. & Shivas, M., 1953, Two cytological forms of Polypodium virginianum in Eastern North America, Nature, 172, 410. Manton, I. & Walker, S., 1953, Cytology of the Dryopteris spinulosa complex in eastern North America, Nature, 171, 1116. Meyer, D., 1952, IJntersuchungen ueher Bastardierung in der Gattuiig Aspleiiium, Bibliotheca Botanica, 30, Heft 123. 98 Sl'ECIES STUDIES IX THE URITISH FJ.OKA Dr. W. B. Turrill said that it had been assumed that tetraploids have arisen from diploids, but could it not happen that diploids arose from tetraploids and other polyploids? Prof. Manton replied that this was a theoretical possibility but the available evidence points in the other direction. Prof. T. G. Tutin said that Dactylis glomerata provided a good example of autopolyploidy, and Prof. Manton said that Biscutella was another good example. Dr. E. F. Warburg enquired whether the lecturer had found both parents in any of these polyploid ferns. Prof. Manton replied that she had — in Polystichum aculeatum. THK I’KOBLKM OF ASI’LENIUM TJUC HOMAN KS 99 THE PROBLEM OF ASPLENIUM TRICHOMANES (Exhibit) .J. D. Lovis (University of Leeds). Asplenium trichomanes L., a common and familiar fern in Britain, is very widely distributed over the surface of the world, occurring in upland districts in temperate regions of every continent. Distribution of Asplenium tncliomanes agg., compiled from floras and tlie herbaria at Kew and the British Museum. Professor I. Manton (1950) discovered that two cytological forms of A. trichomanes occur in Britain. Of these two forms, the diploid (with 2n = 72), was detected only from North Wales, and the Auvergne of France. The tetraploid (2n=144), was found to be common in Britain. Recently, these two forms have been studied in some detail by myself under the guidance of Professor Manton. Both diploid and tetraploid A. trichomanes have been found to be widespread within the range of the aggregate species. The diploid has been identified from other localities in Europe, i.e., from Germany (Meyer, 1952), Switzerland and Norway. It has also been found in Canada (Britton, 1953), from several localities about the Himalayas, and also from Australasia. The tetraploid is known certainly from Canada (Britton, op. cit.) and Europe, and probably ranges more widely through the Northern Hemisphere. In Britain the diploid is much the rarer form; as yet it has been found only in North Wales and the Lake District, although 100 SI'KCIKS STUDIKS IN THK UHITISH i'J.OKA it is most probable that it occurs in Scotland and the Border Country. It does not seem to have any close edaphic preference. This was found to be particularly clear in Norway, where it dis¬ plays rather catholic tastes, occurring both on base-poor rocks and on mica-schist and “olivinstein”, the last a rock of the serpentine group. It will probably transpire that the restriction of the diploid in Britain to submontane levels in mountainous districts is determined by the limits of its climatic tolerance. In contrast, the tetraploid is characteristically, though not exclusively, a limestone fern, and has been able to extend its range eastwards across England into the lowlands, following the construction of mortared walls. Both diploid and tetraploid are highly fertile. Meiosis is regular in both, 36 bivalents being formed in the diploid, and 72 bivalents in the tetraploid. No sign of multi valents has so far been found in the tetraploid. The diploid and tetraploid are genetically isolated by sterility in the triploid hybrid, and therefore represent distinct coenospecies according to the defini¬ tion of Clausen, Keck, and Hiesey (1945). A wild triploid hybrid collected near Dolgelly in September 1952 shows highly irregular meiosis, and complete abortion of its spores. A preliminary analysis of chromosome pairing in this plant showed a rather complex situation. Two of the three genomes present pair, but not quite completely. 34 or 35 associations are formed, of which 1-4 are tri valents. Between 34 and 37 chromosomes remain un¬ paired, and consequently appear as univalents. It is clear from this result that diploid and tetraploid are closely related, though precisely how remains open to question. Breeding work now in progress may help us to determine whether perhaps the tetraploid is an ancient autotetraploid which has so evolved as to have lost the power of multivalent formation, or else is an allotetraploid with our known diploid as one of its parents. Before considering morphological differences between diploid and tetraploid A. trichomanes, it must be emphasised that a basic difficulty in fern taxonomy is that their reproductive structures present very few characters for use by the taxonomist, in compari¬ son with the details of floral structure available in the majority of flowering plants. Many vegetative characters are consider¬ ably affected in expression by environmental influences, but fern taxonomists are frequently forced to use them to distinguish species. It is not easy to distinguish diploid and tetraploid A. tricho¬ manes on the basis of gross morphology alone. There are no obvious qualitative differences, such as sorus position or pinna- shape, which serve to distinguish other species within the A, trichomanes group. The fronds of A. trichomanes are even more susceptible than those of most ferns to modifications by environmental conditions. Scatter diagrams which illustrated this plasticity in frond form Plate 11. 4 m ^ % # »o c. *u p ^ 49 ^0 %**p ^0 0^ 0 Jl 0 c? ^tro 0 ^ n y Asplerihim trichomanes agg. Figs, n and h, permanent acetocarmine prepara¬ tions of spore mother cells in meiosis from a wild triploid hybrid found near Dolgelly, Merioneth. Magnificatinn xiOOO. Figs, c and <1, explanatory diagrams of the preceding showing univalents in outline and pairs and trivalents in black 'I’here are four trivalents in fig. c hut only one in fig. d »■ •* ■ . ^ * < -J* > %, ' <»£■ U i^'. ■i^%n '.' •: -■ -, .^'^A ■"?f'"K- V * . ."i r- ' ' '' . ■_' •- ., ., -V ■ ■ ’' ; '■ li*. , '•-*151 A ■*^' -'' iVs4: '. : ^ ^ 'f ■;L. •"■ -■ -. '* ; ;^*- .• .- . ,■>>*■>>'•, ' ■<■ ■■■■.- -S .^’...: .i4-^ -’'ft " “ T>. •><*. sJ ». •- 4 r"'/. • 0 IS > ♦’ '^• •'-• *-* '-'S^ V.T*.bj»'.> • v "*• 'StCHI^ •- * -V ’ ,;. -/.A ^r?- •■■' '{w* . . - ' -f _» •* ’ . I aV* >:. --I ■>. •; -v. -3-! ' ■ ‘ *■ i -» '■'•I , n ' — - : 4t. '■.r- - ., f ’• V-**. 'V'.i'*' THK PROBLEM OF ASF’LEXIUM TRICHOMANES 101 were exhibited at the Conference, and a sample is included here (fig. 16). These scatter diagrams show that both length and area of pinnae, considered in relation to the distance between pinnae, tend to be greater in the tetraploid than in the diploid, although even when single plants are considered there is an overlap, which makes this character valueless as an absolute criterion. O 1 o 2 O < A3ER FALLS. CAERNARVONSHIRE DIPtOlO * TETRAPLOID O OO ^ o ° X X X 2 i oo O 8 O X o ^ X S X PASS OF LLANBERIS DIPLOID TETRAPLOID O O o O , \ X 3 4 5 O O O o 7 0 AVfRAGE distance BETWEEN PINNAE IN MMS AVERAGE DISTANCE BETWEEN PINNAE IN MMS Fig. 16. Scatter diagrams to show pinna size (length and area) plotted against pinna¬ spacing in individual plants of diploid and tetraploid from two localities. Each entry represents the average for the middle ten pinnae of one frond, the same fronds being used for both area and length measurements. Only one plant of each type is represented in the diagrams for Aber Falls but those for Llanberis contain entries from two diploid plants and one tetraploid. It is possible to distinguish diploid and tetraploid satisfactorily on the basis of spore size. The standard used has been at least 50, preferably 100, determinations of exospore length for each plant. Average spore lengths in diploid plants range from 29-33/x; in tetraploid plants from 35-42/z. A typical histogram derived from spore measurement of a diploid and a tetraploid plant from the same locality is illustrated here (fig. 17). Other characters which are usually affected quantitatively by increase in level of polyploidy, e.g., stomata size and epidermal cell size, would probably also serve as satisfactory^ criteria for distinction, but the measurement of spore size is more convenient in practice, especially for the determination of herbarium material. Conclusion. It is already clear that diploid and tetraploid Asplenium trichomanes display the fundamental characteristics of two dis- 102 SI’KCIES STUDTES IX THE BRITISH FLORA Fiff. 17. Spore length measurements (taken as in inset diagram) of one hundred spores from each of a diploid and a tetraploid plant from the pass of Llanberis. tinct natural species: 1, they are reproductively isolated', 2, pos¬ sess distinctive geographical and ecological distributions’, and 3, can he distinguished morphologically (although admittedly not with ease). It is nowadays generally agreed that the genetic criterion is of primary importance in the delimitation of specific limits within a group. Nevertheless, there is still a tendency to disregard, on considerations of practical convenience, ^natural species which cannot be distinguished with certainty without recourse to the microscope, and which therefore cannot be determined with com¬ plete confidence in the field. But the lack of easy and convenient morphological distinc¬ tions should not be an inducement to ignore evidence that what has previously been regarded as one clearly defined species in Floras and herbaria is in fact composed of two separate natural species. Numerous examples of polyploid series within familiar Linnean species are now known, sufficient for the question of the status of the units within these complexes to represent a taxonomic problem which urgently requires clarification. REFERENCES. Britton, D. M., 1953, Chromosome studies on ferns, Am. J. 40, 575-583. Clausen, J., Keck, D. D. & Hiesey, W. M., 1945, Experimental Studies in the nature of species, Puhl. Carneg. Insfn., no. 564. Manton, I., 1950, Problems of Cytology and Evolution in the Pterido- phyta. Cambridge U.P. THE PROBLEM OF ASPLENIUM TRICHOMANES 103 Meyer, D. E., 1952, Untersuchungen ueber Bastardieriing in der Gat- tung Asplenium, Bihliotheca Bofanicd, 30, Heft 123. Prof. D. H. Valentine enquired whether A. frichomanes as found in Australia diflPered from the European plants. Dr. R. Melville remarked that the distribution of the aggregate species extended to Tasmania and New Zealand. (Since this paper was read to the Conference, Brownlie (1954, Tm/os, Boy. Soc. 82, 665-666), has recorded that in New Zealand Asplenium trichomanes has 2n = 108, i.e., is hexaploid. Further com¬ parative studies are now being made which include this most interest¬ ing plant. J. D. Lovis. June 1955.) 104 SPECIES STUDIES IN THE BRITISH FLORA THE TWO SUB-SPECIES OF ASPLENiUM ADIANTUM-NIGRUM L. IN BRITAIN (Exhibit) Miss M. G. Shivas (University of Leeds). A plant of Asplenium adiantum-nigrum ssp. onopteris (L.) Heufl. was collected in July 1951 at Glencar, Co. Kerry. On investigating this plant cytologically, it proved to be a diploid. Fig. 18 shows a silhouetted frond of this plant with one of a tetraploid plant of A. adiantum-nigrum L. ssp. adiantum-nigrum for comparison. The tetraploid also came from an Irish locality, Graiguenamaugh, Co. Carlow. Fig. 18. Asplenium adiantum-nignim L. (a) ssp. adiantum-nigrum. Graiguenamaugh, Co. Carlow. Tetraploid. (b) ssp. onopteris (L.) Heufl. Glencar, Co. Kerry. July 1951. Diploid. THE DIIYOI’TERIS SPTNUEOSA COMPLEX IN EUROPE lOo THE DRYOPTERIS SPINULOSA COMPLEX IN EUROPE (Exhibit) • S. Walker (University of Liverpool). In modem floras the D. spinulosa complex is represented by three recognised species and the hybrids between them. These may be listed (Christensen, 1905-6, 1906-12): (i) Dryopteris cristata (L.) A. Gray. (ii) Dryopteris spinulosa (Miill.) Watt. (hi) Dryopteris dilatata (Hoffm.) A. Gray. (iv) X Dryopteris uliginosa (Newm.) Druce {=D. cristata X D. spinulosa). (v) D. dilatata x D. spinulosa. D. cristata, D. spinulosa and D. dilatata were first given separate recognition during the latter half of the 18th century, but many times since they have been considered merely as sub¬ species or as varieties of a single species (Newman, 1844 and 1865; Druery, 1912). The three species differ morphologically (viz. pinnation, shape of frond, rhizome habit and colour of ramenta) and in habitat, but are cytologically uniform; each is tetraploid with 2u = 164 (Manton, 1950). Cytogenetical investigation has confirmed D. cristata, D. spinulosa and D. dilatata as distinct species, each being an allo- tetraploid, but interrelated by common ancestral diploids. One of these ancestors is at present represented in Europe by a diploid form of ‘^D. dilatata” and found in Norway, Sweden, Switzerland and Scotland. It is suggested that the diploid ‘*D. dilatata” should be recog¬ nised as a distinct species, though morphological differences from tetraploid D. dilatata are not always evident since the tetraploid itself is very polymorphic. The degree of pinnation and shape of the basal pinnae are a good guide but probably the final dis¬ tinction will rest with spore characters. This work will be published in detail elsewhere. (See W at sonia, 3, 193-209.) Literature cited. Christensen, C., 1905-6, Index Filicum-. Copenhagen. - , 1906-12, Index Filicum, Supplementum I. Druery, C. T., 1912, British Ferns and their varieties. London. Manton, I., 1950, Problems of Gytoloqy and Evolution in the Pterido- phyta. Cambridge. Newman, E., 1844, History of British Ferns, Ed. 2. London. - , 1865, History of British Ferns, Ed. 4. London. 106 SI’ECIRS STUDIES IN THE BRITISH FLORA Mr. J. OuNSTED asked whether a hybrid was known between Tf. cristata and D. dilatata. Dr. Walker replied that this hybrid was not known to occur in nature. Dr. R. Melville enquired whether Dr. Walker had any evidence of a dine in diploid Dryopferis. For comparison he instanced Ulmus coritana, where the two ends of the dine might be taken for different species. Dr. Walker replied that he did not know of a dine but had in mind the possibility that two diploids, which were very different morphologically, might have a common ancestor. The Rev, E. A. Elliot asked what was the relationship of D. aemvla to the complex which had been discussed. Prof. Manton said that it was diploid. She had asked Dr. R. Lloyd Praeger whether in his wide experience of this species in Ireland he had seen anything likely to be a hybrid of afmula, but he replied that he had not. CALTHA IN THE BRITISH FLORA 107 CALTHA IN THE BRITISH FLORA (Exhibit) G. Panigrahi (University of Leeds). Caltha palicstris L. has posed a serious problem to taxonomists over the last two hundred years and, more recently, the genus has proved to be very complex cytologically. Our present knowledge of the somatic numbers of chromosomes of non-British materials of Caltha reveals that there is an aneuploid series includ¬ ing polyploid types with a basic number of 8 (c/. Leoncini, 1951). After Linnaeus founded the species Caltha palustris in 1753 (Linnaeus, 1753), Miller (1768) in this country, was the first to propose Caltha minor for plants having round, crenate, heart- shaped leaves and a smaller flower as a separate species from C. palustris also possessing round, crenate leaves but with a larger flower. All subsequent authors, however, have treated C. minor Mill, as a mere variety or subspecies of C. palustris L. Forster (1807) considered a plant sent to him by Mr. J. Dickson from Scotland to be distinct from C. palustris and described a new species, C. radicans, having triangular, crenate leaves and rooting at the nodes of the aerial, procumbent stems as its chief dis¬ tinguishing characters. Beeby (1887) proposed a variety zetlan- dica, for plants of C. palustris from the Shetland Islands which had radical leaves like those of Caltha palustris but rooted at the nodes like C. radicans. Later authors have considered Beeby’s zetlandica to be a form or variety of radicans on the basis of the rooting nodes. Thus British taxonomists have recognised the existence of at least three taxa, viz. lowland palustris, montane minor and Scottish radicans. They differ as to whether radicans is a separate species or a variety of palustris L. and whether minor is to be treated as a variety of the lowland palustris or as a subspecies including radicans. In this last treatment, Beeby’s zetlandica is submerged in the variety radicans (cf. Clapham 1952). The morphological features believed to separate these taxa refer mostly to vegetative characters. They appear to be quanti¬ tative rather than qualitative in nature and are subject to con¬ siderable variation. It appears that, whether radicans is treated as a distinct species or simply as a variety of Caltha palustris, there is general agree¬ ment that material exactly matching Forster’s description and drawing is rare in the British flora. There is greater inclination amongst those who have studied the species to use rooting at the nodes of the aerial procumbent stems as the chief distinguishing feature of radicans rather than leaf shape. Praeger (1934-35, 108 SPECIES STUDIES IX THE BRITISH FLORA 1951), who considers both these features important, has proposed additional characters to distinguish C. radicans from C. palustris in Ireland. There, according to him, C. radicans of Forster re¬ presents a mass of hybrids between the typical C. palustris and an extreme form, which he refers to as C. radicans sensu Babing- ton. He not only considered that the true radicans “was being hybridised out of existence” in Ireland, but believed that the “varieties” of C. palustris which appear in the British and conti¬ nental floras also owe their origin to this kind of hybridisation. Counts of the chromosomes in the root tips of some British collections have now been made. Although based on a single plant in each case, they are of considerable cytological interest in recording the highest number of chromosomes for Caltha so far reported, and are tabulated below. Cytology of Caltha in the British Flora. Names of taxa Locality 2n 1. C. palustris subsp. palustris Golden Acre 56 (near Leeds, West Yorks.) Dungeon Ghyll 56 (Westmorland) Kettlewell (West Yorks.) c. 80 2. C. palustris subsp. minor var. Bow Fell 64 minor (Westmorland) 3. C. palustris subsp. minor var. Whiteness Parish 72 radicans Mainland, Shetland Isles The present investigation showed that plants representing the three taxa are self-incompatible and that agamospermy does not appear to occur as a reproductive method in Caltha. With only two exceptions, every attempted crossing within subsp. palustris and also those between palustris and each of the varieties, radicans and minor, resulted in the development of mature follicles with abundant seeds in them. In spite of the good seed-setting, however, the percentage germination of these seeds is rather low and shows a wide range of variation. Study of the relatively few hybrids obtained will be extremely interest- ing. In the material which we have examined, more often than not the cauline leaves of Caltha palustris are deltoid or deltoid-reni- form, whereas the radical leaves of the same plant are cordate with CAI/rJlA IX Till*; JJKITISH FLOllA iOl' basal lobes wliicli may be overlapping or parallel to the petiole. Nevertheless the variability of the angular divergence between the basal lobes of radical leaves, their size and crenations of their mar¬ gins, may be striking even in one small population of the lowland iorm of C. palustris. Occasionally individual plants having deltoid radical leaves are met with in populations of subsp. palustris where there is no likelihood of radicans having occurred, at least in recent times. Again, rooting at the nodes of aerial stems is not restricted to Caltha radicans. Some of the lowland palustris from Leeds and Scotland have rooting nodes although possessing cordate radical leaves. Lastly, three of the other characters listed by Praeger (1934- 35) (viz. unbranched rootstock, short lowermost internode and small flowers) also appear to go together in the montane var. minor, which, however, usually possesses heart-shaped radical leaves and shows rootless nodes, both in its natural habitats and in garden culture. It must, therefore, be concluded that each particular morplio- logical character is unreliable on its own for the separation of species on British material of Caltha. Only future studies can show whether there are significant statistical differences between populations in respect of a number of these characters taken together. Does, then, cytology provide any solution to this species pro¬ blem? Frankly, it does not. Caltha is characterised by the possession of an extensive aneuploid series. None of the taxo¬ nomic units has a constant chromosome number of its own. It also appears that there is no correlation of the karyotypes with latitude or with the geographical distribution of the species in general. This lack of correlation between cytology and geography coupled with the great morphological variability of the group, all suggest that for the present Caltha palustris L. may best be treated as a polymorphic species with a complex network of interrelated forms in a polyploid and aneuploid series. Despite the difficulty of reconciling morphology, cytology and geography, however, ecogeographical races within this com¬ plex do seem to exist. Hulten (1950) and some American and Canadian authors generally consider that radicans is a more northern race and is arctic-circumpolar in distribution. On the other hand palustris is more southern and is boreal-circumpolar in distribution. In subalpine situations in England one finds genetically dwarfed forms, which are contained in subsp. minor. Those who would investigate Caltha in the British Isles do so in a region of overlap between the northern radicans and the more southern palustris. This position, which is so difficult for the morphologically minded taxonomists, is stimulating to the student of populations and evolution. 110 SI’ECIKS STUDIES IN THE JHIITISH EEOllA On this ecogeographical basis, then, must rest the justification tor continuing to pursue a quest for the means of separating at least three taxa (races) in the British flora despite possible blend¬ ing between them. There is slight evidence that genetical dis¬ harmony occurs between these forms. Even if this be true, however, obligate outbreeding, coupled with the long life of individual Caltha plants (as long as 50 years, according to Syme, (1863) and the ease of vegetative reproduction, militate against a sharp separation of taxa. Grateful thanks are due to Dr. H. G. Baker for his encourag¬ ing guidance and supervision. Literature cited. Beeby, W. H., 1887, Scut. Nat., N.S., 3, 21-22. Clapham, a. R., 1952, in, Flora of the British Isles. Loudon. Forster, T. F., 1807, Trans. Linn. Soc. Land., 8, 323-24. Hulten, E., 1950, Atlas of the distribution of vascular plants in A. IF. Europe. Stockholm. Leoncini, M. L., 1951, Caryologia, 3, 336-50. Linnaeus, C., 1753, Species PI'antarum, 1, 558. Holmiac. Miller, P., 1768, Gard. Dictionary, ed. 8. Yverdiiii. Praeger, R. L., 1934-35, Dish Nat. J., 5, 98-102. - , 1951, Proc. Boy. Irish Acad., 54B (1), 6. Syme, J, T. B., 1863, English Botany, ed. 3, 1, 49-53. Loudon. STELLAIUA XEMOilUAl AND THK SPECIES CONCEPT 11 1 STELLARIA NEMORUM L. AND THE SPECIES CONCEPT Ch. H. Andreas (University of Groningen). The Species Concept in its relation to the British Flora being the subject of the lectures at this conference, I shall try to say a few words using Stellaria nemorum, a species common to both Great Britain and tire Netherlands, as an example. It is neces¬ sary to state that in the early part of this paper I am using the name in a wide sense, for a closer study of Stellaria nemorum in the Netherlands in recent years has shown that we should divide the species into two taxa. The decision as to the taxo¬ nomic rank of the two divisions justifies the title of my paper. Stellaria nemorum is a plant of rare and local occurrence in the Netherlands. With only few exceptions, known from her¬ baria of earlier years and which I shall not discuss today, it is restricted to the eastern part of the country, mainly to the very south-east (southern part of the province of Limburg) and the middle north-east (near the village of Norg in the province of Drente). When collecting specimens in Limburg and Drente, I found them to be of different appearance, especially on account of the indumentum and the length of the leaf stalks. Further investigation, however, brought to light many more differences in morphology, biology and probably even in ecologj^ and genetics. Moreover, in the Netherlands a geographical barrier between the two taxa apparently exists. Yet we cannot verify such a barrier when studying the distribution of the two groups in a wider sense, that is in their European area. Both forms occur all over Europe; only in Scandinavia one of the tw^o goes a little further north than the other. As it is sometimes difficult to decide between geographic and ecological differences, we feel inclined to ask whether the dis¬ tribution pattern in the Netherlands does not mean an ecological barrier instead of or together with a geographical separation. The two Dutch habitats are very different from an eco- sociological point of view. The southern habitat in Limburg is a Querceto-carpinetum filipenduletosum of the alluvia along little streams. This signifies a very rich, soft, humus-like soil with pH between 6 and 7. The more northern habitat near Norg (Drente) is a transition between Querceto-carpinetum Stachye- tosum and Querceto-sessiliflorae-Betuletum molinietosum, signi¬ fying vegetation on more acid, rough humus-like soil than that of the normal humid Querceto-carpinetum^). Thus, in the light of ecological considerations we might conceive the two taxa as ecotypes or even ecospecies in the sense of Turesson. iMy thanks are due to Dr. V. Westhoff for sociological information. SI’KCIKS STUDIKS IN THK 15RIT1SH FLORA We now turn to a study of biological and genetical barriers. In the Netherlands there is a difference in time of flowering of about three weeks between the two taxa in their natural habitats. This difference might be caused by the distance of nearly 200 miles in a north-south direction between the two localities, as it can be observed in many species in our country. From growing the two types side by side in the ecological garden of the Univer¬ sity of Groningen, however, we learned that this explanation does not hold for Stellaria nemorum, that the difference in flowering time has a genetical base and that it continues to exist under uniform conditions. The phenomenon is of great importance in nature, as it generally will prevent cross-pollination, even where distributional areas of the two taxa might meet, and this is probably not impossible in certain European localities. The next step is to study cross-pollination and to determine whether or not cross-fertilization will result. Experiments have been made thus far on a very small scale only. The first results, however, point to incompatibility between the taxa or even impossibility of hybridization. If, after further investigation, the questions on ecological and genetic barriers could be answered in the affirmative, as we expect, we should not hesitate to give those taxa species rank on the strength of morphological, biological and probably also ecological and genetic considerations. In many cases we might wish species limits to be as obvious. Nomenclature is still confused and discussion is not essential here. The current naming in the rank of subspecies is Stellaria nemorum subsp. montana (Pierrat) Murbeck for our southern and S. nemorum subsp. glochidosperma Murbeck for our northern type. But it is not quite clear from Pierrat’s description which form he actually meant when describing Stellaria montana as a species and I have not yet seen any material named by him. When comparing the species of this genus, we might feel in¬ clined to unite the two subspecies treated above. Many workers would consider them closely related, very often on the strength of external morphological likeness, that is, of their phenotypes — which certainly is not always a good basis for estimating relation¬ ships — and so might come to the conclusion that the two sub¬ species have arisen from one parent-species, or the one subspecies from the other. This seems to be a common way of thinking. In taxonomic, geographical and cytogenetical publications, we find one species described as easily derived from another, still living species, notwithstanding the fact that in many cases the data for such a conclusion apparently are quite insufficient. We cannot do without good hypotheses as stimulating factors in the develop¬ ment of science, but before they have been turned from theories into facts Ave should be aware that they are only hypotheses, not more. Gradual divergence is often seen as a principal process in species formation, but it is not the only one. We have good evidence of the origin of species after interspecific hybridization STELLARIA NEMORUM AND THE SPECIES CONCEPT and chromosome doubling. Modern species concepts centre about intercompatibility, the degree of which, however, need not always be correlated with that of morphological equalness. We cannot always predict results concerning phenotype and hybrid sterility, even in an experiment where we know both parents well. One example is the extravagant hybrids known within the genus Salix. While considering the two taxa of Stellaria nemorum s.l. as different species, we should not give the concept a phylogenetic interpretation. In more modern terminology we should call the two taxa not species, but probably “ecospecies” (Turesson) or “commiscua” (Danser). The two terms have the same meaning and are more or less of the same age (25-30 years). Most workers in the field of plant taxonomy are likely to prefer the term eco- species over Danser ’s commiscuum. Yet it is worth comparing the two terminologies. Though Danser was aware of ecological influences in taxon formation, he laid stress upon the biological aspects of interfertility and intersterility even in the terms “comparium”, “commiscuum” and “convivium” which denoted his taxa. Turesson did the reverse; his terms “coenospecies”, “ecospecies” and “ecotype”, though running more or less parallel with Danser’s, lay stress upon ecological influences. The coenospecies, however, also has a phylogenetic meaning, while in the lower categories of the con¬ vivium on one side and the ecotype on the other, Danser’s term has a wider meaning than the purely ecological indication of Turesson. When preferring, therefore, two of Danser’s terms over those of Turesson, we might also choose the third and use the term commiscuum, which, as an indication of the taxonomic unit he had in mind, is not less valuable than the word ecospecies. We expect the two subdivisions of the present Stellaria nemorum to represent two commiscua, every commiscuum being a group of forms which hybridize freely and yield fertile progeny. Preliminary results on intercommiscual hybridization were negative in the sense that they did not even yield sterile hybrids. To conclude from this that each of the two commiscua, though morphologically much alike, is a comparium of its own, would be highly premature. Fieldwork has to be continued on a larger scale. Though Danser recognized the processes of evolution, it is essential that he kept his species concept free of phylogenetical interpretation, which he often thought too speculative. Instead, he denoted his taxa com¬ parium, commiscuum, convivium in a way which primarily points to biological processes, so closely related to the gaps between the species as we generally know them. Whatever terminology we may choose, we can say that many taxonomists of to-day delimit their species in such a way that they are commiscua or ecospecies. As a consequence, taxonomy is no longer merely theoretical; experimental work can make its con¬ tributions as an objective tool in the hands of taxonomists. We must, however, keep in mind that nature’s ways are always com¬ plex in the field of multiformity in plants. I might point to 114 Sl’ECIRS STUDIES IN THE BRITISH FLORA an article by C'amp and Gilly in Brittonia, 1943, which, I think, is very interesting, distinguishing not less than 12 types of species. Experimental work has shown that hybrids are not always fuUy fertile nor completely sterile. The difficulty of interpreting experimental results lies in their complexity and in the fact that the delimitation of species on the strength of hybridization as well as on other criteria, is not only a matter of quality, but also of quantity. However, we may say that the species concept has gained a sound and biological basis, although the species them¬ selves may change in the course of evolution in a changing w^orld, so that the species of the future may not be the same as those of to-day, or of the past. REFERENCES. Camp, W. H. & Gilly, C. L., 1943, The Structure and Origin of Species, Brittonia^ 4, 323-385. Danser, B. H., 1929, Ueber die Begriffe Komparium, Kommiskuuin und Konvivium, und iiber die Entstehungsweise der Konvivien, Genetica, 11, 399-450. - , 1950, A Theory of Systematics. — Leiden. Hegi, G., 1911, Zwei Unterarten von Stellaria nemorum L., Mitt. d. Bayer. Bot. Ges. 2, 340-341. - , 1911, III. Flora v. Mittel-Europa, 3, 350-352. Murbeck, S., 1890-91, Beitrage zur Kenntniss der Flora von Siidbosnien und der Hercegowina, Acta TJniv. Lund, 27. - , 1899, Die nordeuropaischen Formen der Gattung Stellaria, Bot. Notiser, 1899, 193-218. PiERRAT, D., 1880, Note sur le Stellaria montana Pierrat sp. nov., Comptes-rendus de la Soc. Bot. Bochelaise, 2 (1879), 58 (La Rochelle, 1880). Turesson, G., 1929, Zur Natur und Begrenzung der Arteinheiten, Hereditas, 12, 323-334. Mr. J. E. Lousley enquired whether Dr. Andreas had seen any material of subsp. glochidospernia from Britain. Dr. Andreas replied that she did not know whether this occurred or not*. Mr. R. D. Meiki^e said that Mr. D. N. McVean had told him that glochidosperma did occur, but was confined to Wales. *See Watsonia, 3, 122-6, t. II (1954)— Editor. Cr.INAL VARIATION IN FLOWER SIZE IN LOTUS CORNICULATUS 116 CLINAL VARIATION IN FLOWER SIZE IN LOTUS CORNICULATUS L. (Exhibit) B. A. PouLTER (Royal Botanic Garden, Edinburgh). A series of graphs and diagrams was shown to demonstrate the possible occurrence of a dine in flower size in Lotus cornicu- latus L. within the British Isles. The measurement used as an indication of flower size was the length from the base of the calyx to the tip of the wing, the range being from 8-25 mm. to 20 mm. The axis of the dine was from S.E. to N.W., the smaller-flowered forms being most numerous in S.E. England, and being replaced by larger-flowered forms towards the north and west. Before this dine can be fully substantiated, much more evidence is required. In particular, it will be necessary to deter¬ mine as far as possible to what extent the variation is on a geographical basis, and how much is a response to variation in ecological conditions. Examination of populations growing under varying ecological conditions within small geographical areas may help to elucidate this point. 116 SI’KCIKS STUDIES IN THE J3KITISH FLORA THE SECTION EU-CALLITRICH E IN THE NETHERLANDS (Exhibit) Henriette D. Schotsman (University of Groningen). This exliibit included photographs of plants in various inodi- hcations, drawings of fruits and flowers, distribution maps and cytological peculiarities of the five Callitriche species of the section Eu-Callitriche in the Netherlands, viz. C. hamulata, C. obtusayigula, C. stagnalis, C. platycarpa and C. palustris. Most of these species are very polymorphic, sometimes resembling each other very closely in the vegetative parts. This makes it often rather difficult to distinguish them with certainty in the field without cultivating them and without cytological investigations. On account of the polymorphy many difficulties have arisen in nomenclature. Moreover, varieties of each species have been described. From culture experiments under various conditions it appeared, however, that most of these varieties are modifications only. Based on cytological, morphological, geographical and eco¬ logical characteristics the section Eu-Callitriche in the Nether¬ lands can be divided into the five species mentioned above. C. hamulata Ktzg. As a water form, this species is a winter annual. The plants grow quickly in autunm, producmg linear leaves only. When the water is not too deep, in spring floating rosettes with spathulate leaves are formed. The flowers are totally submerged and consequently pollination also takes place below the water surface. The anther turns towards one of the stigmata and after dehiscence the pollen-grains — without exine — germinate quickly, the pollen tubes forcing their way through the stigmatic tissue. C. hamulata occurs in dune sandy soils and pleistocene sandy soils. The water is slightly acid to neutral. In its chromosomes C. hamulata is different from the other species. The number of chromosomes is 2n = 38; it is the only species having chromosomes with median centromeres. C. obtusangula Le Gall. This species is a perennial and gener¬ ally occurs in slightly brackish water in the coastal region. The rhomboid leaves, the fruit with very blunt edges, the unwinged seeds and the ellipsoidal pollen are characteristic for this species, as shown in the exhibit. The number of chromosomes is 2n= 10. In the province of Zeeland we found plants with a different leaf shape, possessing 2 ASAT-chromosomes. These plants proved to tolerate a higher salt concentration. C. stagnalis Scop. C. stagnalis is not very common in the Netherlands. It generallj^ occurs in valley regions and in estuaries and is rather salt-tolerant, like C. obtusangula. The THE SECTION ELT-CALLITRICHE IN THE NETHERLANDS 117 number of chromosomes is 2n=10; the chromosomes are smaller than those of C. obtusangula. Plants were found with 0, 1 and 2 AS'^7^-chromosomes. C. stagnalis is the most constant species under varying conditions; linear leaves are never formed. C. platycarpa Ktzg. C. platycarpa is a very polymorphic species. Especially as a land form or in shady localities it may resemble C. stagnalis. In winter (C. platycarpa is perennial) linear leaves are formed and in this period it resembles C. hamu- lata and C. obtusangula. It is common in the Netherlands, occur¬ ring in slightly acid to slightly alkaline water on different soils. Only in brackish water is the species absent. The number of chromosomes is 2n = 20. C. palustris L. This arctic-alpine species is rare in the Nether¬ lands. It is a summer annual and occurs in “cowpuddles” in brook- valleys. As a water form, it possesses more or less normal flowers, though the act of dehiscence is often lacking. In well developed forms normal air pollination may occur, but as a land form the species is apogamous. The male flowers are abortive or totally absent, and the stigmata have been nearly wholly re¬ duced. The small, obovate, blackish fruits are characteristic of this species. The number of chromosomes is 2n = 20; these chromo¬ somes are the smallest in this section. 118 SPECIES STUDIES IN THE HRITISH FLORA ECOTYPICAL VARIATION IN AOOXA MOSCHATELLINA L. (Exhibit) H. G. Baker (University of Leeds). Adoxa moschatellina is the only species of the only genus in the isolated family Adoxaceae. Although circumboreal in its distribution it lacks discrete morphological variation and only a single, unsatisfactory variety (var. inodora Clarke from Kash¬ mir) has been described apart from the type. Adoxa has the appearance of a species which has ‘'run out of morphological variation”. It is interesting, therefore, to see if it retains the potentiality of “physiological” variation which may be adaptive. Plants collected from deep, damp crevices in the Yoredale limestone cliffs near the summit of Ingleborough (at c. 2,150 feet) regularly come into leaf and flower later than lowland (wood¬ land) material from several parts of the British Isles when kept in an unheated greenhouse at Leeds. Although, in general, subalpine races of plants come into flower earlier than lowland races when these are grown together in a garden, the reversal in this case may be related to the wood¬ land habitat of the lowland race. There the emphasis is upon early development before the trees come into leaf. Material from about 1,000 feet elevation on Ingleborough (growing amongst Carboniferous limestone rocks amidst scrubby trees) is intermediate in behaviour. There is a complete dis¬ junction between the plants at this elevation and those near the summit. All races shown have the same chromosome number (2u = 36) and have been proved to be self-incompatible. Vigor¬ ous vegetative reproduction accompanied by self-incompatibility is probably responsible for the “shy seeding” of this species. VARIATION IN CENTAURIUM IN WEST LANCASHIRE. 119 VARIATION IN CENTAURIUM IN WEST LANCASHIRE (Exhibit) WiNEFRiDE M. T. O’Connor (University of Liverpool). An investigation of the species of Centaurium on the dunes at Freshfield, Lancashire (v.c. 59), is in progress. The commonest species which occur are Centaurium minus Moench and Cer\- taurium littorale (Turner) Gilmour. Centaurium pulchellum (Sw.) E. H. L. Krause has been recorded but not recently (Green, 1933), while the only known stations for Centaurium latifolium (Sm.) Druce were in this area. The latter was last recorded in 1871 (Green, 1902; Stansfield, 1936) and is now believed to be extinct. This investigation has, therefore, been concentrated on the two former species. Both species are very variable and a list of the described varieties (Gilmour, 1937) together with their discriminatory characters is given in Table 1. The varieties concerned in the populations studied at Freshfield are Centaurium minus var. fasciculare and Centaurium littorale var. occidentale. Centaurium minus is a plant of dry grassland, dunes and clear¬ ings in woods. It is widely distributed in England and Wales but is less frequent in Scotland. Centaurium littorale is much more restricted. It occurs in coastal districts of northern England, North Wales and Scotland. The distribution of both species in Ireland is inadequately known from available herbarium material. Distribution maps are given in figs. 19 and 20. Further details and records with voucher specimens for examination would be wel¬ come. Mixed populations of the two species occur in certain areas of the coast of northern England and Wales, and the populations studied lie within this area. At Freshfield there are distinct populations of each species, and other populations which con¬ tain both species, together with plants which cannot be referred to either. Similar plants were found by Salmon and Thompson (1902) at Ansdell, near Lytham, v.c. 60, and described by \^el- don (1902) as Erythraea lit t oralis intermedia {Centaurium x intermedium (Wheld.) Druce).* Wheldon’s specimens from High- town, v.c. 59, near to Freshfield, are in Herb. Mus. Brit. •Since this paper was submitted it has become clear that the plant originally described by Wheldon as F. UttornHs var. intermedia is in fact a polyploid form of C. littorale. The status of this form will be discussed more fully elsewhere. 120 SPECIES STUDIES IN THE BRITISH FLORA Centaurium Centaurinm TABLE 1. General var. minus ( = var. cen¬ taurium (L.)) 1. Basal leaves obo- vate-oblong. 2. Stem simple or Leaves 1-5 cm. x (4) branched. 18-20 mm., obovate to 3. Stem leaves ellip¬ elliptical, prominent¬ tic oblong minus ly 3-7 veined. Calyx less than cor¬ 4. Flowers sessile or olla-tube at anthesis. subsessile in fair¬ Corolla limb flat. ly compact clus¬ ters. 5. Calyx less than corolla-tube at an¬ thesis. 6. Not scabrid. vnr. littorale (Turner) Gilmonr 1. Stem 2-8 cm. siib- glab rolls. littorale Leaves 1-2 cm. x 3-5 mm., linear-spathii- late to ligulate (nor¬ mally parallel-sided). Indistinctly 1-3 veined. Calyx equalling cor¬ olla-tube at an thesis. Corolla limb concave. 2. Basal h’^s. obovate spathulate; cau- line Ungulate, lan¬ ceolate or ovate- lanceolate. 3. Flowers in lax or ± compact few- flowered cymes. 4. Sepals linear not attenuate, nearly or equal to corolla- tube at anthesis. VARIATION IN CENTAURIUAI IN WEST LANCASHIRE, 121 (after Wheldon and Salmon, 1925) var. fascicular e (Duby) var. suhcapifafum (Corb.) var. suhlitorale (Wheld. & Salm.) 1. Basal leaves ovate- oblong, 2. Simple or branches from i up stem. 3. Stem leaves ellip¬ tic-oblong, often 5- nerved. 4. Firs. numerous, sessile or subses- sile in compact clusters. 5. Calyx less than corolla-tube at anthesis. 6. Not scabrid. 1. Basal leaves broad often suborbicu- lar. 2. Short, stout, 3-10 cm. 3. Stem leaves ellip¬ tic-oblong, often o- nerved. 4. Firs, sessile or sub- sessile in crowded compact heads. 5. Calyx less than corolla-tube at anthesis. 6. Not scabrid. 1. Basal leaves ob¬ long or ovate- spathulate. 2. 10-30 cm. 3. Linear to ovate- lanceolate. 4. Firs, in terminal few-flowered cymes. 5. Calyx not quite as long as corolla-tube at anthesis. 6. Slightly scabrid. var. occidentale (Wheld. & Salm.) 1 . Stem 2-25 cm . Wings usually scabrous-ciliate. 2. Basal leaves linear to linear-obovate or spathulate, cau- line erect ligulate, margins scabrous. 3. Flowers few in fastigiate cymes. 4. Sepals linear, at¬ tenuate, equalling or exceeding cor¬ olla-tube at an¬ thesis, densely scabrous. var. haileyi (Wheld. & Salm.) Gilmour 1. Stem 4-10 cm. (branched plant usually as broad as tall) usually scab¬ rous. 2. Leaves linear to linear-spathulate . 3. Flowers numerous, crowded in com¬ pact heads. 4. Sepals linear, at¬ tenuate, equalling or exceeding cor¬ olla-tube at an¬ thesis, densely scabrous. var. minus (Hartm.) Gilmour ] . Stem 2-5 cm . Glabrous. 2. Basal leaves spath¬ ulate, cauline nar¬ row, linear. 3. Flowers in com¬ pact cymes. 4. Sepals linear, at¬ tenuate, equalling corolla-tube at an¬ thesis, smooth or rarely slightly scabrid. i 122 SPECIES STUDIES IN THE BRITISH FLORA 'd c o ^ c o Ol A iv i c o a o VI . <1-1 c; C !; o s o t: <*-i ^ o ?) O) +J -3 <)j •c e«i (i) 0<5 s _ s i£ *«»d :3 - 3 c e a. < K S ^ o 2 <1-1 S O V fj u C ^ oj £ £5 < •cH ee ^ 3 C/3 w < i3 «» ■^3 £5 a 05 .5f pti o) ?: ni 1^ in w e c3 iCj s. ^ -t-^l ►— £ c o (U 73 3 I— H w 3 cn n a <0 in a> £: H I VARIATION IN CENTAURIUM IN WEST LANCASHIRE. 123 The most useful characters for discriminating between the two species are : — (a) Length, breadth and shape of leaves. In Centaurium littor- ale the leaves are narrow and approximately parallel-sided, in Centaurium minus they are broader and ovate to ovate- lanceolate. (b) Length of inflorescence and number of flowering nodes. These characters give a measure of the difference in the amount of contraction of the inflorescence in the varieties of Centaurium minus and Centaurium littorale at Freshfield. They do not hold for the whole range of variation of the species. Centaurium minus var, fasciculare is characterised by a contracted inflorescence involving a number of nodes of the main axis. Centaurium littorale var. occidentale has a loose cymose inflorescence composed of laterals from only one or two of the main axis nodes. (c) Relative length of calyx and corolla tube. The calyx nor¬ mally equals or exceeds the corolla tube in Centaurium littorale. In Centaurium minus it is shorter than the corolla tube. These characters are scored in populations subjectively de¬ termined as pure or mixed and the results plotted as a scatter diagram using the following indices : — (a) (Leaf length / length from base of leaf to widest point) x breadth of leaf. (b) Length of inflorescence / number of nodes in inflorescence. 3-8 3 0 • • X A A • ^ A A AA 2 AaJa^^AA 0-5 j.o 1-5 2 0 2 5 of lnflor«x<*<)KA ]1>G PROBLEMS OF HYBRIDIZATION AND SPECIES LIMITS IN SOME ERICA SPECIES (Exhibit) Peter A. Gay (University College, London). In considering problems where hybridization is thought to occur, we should first satisfy ourselves that the system is not merely a case of wide and anomalous variation with the two putative species at opposite ends of a variation range. In the case exhibited, Enca cUiaris L. crossing with E. tetralix L., the two taxa have long been recognised, but this in itself is not a sufficient argument. However, when the geographic distribution of the two species is studied, their very different ranges, one tend¬ ing to be northerly and easterly and the other southerly and westerly, with an overlapping range between, it is at least strongly suggestive that they are two “good” species. Their ecological ranges are also somewhat distinct, although in this country at least this seems to be breaking down due to man’s action. Although no strict barriers are demonstrated above, they do exist when considering other aspects of the two species. One barrier to free gene flow is the low fertility of the hybrids. Slides were exhibited to demonstrate this. Differences in flowering time of the two species wiU render the gametes less readily available for fusion with those of the other species. Thus we are deahng with two distinct entities. The putative hybrids usually show in most of their characters a complete gradation between the two parent species, for example, in anther characters (which were demonstrated), leaf shape, corolla shape, branching type. Occasionally unpredictable forms turn up because of the totally new gene combinations. The odd corolla shapes demonstrated illustrate this. Both these states of affairs are characteristic of hybrids. A useful method of studying hybrids is Anderson’s technique involving a Hybrid Index. Nine characters were employed in the scheme here. However, for comparisons of whole populations especially in relation to habitat studies and evolution, a more complex scheme has been devised involving taking the means of both the Hybrid Index and of a new term, the Hybrid Number, for the whole population. This system has proved a very useful method of population analysis where the hybrids are involved. The old problem of the exact status of E. mackaiana Bab. was considered. Morphologically it appears to be a hybrid be¬ tween E. ciliaris and tetralix but it differs greatly from those which we know to be hybrids. Its distribution is not what would be expected from a distinct taxon except by polytopic origin. Its PROBLEMS OF HYBRIDIZATION AND SI’ECIES LIMITS IN SOME ERICA SPECIES 127 pollen fertility is low and of an odd type. It certainly does not seem to be of the same taxonomic category as the other species and a tentative suggestion put forward is that it is something between an autotetraploid and an allotetraploid arising some¬ where from the E. ciliaris / tetralix hybridization complex. Such a hypothesis will explain the above anomalies and also the fact that E. ciliaris was formerly recorded from the vicinity of the Connemara locality of E. mackaiana. However, at the moment it should not be considered as more than a tentative working hypothesis. Prof, D. H. Valentine asked how the chromosome number of Erica mackaiana was related to these of E. ciliaris and E. tetralix. Mr Gay replied that the chromosome number of E. mackaiana was not known. Mr. N. Y. Sandwith enquired about the Dorset hybrid between E. ciliaris and E. tetralix, E. x watsoni. Mr. Gay’’ said that in watsoni pollen fertility was very low, usually less than 1%. He said he did not agree with giving a name to this hybrid which was only one biotype of many. Mr. J. Ounsted pointed out that E. x praegeri was usually regarded as a hybrid between E. mackaiana and E. tetralix, which was interest¬ ing if E. mackaiana was itself of hybrid origin and had low pollen fertility. SI’J'ICIES STUDIES IN TJIE JUIITISH FLOllA THE SPECIES CONCEPT IN EUPHRASIA (Exhibit) P. F. Yeo (University Botanic Garden, Cambridge). Tliis talk deals with the taxonomic problems in Euphrasia, as w^ell as the species concept. Twenty-five species of Euphrasia were recognised by Pugsley in the British Isles. The characters which are used to distinguish them are internode length relative to leaf length, the node at which fiowering starts coupled with number of the branches, the date at which flowering starts, indumentum of leaves and calyx, shape of leaves and leaf teeth, size of flower, and shape of calyx and capsule. Each species has its own set of characters, and differs in several characters from its nearest relatives. The forms distinguished on the basis of these characters have, as a rule, distinct habitats, and more or less extensive distribu¬ tions. These distributions usually show discontinuities, due to discontinuity of suitable habitats, and the discontinuities may be quite considerable ; for example, those which separate Euphrasia montana in Yorkshire and Belgium, or E. frigida in Scandinavia, Scotland, and Ireland. Sympatric species, of which there are often several in an area, generally show differences in habitat which keep the populations of different species apart. Sometimes, however, two species grow together, and in the same habitat. It is then most often the case that one is a diploid and the other a tetraploid. In some cases, though their chromosome numbers are the same, one or both may be minute-flowered forms, and these are habitually self-pollinated. They thus remain distinct, though growing intermixed. Ecology- and distribution support the distinctions made on morphological grounds, and this is the justification for regarding these morphologically defined groups as species. This is not to say that Euphrasia taxonomy is easy, or that the group is not a critical one. There are three sources of taxonomic difficulty, which I shall now describe. The first lies in the fact that a large proportion of the features used to distinguish the species are quantitative, and are thus significantly affected by environmental and genetic variation. Regarding environmental effects, it can be said that all the differences used taxonomically are ones which have a genetic basis. But environment may accentuate them. For example, cultivation in pots suggests that E. pseudokerneri has smaller leaves than E. nemorosa, but this difference may be increased in nature owing to the extreme dryness of the chalk soil in which THE Sl’EClKS COISCEET UN EUEHltASlA i-l) jjseudokerneri usually grows. Probably the character least affected by environment is indumentum. Next comes flower size, which varies much less than the luxuriance of individuals. Another fairly stable character is average date of commence¬ ment of flowering. But this is probably controlled partly by temperature, and an early warm spell may bring a species into flower prematurely. There are two principal influences on all other characters, that is, characters of the vegetative parts. They are shade and nourishment, including water supply. The influence of shading and nourishment is shown by com¬ parison of plants grown in the garden, with those grown in pots in the greenhouse. Greenhouse plants of E. ne^norosa resembled wild ones fairly closely, as is usually the case with Euphrasias grown in the greenhouse. Garden plants, however, were very much more luxuriant, having larger, darker, shinier leaves, and being much more branched. These plants also had shorter inter¬ nodes than wild or greenhouse ones, and their branching started at a lower node. It seems that the supply of nutrients in the pot was approximately what is available in nature, while in the garden it was greater, and that the effect of the greenhouse roof in reducing light intensity was similar to that of the vegetation that surrounds the plants in the wild. The host provided for the garden plants, Plantago lanceolata, was cut back from time to time and produced little or no lateral shading. A fairly wide variation in the appearance of plants can be caused by variation in luxuriance. In luxuriant plants, leaf size and flower size are at their largest, and favourable nutritional conditions seem to bring out the greatest number of differences between forms. When nourishment is really poor the species begin to lose the characters by which they are distinguished and to resemble one another more closely. There are evidently a number of nutritional factors affecting luxuriance. Thus leaf size is not always correlated with the vigour of growth and branching of the plant. Some, which, judging by stature and number of branches, are vigorous plants, have leaves of about normal size, while others of the same stature, or even smaller plants, have relatively large leaves, dark shining green, like those of garden-cultivated plants. Small plants with large healthy leaves are ones that got established late, but when they did so, it was on a particularly favourable host plant. Two nutritional factors are, therefore, the type of host plant, and the time of establishment on the host. Euphrasias exhibit a great deal of variation of this kind because they are annuals and parasites. This means that population studies have to be carried out on cultivated plants. The genetic variation which is normally present in a species will affect its taxonomic characters, for the same reason that environmental variation does, namely that most of the taxonomic characters are quantitative ones. An example of genetic varia- 13U ai'KCiKS iSTUUlES IX THE JHUTISli EJ.OJtA tion is probably to be seen in the differences between West Country E. conjusa, and that in Derbyshire and Staffordshire. The latter tends to be more compact, and robust, not quite so freely branched, and to have larger capsules and flowers. In addition to genotypic variation within a species, there occur forms which diverge too much from recognised taxa to be ii_cluded in any of them. This is the second source of taxonomic difficulty. The exliibit shows examples of this. It is not clear whether the plants from the north of Scotland exliibited ought to be included in E. conjusa, though they have some characters m common with the south-western plant. The form shown from North Wales, however, though it has sometimes been referred to E. conjusa, lacks some of the most important characters of that species. It seems probable that all the easily recognisable British species have been described, and though some of the problema¬ tical forms may possibly have to be described as species this cannot be undertaken without careful study in the field. For example, the form from N. Wales needs to be seen at all ages, its habitat observed, and its range investigated, and attention paid to the possibility of its varying towards E. conjusa in any part of its range. In addition to variation within species, and the occurrence of forms falling outside recognised species, hybridisation causes difficulties. Species with minute flowers are habitually self- fertihsed, the anthers dehiscing and shedding pollen just before the flower opens. But even these can give rise to hybrids by the transference of their pollen to the flowers of larger-flowered species. The species that can most easily hybridise are usually kept apart by their ecological differences, but this obstacle to hybridisation is frequently overcome. Hybrids occur in spite of ecological differentiation, either as a result of two different habitats being adjacent, and perhaps intergrading, or because tolerances over¬ lap. An example of the latter is provided by E. occidentalis, which is a plant characteristic of short turf on cliff tops round the coast. In S.W. England, however, it occurs inland on lime¬ stone. As a result, it meets E. conjusa on Carboniferous Lime¬ stone in the Mendips, and E. pseudokerneri on chalk in Dorset, and in both cases hybrids are produced. Hybrids between species with the same chromosome number are fertile; they can therefore segregate and backcross, so that a great range of forms is possible among the descendants of a cross. Consequently, groups of individuals descended in this way may show a close resemblance to one of the parent species and yet be untypical of it. And it may be that while the taxonomist can recognise this, he could not say with certainty, on the evidence of such individuals alone, what was the other parent, or even whether or not they were hybrids. Another possibility is that a hybrid between species “A” and “B” may resemble “C”. TllK Si*EClE& (jO^'CErX IX EUi’liKAISlA lai All inatance of this was observed at Bettyhill, Sutherland, wliere there occur E. hrevipila var. reayensis, and the north Scottish form of E. nemorosa. The first, which is a very marked variety, perhaps deserving specific rank, has long glandular hairs, and the second is eglandular. Along the roadsides these hybridise and plants with short glands occur. A gathering of these might pass for E. hrevipila, which has short glands. These facts show that certain recommendations to botanists interested in Euphrasia are needed. Firstly, in view of the existence of local forms or hybrids, the attempt should not be made to force every population into some recognised species, buch forms may run down in a key to a particular species. These present a particular difficulty to the beginner. As far as possible one should take one’s conception of a species from specimens from an authentic locality in the species’ main area of distri¬ bution. Unfortunately, in the case of E. confusa, some gather¬ ings determined by Pugsley are widely different from the south¬ western form on which the species is based. Whatever one’s original conception of the species, one should collect and take note of divergent forms. We must be resigned to not being able to give a name to every plant, but the unnameable or untypical populations should not be ignored. Improved knowledge of variation may make possible improvements in taxonomy. The second recommendation is that anyone who collects Euphrasias in a particular area is advised to make a gathering of every form he finds in the district, if possible, whether or not it belongs to a species known to him. If he then submits them to a specialist it will make it easier for the specialist to recognise hybrids and to come to conclusions. The specialist will then find the material more instructive, and will be able to be more informative to the collector. Advice on how to select individuals of Euphrasia is given in Clapham, Tutin, and Warburg’s Flora of the British Isles. I have now described the differences which distinguish species, and the sources of the taxonomic difficulty which makes Euphrasia a critical group, and now wish to sum up the situation regarding the species concept in the group. The facts of close morphological similarity, and ease of hybridisation, apart from creating difficulties for the taxonomist, indicate close relationship between the species. Species boundaries are comparatively close together in Euphrasia, but they have been placed in the most convenient places, coinciding with the most marked morphological and ecological discon¬ tinuities. In addition to the peculiarity of Euphrasia that its species are rather “small”, there are two others, namely great variability and a high rate of endemism. This last is a general phenomenon, though particularly marked in Britain. The existence of many endemics in Britain suggests their recent and rapid evolution. l;i2 Sl’KClKS STUUIKS IN THK IJIUTISH I'LOllA and the variability of the forms is consistent with this. The species tend to be unstable, and the group is probably in a phase of rapid evolution. I have at times tried to envisage the result of lumping Euphrasias, to see whether any more satisfactory groupings would result. One could either reduce the British forms to three species, the absolute minimum, or carry the lumping to some intermediate level. I cannot see that any lumping can reason¬ ably be done. The first course would leave one species, at least, that was extremely heterogeneous, while there is no satisfactory basis for lumping to an intermediate degree. I consider that the present taxonomic situation is broadly satisfactory, though it involves a rather special concept of the species. This narrow view of the species involves less distortion of the concept, as it is generally understood, than would be caused by reducing the number of species to three, with a large array of subspecies. In connection with this talk an exhibit of herbarium sheets was shown, intended to illustrate the following points about Euphrasia confusa : — 1. The species occurs in the Peninsula counties, the Derbyshire and Staffordshire Dales, and in Yorkshire. 2. Forms occur in Scotland showing affinities with the English plant, but which are not identical with it. 3. In N. Wales there occurs a plant which has apparently been referred to E. confusa but which appears to have little in com¬ mon with it. 4. The differences that separate E. confusa from E. nemorosa and E. occidentalis, which are allied to it. 5. These two species hybridise with E. confusa. 6. The kind of differences that occur between individuals in the same population, due mainly to environment. 7. The difference in the appearance of the same form at different ages. Prof. D. H. Valentine asked how the pattern of variation in Britain compared with that in other parts of Europe and whether this country is especially rich in Euphrasia species. Mr. Yeo replied that it seemed that there are quite a lot of endemic species of restricted range in Europe and these seemed to occur especially in mountain areas. Dr. J. G. Hawkes said that in south-east Australia nearly every mountain range had its own series of Euphrasias. Mr. Yeo in reply suggested that this was particularly interesting since the Australian species belonged to a different Section of the genus from the European plants. THE SPECIES CONCEPT IN EUPHRASIA 133 Dr. E. F. Warburg said that he felt far from sure that all the recognisable species to be found in Britain liad been described. As an example he mentioned three distinct plants found in Buckingham¬ shire which hybridise and hare all been referred to E. nemorosa. He felt doubtful whether it was right to treat so many taxa as species and it might be better to lump the British plants under only three species v.’ith a large number of subspecies. This seemed desirable in view of the numerous intermediates and had the merit that it would be possible to put some sort of a name to every gathering. Mr. Yeo replied that from his experience he would have said that all easily recognisable species had now been described and he thought the treatment of existing taxa as species or subspecies was a matter of opinion. 134 SPECIES STUDIES IN THE BRITISH FLORA SOME VARIATION IN SALICORNIA AND ITS SIGNIFICANCE (Exhibit) D. H. Dalby (University College, London) The number and nature of the taxa recognised within the annual section of the genus Solicornia in Britain have varied con¬ siderably according to the differing conceptions as to what con¬ stitutes a species in this difficult genus. The earliest workers considered there were but one or two species ; subsequently the outlook has become more and more critical and the number of supposed species has risen sharply. This variable treatment of the group is a result of the fact that the various characters used in species discrimination show very few correlated discontinuities ; in general variation is continuous between extremes. The main reason for this continuous variation appears to be the extreme plasticity of the plants in response to environmental factors. Some of the effects of environment have so far been deduced from comparative field studies and from plants grown in standardised conditions. Small and more or less simple plants, usually markedly red when in fruit, and with few fertile segments, may be referred to as the gracillima-tyxye (although it is not to be assumed that they are necessarily identical with Woods’ S. gracillima). Such plants may grow in the densely colonised upper saltmarsh pans, whilst plants that are morphologically similar are often to be found along the upper margin of colonies bordering creeks, especially when the soil is rather sandv or stonv. These plants probably represent the ultimately reduced condition in Salicornia induced by intense physical competition in the one instance, and by water shortage during the growing season in the other. There is evidence also of environment having an important effect in determining whether a plant be erect or prostrate in its growth habit. When grown in saline mud in shallow tins, the progeny of about 15 plants showed little variation in growth habit when mature, even though two of the parent plants were prostrate. All the progeny were completely erect, or with a slightly bent or shortly decumbent portion below the bottom node. This bending is due to the seedling falling over at the cotyledonary stage. However, a third prostrate parent plant did produce progeny which were mostly prostrate, indicating that the procumbent habit may sometimes have a genetic basis. Small colony differences can be detected in the field when the observer is familiar with the plants, although these differences may well be difficult or impossible to measure as they often con¬ cern such attributes as branching posture, shininess of the seg- SOME A^ARTATION IN SALTCORNTA AND ITS SIGNIFICANCE 135 ment surface and subtle colour variations. The colonies so far examined in this study can be distinguished one from another by the biometrical treatment of certain characters, and it is likely that most colonies can be separated statistically if a variety of characters be used. Of necessity, the characters that can be treated biometrically are gross and rather crude, but they are none the less real. The colonies themselves, or the component parts of heterogeneous colonies, show varying degrees of separa¬ tion for different reasons. Thus the separation may be due to the differences in chromosome number between diploid and tetraploid plants. It is often found that whilst the plants belonging most nearly to the ramosissima-type (occurring at the upper edge of a colony) prove to be dinloid, those belonging to the dolichostachya- tvpe (at the lower edge) are tetraploid. Another cause may be differences in flowering time, as sometimes occurs between the stricta-type and ramosissima-type plants, the former flowering perhaps a fortnight earlier than the latter. Allied to this is the marked variation in the relative times of expansion of the stig- matic lobes and the stamens. Many plants (as at Blakeney) are definitely protogynous, every flower having receptive stigmas but no externally visible anthers. Other plants (as on Hayling Island) have their anthers and stigmas maturing simultaneously and in mutual contact. In plants such as the latter, self pollina¬ tion is very likely to occur and experimental evidence suggests that Salicornia is frequently autogamous. Seedlings from one parent plant usually resemble each other very closely, but differ from seedlings from other parent plants grown in similar condi¬ tions. The principal characters concerned here are time of ger¬ mination, cotyledon size and shape, presence or absence of antho- cyanin in the hypocotyl and number and origin of the primary branches. Autogamv will lead to the production of pure lines, and must be considered as a cause of local population or colony differentia¬ tion. Salicornia has rather special requirements as far as its natural habits are concerned, namely a freedom from competition and a soil that is both saline and permanently moist. In Britain these habitats will occur naturally only along the coast, distributed in a discontinuous linear manner. Inter-colony gene exchange would thus be limited even if the plants were amphimictic. It seems very likely that it may be impossible to determine specific limits in this genus, as a result of polymorphism and breeding behaviour. Furthermore, rather than that they exist but are obscure, it is possible that these specific limits do not, in fact, exist at all. The exact interpretation of the species concept in such a group is not at all clear, and must await the results of de¬ tailed cytological and ecological investigation. 136 SPECIES STUDIES IN THE BRITISH FLORA PROBLEMS OF SPECIATION IN THE BRITISH SPECIES OF ARUM (Exhibit) C. T. Prime. Two species of Amm are undoubtedly native to Britain, one Amm maculatum L., too well known to need description, and the other A, neglectum (Townsend) Ridley, an aptly named species which has not received the attention it deserves. It was formerly included under A. italicum Miller and was described as var. neglectum by Townsend, 1883, and later raised to specific rank by Ridley (1938). Whether it is really specifically distinct from A. italicum, and whether the A. italicum we find occasion¬ ally in this country is truly native, are two of the problems to which final solutions have not yet been found. A. maculatum is a common plant of soils with a high base status, and it shows a wide range of variation. Some of this variation is associated with differences in chromosome number. The earliest chromosome count for the species is that of Schmucker (1925) who gives the number c. 32 for the diploid plant from German material. More recently Maude (1940) has given 2n=56 and 2n = 84 for British material. The number 2n = 56 has been confirmed for plants from all over the country. J. D. Lovis (unpub.) and H. G. Baker (1949) have also con¬ firmed this figure, and it is certain that most of the plants in Great Britain are of this race. The plants with 2n==84 were recorded from Oxted and Merton, but despite repeated search have not been refound. To rediscover and study this rare plant is one of the first problems concerned with A. maculatum which await solution. Fiff. 22. A. maculatum corms (2n=28) from Tubson,^ StorsEov, Laalanrl. Hagenip (1944) records plants with 2n — 28. A. maculatum is rare or absent over most of Denmark except in woods in the south of Zeeland. These plants show several differences from PROBLEMS OF SPECIATION IN THE BRITISH SPECIES OF ARUM 137 those of this country, differences confirmed by examination of herbarium material from Copenhagen. Firstly, the main shoot of the corm is centrally situated much as in the manner of a crocus corm, and not placed laterally as in most of the British plants (Fig. 22). The leaves of these plants are all unspotted and Dr. Hagerup tells me that plants with spotted leaves are vir¬ tually unknown in Denmark. The leaves are darker than the average, smaller, and show less hastate development when com¬ pared with leaves of British plants. Table I, Leaf size in Danish and English A. maculatum. No. of leaves examined. A v. length. Width. Length/ width. Denmark 70 8*8 cm. 6*6 cm. 1-2 England 74 10-2 5-4 1*9 The stomatal index is variable and within the range of British material. A few plants grown in this country have flowered and the ratio of the spathe length to the basal portion (that enclos¬ ing the flowers) was less than in Amm maculatum (2n=56) and the proportion of spadix to spathe length approximates more closely to that of A. neglectum. Table II. The relative length of spathe and spadix in A. maculatum (2n = 2S) (data from Copenhagen herbarium material) I.,ength spathe spadix 18*0 cm. 5*0 cm. 16-5 6-0 10-0 5*3 9-5 4-6 8-2 4-6 11-8 4-8 12-5 5-7 15-0 6-8 180 7*6 120 5-6 110 60 130 6*5 14-0 6-5 12-0 6-0 15-0 5*6 Average ratio 1 : 2-3 Average ratio in A. maculatum 1 : 3*3 Average ratio in A. neglectum 1: 2-6 The spadix is a uniform cylinder unlike those of the “56” rac(' where the base is always slightly wider than the tip (Fig. 23). 138 ■>l>RriES STUDIES IN THE MRITISH FLDR^ Fi". 23. The shapes of the spadices of A. maculaHim 2r?=28 (left) and 2n=r)() (riSht). The inflorescence axis bears three to four rows of sterile males as in the “56” race, but there are 1-2 rows of sterile females whereas there is usually only one row in the “56” race. This latter difference seems to be confirmed by the herbarium material. The pollen grains are the same size in the two races. Three plants of the Danish race were pollinated in 1951 with pollen from the “56” race and set viable seed. So far (1954) in time of appearance, size of leaf and other characters the seed¬ lings appear indistinguishable from those of the “56” race sown at the same time. The Danish habitats seem to be similar to those of this countI3^ In one beech wood there was a shrub layer of Corylus avellana, Acer pseudoplatanus, and Sambucus nigra, while the ground flora at the time of collection of the seed (July 31, 1952) included Aegopodium podagraria and Circaea lutetiana. An¬ other locality was in scrub consisting mostly of Corylus avellana growing near a pond. The 28 chromosomes of this plant pair as bivalents only and meiosis is regular, so as a working hypothesis one may assume the basic number to be 14. On this assumption, the plants with 2n — 56 is a tetraploid and it behaves as an autotetraploid at meiosis. The plant with 84 chromosomes would be a hexaploid. Arum maculatum 2n=28 on the evidence given above con¬ stitutes a geographic subspecies and is to be described accord¬ ingly. Finally, G. Malvesin-Fabre (1944) quotes A. maculatum as having 64 chromosomes but I have not as yet been able to trace the source of this reference. To pass to the second species, A. neglectum. Firstly, it should be recognised that both this and A. italicum are very distinct plants from A. maculatum. A. italicum and A. neglectum are winter green species, being in full leaf by the end of October, while A. maculatum usually appears in the first week of February. There are several other distinct differences fully described else¬ where. The two species rarely hybridise, but in one instance hybridity has been proved by cytological investigation (J. D. Lovis, unpub.). A. italicum and A. neglectum are closely related species; italicum is distinguished by the narrow dark green shining leaves with widely diverging lobes and marked white veining. Further PROBLEMS OP SPECIATION IN THE BRITISH SPECIES OF ARUM 139 it is a slightly larger plant and the fruits have 2-4 or 5 seeds, while A. neglectum usually has two or three. The chromosome number of the two species is the same (2n = 84) and the two plants are in ter fertile. Ridley (1938) considered that A. neglectum was confined to the coast of N.W. France, the Channel Isles and south and west Britain. All A. italicum plants in this country he regarded as cultivated or escaped from cultivation. Recent searches show that the distribution does not conform to so simple a pattern (Prime, 1954 & 1955). Arum neglectum occurs on the Continent in districts other than north-west France and it is possible that isolated colonies of A. italicum in this country may be native. The most likely answer is that Arum neglectum becomes more com¬ mon as one passes from south to north across the area of dis¬ tribution of the two species, becoming the dominant and possibly the only native in Britain. On this evidence Arum neglectum should be relegated to the status of a geographical subspecies. REFERENCES. Baker, H. G., 1946, see SoAvter, F. A., 1949. Hagerup, 0., 1944, Hereditas, 30, 152. Mauhe, P. F., 1940, Chromosome numbers in some British Plants, New Phyt., 39, 17. ^NIalvesin-Fabre, G., 1946, Remarqnes snr la phytogeo^raphie et la biologie d’Arum italicum Mill, et d’A. maculatum L., /b’or. Soc. Linn. Bordeaux, 93, 112. Prime, C. T., 1954, Arum neglectum, J . Ecol., 42, 241. Prime, C. T., Buckle, O., & Lovis, J. D., 1955, The disti’ibution and ecology of Arum neglectum in Southern England, Prop. B.S.B.I., 1, 287. Ridley. H. N., 1938, On Arum neglectum (Towns.) Ridlev, J. Bof.. 76, 144. Schmucker, Th, 1925, Beitrage zur Biologie und Physiologie von Arum maculatum. Flora, 18 and 19, 460. SowTER, F. A., 1949, Arum maculatum, J. Ecol., 37, 207. Townsend, F., 1883, Flora of Hampshire. London. Dr. E. F. Warburg pointed out that the count of 2r = 84 for Arum macAtlafum might be due to a mistake, particularly as this was the number recorded for both A. neglectum and A. italicum. ]40 SI’RCIES STUDIES IN THE BRITISH FLORA SPECIES PROBLEMS IN RECENT SCANDINAVIAN WORKS ON GRASSES A. Melderis (British Museum (Natural History)). The species has long been regarded as a fundamental unit in taxonomy. Not only taxonomists, but also other biologists such as phyto- or zoo-geographers, geneticists, cytologists, physiologists and others, are concerned with it. All of them are studying organisms and using the nomenclatural system established by Linnaeus for classification or designation of plants or animals. According to this system, which is based on natural affinities between organisms, individuals are classified into various categories such as species, genera, families, etc., which are desig¬ nated by names and described to facilitate identification. Botanical nomenclature is regulated by the International Code. Every plant (a member of a natural species) is denoted by a binomial or a biverbal name (taxonomic species) composed of the name of the genus and a specific epithet to which epithets of intraspecific categories (subspecies, varieties or forms) may be added. Classification, however, is a matter of individual opinion, and in many cases there is no agreement in the use of taxa (taxonomic units). The species of one botanist is the subspecies of another, a variety or form of a third. It is due to the fact that different workers studying material by different methods form different concepts of the taxonomic categories. Efforts to give a precise definition of the species have led to much discussion which is still going on. The concept or definition of the species in Scandinavian htera- ture has been the subject of many arguments and different opinions have been expressed. Thus, according to Du Rietz (1930), the species are “the smallest natural populations permanently separated from each other by a distinct discontinuity in the series of biotypes”. Size, polymorphy and amount of variation in characters have not been considered by him to be important bases for the delimitation of a species, which includes natural units of widely different kinds (cf. also Muntzing, Tedin & Turesson, 1931, Nannfeldt, 1935). Turesson (1922a, b, 1929) has proposed a classification of plants based on the relationships between the genetical composi¬ tion of populations and their environments (genecology). His basic unit is the ecotype, a group of biotypes which is recognizably distinct as a result of the selection of suitable hereditary forms by particular habitats. He introduced the term “ecospecies” — “an amphimict-population the constituents of which in nature pro¬ duce vital and fertile descendants with each other, giving rise to SI’KCIKS riU)]iLKMS IN IIKCKNT SCANDINAVIAN WORKS ON CRASSKS i J 1 less vital or less sterile descendants in nature, however, when crossed with constituents of any other population”. As examples of ecospecies, several species of Phleum, i.e., P. pratense L., P. nodosum L., P. alpinum L. and P. commutatum Gaud., could be mentioned (cf. Nordenskiold, 1945). These species are mor¬ phologically close to one another, but genetically and cytologically are distinctly separated. They form a coenospecies, a group of ecospecies within the genus Phleum. Concepts of the categories proposed by Turesson have been further substantiated by Clausen, Keck and Hiesey (1940, 1945), Gregor and his collabora¬ tors (cf. Gregor, 1944, 1946), etc. Nilsson (1930) has regarded the species as a circle of genotypes (a combination sphere), which, as a population, is approximately constant, but in crosses with other species does not intermix due to incompatibility of parental genomes and sterility or non-vitality of crossing products — hybrids (cf. also Turesson, 1931). Recentlj^ Larnprecht (1949, 1953), in his discussion on the relation between the modern conventional species and the natural species, has pointed out that the conventional species has been used for various, quite different taxonomic categories, such as real or primary species, allopolyploids (addospecies), autoploids (superspecies), types with different structure of the chromosomes (mixtospecies), apomicts and ecotypes. The real species, accord¬ ing to him (1953), is “the totality of all biotypes, characterized by the same alleles of the interspecific genes”, which causes simul¬ taneously the morphological differences and the barrier of sterility between two real species. According to Larnprecht, it should be designated by a binary name, e.g., Pisum sativum L. Of the other five taxonomic categories mentioned above, which in Lamprecht’s opinion should all be kept apart from the real species, only addospecies should be designated by a binary name, accompanied by the names of the species forming it, e.g., Galeopsis tetrahit L., adsp. puhescens Bess. + speciosa Mill. All other categories should be subordinated to the real species and Larnprecht has proposed the following designations for them: for the superspecies — ssp. (can be confused with the taxonomists’ subspecies), e.g., Empetrum nigrum L. spp. hermaphroditum (Lge.) Hagerup (4a;); for the mixtospecies — mxsp., e.g., Pisum sativum L., mxsp. abyssinicum Braun; for the apomicts — apom., e.g.. Taraxacum officinale Web., apom. aculeatum Hagl. (3a;); and for the ecotypes — oect., e.g., Lactuca canadensis L., oect. graminifolia Michx. Unfortunately, only a small portion of the plant kingdom has been sufficiently investigated cytogenetically, physiologically, biochemically, ecologically and anatomically, and therefore it is difficult at present to come to a definite conclusion as to whether the nomenclatural system proposed by Larnprecht has advantages compared with the others. It seems, however, that this nomenclature cannot be considered as complete because, as pointed out by Larnprecht himself, overlapping between various categories can occur. Sri-:C1ES STUDIES IX THE BRITISH FLORA 1 12 All definitions and concepts of species mentioned above are different but they have some substantial similarities. The com¬ parison of these definitions reveals that their authors seem to agree that the species must consist of genetical systems of populations, which are separated from each other by complete, or at least sharp, discontinuities in morphological and physio¬ logical characteristics. These discontinuities are maintained in the following generations due to complete incompatibility or to the production of hybrids with much reduced fertility (cf. also Stebbins, 1950). I agree with Nannfeldt (1938), who says: “The scientific value of the species concept is often vastly overestimated in these discussions, for the whole hierarchy of taxonomical categories of higher and lower rank is a purely practical scheme. Nature is too rich in possibilities for allowing such definitions as leave no room to doubts or subjective opinions.” The present classification of grasses into species is based chiefly on morphological characters of the reproductive and vegetative organs. The structure and arrangement of spikelets and the great diversity in their shapes and peculiar modifications of their separate parts, especially lemmas and paleas, associated with the length of anthers, provide a greater variety of distinctive characters than the vegetative organs such as presence or absence of ligule, its type of leaf-sheath (free, with overlapping margins or tubular, with connate margins), nervation and shape of leaf-blades, peculiarities in formation of shoots (intravaginal or extravaginal), etc. In many instances, however, the morpho¬ logical characters are insufficient for classification because distantly related species can possess similar spikelets, lemmas, anthers, ligules and leaf-blades It has been proved that in some of these cases the anatomical characters of leaves can be of great value in classification. Thus, the structure of the epidermal cells of leaves, particularly the size and shape of these cells and shape Fig. 24. Lower epidermis of the leaf blade : X—Hordeum secalinurn (from Bottesford, E. F. Linton, 2.viii.l877), B — H. hrachyantherum (from Saskatchewan, A. J. Breitung, 4671), C — Seslerla coerulea ssp. calcarea Hegi (from banks of Tees A. T. Wilmott, 2020). D — S. coerulea ssp. uliginosa (from Sddermanland, E. Asplund, 11. vi. 1015). tsIMOClKS rilOHJvKMS IN llECKNT SCANDINAVIAN WOilKS ON CKASSKS 14:5 of their walls, is fairly constant and characteristic in many species, e.g. of the genera Hordeum and Sesleria (Fig. 24). Also the presence or absence of paired siliceous and suberised cells and the disposition of the siliceous cells in the epidermis over the nerves can be used as distinctive characters in the separation of several species within the same genus, e.g. to separate European Hordeum secalinum Schreb. (H. nodosum L. pro parte) from the closely allied Californian H. calif ornicum Covas & Stebbins and N. American H. brachyantherum Nevski. The arrangement of the vascular bundles and the disposition of sclerenchyma have been proved to be valuable characters to distinguish taxa in the genus Festuca. Results of cytogenetic research as well as data obtained in ecological and phyto-geographical studies have given agrosto- logists many important indications regarding the separation and relationship of species. The value of distinctive characters in grasses, as in other plants, is measured by their constancy. A comparison of the characters used for distinguishing species shows that the taxo¬ nomic value of these characters varies from genus to genus. Those characters which are rather constant among species of one genus may be variable in species of another, particularly the absence or presence of awn or of pubescence on the lemma. Thus, species in the genera Poa and Puccinellia are awnless, but in the genera Agrostis, Festuca, Lolium, etc. the awned species usually have awnless forms or vice versa. In some other cases, species with typically glabrous lemmas have forms with hairy lemmas, e.g. Bromus erectus Huds., B. sterilis L., B. secalinus L. and B. lepidus Holmb., Festuca rubra L. and F. ovina L., etc. This parallelism is not confined only to the characters of the lemma, but appears sometimes also in the shape of inflorescence (panicle contracted or open), in the length of the ligule, in the presence or absence of pubescence on culms, leaf-sheaths or leaf- blades. In the older Floras such grasses aberrant in single characters were described as varieties or forms. The taxa have, however, no scientific value. A few extreme forms do not exhibit the pattern of variability sufficiently within a certain species. On the other hand, such forms may provide valuable material for genetical studies (e.g. of gene frequency or of mutation rates in nature). During the last ten years the attention of several Scandinavian botanists has been turned to the study of variation within some polymorphic and critical genera of grasses, such as Calamagrostis, Poa, Deschajnpsia, Phleum, Agrostis, etc. Various methods of modern taxonomy have been used to solve problems involving the origin, delimitation and relationship of the species and varia¬ tion within the species. Many interesting results have been obtained, on which Hylander (1953) based his treatment of cor¬ responding genera in his recently published northern Flora. Much interest has been paid also to the study of plant distribution in Scandinavia which resulted in Hulten’s (1950) comprehensive Atlas of the distribution of vascular plants in V.IF. Europe with Ill Sl’KClKS STUDIES IN THE mtlTlSlI FLOllA maps for every species, including grasses. Data of the cytological investigations have been compiled by Love and Love (1948). A short account of species problems in recent Scandinavian works on grasses, which may be of some interest in relation to those species occurring in Britain, is given below. Calamagrostis. As shown by Nygren (1946), this genus con¬ tains amphimictic (sexual) and apomictic (producing seeds asexually) species. In Scandinavia the sexual species are the following: C. arundinacea (L.) Roth, C. canescens (Web.) Roth, C. stricta (Timm) Koel. {C. neglecta Gaertn., Mey. & Scherb.), C. varia (Schrad.) Host and C. epigejos (L.) Roth, of which the first four are tetraploids (2n = 28), while C. epigejos consists of tetraploid, hexaploid and octoploid forms (2n — 28, 42, 56). These species are characterized by usually X-shaped, dehiscent anthers with well developed pollen. They hybridize one with another and produce more or less intermediate hybrids, which can be recog¬ nized by I-shaped, indehiscent anthers with more or less badly developed pollen. They do not form seeds. The apomictic species in Scandinavia are represented by C. chalybea (Laest.) Fr. (2n = 42), C. lapponica (Wahlenb.) Hartm. (2n = 42-112) and C. purpurea (Trin.) Trin. (2n = 56-91), which all have usually I- shaped anthers. Pollen formation has not been observed in C. chalybea, but occasionally or very rarely it can occur in (J. lapponica and C. purpurea. According to Nygren, the apomixis within this genus has arisen as a consequence of intra- and inter¬ specific hybridization in combination with polyploidy. He has obtained apomicts in crosses between sexual species. Thus, in C. epigejos (2n = 56), an apomictic hybrid with a chromosome number 2n = 42 has been obtained. In the second generation this hybrid has given several plants with a chromosome number varying from 2n=35 to 2n = 12. Some descendants with chromo¬ some numbers 2n=35, 37 and 38 resembled C. varia (cf. Nygren, 1948b). Another apomict with a chromosome number 2n = 56, resembling apomictic C. purpurea, has been synthesized by Nygren (1946, 1948a) in a cross between tetraploid C. canescens (2n = 28) and hexaploid C. epigejos (2n = 42). Nygren (1946) has also succeeded in obtaining C. purpurea by colchicine treatment of seeds of C. canescens. C. purpurea is the most polymorphic species in Scandinavia. Some of its types show remarkable similarities to C. canescens, while others approach in some respects to hybrids between C. canescens, on the one hand, and C. arundinacea, C. epigejos and C. stricta, on the other. Its types with a high chromosome num¬ ber produce hybrids with tetraploid sexual species (cf. Nygren, 1948a). The chromosome number in these hybrids is usually 2n = 70, which, according to Nygren, means that unreduced gametes of C. purpurea with 56 chromosomes have fused with reduced gametes with 14 chromosomes from a sexual species. The behaviour of these hybrids is similar to that of C. purpurea. SI’KCIES I’KOliLKiMS IN KKC'KNT SCIANDINA VIAN WOKKS ON GKASSES J4o and it is very difficult to distinguisli them from each other. In his further experiments with C. purpurea, Nygren (1949b, 1951) has discovered also some facultatively apomictic types with pollen formation in this species. They behave, however, differently with regard to their ability to produce new types. Some of them can increase the variation within the genus by self-fertilization, followed by segregation, or by hybridization with sexual species. The origin of C. chalybea is not clear. According to Nygren (1946, 1948b), it may have arisen from some derivatives of the hybrid C. arundinacea x C. canescens or from a population of the Central Asiatic sexual species C. obtusuta Trin., with which it has previously been confused in Scandinavia. C. lapponica is more uniform than the other apomictic species but it forms some types which show a close affinity to C. stricta and some of its hybrids. The situation within the genus is interpreted by Nygren (1951) as follows : “The combination of sexuality and apomixis in Calamagrostis makes possible a formation of a higher number of distinct forms in nature than does the usual sexual process. The facultative apomictic strains are able to perpetuate themselves by diplospory at the same time as they give rise to new biotypes by hybridizing with other facultative strains or with other species, environments different from those to which purpurea is adapted, to climates outside its range. When the distribution areas of the different species overlap new forms will arise which are fitted to environments different from those to which purpurea is adapted. In this way the species complex may be able to spread into new areas. In the case of C. purpurea it is not possible to keep the sexual and the apomictic formation of biotypes apart as has been done by several authors in other material. The two processes sometimes occur separately, but they often intermix, and in such cases great variation will be the final result.” The British species of Calamagrostis, i.e., C, epigejos, C. canescens, C. stricta and C. scotica (Druce) Druce have not been studied cytogenetically as yet, but a similar situation may be expected also here. C. scotica, which is a very rare species, occur¬ ring in bogs in the northern part of Caithness (Scotland), has been the subject of discussion (cf. Hubbard, 1954). According to Ben¬ nett (1885, 1886), it was considered to be identical with Scandina¬ vian C. lapponica, C. stricta or C. x strigosa (Wg.) Hartm., which is now recognised as a hybrid between C. epigejos and C. neglecta. Hackel, who saw this grass, rejected it as C. x strigosa, and Druce, therefore, referred it as a separate variety to C. stricta (Deyeuxia neglecta) (cf. Druce, 1915) and later raised it to specific rank (Druce, 1926). On examining specimens of C. scotica at the British Museum, the author has observed that a portion of their pollen is angular, shrivelled, lacking plasma, although many anthers are more or less X-shaped and dehiscent. Cytogenetic studies may give some valuable indications regarding the origin of this taxon. It should be noted also, that two striking forms 146 SI'KCIKS STUDIKS IN THK HRITISH FLOKA with features of C. purpurea have been collected by Mr. E. C. Wallace in Scotland (Braemar) and by Miss M. I. Tetley in West¬ morland (Esthwaite). They have broader leaves than in C. canescens, a long ligule, shortly pubescent spikelets and narrow, yellow, I-shaped anthers with badly developed pollen. These forms, as well as those resembling a hybrid between C. canescens and C. stricta (F. E. Crackles, 1953), are in need of further criti¬ cal study based on cytogenetic investigations. Poa. This genus has been a subject of extensive studies for a long time, chiefly due to the economic importance of Poa pratensis s. lat. It has been found that several species have apomictic facultatively sexual and sexual types and are charac¬ terized by a very wide variation in the chromosome number* (cf. Gustafsson, 1947^ for references). As shown by Nannfeldt (1935), the section Stoloniferae, which in Scandinavia is represented by P. pratensis s. lat. and P. arctica R. Br. (not found in Britain), is very polymorphic and seems hardly to contain any uniform species which can be sharply distinguished. In Scandinavia the members of this section hybridize with some species of other sections, i.e., P. alpina L. of the sect. Suh-bulbosae and P. flexuosa Sm. of the sect. Oreinos, which both occur also in Britain. P. pratensis s. lat. in Scandinavia is classified into four distinct ecological subspecies, namely ssp. pratensis with a chromosome number 2n= 50-124, ssp. angustifolia (L.) Lindb. fil. with 2n = 50- 65, ssp irrigata (Lindm.) Lindb. fil. (P. subcaerulea Sm., P. irrigata Lindm.) with 2n = 48-95 (in Scandinavian material) and ssp. alpigena (Fr.) Hiitonen, with 2n = 48-92, which all are re¬ garded by some taxonomists as separate species. British speci¬ mens of ^*P. pratensis ssp. alpigena”, examined by the author, proved to belong to various mountain forms of P. subcaerulea, which is closely allied to ssp. alpigena. It seems that the true ssp. alpigena has not been recorded from Britain as yet. The glumes of P. subcaerulea are 3-nerved as in ssp. alpigena, but they are longer and more- acuminate. In addition, the inflorescence of P. subcaerulea is open, not contracted as in ssp. alpigena, and the spikelets larger. The British P. subcaerulea, which exists in many closely related forms occupying ecologically quite different situations such as sea-shores, dune-slacks between dunes, moun¬ tains, etc., unfortunately, has not been studied morphologically, ecologically, cytologically or embryologically as yet. The embryo- logical situation in the British forms may be similar to that of Icelandic forms investigated by Love (1952). The development of seeds in Icelandic P. subcaerulea (2n= 82-147), according to him, is mainly aposporous, associated with pseudogamy as in P. pratensis. Occasionally, possibly only in plants with a lower *It is interesting to note that Hedberg (1952) has found a small variation in tlie chromosome number 2»=up to 42) in members of the genus Poa growing in East African mountains. All species there seem to be sexual in contrast to the species occurring in the boreal and arctic regions. Sl’KCIKS L’llOBLlOMS IX KKCEXT SCANDINAVIAN WORKS ON GRASSES 147 chromosome number, the seeds are formed also in the normal sexual way, after the fertilization of aposporous egg cells. In addition, in F. pratensis, egg cells with a reduced number of chromosomes are involved in the seed formation. Love (1950) has obtained no indication that a reduction division takes place at the formation of egg cells in Icelandic P. subcaerulea. P. pratensis, on the contrary, has a reduction division, and there¬ fore seeds can be formed also after the fertilization, either by apos¬ porous egg cells or egg cells having a reduced chromosome num¬ ber. The polymorphic P. arctica, with a chromosome number 2n-- 39-92 (in Scandinavian plants), according to Nannfeldt (1940), contains only a few types with distinct and constant characters, in contrast to numerous forms with indistinct features, the limits of which are not sharply defined. These constant apomictic types were originally separated by him as subspecies, but later on (cf. Hylander, 1941, 1945) they were reduced by him to varietal rank, because the whole bulk of the forms has not been sufficiently studied to divide it into similar types. Three of them, namely var. depauperata (Fr.) Nannf., var. elongata (Blytt) Nannf. and viviparous var. stricta (Lindeb.) Nannf. are restricted to a limited area in Dovre mountains (in Middle Norway). Two other varieties have been found in the mountains of N. Scandinavia, i e., var. tromsensis (Nannf.) Nannf. on the Mt. Fl0ya (N. Nor¬ way) and var. microglumis (Nannf.) in Signaldalen (N. Norway) and in eight localities in Lule Lappmark and Tome Lappmark (N. Sweden). Var. caespitans (Simm.) Nannf., occurring on Mt. Hogtinden (N. Norway) and at Nissontjarro and Kaisepakte (N. Sweden), has a wide distribution in the Arctic, from Novaya Zemlya in the East to Baffin Land and Ellesmereland in the West (cf. Nannfeldt, 1940). Recently Nygren (1950a, b), who has studied the cytology and embryology of the Scandinavian mountain species of Poa, has come to some interesting conclusions regarding their polymorphy. According to him, viviparous and non- viviparous P. alpina have a varying chromosome number 2n= 31-57 (in the British viviparous P. alpina only 2n = 35 has hitherto been reported). The non- viviparous form has been found to be sexual. The viviparous form propagates only by bulbils, but produces pollen and can cross with the other species growing in its range (cf. Nannfeldt, 1940; Melderis, 1953). P. fiexuosa, which has been separated by Nannfeldt (1935) from P. laxa Haenke of the Alps, is a sexual species with 2n = 42 and sets very good seed. It forms with vivi¬ parous Po alpina a hybrid P. x jemtlandica (Almq.) Richt. which is so far recorded from Scandinavia, Scotland (Ben Nevis and Lochnagar) and Iceland. Its chromosome number 2n = 37 (in Scandinavian material) indicates that it may have arisen by a fusion of a reduced female gamete (n— 21) of P. fiexuosa with a male gamete (n=16) of viviparous P. alpina (cf. Nygren, 1950a). This hybrid is morphologically uniform, viviparous. 14W Sl'KClKS !STU1)1KS IN THK JllUTi.Sll i'l.OKA sexual as to the embryo-sacs, but does not give seeds. The pollen is very poor. The embryology of non- viviparous, as well as that of viviparous P. prate7isis ssp. alpigena with a chromosome num¬ ber 2n. = 38-81 (recently found on Mt. Paldsa (N. Sweden)) is similar to that of the typical P. pratensis. The viviparous type of ssp. alpigena forms morphologically good pollen which is able to fertilize egg cells of non- viviparous ssp. alpigena or those of other species. It is possible that some hybrids in the Paldsa district may have arisen from a crossing between viviparous ssp. alpigena and non-viviparous P. alpina. A hybrid between non- viviparous ssp. alpigena and viviparous P. alpina (P. x herjedalica H. Smith) is not rare in Scandinavian mountains. It is viviparous and has forms with a chromosome number 2n= 47-80. As it pro¬ duces good pollen and many give seeds in a sexual or in an apos- porous way, Nygren (1950a) presumes that it may be recombined in the offspring of secondary hybrids or from their derivatives. P. arctica var. depauperata and var. elongata, according to Nygren (1950a, b), have a chromosome number varying between 2n=75 and 2n=79 in the former, and between 2n= 68 and 2n = 76 in the latter. Both varieties form rather poor pollen and aposporous em¬ bryo-sacs. They cannot as mother plants give hybrids with other species of Poa due to a selective mechanism of an early first division of the egg cells. Var. microglumis with 2n = 68-82 has well developed pollen. Although it is aposporous, the hybrids with other species can be produced with it as a mother plant. Var. caespitans is aposporous and sexual. It is characterized by a chromosome number 2n = 56 and by completely barren anthers. It can give hybrids which may be persistent by propagation in an asexual way. The viviparous var. stricta^, with a chromosome number 2n ~ 39, has a sexual development of the embryo-sac. Ac¬ cording to Nannfeldt (1940) and Nygren (1950a, b), it must be regarded as a relict of a very old sexual population which seems to have survived the last glaciation in Scandinavia. A recently discovered hybrid between viviparous P. alpina and P. arctica has a similar embryology to P. x jemtlandica mentioned above. On the basis of results obtained in his studies, Nygren (1950a, b) has pointed out that polymorphy in the genus Poa is due to the hybridization processes accompanied by apomixis, as in Calamagrostis, and vivipary as well (cf. also Melderis, 1953). Numerous viviparous and non-viviparous types occurring in the Scandinavian mountains and in the Arctic have arisen from crossings between apomictic, facultatively sexual and sexual forms. The formation of new types seems to be still in progress. Repeated crossings increase the variability within the genus and level the limits between the species and types. Only pure apomicts will remain sharply differentiated. There are many types which cannot be kept apart. A portion of types arising from these processes can propagate only by vegetative means *lts taxonomic status has been i-ecently discussed by Xoi-dlia<>en (lit.Vi). SPECIES PROBEEMS IN RECENT SCANDINAVIAN WORKS ON GRASSES 149 (bulbils, rhizomes, etc.), whilst the others reproduce by seeds formed in an asexual or in a sexual way. In the British mountain flora, P. subcaerulea on the one hand, and P. glauca with P. nemoralis on the other, exhibit a great variation in their external morphology. Viviparous and non- viviparous P. alpina, P. fLexuosa and the hybrid P. x jemtlandica, which are all rare, seem to be more uniform. The hybridogenous origin of P. pratensis has been pointed out by Clausen, Grun, Nygren and Nobs (1951), when they dis¬ cussed some constant new plants, resembling P. pratensis, which were obtained in sexual, segregating progenies of hybrids between P. pratensis and some American species, e.g. P. ampla Merr. and P. scabrella (Thurb.) Benth. According to them, P. pratensis during its evolution and distribution may have absorbed genomes from several species. Akerberg and Bingefors (1953) have observed from a third generation (F3) offspring among plants with 2n = 60-78 a sub- haploid plant with 2r? = 16. This offspring has been derived from a second generation (Fo) plant with 2n=75, the parents of the original cross being apomictic P. pratensis s. lat. with 2n = 50 and apomictic P. alpina with 2n = 37. This plant was morphologically very similar to P. trivialis L., and resembled another subhaploid plant with 2n=18 obtained by Kiellander (1942). Kiellander has indicated that P. trivialis may be one of the plants which have taken part in the formation of P. pratensis. Nannfeldt (1940), who is also in favour of the assumption of hybridogenous origin of P. pratensis ^ and P. arctica as well, has mentioned, on the contrary, that their parents might not be among any existing species. It is interesting to note that the hybrid origin of another species of Poa, namely tetraploid P. annua L. (cf. Nannfeldt, 1937; Tutin, 1951), has recently been confirmed by Tutin (1953), who has synthesized tetraploid plants (2n = 28) with morpho¬ logical features of P. annua from a cross between two diploid species, viz. annual P. infirma Kunth (2n=14) with perennial P. supina Schrad. (not found in Britain as yet). Deschampsia. The distinguishing of D. cespitosa (L.) Beauv. (2n = 26) from D. alpina (L.) Roem. & Schult. (2n = 26, 39, 41, 48, 49, 52) in some cases is remarkably difficult (cf. Hylander, 1953). Both have a basic chromosome number 13 in contrast to 7 in other species, e.g., D. setacea (Huds.) Hack. (2n=14), D. flexuosa (L.) Trin. (2n=:28) and D. bottnica (Wg.) Trin. {2n = 2S). The characters usually given in the keys for separation of D. cespitosa from D. alpina are the following ; in D. cespitosa the spikelets are usually non-viviparous, the awn is inserted at the base of the lemma and the branchlets of the panicle are scabrous; in D. alpina the spikelets are usually viviparous, the awn is inserted at about the middle of the lemma or higher and the branchlets are smooth. The occurrence of a viviparous form in D. cespitosa 150 SPECIES STUDIES IN THE BRITISH FLORA makes them very difficult to distinguish. In metamorphosis of the spikelets in this species the awn is shifted upwards and it may occupy the position which is characteristic for D. alpina. Also the roughness of the branch lets in some cases is proved not to be a constant character. Thus, I have seen a viviparous D. cespitosa from Coire-an-Lochan in Eastemess (Scotland) which has quite glabrous panicle-branchlets. Tlie insertion of the awn in these specimens varies from the basal up to a nearly apical position in the spikelets of the same panicle. According to Nygren (1949a), the vivipary is genetically fixed and its occurrence is due to a change of the normal life rhythm, e.g., variations in the length of day. It has been observed that an increase in the number of chromosomes is accompanied by an increase in the degree of vivipary. Thus, a race of D. alpina with 2n = 26 from Herjedalen (Middle Sweden) is non-viviparous, sexual and produces seeds. A race with 2n = 39 from Spitsbergen has weakly developed bulbils and can also produce seeds (cf. Flovik, 1938). Races with 2n = 41 and more are viviparous. Nygren is in favour of the assumption that these chromosome races are polyploid derivatives of D. cespitosa. Non-viviparous D. alpina from Herjedalen, mentioned above, in its external morphology can be referred either to D. alpina or to D. cespitosa. The two species are closely allied and at present no essential characters are known for separating their viviparous forms. The British viviparous forms of Deschampsia have been discussed by Wycherley (1953). Phleum. In Scandinavia this genus is represented by the same species as in the British Isles: P. phleoides (L.) Karst. (2n= 14, 28), P. arenarium L. {2n= 14), both of the section Chilo- chloa, P. pratense L. (2n = 42j, P. nodosum L. (2n=14) and P. commutatum Gaud. (2n = 28, 56), all of the section Euphleum. P. commutatum was included in older floras, both in Scandinavia and in Britain, under P. alpinum L. (2n=14), which is restricted to the mountains of Central and South Europe and differs from P. commutatum in chromosome number and in the hairy awns of the glumes. Horn af Rantzien (1946), who studied the taxonomy and dis¬ tribution of P. arenarium, has found that this species exhibits only small variation in some morphological characters which are not essential from the taxonomic point of view. In its northern distribution this species is sharply differentiated from the other member of the same section — P. phleoides but this is not so in the Mediterranean area, where many forms of various related species, showing a great ecological and morphological re¬ semblance, have been observed. Nordenskiold (1945) has tried to elucidate the origin of the cultivated hexaploid P. pratense and its relationship to the allied species such as P. nodosum, P. alpinum and P. commutatum. All these species are morphologically closely related and they SPECIES PKOBI.EMS IN RECENT SCANDINAVIAN WORKS ON GRASSES 15] have been treated by some authors, e.g., Ascherson & Graebner (1899), as a single collective species. Some others, e.g., Druce (1932), Ovczinnikov (1934), have divided this collective species into two species, namely P. pratense (inch P. pratense s. str. and P. nodosum) and P. alpinum (inch P. alpinum s. str. and P. com- mutatum). The interspecific crossing experiments carried out by Miss Nordenskiold show'ed that these species are distinctly separated genetically. The results of the crosses are given in a table below* ; Hybrids in F, (first generation) Combinations Chromosome numbers (2n) Fertility Pollen Seed¬ setting prat. 42 X nod. 14 28 (int.), 35 (prat.- like) good good nod. 14 X prat. 42 35, 49 (both prat.- like) good good nod. 21** X prat. 42 around 42 (prat.- like) good good alp. 14 X prat. 42 35 (int.), 49 {alp.- like) good good prat. 42 x alp. 14 — — prat. 42 x com. 28 35 (int.) st. poor com. 28 X prat. 42 49 (nearly int., more like com.) st. poor prat. 42 x com. 56*** 49 (nearly int., more like com.) st. poor nod. 14 X alp. 14 14, 21 (both alp.~ like) poor poor alp. 14 X nod. 14 14 (al'p.-like) poor poor nod. 14 X com. 28 21 (int.) st. st. uod. 14 X com. 56*** 42 (more like com.) good not abundant com. 28 X nod. 14 21 (int.) st. st. nod. 21 X nod. 14 20 (nod.-like) st. st. alp. 14 X com. 28 21 (int.) st. st. com. 28 X alp. 14 21 (int.) st. st. * Abbreviations : vrnt. = P. pratense-, nod. = P. nodosum-, alp. = P. alpinum-, com. = P. commutatum-, int. = intermediate; st. = sterile; — = unsuccessful cross. Numbers after specific names refer to the chromosome numbers (2n). Indications of the morphological features of the hybrids are enclosed in brackets after chromosome numbers. **Triploid P. nodosum (2r?,=21) has been obtained in the progeny of a male- sterile P. nodosum crossed with the typical form. ***Octoplold P. commutatum {2n=56) has been found in an offspring of tetra- ploid P. commutatum from Lapland (N. Sweden). It resembles the typical form, but is somewhat robust but not so tall in growth. 152 SPECIES STUDIES IN THE BRITISH FLORA According to Miss Nordenskiold, P. pratense, P. nodosum and P. alpinum give fertile hybrids in crosses with one another, pro¬ vided that gametes do not possess a too deviating chromosome number. The tetraploid P. commutatum, which shows resemb¬ lance in some morphological features to the diploid P. alpinum, is the most strongly differentiated and in crosses with the other species produces highly sterile hybrids. It seems to be allopoly¬ ploid, although none of the diploid species mentioned above can be considered to have taken a part in the formation of this species. The wide geographical range of this uniform species seems to indicate that it must be very old in spite its polyploidy. Miss Nordenskiold (1949) has also succeeded in producing a hexaploid P. nodosum with features of P. pratense by repeated colchicine treatments of seeds and young seedlings of P. nodosum, followed by crosses between the polyploids obtained. A cross of this hexaploid P. nodosum with the typical P. pratense gave a good seed-setting. It indicates that P. pratense is closely allied to P. nodosum and that it may have arisen from the latter. As result of Gregor & Sansome’s (1930) and Gregor’s (1931) ex¬ periments and her own Miss Nordenskiold is in favour of keeping all the species in question apart. Hylander (1953), however, has treated P. nodosum as a subspecies under P. pratense, because morphologically it cannot always be easily distinguished from P. pratense. As regards P. commutatum, he has followed Miss Nordenskiold, but he has pointed out that the discovery of a race of P. alpinum with features of P. commutatum in the Pyrenees has complicated the separation of these two species. Roegneria. Interspecific hybrids within this genus are not uncommon in the northern part of Scandinavia, where the distribution ranges of these species overlap, and the species come together. In the hybridization processes the following species are involved there: long-awned R. canina (L.) Nevski, short-awned R. mutahilis (Drob.) Hyl., R. borealis (Turcz.l Nevski and R- fibrosa* (Schrenk) Nevski, which are all tetraploids hav¬ ing 2n — 28 (cf. Hylander. 1953). These hybrids are more or less intermediate in external morphology between their parents. Some of them are found onlv as scattered individuals in the fring¬ ing zone between the populations of the parent species, e.g.. R. borealis x R. canina and R. canina x R. fibrosa. Their pollen has been found to be completelv sterile. A recently recorded hybrid between R. borealis and R. mutabilis has about 80% of sterile pollen. Hybrids with R. mutabilis usuallv grow side by side with various morphologically different types, as observed by the author in Lapland and Norrbotten (N. Sweden). Experiments with an artificial hvbrid between these species revealed that it is partially sterile and by back-crossing can give *TliP chr-oiDosome mimher 2r?— 28 Tins Tippti fonrid by tbo niitbor in 7?. I^hroftn 'yrown nt TpsnTn I'votii tbo sooris I'ocoivod from llu' r.otnnicnl (’.nrilmi in I-onlnffrad) and in 7?. hphwii (from Jnmtlnnfl). SPECIES PROBLEMS IN RECENT SCANDINAVIAN WORKS ON GRASSES 153 rise to a hybrid swarm. The hybrid R. borealis x R. canina on the contrary seems to be highly sterile. It did not produce seeds after back-crossing. These facts may indicate that R. mutabilis is more closely allied to R. canina than is R. borealis. The recently described R. behmii Meld. (2n = 28), from Jamtland (Middle Sweden), is closely related genetically to R. canina, but differs from the latter in several essential morphological charac¬ ters, such as relatively shorter and broader leaves, an erect spike, the shape of the glumes, a short awn to the lemma, etc. An artifi¬ cial hybrid between them, however, has been found to be fertile, with a low proportion of badly-developed pollen. It seems that Scottish R. doniana (F. B. White) Meld, has the same behaviour in relation to R. canina as has R. behmii. The intermediates col¬ lected by Mr. J. E. Raven and Dr. S. M. Walters at Inchnadamph in 1953 had well-developed pollen (cf. Raven & Walters, 1954). Agrostis. Four native species of this genus are common both in Scandinavia and the British Isles. They are : A. canina L., A. tenuis Sibth., A. gigantea Roth and A. stolonifera L. These species have been investigated cytogenetically : Scandinavian material by Bjorkman (1951, 1954) and British material by Davies (1953) and Jones (1951, 1953). Cytological evidence obtained by Bjdrkman and Jones shows that Scandinavian and British forms of A. canina can be divided into two larger groups, which differ in chromosome numbers, ecology and shoot morphology. These are : var. fascicularis (Curt.) Sind, and var. arida Schlecht., which have been considered by subsequent authors, e.g. Hylander (1953) and Hubbard (1954), as separate subspecies, the former as ssp. canina [ssp. fascicularis (Curt.) Hyl.] and the latter as ssp. montana Hartm. Ssp. canina, which is diploid (2n=14), has thin leaves, creeping leafy stolons and occurs in damp or wet situations. Bjorkman (1951) has discovered also a tetraploid and a triploid form in Scandinavian material, and he has assumed that they may have arisen due to unreduced gametes. Ssp. montana is a tetraploid (2n=28) with stiffer leaves and scaly underground rhizomes, forming dense tufts. It is a plant of dry habitats, e.g. heaths, grassy hills, mountains, on sandy soil, etc. A pentaploid form has been discovered by Bjorkman (1951) in material from N. Sweden. According to him, some transitional forms between these subspecies have been found in situ, but their chromosome number does not indicate that they are hybrids. Artificial hybrids with 2n = 21 have been obtained in crosses between the two subspecies. As shown by Davies, they are completely sterile. A. stolonifera exhibits a greater variation in chromosome number. Of 900 plants of this species from different countries examined by Bjorkman (1954), about 600 plants were tetraploids (2n = 28), 160 plants proved to be pentaploids (2u = 35) and 135 plants turned out to be hexaploids (2n = 42). In addition 2 plants with aneuploid numbers (2r? = 33 and 41) were found. In 154 SPECIES STUDIES IX THE BRITISH FUORA spite of the fact that these races are cytogenetically distinct they seem not to possess any essential character for their separation. Also the ecology does not give any clue for their classification. According to Bjorkman, the pentaploid and hexaploid types have been found in situ, growing together or separately, side by side with the tetraploids or in localities where tetraploids were absent. All three chromosome numbers have been found in plants with sim lar external morphology. Of the other species of Agrostis, A. tenuis has been detected to be tetraploid (2n = 28), and A. gigantea is hexaploid (2n = 42). Artificial hybrids between Scandinavian or British species of Agrostis obtained by Davies and Bjorkman are shown in the table below.* As shown by Davies, a hybrid between A. tenuis and A. stolonifera sets a low proportion of seed, but the plants of the first generation (FJ from crosses A. tenuis x A. gigantea and A. stolonifera x A. gigantea are quite fertile. A hybrid between A. canina ssp. montana and A. tenuis is found to be moderately ^.Abbreviations : can. = Agrostis canina ssp. canina-, mont.=A. canina ssp. mon¬ tana-, ten.=A. tenuis-, stol.=A. stolonifera-, gig.=A. gigantea-, l)or.=A. borealis Hartm. (not found in Britain); sem.=A. semrvei'ticillata (Forsk.) Christens. ; + =successfiU cross; - =unsuccessfiil cross. SPECIES PROBLEMS IN RECENT SCANDINAVIAN WORKS ON GRASSES 155 fertile and in the next generation it produces about half the num¬ ber of both parent species. The natural hybrids recorded from Scandinavia (cf. Hylander, 1953) and Britain (cf. Philipson, 1937, and Davies) are given in the following table- Combinations Scandinavia Britain A. canina x .4. stolonifera + 4- A. canina x A. gigantea + A. canina x A. tenuis + + .4. gigantea x A. stolonifera 4" A. gigantea x A. tenuis + 4- .4. stolonifera x .4. semiverticil lata 4- .4. stolonifera x A. tenuis + A. borealis x ,4. stolonifera + .4. borealis x A. tenuis + Agropogon littoralis (Sm.) C. E. Hnbbard {Agrostis stolonifera x 4- 1*0 1 ypog on n i o nsp eJien s i .s) Hybrids of Agrostis are not rare in nature and many more of them could be discovered if the wild populations were studied more carefully. Thus Bjorkman (1954) has found 33 natural hybrids between A. gigantea and A.* tenuis in different localities, where the species grow together. Offspring of the hybrids can give rise to hybrid swarms which show great variation in the morphological pattern and in chromosome number, making the delimitation of the species difficult. Anthoxanthum. Love & Love have found that A. odoratum L. in its northern range consists of two types which differ in minor morphological characters but are quite different in their cytology, one being diploid (2n = 10) and the other tetraploid (2n = 20 + 6f). They have separated the diploid type as a distinct species, A. alpinum Love & Love, from the tetraploid A. odoraturr,. A. alpinum has been hitherto recorded from Scandinavia, Switzer¬ land, Iceland and Greenland (cf. Tutin, 1950). It differs from A. odoratum usually in narrower leaves of the vegetative shoots, shorter inflorescence, smaller spikelets, glabrous pedicels and glumes and a comparatively longer awn. A. alpinum from Green¬ land and Iceland seems to be distinct from A. odoratum, but according to Hylander (1953), there is some difficulty in separat¬ ing these species in Scandinavian material. It should be noted that polyploids occur rather widely among grasses in nature (cf. Stebbins, 1950; Heslop-Harrison, 1953). A portion of them is known or presumed to have arisen through hybridization by the addition of both sets of chromosomes from each parent species (amphipolyploidy, allopolyploidy), e.g. 156 SPKCIES STUDIES IN THE BRITISH FLORA Poa annua (2n = 2S) from a cross between P. infirma (2n=14) and P. supina (2n=14) (see p. 00), Spartina townsendii H. & J. Groves (2n— 126), a product of the crossing S. maritima (Curt.) Fernald (2u = 56) and S. alterniflora Lois. (2?t = 70) (cf. Huskins, 1930) and Bromus diandrus Roth {B. gussonii Pari.) (2n = b^) through hybridization between B. rigidus Roth (2n = 42) and B. sterilis L. (2n— 14) (cf. Cugnac & Camus, 1931); all of them are fertile. Some other polyploids may have originated by the doubling or multiplication of the basic chromosome number in the non-hybrid plant (autopolyploidy), c g. Phleum pratense (2n = 42) from P. nodosum (2?^= 14) (see p. 152), Dactylis glomerata L. (2n = 28) from D. aschersoniana Graebn. (2n=14) (cf. Miintzing, 1943), etc. Morphologically, polyploids are difficult to separate from their progenitors and they differ from the latter chiefly in slight quan¬ titative characters. In many cases they are more vigorous in habit, darker green in colour, have larger spikelets and longer anthers. In the identification of dried herbarium specimens size of stomata and pollen-grains often can provide valuable clues for the recognition of morphologically closely related diploids and their polyploid derRatives. Thus, e.g., according to Covas, 1949. the size of stomata in H. secalinum (2n=28) is 44-50/i, in H. cali- fornicum (2n=14) is 30-34/^, and in H. hrachyantherum {2n = 2S) is 42-48y. The size of pollen-grains in the same species is respec- itvely 39-44y, 32-36y, and 39-44a. Genetically, on the contrary, polyploids are well distinguished: they are isolated from their progenitors usually by partial or complete sterility. They can also be ecologically distinct. Usually polyploids are treated taxonomically as separate species or subspecies based on the relative difference in the size of their various parts. When it is not possible to find any essential morphological characters for their separation, poly¬ ploids are not classified as separate taxa, e.g. Setaria glauca (L.) Beauv. (27t=18, 36, 72), Sieglingia decumhens (L.) Bernh. (2ri = 18, 36, 124), Calamagrostis epigejos (2n = 2S, 42, 56), Agrostis stolonifera (2n = 28, 35, 42), Hierochloe odorata (L.) Wahlenb. (2n = 28, 42, 56), Bromus ramosus Huds. (2n=14, 28, 42), etc. There is no doubt that experimental taxonomy greatly assists taxonomists in the solution of species problems involving delimi¬ tation, origin, relationship, etc. Modern taxonomy must be based on close co-operation between all branches of biology. It should be emphasized, however, that all such methods can reveal differ¬ ences existing between plants but they cannot show which taxonomic rank should be given to a certain plant. The final conclusion, which will be more objective if based on data obtained from many different sources of investigation, must be left to the taxonomist. STKCIKS I’llOIMiEMS IX ItECENT SCANDINAVIAN WOUKS ON CUASSKS lo7 REFERENC KS. AKEIIBEKG, E. & Bingefors, S., 1953, Progeny studies in the hybrid Poa pratensis X Poa alpina, Hereditas^ 39, 125-136. Ascherson, P. & Graebner, P., 1899, Synopsis der niitteleuropdischen Flora, 2/1. Leipzig. Bennett, A., 1885, Calamagrostis strigosa Hartm. in Britain, J. Bot., 23, 253. - , 1886, On Calamagrostis strigosa (Hartman) as a British plant and two Carex forms new to Scotland, &c., Trans. & Proc. Hot. Soc., Fdinh., 16, 313-316. Bjorkman, S. O., 1951, Chromosome studies in Agrostis, Hereditas, 37, 465-468. - , 1954, Chromosome studies in Agrostis. II, Hereditas, 40, 254-258. Clausen, J., Grun, P., Nygren, A. & Nobs, M., 1951, Genetics and evolution of Poa, Carnegie Inst. Washington, Yearbook, 50, 109-111. Clausen, J., Keck, D. D. & Hiesey, W. M., 1940, Experimental studies on the nature of species. I. The effect of varied environ¬ ments on western North American plants, Carnegie Inst. ^yashington, Publ. 520. - , 1945, Experimental studies on the nature of species. II. Plant evolution through amphiploidy and autoploidy, with examples from Madiinae, Carnegie Inst. Washington, . Publ. 564. CovAS, G., 1949, Taxonomic observations on the North American species of Hordeum, Madrono, 10, 1-21. Crackles, F. E., 1953, Calamagrostis stricta in S.E. Yorkshire, B.S.B.I. Yearbook 1953, 54. CuGNAC, A. & Camus, A., 1931, Revision du Bronius maximus Desf., d’apres Petude des peuplements naturels. Bull. Soc. Bot. France, 78, 327-341. Davies, W. E., 1953, The breeding affinities of some British species of Agrostis, Brit. Agricult. Bull., 5, 313-315. Druce, G. C., 1915, Deyeuxia neglecta Kunth, var, scotica. Hep. Bot. Soc. A E.C., 6, 172. - , 1926, Deyeuxia scotica Dr. in Hayu-ard’s BotanisPs Pocket- Book. London. - , 1932, The Comital Flora of the British Isles. Arbroath. Du Rietz, G. E., 1930, The fundamental units of biological taxonomy, Svensk Bot. Tidskr., 24, 333-420. Flovik, K., 1938, Cytological studies of arctic grasses, Hereditas, 24, 265-376. Gregor, J. W. & Sansome, F. W., 1930, Experiments on the genetics of wild populations. II. Phleum pratense L. and the hybrid P. pratense L. x P. alpinum L., J . Gen., 22, 373-387. Gregor, J. W., 1931, Experimental delimitation of species, Hew Phyt., 30, 204-217. - , 1944, The Ecotype, Biol. Bev. Cambridge Phil. Soc., 19, 20-30. - , 1946, Ecotypic differentiation. New Phyt., 45, 254-270. 158 Sl’ECIKS STUDIES IN THE BRITISH T'JvOllA Gustafsson, a., 1947, Apomixis in higher plants. 111. Biotype and species formation, Lunds Unix. Arsskr. N.F. Avd. 2, 44. Hedberg, O., 1952, Cytological studies in East African mountain grasses, llereditas, 38, 256-266. Heseop-Harrison, J., 1953, New Concepts in Flowering-Plaiit Taxo¬ nomy. London. [loRN AF Rantzien, H., 1946, Taxononiical and phytogeographical studies in Phleum arenarium L., Lot. Notiser 1946, 364-386. Hulten, E., 1950, Atlas of the Distribution of Vascular Plants in A".TF. Europe. Stockholm. Huskins, C. L., 1930, The origin of Spartina Townsendii, Genetica, 12, 531-538. Hubbard, C. E., 1954, Grasses. London. Hylander, N., 1941, Fbrteckning over Skandhiaviens vdxter. Lund. - , 1945, Nomenklatorische und systematische Studien iiber nor- dische Gefasspflanzen, Uppsala Unix. Arsskr., 7. - , 1953, Nordisk kdrlvdxterflora. Stockholm. Jones, K., 1951, Autotetraploidy in Agrostis canina. Nature, 169, 159- 160. - , 1953, The cytology of some British species of Agrostis and their hybrids, Brit. Agricult. Bull., 5, 316. Kiellander, C. L., 1942, A subhaploid P. pratensis L. with 18 chromo¬ somes and its progeny, Svensk Bot. Tidskr., 36, 200-220. Lamprecht, H., 1949, Systematik auf genetischer und zytologischer Grundlage, Agri. Hort. Genetica, 7, 1-28. - , 1953, Petunia axillaris (Lam.) B.S.P. und ihre Synonyme ih violacea Lindl. und P. inflata R. Erics, mit Betrachtungen zum konventionellen und naturbedingten Artbegriff, ibid., 11, 83-108. JiOVE, A., 1952, Preparatory studies for breeding Icelandic Poa irrigata, llereditas, 38, 11-32. JidvE, A. & Love, D., 1948, Chromosome numbers of Northern plant species, Beykjaxik Unix. Inst. Appl. Scien., Dep. of Agron. Pep. Ser. B., 3, 1-131. Meldeeis, a., 1953, Some parallels between the British and Scandina¬ vian mountain floras, in The Changing Flora of Britain, (edit. J. E. Lousley), 89-104. Muntzing, a., 1943, Characteristics of two haploid twins in Dactylis glomerata, Hereditas, 29, 134-140. JMuntzing, a., Tedin, 0. & Turesson, G., 1931, Field studies and experi¬ mental methods in taxonomy, llereditas, 15, 1-12. Nannfeldt, j. a., 1935, Taxononiical and plant-geographical studies in Poa laxa group, Symb. Bot. Upsal., 1/5. - , 1937, The chromosome numbers of Poa, sect. Ochlopoa A. & Gr. and their taxonomiQal significance, Bot. Notiser, 1937, 238- 254. N.4NNFELDT, J. A., 1938, Poa maroccana Nannf. n. sp. and P. rivulorum Maire & Trabut, two more tetraploids of sect. Ochlopoa A. & Gr., and some additional notes on Ochlopoa, Sxensk Bot. Tidskr., 32, 295-321. SI'KCIKS I’ROHr.EMS IN RECENT SCANDINAVIAN WORKS ON GRASSES 159 - , 1940, On the polyniorpliy of Poa arcti^a R.Br., with special reference to its Scandinavian forms^ Si/mh. Bot. Upsal., 4. Nilsson, Heribert, N., 1930, Synthetisclie Bastardierungsversuche in der Gattung Salix, Lunds Univ. Arsskr. N.F. Avd, 2, 27/4. XoRDENSKioLD, H., 1945, Cyto-genetic studies in the genus Phleum, Acta Agricidt. Suecana, 1, 1-138. - , 1949, Synthesis of Phleum pratense L. from P. nodosum L., Hereditas, 35, 190-202. Nordhagen, R., 1954, Apologi for Poa stricta Lindeb., ^^rensk Bot. Tidskr., 48, 1-18. Nygren, a., 1946, The genesis of some Scandinavian species of Calania- grostis, Hereditas, 32, 131-262. - , 1948a, Further studies in spontaneous and synthetic Calamagrostis purpurea, Hereditas, 34, 113-117. - , 1948b, Some interspecific crosses in Calamagrostis and their evolu¬ tionary consequences, Hereditas, 34, 387-413. - , 1949a, Studies on vivipary in the genus Deschampsia, Hereditas, 35, 27-32. - , 1949b, Apomictic and sexual reproduction in Calamagrostis pur¬ purea, Hereditas, 35, 285-300. - , 1950a, A preliminary note on cytological and embryological studies in arctic Poae, Hereditas, 36, 231-232. - , 1950b, Cytological and embryological studies in Arctic Poae, Symh. Bot. Upsal., 10. - , 1951, Form and biotype formation in Calamagrostis purpurea, Hereditas, 37, 519-532. OvcziNNiKov, P. N., 1934, Phleum, in Komarov, V. L., Flora UBSS, 2, 127-135. Philipson, W. R., 1937, A revision of the British species of the genus Agrostis Linn., J. Linn. Soc. Loud. Bot., 51, 73-151. Raven, J. E. & Walters, S. M., 1954, The Inchnadamph Roegnerias, Proc. Bot. Soc. Br. Isl., 1, 88-89. Stebbins, G. L., Jr., 1950, Variation and Evolution in Plants. London. Turesson, G., 1922a, The species and the variety as ecological units, Hereditas, 3, 110-113. - , 1922b, The genotypical response of the plant species to the habitat, Hereditas, 3, 211-350. - , 1929, Zur Natur und Begrenzung der Arteinheiten, Hereditas, 12, 323-333. - , 1931, The geographical distribution of the alpine ecotype of some Eurasiatic plants, Hereditas, 15, 329-346. Tutin, T. G., 1950, A note on species pairs in the Gramineae, Watsonia, 1, 224-226. - , 1951, Origin of Poa annua L., Nature, 169, 160. - , 1953, Poa annua and its parents, B.S.B.I. Yearbook 1953, 60. Wycherley, P. R., 1953, Proliferation of spikelets in British grasses, V'atsonia, 3, 41-56. SI’KCIKS Sri DIKS IN T IIK 15H1T1S1I FLOHA KiO THE CONFLICT OF CATEGORIES J. Heslop-Harrison (Queen’s University of Belfast). The year 1953 marked the 200th anniversary of the publica¬ tion of Linnaeus’s Species Plantarum, the work taken as the starting point, for the purposes of priority, of the present system of nomenclature for the higher plants. One of the principal reasons for accepting the Species Plantarum as a starting point is that in this work Linnaeus employed consistently binomial nomenclature, a system used sporadically by several of his pre¬ decessors, but not before exploited in a uniform and logical manner. This crystallisation of the system of nomenclature for living organisms is significant in a way not often stressed, for it indicates that the Linnaean concept of the species, itself leading back to that of Ray, has, on the whole, satisfactorily stood the test of time, and has proved applicable in a vastly greater field of organisms than was ever contemplated by Linnaeus himself. Indeed, it is a matter of considerable biological interest that through two centuries the Linnaean taxonomic system should, with the occasional interpolation of intermediate categories, have served fairly adequately the needs of systematics. This in itself is an indication that the actual pattern of organic variation is for the most part of a form which can be fitted into a classificatory system composed of a hierarchy of categories. As various biologists have recognised, this means that a basic feature of the variation of living organisms is the existence of discontinuities at various levels. In Dobzhansky’s words (1941), “. . . . the living world is not a single array of individuals .... but an array of more or less distinct separate arrays, intermediates between which are absent or at least rare. Each array is a cluster of individuals, usually possessing some common characteristics and gravitating to a definite modal point in their variations. Small clusters are grouped together into larger secondary ones, these into still larger ones, and so on in a hierarchical order.” Until relatively recently, sy sterna tists on the whole have simply accepted this state of affairs, and have not been seriously con¬ cerned with re-examining the basic assumptions about organic variation upon which their work depends. Energy has been de¬ voted to the completion of the primary survey of the living king¬ doms using the traditional methods, and discussion has been mainly directed towards the problems of codifying the rules governing nomenclature rather than towards any reassessment of fundamental principles. In the last half century, however, methods other than those of comparative morphology have been applied to the analysis of natural variation. Geneticists, cytologists and others, initially basing their investigations on 'I’llK (!<>NKLlC'r OF CATKGOJllKS K)l existing taxonomic arrangements, liave not inlrequently found these unsatisfactory, and have on occasions come to question the whole existing basis of taxonomic methodology (cf. Darlington, 1951). This has been particularly true in alliances where the rigid taxonomic framework does not fit too well, and where the criteria of classical taxonomy, derived mainly from comparative morphology, produce groupings at variance with those exposed by the methods of the newer disciplines. It is naturally at and about the level of the average Linnaean species that these problems have mostly come into prominence — the level of variation with which the present conference is con¬ cerned. In this contribution, I wish to focus attention primarily upon one matter: how desirable or practicable it is for the basis of orthodox nomenclatural taxonomy to be modified to accommo¬ date the newer conceptions of the sources and nature of organic variation. In the literature of the last two or three decades two divergent tendencies are apparent. Broadly these are as follows : — - 1. To attempt to give the word “species” a particular genetical or biological meaning in sexual groups, and to convert the basis of orthodox (nomenclatural) taxonomy to establish the unit so defined as the sole one to bear the binomial, even if this should lead 'to the abandonment of some of the classical criteria of taxonomy ; 2. To serve the ends of experimentalists by the formation of various special purpose classifications with distinctive categories and criteria wherever such are required within the framework of orthodox taxonomy, allowing the latter to develop and adapt its methods and species criteria to absorb such of the experimen¬ tal data as may prove compatible with its functions and facilities, without necessarily abandoning its basis of comparative mor¬ phology. The popularity of the first of these approaches is considerable, albeit mainly among non-taxonomists. The logic of the argument lying behind it is certainly persuasive, and indeed in some ways incontrovertible. In essence it runs as follows : — (a) The taxonomic system currently in use, consisting of a hierarchy of categories, is workable only because of the existence of variational discontinuities in the array of living organisms. (b) These variational discontinuities arise from certain genetical causes, the bulk of which appear now to be understood in some detail. (c) The road, therefore, to a more “perfect” taxonomy is to take cognisance of the existence of this knowledge of causes, and to define taxonomic categories — particularly the fundamental one, the species — in terms of the basic biological properties which, in effect, provide the mandate for the existence of the variational units which are fitted into them. The difficulty is, of course, that there is little or no agreement as to which biological property shall be given primary importance. Attempts at “genetical” definition of species are in no way new, 1(52 .SI-HCIKS STUDIKS IN TUK JJKITISU FLOKA nor for that matter have they all arisen within the half-century life of the science of genetics. A primary attribute required of a species is that it “breed true”, a genetical property as meaningful to Theophrastus as to any modern Drosophihst. However, if the requirement of genetical uniformity within a species, both between the individuals existing at a given time and in the lineage, is pushed to the extreme, the result is that. the homozygous bio- type becomes the only variational unit which can bear the name. This is the outcome of one line of reasoning, and preposterous as it may seem to us to-day, it was seriously put forward by one of the most distinguished geneticists of the early part of this cen¬ tury, Lotsy (1918). Such a concept is, obviously, untenable in any general application. There are no doubt many taxonomic plant species which fulfil the requirement, being essentially homozygous through many generations of autogamy; examples are to be found among our critical species of ephemerals, in Capsella, Stellaria and the like, in some annual groups, e.g., Euphrasia and probably Salicornia, and also among perennials like autogamous Epipactis and Ophrys. But with allogamous organisms, a genetical species concept such as that of Lotsy can have no meaning, for the processes of gene segregation and recom¬ bination ensure that all individuals will in effect be genetically different to some extent. With these sexual outbreeding organisms, the species of taxonomy normally embraces an assemblage of morphologically similar but not necessarily identical individuals. The Lotsyan concept not being applicable to this situation, the geneticist naturally enquires what property these groups possess which preserves their unity and distinction from others. The answer generally found is, of course, some level of reproductive isolation. The idea that the limits of the morphologically recognisable species coincide largely with the boundaries beyond which hybridisation is impossible or productive of sterile offspring is also certainly an ancient one. It was apparent enough to Lin¬ naeus, and Lindley, rather more than a centurj^ ago, penned a definition of the species in which this notion was given promin¬ ence. This concept, also, can be given an extreme interpretation, and has indeed been given such by some workers in the last half century, primarily by geneticists who have accepted “specific differences” between two groups of organisms as being indicated only by a total — or practically total — inherent sterility barrier. Winge (1938) gives as a geneticist’s species concept, “. . . individuals are specifically different when they are unable to hybridise or when, by crossing, they produce more or less sterile progeny”. Where limitations on interbreeding exist as a normal feature in a plant population due to incompatibility mechanisms, dioecism, etc., strict application of a principle as bald as this would result in the segregation as different species TliK CONFLICT OF CxVTEC:OiUF,S iGo of two individuals of the same sex, or of tiie same incompatibility group ! A related viewpoint has recently been upheld by the Swedish geneticist, Lamprecht (1944, 1948) who, on the basis of extensive experiments with legumes, considers that a strict genetical delinition can be given to the species. In effect, Lamprecht’s view contains a development of one stated by Winge (1938), that all species are characterised by genes which cannot be trans¬ ferred in interspecific crosses or which behave as lethals. In Phaseolus, Lamprecht states that free recombination is possible for all of the genes governing differences between the species P. vulgaris and P. cocoineus except for those governing the position of the cotyledons during germination and the form of the stigma. These genes cannot be recombined in fertile offspring, and for this reason Lamprecht refers to them as “primary species- separating”. Generalising from this finding, he argues that the only true species in nature are those which are separated in this manner, and arrives at a definition of the species as “. . . a unit including all biotypes which differ from all others by at least one common interspecific gene”. Lamprecht’s is probably one of the most radical of the species concepts currently held by geneticists, but a viewpoint almost as extreme has been expressed by Clausen, Keck and Hiesey (1939). These investigators, in developing their “experimental concept of the species”, are prepared to accept only those “. . . internal barriers that are of a genetic-physiologic nature” as species-separating ; those, in other words, which prevent the production of offspring even after artificial cross-pollination, or lead to the sterility of any progeny produced. Where such barriers do not exist between two forms, they must be considered as belonging to the same species, whatever the taxonomist may have done with them. Acting on a principle of this sort, some workers have actually attempted taxonomic revisions of flowering- plant groups and have degraded forms lacking inherent isolating mechanisms from species to lower taxonomic rank. Thus, be¬ cause they show high interfertility in cultivation and are not therefore to be considered “biologically different species”, D. Love (1944) has down-graded the taxonomic species Melandrium ruh- rum Garcke and M. album Garcke to subspecies of M. dioecum (L.) D, Love. There can be no doubt that these thorough -going genetical concepts of species limits, initially attractive because of their apparent simplicity and infallibility, have fallen into some dis¬ repute, mainly because as laboratory or experimental garden concepts they simply do not relate in every instance to the situation in nature, at least in the manner their authors assert. A more subtle approach is that of those who have recognised that barriers other than internal-genetic may operate in nature, and who seek to establish a “biological” species concept. Pro¬ ponents of the biological species have, naturally, much common 1G4 bl’EClKS fciTLDlKfci IN THE BKlTibll ELOllA ground with the advocates oi‘ the genetical species. Tliey point to the fact that it is characteristic of outbreeding sexual organ¬ isms tliat they form discrete or almost discrete groups m nature, these being the species, and argue that the formation and con¬ tinued independent existence ot these species requires the origin and maintenance of some form of reproductive isolation. While the latter may not necessarily be of an mherent genetical nature. It must be a biological attribute of all true species. The type of species definition arising from these considerations is well illustrated by Mayr’s formulation (1940); “A species consists of a group of populations which replace each other geographically or ecologically and of which the neighbouring ones intergrade or interbreed wherever they are in contact or which are potenti¬ ally capable of doing so (with one or more of the populations) in those cases where contact is prevented by geographical or ecological barriers. Or shorter : Species are groups of actually or potentially inter¬ breeding populations, which are reproductively isolated from other such groups.” Biological species definitions such as this one of Mayr have been criticised from various viewpoints, often on the grounds of the general ambiguity or indefiniteness of the terms employed, and also because examples can always be quoted where the application is obscure. This would seem to be inevitable because of the nebulous nature of the concept of reproductive isolation once there is a departure from the relatively secure basis of the “inherent genetical barrier”. Dobzhansky has used the phrase “reproductive isolation” in a sense synonymous with “physio¬ logical isolation” to imply that the factor inhibiting the cross¬ breeding of two forms is not solely their geographical separation from each other. A limitation of this nature is necessary, since two colonies separated by a major topographical feature are of course as effectively inhibited from gene exchange as they w'ould be by any inherent genetical barrier, and all such colonies would have to be given “specific” rank if the simple criterion of actual inter-breeding or non-interbreeding were assumed. For this reason, definitions like that of Mayr quoted above have to incor¬ porate such phrases as “. . . actually or potentially interbreeding populations . . .” But it is just this introduction of the idea of potential capacity for interbreeding which creates difficulty. It disallows, for example, one form of inherently- determined isolating factor which is certainly of considerable importance in the plant kingdom. An inherent genetical difference between two groups, itself concerning in no direct manner the processes of reproduction, may neverthe¬ less act as a reproductively isolating factor through the agency of spatial isolation when it determines differences in habitat pre¬ ference. This is the basic aspect of “ecological isolation”, which, while appreciated by most botanists, seems to be incomprehensible or unacceptable to many zoologists, no doubt as an outcome of THE CONFLICT OF CATEGORIES 165 the greater motility of the organisms with which they deal. In such instances, the potentiality of gene exchange exists while the actuality is absent, and its absence is not due to the fortuitous existence of a geographical hiatus in range, but to distributional differences arising from innate dissimilarity in habitat predilec¬ tions. The British flora provides several examples of forms which are reproductively isolated largely if not entirely by ecological differences. The most fully investigated case is undoubtedly that of Silene vulgaris and S. maritima, studied in detail by Turrill and Marsden-Jones (1928-51). Others were discussed by Clapham at the 1948 Conference of the Society, and by Valentine at the Conference of 1950. Also with the two Melandrium species quoted above ecological isolation is certainly important (Baker, 1948), but here the specialisation of the flowers for different pollen vectors and a difference in time of anthesis contribute to some extent. An example like this reveals a further difficulty which arises when¬ ever a species definition such as that of Mayr is taken as a precept for action as apart from simply an interesting concept for academic discussion. In any particular instance, when it comes to the point of actually determining whether two allopatric populations are “potentially interbreeding”, what tests are to be applied? With plants, one can, of course, grow representatives together in the experimental garden. Artificial cross-pollination, such as practised by Clausen, Keck and Hiesey, simply short- circuits some of the subtle barriers which effectively separate some breeding populations in nature, just as growing the plants together eliminates the spatial isolating factor. In actual fact there is no clear level at which the intervention of the experimen¬ ter must cease. In one instance, cultivation in close proximity will be sufficient to promote crossing; in another cross-pollination may be necessary; in a third, cross-pollination plus an operation to reduce style length; in yet a fourth, perhaps artificial culture of the Fj embryo. Any of a large number of techniques is indeed available to test the “interfertility” of two forms, and each will, of course, give a result valid within its own context. But as to what level shall be taken as marking the end of “potential capacity for interbreeding”, there can be no unequivocal rule. This part of so-called biological species definitions can, therefore, hardly be effectively translated into practice. Proponents of such definitions have to some extent realised this. Mayr, for example, states that reproductive isolation can be an immediate practical test only for sympatric, synchronically reproducing species, and admits that it is often impossible for practical reasons to test to what extent reproductive isolation actually exists between geographically isolated populations. He goes on to suggest that “the conspecificity of allopatric . . . forms, which depends on their potential capacity for interbreeding, can be decided only by inference, based on a careful analysis of the morphological differences of the compared forms”. This he con- 166 SPECIES STUDIES IN THE BRITISH FLORA siders feasible because “ . . . the biological gap between species (reproductively isolated groups) is, in general, correlated with certain morphological differences”. With plants, however, this principle can hardly be regarded as generally applicable. In¬ stances from the British flora in which morphologically well- defined forms retain complete interfertility have been quoted above, and several others could be cited to illustrate the reverse — the intersterility of morphologically closely similar forms. Cardamine pratensis is an interesting case. In this Linnaean species several chromosome races exist, some of which are yet to be sorted out as morphologically distinguishable units (Lovkvist, 1947). Professor Tutin will, I believe, be telling us of intersterile races within Glyceria fluitans, barely recognisable morphologically from each other; here the barrier appears to be on a genic or chromosomal-structural level rather than chromosome-numerical. The case is thus similar to that of Datura stramonium, a species in which morphologically similar, intersterile “sectors” exist, the intersterility being due to the presence of chromosomal inversions and translocations. Even although it cannot be regarded as a general truth that reproductive isolation and morphological divergence are neces¬ sarily correlated, the fact that resort has to be made to such a proposition in developing the “biological” species concept has interesting implications. It is clear that the idea of the species advocated by Mayr and his followers is quite a different matter from the experimental concept of the species laid down by such workers as Clausen, Keck and Hiesey. While ostensibly based upon the principle that specific distinction depends upon repro¬ ductive isolation, it does not assert that species limits can, or necessarily should, be determined by experimental tests for sterility or fertility. It does indeed ultimately admit for allo- patric forms what is essentially an orthodox interpretation of what shall be taken as constituting a species, based upon degree of morphological difference; this is so, in spite of Mayr’s argu¬ ment that degree of morphological difference is only to be em¬ ployed as a yardstick in cases where the presence of reproductive isolation cannot be directly determined. In fact, it seems that in sexual groups as we move away from the crude interpretations of what are species which arise from the acceptance of unmitigated genetical criteria, adding proviso after proviso in the attempt to attain a broader biological basis, the units defined become more and more congruent with the morphological species of orthodox taxonomy. Ultimately bio¬ logical species definitions cease to become definitive altogether: they provide no absolute criteria for circumscribing species, and in effect simply describe some of the biological properties of the “good” species of taxonomy. And this, I think, is generally a sign of their success. The effect of the abandonment of anv strict basis of experi¬ mental delimitation by proponents of the biological species THR CONFLICT OF CATEGORIES 1G7 concept is, simply, to re-establish personal judgment as a primary factor in species discrimination. In a doubtful situa¬ tion, decision is made by reference to a complex mental frame¬ work in which considerations of comparative morphology, ethology, distribution, etc., all play a part. The units of a classi¬ fication based upon this sort of complex foundation, in which the same criteria have different significance in different contexts, are unlikely to be in any way uniform in their genetical pro¬ perties. The need for special purpose sub-classifications based upon single criteria would still arise, as it does against the present background of a primarily morphologically based taxonomy. The construction of special genecological (or “experimental taxonomic”) classifications is the second current tendency which I referred to at the outset. It is perhaps desirable for a moment to consider the philosophical implications of this sort of approach, for it is misunderstanding of these which has led to doubts and suspicions in the minds of taxonomists, and to some misinterpretation of the significance of their activities to taxo¬ nomy on the part of the experimentalists themselves. I hope I will not overlap t-oo much here wdth Mr. Gilmour, who has already discussed some of these matters elsewhere (Gilmour, 1940, 1951). All would, I think, concede the need for a general classifica¬ tion of the living kingdoms, and would to this extent accept the principle laid down in the International Codes of Nomenclature that all organisms should find a place in the general classification. Further, without entering upon the debatable matter of the place of phylogenetical speculation in classification, it appears to be true to state that most satisfactory form of general classification is one based upon maximum correlation of attributes of the organisms and groups of organisms involved; this is certainly the most “natural” in the original sense of Linnaeus and de Candolle, since it allows the greatest number of inductive generalisations to be made about the groupings produced. But “maximum correlation” does not mean “complete cor¬ relation” of attributes. It is self evident that, even in the most natural of arrangements, groups placed together may for some characteristics resemble each other less than they do other, more remote groups. A cross-classification based upon these features is always possible, and may for some purposes be desirable. The fact that we may wish to make it is in no way an attack on the original “natural” arrangement. We may take, for example, the matter of latex production. Tliat a text-book of economic botany places together as latex plants representatives of Compositae, Moraceae, Asclepiadaceae and Euphorhiaceae, is not a criticism of the taxonomic arrangement which separates these plants rather widelv in the general scheme of classification. The same principle of cross-classification can be widened indefinitely: ecological classifications for the whole Phanerogamae are pos¬ sible. and have indeed been formed, based, for example, on life 168 SPECIES STUDIES IN THE BRITISH FLORA form. The classifications of so-called experimental taxonomy can be looked upon in a similar sense, as giving recognition to specific attributes of organisms which, in certain contexts, are of special importance. It is at the point where experimentalists assert that their classifications are necessarily the most “natural” and are tempted upon such grounds to challenge orthodox taxonomic arrangements that their activities become suspect. As with an economic classification of plants, there is simply no inherent need to expect or require that classifications based upon special criteria such as degree of interfertility should coincide with those of orthodox taxonomy, and certainly no justification in insisting that the latter be brought into line where the agreement is low. It is perhaps unfortunate that one of the earliest and best known approaches to the “experimental” treatment of the natural variation of plants, that of Turesson, began with a set of categories two of which bore names based upon the root “-species”. It may be largely due to this simple fact that so great a measure of confusion about the concepts involved in this type of classification has arisen. On the one hand, some taxonomists have reacted against the whole scheme, disconcerted perhaps by the suggestion that one kind of species, the coenospecies, could contain several whole taxonomic genera, or that another, the ecospecies, could be composed simply of a morphologically un¬ differentiated local race characterised perhaps by a chromosome inversion. On the other hand, some enthusiasts have seen in the Turessonian system the foreshadowing of doom for orthodox taxonomy, and have hastened to demand nomenclatural recogni¬ tion of the new categories. Knowing that Mr. Gilmour will be taking this matter of the relationship of nomenclatural and experimental categories con- sidera,bly further, I do not propose to pursue it, except to state that I find myself in full agreement with what I believe he is going to say. I think in the present state of our knowledge, there is nothing whatever to be gained by an attempt to incorporate genecological, biosystematical, experimental taxonomic, or what¬ ever they may be termed, category concepts in the formal taxonomic system. There is a motion to open the door to this before the Eighth International Botanical Congress to be held in Paris this year : I can only hope that its sponsors will think a little more carefully of what they are proposing to do before press¬ ing ahead with it. These remarks extend also to variational categories below the rank of species. The case concerning these has been well expressed by our leading genecologist. Dr. Gregor, in his Presidential Address to the Botanical Society of Edinburgh of 1948, and I recommend all botanists interested in these problems to read it. Before leaving the matter of the experimental classification of plants, a word or two on the methods and purposes of this activity may not be out of place. Since the original proposals of Turesson THE CONFLICT OF CATEGORIES 169 and Denser were put forward, now more than a quarter of a cen¬ tury ago, there has been a considerable development of outlook. Following upon a phase of optimistic attempts to improve the precision of the criteria used in experimental classification, there has been an increasing realisation that the complexity of natural variation is such that no absolute standards can be established for defining the experimental categories, any more than for the categories of the orthodox taxonomic system. It has become clear that several different and equally valid classifications of the same groups of populations may be possible using experimental ‘criteria, so far are the latter from establishing the ultimate relationships. In recognition of this, some of us (Gilmour & Heslop-Harrison, 1954) have been seeking recently to establish a terminological system which would be sufficiently flexible to allow a whole sequence of properties and inter-relationships. However, currently developing among genecologists and evolutionists there is a belief that an “experimental taxonomy” as such may not be required at all. As Gregor has suggested, genecological classifications tend to be summaries of experiment¬ ally or cytologically determined facts about natural plant popula¬ tions which bear upon their origin, structure and properties. To express this information, it maj^- not be necessary to resort to a classificatory approach, and certainly no nomenclatural system is required. It may even be that the carrying over of taxonomic concepts seen in the tendency to define ‘types’ and ‘species’ is actually an impediment to genecological research. An example of this may perhaps be seen in the early history of the ecotype, for there is no doubt that here the overstressing of the ‘type’ aspect long tended to conceal the existence of ecologically con¬ ditioned clinal variation. The problems of genecology are the problems of the inter¬ relationship and interaction of organism with organism, and organism and population with the secular environment, all in the continuum of time. Taxonomic typification has little to do with such a study, for the recognition of stages is but a poor substitute for the investigation of processes. In conclusion I feel obliged to add a word or two on what Dr. Turrill terms the synthetic approach to taxonomy, lest it be thought from the foregoing that I am advocating a retention of the status quo in the science. I began with the postulate that since the Linnaean method had persisted in its essentials for two centuries and had seen a fair measure of success in that period, the presumptions upon which it was based could hardly be fundamentally incorrect. Modern population theory does little to contradict this in any general manner; it does, in fact, provide an explanation for the pattern of variation which fits reasonably well a taxonomic sys¬ tem composed of a hierarchy of categories. The essentially Linnaean species concept which is the basis of our present nomenclatural taxonomy is not, of course, without 170 SPECIES STUDIES IN THE BRITISH FLORA its defects in practice. Perhaps it is hardly necessary to specify the evidence for this to the present audience: it lies in any flora. A day or two ago, while pondering on the topic of this conference, I dipped at random into the new Flora of the British Isles, and leafed through a page or two from the one I first encountered. In six openings, I struck eight families. All told, in the vicinity of the pages I hit upon, almost half of the genera listed bore some indication of confused taxonomy. In several instances, the evidence was of man-made nomenclatural confusion, but in the remainder the problems were inherent in the plants themselves. In about one-third of the cases, the trouble arose from apomixis : * one could, of course, hardly avoid hitting Rubus or Hieracium at least once. In one genus, the difficulty arose apparently from autogamy, in another from possible hybridity. This small study was informative, and an analysis of the entire flora on these lines would no doubt be highly illuminating. It illustrates the interesting point that the bulk of the problems of plant taxonomy in a geographical region the size of ours arise from special forms of reproductive behaviour which, while not peculiar to the higher plants, are much more prevalent among them than, say, among the higher animals. In apomictic and autogamous groups, the species concept of nomenclatural taxo¬ nomy certainly needs overhauling. I have deliberately refrained from discussing these groups in this paper, for in what direction an improvement is to be sought I am not prepared to say. But I commend the whole topic of the reproductive biology of British plants to the Society as a possible subject for a future confer¬ ence or symposium. A striking feature is that for the sexual forms in our flora there is on the whole surprisingly little confusion. Where it does arise, there are two primary sources of discord; firstly, vertically in the taxonomic hierarchy, over the matter of the taxonomic ranks to be awarded to groups of low degrees of differentiation, and secondly, horizontally in the hierarchy, over the number and position of the lines of demarcation in alliances where there is some degree of intergrading variation due to imperfect differen¬ tiation or hybridity. It is here that our orthodox taxonomy has a clear oppor¬ tunity, and indeed an obligation, to develop and adapt its methods, and to make use if necessary of experimental data to produce a superior orthodox taxonomic arrangement. It is here, indeed, that the attitude urged by the proponents of the 'biological’ species concept is likely to be the fruitful one. In dipping into the new flora, I encounter Ranunculus aquatilis. Here there are listed five subspecies, all accepted as species in other recent critical treatments of the genus in Britain, if under different names. Here is an example of variation of treatment in both dimensions of the taxonomic system, and this is surely an instance of an alliance the taxonomy of which would benefit from experimental studies. How much clearer the varia- THE CONFLICT OF CATEGORIES 171 tion pattern would be with some comprehensive and reliable data on phenotypic plasticity, chromosome numbers, reproduc¬ tive biology and the like. Another example worth considering is that of Salix. Here there is every evidence of a wealth of hybridisation among our native forms. Perhaps no orthodox taxonomic treatment will ever be adequate for this genus throughout the Palearctic region, but at least our knowledge of the source of the variational com¬ plexities in the group should guide our attempts to produce one, and should certainly guard us against the excesses into which some essentially non-biological systematists have been led. Chromosome numbers point to where to look for variational foci ; population studies offer a method of detecting them, and bio¬ metrical techniques a means for expressing our findings. Know¬ ledge of breeding behaviour and the extent and effects of hybri¬ disation, already available in the work of Heribert-Nilsson, should warn us against slavish attempts to apply a ‘type’ con¬ cept of the species. They should certainly check — or at least throw into disrepute — the activities of the name conjurers, whose species limits are set by the finest morphological differences they can detect — or fancy they can detect^ — upon herbarium sheets, and who are accordingly prepared to continue to describe and name hybrid segregates as long as they can persuade botanical journals to publish them. References. Baker, H. G., 1948, Stages in invasion and replacement demonstrated by species of Melandrinm, J . Ecol., 36, 96-119. Claphaim, a. R., 1948, Ecology and critical groups, in British flowe.rinq plants and modern systematic methods (ed. A. J. Wilmott). London. Cpaitren j.. Keck, D. D. & Hiesey, W. M., 1939, The concept of the species based upon experiment, Amer. J. Bot., 26, 103-106. Darlington, C. D., 1951, Do the chromosomes fit the species? Nature, 167, 662-3. Dorzhansky, T., 1941, Genetics and the origin of species. New York. Gtlmottr. j. S. L., 1940, Taxonomy and Philosophy, in The New Sys- ternatics (ed. J. Huxley). London. - , 1951, The development of taxonomy since 1851, Adr. Sci., 8, 70-74. Gilmour, j. S. L. & Heslop-Harrtson, J., 1954, The deme termino¬ logy and tbe units of micro-evolutionary change, Genetica, 27, 147-161. Gregor, J. W., 1948, vSome reflections on infraspecific ecological varia¬ tion and its classification, Trans, d? Proc. Bot. Soc. Edinh., 34, 378-91. liAMPRECHT, H., 1944, Die genisch-plasmatische Grundlage der Artbar- riere, Agri Hort. Gen., 2, 75-142. , 1945, Tntra- and inter-specific genes, Agri TJort. Gen., 3, 45-60. 172 SI’ECIRS STUDIES IN THE BRITISH FLORA Lotsy, J. P., 1918, Ou’est-ce qu’nne espeoe?, Arch. Neerl. Sci. Nnf.. Ser. ‘,]h, 3, 57-110. Love, D., 1944, Cytogenetic studies on dioeeions iVlelandrinm, Bot. Not., 1944, 125-213. Lovkvist, IL, 1947, Chi-omosome studies in Cardamine, TIereditaa, 33, 421-422. Marsden-Jones, K. M. & Turrill, W. B., 1928-1951, Researches on Silene vulgaris and S. maritima, Ken: B'nll., 1928-1951. JMayr, R., 1940, Speciation phenomena in birds, Amer. Nat., 74, 249- 278. - , 1942, Systenintic.. New York. Valentine, D. H., 1950, Geographical distribution and isolation in some British ecospecies, in The Study of the T)i. allelomorphic systems, 20 Allium, 25; vineale^ 180 AUomyces, 77; javanicus, 74 Alnus glutinosa, 40 Alston, A. H. G., 91 Alternaria, 79 Anderson’s hybrid index, 84, 126 Andreas, Ch. H., ill— > Anemone nemorosa, 180 Anthoxanthum alpinum, 155; odorntum., 155 Anlrophyum plantagineum, 92 apical buds, 57 apomixis, 49, 75, 88, 141, 170, 178 Arber, Agnes, 28, 50 Arctium, 180 Arenaria leptoclodos, 180; serpylHfoHn, 180 Armstrong, S. F., 22 Arum italicum, 136->; m a cut a turn , 136— >; neglectum, 136— > Ascobolus eguinus, lb Aspergillus, 73; niger, 70 Aspienium odiantum-nigrum, 91, 94, 104; affine, 92; lunulatum, 92; trichomanes , 91, 94, 95, 99-> AthyHum macrocarpum, 92; solenop- teris, 92 Atriplex, 180 autogamy, 134, 162 Bahington, C. C., 22 Raker, H. G., 20, 89, 118, 125 ‘Baldwin’ effect, 48 Beilis perennis, 180 Betula pubescens, 180; verrucosa, 180 Biological Flora, The. 179 biotypes, 140 Biscutella, 98 breeding, 34 Bromus erectus, 143; lepidus, 143; seca- linus, 143; sterilis, 143 bryologists, nature of, 86 Bryophyta, 8&-> bulb structure, 25 Burges, A., 65—^ Burnett, J. H., 32-^ Burrows, B., 83— > Burtt, B. L., 48 Butcher, R. W., 89, 176, 181, 182 Calamagrostis, 143; arundinacea, 144; canescens, 144; chalybea, 144; epige- jos, 144; lapponica, 144; neglecta, 144; purpurea, 144; stricta, 144; varia, 144 Calamintha a seen dens, 180 CalUtriche, 116->; hanmlata, 116; oh- tusangula, 116; platycarpa, 116; palustre, 116; stagnalis, 116 Caltha, 107— >; minor, 107; palustris, 107; radicans, 107; zetlandica, 107 Capseila, 162 Cordamine, 125; pratensis, 166 Cenlaurea, 178 Centaurlum. latifolium, 178; littorale, 178-^; minus, 178^; pulchellurn, l78-> Cephaloziella, 87 Cerastium tetrandrum, 48 Ceylon, fern flora, 92— >■ Chamaenerion angustifolium., 180 ('helianthes farinosa, 92 Cinnamomum, 29 citation, 141 Clapham, A. R., 65, 183 coenospecies, 72 colchicine treatment. 19, 20, 144 Colletotricbum. gloeosporioides, 69 commiscua, 113 comparia, 113 convivia, 113 Convolvulus arvensis, 180 corticium coronilla, 72 cotton, 16 Crepis, 19 Cretaceous period, 29, 31 ( ultigens, 19 cultivation of plants, 25, 128— >, 134, 165, 180 culture experiments, 69, 83 culture growths, 71 Cyclosorus parasiticus, 93; repandulus, 93 Cystopterus dickieana, 91; fragilis, 91 INDEX 185 rytology, 17, 18, 19, 24 , 26, 36, 46 , 64 , 76, 86, 88, 89, 90, 99, 104, 105, 108, 116, 124, 130, 136, 144, 160, 171, 181 cytotaxonomy, 76 cytotypes, 95 Dactylis glomerata, 98 Dahlia, 79 Dalby, D. H., 134 Darlington, C. D., 34, 161 Darwin, C., 33 Datura stmmoniurn 166 deme, 175 Deschampsia, 22, 143; alpina, 149; botl- nica, 149; cespitosa, 40; flexuofia, 149; setacea, 149 Digitalis purpurea, 42 diploids (see cytology) distribution of plants, 90— >, 99, 116, 119, 122, 13fr-> Distribution-Maps Scheme, Tlie, 183 Dobzhansky, T., 160 Draba, 72 dragonfly, 31 Drepanoclados lycopodi aides, 87; seadt- neri, 87; wilsoni, 87 Di'ijoptens abbreviata, 94; borreri, 91, 94; cristata, 105; dilatata, 91, 94, 96, 105; fllix-mas, 91, 94; spinulosa, 105: uliginosa, 105; villarsii, 91 dysploidy (see cytology) ecodeme, 175 ecology, 21 ecological isolation, 165 eeospecies, 113 Eleocharis palustris, 36 Elliot, E. A., 106 Endymion nonscriptus, 180 Eocene beds, 30 epicormis branches, 56 Epipactis, 162 Erica ciliaris, 126; mackaiana, 126; tetralix^ 126; watsoni, 127 Erioxylum aridum, 18 Erodium cicutarium, 180 Eraphila, 72, duplex, 36. 46 Eucalyptus, 55 Euphrasia, 128->, 162; brevipila, 131; confusa, 130; frigida, 128; montana, 128; nemorasa, 128; pseudokeineri, 128 evolutionary development, 26, 66 ('xperimental taxonomy, 174 Festuca ovina, 143; rubra, 143 Fom.es pinicola, 40, 46 fossil evidence, 27, 59 Frag aria, 19; vesca, 180 Fraxinus, 55 Fucus, 46, 83->; cerauoides, 83; imflatus, 46, 83; serratus, 83; spiralis. 83; vesi- culn.sus, 46, 83 Fungi, 65-> Fusarium, 70, 73, 79, 80 Galeopsis tetahit, 141 Galium palustre, 36 garnetophytes, 87 gamodeme, 175 Gay, P. A., 126^ genecodeme, 175 genecology, 140 genetical relationship, 17 genet ical species, 162 genodeme, 175 genotype, 35, 87 Gentianella, 180 germination, 25 Gilmour, J. S. L., 50, 167, 173->, 178 Gleichenia linearis, 92 Gloeosporium, 68 Gossipioides brevilanhim, 17; kirkii, 16; thurberi, 16 Gassypium, 16-> giafting, 17 Gramineae, 22, 140 Gregor, J. W., 168 guard cells, 23 haploids (see cytology) Harland. S. C., 16 Havvkes, J. G., 132 Haygarth-Jackson, A. R., 16 Hebe, 49 Heslop-Harrison, J., 26, 160->, 176, 178, 181 heterokaryosis, 70, 75, 80 hexaploids (see cytology) Hieracium, 80, 170, 178 Hordeum brachyantherum, 143; califor- nicum, 143; nodosum., 143; secalinum, 143 Huxley, J. S., 50 hybrid index, 84, 126 hybridization, 17, 34, 44, 55, 74, 83, 95. 124, 126, 130, 146, 165 iTypolepis punctata, 92 incoinbatibility, 118 infra-specific units, 74 International Code of Nomenclature, 167->, 173, 179 Iris pseudacorus, 36 Jasione montana, 180 Jones, E. W., 86— > ‘Jordanons’, 72 Jurassic age, 31 Koelerin albescens, 24 Lactuca caxiadensis, 141 Lamium purpureum, W, 47 Lampreeht, H., I6:i 186 SPKCTFS STUDIES IN THE BRITISH FLORA lafi‘X plants, 167 leaf spectrum, 56 leaf structure, 22, 1-42— > LeptocJiilus lanceoUitiis, 92 life of plants, llO fjliaceae, 79 Tannaeus, 160 “lanneiins’, 72 Lodfte, S. M., 83 LoJium, 143 Long-field, C., 31 Lotus cornicutntus, 115 Lousley, .T. E., 26, 114 I.ovis, J. D.. 26, 88. 91, 99— > McVean, D. H., 114 I .\falva sylrestris, 180 Manton, I., 90— 99, 106 Matricaria marltirna, 180 Mayr, E., 50, 164 Meikle, R. D., 114 j Mclandiium athvm, 163; (tioccum. 163; | nil) rum, 163 M clarnpr/nim, 180 Molderis, A., 140-^ Melica, 22 I Melville, R., 31, 55->, 103, 106 | Mentha, 180 mierospecies, 178 mixtospecies, 141 Monitia sltophila, 79 Mnnntropa hypophcyia. 36; h tipopith ys, 36 mutatiotis, cytoplasmic, 69; genic, 69 Alycena. 74, 75 \purospora, 69, 79 ynr Sy sterna tics. The, 33. 179 ‘Noiiienclatorialisnr, 71 non-floral cliaracters. 21, 5.5—^, l34-> O'Connor, W. T. M.. 119 Odonata, 31 Oligoceno age. 29 ontogenetic cycle, 56 Ophrys, 162 ()rchis, 79 firiqa/num vulgare, 180 ftrlyin of Species, The, 33 Osmunda reyalis, 95 OnnstPd. J., 20, 106, 127, 182 palaeontology', 27. 59 ranigralii, G., 91 f*anus stipticus, 42 Vaparer duMum,, 180, rhoeas, 180 Paris quadrlfotia , 38 Penicillium., 73 T’ennian age, 31 ‘Petripatellism’, 71 phenotype, 34, 35 Phlevm alpinum, 141, 150; arena riutn , 150; commutatum, 141. 150; nodosum. 141, 150; phleoides, 150; pratense, 141, 150 Poa, 143; arctica, 146; flexuosa, 146; ylauca, 149; jerntlandica, 147; laxa, 147, pratensis, 146; subcaerulea, 147 pollen fertility, 127 pollen in peat, 29 Polygala, 180 Polygonum,, 180 Polypodium, 96; vulgare, 91, 95, 96 polyploidy (see cytology) Polystichum aculeatum,, 98; Icmchitis x aculeatum , 96; setifemm. x aculea¬ tum., 96 Poulter, R. A., 115 Plantago lanceolata , 129; marltirna, 42, 48 Prime, C. T., 136— > Primus insititia, 180; splnosa, 180 P sail iota, 78 Psilocybe bultacea, ITy, coprophila, Tr, scocholm.ica, 75 Pteridophyta, 90— > Pteris ensiformis, 92 Puccinellia, 22, 143; distans, 26; fascicu- lata, 26; festuciformis, 23; marltirna, 23; pseudodistans, 23, 26 Puccinia coronata, 77; graminls, 74. gram inis tritica, 67 Quaternary beds, 29, 30 Quercus, 29; petraea, 55; robur, 55 Banunculus aguatilis, 170; ficaria, 36; flammula, 180 Raven. C. E., 15, 32 Ray, John, 32, 160 leproduction. 87 reproductive biology, 170 Tlhinantlius, 180 Jlibes odoratum x sanguineum, 64 llicinus communis, 42 lloegneria. behmii, 153; borealis, 152; ca.nina, 1.52; fibrosa, 152; mulabilis, 152 nubus, 80, 87, 170 Jlumex obtusifollus, 40 rusts. 67 Salicoi'nia, 25, 134^, 162; gracilUma, 134 Sal lx, 22, 87; a trocinerea, .55; caprea, 36 Salvia horminoides, 180 Sand-wnth, N. Y., 127 Saxifraga. hirsuta x spathularis, 44 Schotsman, H. G., 116-> selective pressures, 67 Senecio aquaticus x jacobaea, 44; galli- cus, 18; inaequidens, 18; squalidus, 18, 20; vulgaris, 18 Sesleria, 143 Sbivas. M. G., 91— >, 104 Sllene maritima, 165; vulgaris, 165 Sonchus asper, 180; oleraceus, 180 rNDKX 187 Solan nm, 19, 79 Spnrtina townsendii, 07 species, definitions of, 15. too— >, 17-i, 178 species limits, 68, 131, \Al) Spergiila arvensis^ 180 Spfiaerncarpus texanns, 87 spore characters, 88 Sporodima grondis, 75 sporophyte, 87 statistical methods, 24, 26 Stellnria, 162; media, 180; nemorum, lll-> stomata, 23 Stratiotes aloideft, 30. 3i, 59 siiperspecies, 141 Sgmpfigtvm., 180; officinale, 44 sytitlK'tic taxonomy, 177 Taraxacum, 49, 178 taxonomy, experimental. 18, 167; mor¬ phological, 178; synthetic, 177 Tertiary beds, 29 tetraploids (see cytology) Thaiictnim, 180 Thelypteris hrunnea, 92: flaceida, 92 Thomas, H. Hamshaw, 27—^. 59 TJnirbetia. thespesioides, 16, 18 tide levels, 83 time slice, 65 topodeme, 175 TrichoJcwno argyracetim,, 78; brevlpes, 78; chrysites, 79; excissnm., 78; humile, 78; melaceiicum, 78; oreinum, 78. paedJdum, 78; palulum, 78; phaeopo- dium, 78: scaiptarolnm , 78; siibpal- veraientum, 78; leireiim, 78 Trifoliurn repens, 42, 47, 49 Trillium harntscfiaticum, 38 triploid (see cytology) Turrill, W. B., 98, 176, 177-> Tutin. T. G., 15. 21->, 166, 181 THmus. 22; carpinifolia, 58; coin tana, lOO eglantissima, 61; glabra, 55; plotii 56; procera, 55 Valentine, D. H., 31, 64, 103, 127, 132, 176 Valeriana officinalis, 36, 46 variants, 70; cryptic, 35; distinct, 35 variation in plants, 21, 32, 68, 84, 105, 111, 115, 118, 119, 129, 134, 137 variation, chromosomal, 37, 39; ecotypi- cal, 118; genic, 40; phenotypic, 21 Veronica, 49; anagallis-aqualica, 40, 48 Vida angnstifolia, 180; saliva, 180 Viola, 49; canina, 125; lacfea, 125; liriniana,, 36, 46 Walker. S.. 64 . 91. 105-^ W'a Iters, S. M., 183 W5arburg, E. F., 133, 139, 181 wheat, 67 VVinge, (”)., 162 W'oodsia alpina x ilrensis. 96 Yeo. P. F.. 128-> Young, D. P., 125 Zea. 19 Zenner, 59 B.S.B.I. CONFERENCE REPORTS o No. 1. BRITISH FLOWERING PLANTS AND MODERN SYSTEMATIC METHODS. Being the report of the Conference on “ The Study of Critical British Groups ” arranged in 1948. Edited by A. J. Wilmott, pp. 102 + 18 half-tone plates. Price 10/-, plus postage 4d. No. 2. THE STUDY OF THE DISTRIBUTION OF BRITISH PLANTS. Being the report of the Conference arranged under this title in 1950. Edited by J. E. Lousley. Demy 8vo., pp. 128, with about 28 plant distribution maps and other illustrations. Price 10/-, plus postage 4d. No. 3. THE CHANGING FLORA OF BRITAIN. Being the report of 'the Conference arranged under this title in 1952. Edited by J. E. Lousley, pp. 204 + 6 half-tone plates and 25 figures in the text. Bound in buckram. Price 15/-, plus 9d postage. No. 4. SPECIES STUDIES IN THE BRITISH FLORA. Being the report of the Conference on “ The Species Concept in its relation to the British Flora ”, held in 1954. Edited by J. E. Lousley, pp. 190 + 2 half-tone plates, and 24 figures in the text. Bound in buckram. Price £1, plus 9d postage. The above are obtainable from the Botanical Society of the British Isles, cfo Department of Botany, British Museum [Natural History), Cromwell Road, London, S.W.7. THE BOTANICAL SOCIETY OF THE BRITISH ISLES Membership is open to all persons, whether amateur or professional, interested in British botany. Societies, Libraries, Museums, Botanical Gardens and similar Institutions may become Subscribers. Subscription rates are : — 1 guinea per annum for Ordinary Members and Subscribers; 10s 6d for Junior Members (under 21 years of age). The journal Watsonia and the Proceedings are each issued twice a year free to members and subscribers, who are also entitled to other important privileges. Full details of these are given in the prospectus to be obtained from the Honorary General Secretary, c/o Department of Botany, British Museum (Natural History"), Cromwell Road, S.W.7.