a sR. a . ht: + 4) OV Oe rn4 iC tt ‘ y ; a m eres +5 SP Se hFat ie Ae BYNCMNTC ute i "9 Pree Beads tke VATE IY 1 Wa ‘ rhs ' SORT niga t ' ” SF Va big AS URN gh es AS t he k aye ae HL t ‘ ih Hy } ewes a anh hAL eS brash: Rerut y pe »\ ni + yy oe yurt 4 a ” | te wtb tal Fe. wo ltte - oS ade, et el et ee oe LP viet Yiletatg 1 > . 0 : Sen a q pu okie . ¥ : _ _ ve . J ‘ a Ms ? AAA : teva a te ea ORES ! 7 “) J ‘ 4 j q 4 t 4 4 . ai 7 ! ts “Mais to, egy OF Mn, Tg “by sa | “i See cox S- & ‘its E A) - y, ca a MiB ts. TAL Wah CHP Pa ge Toth, (1 agit fa iho aN a! My 1, ge x Fe Peas if a ad lt site ae ee Se Le pay Eee ee ea fe ASS te cis Pi oh, mtdere ue ans, a Eeerecese oe A ge viper i aa Sh Aeoen TiN eed sal ava FA a res “es Soe ie wie Puy Fer Pa, at eel OR a: yet WATSONIA VOU RIVAL, OF LEE BOTANICAL SOCIETY OF THE BRITISH ISLES Editors: C. D. K. Cook M. C. F. Proctor N. K. B. Robson Vol. 6 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES AT THE DEPARTMENT OF BOTANY, BRITISH MUSEUM (NATURAL HISTORY), LONDON 1964-68 DATES OF PUBLICATION Part 1, pp. 1-84, 14th July, 1964 Part 2, pp. 85-140, 20th October, 1965 Part 3, pp. 141-204, 13th December, 1965 Part 4, pp. 205-270, 7th November, 1966 Part 5, pp. 271-326, 28th July, 1967 Part 6, pp. 327-403, 27th July, 1968 PRINTED IN GREAT BRITAIN BY HEADLEY BROTHERS LTD I09 KINGSWAY LONDON WC2 AND ASHFORD KENT CONTENTS Part 1: THE GROWTH OF EUPHRASIA IN CULTIVATION. By P. F. YEO NOTES ON THE STIGMA MORPHOLOGY AND FLOWERING BEHAVIOUR IN BRITISH SALICORNIAE. By I. K. FERGUSON INFRASPECIFIC VARIATION IN LATHYRUS NISSOLIA i By ik F. M. CANNON CHROMOSOME NUMBERS OF EPILOBIUM IN BRITAIN. oe PETER H. RAVEN and D. M. Moore STUDIES ON BRITISH PANSIES. I. CHROMOSOME NUMBERS AND POLLEN ASSEMB- LAGES. By A. PETTET STUDIES ON BRITISH PANSIES. IJ. THE STATUS OF SOME INTERMEDIATES BETWEEN VIOLA TRICOLOR L. AND V. ARVENSIS MURR. By A. PETTET MIMULUS HYBRIDS IN BRITAIN. By R. H. ROBERTS STUDIES ON ALCHEMILLA FILICAULIS BUS. SENSU LATO AND A. MINIMA WALTERS. Ii]. ALCHEMILLA MINIMA. By MARGARET E. BRADSHAW BOOK REVIEWS Panes: A REVISION OF THE BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA [FRIES] F. N. WILLIAMS. By P. D. SELL and C. WEsT FRUIT VARIATION IN POLYGONUM PERSICARIA L. By J. TIMSON LEAF MEASUREMENTS AND EPIDERMIS IN POA ANGUSTIFOLIA. BARLING are a a SULPHUR AND THE DISTRIBUTION OF : BRITISH PLANTS. By H.J.M. BOWEN THE ASSESSMENT OF THE TAXONOMIC STATUS OF MIXED OAK Creer a POPULATIONS. By A. CARLISLE AND A. H. F. BROWN DACTYLORHIZA NEVSKI, THE CORRECT GENERIC NAME OF THE DACTYLORCHIDS. By P. F. HUNT and V. S. SUMMERHAYES BOOK REVIEWS By D. M. Part 3. STUDIES ON BRITISH PANSIES. III. A FACTORIAL ANALYSIS OF MORPHOLOGICAL VARIATION. By A. PETTET THE STATUS OF THE PEDUNCULATE AND ‘SESSILE. OAKS IN BRITAIN. By J. E. COUSENS THE DISTINGUISHING CHARACTERS AND GEOGRAPHICAL DISTRIBUTIONS OF ULEX MINOR AND U. GALLI. By M.C. F. PROCTOR DIPLOIDS IN THE GENUS COCHLEARIA. By J. J. B. GILL NATURAL HYBRIDIZATION BETWEEN THE COWSLIP (PRIMULA \ VERIS fie) AND THE PRIMROSE (P. VULGARIS HUDS.) IN BRITAIN. ByS. R. J. WOODELL BOOK REVIEW Part 4. A LIST OF INFRASPECIFIC TAXA OF BRITISH PHANEROGAMS TESTED IN CULTI- VATION. By D. E. ALLEN ae oe THE BREEDING RELATIONSHIPS OF SOME EUROPEAN EUPHRASIAE. By P. F. YEO STUDIES IN RANUNCULUS SUBGENUS BATRACHIUM (DC.) A. GRAY. III. RANUN- CULUS HEDERACEUS L. AND R. OMIOPHYLLUS TEN. By C. D. K. CooK STUDIES ON WELSH ORCHIDS. III. THE COEXISTENCE OF SOME OF THE TETRAPLOID SPECIES OF MARSH ORCHIDS. By R. H. ROBERTS BOOK REVIEW ili 85-105 106-108 109-113 114-119 120-127 128-133 134-139 141-160 161-176 177-187 188-189 190-202 203 205-215 216-245 246-259 260-267 268-269 Pane: FURTHER OBSERVATIONS ON TRIFOLIUM OCCIDENTALE. P. MORISSET NOTES ON MYOSOTIS SCORPIOIDES AGG. By D. WELCH TAXONOMY AND NOMENCLATURE OF THE RADIATE VARIANTS OF SENECIO VULGARIS L. By D. E. ALLEN ON THE IDENTITY OF THE IRISH POPULATIONS OF SISYRINCHIUM. INGRAM. . CLEISTOGAMY IN SPARTINA. By J. Cc BE. HUBBARD TAXONOMIC AND NOMENCLATURAL NOTES ON THE BRITISH FLORA BOOK REVIEWS By D. E. CoomBE and By RUTH Part 6. THE BROMUS MOLLIS AGGREGATE IN BRITAIN. By PHILIP SMITH CHROMOSOME NUMBERS OF ORNITHOGALUM UMBELLATUM L. FROM THREE LOCALITIES IN ENGLAND. By R. CZAPIK COMPUTER MAPPING OF SPECIES DISTRIBUTION IN A COUNTY FLORA. By J: G. HAWKES, B. L. KERSHAW and R. C. READETT PARIETARIA OFFICINALIS AND P. JUDAICA. By C. C. TOWNSEND THE HYBRIDS OF MIMULUS CUPREUS. By R. H. ROBERTS THE INTERSPECIFIC RELATIONSHIPS OF JUNCUS EFFUSUS AND J. CONGLOMERATUS IN BRITAIN. By A. D. Q. AGNEW NOMENCLATURAL NOTES ON THE GENUS ANTHYLLIS I, By i CULLEN . BOOK REVIEWS ERRATA os INDEX TO VOLUME bE iV 271-275 276-279 280-282 283-289 290-291 292-318 319-325 327-344 345-349 350-364 365-370 371-376 377-388 389 390-396 S07 399-403 —————— 560.544 m3 ee NIA JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 6 MAY, 1964 Pt. 1 CONTENTS THE GROWTH OF EUPHRASIA IN CULTIVATION. By P. F. YEO ip: a 1-24 NOTES ON THE STIGMA MORPHOLOGY AND FLOWERING BEHAVIOUR IN BRITISH SALICORNIAE. By I. K. FERGUSON ee te ae ae ea 25-27 INFRASPECIFIC VARIATION IN LATHYRUS NISSOLIA L. By J. F. M. CANNON 28-35 CHROMOSOME NUMBERS OF EPILOBIUM IN BRITAIN. By PETER H. RAVEN and D. M. Moore Ke Be ae a a: bie uF oe 36-38 STUDIES ON BRITISH PANSIES. I. CHROMOSOME NUMBERS AND POLLEN ASSEMBLAGES. By A. PETTET ae sa Es oe ate & 39-50 STUDIES ON BRITISH PANSIES. II. THE STATUS OF SOME INTERMEDIATES BETWEEN VIOLA TRICOLOR L. AND V. ARVENSIS Murr. By A. PETTET x, 51-69 MIMULUS HYBRIDS IN BRITAIN. By R. H. ROBERTS oF =e Bb 70-75 STUDIES ON ALCHEMILLA FILICAULIS BUS. SENSU LATO AND A. MINIMA WALTERS. Ti]. ALCHEMILLA MINIMA. By MARGARET E. BRADSHAW we . 76-81 BOOK REVIEWS a ve ay Ee a on at 5 i 82-84 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH MUSEUM (NATURAL History), LONDON PRICE: TWENTY-FIVE SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES Patroness: H.R.H. THE PRINCESS ROYAL OFFICERS President: J. E. LOUSLEY Vice-Presidents: Dr. S. M. WALTERS, Dr. C. E. HUBBARD, Dr. J. G. DONY, and R. MACKECHNIE Hon. General Secretary: E. B. BANGERTER, c/o Dept. of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7 Hon. Treasurer: J. C. GARDINER, Thrift House, 12 & 14 Wigmore Street, London, W.1. Hon. Editors: (Watsonia) M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT Hon. Meetings Secretary: Mrs. MARY BRIGGS, White Cottage, Slinfold, Sussex Hon. Field Secretary: P. C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Dartford, Kent. Hon. Junior Activities Secretary: P. F. HUNT, The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey. Applications for membership should be addressed to the Hon. Secretary, from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 25/— per part. To be obtained from E. B. BANGERTER, Botanical Society of the British Isles, c/o Dept. of Botany, British Museum (Natural History), London, S.W.7. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/— for Junior Members, and 10/— for Family Members, and the Society’s year runs from January Ist to December 31st. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. THE GROWTH OF EUPHRASIA IN CULTIVATION By P. F. YEO University Botanic Garden, Cambridge ABSTRACT Previous workers found that Euphrasia could be brought to maturity in cultivation whether supplied with a host-plant or not (though better growth was obtained with a host than without), and found also that a wide variety of annual and perennial herbs could act as hosts. The present paper describes chiefly cultures in clay pots, into which the Euphrasias were transplanted as seedlings. Usually growth was better when a host was provided than when it was not, but growth without a host was better than that obtained by previous workers. The development of some plants without a host was much affected by a difference of soil composi- tion. A spurt of growth some weeks after planting with a host was taken to indicate parasitic establishment on the host, and the shorter the distance between the Euphrasia and its host the sooner establishment occurred. Euphrasias grown with certain Leguminosae were particularly vigorous and showed symptoms of a good nitrogen supply. Some severely retarded Euphrasia plants were probably being parasitized by their neighbours. Euphrasias sometimes flowered very early, beginning at a very low node; this appeared to be caused by unusually high temperatures. The early-flowering species were particularly frequently affected in this way, which made it difficult to raise normal plants of these species. Deaths of young Euphrasia plants took place chiefly in cold, dull weather, but were reduced by watering sparingly. Cultures of various Euphrasia species with various hosts showed that certain hosts were good, and that others were bad or probably bad. There were indications of a difference between E. pseudokerneri and E. anglica in their growth on the same hosts. Different populations of the same Euphrasia species sometimes differed in their host-reactions. The host affected percentage establishment, mortality and vigour of the Euphrasias. For cultivation in the open ground Euphrasia plants and hosts were planted in bituminized paper pots, which were embedded in the ground. Very vigorous growth was obtained in this way, using Plantago lanceolata as a host-plant. The parasitic habit is probably not responsible for the taxonomic complexity of Euphrasia, but may be connected with its gregariousness in nature. PREVIOUS WORK The possibility of cultivating haustoria-forming Scrophulariaceae was demonstrated at the end of the last century. The behaviour in cultivation of the three semi-parasitic genera, Rhinanthus, Odontites and Euphrasia, was first studied by Ludwig Koch. His first paper dealt with Rhinanthus, and the second (1891) with Euphrasia officinalis L. He found that seedlings which germinated in pots grew much as wild ones do, if the pots contained old tufts of grass or grass seed sown at the same time as the Euphrasia seed (i.e. autumn). In Rhinanthus minor, however, Koch had found that simultaneous sowing of host and parasite gave retarded growth. He attributed this to the more rapid growth and earlier maturity of Rhinanthus placing a big demand on the host plants while they were still relatively young. Koch also sowed grass seed in pots with Euphrasia seedlings which were 3 cm tall (possibly an error for 3 mm tall). The effect of this was similar to that of simultaneous sowing with Rhinanthus, but worse from the point of view of the Euphrasia plants, which died off after 3 to 4 weeks. Euphrasias with no host-plant grew about as weakly as those with host seed sown after germination; if the seedlings were thinly and evenly distributed, they appeared starved and died off from mid-June onwards; in thick, uneven distributions some plants grew faster than the rest and were found to be attached to their neighbours by haustoria. These plants died if surrounded by few others, but if surrounded by enough others they occasionally flowered. They were prevented from fruiting by being dug up for investigation of the roots. Koch concluded that, for Euphrasia and Rhinanthus, parasitism on an autotrophic host was essential and not merely facultative. wane §=—AING av ithe st Watsonia 6 (1), 1964. ustital lus 2 P, F. YEO Wettstein (1896) at first came to the same conclusion. He reported that seedlings of E. rostkoviana germinating in troughs in the open, in which grass seed had been sown the previous year, produced quite vigorous plants which flowered and had numerous haustoria. With no hosts the plants were dwarf, produced up to ten pairs of leaves, formed no flowers and had no haustoria. Later, however, among many seedlings grown together in pots or pricked out singly in pots after germination, he brought a few plants to the stage of flowering or fruiting (Wettstein 1897). Some of them were without haustoria, while others had haustoria on the roots of other Euphrasias. Wettstein now concluded that, though parasitism on an autotrophic host is necessary for vigorous development, individuals can flower and fruit, either alone or when behaving autoparasitically. Heinricher’s four papers on ‘The Green Semiparasites’, describe a lot of work on Euphrasia. It is first necessary, however, to describe an experiment with Odontites odontites (O. verna) (Heinricher 1898a). Seed was sown in a pot on 2 March 1895, and the largest plants were 2-5 to 3 cm tall by 27 April. On 29 April, it was found that seedlings of a closely-sown group had haustorial initials at the points where their roots were in contact. Two other cultures, thickly-sown, gave the following results: in one there were 36 plants on 2 August, with from 3 to 12 pairs of leaves; two plants had flower buds and branch initials. On 20 September the strongest plant was 20 cm tall, had produced 20 flowers, and its lowest capsules were ripe. In the other thickly-sown culture the plants, on 3 August, were from 2 cm tall, with four leaf-pairs, to 7-8 cm tail, with 2-3 flowers. The plants did not grow much more, and produced up to two or three fruits each. The wide individual variation of the first culture is attributed to a long period of germination, enabling some plants to get ahead by parasitizing others; the uniformity of the second 1s attributed to germination taking place over a short period. These results are similar to those obtained by Koch with uneven and even sowings respectively, using Euphrasia seed. In addition to Odontites, Heinricher (1898a) got Euphrasia stricta Lehm. to flower with no host in a thickly-sown pot or garden soil. Germination began on 17 March 1896, and 70 plants were present on 15 May. Some plants that were surrounded by many others grew more strongly, and the strongest one started to flower on 8 August when 32 mm tall. Eight more were in flower on 13 September. The smallest flowering plants had stems 7-10 mm tall, their leaves were scarcely half the area of’those of the largest, and their flowers were a half to a third the size. In two other pots, with 30 and 36 seedlings respectively, no plants came into flower. Later, experiments were carried out with EF. rostkoviana (Heinricher 18985). In one pot the seedlings were abundant, and in places dense, on 17 March; flower buds appeared in May but none opened; more were formed in July, and the plants produced up to 20 leaf- pairs, but only one eventually flowered, although 80-90 had germinated. In each of five further pots, one seedling came up; these died one by one, from May to August, all without flowering. The growth of a third species, EL. minima Lam., in the absence of a host plant, was also investigated (Heinricher 18985). In one pot 27 seedlings were present in June 1898, but they were mostly chlorotic and were attacked by aphides, and none flowered. A similar number of seedlings appeared in this pot in 1899, and six dwarf plants flowered in May. E. minima was also sown outdoors on a plot of gravelly, stony soil. Most plants flowered in June; three examples which were growing 2 cm apart from one another but isolated from other plants were illustrated; they were about 2 cm tall and began flowering at the third node. Another similar small isolated plant was found and its roots examined. Lateral roots had penetrated and formed many branchlets in rotting wood-fragments in the ground; root hairs were plentiful in places, but no haustoria were found. These dwarf plants of E. minima, and also those of E. stricta mentioned earlier, resem- bled wild plants of E. minima seen in late August at Hiihnerspiel, Brenner, growing on stony ground at a distance from other plants. Heinricher concluded that E. rostkoviana, with one out of 90 plants flowering with no host, was the most advanced parasite, E. stricta, with nine out of 70 flowering with no Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION 3 host, was less advanced, and E. minima was the least advanced. In fact, Heinricher did not grow approximately similar numbers of plants per pot of each species in the same year and with the same soil. Nor was there any duplication of pots, and his conclusions, though possibly correct, have little foundation. In any case, such a comparison would be very difficult to make because the three species probably have different optimal soil conditions. The occurrence of wild populations of E. minima with no apparent access to hosts, which Heinricher reports, does, however, indicate that this species is sometimes autotrophic in nature. Heinricher (18985) states that, in having plenty of root hairs and in being relatively independent, E. minima resembles Odontites odontites (O. verna). Heinricher also raised Euphrasia to the flowering and fruiting stages with host-plants, finding that the parasites were more vigorous with hosts than without, and more of them flowered. Pot cultures on indoor window-ledges suffered from unsatisfactory conditions, the best results being obtained with E. stricta. About 100 seeds of E. stricta were sown in a large pot of garden soil with seed of Trisetum flavescens on 27 February 1896 (Heinricher 1898a). Stellaria, Capsella and Polygonum came up as weeds, and most Euphrasia plants were closer to weeds than to Trisetum. Some plants evidently became established; the number of flower initials formed ranged from 1 to 16. The weakest plants were weaker than the strongest in the cultures with no host. Better results were obtained outdoors by sowing seed in autumn on plots planted with tufts of Carex, Luzula and Gramineae (Heinricher 1898a). At a suitable time Euphra- sias and hosts were dug up and haustoria were found on the host roots. Seedlings of E. salisburgensis were abundant on all plots on 14 April 1897 and flowering reached its peak at the end of June. Haustoria were found on: Carex alba Scop. Luzula spadicea (All.) Lam. & DC. Carex firma Host Trisetum distichophyllum (Vill.) Beauv. Carex sempervirens Vill. Sesleria caerulea Scop. Carex brachystachys Schrank & Moll. Poa alpina was also used, but it was not investigated for haustoria. It was found that the hosts tended to shade out the Euphrasias, while vigorous development occurred when the Euphrasias were not shaded but were within reach of the host roots. Under the conditions of this experiment, the host species therefore varied in suitability as hosts according to their habit, caespitose species with short leaves being best. The following results were obtained with weeds acting as hosts: extraordinarily vigorous growth on Capsella bursa-pastoris ; vigorous growth on Trifolium pratense; better growth on young seedlings than on larger plants of Dipsacus fullonum; giant plant on Senecio vulgaris, similar to result with Capsella; one plant vigorous on Sonchus oleraceus, lower parts shaded by host rosette and. etiolated, upper parts normal; very vigorous development on Veronica peregrina and V. persica. In some cases a Euphrasia plant was found to have attachments to more than one host- plant simultaneously—in one case, for example, to Carex alba, Poa annua and Trifolium pratense. Sowings of E. rostkoviana and E. minima were made under similar conditions, and gave very similar results. Heinricher (1910) concluded that cultivation of Euphrasia was easy, provided the seed was sown in summer or autumn, the density of growth of the host-plant restricted, and good illumination ensured; a wide range of species could be used as hosts. Euphrasia rostkoviana was also grown by Neidhardt (1947). The numbers of plants per pot were not given, but numbers of seeds sown were, and germination was up to 40%. Neidhardt tried one, two, five or 100 seeds per pot with no host, one, 100 and over 100 with Watsonia 6 (1), 1964. 4 Pak. YEO Poa annua, and over 100 with P. nemoralis and Agrostis tenuis. With no host, no Euphrasias flowered; when there were 100 in a pot some grew more strongly if in a group of several close together; with one per pot, plants attained a height of 1 to 3 cm and were chlorotic; with two and five per pot they were similar but reached 5 to 6 cm. With host-plants present and many seedlings in a pot, the Euphrasias became more vigorous after 4 weeks. If the host was regularly cut back, plants reached 25 cm and began to flower at the beginning of August. If it was not cut back, the Euphrasias developed slowly, and died in June without flowering. With one seedling per pot and Poa annua the Euphrasias were vigorous, much- branched and up to 15 cm tall, and they began to flower at the beginning of August. Neidhardt also got E. rostkoviana to grow in grass in the garden of the Hamburg- Fuhlsbuttel school, in a part of Germany where the species is not native. Germination took place in April, plants were similar to those of pot cultures and they flowered at the beginning of August. Success was not achieved in all cases, and Neidhardt states that the ground should not be too solid and should be as moist as possible, and that the hosts should be as low- growing as possible and not too dense. Recently Wilkins (1963) has given comparisons of dry weights of Euphrasia grown with four hosts and no host. All the Euphrasia samples reacted in much the same way to the hosts, and all hosts gave superior growth to no host, the difference being marked for all except one of the Euphrasias. Successful cultivation was achieved at least once with seven other hosts. Many of the potted plants were planted out in the garden when the roots had bound the soil together sufficiently. Establishment on the host was estimated by measuring the span of the largest leaf-pair; with Trifolium repens the time required for establishment was about 7-10 weeks. METHODS OF CULTIVATION The simplest method of growing Euphrasia, in my experience, is to dig up turf in which the plants are growing and plant it in pots. In 1952, I collected plants of many species by this method from early May until mid- September. In the great majority of samples collected in this way some of the Euphrasias continued to grow and flower. In another instance a turf was dug up in November, and Euphrasias germinated and flowered in it the following year. In such cultures the host- plants have to be cut back to prevent them smothering the Euphrasias; this is chiefly necessary in the early stages, as the host-plants later become pot-bound. A second method is to sow Euphrasia seed in pots and, at the same time or later, sow seed of host-plants in the same pots. Alternatively, germinated seedlings or rooted portions of host-plants can be added to the pots. This method was used for E. pseudokerneri in 1951-52. Germination in Euphrasia, however, is always rather uncertain (Yeo 1961). Wastage of pot capacity and host-plant supplies can be avoided and the number of plants per pot controlled, if the Euphrasia seedlings are transplanted after germination. This method can be used either for pot culture or for cultivation in open ground, and it is possible to use seedlings germinated in cultivation or collected wild in March, April or May. Transplanting hosts and Euphrasias, at approximately the same time, into the pots in which they are to be grown has been my standard method of culture. For cultivation in open ground the Euphra- sias and their hosts are planted in bituminized paper pots which are then planted in the soil and gradually rot. For almost all cultures the potting soil used was John Innes compost no. 1, or slight variants of it. PRESENCE AND ABSENCE OF HOST-PLANTS The growth of several species of Euphrasia with and without hosts has been compared. In 1952, E. pseudokerneri (serial no. E42), from S. Lincolnshire, was grown from seed with several different hosts, and with no host. Diagrams of some of the plants can be seen in Fig. 1, p. 8. Flowering occurred without a host, but the diagrams show that these plants were Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION 5 far weaker than some, at least, of those with hosts. Similar results were obtained with seedlings of E. pseudokerneri (E66) collected from Box Hill, Surrey, in April 1952. In 1955, E. nemorosa (E507), collected from Cambridgeshire, was grown from seed in pots with four separate host-plants, and without a host. The weights of the plants after drying (Table 5, p. 12) show distinctly superior growth with three of the four hosts, and that the largest individual with a host was over five times the weight of the largest with no host. Two plants from this experiment are illustrated in Fig. 3, p. 21. The one grown with no host shows a very gradual increase in the size of the leaves on the main stem, and has few branches, which are confined to the upper cauline nodes and were still short at the time TABLE 1. Length of Stem and Branches of Euphrasia. (Eight Euphrasias per pan, 6-8 survived in each.) Mean (cm) Range Mean (cm) Range 1957 4 Medicago lupulina, 8 in. pan No host, 8 in. pan rostkoviana, E669, v.c.49 = 10-4 5-13-5 10-3 1-13-5 1962 2 Sagina procumbens, No host, 5 in. pan 2 Trifolium occidentale, 8in. pan | | occidentalis, E922, v.c.75 33-0 10- 71 5:9 4-8 nemorosa, E919, v.c.48 .. 49-7 8— 87 15-0 9-30 confusa, E907, v.c.45 40-5 22— 84 17207) 7-17 confusa, E914, v.c.45 12-3 11-160 9-5 4-14 brevipila, E917, v.c.45 .. 52:4 21-149 8:3 2-14 rostkoviana, E916, v.c.45 42-1 24— 60 6°3 4-9 anglica, E913, v.c.45 59-9 27— 94 14:5 6-23 1963 As 1962 No host, 8 in. pan confusa, E929, v.c.26 ... ge 52°4 16-103 8-9 2-21 brevipila E940, v.c.98.. a 50-0* 32— 77 16:0 10-26 * Four plants; the other four plants in the pot had a similar size range, but were pressed later. the plant was pressed. In the plant with a host, large leaves were formed lower down the stem, but near the apex the leaves were smaller, probably in correlation with the great development of the branches. Further results are presented in Table 1.* The beneficial effects of providing a host are evident in all samples except the first. Even here, some benefit can be inferred, if it is assumed that the Euphrasia, had it not been parasitically established, would have been adversely affected by the competition of the Medicago. In fact some of the plants of E669 with Medicago showed visible indications of rather superior growth, compared with that with no host, which are not shown by the measurements. In view of the postulated competitive effect of the host the comparison in 1962 between plants with a host in 8 in. pans and plants with no host in 5 in. pans is probably not unfair. In fact, growth of host-less plants in 5 in. pans was much the same as in 8 in. pans. An apparent indication of establishment on the host is that the apex of the plant becomes a fresher green, probably because of a more rapid expansion of the youngest leaves, which soon become conspicuously larger than their predecessors. After this, new leaves are formed at a greater rate, and usually the branches begin to grow. There is a great contrast between such plants and those with an apparently unsuitable host, which may remain almost static and are likely to die. * Euphrasias often produce some flexuous capillary branches at the base; in all cases where branches have been measured these have been ignored. Watsonia 6 (1), 1964. 6 P. F. YEO In general, pot-grown Euphrasias which appear to be established on the host resemble well-developed wild Euphrasias; those that do not appear to be established, or are not provided with a host-plant, are inferior and resemble the starved-looking Euphrasias that can also be found in nature. The time from germination to establishment is usually 4 to 10 weeks, but may be as much as 17. PROXIMITY TO HOST-PLANTS In 1955, E. nemorosa (E507), from Cambridgeshire, was grown in pots at varying distances from the host-plant. A correlation appeared between the time required for apparent establishment on the host and the distance between host and parasite. Euphrasias planted 2-5 cm from the host appeared to be established about a week earlier than those 4-5 cm from it, and about a fortnight earlier than those 6-5 cm distant. Another effect of distance was that, of the Euphrasias planted 6-5 cm from the host, a considerably lower percentage became established than was the case with those planted closer to the host. LEGUMINOUS HOSTS Comparisons of various host-species are given later, but the effect of some Leguminosae was distinctive and will now be described separately. The effect showed itself in the great vigour of the Euphrasia plants and in their dark green, often glossy, leaves. It was first seen in Euphrasias grown in their own turf in 1952. With EF. anglica (E67 and E76), from Surrey, the legume concerned was Medicago lupulina, with E. occidentalis (E73), from S. Hampshire, it was Trifolium repens, and with E. pseudokerneri (E75), from Cambridge- shire, it was Anthyllis vulneraria. Each of these Euphrasias was also growing in other turves (either in the same pot or in separate ones) containing few or no Leguminosae, and here they were much less vigorous and were paler green. MM. /upulina, with a slight admixture of Anthyllis vulneraria, had the same good effect on many plants of E. nemorosa (E74), growing in a turf collected near Cambridge. Table 10, p. 17, shows that four Leguminosae were used as hosts more than once in pot culture, and that three of them usually gave medium or good growth. With these three the Euphrasia plants almost always had particularly dark green foliage. In their luxuriant growth and dark green leaves, Euphrasias growing with suitable Leguminous species showed symptoms of a good supply of nitrogen, which they must have been getting either through their haustoria or through the soil after its release by the legumes. The former hypothesis seems more probable, in view of the differences existing between individual Euphrasias in separate turves in the same pot. The legume which Table 10 shows was a poor host was Trifolium dubium. Its unsuit- ability was not caused by an absence of bacterial root-nodules, however, for these were seen to be present when the seedlings were planted out with the Euphrasias. MUTUAL PARASITISM Koch (1891) and Wettstein (1897) reported that Euphrasia plants can parasitize one another, and their work has been summarized in another paper (Yeo 1961), where their observations on haustorial connections between individual Euphrasia plants were confirmed. The cultural effect of this—the markedly superior development of a few plants in a pot containing many seedlings—has also been observed by the present writer. A further cultural effect, probably due to the same cause, is that occasionally some, but not all, of the Euphra- sias in a pot provided with a host fail to show signs of establishment, or fail to grow much after establishment. For example, a plant of E. anglica (E180), from S. Somerset, and a plant of a form of E. micrantha (E231), from E. Donegal, each growing in a small pot with Sagina apetala, remained very small, while the other three Euphrasias in each pot grew quite vigorously. Further, a plant of E. nemorosa (E897), from W. Sussex, apparently established on 8 June 1961, had only 16 cm of stem and branches when pressed on 9 September, while by this date its nearest neighbour (growing about 3 cm away and established only a week earlier) had produced 290 cm. Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION 7 TIME OF FLOWERING The various species of Euphrasia are characterized by their time of flowering. Examples of late-flowering species are EF. pseudokerneri and, in at least part of its range, E. nemorosa. These usually start flowering in late July or early August, at about the 10th to 16th node. Most species, on the other hand, begin to flower about mid-June at the sixth to tenth node; these are mountain meadow types (E. brevipila and E. rostkoviana) and pasture species (E. confusa, E. anglica, etc.). The high mountain species, E. frigida, E. rivularis and E. cambrica, usually begin flowering in June or July at the second to sixth node. It will be seen from Fig. 1, however, that certain plants of E. pseudokerneri (E42) cultivated in 1952, some with no host and some with Pelargonium = hortorum, began flower- ing at the third to fifth node. The first flower buds were seen on 28 March, and the first flowers opened shortly before 21 April. The plants completed a very brief life cycle and died in a natural way and not from disease. This precocious flowering was evidently due to the very high temperatures prevailing in March in the greenhouse, which often exceeded 25° C. This suggests that temperature is one of the main factors controlling flowering in Euphrasia, and that given sufficiently high temperatures E. pseudokerneri can be made to flower at about as low a node as the high mountain species. On 21 March, these plants that flowered and died early had three or four leaves with visible teeth, whereas the plants numbered three and seven in Fig. 1, had only nought totwo. It seems, therefore, that, being at a sufficiently early stage of development in March, these two plants were unable to respond to high temperatures by premature flowering and death. In fact, although one of them was without a host plant, both were ultimately able to develop more fully than those that appeared more vigorous in March. Other plants of E. pseudokerneri (E42), grown as part of the same experiment but with Plantago lanceolata for host, flowered rather less precociously—early in May, at the fifth to seventh node—and they survived until July or August. Two plants of the same population of E. pseudokerneri, cultivated in 1952 in their own turf, flowered from the sixth to eighth node, then produced branches at higher nodes, and then flowered again. This was evidently the result of temperature fluctuations, and it shows that the conversion of a shoot from the vegetative to the flowering state is reversible. Precocious flowering of very young plants after hot weather has occurred repeatedly in my greenhouse cultures. It is generally difficult to get naturally early-flowering forms to become established before coming into flower and, once flowering has begun to drain the reserves of tiny plants, there is little chance of subsequent establishment inducing vigorous growth. Thus in 1961 20 plants of a form of E. brevipila (E870), from Sweden, were grown with various host-plants. Sixteen Euphrasias had flower buds visible by mid-May. All but three of these produced one or two flowers and were then pressed, as they did not appear to be capable of producing any more; they were all about | cm tall. One of the three exceptions was a little larger and produced five flowers. The other two appeared to become established on the host, and were moderately vigorous. Precocious flowering of Euphrasia sometimes takes place in nature, presumably as the result of an early hot spell. E. pseudokerneri and E. nemorosa were found flowering in Bed- fordshire on 15 June 1952 and in Cambridgeshire on 12 June 1957, high temperatures having been recorded in Cambridge in mid-May 1952 and at the end of May in both 1952 and 1957. All these early-flowering Euphrasias appeared to be established on host-plants, and had begun flowering at the fifth to seventh node. EFFECTS OF SOIL No investigation was made of the influence of soil on growth, but the following observations suggest that small differences in soil constitution may be important for seedling Euphrasias. Seedlings of Euphrasia anglica (E240) were collected in Leicestershire on 28 February 1953. They were intended to be a reserve for cultivation experiments but were not needed Watsonia 6 (1), 1964. 8 Pb YEO for this purpose. Three or 4 days after collection, about 120 seedlings were removed from their turf and planted in two pots (numbered 1 and 2) in approximately equal numbers. On 22 May pot | had 42 seedlings and pot 2 had 39. In pot 1 the two smallest plants were pale green throughout, while the rest were medium green, but paler at the top; all the plants in pot 2, however, were distinctly chlorotic, the upper leaves being quite yellowish, and all the leaves appeared slightly fleshy. On 12 June, two plants in pot | had their first flowers open, three more were ready to bloom shortly, and others had flower buds. In pot 2 all the plants were chlorotic and no flower buds were visible. h a c e g i 13 July 4 July 1 August 1 August Fig. 1. Diagrams of Euphrasia pseudokerneri plants. (a)-(c) without a host; (d)-(g) with Pelargonium xX hortorum; (h) with Carex caryophyllea; (i) with Bromus erectus. Horizontal bars = nodes; cross-pieces on them show leaf-tooth number. Sloping bars = branches. Forks = flowers. Un-dated diagrams represent plants that had completed their development. The poorer condition of the plants in pot 2, was probably due to differences of soil. The soil in pot 2 appeared lighter in colour and had less peat at the surface than that in pot 1; also, unlike the soil in pot 1, it had mosses and green algae developing on it by 22 May. Euphrasia plants sometimes show chlorosis before establishment and, using John Innes compost, this is especially prevalent in E. micrantha, E. anglica and E. rostkoviana. WATERING During the spring and early summer, Euphrasias in pots in an unheated greenhouse are liable to die if there is a spell of cold, heavily overcast or damp weather. The entire plant wilts, and dies in a day or two; this can occur in unestablished or in vigorously growing Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASTA IN CULTIVATION 9 plants that have been established for a week or so. The best watering policy seems to be to keep the Euphrasias rather dry during April and the first part of May, as they endure this well, and then to water more freely to encourage growth. After watering, the plants and the soil surface should be encouraged to dry off quickly; and no water at all should be given in cold dull weather. TABLE 2. Length measurements of Euphrasia pseudokerneri grown in 1952. Pot Length in cm of stem No. Host plus branches of each plant E42 1 None .. as Ay ae ae se COR As Sis LoS) 2 Festuca ovina . es Ne = en ZO 16s. 95.65,35.0% 5 3 Koeleria cristata vA ; Soa | 4 Festuca ovina and Keeleria Grisrata! WO PAG 32, 119.5 5 Pelargonium X hortorum .. bis ODD Deal 6 Plantago lanceolata .. an re eI e Oe 4s 7 Prunella vulgaris oa teehee 8 Plantago lanceolata and Prunella magne. 32, 28, 0:5 9 Thymus drucei oe aie OFS, .0 2521035 10 Festuca ovina and. Thymus deuce ant Re E70 1 Original turf .. uy: ay ai an OO! 60. 374) 25s lane 2 Carex caryophyllea Se ae .. 64 3 Bromus erectus i a ur = ew LS 4 Sieglingia decumbens. . i igs .. 63, 30 5 Hieracium pilosella .. isn uy AC (Sw GROWTH OF E. PSEUDOKERNERI, E. ANGLICA AND E. NEMOROSA WITH VARIOUS HOST-PLANTS The growth of FE. pseudokerneri with various hosts was compared in 1952 by cultivation in pots. A group of pots 54 in. in diameter was prepared in November 1951; the soil consisted of loam covered by a layer of sand. (It was hoped that it would be possible to trace haustorial connections easily in the sand layer, but there was little branching of roots in it as it turned out.) Seeds or plants of the host species were sown or planted in each pot, together with seed of the Euphrasia (E42), from S. Lincolnshire. Germination of the Euphrasia occurred in January, February and March. On 29 April 1952, seedlings of E. pseudokerneri (E70) were collected from the same locality and some of them were planted, at the same time as their host-plants, in pots 34 in. in diameter. The rest of these seedlings were left in their turf, and this was also potted up; other species in the turf were Carex caryophyllea, Thymus drucei and three or four species of grass. The development of the plants was regularly recorded, the dates of appearance of the leaves, branch initials and flower buds being noted, in addition to the leaf tooth number; notes were made later to show which branches developed and which flower buds opened. Some of these observations are reproduced in Fig. 1. The E42 and E70 plants were eventually pressed and the total length of stem and branches measured (Table 2). The plants were pressed on various dates but, with a few exceptions, all appeared to have completed or nearly completed their growth. It is clear that some plants were good hosts; there is one inconsistent result in that the Euphrasia plant with Festuca and Thymus did not develop, although those in the two other pots containing Festuca developed quite well. The dwarf plant with Koeleria, when dug up on 5 June, was found to have its roots surrounded by those of the grass, and a dwarf plant with Thymus, dug up dead on 10 May, Watsonia 6 (1), 1964. 10 Leslee 1 aO) was found to have a few detached haustoria, suggesting that parasitic union had been effected but had not proved beneficial, presumably because the host failed to provide what the Euphrasia required. The occurrence of some feeble plants in pots with apparently good hosts has been discussed in general on p. 6. It is known, however, that in some of the E42 plants growth was prevented by precocious flowering. TABLE 3. Number of Euphrasia plants of different degrees of vigour with various host species E. pseudokerneri (£226), Surrey Grade Host 1 2 3 4 5 6 Poor Moderate Good Anthyllis vulneraria a fe Medicago lupulina Ee Ae 1 Hieracium pilosella Ae i 1 1 Thymus drucei .. *) Ne 1 1 1 Dactylis glomerata ae Se 1 Agrostis gigantea Aa es 1 1 Festuca ovina Luzula campestris Achillea millefolium Sieglingia decumbens ay, Carex flacca Wi: ya aa 1 Brachypodium pinnatum Koeleria cristata. . ae me 1 Bromus erectus .. Plantago lanceolata Carex demissa Carex caryophyllea Carex pilulifera . . et me LQ WO —_— NRF Ee wWN Ww N a aoe ms pet et pet ees WN NN = —& E. anglica (E157), Surrey Grade Host Poor Moderate Good Luzula campestris ae a 1 Agrostis gigantea ai in 2 1 Anthyllis vulneraria sh he 1 Brachypodium pinnatum ale 1 Medicago lupulina Carex caryophyllea Carex pilulifera .. Achillea millefolium é Helianthemum chamaecistus .. 1 Hieracium pilosella Bromus erectus .. ; = 4 Carex flacca La ae ag 2 Carex demissa .. Re au DD, Thymus drucei .. Ass ie 1 = = W = eS = me LO SS PS FE OS Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION 11 The Euphrasia plants grown from seed took 72-95 days from germination to establish- ment. Those grown from wild-collected seedlings became established 40-60 days after potting-up with the host-plants. In 1953, E. pseudokerneri (E226), this time collected in Surrey, was again grown with various hosts and compared with E. anglica (E157), also from Surrey. Four Euphrasia seedlings were planted in each pot, the pots being 33-4 in. in diameter. For each species of Euphrasia there were two pots with each of 21 host species. There were two host-plants per pot, except in the case of Achillea millefolium where there was only one per pot. The hosts were of known origin and were in the form of young seedlings or portions of newly- divided plants. The pots were plunged in an ash-bed in a sheltered but sunny position at Leicester. The Euphrasias and hosts were planted out from 28 to 31 March. By 20 April, many Euphrasias were dead and many damaged, apparently because of frost. Replacements were made, but not all the pots could be brought up to their full numbers. Most of the Euphrasias were pressed on 20 September, but a few were pressed on | and 3 September. They were graded into six sizes; a ‘type-plant’ of E. pseudokerneri was chosen for each size, and the other plants of this species were then classified into the six grades by comparing them with the ‘type-plants’. Silhouettes of two of the latter are shown in Fig. 2, p. 12. The plants of EF. anglica were comparatively sparsely branched and these were graded according to their length of stem and branches, as indicated in Table 3, where the performance of both Euphrasia species is shown. The number of host species is less than 21 because, with some hosts, all the Euphrasias died. Owing to the small numbers of Euphrasias with each host species, it is inadvisable to conclude, from the results of this experiment, that any host is a poor one (though some are clearly good). Thus, the previous year EF. pseudokerneri had grown well on Plantago lanceolata and on Carex caryophyllea, and on some occasions E. anglica has grown well on Medicago lupulina. It is shown in the experiment with E. nemorosa described next that some hosts, though they produce good growth in established plants, are also liable to give a high proportion of failures to become established. This factor could be responsible for the variations of behaviour with the hosts just mentioned. It cannot be claimed with confidence that £. pseudokerneri and E. anglica differ in their host reactions, but the table shows that, on the whole, /. pseudokerneri grew better with dicotyledonous plants than with monocotyledons, whereas this superiority of dicotyledons was not evident in the case of E. anglica. In 1955, many plants of E. nemorosa (E507), from Cambridgeshire, were grown in 54-in. diameter pots with four host species, and with no host (=N#). The four host species were: Hieracium pilosella (=H), Plantago lanceolata (=P), Bromus erectus (=Bs), Medicago lupulina (=M). There were 1-6 Euphrasias per pot, and either one or three host-plants in the H, P, Bs and M pots. The Euphrasias and hosts were potted up as young seedlings. The Euphrasias were placed near the edges of the pots, and hosts were placed in the middle, except when only one Euphrasia and one host were present. During growth, the host-plants were cut back where they tended to shade Euphrasias. The potting up was done on 8 and 9 April. Some replacements were already necessary on 10 April, and further deaths were observed throughout the period of cultivation. Observations were made on all plants about once a week from 8 May to 4 July. On 13 August a note was made of the survival or death of the plants, and the following day all were pressed. The appearance of two of the Euphrasias can be seen in Fig. 3, p. 21. The overall mean weight of the plants with each host species is given in the fourth line of Table 4. H and P gave high mean weights, M a fairly high mean, and Bs a slightly smaller mean than Watsonia 6 (1), 1964. 12 Paka YEO Fig. 2. Plants of Euphrasia pseudokerneri, x 4. Left, grown with Hieracium pilosella (size-grade 5). Right, grown with Anthyllis vulneraria (size grade 6). TABLE 4. Mean weights (in g) of Euphrasias after pressing. (The number of plants in each class is given in brackets.) Number of plants per pot Euphrasias Hosts H P Bs M NH ior 2 1 -33 (8) -48 (4) "07 1G) -23 (4) -15 (42) 3 or 4 1 or3 -41 (10) -49 (8) -08 (4) -28 (14) -14 (10) 5 or 6 ivors “32)1©) Tle) “13 © 25) (i) ‘08 (6) 1 to 6 iors -36 (23) 39a) -10 (13) -26 (29) -13 (28) TABLE 5. Variation of Euphrasia according to host species H jz Bs M NH Percentage established 3% Ms V3 64% 64% 16% — Average wt. of plants (g) -36 139 -10 -26 “Ad Standard deviation pe fy -374 363 tS -151 -068 Wt. of largest individual (g) 53 1-6 1-3 33} 27 2 Mortality, 8 May4 July .. ee 20% SYA 41VY% ay 16% Mortality, 8 May—13 August 3b 237, 40% SEA IPA 167, Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION 13 NH. The result with no host has been discussed on pp. 5 and 6. Many of the Euphrasias with Bromus erectus appeared to become established in the period 8 May to 21 June. Presumably they received little benefit from this host, and suffered competition from it. Establishments occurred mainly during the first 4 weeks of the 8-week observation period. It was found that there was little variation in the time of establishment on the different hosts, but the total percentage of Euphrasia plants that became established varied according to host (Table 5). Table 5 also shows that the mortality of the Euphrasias was low with the hosts M and H, and high with P and Bs; M and H also showed better establishment. This bears out an impression obtained during my work on Euphrasia that plants which are not established on the host are more liable to die than those which are. The very low mortality with no host (none between 4 July and 13 August) supports the theory that, if Euphrasias are in the same pot as a fully autotrophic plant from which they receive no benefits by parasitism, they will be at a severe disadvantage compared with Euphrasias in a pot by themselves. For experimental work it is desirable to have a high percentage of plants becoming established, a high average weight, little variation in weight, and low mortality. Euphrasias with Bs and NH show little variation in weight, but both are unsatisfactory in their low average weight, as is Bs in its high mortality. P has the highest average weight, but its establishment percentage is relatively low and its mortality high. H and M present the best combinations of characters. Since plants weighing 0-2 g and upwards are quite well developed, M’s lower average weight than H’s is not a serious disadvantage, and it is more than counterbalanced by a smaller variation in weight and by lower mortality. The experiment, therefore, shows that M is the most useful host plant of those tested. FURTHER HOST COMPARISONS INVOLVING MEDICAGO LUPULINA I have generally used Medicago lupulina as the host on which to raise my Euphrasias. However, this has been less successful with species from the West, the North and the mountain areas of Britain than with the south-eastern species. In the case of E. scottica this could have been due to its being an inhabitant of wet places, and one year some plants of a sample of E. scottica (E782), from W. Ross, were grown in a pot embedded in wet peat, but this special treatment produced little or no improvement in their growth compared with the rest of the sample. In 1960, when several northern and western species became available, samples were grown in a particularly cool and shady situation with M. lJupulina, but again with poor results. At the same time some of these species, together with the southern E. anglica and the widespread EF. confusa, were grown with various hosts in an unshaded greenhouse. For these, Poa pratensis and (as in the shady position) M. /upulina were very poor hosts, but quite good results were obtained with Sagina procumbens for E. anglica (E849), E. brevipila (E806), E. confusa (E803, E834), from Yorkshire, and E. rostkoviana (E855), while Trifolium occidentale D. E. Coombe gave good results with E. brevipila (E860) and E. rostkoviana (E857), both from Kirkcudbrightshire. In addition, some samples were planted with /. lupulina in the greenhouse, and gave slightly worse results than in the shade; these were E. borealis var. zetlandica Pugsl. (E851), E. curta (E853) and E. rotundifolia (E845). These results suggested that temperature and humidity were not the main causes of difficulty in growing these plants, and that unsuitability of host-plant was probably the main cause. Plants grown in 1961, largely from the same samples as were used in 1960, gave the results shown in Tables 6 and 7; these show, in most cases, a marked superiority of Sagina, Trifolium, or a mixture of the two, over Medicago. Two stocks of M. lupulina are referred to in Table 6, the new one being superior to the old for E. confusa x E. occidentalis and for E. nemorosa. The old stock of this host had been maintained and reproduced annually from seed since 1952, and the new one was obtained in 1961 by collecting seedlings from the herbaceous bed for Leguminosae in the University Botanic Garden, Cambridge. The new Watsonia 6 (1), 1964. 14 Pb MEO TABLE 6. Total length (in cm) of stem and branches of Euphrasias grown with various host-plants, 1961. (No. before colon = no. of surviving Euphrasias; no. after colon = mean length; no. in brackets = length of largest plant. Every pot contained ten Euphrasias and four host-plants.) E. confusa X E. brevipila occidentalis E.anglica E.nemorosa E. micrantha E806 E870 E900 E849 E897 Es9] Vice-county 65 (Sweden) 99 6 TS if Medicago lupulina (old stock) S330 7/ 5)a) Pots) 5:6 4:13-3 4:49 -3 PIAS) 3 pots (12) (4) (8) (21) (79) (12) M. lupulina (new stock) 2:14 Slits 29 2:14 4:92°8 4:7-3 3 pots (15) (4) (75) (16) (138) (22) Sagina procumbens 4:53-3 5:4°8 3:54-3 3:24-7 5:120 0:— 3 pots (85) (14) (88) (43) (249) Trifolium occidentale 4:19-8 S283 5:13-4 SEZ/ES SED 28 33655 3 pots (34) (10) (32) (41) (134) (74) stock was more vigorous, had larger leaflets and was less affected by heavy attacks of red spider mite. It can be seen from Table 6 that E. micrantha failed to survive with Sagina as host, but in 1961 E837 grew very well on Sagina procumbens seedlings which germinated in its original seed pot. E. micrantha is particularly liable to show signs of ill-health before establishment, and this no doubt contributed to the poor results shown in these tables. It thus seems to be the case that there are many species of Euphrasia (particularly the northern ones) for which M. /upulina is not a favourable host-plant. For some Euphrasia species, however, even if Medicago is usually unsuitable, there are certain populations which will grow well on it. For example, quite good plants of E. brevipila and E. rostkoviana have sometimes been grown on Medicago, and very vigorous plants of E. brevipila from Poland have also been grown on it. TABLE 7. Total length (in cm) of stem and branches of Euphrasias. (Explanation as for Table 6.) E. borealis var. E. E. E. 1B zetlandica __ rotundifolia scottica curta rostkoviana E851 E845 E864 F853 E8s55 Vice-county 112 108 73 73 73 1960 Shady frame with Medicago lupulina 0:— Sly. 0:— 5:6 4:7°3 (2) (10) (8) 1961 Greenhouse with Sagina pro- cumbens mixed with Trifolium occidentale 7:7:6 3:5) 2:31 6:14 1:90 (24) (8) (38) (20) (90) Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASITA IN CULTIVATION 15 CULTIVATION OF THREE EUPHRASIA SPECIES WITH FOUR HOST SPECIES In this experiment Euphrasia borealis (E329), from NW Yorkshire, was grown indoors and outdoors, and E. nemorosa (E417), from Staffordshire, and E. occidentalis (E351), from W. Cornwall, were grown indoors only. Medicago lupulina (=M), Plantago lanceolata (=P) and Trifolium dubium (=T) were used as host-plants for the Euphrasias indoors. Each Euphrasia species and each host species was represented by one pan, and each pan contained four seedlings of the host- plant near the centre, and eight seedlings of Euphrasia towards the edge; in addition, there was a second pan of E. occidentalis with Medicago, provided with six host seedlings. A total of six Euphrasias died and were replaced in the first 3 weeks. E. borealis was grown outdoors with the same three hosts and with Briza media (= Ba), by the method described on p. 4. Each pot was planted with one Euphrasia and one host seedling on 11 May 1954. Replacements of dead Euphrasias were made from time to time until 30 May. On 2 June the pots were embedded in the soil of the experimental ground; there was a row of ten pots with each host, and the rows were placed 2 ft apart. On 11 July observations were made on the state of the plants outdoors, and on 3 September, all but four very small ones were pressed. Stem and branch lengths of the indoor and outdoor Euphrasias, after pressing, are given in Table 8, which shows the average size of the plants and, for the indoor cultures, the aggregate growth in each pot. (E. borealis and E. occidentalis are naturally more sparingly branched than E. nemorosa.) The growing period of E. borealis indoors with M was 21 days shorter than with P, but five or six of the eight plants with 7 had more or less finished growing when the plants were pressed on 23 July. E. occidentalis grew very much better with M in pot 2 than in pot 1. Pot 2, however, had been set up much earlier, and the Euphrasias had become established on the host 8 weeks earlier than those in pot 1. The death of all the indoor plants of E. borealis with T is probably accounted for by the un- satisfactoriness of this host. TABLE 8. Total length (in cm) of stem and branches of Euphrasias. (No. before colon = no. of surviving Euphrasias; no. after colon = mean length.) Total length for each pot Ba M JP Ee M iF Er occidentalis (pot 1) — ie sal 4:7°5 D375 74 30 7 occidentalis (pot 2) — 5:28 — — 140 — — nemorosa — 7:65 5:178* Tal: 453 890 84 borealis (indoors) — 8:15 5) P85 0:— 119 139 —_ borealis (outdoors) 0:— 4:10 5:42 322 — — — * Length estimated from the measurements of selected branches. This indoor cultivation can be summed up by saying that M and P were good hosts for E. nemorosa and E. borealis, M was fairly good for E. occidentalis, and T was very poor for all the Euphrasias. It will be noted that more Euphrasias of all species survived with M than with P. The difference between M and P in the results for vigour and survival of E. nemorosa is thus the same as that obtained with a different population of FE. nemorosa in NO55; Though P was a better host for E. borealis outdoors than in, M showed the reverse behaviour. Possibly, in open soil, the roots of M branch at too deep a level to benefit the Euphrasias. Four out of five outdoor plants of E. borealis with P were very vigorous with large Watsonia 6 (1), 1064. 16 Pa uYEO or very large leaves. A sixth plant was also vigorous, but, owing to some accident, only a fragment of it was alive at the time of pressing, so that it could not be measured. The observations that were made on 11 July on the plants of E. borealis in this trial are given in Table 9; here, as in Table 8, P appears as the best host for E. borealis outdoors, M is again second best, and 7 and Ba are both very poor. TABLE 9. Observations on E. borealis outdoors, 11 July, 1954. Host Trifolium Plantago Briza Medicago dubium lanceolata media lupulina Number alive oe aie ae: ae 8 10 5) 9 Number apparently host-established 0 8 or 9 0 3 possibly Number flowering oe ute uid 3 8 3 5 GARDEN CULTIVATION In 1953, at Leicester, a considerable number of plants of E. nemorosa and E. stricta was grown with Plantago lanceolata raised from seed (Yeo 1962). The bituminized paper pots (see p. 4) were plunged in the experimental plot on 31 March to 2 April. Between 22 and 26 April many dead plantains and Euphrasias were replaced. For this purpose the pots were lifted out of the ground. Of the 448 plants present after replacements had been made, 337 were alive on 19 July. Pronounced signs of establishment on the host in a few plants, and slight symptoms in many others, were seen on 12 May, 7 to 8 weeks after potting up. By 19 July only two plants out of 337 were not established. Nearly all the plants that were healthy became extraordinarily large-leaved and vigorous. Generally they were much more luxuriant than the most vigorous plants to be found in nature. Most of them were similar in size to a plant of E. nemorosa grown in the same way in 1956 which weighed 21 g after drying, and their upper cauline leaves were commonly 16-19 mm long and 16-18 mm wide. They may be contrasted with another plant of E. nemorosa, grown in a greenhouse, which was about as large as the normal maximum for wild plants of this species of Euphrasia, and which weighed only 1-6 g (with a stem-plus-branch length of 214 cm). The host-plants were also very vigorous and were several times cut back severely to stop them from covering the Euphrasias and to check their growth somewhat. The last time the severe cutting-back was done, some Euphrasias were getting bushy, and the plantains overshadowed by them did not recover their vigour. Subsequently, these Euphrasias appeared to suffer somewhat. It appears that, when the plantain is liable to be over- shadowed, some of its leaves that project out beyond the shadow of the Euphrasia should be left untrimmed. Many of the Euphrasias that died did so in July, after establishment on the host and after a period of wilting. It was thought that the wilting was caused by the fraying of the base of the stem, which resulted from the twisting of the stem as the head of the plant blew about in the wind. The movement of the plants in the wind doubtless brought on the wilting symptoms, but the stems may previously have been weakened by attacks of damping- off fungus at the seedling stage, as weaknesses at the bases of the stems were found in E. nemorosa cultivated in the greenhouse in 1955. The weak zone in these plants was dis- coloured and thinner than normal, and it sometimes broke when a plant was pulled up for pressing. In the absence of wind, the weakness had no effect on the vigour of the plants, some of the largest having the weakest stems. Damping-off had been particularly troublesome among the young seedlings of this population. Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION TABLE 10. Host-plants and Euphrasia species grown with them. 17 (Country of origin of Euphrasia indicated if not Great Britain or Ireland.) Clay-pot cultivation Ranunculus repens E. anglica Coronopus squamatus E. stricta Lehm.. . E. brevipila Helianthemum chamaecistus E. anglica Sagina apetala E. nemorosa E. confusa E. rostkoviana E. anglica Sagina procumbens E. micrantha E. occidentalis E. nemorosa E. stricta Lehm.. . E. confusa E. brevipila E. rostkoviana aS E. anglica ee Pelargonium x hortorum Bailey E. pseudokerneri Medicago lupulina E. micrantha . Scottica ape . Scottica X E. confusa . occidentalis - hemorosa S mm Fae . stricta Lehm.. . pseudokerneri confusa . borealis . brevipila . rostkoviana . Montana . hirtella re mm wash . curta var. rupestris Pugsl. .. No. of samples in which the indicated degree of vigour Poor 1 (Germany) 3 6 (2 Sweden) 1 1 was attained Medium 1 1 (Channel Is.) a (Channel Is.) 1 1 (Sweden) NY - = = 6 (1 France) 1 (Poland) DM 1 6 (1 Poland) 2 (1 Italy) 1 (France) Good 1 (Poland) 1 3 (1 North America) 1S) (2 France) 12 (1 France) 5 (2 France, 2 Germany, 1 Channel Is.) 6 6 4 (1 Sweden) 2 (1 Poland) Watsonia 6 (1), 1964. 2 18 Bo bay YEO TABLE 10—cont. No. of samples in which the indicated degree of vigour was attained Clay-pot cultivation Poor Medium Good Medicago lupulina—cont. E. anglica a Re te 1 6 5 E. anglica X E. nemorosa sys a 1 E. vigursii Hes a8 ue ae 1 E. salisburgensis. . ae ie ws 1 1 (France) (Austria) E. salisburgensis var. hibernica Pugsl. 1 Trifolium repens ae ae we E. stricta Lehm. a mu ie 1 (Poland) Trifolium occidentale Coombe E. micrantha... bins ay! i 1 E. nemorosa A ase ae ee 1 E. confusa a be ae Ss 1 E. borealis ann et ae a 1 E. brevipila ce xi Bh ye 1 1 (Sweden) E. rostkoviana .. Bas ees af 1 E. anglica oe an ee ah 1 Trifolium dubium E. occidentalis .. aye ie, a 1 E. nemorosa Me se ie oe 1 E. borealis a ws << ae 1 Anthyllis vulneraria E. foulaensis a ine ie 1 E. curta. var. rupestris ‘Spell, Ws 1 E. occidentalis .. ze ate cs 1 E. pseudokerneri aN a is 1 E. anglica a Be 5S E. confusa a ae a ms 1 Epilobium parviflorum E. rostkoviana .. at ea on 1 Calluna vulgaris E. micrantha... ; te 1 1 Calluna vulgaris and Erica cinerea afaedl | E. micrantha .. he Ab a 1 Thymus drucei E. occidentalis .. ae ite at 1 E. pseudokerneri a ne ae 1 1 E. anglica £3 ays te Be 1 E. salisburgensis. . Me ae ih 1 (Austria) —_ Prunella vulgaris a at se E. pseudokerneri A We ys 1 Plantago lanceolata .. By oe E. occidentalis .. se se ae 1 1 E. nemorosa ; ay a aw E. pseudokerneri ie a os 1 E. borealis E. anglica Plantago coronopus E. occidentalis See HN = Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION TABLE 10—cont. No. of samples in which the indicated degree of vigour was attained Clay-pot cultivation Poor Medium Good Achillea millefolium E. pseudokerneri a oe Bis E. anglica Ae HH ss ee 1 Leontodon autumnalis E. foulaensis E. occidentalis E. anglica E. vigursii Hieracium pilosella E. nemorosa - “o oe bie 1 E. pseudokerneri ee ae na 2 Luzula campestris E.micrantha .. Bap ate oy. 1 E. occidentalis .. AS Me Be 1 | E. pseudokerneri E.rostkoviana .. he 7 al 2 | Dy 1 E. anglica a? r ae vA i E. rivularis es A Re 3 1 Carex demissa E. pseudokerneri ie He ie 1 E. anglica or Se ae Se 1 Carex flacca E. pseudokerneri ig. ae ie 1 E. anglica sa ae. ihe Bs 1 Carex pilulifera E. micrantha... ae be aie 1 Carex caryophyllea E. pseudokerneri 1 E. confusa is vy sf By, 6 1 E. anglica ae Le ane ie 1 Sieglingia decumbens E. occidentalis .. 3 a he 1 E.nemorosa .. ae oe ba: 1 E. pseudokerneri ~ ip ” 1 1 Festuca ovina E. pseudokerneri ae ae: we 1 1 Poa pratensis E. confusa i we ne en 2 E. rostkoviana .. ah ae a 1 Poa trivialis E. micrantha... ys a xy 1 E. pseudokerneri 7 oe 1 Dactylis glomerata E. pseudokerneri 6 A S 1 Bromus erectus E. occidentalis E. nemorosa ve bee Le E. pseudokerneri iA oe a 1 E. anglica Brachypodium pinnatum ae ci E. pseudokerneri. . ae abe we 1 E. anglica ae ee te val 1 m= CON = —_ a ee Watsonia 6 (1), 1964. 20 PE YVEO TABLE 10—cont. No. of samples in which the indicated degree of vigour was attained & Clay-pot cultivation Poor Medium Good Koeleria cristata E. micrantha ; E. confusa X E. micrantha E. curta var. rupestris Pugsl. E. nemorosa E. pseudokerneri E. confusa Agrostis tenuis | E. occidentalis .. = sh i 1 E. nemorosa as Bie a sue 1 Agrostis gigantea E. pseudokerneri a oe ue 1 E. anglica che pes ae Lok 1 ReNe BRN Cultivation in open ground Poor Good Very Good Medicago lupulina E. borealis ae mh ae a 1 Trifolium dubium E. borealis AS Br a 53 1 Plantago lanceolata E. occidentalis .. i ce 4s, 1 (France) E. nemorosa Ah fe ie a | 9 E. pseudokerneri we ay. Be: 1 | E. confusa bi “ie ap He 74, E. stricta Lehm. ce oe ae 1 E. borealis a Bt ue pe 1 Luzula campestris E. anglica zu a a EN 2: Briza media E. borealis a 5 a ae 1 Koeleria cristata E. micrantha form a bk ae 1 E. vigursii ae tee an a 1 Other Euphrasias were also tried with other hosts in 1953. Two Euphrasias were planted in each bituminized pot, and the hosts were pieces of divided plants with some roots present; one piece per pot was used. The most successful populations were two of E. anglica (E168 and E219), from Leicester- shire, each represented by ten pots with Luzula campestris. On 31 May (about 8 to 9 weeks after potting up) many of these Euphrasias showed signs of establishment on the host, and on 21 June nearly all were established. The plants grew vigorously, having fairly large leaves and numerous branches. They did not attain anything like the size of the largest plants of E. nemorosa grown outdoors with Plantago lanceolata, however, probably because of their lower node of flowering. On Koeleria cristata, a form of E. micrantha (E231), from E. Donegal, made little progress. On 26 June, two plants out of 16 showed signs of being established on the host, Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION 21 though five others (apparently not established) were flowering. On 31 August, three out of the 12 survivors may have been established, but none was vigorous. In E. vigursii (E197), from S. Devon, the first signs of establishment on the same host were seen (in one plant) on 18 May, but it was not until 30 July that nearly all were established. They were, therefore, about 5 weeks behind E. anglica in becoming established. They eventually became about as vigorous as the plants of their parent wild population, and they varied quite considerably in vigour. The same method of cultivation was used at Cambridge in 1954. In one experiment, populations of E. nemorosa and E. confusa were grown; mortality was very heavy, but the survivors grew very vigorously on Plantago lanceolata grown from seed. In 1956, the method was again used to see whether better survival could be obtained at Cambridge. E. nemorosa (E608), from Cambridgeshire, was grown in four rows of ten pots each. The 40 pots were each planted with one Euphrasia and one Plantago lanceolata seedling on 4 April. The rows of pots were then planted in the garden, each on a different date. However, the mortality of the Euphrasias awaiting planting in the garden was nearly as severe as that of those already planted out, so that late planting-out gave little advantage. In fact, there were again very few survivors, but most of them grew extremely vigorously. LIST OF HOST-PLANTS The host-plants employed in all my experiments are listed in Table 10, together with a rough indication of how well various Euphrasia samples grew on them. This indication is based on the most vigorous Euphrasia plant raised in each case. Fig. 3. Plants of Euphrasia nemorosa after pressing on 14 August 1955, x 4. Left, one of two plants grown in a pot with no host; weight 0:2 g. Right, one of three plants grown in a pot with one plant of Medicago lupulina; weight 0-7 g. Watsonia 6 (1), 1964. 22 PE RYEO Particularly poor hosts seem to be Leontodon autumnalis and Koeleria cristata; Grami- neae in general seem to be mediocre hosts, while some of the best hosts are to be found among the Leguminosae. Indications that Medicago lupulina may be a better host for some species of Euphrasia than for others have already been mentioned; Luzula campestris may be a similar case, for it appears to be better on the whole for the species of the diploid Series Hirtellae (EZ. anglica, E. rostkoviana and E. rivularis) than for the tetraploid species of two other Series. TABLE 11. Growth of Euphrasias without host-plants. Date when Pot No. of Height No. not No. Species pressed diam. Euphrasias inmm flowering fruiting inches micrantha, E185A, S. Somerset Aa 2/9/53 54 fairly 20-40 few rather numerous few occidentalis, E192, E. Cornwall Wy 2/9/53 54 fairly 6-15 few many numerous DADOtS) aoe Be i rn ie 2/9/53 53 very 2-15 many many numerous occidentalis, E351, W. Cornwall 9/7/54 3 5 45-80 0 5 2 pots, ia; a es i 7 9/7/54 3 18 15-55 4 10 nemorosa, E196, S. Devon .. WY 2/9/53 5% extremely 10-50 many moderate numerous no. nemorosa, E421, Surrey He ie 9/7/54 3 4 10-40 3 0 2 pots. 5% ie as ie 3 9/7/54 3 9 8-20 5 3 nemorosa, E417, Staffordshire .., 27/8/54 3 15 17-35 11 0 confusa, E172, S. Somerset .. a 2/9/53 53 fairly 7-35 few moderate numerous no. confusa, E208, Derbyshire .. oe 2/9/53 54 numerous 6-36 few many brevipila, E325, W. Sutherland oa 9/7/54 3 6 7-30 1 5 brevipila var. reayensis Pugsl., E308, W. Sutherland .. Eat ..| 24/7/54 3 6 37-70 1 1 rostkoviana, E239, S. Kerry .. ..| 26/8/53 54 15 12-65 2 113) anglica, E180, S. Somerset BIEN 2/9/53 54 fairly 10-40 several many numerous BEHAVIOUR OF EUPHRASIA WHEN GROWN WITHOUT A HOST A silhouette of a Euphrasia plant fruiting in the absence of a host-plant can be seen in Fig. 3. Table 11 shows how the surplus seedlings of Euphrasia grew in their seed- pots without hosts in 1953 and 1954. The seedlings were too crowded to be fairly compared with plants grown in the presence of a host. However, the table shows that fruiting took place in nearly all samples. The capsules contained perfectly normal-looking seeds. The conditions of growth of three samples, E351 (E. occidentalis), E417 and E421 (both E. nemorosa), were closely similar, but the first grew much better than the other two, being evidently less affected by the lack of a host. E. nemorosa and E. confusa were the only species which produced branches and these are normally the most branched species. DISCUSSION Euphrasias in cultivation are decidedly more subject to disease than autotrophic plants from similar habitats. This applies both before and after establishment, but it varies greatly from one population to another and also from species to species. Careful treatment can minimize losses to some extent and, though some losses must be expected, there is no serious obstacle to the cultivation of the majority of Euphrasia species for experimental purposes, even when this requires the pricking-out of young seedlings. Watsonia 6 (1), 1964. THE GROWTH OF EUPHRASIA IN CULTIVATION ps A wide range of possible host species for Euphrasia in cultivation was reported by Heinricher and this has been amply confirmed by my cultures. This lack of host-specificity has also been demonstrated in the field by Wettstein (1897) and Crosby-Browne (1950). These facts make it appear unlikely that the existence of numerous critical species in Euphrasia is related to host-specificity. It may be noted that Euphrasia micrantha can grow on a number of hosts and not only on Calluna vulgaris and Erica cinerea, though it is rarely found in nature except in close proximity to one or both of these species. Certain species appear to be unsuitable as hosts for all species of Euphrasia tried, while others vary in suitability according to the species of Euphrasia. There are a number of possible factors involved in the unsuitability of host-plants, though it has not been possible to investigate these in most of the cultures described. Euphrasias may be prevented from parasitizing plants growing with them by the inability of the haustoria to penetrate their roots, owing to their physical structure or to physiological obstacles. There might also be physiological obstacles to the formation of the haustoria themselves, or to the absorption of nutriment once the haustoria had penetrated. These obstacles to absorption could be complete or only partial. Extraction of nourishment must also be affected by the quantities of food materials available in the host. Most of these factors could be influenced by the age, stage of growth and vigour of the potential host. If penetrability of roots varies inversely with their age, some parts of the root system of an individual plant will be more impervious than others; one can imagine the susceptible zone of a root system growing rapidly away from a Euphrasia plant, so that the parasitic attack becomes increasingly unsuccessful. On the other hand, a host-plant which continually produces new roots from the crown will continue to be susceptible to attack by a Euphrasia growing near it. The possibility also exists of intraspecific variation in susceptibility on the part of hosts and in parasitic capacity on the part of the Euphrasias; such variation might affect whole populations or only individuals. There seems to be little connection between the suitability of a host and its systematic position. It would seem that the good growth of Euphrasia on most Leguminosae is con- nected with the nitrogen-fixing activities of the root-nodules of these plants, notwithstanding the clearly established unsuitability of nodulated Trifolium dubium. -Euphrasias grown in cultivation normally retain their important taxonomic characters, and it is therefore clear that modifications caused by hosts are not the cause of the systematic complexity of the group. In fact, the hosts appear to cause only variation in vigour. Late establishment may have a marked effect on the habit of a Euphrasia, but this can occur with any host on which establishment is possible. The marked effect of a soil difference on the growth of Euphrasia without a host probably explains the generally greater success (compared with that of earlier workers) that I have had in growing Euphrasia in this way. It may also explain some of the variation in the results of Heinricher and Wettstein which led to a public dispute between them (Heinricher 1898c, Wettstein 1898). The parasitism of Euphrasia may contribute to its marked gregariousness in nature. One sometimes finds dense isolated colonies, apparently consisting of hundreds of plants packed into a few square feet, with very few outliers. Weight of numbers may help them to compete with other vegetation, and where the density falls below a critical value the Euphrasias may, under some conditions, be entirely eliminated. The decline of an artificial colony in my garden following a disturbance which buried deeply many of the seeds, suggested that seedling density was important; this may act through autoparasitism allowing a sufficient number of seedlings to survive a difficult period early in their lives in a sufficiently vigorous condition. ACKNOWLEDGMENTS The work described in this paper was commenced under a research scholarship of the University of Leicester and with the help of Professor T. G. Tutin. Professor Tutin, Watsonia 6 (1), 1964. 24 P. F. YEO Mr. E. K. Horwood and numerous other persons and institutions have kindly supplied seed and information on localities from which I was able to collect Euphrasia seed. Postscript.—In 1963 a plant of Stellaria media coming up in a pot of E. nemorosa seedlings led to very quick establishment and remarkably vigorous growth. This host, and other quick-growing weeds (see p. 3), may be useful in the cultivation of the more precocious- ly flowering Euphrasias. As, however, the host in this instance quickly declined in vigour, they should perhaps be combined with other, slower-growing species. REFERENCES CrOSBY-BROWNE, A. J. (1950). The root parasitism of Euphrasia Salisburgensis Funk. Watsonia 1, 354-355. HEINRICHER, E. (1898a). Die Griinen Halbschmarotzer, I. Jb. wiss. Bot. 31, 77-124. HEINRICHER, E. (18985). Die Griinen Halbschmarotzer, II. Jb. wiss. Bot. 32, 389-452. HEINRICHER, E. (1898c). Gegenbemerkungen zu Wettsteins Bemerkungen tiber meine Abhandlung “Die Griinen Halbschmarotzer I.”’ Jb. wiss. Bot. 32, 167-174. HEINRICHER, E. (1910) Die Aufzucht und Kultur der parasitischen Samenpflanzen. Jena. Kocu, L. (1891). Zur Entwicklungeschichte der Rhinantheen (II, Euphrasia officinalis L.), Jb. wiss. Bot. 22, 1-34. NEIDHARDT, G. (1947). Euphrasia rostkoviana Hayne, Der Augentrost. Pharmazie 3 Beih., 1 Erganzungsbd. WETTSTEIN, R. VON (1896). Monographie der Gattung Euphrasia. Leipzig. WETTSTEIN, R. von (1897). Zur Kenntniss der Ernahrungsverhaltnisse von Euphrasia Arten. Ost. bot. Z. 47, 319-324. WETTSTEIN, R. VON (1898). Bemerkungen zur Abhandlung E. Heinricher’s ““Die Griinen Halbschmarotzer I, ... Jb. wiss. Bot. 31, 197-206. WILkIns, D. A. (1963) Plasticity and establishment in Euphrasia. Ann. Bot. Lond. 27, 533-552. YEO, P. F. (1961). Germination, Seedlings, and the formation of Haustoria. Watsonia 5, 11-22. YEO, P. F. (1962). A Study of Variation in Euphrasia by means of outdoor cultivation. Watsonia 5, 224-235. Watsonia 6 (1), 1964. 2 PEATE?! Inflorescences of annual tetraploid Salicornia spp., * 7 approx. (a) Variation in stigma morphology. (6) In the two lower lateral florets the stigmas are protruding before the anthers; in the central florets immediately above, the anthers are dehiscing in contact with their stigmas. (c) Simultaneous development of anthers and stigmas in the lower central floret and in the dissected floret. (7) A further inflorescence showing simultaneous development of anthers and stigmas. { facing page 25 NOTES ON THE STIGMA MORPHOLOGY AND FLOWERING BEHAVIOUR IN BRITISH SALICORNIAE By I. K. FERGUSON Trinity College, Dublin ABSTRACT Moss transferred S. perennis Mill. to the genus Arthrocnemum. This paper sets out to show that two of the characters, the shape of the stigma and the sequence of maturation of the reproductive organs, that Moss used to separate the genera Salicornia and Arthrocnemum are very variable in plants from the British Isles. If this is accepted it would seem that the criteria for the definition of the genus Arthrocnemum will need reconsideration once more. INTRODUCTION In a posthumously published paper primarily concerned with South African Chenopo- diaceae Moss (1948) transferred S. perennis Mill. to the genus Arthrocnemum, and I under- stand that the combination A. perenne is being taken up in Flora Europaea. This paper draws attention once more to the confused and unsatisfactory position with regard to the delimitation of these two genera. Arthrocnemum was first proposed by Moquin (1840), who included S. perennis Mill. in his original delimitation of the genus. Moquin uses the absence of albumen in the seed and of wings on the perianth, together with the mode of insertion of the florets, as characters to separate the genus from Salicornia. It seems doubtful however, whether these characters possess the validity which he attributed to them as they have been adopted by few subsequent botanists. The absence of albumen in the seeds does not seem to be constant throughout the traditional limits of the genus (Moss 1954). De Fraine in an extensive investigation of the anatomy of the genus Salicornia makes no reference to any difference in the insertion of the florets between S. perennis Mill. and the annual species in the British Isles. In my own observations I have found that small differences do exist between S. perennis Mill. and the annual species of Salicornia in the British Isles in the shape of the florets and in the degree to which the central floret separates the laterals and also in the arrangement of the wings of the perianth. However, the former characters are very variable among the annual Sali- corniae that I have seen and it is doubtful whether they are sufficiently constant to serve as a basis for generic separation. In a later paper Moss (1954) redefined the genera according to the following criteria. Arthrocnemum: perennial; protogynous; stigma bifid; some of the branches not terminating in an inflorescence. Salicornia: annual; protandrous; stigma tufted; branches all terminating in an in- florescence. Salicornia perennis Mill. certainly differs from the annual species of the British Isles in the first and last of these four characters. This paper sets out to show, however, that in the shape of the stigma and in order of maturation of reproductive organs of the flower, no clear separation is possible. If this is accepted it would seem that the criteria for the definition of the genus Arthrocnemum will need reconsideration once more. I hope, at a later stage, to undertake a limited investigation of these criteria throughout the range of the two genera. STIGMA MORPHOLOGY The stigma in S. perennis Mill. may be bifid or trifid; the branches are usually about 1 mm in length and often persistent after fertilization (Fig. 1). In annual species of Sali- cornia the stigmas are variable even within the same plant, being ‘tufted’ (that is very shortly branched with branches often unequal in length), bifid, or trifid, and often persistent 25 Watsonia 6 (1), 1964. 26 I. K. FERGUSON (Fig. 1 and Plate 1a). The stigmas of annual material collected in the field are usually shorter than those of S. perennis, with branches usually about 0:5-0-7 mm in length. Tetraploid plants cultivated from seed from six populations resembling S. /utescens Ball & Tutin, however, produced stigmas with very long branches of more than 1 mm in length and in- distinguishable from those of S. perennis (Fig. 1). In diploid plants referable to S. europaea agg. and S. pusilla Woods the branches of the stigmas are usually shorter and more slender than those of S. perennis but are very variable, and in cultivation they may be indistinguish- able from those of S. perennis. Fig. 1. Stigmas of Salicornia, x 10 approx. (a) S. perennis. (b) from an annual tetraploid plant in cultivation. (c) and (d) from annual tetraploid plants in the field. These observations are in harmony with the illustrations in Smith’s English Botany, where S. perennis (t. 1691) and an annual species closely resembling S. /utescens Ball & Tutin (t. 415) are both depicted with trifid stigmas, though the branches of the former are longer. A third illustration (t. 2475) shows another annual species with a tufted stigma. Tutin (1952) describes the stigmas in S. perennis as bifid and those in the annual species as tufted. Later (Tutin 1962) he retains the description of the stigma in S. perennis as bifid but omits any description of those in the annual species. FLOWERING I have found from observations in the field that S. perennis Mill. is protogynous; this is in agreement with Moss (1954) and Dalby (1962). Among the annual species I have observed that in tetraploid plants resembling S. lutescens Ball & Tutin and S. dolichostachya Moss, both in the field and in cultivation, either the stigmas protrude just before the undehisced anthers are exserted (Plate 1b, seen in the two lateral florets of the lower segment) or the anthers and stigmas:emerge simul- taneously (Plate 1c and d). It thus appears that some annual Salicorniae are weakly protogy- nous or homogamous. This view to some extent supports Dalby (1962) who says ‘usually it seems that Salicornia is weakly protogynous and sometimes it may be markedly so (as for example S. perennis). Plate 1b further supports Dalby’s view ‘that it is very likely that many of the annual forms are self-pollinated in nature, as ripe dehiscing anthers may be seen in contact with presumably receptive stigmas, and their pollen spilling on to the stigmatic papillae (see Fig. 1)’. However, his figure does not show this very clearly. In some diploid plants I have confirmed the observations of Ball & Tutin (1959) that the flowers may be cleistogamous. In others the anthers are exserted after dehiscence; it is difficult, however, to observe the sequence of development in either of these conditions. Further observation is required. I have observed that plants in cultivation and in the field may sometimes have unisexual flowers, either of the organs failing to develop. I have also observed that the stigmas fre- quently elongate and persist after fertilization. Elongation must occur after fertilization has taken place as the stigmatic papillae have always been found to have collapsed when it commences. These phenomena may have contributed to misinterpretation in the past of the sequence of development of the anthers and stigmas. Watsonia 6 (1), 1964. STIGMA MORPHOLOGY IN BRITISH SALICORNIAE 27 ACKNOWLEDGMENTS I wish to express my thanks to Prof. D. A. Webb for assistance and advice. The work was done during the tenure of postgraduate awards from the Department of Education and the Board of Trinity College, Dublin. REFERENCES BALL, P. W. & TuTin, T. G. (1959). Notes on annual species of Salicornia in Britain. Watsonia 4, 193-205. DALpy, D. H. (1962). Chromosome number, morphology and breeding behaviour in the British Salicorniae. Watsonia 5, 150-162. DE FRAINE, E. (1912). The anatomy of the genus Salicornia. J. Linn. Soc. (Bot.) 41, 317-348. Mogutn, A. (1840). Chenopodearum monographica enumeratio, 111-116. Paris. Moss, C. E. (1948), in Barker, W. F., et al., Plantae novae africanae, ser. 28. J. S. Afr. Bot. 14, 29-40. Moss, C. E. (1952), in Adamson, R. S. The species of Arthrocnemum and Salicornia in southern Africa. J. S. Afr. Bot. 20, 1-22. SmiTH, J. E. (1807, 1815). English Botany, vols. 6, 24 & 35. London. TuTIN, T. G. (1952), in Clapham, A. R., Tutin, T. G. & Warburg, E. F. Flora of the British Isles. ed. 1. Cambridge. TuTin, T. G. (1962), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. ed. 2. Cambridge. Watsonia 6 (1), 1964. INFRASPECIFIC VARIATION IN LATH YRUS NISSOLIA L. By J. F. M. CANNON Department of Botany, British Museum (Natural History) ABSTRACT The varieties described for Lathyrus nissolia by Continental authors are reviewed and their findings are clarified with the aid of cultural experiments and statistical techniques. Two well-marked varieties are recognized and their distributions are plotted on a map. Only one variety appears to occur naturally in the British Isles. Some preliminary observations are made on the breeding biology of the species and problems demanding further field study are indicated. As Lathyrus nissolia is a very distinct species which is clearly separated from the rest of the genus by its grass-like phyllodes (hereafter referred to as leaves), its circumscription has been understood from early times and its taxonomy is free from troubles at the specific level. When, however, its infraspecific variation is examined a great confusion of taxa is found. This is due in part to the rather subtle nature of the characters involved and partly to the consideration by some of the early authors of too little material, both in the herbarium and in the field. My own interest in the subject was aroused by the fact that although the major Continental Floras all recognized varieties of this species, very little notice had been taken in British botany of their conclusions. Lousley writing in the Botanical Exchange Club Report for 1934 was almost alone in drawing attention to the Continental work and his note does not seem to have received the attention it deserved. The reason for this lack of interest by British botanists is now clear, since as a result of this study it will be seen that only one infraspecific taxon is represented in this country and the absence of interest was therefore probably occasioned by frustration rather than neglect. HISTORICAL A. Kerner was the first to draw attention to the variation within the species when he described Lathyrus gramineus in 1863. This he believed to be a related species which differed from L. nissolia by ‘petiolis angustioribus concavis et leguminibus germinibusque glaber- rimis’. Thus, by implication, Kerner suggested that L. nissolia (sensu stricto) has broader leaves and a hairy pod. As will be shown later, this was an error which was perpetuated by later authors till the present time. L. gramineus A. Kerner is a later homonym of L. gramineus Gray, which is itself a nomenclatural synonym of L. nissolia L. Uechtritz, writing a letter to the editor of the Oesterr. Bot. Zeitschr. in 1864, pointed out that Orobus (Lathyrus) nissolia was quite a variable plant, especially with regard to leaf width. He suggested that Kerner’s L. gramineus was a variant of this species and went on to point out that the glabrous-podded forms are of frequent occurrence throughout the range of the species, to an extent that some authors of Floras have described it as having glabrous pods. He then made a short reference to a form which he referred to as O. nissolia genuinus which had a short fine pubescence. In making these observations, Uechtritz showed that he had an appreciation of the variability of this species unequalled by any of the older authors. Subsequent authors have considered that this letter includes a valid publication of the name O. nissolia genuinus and have presumed the trinomial to be of varietal rank. It seems clear to me, however, that he only used genuinus meaning ‘in the strict sense’ and did not visualize it as a formal nomenclatural unit. Freyn (1878) in a paper on the flora of Southern Istria provided a varietal epithet for the forms represented by Kerner’s L. gramineus. He noted that his var. glabrescens had a narrow leaf and a glabrous or almost glabrous pod. In doing this he undermined the excellent observations about variation in leaf width made earlier 28 Watsonia 6 (1), 1964. VARIATION IN LATHYRUS NISSOLIA 29 by Uechtritz and paved the way for later authors to further confuse the situation. In 1885 Uechtritz also provided a varietal epithet (var. /iocarpus) for the glabrous-podded forms, citing L. gramineus A. Kerner as a synonym. In his Flora of Lower Austria (1893) G. Beck recognized two varieties: a pubescens G. Beck which he described as being hairy-podded and 8 gramineus G. Beck which has a glabrous or somewhat rough pod. He cited var. glabrescens Freyn as a synonym of the latter. In the addendum to the Flora, Beck provided an additional name (var. puberulus) for var. pubescens. Since he gave no reason for this action, we must regard this as the publication of a later superfluous nomenclatural synonym. Rouy, in Flore de France (1899) provided two further varietal epithets with the following descriptions: lanceolatus Pétioles lancéolés-linéaires (4-10 mm de large); plante plut6t robuste. linearis (L. gramineus Kern.) Pétioles étroitement linéaires (2-3 mm de large); plante plutot gréle. These descriptions are important as they draw attention for the first time to a possible slight difference in leaf shape. Unfortunately Rouy was influenced by Kerner’s early re- marks on the narrowness of the leaves of the glabrous-podded forms and introduced the size restrictions shown above. In his general description Rouy refers to the pods being “pubescents ou presque velus’. It is strange that he made no reference to the pods of the varieties. Ascherson & Graebner also writing in 1899 proposed a forma gramineus Aschers. & Graebn. and referred to its glabrous pods without mentioning the leaves. Hegi (1924) recognized two varieties, using the epithet genuinus Uechtr. (‘pod hairy, leaves moderately broad’) and g/abrescens Freyn (‘pod glabrous or somewhat rough, leaves smaller about 3 mm broad’). Finally Fiori in his Flora of Italy of 1925 proposed a var. typicus, with linear-lanceolate leaves 4-10 mm broad and with hairy pods, and a var. gramineus with linear leaves 2-3 mm broad and with glabrous pods. He cited var. g/abrescens Freyn and var. linearis Rouy as synonyms of the latter and derived its epithet from Gray and Kerner. As we have already seen, L. gramineus Gray is a direct nomenclatural synonym of L. nissolia L. My own studies in this field originated in an attempt to apply this maze of observations to the British material. Much time was spent looking for hairs on the pods of the broader- leaved forms, but it eventually became apparent that this kind of plant was not represented in the British flora (except by one specimen at the British Museum (Isle of Wight 1916) which, must, I think, be considered to be of adventive origin or a curatorial error). In considering the interpretation of the glabrous-podded variety, I was for a long time misled by the in- sistence of some Continental authors that narrow leaves went with glabrous pods. The situation seemed especially difficult as the British material included some conspicuous plants with exceptionally broad leaves of up to | cm in width. Early attempts to apply the var. g/abrescens concept to British plants suggested that there might be two glabrous- podded varieties, but eventually it became clear that a continuous range of variation existed both in this country and on the Continent. Some consideration was then given to finding additional reliable features to support the pod-pubescence character. At this stage Uechtritz’s excellent remarks of 1864 were unknown to me, but inspection of the material at the British Museum and Kew suggested that a slight but constant difference in leaf shape was associated with the pod character. The plants with hairy pods had narrow-lanceolate leaves, while those with glabrous pods had linear-lanceolate leaves. As already noticed this was suggested by Rouy but he confused the issue by also suggesting an absolute difference in leaf width. In order to express this difference statistically a ratio was devised whereby the leaf length was divided into six equal parts and the width at one sixth from the apex (W,) was divided by the width at one third from the apex (W,). This is referred to hereafter as the /eaf ratio. Fig. 1 shows a simple plastic scale which was made to facilitate the division of the leaves into six equal parts. By aligning the mid-rib of the leaf parallel to one of the horizontal lines of the scale, so that Watsonia 6 (1), 1964. 30 J. F. M. CANNON its tip and base touched outermost sloping lines, the one-sixth and one-third points could quickly be found with an accuracy adequate for the present purpose. As a pilot experiment fifty leaves each were measured from glabrous and hairy-podded specimens in the British Museum herbarium and average ratios worked out for them. The hairy-podded plants had a ratio of about 0-5, while the ratio of the glabrous-podded forms was about 0-7. At this point it became essential to clarify the nomenclatural position in Fig. 1. Diagram showing the use of a simple transparent plastic scale for finding the positions for the measurement of the leaf-widths used in the /eaf ratio determinations. respect of these two varieties. Continental authors have always referred to the hairy-podded variety as being typical L. nissolia but when the Linnaean type was examined (Herb. Linn. no. 905/2) it was found to fit very exactly into the glabrous-podded group; the correct name for these plants is therefore var. nissolia. LL. nissolia var. pubescens G. Beck (1893) is the correct name for the hairy-podded variety. A synonymy of varietal names is given below: LATHYRUS NISSOLIA L. VAR. NISSOLIA Lathyrus gramineus A. Kerner, Oesterr. Bot. Zeitschr. 13, 188 (1863) non Gray (1821). . nissolia var. glabrescens Freyn, Verh. K. K. Zool-bot. Ges. Wien 27, 325 (1878). . nissolia var. liocarpus Uechtritz, Jahres-Bericht. Schlesischen 62, 310 (1885). L. nissolia var. gramineus G. Beck, Fl. Nied-Oesterr., 882 (1892). L. nissolia var. linearis Rouy in Rouy & Fouc., Fl. France 5, 253 (1899). L. nissolia forma gramineus Aschers. & Graebn., Fl. Nordost. Flachl., 99 (1899). hh Fig. 2. Graph showing the relation of mean leaf ratio and mean pod vein number for eleven populations. The lines represent twice the standard error of the means. & = Linnean type specimen. Fig. 3. Graph showing the relation of mean leaf ratio and mean pod diameter for eleven populations. The lines represent twice the standard error of the means. & = Linnean type specimen. Fig. 4. Graph showing the relation between mean pod vein number and mean pod diameter for eleven populations. The lines represent twice the standard error of the means. & = Linnean type specimen. Watsonia 6 (1), 1964. VARIATION IN LATHYRUS NISSOLIA as 7 _ (=) var, pubescens oO o var. nissolia awk 7 Leaf ratio Fig. 2 Pod vein number =~ 0-6 0-8 375 € var. pubescens = 3°5 oD 59°25 E var. nissolia sS 3:0 — no} aah a & — 0:7 0:8 Leaf ratio 0:6 Fig. 3 var. pubescens var. nissolia Pod diameter (mm) no on 6 7 8 9 Pod vein number Fig. 4 Watsonia 6 (1), 1964. 10 31 32 J. F. M. CANNON L. nissolia subsp. amanus Rechinger. Ark. fér Bot. 5, 268 (1960). (See discussion later in this paper.) LATHYRUS NISSOLIA VAR. PUBESCENS G. Beck, op. cit., 882 (1892). Lathyrus nissolia var. puberulus G. Beck, op. cit., 1329 (1893). Illegitimate substitute name. L. nissolia var. lanceolatus Rouy in Rouy & Fouc., Fl. France 5, 253 (1899). L. nissolia var. typicus Fiori, Nuova Fl. Anal. Ital. 1, 909 (1925). EXPERIMENTAL TAXONOMY Having reached these conclusions by more or less traditional methods, it seemed clear that cultural methods with statistical investigation would provide the best means of further clarifying the separation of these varieties. Through the courtesy of the directors of several botanic gardens and by direct collection in the field, seeds of ten strains of L. nissolia were obtained for culture the following season. The seeds were sown directly in the ground in patches so that clumps would be formed by the growing plants. Previous experience had shown that germination is very slow and irregular unless the testa is filed before the seeds are sown. The following characters were observed in the living plants. The foliage of var. nissolia is rather yellowish-green compared to that of var. pubescens which is relatively glaucous bluish-green. I have not been able to detect this colour difference in herbarium material. The colour difference is most noticeable when the plants are massed in clumps, under these conditions the plants can be separated by this character at a distance of several yards. In my experience plants of var. nissolia are in full flower 4 months plus after planting, while plants of var. pubescens sown on the same day come into flower 3 months plus after planting. The generally quicker growth rate of var. pubescens is noticeable at all stages when the plants are grown side by side. Var. nissolia is very variable in height, from 10 cm dwarfs to 100 cm giants. Its stem is often nearly simple, sometimes with a second main branch from near the base. Var. pubescens on the other hand is rather uniform in stature, varying between about 25-35 cm. It frequently has a rather bushy growth habit with up to six or seven stems of equal magni- tude rising from the base. In addition to the leaf ratio character already described, two other characters were tested statistically on the cultured populations. These were (1) The number of major veins on a pod valve at a position half way along its length (Pod vein number) and (2) the diameter of the pod measured at the same position. For each culture 50 leaves and 50 pods were collected, only one of each being taken from each plant. Only lower leaves from fully developed plants were used and likewise only fully grown but un- ripened pods were selected. The results of the analysis are shown graphically in Figs. 2, 3 and 4. In each case the mean value for the 50 samples is plotted together with lines repre- senting twice the standard error of the mean which is used as an indication of the variability within each population. Some measurements were also made of the Linnean type specimen and figures representing this plant are included on the graphs. In each case its clear association with the glabrous-podded variety can be seen. LATHYRUS NISSOLIA SUBSP. AMANUS RECHINGER In 1952 Dr. K. Rechinger kindly drew my attention to his then unpublished subsp. amanus from the Amanus range in northern Syria. This is founded on Haradjian nos. 222, 266 and 275 in the Delessert Herbarium at Geneva. Rechinger’s description reads “Differt a planta typica pedunculo folium fulcrantum aequante vel eo longiore’. I later had the opportunity of examining the type material and was unable to find any additional features by which it could be separated from var. nissolia. While | have not seen another specimen with such relatively long peduncles, I have seen plants ranging widely between the var. amanus condition and the more normal state. For these reasons I prefer to regard the Amanus material as a highly localized population of the type variety. Judging from the Watsonia 6 (1), 1964. VARIATION IN LATHYRUS NISSOLIA 313) material that has reached the major European herbaria from Asia Minor, Lathyrus nissolia is a rare and scattered plant in this region. I would therefore suppose that this long- peduncled population has developed in the seclusion of the Amanus mountains and the character has become established in the population through the habitual self-pollination which I believe is normal in this species. Fig. 5. Map showing the distribution of the varieties of Lathyrus nissolia. @ = var. nissolia, < = var. pubescens. DISTRIBUTION The accompanying map (Fig. 5) shows the distribution of the two varieties throughout the known range of the species. The British range is adapted from the map in the Aflas of the British Flora while the remainder has been plotted from the records provided by specimens which I have examined myself. It will be noted that while the distribution is sympatric over the greater part of the range, the British Isles only has var. nissolia with the exception of one record of var. pubescens from the Isle of Wight. A specimen in the British Museum Herbarium forms the basis for this record. The label provides no additional information other than the date 1916. In view of the lack of any other records and absence of var. pubscens from the north-west part of France, I am disposed to regard the Isle of Wight record as either an adventive stray or else the result of some curatorial error involving a mixing of labels. Apart from a broadly Continental trend in the distribution of var. pubescens, compared with the exploitation of more oceanic conditions by var. nissolia, no obvious distributional patterns can be seen from the map. CHROMOSOME NUMBER The chromosome number of this species has been reported as being 27 = 14 on several occasions including Simonet (1932) and Senn (1938). The vast majority of Lathyrus species have been found to be diploids with the base number 7; only four species have been recorded as polyploids. One of the most interesting of these was reported by Marks in 1950. He found that a tetraploid form of Lathyrus pratensis was common in this country. A detailed cytological investigation or L. nissolia would very probably prove rewarding. Watsonia 6 (1), 1964. 3 34 J. F. M. CANNON GENERAL SUMMARY AND CONCLUSIONS One of the most important observations resulting from this study-is that individual plants can always be placed with comparatively little difficulty into one or other of the two varieties. Plants with markedly intermediate characters do not seem to occur and I have not detected any hybrids among the very large number of sheets that I have examined. Attempts to produce artificial hybrids under cultural conditions were not successful. My observations suggest that this species is habitually self-pollinated—despite the apparent adaptation of its flowers towards entomophily. Buds enclosed in bags to prevent the access TABLE 1 Character var. nissolia var. pubescens 1. Pods Glabrous or somewhat With short bristly hairs scabrous 2. Leaf shape Linear lanceolate (leaf ratio Narrowly lanceolate (leaf ratio about 0-5) about 0-7) 3. Mean pod diameter About 3:0 mm About 3:5 mm 4. Mean pod vein number About 6 About 10 5. Colour of living plant Relatively yellowish-green Relatively blue-green 6. Growth rate In full flower four months In full flower three months plus after plus after planting planting 7. Habit Very variable in height, stem Relatively uniform in height and frequently often simple with a rather bushy habit with several stems of equal size of insects always set seed perfectly normally, while emasculated flowers were never found to set any seed, although unprotected by bags. I have never observed the flowers being visited by any insects, in the wild or in cultivation. These observations are confirmed by Kirchner who says ‘. . . the flowers frequently do not open at all but nevertheless set healthy fruits, being cleistogamously fertilized’. Knuth does not record any insect visitors to these flowers in his Handbook of Flower Pollination. This habitual self-pollination presumably provides an explanation of why hybrids apparently do not occur and also why the varieties remain distinct despite the very large overlap in their ranges. The question remains as to what status in the taxonomic hierarchy should be accorded to these taxa. Since interbreeding does not seem to occur and since the individuals can always be assigned to a variety without undue difficulty, it would seem that they might well be regarded as distinct species or at least sub-species. But since the evidence from genetics and cytology remains to be clarified, I prefer for the present to retain the existing varietal status, for which validly published names are available, rather than to further burden the literature with additional names at a new status which may prove undesirable after further studies have been executed. J am content for the present to draw attention to the interesting infraspecific variation of this plant, with the hope that other workers will make observations on its biology in the field, especially in those areas where both varieties are known to occur. It would be interesting to know if the difference in flowering times encountered under cultural conditions holds good when the varieties are growing in the same area in the wild and if natural hybrids ever occur in regions where both varieties are present. The characters which separate the two varieties are summarized in Table I. ACKNOWLEDGMENTS I am greatly indebted to my wife who took a large share in the biological and statistical phases of this study. In addition I am grateful to Dr. K. B. Blackburn, Mr. E. B. Bangerter and Dr. A. Melderis for their help and interest in the earlier phases of the work and Miss S. Gould who assisted with the preparation of the map and graphs. In addition to the material Watsonia 6 (1), 1964. VARIATION IN LATHYRUS NISSOLIA 35 in the British Museum, I have been privileged to examine the collection at the Royal Botanic Gardens, Kew, and to receive loans from the following institutions through the courtesy of their respective directors: Florence, Geneva, Leiden, Lyon, Paris, Rome, Vienna and Zurich. REFERENCES KircHNER, E. O. O. (1888). Flora von Stuttgart: 515. KNuTH, P. (1908). Handbook of Flower Pollination. Trans. Ainsworth Davies 2, 337. LousLey, J. E. (1935), in Distributor’s Report for 1934. Rep. Bot. Exch. Club. 10, 964. Marks, G. E. (1950). Referred to in Rep. John Innes Hort. Inst., 1949, 7. PERRING, F. H. & WALTERS, S. M. (1962). Atlas of the British Flora. London. p. 116. SENN, H. A. (1938). Contribution for a Revision of the Genus Lathyrus. Amer. Journ. Bot. 25, 73. SIMONET, M. (1932). Numeérations Chromosomiques dans les Genres Baptisia, Thermopsis et Lathyrus. Comptes Rendus Acad. Sci. Paris 195, 739. Watsonia 6 (1), 1964. CHROMOSOME NUMBERS OF EPILOBIUM IN BRITAIN By PETER H. RAVEN and D. M. Moore Division of Systematic Biology, Stanford University, Stanford, Calif., U.S.A. and Botany Department, The University, Leicester ABSTRACT The 14 species of Epilobium occurring in Britain are shown to have 2n = 36. Chromosome numbers of E. lanceolatum and E. tetragonum subsp. lamyi are given for the first time and a further 11 species have not hitherto been counted from British material. The artificial F,; hybrid E. montanum x adenocaulon is of reduced fertility, but shows no meiotic irregularities. The réle of self-pollination in reducing interspecific hybridization is stressed. Chromosome numbers have been reported for the majority of European species of Epilobium and they were summarized by Love & Léve (1961, p. 256-7), but there seems to have been only one published record dealing with British plants. Rutland (1941, p. 210) gave the somatic chromosome number of Epilobium palustre as 2n = 36 without specifying the locality from which the material originated. A few of the counts reported in the present paper were mentioned by Clapham (1962). All counts reported here were made during studies of meiosis in pollen mother cells, the anthers being fixed in 1:3 acetic alcohol and squashed in acetocarmine. With the exception of EF. alsinifolium, all the chromosome numbers were determined by the senior author during the tenure of a United States National Science Foundation post-doctoral fellowship spent in London during 1960-61. The authors are grateful to Professor D. H. Valentine, Dr. C. R. Metcalfe and Dr. K. Jones for the facilities afforded Raven while he was making his cytological studies and to Mr. A. Eddy for collecting plants of E. alsinifolium. Vouchers for the chromosome counts are deposited in the herbarium, Depart- ment of Botany, British Museum (Natural History) and in the Dudley Herbarium, Stanford University. Details of the collections from which plants were counted are given in Table 1. TABLE 1. Chromosome vouchers of British species of Epilobium. E. adenocaulon Hausskn. MIDDLESEX, Chiswick, Raven 18060 (garden progeny). SURREY, weed by the herbarium, Royal Botanic Gardens, Kew, Raven 16086. E. alsinifolium Vill. WESTMORLAND, Moorhouse Nature Reserve, altitude c. 1,800 ft, Moore 464 (garden progeny): E. anagallidifolium Lam. WESTMORLAND, at head of Knock Ore Gill, by road to Radar Station, altitude c. 2,500 ft, Raven 16277 (garden progeny). angustifolium L. MIDDLESEX, Twickenham, Raven 16490. hirsutum L. DERBYSHIRE, along hedgerow 2 miles north of Derby, Raven 16194. lanceolatum Seb. & Mauri. SURREY, weed by the herbarium, Royal Botanic Gardens, Kew, Raven 16087. montanum L. OXFORDSHIRE, Bagley Wood, Raven, 15915. nerterioides A. Cunn. KENT, Bayham Abbey, Raven 16168 (progeny grown at Chelsea Physic Garden). obscurum Schreb. WESTMORLAND, Glencoyne Wood, Ullswater, Raven 16226. palustre L. WESTMORLAND, Glencoyne Wood, Ullswater, Raven 16225. . parviflorum Schreb. SURREY, weed by the herbarium, Royal Botanic Gardens, Kew, Raven 16488. . roseum Schreb. SURREY, weed by the herbarium, Royal Botanic Gardens, Kew, Raven 16193. . tetragonum L. subsp. lamyi (F. W. Schultz) H. Lév. SURREY, Ewell, Pennington in 1961. . tetragonum subsp. tetragonum. KENT, sand pit behind Covers Farm, Westerham, Raven 16485. Pree h hb eh 36 Watsonia 6 (1), 1964. CHROMOSOME NUMBERS OF EPILOBIUM 37 All collections of Epilobium cited in Table 1, formed 18 bivalents at meiotic metaphase I; no meiotic irregularities were observed. All numbers except that for E. palustre are being reported from Britain for the first time; E. /anceolatum and E. tetragonum subsp. lamyi do not appear to have been counted before. The list includes all British species of Epilobium, including the two introduced ones, EF. adenocaulon and E. nerterioides. The segregate genus Chamaenerion is not recognized here for reasons discussed in detail elsewhere (Raven 1962a, Ye aad oe Fig. 1. Chromosome pairing at first meiotic metaphase in the artificial F, hybrid between E. montanum (Turkey) and E. adenocaulon (Middlesex) showing 18 bivalents. Sp DISCUSSION These additional chromosome numbers re-emphasize that barriers to hybridization in Epilobium sect. Epilobium (Lysimachion) are not a result of aneuploid differences in chromo- some number. Judging from reports in Britain and on the Continent, hybrids are probably possible between all combinations of species except those involving E. angustifolium and E. nerterioides. Many if not most of these hybrids are at least partly fertile. Meiosis in interspecific hybrids in Epi/obium has been studied by Hair (1942) and by Lewis & Moore (1962). These authors studied natural hybrids and in no case were meiotic irregularities observed. In an effort to extend their observations, we artificially duplicated one of the hybrids frequently encountered in Britain, that between E. montanum and the introduced North American E. adenocaulon. Our strain of E. adenocaulon was from Chiswick (see Table 1), whereas our seeds of E. montanum were from Turkey (Prov. Bolu, Ala Dag, alt. 700 m, Khan, Prance & Ratcliffe 515). Both of the parents consistently formed 18 bivalents at meiotic metaphase I. The hybrid plants (Raven 18557) were more vigorous than the parents, with darker, larger flowers, irregularly lobed stigmas, and the short capsules caused by a low seed set (c. 10 per cent. and below). Judged by staining in aceto- carmine, their pollen was approximately 15 per cent. fertile (based on 200 pollen grains from each of four F, plants). Just as in the cases of natural hybrids analyzed previously, meiosis in this hybrid was quite regular (Fig. 1). The reduced fertility certainly may help to maintain species distinctions in such a case, but we are still in large measure faced with the problem, posed by Valentine (1951), as to how the species maintain themselves in the absence of strong genetical, geographical and ecological barriers to interbreeding. The relatively small percentage of hybrids found in mixed populations of two or more species of Epi/obium in Britain suggests that a part of the answer may be in the high degree of self-pollination characteristic of most species of the genus. Of the 14 British species of Epilobium, only two are outcrossed to any appreciable degree—E. hirsutum and the very distinct E. angustifolium, which is presumably incapable of forming hybrids with the other species. In all other species found in Britain, the anthers surround the stigma and shed pollen directly on it, and the high degree of autogamy resulting from this doubtless tends to decrease sharply the proportion of hybrids that might otherwise be expected. This explanation was adduced by Lewis & Moore (1962) for the rarity of hybridization between the North American species E. adenocaulon and E. brevi- stylum Barbey in the Rocky Mountains. These authors also postulated that the hybrids they were studying were genotypically not so well adapted as the parental species to the communi- ties in which the parents grew, and consequently they were less likely to become established Watsonia 6 (1), 1964. 38 PETER H. RAVEN and D. M. MOORE in the undisturbed habitats available. In Colorado, where Lewis & Moore studied hybridi- zation in the genus, the summers are long and dry so that the habitats occupied by Epilobium are of rather limited extent, normally consisting of wet meadows and streamsides where the plants are members of relatively stable communities. In Britain, on the other hand, the climate is generally much moister throughout the year and many species of Epilobium are decidedly weedy, frequently growing in ecologically open communities in which there is a rapid turn-over of individuals. Even under these conditions the proportion of hybrids still remains extremely low. This tends to suggest that habitual autogamy and, possibly, lowered cross-compatibility (Valentine 1951, p. 86) may be far more important in main- taining the species as discrete units, at least for the ‘weedy’ species, although lowered hybrid fitness probably operates when interspecific crosses do occur. The evidence presented here further emphasizes that the gross chromosomal re- arrangements that have been so important in the evolution of the tribe Onagreae have apparently had no part in the differentiation of species in other tribes of the family (Lewis & Moore 1962, Raven 1963 and unpubl.). REFERENCES CLAPHAM, A. R. (1962). Onagraceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles, Ed. 2, 469-483. Cambridge. Har, J. B. (1942). The chromosome complements of some New Zealand plants. I. Trans. Roy. Soc. N.Z. 71, 271-276. Lewis, H. & Moore, D. M. (1962). Natural hybridization between Epilobium adenocaulon and E. brevistylum. Bull. Torrey Bot. Club 89, 365-370. Love, A. & Love, D. (1961). Chromosome numbers of central and northwest European plant species. Op. Bot. Lund. 5, 1-581. RAVEN, P. H. (1962a). The genus Epilobium in Turkey. Notes Roy. Bot. Gard. Edinb. 24, 183-203. RAVEN, P. H. (19625). The genus Epilobium in the Himalayan region. Bull. Brit. Mus. (Nat. Hist.) Bot. 2, 325-382. RAVEN, P. H. (1963). Circaea in the British Isles. Watsonia 5, 262-272. RUTLAND, J. P. (1941). The Merton catalogue. A list of chromosome numbers of British plants. Supplement 1. New Phytol. 40, 210-214. VALENTINE, D. H. (1951). Geographical distribution and isolation in some British ecospecies, in Lousley, J. E. (ed.), The Study of the Distribution of British Plants, 82-90. London. Watsonia 6 (1), 1964. STUDIES ON BRITISH PANSIES I. CHROMOSOME NUMBERS AND POLLEN ASSEMBLAGES By A. PETTET Department of Botany, University of Southampton* ABSTRACT The chromosome counts and pollen assemblages of a range of British pansies are given. With the exception of what is believed to be an aberrant count for Viola nana Corb. (V. kitaibeliana in Dandy, 1958), the chromosome counts confirm those cited by past workers. The polymorphic pollen assemblages which can be characterized by the most abundant grain type and the associated subsidiary types show that the taxa can be divided into two main groups. In the case of the two variable and often confused species, V. tricolor and V. arvensis, differences in their pollen assemblages confirm their separation on the basis of chromosome counts and such is the agree- ment between the two criteria that the former should be a reliable criterion for the separation of the two taxa. 1. INTRODUCTION The Melanium subgenus of Viola, comprising those plants better known as pansies, forms a very variable group not readily divisible into coherent taxa. Past treatments of the complex of variation surrounding V. tricolor sensu lato have ranged from the acceptance of a few rather variable species to the erection of many, narrowly-defined microspecies. The cytogenetical studies on European pansies by Clausen (1921, 1922, 1924, 1926, 1927, 1931) and the cytological investigations of British plants by Fothergill (1944) have, however, done much to clarify and stabilize the species concept in the subgenus. The Jordanian microspecies recognized by Drabble (1909, 1926, 1927 a-c, 1928) were shown to be in- distinguishable on the basis of chromosome number and chromosome morphology, although Drabble’s major groupings of these—Luteae, Curtisiae, Tricolores, Arvenses and Nanae— could be characterized by such features. On this basis the Tricolores are not separable from the Curtisiae, or the Luteae from the Nanae. The latter pair are, nonetheless, readily divisible on morphological features. As Clausen (1927, 1931) pointed out whilst discussing the concept of species within the subgenus Melanium, it is quite impossible to give a definition covering all cases. Morpho- logical differences coupled with complete or partial sterility fail to delimit all taxa satis- factorily because there are some morphologically uniform taxa which contain several intersterile cytotypes. On the other hand, differences in chromosome number used as the only means of delimiting species also fail because of the limited association between morphological appearance and chromosome number. In the particular case of the tricolor— arvensis group of annuals, Clausen (1922) recommended the retention of two species, V. tricolor L. and V. arvensis Mutr., as the basic units, mainly on the grounds that two different chromosome numbers were associated with two reasonably well-defined morphological types which had differing edaphic ranges. It is sometimes felt, however, that this division into two species is not always one easy to maintain, for their limits are blurred by a rather complex variation which itself may be further complicated by the reputedly frequent occurrence of hybridization and introgression of the two species. During an investigation into the possibility of frequent interspecific hybridization between V. tricolor and V. arvensis a considerable number of observations were made on the cytology and pollen morphology of a representative range of British pansies. Since these observations emphasized the distinctness of the two taxa within the British Isles * Now at Department of Botany, University of Khartoum. 39 Watsonia 6 (1), 1964. 40 A. PETTET and are pertinent to the problem of hybridization to be discussed in detail elsewhere, they form the basis of the present paper. There is, perhaps, little need to stress the value of cytology in understanding the com- plex variation in the pansies, but the potential value of pollen morphology appears relatively unknown. | Wittrock (1897) was the first to describe the quantitative differences in the polymor- phic pollen assemblages of V. tricolor and V. arvensis and to use them as an additional character in discriminating between the two taxa. These differences were subsequently mentioned by Clausen (1922) although he made little use of the character in his final analysis of the taxa. Since then the taxonomic value of the character has been overlooked with the one recent exception when Mullenders & Mullenders (1957) used it as a criterion for deciding whether V. maritima Schweigg. (= V. curtisii Forst., V. sabulosa Bor.) should be considered as a species distinct from V. tricolor subsp. tricolor. Before detailing the observations on the cytology and pollen morphology of this group of plants—where there has been considerable difference of opinion over specific limits in the past—it would be wise to define the limits of taxa used in this account. With the exception of certain modifications in application the specific names used here largely follow the usage of Warburg (1952) as the observations were originally made within this frame- work. The main exception concerns the tricolor—arvensis complex where much of the work was concentrated and where the three groups of plants were distinguished as follows: V. tricolor: Flowers blue, purple or predominantly so; upper petals longer than sepals (usually more than x 14); stylar flap large and relatively conspicuous. V. arvensis: Flowers all cream or occasionally with upper petals tipped with blue-purple; upper petals shorter than, or sometimes as long as, the upper sepals; stylar flap so reduced as to be, or appear to be, absent. Intermediates between V. tricolor and V. arvensis (subsequently abbreviated to ‘AT-intermediates’): Flowers cream, or predominantly so; upper petals longer than upper sepals (usually x 14-14); stylar flap present but approximately intermediate in size between that of V. arvensis and V. tricolor. The AT-intermediates formed a rather arbitrary group which was meant to cover all those plants not readily ascribable to V. tricolor and V. arvensis, and which might be considered to be putative hybrids. Observations on these and experimentally-produced hybrids between V. tricolor and V. arvensis will be the subject of a following paper. It will be seen that V. tricolor, as here defined, is more or less equivalent to V. tricolor subsp. tricolor of Warburg (1952); in consequence, V. tricolor subsp. curtisii of the latter is here referred to as V. curtisii. 2. CHROMOSOME COUNTS Previous chromosome counts for taxa represented in the British Isles made on British and other European pansies by Clausen (1921, 1926, 1927, 1931) and Fothergill (1944) may be summarized as : V. lutea 2n — 148 V. curtisii 2n = 26 V. tricolor 2n == D6 V. arvensis Dn = 34 V. nana Qn — 48 It would thus appear that chromosome number offers an important criterion for the separation of V. arvensis from V. tricolor, and V. arvensis from the V. kitaibeliana complex (includes V. nana), two pairs of taxa difficult to delimit satisfactorily on purely morpho- logical criteria. Counts reported by Fothergill (1944) for V. variata var. sulphurea (2n = 26, 2n = 28) and V. contempta (2n = 40) refer to AT-intermediates as defined above and will be discussed in another paper. Watsonia 6 (1), 1964. CHROMOSOMES AND POLLEN OF BRITISH PANSIES TABLE |. List of chromosome counts. 41 lutea curtisii tricolor arvensis nana Reference Number of Population L18 L27 L28 L50 L52 L58 TC44 TC67 TC68 TC69 TC81 TC83 Locality Moorhouse, Westmorland Malham Tarn, Yorkshire Combs Dale, Derbyshire Widdybank Fell, Teesdale, Co. Durham Schiehallion, Perthshire Wardlow Hay Cop, Derbyshire Aberffraw, Anglesey. Kenfig Burrows, Glamorgan Broughton Burrows, Gower, Glamorgan Llangennith Burrows, Gower, Glamorgan Castlegregory, Co. Kerry, Ireland Corbally, Co. Waterford, Ireland T10 T14 T20 T23 25 T34 Wy) T80 Al A2 A5 A6 AT7(A) nr. Aberystwyth, Cardiganshire nr. Durham, Co. Durham Orpington, Kent nr. Aberfeldy, Perthshire Lawers Burn, Perthshire Feltham, Middx. High Force, Teesdale, Co. Durham Yetholm, Roxbs. Southampton Common, Hampshire Highfield, Southampton, Hampshire Swaythling, Southampton, Hampshire Sholden Downs, nr. Deal, Kent Kingsdown, nr. Deal, Kent nr. Aberystwyth, Cardiganshire nr. Aberystwyth, Cardiganshire nr. Durham, Co. Durham Frome, Somerset Porton, nr. Salisbury, Wiltshire nr. Southampton, Hampshire Long Down, New Forest, Hampshire Kempshott, nr. Basingstoke, Hampshire Stoborough, nr. Wareham, Dorset nr. Hursley, Hampshire nr. Hursley, Hampshire nr. Rockland Broad, Norfolk nr. Perranporth, Cornwall St. Martin’s, Isles of Scilly St. Mary’s, Isles of Scilly Chilworth, nr. Southampton, Hampshire Twyford, nr. Winchester, Hampshire Arne, Dorset St. Mary’s, Isles of Scilly nr. Eastleigh, Hampshire nr. Romsey, Hampshire nr. King’s Somborne, Hampshire nr. Corhampton, Hampshire St. Brelade’s Bay, Jersey St. Quen’s Bay, Jersey Tresco, Isles of Scilly Bryher, Isles of Scilly (1) 24 24 24 24 24 24 13 13) 13 13 13 13 1185 13 13 13 13 13 8) 13 7/ 17 7 17 17 17 17 17 17 I 7/ 17 17 L7/ 17 17 17 7/ Ha) 1 7/ NY 17 7 17 lig 17 ily 17 17 24 24 24 24 RT (2n) Chromosome number PMC Watsonia 6 (1), 1964. 42 A} PERTED (a) Methods For reasons of practical convenience most chromosome determinations were made from squash preparations of PMC meiosis although confirmatory observations were some- times made on similar preparations of root-tip mitosis. In the latter case root-tips were pretreated for approximately four hours with a satur- ated solution of 8-hydroxyquinoline before fixation in 1:3 acetic alcohol and stained with acetic-orcein as in the technique of Tjio & Levan (1950). When PMC meiosis was used satisfactory stages in division were found when the flower buds were c. 2 mm long and almost invariably occurred during the late morning. Flower buds were fixed in Carnoy’s solution for approximately + hour and then were transferred to 1:3 acetic-alcohol, containing a little ferric acetate or chloride, and were usually stored at —15° C in this solution until needed. The PMCs were subsequently stained with acetocarmine solution. Observations were usually made on temporary slides ringed with a rubber solution to avoid the loss of contrast in staining sometimes experienced when the slides were being made permanent by the dry-ice method (Conger & Fairchild 1953). In most cases counts were made from several cells on each slide and, wherever possible, repeated on 1-5 plants in each population examined. Whenever a population was morpho- logically variable, counts were also made from as many of the main types as possible. (b) Observations Details of the counts obtained are given in Table | and illustrated in Figs. 1 and 2. With but a single exception they agree with those already listed. The exception concerns a single plant of V. nana from six originally collected from Tresco, Isles of Scilly, and trans- planted to the gardens at Southampton University. This plant had 27 = 50, instead of 2n = 48 as in the other five examples from the same population. The count was repeated on three different buds from the same plant and in all of these meiosis appeared perfectly Fig. 1. Meiosis in PMCs. (a) V. lutea (L50): Prometaphase of second division, n = 24; (b) V. tricolor (T14): Second metaphase, n = 13; (c) V. arvensis (A75): Late stages of first anaphase, nm = 17; (d) V. curtisii (TC44): Late stages of first anaphase, n = 13. Watsonia 6 (1), 1964. CHROMOSOMES AND POLLEN OF BRITISH PANSIES 43 Fig. 2. Meiosis in PMCs of V. nana (N64). (a) Second metaphase, n = 24; (b) First metaphase with 25 bivalents, n = 25. regular with the formation of 25 bivalents and without the expected occurrence of unpaired chromosomes or chromatid bridges. Since this particular plant was morphologically in- distinguishable from others having chromosome numbers typical of V. nana and the count could not be repeated in other plants raised from seed collected from the original population, it was probably of little significance, being merely an aberrant chromosomal form which may have arisen as a result of primary nondisjunction of chromosomes or misdivision of a centromere in the previous generation. Incidental observations on chromosome behaviour at meiosis have indicated some differences between the taxa studied. In many plants of V. arvensis 0-5-1-0% of the PMCs had two (or four) univalents which had failed to pair and were either lagging on the meta- phase plate or lying disorientated in the cytoplasm (Fig. 3). Chromatid bridges were re- corded only twice in this species. In plants of V. nana the only irregularities observed were one or two chromatid bridges at first anaphase in many PMCs of a single plant of N65; otherwise meiosis was perfectly regular. In this connection it is interesting to note that Clausen (1931), describing material of this taxon from Jersey, also observed ‘slight ir- regularities’ although he stresses that the plants had a ‘fairly constant chromosome number’. He gave, however, no indication of the nature of these irregularities or how constant the chromosome number was. In V. tricolor the frequency of chromosomal irregularity varied considerably from population to population. Although most plants examined had regular meiosis, others showed well-defined irregularities. Dicentric bridges and acentric fragments, suggesting the presence of large inversions, were observed in nearly all plants examined from T10 (Fig. 3). In both T25 and T79 some plants had chromatid bridges visible at first anaphase in most PMCs but poor staining made it difficult to be sure of their cause. As in V. arvensis, disorientated univalents were rare, being found with a similar frequency. The situation in both V. Jutea and V. curtisii was roughly parallel to that in V. tricolor. Plants from the majority of populations had regular meiosis whilst others in certain popu- lations, notably L50 and L58 (V. lutea) and TC44 (V. curtisii) had high frequencies of chromatid bridges. Watsonia 6 (1), 1964. 44 A. PETTET These differences in frequency of chromosomal irregularities might be associated with the breeding system of the taxa involved. Thus, meiosis in the predominantly inbreeding. plants of V. arvensis and V. nana was most nearly always regular, whilst in the outbreeding plants of V. tricolor, V. lutea and V. curtisii chromosomal irregularities were more frequently observed. Presumably the nearly complete genetic homozygosity of the inbreeders would be responsible for their meiotic regularity, whereas the greater chromosomal heterozygosity in the outbreeders was responsible for their chromosomal irregularities. Fig. 3. Some irregularities in PMC meiosis. (a) V. arvensis (A36): Second prophase with univalent and fragment (X) in cytoplasm; the upper complement contains 16 chromosomes, the lower 17 chromosomes.. 2n = 34. (b) V. arvensis (A36): Prometaphase of second division with chromatid bridge persisting from first division; disorientated univalent (Y) from first division to one side; each complement has 17 chromosomes but the number has been maintained by the precocious division of the other univalent (X and X!). 2n = 34. (c) V. tricolor (T10): Late stages of first anaphase with two chromatid bridges and two acentric fragments (X and X!). 2n = 26. (d) V. curtisii (TC44): First anaphase with two chromatid bridges. 2n = 26. (e) V. arvensis (A37): Second anaphase with a chromatid bridge between two chromosomes of upper complement; a univalent from the first division lying disorientated in cytoplasm; upper group of cells with 16 pairs of chromosomes, lower with 17 pairs of chromosomes. 3. POLLEN POLYMORPHISM Specific references to the pollen morphology of V. tricolor and its immediate allies. contain a number of confusing statements. Faegri & Iversen (1950) place V. tricolor into their group, the Stephano-colporatae, i.e. pollen characterized by four furrows with median pores. Wittrock (1897), comparing pollen of different species of pansies, stated that V. Watsonia 6 (1), 1964. CHROMOSOMES AND POLLEN OF BRITISH PANSIES 45 tricolor had mainly 4-colpate grains with some 3- and 5-colpate grains, whereas V. arvensis had mainly 5-colpate grains with a few 4-colpate ones. On the other hand, data for these same two taxa given later by Erdtman (1952) would suggest they are not so readily distinguishable since both are recorded as having assemblages of 4-colpate grains, but with the occasional 5-colpate grains in V. arvensis. Counts given by Mullenders & Mullenders (1957) show V. tricolor subsp. tricolor with mostly 5-colpate and a few 4- and 6-colpate grains, and V. maritima Schweigg. (= V. curtisii Forst., V. sabulosa Bor.) with mostly 4-colpate, some 5- colpate and a few 3-colpate grains. (a) Methods Wherever possible, slides were made of a small sample of plants from each population with pollen from one flower of each representative plant being mounted on a slide. For good staining and ease of mounting a slight modification of the methylene-green/glycerine- jelly method of preparing pollen grains (Wodehouse 1935) was used. Pollen grains were first stained on the slide with a weak solution of eosin in absolute alcohol and this was allowed to evaporate almost to dryness before the grains were mixed in a warm solution of methylene green in glycerine jelly and then covered with a coverslip. As the grains expanded fully before the jelly set, their furrow configuration was immediately obvious. Pollen from both fresh and dried flowers was treated in this manner although pollen from the former was the most convenient to handle since the anthers had merely to be tapped on the slide to release the pollen. Anthers from herbarium specimens of less than a year’s standing could be similarly treated after their dissection from the flower since the pollen still expanded quite readily in the glycerine jelly. Pollen from older or badly prepared herbarium specimens did not expand so readily, so the technique was slightly altered. The flowers were dissected to expose the anthers and then left overnight in an humid atmosphere to allow the grains to swell before the staining and mounting. Such material gave slightly inferior preparations but there was usually little difficulty in identifying the different pollen types. The pollen assemblage of each plant was assessed by counting the different grain types on the slide. Both fully developed and aborted grains were counted over the entire field under the coverslip to avoid possible bias due to the differential movement between the larger, fully-developed grains and the smaller, aborted ones through the warm glycerine jelly on mounting. This usually entailed the counting of 300—-1,200 grains. (b) Observations The counts for all plants of the populations studied are summarized in Table 2 and Fig. 4, which give the frequencies of the different pollen grains as mean percentages for the population. More detailed data for the individual slides are available in Pettet (1960). These counts showed that the taxa represented by these populations could be divided _ into two groups according to the most abundant grain type and the less frequent types associated with it: (i) V. lutea, V. curtisii and V. tricolor: mostly 4-colpate grains, with some 5- and a few 3-colpate grains. Although 6-colpate grains normally did not occur in this group, a few examples were found in several plants from L50 (V. Jutea) and TC68 (V. curtisii). In all cases their frequencies were very low, being of the order of | per 200-300 grains, and their appearances were of abnormal or unreduced pollen. They were larger than normal grains and had thickened exines and often poorly developed colpae. Two exceptional assemblages were also found. One plant in L50 had 71% of its assem- blage composed of 5-colpate grains (cf. second group below), whilst another plant of L58 (V. lutea) had a greater proportion of 3-colpate grains than of 4-colpate ones. Watsonia 6 (1), 1964. A. PETTET 46 L:0 -0:0 1-0 9:0 -0:0 1-0 8:0 -0:0 1-0 8°0 -0-0 c:0 ¢:0 -0:0 £:0 UOTJEWIOJUT ON UOMeWIOJUT ON v:0 -0-:0 1-0 6-1 -0-:0 €:0 UOTWVWIOJUT ON UOI}CULIOJUI ON, UOTJEWIOJUL ON €:0 -0:0 1-0 VL -0:0 L:0 I-¢c -0:0 9:0 6:0 -0:0 c:0 UOTJCULIOJUT ON 8-T -0-0 v:0 L:0 —L:0 L:‘0 6:0 -0:0 v:0 6:0 -0:0 €:0 L:{t -0:0 ¢:0 9-cI-0:0 6-1 6:1 -0-0 9:0 Saspuaosad aspjuadsad Ul ASUDY uDIWFG SUIDASOANLJN Th SSS aoe aba ann A oy ai co | oo 8 coooocooqoqo oon o& DAOnnonnannanst ta ai I val SON NAN © ~~ MNOMnMN sO es Ot CO = ™~ ne ease loa) = ° N oO anon ‘Th-6: 69-6: ~ Ov + +naAA i6Vaxon *LC=6° SaspJuadsad Ul ASUDY o£ i 0) v0-98 I9-€T = | E°S 6c:0 OS- pL I@-S7 = ¢:8P COA cv: L8 €9:8 — 6-1 08-ZL LE-16 LL-O = 8°C 66:0 61-CL 98-97 as vl cS :0 8-68 99-6 ss 6:1 LS:0 v0: v6 6£°¢ all 0-1 90-1 18-S6é is {Lote cane [-¢ 98-1 17-68 €6°S = 8-€ OI-C 88-16 cO-9 = (6G vO-T¢ 68-89 L0:0 i (3,96 €v-0 OL: 68 L8-6 a vv c£-0 C8: 16 OL-L a (G2%3 90-7 Iv-98 iso] = 8-81 =: 96°£ C6: S6 ¢l-0 SS, = ¢I-8 LEG 87-0 8°C aa 69-81 Seats = ey . 69-8 C6: 06 tv-0 pats ae VS: OV: L6 90-0 1-6 £0-0 (heo/e}| 8°-C8 L0:0 VoD =a €8-9 67-68 89-¢ 8: VE 90-0 vL:-9T LL-C8 tv-0 9-97¢ = vs: 1 €°-06 v9-L [Efe v0-0 CS-L C8-68 SSG aspjuaosad ajvd]0I-9 ajvdjoo-¢ ajvdjoo-p ajvdjor-¢ uvaP sada] unas uayjod fo Aduanbadf asvjuadsad uvapy UOlJAOGD UA]JOd suonendod snowed jo sose[quiossy us][Od ‘7 ATAV |, syonplalpuy fo AAQqUnN 97V (V)OLV O1V (WILV TV O8.L SOL DILL OIL 801 89).L LOD VIOL 8c OSI uoipjndod fo saqunu INUASA{IY DUuDdU SISUIAAD AOJOIIAJ NS1JANI pain] Watsonia 6 (1), 1964. CHROMOSOMES AND POLLEN OF BRITISH PANSIES 47 (ii) V. arvensis and V. nana: mostly 5-colpate grains, with some 4- and 6-colpate grains. Tri-colpate grains were never recorded in this group but the frequency of the 6-colpate grains suggests that the latter correspond to them here. The 4-colpate grains were normally more frequent than the 6-colpate grains except in A10(A) (V. arvensis) where the latter were found in consistently higher percentages. A single plant with an exceptional assemblage resembling that of group (i) was recorded in N64 (V. nana). 100% 3-COLPATE OR 6-COLPATE ‘ 80% 90% S-COLPATE Fig. 4. Scatter diagram of mean percentage frequencies of grain types for populations listed in Table 2 l= Vlutea; @ = V. curtisii; O = V. tricolor; @ = V. arvensis; * = V. nana. The scatter-diagram of mean frequencies of pollen types for each population (Fig. 4) shows the distinctness of these two groups. Within the left-hand group the assemblages of V. curtisii and V. tricolor are very similar and not readily distinguishable; V. /utea may differ from both by having a slightly higher frequency of 3-colpate grains with a corres- ponding reduction of 5-colpate grains but the data are too few to be certain. Within the right-hand group V. arvensis and V. nana are basically similar although V. nana may have a tendency towards a lower frequency of 5-colpate grains than V. arvensis, but the latter is rather widely-ranging in this respect. Although within any one population the assemblages were relatively uniform in com- position, differences from plant to plant did occur, but, with the exception of those already mentioned, these were not sufficient to cut across the limits of the groups defined above. It may be supposed that this variation was controlled partly by genetic differences and partly by environmental differences between individual plants. That this was not the whole explanation was shown by some chance counts on a series of duplicate slides. These are given in Table 3 and represent counts of slides made from different flowers of the same plant collected at the same time. The differences between duplicates in the ratio of “most frequent grain type’ to ‘rest of assemblage’ were mostly significant when tested statistically. This means that the differences were unlikely to have been the outcome of random sampling and Suggests, perhaps, the conditions experienced by flowers on different branches of the same plant were sometimes sufficient to cause the small anomalies. It is, of course, unlikely that Watsonia 6 (1), 1964. 100:0> d { = LL+% ETL a OI q JUSIOYIP APUBOYIUSIC = TL: 8 87-ST = 68S B LS61 “349g ceV 100:0<10:0> d { ve-0 99-66 = = £67 q jusIEyIp AjuBoyTUsIg || €S-€ IS -S6 96-0 = v79 e 9561 ‘90 67V 5 | 100-0 <10:0> d { Cr: l 80-€6 Lv-s = 179 q JUDSIOYIP A[JUBSYTUBIC IL-0 Lv-88 78-01 = Sty | e 9661 “3dag S/ELV 100:0> d { = 9-89 9€-1E = | 68£ q JUDIOYIP APUBSYIUSIS | — 65° +8 €p-SI = CL8 B 9661 ‘Ideas v/ElV = | 0£-0 £6-98 LL-TI = 67E | q UdIOIp AULOYIUSIS 10 { | ea emer rN = 18-€8 61-91 = ISv e 9661 ‘3dag CHIGING EH aay A = 96-98 by: el = ONS q JUSIOYIP AJULOYIUBIS 10 < eee SN { IT-0 Cv-68 by: Ol = 676 e 966] “3d9g 7/Z1V = 89-67 ZE-OL = vLe q JUDIOYIP ATUCOYIUSIS 10 BEE eas a { = 70-672 86-OL = €6I e 9661 220 7/vIL 10:0 d { = vI-L ZL: C6 71-0 Tp q JUSIOYIpP ATUPSIUSIC = 67-01 8-68 €1-0 COL e 9S6[ ‘des PL/OLL ajvdjo2-9 ajpdjoj-¢ ajvd]0I-p ajvdjoo-¢ padoos Aaqunn Jada] aouvoyfiusis — SulDAs aps aod Juv] sada] uipas uayjod fo Aauanbadf aspjuaosag ]D1OL a ee Ne Sapl]s oyeoT[dnp Jo sose[quissse UsT[od Ul UOTeLIeA “¢ ATAVL, 48 Watsonia 6 (1), 1964. CHROMOSOMES AND POLLEN OF BRITISH PANSIES 49 differences between flowers such as these would ever cause the assemblage of V. arvensis to resemble that of V. tricolor, or vice versa. Data on the incidence of aborted grains and micrograins are also summarized in Table 2. Aborted grains were easily distinguishable from normal ones. They were smaller, contained little or no cytoplasm and in consequence lacked the granular appearance of the normal grains; the furrow configuration was present but usually inconspicuous because of their shrunken state; with methylene blue and eosin they usually stained a blue-green in contrast to the purple colour of the normal grains. Micrograins were somewhat similar to the aborted grains but different in that they were much smaller, lacked furrows, had very thin walls and consequently stained very lightly. Whereas the aborted grains were merely pollen grains which had failed to develop completely and mature, the micrograins frequently owed their origin to the aggregation of cytoplasm around lagging and disorientated chromosomes or chromosome fragments. Since, then, micrograins are not strictly pollen grains they were not included in the total grain count for the calculation of percentage pollen abortion. Their incidence was expressed, however, as a percentage of the total grain count to indicate very roughly the incidence of chromosomal abnormalities during pollen formation. A perusal of the data given in Table 2 will show that there was a tendency for the mean frequency of pollen abortion to be higher in the predominantly outbreeding taxa, V. lutea, V. curtisii and V. tricolor, than in the predominantly inbreeding taxa, V. arvensis and V. nana. This is probably a reflection of the greater heterozygosity for chromosomal inter- changes in the former. (The anomalous high values of N45 were probably caused by the very heavy aphid infestation of the plants prior to pollen collection.) In contrast, the mean incidence of micrograins is more or less uniformly low throughout the range of populations where information is available. 4. DISCUSSION A comparison of the chromosome counts with the type of pollen assemblage for the different taxa shows a certain agreement between the two sets of data. It should be noted that the two very variable and often confused species, V. tricolor and V. arvensis, are readily separable on the basis of both these characters. In fact, the close association between the two characters means that a determination of the pollen assemblage would be sufficient to place any doubtful specimen into one of the two species. Details of this character are now to be found in the second edition of Clapham, Tutin & Warburg’s Flora (Warburg 1962). This should be of particular use with herbarium specimens where identification is often difficult because of poor preservation or undue fading of the flowers, and where counting of chromosomes cannot be contemplated. In the case of the closely related taxa, V. tricolor and V. curtisii, neither the chromosome numbers nor the pollen assemblages offer grounds for their separation into two distinct species. This is contrary to the statement of Mullenders & Mullenders (1957) who maintained they were separate species largely on the basis of pollen assemblages. In the light of this present work it seems that, although the Mullenders emphasized their plants to be V. tricolor subsp. tricolor, they were in fact dealing with V. arvensis or something closely related to it. This is supported by the ecological details given by them which indicate that most of the plants were growing in calcareous soils. In Britain and Scandinavia at least, V. tricolor subsp. tricolor is confined to soils ranging from acid to neutral, whereas V. arvensis is characteristically found on calcareous soils (cf. Ferdinandsen 1918, Clausen 1922). Study of the AT-intermediates to be discussed in detail in another paper also indicates how this confusion probably arose. The figures given in Erdtman (1952) showing that V. arvensis has 4-colpate with a few 5-colpate grains must also be based on a similar confusion over specific limits since this type of assemblage is found in V. tricolor and not V. arvensis. There is little reason for be- lieving that Scandinavian plants of the latter species are any different from British examples in this respect, especially since Wittrock’s original diagnoses of pansies using this character Watsonia 6 (1), 1964. 4A 50 A. PETTET refer in particular to Scandinavian plants, and they are in complete accordance with the present findings. Finally, in addition to recommending the use of pollen assemblages for decisions on critical plants of V. tricolor and V. arvensis, the uncertain value of a high level of pollen abortion as a criterion for putative hybridity should also be emphasized. Clausen (1931) has already shown most interspecific hybrids to be relatively fertile, and very few to be completely sterile, whilst the data given above show that a certain amount of pollen abortion could be expected in at least some examples of the outbreeders, V. tricolor, V. curtisii and V. lutea, without the complications of interspecific hybridization. ACKNOWLEDGMENTS The work embodied in this paper was carried out during the tenure of a Nature Conservancy Research Studentship for which I express my thanks. I also wish to thank Dr. J. M. Lambert for her advice and encouragement throughout this work. I am also grateful to those who sent me samples of seed and living plants. REFERENCES CLAUSEN, J. (1921). Studies on the collective species Viola tricolor L. Preliminary notes. Bot. Tidsskr. 37, 205. CLAUSEN, J. (1922). Studies on the collective species Viola tricolor L. Il. Bot. Tidsskr. 37, 363. CLAUSEN, J. (1924). Increase of chromosome numbers in Viola experimentally induced by crossing. Hereditas, Lund 5, 29. CLAUSEN, J. (1926). Genetical and cytological investigations on Viola tricolor L. and V. arvensis Murr. Hereditas, Lund 8, 1-156. CLAUSEN, J. (1927). Chromosome number and the relationship of species in the genus Viola. Ann. Bot., Lond. 41, 677-714. CLAUSEN, J. (1931). Cytogenetic and taxonomic investigations on Melanium violets. Hereditas, Lund 15, 219-308. Concer, A. D. & FAIRCHILD, L. M. (1953). A quick-freeze method for making smear slides permanent. Stain Tech. 28, 281-283. DANDY, J. E. (1958). List of British Vascular Plants. London. DRABBLE, E. (1909). The British Pansies. J. Bot., Lond. 47, Suppl. 2, pp. 1-32. DRABBLE, E. (1926). Notes on the British Pansies. The arvensis series. J. Bot., Lond. 64, 263. DRABBLE, E. (1927a). Notes on the British Pansies. The tricolor series. J, Bot., Lond. 65, 42-53. DRABBLE, E. (19275). Notes on the British Pansies. The saxatilis series. J. Bot., Lond. 65, 167-171. DRABBLE, E. (1927c). Notes on the British Pansies. The /utea-curtisii series. J. Bot., Lond. 65, 213-219. DRABBLE, E. (1928). Notes on the British Pansies. Viola nana. J. Bot., Lond. 66, 129-132. ERDTMAN, G. (1952). Pollen morphology and plant taxonomy. Angiosperms. Stockholm. FAEGRI, K. & IVERSEN, J. (1950). Textbook of modern pollen analysis. Copenhagen. FERDINANDSEN, C. (1918). Undersggelser over danske Ukrudtsformationer paa Mineraljorder. Copenhagen. FOTHERGILL, P. G. (1944). Studies in Viola. 1V. The somatic cytology and taxonomy of our British species of the genus Viola. New Phytol. 43, 23-35. MULLENDERS, W. & MULLENDERS, E. (1957). Les pollens de Viola tricolor L. tricolor et de Viola maritima Schweigg. Bull. Soc. Roy. Bot. Belge. 90, 5-12. PeTTET, A. (1960). Variation within the British representatives of the Melanium subgenus of Viola. Ph.D. Thesis, University of Southampton. Tyo, J. H. & LEVAN, A. (1950). Quadruple structure of the centromere. Nature (Lond.) 165, 368. WarsurG, E. F. (1962). Violaceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F. Flora of the British Isles, Ed. 2. Cambridge. WITTROCK, V. B. (1897). Viola-studier. I. Morfologisk-biologiska och systematiska studier 6fver Viola tricolor L. och hennes narmare anforvandter. Acta Hort. berg. 2, 1-142. WODEHOUSE, R. P. (1935). Pollen Grains. Ed. 1. New York. Watsonia 6 (1), 1964. STUDIES ON BRITISH PANSIES Il. THE STATUS OF SOME INTERMEDIATES BETWEEN VIOLA TRICOLOR L. AND V. ARVENSIS MURR. By A. PETTET Department of Botany, University of Southampton* ABSTRACT Plants morphologically intermediate between Viola tricolor L. and V. arvensis Murr. have, in the past, been interpreted either as hybrids between the two species or as ‘stabilized’ hybrid segregates. Several recent gatherings of such intermediates (AT-intermediates) and putative hybrids have been studied and compared with experimentally produced F, and Fo hybrids. Observations on the morphology and cytology of the experimentally produced hybrids largely confirm previous accounts. Pollen assemblages, not previously examined in any detail, were shown to be variable in both generations, ranging in character from that of V. tricolor to that of V. arvensis. Percentage pollen abortion was generally below 40% with a mean value of 22-8 % for the F, and 24-4% for Fy. The incidence of micro-grains was usually greater than 2% with the mean value dropping from 7-4% in the F; to 3-:9% in the Fo. One gathering of intermediates from a mixed population of V. tricolor and V. arvensis was shown to be indistinguishable from experimentally produced hybrids. AT-intermediates from populations composed otherwise of V. arvensis or entirely of the intermediate form were shown to be indistinguishable from V. arvensis on the basis of chromosome number and pollen morphology. Observations on successive generations of AT-intermediates grown in an experimental garden showed the original gatherings were composed of two general elements, (i) plants losing their intermediate features and becoming indistinguishable from the small-flowered V. arvensis, and, (ii) plants retaining their characteristics which were presumably genetically controlled. The first group simulate AT-intermediates through phenotypic plasticity. It is suggested that the second group represent either the original outbreeding forms of V. arvensis or intermediate stages in the development of the inbreeding régime typical of V. arvensis, rather than ‘ stabilized’ hybrid segregates. In consequence of this it seems advisable to widen the circumscription of V. arvensis to include these extreme morphological forms. 1. INTRODUCTION During a study of the cytology and pollen morphology of some British pansies (Pettet 1964) it was found convenient to divide the Viola tricolor—V. arvensis complex of annual pansies into three main groups, viz. (i) V. tricolor, (ii) V. arvensis, and (iii) intermediates between these, designated ‘AT-intermediates’. The last-mentioned had cream, or pre- dominantly cream flowers, the upper petals longer than the upper sepals (petals usually 14-14 xX the length of the sepals) and stylar flap intermediate in development between those of V. tricolor and V. arvensis. In brief, the AT-intermediates had flowers resembling those of V. tricolor, except that they were cream-coloured, had a smaller stylar flap and a somewhat reduced ratio of upper petal to upper sepal. As was previously stated, the AT- intermediates so defined should not be considered a distinct entity—only one of convenience encompassing those plants not readily ascribed to either V. tricolor or V. arvensis. Some individuals can be equated with the cream-flowered species or varieties of Drabble’s Tricolores (Drabble 1909, 1927a), e.g. V. contempta Jord., V. variata Jord. var. sulphurea Drabble, but others are not so readily identified with these microspecies. Wittrock (1897), Kristofferson (1923) and Clausen (1922, 1926) have reported similar plants from a number of different places on the Continent and a study of specimens in the herbaria of Kew and * Now at Department of Botany, University of Khartoum. 51 Watsonia 6 (1), 1964. 4B a2 A PETIET the British Museum (Nat. Hist.) has shown that AT-intermediates occur not infrequently in Britain. Clausen (1922, 1926) examined a number of intermediates occurring in mixed popula- tions of V. tricolor and V. arvensis and showed them to be hybrids on cytological evidence and on comparison with experimentally produced hybrids. Others occurring without either parent were found to be cytologically regular and these Clausen suggested were ‘stabilized’ hybrid segregates of long standing. By way of confirmation for this conclusion Clausen showed how, after a few generations, some segregates of intermediate morphology derived from the interspecific cross V. tricolor “Ou AUDI (uonesoucs 1.7) -8EV IvV LEV 9EV OV CV ‘ou uoljpjndog a (Aussoid svt pue [/ZTW SUIpnfoxe) soyeIpowioUI-[W JO SUTeIZ-oJOIWU pue UOIIOge UdT[Od Jo ssoUSpIOU! pu soSe[quiesse UaTOg ‘OT ATaV I. Watsonia 6 (1), 1964. 65 INTERMEDIATES BETWEEN VIOLA TRICOLOR AND V. ARVENSIS SS SS Se Save al a i Aa a) ©) 62 SQ oS 2) OS S&S oS) SSeS] (>) ooeooeowrwoa yToorwrt-w3s= SUIDAS-OADIU f0 aspjuadsag C6: hee CO: VC: GS: Be: 06: GE: Oil: CI: IV: ve 66: ve: LV: QE NOwnwnwwewanas N = mM m~wwrmroo=naNnm wo SUIDAS pajsoqv | f0 aspJuansag pajunos SUIDAS fo Jaqunn ae Iss as 69: = Be: eS 9S: 91-0 ev: = 86° = Wye es VC: = Sv: “OS cer v6: = 09 ajvdjoo-9 add} uipss-uajjod fo Aduanbadf saspjuad 8V 69 6S cL cL EE OL Die OL 9¢ 61 LC Sc (G4 61 ajvdjoo-¢ 69° “Of Ie (Oy vv: cv: E6, 08: co: Ge; Ee 6V- 88- Sv: $6: OV: 90: 6L ajvd]0I-y TS Ov 9¢ LC 6¢ tc 6¢ ev 08 cL vL LS 6V 08 ajvdjoo-¢ Jad OSE VLS LL9 9¢¢ 1¢9 Ice 861 Oes 099 S99 ISO'T C17 T S6ET S7O'T ISL CVE padsors SUuIDAS fo 4aquUunny om ome tet eet eee et eet At “ou aps Aussoid sit pue [/Z7]V jo suleis-o191l JO UOTJIOQe Us[Od JO SsoUDpIOUL pue Ssose[quiosse Ud[Og “[] ATAV] re AaNnNTN OM OD © ‘OU JUD] ed td W/V UONDAQUAL) = \O ox =— ~~ = — \© S ~ S x S 66 A. PETTET Counts from most plants of this generation showed assemblages similar to those of V. arvensis but plants 1, 8 and 10 had assemblages midway between the two main types. In view of the comparative uniformity in gross morphological features of the various generations of this line and the absence of any similar variation in counts from other plants, except perhaps in those of hybrids, this variation was unexpected. Unlike the hybrids, however, the variation here was not associated with cytological disturbances for, as shown below, chromosome behaviour at meiosis was regular. It is thus difficult to explain these changes, especially as the genetical basis of pollen polymorphism is not fully understood. The variation in the assemblages recorded for the parent plant from the original population, and the comparative uniformity in the reversed assemblages of the F., suggest the differences here are not primarily controlled by genic segregation. Most of the variation must instead be explicable on the grounds of the plasticity of pollen morphology of this particular line of plants—a unique situation not observed elsewhere. TABLE 12. Chromosome numbers of AT-intermediates Population | Plant Chromosome number no. number n(PMC) —_2n (RT) A2 , 2A 17 34 AS 6 | 17 — Al2 1 17 — AT30.. |» ol=4 17 — A36 | 11 | 7, 34 | 12 | 17 Bs A37 | 10 | 17 = A38i 1 | 17 = AT41 1-5 | 17 = (ii) Cytology All plants collected from the field were examined cytologically and their chromosome numbers are recorded in Table 12. In every case the number is identical with that of V. arvensis and, like the latter, meiosis in all plants was regular with the formation of seventeen bivalents. Only very occasionally were PMCs found with a pair of unorientated univalents in the cytoplasm, or more infrequently still, with chromatid bridges present at first anaphase. However, these were no more frequent than to be observed in most plants of V. arvensis. 4. DISCUSSION From the observations elaborated above it is obvious that not all the AT-intermediates are of recent hybrid origin even though they resemble experimentally produced hybrids between V. tricolor and V. arvensis in gross morphology. In fact only those plants taken from a mixed population of the two putative parent species were found to be F, hybrids. The rest, from populations where at least V. tricolor was absent, were of two general sorts, viz. firstly, plants of V. arvensis simulating intermediates, and secondly, genetically determined (presumably) AT-intermediates. These two groups are both of interest in the unravelling of the taxonomic problems of the annual pansies, but for different reasons. The first shows that phenotypic plasticity occurs in populations of V. arvensis and may occasionally be sufficient to blur the limits of the species and cause confusion in the naming of plants. Any doubts about such plants, however, should be easily resolved by an examina- tion of the pollen morphology. The second group, on the other hand, raises problems over interpretation and perhaps, incidentally, over the recent evolution of the annual pansies. On the basis of pollen morphology and cytology—perhaps more reliable indicators Watsonia 6 (1), 1964. INTERMEDIATES BETWEEN VIOLA TRICOLOR AND V. ARVENSIS 67 of affinity in this instance than gross floral morphology—the genetically determined AT- intermediates appear most closely allied to V. arvensis. This suggests that two possible interpretations of their origin need be considered. Either they arise from past hybridization or there has been confusion over the real limits of V. arvensis and they represent extreme variants of this species. Past hybridization may have involved intraspecific or interspecific crosses. The possibility of intraspecific crosses in V. arvensis (cf. Clausen 1926), coupled with chromosome instability and loss, being responsible can be eliminated since this would lead to a reduction below 2” = 34. On the other hand these plants may be ‘stabilized’ hybrid segregates following interspecific hybridization and subsequent introgression as envisaged by Clausen, but this raises a number of difficulties. There is no doubt that hybridization between V. tricolor and V. arvensis does occur for, apart from the F, hybrids mentioned here, there are several reports of naturally occur- ring hybrids by other workers, e.g. Wittrock (1897), Clausen (1922, 1926). The question to be really considered is whether the occurrence of hybridization materially alters the variation patterns of the two species involved. Normally V. tricolor and V. arvensis do not grow together. Ferdinandsen (1918) and Clausen (1922) have both shown the two species to have their optimal developments on different soil types. V. tricolor occurs most frequently on acidic soils, V. arvensis on basic soils, with only a slight overlap on neutral to weakly acid soils. This edaphic differentiation was thought by Clausen to be sufficient to prevent hybridization occurring frequently enough to break down entirely the specific limits of the two species. When the two species hybridize, however, the F, is fertile and may produce much seed. Thus it would seem that only a few mixed populations where hybridization occurs would be sufficient to produce a whole range of segregates of intermediate morphology, providing these were not at a marked competitive disadvantage in the population. In the only mixed population seen during the present study F, hybrids alone were found. This may have been quite fortuitous as the two species may not have been growing together until the previous year, providing too little time to produce the second generation. However, observations on F, and subsequent hybrids grown in the experimental garden showed them to be rather weak and suggest that under field conditions they may be at a distinct dis- advantage—an assessment which is borne out by Clausen’s observations on similar hybrids. Even if hybrids can compete successfully in the field introgression will not necessarily occur. The F, and F, hybrids have been found to be capable of at least some self-pollination and under field conditions it seems likely that they would be largely self-pollinating. Should backcrossing occur it is more likely to be to the tricolor parent, which is a fairly well adapted outbreeder, than to the arvensis parent which, with its reduced flowers, is virtually always self-pollinating. As Clausen has already shown, and as has been observed in this study, the chromosome number in successive hybrid generations most frequently reduces from 2n = 30 of the F, towards 2n = 26 as a result of loss of the extra arvensis chromosomes. Yet no AT- intermediates were discovered with these lower numbers. Increase in chromosome number beyond 2” = 34 through irregular segregation of chromosomes has also been reported by Clausen, and the chromosome number (2n = 40) quoted by Fothergill (1944) for V. con- tempta Jord., a yellow-flowered Tricolores pansy, could possible be due to this. However, the increase of chromosome number reported by Clausen was of rare occurrence and was associated with chromosome instability. Consequently it is unlikely to have lasting effects on the variation pattern of these pansies. The production of plants with a stable chromosome number of 2” = 34 from hybrid segregates requires repeated backcrossing to the arvensis parent but this is unlikely to produce AT-intermediates as described above. It is known (Clausen 1926) that a number of inhibitor genes producing reduction of floral parts are present on the extra chromosomes of V. arvensis. It is therefore likely that backcrossing to the typical, small-flowered V. arvensis would thus produce a small-flowered arvensis-type plant rather than the present AT- intermediates. Watsonia 6 (1), 1964. 68 A. PETTET The interpretation of the AT-intermediates as extreme variants of a rather variable V. arvensis, whose exact limits have been confused in the past, raises fewer difficulties. The basis of the differences between V. tricolor L. and V. arvensis Murr. rests with Murray’s reference (Murray 1770) to Haller’s distinction (Haller 1745) between the two taxa: tricolor: ‘flore calyce duplo longiore’ arvensis: ‘flore calyce paulo majorr’ Somewhat incorrectly this distinction has come to be accepted as V. tricolor with ‘upper petals longer than upper sepals’ and V. arvensis ‘with upper petals shorter, or as long as, upper sepals’ (e.g. Clausen 1922, Drabble 1909, 1926, 1927a, 1927b, Warburg 1952). This general difference between the two taxa is associated with their reproductive behaviour. V. tricolor is well adapted to cross-pollination and has large, showy flowers with a well- developed stylar flap, whilst V. arvensis, as normally understood, is well-adapted to self- pollination and has reduced petals and a very much reduced stylar flap. The distinguishing features of V. arvensis are characteristic of successful exploiters of the ephemeral habitat offered by arable farmland, i.e. short life-cycle, increased flower production and reduced flowers promoting self-pollination (cf. Stebbins 1957). It is reasonable to assume, therefore, that V. arvensis (as normally understood) has developed from an outbreeding pansy mainly by selection of mutant genes present on the extra arvensis chromosomes which inhibit flower development to varying degrees. In view of the marked genetic similarity of the two species, as reflected in the pairing behaviour of the chromosomes in the F, hybrid, the original outbreeding V. arvensis also would resemble closely the present-day V. tricolor. Therefore one can consider that the morphologically constant, and presumably genetically determined, AT-intermediates, A12/1, A30, A41 and A37, represent either forms closely allied to the original outbreeding V. arvensis from which the small-flowered in- breeding V. arvensis was developed, or represent the remnants of an intermediate stage between the two. The present variation between populations of the small-flowered V. arvensis can therefore be traced back to those differences originally present in the heterozy- gous, outbreeding progenitors and which have since become segregated out and distributed into the now almost completely inbreeding lines. Such lines are usually morphologically uniform and the presence of the morphologically constant AT-intermediates in these populations must result from the mixing of two separate populations. On the basis of the available evidence it seems an economy of hypotheses to consider the cytologically regular and non-segregating AT-intermediates as being extreme variants of V. arvensis rather than ‘stabilized’ hybrid segregates. Taxonomically these plants are best considered as representatives of a variable V. arvensis Murr. since the observed variation suggests it would be possible to find a complete range in morphology from V. arvensis, sensu stricto, to the most extreme AT-intermediates defined above. V. arvensis would then be taken to include: annual pansies with cream, or predominant- ly cream, never blue or purple, flowers; upper petals usually shorter than, or as long as, the upper sepal, although occasionally longer, up to about x 12; stylar flap usually reduced and appearing absent but sometimes developed enough to be seen with a hand-lens; pollen assemblage composed predominantly of 5-colpate grains with a few 4-colpate grains and occasionally some 6-colpate grains; chromosome number 2” = 34. REFERENCES CLAUSEN, J. (1922). Studies on the Collective Species Viola tricolor L. Il. Bot. Tidsskr. 37, 363. CLAUSEN, J. (1924). Increase of Chromosome Numbers in Viola experimentally induced by crossing. Hereditas, Lund. 5, 29. CLAUSEN, J. (1926). Genetical and cytological investigations on Viola tricolor L. and V. arvensis Murr. Hereditas, Lund. 8, 1-156. CLAUSEN, J. (1931). Cytogenetic and Taxonomic investigations on Melanium violets. Hereditas, Lund. 15, 219-308. DRABBLE, E. (1909). The British Pansies. J. Bot., Lond. 47, (Suppl. 2), 1-32. Watsonia 6 (1), 1964. INTERMEDIATES BETWEEN VIOLA TRICOLOR AND V. ARVENSIS 69 DRABBLE, E. (1926). Notes on the British Pansies. The arvensis series. J. Bot., Lond. 64, 263. DRABBLE, E. (1927a). Notes on the British Pansies. The tricolor series. J. Bot., Lond. 65, 42-53. DRABBLE, E. (1927). Notes on the British Pansies. The saxatilis series. J. Bot., Lond. 65, 167-171. FOTHERGILL, P. G. (1944). Studies in Viola. [V. The somatic cytology and taxonomy of our British species of the genus Viola. New Phytol. 43, 23-35. HA ter, A. (1745). Flora Jenensis. Jena. KRISTOFFERSON, K. B. (1923). Crossings in Melanium Violets. Hereditas, Lund 4, 251. Murray, A. (1770). Prodromus designationis Stirpium Gottingensium. Goettingen. PetTeT, A. (1960). Variation within the British representatives of the Melanium subgenus of Viola. Ph.D. Thesis, University of Southampton. PeTTET, A. (1964). Studies on British Pansies. I. Chromosome Numbers and Pollen Assemblages. Watsonia 6, 39-50. STEBBINS, G. L. (1957). Self-fertilization and population variability in the higher plants. Amer. Nat. 91, 337-354. WARBURG, E. F. (1952). Violaceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles, Ed. 1. Cambridge. Ed. 2, 1962. WITTROCK, V. B. (1897). Viola-Studier. I. Morfologisk-biolgiska och systematiska studier 6fver Viola tricolor L. och hennes narmare auforvandter. Acta Hort. berg. 2, 1-142. Watsonia 6 (1), 1964. 5 MIMULUS HYBRIDS IN BRITAIN By R. H. ROBERTS ABSTRACT The hybrid M. guttatus x luteus has been synthesized and shown to be identical with naturally-occur- ring plants, which have been found to be widespread in Britain. Plants of the putative hybrid M. guttatus < cupreus have also been found to occur in several northern localities. INTRODUCTION According to Loudon (1872) species of the genus Mimulus were first brought into the British Isles just over 200 years ago. They were introduced from the American continent and became popular as border and pot plants. A few of them became naturalized: M. guttatus DC., said to have been introduced in 1812, was already established as an escape by 1830 (Warburg 1962); it is now widely distributed on the banks of streams and rivers, and is recorded from all but a dozen of the vice-counties in Britain. M. Juteus L. and M. moschatus Dougl. ex Lindl. have also become naturalized, but far less commonly; and more recently a fourth species, M. cupreus L., has been reported as a widespread escape in the English Lake District (McClintock 1957). M. guttatus is a + pubescent perennial, often glandular-pubescent, particularly on the upper part of the flowering stem, the pedicel and the calyx. It has a bilabiate yellow corolla, marked with small, red spots in the throat, and with the lower lip longer than the upper. It is a native of western North America, where it is very variable, being made up of many polymorphic populations growing in wet places (Vickery 1959). M. luteus on the other hand is completely glabrous except inside the calyx and corolla, the latter usually marked with large, reddish-brown spots on the lobes, in addition to small ones in the throat. The lips of the corolla, moreover, are nearly equal. It is a smaller, more prostrate plant than M. guttatus and is a native of Chile. M. cupreus is described as being very similar to M. Juteus, differing from it only in its compacter habit, smaller leaves and copper-coloured flowers, which may or may not have small, dark spots in the throat. Like M. Juteus, under which it was formerly placed as a variety (Bailey 1927), this also is a native of Chile. M. moschatus is a densely and viscidly-hairy plant with a much smaller, pale yellow, normally unspotted corolla, and smaller leaves of a paler green. It originates from western North America. THE NATURALIZED PUTATIVE HYBRID Despite the descriptions given in the floras many plants of Mimulus have proved difficult to identify and have often been misrecorded. For example, at several places in Caernarvonshire plants were found which could not satisfactorily be placed under either M. guttatus or M. luteus; they showed considerable variation in pubescence and corolla marking, many of them being tall, robust plants, with large, heavily-blotched corollas. In 1960 the intermediate character of a plant growing at Harlech, Merioneth, aroused the suspicion that it might be of hybrid origin—a view which was further strengthened when its pollen was found to be highly sterile. Consequently, similar plants from as many other localities as possible were examined. In many of these places the plants were so uniform as to suggest that they were, in fact, clones, and that propagation by seed did not occur there. Altogether, plants with highly sterile pollen from seven localities were taken into cultivation and, during the summer of 1961, pollen of four flowers from each of them was examined and an estimate made of the percentage of full pollen grains. This was done by mounting 70 Watsonia 6 (1), 1964. MIMULUS HYBRIDS IN BRITAIN 71 the pollen from undehisced anthers in a few drops of aceto-carmine, when the full grains took up the stain and could readily be counted. The results (Table 1) show that the pro- portion of full grains varied from plant to plant and even from flower to flower on the same plant. TABLE 1. Percentage of full grains in the pollen of the naturalized hybrid M. guttatus < luteus. Samples are from undehisced anthers of four flowers from each plant Flower no. Locality 1 2 3 4 Harlech 19 13 8 6 Bethesda 6 4 8 6 Llanllechid 0 3 8 14 Groeslon 1 0:5 1 10 Machynlleth 4 1 4 8 Nant y Garth 6 3 3 2 Penmachno 4 1 2 6 The plants from different localities were found to be very heterogeneous: they differed considerably in corolla size and marking, the latter varying from some which differed little from M. guttatus, to others in which the lobes were beautifully marked all over with a deep red or reddish-brown pattern of blotches and spots. But all of these plants possessed short, frequently very sparse pubescence, mainly on the upper parts of the flowering stems (particularly on the leaf bases and the nodes), on the pedicels and the calyces. Herbarium specimens of Mimulus from various localities in Britain were also examined and a very large number of these proved to be similar to the putative hybrid and to have highly sterile pollen. It therefore appeared that these plants were surprisingly widespread and a short series of experiments was carried out in an attempt to get further light on them. EXPERIMENTAL POLLINATIONS In order to attempt the artificial synthesis of the hybrid a search was made locally for M. guttatus and M. luteus. M. guttatus was found at a few places in Caernarvonshire; a plant from one of these, and another from Dolgellau, Merioneth, were grown for experi- mental crossing. The search for M. luteus was less successful. Although this species had been recorded from two Welsh vice-counties (Hyde & Wade 1957), when the specimens supporting these records were examined, both were found to be of the putative hybrid. Eventually, a plant of M. luteus was obtained from a garden in Evesham, Worcs., and, later, seeds were collected from a herbarium specimen of M. luteus gathered in Caithness by D. P. Young in 1959. When these seeds were sown on damp blotting-paper they germi- nated in about 4 or 5 days and the seedlings were transplanted into pots of sterilized soil and grown on. With these materials and the seven plants of the putative hybrid the following polli- nations were carried out. Except in the case of the trial selfings, the flowers were first emasculated to prevent self-pollination and each flower covered with a cellophane bag, which was kept in place until the style had withered. (a) Four flowers on each of the putative hybrid plants were selfed, but a few were pollinated with pollen from another of the hybrid plants. (b) Several flowers on each hybrid plant were pollinated with M. guttatus and several with M. luteus. Watsonia 6 (1), 1964. 72 R. H. ROBERTS (c) A number of flowers of M. guttatus were pollinated with M. /uteus and a number of the reciprocal pollination also made. After each pollination had been carried out the flower was labelled and the cellophane bag replaced. RESULTS The success of each pollination was estimated by gathering each mature ovary, whether it had developed into a capsule or not, and storing in a separate container to dry, after which it was examined for any seed set. TABLE 2. Percentage of full grains in the pollen of the syn- thesized hybrid M. guttatus x< luteus. Samples from four flowers per plant, referred to as H1, H2, etc.; R1 is the plant of the reciprocal cross Flower no. Plant 1 2 3 4 H1 5 3 4 4 H2 26 1 22 q H3 0:5 1 0:2 0 H4 3 2 5 4 H5 6 6 6 3 H6 12 13 16 31 H7 20 18 5 2 H8 2 1 1 1 H9 5 4 4-5 4 H10 0:5 1 1 0 Hil 5 35 1 6 H12 0 2 2 2 H13 5)°5) p 5) 6 H14 5 3 3°5 3)°5) H15 1 0:4 0 4 R1 9 5 2 3 The results may be summarized as follows. (a) In all the ovaries from the putative hybrids, whether selfed, or pollinated with each other, with M. guttatus or M. luteus, not one seed was set. (b) The capsules of M. Juteus which had been pollinated with M. guttatus were moderately well developed and, before being opened, gave the impression that fertilization had taken place. When they were examined, however, with one sole exception, they were found to contain empty testae. The exception was one capsule of the M. Juteus plant from Evesham, which was found to contain two well-developed seeds out of 466, the others being empty testae like those in all the other capsules. One of these two seeds germinated and grew into a vigorous plant (R1 in Table 2). (c) Capsules from M. guttatus pollinated with M. /uteus had a good set of seed, over 800 were counted in one capsule, but the seeds were variable in size compared with those of the parent species. Percentage germination in these was fairly good (c. 30%). Fifty-five ' of the resulting seedlings were transplanted into sterilized soil and grown to maturity. The synthesized hybrid plants Fifteen plants of the artificially synthesized hybrid were chosen at random and during 1962 and 1963 these were subjected to the same programme of pollinations as that carried out on the naturally-occurring hybrid. The results were identical: neither selfing, pollination Watsonia 6 (1), 1964. MIMULUS HYBRIDS IN BRITAIN 73 with one another, with M. guttatus, nor with M. luteus produced a single seed. The sole plant (R1) of the reciprocal cross gave the same results. _ Pollen from four flowers of each of these 15 plants and of the hybrid R1 was examined and the results (Table 2) agree very closely with those from the naturally-occurring hybrid. Other effects of pollination were observed. When hybrid plants with a very low pro- portion of full pollen grains were selfed there was no observable swelling of the ovaries or inflation of the calyces. But when those having higher proportions of full pollen grains (e.g. H2, H6 and H7) were selfed there was often a marked enlargement of both capsule and calyx. Pollination of the synthesized hybrid plants with M. guttatus also produced the same effects, but to an even greater degree; while pollination with M. luteus resulted in the most pronounced effects of all. The results, summarized in Table 3, suggest that these effects, which were also observed in the naturalized hybrid plants, may be due to hormone action following pollination. TABLE 3. Capsule development (mean of length x breadth in mm) in hybrid Mi- mulus after pollination with pollen from different sources Pollen source: M. luteus M. guttatus Hybrid 44-8 26:2 17:0 Variation in the synthesized hybrid plants The plants of the synthesized hybrid grown to maturity exhibited considerable variation in several morphological characters. In flower colour and density of pubescence they showed varying degrees of intermediacy between the parents, but the length of the hairs was usually less than half those of M. guttatus; in some cases the pubescence was so short and sparse as to be easily overlooked—as indeed it seems to have been in many specimens of the naturalized hybrid, which have consequently often been mistaken for M. /uteus. Stature, leaf-size and corolla-size were frequently much larger than in either of the parents, but about a dozen of the 55 hybrid plants were markedly lacking in vigour; although grown under the same conditions as the others, some of these only attained a stature of 5 to 8 cm and either bore minute flowers or failed to come into flower at all. DISCUSSION AND CONCLUSIONS The experimental pollinations have shown that the cross M. guttatus2 x M. luteus 3 occurs readily and gives a reasonably good seed-set, from a proportion of which highly variable F, plants result; the majority are large and vigorous. These F,s are completely sterile and are identical with the naturalized sterile plants, which are therefore presumed to be of the same parentage. The reciprocal cross M. luteus 2 x M. guttatus 3 appears to be very difficult to obtain, only one plant being produced out of many such pollinations; but this single plant was very vigorous and morphologically identical with the others. Like them, too, it was completely sterile. The synthesized hybrid plants also proved to be as winter-hardy as the naturalized ones. All the cultures were left outside throughout the severe weather conditions of the winter of 1962-63; whilst all the plants of M. luteus died, those of M. guttatus and both lots of the hybrid survived and grew vigorously in the following summer. This may well be the explanation of the widespread occurrence of the hybrid in the absence of the M. luteus parent. The reason for its presence in so many places is problematic: whether it is due to its deliberate production by nurserymen in the past, or to accidental crossing when both species were grown in the same gardens. Watsonia 6 (1), 1964. 74 R. H. ROBERTS LOCALITIES FROM WHICH PLANTS OF THE HYBRID MIMULUS GUTTATUS X LUTEUS HAVE BEEN SEEN V.c. 3, SOUTH DEVON: near Bridestow, 1953, N. Y. Sandwith (NMW); near Lydford, 1939, D. P. Young (herb. D. P. Young). V.c. 17, SURREY: Shere, 1935, D. P. Young (herb. D. P. Young). V.c. 35, MONMOUTH: Llanover Forest, Abercarn, 1954, R. E. Stumbles (NMW). V.c. 40, SHROPSHIRE: Oswestry, 1954, Miss Vera Gordon (herb. V. Gordon). V.c. 45, PEMBROKE: Aberfelin, Trevine, 1954, Mrs. M. Barnes (NMW). V.c. 46, CARDIGAN: Eglwys Fach, 1961, P. M. Benoit. V.c. 48, MERIONETH: Harlech, 1960, R. H. Roberts & P. M. Benoit. V.c. 49, CAERNARVON: Llanllechid, 1960, R.H.R.; Nant y Garth, 1960, R.H.R.; near Groeslon, 1960, R.H.R.; Tregarth, 1959, Mrs. I. Roberts (herb. R. H. Roberts); Gerlan, Bethesda, 1961, R.H.R.; same, 1954, Miss V. Gordon (herb. V. Gordon); near Waun Fawr, 1962, R.H.R.; near Penmachno, 1962, M. Morris; Cwm y Glo, Llanberis, 1941, R. D. Tweed & N. Woodhead (herb. Univ. College of N. Wales, Bangor). V.c. 50, DENBIGH: Llangollen, 1954, Miss V. Gordon (herb. V. Gordon). V.c. 52, ANGLESEY: near Pentraeth, 1961, R.H.R. V.c. 66, DURHAM: Teesdale, 1958, Miss U. K. Duncan (herb. U. K. Duncan). This had been determined as the putative hybrid by D. McClintock. V.c. 89, EAST PERTHSHIRE: bank of R. Tay, near Dunkeld, 1958, D. P. Young (herb. D. P. Young). V.c. 90, ANGUS: Wester Denoon, Sidlaw Hills, 1958, Miss U. K. Duncan (herb. U. K. Duncan). V.c. 97, WESTERNESS: Dalwhinnie, 1954, H. J. M. Bowen (herb. H. J. M. Bowen). V.c. 105, WEST ROSS: near Applecross, 1957, Miss V. Gordon (herb. V. Gordon). V.c. 108, WEST SUTHERLAND: Lochinver, 1961, Miss V. Gordon (herb. V. Gordon). OTHER HYBRID PLANTS Two specimens of Mimulus sent to me by Miss U. K. Duncan (Tarfside, Glen Esk, Angus, v.c. 90; near Fettercairn, Kincardine, v.c. 91) and one sent by Dr. D. P. Young (River Tay, near Dunkeld, East Perthshire, v.c. 89) had copper-coloured flowers with few or no small dots in the throat of the corolla; the calyces and pedicels were also minutely, but rather densely pubescent, and their pollen had a high proportion of empty grains. From their morphology these plants appeared to be the hybrid M. guttatus < cupreus. Similar specimens, except that they possessed a petaloid calyx, had been collected by Mr. R. C. Palmer and Mr. W. Scott (near Scalloway, Shetland, v.c. 112) and by Miss V. Gordon (near Inverkirkaig, West Sutherland, v.c. 108). Miss Gordon (in /itt.) informs me that this plant occurs in several localities in Sutherland. In 1962 a rooted piece of the Scalloway plant was obtained and some observations were made on it during the following summer. Pollen from 6 flowers was found to have from 50% to 60% of full grains—a much higher proportion than in that of the M. guttatus x luteus hybrid. Ten flowers were selfed and the resulting seed-set found to be extremely low: from 0 to 3 seeds per capsule. Only twelve seeds were obtained altogether, but these germi- nated fairly readily and the seedlings were transplanted into sterilized soil. Some of these seedlings had small and malformed cotyledons, while some lacked chlorophyll. None of them made any growth beyond the cotyledon stage, and even those with normal, green colouring had all died within a fortnight of germination. The evidence so far available thus favours the view that all these plants possessing copper-coloured flowers and a very short pubescence are also of hybrid origin, the presumed parents in this case being M. guttatus and M. cupreus. Nevertheless, the artificial synthesis of this hybrid is desirable before any final pronouncement can be made on them. Watsonia 6 (1), 1964. MIMULUS HYBRIDS IN BRITAIN 75 ACKNOWLEDGMENTS I wish to record my thanks to Mr. P. M. Benoit for his helpful comments and for obtaining much living material for me; to Mr. M. Morris, Llandudno, and Mr. Walter Scott, Scalloway, who also sent me living material of Mimulus species and hybrids. Thanks are also due to Dr. H. J. M. Bowen, Miss U. K. Duncan, Miss Vera Gordon, Mr. R. C. Palmer and Dr. D. P. Young; to the Keepers of the herbaria at the National Museum of Wales, Cardiff, and at the Botany Dept., University College of Wales, Bangor; and to the authorities of the British Museum (Natural History) for assistance in checking some references. Lastly, to my daughter, Anne Roberts, whose insistence on obtaining a specimen of the Harlech hybrid first aroused my interest in it. REFERENCES Bamey, L. H. (1927). The Standard Cyclopedia of Horticulture, New York. Hype, H. A. & WApg, A. E. (1957). Welsh Flowering Plants, ed. 2. Revised by A. E. Wade. Cardiff. Loupon, J. C. (1872). Encyclopaedia of Plants. London. McCLiinTock, D. & Fitter, R. S. R. (1956). The Pocket Guide to Wild Flowers. London. Supplement by D. McClintock (1957). VICKERY, R. J. (1959). Barriers to Gene Exchange within Mimulus guttatus (Scrophulariaceae). Evolution 13, 300-310. WARBURG, E. F. (1962), in Clapham, A. R., Tutin, T. G. & Warburg, E. F. Flora of the British Isles, ed. 2. Cambridge. Watsonia 6 (1), 1964. STUDIES ON ALCHEMILLA FILICAULIS BUS. SENSU LATO AND A. MINIMA WALTERS. Ul. ALCHEMILLA MINIMA By MARGARET E. BRADSHAW Department of Extra-Mural Studies, University of Durham ABSTRACT By transplant experiments and by growing seed progenies, it has been shown that A. minima differs from the dwarf ecogenodemes of A. filicaulis in several morphological characters, in habit and in ecological requirements. The chromosome numbers of both are within the same range, 2n = c. 103-108. A. minima occurs on Whernside and Ingleborough in north England and is endemic. The dwarf ecogenodemes of A. filicaulis are known from the north Pennines only and may also be endemic. INTRODUCTION The third taxon in this critical group within the A. vulgaris aggregate is A. minima Walters. This small plant is very similar to the dwarf ecogenodemes of A. filicaulis, sensu lato, found in close-grazed montane grassland in the north Pennine hills (Bradshaw 1963a, p. 316). Leaves from field collections of both are shown in Fig. 1. The plants are difficult to separate in the field and it was hoped that the diagnostic characters could be better established by cultivation experiments. Also, A. minima was known only from Ingleborough, yet dwarf A. filicaulis plants are plentiful in other hill areas. Cultivation experiments were needed to show if any of these plants were referable to A. minima. MORPHOLOGY AND HABIT The morphological characters of A. minima are most clearly appreciated when com- pared with the dwarf ecogenodeme of A. filicaulis. Transplants of both from the field were grown in pots and boxes (as described by Bradshaw, 1963a, p. 308) for one year under similar conditions in the Botanic Garden at Durham. Some differences in the general appearance can be seen in Plate 2 and measurable differences are given in Tables 1 and 2. TABLE 1. Measurements of plants cultivated in pots A, filicaulis Characters A. minima dwarf ecogenodeme (25 plants) (29 plants) 1. Petiole length: lamina length 0-8 + 0-03 1:16 + 0-03 2. Lobe length: lamina length 0-51 + 0-01 0:4 + 0-009 3. Depth of incision between lobes: lamina length 0-21 + 0-02 0-14 + 0-006 (a) Rhizome and habit The rhizomes are more slender in A. minima and longer because of the longer inter- nodes. In A. minima the leaves are more or less prostrate giving a plant a few centimetres high; in A. filicaulis the leaves are more erect, so that the length of the petiole gives an approximate measure of the height of the plant (Table 2). 76 Watsonia 6 (1), 1964. PLAINS, 2 i z : i 550.289 Plants from closely grazed grassland after 9 months cultivation. Top row: A. minima from Ingleborough area. Bottom row: A. filicaulis from Mickle Fell and Moor House National Nature Reserve. [ facing page 76 Wi ALCHEMILLA MINIMA ‘(T X) 198 yded UI JOPIO SUIS OY} Ul PosuRIIE dv SdARd] OY], “UOTWVANINS JO\Je pue s1OJaQ—soaIdeds Yoee JO us}j—syuRId AJUIM} WOT Pd}d9[[Od SoAvd] aIZUIS “| “SI4 UONCAT[NS Ide synpoyyf “V UOTPATI[ND I9e UOT}D9][09 PIey UOI}D9]]OO pPyoy DUIIUIL “YY DUIUNU "PY snooty Watsonia 6 (1), 1964. 78 MARGARET E. BRADSHAW (b) Stipules and base of the petioles In A. minima these are usually without the wine-red pigment usually found in A. filicaulis. TABLE 2. Measurements of plants cultivated in boxes. A. filicaulis Characters | A. minima dwarf ecogenodeme (10 plants) (3 x 10 plants) 1. No. of shoots 4-7 + 0-45 5-7 + 0:45 | 71-3 = 0-44 | 6-4 + 0:67 2. Length of petiole (cm) 1:9+ 0-17 10:0 + 0-40 | 8-7 + 0-40 9-2 =e? 3. Breadth of leaf (cm) 2:3 +0-18 5-1 + 0-10 | 4-7 Se 4-4+4 0-21 4. No. of inflorescences 1577222722 3-8 + 0°39 1:9 + 0-47 3-2.== 0°66 (c) Leaf lamina In A. minima the leaves are usually blue-green, and the lamina of mature leaves tends to be recurved and rather stiff with lobes which are V-shaped in cross section. A. filicaulis is less blue, often pure green, and the lobes are more open and the lamina flatter and not stiff. These differences can be seen in Fig. 2. The leaves of A. minima are much smaller with five full lobes, while in A. filicaulis another pair of basal lobes is more frequently well developed. The longer lobes of A. minima have 9(11) teeth on the median lobe while A. filicaulis has 11(13). (d) Indumentum The indumentum of the upper and lower surfaces of the leaf lamina were scored separately in five classes, one being almost glabrous, five very hairy. The flower-stems were scored in seven grades, six being least hairy and 0 most hairy (Bradshaw 1963a, p. 306). The hair characters of A. filicaulis have been considered in the first paper of this series. Dwarf ecogenodemes of subsp. filicaulis and subsp. vestita are known, and some of the former were almost glabrous when collected in the field. The leaves of A. minima were almost all class I on the upper surface and in only one case were the leaves more hairy (classes 3 and 4 respectively). An analysis of the hairiness of the flower-stem comparable with that of A. filicaulis was difficult because of the smaller number of internodes in A. minima, four or five against six or more. The flower-stems of 25 plants of A. minima were examined. In all cases the first internode was glabrous (this is often so in A. filicaulis) and frequently the second also. The whole flower-stem in ten plants was grade 6, the same as subsp. filicaulis; those of the other plants were hairy throughout, 12 were placed in grades 0-3, and three in grades 4-5. Clearly there is variation in the indumentum similar to that in A. filicaulis. CHROMOSOME NUMBERS Chromosome counts were obtained of pollen-mother-cells of two plants of A. minima from different localities on Ingleborough (c. 34/750747). These were as follows: 2n = c. 105 and c. 108 for one plant, and for the other 2n = c. 103, c. 98 and c. 99 (the last two values Watsonia 6 (1), 1964. ALCHEMILLA MINIMA 79 TABLE 3. Associates of Alchemilla minima. The figures indicate cover/abundance on the Domin scale in quadrats 25 cm x 25 cm in area. X = present in the same community near the quadrat. + = present in the quadrat. Localities 1 Ingleborough, 34/714733, 950 ft, W aspect 2-7 Ingleborough, c. 34/721753, 800 ft, NNW aspect 8-9 Whernside, c. 34/734798, c. 1,600 ft, SE aspect 10-14 Ingleborough, c. 34/737745, c. 2,000 ft, S and W aspect — =) = bod, —_ N = Ww — _ Localities 1 D, 3 4 5 6 7] 8 9 Alchemilla glabra = 1 A. minima a 1 A. xanthochlora - — Xe 1 1 —= pe | —_ —_ | WwW N | wW | Ww | Wo x WwW | Achillea millefolium - Bellis perennis ate Cardamine pratensis - Carex caryophyllea =e NWN 1 re i i ee ce | U2 1 N WN AN ! WN he | C. panicea Campanula rotundifolia = = Cerastium holosteoides | —- ~ 1 ~ 1 - ~ 1 Erophila verna — = = = Euphrasia officinalis + 1 - - 1 1 1 2, Galium sterneri G. verum Leontodon autumnalis Lotus corniculatus Luzula campestris Plantago lanceolata Potentilla erecta P. sterilis Poterium sanguisorba Prunella vulgaris Ranunculus acris R. bulbosus R. repens Rumex acetosa Sagina procumbens S. nodosa Taraxacum spp. - Thymus drucei ap Trifolium repens Sa 1 2 Veronica chamaedrys ~ Viola riviniana - Agrostis tenuis 4 Anthoxanthum odoratum| — Briza media 3 Cynosurus cristatus 3 3 5) 177) ~ | Nn | —_ NS | | | Re | NW ] NWN |! Ne | —— NN |] — ba | wN | | | | | ] ] | WSs istea ste ie tees! N | N | Ww | | | | Ne | We | he | b+++i4¢4+14+41 I 1S) | N | — | NY ! | eee | N * | mw il | 1 N — | +li++i ti | | | | j— | | | N | | | | N | — | | ay | | | | =e | — pa | | | | | | lel Nv | NN | ] | | | me N ! Ww | | WwW | = | Woe | - | uw ~ | whe | wp | | be | Nn | Deschampsia cespitosa Festuca ovina Helictotrichon pratense ~ Koeleria cristata — ee, | | | 1 Nw | 1 COoNNDYW 1 ~ nN | 1 oN | — - | + Noof | |] Oe | | |] cof WwW | N | | | N | | | Additional species occurring in one quadrat only: Alchemilla glaucescens (5), A. filicaulis subsp. vestita (1), Hypochaeris radicata (1). Juncus effusus (14). Leontodon hispidus (1), Veronica officinalis (6), V. serpylli- folia (14), Holcus lanatus (7), Nardus stricta (4), Sieglingia decumbens (3). Watsonia 6 (1), 1964. 80 MARGARET E. BRADSHAW are probably too low). These are within the range for A. filicaulis (2n = 101-110) and near the values obtained for a dwarf montane subsp. vestita plant from Mickle Fell (35/809244), which were 2n = c. 105 and c. 109 (Bradshaw, 1963b, p. 323). DISTRIBUTION So far A, minima has been recorded from two mountains only, Ingleborough and Whernside (34/77 and 78) in the Craven area. All the dwarf plants collected from other montane pasture areas became larger in cultivation and belong to A. filicaulis (Bradshaw 1963a, p. 316). Walters (1949) first recorded A. minima from the east side of Ingleborough at about 1,000 ft (310 m) altitude, and around the 2,000 ft contour on Simon Fell. Now, A. minima has been found in suitable habitats down to 750 ft (230 m) on the north west slopes of Ingleborough. Recently, more plants were found on Whernside at 700 ft (220 m) and 1,000 ft and between 1,600 ft (500 m) and 1,800 ft (560 m). In all the localities A. minima plants were plentiful. ECOLOGY A. minima appears to be confined to the intensively grazed Festuca-Agrostis grassland areas on or just below the limestone strata of the Carboniferous Limestone Series of the Ingleborough district. All habitats were characterized by the high moisture content of the soil. At the lower altitudes (700 ft—c. 1,400 ft) this is maintained by seepage from flushes below the limestone; above this (c. 2,000 ft) presumably the higher humidity and rainfall are sufficient since A. minima grows amongst the limestone boulders where surface water is absent. In all cases the soil was a moist, slightly sticky, dark-brown peaty loam. the lower and wetter habitats were less acidic (pH 6-5-7-3) than those at c. 2,000 ft (pH 5-2-6-1). No doubt, this is due to the more base-rich flushing by water from the limestone. The associated species are listed in Table 3. Festuca ovina was present in all the sample areas and Agrostis tenuis, Cynosurus cristatus, Trifolium repens, Prunella vulgaris, Plantago lanceolata and Euphrasia officinalis agg. were present in most. The characteristic blue-green appearance of the lower habitats is caused by the high frequency of Carex panicea; Leont- odon autumnalis is a constant species and the only records of Poterium sanguisorba are at these lower altitudes. Carex panicea is absent from the higher altitudes but several other species are more common there: these are Achillea millefolium, Bellis perennis, Cerastium holosteoides, Sagina procumbens and Thymus drucei. CONCLUSIONS This work fully supports Walters’ (1949) original treatment of A. minima as a distinct species. The cultivation experiments have emphasized the diagnostic characters which he described, and revealed some other contrasts with the dwarf plants of A. filicaulis; these are set out in Table 4. Although typical plants can be identified with a fairly high degree of confidence, plants are found which are less distinct (cf. Fig. 1) and can be determined with certainty only after cultivation. Details of the ecological preferences and associated species of A. minima have been given. Apparently suitable habitats in closely grazed Festuca-Agrostis grassland are abun- dant in the Pennines and may be found elsewhere. Only a relatively small number of these have been examined, and the recent discoveries of A. minima some miles from the early localities on Ingleborough suggests that it may be more widespread, at least in the Craven district. Some similar habitats in the north Pennines are occupied by the dwarf ecogenodeme of A. filicaulis. Curiously, no plants of A. filicaulis were collected from the populations on Ingleborough although A. filicaulis does occur in other habitats in the area. This raises interesting points about the evolution and history of these two species. These will be con- sidered in relation to other members of the A. vulgaris aggregate in the fourth paper of this series. Watsonia 6 (1), 1964. ALCHEMILLA MINIMA TABLE 4 Character A. minima dwarf A. filicaulis Rhizome long and slender, little branched, | short and thicker, often much Base of plant and stipules Petioles Lamina Lobes Teeth on median lobe Flower-stems Indumentum of whole plant BrApDsHAw, M. E. (1963a). Studies on Alchemilla filicaulis Bus., sensu lato, and A. minima Walters. I. internodes long not wine-red usually about length of leaf lamina usually blue-green; lamina pleated and recurved (lobes half open, V-shaped in cross section) usually 5, 4 radius of lamina, deep incision between lobes usually 9 several, very slender, usually with 5 internodes variable, usually less hairy than A. filicaulis subsp. vestita REFERENCES branched, internodes short wine-red, sometimes rather faint longer than leaf lamina not so blue, usually pure green; whole lamina flat usually 7, less than 4 radius of lamina, shallow incision be- between lobes usually 11 few, less slender, usually with 6 internodes variable, subsp. filicaulis, subsp. vestita and intermediate states Morphological variation in A. filicaulis, sensu lato. Watsonia 5, 304--320. BrapsHAw, M. E. (19635). Studies on Alchemilla filicaulis Bus., sensu lato, and A. minima Walters. II. Cytology of A. filicaulis, sensu lato. Watsonia 5, 321-326. WALTERS, S. M. (1949). Alchemilla vulgaris L. agg. in Britain. Watsonia 1, 6-18. Watsonia 6 (1), 1964. BOOK REVIEWS Flora of the British Isles. A. R. Clapham, T. G. Tutin and E. F. Warburg. J/lustrations, Part III, Bora- ginaceae—Compositae. Drawings by Sybil J. Roles. Pp. 115, 84 x 114 in. Cambridge University Press, 1963. Price 32s. 6d. Miss S. J. Roles’s illustrations to ‘CTW’ are by now familiar to British botanists. This is the third of four parts, covering the remaining dicotyledonous families. The drawings, arranged four to a page, are designed as a companion to the descriptions in the Flora. * . . . The intention is to provide a visual impres- sion of the habit, and a selection of the chief features of the plants . . . ’ Almost all have been made from fresh specimens. These drawings are less detailed than either Miss Ross-Craig’s much more ambitious series, or the illus- trations to R. W. Butcher’s New Illustrated British Flora; indeed, from their stated intention, they cannot fairly be compared with either. A fairer comparison is with Fitch & Smith’s illustrations for Bentham & Hooker’s flora. The policy of drawing from living specimens has resulted in a freshness and naturalness lacking from many of Fitch’s drawings. It has its pitfalls: several of the plants look wilted (e.g. Asperugo, Galium debile) and in other cases the specimen chosen seems poor or even quite untypical (e.g. Cymbalaria muralis, Linaria arenaria). In trying to portray habit Miss Roles has attempted a difficult task. Some of the drawings are notably successful; the Pinguicula spp., Origanum, most of the Lamium spp., the Plantago spp., Viburnum opulus, Sonchus asper, Chrysanthemum segetum and many others are instantly recognizable. In some other cases the habit is not well caught (e.g. Galium saxatile, Stachys sylvatica, Bartsia alpina, Dipsacus fullonum), and sometimes the distinguishing characters of related species are not well brought out (e.g. Leontodon hispidus and L. taraxacoides, some Euphrasia spp.), perhaps most often through a combina- tion of unselectiveness and understatement. To capture the character of a plant at this scale may demand an approach akin to caricature; one is often driven to admire the effectiveness and economy of Fitch’s little sketches carried out in a much less flexible medium. A set of illustrations like these is inevitably a compromise, in which clarity and detail (and to some extent accuracy) are set against portability and cost. Are the illustrations to be used to assist identification in the field, or are they to be studied at leisure at home or in the herbarium? It is possible to have the worst of both worlds, and one wonders if the right compromise has been chosen here. For anyone who simply wants help in identification there is much to be said for a book like Fitch’s illustrations, or Thommen’s useful Taschenatlas der Schweizer Flora, which crams over 3,000 thumb-nail sketches of Swiss plants into a single cheap pocketable volume. Nevertheless, Miss Roles’s drawings will be found an invaluable companion with a vasculum of plants at the end of the day, and they will give much pleasure besides. The drawings themselves are essentially accurate, and their delicate style is admirably set off by good paper, and excellent typography and layout. This is altogether an attractive volume. M. C. F. PROCTOR North Atlantic Biota and their History. Edited by Askell Love and Doris Love. Pp. xii + 430, with 13 plates and about 180 figs. and tables in the text. Pergamon Press, Oxford. 1963. Price £5 net. This is the report of a symposium held at the University of Iceland, Reykjavik, under the auspices of the University of Iceland and the Museum of Natural History, in July 1962. The book contains an introduction followed by 27 papers read at the symposium. More than half of these are devoted to studies of higher plants, and the contributors include such well-known workers as T. W. Bocher, E. Dahl, K. Faegri, E. Hulten, Askell and Doris Love, J. A. Nannfeldt and R. Nordhagen. The papers cover a variety of topics but the main theme is the study of species, seemingly identical, which occur on both sides of the Atlantic Ocean. Considerable attention is paid, also, to the geological inter- pretations of the north Atlantic climates of the past. Professor Hultén’s contribution on phytogeographical connections of the north Atlantic is based on a study of distribution maps of all plants known to occur in Iceland, Greenland, Spitzbergen, Jan Mayen. Bear Island and the Faeroe Islands. These are arranged into 24 groups, e.g. Circumpolar, or near circum- polar species which do not occur in Greenland or central Europe; circumpolar plants occurring in Green- land but lacking in central Europe, etc. An analysis of this data leads to the conclusion that phytogeo- graphical conditions around the north Atlantic give poor support to the hypothesis that a land-bridge could have existed in Quaternary or late Tertiary times. 82 Watsonia 6 (1), 1964. ee ere a ee ee gS ay tee SE a” Oe a oe ee eS es ve BOOK REVIEWS 83 In his paper, Professor Nannfeldt endeavours to illustrate the closer floristic connections of the Scandes with Iceland and Scotland than with the Alps. DrE. Dahl discusses plant migrations across the north Atlantic Ocean including plants transported across the ocean by man. The western amphi-atlantic element in Europe and the eastern amphi-atlantic element in America, the long-distance dispersal hypothesis and the theory of a land connection between Europe and America are also reviewed. Doris Love contributes a fascinating paper on the dispersal and survival of plants. Very few species are really adapted for long-distance dispersaJ. Such mammals as polar bears and foxes subsist mainly on animal food and are thus excluded as possible plant dispersers. Birds travel freely between land masses and are active as transporters for both ingested and externally carried seeds and plant fragments—ducks, geese and swans in particular are mentioned. A table gives interesting data on the various species of birds and their food plants, together with geographical information on nesting and winter areas. It is concluded, however, that there is a relatively small chance for any large numbers of plants to have been dispersed by birds between North America and the British Isles, and the hypothesis that Sisyrinchium and Eriocaulon were introduced into Ireland from North America attached to the bodies of pink-footed geese is disputed on the grounds that the American and Irish plants are quite distinct species. Dispersal by wind, and by sea currents are also discussed. The author’s conclusion is that the flora of the north Atlantic area is old, well established, and mostly relict from a time preceding the Pleistocene Ice Age. O. Gjaerevoll in dealing with the survival of plants in nunataks in Norway during the Pleistocene glaciation refers to the strange distribution of Artemisia norvegica—the Urals, southern Norway and a single locality in Scotland. Other Arctic species selected for discussion include Papaver radicatum subsp. subglobosum, Arenaria humifusa, Draba crassifolia, Saxifraga aizoon, Potentilla hypoarctica and Carex scirpoidea. This unusual concentration of rare species in southern Norway is used to illustrate the argument that there must have been quite extensive ice-free areas in that part of the country. E. Einarsson reviews the elements and the affinities of the Icelandic flora. Between 430 and 440 species of vascular plants are considered to be native to Iceland, about 97% of these are found in Scandinavia, and c. 85% are found also in the British Isles. It is concluded that more than half of the Icelandic species have a Boreal distribution, whereas Arctic-Alpine species comprise of c. 33% of the flora. There are many other interesting papers dealing not only with higher plants, but with the maximum extent of the Pleistocene glaciation, palynology and geology. The entire work is of a very high standard and is well printed on excellent paper. It is an essential work of reference to botanists interested in the phytogeography of the north Atlantic area. D. H. KENT Drawings of British Plants. Part XVIII, Compositae (4). Stella Ross-Craig. 76 plates. G. Beil & Sons Ltd., London, 28 March 1963. Price 10s. 6d. Part eighteen of Stella Ross-Craig’s series of drawings deals with the remaining ten genera of the family Compositae. Owing to their asexual means of reproduction the species of two of these genera, Hieracium and Taraxacum, are notoriously difficult taxonomically. In her introduction to Hieracium Miss Ross-Craig does not mention apomixis but writes: ‘The variation within the species of Hieracium can sometimes be so great that I think many of the specific names which have been proposed for the variants could quite happily be forgotten.’ It is unfortunate that the authoress did not seek advice from a specialist in the most difficult group. As it is, her selection of species, her identification of the species chosen, and her nomenclature and citation of authors all leave very much to be desired. In the latest list of British Hieracia (Sell & West in Dandy, 1958), 234 species in 17 sections are listed. That list follows closely the Monograph by Pugsley (1948), the cited specimens in which Miss Ross-Craig has used for her illustrations. In the book under review, 20 species have been selected to illustrate the genus. The section Alpina, which is mainly confined to the mountains of Scotland, is represented by four species, the section Vulgata subsect. Eu-Vulgata and the section Sabauda by three species each, and nine other sections by one species each. The very large groups of the section Vulgata, namely subsections Bifida, Glandulosa, Sagittata and Caesia, in which are placed a very large percentage of the British Hieracia, are tTepresented by only one species of the subsection Glandulosa. No species at all are given for the sections Oreadea, Alpestria, Auriculina, Collinia and Praealtina. The three latter groups contain only introduced species, but the widespread H. brunneocroceum could lay much greater claim to inclusion than the very rare introduced H. amplexicaule of the section Amplexicaulia. 1 can make nothing of the order in which the species are arranged, the oddest thing being the putting of H. amplexicaule between two species of the section Vulgata. A close examination of the individual plates reveals several points of criticism, some general and some detailed. The relative sizes of the capitula seem to give the wrong impression, as anyone can clearly observe if Watsonia 6 (1), 1964. 34 BOOK REVIEWS they compare those of H. calenduliflorum, lachenalii and pellucidum (? exotericum) with the actual plants. Colours are notoriously difficult to describe, but those given for the leaves and ligules rarely fit in with my idea of the living plant. In H. holosericeum, lingulatum and anglicum the tips of the ligules are usually ciliate. In H. holosericeum the parallel-sided involucral bracts are not emphasized. The illustration of H. pellucidum seems to be a better fit for one of the exotericum aggregate. That of H. vulgatum may be of that species but it is a very uncharacteristic example of a widespread and easily recognizable plant. The leaves are too deeply dissected and are certainly never glaucous. In the plate of H. lachenalii, Fig. F. seems to be an entirely different plant (? species) from the rest of the figures which are characteristic of the plant they are meant to represent. The example of H. tridentatum is what Dr C. West and I would call H. calcaricola (F. J. Hanb.) Roffey. H. tridentatum Fries sensu stricto does not occur in the British Isles. Pugsley distinguished A. perpropinquum and H. bladonii on whether the eglandular or glandular hairs were dominant on the involucral bracts. All intermediates are possible and both extremes can be found on the same plant, and for this reason these species were lumped by Dr C. West and myself in Dandy’s Check List. These illustrations do not even show the two extremes and are clearly referable to the same taxon, the correct name for which is H. perpro- pinquum (Zahn) Druce. The genus Taraxacum has not yet received a similar treatment to that of Hieracium in this country, and the four species here shown represent the four main groups of microspecies recognized by Continental botanists. T. officinale and T. laevigatum are excellent illustrations of microspecies within two of these groups. That of 7. palustre is a composite one. The plant in the foreground probably belongs to the aggregate because of its broad involucral bracts, which are, however, only loosely appressed. The two leaves in the background are characteristic of T. palustre (Lyons) DC. sensu stricto. It would be interesting to know to which plant the dissected parts belong. T. spectabile seems from the involucral bracts to be correctly named, although the leaves appear to be those of a late-flowering, and therefore uncharacteristic, example. The species of the other genera are excellently illustrated and show clearly how this group of planis, which prove especially difficult for the beginner, can be identified much more easily by a careful examination of the dissected parts. In these species, also, the relative sizes of the capitula are much better shown, for ex- ample in the three species of Hypochaeris. The standard of these drawings is high and it is difficult to find fault with the artist; indeed, if the parts of the plants had not been so meticulously drawn the above comments could not have been made. P. D. SELL Watsonia 6 (1), 1964. A BIBLIOGRAPHICAL INDEX OF THE BRITISH FLORA COMPILED BY N. DOUGLAS SIMPSON, o.a., F.L.s. The purpose of this work is to provide references to sources of information whereby flowering plants, vascular cryptogams and Charophytes found in Britain may be identified, their history traced and their geographical range determined. In addition, information is provided on plantlore, local names, poisonous plants and weeds. The work has been in preparation for nearly twenty years and contains over 65,000 entries, including references to books, articles and manuscripts relating to the flora of the British Isles from the fifteenth century to the present time. Demy 4to, bound, 448 pages in double column. Limited Edition of 750 copies. PRICE £3 I5s. (postage extra) Obtainable from the compiler at ‘Maesbury’, 3 Cavendish Road, Bournemouth, Hants. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANpby for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf of a sub-committee of the Society appointed for this purpose. Sir George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and have become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8vo, 176 +-xvi pages. Bound in cloth. Price: 10/- (postage extra). Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. Joseph Dalton HOOKER W. B. TURRILL The life, travels and work of the great Victorian botanist who was Director of the Royal Botanical Gardens, Kew, from 1865 to 1885, and who first applied Darwin’s theories of evolution by natural selection to the plant kingdom. Dr. Turrill, a past Keeper of the Herbarium s his contributions to botany, and some new Afacts about his achievements emerge. and Library at Kew, has drawn from Hooker’s own journals for this authorita- tive and informative biography. ‘An excellent study of a great botanist, and a well-balanced assessment of his scientific researches. —Times Educational * Supplement. ‘A first-class account of Hooker’s life and work.’—Discovery. NELSON IRISH NATURALISTS’ Charles DARWIN Sir GAVIN de BEER One of the greatest living authorities on Darwin and the theory of evolution by inatural selection gives a strikingly clear picture of the way in which Darwin’s f brilliant mind gradually moved to his revolutionary theories. Every aspect of Darwin’s varied work is studied, including ‘A valuable contribution to the history fof science and will certainly help to | illuminate evolutionary thinking today.’— Sir Julian Huxley. ‘Readable, fresh and scholarly. It should /immediately take its place as the standard # work on the subject.’— Nature. 16 half-tone plates. 21s 16 half-tone plates. 21s JOURNAL A Magazine of Natural History Published Every Quarter by the I.N.J. Committee Edited by Miss M. P. H. KERTLAND, M.Sc. with the assistance of Sectional Editors Annual Subscription, 10/- post free Single Parts, 3/6 All communications to be addressed to :— The Editor, Department of Botany, Queen’s University, Belfast INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the taxonomy or distribution of British vascular plants are invited from both members of the Society and others. They should be typewritten, with wide margins, double spaced, on one side of the paper only; contributors are recommended to keep a carbon copy of their typescripts. The form adopted in recent parts of Watsonia should be used for layout, headings, citations and references. Contributors are urged to avoid very complicated hierarchies of headings and sub-headings, and to check carefully the consistency of those that they use. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British flowering plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (N.H.) and B.S.B.I., 1958), and may then be cited without authorities. Otherwise, authors of names must be cited, at least on the first occasion where they appear in the text. Except for citations of the place of publication of plant names, full references should be listed in alphabetical order of authors’ names at the end of the paper. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals, 3rd ed. (1952). References to herbaria should include the abbreviations given in British Herbaria | (BS.B.I., 1958) and Index Herbariorum 3rd ed. (1.A.P.T., 1956). Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new taxa, and the main conclusions. Line-drawings should be boldly drawn in Indian ink on Bristol board or similar smooth white card, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. Graphs can be reproduced from originals on graph paper with faint b/ue ruling (not grey or any other colour), but all lines to appear on the finished block must be inserted in Indian ink. Lettering on line-drawings and graphs should be inserted lightly in pencil, and will be finished in uniform style. If an illustration includes plant-names or place-names, it is advisable to type these clearly on a separate sheet of paper. Photographs can only be accepted where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable for the necessary reduction. It should be remembered that the finest detail on the originals may be lost even on the best half-tone blocks. If in doubt about the citation of names or references, or the presentation of illus- | trations or tabular matter, contributors are advised to consult the editor before submitting their typescripts. Authors will normally receive both galley-proofs and page-proofs. Particular care should be given | to the thorough checking of references and tables. Alterations in page-proof should be avoided as far as possible, and any words or phrases deleted replaced by others of equal length. Twenty-five separates are given free to authors of papers. Further copies may be obtained at the printer’s current price, and must be ordered when the proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to the Editor, Dr. M. C. F. Proctor, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. PROCEEDINGS OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Published Twice Yearly Price 10/- per part - Obtainable from E. B. BANGERTER, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. A DARWIN CENTENARY Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the very successful Conference arranged by the Botanical Society of the British Isles in 1959. The contributors include Lady Barlow, G. D. H. Bell, Prof. T. W. Bocher, Prof. W. H. Camp, Prof. N. Hylander, Prof. I. H. Manton, Prof. T. G. Tutin and Prof. D. H. Valentine. Demy 8vo, 140 pages, with 7 half-tone plates and 11 maps and figures in the text. Bound in buckram. LOCAL FLORAS Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read _ at the Conference held by the Botanical Society of the British Isles in 1961. The — contributors include D. E. Allen, Dr. H. J. M. Bowen, Dr. J. G. Dony, J. S. L. Gilmour, Prof. J. G. Hawkes, D. H. Kent, J. E. Lousley, D. McClintock, R. D. Meikle, Dr. F. H. Perring, R. C. Readett and C. C. Townsend. Demy 8vo, 120 pages, with 2 half-tone plates, and figures in the text. Bound in buckram. PRICE £1 (postage extra) THE CONSERVATION OF THE BRITISH FLORA Edited by E. MILNE-REDHEAD The Report of an important and most successful Conference held by the Botanical © Society of the British Isles at Durham in April 1963. This report includes most of the papers read at the Conference, together with a verbatim account of the discussions | which followed. Demy 8vo, 80 pages. Bound in buckram. PRICE 15/- (postage extra) The above items are obtainable from E. B. Bangerter, c/o Dept. of Botany, British | . Museum (Natural History), Cromwell Road, London, S.W.7. Printed by Headley Brothers Ltd 109 Kingsway London WC2 and Ashford Kent Ws 4 Ber. WATSONIA JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 6 MAY, 1965 Pt 2 CONTENTS A REVISION OF THE BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA. [Fries] F. N. WiLuiAMs. By P. D. SELL and C. West... .. a 85-105 FRUIT VARIATION IN POLYGONUM PERSICARIA L. By J. TIMSON ... of 106-108 LEAF MEASUREMENTS AND EPIDERMIS IN POA ANGUSTIFOLIA. By D. M. BARLING re oe ae ar a as AG a a 109-113 SULPHUR AND THE DISTRIBUTION OF BRITISH PLANTS. By H. J. M. BOWEN .. 114-119 THE ASSESSMENT OF THE TAXONOMIC STATUS OF MIXED OAK (QUERCUS spp.) POPULATIONS. By A. CARLISLE and A. H. F. BROWN .. ay = 120-127 DACTYLORHIZA NEVSKI, THE CORRECT GENERIC NAME OF THE DACTYLORCHIDS, By P. F. Hunt and V. S. SUMMERHAYES ae ie ae He 128-133 BOOK REVIEWS ie ae Ae ae as sigs aa es oe 134-139 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BoTANY, British Museum (NATURAL History), LONDON PRICE: TWENTY-FIVE SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES OFFICERS President: Dr. E. F. WARBURG Vice-Presidents: Dr. S. M. WALTERS, Dr. C. E. HUBBARD, Dr. J. G. DONY, and R. MACKECHNIE. Hon. General Secretary: E. B. BANGERTER, c/o Dept. of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. Hon. Treasurer: J. C. GARDINER, Thrift House, 12 & 14 Wigmore Street, London, W.1. Hon. Editors: (Watsonia) Dr. M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) D. H. KENT. Hon. Meetings Secretary: Mrs. MARY BRIGGS, White Cottage, Slinfold, Sussex. Hon. Field Secretary: P. C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Dartford, Kent. Hon. Membership Secretary: Mrs. J. G. Dony, 9 Stanton Road, Luton, Bedfordshire. Applications for membership should be addressed to the Hon. Membership Secretary, from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 25/- per part. To be obtained from D. H. Kent, 75 Adelaide Road, West Ealing, London, W.13. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/— for Junior Members, and 10/— for Family Members, and the Society’s year runs from January Ist to December 3ist. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to — 4 D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. A REVISION OF THE BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA [FRIES] F. N. WILLIAMS By P. D. SELL and C. WEST The Botany School, Cambridge For the following account we have attempted to examine all the available herbarium material of the British species of the Section Alpestria and to consult all the relevant literature. We are grateful to the authorities in charge of the British Museum (Natural History) (BM), Royal Botanic Garden, Edinburgh (E), Royal Botanic Gardens, Kew (K), University College Leicester, University of Liverpool (LIVU), Marlborough College (MACO), University of Oxford (OXF), University of Sheffield and the South London Botanical Institute (SLB) for the loan of specimens from their herbaria; to C. J. Cadbury, A. G. Kenneth, J. E. Lousley, R. C. Palmer, Miss C. M. Rob, W. Scott and N. D. Simpson for loans from their private herbaria and for information about the plants in situ; to Dr. Carl E. Alm of Uppsala for information on the type material of H. dovrense; to Erik Almquist of Uppsala for general information on the section in Scandinavia; and to N. Y. Sandwith for checking the latin descriptions. We are grateful also to the authorities of the British Museum (Natural History) for the photograph of the specimen of H. hethlandiae which is in their herbarium, to F. T. N. Elbourn and Miss J. Chenery for taking the re- mainder of the photographs, which are of specimens in CGE, and especially to Professor H. Godwin who has made available a grant from the ‘Brooks Fund’ towards the cost of their publication. Section ALPESTRIA [Fries] F. N. Williams (1902) 82; F. J. Hanb. in Bab. (1904) 263; W. R. Linton (1905) 71; Pugsley (1948) 219 (includ. Series Dovrensia, Series Protracta pro parte exclud. type, Series Subfoliosa and Series Spectabilia); A. R. Clapham in A. R. Clapham, T. G. Tutin and E. F. Warburg (1952) 1144; ed. 2 (1962) 907 (Lectotype species H. dovrense Fries). Hieracium—Archhieracium—Pulmonaria—A lpestria Fries (1862) 102. Section Prenanthoidea group species H. carpathicum Besser in Zahn (1921) 809. Plants without basal leaves, with basal leaves withered at the time of flowering, or rarely with basal leaves at the time of flowering. Cauline leaves usually fairly numerous, 2-19, rarely thickened at the margins, at least the upper--rounded at the base and semi- amplexicaul. Inflorescence cymose-paniculate. Capitulas dark, with broad, --obtuse phyllaries, which are usually almost without stellate hairs. Cypselas dark. Flowering in July and August. The species of the Section Alpestria are most closely allied to those of the Sections Prenanthoidea and Foliosa. They can be distinguished from the former by their darker cypselas and usually fewer leaves, and from the latter by their fewer cauline leaves, longer internodes and by the leaves having less thickened margins. About 200 species have been described in this Section, five from Greenland, 22 from Iceland, three from the Faeroes, 18 from the British Isles (one of which also occurs in the Tatra Mountains) and the remainder from Scandinavia. At least another five species from the mountains of Central Europe should also be included in this Section, but in that region the distinction between the species of the Section A/pestria and the Section Prenanthoidea is not clear. In the British Isles all except five of the eighteen species are confined to the Shetland Isles, where they are endemic. Of the remainder, H. dovrense is locally abundant in Scan- dinavia and occurs in one isolated locality in Sutherland; H. carpathicum is locally dis- tributed in the Tatra Mountains and also occurs in a few isolated localities in Perth and 85 Watsonia 6 (2), 1965. 1 86 Pp. DD, SELL and7@. WEST Forfar; H.dewarii is an endemic species, rather widespread in central and southern Scotland; H. solum is endemic to a single locality in Kintyre; wee H. mirandum is endemic to a single locality in mid-west Yorkshire. In the Shetlands they are, with one exception, restricted to coastal localities in- accessible to sheep. From the field experience of several botanists the one exception, H. vinicaule, occurs, sometimes in quantity, in places easily reached by grazing sheep. This probably accounts for the fact that this species is the most widespread of the Section Alpestria in the Shetland Isles. In Scandinavia they are apparently widespread and often abundant in grassland and under trees in mountainous regions. Pugsley (1948) divided the British species of the Section Alpestria into four Series, all of which are nomina nuda. These Series are almost entirely artificial, and we fail to under- stand why the species were so grouped. In our opinion no useful purpose is served by subdividing the Section, although some species are much more closely allied than others. H. attenuatifolium, dilectum, hethlandiae, northroense, pugsleyi, subtruncatum and vinicaule have many characters in common. H. carpathicum, dewarii and mirandum are also closely allied and are very distinct from the rest; in fact they closely approach the Section Foliosa. H. australius and dovrense have several characters in common, but the six remaining species, H. breve, gratum, praethulense, difficile, solum and zetlandicum have no close allies in the British Isles. Until recently the study of the Shetland species of the Section Alpestria was based mainly on the collections made by W. H. Beeby and by G. C. Druce. The definition of the species and varieties made by Beeby is clear and precise, and the identification of his specimens consistent. On the other hand, an examination of the specimens in Druce’s herbarium suggests that he did not understand the taxonomy of the group. Moreover, some of the species, according to the labels attached to the specimens, are recorded from localities where they have never been seen by other botanists. Also, Druce (1928) 71, no. 419: 1635d, published without a diagnosis and without annotating a specimen the name H. crocatum var. ronasii from *Zetland’, which presumably refers to a taxon in this Section. One cannot even guess to which taxon he is referring since all the taxa described at that time are covered by names in his List, and no new taxon from Ronas Voe (assuming he is describing a plant from that locality, i.e. ronasii) is represented in his herbarium. Pugsley (1948) based his account of the Section on that of Beeby; we have followed the same course. In the interests of the conservation of our native flora we should like to point out that the species of this Section, which are very local and in small numbers, should not be collected without good scientific reason. They are well represented in public herbaria, and anyone wishing to become familiar with them should examine the specimens to be found there. The collection at the University of Cambridge (CGE) is the most complete, every British species being represented. Here are deposited all the authors’ specimens, including material cultivated in the Botanic Garden at Cambridge. The descriptions in this paper are based on a more precise terminology than is usual, a key to which will be published at a later date; for example, short, medium and long, when applied to hairs, refer to a definite range of length. As the terminology is that used in general practice, the terms having been only more clearly defined, it has been decided not to hold up the paper until the definitions are published. In all the descriptions the most important characters have been put in italic so that they can be read off to give a precise diagnosis. KEY TO THE SPECIES 1. Leaves up to four times as long as broad 2 Leaves more than four times as long as broad 16 2. Ligules strongly pilose-tipped 3 Ligules glabrous 4 3. Phyllaries 7-9 mm. long, with few stellate hairs at base 17. dewarii Phyllaries 9-11 mm. long, with numerous stellate hairs especially on the margins 18. carpathicum Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 87 4. Cauline leaves up to five 5 Cauline leaves more than five 10 5. Pbyllaries 11-13 mm. long 3. solum Phyllaries 8-11 mm. long 6 6. Styles yellow when fresh 1. zetlandicum Styles discoloured i 7. Leaves less than 2:5 times as long as broad 2. hreve Leaves more than 2:5 times as long as broad 8 &. Leaves dentate; phyllaries densely hairy 7. dovrense Leaves entire or remotely denticulate; phyllaries sparingly hairy 9 9. Cauline leaves 2-4 (—-7), about three times as long as broad 4. gratum Cauline leaves 4-8, about four times as long as broad 5. difficile 10. Leaves never more than 18 mm. broad 9. hethlandiae Leaves more than 18 mm. broad 11 11. Styles yellow when fresh 112 Styles discoloured 13 12. Peduncles usually glabrous 8. praethulense Peduncles usually hairy 13. subtruncatum 13. Leaves not more than 2:5 times longer than broad 2. breve Leaves more than 2-5 (usually more than 3) times as long as broad 14 14. Leaves ovate, abruptly contracted at the base 16. mirandum Leaves elliptic, gradually narrowed at the base 15 15. Leaves dentate 7. dovrense Leaves finely denticulate or subentire 6. australius 16. Plant very hairy; phyllaries 1-5-2 mm. wide 6. australius Plant less hairy; phyllaries 1-1-5 mm. wide 5. difficile 17. Phyllaries with many unequal glandular and few simple hairs 18 Phyllaries with few glandular and without or with few to numerous simple hairs 20 18. Leaves under 20 mm. broad 10. attenuatifolium Leaves over 20 mm. broad 19 19. Plant usually under 35 cm. high; leaves usually closely spaced and regularly denticulate, abruptly decreasing in size half-way up the stem 14. northroense Plant usually over 35 cm. high; leaves usually less closely spaced and less denticulate, gradually decreasing in size up the stem 15. vinicaule 20. Phyllaries with numerous simple hairs 11. pugsleyi Phyllaries with few or no simple hairs 21 21. Styles yellow when fresh aD Styles discoloured 23 22. Leaves gradually narrowed at both ends, with an acute apex; peduncles glabrous or with an occasional simple hair 12. dilectum Leaves usually abruptly narrowed at base, with an obtuse to subacute apex; peduncles with few simple hairs 13. subtruncatum 23. Cauline leaves °—16 24 Cauline leaves 5—9 25 24. Leaves with long acute apex 12. dilectum Leaves with obtuse or subacute apex 14. northroense 25. Phyllaries 9-10 mm. long 5. difficile Phyllaries 10:-5—13 mm. long 26 26. Leaves usually broadest in the middle and gradually tapering at both ends 9. hethlandiae Leaves usually broadest below the middle and with a broader base 3. solum 1. HIERACIUM ZETLANDICUM Beeby (1891) 243 (Lectotype: Sand Voe, Northmaven, Shetlands, 4 July 1889, W. H. Beeby no. 1044 (BM)); (1892) 54; F. N. Williams (1902) 126 (exclud. Roeness Voe* et Sutherland loc.); Edmondston (1903) 71; F. J. Hanb. in Bab. (1904) 263 (exclud. Sutherland loc.); W. R. Linton (1905) 73; Beeby (1908) 113; Druce (1922) 499; (1925) 646; Roffey (1925) 29, no. 1199 (but one v.c. only (v.c.112)); * The modern official spelling is ‘Ronas Voe’. In this paper we have always given the spelling used by the author in question. SMT HST 4 Ae 2 i isTiTuTion ¥¥! Watsonia 6 (2), 1965. 88 P. D. SELL and C. WEST Druce (1928) 70, no. 201 (1619) (but one v.c. only (v.c. 112)); Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 222; P. D. Sell & C. West in Dandy (1958) 127, no. 558: sp 70, JeLAms, 321, : H. demissum subsp. zetlandicum (Beeby) Zahn (1921) 855. Phyllopodous, or rarely with basal leaves withering at the time of flowering. Stem to 40 (—50 cult.) cm. high, slender, flexuous, often purplish, with numerous, medium to long, soft simple hairs at the base, which become gradually fewer upwards, the uppermost sometimes dark-based, usually stellate-hairy throughout, but sometimes sparingly so or not at all near the base, some plants with an occasional, very short, blackish glandular hair near the inflorescence. Leaves deep green, often suffused with purple; the radical 3-6; primordial small, subrotund, finely denticulate; the later variable, usually broadly elliptic, obtuse-mucronate to acute, narrowed below, denticulate to shallowly but sharply serrate, or ovate, obtuse-mucronulate, cuneate-based and denticulate, or in large specimens nearly rhomboidal, acute, cuneate-based and irregularly serrate, the teeth interspersed with denticulations, all with numerous, unequal, short to long, subsetiform simple hairs on both surfaces and on the margins, petioles winged, with few to many long simple hairs; cauline 2-4 (-6 cult.); the lower elliptic to elliptic-lanceolate or sometimes ovate, acute (occasionally obtuse- mucronate), cuneate-based with a winged petiole, denticulate to sharply serrate; the median ovate to ovate- lanceolate, acute to acuminate, cuneate- or rounded-based and semiamplexicaul, denticulate (sometimes finely serrate); the upper linear-lanceolate, linear or bractlike; all cauline glabrous or with a few, medium simple hairs above, with medium simple hairs beneath and on the margins, sometimes with sparse stellate hairs on one or both surfaces. Inflorescence compactly cymose, with 1-7 (30 cult.), often geminate capitula; peduncles short to medium, suberect, with numerous stellate hairs, numerous short, black glandular hairs and a few, medium, usually dark-based simple hairs. Capitula cylindrical in bud. Phyllaries 8-9 (—10 cult.) mm. long, 1-2 mm. wide, incumbent in bud, linear-lanceolate, inner--acute, sometimes abruptly narrowed near the apex, pale green, sometimes with darker centre, outer shorter, less acute, blackish-green; all with scattered stellate hairs, numerous, black, unequal, very short to short glandular hairs and sometimes an occasional, pale, short, dark-based simple hair. Ligules glabrous-tipped, deep yellow. Styles yellow, becoming discoloured in dried specimens. Margins of the receptacle-pits sharply dentate. Cypsela dark, 3-5-4 mm. long. H. zetlondicum was originally discovered by W. H. Beeby on fine pasture above Sand Voe, Northmaven, in 1889. Two years later he found it in North Roe in the same parish. In 1908 he remarks that ‘it is confined to a tract near North Roe about two by one miles, extending from Burga Taing northwards to Benegarth, but plentiful in many places within this area’. Beeby’s records are represented by his exsiccatae nos. 1044, 1082, 1083, 1084, 1649 and 1660, which are to be found in several of the larger herbaria. It has since been collected in this area by R. J. Burdon and by M. L. Wedgwood (1921), G. C. Druce (1921 and 1924), J. E. Lousley (1950), C. West no. 53/10 (1953), N. D. Simpson no 53/760 and A. H. G. Alston (1954), W. Scott (1956) and by C. J. Cadbury and by W. Scott (1961). All the above gatherings have been made in grassy fields near North Roe. C. West noted that much of the area is now cultivated and the distribution of the species may be very restricted. In 1963, W. Scott found it in an entirely new locality, on the west side of Snarra Ness, Mainland. In 1898 E. S. Marshall identified plants from Farr Bay, Kirtomy and Roeness Voe (Beeby no. 1092) as H. zetlandicum. This treatment was followed by F. N. Williams (1902) and F. J. Hanbury (1904). W. R. Linton (1905) pointed out that the Sutherland plant was in his opinion a form of H. caledonicum F. J. Hanb., and Beeby (1908) described the Roeness Voe plant as a new species, H. breve. K. H. Zahn (1921) considered that the Sutherland plant was intermediate between H. caledonicum and H. scoticum F. J. Hanb., and J. Roffey (1925) described it as a new species, H. pseudozetlandicum. H. W. Pugsley (1948) placed it as a variety under H. caledonicum. In our opinion some of the specimens can be referred to H. scoticum, others to H. caledonicum. \t is difficult to understand how this Sutherland plant could ever have been referred to H. zetlandicum or even compared with the specimens later described as H. breve. As the differences between H. zetlandicum and the Sutherland plants had been clearly pointed out by W. R. Linton in 1905, the reason for recording H. zetlandicum for two vice-counties by Roffey (1925) and by Druce (1928) is not clear. H. zetlandicum is quite unlike any other British species, but it is closely allied to the Faeroese H. ostenfeldii Dahlst. and H. hartzianum Dahlst., and to the Icelandic H. arrostocephalum Omang. H. ostenfeldii has a less hairy stem, no simple hairs on the peduncles and phyllaries, and larger, broader and less sharply toothed leaves; H. hartzianum has longer petioles, broader and usually more sharply toothed basal leaves, no simple hairs on the peduncles, and more stellate hairs on the phyllaries; and H. arrostoce- phalum has leaves with more stellate hairs and almost without simple hairs, and more stellate and simple hairs on its narrower, more acute phyllaries. These differences are small, and it would be interesting to bring the four species under cultivation in the same environment. In cultivation H. zetlandicum becomes much Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 89 larger and more robust, its leaves more numerous and more deeply toothed and the petioles longer. The inflorescence is more ample and the peduncles usually have more simple hairs. It may be noted that Beeby (1891) remarks: “‘The taller plants mentioned occurred towards the bottom of the hillside; and one specimen found quite at the bottom reached the height of 16 inches and bore 10 flowers; this was growing on a patch of waste ground and was evidently abnormal.” Thus it seems that the dwarf, slender plant found in the typical habitat could be rather different if found in a more favourable situation. It should be noted, however, that the majority of its characters are constant and that it is sharply demarcated from all except the three species mentioned above. 2. HIERACIUM BREVE Beeby (1908) 112 (Lectotype: rocks, north side of Roeness Voe, Main- land, Shetlands, 9 July 1889, W. H. Beeby no. 1043 (BM)); Druce (1922) 499; Roffey (1925) 29 no. 1194; Druce (1928) 70 no. 196; Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 226; P. D. Sell & C. West in Dandy (1958) 127 no. 558: 1: 174. PLATE 3b. [H. zetlandicum sensu E. S. Marshall & Shoolbred (1898) J. Bot. (Lond.) 36, 172 (quoad Beeby no. 1092); sensu F. N. Williams (1902) 126 (quoad Beeby no. 1092).] Aphyllopodous or with basal leaves withering at the time of flowering.* Stem 15-33 (40 cult.) cm. high, robust or slender, markedly striate, with dense, long simple hairs especially below, those of the upper part dark-based, with a few stellate hairs above, without glandular hairs. Leaves deep green, sometimes becoming copper-coloured with age; the radical usually few, or occasionally up to 9; the primordial subrotund, subentire, cuneate-based to a broad, winged petiole; the Jater more or less broadly elliptic to elliptic-rhom- boidal, obtuse-mucronulate to acute, remotely denticulate or occasionally with a few, shallow, acute, patent teeth, abruptly contracted to a rounded or cuneate base, petioles broadly winged; the cauline 4-11 (-13 cult.), the Jower ovate-lanceolate to ovete, rarely ovate-rhomboidal, obtuse to acute, denticulate or with 2-3, short, acute, patent teeth on each side, cuneate-based, subpetiolate or sessile, the upper smaller, ovate- lanceolate, acute, toothed as lower, the base rounded and semi-amplexicaul; all with straight, medium, bulbous-based, subsetiform simple hairs on both surfaces and the margins (in cult. specimens sometimes very few), and on the petioles where they are similar but longer. Inflorescence compactly cymose (in cultivated specimens the branches more elongate), with 2-6 (13 cult.) capitula, acladium with medium peduncle; peduncles medium, suberect, with few stellate hairs, fairly numerous, short, patent simple hairs and occasionally a few, very short glandular hairs. Phyllaries 8-10 mm. long, 1:75—2 mm. wide, incumbent in bud, blackish-green, inner with paler margins, linear-lanceolate, obtuse, rarely with a few stellate hairs at base, with numerous, very short, glandular hairs, sometimes with a few, short simple hairs. Ligules yellow, glabrous tipped. Styles discoloured. Margins of the receptacle pits long-dentate. Cypsela dark, 3-3-5 mm. ong. H. breve was first collected by W. H. Beeby (nos. 1043 and 1092) on the north side of the Roeness Voe, Shetland, in 1889 and was referred by E. S. Marshall and F. N. Williams to H. zetlandicum Beeby. Beeby, however, realized it was not his H. zetlandicum, and described it as a new species. It was seen in 1920, in the same locality, by G. C. Druce who states that he saw two plants. There is a single specimen in Herb. Druce which is, however, atypical in having a nearly glabrous upper surface to the leaves and a long branching inflorescence. It matches well a cultivated specimen of Beeby’s, and it is possible that Druce brought back a root and cultivated it. In general appearance these cultivated specimens resemble H. zetlandicum, but in detail they are very different. 1t was rediscovered after a long search by W. Scott in 1963. No species that at all resembles H. breve has been seen. 3. Hieracium solum P. D. Sell & C. West, sp. nov. PLATE 4a. Holotype: Loch an Dughaill, Kintyre, v.c. 101, Grid Ref. 16/805806, 11 July 1962, A. G. Kenneth (CGE). In Sectione Alpestribus nulli speciei valde affinis. Planta pseudophyllopoda. Caulis ad 50 cm. altus, flexuosus, fere purpureus, basi pilis paucis medio- cribus vel longis pallidis interdum nigribasibus simplicibus, superne stellatis dispersis, glanduliferis nullis vestitus. Folia caesia, interdum purpureo-tincta; basalia late elliptica, anguste elliptica vel oblonga, obtusa, subintegra, basi cuneata, sessilia, semiamplexicaulia; caulina 3-8, interdum ramos abortivos ex axillis emittentia, lanceolata, acuta, subintegra, basi rotundata sessilia semiamplexicaulia; omnia in marginibus et raro in pagina pilis dispersis, pallidis, mediocribus vel longis, simplicibus vestita, et subtus secus costam et raro in pagina pilis dispersis, pallidis, mediocribus vel longis simplicibus vestita, interdum pilis simplicibus carentia, interdum marginibus puberulis, interdum supra vel utrinque pilis stellatis dispersis induta. Anthela compacta * Rarely phyllopodous. Watsonia 6 (2), 1965. 90 P. D. SELL and C. WEST cymosa, nonnunquam ramos ex axillis foliorum caulinorum superiorum emittens, capitulis 3-7 (-11). acladio pedunculo mediocri usque brevi praedito; pedunculi suberecti, mediocres vel longi, pilis stellatis, raro simplicibus vel glanduliferis brevissimis solitariis vestiti. Squamae 11—13 mm. longae, 1-5—2 mm. latae, in gemmis incumbentes, atro-virides, interiores marginibus pallidioribus, lineari-lanceolatae, obtusae, pilis stellatis paucis dispersis, aliis simplicibus dispersis obscuris mediocribus, aliis glanduliferis brevissimis nigris dispersis vestitae. Ligulae luteae, glabrae. Styli obscuri. Cypselam non vidimus. Pseudophyllopodous. Stem to 50 cm. high, flexuous, usually purplish, with a few, medium to long, pale, sometimes dark-based simple hairs at the very base, with scattered stellate hairs above, without glandular hairs. Leaves caesious, sometimes purple-tinged; the basal broadly elliptic, narrowly elliptic or oblong, obtuse, subentire, cuneate to a sessile, semiamplexicaul base; the cauline 3-8, often with abortive branches in the axil of the leaves, lanceolate, gradually tapering from a rounded, sessile, semiamplexicaul base to a long acute apex, subentire, with scattered pale, medium to long simple hairs on the margins and midrib, rarely beneath, sometimes without simple hairs, sometimes with puberulous margins, sometimes with scattered stellate hairs on one or both surfaces. Inflorescence compactly cymose, often with branches from the axils of the upper cauline leaves, with 3-7 (11) capitula, acladium with short or medium peduncle; peduncles suberect, medium to long, with stellate hairs, occasionally with a solitary, very short simple or glandular hair. Phyllaries 11-13 mm. long, 1-5—2 mm. wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, obtuse, with a few scattered stellate hairs, with scattered dark, medium simple hairs and scattered, very short, black glandular hairs. Ligules yellow, glabrous. Styles discoloured. Cypsela not seen. This very distinct species, which has no close allies, is known only from the type locality and Cruach Lusach nearby. In both localities it is very scarce. Although it is placed between H. breve and H. gratum, the only character it has in common with these two species is its few cauline leaves. The epithet solum is applicable in two senses, first because this species has no close allies, secondly because its only known localities are isolated ones for a species belonging to the Section A/pestria. 4. HIERACIUM GRATUM P. D. Sell & C. West (1955) 236 (Based on H. demissum var. pulchelliforme W. R. Linton); in Dandy (1958) 127 no. 558: 1 : 169. PLATE 4b. [H. pulchellum sensu Beeby (1889-1890) 35; sensu F. N. Williams (1902) 162; sensu Edmondston (1903) 72; sensu F. J. Hanb. in Bab. (1904) 263.] H. demissum var. pulchelliforme W. R. Linton (1905) 73 (Lectotype: Unst, Shetland, 16 August 1886, W. R. Linton in Herb. W. R. Linton (LIVU)); Druce (1922) 499; Roffey (1925) 29 no. 1198b; Druce (1928) 70 no. 200b. H. dovrense subsp. demissum var. pulchelliforme (W. R. Linton) Beeby (1908) 114. H. demissum subsp. demissum var. pulchelliforme (W. R. Linton) Zahn (1921) 855. H. pulchelliforme (W. R. Linton) Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 221; non H. pulchelliforme Dahlst. ex Omang (1944) Nytt. Mag. Naturvidensk (Oslo), 84, Hie [H. pulchellum var. vestitum Lindeb., nomen in Herb.] Phyllopodous or with basal leaves withering at time of flowering. Stem 15-35 (40) cm. high, slender, flexuous, markedly striate, usually purple below, with long, pale, wavy simple hairs below, becoming nearly glabrous upwards, but usually with some simple hairs at the nodes, with scattered stellate hairs throughout, without glandular hairs. Leaves pale green, often suffused beneath and tinged above with purple; the radical few, soon dying off; the primordial when present broadly elliptic, obtuse-mucronulate, with a small den- ticulation on each side, cuneate to a winged petiole; the / :ter narrowly to broadly elliptic or ovate-lanceolate with an obtuse-mucronulate to subacute apex, with up to 4 sharp, patent denticulations on each side; -+-attenuate to winged petioles; the cauline 2-4 (—7), rapidly decreasing in size, the lower elliptic to lanceolate, obtuse-mucronulate to acute, toothed as basal, narrowed to winged, semiamplexicaul petioles; the median ovate-lanceolate or lanceolate, obtuse to acute, the apex sometimes tapering from near the middle, abruptly contracted to a narrow, sessile, semiamplexicaul base, toothed as lower; the upper lanceolate or ovate- lanceolate, acute, apex tapering from the middle, entire or with up to 3 denticulations on each side; all with long, pale, subsetiform simple hairs on both surfaces and on the margins, the upper cauline sometimes only sparingly so, and sometimes sparsely stellately hairy beneath. Inflorescence compactly cymose, with 1-6 capitula, sometimes with one or two branches from the upper leaf-axils; peduncles usually long, slender and suberect, sometimes with several bracts, with stellate hairs, with or without a very occasional short simple hair or very short glandular hair. Phyllaries 9-11 mm. long, c. 1:5 mm. wide, incumbent in bud, blackish- green, the inner with paler margins, linear-lanceolate, obtuse or occasionally acute, with a few stellate hairs Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 91 near the base, usually sparingly covered with short to medium, pale, dark-based simple hairs, and slender, very short, dark glandular hairs, sometimes, however, the simple hairs are numerous, sometimes nearly absent. Ligules yellow, glabrous. Styles discoloured. Margins of the receptacle pits fimbriate-dentate. Cypsela dark, 2-5-3 mm. long. HA. gratum was first found by W. H. Beeby (no. 608) and W. R. Linton near Burrafirth, Unst, Shetlands, in 1886. It was sent to Lindeberg, who referred it to H. pulchellum Lindeb. var. vestitum Lindeb. The varietal name does not seem to have been published and is only to be found on herbarium sheets. H. Dahlstedt was of the opinion that it was not AH. pulchellum Lindeb., but a form very closely allied to H. demissum (Stromfelt) Dahlst. In 1905, W. R. Linton described it as a new variety, pulchelliforme, of H. demissum, which was raised to specific rank by H. W. Pugsley in 1946. As this name was a homonym we were forced to give it a new one, H. gratum, chosen because of the pleasing appearance of the plant. This species has been re-collected by W. A. Shoolbred (1892), F. J. Hanbury (1894), G. C. Druce (1920), R. J. Burdon (1921), J. E. Lousley (1950), C. West no. 53/13 (1953), R. C. Palmer (1956), C. J. Cadbury (1960) and W. Scott (1962). In our opinion this species bears little resemblance to either H. pulchellum or H. demissum. \t shows no close affinity to any British species, but is closely allied to the Icelandic H. elegantiforme Dahlst. It differs from that species, however, in having nearly glabrous, not hairy, peduncles and larger flowers. In 1953 it was sparingly distributed on steep, grassy slopes on both sides of the Burrafirth. 5. Hieracium difficile P. D. Sell & C. West, sp. nov. PLATE Sa. Holotype: Bank of stream above the Bay of Okraquoy, Mainland, Shetlands, v.c. 112, Grid. Ref. 41/4431, 17 July 1964, R. C. Palmer, N. D. Simpson and C. West (CGE). Ab H. australiore (Beeby) Pugsl. pilis paucioribus vestitis squamisque angustioribus differt. Folia basalia florendi tempore marcescentia vel delapsa. Caulis ad 70 cm. altus, flexuosus, purpureus praesertim subtus, pilis vestitus aliis simplicibus longis flexuosis interdum nigribasibus ubique sed supra tantum dispersis, aliis stellatis ubique, glanduliferis nullis. Folia mediocriter viridia, interdum purpureo- tincta; basalia late elliptica, obtuso-mucronulata, subintegra, e basi ad petiolum brevem alatum attenuata, utrinque et in marginibus pilis simplicibus densis, mediocribus, pallidis, subsetiformibus vestita; caulina 4-8, remota, a basi caulis sursum gradatim minora, interdum ramos abortivos ex axillis emittentia; inferiora elliptica, acuta vel longe acuta, remote et leviter denticulata, sensim ad basem sessilem semiamplexicaulem attenuata vel petiolo alato semiamplexicaule, ut folia basalia vestita sed pagina superiore pilis simplicibus paucioribus induta; swperiora lanceolata, Jonge acuta, e basi rotundata, sessilia semiamplexicaulia, utrinque pilis simplicibus mediocribus dispersis et interdum aliis stellatis vestita. Anthela anguste et compacte cymosa, capitulis 3-12, acladio pedunculo mediocri praedito; pedunculi suberecti, pilis stellatis, interdum pilo unico simplici brevi nigribasi et nonnunquam pilis glanduliferis brevissimis perpaucis obscuris vestiti. Squamae 9-10 mm. longae, 1-1-5 mm. latae, in gemmis incumbentes, atro-virides, interiores marginibus pallidioribus, lineari-lanceolatae, obtusae vel subacutae, basi pilis stellatis dispersis, alioquin pilis gland- uliferis brevibus vel brevissimis obscuris dispersis et nonnunquam aliis simplicibus solitariis brevibus nigribasibus vestitae. Ligulae luteae, glabrae. Styli obscuri. Cypsela obscura, circa 4 mm. longa. Basal leaves withering before or at the time of flowering. Stem to 70 cm. high, flexuous, purplish especially below, with long, wavy, pale, sometimes dark-based simple hairs throughout, but only scattered above, with scattered stellate hairs throughout, without glandular hairs. Leaves medium green, sometimes purplish tinged; the basal broadly elliptic, obtuse-mucronulate, subentire, attenuate to a short winged petiole, with dense medium, pale, subsetiform simple hairs on both surfaces and the margins; the cauline 4-8, widely spaced and gradually decreasing in size, sometimes with abortive branches in the axils of the leaves; the /Jower elliptic, acute to long acute, remotely and shallowly denticulate, gradually narrowed to a broad, sessile semi- amplexicaul base or with a winged, semiamplexicaul petiole, hairs as on basal but fewer in number on the upper surface; the upper lanceolate, gradually tapering from a rounded, sessile, semiamplexicaul base to a long acute apex, with scattered, medium simple hairs and sometimes with stellate hairs on both surfaces. In- florescence narrowly and compactly cymose, with 3-12 capitula, acladium with medium peduncle, peduncles suberect, with stellate hairs, sometimes with an occasional, short, dark-based simple hair and a very short, dark glandular hair. Phyllaries 9-10 mm. long, 1-1-5 mm. wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, obtuse to subacute, with scattered stellate hairs at the base, with scattered short and/or very short, dark glandular hairs, and an occasional, short, dark-based simple hair. Ligules yellow, glabrous. Styles discoloured. Cypsela dark, about 4 mm. long. This species was discovered by W. Scott in 1962, and collected later that year by R. C. Palmer. In 1964 R. C. Palmer, N. D. Simpson and C. West examined the colony and obtained additional material. We have given it the specific epithet difficile, not only because of the difficulty in placing it in its correct Watsonia 6 (2), 1965. 92 | P. D. SELL and C. WEST position in the Section Alpestria, but also because the variation in hair clothing, even on the same plant, makes it difficult to provide a precise diagnosis. For example, the leaves of the holotype have stellate hairs on both surfaces whereas the leaves of the other specimens have not. Normal plants of H. difficile closely resemble H. australius, but differ by being less hairy and having narrow phyllaries. A plant from this population, which seems to us to be conspecific, has a basal rosette and closely resembles H. gratum in appearance, but it is immediately distinguishable by its more numerous, less denticulate cauline leaves. 6. HIERACIUM AUSTRALIUS (Beeby) Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 224. PLATE 5b. [?H. crocatum sensu Tate (1866) 6.] [H. dovrense sensu Beeby (1889-1890) 35; sensu F. N. Williams (quoad Shetland loc. exclud. Beeby no. 1151) (1902) 161; sensu Edmondston (1903) 72; sensu F. J. Hanb. in Bab. (1604) 264; sensu W. R. Linton (quoad Shetland loc.) (1905) 72.] H. dovrense subsp. demissum var. australius Beeby (1908) 114 (Lectotype: Rocks, Loch of Cliff, Unst, Shetlands, 27 July 1886, W. H. Beeby no. 609 (BM)). H. demissum var. australius (australe) (Beeby) Druce (1922) 499; Roffey (1925) 29 no. 1198c; Druce (1928) 70 no. 200 c [1621]. [H. polycomum sensu Dahlst. (quoad Ronas Voe loc.) (1926 for 1925) Rep. Bot. Soc. & Exch. Cl. Brit. Isles, 7, 997.] |Z’. polycomatum sensu Druce (quoad Ronas Voe et Cliva Hill loc.) (1928) 71 no. 236.] Aphyllopodous or with basal leaves withering at the time of flowering. Stem to 65 cm. high, usually robust, markedly striate, purplish below, usually with long, pale simple hairs below, sometimes throughout, the hairs often deciduous above, with a few stellate hairs just below the inflorescence, without glandular hairs. Leaves pale green, sometimes with purplish spots; the radical soon drying up and dropping off, elliptic to broadly elliptic, obtuse-mucronate to acute, denticulate, narrowed below to winged petioles, with pale, medium, subsetiform, bulbous-based hairs on both surfaces, the margins and petioles; cauline 5/2 (21 cult.), gradually decreasing in size, the lower more or less elliptic, obtuse-mucronate to acute, subentire or sharply denticulate, attenuate to a winged petiole, clothed as basal, the median very variable, elliptic, lanceolate or oblong-lanceolate (sometimes curved in the shape of a crescent), acute, sometimes gradually narrowed from the middle, usually with sharp denticulations, gradually narrowed to a semiamplexicaul base, hair-clothing as on lower, but sparser, some specimens with abortive branches growing in the axils of the leaves, the upper ovate-lanceolate, gradually narrowed from a rounded base to an acute apex, denticulate or some- times with a few sharp teeth, usually semiamplexicaul, nearly glabrous above otherwise as lower, the upper- most sometimes with scattered stellate hairs beneath. Inflorescence compactly cymose, in cultivation laxer and cymose-corymbose, with 2—10 (80 cult.) capitula, the median acladium subtended by the branches; peduncles suberect, stellately hairy, with an occasional, short, dark-based simple hair. Phyllaries 9-10-5 mm. long, 1-5—2 mm. wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, obtuse, with a few stellate hairs near the base, with many, very short, pale and dark, glandular hairs, and few to many, usually short, rarely medium, pale, dark-based simple hairs, sometimes without simple hairs. Ligules yellow, glabrous. Styles discoloured. Margins of the receptacle pits long-dentate. Cypsela dark, 3-3-5 mm. long. The first record of this species was probably that of Tate (1866), who recorded as H. crocatum plants from Burrafirth and Loch-a-Cliff* which were almost certainly this species. The earliest specimens we have seen, however, are those collected in 1886 at Loch-a-Cliff by W. H. Beeby (no. 609), who named them H. doy- rense. In 1908, however, it was recognized as being distinct from that species, and was described by W. H. Beeby as a new variety, australius, of H. dovrense subsp. demissum. It has subsequently been found in that area by W. H. Beeby no. 859 and 861 (1887), E. F. Linton set no. 156 (1891), F. J. Hanbury (1894), R. J. Burdon and G. C. Druce (1920 and 1921), J. E. Lousley (1950), C. West no. 53/12 (1953) and R. C. Palmer and W. Scott (1962). There are specimens of this species from Ronas Voe collected by G. C. Druce (OXF) labelled H. polycomum, and from Cliva Hill in 1920 and 1921 (BM) by the same collector labelled @. polycomatum var. thulense. H. australius is quite distinct from H. dovrense Fries, H. demissum (Strémfelt) Dahlst., and H. poly- comum Dahlst. (H. polycomatum (Zahn) Druce) with which it has been confused. The only species closely allied to H. australius is H. difficile. * The modern official spelling is Loch of Cliff. Watsonia 6 (2), 1965. y Miia, al I Sin la BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 93 7. HIERACIUM DOVRENSE Fries (1848) 128 (Provisional lectotype: Norveg. centr. a Dovre in Finmarkiarn copiose, M. N. Blytt, sent out as H. cydonaefolium in Fries Set 11 no. 12 (BM)); E. S. Marshall & Shoolbred (1898) 172; F. N. Williams (1902) 161 (quoad Ben Loyal* loc.); W. R. Linton (1905) 72(quoad Ben Laoghal* loc.); Roffey (1925) 29 no. 1190 (one v.c. only); Druce (1928) 70 no. 192 (1617) (one v.c. only); Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 221; P. D. Sell & C. West in Dandy (1958) 127 N@ws58> 1 7168. PLATE 6a. H. boreale 8 2latifolium Blytt (1838) 304 (Lectotype: Lierne ved Drivstuen, 1835, M. Blytt no. 110 (O)). H. dovrense var. vulgare Lindeb. in Blytt (1874) 668. H. dovrense subsp. eudovrense Omang in Blytt (1906) 760. H. carpathicum subsp. dovrense (Fries) Zahn (1921) 810) (exclud. all Brit. loc. except Ben Laoghal). Aphyllopodous or with basal leaves withering at the time of flowering. Stem to 50 cm. high, robust, flexuous, slightly purple tinted below, with long, pale simple hairs throughout, sometimes sparsely, sometimes densely, sometimes the hairs deciduous, sometimes with a few stellate hairs above, with an cccasional, very small glandular hair. Leaves pale green, radical few (in British specimens seen, soon disappearing); primor- dial elliptic, obtuse-mucronulate, slightly denticulate, attenuate to the petiole; the later (from Scandinavian specimens) elliptic, obtuse-mucronulate, remotely denticulate to coarsely dentate, attenuate to winged petioles which are broadened below, midrib very prominent, nearly glabrous above, with a few, pale, medium simple hairs below, margins and petioles with medium, pale simple hairs; cauline 4-8, slowly decreasing in size upwards, the lower and median elliptic, obtuse-mucronulate to acute, denticulate to sharply and irregularly dentate, the teeth ascending mammilliform, the lower attenuate below to winged semi-amplexicaul petioles, the median to a sessile or abruptly contracted semiamplexicaul base, clothing as radical, the upper lanceolate, acute to acuminate, denticulate or shallowly dentate, cuneate to a sessile, semiamplexicaul base, glabrous except on the midrib below. Inflorescence rather compactly cymose, usualiy with 3-9 (-12) capitula, acladium geminate, peduncles medium, suberect, slender, stellately hairy, otherwise variable in clothing with few to numerous, short to medium, dark-based simple and very short, black glandular hairs. Phyllaries 9-11 mm. long, 2-2-5 mm. wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, obtuse, sparingly stellately hairy with a tuft at the apex, with many, short to medium, dark-based simple and very short, black glandular hairs. Ligules yellow, glabrous. Styles discoloured. Margins of the receptacle pits long-dentate. Cypsela dark, 2-2-5 mm. long in Scandinavian specimens, cypselas from British plants not seen. H. dovrense was described by E. Fries in 1848 from several alpine localities in Norway, particularly in the Dovre area. He states that he has seen a specimen of this species from Scotland, but the specimen has not been traced. In 1896 E. S. Marshall and W. A. Shoolbred gathered specimens at Glen Roy, Inverness, which were identified as H. dovrense. These were later sent out in Linton’s Set no. 46 under that name. They are, however, H. dewarii Syme. The following year Marshall and Shoolbred collected the true plant on Ben Loyal, which is its only known British locality. It has subsequently been seen by J. E. Raven in 1952 and by C. West in 1953. Specimens collected by W. H. Beeby and F. J. Hanbury in Unst, Shetland, were identified as H. dovrense, but they belong to a different species, H. australius (Beeby) Pugsley. Specimens collected by E. S. Marshall at Tain in 1890 and labelled H. dovrense are H. subrude (A.T.) A.T. The typification of H. dovrense Fries is rather complicated. There is no specimen that can be regarded as a type in Fries’ own herbarium at Uppsala. At Oslo there are four syntypes of H. boreale ? var. latifolium Blytt, which is cited in the synonomy of Fries’ original account. Although these specimens are all typical H. dovrense, there is no evidence on the sheets that they were seen by Fries. Fries’ Exsiccatae Herb. Norm. XI : 12 sent out as H. cydonaefolium is cited by him in his original account of H. dovrense, and can be regarded as type material. All the specimens we have seen of this set number are badly damaged, but they are in our opinion certainly referable to the taxon regarded as H. dovrense. We provisionally select the sheet in the British Museum (Nat. Hist.) as the lectotype, but recommend that should a better specimen be found, a fresh designation should be made. K. H. Zahn (1921) describes seven subspecies of his Grex H. dovrense, all of which come from Scan- dinavia. We have not seen specimens of these, but from the descriptions they seem to differ little from H. dovrense Fries in its typical form. No British species is closely allied to H. dovrense. * Ben Loyal, not Ben Laoghal, appears to be the correct modern spelling for the name of this mountain. Watsonia 6 (2), 1965. 94 P. D. SELL and C. WEST 8. HIERACIUM PRAETHULENSE Pugsley (1948) 226 (Holotype: rocks on north side of Roeness (Ronas) Voe, Mainland, Shetlands, 3 August 1891, W. H. Beeby no. 1115 in Herb. F. J. Hanbury (BM)); P. D. Sell & C. West in Dandy (1958) 127, no. 558:1:175. PLATE 6b. [H. auratum sensu Beeby (1892) 55.] H. auratum var. thulense F. J. Hanb. (1894) 232 (Lectotype: rocks on north side of Roeness Voe, Mainland, Shetlands, 3 August 1891, W. H. Beeby no. 1115 in Herb. F. J. Hanbury (BM)); F. J. Hanb. in Bab. (1904) 267; W. R. Linton (1905) 89; Beeby (1908) iS. [H. aestivum subsp. crocatum sensu Zahn (quoad Shetland loc.) (1922) 922.] H. crocatum var. thulense (F. J. Hanb.) Druce (1928) 71, no. 245e (cf. Rep. Bot. Exch. Cl. Brit. Isles 8, 879 and 880 (1929 for 1928)). [H. praethulense Pugsley in A. R. Clapham (1946) 346 nom. nud.] Hypophyllopodous or aphyllopodous or with basal leaves withering at the time of flowering. Stem to 60 cm. high, usually rather slender but some large plants robust, flexuous, markedly striate, purplish below, sometimes throughout, with long, pale simple hairs throughout, but decreasing in density upwards, some- times without simple hairs except at the very base, with scattered stellate hairs in the upper half, without glandular hairs. Leaves pale green, often suffused with purple; the radical few or absent, usually broadly elliptic and sometimes almost subrotund, apex rounded or obtuse, remotely denticulate, narrowed (sometimes abruptly) to winged petioles, with scattered, medium simple hairs or glabrous above, with long simple hairs beneath and on the margins; the cauline 7-12 (—16) (in cultivation with numerous other leaves on the branches), the /Jower variable, elliptic, ovate or oblong-lanceolate, obtuse-mucronulate to acute, remotely - denticulate, attenuate to a winged petiole or abruptly narrowed to a sessile, semiamplexicaul base, clothed as basal, the median elliptic or lanceolate, acute, remotely denticulate, with a rounded or subcordate semiamplexicaul base, glabrous above, with few simple hairs beneath, the margins with simple hairs or scaberulous; the upper ovate-lanceolate to lanceolate, gradually narrowed to an acute apex, entire or with an occasional denticulation, base rounded or subcordate, semiamplexicaul, glabrous above, with few simple and stellate hairs beneath, margins ciliate or scaberulous. Inflorescence compactly cymose (in cultivation cymose-corymbose), with 2—9 (100 cult.) capitula, with medium acladium; the peduncles slender, suberect, medium to long, sparingly stellately hairy and sometimes with a few, pale, short simple hairs, without glandular hairs. Phyllaries 9-10 mm. long, (1—) 1-5—2 mm. wide, incumbent in bud, blackish-green, inner with paler margins, linear-lanceolate, obtuse, with scattered stellate hairs, few to many, very short to shart glandular hairs and occasionally a few, short simple hairs intermixed. Ligules yellow, glabrous. Styles yellow, sometimes discoloured when dry. Margins of the receptacle pits shortly and irregularly dentate. Cypsela dark, 4-4: 5 mm. long. H. praethulense was first found by W. H. Beeby in 1889 (nos. 1041 and 1042), and re-collected by him in 1891 (nos. 1108, 1109, 1111, 1112, 1114, 1115 and 1116). It has since been collected by G. C. Druce (1920), J. E. Lousley (1950), C. West no. 53/8 (1953), C. J. Cadbury (1960 and 1961) and W. Scott (1961 and 1963). It is a frequent plant on the north side of Ronas Voe and near Mavisgrind. In 1892 Beeby called it H. auratum Fries; two years later F. J. Hanbury described it as a new variety, thulense, of that species. It was included under H. aestivum subsp. crocatum by Zahn, and both Roffey and Druce transferred the varietal name thulense to H. crocatum Fries. Pugsley quite rightly gave the plant specific status. In general appearance it resembles the Icelandic H. phrixoclonum Omang, but that species differs in having fewer cauline leaves, more glandular hairs on the peduncles, dark styles and smaller achenes. It is also closely allied to the Icelandic H. halfdanii Oskarss. from which it differs in being more hairy throughout and in having numerous fine glandular hairs on the peduncles. In cultivation, H. praethulense often becomes a very large plant with a widely spreading inflorescence with many capitula. It then has a superficial resemblance to H. latobrigorum (Zahn) Roffey (H. auratum auct.), which is perhaps the reason why Beeby and Hanbury originally placed it under that species. The shape and clothing of its phyllaries, however, are very different. The leaves in cultivated plants are far less amplexicaul than in the wild plant, a point which suggests that one of the characters on which the Section Alpestria is based may be artificial. The plants which Druce recorded as H. polycomum Dahlst., and which Pugsley referred to H. praethu- lense, are in fact AH. australius (Beeby) Pugsley. 9. HIERACIUM HETHLANDIAE (F. J. Hanb.) Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 224; P. D. Sell & C. West in Dandy (1958) 127 no. 558: 1: 171. PLATE 7a. Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 2) [H. truncatum sensu Beeby (1892) 55.] H. dovrense var. Hethlandiae F. J. Hanb. (1894) 232 (Lectotype: rocks at the foot of Cliva Hill, near Brae, Mainland, Shetlands, 14 August 1891, W. H. Beeby no. 1107 in Herb. F. J. Hanbury (BM)); F. J. Hanb. in Bab. (1904) 264; W. R. Linton (1905) 72; Druce (1922) 499 (exclud. Burn of Quoys loc.). [H. dovrense sensu F. N. Williams (1902) 162 (quoad W. H. Beeby no. 1151).] H. dovrense subsp. humidorum var. Hethlandiae (F. J. Hanb.) Beeby (1908) 114. H. aestivum subsp. crocatum subvar. Hethlandiae (F. J. Hanb.) Zahn (1922) 923. H. crocatum forma Hethlandiae (F. J. Hanb.) Roffey (1925) 29, no. 1243. H. crocatum var. Hethlandiae (F. J. Hanb.) Druce (1928) 71, no. 245c. (1635). Aphyllopodous or with basal leaves withering at the time of flowering. Stem to 40 cm. high, slender, flexuous, purplish below, often throughout, with long, pale simple hairs only at the very base and then often only sparingly, with --_numerous stellate hairs especially above, without glandular hairs. Leaves 5-9, caesious medium green, often purple-tinted especially beneath and on the margins, the upper rapidly decreasing in size; the lowest broadly elliptic, +-obtuse, subentire, attenuate to short, winged, semiamplexicaul petioles, glabrous or with pale, medium, subsetiform, bulbous-based simple hairs above and on the margins, petioles with medium, pale, simple hairs; the /ower cauline narrow-elliptic, gradually narrowed at both ends, apex acute, denticulate, with short winged petioles, or sessile and semiamplexicaul, glabrous or with pale, medium simple hairs on both surfaces and the margins; the median as lower, but sometimes broader and more rounded at the base with a long acute apex, sessile, glabrous or nearly so on both sides, margins puberulous, or with a few, short simple hairs; the upper gradually narrowed to an acute, apiculate tip from a broad, rounded, semiamplexicaul base, glabrous except for a few stellate hairs below. Inflorescence compactly cymose, with 2-4 capitula, acladium with medium peduncle; peduncles medium, suberect, slender, stellately hairy, usually without simple hairs, without glandular hairs. Phyllaries 10:5-12 mm. long, 1:5—1-75 mm. wide, incumbent in bud, blackish-green, inner with paler margins, linear-lanceolate, obtuse, with stellate hairs scattered over the whole surface, with very short to medium, dark glandular and some short to medium, dark simple hairs. Ligules yellow, glabrous. Styles discoloured. Margins of the receptacle pits long-dentate. Cypsela dark, 3—5 mm. long. This species was first described as a variety of H. dovrense Fries by F. J. Hanbury (1894), from material collected by W. H. Beeby (no. 1107) at Cliva Hill. It was later placed by some authors under H. crocatum Fries. It is a rare plant, and has only been collected since by W. H. Beeby no. 1151 (1896), R. J. Burdon (1920), G. C. Druce (1920 and 1921), J. E. Lousley (1950), C. West no. 53/4 (1953) and C. J. Cadbury and W. Scott (1960). A sheet in Herb. Mus. Brit. (BM), collected by R. J. Burdon at Roeness Voe on the 28 July 1920, contains a mixed gathering of two plants of this species and one of H. vinicaule. H. hethlandiae needs to be refound at this locality. Both W. H. Beeby (1908) and G. C. Druce (1922) remark that it grows in some plenty on Cliva Hill rocks, but C. West (1953) found that much of the cliff had been blasted away for road repair, and that the plant might be in danger of extinction. In 1960 Cadbury and Scott saw eight plants. Druce (1922) identified plants from the Burn of Quoys as H. hethlandiae, but we consider these plants to be a distinct species, H. dilectum. H. hethlandiae does not seem to be closely allied to any of the British species of the Section Alpestria. It was likened to the Scandinavian H. dovrense subsp. humidorum Almq. ex Elfstrand by Elfstrand, but that taxon, although closely allied to H. hethlandiae, differs in its long-petioled lower leaves, smaller capitula and markedly pale-edged phyllaries. H. hethlandiae has a superficial resemblance to H. sparsifolium Lindeb. (H. pseudoprotractum Pugsley), which grows with it; it can easily be separated, however, by its broad-based, semi-amplexicaul leaves and different clothing of the inflorescence. 10. Hieracium attenuatifolium P. D. Sell & C. West, sp. nov. PLATE 7b. Holotype: Laxo Burn, Mainland, Shetlands, August 1953, C. West no. 53/1 (CGE) [H. demissum var. pulchelliforme sensu Druce (1925) 646.] Ab H. dilecto P. D. Sell & C. West, H. pugsleyi P. D. Sell & C. West et H. subtruncato Beeby Squamis capitulorum pilis glanduliferis densis indutis distinguitur. Planta aphyllopoda. Caulis ad 90 cm. altus, flexuosus, -+-purpureus, pilis simplicibus longis pallidis superne saepe deciduis, per bases pilorum deciduorum trunciformes tandem scaberulus, aliis stellatis paucis, glanduliferis nullis. Folia caulina 6-15, saturate vel pallide viridia, saepe purpureo-tincta; ima late elliptica, rotundo-obtuso-mucronulata, subintegra vel remote denticulata, basi cuneata vel in petiolum attenuata; inferiora elliptica, interdum late elliptica, obtuso-mucronulata vel subacuta, remote denticulata vel raro dente levi acuto solitario praedita, basi gradatim in petiolum late alatum angustata; media anguste ad late elliptica, Watsonia 6 (2), 1965. 96 P. D. SELL and C. WEST -+-acuta, remote denticulata vel raro dentibus levibus acutis paucis praedita, basi angustata vel abrupte contracta, sessilia, semiamplexicaulia; superiora lanceolata vel lineari-lanceolata, acuta, subintegra vel bracteiformia; inferiora et media utrinque et in marginibus pilis simplicibus mediocribus pallidis, praeterea media raro subtus pilis stellatis paucis, superiora raro pilis simplicibus paucis et utrinque pilis stellatis dispersis vestita cum marginibus costisque puberulis. Anthela laxe cymosa, capitulis 2-7 (ad 13 cult.), acladio pedunculo longo praedito; pedunculi vulgo longi, pilis stellatis dispersis, saepe aliis simplicibus numerosis mediocribus nigribasibus, aliis glanduliferis paucis brevissimis atris vestiti. Sguamae 10-12 mm. longae, 1-5-2 mm. latae, in gemmis incumbentes, exteriores atro-virides, brevissimae, triangulari-ovatae, mediae marginibus pallidioribus, lanceolatae, interiores valde pallidae et angustiores, omnes obtusae, pilis stellatis paucissimis, aliis glanduliferis inaequalibus brevissimis usque mediocribus nigris, aliis simplicibus paucis mediocribus obscuris vestitae. Ligulae pallide luteae, glabrae. Styli lutei, in sicco obscuri. Alveoli receptaculi margine breviter dentati. Cypsela obscura, 4-5—5 mm. longa. Aphyllopodous. Stem to 90 cm. high, flexuous, --purple, with long, pale simple hairs above, often becom- ing scaberulous from the bases of the deciduous hairs, with a few stellate hairs, without glandular hairs. Cauline leaves 6-15, deep or pale green, often suffused with purple; the /owest broadly elliptic, rounded- obtuse-mucronulate, subentire to remotely denticulate, cuneate- or attenuate-based to winged petioles; the lower elliptic, sometimes broadly, obtuse-mucronate to subacute, distantly denticulate or occasionally with one or more sharp teeth, gradually narrowed below to broad winged petioles; the median narrow to broadly elliptic, --acute, distantly denticulate or rarely with a few, sharp, shallow teeth, narrowed or abruptly con- tracted to a sessile, semiamplexicaul base; the upper lanceolate to linear-lanceolate, acute, subentire or bractlike; the lower and median with medium, pale simple hairs on both surfaces and the margins, the median occasionally with a few stellate hairs beneath, the upper with or without an occasional simple hair, with scattered stellate hairs on both surfaces, margins and veins puberulous. Inflorescence loosely cymose, with 2—7 (—13 cult.) capitula, acladium with a long peduncle; peduncles mostly long, with scattered stellate hairs, usually with numerous, medium, dark-based simple hairs and an occasional very short to short, dark glandular hair. Phyllaries 10-12 mm. long, 1:5—2 mm. wide, incumbent in bud, the outer blackish green, very short, triangular-ovate, the median with paler margins, lanceolate, the inner similar to the median but paler and rather narrower, all obtuse, with very few stellate hairs, numerous, unequal, very short to medium, dark glandular and a few, medium, dark simple hairs. Ligules pale yellow, glabrous. Styles yellow, darkening when dry. Margins of receptacle pits shortly dentate. Cypsela dark, 4-5—5 mm. long. G. C. Druce was the first person to mention this species from the Laxo Burn, after,his third visit to the Shetlands in 1924, and we have seen three sheets containing eight very poor specimens which he gathered. In his Additions to the Flora Zetlandica (1925), Druce refers this plant to H. demissum var. pulchelliforme W.R. Linton, and his specimens are so labelled. This species bears not the slightest resemblance to H. gratum P. D. Sell and C. West (= H. demissum var. pulchelliforme W. R. Linton). In 1953, C. West no. 53/1, found it growing in abundance on the grassy banks of the Laxo Burn, where it has since been collected by N. D. Simpson no. 53/762 (1954), W. Scott (1956) and C. J. Cadbury (1962). In 1954 Simpson (no. 54/291) gathered it from sea cliffs at West Burrafirth, Sandsting, where it was collected again by W. Scott in 1963. The plants from Sandsting differ from those of the Laxo Burn in having slightly narrower phyllaries clothed with slightly longer glands, but these differences are negligible. Plants collected by W. Scott at Bousta, Sandness, 29 July 1962, seem to belong to this species, but they have fewer leaves and apparently darker styles. H. attenuatifolium is closely allied to H. subtruncatum Beeby, H. pugsleyi P. D. Sell & C. West and H. dilectum P. D. Sell & C. West. From these three species it can be distinguished by the dense glandular hairs on its phyllaries, on which there are only a few simple hairs. We have called it attenuatifolium because it can readily be distinguished from H. subtruncatum Beeby, which it strongly resembles in appearance, by the narrower more attenuate bases of the leaves. 11. Hieracium pugsleyi P. D. Sell & C. West, sp. nov. PLATE 8a. Holotype: grassy banks of burn, Weisdale, Mainland, Shetlands, August 1953, C. West no. 53/14 (CGE). Ab H. subtruncato Beeby squamis capitulorum pilis simplicibus numerosis vestitis, ab H. attenuatifolio P. D. Sell et C. West capitulis minoribus, squamis capitulorum pilis glanduliferis sparsis et forma foliorum basalium differt, praeterea ab H. dilecto P. D. Sell et C. West caulibus pedunculisque pilis simplicibus praeditis et internodiis longioribus distinguitur. Planta aphyllopoda. Caulis ad 60 cm. altus, gracilis, flexuosus, purpureus, perpilosus, pilis longis pallidis simplicibus, superioribus nigribasibus, superne pilis stellatis paucis, nonnunquam glanduliferis sparsis brevissimis vestitus. Folia caulina 8-12 (—15), pallide viridia, praecipue infra purpureo-tincta; inferiora Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 97 elliptico-oblonga ad oblonga, obtuso-mucronata ad subacuta, subintegra ad minute denticulata, basi in petiolum brevem alatum attenuata; media elliptico-oblonga ad oblonga, -+-acuta, subintegra ad regulariter denticulata, basi rotundata, sessilia, semiamplexicaulia; superiora lanceolata ad ovato-lanceolata, acuta, subintegra, basi rotundata, sessilia; omnia utrinque et in marginibus pilis pallidis mediocribus simplicibus vestita, superiora inferne pilis stellatis paucissimis vestita. Anthela cymosa, nonnunquam ramos longos foliosos ex axillis foliorum caulinorum emittens, capitulis 3-9 (22), acladio pedunculo brevi usque mediocri praedito, ramis gracilibus suberectis; pedunculi breves usque mediocres, nonnunquam geminati, pilis aliis stellatis paucis, aliis simplicibus numerosis mediocribus basi obscuris, aliis glanduliferis raris brevissimis obsiti. Squamae 9-10 mm. longae, 1-5—1-75 mm. latae, in gemmis incumbentes, atro-virides, nonnunquam apicibus rubescentibus, interiores marginibus pallidioribus, lanceolatae, obtusae, pilis aliis stellatis rarissimis, aliis simplicibus numerosis mediocribus basi obscuris, aliis glanduliferis paucis vel multis brevissimis obscuris vestitae. Ligulae luteae, glabrae. Styli Jutei. Alveoli receptaculi margine breviter dentati. Cypsela obscura, 3-3-5 mm. longa. Aphyllopodous. Stem to 60 cm. high, slender, flexuous, purple, with dense, long, pale simple hairs throughout, those of the upper half with dark bases, with few, scattered stellate hairs, sometimes with a few very short glandular hairs. Cauline leaves 8—12 (—15), pale green suffused with purple especially on the lower surface; the Jower elliptic-oblong to oblong, obtuse-mucronate to subacute, subentire to minutely denticulate, base attenuate to a short, winged petiole; the median elliptic-oblong to oblong, --acute, subentire to regularly denticulate, base rounded, sessile, semiamplexicaul; the upper lanceolate to ovate-lanceolate, acute, subentire, with a rounded, sessile base; all with pale, medium simple hairs on both surfaces and the margins, the uppermost with very few stellate hairs beneath. Inflorescence cymose, sometimes with long leafy branches in the leaf-axils, with 3-9 (22) capitula; acladium with short to medium peduncle, with slender, sub- erect branches; peduncles short to medium, sometimes geminate, with scattered stellate hairs, numerous, medium, dark-based simple hairs and a few, very short glandular hairs. Phyllaries 9-10 mm. long, 1-5—1-75 mm. wide, incumbent in bud, blackish green, sometimes reddish at the tip, the inner with paler margins, linear-lanceolate, obtuse, with very few stellate hairs, numerous, medium, dark-based simple hairs and few or many, very short, dark glandular hairs. Ligules yellow, glabrous. Styles yellow. Receptacle pits shortly dentate. Cypsela dark, 3-3-5 mm. long. This species was discovered by C. West (no. 53/14) in 1953 on grassy banks by the Weisdale Burn, Mainland, Shetlands, where it was re-collected in 1961 by C. J. Cadbury and in 1962 by W. Scott. It has also been collected at the Laxo Burn, Lunnasting by N. D. Simpson (no. 54.762) in 1954, and near Otters Wick, Yell, by C. J. Cadbury and by R. C. Palmer in 1961. In 1963, W. Scott collected it at the Burn of Laxdale near Cunningsburgh and at Tumblin, Sandsting. It is closely allied to H. subtruncatum Beeby, H. attenuatifolium P. D. Sell & C. West and H. dilectum P. D. Sell & C. West. From H. subtruncatum it differs in having numerous simple hairs on the phyllaries, from H. attenuatifolium by its smaller heads, less glandular hairy phyllaries, and differently shaped basal leaves, and from H. dilectum by its hairy stems and peduncles and its longer internodes. It is thought appropriate to dedicate this plant to the late Mr. H. W. Pugsley, whose contributions to our knowledge of the British species of Hieracium are outstanding. 12. Hieracium dilectum P. D. Sell &. C. West, sp. nov. PLATE 8b. Holotype: Burn of Quoys, Catfirth, Mainland, Shetlands, August 1953, C. West no. 53/2 (CGE). [H. dovrense var. hethlandiae sensu Druce (quoad Burn of Quoys loc.) (1922) 499.] Ab H. attenuatifolio P. D. Sell et C. West et H. pugsleyi P. D. Sell et C. West squamis capitulorum sparsius vestitis differt, praeterea ab HA. subtruncato Beeby foliis longioribus angustioribus acutioribus distinguitur. Planta aphyllopoda. Caulis ad 55 cm. altus, flexuosus purpureusque, raro glaber, plerumque pilis sim- Dlicibus paucis (raro numerosis) mediocribus pallidis praesertim in nodis, aliis stellatis dispersis, superne pilis glanduliferis brevissimis paucis vestitus. Folia caulina 9-16 (—20), pallide viridia, saepe purpureo-tincta praesertim inferne; inferiora longa anguste elliptica vel interdum linearia, obtuso-mucronulata vel subacuta, regulariter remoteque denticulata, basi in petiolum brevem late alatum attenuata, semi-amplexicaulia; media longa, anguste elliptica vel interdum linearia, acuta, remote denticulata, interdum dente acuto parvo praedita, basi breviter angustata vel cuneata vel rotundata, semiamplexicaulia, sessilia; superiora longe lanceolata, longe acuta, remote denticulata vel integra, sessilia, basi rotundata, semiamplexicaulia; omnia glabra vel utrinque et in marginibus pilis simplicibus paucis usque nullis pallidis, mediocribus vestita, interdum costa subtus dense pilosa, media atque superiora saepe pilis stellatis utrinque dispersis induta. Anthela laxe cymosa, capitulis 1-10, acladio pedunculo mediocri usque longo praedito; pedunculi mediocres usque longi, suberecti, pilis stellatis dispersis, interdum scaberuli et raro pilo simplici pallido mediocri vestiti. Squamae Watsonia 6 (2), 1965. 98 P. D. SELL and C. WEST 9-11 mm. longae, 1-25—1-5 mm. latae, in gemmis incumbentes, atro-virides, mediae et interiores marginibus pallidioribus, lineari-lanceolatae, obtusae, pilis stellatis dispersis, aliis simplicibus brevibus vel brevissimis obscuris dispersis, aliis glanduliferis brevibus vel brevissimis obscurts dispersis vestitae. Ligulac luteae, glabrae. Styli lutei vel obscuri. Alveoli receptaculi margine dentato-fimbriati. Cypsela obscura, 4—4:4\ mm. longa. Aphyllopodous. Stem to 55 cm. high, flexuous, purple, rarely glabrous, usually with a few (rarely more numerous), pale, medium simple hairs especially at the nodes, with scattered stellate hairs and a few very short glandular hairs above. Cauline leaves 9-16 (—20), pale green, usually suffused with purple, especially below; the lower narrowly elliptic or sometimes linear, obtuse-mucronulate to subacute, regularly and remotely denticulate, attenuate below to short, broad-winged, semi-amplexicaul petioles; the median narrowly elliptic or sometimes linear, acute, remotely denticulate with an occasional small sharp dentation, shortly narrowed to a cuneate or rounded, semiamplexicaul, sessile base; the upper long-lanceolate, long-acute, remotely denticulate or entire, sessile with a rounded, semiamplexicaul base; all glabrous or with few to many, pale, medium simple hairs on both surfaces and the margins, sometimes with very numerous simple hairs on the midrib beneath, the median and upper often with scattered stellate hairs on both surfaces. Inflorescence loosely cymose with 1-10 capitula, acladium with medium to long peduncle, peduncles medium to long, suberect, with scattered stellate hairs, sometimes scaberulous and occasionally with a solitary, pale, medium simple hair. Phyllaries 9-11 mm. long, 1-:25—1-5 mm. broad, incumbent in bud, blackish-green, the median and inner with paler margins, linear-lanceolate, obtuse, with scattered stellate hairs, scattered short or very short, dark simple hairs, and scattered very short or short, dark glandular hairs. Ligules yellow, glabrous. Styles yellow to discoloured. Margins of the receptacle pits fimbriate-dentate. Cypsela dark, 4-4-5 mm. long. This species was first collected in 1908 in the ravine of the Burn of Quoys, Catfirth, Mainland, Shetlands, by W. H. Beeby (no. 1664). Although Beeby does not mention this plant in any of his papers on the Shetland Hieracia, a specimen in his herbarium is labelled H. dovrense var. hethlandiae F. J. Hanb. A plant of this same gathering in Herb. E. S. Marshall (CGE) is similarly labelled, and the Burn of Quoys is given as a locality for H. hethlandiae in Druce’s Flora Zetlandica. In 1953 C. West (no. 53/2) collected a series of specimens from this locality, as did W. Scott in 1956, and these were identical with Beeby’s plant. The same plant was collected by C. West (no. 53/15) from the Laxo Burn in 1953. In 1959 W. Scott found it on rocks one-and-a-half miles north-east of West Burrafirth, Sandsting, where it was re-collected by C. J. Cadbury in 1960. Cadbury also collected it in two other localities, the Burn of Crookadale, Catfirth, Mainland, in 1960 and 1961, and near the South Wick of Sound, near Setter, Yell, in 1961. In 1962 W. Scott found it on rocks above the Bay of Brenwell, near Sandness, and in 1963 at Channer Wick in south Mainland. A plant collected on 29 July 1961 by R. C. Palmer from sea cliffs at the West side of Whale Firth, near Grimister, Yell, differs only in having few (six), widely spaced cauline leaves. The plants from Crookadale have narrower (almost linear) leaves than those from the other localities; this character is retained in cultivation. H. dilectum is very different from H. hethlandiae (F. J. Hanb.) Pugsley and has a closer affinity with H. subtruncatum Beeby, H. attenuatifolium P. D. Sell & C. West and H. pugsleyi P. D. Sell & C. West. It can be distinguished from H. pugsleyi and H. attenuatifolium by the sparse indumentum of its phyllaries, and from H. subtruncatum by its longer, narrower and more pointed leaves. The use of the epithet dilectum for this beautiful plant needs no explanation. 13. HIERACIUM SUBTRUNCATUM Beeby (1908) 114 (Lectotype: near the sea, Hamar Voe, Northmaven, Shetlands v.c. 112, 10 August 1891, W. H. Beeby no. 1105 in Herb. F. J. Hanbury (BM_); Druce (1922) 500; Roffey (1925) 29, no. 1192; Druce (1928) 70, no. 194; Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 225; P. D. Sell & C. West in Dandy (1958) 127, no. 558:1:173. PLATE 9a. [H. friesii sensu Beeby (1892) 55 (exclud. Loch of Cliff loc.); sensu Roffey (1925) 29, no. 1224 (quoad Shetland loc.); sensu Druce (1928) 71, no. 226 (quoad Shetland loc.).] [H. truncatum sensu F. J. Hanb. in Bab. (1904) 264; sensu W. R. Linton (1905) 75 (quoad Shetland loc.).] [H. rigidum var. friesii sensu W. R. Linton (1905) 82 (quoad W. H. Beeby no. 1040); sensu Beeby (1908) 115; sensu Druce (1922) 500 (exclud. Loch of Cliff loc.).] [H. carpathicum subsp. truncatum sensu Zahn (1921) 811 (quoad Shetland loc.).] [H. levigatum subsp. Friesii sensu Zahn (1922) 887 (quoad Shetland loc.).] H. subtruncatum var. glussburnense Pugsley (1948) 225 (Holotype: Banks by the Gluss Burn, near Ollaberry, Mainland, Shetlands v.c. 112, 5 July 1889, W. H. Beeby no. 1040 (BM)). : Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 99 Aphyllopodous. Stem to 75 cm. high, flexuous, usually purplish below, sometimes throughout, usually with pale, medium to long simple hairs throughout, those above dark-based, often densely so below, some- times only sparingly so or scaberulous, with scattered stellate hairs throughout, sometimes more dense near the inflorescence, without glandular hairs. Cauline leaves 8—20, gradually decreasing in size, pale to medium green, often suffused with purple; the Jowest broadly elliptic or oblong, obtuse, subentire or denticulate, narrowed (sometimes abruptly) to a sessile base or occasionally shortly petiolate, with numerous, pale, sub- setiform, bulbous-based, medium simple hairs on both surfaces and the margins; the median lanceolate to oblong, subacute, denticulate or with a few shallow teeth intermixed, contracted to a rounded, semi-amplexi- caul base, clothed as lower except that the hairs are fewer, the upper surface is sometimes glabrous, and the lower surface sometimes has scattered stellate hairs; the upper lanceolate to ovate-lanceolate, often gradually narrowed from the middle to an acute apex, denticulate, sometimes with a simple shallow tooth, base broad, rounded and semiamplexicaul, clothed like the median. Inflorescence narrowly and laxly cymose with 1-13 capitula, acladium with a long peduncle, peduncles long and slender, suberect, with few to numerous stellate hairs, few to numerous, usually dark-based, medium simple hairs, sometimes only the dark bases remaining, and occasionally a few, very short, dark glandular hairs. Phyllaries 9-11 mm. long, c. 1:5 mm. wide, incumbent in bud, blackish-green, inner with paler margins and often reddish tipped, linear-lanceolate, obtuse, without stellate hairs, usually with a few, very short, dark glandular hairs and solitary, dark, very short simple hairs. Ligules yellow, glabrous. Styles yellow. Margins of the receptacle pits shortly dentate. Cypsela dark, 3-5-4 mm. long. This species was originally collected on the banks of the Eala Water,* Northmaven, by W. H. Beeby in 1891 and 1896, and is represented by his exsiccatae nos. 1105, 1106, 1153 and 1154. It was referred by M. Elfstrand and E. F. Linton to the Scandinavian H. truncatum Lindeb. In 1908, however, Beeby, on the advice of W. R. Linton, described it as a new species, H. subtruncatum. It is indeed very different from H. truncatum, which has a nearly glabrous stem and foliage, petiolate basal leaves and a more spreading inflorescence. Later gatherings from this locality have been made by G. C. Druce (1920 and 1921), M. L. Wedgwood (1921), J. E. Lousley (1950), C. West no. 53/6 (1953), A. H. G. Alston, and N. D. Simpson no. 54/745 (1954), and W. Scott (1956). In 1953 C. West found this species on the banks of a stream on the north side of Ronas Voe, and W. Scott sent us a specimen from the same locality in 1957. There is a sheet in the Oxford herbarium from Cliva Hill, collected by G. C. Druce in 1921; and W. Scott found it near there at Islesburgh in 1963. Beeby (1902) quotes Hanbury as having identified a plant from Cliva Hill as H. truncatum, but the Beeby specimen in Herb. Mus. Brit. is H. hethlandiae. W. Scott collected a specimen in 1959 along the NE arm of the Mangaster Voe, which is probably this species, and another collected nearby in 1963 which is certainly it. In 1963 W. Scott found it at Channer Wick, which extends its range to the southern part of Mainland, and in the same year collected it at Hamna Dale, near Lunning, Lunnasting. In 1889 and 1891 Beeby collected specimens from the banks of the Gluss Burn, near Ollaberry, Main- land (nos. 1040 and 1104) which, together with the Eala Water plant, were referred by Hanbury to H. friesii Hartm. When Beeby described H. subtruncatum, however, the Gluss Burn plant, on the advice of W. R. Linton, was left under H. friesii. Here it remained until Pugsley (1948) pointed out its affinities with H. subtruncatum, and described it as variety glussburnense of that species. It has more recently been gathered by C. West no, 53/16 (1953), and by A. H. G. Alston and N. D. Simpson no. 53/746, and from their fine series of specimens we consider that it is distinguishable from the type only in the insignificant character of having fewer and shorter simple hairs on the stem and peduncles. The amount of purplish colouring is very variable in plants from all populations of H. subtruncatum. In 1892 (J. Bot. (Lond.) 47, 51) A. Ley recorded a plant from the Mellte Glen, Brecon as H. truncatum forma. This plant was brought to Ley in an undeveloped state and grown by him in his garden. No specimens gathered in the wild state exist, but we have seen the cultivated specimens. They are certainly neither H. subtruncatum, Beeby nor H. truncatum Lindeb., and do not match any described British species. Until wild material is available it is difficult to say whether it is, in fact, a new species. H. truncatum sensu F. B. W. White (1898) 203 according to his specimen in BM is H. subcrocatum (E. F. Linton) Roffey. H. subtruncatum is closely allied to H. pugsleyi P. D. Sell & C. West, H. attenuatifolium P. D. Sell & C. West and H. dilectum P. D. Sell & C. West. From H. pugsleyi and H. attenuatifolium it differs by the sparser indumentum of its phyllaries, and from H. dilectum by its shorter, broader and less pointed leaves. No extra-British species that is closely allied to H. subtruncatum has been seen, H. truncatum Lindeb. being easily distinguishable, as mentioned above, while H. friesii Hartm. which is quite different, is perhaps best placed in the Section Tridentata. * The modern official spelling is Eela, not Eala. Watsonia 6 (2), 1965. 100 P. D. SELL and C. WEST 14. HIERACIUM NORTHROENSE Pugsley (1948) 228 (Holotype: hillside, North Roe, North- maven, Shetlands, 30 July 1891, W. H. Beeby no. 1085 (SLBI) (the label also contains the date 11 August 1891 and no. 1086)); P. D. Sell & C. West in Dandy (1958) 127, NO: Do 77 PEATE 9b: [H. crocatum sensu Beeby (quoad North Roe loc.) (1892) 55.] H. crocatum var. congestum Beeby (exclud. no. 1634) (1908) 115 (Lectotype: hillside, North Roe, Northmaven, Shetlands, 30 July 1891, W. H. Beeby no. 1085 (SLBI); Druce (exclud. Holm in Burga Water loc.) (1922) 500; Druce (1928) 71 no. 245e. H. congestum (Beeby) Roffey (1925) 29, no. 1246, non Freyn (1891) Oest. Bot. Zeit., 41, 56. [H. northroense Pugsley in A. R. Clapham (1946) 34 nom. nud.] Aphyllopodous or with basal leaves withering at the time of flowering. Stem 17-35 (45) cm. high, flexuous, with numerous, pale, medium simple hairs at the base, usually less numerous above or scaberulous, with scattered stellate hairs, without glandular hairs. Leaves pale to medium green, often reddish-purple on the margins and veins; the basal elliptic, obtuse, regularly denticulate, attenuate to a short, winged petiole; the cauline (7—) 9-15, suddenly decreasing in size about half-way up the stem; the lower lanceolate to broadly elliptic, obtuse to acute, with numerous, fine, closely placed denticulations, narrowed to a semiamplexicaul base; the upper linear-lanceolate to lanceolate (to ovate-lanceolate in cultivation), subacute to acute, denticulate as the lower, base sessile, rounded and semiamplexicaul; all with short to medium, pale simple hairs below and on the margins, the lower sometimes have them above, all with numerous stellate hairs on the lower surface, the upper with stellate hairs on both surfaces. Inflorescence cymose, laxly so in cultivation, with 2-9 (—20 cult.) capitula, acladium with a medium peduncle; peduncles medium to long, suberect, with scat- tered stellate hairs, with a few, very short, dark glandular hairs and occasionally a few, pale, medium simple hairs. Phyllaries 9-11 mm. long, 1-5—2-0 mm. wide, incumbent in bud, green with a dark centre, lanceolate, obtuse, with scattered stellate hairs, with a few, very short or short, dark glandular hairs and sometimes a few pale, medium simple hairs in a row down the centre and at the base. Ligules yellow, glabrous. Styles discoloured. Margins of the receptacle pits shortly dentate. Cypsela dark, about 3-5 mm. long. H. northroense was discovered by W. H. Beeby in 1891 (nos. 1085 and 1086). There is no specimen to prove that Druce saw it on his visits to the Shetlands, even though he states (1922, p. 466) that he had re- gathered all Beeby’s critical plants. In the last few years several botanists have searched carefully for it without success. In 1959, however, C. J. Cadbury collected seeds from a plant growing at North Roe which he sent to us to cultivate. The plants cultivated from this seed in the Botanic Garden at Cambridge flowered freely in 1961, 1962 and 1963. They were H. northroense. The colony found by C. J. Cadbury has since been examined by R. C. Palmer, W. Scott, N. D. Simpson and C. West, and additional material obtained. It is very closely allied to H. vinicaule, but differs in its more dwarf, compact habit, more numerously denticulate shorter leaves, which are less drawn out at the apex, and in its paler, less densely clothed phyl- laries. The specimens cultivated by Beeby in 1894 and by us in 1961 and 1962 were taller, the ieaves even broader in proportion to their length, and the denticulations more pronounced. Our cultivated specimens in 1963, however, exactly matched Beeby’s original wild material. 15. HIERACIUM VINICAULE P. D. Sell & C. West (1955) 236 (based on H. vinaceum (Beeby) Pugsley, non Johansson & Samuelsson); in Dandy (1958) 127, no. 558:1:176. PLATE 10a. [H. crocatum sensu Beeby (quoad 3 loc. north shore Roeness Voe) (1892) 55.] [H. strictum sensu W. R. Linton (quoad W. H. Beeby no. 1156) (1905) 85.] H. crocatum var. vinaceum Beeby (1908) 115 (Lectotype: Roeness Voe, Northmaven, Shetlands, 1891, W. H. Beeby no. 1087 (BM)); Druce (1922) 500; (1928) 71, no. 245f. H. strictum var. humilius Beeby (1908) 115 (Lectotype: Gelli Burn, Hillswick, Northmaven, 13 August 1896, W. H. Beeby no. 1156 (SLBD); Druce (1922) 500; Roffey (1925) 29 no. 1237b; Druce (1928) 71 no. 239b. H. congestum forma vinaceum (B eby) Roffey (1925) 29 no. 1246. [H. platylepium sensu Druce (1928) 71, no. 246.] H. vinaceum (Beeby) Pugsley in A. R. Clapham (1946) 346; (1948) 227; non Johansson & Samuelsson (1923) Dalarnes Hieracia Silvaticiformia 90. | Aphyllopodous. Stem to 70 cm. high, flexuous, purple below, sometimes throughout, usually with scattered, pale, medium simple hairs throughout, with few stellate hairs especially above, sometimes with a few, very short, dark glandular hairs near the inflorescence. Leaves all cauline, 10—19 (—30 cult.), medium or Watsonia 6 (2), 1965. PLATE 3 ce ad a] aa “AQa0g Adatg WHLIDIALH (q) tpeope uted Crore x tet et eeyy & Yr ee ee rf Hb) pexueses HEUSIEW “6 3 4 9H & 7 OW Sv AG3I8 HOM 4 ida Af tO hee 7 Ye 34” 3 : SL tating are Oy ters ic naieays 8: ett We arby / LX TW Dp ee ngs Dd bay sean Ne ape ey) % LO ETE YO IH MIRUIG TEIID og ‘AqQoog WHoIpUubjjaz UN1IDIALIFY (V) TyGl HAMAR anny Seri pele WIN SRI91py IPeYsseLy “¢ “9 SiG} PRIVeSBIg ey qua SV¥ AGSISS HOM CONOR Why y Seg (re wren ee) Pe, £ coger OA URINE T2880 a Ag 7) eatin Pe bee cae ee ae Ah: [facing page 100 "ISOM “OD BY [19S ‘A ‘d Wnjoss wnisvda1f]T (q) “adA}OJTOH “ISOM “OD ® [19S (Cd ‘d Whos wnlovsalfy (dD) m1 we yy . ita att EEG. & QNRE MG 6 oN QEsVUSY’ oh ON ROT {AL PIM BISROT sgnoswetd? Seaeey "parnoj[oos Tp BaTsAS E56) SING : "ERO T/L/TX Tere ayen ES *s09G08/9T *B°D “SBIV TR BYE *3SUQ “th sedtaing : ee ‘yore wyetg Tatane omy fq 6. sto Aewwtg ‘Tryeaeng we qoo7y ce SS YEIGTH PHE GyR204 "tat A ARGU PU AIRS +21 : VER based Bae aces aur a acremmneoneeeil tad i" ; ‘wejs¢ Bayauxe seeking “AURGUET NS PROFIT ERO Ins MICOS TORT 4 ‘a { + TH S3%Si HAlLied QEIUED “ALLY “G45 SIUSi HSILISY HEIUED AWS] 845}, IAN D é acyl PIRI 99 27} WS Od ‘ Pegs Ce nel wegen PLATE 4 GHYERG £89 OAL i we w PARSE BS poner ‘Ag[ssnd (Aqsog) smyvajsnp wiilavsalpf{ (q) “Od AJO[OH “ISAM “DO ® [19S °C ‘d Alouffip wniovsaipy (v) SN ROD Yao 7S on wesdats yee “HeD ‘4 FOOTE OTE se “SENET og , WIETALSLE ; : “ss 8 “os To-worper : : TEPH/TY “HD . ‘ RES SSRERSS Puy psang ‘pur purm_ ‘fonbsaxg ay a year SPSLARMD ang oF Jo Seg ayy eange ower4e jo sueg TENG HY Put 4ye2s4 { x4 FEL Pus Ayer 4 2 A AotePay /dQood) EXYREN SAS Ear > : TSOR"D F TSR Hy STESTIGED SH poRte ty S3TS1 HSiLiVE GSIUED “ALE “ay S33Si HSiLIe8 ARIUET AUC} -GaaH4 PLATE 5 PLATE 6 LR Oe S181 peIuesesy Weysaepy “sy giay thi Gr (ty nape pe) tt FOI 8-H od Lag brn) FIT BVE RBBAS HOM £7 7) Fe G99 eg ey ey Fe SA UIE TID Aus “aware -ap -wmpaswe 44 fe praying HANS 2 e66y PARI us a4 i eis bay bay ie bal PEO AEN IBID “SOIL asuasaop wniavdalf] (V) oN aaAEe *g* fs HOD 2EUESS arse Biot ker hs che tyekory wag $*33 DOO'E TOaE FF BOS aqaca oy; WO seBpst JJtT> A400H yEngey poe AYLI} "933 A *wparg Boustasp i SRE | S318) HSIL148 qmUty AU GOH * PLATE 7 “adA1OJOH “ISOM 'O ® 19S °C ‘d “nyofipnuasp wniov. *SMOTTAL CQUGS < ‘ S SESh Se day SCQANEES AVOH SSPE TE EO TPUOTERY GEN QW! YF FO KAUAG douse FERQQY AVE GHHg 2ki SA TRAM Dae SIIB HSL ys & “ “GREED ARG] QAAH 11H (q) ‘adAjoypo7T “Agissng ‘quey ‘ff 4) aipunjyjoy wniovial] (VD) QE OR § ES) Ay EO ON hy x STS ae KH RPA CES FRR. e ¥ rer og @ Tieeta’s Eset (eeer) wer Poe ‘ang HEL *QuRR SETRUNTUSeT “TEA SSUGTACE °F (suhT} See PRREE TTOOe "ROL Seyesng (1quey) BOERS PHISH ER TSEIS TY SILOS! ee apd AF ODD font WAMVRSAN AEUSI THREES) RAERY BERLE 3 ae (eee em as Mem nGe ecco. ; BRE HALOS; BIBI “f - PLATE 8 ‘adA1O[OH{ “189M “O F [19S °C ‘d WnoajIp wntap4alH (9) ts ON 989276 1S Cook 1 Sieg Cote f{y tye ERUATRONE “UTT1 499 {gkoth go wing a4 go smug Asean, caevigngs pur 419e20} Zk KA 28On°O % TTA ‘ata PRASOT ED mmpousont S215) HStL1¥S ee goer aor ai ‘adA1OJOH “ISOM ‘OD ® [19S ‘Cl ‘d Majsénd winiopialH] (vd) WG OS 380% *D 1D tarred woteas Necgy + Bey LEGER JAN DSR euE “eTUTER Yama OTARSTAL SH OO sxe LeReay cme pur heyaoy Zk 4eouD B Tleeta'a Fas Taend moe Hy $3151 HSiL1¥s goa ELUED) AEE “OO PLATE 9 Re ‘AQ|SSNg ASUIOsYJIOU WNIIDAIIH (G) “Aqoog WNjDIUNAJGNS WUN1LIDAIFT (DV) ¢ Py Lay *s 42D coe pe sages # KeoIet es Boat BSBA US 33 Se} W359 Soar < See Bad Las} eS kre br “dhe MUU SEIS IRY “ES Teous ¢ Ec oy Sera “303 sys A insehar aiataentl re # ee oo . RTO AUQ I0 Usry 33 JO <3 SSoete . BINS} PVE AgipEIEy eh} wa STV ABSSS HM 2 éy- L- ce acer SHPSSERISGHS BOTS Eee car? can a SIVSi HSILIUG YUIBEA sia 4 (1p hae PR IT, rte Be we —) Gorge pry hen - Tee “6 SE RD Cra pee ee OA ‘RIESE JAD Ly | en = _ bib) parienedd ETDNT a ese “Sa WH ‘ad AJO]OH “189M (OF 19S °C ‘dd Wnpunau winiov1al (q) ISOM “OF [19S °C ‘d a/npoiuia winiovsalfx] (V) + lat renee oe | £E AES Soh paar *y HED ; BAL, ” we ON 1) ERIE IT BOTY 89 a yes 4 y 6f6i — aseuImEg thi * Bee ALLL FD } LR INO SOD encom LULA br LG LOE WS rep ang tv t% eo woos aT) TIS 30 vymg Sosa Fane PARE TB? LAO 4OOAS &s TB PAeA BI aE 4th PTY fae AYCIOY to, wh SSE es LGGIRA (SORE LODE. BEI BEB TURN OA BOOM STE BAUS $3751 HSI, ge WSILIHY BPE EN OPH GC, “ABSSY 9554 PLATE 10 PLATE I1 ‘Jossag wnolyipdivd wnigvdialf] (q) ‘OUAS UDMap WNlIDAA (V) ~ "ps FSS ess “HP vey ~ = bees yr ARAL 2 Ft rege eer) PING = yer OS ZOOS” ; amehige HVE es i an dd te : OH whey PORTO Bn| py AG ONE : ; SK PARESIIL} af RSG anions 615} POIueserg HEYSIR SS GH 616} Pewupso.sg beads HegseeW °6 9 92H WEE SEE ss * = + t _ os 7 ¢ 7 : i ; 7 ~~ Wits \ } - i a aed > foe kee oa o be ‘ary § Mi J ’ me Bee = : i - kK —- , Mies gyi Mout ‘ Pt ae? Wie ‘ z= ae t 7 P 7 a ean 3 z F ; ait i . ae 7 se - 7 - : Pal : a v = & BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 101 dark green, slowly decreasing in size upwards; the lower narrow-elliptic to elliptic-oblong, --acute, obscurely denticulate or with an occasional shallow tooth, attenuate below to short or medium, narrowly winged petioles, with sparse, pale, short to medium simple hairs on both surfaces, the margins and the petioles; the median and upper lanceolate, linear-oblong or narrow elliptic, subacute to acute, subentire to shallowly and remotely denticulate or occasionally dentate, abruptly or gradually narrowed below to a semiamplexicaul base, clothed like the lower except that simple hairs are sometimes absent on the upper surface, and stellate hairs are frequent below and rare above, and the margins are sometimes only scaberulous. Inflorescence narrowly and laxly cymose, with 2-42 (-350 cult.) capitula, with long, slender, suberect branches exceeding the medium peduncled acladium; peduncles medium to long, suberect, with scattered stellate hairs, and with dark-based, short to medium simple hairs and very short, dark glandular hairs in varying proportions, one or the other type of hair sometimes absent. Phyllaries 9-10-5 mm. long, 1:25—-1-5 mm. wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, sometimes narrowed above but always obtuse, with few scattered stellate hairs, with+-numerous, very short to short, dark glandular hairs and usually a few, medium, dark-based simple hairs. Ligules yellow, glabrous. Styles discoloured. Margins of the receptacle pits shortly dentate. Achenes dark, 4-4-5 mm. long. H. vinicaule was first found in 1891 by W. H. Beeby (nos. 1087, 1088 and 1089) at Roeness (Ronas) Voe, Northmaven, and referred to H. crocatum Fries. A sheet of no. 1088, now in Herb. Hanbury (BM), was sent to Elfstrand, who identified it as H. angustum var. platylepium Dahlst. In 1908 Beeby described it as a variety, vinaceum, of H. crocatum Fries, and Druce accepted this treatment. Roffey considered it to be a form of H. congestum (Beeby) Roffey. It has since been collected in this locality by R. J. Burdon and G. C. Druce (1920), J. E. Lousley (1950), C. West no. 53/5 (1953), A. H. G. Alston and N. D. Simpson no. 54/731 (1954), R. Palmer (1955), W. Scott (1956) and C. J. Cadbury (1961). In 1896 Beeby collected specimens (no. 1156) around Hillswick, and yet others (no. 1633) by the Hamari Water, which Dahlstedt suggested were a form or variety of H. strictum Fries. Beeby then named the plants H. strictum var. humilius, this treatment being accepted by both Druce and Roffey. Further material has been collected from these localities by C. West no. 53/17 (1953), R. Palmer (1956) and W. Scott (1956). Pugsley (1948), after examining all the available material of H. crocatum var. vinaceum and H. strictum var. humilius, united them under the specific name, H. vinaceum. On comparing material from these localities we found that it could be separated into two groups, but that no clear-cut characters could be found that would distinguish them. The leaves, even in dried specimens, seemed to be a slightly different colour, while var. humilius had a more strict stem and more regularly denticulate leaves; however, a series of specimens from the two areas must be consulted before these small differences can be appreciated. In the few cultivated specimens we have seen, the var. humilius retains its characters while var. vinaceum looks much more like var. humilius. In view of this, and of the fact that the characters they have in common are much greater than their differences, we have followed Pugsley in uniting them as one species. We have given the plant a new name because vinaceum had already been used for another species belonging to an entirely different Section. The epithet vinicaule was chosen because of the wine-coloured stems of many of the plants. The range of H. vinicaule has recently been greatly extended. In 1958 W. Scott found this species near Muckle Roe Bridge, in 1960 R. C. Palmer found it at Swining Burn, Lunnasting, in 1961 W. Scott found it in Yell at the lower end of North Burn, in 1961 C. J. Cadbury found it at North Roe, Mainland, and in 1963 W. Scott found it at Kels Wick, Lunnasting. Specimens collected by R. C. Palmer in 1960 on a holm in the Loch of Lumbister, Yell, probably belong to this species, but show slight differences in leaf shape and type of inflorescence. This is the tallest of the Shetland species of Alpestria, and although Pugsley gives the number of capitula as 2-4 (-9), C. West found that this number is often exceeded. In fact, a plant seen by the roadside near Hillswick, had 19 cauline leaves and 42 capitula. In cultivation at Cambridge, plants were produced with 30 cauline leaves and over 350 capitula. In its natural habitats H. vinicaule grows on steep grassy slopes of burns and sometimes near roadsides. The only closely allied species is H. northroense, from which it can be distinguished by its taller habit, less denticulate, longer leaves, and by its darker, more hairy phyllaries. 16. Hieracium mirandum P. D. Sell & C. West, sp. nov. PLATE 10b. Holotype: By aqueduct near Well House, Stean, Mid-W. Yorks., v.c. 64. Grid. Ref. 44/0774, 29 July 1962, Miss C. M. Rob (CGE). Ab H. dewarii Syme ligulis glabris, foliis ovatis, pedunculis pilis glanduliferis carentibus differt. Planta foliis basalibus florendi tempore marcescentibus. Caulis ad 38 cm. altus, flexuosus, inferne purpureus, ubique pilis simplicibus longis pallidis, superne paucioribus nigribasibus pilosus, pilis stellatis Watsonia 6 (2), 1965. 3 102 P. D. SELL and C. WEST glanduliferisque nullis. Folia mediocriter viridia, inferne caesia; basalia elliptica vel ovata, subacuta, denticulata, basi in petiolum mediocrum anguste alatum attenuata, utrinque et in marginibus pilis simplicibus mediocribus pallidis dispersis vestita; caulina 6—7, ovata, subacuta vel acuta, denticulata vel dentibus paucis brevibus praedita, inferiora basi cuneata petiolis brevibus, superiora basi rotundata, sessilia, semiamplexicaulia, superne glabra, inferne et in marginibus pilis simplicibus mediocribus pallidis paucis vestita. Anthela com- pacte cymosa, capitulis 3-5, acladio pedunculo mediocri praedito; pedunculi suberecti, mediocres, pilis stellatis densis, aliis simplicibus numerosis nigribasibus mediocribus, glanduliferis nullis vestiti. Squamae 9-10 mm. longae, 1-5—2 mm. latae, in gemmis incumbentes, atrovirides, interiores marginibus pallidioribus, lineari-lanceolatae, obtusae, praecipue basi pilis stellatis paucis, aliis glanduliferis brevibus et brevissimii atris, necnon aliis simplicibus paucis brevibus nigribasibus vestitae. Ligulae luteae, glabrae. Styli obscuri. Cypselam non vidimus. Basal leaves withering at the time of flowering. Stem to 38 cm. high, flexuous, purplish below, with pale, long simple hairs throughout, those above fewer and dark-based, without stellate or glandular hairs. Leaves medium green, caesious below; the basal elliptic or ovate, subacute, denticulate, attenuate to a medium, narrowly winged petiole, with scattered, pale, medium simple hairs on both surfaces and margins; the cauline 6-7, ovate, subacute to acute, denticulate or with a few shallow teeth, the lower cuneate with a short petiole, the upper rounded, sessile, semiamplexicaul, glabrous above with few, pale, medium simple hairs below and on the margins. Inflorescence compactly cymose, with 3—5 capitula, acladium with medium peduncle; peduncles suberect, medium, with dense stellate hairs and numerous, dark-based, medium simple hairs without glandular hairs. Phyllaries 9-10 mm. long, 1-5—2-0 mm. wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, obtuse, with occasional stellate hairs mainly at the base, with very short to short, dark glandular hairs and a few, short, dark-based simple hairs. Ligules yellow, glabrous. Styles discoloured. Cypsela not seen. This species is known only from two plants taken from the type locality, which is so remote from any other locality for an Alpestrian species that the specific epithet mirandum has been used. It was searched for in 1963 and 1964 without success. However, in view of the distinctness of the plant, and of the remoteness of the locality from that of any other species of the Section Alpestria, we have described it from what we should otherwise have considered to be insufficient material. It is most closely allied to H. dewarii, from which it can be distinguished by its glabrous ligules, ovate leaves, and peduncles without glandular hairs. 17. HIERACIUM DEWARII Syme (1878 for 1876) 23 (Lectotype: Linmill, Clackmannan, July 1876, J. T. Boswell-Syme in Herb. Boswell-Syme (BM)); (1879) 211; F. B. W. White (1898) 203; F. N. Williams (1902) 162; F. J. Hanb. in Bab. (1904) 265; W. R. Linton (1905) 73; Druce (1928) 70 no. 197 (1620); Roffey (1925) 29 no. 1195; W. Young (1936) 92; Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 229; A. R. Clapham in A. R. Clapham, Tutin, T. G. & E. F. Warburg (1952) 1145; ed. 2 (1962) 907; P. D. Sell & C. West in Dandy (1958) 128, no. 558:1:178. PLATE Ila. [H. dovrense sensu F. N. Williams (quoad Glen Roy loc.) (1902) 161; sensu W. R. Linton (quoad Glen Roy loc.) (1905) 72.] H. carpathicum subsp. dewarii (Syme) Zahn (1921) 812. Basal leaves withering before or at the time of flowering. Stem to 110 cm, high, flexuous, occasionally purplish below, with long, pale simple hairs throughout, less densely so above, the upper sometimes dark- based, with scattered stellate hairs above, usually without glandular hairs, but occasionally with a very short glandular hair near the inflorescence. Leaves 6—20, all except the lowermost widely spaced, gradually decreasing in size, Medium green, midrib prominent; the lowermost broadly elliptic or obovate, obtuse-mucronulate to acute, entire or remotely denticulate, attenuate to winged petioles; the lower and median similar to the lowermost but never obovate, and the petioles are either short or absent; the upper ovate or ovate-lanceolate, acute to long acute, entire or denticulate, with a rounded, sessile, semiamplexicaul base; indumentum variable, usually very hairy on both surfaces and the margins with pale, subsetiform, medium simple hairs, but sometimes the hairs are fewer or nearly absent. Inflorescence narrowly and compactly cymose, with —17 (—35 cult.) capitula, acladium with a short to medium peduncle; peduncles medium, suberect, with dense . stellate hairs, with more or less numerous dark-based, short simple hairs and --numerous, very short, dark, fine glandular hairs, the number of simple and glandular hairs, and the proportions of one to the other, varying greatly. Phyllaries 9-10 mm. long, usually c. 1-0 mm, wide, incumbent in bud, blackish-green, the inner with paler margins, linear-lanceolate, obtuse or rarely subacute, with scattered stellate hairs towards Watsonia 6 (2), 1965. BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 103 the base and on the margins, and with a tuft at the apex, with numerous short or very short, dark glandular hairs and few to many, short to medium, dark-based simple hairs. Ligules yellow, with very short simple hairs at the apex. Styles discoloured. Margins of the receptacle pits fimbriate-dentate. Cypsela dark, 3-5-4 mm. long. Distribution: V.c. 87. W. PERTH: 26/89, Menstrie Glen, 8 August 1882, J. Groves (BM); 26/99, Linmill, 1875, T. Drummond (CGE), 1876, J. B. Syme (BM, CGE), 1876, T. Drummond (BM, OXF), Glen of Sorrow near Dollar, 1876, J. B. Syme (BM) and 1877 (K); 27/42, Am Binnein, 1889, F. J. Hanbury (BM); 27/90, Glen Devon, 1876, J. B. Syme (herb. J. C. Melvill) and 1881 (BM), 1876, T. Drummond (CGE, K), Glen Quay, 1876, J. B. Syme (BM, CGE, OXF) and 1881 (BM, CGE), 1949, J. E. Lousley (K), 1959, P. D. Sell nos. 59/218 and 59/224 (CGE), N. D. Simpson no. 59/086 (herb. Simpson) and C. West (CGE). Lethensdene, Clackmannan, 26 August 1875, T. Drummond. V.c. 88. MID-PERTH: 27/32, Strath Fillan, 1891, E. S. Marshall (CGE and herb. J. C. Melvill), 1917, R. J. Burdon (CGE, K, OXF, herb. Univ. Sheffield, YRK), Crianlarich, 1889, E. S. Marshall (CGE), 1891, E. S. Marshall (BM), Glen Falloch, 1889, E. S. Marshall (BM, SLBI), 1914, E. S. Marshall no. 3996 (CGE, E), 1915, H. W. Pugsley (BM); 27/43, Ben Heasgarnich, June 1952, D. McClintock (herb. J. E. Lousley); 27/52, Lochearnhead, 1891, F. Buchanan-White (BM), By the Dochart, 1935, H. W. Pugsley (BM); 27/53, Killin, 1844, C. C. Babington (CGE)), 1848, I. B. Balfour (BM), 1888, F. J. Hanbury (BM), 1891, E. S. Marshall (CGE), 1891, E. F. Linton (Set of British Hieracia no. 47) (BM, CGE, K), 1891, W. R. Linton (OXF), 1929, H. W. Pugsley (BM); 27/64, Glen Lyon, 1913, E. S. Marshall (BM, CGE, E), Ben Lawers, 1887, E. S. Marshall (CGE, SLBD, 1888, F. J. Hanbury (BM, OXF), 1900, F. R. Tennant (CGE), 1951, T. G. Tutin (herb. Univ. Leicester); 27/74, Fortingall, 1891, W. R. Linton (LIVU), 1913, E.S. Marshall (CGE), 1913, C. E. Moss (CGE); 27/75, Keltney Burn, 1913, E. S. Marshall (CGE). V.c. 89. E. PERTH: 27/55, Loch Rannoch, C. E. Palmer (OXF). V.c. 96. E. INVERNESS: 28/70, By the River Spey, Kingussie, coll. unknown (BM). V.c. 97. W. INVERNESS: 27/06, Corran near Onich, 1913, N. D. Simpson no. 13/188 (herb. Simpson); 27/17, Glen Nevis, 1887, E. S. Marshall (CGE); 27/27, Choire Coille, 1896, E. S. Marshall and W. A. Shoolbred (as Linton’s Set of British Hieracia no. 47) (BM, CGE, herb. Leicester, herb. J. C. Melvill, OXF); 27/28, Glen Roy, 1896, E. S. Marshall and W. A. Shoolbred (as Linton’s Set of British Hieracia no. 46) (BM, CGE, E, K, OXF). V.c. 98. ARGYLL: 27/12, Dalmally, 1888, G. C. Druce (OXF), 1910, E. S. Marshall (BM, CGE), Loch Awe, 1898, H. J. Riddelsdell (BM); 27/15, west side of Buchaille Etive Beag, 1893, E. S. Marshall (BM, CGE, SLBI); 27/20, Ben Donich, 27 June 1962, A. G. Kenneth; 27/24, Clach Leathad, 1889, F. J. Hanbury (BM); 27/25, near Kingshouse, 1889, E. S. Marshall (BM, CGE), Cam Glen, 1899, C. Bucknall (K). V.c. 99. DUNBARTON: 27/20, Arrochar, Loch Long, 1842, C. C. Babington (CGE). H. dewarii was described in 1878 by J. T. Boswell-Syme from specimens gathered in Perth, Clack- mannan, Dunbarton and Stirling. The description is a good one, and all the original specimens belong to the same taxon. A lectotype has been selected from the many syntypes in the Boswell-Syme herbarium. H. dewarii seems to have caused little trouble taxonomically, and the great majority of the specimens we have examined have been correctly identified. Before it was described as a new species in 1878 it had been in- correctly referred by J. Backhouse and others to H. strictum Fries. Ley’s record (loc. cit.) of H. dewarii from an old lime kiln above Chapel-le-Dale, Yorks, is based on weak specimens of H. latobrigorum (Zahn) Roffey. Zahn (1921) considered H. dewarii to be a subspecies of H. carpathicum grex H. truncatum Lindeb., but it is much more closely allied to grex carpathicum Besser. Its usual habitat is on the banks of streams, but occasionally it occurs on rocky ledges. 18. HIERACIUM CARPATHICUM Besser (1809) 154 (typification by original description and by constant application of the name to the same taxon. No original specimens traced). PLATE 11b. H. dovrense var. spectabile E. S. Marshall (1894) 215 (Lectotype: Bank above the Lochsie Burn, in upper Glen Shee, E. Perth, c. 1250 ft, 28 July 1892, E. S. Marshall no. 896 (CGE); F. J. Hanb. in Bab. (1904) 264; F. B. W. White (1898) 203. Watsonia 6 (2), 1965. 104 P. D. SELL and C. WEST H. perthense F. N. Williams (1902) 163 (there is no evidence that Williams ever saw any material, and it is best regarded as a new name. for H. dovrense var. spectabile E. S. Marshall, with the same type); W. R. Linton (1905) 75; Roffey (1925) 29, no. 1196; Druce (1928) 70 no. 198 (1624); Pugsley in A. R. Clapham (1946) 346; Pugsley (1948) 230; P. D. Sell & C. West in Dandy (1958) 128 no. 558:1:179. H. carpathicum subsp. perthense (F. N. Williams) Zahn (1921) 815. Basal leaves withering before or at the time of flowering. Stem to 90 cm. high, flexuous, sometimes purplish at the very base, with long, pale simple hairs throughout but less dense above, the upper sometimes dark-based, with numerous stellate hairs above, without glandular hairs. Leaves 6-17, gradually decreasing in size, dark green, with a prominent midrib; the lowermost broadly elliptic or obovate, rounded-obtuse- mucronulate, entire or remotely and minutely denticulate, attenuate below to short, broadly winged petioles; the lower and median broadly elliptic, elliptic or ovate-lanceolate, obtuse to acute, denticulate to finely, sharply and remotely dentate, sessile, semiamplexicaul; the upper ovate-lanceolate or ovate, acute to long-acute, denticulate or finely and sharply dentate, with a rounded, sessile, semiamplexicaul base; all glabrous or nearly glabrous above, with medium, pale simple hairs on the lower surface and margins, the upper some- times with scattered stellate hairs on the lower surface. Inflorescence narrowly and compactly cymose, with 3-12 (20 cult.) capitula, acladium with medium to long peduncle; peduncles medium to long, suberect, densely stellately hairy, usually with a few, short to medium, pale, dark-based simple hairs, rarely with a very short dark glandular hair. Phyllaries 10-13 mm. long, 1-1-5 mm, wide, incumbent in bud, blackish-green, inner with paler margins, linear-lanceolate, obtuse, stellately hairy, densely so on the margins, sometimes with a Slight tuft at the apex, with dense, short to medium, pale, dark-based simple hairs and numerous, very short to short, dark, slender glandular hairs. Ligules yellow, with dense, very short hairs at the apex. Styles dis- coloured. Margins of the receptacle-pits fimbriate-dentate. Cypsela dark, 44:5 mm. long. In 1894 E. S. Marshall described a new taxon, H. dovrense var. spectabile, from specimens collected in 1892 from two places near Glen Shee, E. Perth. It has since been found there only by J. E. Raven (1953) and P. D. Sell no. 56/38 (1956). On both occasions only a few plants were found. In 1889 E. F. Linton had collected specimens from Clova which proved to be the same species. It has not been recorded again from that locality. Cultivated specimens of the original gatherings from both localities were distributed as Linton’s Set of British Hieracia no. 72. One of these cultivated gatherings was presumably seen by Williams (1902) when he raised the variety spectabile to specific rank, but we can trace no specimen annotated by him. The new epithet perthense was used owing to the earlier use of spectabile for another species. We have regarded H. dovrense var. spectabile as the basionym of H. perthense, the lectotype of the variety therefore being the lectotype of the new epithet. All gatherings have been from grassy riversides. Zahn (1921) placed H. perthense under H. carpathicum as a subspecies, but we are unable to distinguish it from his subsp. carpathicum. The type of H. carpathicum cannot be traced, but all material identified as that species which we have examined belongs to the same taxon, and agrees with the description published by Besser. H. carpathicum is at present known only in a few localities in the Tatra Mountains of east central Europe and at the above-mentioned localities in central Scotland. Such a distribution suggests that mor- phologically similar plants may have arisen independently in different localities. No closely allied species in the Section is found on the main land-mass of Continental Europe, its closest allies being in Scandinavia. In Scotland it is closely allied to H. dewari, from which it can be distinguished by its more dentate leaves, longer capitula and stellately hairy phyllaries. REFERENCES Beesy, W. H. (1889-1890). On the Flora of Shetland. Scot. Nat. NS, 4. Beesy, W. H. (1891). A New Hieracium. J. Bot. (Lond.), 29. Beesy, W. H. (1892). On the Flora of Shetland. Ann. Scot. Nat. Hist. Beesy, W. H. (1908). On the Flora of Shetland. Ann. Scot. Nat. Hist., no. 66. Besser, W. S. J. G. (1809). Primitiae Florae Galliciae Austriacae utriusque. Vienna. Biytt, M. N. (1838). Botanisk Reise i Sommeren 1836. Nyt. Mag. f. Naturvid. CLAPHAM, A. R. (1952), in Clapham, A. R., Tutin, T. G., and Warburg, E. F. Flora of the British Isles. Cambridge. Ed. 2. 1962. Cambridge. Druce, G. C. (1922 for 1921). Flora Zetlandica. Rep. Bot. Soc. Exch. Cl. 6 Suppl. Druce, G. C. (1925 for 1924). Additions to Flora Zetlandica. Rep. Bot. Soc. Exch. Cl. 7. Watsonia 6 (2), 1965. — BRITISH SPECIES OF HIERACIUM SECTION ALPESTRIA 105 Druce, G. C. (1928). British Plant List. Ed. 2. Arbroath. EpMonpsTON, T., revised by SAxBy, C. F. A. (1903). Flora of Shetland. Ed. 2. Edinburgh and London. Fries, E. (1848). Symbolae ad Historiam Hieraciorum. Upsaliae. Fries, E. (1862). Epicrisis Generis Hieraciorum. Upsaliae. HAnpury, F. J. (1894). Notes on British Hieracia. J. Bot. (Lond.), 32. HAnbBuvrRY, F. J. (1904), in Babington, C. C. Manual of British Botany. Ed. 9. London. LINDEBERG, C. J. (1874), in Blytt, M. N. Norges Flora. Christiania. Linton, W. R. (1905). An Account of the British Hieracia. London. MarsHALL, E. S. (1894). New variety of Hieracium dovrense Fries. J. Bot. (Lond.), 32. MARSHALL, E. S. & SHOOLBRED, W. A. (1898). Notes of a Tour in North Scotland, 1897. J. Bot. (Lond.), 36. OMANG, S. O. F (1906), in Blytt, M. N. Haandbog i Norges Flora. Christiania. Pucs.ey, H. W. (1946), in Clapham, A. R. Check List of British Vascular Plants. J. Ecol. 33. Puacs.ey, H. W. (1948). A Prodromus of the British Hieracia. J. Linn. Soc. Lond. (Bot.). 54. Rorrey, J. (1925a). London Catalogue of British Plants. Ed. 11. Rorrey, J. (1925b). The Hieracia of the London Catalogue. J. Bot. (Lond.), 63. SELL, P. D. & WEsT, C. (1955). Notes on British Hieracia I. Watsonia 3. SELL, P. D. & WEsT, C. (1958), in Dandy, J. E. Check List of British Vascular Plants. London. Syme, J. T. B. (1878 for 1876). Rep. Bot. Soc. Exch. Cl. Syme, J. T. B. (1879). Description of Hieracium Dewari, a New Species. Trans. Bot. Soc. Edinb. 13. Tate, R. (1866). Upon the Flora of the Shetland Isles. J. Bot. (Lond.), 4. Waite, F. B. W. (1898). The Flora of Perthshire. Edinburgh. WILLIAMS, F. N. (1902). Prodromus Florae Britannicae. 1. Brentford. Youna, W. (1936). A List of the Flowering Plants and Ferns recorded from Fife and Kinross. Edinburgh. ZAHN, K. H. (1921-1923), in Engler, A. Pflanzenreich. 1V, 280. Compositae-Hieracium. Leipzig. Watsonia 6 (2), 1965. FRUIT VARIATION IN POL YGON UM PERSICARIA L. By J. TIMson Botany Department, University of Leicester* ABSTRACT The variation found in the fruits of Polygonum persicaria L. is described and analysed. A possible explanation of the variation is discussed. A new fruit type (tetragonal), believed to be previously unrecorded, is described. INTRODUCTION The ‘seeds’ of Polygonum persicaria are in fact the fruits (nuts) which fall from the parent plant with the dead perianth still attached. Styles (1962) found the relationship between the length of the nut and the length of the persistent perianth a useful character in section Polygonum, but it would appear to be of no value in section Persicaria. The phenomenon of polymorphism of fruits or seeds is not common. The best-known example outside the Polygonaceae is Rhynchospora, which has fruits which are usually biconvex and rarely trigonous. In other members of the Cyperaceae, e.g. Carex, Eleocharis and Schoenoplectus the number of stigmas and hence the shape of the fruit is used as a species-separating character. Within the Polygonaceae only the genus Polygonum is recorded as having polymorphic nuts and within this genus polymorphy occurs only in sections Persicaria and Echinocaulon. In P. persicaria the two fruit shapes found are: (i) lenticular or biconvex to planoconvex which arise from flowers with two styles and lenticular ovaries; (ii) angular or trigonous with three concave faces which arise from flowers with three styles and trigonous ovaries. There is no apparent correlation between the position of the flower in the inflorescence and the type of nut produced. Both fruit shapes may be found on the same plant, and although considerable variation has been found in the relative numbers of each type, no plants have been found which have nuts of only one type. This situation may be com- pared with that found in P. Japathifolium in which the percentage of trigonous nuts is always very low and where plants may readily be found on which there are only biconcave nuts (Timson 1963). METHODS The length and maximum breadth of the nuts (without perianth) were measured in millimetres to the nearest 0-1 mm, each set of measurements being on a sample of mature nuts usually 5O or more in number. An index was obtained by dividing the length by the breadth. The arithmetic mean, standard deviation, and the range of these measurements were calculated. The average nut weight in mg of a sample was obtained by weighing a known number of nuts, usually 100 and never less than 30. RESULTS The averages of the data for all the nuts of P. persicaria measured are given in Table |. From this it is clear that the British and foreign material are essentially similar. The comparative column for P. lapathifolium shows that the two species are significantly different in index and in percentage of trigonous nuts (see also Timson 1963). Styles . (1962) found that fruit length was a useful character in section Polygonum for the separation * Present address: Department of Chemistry and Biology, Hatfield College of Technology, Hatfield * Herts. 106 Watsonia 6 (2), 1965. FRUIT VARIATION IN POLYGONUM PERSICARIA L. 107 of the species, but the data presented here suggest that it is of little value in section Persicaria. In this section the nuts not only show a considerable range of size on a given plant, but also the average lengths for the different species are very similar. Using nuts collected in 1958 near Ryston, Norfolk, plants were grown from nuts of known shape in similar conditions and the percentages of trigonous nuts determined. This was repeated over three seasons and the results are given in Table 2. From these results it seems that the shape of the nut from which the parent plant has been grown can affect the relative proportions of trigonous and biconvex nuts. The age of the parent plants also seems to affect the relative abundance of trigonous nuts. This was discovered when the nuts of a number of plants which had been allowed to grow old (i.e. leaves dried up, stem very dry and red), in the greenhouse were examined. They were found to be 73 per cent trigonous and 26 per cent biconvex. A third shape, not previously recorded, was present in small numbers (1 per cent)—four-sided with four concave faces (tetragonal). There are few records in the literature of the weight of nuts of P. persicaria. Simmonds (1945) gives 2:3 mg; Korsmo (1934), 2:7 mg; and Pammel (1910), 1-41 mg. It is difficult to compare these figures with those given here because the authors do not say TABLE 1. Combined data for all fruits measured of P. persicaria. P. persicaria P. lapathifolium British Foreign Total Length mean 2:61 2-05 2-33 2-64 s.d. 0-11 0-14 0-12 0-13 range 2:0-3:2 1-7-3-0 1-7-3-2 2:0-3-3 Breadth mean 2-01 1-99 2:00 e721 s.d. 0-12 0-12 0-12 0-13 range 1-4-2-7 1-5-2-4 1:4-2-7 1:7-2:9 Index mean 1-31 1-28 1-30 1-17 Sid. 0-07 0-06 0-07 0-045 range 1-1-1-6 1-1-1-6 1-1-1-6 1-0-1-5 Shape (%) biconvex 76 80 78 99-2 (biconcave) trigonous 24 20 2p) 0-8 Average weight (mg) Day 1-93 Zl 2 Si TABLE 2. Changes in the proportions of trigonous nuts in plants grown from nuts of known shape. % Trigonous Year From trigonous From biconvex From total % of total from plants grown from trigonous nuts 1958 = poe 24 ike 1959 As, 26 25 63 1960 60 34 43 64 1961 45 8 26°5 78 Watsonia 6 (2), 1965. 108 J. TIMSON whether or not the persistent perianth was removed before weighing. If, as is probable, it was not removed, it is difficult to be sure that only mature nuts have been weighed. For this reason the results given in Table 3 are for mature nuts without perianth. TABLE 3. Weights of fruits of P. persicaria. Year of growth Grown in Shape Average weight (mg) Origin: BuDE, CORNWALL 1959 Nature Total pA | 1960 Garden Biconvex 2:64 Fes of Trigonous 2j0777. ns 48 Total 2-93 1961 Greenhouse Biconvex 2-36 i “fi Trigonous 2-61 a x Total 2-49 Origin: KEELE 1959 Nature Total 1-69 1960 Garden Biconvex 2-26 Mes Ae Trigonous 2-74 * - Total 2-50 1961 Greenhouse Biconvex i167, a — Trigonous 2-96 a os Total Pe | There is considerable variation in the results obtained and it seems probable that the environment of the parent plant is the main factor. Thus nuts harvested in nature where the parent plant is in competition with other plants were usually lighter than those grown in cultivation where the plants were spaced out and competitors removed. Those produced in the garden were usually heavier than those produced in the greenhouse, probably because more space per plant was available in the former. It is clear that the trigonous nuts are normally heavier than the biconvex nuts produced on the same plant. DISCUSSION In P. persicaria the control of nut shape appears to be complex. The main conclusions emerging from the results are that: 1. It is possible to select to some extent for nut shape; 2. The nuclear genes do not completely control the nut shape, since the three kinds of nut may occur together on one plant. 3. More trigonous nuts are produced as the plant becomes senescent. From this it seems that the production of trigonous nuts is caused by some interaction between genetic and other factors; possibly their production is favoured by a breakdown of some metabolic mechanism which occurs more frequently during senescence. ACKNOWLEDGMENT My grateful thanks are due to Professor T. G. Tutin, under whose supervision this work was carried out, for his advice and encouragement. REFERENCES Korsmo, E. (1934). Explanation to Prof. Dr Korsmo’s Weed Plates, Series 1. Oslo. PAMMEL, L. H. (1910). Weeds of the Farm and Garden. New York. Stmmonps, N. W. (1945). Biological flora of the British Isles. Polygonum L. J. Ecol. 33, 117-143. STYLES, B. T. (1962). The taxonomy of Polygonum aviculare and its allies in Britain. Watsonia 5, 177-214. Timson, J. (1963). The taxonomy of Polygonum lapathifolium L., P. nodosum Pers., and P. tomentosum Schrank. Watsonia 5, 386-395. Watsonia 6 (2), 1965. LEAF MEASUREMENTS AND EPIDERMIS IN POA ANGUSTIFOLIA By D. M. BARLING Biology Department, Royal Agricultural College, Cirencester ABSTRACT In Poa angustifolia and P. subcaerulea leaf-blade length on fertile tillers increases and then decreases to the flag leaf. The third leaf-blade down from the panicle is usually longest. Leaf sheaths increase progres- sively to the flag leaf. Total leaf length is greatest in the leaf below the flag leaf. The per cent. contribution of sheath in the leaf increases to the flag leaf and sterile tillers show the reverse. Individual tillers may show modified patterns. The leaf blade epidermis of P. angustifolia was studied, and the lower of sterile tiller leaves contained none or very few stomata and were of a xeromorphic type, whilst the leaves of the fertile tiller showed stomata increasing in quantity to the flag leaf. Inflorescence initiation is in December-— January when the leaves showing the above patterns on fertile tillers are all in the primordial stage. INTRODUCTION The leaves are of taxonomic importance in the Gramineae, for where floral character- istics of closely related species are very similar, vegetative features are commonly used as aids to identification. There is considerable variation in leaf dimensions in grass species both in the vegetative tiller (Langer 1958), and also in the fertile tiller (Borrill 1959 and 1961a). Leaf blade length is known to be affected by temperature, daylength and other factors (Langer 1954, Stuckey 1942), particularly in the sterile tillers. The variation in fertile tillers was suggested by Evans (1949) to be the result of apical dominance (Meyer & Anderson 1939). Borrill (1959) has suggested that it is due to the development of inflorescence primordia when the longest leaf blade has emerged and that subsequent leaf blades are shorter due to decrease in cell extension, not number. This work has done much to clarify the factors associated with varying leaf blade dimensions in the grasses. The culms of grasses also show considerable variation in internode length with increasing length up the shoot (Evans 1927, Cooper 1956, Barling 1959). Sgrensen (1953) has used epidermal characters with species of Puccinellia found in Greenland and so have Burr & Turner (1933), Tutin (1955), Melderis (1955), and Borrill (1961b) in other species, whilst Prat’s (1932 and 1948) work has thrown much light on the epidermal characteristics of the grasses. The data presented here are chiefly from transplants of Poa angustifolia and P. sub- caerulea grown in 8-in. pots of calcareous loam immersed in washed gravel at Cirencester (lat. 51° 43’ N., long. 1° 57’ W.) at an elevation of 440 ft. (130 m), under natural daylength. STERILE TILLERS The contribution of leaf blade and sheath to total leaf length in P. subcaerulea (Table 1) is typical. In the early part of the season the relative proportion of blade increases up the tiller. This trend may be reversed in leaves formed in the autumn when the blades shorten probably as a result of decreasing photoperiod. The sheaths may also decrease in length later in the season. The blade is always asymmetrical, with a marginal indentation near the apex. FERTILE TILLERS The leaf measurements from transplants of P. angustifolia and P. subcaerulea have shown a characteristic and similar pattern in the relative development of leaf blade and sheath as well as in total length (Table 2). The third leaf blade down from the panicle is usually the longest with progressively shorter blades above and below. However, the 109 Watsonia 6 (2), 1965. 110 D. M. BARLING TABLE 1. P. subcaerulea: % contribution of blade and sheath to leaf length in early (EB) and late (L) growth of sterile tillers. Origin Time Leaves in ascending order ie 2: 3: 4. Slapton Sheath E 55-56 38-04 28-48 22°73 Blade 44-44 61-96 iiiicays TIe2T Craig-y-Llyn Sheath E 66:51 25-30 2273 — Blade 53-49 74-70 VT 27 — Bwllfa wall Sheath L 30-46 25°32 22-15 23°23 Blade 69-54 73-68 V-20 66-67 Montrose Sheath L 80-65 64-10 37-50 28-53 Blade 19°35 35-90 62-50 61-47 TABLE 2. Average values of leaf and internode lengths for P. angustifolia and P. subcaerulea (mm.). Blade Sheath Total Internode P. angustifolia F 32-86-+2:35 102:29+4:-10 135-15+6:81 199-57+ 9-20 (transplants 1958) 2 69-43+3-10 93 -43+3-25 162-86-+6:-63 127-71+ 6:71 3 81:57+3-90 67:29+3:-20 148-86+6:90 70-43+ 6:80 4 70:29+3:-40 33-71+2-91 104-00+9-50 14-57+ 1-21 P. angustifolia F 36:39+2-65 121-74+4:-97 162:44+6:-63 323-78+11-:70 (wild plants) 2 71-13+3:-30 114-65+3:-98 185-65+6:44 164:70-+7:40 3 92-87+4:21 75-39+3:-77 164-09-+7-00 73-00+7:-01 4 61-00+6-30 51-78+3-25 112-91+9-09 18-22+0-77 P. subcaerulea F 35-50 96-43 131-93 149-86 (transplants 1958) 2 70:50 81-50 152-00 79-79 3 75-64 53:57 129-21 28:71 4 72-07 41-64 113-71 15-29 pattern within individual tillers is not necessarily constant as sometimes the second or fourth leaf blade is longest. The leaf sheaths consistently increase in length up the culm and, like the blades, show considerable variation in length from plant to plant. The culm itself shows the usual pattern of increasing internode length up to the panicle. The sheaths of the lower nodes are often longer than the internodes above them, whilst those of the top node are normally shorter, though exceptions to the latter are common. The ratio of sheath : blade is lowest at the flag leaf and increases down the fertile tiller (Table 3). Whilst these characteristic patterns are seen in leaf parts, total leaf length is greatest in the leaf above the one with the longest blade (Table 2). A conspicuous feature of P. angustifolia is the difference in leaf characters in a tiller that is sterile in one season and fertile in the next (Barling 1959). The sterile leaves are long and narrow and of a xeromorphic type, whilst those on the fertile tiller are shorter but much broader (Fig. 1). Watsonia 6 (2), 1965. LEAF MEASUREMENTS IN POA ANGUSTIFOLIA 111 TABLE 3. Ratio of sheath : blade, and % sheath in the four top leaves of fertile tillers. angustifolia ratio % F 0-32 75-68 . 2 ; 57-36 3 1-21 45-20 + 2-10 32-41 subcaerulea ratio % Pos 37/ 73-09 0-87 53-61 1-41 41-45 Logs) 36-61 LEAF-BLADE EPIDERMIS IN P. angustifolia In P. angustifolia the lower epidermis of the leaf-blade from sterile tillers had cells with folded walls and cork cells throughout the inter-nerve area. The nerves had narrower cells and silica-suberose couples, though there was variation in cell size in the inter-nerve areas. The major feature of the lower epidermis was the absence or limited occurrence of stomata (Fig. 2 and Table 4). The upper epidermis had more or less smooth-walled cells with the stomata in rows and asperites prominent (Fig. 1). Long unicellular hairs or asperites were found in the inter-nerve epidermis. 150 130 110 & 90 Ee = 2 Soe 50 e 30 3 Ae o = 10 Megs 515 4° 3 2 flag Sterile tiller fertile tiller Fig. 1. _ Leaf blade length and width (broken line) sterile and fertile con- | dition of a tiller of Poa angustifolia. + Inflorescence initiation. Watsonia 6 (2), 1965. a Lower epidermis sterile tiller blade Upper epidermis sterile tiller blade ee Lower epidermis flag leaf 0-1mm Upper epidermis flag leaf Fig. 2. Poa angustifolia leaf blade epidermis. Upper diagrams from sterile tiller leaves. Lower diagrams from flag leaf. 112 D. M. BARLING On the fertile tillers, by contrast, the lower epidermis of the flag leaf had numerous stomata (Table 4). The upper epidermis was of the type described for sterile leaves and the leaf margins showed asperites in all cases. As there was such a marked difference in the epidermal characters of the flag leaf blade and sterile tiller blades, this feature was investigated in all the leaves of the flowering tiller. The blade below the flag leaf showed stomata on the lower epidermis, but the third leaf down, which was longest, varied con- siderably. In some tillers there were abundant and well-distributed stomata, in others they were limited to the region near the mid rib. The fourth leaf down showed very few stomata on the lower epidermis in the majority of cases. TABLE 4. Stoma density in P. angustifolia and P. subcaerulea transplants. Stomata/0-7 mm? Upper epidermis Lower epidermis P. angustifolia Deer Park No. 1 Flag leaf 36:92+3-72 9-96+0-76 Sterile tiller leaf 63-76+3-68 0-00+0-00 Daneway 18 Flag leaf 58 -88+3-16 18-96+0-84 Sterile tiller leaf 59-08+4-18 0:04+0-04 P. subcaerulea Cefn-y-Gyngon Flag leaf 28-12+1-07 10-20+0-82 Sterile tiller leaf 46-36+2-78 10-40+0-70 INFLORESCENCE INITIATION In P. angustifolia the inflorescence primordia are first found in late December, but are easier to find in January and February. During this period the field populations tend to lose all foliage through winter burn, and the leaves that emerge on these fertile tillers in spring are to be found as primordia beneath the panicle primordium. The stages of primordial development in P. pratensis have been illustrated by Evans (1949). The stages in P. angustifolia show that the vegetative apex with its leaf primordia first elongates and develops large folds, whilst later these show deeper folds, the beginnings of branching. At this and subsequent stages the number of potential spring and early summer leaves can be counted and usual numbers are not more than five. Similar features are to be found in P. subcaerulea. DISCUSSION Borrill (1959) demonstrated the influence of the development of the inflorescence primordia on blade length in G/yceria, and noted that there is occasional disturbance by external factors. He found that the primordia developed as the longest leaf emerged, subsequent leaves being shorter, and in the species studied inflorescence initiation occurred in spring. In P. angustifolia there is winter initiation and at this time the leaves in the bud are usually four to five in number and are the ones that emerge on the extending culm in spring and early summer. These leaves show the same pattern as found by Borrill, although, unlike the leaves of G/yceria, they all emerge after inflorescence initiation. Leaf-sheath length is closely correlated with increases in stem internode, but total leaf length shows no clear association with inflorescence initiation or internode length. It is also clear that the percentage contribution of the blade and sheath to total leaf length decreases and increases respectively in succeeding leaves. The annual species Lolium temulentum (Borrill 1961a) shows similar increases in sheath length, but total leaf length also increases to the flag leaf which is different from that found in the present species which Watsonia 6 (2), 1965. LEAF MEASUREMENTS IN POA ANGUSTIFOLIA 113 are, of course, perennials. Probably differing patterns occur in different Gramineae, related to timing of inflorescence initiation and perenniality of species. The epidermis of the lower surface of the narrow leaf blades shows an absence of or very few stomata, and it is reasonable to assume that this is a xeromorphic feature tending to increase its adaptability to dry areas, since the closure of the leaf by folding of the upper surface will effectively reduce water loss. On the fertile tiller, however, the narrow leaf blades with xeromorphic features change to broad leaf blades in which there is a progressive increase of stoma number in the lower epidermis. This reversal of xero- morphic characters is in striking contrast to many dicotyledons, and is possibly associated with the important part played by the leaves of the fertile tiller, particularly the flag leaf, in the filling of the caryopsis, as has been demonstrated in barley (Archbold 1945; Watson ef a/. 1958). Further, these broad leaves of the flowering tiller would be active in British conditions in the relatively moist periods of spring and early summer when water economy is not critical. REFERENCES ARCHBOLD, H. K. (1945). Some factors concerned in the process of starch storage in the barley grain. Nature, Lond. 156, 70-73. BarRLING, D. M. (1959). Biological studies in Poa angustifolia. Watsonia 4, 147-168. BorriL1, M. (1959). Inflorescence initiation and leaf size in some Gramineae. Ann. Bot. Lond. N.S. 23, 217-227. BorriL1, M. (1961a). The developmental anatomy of leaves in Lolium temulentum. Ann. Bot. Lond. N.S. 25, 1-11. BorriLi, M. (1961b). Dactylis marina Borrill, sp. nov., a natural group of related tetraploid forms. J. Linn. Soc. (Bot.) 56, 368, 431-439. Burr, S. & TURNER, D. M. (1933). British Economic Grasses. London. Cooper, J. P. (1956). Developmental analysis of populations in the cereals and herbage grasses. I. J. Agric. Sci. 47, 262-279. Evans, M. W. (1927). The life history of Timothy. U.S.D.A. Dept. Bull. 1450. Evans, M. W. (1949). Kentucky Blue-grass. Ohio Expt. Sin. Res. Bull. 681. 1-52. HUBBARD, C. E. (1954). Grasses. Penguin Books, London. LANGER, R. H. M. (1954). A study of leaf growth in Timothy. J. Brit. Grassl. Soc. 9, 275-284. LANGER, R. H. M. (1958). A study of growth in swards of Timothy and Meadow Fescue. J. Agric. Sci. 51, 347-352. Me Lperis, A. (1955). Species problems in recent Scandinavian works on grasses. In Fousley, J. E. (ed.) Species Studies in the British Flora, 141-159. B.S.B.I., London. Meyer, B. S. & ANDERSON, D. B. (1939). Plant Physiology. D. Von Nostrand Co. Inc., New York. PraT, H. (1948). Histo-physiological gradients in plant organogenesis. Bot. Rev. 14, 603-643. SORENSEN, T. (1953). A review of the Greenland species of Puccinellia Parl. Medd. Gronland 136, 5-146. STuCcKEY, I. H. (1942). Some effects of photoperiod on leaf growth. Amer. J. Bot. 29, 92-97. THORNE, G. N. (1959). Photosynthesis of lamina and sheath of barley leaves. Ann. Bot. Lond. 23, 365-370. TuTIn, T. G. (1955). Some problems in plants with reduced floral structure. In Lousley, J. E. (ed.) Species Studies in the British Flora, 21-26. B.S.B.I., London. WATSON, D. J., THORNE, G. N., & FRENcH, S. A. W. (1958). Physiological causes of differences of grain yield between varieties of barley. Ann. Bot. Lond. N.S. 11, 321-352. Watsonia 6 (2), 1965. SULPHUR AND THE DISTRIBUTION OF BRITISH PLANTS By H. J. M. BOWEN The University, Reading ABSTRACT The extent of recent additions of sulphur from fertilizers and atmospheric pollution to British soils is pointed out. Gymnosperms are especially sensitive to sulphur pollution, while crucifers and some other plants are unusually tolerant. It is possible that the diminution of some species, such as Juniperus communis, in Britain is a result of atmospheric pollution, as is the spread of many crucifers such as Cardaria draba. The distribution maps of many alien crucifers can be correlated with a map of sulphur dioxide pollution. Many native crucifers restricted to coastal habitats may be so restricted because their requirement for sulphur is satisfied by sea-spray. Sulphur is a common element in the soil which is essential for the growth of plants. During the present century the amount of sulphur in British soils and plants has increased, partly from the use of fertilizers, but more substantially from pollution of the atmosphere. The present note attempts to correlate some recent changes in the distribution of plants with sulphur pollution in Britain, and suggests that further studies of the problem should be made. SULPHUR IN SOILS Sulphur is universally present in British soils, largely as the sulphate anion, but there are no data from which a map of sulphur abundance could be drawn. In the north and west of Britain, on soils derived from igneous rocks, sulphur is four or five times more abundant in soils derived from basalt and other basic rocks than it is in granitic soils. It is also unusually abundant in regions where certain minerals are common, as in parts of Cornwall. South and east Britain are mostly covered by soils derived from sedimentary rocks. In these rocks the relative abundance of sulphur increases in the order sandstones, limestones, shales and coal deposits: sulphur deficiency is most likely to occur in sandy soils. Soils unusually rich in sulphur are found near open-cast coal mines and above clays containing the sulphide of iron (iron pyrites). Calcium sulphate or gypsum occurs in large amounts in triassic clays in a narrow band stretching from Watchet in Somerset to Hawton in Nottinghamshire, while strontium sulphate or celestine occurs in a small area near Bristol, including the Avon gorge. Barium sulphate or barytes is found very locally in several parts of Britain, for example, the Derbyshire dales and Ben Bulben in Sligo. Sulphur constitutes 4 per cent of the solids in fresh water, and 2:5 per cent of the solids in sea water, so that it must be unusually abundant in salt marshes and on cliffs and beaches exposed to salt spray. Thomas et a/. (1950) have shown that halophytes usually contain far more sulphur than do other plants. Agricultural soils are subjected to a steady drain on their sulphur content, since a ton of crop usually contains between 1-5 and 10 kg of sulphur. During the present century most agricultural land in Britain has received heavy dressings of sulphur in fertilizers, either as ammonium sulphate, potassium sulphate or more often as super- phosphate, which contains 12 per cent of sulphur from the sulphuric acid used to make it. The amount of sulphur applied to agricultural land in Britain each year is about 130,000 tons, corresponding to 10 kg/hectare. Some artificial soils, such as colliery dumps and railway cinder tracks, contain sufficient sulphur to restrict plant growth. SULPHUR IN THE AIR The extent of atmospheric pollution in Britain has been described by Meetham (1956), - and the world problem has been summarized by Junge (1963). The pollution arises largely 114 Watsonia 6 (2), 1965. , SULPHUR AND THE DISTRIBUTION OF BRITISH PLANTS 115 from burning coal, and to a minor extent from smelting ores and making sulphuric acid. Coal contains from 0-5 to 4 per cent of sulphur, but much of this is oxidized to sulphates which remain in the ash after the coal is burnt. The remainder produces sulphur dioxide, and about five million tons of this gas are set free in Britain every year. As a result all big cities and industrial areas in Britain are surrounded by belts of air polluted with sulphur dioxide, as shown in Fig. 1. Polluted air may contain up to 200 pg sulphur dioxide per cubic metre, as against 1-3 pg in the purest air to be found on the earth today. The mean time each molecule of sulphur dioxide remains in the air before it is absorbed by plants or washed out by rain is 4 days. During this time the gas may be blown well out into the surrounding countryside. The ultimate effect of this pollution is to add between 5 and 400 kg of sulphur to each hectare of soil in Britain every year. SULPHUR IN PLANTS Plants need sulphur for many aspects of growth, and absorb it either as the sulphate ion through the roots, or as gaseous sulphur dioxide from the air. Sulphur deficiency has never been reported in Britain, though it is known in Nyasaland and western North America (Wallace 1961): it gives rise to chlorosis of the younger leaves in the same way as a shortage of iron does. Plants suffering from excess of sulphur in the soil are seldom noticed. Excessive amounts of sulphur dioxide in the air are harmful to plants, and markedly reduce their growth rate. The gas causes leaf necrosis and accelerates leaf senescence, and may cause evergreen species such as I/ex aquifolium and Ligustrum vulgare to become deciduous (Scurfield 1960). There are wide variations in tolerance between different species. Analytical data for sulphur in plants have been given for example by Beeson (1941), Thomas et al. (1950), and Spector (1956). Conifer leaves contain only about 0-1 per cent of the element, grasses contain 0-05—-0-2 per cent, while most other Angiosperm leaves contain 0-2-0-4 per cent. Exceptionally large amounts of sulphur are found in species of Cruciferae and the genus Allium (0:5-1-5 per cent), halophytes such as Salicornia and Suaeda (2-3 per cent), and Cuscuta europaea (1 per cent); the last finding needs confirmation. The biochemistry of sulphur compounds in plants has been reviewed by Kjaer (1963). He points out that all species of Cruciferae, together with members of the related families Resedaceae and Capparidaceae, contain thioglucosides (mustard oils), which are responsible for the pungent taste of such well-known condiment plants as radish, horseradish, cress, watercress and the mustards. Species of the genus A//ium contain no thioglucosides, but instead have a wide range of unusual sulphur compounds, including the lachrymator propenyl-sulphenic acid. Other genera which have been shown to contain unusual sulphur derivatives include Equisetum, Athyrium, Pteridium, Petroselinum, Lactuca, Petasites and Asparagus. Much remains to be discovered in this field, but it is clear that the Cruciferae and the genus Allium need more sulphur than other plants and are adapted to high concen- trations of the element. This implies that these groups may have evolved in regions rich in sulphur, but such speculations will not be followed up here. SPECIES ABNORMALLY SENSITIVE TO SULPHUR POLLUTION Sulphur dioxide is known to be absorbed almost entirely by the leaves, but very few studies of chronic exposure of plants have been made (Bleasdale 1959). Hence it is not possible to say which species are most sensitive to sulphur pollution, though it seems that Gymnosperms and grasses are more sensitive than other plants (Katz & Shore 1955, Scurfield 1955, 1960). Since pollution reaches its maximum levels in winter, its effects are usually most severe on evergreen trees, and on herbaceous plants which overwinter with functional leaves. Comparison of Fig. 1 with the distribution maps in the A‘las of the British Flora which show records before and after 1930 (Perring & Walters 1962) is suggestive. Unfortunately it is difficult to disentangle the effects of pollution from two other major factors responsible for recent changes in the distribution of British plants— Watsonia 6 (2), 1965. 116 H. J. M. BOWEN e70% 5056; Vy), 40-80 Fig. 1. Average amount of sulphur dioxide in the air in winter, in micrograms per cubic metre. After Meetham (1956). the rise in mean winter temperatures which has caused many Northern species such as Lycopodium selago to retreat northwards or westwards, and the artificial drainage of wet habitats. However, there is sufficient correlation between areas of gross pollution and recent extinction to justify further studies on the following species: Lycopodium inundatum, Juniperus communis, Hypericum elodes, Radiola linoides, Drosera spp., Gentiana pneumonanthe, Colchicum autumnale, Paris quadrifolia, Neottia nidus-avis and Rhynchospora alba. The majority of these are plants of bogs or wet heaths, whose decrease may be largely due to drainage. Of the remainder, the Juniper (Fig. 2) has aroused interest because of its unexplained diminution on the Chilterns and in County Durham (Fitter 1964, Heslop-Harrison 1962). The decrease of the three monocotyledonous woodland species has been less spectacular and so has not been commented on. Studies of the chronic effects of sulphur dioxide on these species would be needed to establish whether the correlation is a true one. SPECIES TOLERANT OF SULPHUR POLLUTION The maps given in the Atlas of the British Flora show that the distributions of the following species are fairly well correlated with regions of industrial pollution: Artemisia absinthium, Barbarea stricta, Bunias orientalis, Cardaria draba (Fig. 3), Descurainia sophia, Erucastrum gallicum, Erysimum cheiranthoides, Lepidium ruderale, Rapistrum rugosum, Senecio squalidus, S. viscosus, Sisymbrium altissimum and S. orientale. All of these are plants of waste ground and are commonly found growing on soils containing large amounts of coal-ash, such as railway tracks, yards and refuse-tips, though Descurainia, Erysimum and Rapistrum occur as weeds in arable land as well. All are introduced plants, and one, Senecio squalidus, is native to the sulphureous region of Sicily. Rarer plants confined to similar habitats in Britain include Artemisia verlotorum, Astragalus cicer, A. odoratus, Hirschfeldia incana, Rorippa austriaca and several other crucifers. Five other species | (Cerastium atrovirens, Cochlearia danica, Corrigiola litoralis, Diplotaxis muralis and Watsonia 6 (2), 1965. SULPHUR AND THE DISTRIBUTION OF BRITISH PLANTS 117 B wf JUNIPERUS COMMUNIS L. ° Juniper © 1930 onwards © Before 1930 Records from other counties probably introductions Fig. 2. Distribution map of Juniperus communis (Perring & Walters 1962). Note the disappearance since 1930 from polluted regions. eae a ERE See Fig. 3. Distribution map of Cardaria draba (Perring & Walters 1962). Most of the records outside polluted regions are from near the coast. Watsonia 6 (2), 1965. 4a 118 H. J. M. BOWEN D. tenuifolia) should also be mentioned in this group. They are natives of the sea coast which have spread inland along railway tracks, but have not colonized unpolluted soils to any extent. All these species can be said to be tolerant of sulphur pollution, and the high per- centage of crucifers is noteworthy. As mentioned above, crucifers store much of their sulphur as thioglucosides, but the species of Caryophyllaceae, Papilionaceae and Com- positae are believed not to do this. Biochemical examination of Artemisia, Astragalus and Corrigiola for sulphur compounds should be carried out. FURTHER PECULIARITIES IN THE DISTRIBUTION OF CRUCIFERAE AND ALLIEAE Sulphur may also be important in maintaining the abundance of some of the bad cruciferous weeds in Britain, such as Brassica nigra, Sinapis arvensis, S. alba, Raphanus raphanistrum, Thlaspi arvense and Capsella bursa-pastoris, as well as the crow garlic, Allium vineale. The large amounts of sulphur added to the soil in fertilizers may stimulate cruciferous weeds at the expense of others, and this should be tested experimentally. In nature, coastal habitats are rich in sulphur and this may account for the high percentage of cruciferous species restricted to the sea-coast. If one compares the British representatives of other large herbaceous families in this respect, as in Table 1, one finds that only Allieae and Caryophyllaceae have a similar proportion of coastal species. TABLE 1. Proportion of coastal species in various families of British plants. Family No. of native species % of coastal species Cruciferae Si 26:4 Allieae 8 25:0 Caryophyllaceae 71 18-3 Umbelliferae 53 11-3 Ranunculaceae 36 228 Scrophulariaceae 50 2:0 Labiatae 47 0 Coastal habitats are, of course, even richer in chlorine than they are in sulphur, but the two habitats of Dip/otaxis suggest that the latter element is more important as far as crucifers are concerned. The question arises as to whether some of our native crucifers and garlics are limited to soils relatively rich in sulphur. This is a difficult problem to which no definite answer can be given. A number of species of Arabis, Cardaminopsis, Draba, Hornungia and Thlaspi, together with Allium oleraceum, A. schoenoprasum and A. sphaerocephalon, are more or less localized to rocky soils in regions containing unusual minerals. The association of Arabis stricta, Hornungia petraea and Allium sphaerocephalon in the Avon gorge near Bristol may have arisen because of the sulphate minerals there, but there are other possible reasons. The area round Matlock in Derbyshire is noted for the occurrence of Arabis hirsuta, Thlaspi alpestre, Draba muralis, D. incana, Hornungia petraea, Cardamine impatiens and Allium oleraceum, as well as the mineral barytes or barium sulphate. However, it also yields minerals containing large amounts of fluorine, lead and zinc, among other elements, and the effects of specific minerals on these plants have not been tested. There is obviously a rich field here for future research. REFERENCES BEESON, K. C. (1941). The mineral composition of crops with particular reference to the soils in which they were grown. U.S. Dept. Agric. Misc. Pub. 369. BLEASDALE, J. K. A. (1959). The effects of air pollution on plant growth. Sympos. Inst. Biol. 8, 81. Fitter, R. S. R. (1964). Personal communication. Watsonia 6 (2), 1965. SULPHUR AND THE DISTRIBUTION OF BRITISH PLANTS 119 HesLop-HArRrIsOoN, J. W. (1962). Notes on the juniper in Durham. Vasculum, 47, 30. JunGE, C. E. (1963). Air Chemistry and Radioactivity. Academic Press, London. Katz, M. & SHoreE, V. C. (1955). Air pollution damage to vegetation. J. Air Pollution Control Ass. 5, 144-154. Kysarr, A. (1963). Sulphur compounds in plants in M. T. Swain (ed.). Chemical Plant Taxonomy. Academic Press, London. MEETHAM, A. R. (1956). Atmospheric Pollution. Ed. 2. Pergamon Press. PERRING, F. H. & WALTERS, S. M. (1962). Atlas of the British Flora. Nelson, London. SCURFIELD, G. (1955). Atmospheric pollution considered in relation to horticulture. J. Roy. Hort. Soc. 80, 93-101. SCURFIELD, G. (1960). Air pollution and tree growth. Forestry Abstr. 21, 339-347. Spector, W. S. (1956). Handbook of Biological Data. Arnold, Philadelphia. Tuomas, M. D., HENDRICKS, R. H. & HILL, G. R. (1950). Sulphur content of vegetation. Soil Sci. 70, 9-17. WALLACE, T. (1961). The Diagnosis of Mineral Deficiencies in Plants. Ed. 3. H.M.S.O., London. Watsonia 6 (2), 1965. THE ASSESSMENT OF THE TAXONOMIC STATUS OF MIXED OAK (QUERCUS spp.) POPULATIONS By A. CARLISLE and A. H. F. BROWN ABSTRACT The oaks (Quercus spp.) on slate, limestone and peat sites in Roudsea Wood, North Lancashire, were subjected to a taxonomic analysis using both the Hybrid Index and Pictorial Scatter Diagram methods. Both methods showed that Quercus petraea (Mattuschka) Liebl. predominates on the slate site, while the oaks on both the limestone and peat sites are very mixed, with a well-developed Q. robur L. component. The Hybrid Index method over-estimated the morphological intermediacy of the trees, while the Pictorial Scatter Diagram method under-estimated the O. robur component. The scatter diagrams suggested that hybridization and introgression had occurred between Q. petraea and Q. robur. INTRODUCTION Until recently there has been confusion about the identification of Quercus robur L. and Q. petraea (Mattuschka) Liebl. in Britain, partly due to contradictions in the literature and partly to the fact that in Britain many of the oaks exhibit intermediate characteristics. Jones’ (1959) paper has gone a long way towards clarifying the characteristic morphology of the two species, but even so, apart from the recent work by Cousens (1962a, 1962b, 1963) the descriptions of British oak populations have generally been qualitative. In funda- mental work on the autecology and nutrient cycles of oak woodlands, of which this study forms a part, it is essential to have a quantitative description of the individuals and popula- tions if comparisons are to have any real meaning. The present investigation was undertaken to examine two methods of quantitative classification of individual oak (Quercus spp.) trees and mixed populations of Quercus petraea and Q. robur. SITE DESCRIPTION The investigation was carried out in Roudsea Wood in North Lancashire (SD(34)/331828). The woodland, which is owned by Mr. R. E. O. Cavendish, covers 116 hectares (287 acres) of which 45-3 hectares (112 acres) are currently leased by the Nature Conservancy. The mean annual rainfall is 136 cm (53-5 in), and the mean monthly temperature ranges from 4-2 to 14-8° C (Ordnance Survey 1949; Meteorological Office, 1952). The woodland includes areas of Upper Silurian slates of the Bannisdale series, carboniferous limestone, and peat, with local glacial and fluvio-glacial deposits. The oak populations were examined on all these sites. The soils on the slate area are well drained brown earths of low base status, generally > 0-5 m deep, with a pH of 3-6-4:3 at a depth of 5-10 cm, and humus of the moder type (Kubiena 1953). The tree cover on these sites is irregular oak high forest, 70-80 years old, and the ground flora is dominated by Deschampsia flexuosa. The soils of the limestone site are well drained, shallow (< 0:5 m) brown earths of higher base status than the slate soils, and are more variable owing to patches of glacial material containing slate debris. The pH at 5-10 cm depth is 4-3-8-0, and the humus type is similar to a mull. The tree cover on the limestone area consists mainly of dense thickets of oak, ash (Fraxinus excelsior), birch (Betula spp.), and hazel (Corylus avellana) coppice, with scattered standards of the first three. This woodland has a rich ground flora dominated by Mercurialis perennis and Brachypodium sylvaticum. The peats are deep (> 1 m), poorly drained, contain 10 per cent of ash (550° C) per unit dry (105° C) weight, and have a pH of 3:2-3:8 at 5-10 cm depth. The woodland on the 120 Watsonia 6 (2), 1965. ASSESSMENT OF MIXED OAK POPULATIONS 121 peats consists of irregular, uneven-aged birch/Scots pine (Pinus sylvestris) woodland with scattered oak standards more than 50 years old. The ground flora is dominated by Calluna vulgaris and Vaccinium myrtillus. There are no records of planting in Roudsea Wood earlier than 1953, but it is not certain that all the oak is completely natural and indigenous. These woodlands are probably the end product of a number of coppice rotations originating mainly from indigenous, natural maiden trees. METHODS Two methods were used to assess the taxonomic status of individuals and populations of oak in the present investigation, both of them being what Anderson (1949) terms ‘polygraphic analyses’: I The Hybrid Index (HI) Method; II The Pictorial Scatter Diagram (PSD) Method. The main survey of the oak populations in Roudsea Wood was carried out using the HI method. Sub-samples (n = 20) of the populations were assessed by both methods. I. THE HYBRID INDEX METHOD This method, explained in detail by Anderson (1936, 1949), is useful for describing complex populations where two species have hybridized and backcrossed to produce plants of intermediate morphology (Anderson 1936, Raunkaier 1925, Riley 1938). It consists of calculating for each sample tree a numerical index, which describes the degree to which overall morphology of the individual is typical of one or other of the species. When frequency distributions of these indices, calculated for a sample of plants from a population, are drawn as histograms, they give a visual description of the extent to which the components of the population are intermediate or typical of the species. In the present investigation the calculation of the Hybrid Index was modified to permit different numbers of characters being examined in individual sample trees. In the calculation of the ordinary HI for a number of individual plants it is essential to keep the number of characters examined constant, otherwise the indices are not comparable. This is not always possible, and if strictly adhered to could introduce a bias into the sampling. Q. robur tends to produce seed more readily than Q. petraea (Jones 1959), and if all trees without acorns were rejected from the sample there would be a tendency to sample Q. robur more often than Q. petraea in a mixed population. In this instance, therefore, a variable number of characters must be accepted, and to overcome this difficulty, the index was calculated as a Percentage Hybrid Index. As many as possible of the selected characters were examined, assigned individually to the three categories Q. robur, or intermediate, or Q. petraea, and the number of characters in each category summed. These three sums were then calculated as percentages of the total number of characters examined, and the percentages multiplied by the factors 0, } and 1 respectively. The summation of these products gave the Percentage Hybrid Index (PHI) which has a range of 0-100 (0 = Q. robur, 100 = Q. petraea) and is independent of variations in the total number of characters recorded. The characters used in the estimation of the PHIs of the oaks in Roudsea Wood were based on Jones’ definitions (1959). To begin with, 23 characters were examined by two independent workers examining the same set of specimens. Persistent disagreement occurred when attempting to categorize leaf glossiness, colour and texture, and these characters were discarded. The remaining 20 characters are defined in Table 1. In view of the current lack of knowledge on the relative importance of individual oak characters, no weighting was used in the calculation of the PHI. Roudsea Wood was divided into three parts, on the basis of the main soil types: slate soils, limestone soils, and peat. A random sample of 118 trees was taken on the slate site, 54 on the limestone, 48 on the peat. More than half the oaks in the wood are on the Watsonia 6 (2), 1965. 4B 122 A. CARLISLE and A. H. F. BROWN TABLE 1. Morphological characters of Quercus spp. used in the assessment of percentage hybrid indices. Character Stem persistence Angle between branches Crown density Bark Buds (terminal) (a) Size (b) Form Leaves (a) Shape (b) Lobe number* (c) Lobe depth (d) Lobe regularity (e) Venation (f) Base auricle Q. robur Not persistent 1455 Open (foliage in clusters) Thick elongated units Small (< 5 mm long) Obtuse tip Obovate, widest in apical third of leaf <5 Deep (> 50% of half leaf width) Irregular Some veins to the between lobes Strongly auriculate, with the sinuses Q. petraea Persistent < 45° Dense (foliage uniformly dis- tributed) Thin rectangular units Large (> 5 mm long) Acute tip Ovate, widest near mid-leaf SSS) Shallow (< 50% of half leaf width) Regular None to the sinuses Cordate, or slightly indented auricle lobes overlapping the petiole (g) Petiole < 7mm > 13 mm (h) Simple hairs on abaxial surface Few or absent Abundant (i) Stellate hairs on abaxial surface Absent Present Acorns (a) Colour (when ripe) Pale fawn Dark brown (b) Stripe (when ripe) Olive green None (c) Form Elongated Rounded (d) Peduncle lengtht > 30 mm < 20 mm (e) Peduncle pubescence Glabrous Clustered hairs * Jones gives the leaf lobe number of Q. robur as 3-5 (-6), and for Q. petraea as 5-6-8. +t Peduncle length taken as peduncle length to the terminal flower or bract. slate site, and this area was sampled most intensively. Even so the sample of the trees on the slate site was less than a 10 per cent enumeration. On the limestone and peat more than half the total number of oaks present were sampled. The trees were sampled in the autumns (August to September) of 1961 and 1962. Stem persistence, branch angle, bark form and crown density were all recorded in the field. Ten sample shoots were taken from the upper half of the south side of the crown of each tree, all the shoots being free from shading. Lammas shoots were not included. Acorns were taken where possible, but a tree was not discarded from the sample because acorns were absent. Samples were examined either the same day or the day after collection. Ten leaves were taken at random from the pre-Lammas growth, and as many acorns as possible (up to 10) were examined. Decisions were made as to whether each character was typical of Q. robur or Q. petraea, or inter- mediate, and the PHIs calculated (Figs. 1-3). There was a tendency to regard a character which was only slightly atypical of one or other of the species as intermediate. This was particularly so where a character was defined qualitatively (e.g. leaf lobe regularity, acorn shape). There is therefore a tendency for the PHI method to over-emphasize the intermediate character of a tree. Watsonia 6 (2), 1965. ASSESSMENT OF MIXED OAK POPULATIONS 123 (1) (2) 30 Slate (n = 118) 30 Limestone (n = 54) 4 Be 20 sy g : vo = 3 S19 S Fes 3 0 0 0 10 20 30 4050 60 7080 90 100 0 10 203040 50607080 90 100 Qr. AP PETS @ip: Q.r. P.H.I. Q.p. 30 (3) = Peat (n = 48) > 20 g 2 10 oy 0 0 10 20 30 4050 6070 8090 100 Qo Paeete Q.p. Figs. 1-3. The frequency distributions of Percentage Hybrid Indices (PHI) of oaks on slate, limestone and peat sites in Roudsea Wood. We examined samples from 10 trees twice within a period of a week, and the two PHI estimates differed by a mean of +2-6 PHI units with a maximum difference of 6:0 units. The method can, therefore, be repeated with an acceptable degree of error. Il. PICTORIAL SCATTER DIAGRAM This method, based upon a technique described by Anderson (1949), has been used by Cousens (1962a and b) for surveys of oak populations in Britain. The principle of the method is that when, in a population of two species and their hybrids, two continuously variable morphological attributes by which the species are known to differ are plotted graphically, and each co-ordinate coded for subsidiary characteristics, the resultant scatter diagram gives a visual description of the population. The pattern of the scatter may be interpreted in terms of the degree of hybridization and introgression, and different popula- tions can be compared visually. Points for the two species tend to concentrate in two separate zones, those for individuals with intermediate characteristics lying outside and between these zones. The Pictorial Scatter Diagram (PSD) method was used on a sample of 20 trees taken at random on each of the slate, limestone, and. peat sites. Ten shoots were taken from the upper half of the south side of the crown of each tree. Lammas growth was excluded from the sample. As many acorns as possible (up to 10) were included in each sample. Five leaves were taken from the mid-zone of each of the 10 shoots, mixed, and 10 of these 50 leaves taken at random. The petiole length, total leaf length, leaf auricle type and abaxial stellate pubescence were assessed for these 10 leaves, and the petiole length expressed as a percentage of total leaf length as suggested by Cousens. The acorn peduncle lengths to the first bract or flower were measured and the mean calculated. The mean petiole per cent. and mean peduncle length were plotted on logarithmic paper, and the points coded for leaf auricle type and abaxial stellate pubescence (Figs. 4-6). The same 20 trees were subjected to a PHI analysis (Figs. 7-9). This PSD procedure differed from that used by Cousens in that 10 leaves per tree Watsonia 6 (2), 1965. A. CARLISLE and A. H. F. BROWN 124 ‘(Q-p ‘SSI{) SWRISCICG 19}189G [P110}01g OY} JON1}SUOD 0} pasn POOM Baspnoy UI sous jvod puv suojsowl] ‘aye[s Woy sojdures Yeo oY} JO ([Hd) SedIpuy PUqhH oseJUIOIOg oY} JO suOINgIsIp AQuonboly ou “6-1 “S8Iyq me Bh wy) “Ter vd 3) ‘d-% “{> Hed bal 3) 700 9) i. on 5 re | se) OOT 06 08 OL 09 OS OF OF O@ OT O OOT 06 08 OL 09 OS OF O€ O¢ OL 0 OOT 06 08 OL 09 OG OF OF OZ OT O OT 1 -y Oly 4 ce ©, A) a © Ke) = Ko} c (~ ove 02 023 ® 9° A Oo << = < OES One og 8? (0Z = U) 1vagd (0Z = U) QUOJsSOWT'T (Og = U) 2711S (6) OF (8) Ov (1) OP "AJoATO0dse1 jUSsqe pue yeoM “podoaaop [jam s[d1Ine Jeo] = oC ‘P “P-Ajoat}oodso1 yUasqe pue asieds ‘juepuNqe sovJins Jeo] jelxeqe uO aouaosognd o4eI[91S = © ‘@ ‘@ ‘pooMfLespnoy UI soyIs yeod pue oUOJsSoUNT] ‘aye[S UO SyYeO OJ SUIBISLIG IO}eOg ]eHOIg ‘9-p ‘S3Iy (iW) YYsUaT dJouNpsg uray (WIut) YysUueT sTOUNpeg uUvsy (WU) YYsUa'T aTOUNpdd UII 0S OF O€ 0¢ OT G 0 0S OF OF 0Z OL G 0 0S OF OF 02 OT G 0 Z 2 < 9) 90) oO 5 S = 'U % a) 0 a ? sec S o ? 10) 6S or? © te es do le = 8 fe ® a SS ae S (0Z = u) yo = (0¢ = u) ouojsowry a i 0Ges 07s rs (9) (S) Watsonia 6 (2), 1965. ASSESSMENT OF MIXED OAK POPULATIONS 125 were used instead of two; the mean acorn peduncle length was used instead of maximum peduncle length; and the peduncle stoutness and pubescence were not included in the code on the scatter diagram. RESULTS The taxonomic status of the oak populations The PHI frequency distributions (Figs. 1-3) show differences between the oak popula- tions on the different sites. On the slate site, the oaks are mainly trees with marked Q. petraea characteristics, and there are very few Q. robur. The oak populations on the limestone and peat are more mixed with modes near both the Q. robur and Q. petraea ends of the scale. There is a strong Q. robur component in the woodlands on the peat site. The scatter diagrams (Figs. 4-6) also show that the oak populations on the slate, limestone and peat sites are different. On the diagram (Fig. 4) for the slate site, there is a concentration of points in a zone where the mean petiole length per cent. is high and the mean peduncle length low, the leaves having well-developed stellate hairs and no leaf auricles, all of which indicates that most of the trees are Q. petraea. There are three intermediate points. This suggests a predominantly Q. petraea population containing a few trees with intermediate characteristics. The diagram (Fig. 5) for the limestone site suggests a very mixed population ranging from trees with high mean petiole percentages and low mean peduncle lengths (Q. petraea) to trees with low mean petiole percentages and high mean peduncle lengths (Q. robur), with many intermediates. The Q. petraea trees all have at least a few stellate hairs and no auricle, and all the points in the Q. robur zone are associated with trees having well-developed auricles and few or no stellate hairs. The diagram for the peat site shows that while there are a few trees in or near the Q. petraea zone on the diagram, and a few more or less in the Q. robur zone, most of the trees are intermediate in their petiole and peduncle characteristics. Near the Q. petraea zone some of the trees have weakly-developed auricles. The scatter diagrams reflect the Q. petraea component suggested by the PHI frequency distributions, but the former do not suggest as strong a Q. robur component as the latter. As Q. robur apparently prefers base-rich fertile soils and QO. petraea the more acid, less fertile soils (Jones 1959, Moss 1914), the distribution of the species on the slate and limestone sites in Roudsea Wood, with Q. petraea predominating on the slate and a well-represented Q. robur component on the limestone, could perhaps be interpreted in terms of these soil preferences. However, it must be remembered that it is possible that some oak has been planted in Roudsea Wood, although there is no evidence of such plantings in the estate records. It seems most unlikely that oak has ever been planted on the peat sites and more likely that the trees have arisen naturally. If any oaks survive at all on these peats, it would be expected that Q. robur, with its tolerance of soil waterlogging (Jones 1959) would compete successfully with the Q. petraea. Hybridization and introgression So far, all the trees with characteristics which are neither markedly those of Q. petraea nor Q. robur have been termed “intermediates”, and no assumptions have been made that they are necessarily due to hybridization. Dengler (1941) suggests that Q. robur and Q. petraea are able to hybridize to a certain extent, and it is likely that hybridization occurs in field conditions. Cousens (1962a and b) has discussed the possibility of intro- gression occurring. The PHI frequency distributions give little information about hybridization and introgression other than showing that the populations are mixed, with many intermediates which may or may not be hybrids. On the other hand, PSDs provide a certain amount of information on this point. The patterns of the scatter diagrams in Figs. 4-6 are similar to those described by Anderson (1949) as being associated with populations of plants where hybridization and introgression have occurred. The scatter diagram for the slate Watsonia 6 (2), 1965. 126": A. CARLISLE and A. H. F. BROWN site suggests a population of Q. petraea that is almost pure but slightly introgressed. The other scatter diagrams (Figs. 5-6) suggest strongly: introgressed populations, possibly with some FI hybrids. These data lend support to Cousens’ (1962a and b) view that introgression occurs in British oak woodlands. A comparison of the Hybrid Index and Pictorial Scatter Diagram methods Although the two methods give similar general pictures of the taxonomic status of the oak populations in Roudsea Wood, these pictures differ in detail. The closest agreement is in the Q. petraea component. The marked Q. petraea tendencies of the slate population as shown by the PHI frequency distribution histograms (Figs. | and 7) are also reflected by the pronounced concentration of points in the long petiole/short peduncle zone on the PSD in Fig. 4. In the PHI histograms in Figs 2-3 and 8-9 there are indications of a strong Q. robur component which are not reflected in the corresponding PSDs in Figs. 5-6. Either the PHI method tends to over-estimate the Q. robur component or the PSD method under-estimates it. It has already been mentioned that there is a tendency in the PHI method to over-emphasize the intermediacy of a specimen. It seems unlikely, therefore, that this method will over-emphasize the Q. robur (or Q. petraea) components, and more likely that the PSD is under-estimating the Q. robur component. This under-estimation may well be due to the fact that Q. robur apparently is much more variable than Q. petraea (Weimarck 1947). This greater variability may extend to the petiole and peduncle lengths, so that unless large numbers of trees are examined, there may not be clear evidence of a concentration of points in any particular zone at the small petiole/long peduncle end of the scale, even though a PHI analysis suggests an almost pure population of Q. robur. Cousens (1962a and b) designates concentration centres on his scatter diagrams, using petiole length per cent. and maximum peduncle length. He found that none of the many oak woodlands examined by him was pure Q. robur, i.e. not all the points came within the limits of the Q. robur concentration centres. It is possible that this apparent absence of pure Q. robur woodlands may be due to this method under-estimating the Q. robur component. On the scatter diagrams, it will probably be necessary to allow more latitude for Q. robur than Q. petraea if concentration centres are delineated. The two methods, when used simultaneously in an oak population survey, provide complementary information. The PHI method is based upon more characters; the histo- grams give an overall picture of the population. The PSD method, on the other hand, provides information about hybridization and introgression. Where extreme purity of a species is needed, as, for example when seed is needed for autecological studies, the PHI method is to be preferred as it is based on more characters and tends to over-estimate the intermediacy and under-estimate the purity of the trees. For a tree to have a PHI of <10 or >90 it has to be very typical of the species indeed. Modifications of the oak population survey methods In future surveys using the PHI method, it will be better to use only the characters of bud size and form; leaf shape, lobe number, lobe depth, lobe regularity, venation, auricle, petiole, abaxial stellate pubescence; and acorn colour, stripe, form and peduncle length (Table 1). The petiole length character should be expressed as a percentage of the total leaf length as suggested by Cousens (1962a and b), and defined as <6-5 per cent for Q. robur and > 9-5 percent for Q. petraea as suggested by the scatter diagrams in Figs. 4—6. The mean value should be based upon 10 or more leaves taken from the middle of shoots from the south side of the upper half of the crown where possible. The peduncle length should be expressed as the mean length of the peduncle to the first bract or flower, based upon 10 or more peduncles per tree where possible. Tentatively it is suggested that a mean peduncle length of > 20 mm is regarded as a Q. robur characteristic and < 6 mm as a Q. petraea. characteristic. Watsonia 6 (2), 1965. a i = ASSESSMENT OF MIXED OAK POPULATIONS 127 The main weakness of both the methods is our lack of knowledge about the variations of the individual characters within pure stands of Q. petraea and Q. robur. Individual morphological characters may have considerable latitude in the pure species types. If this is so, at least some characters which are now regarded as morphologically intermediate may occur in pure types. At the moment, therefore, estimates of the taxonomic status of these oaks probably under-estimate the purity of the trees. The species are at present defined in terms of multiple qualitative data so that more precise statistical methods of discrimina- tion, such as Fisher’s discriminant analysis (Fisher 1950), cannot be used. ACKNOWLEDGMENTS The authors wish to thank Mr. M. D. Mountford for his valuable comments, and the many people who assisted with the collection of the samples. REFERENCES ANDERSON, E. (1936). Hybridization in American Tradescantias. Ann. Mo. bot. Gdn. 23, 511-525. ANDERSON, E. (1949). Introgressive Hybridization. London and New York. Cousens, J. E. (1962a). Oak population studies in Scotland: I. Variation of some important diagnostic characters of Quercus petraea (Matt.) Liebl. and Q. robur L. Forestry Commission Report on Forest Research for the year ended March 1961, 109-113. Cousens, J. E. (1962b). Notes on the status of sessile and pedunculate Oaks in Scotland and their identification. Scot. For. (J.) 16, 170-179. Cousens, J. E. (1963). Variation of some diagnostic characters of the sessile and pedunculate oaks and their hybrids in Scotland. Watsonia 5, 273-286. DENGLER, A. (1941). Bericht tiber Kreuzungsversuche zwischen Trauben- und Steileiche (Quercus sessiliflora Smith u. Q. pedunculata Ehrh. bzw. Robur L.). Mitt. H.-Goring-Akad. Dt. Forstwiss. 1, 87-109. FisHer, R. A. (1950). Statistical Methods for Research Workers. Edinburgh. Jones, E. W. (1959). Biological flora of the British Isles. Quercus L. J. Ecol. 47, 169-222. KuBIENA, W. L. (1953). The Soils of Europe. London. METEOROLOGICAL OFFICE (1952). Climatological Atlas of the British Isles. London. Moss, C. E. (1914-20). The Cambridge British Flora. Cambridge. ORDNANCE SuRVEY (1949). The Map of the Average Annual Rainfall of the British Isles. London. RAUNKAIER, C. (1925). Ermitageslettens Tjgrne. Kgl. Danske Videnskab. Biol. Meddel. 5, 1-76. Rivey, H. P. (1938). A character analysis of Iris fulva, I. hexagona var. giganticaerulea and natural hybrids. Amer. J. Bot. 25, 727-738. WEIMARCK, H. (1947). Distribution and ecology of Q. petraea. Botanisker Notiser, 189-206. Watsonia 6 (2), 1965. DACTYLORHIZA NEVSKI, THE CORRECT GENERIC NAME OF THE DACTYLORCHIDS By P. F. Hunt and V. S. SUMMERHAYES Royal Botanic Gardens, Kew ABSTRACT The correct generic name for the dactylorchids (marsh and spotted orchids) is shown to be Dactylorhiza Nevski. A list of species of Dactylorhiza is given and the subspecies occurring in the British Isles are indicated. Several new combinations at specific and subspecific rank and five new bigeneric hybrid formulae are published for the first time. In his Species Plantarum (939-944, 1753) Linnaeus divided the genus Orchis into three parts based on the morphology of the roots, namely: Bulbis indivisis, Bulbis palmatis and Bulbis fasciculatis. Some time later, Necker, in his Elementa Botanica (3,129, 1790), raised these groups to generic level although he actually used the category name “species naturalis’ for them. Orchis L. was retained for Bulbis indivisis whilst Bulbis palmatis and Bulbis fasciculatis became Dactylorhiza Necker. The next important treatment of the genus was by Klinge, in 1898 (Acta Hort. Petrop. 17, 148). He recognized two subgenera, namely Eu-orchis, into which he placed the Linnaean Bulbis indivisis, and Dactylorchis which included Bulbis palmatis. This classification was adopted by many later workers, but in 1935, Nevski, in his account of the Orchidaceae for the Flora URSS, substituted Necker’s name Dactylorhiza for the second of Klinge’s sub- genera on the ground that it was earlier than Dactylorchis Klinge. Nevski also seems to have excluded Linnaeus’s Bulbis fasciculatis, at least by implication. Two years later, however, Nevski evidently decided that the two subgenera were better treated as distinct genera and adopted the generic name Dactylorhiza, making a new combination, D. umbrosa (Kar. et Kir.) Nevski (Acta Inst. Bot. Acad. Sci. URSS ser. 1, 4, 332, 1937). This generic name is obviously based on Orchis subgen. Dactylorchis Klinge although naturally Nevski attributed it to Necker. Ten years later still, in 1947, Vermeulen published his well-known Studies on Dactyl- orchids in which he raised Klinge’s subgenus Dactylorchis to generic rank as Dactylorchis Vermeul., making many of the necessary new combinations. In this book, Vermeulen quite arbitrarily dismissed the name Dactylorhiza as not being a synonym of Dactylorchis because the two concepts were not co-extensive. It is true that Dactylorchis is not identical with Dactylorhiza as understood by Necker but this does not preclude the two names from being synonymous. Vermeulen evidently overlooked Nevski’s use of the name Dactylorhiza in its narrower sense, which was identical with his concept of Dactylorchis. Unfortunately, owing to the relative obscurity of Nevski’s publication contrasted with the obviously wide distribution of Vermeulen’s book, the generic name Dactylorchis has been adopted by some botanists whereas Dactylorhiza was almost completely ignored until 1959. In that year Bullock (Taxon 8, 46) in a paper recommending the rejection of Necker’s names, cited the case of Dactylorhiza versus Dactylorchis as an example of the possible confusion arising if this course were not taken. Necker’s names were arbitrarily designated as unitary specific names and rejected under article 20 of the International Code of Botanical Nomenclature in 1959 but this does not affect the legitimacy of Nevski’s use of the name. Dactylorhiza Nevski is validated by reference to Necker’s description which was effectively published and by reference to Orchis L. subgenus Dactylorchis Klinge which also is accompanied by a description. Although Nevski was wrong in using the name Dactylorhiza at subgeneric level he was | 128 Watsonia 6 (2), 1965. DACTYLORHIZA NEVSKI 19 quite in order when using it for a genus. The correct generic name for the group of Orchis designated by Klinge as subgenus Dactylorchis is therefore Dactylorhiza Necker ex Nevski (or simply Dactylorhiza Nevski) which has ten years’ priority over Dactylorchis Vermeul. The first worker to follow Nevski was Borsos (Acta Bot. Acad. Sci. Hung. 5, 321, 1959) in her “Dactylorchis fuchsii et son affinité dans les flores Hongroise et Carpathique’’. She dealt with the names Dactylorhiza and Dactylorchis and pointed out that Dactylorhiza has priority. She then made several new combinations but unfortunately stated that they were intended to be valid in both Dactylorhiza and Dactylorchis. This is contrary to article 34 of the Code and consequently none of her names can be accepted as validly published in either genus. The next year, 1960, Sod summed up the situation to date in his “‘Synopsis Generis Dactylorhiza (Dactylorchis)” (Ann. Univ. Sci. Budap. de Rolando E6tvés nom. sect. Biol. 3, 335) and made many new combinations in Dactylorhiza. Although he clearly accepts Dactylorhiza as the correct name he thinks that as Dactylorchis has been used by a number of authors it should be conserved. Unfortunately many of these new combinations do not comply with article 33 of the Code which states that a new combination or a new name is not validly published unless the basionym is clearly indicated with a full and direct reference to its author and original place of publication, including page reference and date. Sod was informed by us of this error and in 1962 he published a separate paper entitled ““Nomina Nova Generis Dactylo- rhiza”’ in order to validate his new combinations and new names. There seems to be little doubt, in view of the evidence from both morphological and cytological studies and the incidence of hybridization in nature, that Nevski, Vermeulen, Borsos, Soo and others are correct in considering the dactylorchids to belong to a genus distinct from Orchis proper which is typified by O. militaris L. and contains species such as O. mascula (L.) L. The dactylorchids are readily distinguishable by their foliaceous bracts (Orchis proper usually has chaffy, membranous bracts), hand-like (palmate) tubers (testicle- like in Orchis) and basic chromosome number of 20 (Orchis n = 16, 18 or 21). The general habit is also different in the two groups, the developing inflorescence in Orchis proper being covered by the spathe-like uppermost stem bract until just before the buds open, whereas in the marsh and spotted orchids the apical flower buds of the inflorescences are exposed as soon as the spike appears above ground. In order to emphasize our acceptance of the existence of a separate genus for the marsh and spotted orchids and of the correct name, Dactylorhiza, for this genus, we repeat here the new combinations of all the species that Sod recognizes in Dactylorhiza. We are not, however, expressing any opinions as to the status of any of the taxa that do not occur in Britain, nor do we agree entirely with the details of his classification. For various reasons we have found it necessary to make some new combinations ourselves; these include cases where Soo has still not made the combinations in a valid manner. With regard to the British species, which are indicated by asterisks, we have added what we consider to be the subspecies occurring in Britain. We do not think that it is possible at present to recognize varieties in the British species of Dactylorhiza. The taxonomic positions of Orchis francis- drucei Wilmott and of Dactylorchis majalis subsp. cambrensis Roberts are not yet clear and we have therefore omitted them from this account. We also give a synonymy of the names used in the Flora of the British Isles by Clapham, Tutin & Warburg, edition 1, 1952 (abbreviated as CTW) and in the List of British Vascular Plants by Dandy, 1958 (abbreviated as DANDY). We add those of the Flora of the British Isles CTW, edition 2, 1962, where the nomenclature or taxonomy differs from that of Dandy. In all these cases we quote the author citations of these synonyms as given in the books cited. 1. DACTYLORHIZA IBERICA (M. Bieb.) Sod. 2. DACTYLORHIZA SAMBUCINA (L.) Soo. 3. DACTYLORHIZA ROMANA (Seb. & Maur.) Soo. Watsonia 6 (2), 1965. 130 4. Oo COND latifolia L., no universally accepted decisions have been reached. At some time or another the variously referred to O. latifolia L., together with less well known species. On careful Watsonia 6 (2), 1965. . *DACTYLORHIZA CRUENTA (O.F. Muell.) Soo. . DACTYLORHIZA PSEUDOCORDIGERA (Neum.) Soo. . DACTYLORHIZA SALINA (Turcz.) S00. . DACTYLORHIZA SANASUNITENSIS (Fleisch.) Soo. . Dactylorhiza cilicica (Klinge) P. F. Hunt & Summerh., comb. nov. . Dactylorhiza kotschyi (Rchb.f.) P. F. Hunt & Summerh., comb. nov. . Dactylorhiza osmanica (Klinge) P. F. Hunt & Summerh., comb. nov. . DACTYLORHIZA UMBROSA (Kar. & Kir.) Nevski. . DACTYLORHIZA PERSICA (Schltr.) Soo. . DACTYLORHIZA GRAGGERIANA (S06) Soo. . DACTYLORHIZA HATAGIREA (Don) Soo. . Dactylorhiza majalis (Rchb.) P. F. Hunt & Summerh., comb. nov. P. F. HUNT and V. S. SUMMERHAYES *DACTYLORHIZA INCARNATA (L.) S00. Dactylorchis incarnata (L.) Vermeul. DANby 6437/3. Orchis strictifolia Opiz CTW. (a) subsp. INCARNATA Dactylorchis incarnata (L.) Vermeul. subsp. incarnata DANDY 643/3/a. Dactylorchis incarnata (L.) Vermeul. subsp. gemmana (Pugsl.) H.-Harrison f. DANDY 643/3/e. (b) subsp. ochroleuca (Boll.) P. F. Hunt & Summerh., comb. nov. Orchis incarnata subsp. ochroleuca (Boll.) | Schwarz. Mittel. Thiiring. Bot. Ges. 1 (1), 94 (1949). Orchis incarnata L. var. ochroleuca Boll. Arch. Ver. Nat. Meckl. 14, 307 (1860). Dactylorchis incarnata (L.) Vermeul. subsp. ochroleuca (Boll.) H.-Harrison f. DANDy 643/3/d. (c) subsp. PULCHELLA (H.-Harrison f.) Soo. Dactylorchis incarnata (L.) Vermeul. subsp. pulchella (Druce) H.-Harrison f. DANDY 643/3/b. Orchis strictifolia Opiz subsp. strictifolia var. pulchella (Druce) Clapham CTW. (d) subsp. COCCINEA (Pugsl.) Sod. Orchis latifolia L. var. coccinea Pugsley. J. Linn. Soc. Bot. 49, 578 (1935). Dactylorchis incarnata (L.) Vermeul. subsp. coccinea (Pugsl.) H.-Harrison f. DANDY 643/3/c. Orchis strictifolia Opiz subsp. coccinea (Pugsl.) Clapham CTW. Dactylorchis incarnata (L.) Vermeul. subsp. cruenta (O.F. Muell.) Vermeul. DANDy 643/3/d. Orchis cruenta O.F. Muell. CTW. Orchis cilicica (Klinge) Schltr. Feddes Rep. Sonderbeih. A, 1, 178 (1927). Orchis orientalis Klinge subsp. cilicica Klinge in Acta Hort. Petrop. 17 (1) 36, (1898). Orchis kotschyi (Rchb.f.) Schltr. in Fedde Rep. 19, 48 (1923). Orchis incarnata L. var. kotschyi Rchb.f. in Orch. Fl. Germ. Recens. 53 (1851). Orchis osmanica (Klinge) G. Camus in Icon. Orch. Europ. 222 (1929). Orchis orientalis Klinge subsp. osmanica Klinge in Acta Hort. Petrop. 17(1), 188 (1898). Orchis majalis Rchb. in Pl. Crit. 6, 7 (1828). Although there have been many papers dealing with the supposed identity of Orchis three entities we now call D. incarnata, D. praetermissa and D. majalis have been DACT YLORHIZA NEVSKI 131 consideration of the evidence we agree with many taxonomists that O. Jatifolia L. is best looked upon as a nomen confusum and therefore to be disregarded (Art. 69). Sod, however, uses Dactylorhiza latifolia for what we here call Dactylorhiza majalis. Three other names published between O. latifolia L. (1753) and O. majalis Rchb. (1828) have been quoted in the major monographs as synonyms of O. Jatifolia. Vermeulen and others who have accepted majalis in preference to Jatifolia have not, however, stated why they did not use any of the above mentioned synonyms. We have therefore looked into these names with the following results: (i) Orchis comosa Scop. Fl. Carn. ed. 2, 2, 198 (1772). This is based on Bauhin’s Palmata vilissima (Hist. Plant. Univ. 2, 776 (1651)). The identity of this plant is very doubtful but it may possibly be equal to Dactylorhiza latifolia (L.) Sod subsp. alpestris (Pugsl.) Sod, which, however, may prove to be distinct from D. majalis proper. (ii) Orchis palmata Gilib. Exercit. Phytol. 2, 479 (1792). This is not a binomial name, being given as Orchis palmata rubra nectarii labio maculato. The great majority of names in this book consist of two words but these are to be regarded as abbreviated phrase names and not Linnaean binomials. (iii) Orchis fistulosa Moench, Meth. 713 (1794). In this case Orchis latifolia L. is given as a synonym; O. fistulosa is therefore illegiti- mate. From the above it is clear that no one of the above three names can be used and we are consequently adopting the epithet majalis since it is the earliest legitimate one which can be identified unequivocally. 17. *Dactylorhiza kerryensis (Wilmott) P. F. Hunt & Summerh., comb. nov. Orchis kerryensis Wilmott in Proc. Linn. Soc. Lond. Sess. 148, 126 (1936). Dactylorchis majalis (Rchb.) Vermeul. subsp. occidentalis (Pugsl.) H.-Harrison f. Danby 643/6. Orchis occidentalis Pugsl. subsp. kerryensis (Wilmott) Clapham CTW. 18. DACTYLORHIZA CORDIGERA (Fries) Sod. 19. DACTYLORHIZA CATAONICA (Fleisch.) Holub. Preslia 36, 252 (1964). Orchis cataonica Fleisch. in Ann. Nat. Hofmus. Wien 28, 34 (1914). Orchis caucasica (Klinge) Medvyedev. Acta Hort. Tiflis 18, 271 (1926), non Regel (1809). Orchis cordigera Kraenzl. Fedde Rep. Beih. 65, 39 (1931), non Fries (1842). Orchis euxina Nevski in Komarov F. URSS 4, 709 (1935). Dactylorhiza euxina (Nevski) Soé and D. caucasica (Medvyedev) Sod were simul- taneously published as alternative interchangeable names for this concept in ‘Nomina nova generis Dactylorhiza’. According to Article 34 of the Code both must be regarded as not validly published. Sod, although including Orchis cataonica as a synonym of this concept and making it a variety, overlooked the fact that it is the earliest available specific epithet. 20. *DACTYLORHIZA PRAETERMISSA (Druce) Soo. Dactylorchis praetermissa (Druce) Vermeul. DANDY 634/4. Orchis praetermissa Druce CTW. 21. *DACTYLORHIZA PURPURELLA (T. & T.A. Steph.) Soo. Dactylorchis purpurella (TY. & T.A. Steph.) Vermeul. DANDy 643/5. Orchis purpurella T. & T.A. Steph. CTW. 22. DACTYLORHIZA ARISTATA (Fischer ex Lindl.) Sod. 23. DACTYLORHIZA LAPPONICA (Laestad. ex Rchb. f.) Sod. 24. *DACTYLORHIZA TRAUNSTEINERI (Sauter) Soo. Dactylorchis traunsteineri (Sauter) Vermeul. DANDY 643/7. Orchis traunsteinerioides (Pugsl.) Pugsl. CTW. 25. DACTYLORHIZA ELATA (Poir.) Soo. 26. DACTYLORHIZA FOLIOSA (Vermeul.) Sod. Watsonia 6 (2), 1965. 132 P. F. HUNT and V. S. SUMMERHAYES 27. *DACTYLORHIZA MACULATA (L.) Soo. Dactylorchis maculata (L.) Vermeul. DANDY 6437/2. (a) subsp. ericetorum (Linton) P. F. Hunt & Summerh., comb. nov. Orchis maculata L. subsp. ericetorum Linton in Fl. Bournemouth 208 (1900). Dactylorchis maculata (L.) Vermeul. subsp. ericetorum (Linton) Vermeul. DANDY 643/2/b. Orchis ericetorum Linton CTW. (b) subsp. RHOUMENSIS (H.-Harrison f.) Soo. Dactylorchis maculata (L.) Vermeul. subsp. rhoumensis (H.-Harrison f.) H.-Harrison f. DANDY 643/2/a. Dactylorchis fuchsii (Druce) Vermeul. subsp. rhoumensis (H.-Harrison f.) Clapham CTW ed. 2. Orchis fuchsii Druce subsp. rhoumensis H.-Harrison f. CTW ed. 1. 28. DACTYLORHIZA SACCIFERA (Brongn. ex Bory) Soo. 29. *DACTYLORHIZA FUCHSII (Druce) Soo. Dactylorchis fuchsii (Druce) Vermeul. DANDy 643/1. (a) subsp. FUCHSII. Dactylorchis fuchsii (Druce) Vermeul. subsp. fuchsii DANDY 643/1/a. Orchis fuchsii Druce CTW. (b) subsp. OKELLYI (Druce) Soo. Dactylorchis fuchsii (Druce) Vermeul. subsp. okellyi (Druce) Vermeul. DANDY 643/1/b. Orchis okellyi Druce CTW in obs. (c) subsp. HEBRIDENSIS (Wilmott) Soo. Dactylorchis fuchsii (Druce) Vermeul. subsp. hebridensis (Wilmott) H.-Harrison f. Danby 643/1/c. Orchis fuchsii Druce subsp. hebridensis (Wilmott) Clapham CTW. BIGENERIC HYBRIDS We are taking this opportunity of publishing some bigeneric hybrid names in which Dactylorhiza is one of the parent genera. In our opinion such hybrid names should be regarded as mere formulae and no descriptions should be needed for validation, which is effected solely by stating the parent genera concerned. Being formulae and therefore not tied to the rules of priority (except that when the same two generic names have been combined in various ways the earliest ought to be chosen), hybrid generic names should be altered whenever the generic position of one or other of the parent species alters. This should not, however, apply if the change is merely nomenclatural but this question does not arise here as we know of no bigeneric names formed with the name Dactylorchis used for one parent. | The necessary new formulae consequent upon accepting the genus Dactylorhiza are as follows: 1. x Dactylocamptis P. F. Hunt & Summerh. (Dactylorhiza x Anacamptis L. C. Rich.) Orchis L. can also cross with Anacamptis; these are called « Anacamptorchis G. Camus 1892 (syn. x Orchidanacamptis Labrie 1927) 2. x 0: 5) Ovate-lanceolate (width/length 0- 3-0: 5 Lanceolate Res (VPS) ve 3 Shape of stipule Pinnate Pinnate—palmate Palmate é 4 Shape of stipule mid-lobe Leafy, ovate Guidtiilensth = = Gs 5) Ovate—lanceolate (width/length 0-3-0-5) Not leafy, lanceolate ae < 0-3) 5 Colour of lateral petals Cream ; Pale blue or blotched Deep blue : 6 Upper petal/upper sepal Petals shorter than cepa Petals as long as sepal Petals longer than sepal be 7 Spur/sepaline appendage Spur shorter than sepaline appendage Spur as long as sepaline appendage Spur longer than a appendage 8 Honey streak Absent ; Present, Gaocatiched: Present, branched 9 Stylar streak Absent Intermediate .. Distinct 10 Stylar flap Absent Intermediate .. Distinct iy Pollen assemblage More than 60% of grains 5- eolpate) 40-60% of grains 5-colpate Less than 40% of grains 5-colpate WNrF WNP WNP WN RK WN RK WN RH WN KH WN HK WN KH WNeK N The computational stages in the analysis of the two sets of data (R-technique) are shown in the flow-diagram (Fig. 2). As will be seen a component analysis as well as a factor analysis (sensu Kendall 1950, Rao 1955) was carried out for each set of data. Most of the computation involved was carried out by the Ferranti ‘Pegasus’ electronic computer of Southampton University, although, at the start of the study in 1959, the standardization of data and calculation of correlation matrices had to be undertaken with a hand-operated calculating machine as these were not yet programmed for. This was practical for the qualitative data where relatively few attributes were involved but the quantitative data had to wait until the whole analysis had been fully programmed. Rotation of axes for ‘simple structure’, sensu Thurstone (1947), was not attempted. 3. RESULTS (a) Qualitative data Inspection of the residual matrices during a centroid analysis of the correlation matrix of the qualitative attributes showed that at least 2, and probably 3, factors were significant. The factor-loadings of the attributes for these three factors are given in Table 4 and their configuration in the three dimensional common factor space shown graphically in Fig. 3. These show that the factor-loadings are orthogonal to the axes and that they fall into two Watsonia 6 (3), 1965. 144 A. PETTET TABLE 3. Measurements used as quantitative attributes. Attribute No. Type Measurement 1 Foliar attributes Length of flower stalk (measured when fruit mature) 2 Width of leaf 3 Length of leaf 4 Length of leaf stalk 5 Number of notches on leaf (left side) 6 Stipular attributes Length of stipule base 7 Width of stipule base 8 Length of stipule mid-lobe 9 Width of stipule mid-lobe 10 Length of stipule mid-lobe stalk 11 Number of teeth on stipule mid-lobe (left) 12 Number of lobes on left side of stipule 13 Number of lobes on right side of stipule 14 Length of upper left lobe of stipule 15 Width of upper left lobe of stipule 16 Petaline attributes Width of lower petal 17 Length of lower petal (including length of pollen chamber) 18 Length of spur 19 Width of lateral petal 20 Length of lateral petal 21 Length of limb of lateral petal 22 Width of upper petal 23 Length of upper petal 24 Sepaline attributes Length of lower sepal 25 Width of lower sepal 26 Length of appendage of lower sepal Pi Length of lateral sepal 28 Width of lateral sepal 29 Length of appendage of lateral sepal 30 Length of upper sepal 31 Width of upper sepal 32 Length of appendage of upper sepal groups. The majority of the attributes have high positive loadings for Factor I, low loadings for Factor II, and variable loadings for Factor III. Two attributes, i.e. the shape of the leaf and the shape of the stipule mid-lobe, form the second group with relatively low negative loadings for Factor I, high positive loadings for Factor H, and negligible loadings for Factor III. The configuration of the factor specifications of plants in the common factor space is also shown graphically in Figs. 4(a)-(c). The distribution of plants within this space suggests the presence of 2, perhaps 3, groups of plants largely through the effect of Factor I. The upper group consists of the Luteae, Curtisieae, Saxatiles and the blue-flowered Tricolores, while the lower group consists of the Arvenses and Nanae with the cream-flowered Tri- colores somewhat separated from them and perhaps composing a third group. Factor I, therefore, serves to separate the outbreeding from the inbreeding pansies—not un- expectedly in view of the high loadings that the attributes concerned with reproductive organs received from this factor. Factor II does not resolve the plants into any of the major categories so far recognized. Factor III, on the other hand, tends to separate the blue- flowered Tricolores from the Curtisieae in the upper group and the cream-flowered Watsonia 6 (3), 1965. FACTOR ANALYSIS OF BRITISH PANSIES 145 2h, 27, 30 i+ “——} Bol, UPPER PETAL Fig. 1. Measurements used in analysis of quantitative data. The numbers refer to attributes listed in Table 3. Tricolores from the Nanae and, to a lesser extent, the Arvenses in the lower group. The tilt of the cream-flowered Tricolores and Arvenses suggests that within these the Factors I and III are not acting completely independently of each other. : These factors are algebraic abstractions necessary for the solution of various matrix equations. This is not normally a unique solution and in consequence, the factors cannot strictly represent the operation of biological factors (in the broadest sense). None the less their effects may have a parallel to those of biological factors and from an examination of the configurations of factor loadings and factor specifications it is possible to suggest a biological interpretation of the variation. In the present case Factor I has the overriding effect and most attributes have high loadings for it, in particular, the presence/absence of stylar-flap, the type of pollen assem- blage and upper petal/upper sepal ratio. Since it brings about the separation of groups on the basis of reproductive behaviour, Factor I can be compared with the genetic complex responsible for the evolution of the inbreeding pansies from the outbreeding ones. It may also be considered the ‘general taxonomic factor’, since the attributes affected by it are important taxonomically. Watsonia 6 (3), 1965. 146 A. PETTET standardisation of data correlation matrix estimation of the number of factors and calculation of communalities component loadings of attributes factor loadings of attributes component specifications of plants factor specifications of plants ~ FACTOR ANALYSIS COMPONENT ANALYSIS Fic. 2. Flow-diagram of computational stages in factorial analysis. TABLE 4. Factor-loadings for qualitative attributes for 3 common factors. Attributes Factors I II lil 1 +0-70 +0-12 +0-32 DZ —0:28 +0-75 +0-02 3 +0-60 —0Q-07 +0-49 4 —0-02 +0-74 —0-02 5) +0-63 +0-09 —0-14 6 +0-80 —0-02 —0-22 7 +0-61 —0-03 +0-23 8 +0-57 —0-06 —0-42 9 +0-55 +0-03 —0-08 10 +0:96 - +0-10 —0-13 11 +0-94 +0-09 +0-02 The second factor which affects virtually only the shape of the leaf and the stipule mid-lobe is a ‘leaf-shape factor’ acting--independently of Factor I. Previously this was identified as a morphogenetic factor associated with the position on the axis of the node scored for the analysis (Pettet, in Williams & Lambert 1961). In all taxa the shape of the leaf and the stipule mid-lobe varies from the base of the plant to the apex and, to avoid unnecessary complications, the node taken for scoring was always taken near the top of the plant. Those chosen in plants with few nodes per branch would, of course, not be at the strictly comparable stage as those chosen in plants with many nodes per branch. It was originally thought that the analysis was revealing this difference as Factor Il, but now this Watsonia 6 (3), 1965. FACTOR ANALYSIS OF BRITISH PANSIES 147 1.0 T +1, ay or] vo pA] (c) Fig. 3. Factor-loadings of qualitative attributes in common factor space. 1 = annual/perennial; 2 = shape of leaf; 3 = shape of stipule; 4 = shape of stipule mid-lobe; 5 = colour of lateral petals; 6 = upper petal/ upper sepal; 7 = spur/sepaline appendage; 8 = honey streak; 9 = stylar streak; 10 = stylar flap; 11 = pollen assemblage. is considered not to be the case. The difference in leaf and stipule mid-lobe shape is not very great after a few nodes from the base and would be too small to account for the effect of Factor II which, on closer study, appears independent of the number of nodes on the plant axis. The inbreeding forms, which are generally many-noded, have as wide a range with respect to this factor as the outbreeders which include plants with a varying number of nodes, while the factor does not separate the blue-flowered Tricolores with many nodes from the Luteae and Curtisieae with a smaller number of nodes. The existence of this ‘leaf-shape factor’ and its independence of the first factor have been largely responsible for the recognition of many of the microspecies by Drabble (1909, 1926, 1927), particularly within the inbreeding or partially inbreeding taxa where emphasis has been laid on leaf and stipule mid-lobe shape. Factor III is of low order and it is rather difficult to suggest its biological parallel. It affects most of the attributes showing high loadings in Factor I while the factor specifications Watsonia 6 (3), 1965. 148 A. PETTET ° es | & ) = in .o) Vv o 2 6 Py 2 ca : | our ® & "1 5 3 - | Vv Pos et pe eee eee ee 2.0 —- 1.0 +! + Lae ce lee i + V, Vv Vv ° O | Oo Vv + ® ¥ @ 7 ®@ re) =1.0 ro) re) @ et + | e ® x j ® Tee «c) Fig. 4. Factor specifications of plants for common factor space. (a) Factor I/Factor II; (6) Factor 1/Factor III; (c) Factor II/Factor III. 7] = Luteae; <> = Saxatiles; \/7 = Curtisieae; = © blue-flowered Tricolores; + = cream-flowered Tricolores; (} = F, hybrid, V. tricolor (blue-flowered) x V. arvensis; @ = Arvenses; * = Nanae. show some separation of the blue-flowered Tricolores from the Luteae, Curtisieae, Arvenses and Nanae, paralleling the edaphic ranges of these plants. The latter groups which are affected by Factor III in a similar manner, are plants of neutral to basic soils whereas the blue-flowered Tricolores are plants of acidic to neutral soils and the cream-flowered Tricolores probably occur mainly in neutral soils (although information on the edaphic ranges of these plants is scanty). This suggests a connection between certain combinations of attributes and habitat preferences and one which might have been demonstrated more convincingly by the inclusion of ecological data had these been available for all plants. Watsonia 6 (3), 1965. FACTOR ANALYSIS OF BRITISH PANSIES 149 The component analysis of the complete correlation matrix gave much the same configuration of component specifications as factor specifications (Figs. 5(a)-(c)). The major groups are, however, much less distinct because of the increased range along the axis of the first component and a corresponding contraction along the axes of the second and third components. This difference reflects the greater part of the total variance accounted for by the first component and the smaller parts by the second and third components. a ii ii a O O le S| aie Qo © Ee 4 O win ee 2 eels “ +03 S i) 2 Peay Vv % Vv i Vv i 7% + ©) O @) I a ont fe I Ce es Se mp $j 19 ~0E i Ort +05 +1.0) (69 amine a) Vise at | - . +). +} 4 @ | : @* 34 ® © e ar @ ° o ae of e ar V3 oF 8 ve ©$ e ee ® | ie ae (a) (b) I “41.0 40.5 ® eV % bedi Con O + tomes oO Vv Vv i eee SS SS a ae 14.0 -0.5 ae a +05 +1.0 Ta. (c) Fig. 5. Component specifications of plants for first three components. (2) Component I/Component II; (6) Component I/Component III; (c) Component II/Component If. (] = Luteae; <_> = Saxatiles; A, = Curtisieae; © = blue-flowered Tricolores; + = cream-flowered Tricolores; @ = F, hybrid, V. tricolor (blue-flowered) x V. arvensis; @ = Arvenses: * = Nanae. Watsonia 6 (3), 1965. 150 AS PETTED The difference in the effect between the first component and the first factor can be seen in Fig. 6 where the specifications for these alone have been plotted. The factor specifications show a complete break about the middle, separating the two main groups, and show, what is perhaps more obvious here than before, that the first factor alone is insufficient to show any definite break between the cream-flowered Tricolores and the Arvenses although a short steepening in the lower group can be seen about a third of the way from the left. In the component specifications there is no similar break in the slope although there is some steepening in the position corresponding to the break in the factor specifications. The two types of specifications can be compared with the upper line which represents the distribution of total attribute scores of each plant. As was first pointed out by Valentine (personal communication) this distribution resembles that of the first component specifications except for some minor changes of position. This is in some ways fortuitous—the resem- blance arising because most attributes had high positive loadings for the first component. Had there been fewer of these high loadings for the first component there would have been little resemblance between the two. _(b) Quantitative data In the factor analysis of quantitative data four common factors were extracted. The . factor loadings of the attributes for these four factors are given in Table 5 and shown graphically in Fig. 7. The only clear-cut separation of attributes occurs for Factor II where TABLE 5. Factor-loadings for quantitative attributes for 4 common factors. Factors Attributes I II Til IV 1 +0-502 —0-076 —0-381 —0-250 p) +0-652 +0-168 —0Q-535 —Q-212 3 +0-704 +0-282 +0-073 —0-371 4 +0-451 +0-448 —0-444 —0-368 5 +0-510 +0-161 —Q-187 +0-278 6 +0-640 +0-186 +0-460 —0Q-147 7 +0-781 +0-143 +0-058 +0-276 8 +0-702 +0-439 —Q-397 —Q-215 9 +0-659 +0-376 —0-483 +0-209 10 +0-196 +0-372 —0-633 +0-013 1] +0-386 +0-400 —0-372 +0-565 12 +0-513 +0-219 +0-078 +0-579 13 +0-372 +0-102 +0-333 +0-556 14 +0-584 +0-282 —0-414 —0-047 15 +0-592 +0-018 —0-272 +0-231 16 +0-313 —0-876 —Q-193 +0-078 i7/ +0-257 —Q-790 —0-264 —0-035 18 +0-158 —Q-730 —0-160 —Q-175 19 +0-358 —0-862 —0Q-154 +0-182 20 +0-186 —Q-938 —Q-199 +0-004 21 +0-299 —0-628 +0-152 —0-047 22. +(0-337 —0-864 —Q-182 +0-179 23 +0-149 —Q-957 —Q-168 +0-001 24 +0-846 —Q-092 +0-215 —Q-175 D5) +0-715 —Q-107 +0-283 —0-060 26 +0-816 —Q-075 +0-205 —0-052 Pa +0-854 —0-146 +0-233 —0-098 28 +0-743 +0-049 +0-319 —Q-022 29 +0-813 —0-023 +0-300 +0-008 30 +0-861 —0-007 +0-274 —0-160 31 +0-673 +0-040 +0:-221 —0-082 32 +0-834 +0-128 +0-:263 +0-031 Watsonia 6 (3), 1965. 151 FACTOR ANALYSIS OF BRITISH PANSIES ‘ORUCN] = 4 :SOSUDAIY = @ ‘sisuaddD *A X (paroMo-onyq) 40j09141 (4 “priqky by = © ‘sosO[Oo], PatoMOy-wievsto = + ‘{so1of[ooi], posomoy-ontq = O ‘ovoisnind = /\ ‘sopexeg = <_> ‘ovayn] = [J ‘eouarusauo0d Joy AyyeyUOzIIoy pooeds ore sjuvrjd oy, ‘suoreoyToads 10jORy ISI OY} ‘MOI JOMOT OY} puv ‘sUOT}VOyIOoOds JUOUOdWIOD jsIY OY} ‘MOI Q[Pplu ay} {soIOOs 9)NgqII1}}e JO suns jUaSoIdeI MOI Joddy ‘suONeoyIoods JoJORJ ySIY puv JUSUOdWIOS JsIY YIM SorOOS ayNqII}}e JO Stuns Jo UOSTIVdWOD *9 “SIF oh Sele @ XO Tote, ++ y 0 i 0 0'0 Z af -- eli = O06 O0G9 © O20 00ceeceeds 4o © OOA teccece, O SO4AAO goad AA Cex OA0cocnnAacg S "+ 00 +O, g‘i+ 00 O+ C09 “01+ ah 3 01+ A000 Be A0044A1A000,, Se DA e+ =a 0'\+ ALLA INT eeeee OOO ot 66 406066+4+4+ NS @@+ DS 02+ a ++0O oz+ -s i *++ oo) ++0A ° \o A AS} +OO0AA S OOAAA “< O000OO04An = AO OOAAODO oC+ | | AOOO oe 152 A. PETTET Oo . | oi 08 i B_19 If | 10 Fig. 7. Factor-loadings of quantitative attributes in common factor space. Mi = foliar attributes; CJ = stipular attributes; O = petaline attributes; @ = pedicel and sepaline attributes. Watsonia 6 (3), 1965. i | een FACTOR ANALYSIS OF BRITISH PANSIES 153 the petaline attributes (high negative values) are separated from the sepaline attributes (low negative and positive values) and the stipular and foliar attributes (moderately low positive values). The second cluster of attributes is partially resolved by Factor III which partly separates the stipular and foliar attributes from the sepaline attributes. The configuration of these attributes in the four dimensional common factor hyper- space can be examined in successive permutations of three dimensional space. The effect of Factor II can be seen in I/II/IIH, I/T1/1V and I/II/[V common factor spaces. In I/II/III and I/II/IV the attribute loadings separate into two groups, viz, a separate cluster of petaline attributes orthogonal to the three axes and a cluster of sepaline, stipular and foliar attributes forming a group oblique to all 3 axes. In II/III/{V space, the cluster of petaline attributes forms a group orthogonal to the axes very close to the positive end of the axis of Factor II, while the sepaline attributes form a small cluster near the centre of the three axes and the stipular and foliar attributes are spread out into a relatively flat disc roughly in the plane of the Factor II axis. The I/III/IV space has but a single cluster of attribute loadings oblique to all three axes with the sepaline attributes forming a relatively denser node towards the positive end of the axis of Factor I and the petaline attributes forming a similar node at the base of the cluster. The configuration in the common factor space would seem to indicate, then, that there are three groups of attributes, i.e. sepaline, petaline and foliar-stipular attributes— the last mentioned not being resolvable on the basis of the four factors extracted. The biological factors affecting these groups of attributes will clearly be of a morphogenetic nature since the analysis was based on measurements alone. Factor I affects all attributes positively, the attributes of the green organs having middling to high loadings and the attributes of the non-green, petaline organs having generally low loadings. The factor could be called a ‘general growth factor’ but, since the sepaline characters show the highest loadings for this factor and the foliar-stipular attributes have lower loadings with indications of interdependence between this factor and Factor II, it is possible to envisage this factor as the ‘sepal-growth factor’. Factor II, on the other hand, could be considered the ‘petal- suppression factor’ since the petaline attributes have high negative loadings for it whereas the other attributes have generally low loadings, except for some stipular and foliar attributes with positive loadings of middling magnitude. The effects of Factors I and II emphasize the difference between the sepaline and petaline attributes. The former have the highest loadings for Factor I and the lowest loadings for Factor II; the latter have the lowest loadings for Factor I and the highest for Factor II. This demonstrates that variation in these two parts of the flower is produced by two different, independent factors and confirms the value attached by taxonomists to general flower size and the petal-sepal ratio. Factor III would appear to be a foliar-stipular factor affecting mainly stipule shape and leaf size. Its highest loadings are on the length of the stipule base (positive), width of leaf (negative), and length of the stalk of the stipule mid-lobe (negative). Similarly, Factor IV seems to be a foliar-stipular factor affecting mainly the relationship between leaf size and degree of dentation, stipule size and amount of lobing. The factor specifications of the plants for the four factors show a somewhat different pattern from that of the previous analysis of ‘qualitative data’ and clearly demonstrate the overlapping variation of related taxa when morphometric data alone are considered (Figs. 8(a)-(f)). The different factors only partially resolve the recognized taxa. As might be expected the best resolution occurs with Factor II, the ‘petal-suppression factor’, which partially ‘stratifies’ the taxa into relatively narrow segments, going from positive to negative in the following order, Nanae, Arvenses, cream-flowered Tricolores with the Curtisieae, Saxatiles with the blue-flowered Tricolores, Luteae. Factor 1, the ‘sepal-growth factor’, does not appear particularly important but partly separates the smaller Curtisieae from the blue- and cream-flowered Tricolores which are themselves partially resolved by Factor IV. The Nanae are almost separated from the Arvenses by Factor I, as well as Factor I, confirming the idea that the small size of the plant and the very small petals are almost adequate features for separating the group from the Arvenses. Watsonia 6 (3), 1965. a aT A. PETTET 154 poloMopeuresis = + (J) ot- sii 8 ~ - — ~ —* ‘oBvUEN] = y {SOSUDATY = @ ‘sisuaduDv “4 x (po1sMOY-onjq) 4ojod141 “4 ‘praqhky ly = © ‘sor0jos1i |, SSOIO[ODII[, PoIoMoy-en[tq = O {TIL 1010¥,4/TT 10398, (P) SAT 10108.4/J 103984 (9) S]IT 101984 /[T 1019k4 (9) ST] 101984 /T 10}0e,4 () ‘90edS I0}ORJ UOWIUIOD IOF s}uK[d Jo suoTywoytoods 1OJOV{ gs “SIA SovoIsnINg = A ‘sopyexeg = <_> (2) Her I Qo O+ o2- Oo (oe) (ome) oO (eo) A Me) bt o 0 ro) (o) 18) SA 8 24\c6 fe + Sey Xe) ao We Oe es ou ttt Saw 4 + i ou; oe; a Al at es Al ee Od @ oe @® Bite vs & & & id e Eg > e @ el (9) fovonyT = “AT J0pey/IIT rowey (Sf) SAT 103984/T] 10}Oe4 (2) (P) nf- all a] ov 2] Oo (o} Lexe) oO 9 > (5) 4A i On4 Oo ou 0% Onaet e a A 7 A vAN spas eee Say “ Oe eee 0G @ rN 4 + I ® “oa? eM o.» *es © ® @ @ Bd 3 t ® 6% ou off Vatsonia 6 (3), 1965. FACTOR ANALYSIS OF BRITISH PANSIES 155 Inspection of the configuration of the factor specifications in the four dimensional hyperspace by means of different permutations of three dimensional spaces shows roughly tetrahedral or spherical clusters of plants. In the I/II/III space, perhaps the most important 3-factor space, the roughly tetrahedral cluster shows very clearly the stratification along the II axis mentioned above, as well as the partial separation of the Curtisieae (negative) from both cream- and blue-flowered Tricolores (positive) and the Saxatiles (positive), and the Nanae (negative) from the Arvenses (negative and positive). The Arvenses group forms a fairly distinct segment with the approximate boundary between it and the other groups lying in the +0-5 plane of the II axis. The tetrahedral cluster appears slightly tilted with respect to all axes, suggesting some interdependence of all three factors within the plants. The stratification along the IJ axis is also to be seen in the roughly spherical clusters of the I/II/1V and 11/II1/IV spaces although the change of factors occasions some small alterations of relative positions of the various groups. Within the I/III/IV space there is a less obvious sorting within the cluster. The cream-flowered Tricolores, blue-flowered Tricolores, Cutisieae and Luteae tend to occupy different positions towards the periphery of the spherical cluster but there is a great volume of overlap at the centre and the Arvenses are distributed more or less throughout. In the configurations involving Factor IV there is a fairly sharp truncation of the cluster in the plane of the IV axis at about —1-5, beyond which there are no other plants except for a single extreme example of the Arvenses. This exceptional plant, No. 19 of Table 1, is one with exceptionally large vegetative parts. The leaves and stipule mid-lobes were unusually long and wide and their petioles, like the flower-stalk, were also unusually long. Since all these attributes received moderately high negative loadings for Factor IV, the plant is situated towards the negative end of the IV axis. The abrupt termination of the main cluster suggests that a segment of the variation in leaf and stipule characters is missing. Plants with large leaves and stipules like No. 19 have not been included in the analysis, presumably because for some reason they are rare in Britain and consequently were not available in any numbers when plants were selected for the analysis. The corresponding component specifications of plants are illustrated in Figs. 9(a)-(f). Since most of the total variance has been extracted by the first component, and a smaller amount by the second, and little by the succeeding components, the distribution of these specifications is less revealing of the variation than that of the factor specifications. The configuration in the I/II/IIJ component space is rather similar to that of the I/II/III factor space and shows a similar distribution of taxa with the stratification along the II axis _ emphasized in IJI/III by the relatively low values for component HII. The generally low values for components III and IV obscure the distribution of taxa in the other three dimensional spaces. DISCUSSION There were two aims in carrying out this factorial analysis of British pansies: firstly, to determine the probable factors (in the broadest sense) producing the variation, and secondly, to see if the plants separated into discrete clusters corresponding to taxa already recognized. The first has been largely realized; the second has been less conclusive— clear-cut clustering of specifications within the common factor space was not obtained in every case. The condensing of the variability by component analysis was also far less successful than expected. There are reasons, of course, for not expecting a complete resolu- tion of the variation into groups corresponding to the recognized taxa. To call the group a ‘critical’ one is tantamount to saying there are no obvious discontinuities in the pattern of morphological variation. In the particular case of the pansies this has been overcome by attaching considerable weight to cytological and ecological features, and as criteria these admit overlap in morphological features. Before discussing the main aspects of the analyses the limitations involved in the method should be stressed. The results apply to particular sets of attributes for a particular selection of plants. The latter feature—the selection of plants—seems to be a limitation of minor Watsonia 6 (3), 1965. ——— ee — ———— — ———--— ——— ~——_—~ . “S00 Nl = ae fsosusAlyY = @ ‘sisuadav “4 < (perOMOP-on]q) 4oj09147 “4 ‘praqdy 1a = © ‘sosopoolsry, poloMoy-wivsio = + ‘fsolo[Ools1y, postoMoy-onjq = © ‘ovsisning = /A fsopyexeg = <_> ‘fovoiny = [] ‘AT JUoUOdWIOD/TIT }UoUOdWIOD (f/f) SAY JUouOdWIOD/[]T JUOUOdWIOD (a) ‘]]]T JUSUOdWIOD/IT yUoUOdWIOD (p) * AT yUouodWoD/T quouoduioy (9) ‘JJ yUouodwog/t youodwoD (9) ‘{[{ JuouodwoD/] JusUOodWIOD (v) ‘syusUOdWIOD moj ysI JOJ sjuejd Jo suonmeoyioods jusuodwioD 6 ‘“sI-{ (S) (2) (p) oor I sate IL roc- I a} 2] vor- Oo 0'0\- C vy PD -@ oO AO a @ ‘3 aed q 12 ee o Yea ne 6 e Ot © 01- 0'02- ee pet eae ome tee Sous HAY be aE iN Al ys Al vn E50" il ) ) ® e® eo © “Re ok eg @, e x» € \ vor e Ke ee a 3 oer eo 6 ht i 00l+ I outs I < (9) (9) (v) oo~ I ce ] o0c i{ % a ‘00e- * an A 5 ke & Al cP 4 © Aa “4 # as e O1-~ © ~e @e@e + 001 ¢ 6 soot @ _ uae) £.) H x oo 4 oO A fo) Y oer ovr+ ou+ Leh pe eu re a+ o+ oul + “e OW tore 0'02- 0vt- ome Ae " Cy bad [ eae : ve oa ou, Al + +O Q A il ® 9 He. i I ec et Q 001 2 I ® 42 I ry of + ® e ® at 4° sing, @ ee © | oe 0 + a° @@® e é @ @ Ne) b ° fe) a oer ant + ore. e @ @ fe) @ Ye) a + + vot ell Cx [ © raal| Watsonia 6 (3), 1965. FACTOR ANALYSIS OF BRITISH PANSIES LS y7 importance. Initially the number of plants was limited by having to carry out the first phases of the computation using a hand-operated calculating machine. To overcome this a careful selection of plants was made to represent, as far as possible, the full range of variation present in the British Isles. It should also be remembered there is a ‘diminishing returns’ relationship to be considered and beyond a certain stage extra plants do not materially affect the picture revealed by the analysis. It is believed that the number of plants used here represents approximately the optimum number for what was expected of the analysis. The results of the two parallel analyses (of ‘qualitative’ and ‘quantitative’ characters) show that the content of attribute sets is of importance in affecting the picture obtained. At first glance the results derived from the two analyses are quite different, but closer examination shows underlying similarities. In the factor analysis of the qualitative data many of the attributes used were associated with reproduction and it is, therefore, not surprising that the analysis gave a clear-cut division between outbreeders and inbreeders in the first factor specifications. There was no marked resolution with respect to Factors II and III but this might have occurred had more attributes been used. The component analysis produced a similar distinction but the retention of specific variances obscured the difference. This primary division on reproductive behaviour reflects the probable recent evolu- tionary trends in pansies, some details of which have already been discussed (Pettet 1964b). It is apparent that ecological adaptation and subsequent isolation has been accompanied by a certain amount of morphological differentiation which has, in some cases, been associated with cytological changes. These ancestral types were entirely outbreeders and one recent, but important, change has been the development of an inbreeding system associated with the exploitation of transitorily favourable habitats, e.g. arable farmland and dry sand dunes. This change from one system to the other has been picked out by the analysis. In addition, an incidental feature of Factor I has been the demonstration of the high degree of correlation between most of the attributes. Of particular interest is the correlation of stylar-flap and pollen assemblage with the upper petal-upper sepal, confirming the value of these features in distinguishing between V. tricolor and V. arvensis (Pettet 1964a). The division of outbreeders from inbreeders obtained in the first analysis is not so clear cut in the case of the quantitative data. These attributes are, of course, of a different nature, expressing differences in the dimensions of floral, foliar and stipular parts. Factor analysis produces no clear breaks in the configuration of plant specifications, instead it _ produces a large indistinct cluster. This corresponds with the general impression of the group where variability is so great in such attributes that most taxonomic characters involve comparisons between different parts of the same plant such as was analysed in the first analysis. It also serves to demonstrate the small degree of morphological divergence between the various taxa—perhaps not remarkable in itself since all taxa are exploiting approximately similar niches in the communities to which they belong. Had there been marked divergence this analysis would not, of course, have been undertaken! The stratification along Factor II in the I/II dimensions is roughly comparable to the arrangement along Factor I of the factor analysis of qualitative data. Since Factor I is the ‘sepal-growth’ factor and Factor II is the ‘petal-suppression’ factor, this confirms the emphasis of reproductive characters as important and consistent features of the variation pattern. The foliar-stipular factors of III and IV show even greater overlap and less discrete grouping of the taxa which partly confirms the idea that foliar and stipular characters have _ limited taxonomic value because of their wide and intricate variability. An interesting point shown by the analysis of quantitative attributes is that the Arvenses exhibit the greatest range within the common factor space. This no doubt arises because the Arvenses represent many relatively homozygous, inbreeding lines developed from heterozygous, outbreeding ancestors and which, in consequence, show a greater range of variation than in the related outbreeding pansies. The related AT-intermediates, here classified under the cream-flowered Tricolores, have already been indicated as representing Watsonia 6 (3), 1965. 158 Aa PERE’ the remnants of these ancestral outbreeding arvensis pansies (Pettet 1964b). This would explain their position in the factor spaces of the two analyses, i.e. lying between the in- breeding Arvenses and the remaining outbreeding pansies in both cases. The morphological similarity between cream-flowered Tricolores and the hybrid, V. tricolor x V. arvensis, is shown by their similar positions in the common factor space of the qualitative attributes. In the case of the quantitative attributes the hybrid shows a close affinity to the Arvenses, especially in the case of the second factor. Factorial analysis should not be considered an end in itself. If done at the beginning of a study into the variation of a group it should suggest immediately the more important aspects to be concentrated on. In the present case the analyses were carried out towards the end of an experimental study on pansies in which the problem of the AT-intermediates had arisen incidentally during a general survey of the group. Had the analyses been done earlier the problem posed by these plants would have been pin-pointed immediately. One gratifying feature of the analysis has also been the way in which the various aspects of the variation pattern were thrown into emphasis at one and the same time—aspects which had been laboriously formulated by other methods. ACKNOWLEDGMENTS I am extremely grateful to Professor W. T. Williams who not only introduced me to the method of factorial analysis, but helped with computation and provided a never failing source of helpful advice and criticism. I also wish to thank Dr. M. Dale for help with computation of the quantitative analysis and the Keeper of Botany, British Museum (Natural History) for allowing me to study specimens there. This work was started whilst I held a Nature Conservancy studentship for which I record my thanks. REFERENCES CATTELL, R. B. (1952). Factor Analysis. Harper, New York. CLAUSEN, J. (1926). Genetical and cytological investigations on Viola tricolor L. and V. arvensis Murr. Hereditas, Lund 8, 1-156. CLAUSEN, J. (1927). Chromosome number and the relationship of species in the genus Viola. Ann. Bot., Lond. 41, 677-714. CLAUSEN, J. (1931). Cytogenetic and taxonomic investigations on Melanium violets. Hereditas, Lund 15, 219-308. Danpy, J. E. (1958). List of British Vascular Plants. London. DRABBLE, E. (1909). The British Pansies. J. Bot., Lond. 47, Suppl. 2, pp. 1-32. DRABBLE, E. (1926). Notes on the British Pansies. The arvensis series. J. Bot., Lond. 64, 263. DRABBLE, E. (1927). Notes on the British Pansies. The tricolor series. J. Bot., Lond. 65, 42-53. HARMAN, H. H. (1960). Modern Factor Analysis. University of Chicago Press, Chicago. PeTTeT, A. (1964a). Studies on British Pansies I. Chromosome numbers and pollen assemblages. Watsonia 6, 39-50. Pettet, A. (1964b). Studies on British Pansies. II. The status of some intermediates between Viola tricolor L. and V. arvensis Murr. Watsonia 6, 51-69. SOKAL, R. R. & SNEATH, P. H. A. (1963). Principles of Numerical Taxonomy. Freeman, San Francisco. THomMson, G. H. (1951). The Factorial Analysis of Human Ability. Houghton Mifflin, Boston, & Riverside Press, Cambridge. THURSTONE, L. L. (1947). Multiple-factor Analysis. University of Chicago Press, Chicago. WARBURG, E. F. (1952). Violaceae, in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles, ed. 1. Cambridge, ed. 2, 1962. WILLIAMS, W. T. & LAMBERT, J. M. (1961). Multivariate methods in Taxonomy. Taxon 10 (7), 205-211. Watsonia 6 (3) 1965. 159 FACTOR ANALYSIS OF BRITISH PANSIES APPENDIX 1. Scores of qualitative attributes. Attribute No. Plant No. Sums of scores 11 10 iu 17 18 19 18 15 13 14 16 16 16 7 16 16 18 16 14 7, 16 13 18 18 18 16 18 16 17 16 18 i) 22 10 11 12 a) 14 15 16 17 18 19 20 21 N 22 23 24 25 26 Di 28 29 30 31 22 DI 32 33 20 20 28 N 34 35 36 Si) 38 39 *40 18 27 21 15 23 4] PU 42 26 24 18 18 19 20 18 27 43 44 45 46 AT 48 49 50 51 29 29 ay cont. Watsonia 6 (3), 1965. 160 A. PETEET APPENDIX 1—cont. Plant No. Attribute No. Sums of scores 53 1 1 2 1 2 3 3 25 1 3 3 22 54 1 1 2 3 1 3 3 pz 3 3 3 25 55 2 2 2 3 3 3 2 3 1 3 3 27 56 2 2 2 3 1 3 3 3 nex yd 3 3 27 Vow 1 3 2 3 2 3) 3 v2 1 3 3 26 58 2 2 3 3 3 3 3 2 3 3 3 30 59 2 1 3 3 3 3 3 3 3 3 3 30 60 2 1 3 3 3 3 3 3 3 3 3 30 61 2 1 2 2 3 3 3 3 3 3 3 28 62 2 1 2 1 1 3 3 2 1 3 3 22 63 2 2 3 2 1 3 3 2 2 3 3 26 64 2 2 2 3 1 3 3 2 2 3 3 26 65 2 1 3 3 1 3 3 2 3 3 3 27 66 2 2 3 2 3 3 3 2 3) 3 3 29 67 2 3 3 3 3 3 3 2 2 3 3 30 68 2 2 2 2 3 3 3 2 3 3 3 28 69 2 1 3 2 1 3 3 2 3 3 3 26 70 2 1 2 2 1 3 2 2 3 3 3 24 71 p) 2 3 3 1 3 3 3 3 3 3 29 WZ p2 1 3 2 3 3 3 3 3 3 3 29 73 2 2 3 3 1 3 3 3 2 3 3 28 74 2 1 2 2 1 3 3 2 3 3 3 25 75 2 2 2 2 3 3 3 3 3 3 3 29 76 2 2 2 1 1 3 3 3 3 3 3 26 id 1 3 3 3 2 1 3 2 1 1 1 21 78 1 1 3 1 2 1 3 2 1 1 1 17 * Although Drabble named No. 40 as V. vectensis it does not in fact agree with the description of the species. + The herbarium specimen of No. 57 is an annual notwithstanding the published accounts of this plant. Watsonia 6 (3), 1965. | meSQegnnns: ~SnNSornrann< Qpatatatam QS OBS Seee:s ~onr~n m™NM ¢ wn ANAMNAMAAS SLeesegvee;s Nr OM m~ ome RZ loawnwwocnn: | AA\2QOQoeae 4 SSnnunNnSS< Q NNMNMNMNMNNNM TFT aOanoNN NM CO Or NrNNANhe NH: a lAaAarnannaane MAAS Oeawoea, ~SSOHnSrnHMS NAN | enconnwor oo = A Yr Se enero ao om MNnNONnNNM—- NO ¢ a | Att anat ot aoannnNoOoN OC OS Sa = or oom a | ARNANANMNA N}onnnamovwe — ae co = SIQ29 06 6 _|eeesess8seE TOONTHEHE ~~ £ S APPENDIX 2. Measurements of quantitative attributes (in mm), Attribute No. 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 10 No. Plant esnnn 0 eseoense woescenuneocsooesosnnnansssunsoon Ra eseseenesseesesssaassrssss egeasessesssenesssezesasaicerssseussenesenussesseas 2 Sas Oe Se Ae A NAGA AO ROO S MAS HAS DS SOON SO OC Se AC ie eee Ie ral 0 Peel a) Seeununesounsneoonunnssenesneusunsonny Pale) ree CSS SR ARERR AS SASSO SSSSAA SAR AR SSF SR GRRARR ARE AeR CH SAHSCHRGLSRGAASRARAASKAABaS AaaAmansansaaaie ARAMARK ARK RRKRANAARNAANAAAANAAAR=RAN RRA AAR ATA MNATRN RRB RANA NR N RNR S eesces enue mwneneonussocsessuuneose esesesesesesssensvercusneesasassaaesesessacusrssensusseaaagsssecasssssussssssse ate es Ce Ce Oss oa So COC CC ACS Abe GO CONC O Lanes =) 0 al enoousoocoson eoseeosononosenonnsnnnsonnson SS SSSSSS See SSS SSSR SeaR SAS SaQSSSSsses Sass sgssrssgssasssssscgeasadsnaggeassenaagese OOS So Rae SO AA AES Mae KIA OAD SAA AAG AMA HAO AAAOOVOAIe aaa e SAGAR sea Se ‘0 mo ” 1 ooo menseonscenesoooneuuunsssooneooeneosoounNy Pal CS OERESSRSRRS REALS SONS SK SSL SSS SSSSSSSLSLS SK SSK SSSSALRLKLAASSAASSASASACLaassne AHANAAMBAAN Raa AN RR NR RANA RMAANAAANRRANARANAARANR ANN R RNR RRR RRR sean eee seh 0 a) nenon ) oo wescosesenenessoosos eouvucoenonscoso eeseseesssusseesencsseansesnsessqsrecssasssaagsqsesaassasagasssasassaaacascacasa SCASHDHORHDOSSOHSHHD OETRRHDSRR ODE HER ORHOSOCHDDESHE HOR RR ANDRE HE oMUroOrrernesnonnreroernonnnnn 0 0 0 nennnonooun wonunesesoouseneoeansoonnyg weonneon esseeecssusessssesesssagausecrsseseenueensseuseesseanerssarscsasaggrenesesgeesae Zo Sabo aS SA FOE HINA DARIO AD SAO FOR SOA DOO BOAO SOMOAOA OOO Se AMAA OGM Sie Palade) ono e wnnnooono wnonSSeNnNSSoHONONOSONSSNUSSNSSSOSOOeNS 10 Ra Saeversssrasseresssesqeersesnesssausneasngsauscanasksscascsagarsskasaseaaasaeas ABAAAMADMARAARAARAAAAATHANAARAMAMMAAAAACAAMAAAACTAAAAAAMARARANAAARARA AA AA AAS ARAN al So a mwnsosnesennnncoonunaeneoocsonose nonneseouscosos Pal SSSSCSeSSSSeSRZeSgSseRAeSg GSA QseaQeAaSeesagssaeggsseeaasSs SACRA saassagssserRasesegaas BSR SMAMHAARRERSSASAAALAANAAABDAAKDADDGADAD AAD MDAAANNOAAARNDAHADMDEASSHNHN SSK Arar Nnnonoe e Pale] S > ral Pa wnusonensos ° oes enonneoocsoonsossoooenssososonuneon SSSRARSSSSRLSSSHSSGSLSSLSSA GL SSR SASS SSSSSSSSSSSSRALALLAGSRAGRSASSSASSSHSALL SAA SHTAROR ASR SCHTHETEHASCHEHHADROSHTHAAANDEALOAANDS HHH HDSSSHMSHEATHTMANANDASCMDOMARHAMMAHOAMHAOAT nea Pal to ro) een unoconueoueosesoooeooeSoeSoOoNSeNSo ON BSQSSSHRS RLSM SLSSSSR SELL SSSSLSSSLSSSSSURASSSSSSSASAASSSLSASASASSSRSRSSSLSASAL ADHDAFTMAPEAMDAAAAAMDMMAATAMMATMTAMAAMAMNAARK NR SHHOR ADDO RMDODOORMNEO MR HOH-ODtTOTHENEESOANNNEAN neo ooo wnonon or) reir) =) eceoosescoseo wocoeossssosoosos ecoeonuscoue RELSSSSASSSSSSSRLSLSLSSSLSSSSRLSSSSsSAagsssssSgsanueggggassSsSsSasAggassgassssarsses ANMAMAAADAGDAAMAMNAAAGDAMAAAMANMANATAMNAGDATMAMANTENTAAMAMMMMAMNTIEMEMTMMMMMMMMENTIEMANMMAA aunoocoonucosoooone ral nuonsesoeonensoonnoe 4 eononenesenesoseooounocse eonees SRSSSSLSSSSSSSLSSSSASAGSSS LLL SSSSSLASGS SAAS SASL S SSSSAlRLSKS SSK SRsssarasagssagsa MARTHETAGTHEMAAAATHESESSTSENMOHETFANDHROATHE A SORMOARDAAHASSCSCHEDHZAMPAALPABHDAKRHOFTOSOGCOMAHOWDODOOONAN ennsonvounnonoenounscononnonnneue nao 0 Paina) ~ ucoenunsconenocsosonoscenoseson PaveresaeeecseersensaeseresreseunassesseqeesssuanaasessceagesseRgsssasRgeeasasse AAFSEMENOTPTTMDAADAANTDAAMNMAMANMMTANMHRORMHEOOSNOSCORHEORHESHEHNAPNHOENHNAOCTHNNTTHOSCHENNHEEANR onenunncso ecounsunsoounounun e oon Rael a wnenesoounoneoun nennoonu SRSLLASSSSSSSSRLSASSGSAALASSASSSSSSSCSSRARGRSgsesseeesessggecAgAesaegagsscscsaassge ABMAMIUMTMAMAMMAMDMAMMAMMSTMMMMNMENTHEMNTNtTEAtTtemensentetHNHteHe Nt NOT AAMT HMHTHAMMANG wbooonsonsssosooaSoSooNnonNnSoOonSOoONSoONSeNSOSSeSSSSeeNnIN 2 en io encoce aeeeasnsescecs ces asenocsrsasne casos sted. saceqsenassesseastesereassacsag BEA Koh AsSMmar Sr MBABDAAARK HE SHAR SAU MTSU AANMSMAMAGIAAASAPTIAANMAOHAMAADCHSAAAMHENTRZOONO eencoeusesoeosoncooenunonesseoonunenosoesounesoSsOoS 2 to exes enn BBASSSLSSSSSSRASSCASASSSSsSqesegAQsAagrasAaaagassAssAssaaassesAssasassagagassasads MER R HOHE KR SEHEHEHESSCSOHEHR SHER NHODOSUROSHRGAAODSONSSHA AA AMADAMTARAANAHHSRONHreCATNHTTONONT eeneou wnsonscososounuouNnusoensonuNnenS nan woesnnen wenoun Seienisekesueeereses seeker eeeMssexserasessseneneesesueeeresseransseseauesesasersss Te eet wat tl OOOO dt OO OO OO Ol dd ed en ee a ey ensoceoeososeoseoeneunsossesooeesspaeesoeSsSeSessesso eo esooe ° nooo enoeoe eseesssescresessssssssssuskesesessssesseeesesssseassssussssasesssseegsasasaseass SR OH AS SSS BOOS OOS DE OR SOQ OOS BEDS AOS SOAS RODS BE SOROS SSSR OSS OOS GO OE OAS ODS OOS ATOR esesoeseseseesesceesSeeseepeSeesesap ee SeSseSseses9999 esesooesooeseoe ecsesesses Seeesesesssssssssssesssessssesssescssssessssssssesesssssssssssssssssesssssssse AAAAAMAAMANAFARAARRSHTANTHAATATANAAAATAAARAAATAAAAAMABATAAAAAAAAAANAAAATAS AAAS aasa esosseso eesosososcoeo eeesocoos ecescoscos ecsooesesesosesosese ese ssssseesssesssssssssssssesssssessssssesssssssssssssessssssssssssssssssssssssss MHAtHHOSOHHOSHANHT THUONG tHE NTT EHEHT THE NHOMHHOTHE SOHNE THOttetEnNtTANAMntTHETEMANMtOtTITNOTNNAH eeooos00 eee = esse eesoooo eesescece ecosoece eceessscso e sssesesessssssesssssssessssssesesessssssesesssesssssssssssssssssssssssssssssss SS BOS AM DS LAMAN SO OS LODO DANS SOSA es DO DARN Oa SO SS SSMS OOS LOSS OOOO SOOSLOSS eeeesseooecs e S) e e esesoosco ° e eose ecees ececesse e ssesesssssssssssssssesssessssssssssssessses sess ssssss ss sss es ssessgssssssssssssa SHESASMSSAAKAMALVSMATSOSOHPTOSSSSOSSOSCOHASS ESS TANAANSAMASGSSOSOVSSSOHSNSONHTSSASDSOASSOSCSSSOSre wesoosooounusounlooSoo esse SeOouNSoSeo en enennesooonnene4 Reeecsssseteressssssssseussssaekessssseaaeesesessssenvenseeguesseneaesssssacscae AR TPTSHSTPTSOPPTAANTOAHHUHAGDHMATAMANTHOH AMEE MHTTATTTHNDOFTARMAOATAAMARAMHAARAOAAAANATATAT ecoessoeooceseeoeoe oo ono oS Oe eSoe oes y esooeseunesesoen ) SSSSSSSSSSSRSSSSSSSSSLSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSLASSSASSSASRSSKSSSALssa AHROONDMTHTASCMASOIMNHORHBDARMMACMIMHARAAHORHTOSCRHMDHTOMA KN BDHMOH HK DAONATOASCASCHNNNANBAASCTOOT OT eS SSR SS] SAS RO SN OS eae anata Saas SP Cass aaa- asa See Si SS sooososeonoooseSsosoSoeuNnSeooOoSsSeeNnSoSD eeoceununonennsosn SSSSSSSSSSSRSSSSSSSSSLNSSSSSsSsusssasssasssssSssSesesssrssgssaaaggseaeanseasasa AMDAANTHAPFANASSATAAMAAAFAAATAAAMDATTFAATAAAMHAATFANMNAAATATAMAMASTAANATAEROT SAAR AOTAN TT eose eescoseessesouossosossS eo eo ooo ensosonus ssssesesessssassssssseesssasssssesssssssseessssaesseeessasssasasasssaesasacs3ss SPHASOOSCHAANTIAMHAAMSKTAHOTAASHHSSHAHONMTENAANHSTHOOYTHTIANHTOFAMAANTATAAANATAT AAI So eS e eo eseocoocso esses ees ecsesso SSSSSSSSSSSsSsssssss ss sss sss sss ssss sss sss sess sss sss sss esses ss sssssssssssssssssss SS$tonntsrrsrsrnsttHstnntgostnnseattannnnonntnsttet¢etnnonuntannttynronnttttrtntttrnotse TTA So eos e eos eosoe eeo escsssesooeooesesoseseS on ° esssssssssessssssesssssessesesessssssssssasssssssssassssssssssssssassssssesss SAO SSS SAS ROMEO MERN MOD GO DER DOS R ANCA SOS ODOR GOANSE ATEN SS BADER GE MONTE TMEONT OOS oso eesesco eo 2 cos r=} =) esoesceo eecsesees e SSSSSSSSSSSSSSSSSSSSeSsSsssSSSSSSSASssAsSsSsSSsSSsgssassSssssAassassassssSssassgssssass Ht DmOSOSHDRDONTOMOKDOG MMA THEARTHADORMMAGTHARDMASOSHNOHHADEANTENARSMAAAT LET ABAHASCHTMNNTNVO DRA AAMAS SS SEAR ANA AANA MARAE TON RMA Meee QASSesetaaagtsasasas Sit aaF Sasa 2 nesnnooounoeosseocssoseyesouey4 SSSSSSSSSSSLSSSSSSSSSRSSSSSSSG SVS SS SSASSSSSSeUSSSAGSARSsSVgqggseeAgsggsasasaagsaegas SRAARAMSSCABSSTHAHOERAA RENE TSOHREARONSSH ME NERSAR BE BDHSHRADVHDTPTOBAIANIDAMNASCHOSOTENTONNAM™ MIM co nm DAO AOHRATAG motaac =< =6¢ BRHADSAFOHR MSU MNAMDMAEARAASYOSeoS AS aANmEHOoAALS SSSSSSSEASSSASSRSSVASSSSSSRSSLABSSAASTSSHSFRRSSESHEASAGSESSSESGSSISSAIGA FSSHSLAaSas = = = = weraancmanmaoenwacn NITEM BMBAOHRAANMZOLCNLASGRAHAASARAARSASABSSRASSTISITISESSSRAGSAZSS Ss S32 [facing p. 160 ‘atsonia 6 (3), 1965. THE STATUS OF THE PEDUNCULATE AND SESSILE OAKS IN BRITAIN By J. E. COUSENS Department of Forestry and Natural Resources, University of Edinburgh ABSTRACT Introgression in the pedunculate oak (Quercus robur L.) and sessile oak (Q. petraea (Matt.) Liebl.) in Scotland was known to be so extensive that neither could be defined satisfactorily. Population studies were therefore carried out in Eire to clarify the status of the sessile oak and in south-east England and in the flood- plain oak forests of Yugoslavia in an attempt to find good pedunculate oak. Each of these collections revealed some evidence for introgression. A method of analysis has been devised to distinguish between currently active introgression and a condition due either to incomplete evolutionary divergence or to old introgression now largely assimilated into the gene pool. Q. petraea is a relatively good species in Eire and possibly in southern England too. Q. robur is diagnosed as a species, possibly as homogeneous as Q. petraea at one time, but, if so, introgressed by a number of related species during the last glaciations or earlier; it arrived in post-glacial Britain already a very variable species and was subjected to a little secondary in- trogression by Q. petraea. Introgression of both species increases northwards into Scotland. The much greater introgression of Q. petraea in Scotland has been made possible by the practice of planting Q. robur to maintain the stocking of petraea coppice woods. The observed variation in the selected diagnostic characters in the Yugoslavian and Irish collections is outlined to provide a provisional definition of Q. robur and Q. petraea in Britain. INTRODUCTION A study of the partially interfertile Quercus robur and Q. petraea in Scotland revealed considerable reciprocal introgression (Cousens 1963). In a few populations of Q. petraea the selected diagnostic characters were reasonably homogeneous; but even these populations were a little suspect since characters such as ‘veins to sinus of leaf’ occurred among them. In the absence of satisfactory definitions of either species it was not possible to estimate the extent of introgression. This paper describes subsequent studies to clarify the status of these two species by comparison with populations elsewhere. The methods of sampling and specimen collection have already been described (Cousens 1963). QUERCUS PETRAEA IN EIRE The forests of Ireland have been more completely destroyed than those of Scotland or England. By A.D. 1600 about one-eighth of the country still carried forest but during the next two centuries their destruction was greatly accelerated when iron-smelting with charcoal was developed on a large scale (McCracken 1963). Labour costs were relatively low and it became profitable to import iron ore from Wales or Spain and smelt it in or near the woods; when an area of forest was exhausted the smelters moved on to another centre. The only reports of attempts at forestry management come from Co. Wicklow which had exported oak billets to Wales prior to this period. In spite of the acknowledged fact that coppice-oak billets gave best results in smelting, the short-term approach, exploitation of existing woodland, seems to have prevailed. The situation in Ireland was thus very different from that in Scotland or England where there had been a long history of intensive coppice management and by implication a great deal of planting of oak to replace standards and old coppice stools. Tree planting on any scale is not recorded in Ireland till the sixteenth century and even then was mostly within the Pale (around Dublin). The Williamite wars discouraged planting through much of the seventeenth century and by the eighteenth century planting was mainly of exotic species, largely larch and pine (McEvoy 1943). All these plantations 161 Watsonia 6 (3), 1965. 162 J. E. COUSENS were included in the 2 per cent. of Ireland that carried woodland in 1840 (McCracken 1963). At the end of the Napoleonic wars (about 1820) charcoal smelting became unprofitable with the advent of coal smelting and a majority of the stands of oak to be found in Eire today are derived from woodland last felled about that time. Both Q. petraea and Q. robur occur in Eire today but varying opinions have been expressed about the status of Q. robur. According to McEvoy (1943) it is native and was the main oak of the central limestone area which was not overlain by bog. The only extensive remnants of oak woodland are all peripheral to the central area, mainly on siliceous rocks and of undoubted Q. petraea affinity. The populations sampled include well documented localities for Q. petraea in the Wicklow mountains, in the south-west near Glengariff and Killarney and in the west near Pontoon and Sligo (Praeger 1934). Fig. 1 shows the location of all the major samples. 2 5 DONEGAL SLIGO "11 : “10 9 GALWAY DUBLIN is 2 sf 3g WICKLOW ¢ LIMERICK 5 WATERFORD oKILLARNEY Guc 8 CORK 7 5 Fig. 1. EI[RE—location of populations sampled. ANALYSIS OF THE IRISH DATA The main analysis was confined to six diagnostic characters; contrary to opinions expressed on p. 283 of the previous paper (Cousens 1963) it will be argued later that all are probably under polygene control. Two of these characters, Petiole per cent. of leaf length and Peduncle length to first flower bract, are readily recorded as continuous variables and have been called the Primary Characters. The other four, Auricle type, Abaxial leaf pubescence, Peduncle diameter and Peduncle pubescence, are each classified as nominal Q. robur, indeterminate or nominal Q. petraea on the basis of multiple correlation within the Scottish material: these are the Secondary Characters. The presence of such a correlation in the Scottish material (which exhibited complete intergradation between the species) confirms the diagnostic value of these characters. Specimens in which all four characters were within the nominal Q. robur or Q. petraea range were called the Theoretical Species Types (or TSTs). By definition then the TSTs show less than the normal species variation for the four characters from which they were derived. It follows that specimens differing from the TSTs to a certain degree may still represent only the normal variability of the species. To test this the range of variation of the Primary Characters for TST and non-TST specimens may be compared. The ranges should be similar if both are part of a non- introgressed population. It has been shown for the Scottish collections that if specimens are classified according to the degree of their difference from the TSTs, then the plotted Watsonia 6 (3), 1965. PEDUNCULATE AND SESSILE OAKS 163 mean values of the Primary variables for each class fall on a curve linking the two TSTs. This curve has been called the Introgression Path and is shown in Fig. 2. The degrees of difference from the TST can be compiled as in Hybrid Index scoring but the resultant index indicates only the degree of heterogeneity. Determination of the level at which such an Heterogeneity Index is indicative of hybrid origin requires a knowledge of the normal 13 12 te ia 11 PETRAERS, } 10 Vag ene na N 9 \ Da 9 8 po ee 8 x“ 0 67 S 7 4 ™“ 3 ‘os es a6 X ~Aur3 6 5 4 2 10 20 Peduncle Length to first bract— mm Fig. 2. The INTROGRESSION PATH constructed from the mean peduncle and petiole values for the nine combination classes P(TST), P2, P3, P4, X, R4, R3, R2, R(TST) (Scottish data). species variability. The Eire data are first presented in this way (Table 1). Note that the percentage of TSTs is uniformly high while, with one exception, the Heterogeneity Indices are uniformly low. The ‘B’ collection from Croniebyrn in the Wicklow mountains is the exception and it happens to be the only sample from an area known to have been under long-standing coppice management. The proportion of truly intermediate specimens (i.e. classes P4, X and R4) is less than 2 per cent. Fig. 3 shows the distribution of the values for the Primary Characters. Note that the index is largely determined by values for TSTs and that there is no obvious trend in the direction of the Introgression Path. The greatest concentration of non-TST values does lie on the Introgression Path, however, indicating the possibility of slight introgression. DISCUSSION OF THE IRISH RESULTS It may be objected that the Irish samples, being selective for good Q. petraea, cannot be validly compared with those from Scotland. Table 2 lists the equivalent Scottish aff. Q. petraea samples, i.e. those with no specimens showing more than four degrees of difference from the Q. petraea TST. Note that the proportion of truly intermediate class specimens is higher at 7 per cent. and that the overall Heterogeneity Index is nearly twice that for the Eire collections. That these differences are likely to be very significant can be seen more clearly when the Primary variables are compared in Secondary Character combination classes. Fig. 4 is an enlargement of the Q. petraea end of the Introgression Path (Fig. 2) with the values for the Eire collections and the Scottish equivalents super- imposed on it. When individual values are plotted there is extremely wide scatter; means Watsonia 6 (3), 1965. 164 J. E. COUSENS TABLE 1. Range and frequency of secondary character combinations in the major Eire population samples. Degrees of difference from petraea TST Height No. Locality 0 pdiee: 3 4. 5.46 above P P2 P3 P4 xX R4 R3 Total HIt _ sea-level feet 1. Ballycoyl Re ik cu regal 3 1 21 0-52 500 2. Powerscourt, Deer Park .. 16 1 23 0-35 500 3. Glendalough, Lugduff Wood 13 a 20 0-35 950 4. Vale of Clara: Croniebyrn A bas ote al Date 15 0-40 450 Croniebyrn B sie - 4 Calera? aa | 12 1-08 450 5. Shillelagh, Culliton Estate .. 13 ie ES 16 0-20 400 6. Portlaw. Curraghmore Estate 23 cs ee ae | Ll 33 0-42 150 7. Glengariff Estate .. eae 3) Dea 1 33 0-39 50 8. Killarney, Tomies Wood .. 37 Siero 2 55 0-45 250 9. Connemara, Ballynahinch .. 12 4 1 7/ 0-35 100 10. Lough Conn, Pontoon eee ti) al 25 0-25 150 11. Lough Gill, Cullendra i eS 6 1 30 0-30 200 12. Lough Eske, Ardnamona Estate WF uF ie tot elle Thee 205) 29 0-59 200 Totals .. a " mit 22902 162. 27> 4 2 A a 329 0-42 Per cent. =. ae Le OS Sel 2 One * Specimens taken on the margin of the wood in each case; they do not form a continuous series with those in the wood and have been treated as extraneous to the populations sampled. Tt Heterogeneity Index. 1_TST 20 iS ia Concentration centre for Se) 2 1 1 non TST values italicised : TST Het eS 3 2 1 Is P2 Dames 7A ie 2 1 14 A513 5 PA 3 1 1 12 Sy "EST =< AV vip 22 aural ob gue Sennaeis Wana 35. l3 3 eae 2 ea Say é 8 eS 3 —— seclaty egreesieh a 12 FoqIST Pee =ic RAD S 6 joel on F2 4 5 N.B. Marginal values - t t q Petiole % up- Peduncle Length down 2 46 SE (© IRM A ite Son F SMA FSS AO) Peduncie Length to first bract-mm Fig. 3. Distribution of the combined Petiole % and Peduncle Length values for the Eire collections. Watsonia 6 (3), 1965. PEDUNCULATE AND SESSILE OAKS 165 TABLE 2. Range and frequency of secondary character combinations in Scottish samples within an equivalent range. O.S. ‘Serial Locality Degrees of difference from petraea TST Grid No. Oa on A Total HI Pip de es Es ox NJ 3. Elgin Town Wood Bens TMZ 9 0-78 NH 3. Locan a’ Chuillin 2 De aD. 6 1-67 NM 4-5 Loch Sunart .. Aer Se 2h 13 1-46 NN 5 Loch Arkaig 2510 1 23 0:56 NN 7 Spean Bridge Ow AS 18 28 NN 9 Glen Falloch .. | he eapBP 2 ir eel Ao | 6 1-50 NN 23 Glen Goil BEE NBS etars| 1 13 1t23))| NR 1 Claonig AriiB 1 8 0-88 NS 1 Menstrie Burn Jas 6 0-50 NS 7 Ross Wood ST Sim SO) eae bk 10 1-20 NT 8 Cheviot 7 Vi 6 0:33 NT 15 Aikieside Wood 9 62 11 0-18 NT 18 Jedforest remnant SES DOs Deeg Sil 0-71 NT 21 += Ettrick remnant SMe ssi] 11 1-00 ) NX 1 Burnfoot Wood See bv) 11 0-55 NX 2 Currarie Glen 11 46 17 0-35 NX 3. Cairnryan Gu 2 8 0-25 NX 4 Penkiln Burn Go 2 8 0-25 NX _ 7, 8, 11 Glentrool 20° 6 26 0-23 Totals .. a - S40 128, 69) 26, 4s A 241 0-74 Per cent. ape Ne Bae 53 29) Eh 76. 2 100 13 eae a See 12 ce a S(t) _ LE | aP-x(e) ~ } Neath 5 i S petraea P2€) ol STS) \ 11} Concentration . Centre . (Scottish Data) SS x e 10 6) (S197) VST 4 100 Fig. 6. Distribution of combined Petiole % and Peduncle Length values for collections of robur affinity in south-east and central England. Watsonia 6 (3), 1965. PEDUNCULATE AND SESSILE OAKS 169 well to the left of the main concentration centre for Q. robur TSTs. Few of the peripheral values are TSTs and as in the Scottish data all the very high peduncle values are from non-TST specimens. These abnormally high values have a marked effect on the means of the combination classes to which they belong. In Fig. 7 their trend, with and without these high values, is shown and related to the Introgression Path. The introgressive trend is condensed but this might be expected in a series of collections from populations selected as most likely to be good Q. robur. It seemed that it should be possible to find better Q. robur than this somewhere in Europe. 7 es . Go SS 1 ie eee eee LS PN 6 \ AR3 R4_ 37 R4. 39 A D5. aR | Introgression path—broken line No. % No, % ] Tse ENE PERO) 48 23 16 22 127 62 48 65 23.11 18 el Peduncle Length to first bract -mm All No 3 27 Aad) SOW May © Get ig 3 2 206 hm. 1 13,702), 28220 “9. 4 1 1 100 TST No 4 18320 eS qgenay 3 74 ° 5 24°27 90 “15.4 4 100 Fig. 9. Distribution of combined Petiole °% and Peduncle Length values for all Yugoslav robur collections. average (Table 5). Fig. 9 shows a very wide scatter of values but no obvious concentration of values along the robur-petraea introgression line. In Fig. 10 the distribution patterns for various components are shown as simplified scatter diagrams. The first four diagrams show that divergence from the Q. robur TST is associated with a reduction in peduncle length but there is no tendency for the petiole per cent. to rise at the same time. The variability cannot therefore be ascribed to intro- gression by a species like Q. petraea which has high petiole per cent. and low peduncle length values. In the TST diagram there is a suggestion of a lower concentration centre (peduncles 20-25 mm to first bract). A lower concentration centre appears more markedly in the diagram for specimens with pubescent peduncles; but it is now in the 15-20 mm range. Such a pattern could have arisen through introgression by a species with short petioles like Q. robur but with short pubescent peduncles. In the TST diagram there is also the suggestion of an upper concentration centre in the range 55-60 mm for peduncle length to first bract. This could indicate introgression by a species with short petioles like Q. robur but longer peduncles. There are, in fact, two species of oak in this region which conform to these specifications, namely, QO. conferta Kit. and Q. pedunculiflora K. Koch; both belong to Section Robur of the genus and presumptive hybrids of both with Q. robur have been found in Bulgaria (K. Kostov, personal communication, 1964). One explanation of the great variability of Q. robur in Yugoslavia and elsewhere may lie in its introgression by a number of related species. There is no evidence of currently active introgression. If it occurred, the most likely period would be during the enforced migrations caused by the Pleistocene glaciations of northern Europe. The theory is an attractive one in that it could be held to account for several characters which appear occasionally in Q. robur but more consistently in related species—for example, forms with (a) very long peduncles or secondary leaf lobing (Q. pedunculiflora), (b) very deep regular lobing (Q. conferta), (c) abaxial stellate pubescence (most members of Section Robur). For the present the Yugoslavian collections merely indicate that Q. robur is, or has become, an exceedingly variable species. Watsonia 6 (3), 1965. 172 J. E. COUSENS Petiole % Totals 1 2 1 1 16 446212 48 3 8 R —Robur TST N&O CO ee pe N NaN hm Ue — N B&B BD COO © — R2—specimens showing one difference eel WN oo) eS Cone eS R3—specimens showing two differences R4—specimens showing three differences N&O CO N — — — — — N All specimens with pubescent peduncles N BO CO NO U2 N oN Fe 0°10 20 §30" 4050 60" 70° "SO +au Peduncle to first bract—mm. Fic. 10. Simplified scatter-diagrams for various components of the complete Yugoslav data. GENERAL DISCUSSION AND CONCLUSIONS The pattern of distribution of petiole-peduncle values is very similar in outline in the Yugoslav and English Q. robur collections (cf. Figs. 6 and 9); equally the pattern for the TSTs only in Yugoslavia, England and Scotland is very similar (Fig. 11). All show a modal value for Petiole per cent. of between 4 and 5 and all are bi-modal for peduncle length, with peaks at 20-30 mm and around 40 mm; the Yugoslavian and English diagrams also have suggestions of a third peak in the upper fifties. In addition each collection is characterized by a few non-TST specimens with abnormally long peduncles. If Yugoslavian populations were transferred to south-east England and subjected to a little secondary introgression from Q. petraea, they would probably be indistinguishable from present populations in that part of England. The scale of introgression by Q. petraea would be roughly that Watsonia 6 (3), 1965. PEDUNCULATE AND SESSILE OAKS 173 Petiole % Totals Yugoslavia 8 | 2 6 1 l 16 4 ie Gah al (9) 72 he? 48 2 1 3 8 74 England N # A CO — N=_ = Scotland NR DD C pot eet 0 10 20 30 40 50 60 70 80 — Peduncle to first bract—millimetres. Fig. 11. Scatter diagrams for the robur TST in various collections. suggested by Jones (1959, p. 198) in southern England, with a slight increase northwards into Scotland. It seems reasonable to conclude that the Q. robur that arrived in post- glacial Britain was very like present-day Q. robur in Yugoslavia; that it was even then a very variable species; and that it must be accepted now as a species very difficult to define. In the previous paper the prevalence of Q. petraea forms of pubescence on the leaves and peduncles of Scottish hybrid specimens was tentatively ascribed to partial dominance by Q. petraea. It is now apparent that Q. robur already has some genes for such pubescence and there is no reason to believe that these characters are other than under polygene control. It should be emphasized that none of these major collections were planned as repre- sentative samples of oakwoods in the regions concerned. Generally the aim was either to determine the range of variation in what were thought to be the most ‘natural’ woodlands (Irish and earlier Scottish collections) or to sample woodland reputed to consist of good Q. robur or Q. petraea (Eire, Yugoslavia, southern England and the later Scottish collec- tions). Infertile specimens, which were ignored in the general analyses, are more likely to be hybrids than fertile specimens in areas where current introgression has been demon- strated (Cousens 1963, Table 6). Mixed populations have also been ignored. Overall there has been a bias against demonstrating the full extent of any introgression. There seems to be a cline of increasing introgression northwards in England and Scotland. Planting of Q. robur in Scotland may explain the difference between Irish and Scottish Q. petraea populations but planting has been equally prevalent in England and another explanation is required for the increasing introgression northwards. Fig. 12 con- trasts the extremes as exemplified by samples from Kent and the Central Highlands of Scotland; both have a long history of intensive forest management. In Ham Street Nature Reserve Q. robur and Q. petraea are both reasonably good as now redefined (Table 6), while in the Loch Katrine sample a majority of the forms were intermediate and there were no TSTs. It has been suggested, but not demonstrated, that cross-pollination should occur more frequently in Scotland because the growing season is Watsonia 6 (3), 1965. 174 J. E. COUSENS Petiole % e © @© Specimens differing from either TST in 15 ® two or more characters — presumptive hybrids 6 @ @ aff. petraea O aff. robur --® -@cc ‘\ PETRAEA 7) @ 10 Bb. ,@% @ i 2 ® O O O O © oO 00 One OO 2 \ 5 (OX) IEG 4\ O Conn | ® HAM STREET Py ee. Oo NATURE RESERVE Oo KENT 6) a SS AO (OS \ PETRAEA/ 10 7@ ¥ (0) © © (0) i 2 © © 2 ©@ to @ L-tc’\ Peceine LOCH KATRINE Vas NORTH SHORE Woops CENTRAL HIGHLANDS SCOTLAND 10 20 30 40 50 60 70 Peduncle Length to first bract-mm Fig. 12. Scatter Diagrams for samples from two woods with a long history of intensive forestry management, one from south-east England, the other from Scotland. shorter and there is greater overlap in the flowering times of Q. robur and Q. petraea. Such an hypothesis is an essential part of any explanation of the increased introgression northwards. By itself it cannot explain the striking increase in the introgression of Q. petraea, considering the low level of interfertility found by Pyatnitski and Dengler under the optimal conditions of artificial cross-pollination. However, their crossing experiments are not necessarily conclusive; a few selected individuals of each species only were con- cerned and in one year the robur 2 x petraea 3d cross reached a level of 15 per cent. interfertility. If O. robur is such a variable species, the degree of interfertility with Q. petraea may be very variable too. Crossing trials with a much wider range of parental genotypes would clear up this point. Some indirect evidence might be obtained by comparing the degree of introgression of Q. petraea in Devon and Wales with that in Scotland. Assuming, then, that rather more first generation hybrids are likely to appear in the north wherever Q. robur and Q. petraea occur together, it is obvious that the extensive planting of Q. robur in Q. petraea areas (bringing the species together) and intensive coppice management (shortening the generation time) could have greatly accelerated the introgression of Q. petraea by Q. robur. Q. petraea was seldom used for planting and this factor alone is sufficient to account for the great difference in degree of introgression of Watsonia 6 (3), 1965. = SSS ee ee ee Eee PEDUNCULATE AND SESSILE OAKS 175 Q. robur and Q. petraea in Scotland. The very considerable introgression of Q. robur by Q. petraea in Scotland previously postulated (Cousens 1963) must now be considered illusory; most of the heterogeneity in Q. robur forms was probably there before the species arrived in Scotland. The validity of most of these conclusions could be checked by crossing experiments but as the generation time would be of the order of 10 to 20 years at least, it would be a very long time before the genetic basis of variability could be established. Table 6 shows the range of variation observed in the Yugoslavian Q. robur and Irish QO. petraea populations sampled. Ranges overlap for all the selected diagnostic characters but the species are separate entities in terms of petiole per cent. and peduncle length combined (see Fig. 9). Using these definitions it should be possible to diagnose good populations of either species. Individual trees, however, can only be identified as good Q. robur or Q. petraea when all four secondary characters are typical and petiole and peduncle values lie within the ranges shown in Fig. 9; this means, for example, that only about 25 per cent. of Scotland’s oaks can be positively identified. ACKNOWLEDGMENTS I am indebted to all those who assisted me in planning or making the collections; to many members of the Forestry Division of the Department of Lands in the Republic of TABLE 6. The observed range of variation of selected diagnostic characters for Q. robur (Yugoslavian collections) and Q. petraea (Irish collections). Diagnostic character Peduncle length to first bract— mm. Total Peduncle Length Petiole °% of leaf length Auricle Type* Stellate Pubescence abaxially on lamina (small, prostrate) or along midrib (large, erect) Peduncle Pubescence below first bract—in vicinity of first bract if peduncle very short (recorded from July onwards) Peduncle Diameter at narrowest point below first bract—if elliptical, mean of two diameters Q. robur (6)—15—22—38—60—(88) 14—35—55—80—(123) (1)—2—_4—6— (9) Medium—Strong Both types absent 8%—one or both types repre- sented by a few stellate hairs 1°%—small but obvious amount of both types Glabrous or with a few scattered hairs (equally common) Forms with general pubescence may reach 30% in some popu- lations I mm or less 11%—c. 1:5 mm Less than 2% reaching 2 mm Q. petraea 0—/—4—(20) 1—/0—26—(63) (53)—9—/2—15—(20) Nil—Weak—(Medium) Both types well developed 8 %—one or both types deficient 1 °—both absent At least patches of dense pubes- cence 4%—only a trace of pubescence Less than 1% glabrous 2 mm or more 20%—c. 1-5 mm Less than 7% as little as 1 mm Notes. Modal values and commonest forms in italics Parentheses indicate fairly isolated extreme values. Peduncle characters from longest peduncle on specimen. Leaf characters from two mid-shoot leaves on a representative sun-shoot taken from the lower crown on the south or south-east side of the tree. * For a description of auricle types see Cousens (1962). Watsonia 6 (3), 1965. 176 J. E. COUSENS Ireland, of the Nature Conservancy in south-east England and of the Forestry Faculties of the Universities of Zagreb and Belgrade in Yugoslavia. I am particularly grateful to Dr. E. W. Jones for his interest, encouragement and advice at critical stages in this work; and to Professor J. N. Black and Dr. J. W. Gregor, who kindly read and criticized the original manuscript, for their many valuable suggestions. The work was carried out with the aid of grants from the Forestry Commission. REFERENCES Cousens, J. A. (1962). Notes on the status of the sessile and pedunculate oaks in Scotland and their iden- tification. Scot. For. 16, (3). Cousens, J. E. (1963). Variation of some diagnostic characters of the sessile and pedunculate oaks and their hybrids in Scotland. Watsonia 5, 273-86. DEKANIC, I. (1962). Utjecaj podzemne vode na pridolazak i uspijevanje sumskogdrveca u posavskim sumama, kod Lipobljana. Glasnik za sumske pokuse 15, 5—102. GLAVAC, V. (1961). O vlaznom tipu hrasta luznjaka 1 obicnog graba. Sum. List 9-10, 342-7 Zagreb. JONES, E. W. (1959). Biological Flora of the British Isles. Quercus L. J. Ecol. 47. McCrRACKEN, E. (1963). Irish woodlands, 1600 to 1800. Quart. J. For. 57 (2). McEvoy, T. (1943). Some Irish native woodlands: an ecological study. Thesis M.Agr.Sc., National University of Ireland, Dublin. PRAEGER, R. LL. (1934). The Botanist in Ireland. Dublin. Watsonia 6 (3), 1965. eA ee Ee THE DISTINGUISHING CHARACTERS AND GEOGRAPHICAL DISTRIBUTIONS OF ULEX MINOR AND ULEX GALLII By M. C. F. PROocToR Department of Botany, University of Exeter ABSTRACT The lengths of the petals and calyx provide the most satisfactory characters by which Ulex minor and U. gallii may be separated. Standard-length rarely exceeds 12 mm. in U. minor, and is rarely less than 12 mm. in U. gallii; population means for calyx-length are less than 9:5 mm. in minor and greater than 9-5 mm. in gallii. Calyx-length remains virtually unchanged from flowering time to the dehiscence of the fruits, and is the most generally useful diagnostic character. All other characters examined show greater overlap between the species. Both species are genetically variable and very plastic in vegetative characters, which are con- sequently of little use for identification. The two species are almost perfectly vicarious in their distribution within the British Isles; their main areas are separated by a line running roughly through Weymouth, Salisbury, Marlborough, Oxford and Sheffield—U. gallii to the west of the line, U. minor to the east. Ulex minor and U. gallii are not generally regarded as critical species. Nevertheless, they are widely misunderstood. Most Floras agree that U. gallii is a larger plant than U. minor, and quote a number of supposedly qualitative characters separating the two species. But there is little agreement beyond this, and different floras emphasize different diagnostic characters and give widely differing figures for such characters as plant size, spine length and seed number. The problem arises because most of the distinguishing characters are quantitative, while the species are highly plastic. There is rarely serious difficulty in distinguishing either species from U. europaeus. That species has very much hairier calyces, so that the young buds have a conspicuous brown furry covering, bracteoles considerably wider than the pedicel, and primary spines which often bear branches (and flowers) a considerable distance from their bases. U. europaeus usually starts to flower in early autumn, bearing flowers sporadically throughout the winter whenever the weather is mild, and has its main flowering season from March to May. The other two species, by contrast, flower in late summer and early autumn, and are past their best by the end of October, though the odd flower may straggle on until New Year. Ecologically, too, U. europaeus is quite distinct. It is a plant of poor pastures, road- sides, scrub, wood margins, sea cliffs, etc., on poor soils, often where there has been some disturbance of the habitat. It is commonest on more or less acid soils, though it occurs even on thin rendzina soils over chalk and limestone. It is seldom found in quantity, or growing with any degree of luxuriance, in the extensive areas of undisturbed heath which are the most characteristic habitat of U. minor and U. gallii. My own difficulties in placing many plants of U. minor and U. gallii convincingly in one species or the other prompted the investigation of which the results are set out in this paper. In the late summer and early autumn of 1958 I collected samples of the two species from a number of localities scattered over the south of England: usually single flowering shoots from ten individual plants in each population. With some additional samples collected since, these have provided the data on flower dimensions given below. The samples are small, and I emphasize that in collecting them I was on the look-out for inter- mediate and ‘difficult’ plants and populations. In the event, they proved more than adequate to show that the two species are quite clearly distinct. This provided a basis on which to evaluate the other characters of the species, both in the field and in the herbarium, and to re-examine their geographical distribution in the British Isles. 7/7/ Watsonia 6 (3), 1965, wo 45 eee O oe gas 42% a5 els eset gg on Oa ‘eal g : QE : 3 ; aa : en , [S) Atel | “= O Sas : a Toe fe” oe } oo : : Sey x Od ; 7) o : Be 8 O) BR es 5 : ©) FG Ryu : ov , a SS) tel) te | Mees : = rm 2 ce = ° | Ob 28am | | O iy ae a | | 5 @ SEES ° ~- § eh= Fs 3 aio as] (cD) Lee 7 a \ y Ee ; fo) 9 Bs | 3s 8 : i 7 = 5 3 aU : G is aS ij + Or, oe ra oN , QL : , . Y a6 Os a2 | : : =: Oo) wo | t os = S | = ograms i ns of the hist ULEX MINOR AND ULEX GALLII 179 any of these flower parts show a clearly bimodal distribution. The two peaks for standard- length are almost completely separated; intermediate values occur with greater or less frequency in the other three cases. However, if population means are considered rather than measurements of individual flowers, the populations fall quite sharply into two groups, whichever flower part is measured (e.g. calyx, Fig. 2). Mean standard-lengths for the populations are plotted on a map of the south of England in Fig. 7. Populations with a mean standard-length less than 12 mm are confined to an arza from Dorset eastwards; to minor gallii Fig. 2. Frequency histograms of mean calyx length in populations of Ulex minor and U. gallii; same material as Fig. 1 (usually 10 measurements per population). the west all populations have a standard-length of 12 mm. or more. The large-flowered western plant is U. gallii, while the small-flowered eastern plant is U. minor (Fig. 3). U. gallii usually has the wings rather strongly curved, and longer than the keel, though occasional plants are found with them shorter. In the collected samples the mean length of the wings is nearly 1 mm. longer than the keel. U. minor is more variable in these characters. In some populations the wings are nearly straight, and regularly shorter than the keel. On the other hand, there are many populations of undoubted minor which have the wings strongly curved and regularly longer than the keel. Taking the collected samples of minor as a whole, the mean lengths of wings and keel scarcely differ. (b) Flower colour The population samples collected in 1958 were scored for flower colour by matching the flowers with the standard tints in the R.H.S. Horticultural Colour Chart (British Colour Council 1938). Although there is considerable overlap the two species are fairly sharply separated (Fig. 4), with aureolin and lemon yellow the most frequent colours for U. minor, and buttercup yellow and indian yellow the most frequent colours for U. gallii. (c) Bracteole size Rothmaler (1951) considered the size of the small bracteoles just below the calyx to be the best critical character separating the two species. In the material I have examined the bracteoles are usually about 0-6-0-8 mm. long x0-4-0:6 mm. wide in minor, and 0-5-0 -8 mm long x0-6-0-8 mm. wide in gallii. They seem to provide a fairly satisfactory separation between the two species, but not easy characters to work with, as they are small and often thickly covered with hairs which make accurate measurement difficult. Watsonia 6 (3), 1965. M. C. F. PROCTOR 180 ‘UMBIP USUIIOSdS SY} ULT} SSUIM JOB8UO] A[QUIOPISUOD SALY UdIJO JOUIU ‘7 JO SIOMOTA “¢-7 X “QUBII) 17j08 "Q pue (3Jay) soul xajf JO SIOMOT *€ “SIA Watsonia 6 (3), 1965. ULEX MINOR AND ULEX GALLII 181 (d) Fruiting characters Pod measurements and seed counts from a number of localities are summarized in Fig. 6. Calyx length still remains a reasonably satisfactory distinguishing character, with the population means for the two species falling into two quite clear groups, though indi- vidual calyx measurements show considerable overlap. Pod length and breadth give a less satisfactory separation, and seed and ovule number none at all. There are indications that gallii averages slightly more seeds per pod than minor, but seed number is very variable in both species. It is interesting that ovule number shows even less difference between the species, and no less variation. It is often said that the pods of U. gallii dehisce in spring while those of U. minor persist for almost a year. The seeds ripen at much the same time in both species (May), and the time of dehiscence probably varies a good deal with weather conditions in different years. Undoubtedly the pods of U. gallii do tend to dehisce somewhat earlier, probably because of their rather larger and firmer valves, but there is not a clear-cut difference between the species in this respect. 5 BUTTERCUP YELLOW 6 INDIAN YELLOW 7 SAFFRON YELLOW 2 CANARY YELLOW 4 LEMON YELLOW 2 —_ O WwW ag =) < a) minor gallii Fig. 4. Frequency histograms of flower colour, scored on RHS chart, of Ulex minor and U. gallii. In most cases the flower colours were close to the saturated hues; for convenience of representation in the few instances in which the flowers departed markedly from saturation 6/1 has been regarded as 6, and so on. 1958 material. (e) Size of vegetative parts Both species are extremely variable in size. In areas of poor short heath U. minor may grow as procumbent mats barely 5 cm. high; in roadside scrub and around wood margins in the New Forest and the Isle of Wight I have seen it as high as 1-5 m. U. gallii is a larger plant. Under poor conditions it may be only 10 cm. high, while bushes on neighbouring roadsides grow to 1-1-5 m.; old plants in scrub or open woodland may reach 2 m. The variation is partly genetic and partly due to plasticity. Plants of U. gallii from an exposed hillside at Mynytho in the Lleyn peninsula grown for five years in Exeter reached only Watsonia 6 (3), 1965. 182 M. C. F. PROCTOR 25 cm., while plants from a heath near Woodbury, Devon, reached 100 cm.; U. minor from near Wareham formed a sprawling mat no more than 10 cm. from the ground while material of the same species from the New Forest reached 95 cm. The primary spines are also very variable in length, with a large overlap between the two species; Fig. 5 shows the distribution of spine length in the material used for the flower samples. The two species show widely differing means, but a histogram of the combined data is not convincingly bimodal. Curved primary spines are often given as a character for U. gallii. It is true that U. gallii usually has rather strongly curved spines, especially in well-grown material, while the most poorly-grown (and to most people most typical) minor usually has straight spines. But gallii often has straight spines, and well-grown minor often has them strongly curved. The character is least helpful in those plants which are most likely to be difficult to deter- mine. Much the same applies, mutatis mutandis, to the rigidity of the spines of gallii, which Tutin (1962) uses as a key character. (f) Other characters A number of other qualitative characters have been suggested. Red veining of the standard has been said to be characteristic of gallii. | have never observed this in any of the populations of gallii I have examined in the field. The veins of the petals do tend to go brown in badly dried specimens, and this affects the larger flowers of gallii more than the small flowers of minor. I assume that the character was drawn from herbarium material in which some specimens were affected in this way. The calyx teeth are sometimes described as ‘connivent’ in minor and ‘spreading’ in gallii. I am not clear what is meant by this. The calyx is deeply divided into an upper and a lower lobe. The flower as a whole opens more widely in gallii than in minor, especially after pollination. The individual calyx teeth at the ends of the lobes seem to offer no taxonomically useful characters. Fig. 5. Frequency histograms of the length of the longest primary spine on the flowering shoot of Ulex minor and U. gallii; 1958 material. Conventions as in Fig. 1. DISCUSSION OF THE CHARACTERS The characters considered above are all correlated, as might be expected if we are dealing with two distinct species. The best diagnostic characters are those that divide the populations most clearly into two groups, in this case the measurements of the flower parts. Standard-length is the easiest and most effective; measurement of the standard of a single flower will almost always place a plant correctly. Unfortunately its use is confined to the relatively short flowering season. Calyx-length shows more overlap, and it may be necessary to measure a number of calyces to arrive at a certain identification, but it has the great advantage of being a character which is available at almost any time of year. Mean calyx- length of 9 mm. or less indicates minor; a length of 10 mm. or more indicates gallii. There is virtually no change in calyx-length from flowering time to the dehiscence of the pods nine or ten months later, and calyces often persist even longer than that. Watsonia 6 (3), 1965. 183 ‘ONIVA UBIUT OY} JO OPIS JOYyj}IO UO UONLIADP prepUL}s OY} 991M} 0} [eNbos souK\sIp e Soul] UIY) pue “UO oY} JO IOIIO pIepuL}S OY} 901M} 0} [eNbo soURIsIP B UMBIP oI SIeq YOTY) :suBOLU uonendod oyeorput soposr “JOWUUNS A[IeO UI 117/08 "7 pure soul Xaj-Q JO SJO\OeIeYO SUTjINIJ pue xXATVD “9 “BIA ULEX MINOR AND ULEX GALLII eee ee ee pe ee a pe ea See = ——_ b+ ——_——— —-O-——_ +———«Q2———+ -——=Qxe——— rr K——— +—_— Qt ee — $$$ -—O— ro -o— -——_==——_ -——_=O=—— —-O0=— -——_=O=—— p= Js OO -———Oe— Oe —__—_ HOH |} at $s $$ ——0=— + O==— tO SS SS EE (ee ee —EE— (Ee a |e) 9 S 7 € Z t WJLUG / € WwWOl 8 9 wuz! Ol 8 9 YSEWNN 3JINAO Y3SEWNN O33S HLGV3Yd GOd HLONAT GOd HLON31 XAIWS (Buide4-S) 'UMOG P4OJEEM (BulDe}-N) G90 XS ‘UOAaG'UMOG P4JOJEEM LLOL XS ‘UOAIG'IO)Ae}Y JeaU LLEB XS “UOAIG ‘YEAH UOYYBiuy YBiajpny9 (W2 00|'2 ‘aPIspDO4) ‘UOPJeH }eAI9 £868 XS UOAeq‘uopjey yeas9 G870 AS “UOAeq‘UOWLWOD UO0jJIIg QIEL 1S “assawos‘uojBuNjam Jeau €G0G ‘21S J8S49WOS ‘ApPlig Jeau G86 AS 38S40q YOOW PUe}JeH (q)'eu0}speojg Jeou (©) 9610 ZS 19S40q ‘aUO\speojg Jeau Wav9 Xx31N 98SZL AS J9S4JOG'YUIESH NEMWJIeM G896 AS }28S40g JOOW pue)}JeY 686E ZS 1YBIM JO ajs|'UOWLUOD poomBulN 06872 ZS IUBIM JO 215|'}S9404 JSINYAIEY IBIS ZS ‘J4BIM JO als} ‘UMOg Ye2I1g 961Z ZS ‘SJUBH ‘UOWLWO) UO}MaYD 86IZ ZS “SJURH'YJEAH Ul} B60E ZS ‘S}UBH/SEWIOH Aa}41US OOSE NS ‘SIUEH'4IEMSSOD Apey Z1ZZ NS “SJUCH‘BU0}S SNjny JeaU YONIW X31N Watsonia 6 (3), 1965. M. C. F. PROCTOR 184 ‘| ‘SI St [elioyeul oes JOUIM */) OTE SITY} ULY) SSO] SURO YIM soy} ‘17/05 "7 91 SIOW IO “UU QZ] JO UROL Y}IM suonendod ‘purlsuq UIOYINOS UI 17/03 "| Pue AOUIU Xaj/) JO suonrindod ur y)8ue|-piepur}s uvow Jo UOTNINSIG */ “S14 ®) G0 6 OC ea 431S3HIYOO ag D O NOLH9INa = olns) NOLNAWL ZA UALSIHONIM CS O a ° SNIMOM age : ONIGVSY A HONOYOE YW 08-04 0O 06-08 O oo-06 O HLNOWAId O'll-0'01 O ozi-ou OC O€l-02I O O'"l-O'€l 0'SI- O° 0'91-0'SI 041-091 Watsonia 6 (3), 1965. ULEX MINOR AND ULEX GALLII 185 GEOGRAPHICAL DISTRIBUTION There has never been any question that U. gallii is a western species by contrast with U. minor which has its headquarters in south-east England. The striking fact that emerged in the course of the present work was the remarkable clarity of this pattern; in Britain the two species are almost perfectly vicarious. In the course of my field work this became very obvious in Dorset, where the areas of the two species are in contact. U. minor is abundant in the heaths on Eocene sands and gravels in the Poole basin, where U. gallii is very local and sporadic, and—apart from a substantial outlying area north-west of Bournemouth— almost always on roadsides. But farther west, on clay-with-flints over the chalk and on Cretaceous and Jurassic sands, U. minor is apparently completely absent, and every population I examined was U. gallii. In Hampshire and Wiltshire too, apart from one or two isolated localities west of Salisbury, U. minor seems to extend westwards only as far as the western limit of the Tertiary beds. Field records and herbarium material show that the main areas of the two species are separated by a line running roughly through Wey- mouth, Salisbury, Marlborough, Oxford and Sheffield (Fig. 8). é é é é é é é 2 é Ou é r é 300 400 500 Fig. 8. 10 km. distribution of Ulex minor and U. gallii in the area of contact of the two species in southern England. All marginal and outlying records for both species are based on herbarium specimens or field records which I have been able to verify personally. Records of the two species within their main areas have in general been accepted without question. Most of the apparent anomalies to this pattern that I encountered turned out to be due to errors in identification. Thus all the specimens I have seen labelled U. minor from Wales and Cornwall have been dwarf forms of U. gallii, usually from exposed places near the sea. Large bushes of U. minor have often been mistakenly named U. gallii; with good gallii for comparison there is seldom any difficulty in naming them correctly. However, both species have outliers within the area of the other. U. gallii has a sub- stantial detached area in a narrow strip along the coast of Norfolk and Suffolk; farther inland U. minor is apparently the predominant species in the heaths. U. gallii appears to have a further small outlier in Kent. Apart from these outliers, small colonies of U. gallii occur in a good many other localities, usually along roadsides and in other disturbed Watsonia 6 (3), 1965. 186 M, Cy Es -pPROCTOR habitats where it is probably a relatively recent arrival. U. minor has several stations near Carlisle. At the two I have visited, Sowerby Wood (NY c.365523) and Hosket Hill near Kirkandrews-upon-Eden (NY c.347575), U. minor is growing in relic areas of sandy heath very reminiscent of south-east England, and U. gallii is not present in the immediate vicinity. Good specimens of U. minor also exist from Slieve Donard, Co. Down (S. A. Stewart, 22 July 1879, DBN) and from between Holy Loch and L. Eck, Argyll (Wm. Neathead (?), 1 September 1884, BM). I have not visited either locality and both need confirmation; a note on the Argyll sheet suggests that the plant—which was apparently present in quantity—may possibly have been introduced from the south of England with rhododendrons and rare conifers.* The clear geographical separation of U. minor and U. gallii in England parallels strikingly the situation demonstrated in north-west France by Des Abbayes & Corillion (1949) and Corillion (1950a, 1959). There, U. gallii is abundant to the exclusion of U. minor in Brittany west of a line roughly from Dinard to Vannes, and in a small area at the tip of the Cotentin peninsula, with only small outliers farther east. U. minor is widespread outside this area in western and central France. In the Basque country and in north-west Spain there are indications of a repetition of a similar pattern. Perhaps the most likely explanation of this remarkable distribution pattern is that the two species are so closely similar in ecological requirements that they are competing for the same ecological niche, and cannot coexist (except transitorily) in the same vegetation. With this in mind, it would be worth making systematic observations over a period of years in localities where the two species do occur together. At the moment it is hard to say whether the pattern is determined by present conditions of soil or climate, or whether it reflects past conditions, and has been ‘frozen’ by the fragmentation of heathland. HYBRIDIZATION Corillion (1950b) reported hybrids between U. minor and U. gallii from heaths in the zone of contact of the two species in Brittanyy. As far as I know, this hybrid has never been reported from England (see Millener, 1952), and hybridization does not appear to enter as a complication into the problem of distinguishing the two species. It is possible that some hybridization may take place where the two species meet; I have not looked critically for evidence of it. Hybrid individuals might be expected to show intermediate characters in the size of flowers and other parts—perhaps with some modification due to hybrid vigour. As U. minor has 2n = 32 and U. gallii 2n = 80 (Castro 1941, 1943) the most satisfactory evidence of hybridity would be provided by chromosome number, and perhaps sterility. ACKNOWLEDGMENTS I would like to thank all the correspondents, too numerous to mention individually, who so readily answered my questions about Ulex specimens and localities. REFERENCES BRITISH CoLouR CounciL (1938), in collaboration with the Royal Horticultural Society. Horticultural Colour Chart. London. Castro, D. DE (1941). Algunas contagens de cromosomas no genero Ulex L. Agron. Lusit. 3, 103-13. Castro, D. DE (1943). Contribugaéo para ho conhecimento cariologico dos generos Ulex L., Stauracanthus Link e Nepa Webb. Agron. Lusit. 5, 243-49. CoRILLION, R. (1950a). Contribution a l’étude de la répartition de 1’Ulex gallii Planch., sur le littoral du nord de la Bretagne. Bull. Soc. Sci. Bretagne, 24, 97-9. * The occurrence of the species in the following vice-counties may be taken as satisfactorily established : U. minor: 7—17, 20—28, 32, 70. U. gallii: 1-9, 15-17, 25, 27, 34-53, 55—60, 63-65, 69-73, H 1-13, 15, 16, 19-21, 27, 32, 37, 38. Confirmation of either identification or status in other vice-counties is desirable. + See also Lambinon, J. (1962). Note sur les Vlex du Massif Armoricain. Lejeunia NS 9, 64-66. Watsonia 6 (3), 1965. Sa | ULEX MINOR AND ULEX GALLII 187 CoRILLION, R. (1950b). Phanérogames intéressantes pour la Bretagne (IV). Bull. Soc. Sci. Bretagne, 25, 131-40. CORILLION, R. (1959). Nouvelles précisions sur la peparHeon de |’ Ulex gallii Planch. en Bretagne. Bull. Soc. Sci. Bretagne, 34, 233-6. Des ABBAYES, H. & CORILLION, R. (1949). Sur la répartition de I’ Ulex gallii Planch. et d’Ulex nanus Sm. dans le Massif Armoricain. C.R. Som. Séances Soc. Biogéog. 229, 86-9. MILLENER, L. H. (1952). Experimental studies on the growth forms of the British species of Ulex L. Ph.D. thesis, University of Cambridge. ROTHMALER, W. (1951). Revision der Genisteen. I. Monografien der Gattungen um Ulex. Bot. Jb. 72, 69-116. TuTIn, T. G. (1962), in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles. Ed. 2, Cambridge. Watsonia 6 (3), 1965. DIPLOIDS IN THE GENUS COCHLEARIA By J. J. B. GILL Department of Genetics, University of Liverpool ABSTRACT Three new counts of diploid Cochleariae are reported. (C. aestuaria (Lloyd) Heywood 2n = 12, C. alpina Wats. 2n = 12 and C. arctica Schlecht 2n = 14.) It is suggested that the previous count for C. alpina of 2n = 28 is wrong and was probably made on a mis-identified specimen of C. officinalis L. containing accessory chromosomes. INTRODUCTION The chromosome counts in the genus Cochlearia have been reviewed recently by two authors (Saunte, 1955, and Léve & Léve, 1961). Both these reviews show the genus to consist of polyploid series based on the numbers x = 6 and x = 7. The x = 6 series contains the diploid (2n = 12) C. pyrenaica DC., the tetraploid (2n = 24) C. officinalis L., the hexaploids (2n = 26) C. micacea Marshall and C. polonica Borb. and the octoploid (2n = 48) C. anglica L. Decaploids (2n == 60) ascribed to C. anglica have also been reported (Lévkvist, 1963 a & b). In the x = 7 series there have been reported the diploids (2n = 14) C. groenlandica L., C. oblongifolia DC., and C. scotica Druce, the tetraploid (2n == 28) C. alpina Wats. and the hexaploids (2m = 42) C. danica L. and C. tatrae Borb. (For all references except those which are individually noted see Saunte (1955) and Léve & Léve (1961).) In the course of a_ cytogenetical investigation of the genus it has been shown that C. aestuaria (Lloyd) Heywood and C. alpina (Wats.) are diploids of the x = 6 series and C. arctica Schlecht is a diploid of the x = 7 series. METHOD The plants were grown in pots in a cold frame and mitotic chromosome preparations were obtained from root tips pre-treated for four hours in a saturated solution of para- dichloro-benzene, fixed in | : 3 acetic alcohol, stained by Feulgen’s method and squashed in aceto-carmine. Meiotic preparations were made from pollen mother cells fixed in Carnoy’s fixative and stained and squashed in aceto-carmine. In all cases the preparations were made permanent by the dry-ice method (Conger & Fairchild, 1953) but were mounted in Canada balsam instead of euparal. RESULTS Seed of C. aestuaria was obtained from Galicia in Northern Spain. Only one plant developed from this seed and squashes of root tips from this plant showed it to be a diploid of the x = 6 series (2n = 12). The somatic chromosomes (Plate 12, A) vary after pre- treatment from 1—2y in length and all have more or less median centromeres. The longest chromosome forms about 10 per cent. and the shortest about 6-5 per cent. of the total length of the karyotype. The somatic count was verified from PMC squashes which showed the presence of 6 bivalents (Plate 12, B). Two samples of English C. alpina were obtained, one from Malham Tarn Field Centre and the other from Helvellyn. Both samples gave chromosome counts of 2n = 12 but the populations also contained plants with accessory chromosomes. Root-tip squashes of 17 plants from Helvellyn showed that 16 of them were normal with 2m = 12 (Plate 12, C) but one was 2n = 12-+1 accessory with the extra chromosome distinguishable from the other members of the karyotype by its smaller size (Plate 12, D). The normal somatic chromosomes of this material were very similar to those of C. aestuaria in that they varied from 1I—2u in length and were more or less metacentric. The longest formed about 11 per cent. and the smallest about 6-5 per cent. of the total length of the 188 Watsonia 6 (3), 1965. ReAte 2 EXPLANATION OF PLATE A. Root-tip cell of C. aestuaria showing 2n = 12. B. PMC of C. aestuaria showing 6 bivalents. C. Root-tip cells of C. alpina from Helvellyn showing 2n = 12. D. Root-tip cell of C. alpina from Helvellyn showing 2n = 12+1 accessory. E. PMC of C. alpina from Helvellyn showing 6 bivalents. G. Root-tip cell of C. alpina from Malham showing 2” = 12+-1 accessory. H. Root-tip cell of C. alpina from Malham showing 2” = 12+ 2 accessories. J. PMC of C. alpina from Malham showing 6 normal bivalents+paired accessory chromosomes, the accessories being indistinguishable from the normal bivalents. K. PMC of C. alpina from Malham showing 6 bivalents+ 3 non-associated accessories. L. PMC of C. alpina from Malham showing 6 normal bivalents+1 pair associated accessories-+ 1 single accessory. M. PMC of C. alpina from Malham showing 6 normal bivalents (one slightly out of focus)+1 pair accessories+ 2 non-associated accessories. N. Root-tip cell of C. pyrenaica showing 2n = 12. O. PMC of C. pyrenaica showing 6 bivalents. P. Root-tip cell of C. arctica showing 2n = 14. R. PMC of C. arctica showing 7 bivalents. All x 1,000. . [ facing page 188 DIPLOIDS IN THE GENUS COCAHLEARIA 189 karyotype. No accessory chromosomes were visible in any of the PMCs of this material examined, all showing six bivalents only (Plate 12, E). Accessory chromosomes are extremely common in the Malham population and root- tip squashes of 20 plants gave the following results 2n = 12 (2 plants), 2n = 12++1 (14 plants) and 2” = 12+2 (4 plants). The accessories were again distinguishable from the normal members of the karyotype by their smaller size (Plate 12, G and H). The normal karyotype members again varied from 1—2y in length and had median centromeres. The longest formed about 11 per cent. and the shortest about 7 per cent. of the total karyotype length. Accessories were present in all the PMCs of Malham material examined and varied in number from | to 4 per cell. When more than one were present they could be paired so that the accessory pair could not be distinguished from the other pairs (Plate 12, J), not associated at all (Plate 12, K) or associated in such a way that the accessory pair was evident (Plate 12, L and M). Frequently when the accessories were associated their separation at anaphase I was completed before that of the normal pairs. The somatic chromosomes of all the above collections were compared with those of C. pyrenaica from South-West Westphalia (Plate 12, N). No differences were apparent. In C. pyrenaica the chromosomes again vary from 1-2 in length and are all approximately metacentric. The longest forms about 10-5 per cent. and the smallest about 6-5 per cent. of the total karyotype length. No accessories were present in any of the root-tip cells examined and all the PMCs showed only 6 bivalents (Plate 12, O). Seeds of C. arctica were obtained from Leningrad. Cytological examination of the plants developed from this seed showed the plant to be a diploid of the x = 7 series. All root-tip squashes gave a count of 2m = 14 (Plate 12, P). The chromosomes, like those of the x = 6 series varied in length from about 1—2y and all were approximately metacentric. This count was verified from PMC squashes which showed 7 bivalents (Plate 12, R). CONCLUSIONS The only previous report of a diploid Cochlearia of the x = 6 series was that of C. pyrenaica (Rohner, 1954). The range of this group is therefore extended. All counts of diploids of the x = 7 series come from farther north and it appears that the x = 6 diploids have a more southerly distribution than those of the x = 7 series. The count previously reported for C. alpina is 2n = 28 (Crane & Gairdner, 1923). Saunte (1955) has, however, pointed out that Crane & Gairdner’s count could be inter- preted as 2n = 24+4 accessories because one of their illustrations shows 4 chromosomes which are smaller than the other karyotype members. If their count is interpreted as 2n = 2444 accessories, then their specimen can be recognized as C. officinalis with accessory chromosomes. This occurrence of accessory or supernumerary chromosomes in C. officinalis is common (Saunte, 1955; Lévkvist, 1963a, 1963b). The true identity of Crane & Gairdner’s material cannot, however, be determined as they do not give the locality of their collection and it has therefore been impossible to re-collect. It has also proved impossible to trace any of their voucher specimens. It is, however, suggested that the true chromosome count for C. alpina is 2n == 12. REFERENCES ConGER, A. D. & FAIRCHILD, L. M. (1953). A quick-freeze method for making smear slides permanent. Stain Tech. 28, 281-3. CRANE, M. B. & GAIRDNER, A. E. (1923). Species crosses in Cochlearia, with a preliminary account of their Cytology. J. Genet. 13, 187-200. Love, A & Love, D. (1961). Chromosome numbers of central and northwest European plant species. Op. Bot. Lund. 5, 1-581. L6vxkvisT, B. (1963a). Nagot om de skanska Cochlearia-arterna. Bot. Not. 116 (2), 326-30. L6vkvisT, B. (1963b). Taxonomic problems in aneuploid complexes. Reg. Veg. 27, 51-7. ROHNER, P. (1954). Zytologische Untersuchungen an einigen schweizerischen Hemi-oreophyten. Mitt. Naturf. Ges. Bern. 11, 43-107. SAUNTE, L. H. (1955). Cyto-genetical studies in the Cochlearia officinalis complex. Hereditas 41, 499-515. Watsonia 6 (3), 1965. NATURAL HYBRIDIZATION BETWEEN THE COWSLIP (PRIMULA VERIS L.) AND THE PRIMROSE (P. VULGARIS HUDS.) IN BRITAIN By S. R. J. WOODELL Botany School, University of Oxford ABSTRACT A population comprising Primula vulgaris, P. veris and hybrids between them has been found in a wood near Oxford. In contrast to the common situation, where only first generation hybrids occur, second generation and later segregants, including backcrosses to each species, have been found. In most cases where these two species meet, genetic and temporal barriers to hybridization are reinforced by ecological isolation. In this wood P. veris is growing in an atypical habitat in close proximity to P. vulgaris. The ecological barriers have thus been lowered, and some hybridization has resulted from the proximity, combined with frequent disturbance. So far as can be seen little introgression has taken place. Thus, although conditions appear to be very favourable for introgression, the two species remain almost entirely distinct. INTRODUCTION It has been shown by Valentine (1955) that there is considerable seed incompatibility between Primula vulgaris and P. veris. The cross between these species is successful only when P. veris is used as seed parent, and the maximum germination obtained has been 37 per cent. Woodell (1960) showed that when the cross is made in this direction, seed development is abnormal; the inner integument is hypertrophied, and the endosperm is reduced in quantity, with a concomitant reduction in embryo size. In the reciprocal cross the seeds are large, with a poorly developed integument, a minute quantity of highly aberrant endosperm, and an undifferentiated embryo which dies early. The F, hybrids from the successful cross are moderately fertile; Valentine (1955) found a mean pollen fertility (percentage grains stained in aceto-carmine) of 32 per cent. with a range from 14-5 to 45 per cent. In the wild the number of seeds per capsule in putative F, plants is about 20 per cent. of that in the two parents. Valentine (1955) obtained experimental backcrosses to each parent with a partial restoration of fertility. Germination of the backcross to P. veris ranged up to 75 per cent., compared with up to 41 per cent. for the backcross to P. vulgaris. Though the putative F, appears in almost all instances where the two species grow near to each other in the wild, it rarely forms more than 1-2 per cent. of the population. Back- crosses are apparently very rare. Until the population to be discussed here was found, I had only seen three plants which were clearly backcrosses. This has been the experience of other workers (Harrison 1931, Clifford 1958). This apparent rarity of backcross or F, plants is rather surprising, since field observation shows that the F, hybrids set seed, and they also provide pollen for fertilization of the parent species. Clifford (1958) attempted to detect introgression between these two species by com- paring means and variability of the two species in ‘pure’ and ‘mixed’ populations. There was no detectable introgression, but Clifford showed that the wild F, plants were more variable than some artificial ones from the garden, and suggested that hybrid segregates other than F, were present, but that they resembled the F, so closely that they had been included among them by him. I shall comment on this view later. Mowat (1961) studied some mixed populations in Fife, using a simplified hybrid index technique. Her results suggest that in one population some sort of hybrid swarm was present. This population would be worth further investigation. In the spring of 1963 a large mixed population was discovered in a wood ten miles 190 Watsonia 6 (3), 1965. HYBRIDIZATION BETWEEN COWSLIP AND PRIMROSE 191 east of Oxford, and in one small area in particular, the scene of recent coppicing, there was a mixture of many hybrid types as well as apparent Fs. It is the plants from this small area that are to be discussed here. MATERIALS AND METHOD The area of the woodland floor that contained the hybrid swarm was marked out with string, and every plant within this area was sampled. No selection of plants was made. The largest flower and the largest leaf was taken from each plant. On the fresh material the following five characters were measured (see Fig. 1): (1) corolla diameter, (2) petal width, (3) calyx circumference, (4) calyx tooth length, (5) leaf angle. In addition four qualitative characters were scored: (6) throat pattern, (7) calyx hair, (8) leaf hair, and (9) pedicel hair. For comparison, a characteristic population of primrose from Marley Wood, Berks, and one of cowslip from Dickleburgh, Norfolk, were scored similarly. Fig. 1. Diagram showing measurements made (a) corolla diameter (b) petal width (c) calyx circumference (d) calyx tooth length (e) leaf width (f) leaf length (g) leaf angle After scoring, the flowers were preserved in 70 per cent. alcohol for subsequent deter- mination of pollen fertility. The frequency of ‘good’ pollen was estimated by teasing out anthers in acetocarmine, and counting the percentage of rounded, stainable pollen grains. It is realized that the characteristic populations may have themselves been introgressed in the past; indeed in Britain it is very possible. However, one is comparing an obviously hybrid population with populations that are clearly referable to the parent species; i.e. comparing relative, not absolute differences. A population at Cockayne Hatley, Cambs, was sampled in two successive years at different times in the season, and comparison of the measurements indicated differences between the two years (Table 1). Consequently a population was sampled at weekly intervals through the following spring to determine seasonal differences in the characters used. The results, shown in Table 2, indicate that there are seasonal differences, and that care must be taken to sample at the height of the season when dimensions are likely to be more stable. The populations used in this study were sampled in late April. It should be noted here that Clifford (1958) pooled his results from many populations, and therefore probably masked any environmental differences between his populations. His Watsonia 6 (3), 1965. 192 S. R. J. WOODELL TABLE |. Means and standard errors of certain characters in the samples of 30 pin and 30 thrum plants of Primula vulgaris from the same population (Cockayne Hatley Wood, Cambs.) in two successive years. Date Corolla diam. Petal width Calyx circ. Calyx tooth-length Pin 22.4.59 34-6 +-097 12-5=--035 19-8 +-060 7°6 +-041 23.3.60 28-7 +-109 10:8-+ -061 15-2 +-080 5:9 +-032 Thrum 22.4.59 34-03 + -123 12-7+-067 18-18+-051 7°37+:-041 24.3.60 30-1 +-090 10-:7+:061 16-0 +-055 6-3 +-038 TABLE 2. Means and standard errors of certain dimensions in samples of pin and thrum plants of Primula vulgaris in each of eight successive weeks in one year. Date No. of plants Corolla diameter Petal width Tooth length Pin 31.1.61 8 26:00+0-44 9-50+0:18 5-75+0:-14 7.2.61 14 25-80+0-19 10-:50+0-07 6-25+0-06 14.2.61 18 27-66+0-22 10-77+0-10 5-64+0-04 20.2.61 20 32 15-0 12-55+0-06 6:03+0-05 28.2.61 29 31-17+0-14 12-36+0-06 5-83+0-03 6.3.61 30 30-73-+0-16 11-22+0-05 6-55-+0-03 14.3.61 28 28-63-+40-14 11-59=-0-05 6-13+0-04 24.3.61 29 32-23+0-18 12-95-=+-0-07 7°-21+0:05 Thrum 31.1.61 7 29 -50+0-36 10-50-+0-21 6°75 +0-07 e276 10 26-50+0-15 9-50+0-10 6-25+0-07 14.2.61 13 27-69=+-0-29 10-69+0-22 5-92+0-08 20.2.61 23 30-70+0-14 11-48+0-07 6-09-+40-04 28.2.61 28 29-48+0-19 10-36+0-05 6:04-40-03 6.3.61 30 30-40+0-13 11-62-+0-06 6:83-0:04 14.3.61 28 28 -91+0-19 11-96-+0-07 6-17+0-06 24.3.61 29 34-90+0-09 12-59+0-03 7-15+0-03 garden experiment to test the effects of environment was based on too few families to provide a reliable indication of environment influences. The results in Tables 1 and 2 indicate that pin and thrum flowers are very little different in their dimensions, and they have been considered together in the scatter diagrams. RESULTS Comparison of ‘characteristic’ and mixed populations The populations of P. veris from Dickleburgh, and of P. vulgaris from Marley Wood are plotted on the scatter diagram in Fig. 2 and the Boarstall Wood population in Fig. 3. It can be seen that there is a clear separation of the two species, using the characters chosen, particularly corolla size and calyx tooth length. Although these two species are so distinct in the field that no one could confuse them, individual quantitative measurements that distinguish them absolutely are not easy to find. However, when the scatter diagram for the Boarstall Wood population (Fig. 3) is compared with those for the species (Fig. 2) it is clear that here the distinctions have broken down, and there is present a series of types ranging between the species. Watsonia 6 (3), 1965. HYBRIDIZATION BETWEEN COWSLIP AND PRIMROSE 193 42 bey vy vy MARLEY WOOD * pe My ¥sy 38 i me =e : y at = s ¥ A 26 x ag 5 Y% 22 ® O : z ; ® r ) ; Lon di, DICKLEBURGH é 4¢8 ry e yy ee? 14 @eg $e¢ a6 TOOTH LENGTH (@) 1 2 3 4 fa) Fig. 2. Scatter diagram representing populations of Primula vulgaris from Marley Wood, Berkshire, and P. veris from Dickleburgh, Norfolk. (For key see Fig. 3.) The data on the scatter diagram are plotted so as to show five groups of plants, dis- tinguished on the basis of the marking of the throat and flower colour (termed together, ‘throat pattern’). These five groups are as follows (see key to Fig. 3): C. Flowers of typical cowslip colour and throat marking; bright yellow, with distinct, deep orange spot at the base of each petal. | CCP. Flowers larger than those of cowslip; slightly paler yellow, inflorescence less nodding, orange spot less distinct. ; CP. Flowers intermediate in size between cowslip and primrose, colour medium yellow. Orange throat spot diffuse. PPC. Flowers large, almost as large as those of primrose; pale yellow, with very diffuse , and pale orange spot at base of each petal. P. Flowers pale sulphur-yellow, very large, with greenish marking at throat of corolla tube. Typical primrose flower colour and pattern. These groupings were made before plotting the diagram, and it is interesting to note that the quantitative measurements place the plants in the positions that might be expected if the throat pattern distinctions are correct. Fig. 3 does give the impression that the putative F, plants are nearer to P. vulgaris than to P. veris. This is, I think, a chance effect, and not an indication that the plants that I have classified as F, individuals are in fact backcrosses to P. vulgaris. This opinion is based partly on the evidence of pollen fertility, which will be discussed later, and partly on a Watsonia 6 (3), 1965. 194 S. R. J. WOODELL 38 ra y BOARSTALL WOOD y ae mixed xs é y W f = A y & CALYX CIRCUMFERENCE. E é dx pis Vix ae : ee ae = ifty ee 4 LEAF ANGLE. 2 Xx} > 2 @ O22 45-62 ® jo é 6 Be CO dof che a & © ° Gea 5 ece ee ee j & @ & eé éedde oe e ppc 14 & eege ceed p $ & @ ee TOOTH. LENG? 0 } 2 3 4 5 6 7 8 9 Fig. 3. Scatter diagram representing the hybrid population from Boarstall Wood, Buckinghamshire. comparison with artificial hybrids. There appear to be more backcrosses to P. vulgaris than to P. veris, and this may be pure chance or the result of the ecological situation. This will be discussed later. Fig 4 shows histograms that have been constructed using four of the characters measured, and again they illustrate that whereas some characters give good separation between these two species, others do not. Throat pattern, being subjectively classified, naturally gives the best distinction. The histograms for the Boarstall Wood population show that there is a range, varying for different characters, between the two parental forms, and that in some cases hybridization has resulted in increased variability, e.g. leaf angle. Fig. 5 is a histogram constructed from a hybrid index, the characters used being set out in Table 3. It is clear from this that there is a peak in the histogram representing the F,, and again that there is more apparent variation toward P. vulgaris than to P. veris. Pollen fertility The pollen fertility of all the individuals was measured from the three populations, except for three plants in Boarstall Wood, which lacked anthers. The populations of the parent species show a consistently high pollen fertility (Table 4) suggesting that they have not been involved in interspecific hybridization. Indeed, even if it has occurred in the past, the individuals that remain are very close to the recurrent parents, and are genotypically balanced. The Boarstall Wood population has a wide range of pollen fertility, and the results are set out in detail in Table 4. They are of interest for several reasons. Firstly, they show that Watsonia 6 (3), 1965. 195 HYBRIDIZATION BETWEEN COWSLIP AND PRIMROSE ‘sofdures woryefndod se14} oy) Ul S1a}SeIeYD INO} JO} sUONNgINsIp Aouanber{ “py ‘B1q 7 NUALLVd HYTIONV AVAT OLLVY d/T HLOOL YWHLAWVIG WIIONOO LVOUHL 5 ls SEOs Bee oe BSS 8 ese sss poxtwi COOMA TIVILSYVOd 0 7 ol 07 O€ Ov sues[na oe CGOOM AYTAVW OZ HOANAA TADIA 0 ol + 02 Of or SIIOA is ool Watsonia 6 (3), 1965. 196 S. R. J. WOODELL DICKLEBURGH MARLEY WOOD a P. veris P. vulgaris 40 30 20 10 30 BOARSTALL WOOD mixed 0-2 3-5 6-8 9-N 12-14 15-17 18-20 21-23 24-26 27-29 30-32) 33-35 Fig. 5. Frequency distribution of scores calculated using a hybrid index. Characters used to construct the hybrid index are given in Table 3. TABLE 3. Characters used in construction of hybrid index. veris vulgaris Character Score 0 1 2 3 4 5) 6 7 8 9 1. Corolla diameter (mm.) 12-14 15-17 18-20 21-23 24-26 27-29 30-32 33-35 36-38 39-41 2. Calyx circumference (mm.) 24-22 21-19 18-16 15-13 12-10 3. Calyx tooth length (mm.) 23 34 45 56 67 7-8 8-9 9-10 4. Throat pattern CRG) Ce (©P, Pe 5. Calyx hair Cia CR er’ (©@P 7k 6. Pedicel hair Cee Gey Ce (O)Re= Pe 7. Leaf hair Cw CB)ix (CPR*en(C) Pai ae Note: In characters 4-7, C = P. veris; C(P) = putative backcross to P. veris; CP = putative F,; (C)P = putative backcross to P. vulgaris; P = P. vulgaris. The minimum score, representing ‘pure’ P. veris, would be O, that representing ‘pure’ P. vulgaris would be 36. the plants regarded as parent species individuals, using the throat pattern criterion, have high fertility, equivalent to that of the plants from ‘characteristic’ populations. Secondly, the putative F, plants have a low average fertility, and thirdly, the putative backcross plants have an average fertility much higher than that of the F, individuals, and, with one excep- tion, an individual value above the majority of individual F, plants. This indicates that pollen fertility, at least, is rapidly restored to a level not falling very far short of that of the parents; in some cases to 100 per cent, even in recognizably hybrid plants. DISCUSSION The data presented in this paper indicates that under some circumstances P. vulgaris and P. veris hybridize fairly freely, and that as well as F, plants, a proportion of backcross Watsonia 6 (3), 1965. HYBRIDIZATION BETWEEN COWSLIP AND PRIMROSE 197 TABLE 4. Pollen fertility of all plants sampled from Marley Wood, Dickleburgh and Boarstall Wood. Species Locality No. of plants Mean of fertility Individual values P. vulgaris Marley Wood 60 99-97 99, 99, 58 at 100 Boarstall Wood 60 99 -98 99, 59 at 100 P. veris Dickleburgh 42 99-93 98, 99, 40 at 100 Boarstall Wood 39 99-95 99, 99, 37 at 100 P. veris X vulgaris Boarstall Wood 20 43-45 9, 17, 20, 24, 31, 33, 33, 35, 38, 40, 45, 46, 50, 50, 51, 60, 60, 62, 63, 65, 70 Putative backcross to P. veris Boarstall Wood 6 88-15 63, 70, 98, 99, 99, 100 Putative backcross to ; P. vulgaris Boarstall Wood 12 77-33 2, 62, 63, 72, 75, 80, 85, 95, 95, 96, 99, 100 or F, segregants may be formed. This is in contrast to the situation normally found, where the barriers to hybridization between these species appear to be strong. In this discussion I will consider what these barriers may be; why they have broken down to some extent in this wood; and further, whether the data show that introgression is occurring between the species. A consideration of the behaviour of these two species suggests that any large amount of hybridization would be unlikely. As has been shown by Valentine (1955) and Woodell (1960) the cross between them is only successful when P. veris is the seed parent (see intro- duction). The low viability will be unlikely to result in many hybrid plants in the field, and in most populations they form a minute percentage of populations where the species come into contact; usually less than | per cent. (Clifford 1958, and personal observation). Even this number is likely to give a false impression of the amount of hybridization occurring, since there is some evidence from garden cultivation of these plants that the F, hybrid almost always outlives the parent species. Of course, extrapolation of this situation to field conditions is dangerous, but it is suggestive. Thus it is possible that even less hybridization occurs, since hybrids will be over-represented by virtue of living longer. A second isolating factor is flowering period. In the Oxford area P. vulgaris flowers from early spring until May, while P. veris flowers from mid-April to June so that the over- lap in flowering periods is relatively small, and although pollinating insects have been observed visiting both species in the same area, I have no idea how frequent this is. Apparent F, hybrids do have a flowering season that overlaps both parent species, so once they are established there are opportunities for backcrossing to take place. Experimental work on pollinating insects is now in progress, and may lead to a better knowledge of the amount of pollen transfer that takes place between these species. In my view habitat differences constitute the most important isolating factors. In England, P. veris is a plant of unshaded or lightly shaded habitats, rarely found in woodland, though often along the margins of woods and in broad woodland rides. P. vulgaris is a plant of much more mesic environments, found in open conditions in western districts, but in eastern England mainly a woodland plant. Normally the species only meet along wood margins, in hedgerows, and along woodland rides and paths. Thus the opportunities for pollen transfer are not great. As Clifford (1958) pointed out, these meeting places are often disturbed, and it is in disturbed habitats that introgression is favoured (Anderson 1949). But very often the disturbance is short-lived, and here lies one possible reason for such limited hybridization. If a species cross takes place, and the cowslip thus pollinated sets seed, it is unlikely that the plants resulting from such seed will flower before their second year, even under very Watsonia 6 (3), 1965. 198 S. R. J. WOODELL favourable conditions. It may well be later than this. In many areas where F, plants occur, such as woodland rides after clearance has taken place, hedgerows after ditching, etc., the ground usually becomes rapidly grassed over after a couple of years, and the establishment of any new plants, either of the species or of hybrids, will be greatly reduced. Hybrids are likely to be ill-adapted and at a great selective disadvantage. The fact that the F, plants are likely to be with the cowslip, i.e. in the grassy areas, will increase this effect. In meadows which are grazed, and therefore subject to some disturbance, backcross or F, plants are very rare, indicating that even here the hybrids are ill-adapted. Harper & Chancellor (1959) and Cavers & Harper (1964) have demonstrated that even in meadowland conditions where they occur naturally, it is difficult to establish several species of Rumex from seed, and Sagar & Harper (1964) have shown that the same holds true for Plantago. Such difficulties of establishment are likely to be much greater for hybrids. There are therefore fairly strong genetic barriers between these two species, operating after fertilization; there is some temporal isolation, and there is very considerable ecological isolation. From a consideration of these isolating factors, one could perhaps forecast that the most important factor leading to an extension of hybridization would be a breakdown of the ecological isolation. If one considers a mixed population in grassland, where prim- rose has emerged from its usual shaded habitat, the establishment of hybrids would be inhibited in the manner suggested above. On the other hand, if cowslips occurred in wood- land, then the hybrid seed would be given a greater chance of establishment, since ground cover in woodland is less complete. (It is relevant here to note that in woods containing Oxlip (P. e/atior) seed is often set, and seedlings are common, indicating a greater chance of seedling establishment in woodland.) Further, if that woodland is disturbed not once, but frequently, then the proximity of the two species, together with regular disturbance, might be expected to favour an increase in hybridization. This situation is just that which exists in Boarstall Wood. Boarstall Wood is an oakwood (Quercus robur), with well-spaced oaks, and a very dense understory of coppiced hazel (Corylus avellana). The soil is yellow-brown clayey loam, with a pH of from 5-0-5-7 (four samples) and a cation exchange capacity of 26:1 me/100g. Under the dense shrub layer the field layer is sparse. Immediately adjoining the sample area, in 1964 both P. veris and P. vulgaris were present, with occasional hybrids. So few were in flower here that no estimate of the percentage of hybrids could be obtained in 1964. Table 5 gives a list of species present, in a 10 sq m area, estimated on the Domin cover- abundance scale. Also given is a list from a 10 sq m area in the region sampled for Primula. This part of Boarstall Wood had been cleared completely of undershrubs in the winter of 1962-63. The great increase in ground cover is evident from the Table. Yet there was still a good deal of bare ground, which presumably was providing potential sites for establishment of hybrids. Some weedy species had already invaded (Plantago major, Rumex sanguineus). At the time of writing, the area of the wood in question has been completely cleared of trees and shrubs, and the operation has been accompanied by tremendous disturbance. It will be watched with interest over the next few years. There is no difference in ground flora associated with P. veris and P. vulgaris; at least no obvious difference. The two species are interspersed with each other over the sample area, which extends for about 150 m inward from the eastern edge of the wood. This is a contrast to the situation described by Valentine (1948) for P. elatior and P. vulgaris in Buff Wood, Cambs. In Boarstall Wood the cowslip is in an atypical habitat, and its invasion of the wood may have been correlated with an earlier complete clearance. An examination of the felled oaks in early 1965 disclosed that they were all the same age; about 100 years. This suggests that 100 years ago the area was cleared and replanted. Such an occasion may have facilitated the invasion of P. veris. Once established, its survival will have been helped by regular coppicing. Cowslips, identifiable by their leaf shape, were present under the 10-12 year old shrub canopy adjacent to the sample area. In recently coppiced areas they Watsonia 6 (3), 1965. HYBRIDIZATION BETWEEN COWSLIP AND PRIMROSE 199 TABLE 5. Species present in a 10 sq m area of Boarstall Wood, (a) in an area untouched for about ten years (b) in an area coppiced 18 months previously. Estimates of cover-abundance on the Domin Scale. (a) (6) Corylus avellana 9 3 Euonymus europaeus Ligustrum vulgare Lonicera periclymenum Prunus spinosa Rosa arvensis R. canina agg. Rubus fruticosus agg. Thelycrania sanguinea mBANN W W w Ajuga reptans 2) Carex sylvatica 3 Cirsium vulgare Deschampsia cespitosa 3 Dryopteris filix-mas 2 Galium aparine Geum urbanum 1 Hypericum tetrapterum Juncus communis 1 Plantago major Potentilla sterilis 1 Primula veris 2 P. vulgaris 3 Hybrids 2 Rumex sanguineus CoN Wh m= NW wWwWN = Ph = dH = Atrichum undulatum 2 Cirriphyllum piliferum 2 Eurhynchium praelongum 4 E. striatum Zz. 1 E. swartzil 1 Fissidens taxifolius 1 Funaria hygrometrica 3 (on site of fire) Mnium undulatum 1 1 Thamnium alopecurum 2 Thuidium tamariscinum 4 3 Bare ground 8 5 flower freely, and the opportunities for hybridization, establishment of hybrids, and maintenance of cowslips in the population are obvious. The disturbance caused by clearing leaves bare ground that persists for several years. It might not be thought surprising that hybrids other than F, forms are found in such a situation. Clifford (1958) suggested that selection was maintaining each of the species and a third phenotype resembling the F, hybrid. He based this view on two facts; that in wild popula- tions only F, hybrids are normally found and that the wild putative F, plants are more variable than artificial F, plants grown in a garden. In my opinion, these facts cannot be taken to suggest that selection is operating in this manner. My reasons for this are as follows: 1. The artificial F, plants that were used as a basis for comparison of variability were from four families, grown in the same garden. The wild plants were from 15 field popula- Watsonia 6 (3). 1965. 200 S. R. J. WOODELL tions. One would expect the wild hybrids to be more variable, even if they are all true F; plants. They will have come from many more than four crosses; they will have been growing in a variety of environments, and we have seen that environmental effects are demonstrable. It is likely that differences in variability of the order shown by Clifford could have been produced by these two causes alone. 2. For selection to act in the way suggested by Clifford, one would have to suppose that the F, plants were adapted to a set of conditions that prevail almost everywhere that the two parent species meet, and that this intermediate habitat was in fact almost exactly inter- mediate; an “F, habitat’ in fact. This is clearly not so, and F, hybrids can be found in quite a wide variety of conditions. What is often true of F, hybrids is that because they can only be produced from cowslip mothers, they generally occur in habitats more like those of cowslip than those of primrose. One would expect selection to favour backcrosses to cowslip, if anything. 3. For reasons stated above, ecological conditions appear to be operating against the establishment of hybrids other than F, plants, and the widespread presence of these F, plants is more a result of this than of positive selection for them. A study of the pollen fertility of the putative F, plants found by Clifford in the wild might have given a better indication of their true nature. The pollen fertility results from the populations studied here indicate that there is a considerable loss of fertility in the first generation hybrid, followed by a restoration to a high level in later generations (cf. Moore 1959, Heiser 1951, Gajewski 1957). Peto (1933) found in hybrids between Festuca pratensis and Lolium perenne that a much reduced pollen fertility in the first generation was restored, in some cases to a level near to that of the parents, in the backcrosses. He found among his backcrosses one or two plants with very low pollen fertility or complete sterility. It is interesting to note that in my sample one of the putative backcrosses has a pollen fertility of 2 per cent. (Table 4). Finally we have to consider whether introgression is taking place in this situation where everything seems to be in its favour. In an attempt to determine this I have extracted from my original sample in Boarstall Wood, which was a complete sample from a circumscribed area, the plants which were, on subjective grounds, ‘characteristic’ primroses or cowslips, and compared their variability with that of the ‘characteristic’ populations from Dickle- burgh and Marley Wood. The results are set out in Table 6. The results in this Table indicate that while thrum P. vulgaris is significantly more variable in Boarstall Wood than in Marley Wood, the species in Boarstall Wood are not in general more variable. Introgression, in the sense of Anderson (1949) is the transfer of genes across an in- complete genetic barrier, and the establishment of well adapted backcross types, resembling TABLE 6. The variances for selected characters of plants in ‘pure’ and ‘mixed’ populations of Primula vulgaris and P. veris. Corolla Calyx Calyx tooth Leaf length| Species Locality diameter circumference length breadth ratio P. veris thrum Dickleburgh 4-18* 9-05 0-38* 0-053 Boarstall 1-82* Dal 0-86* 0-042 P. veris pin Dickleburgh 3122 4-45 0-48 0-117 Boarstall 2°67 6°12 0-68 0-121 P. vulgaris thrum Marley Pepe Die 0-42** 0-39*** Boarstall 1190s 2) 26.0 0. 13"5— P. vulgaris pin Marley less 2°91 0-65 0-09 Boarstall tS 4-72 0-72 0-13 * — differences significant at 5 per cent. level. ** — at 1 per cent. level. +27 — atl pen centnlevele Watsonia 6 (3), 1965. HYBRIDIZATION BETWEEN COWSLIP AND PRIMROSE 201 either or both parents. If such has happened, one would expect greater variability in the parent species from an area where introgression has occurred. The results from Boarstall Wood indicate that with the exception of thrum primrose, they are hardly more variable. There are two alternative interpretations of this situation. Either introgression has not proceeded very far, or it is so complete that the methods used for detecting variability, and the characters selected, do not show it. Let us examine these alternatives a little more closely. When introgression has occurred in the past it is often difficult to detect (Stebbins, 1950). In view of the strength of stabilizing selection this is not perhaps surprising. If one accepts this, then one cannot rule out the possibility that the small amount of extra varia- bility in the Boarstall Wood primroses is a real indication of complete introgression. One does not know what characters in these two species are of real selective value, and those used in the analyses were selected merely because they were those that gave the clearest distinction between the species. Even among these there are some partially common to both species, and the detection of introgression in such cases is especially difficult. The presence of a number of individuals that are either second generation segregates or backcrosses indicates at least that introgression is in progress. (These individuals are more likely to be backcrosses in view of the likelihood that fertilization of hybrids will take place mainly by parental pollen.) Without knowing the detailed history of this wood it is difficult to know how often disturbance has allowed a burst of flowering and subsequent hybridization. The evidence at hand does not seem to be conclusive enough to enable one to decide with certainty whether introgression is well advanced or in an early stage. It does not really matter; what does matter is that it is clear that introgression does not occur readily between these two species. It requires very exceptional habitat conditions, and these have been met in Boarstall Wood. This area will be watched closely from now on, in an attempt to follow the course of events in the train of the recent catastrophic disturbance. Has introgression occurred elsewhere between these two species? Literature records for continental Europe are very scanty, but they indicate that backcrosses are rare (Pugsley 1927). In Britain, although Harrison (1931) refers to the presence of occasional backcrosses, he states that the vast majority of hybrids in the field are first generation progeny. The only two investigations of possible introgression have been those of Clifford (1958) and Mowat (1961). It is not surprising that Clifford failed to detect introgression, in view of the fact that he pooled material from different localities, thus masking seasonal and environmental differences. The present investigation indicates the importance of individual habitat con- ditions. Mowat’s data were based on populations from three localities in Fife. She used a hybrid index based on ten characters, each being given a score of 0 for typical P. vulgaris and 1 for P. veris; thus ‘typical’ P. vulgaris would have a score of 0 and ‘typical’ P. veris one of 10. One of Mowat’s populations appeared to be ‘characteristic’ P. vulgaris. Her third population had 43 per cent. of the plants differing from the P. veris standard by one character, and she took this to be evidence of introgression. I think that this deviation is unlikely to reflect more than the usual variation pattern; characters distinguishing these species are not of the ‘all or none’ type that Mowat has suggested. Her second population clearly contained hybrids, and although they are grouped in such a way that suggests that most are F, individuals, there may be a few backcross types. In the absence of pollen fertility data it is difficult to say whether any introgression had occurred. It would thus appear that hybridization between primrose and cowslip beyond the first generation is very rare. The internal isolating barriers are strong, and unless a par- ticular and very unusual set of ecological conditions operates, the possibility of backcrossing and subsequent introgression is remote, in Britain at least. Even in such a favourable situation as that in Boarstall Wood, there is little evidence for the occurrence of intro- gression. Watsonia 6 (3), 1965. 202 S. R. J. WOODELL ACKNOWLEDGMENTS I am grateful to Professor D. H. Valentine and Dr. D. M. Moore for their helpful criticisms of the manuscript, and to Miss Janet Bell for her technical assistance. This work was carried out while I was in receipt of a research grant from the Nature Conservancy, to whom my thanks are due. REFERENCES ANDERSON, E. A. (1949). Introgressive Hybridization. John Wiley & Sons, New York. Cavers, P. B. & Harper, J. L. (1964). Biological Flora of the British Isles. Rumex obtusifolius L. and R. crispus L. J. Ecol. 52, 737-66. CLIFFORD, H. T. (1958). Studies in British Primulas VI. On Introgression between Primrose (Primula vulgaris Huds.) and Cowslip (P. veris L.). New Phytol. 57, 1-10. GAJEWSKI, W. (1957). A cytogenetic study on the genus Geum L. Monographiae Botanicae 4, 415 pp. Warsaw. Harper, J. L. & CHANCELLOR, A. P. (1959). The comparative biology of closely related species living in the same area. IV. Rumex: Interference between individuals in populations of one or two species. J. Ecol. 47, 679-95. HARRISON, J. W. H. (1931). The Northumberland and Durham Primulas of the Section Vernales. Trans. Nth. Nat. Un. 1, 48-77. HEIser, C. B. (1951). Hybridization in the annual sunflowers. Helianthus annus x H. debilis var. cucumeri- folius. Evolution 5, 42-51. Moore, D. M. (1959). Population studies on Viola lactea Sm. and its wild hybrids. Evolution 13, 318-32. Mowat, A. B. (1961). An investigation of mixed populations of Primula veris and P. vulgaris. Trans. Bot. Soc. Edinb. 39, 206-11. Peto, F. H. (1933). The cytology of intergeneric hybrids between Festuca and Lolium. J. Genet. 28, 113-56 PuGsLey, H. W. (1927). Primula hybrids. J. Bot. Lond. 65, 351. SAGAR, G. R. & HARPER, J. L. (1964). Biological Flora of the British Isles. Plantago major L., P. media L., and P. lanceolata L. J. Ecol. 52, 189-222. STEBBINS, G. L. (1950). Variation and Evolution in Plants. Oxford. VALENTINE, D. H. (1948). Studies in British Primulas II. Ecology and taxonomy of Primrose and Oxlip (Primula vulgaris Huds. and P. elatior Schreb.). New Phytol. 47, 111-30. VALENTINE, D. H. (1955). Studies in British Primulas IV. Hybridization between Primula vulgaris Huds. and P. veris L. New Phytol. 54, 70-80. WOODELL, S. R. J. (1960). Studies in British Primulas VII. Development of normal seed and of hybrid seed from reciprocal crosses between P. vulgaris Huds. and P. veris L. New Phytol. 59, 302-13. Watsonia 6 (3), 1965. BOOK REVIEW Vascular Plants of the Pacific Northwest. C. L. Hitchcock, A. Cronquist, M. Ownbey and J. W. Thompson. Part 2: Salicaceae to Saxifragaceae, by C. L. Hitchcock and A. Cronquist. Drawings by Jeanne R. Janish. Pp. 597. University of Washington Press, Seattle. 1964. Price $15.00. This is the fourth volume to appear of a flora covering the State of Washington, northern Oregon, Idaho north of the Snake River plains, the mountainous western part of Montana, and southern British Columbia. The parts are being published in reverse order and only one, the fifth, remains to come. The previous volumes were Part 5, Compositae, by Cronquist (1958); Part 4, Ericaceae through Campanulaceae, by Hitchcock, Cronquist and Ownbey (1959); and Part 3, Saxifragaceae to Ericaceae, by Hitchcock and Cronquist (1961). The families, with some exceptions, follow the Engler system and every species is beautifully illustrated, often with its varieties and also with analytical drawings of important parts such as flowers, fruits or types of indumentum. Among the great merits of this flora are the indented keys, the elaborate descriptions and synonymy, the valuable critical notes, the treatment of varieties and intergrades, and the comments on the horticultural value of many taxa. Type collections are cited, even of taxonomic synonyms, and the derivation of generic names is given. Much less acceptable to most of us here is the placing of both genera and species in alphabetical order. Dare we whisper, this seems scarcely more intelligent than the similar arrangement of the covers in a herbarium. It makes a mockery of the indexes at the end of the volumes, where synonyms (in italics) and the relative accepted names (in roman) are ranged in strange parallel vertical columns, page references being confined to accepted specific names (a minority), those of families and genera, and common names. A further disadvantage of this system is that the illustrations also appear in alphabetical order, so that the drawings of closely allied species may well be on different pages. The present volume, since it completes the treatment of the Dicotyledons, begins with two keys (synop- tical and artificial) to the orders and families dealt with in Parts 2-5: the synoptical key amplifies the phylogenetic chart based on Cronquist’s system which is included in the introduction to this Part. Among the important families of which we find accounts in this volume are the Cruciferae, Ranunculaceae, Cras- sulaceae, Polygonaceae, Caryophyllaceae, Portulacaceae, Chenopodiaceae, Salicaceae and the order Urticales. The treatments of critical genera are often wisely conservative and, now and then, one is reminded of remarks or confessions by contributors to Flora Europaea: for instance, Erysimum seems as difficult in this area as in Europe and it seems that part of the confusion results from the maintenance of several taxa that represent no more than extremes in a nearly continuous series of variation; the genus is badly in need of monographic study. But, in contrast to Flora Europaea, the rank of variety is retained for the more common kind of infraspecific taxon, that of subspecies being restricted to a ‘hierarchical category (i.e. a group of varieties)’. British students of Amaranthus may not like the reduction of A. blitoides S. Wats. to A. graecizans L., and among other points to be noted by British botanists are the maintenance of generic status for Tillaea and the appearance of Chenopodium pratericola Rydb. as a synonym of C. leptophyllum (Mogq.) S. Wats. var. oblongifolium S. Wats. N. Y. SANDWITH 203 Watsonia 6 (3), 1965. SS S——..eNE~ A BIBLIOGRAPHICAL INDEX OF THE BRITISH FLORA COMPILED BY N. DOUGLAS SIMPSON, ™.A., F.L-s. The purpose of this work is to provide references to sources of information whereby flowering plants, vascular cryptogams and Charophytes found in Britain may be identified, their history traced and their geographical range determined. In addition, information is provided on plantlore, local names, poisonous plants and weeds. The work has been in preparation for nearly twenty years and contains over 65,000 entries, including references to books, articles and manuscripts relating to the flora of the British Isles from the fifteenth century to the present time. Demy 4to, bound, 448 pages in double column. Limited Edition of 750 copies. PRICE £3 15s. (postage extra) Obtainable from the compiler at ‘Maesbury’, 3 Cavendish Road, Bournemouth, Hants. ROYAL IRISH ACADEMY Robert Lloyd Praeger Fund for Field Natural History Grants, not normally exceeding £40 in any one year, will be awarded for field work relevant to the Natural History of Ireland. Application forms, which should be returned before February 15 of next year, may be obtained from The Secretary, Royal Irish Academy, 19, Dawson Street, Dublin, 2, who will also be glad to give further information. PROCEEDINGS OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Published Twice Yearly Price [5/- per part Obtainable from D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. IRISH NATURALISTS’ JOURNAL A Magazine of Natural History Published Every Quarter by the I.N.J. Committee Edited by Miss M. P. H. KERTLAND, M.Sc with the assistance of Sectional Editors Annual Subscription, 10/- post free Single Parts, 3/6 All communications to be addressed to :— The Editor, Department of Botany, Queen’s University, Belfast we, “= WATSONIA >? A 4 JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editor: M. C. F. PROCTOR, M.A., Ph.D. Vol. 6 SEPTEMBER, 1966 Pt. 4 CONTENTS A LIST OF INFRASPECIFIC TAXA OF BRITISH PHANEROGAMS TESTED IN CULTIVATION. By D. E. ALLEN 205-215 THE BREEDING RELATIONSHIPS OF SOME EUROPEAN EUPHRASIAE. By P. F. YEO 216-245 Srupres IN RANUNCULUS SUBGENUS BATRACHIUM (DC.) A. Gray. III. RANUNCULUS HEDERACEUS L. AND R. OMIOPHYLLUS TEN. By C. D. K. Cook 246-259 STUDIES ON WELSH ORCHIDS. III. THE COEXISTENCE OF SOME OF THE TETRAPLOID SPECIES OF MARSH ORCHIDS. By R. H. ROBERTS .. 260-267 BOOK REVIEW 268-269 PUBLISHED AND SOLD BY THE BOTANICAL SOCIETY OF THE BRITISH ISLES c/o DEPARTMENT OF BOTANY, BRITISH MuszUM (NATURAL History), LONDON PRICE: TWENTY-FIVE SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES PATRON HER MAJESTY QUEEN ELIZABETH THE QUEEN MOTHER OFFICERS Vice-Presidents: DR. C. E. HUBBARD, Dr. J. G. DONY, R. MACKECHNIE and E. MILNE-REDHEAD. Hon. General Secretary: E. B. BANGERTER, c/o Dept. of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. Hon. Treasurer: J. C. GARDINER, Thrift House, 12 & 14 Wigmore Street, London, W.1. Hon. Editors: (Watsonia) Dr. M. C. F. PROCTOR, Hatherly Biological Laboratories, Prince of Wales Road, Exeter. (Proceedings) E. F. GREENWOOD, M.Sc. Hon. Meetings Secretary: Mrs. MARY BRIGGS, White Cottage, Slinfold, Sussex. Hon. Field Secretary: P. C. HALL, 6 Johns Close, Gorsewood Road, Hartley, Dartford, Kent. Hon. Membership Secretary: Mrs. J. G. DONY, 9 Stanton Road, Luton, Bedfordshire. Applications for membership should be addressed to the Hon. Membership Secretary, from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 25/— per part. To be obtained from D. H. Kent, 75 Adelaide Road, West Ealing, London, W.13. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/— for Junior Members, and 10/— for Family Members, and the Society’s year runs from January Ist to December 31st. All subscriptions should be paid to the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. A LIST OF INFRASPECIFIC TAXA OF BRITISH PHANEROGAMS TESTED IN CULTIVATION By D. E. ALLEN In recent years interest in infraspecific taxa has greatly declined among British botanists both amateur and professional. This is undoubtedly a reflection of the scanty, if not non- existent, treatment of named variants below subspecific rank in the standard floras and check- lists published since the war, in understandable reaction to the excessive naming of trivial variants by earlier generations and the accompanying nomenclatural chaos. Modern experimental work has in many cases exposed the futility of attempting to give names where the variation is more or less continuous in character. It has also shown how inevitably mis- leading taxonomic treatment can be in terms of genetics, limited as it must be to readily discernible external differences and so liable to recognise a highly conspicuous variant due to the effect of a single gene, while at the same time bound to ignore different chromosome types which, as in Lathyrus pratensis (Larsen 1957), cannot be separated morphologically with any consistency. While recognizing these limitations, however, there may still be a case for keeping up many of the existing taxa, provided this is done with more discrimination than in the past. Surprisingly little is known about the British distribution of even the more conspicuous of these variants and, to put the argument at its lowest, unless they receive recognition in the form of Latin epithets they are likely to continue to pass largely unrecorded, particularly by amateur botanists, to whom a great part of such recording must inevitably fall. On scientific grounds the task of working out the distribution of even relatively trivial variants is surely justifiable if it results in greater recognition and more detailed knowledge of, for example, polymorph-ratio clines (as demonstrated by New (1958) in Spergula arvensis L.), of the spread of certain variants (such as the inland radiate forms of Senecio vulgaris L.), of variation betraying cryptic introgression between species, and of the existence of local races which, perhaps representing incipient subspecies, contribute to an understanding of the phytogeographical relationships of different areas. Basic to such work, however, is more precise and extensive information on the very validity of many taxa that have been described. Oenanthe lachenalii var. minima Camus & Rouy, for example, is a dwarf variant reported from maritime localities in Co. Cork, Co. Wexford, Anglesey and the Isle of Man ; it has never, apparently, been tested in cultiva- tion and so it remains uncertain whether it is an interesting race peculiar to the shores of the Irish Sea or a mere ecad. On the other hand, far too much has been written in the past and records are still today being published for variants (such as the terrestrial state of Polygonum amphibium L.) that have been exposed, more or less conclusively, as ecads—and thus surely do not merit taxonomic recognition. Because of the absence of any list of reported experiments, many variants have been tested repeatedly, while others whose status could have been much more usefully checked have been continually ignored. If it serves no further purpose than to redirect efforts in future, therefore, the list that follows may prove to be of use. The list has been compiled over fifteen years primarily from British publications, and except possibly for published reports of British experiments it cannot be regarded as exhaustive. It is confined to variants that have received taxonomic recognition (with one or two exceptions) and to those that have been reported as occurring in the British Isles. Albinos and semi-albinos have been ignored, as these have only partially and inconsistently had Latin epithets bestowed on them. A number of names new to British botanists in- evitably appear, but in general no attempt has been made to establish earliest legitimate names, and new combinations have been deliberately avoided, not least because general 205 Watsonia 6 (4), 1966. 206 Dy EVALLEN agreement has yet to be reached on the appropriate infraspecific rank (if any) to be given to each category of variant. The list, inevitably, has a number of pitfalls and must be used with caution. Many of the experiments reported on are, by modern standards, crude in the extreme, and most of those performed (by all except specialists in genetics) before about 1930 probably need re- checking. Some of the results reported in the last century (for example, Anagallis arvensis var. azurea and Leontodon hispidus var. glabratus) are almost certainly misinterpretations resulting from unawareness of the existence of heterozygotes. Too limited a progeny and too short a period of observation may also have led to errors. In many cases, as observed in Plantago coronopus by Bocher, Larsen & Rahn (1955) and in Mentha by Morton (1956), cultivation has a “‘levelling-out”’ effect, the variation noted in the field being still recogniz- able but greatly reduced. Sometimes, too, the variation may differ in intensity from year to year, as reported by Riddelsdell (1948) in cultivated plants of Viola odorata var. sulfurea. Most important of all, non-genetic effects can be closely simulated, perhaps even exactly duplicated, by hereditary variants (genocopies), so that it is sometimes impossible to be sure without breeding experiments whether a particular variant is genotypic or phenotypic. Cook (1962) has described an instance of this in alpine specimens of Ranunculus trichophyllus Chaix, some retaining their original character when cultivated in lowland conditions, while others come to resemble lowland material. A similar situation in Veronica anagallis- aquatica L. has been recorded by Burnett (1955). Examples in the list of conflicting data which may be explicable on this basis are : Caltha palustris var. minor Chenopodium rubrum f. pseudobotryodes Sarothamnus scoparius subsp. maritimus Aethusa cynapium vat. agrestis Heracleum sphondylium var. angustifolium Plantago lanceolata var. capitata P. coronopus var. pygmaea. Test results are placed in the list in one of four categories abbreviated as follows : CC Constant “‘in cultivation’’, i.e. all cases in which raising from seed is not explicitly mentioned, ranging from elaborate transplant experiments to mere transfers from the wild to a garden. RC Reverted to type “‘in cultivation’’. CS Constant when grown from seed. RS _ Reverted to type when grown from seed. Fuller genetical details can be obtained in many cases by referring to the original source. Unless a horticultural origin is clearly indicated, experiments reported in British journals can almost without exception be assumed to have been carried out on wild material of Britannic origin, although very often this has not been specifically stated. It is possible that results reported on non-Britannic material may not invariably be applicable to taxa bearing the same name in this country, due to a differing genetic make-up or more simply to taxonomic confusion. Equally, results may conceivably vary for apparently identical variants originating from different areas even within the British Isles. Of all the infraspecific variants ever reported as British, the great majority of which are cited in G. C. Druce’s British Plant List (Ed. 2, 1928), probably less than a third are now revealed to have been tested so far by cultivation. CALTHA PALUSTRIS var. MINOR (Mill.) DC. CC—Christy, Mag. Nat. Hist. (1833) 54 ; Graebner ex Druce, New Phytol. (1912) 357. RC—Hunnybun ex Moss, Camobr. Brit. FI., 3, (1920) 105 ; var. GUERANGERII (Bor.) Syme. CC—Linton, Rep. Bot. Exch. Club for 1893, 399. For cultivation experiments on the species as a whole see Reese, Planta (1954) 203-268. Watsonia 6 (4), 1966. a ie — INFRASPECIFIC TAXA TESTED IN CULTIVATION 207 TROLLIUS EUROPAEUS var. RELICTUS (H.-Harrison) Druce. CC—Heslop-Harrison, Trans. Nat. Hist. Soc. Northumb., Durh. and Newc. (1917) 136. ANEMONE NEMOROSA var. PURPUREA DC. CC—Butler ex Riddelsdell, Hedley & Price, Fl. Glos., 4 (1948). RANUNCULUS ACRIS subsp. ACRIS (R. boraeanus Jord.) and subsp. FRIESIANUS (Jord.) Rouy & Fouc. (R. vulgatus Jord. ex Bor.; R. stevenii auct.). CC—Drabble, Rep. Bot. Soc. and Exch. Club for 1930, 474; var. pumiLUS Wahlenb. (subsp. borealis (Trautv.) Nyman?). CC—Druce, J. Bot. (1890) 40 ; Graebner ex Druce, New Phytol. (1912) 357 ; Turesson, Hereditas (1925) 179-184 ; Bécher, Dansk Bot. Arkiv (1945), 12 (3), 1-16. R. REPENS var. PROSTRATUS (Poir.) Gaud. (var. subacaulis Bréb.; R. reptabundus Jord.). CC—Shoolbred, Rep. Bot. Soc. and Exch. Club for 1913, 444, and for 1915, 309. R. BULBOSUS var. VALDEPUBENS (Jord.) Corb. (var. dunensis Druce? ; var. maritimus Lange ?). CS—Jordan, Ann. Soc. Linn. Lyon (1861) 450. CC—Drabble, Rep. Bot. Soc. and Exch. Club for 1931, 736 ; Rossiter, Proc. Bot. Soc. Br. Is. (1956) 78. R. ARVENSIS var. ETUBERCULATUS Sér. (var. inermis Nees ; var. reticulatus (Schmitz & Regel) Rouy & Fouc.). CS—Godron, Mém. Acad. Stanislas (1873) 77 ; Hoffmann, Bot. Zeit. (1884) 244. Recessive—Saunders in Bateson, Saunders & Punnett, Rep. Evol. Comm. Roy. Soc. (1904) 54-55. R. SARDOUS var. PARVULUS (L.) Lange. RC—Hailstone ex Wilkinson, Rep. Yorks. Phil. Soc. (1895) 12. R. FLAMMULA subsp. FLAMMULA var. MAJOR Schult. CC—Padmore, Watsonia (1957) 20 ; var. RADICANS Nolte (var. tenuifolius Wallr.). RC—Beeby, J. Bot. (1887) 370 and Scot. Nat. (1888) 210 ; Padmore, Watsonia (1957) 25-26 ; subsp. minimus (A. Benn.) Padmore. CS—Padmore, Watsonia (1957) 23 ; subsp. scoTicus (E. S. Marsh.) Clapham. CC—Marshall, J. Bot. (1889) 230 ; Padmore, Watsonia (1957) 23. R. FICARIA var. INCUMBENS F. W. Schultz. CS : more or less recessive—Marsden-Jones & Turrill, J. Genet. (1952) 530. NYMPHAEA ALBA f. TERRESTRIS Schuster. RC—Brand ex Hegi, J/l. FI. Mitteleur., ed. 1, 3, 44 (1911). PAPAVER RHOEAS var. STRIGOSUM Boenn. Partially CS—Dixon, J. Bot. (1892) 309. Single dominant gene—Winge, Bull. Acad. Sci. URSS, Genet. (1932) 115-120 ; var. PRYORII Druce. Dominant—Philp, J. Genet. (1933) 194 ; var. CHELIDONIOIDES O. Kuntze. Dominant, probably two factors involved—Philp, J. Genet. (1933) 197-199. BRASSICA RAPA var. BRIGGSIT H. C. Wats. ex Briggs. CS—Varenne, J. Bot. (1881) 360 ; Watson ex Brown, Engl. Bot., ed. 3, Suppl. 22 (1892). RAPHANUS RAPHANISTRUM var. OCHROLEUCUS (Stokes) Peterm. (var. flavus (S. F. Gray) Schiibler & Mart.). CS—Druce, Rep. Bot. Soc. and Exch. Club for 1917, 210. LEPIDIUM HETEROPHYLLUM var. ALATOSTYLUM (Towns.) Thurston & Vigurs. CS—Townsend., J. Bot. (1903) 97-98. : COCHLEARIA DANICA var. PRAECOX Le Jolis. CC—Wheldon, Rep. Bot. Soc. and Exch. Club for 1916, 557. CARDAMINE PRATENSIS var. UNIFLORA Sternb. & Hoppe (var. acaulis (Berg) Hallier). RC— Sternberg & Hoppe. Denkschr. Kgl. Baier. Ges. Regensb. (1815) 157. C. IMPATIENS var. POTERIIFOLIA Druce. CC—Stuart ex Druce, Rep. Bot. Soc. and Exch. Club for 1927, 301. BARBAREA VULGARIS var. TRANSIENS Druce. CS—White, Rep. Bot. Soc. and Exch. Club for 1923, 374. var. with variegated leaves. Single recessive gene—Tilney-Bassett, J. Hered. (1963) 545. CARDAMINOPSIS PETRAEA Var. GRANDIFOLIA (Druce). CC—Druce, Rep. Bot. Exch. Club for 1888, 200, J. Bot. (1890) 40 and Ann. Scot. Nat. Hist. (1892) 255. NASTURTIUM OFFICINALE var. SIHFOLIUM Reichb. RC—Howard & Lyon, J. Ecol. (1952) 228.. SISYMBRIUM OFFICINALE var. LEIOCARPUM DC. CS—Jordan, Amn. Soc. Linn. Lyon (1861) 507. Watsonia 6 (4), 1966. 208 D. E. ALLEN ARABIDOPSIS THALIANA Var. BREVICAULIS (Druce). CS—Wilson ex Druce, Rep. Bot. Soc. and Exch. Club for 1923, 27. CAMELINA SATIVA Var. INTEGRIFOLIA Wallr. Double recessive—Tedin, Hereditas (1925) 290. RESEDA LUTEOLA Var. CRISPATA (Link) Miiller Arg. CC—Drabble, Rep. Bot. Soc. and Exch. Club for 1931, 738. VIOLA ODORATA Var. IMBERBIS (Leight.) Hensl. CC—Leighton, F/. Shropshire, (1841) 116. CS—White, Fl. Bristol (1912) 171 ; var. RUBROPURPUREA Greg. CC—Gregory, Rep. Bot. Soc. and Exch. Club for 1917, 148B ; var. SUBCARNEA (Jord.) Parl. CS—Gregory, British Violets, (1912) 4 ; var. SULFUREA (Cariot) Rouy & Fouc. CC—Riddelsdell ex Riddelsdell, Hedley & Dae. FI. Glos. (1948) 68. V. HIRTA subsp. HIRTA f. ROSEA Beeby. CS—Hanbury & Marshall, Fi. Kent (1899) 46 ; subsp. CALCAREA (Bab.) E. F. Warb. CC—Baker ex White, F/. Bristol, (1912) 175. V. RIVINIANA f. VILLOSA Neum., Wahlst. & Murb. RC—White, FI. Bristol (1912) 176 ; Gregory, British Violets, (1912) 59 ; subsp. RIVINIANA and subsp. MINOR (Greg.) Valentine CS—Valentine, J. Ecol. (1939) 420 and New Phytol. (1941) 191. CC—Dizerbo, Bull. Soc. Sci. Bretagne (1960) 103. V. REICHENBACHIANA var. PUNCTATA (Rouy & Fouc.) Wilmott (V. canina var. maculata Boenn. ?). CC—Gregory, British Violets, (1912) 40. HYPERICUM PERFORATUM var. ANGUSTIFOLIUM DC. CC—Westcott, Phytol. (1843) 427. H. PULCHRUM var. PROCUMBENS Rostr. CC—Beeby, Scot. Nat. (1888) 212 ; Druce, Rep. Bot. Soc. and Exch. Club for 1921, 481. HELIANTHEMUM CHAMAECISTUS f. PARVIFLORUM Druce. CC—Anon. ex Druce, Ann. Scot. Nat. Hist. (1911) 98 and Rep. Bot. Soc. and Exch. Club for 1920, 16. subvar. PALLIDUM Druce. CC—Drabble, Rep. Bot. Soc. and Exch. Club for 1931, 738. H. CANUM subsp. LEVIGATUM Proctor. CS—Proctor, Watsonia (1957) 31, 37. SILENE VULGARIS var. HIRSUTA (S. F. Gray). (var. pubescens DC.). Dominant, several genes involved—Marsden-Jones & Turrill, Kew Bull. (1931) 125 ; var. RUBRA (Ramond ex Sér.). Single dominant gene—Turrill, Brit. Plant Life, (1948) 198. S. DIOICA subsp. ZETLANDICA (Compton) Tutin. CS—Baker, J. Ecol. (1947) 284 and New Phytol. (1948) 134. CC—Darimont ex Lambinon, Bull. Soc. Sci. Bretagne (1962) 100. Dwarf var., probably this : CC—-McNab, Proc. Bot. Soc. Edinb. (1838) 57 ; var ALPESTRIS (Fries ex Blytt). Single recessive gene—Saunders in Bateson & Saunders, Rep. Evol. Comm. Roy. Soc. (1902) 15-19 ; De Vries, Species and Varieties (1906) 283 and 585. AGROSTEMMA GITHAGO var. HIEMALIS Compton. CS—Nathansohn, Jahrb. Wissensch. Bot. (1913) 125-153. CERASTIUM FONTANUM subsp. TRIVIALE (Link) Jalas var. PRATENSE (Diard). CC—Drabble, Rep. Bot. Soc. and Exch. Club for 1931, 739 ; var. HANBURYENSE (Druce). Druce, Rep. Bot. Soc. and Exch. Club for 1930, 258. STELLARIA NEMORUM subsp. GLOCHIDISPERMA Murb. CS—Peterson, Bot. Not. (1936) 323. S. MEDIA var. GYMNOCALYX Trautv. Single dominant gene—Peterson, Bot. Not. (1936) 312. S. NEGLECTA var. ELISABETHAE (F. W. Schultz) Bég. Single dominant gene—Peterson, Bot. Not. (1936) 322 ; var. DECIPIENS E. S. Marsh. Simply recessive, 1-2 factors in- volved—Peterson, Bot. Not. (1936) 320. SAGINA MARITIMA var. CILIATA Norsdt. CS—Elliston Wright, J. Bot. (1935) App. 8. S. PROCUMBENS var. SPINOSA S. Gibs. RS—Elliston Wright, J. Bot. (1935) App. 6. SPERGULA ARVENSIS var. SATIVA (Boenn.) Mert. & Koch. CS—Griffiths, J. Bot. (1922) 229. Single gene without dominance ; no genetical association between seed-coat and hairiness characters—New, Ann. Bot. (1959) 25-26 ; var. NANA E. F. Linton. CS— Linton, J. Bot. (1907) 380. CHENOPODIUM ALBUM f. CYMIGERUM (Koch) Schinz & Thell. (C. viride auct.). CS—Boreau, Fl. Centre Fr., 2, (1857) 545. C. URBICUM var. INTERMEDIUM (Mert. & Koch) Koch. CC—Koch, Syn. Fl. Germ., ed. I, (1837) 603. CS—Bickham, Rep. Wats Bot. Exch. Club for 1905-06, 63. Watsonia 6 (4), 1966. INFRASPECIFIC TAXA TESTED IN CULTIVATION 209 C. RUBRUM f. PSEUDOBOTRYODES (H. C. Wats. ex Syme) (Druce. RS—Bromfield, Phytol. (1849) 751 ; Watson, J. Bot. (1869) 142 ; Hume ex Lousley, Rep. Bot. Soc. and Exch. Club for 1932, 444 ; Elliston Wright, Rep. Bot. Soc. and Exch. Club for 1932, 267. But see Rep. Bot. Soc. and Exch. Club for 1945, 165, for an instance where apparently CS. HALIMIONE PORTULACOIDES var. PARVIFOLIA (Rouy). CC—Chapman, Ann. Bot. (1937) 305-309. MALVA MOSCHATA var. INTEGRIFOLIA Ley. CC—Borrer ex Woods, Tourist’s FI. (1850) 62. LINUM CATHARTICUM var. DUNENSE Druce. RC—Drabble, Rep. Bot. Soc. and Exch. Club for 1931, 741. GERANIUM PRATENSE var. LILACINUM Celak. CC—Heginbotham ex Grose, FI. Wilts. (1957) 173. G. ENDRESSII var. THURSTONIANUM Turrill. CC—Rilstone, Rep. Bot. Soc. and Exch. Club for 1937, 653. G. SANGUINEUM var. MICRANTHUM F., B. White. CC—Bennett, Sci. Gossip (1892) 199 ; var. PROSTRATUM (Cav.) Pers. (var. decumbens Druce). CC—Lawson ex Ray, Fascic. Brit. (1688) ; Syme, Engl. Bot., ed. 3, 2, (1866) 192 ; Druce, Rep. Bot. Soc. and Exch. Club for 1923, 32 ; Bolam, Trans. Nat. Hist. Soc. Northumb., Durh. and Newc. (1923-26) 84 ; N. D. Simpson ined. Mostly CC but merely extreme of a more or less continuous geo-ecocline—Bocher & Lewis, Biol. Skr. K. Danske Vidensk. Selsk. (1962), 11 (5), 1-25 ; var. LANCASTRIENSE (With.) S. F. Gray. Corolla colour: CS—Syme, Engl. Bot., ed. 3, 2, (1866) 192. Recessive—Sansome, J. Genet. (1936) 361. Stansfield, Gard. Chron. (1928) 208, describes a case of wholesale reversion after forty years of constancy in cultivation. G. ROBERTIANUM subsp. CELTICUM Ostenf. CS—Ostenfeld, Rep. Bot. Soc. and Exch. Club for 1919, 552 ; Baker, Watsonia (1956) 276 ; subsp. MARITIMUM (Bab.) H. G. Bak. CS— Wilmott, J. Bot. (1921) 95 ; Baker, Watsonia (1956) 275. G. PURPUREUM subsp. FORSTERI (Wilmott) H. G. Bak. CC—Pugsley, Rep. Wats. Bot. Exch. Club for 1927-28, 427. CS—Baker, Watsonia (1955) 165 and New Phytol. (1957) 189, ERODIUM MOSCHATUM var. MINUS Rouy. RC—Miller ex White, Proc. Bristol Nat. Soc. (1928) 53. FE. CICUTARIUM subsp. CICUTARIUM and subsp. DUNENSE Andreas. CS—Larsen, Biol. Medd. K. Danske Vidensk. Selsk. (1958), 23 (6), 17-19. OXALIS ACETOSELLA var. CAERULEA Pers. (var. subpurpurascens DC.). CC—Leighton, F/. Shropshire, (1841) 196. CS—Wolley-Dod, FI. Sussex (1937) 100. EUONYMUS EUROPAEUS var. LEUCOCARPUS Druce. RS—Roper, Rep. Bot. Soc. and Exch. Club for 1924, 710. ULEX EUROPAEUS var. STRICTUS (Mackay) Webb (U. hibernicus G. Don ex Loud.). CS— Cameron, Phytol. (1841) 76. CS, but not invariably—Babington ex Syme, Engl. Bot., ed. 3, 3, 5 (1866). SAROTHAMNUS SCOPARIUS subsp. MARITIMUS (Rouy) Ulbr. RS—Mitten ex Reid, Origin Brit. Fl., (1899) 4. CS—Jackson, Gard. Chron. (1939) 387 ; Bécher, Bot. Tidsskr. (1955) 129 ; Adams, Watsonia (1957) 17 ; P. F. Yeo ined.; Alderney material CC—M. C. F. Proctor ined.; Dungeness plants CC but genetically distinct from Lizard and Pem- brokeshire plants—Gill & Walker ex Ford, Ecol. Genetics, (1964) 293. ONONIS REPENS var. HORRIDA Lange. CS—Bennett, Rep. Bot. Exch. Club for 1889, 245. CS, but spines disappeared subsequently—Henslow, Heredity of Acquired Characters in Plants, (1908) 33 (as O. spinosa L.). TRIFOLIUM ARVENSE L. Stature, head-length, etc. genotypic, but described variants merely extremes of clines—Bocher, Larsen & Rahn, Biol. Skr. K. Danske Vidensk. Selsk. (1955), 8 (3), 13, 26. T. REPENS var. CARNEUM S. F. Gray (var. rubescens Sér.). Single dominant gene—ex Turrill, Brit. Plant Life, (1948) 206. For genetics of other, unnamed variants see Daday, Heredity (1954) 61-72. Watsonia 6 (4), 1966. 210 D. E. ALLEN LOTUS CORNICULATUS var. HIRSUTUS Koch. CC—Larsen, Bot. Tidsskr. (1954) 206 ; var. CRASSIFOLIUS Pers. CC, only with respect to fleshiness of leaves—Larsen, Bot. Tidsskr. (1954) 206. ‘ONOBRYCHIS VICIIFOLIA Scop. Indigenous chalk-down form : CC—Grose, Fl. Wilts. (1957) 206. VICIA SYLVATICA var. CONDENSATA Druce. CS—Druce, J. Bot. (1911) 235. LATHYRUS MONTANUS var. TENUIFOLIUS (Roth) Garcke. CC—Don, Mem. Wernerian Nat. Hist. Soc. (1821) 302. RS—Koch, Syn. Fl. Germ., ed. 3, (1857) 176. CS—Anon. ex Druce, Rep. Bot. Soc. and Exch. Club for 1915, 337. FILIPENDULA ULMARIA var. DENUDATA (Hayne). CC—Beeby, Rep. Bot. Exch. Club for 1887, 170 ; Yapp, Ann. Bot. (1912) 840. CS—Druce, New Phytol. (1911) 311. RUBUS IDAEUS f. OBTUSIFOLIUS (Willd.) W. Wats. (var. obtusifolius (Willd.) Focke. var. anomalus Arrhen). Single recessive gene—Crane & Lawrence, J. Genet. (1931) 254 ; var. ASPERRIMUS E. Lees. Single dominant gene—Crane & Lawrence, Genetics of Garden Plants, (1938) 130. POTENTILLA ANSERINA var. SERICEA (Hayne) Koch. RC—Trail, Ann. Scot. Nat. Hist. (1906) 38. P. REPTANS var. MICROPHYLLA Tratt. CC—Anon. ex Druce, F/. Berks., (1897) 192. RC— Fraser, Rep. Bot. Soc. and Exch. Club for 1926, 260. FRAGARIA VESCA var. BERCHERIENSIS (Druce) Druce. CC—Von Solms Laubach ex Druce, J. Bot. (1910) 198 ; f. ALBA (Ehrh.) Rydb. (var. /ewcocarpa Druce). Recessive—Richard- son, J. Genet. (1923) 152. CRATAEGUS MONOGYNA var. XANTHOCARPA Lange (var. aurea (Hort. ex Loud.) Druce). CS—Druce, Rep. Bot. Soc. and Exch. Club for 1911, 91. SEDUM TELEPHIUM var. FABARIA (Koch) Kirschleger. CC—Britton, J. Bot. (1932) 317. PARNASSIA PALUSTRIS var. CONDENSATA Travis & Wheldon. CC—Graebner ex Druce, New Phytol. (1912) 359 ; Travis & Wheldon, J. Bot. (1913) 88. RC—Salisbury, Downs and Dunes (1952) 278. PEPLIS PORTULA var. LONGIDENTATA Gay. CS—Gay ex Allen, Watsonia (1954) 86. EPILOBIUM HIRSUTUM var. VILLOSISSIMUM Koch. CS—Compton, Rep. Bot. Soc. and Exch. Club for 1913, 469. E. TETRAGONUM f, STENOPHYLLUM Hausskn. CS—Druce, Rep. Bot. Soc. and Exch. Club for 1913, 469. E. PALUSTRE var. LAVANDULAEFOLIUM Lec. & Lam. CC—Marshall, J. Bot. (1910) 122. E. ANGUSTIFOLIUM var. BRACHYCARPUM Leight. CC—Syme, Engl. Bot., ed. 3, (1866) 4, 9. CS—Ockendon & van den Driessche, Proc. Bot. Soc. Br. Is. (1963) 172. HEDERA HELIX var. ARBORESCENS Loud. RS—Turesson, Bot. Not. (1961) 443. ANTHRISCUS SYLVESTRIS (L.) Hoffm. All leaf forms CS but form a continuous clinical series—Petersen, Dansk Bot. Arkiv. (1915) 1 (6), 149. Dissected-leaved extreme (Chaero- phyllum sylvestre var. angustisectum Druce). CS—Hall, J. Bot. (1928) 25. PIMPINELLA SAXIFRAGA Var. DISSECTA (Retz) Spreng. CS, but only one extreme of a topocline —Peterson, Bot. Tidsskr. (1921) 240. AETHUSA CYNAPIUM var. AGRESTIS Wallr. RS—Hoffmann, Bor. Zeit. (1878) 273. CS, though other dwarf plants that might pass under this name RS—Weimarck, Bot. Not. (1945) 351-380 ; var. GIGANTEA Lej. (var. elata (Friedl.) Gaud.). CS—-Weimarck, Bot. Not. (1945) 380. HERACLEUM SPHONDYLIUM var. ANGUSTIFOLIUM Huds. RC—Hudson, Fl. Anglica, ed. 2, (1778) 117 ; Bagnall, F/. Warwicks., (1891) 128. CS—Green ex Grose, Fl. Wilts., (1957) 304. DAUCUS CAROTA subsp. GUMMIFER Hook f. CS except for stature—J. E. Lousley ined.; Nehou, Bull. Soc. Sci. Bretagne (1961) 97-98. MERCURIALIS PERENNIS var. SALISBURYANA Mukerji. CS—Mukerji, J. Ecol. (1936) 42. EUPHORBIA EXIGUA var. RETUSA Roth. CS—Dunn, Rep. Rugby Sch. Nat. Hist. Soc. (1897) 43. Watsonia 6 (4), 1966. INFRASPECIFIC TAXA TESTED IN CULTIVATION ZL POLYGONUM AMPHIBIUM var. AQUATICUM Wallr. and var. TERRESTRE (Leyss.) Boenn. RC— Massart, Bull. Jard. Bot. Bruxelles (1902) 76. P. LAPATHIFOLIUM L. Numerous variants, named and unnamed, CS—Danser, Rec. Trav. Bot. Neerl. (1921) 125-210 ; Timson, Watsonia (1963) 387. RUMEX CRISPUS var. ULIGINOSUS Le Gall (var. elongatus Trimen & Ley, non Guss.). CC— Drabble, Rep. Bot. Soc. and Exch. Club for 1931, 753 ; Lousley, J. Bot. (1935) 256-260. Most trivial variants of this species CS—Lousley, Rep. Bot. Soc. and Exch. Club for 1941-42, 556. URTICA PILULIFERA var. DODARTII (L.) Aschers. CS—Bickham, Rep. Wats. Bot. Exch. Club for 1905-06, 63. Single recessive gene—Crane in Huxley, New Systematics (1940) 531. QUERCUS ROBUR var. CRISTATA A. Henry. CS—Yates ex Grose, FI. Wilts., (1957) 504. SALIX REPENS L. Leaf and habit variants. RC—Howitt & Howitt, F/. Notts., (1963) 184. CALLUNA VULGARIS var. ERIKAE Aschers. CC—Graebner ex Druce, New Phytol. (1911) 315. ERICA TETRALIX var. FISSA Druce. RC—Hall, J. Bot. (1928) 25 ; Rayner, Suppl. Fl. Hamp- shire (1929) 67. E. VAGANS var. KEVERNENSIS Turrill. RC—Turrill, Kew Bull. (1922) 176. VACCINIUM VITIS-IDAEA var. PUMILUM Hornem. (subsp. minus (Lodd.) Hultén). CC— Nicholson ex Beeby, Scot. Nat. (1890) 215. PRIMULA FARINOSA var. LITTORALIS H.-Harrison. CS—Heslop-Harrison, Vasculum (1921) 22. P. VULGARIS f. CAULESCENS (Koch) Schinz & Thell. RC—De Vries, Species and Varieties (1906) 635. CC—White, FI. Bristol, (1912) 493. ANAGALLIS ARVENSIS var. AZUREA Wilmott. RS—Grindon, Phytol. (1842) 130 ; Henslow, Phytol. (1849) 651 ; var. PALLIDA Hook. f. CS—Pugsley, Rep. Wats. Bot. Exch. Club for 1927-28, 435. Corolla colour forms : scarlet dominant over pink, pink over white, white over purple, purple over blue—Marsden-Jones & Weiss, Proc. Linn. Soc. London (1938) 146-154. Allen, Proc. Bot. Soc. Br. Is. (1954) 157, cites data suggestive of linkage in these colour forms. Var. VERTICILLATA Diard. CS—Diard, Cat. Pl. Rais. Pas-de-Calais (1852) 130. PULMONARIA OFFICINALIS var. IMMACULATA Opiz (subsp. obscura (Dumort.) Murb.) CC— Marshall, J. Bot. (1908) 288. LITHOSPERMUM OFFICINALE var. PSEUDOLATIFOLIUM C, E. Salmon (var. /atifolium Gremli?). CC—Salmon, J. Bot. (1906) 367 and Rep. Bot. Exch. Club for 1908, 390. CS—Marshall, Rep. Bot. Soc. and Exch. Club for 1910, 579. HYOSCYAMUS NIGER var. PALLIDUS (Waldst. & Kit. ex Willd.) Fries. Recessive—De Vries, Species and Varieties, (1906) 283. CS—Jackson, J. Bot. (1910) 283. SOLANUM DULCAMARA var. MARINUM Bab. CC—Turesson, Hereditas (1922) 107, 229-231. CS—Marsden-Jones & Turrill, J. Ecol. (1945) 80. S. NIGRUM var. FLAVUM Dumort. (var. chlorocarpum (Spenn.) Boiss.; S. /uteovirescens C. C. Gmel.) Single recessive gene—De Vries, Species and Varieties, (1906) 282, 298. CS—Britton, Rep. Bot. Soc. and Exch. Club for 1935, 94. DATURA STRAMONIUM var. TATULA (L.) Torr. Single dominant gene—Saunders in Bateson & Sanders, Rep. Evol. Comm. Roy. Soc. (1902) 21-32, Safford, J. Hered. (1921) 178- 190 ; var. INERMIS (Jacq.). Single recessive gene—Naudin, Nouv. Arch. Mus. (1865) 41; Saunders in Bateson & Saunders l.c. LINARIA VULGARIS var. LATIFOLIA Bab. RC—Watson, Cybele Brit., 3, (1852) 473. CHAENORHINUM MINUS var. PRAETERMISSUM (Delastre) Rouy. Single dominant gene— Champagnat, Bull. Soc. Bot. Fr. (1952) 301-304. SCROPHULARIA NODOSA var. PALLESCENS D6ll (var. bobartii Pryor). CC—Baxter, Brit. Phaenog. Bot., (1835) 385. DIGITALIS PURPUREA Var. NUDICAULIS Saunders. Single dominant gene—Saunders, J. Genet. (1918) 215-228. VERONICA OFFICINALIS var. INTEGRA (Druce) Druce. CC—Smith, Engl. FI., (1824) 1, 23 ; Druce, Ann. Scot. Nat. Hist. (1911) 170. Differences in leaf shape and pubescence CS Watsonia 6 (4), 1966. 212 D, BEVALLEN but form a more or less continuous geo-ecocline—Bécher, Dansk Bot. Arkiv (1944), 11 (7), 1-20. . SERPYLLIFOLIA Subsp. SERPYLLIFOLIA var. OBSCURA Drabble & Salmon. CC—Drabble & Salmon, J. Bot. (1929) 213 ; subsp. HUMIFUSA (Dickson) Syme. CC—Marshall, Rep. Bot. Soc. and Exch. Club for 1917, 235. . ARVENSIS var. NANA (Lam.) Poir. (var. eximia Towns.). CS—Townsend, F/. Hampshire, ed. 1, (1883) 232. CC—Druce, FI. Berks. (1897) 377. . PERSICA var. ASCHERSONIANA (Lehm.) C. E. Britton and var. CORRENSIANA (Lehm.) C. E. Britton. CS—Lehmann, Osterr. Bot. Zeitschr. (1909) 255 . POLITA var. POLITA (subsp. /udwigiana Lehm.) and var. THELLUNGIANA (Lehm.) Hayek & Hegi. CS—Lehmann, Osterr. Bot. Zeitschr. (1909) 258. . AGRESTIS var. MICRANTHA Drabble. CS—Drabble, J. Bot. (1926) 25 and Rep. Bot. Soc. and Exch. Club for 1932, 442. MENTHA PULEGIUM Var. ERECTA Mill. RC—Fraser ex Salmon, F/. Surrey, 515 (1931). ORIGANUM VULGARE Var. VIRENS Gren. & Godr. CS—Larsen, Biol. Skr. K. Danske Vidensk. Selsk. (1960), 11 (3), 42. BETONICA OFFICINALIS var. NANA (Druce). CC—Druce, Rep. Bot. Soc. and Exch. Club for 1918, 299 ; C. D. Pigott ined. LAMIUM PURPUREUM Var. LUMBII Druce. CS—Lumb, Rep. Bot. Soc. and Exch. Club for 1914, 156. GALEOPSIS TETRAHIT var. NIGRICANS Bréb. RS—Britton, Rep. Bot. Soc. and Exch. Club for 1927, 584 and J. Bot. (1928) 44. TEUCRIUM SCORODONIUM var. ACROTOMUM L. B. Hall. CC—Hall, Rep. Bot. Soc. and Exch. Club for 1929, 236. PLANTAGO MAJOR var. ROSEA Boenn. (var. bracteata Druce). CS—White, F/. Bristol (1912) 501 ; Orme ex Druce, Rep. Bot. Soc. and Exch. Club for 1928, 634 ; var. MINOR Roth (var. minima (DC.) Schlecht., var. intermedia auct.) RS—Schultz, Arch. Fl. Fr. Allem. (1869) 376. RC—Marsden-Jones & Turrill, J. Ecol. (1930) 372-375 ; var. MAXIMA Schlecht. CS, perhaps recessive—Griffiths, J. Bot. (1922) 230; var. SCOPULORUM Fries (subsp. intermedia (Godr.) Lange). CC—Boreau, Fi. Centre Fr., ed. 3, (1857) 2, 537 ; Turesson, Hereditas (1925) 160 ; Koch, Ber. Schweiz. Bot. Ges. (1928) 45-61. P. LANCEOLATA var. ANTHOVIRIDIS W. Wats. CS—Watson, J. Bot. (1937) 231-233. var. CAPITATA Schum. (var. sphaerostachya Mert. & Koch var. lanuginosa Boenn.; P. erio- phora Hoffmgg. & Link ; var. capitellata Sonder ex Koch). RC—Lawson, Bot. Gaz. (1849) 35 ; Druce, Rep. Bot. Soc. and Exch. Club for 1910, 583 and J. Bot. (1911) 235. RS—Griffiths, J. Bot. (1922) 229. CS as one extreme of an ecocline—Bécher, Dansk Bot. Arkiv (1943), 3, 1-18. CC—Nehou, Bull Soc. Sci. Bretagne (1955) 113-133 ; var. ALTISSIMA Schum. (var. sylvatica Pers.; var. major Syme). CS as one extreme of an ecocline—Bocher, l.c. P. MARITIMA var. LINEARIS Syme. CC—Hosking ex Davey, Fil. Cornwall, (1909) 372 and ex Williams, Prodr. Fl. Brit. (1910) 363 ; var. PUNCTATA Neum. Single dominant gene— Gregor, New Phytol. (1938) 31. Other variants CS, but part of more or less continuous ecoclines—Gregor, New Phytol. (1938) 15-49, (1939) 293-322. P. CORONOPUS var. MARITIMA Gren. & Godr. (var. ceratophyllon Rapin). CS—Wheldon, Rep. Bot. Soc. and Exch. Club for 1915, 366; Turrillin Wilmott, Brit. Fl. Plants and Modern Syst. Methods, (1949) 18. CC—M. E. Gillham ined ; var. PYGMAEA Lange (Astero- geum laciniatum var. capitatum 8. F. Gray). RC—Salisbury ex Moss, Rep. Wats. Bot. Exch. Club for 1914-15, 505 and in Huxley, New Systematics, (1940) 333 ; Watson ex Thurston & Vigurs, Suppl. Fl. Cornwall, (1922) 114. RS to some extent—Gravis, Bull. Biol. Fr. Belg. (1932) Suppl. xiv, 66. Some populations RS, others CS, forming one extreme of an ecocline—Bécher, Larsen & Rahn, Hereditas (1955) 440-442. CAMPANULA GLOMERATA var. MINOR S. F. Gray (var. nana C. Bailey). OSS t om Phytol. (1849) 527 ; Druce, Fl. Berks., (1897) 325. ; 0 E474 x E493 10 14 0 5) 0 nemorosa X rostkoviana E474 x E520 a — — Dy 10 3 pseudokerneri < anglica E42 x E76 1 0 0 — — — — — — 2 0 0 EI B16 Watsonia 6 (4), 1966. 220 PAE YEO The type of bad seed produced depended on whether the diploid or the tetraploid was the ovule parent. In the first case, the capsules developed normally and dehisced, though sometimes rather belatedly. As a rule they contained normal numbers of full-sized but empty, or nearly empty, testas. When the ovule parent was tetraploid, full-sized empty seeds were not produced. Many of the capsules were undeveloped, and doubtless contained minute empty ovules, which were seen in a number of cases though not usually counted. Occasionally, partially developed seeds were found; these had considerably undersized testas but some development of contents, though much less than in a normal seed, even when it nearly filled the testa. Usually several such seeds occurred in the same capsule which was itself partially developed (for example, the 24 bad seeds of E. nemorosa x E. anglica, E474 x E493, were produced by only three of the fifteen flowers pollinated). Seeds of this type were formed in all the crosses shown in the lower part of Table 3 except the two between EF. pseudokerneri and E. anglica (they also appear in a number of later tables under the heading ‘partly filled’). A possible interpretation of these results is that, when the diploid is female, fertilization takes place readily and seed-development begins; but although the testa reaches full size there is practically no development of the embryo and endosperm. The initiation of seed- development seems to stimulate the development of the capsule. When the tetraploid is the female, fertilization perhaps takes place more rarely, but when it does the development of seed-contents goes further than in the reverse cross. Perhaps, therefore, there is a better chance of getting a triploid from the cross with the tetraploid as female, although at first the reverse type of cross seemed more promising. All the apparently good seeds from these pollinations were sown. Where selfing was thought to have been impossible one seed (of the cross E. anglica < E. nemorosa, E180 x E196) germinated but the seedling died at an early stage. A total of 58 seeds was sown where there was a possibility of selfing; of these 14 germinated and ten plants were raised, all of them like the seed-parents, and presumably resulting from accidental self-pollination. This failure to obtain hybrids between diploid and tetraploid species is not surprising in view of the fact that I have only once found a triploid in nature; this was a hybrid between E. anglica and E. micrantha (Yeo 1954, 1956). The seed production in crosses between parents alike in chromosome number is given in Table 4. There are two types of cross here, namely those between two populations of the same species (E. anglica) and those between different species. The total seed production was moderate, and so also was the production of good seed, since only about one-ninth of all seeds were bad. Most of the bad seeds that were produced were the minute empty ones described previously, but in the cross E. occidentalis x E. salisburgensis var. hibernica, E192 x E230, two of the bad seeds were half-filled and large; in the similar cross E. nemorosa X E. salis- burgensis, E608 x E515, there was one large empty seed in a capsule in which the good seeds varied in size and shape; and in the reverse cross between these species, E515 x E608, one bad seed had contents but was only about half the normal thickness, while some of the seeds classed as good in the same capsule were also rather thin. In E. anglica x E. anglica, E663 x E649, four of the total of ten bad seeds were large empty ones. In the cross E. pseudokerneri x E. micrantha, E609 1239 (Mean per capsule = 8-2) (Mean per capsule = 8-3) *Some other seeds lost, or probably lost. tFrom Germany. tFrom France. TABLE 5. Germination of hybrid seed, and plants raised from it. (Parents alike in chromosome number) No. of No. of Plants raised seeds sown seedlings hybrid non-hybrid Inter-population crosses of FE. anglica 139 30 DS 0 Interspecific crosses within Subsection Ciliatae 203 45 28 0 Interspecific crosses between Subsections Ciliatae and Angustifoliae 159 30 12 1 was the interpretation quite clear. No definite irregularities were seen, and it is possible that meiosis was normal in this hybrid. It seems clear that, if there were any irregularities, they could not have involved more than one bivalent. E. pseudokerneri < E. scottica, E42 x E120. About 65 cells derived from three plants were examined. The preparations were slightly better than those of the preceding hybrid. It was rarely possible to get a complete interpretation of a cell, but it seems that meiosis may occasionally have been normal, as it was not always possible to detect irregularities. Usually, however, univalents were visible at Metaphase I (Fig. 1), most frequently two in number, and six at the most. Very little was seen of other stages of meiosis, but in one Anaphase II a group clearly consisting of 22 chromosomes was counted, and at Telophase Watsonia 6 (4), 1966. ~ 222 Po FY EO II it was noted in one cell that two chromosomes had not quite joined their nearest re- forming nuclei, while in another more advanced cell there were no stray chromosomes. © SED oO ‘ § a Igo é : O Fig. 1. Metaphase I of meiosis in the F, hybrid Euphrasia pseudokerneri Xx E. scottica, E42 x E120, showing 20 bivalents and 4 univalents. x 1100. E. salisburgensis var. hibernica x E. occidentalis, E230 x E192. Notes were made on about 110 cells derived from two plants. Most of the preparations were good, and definite interpretations of some cells were obtained. All stages of meiosis were seen. There were always many univalents present at Metaphase I (Fig. 2); these were counted as accurately as possible, and the results for 40 cells were as follows: 1 cell had 12 univalents 4 cells had 16 on Gate fe ke a Ones 20 x LO ree) Ks GF flee Mes sy 39) h: lcelim es . Where an odd number of univalents was visible, one was added to the count, on the assump- tion that they always occur in pairs. When the chromosomes resulting from the disjunction of bivalents were congregating at the poles at late Anaphase I, some of the univalents came into the equatorial region and divided (Fig. 3), while others were to be seen in the peripheral region and probably did not divide. Finally, all chromosomes joined one or other of the Telophase nuclei. Fig. 2. Metaphase I of meiosis in the F, hybrid Euphrasia salisburgensis var. hibernica < E. occidentalis, F230 x E192, showing 12 bivalents and 20 univalents (mostly carried away from the equatorial region in squashing the cell). x 1100. Watsonia 6 (4), 1966. —————— es BREEDING RELATIONSHIPS OF EUPHRASIAE 223 Fig. 3. Anaphase I of meiosis in E230 x E192. Univalents dividing in the equatorial region; the rather indefinite outlines which the univalents showed are not indicated by this method of drawing. x 1100. At Metaphase II the chromosomes that divided at the first division of meiosis did not line up on the equator but remained scattered and appeared as laggards at Anaphase II (Fig. 4). However, nearly all the laggards eventually joined a nucleus (Fig. 5). Such an irregular meiosis can rarely lead to the production of a pollen grain with a normal set of chromosomes, and any grains which contain 22 chromosomes may well lack some members of the normal set and possess others in duplicate. Some counts were made at Anaphase II and Telophase II which confirmed the inconstancy of the final chromosome numbers. In one cell there were probably 24, 21, 22, and 20 chromosomes in the four nuclei, plus one stray chromosome; other chromosome counts for individual nuclei were 16 or 17, 18, 22 or 23, 24, and for cells at Anaphase II, in which the nuclei could only be counted in pairs, Al + 47, 39 + 49, 36 + 52,40 + 48, 42 + 46. Fig. 4. Anaphase II of meiosis in E230 x E192. (As the figure shows 89 bodies, it may be that some other matter had been mistaken for a chromosome.) x 1100. Production of normal pollen in the hybrids Pollen counts of hybrids and their parents were obtained by breaking up the anthers in cotton-blue in lactophenol. The numbers of good and bad grains were counted, the good ones being rounded and darkly stained, and the bad ranging from minute to fairly large without contents, or having darkly stained contents but being markedly undersized. These counts were made during July and August 1954; on 6 October 1954, it was discovered Watsonia 6 (4), 1966. 224 PUP YEO @ 385 a v6 EE O€§. re) (e) Guns O D%o 0 ire 2 2} Fig. 5. Late Anaphase II of meiosis in E230 x E192. Nearly all the chromosomes have joined one of the re-forming nuclei. x 1100. TABLE 6. Production of normal pollen in species and hybrids of Euphrasia. Production of normal pollen (approx.) per cent. Parents pseudokerneri E42 x E70 100 E226 95 occidentalis E192 ? 100 salisburgensis var. hibernica E230 ? 100 Cross within one species anglica x anglica E76 x E168 100 Crosses within one Series pseudokerneri < nemorosa E42 « E154 ? 100 pseudokerneri X< occidentalis E226 x E192 ? 90-95 Crosses between Series pseudokerneri < brevipila E70 « E81 85 pseudokerneri * scottica E42 «x E120 (ols) brevipila x scottica E81 x E120 85 micrantha < confusa E185A x E183C 85 Cross between Subsections salisburgensis var. hibernica x occidentalis E230 x E192 17-24 that all the counts had been lost and all that remained was a list of pollen samples taken. However, I could remember most of the counts approximately and at once wrote them down. They are given in Table 6. In the hybrid E. pseudokerneri « E. scottica pollen counts were first made from three plants; two of them produced about 70 per cent of normal pollen, and the third produced considerably more. Additional pollen counts were made on anthers collected about two weeks later from two of these plants in order to see whether the discrepancy would be repeated. Both plants gave very much lower proportions of normal pollen than before. Watsonia 6 (4), 1966. BREEDING RELATIONSHIPS OF EUPHRASIAE 225 It seemed, therefore, that the production of normal pollen was lower at the later date. If this was so, pollen fertility must fluctuate and one must not expect to assign fixed values to the production of normal pollen in hybrids. Probably, counts showing little or no bad pollen in plants that are not hybrids are more reliable. Table 6 shows a fairly good corres- pondence between pollen fertility and the affinity of the hybrid’s parents although, in view of what has been said, such a conclusion must be regarded as tentative. Finally, it should be pointed out that these counts are of visually normal pollen, and that the correlation between this and physiologically good pollen is not known. TABLE 7. Seed production and germination of seed of F, hybrids when selfed, open-pollinated or back- crossed, 1954-1958. Seeds produced No. of ——__________—_———._ No. of Pareniace lof hivbrid Polli- plants No. of | Large good Germi- nation of the capsules Small empty, seeds per nation ‘ Good hybrid empty or partly capsule filled Species in same Series pseudokerneri * nemorosa E609 x E608 selfed 1 2 15 5 0 ee) + E609 x E608 x E609 2 1 4 24 3 1 6 aim pseudokerneri xX occidentalis E226 x E192 selfed 1 6 66 1 3 11 a anglica x hirtella E616 x E559 selfed 1 23 57 70 7 2 or E616 x E559 x E6163 1 1 5 0 5 5 oF Species in different Series, same Subsection micrantha xX confusa E185 x E183 selfed 1 28 227 5 7 8 — pseudokerneri x brevipila E70 x E81 selfed 1 8 85 2 3 10-5 — E609 x E623 selfed Dy 17 174 19 5) 10 aie E609 x E623 x E6992 1 4 45 Z, 0 11:3 a (pseudokerneri) pseudokerneri X scottica E42 x E120 selfed 3 13 100 34 18 35) ~ pseudokerneri X micrantha E609 x E185 selfed i 17 100 99 4 6 ain E609 x E185 x E185 Jd 1 4 23 24 1 6 sie E609 x E185 x E609 2 1 4 18 26 0 4°5 ar E609 x E185 x E699 2 1 1 9 2 0 9 —_ Species in different Subsections salisburgensis var. hibernica X occidentalis E230 x E192 selfed 2 13 8 4 2 146 17 2 48 23 E608 x ‘F,’ plant 7 open-poll. 1959 D5 Dilly ys 97 5 X N seb . ? %, kh yy ea Sethi tiles 2+ \ 5 i} oF “ ie r | 8 : “ (L © e 3 > | aes ‘eo 4 te & | & @@ ou, ee 2 ) % & @& @©¢ | 9 2 % 0,» % a s. oe 4 | y% 8 a ” es 72, i : $ t*s2 6 2 % a - °,7 |) be pono, b ss 8 { g PE La “gta é WS, ~f bp Fig. 2. Diakinesis in aneuploid populations of S. bermudiana L. x 500. A. Tolland State Forest, Massachusetts. 27 = 84. a B. Clinch River Valley, Tennessee. 2n = 88. C. Near Wartburg, Morgan Co., Tennessee. 2n = 90. | D. Chickahominy River Bridge, Charles City Co., Virginia. 2n = 90. Watsonia 6 (5), 1967. { IRISH POPULATIONS OF SIS YRINCHIUM 287 No populations of this species having 2n = 96 were recorded in the course of this investi- gation, but a count from Louisiana by Oliver & Lewis (1962) gave this number. It would therefore seem reasonable to assume that the counts recorded form an aneuploid series derived from the dodecaploid number 2” = 96. The variation in chromosome number could not be correlated with observed morphological variability. S. montanum Greene var. crebrum Fernald. 2n = 96 Springfield Reservoir, Massachusetts. 2n = 96 Near Bondville, Vermont. 2n = 96 South of Lake Catherine, Vermont. Bowden (1945) also reported 2n = 96 for this species. S. mucronatum Michx. 2n = 32 Chapel Hill, North Carolina. 2n = 30 East of Chyde, Haywood Co., N. Carolina. 2n = 32 Magruder, Virginia. 2n = 32 Campus of Lynchburg University, Virginia. It was unfortunately impossible to obtain counts from more northerly plants, but Oliver & Lewis (1962) record 2n = 32 for material of this species from Canada. Hybridization between S. mucronatum and the other two species proved impossible. No seeds were ever produced. Hybridization between the two dodecaploid species was possible, however, and hybrids were formed which were completely sterile, usually pro- ducing aborted capsules, and never yielding any seed. Meiosis in these hybrids was highly irregular, giving a maximum of 14 bivalents. This could be taken to indicate that the two species had only about one third of their chromosomes in common. In spite of these wide cytological differences, there is a superficial morphological resemblance between these three species. Obviously, this does not indicate a close relation- ship. SISYRINCHIUM IN IRELAND Morphologically, plants from various parts of Ireland are very uniform when grown under similar conditions in an experimental greenhouse. It would seem therefore that the apparent variability in branching shown in herbarium specimens, and commented on by Northfield (1961), is caused by varying conditions of grazing and land use. Cytologically too, they show no variation, as the following counts indicate (Fig. 3). 2n = 88 Cloonee Loughs, Co. Kerry. 2n = 88 Caragh Lough, Co. Kerry. 2n = 88 Woodford River, Co. Galway. 2n = 88 Killybegs, Co. Donegal. my 84 @ of 37 af Wee 2 atv 2 4 “ wat Ys ei oy" Bh, 3 fel Fig. 3. S. bermudiana L. Irish populations. 2n = 88. x 670. A. Metaphase. B. Early anaphase, Woodford River, Co. Galway. These counts were each based on two preparations of two plants, and confirm the count made by Northfield (1961), though they conflict with that of Léve & Léve (1958), who record 2n = 64 as the number for the Cloonee Loughs population. Watsonia 6 (5), 1967. 2A 288 RUTH INGRAM The morphological comparison of the Irish and the American species is summarized in Table 1. It is obvious that the Irish population is hardly if at all distinguishable from the American species S. bermudiana L., and the results of the present investigation do not justify the proposal by Love & Love (1958) that the Irish plant should be designated as a new species, S. hibernicum. Cytologically too, the Irish plant fits into the aneuploid series exhibited by the American species. The Irish plant must therefore be referred to the species S. bermudiana L. SISYRINCHIUM IN GREENLAND It is interesting to note that S. bermudiana L. is not the species described by Bécher (1948) as occurring in Greenland. Bocher refers to the Greenland species as S. montanum Greene, which is a species characteristic of the extreme north of the American continent. The plant illustrated by Bécher is simple stemmed, and is described as having light blue flowers. The chromosome number of this plant is 2n = 32 (Boécher & Larsen 1950). Whatever may be its true taxonomic position, it is obviously not the same as the Irish species, nor is it S. montanum var. crebrum. SISYRINCHIUM IN EUROPE The status and extent of Sisyrinchium in Europe is a topic beyond the scope of this work, but it may be pointed out that the plant illustrated by Hegi (1939, Fig. 631) is a simple stemmed species, closely resembling S. montanum var. crebrum. Moreover, this seems to be the species most commonly cultivated as a garden plant in Europe. Certainly, observation of garden material in the British Isles has shown that although a great variety of specific names are in use, in fact most of the garden material can be referred to the taxon S. montanum var. crebrum. Although the names S. bermudiana or S. bermudianum are frequently used, I have never seen S. bermudiana L. (sensu stricto) cultivated, and no material referable to this taxon has been received from any British or continental Botanic Garden. Indeed this species has little horticultural merit, so perhaps its absence is not surprising. THE STATUS OF SISYRINCHIUM BERMUDIANA IN IRELAND The question of whether Sisyrinchium is native in Ireland, or whether it has been recently introduced, is one that was mentioned in the introduction. Although the present investigation does not justify the drawing of any firm conclusion on this question, the following points have some relevance. (1) Morphologically and cytologically the Irish and American populations of S. bermudiana are indistinguishable, although only one chromosome type is present in Ireland, whereas a wide aneuploid series is present in America. This would be compatible with the theory that S. bermudiana in Ireland originated with the introduction of a single biotype. No breeding difficulties would arise if this were the case, as the species is self-compatible, and is fairly often self-pollinated. However, the possibility that the Irish population repre- sents an extreme reduction of a previously wider distribution cannot be ruled out. (2) S. bermudiana is unlikely to have arisen as a garden escape in Ireland, as it often occurs well away from habitation, and is not likely to have been cultivated anyway. (3) If Sisyrinchium was introduced by birds, Greenland is unlikely to have been a ‘half way house’ for the species, as suggested by Heslop-Harrison (1952), since the Greenland plant is a different species from the Irish one, which has never been recorded there. (4) Evidence which will be presented in a later paper indicates that the Irish plants are unlikely to be a recent introduction, because the Irish and American populations of S. bermudiana show cryptic genetic differences which it is difficult to imagine are of recent development. On a balance of probabilities, it may provisionally be suggested that the Irish population originated directly though not very recently from the American population of S. bermudiana. Watsonia 6 (5), 1967. | | IRISH POPULATIONS OF S/S YRINCHIUM 289 ACKNOWLEDGMENTS I wish to thank Dr. J. L. Crosby for his supervision of this work and his constructive criticism of this manuscript. I am grateful also to the technical staff of the Botany Depart- ment, Durham University, and to all those in America who made my visit there worthwhile. The work was carried out whilst in receipt of a N.A.T.O. studentship, which was held in the Botany Department of Durham University. REFERENCES BOcHER, T. W. (1948). Contributions to the Flora and Plant Geography of W. Greenland. 1. Selaginella rupestris and Sisyrinchium montanum. Medd. om Gronland, 147 (3), 12-14. BOcHER, T. W. & LARSEN, K. (1950). Chromosome numbers of some Arctic or Boreal Plants. Medd. om Gronland, 147 (6), 12. BowDEN, W. M. (1945). A List of Chromosome Numbers in Higher Plants. Am. J. Bot., 32, 81-92. FERNALD, M. L. (1946). Identifications and Re-identifications of N. American plants. Rhodora, 48, 137-162. FERNALD, M. L. (1950). Gray’s Manual of Botany. Ed. 8. American Book Co., New York. Hecai, G. (1939). Illustrierte Flora von Mittel-Europa, 2. Carl Hanser Verlag, Munchen. HESLOP-HARRISON, J. (1952). The North American and Lusitanian elements in the Flora of the British Isles. in The Changing Flora of Britain, Ed. J. E. Lousley. B.S.B.I., London. INGRAM, R. (1964). The taxonomy and cytology of the genus Sisyrinchium. Ph.D. Thesis, University of Durham. LAUTERBORN, R. (1927). Mitteil. des badisch. Landesvereins f. Naturkunde usw., N.F. 2, Quoted from Hegi (1939). Love, A & Love, D. (1958). The American element in the Flora of the British Isles. Bot. Not., 111, 376-388. NORTHFIELD, J. M. (1961). ‘The Cytotaxonomy of the genus Sisyrinchium.’ M.Sc. Thesis, University of Durham. OLiver, R. L. & Lewis, W. H. (1962). Chromosome numbers of Sisyrinchium in Eastern North America. Sida, 1 (1), 43-48. PRAEGER, R. L. (1934). The Botanist in Ireland. Hodges, Figgis and Co., Dublin. SHINNERS, L. H. (1957). Sisyrinchium bermudiana L. instead of S. angustifolium Mill. Rhodora, 59, 159-160. Watsonia 6 (5), 1967. CLEISTOGAMY IN SPARTINA By J. C. E. HUBBARD The Nature Conservancy, Furzebrook, Wareham, Dorset ABSTRACT Cleistogamous spikes enclosed in the leaf sheaths below the terminal inflorescence have been observed in the forms of Spartina < townsendii (sensu lato) at many localities in southern England; their structure and position are briefly described. Seed has been found in such spikes in only a few isolated cases. It is first necessary to clarify the nomenclature of Spartina townsendii (sensu lato) to be used in this account. Following the recent cytological investigation by Marchant (1963, 1964) and in accordance with Hubbard (1965) the following names are used: Spartina < townsendii (2n = 62), Spartina X (2n = 120 or more), Spartina 2n = 90 and Spartina 2n=76. In August 1963 several unusual specimens of Spartina X were collected from marshland adjoining the north-west of Brownsea Island, Poole Harbour, Dorset. They possessed a single cleistogamous spike arising immediately above the upper node and enclosed by the leaf sheath. The spike was independent of the terminal inflorescence and 25 cm below it. On revisiting the area in July 1964 many examples were found and a search along the shoreline marshes of the harbour revealed this phenomenon to be a common occurrence. Subsequently numerous cleistogamous specimens were discovered in every locality visited between Poole and Pagham on the south coast and at Bridgwater Bay and Sand Bay in the Severn Estuary (22 localities). In addition, inflorescences of Spartina x townsendii bearing cleistogamous spikes were found at twelve of the localities visited, including Hythe in Southampton Water and Sand Bay in the Severn Estuary. The cleistogamous condition also occurred in Spartina 2n = 76 and Spartina 2n = 90 but was not found in populations of Spartina alternifiora Lois. and Spartina glabra Muhl. from Southampton Water or Spartina maritima (Cutt.) Fernald from Hayling Island. The cleistogamous spike of Spartina X and Spartina xX townsendii is structurally similar to the spikes comprising the terminal inflorescence although the portion of the axis below the lowest spikelet may become greatly extended. There appears to be no reduction or modification in the floral structure of the spikelets. In every specimen examined, the upper internode was grooved from panicle to node on the face adjacent to a cleistogamous spike, but cylindrical and more easily broken when the spike was absent. The development of the spike was usually out of phase with that of the terminal inflorescence and it was possible to find a range of developmental stages from 0-1 cm in length to fully- formed ones. The position of the cleistogamous spike relative to the terminal inflorescence varies according to the expansion of both the internode and the axis below the spike, making a range of positions possible. The spike may therefore be 30 cm below the terminal inflores- cence and entirely enclosed by the leaf sheath or occasionally it may emerge and appear to be part of the inflorescence. In the latter case cleistogamous spikes may sometimes be found at the lower nodes and up to three additional spikes have been discovered on a large specimen (Plate 13). Cleistogamy occurs more frequently with increase in size and it is especially evident in the larger plants of Spartina X at Bridgwater Bay where it is possible to find terminal inflorescences containing up to twenty-two spikes. Herbarium specimens of Spartina X, bearing the characteristic groove on the internode below the terminal inflorescence, have been seen in the herbaria of the Royal Botanical Gardens, Kew (K), the British Museum (BM) and in the private collection of Mr. N. D. Simpson. 290 Watsonia 6 (5), 1967. REATE 18 Specimen of Spartina X from Poole Harbour (August 1964). The leaves have been removed to show cleistogamous spikes arising from 3 internodes. The axis of the 2nd spike is fused to the adjacent internode which is damaged at its base. The 3rd spike has failed to develop and is less than 2 mm in length. oa [facing page 290 * a 7 na ul i i i! CLEISTOGAMY IN SPARTINA 291 An abnormality often associated with the presence of a cleistogamous spike is the fusion of its axis to the internode of the terminal inflorescence and in extreme cases the culm is fractured. During elongation of the internode in Spartina x townsendii the extremely thin axis of the cleistogamous spike is frequently broken just above the node and the severed spike is carried up with the terminal inflorescence. Retention of moisture by the leaf sheath may lead to rotting of the mature spike. Specimens have been collected from Bridgwater Bay and Poole Harbour in which two culms bearing separate leaf sheaths have arisen from the upper node. Although hundreds of specimens were examined during 1964, only in a few isolated cases was seed collected from cleistogamous spikes of Spartina X and Spartina 2n = 76. I wish to thank Miss J. Perkins and Miss G. Shannon for their assistance in collecting and processing the specimens. REFERENCES HUBBARD, J. C. E. (1965). Spartina marshes in southern England. VI. Pattern of Invasion in Poole Harbour. J. Ecol. 53, 799-813. MARCHANT, C. (1963). Corrected chromosome numbers for Spartina > townsendii and its parent species. Nature, Lond. 199, 929. MARCHANT, C. (1964). In J. M. Lambert’s ‘The Spartina Story’. Nature, Lond. 204, 1136-1138. Watsonia 6 (5) 1967. 2B TAXONOMIC AND NOMENCLATURAL NOTES ON THE BRITISH FLORA The following notes have been prepared to validate new names and combinations of taxa to appear in the forthcoming Critical Supplement to the Atlas of the British Flora. At the same time some taxonomical notes on taxa not dealt with in the current British literature have been included. Each of the notes should be attributed to the authors whose names appear at the end of each account, unless it is otherwise stated in the text. I am extremely grateful to those authors who have so kindly contributed, and especially to J. E. Dandy for help with the nomenclature. I am also indebted to H. Gilbert Carter for help with the latin descriptions and to W. T. Stearn for checking them. P. D. SELL ATHYRIUM DISTENTIFOLIUM Tausch ex Opiz, Tent. Fl. Crypt. Boem. 1, 14 (1820). var. flexile (Newm.) Jermy, comb. nov. Basionym: Pseudathyrium flexile Newm., Phytol. 4, 974 (1853). Athyrium alpestre var. flexile (Newm.) Milde, Fil. Eur., 53 (1867). Athyrium flexile (Newm.) Druce, Brit. Pl. List, ed. 2, 137 (1928). This plant requires further study; it appears to have a wider ecological range than var. distentifolium and is likely to be more common than at present thought. The much- branched rhizome gives rise to narrow fronds of a delicate texture mostly growing prostrate along the ground, even in sheltered situations. The lower pinnae are characteristically reflexed or drooping. A. C. JERMY POLYPODIUM VULGARE L. subsp. SERRATULUM Arcangeli, Comp. FI. Ital., 809 (1894). . vulgare var. serratum Willd., Sp. PI. 5 (1), 173 (1810). . australe Fée, Mém. Fam. Foug. 5, 236 (1852). . serratum (Willd.) Sauter in A. Kerner, Sched. Fl. Exsicc. Austr.-Hung. 2, 150 (1882), non Aublet, Hist. Pl. Guiane, 2, 962 (1775). . vulgare subsp. serratum (Willd.) H. Christ, Foug. Alp. Marit. 2 (1900). subsp. VULGARE. . vulgare L., Sp. Pl., 1085 (1753). subsp. PRIONODES Rothm., Mitt. Thur. Bot. Ver., 38, 106 (1929). P. interjectum Shivas, J. Linn. Soc. (Bot.) 58, 28 (1961). In Clapham, Tutin & Warburg, Fi. Brit. Isles, ed. 2, 40 (1962) three species are recog- nized within the Polypodium vulgare aggregate. Elsewhere the three possible hybrids between these species have also been recognized on morphological characters, and inferred from the presence of the parents growing together where the intermediates occur. As the occurrence of such intermediates makes field identification of material impossible with certainty, it seems better to treat them as subspecies. The correct nomenclature is as above. F. H. PERRING P. D. SELL as} ase Pasenss ne) CERASTIUM FONTANUM Baumg., Enum. Stirp. Transs. 1, 425 (1816). Usually a short-lived perennial, rarely annual herb up to 60 cm high, with short basal non-flowering shoots. Leaves oblanceolate, lanceolate, elliptical or ovate-oblong, subacute, sessile. At least the upper bracts with scarious margins. Petals equalling or some- what exceeding the sepals. Capsule usually curved. Seeds 0-4-1-3 mm, tuberculate. 292 Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 293 subsp. scoticum Jalas & Sell, subsp. nov. Holotype: Little Kilrannock, Forfarshire, 1 August 1872, J. Sadler (CGE). [C. triviale var. alpinum auct., non Mert. & Koch (1831), nec C. vulgatum *alpinum Hartm. (1849).] Planta ut videtur perennis, 3-12 cm alta; ramuli steriles breves, apice parvorosulati, vix vel raro radicantes; caules floriferi adscendentes, internodiis plerumque 1-5-2-5 cm longis, ubique pubescentes vel interdum lineis duabus exceptis et praecipue basin versus fere glabri; pili glanduliferi omnino desunt. Folia caulina sessilia, parva, lanceolata, (6—) 7-8 (-10) mm longa, 2-3 mm lata, plerumque 3-3-5—plo longiora quam latiora, apice obtusa vel subacuta, utrimque sparse pilosa vel subglabra, marginibus pilosis nervis medianisque subtus prominentibus. Bracteae herbaceae vel superiores apicem versus scarioso-marginatae. Inflorescentia confertim (1—) 3 (-5) flora, pedicellis tenuibus calyce subduplo longioribus. Calyx basi truncatus, (4: 5—) 5-5 (—6:5) mm longus, sepalis lanceolatis acutis margine scariosis, parce pilosis; basis sepalorum nervo mediano distincte carinata. Petala calyce c. 1-4 plo (usque ad 1-7 plo) longiora. Capsula leviter curvata vel fere recta, calycem vix superans vel usque ad 10 mm longa, dentibus rectis margine revolutis. Semina 0-8—-1-0 mm longa, verrucosa, verrucis c. 50 p altis et 125 » longis. Habitat in partibus alpinis montis scotici ‘Clova’ dicti, c. 910 m supra mare, ubi ut videtur endemica est. subsp. TRIVIALE (Murb.) Jalas, Arch. Soc. Zool.-Bot. Fenn. Vanamo 18 (1), 63 (1963). . viscosum L., Sp. PI., 437 (1753), nom. ambig. . vulgatum L., Amoen. Acad. 4, 101 (1759); non FI. Suec., ed. 2, 158 (1755). . holosteoides Fries, Nov. Fl. Suec. 4, 52 (1817). . triviale Link, Enum. Hort. Berol. Alt. 1, 433 (1821), nom. superfl. illegit. pro C. viscosum L. . vulgare subsp. triviale Murb., Bot. Not. 1898, 252 (1898). caespitosum subsp. eucaespitosum Graebner in Ascherson & Graebner, Syn. Mitteleur. FI. 5 (1), 638 (1918). . holosteoides subsp. glabrescens (G. F. W. Meyer) Méschl, Bot. Not. 1948, 375 (1948). . holosteoides subsp. pseudoholosteoides Méschl, Bot. Not. 1948, 364 (1948). BQ. DQAQAQ UE Flowering stems up to 50 cm high, pubescent to nearly glabrous, with or without glandular hairs. Leaves normally 10-25 x 3-10 mm, hairy. Sepals 3-7 mm long. Petals shorter than to slightly longer than sepals. Capsule up to 12 mm long. Seeds 0-4-0-9 mm, tubercles 15-40 x 15-40 p. Throughout the British Isles. In much of Continental Europe, but often replaced by other subspecies in mountainous districts. C. fontanum subsp. scoticum clearly differs from all the subspecies recognized within C. fontanum by Jalas (Arch. Soc. Zool.-Bot. Fenn. Vanamo 18 (1), 61-64 (1963); Fl. Eur. 1, 142 (1964)). From subsp. triviale (Murb.) Jalas it can be distinguished by its longer petals and larger seeds which have larger tubercles, and by the basal part of the sepals which seem to be more prominently keeled. In its seed and petal characters it is much closer to subsp. scandicum Gartner from Fennoscandia, and to subsp. fontanum from the Alps and Carpathians, but they are much taller plants with larger sepals and capsules and even larger seeds (up to 1:3 mm). Subsp. hispanicum Gartner from the Pyrenees and S. Spain has seeds of the same size, but it also is a taller plant and the tubercules of the seed are flattened and look quite different. In general appearance subsp. scoticum somewhat resembles small individuals of C. alpinum, but the long, soft, entangled ‘alpinum-hairs’ are lacking and the flowers are conspicuously smaller in all their parts. In addition to the numerous specimens in herbaria of subsp. scoticum from the mountains of the Clova region, specimens collected on coastal rocks at Strathy Point, W. Sutherland on the 30 July 1915 by E. S. Marshall (No. 4076), differ only in having the tubercles of the seeds 70-100 » x 50-100 p, and are best placed in this subspecies. Subsp. triviale is a most variable plant which tends to form local recognizable races which are geographical and/or ecological. This type of variation, is, no doubt, mainly due Watsonia 6 (5), 1967. 294 NOTES ON THE BRITISH FLORA to the predominance of autogamy, and the utmost care is needed before new subordinate taxa are described. Var. triviale (C. holosteoides sensu Clapham, non Fries) is the common plant of the lowlands both in the British Isles and on the Continent of Europe. In the British Isles var. holosteoides (Fries) Jalas (C. holosteoides Fries; C. holosteoides var. glabrescens (G. W. F. Meyer) Hylander) seems to be a plant of wet places and riversides of the north. It also occurs in southern Sweden and Denmark and has been recorded from the Alps. A plant collected by W. H. Beeby (No. 404) in wet meadows by the Thames at Putney in 1887 differs only in having obtuse, not acuminate hairs (C. holosteoides subsp. pseudoholosteoides Moschl). A variable group of rather dwarf plants with medium sized seeds from alpine and coastal localities may perhaps be referred to C. triviale var. serpenti- num Syme. They somewhat resemble var. serpentinicola (Domin) Jalas and var. kajanense (Kot. & Salmi) Jalas from Fennoscandia. Plants from a number of widespread localities in wet places in the British Isles which are like var. triviale, but have larger flowers and seeds, seem to merit recognition as a distinct variety. C. triviale var. pentandrum Syme is a curious annual plant that warrants further research. A full account of these varieties will be published at a later date. J. JALAS P.) DY SEE ARENARIA SERPYLLIFOLIA L. subsp. LEPTOCLADOS (Reichenb.) Nyman, Consp. 115 (1878). A. serpyllifolia var. tenuior Mert. & Koch in Rohling, Deutschl FI., ed. 3, 3, 266 (1831). A. serpyllifolia var. leptoclados Reichenb. in Reichenb. & Reichenb. fil., Icon. Fl. Germ. 5, 32 (1842). A. leptoclados (Reichenb.) Guss., Fl. Sic. Syn. 2, 824 (1845). A. serpyllifolia subsp. tenuior (Mert. & Koch) Arcangeli, Comp. FI. Ital., 101 (1882). Inflorescence lax. Capsule straight-sided, less than 3 mm long and 1:5 mm wide, possible to indent without fracture when mature. Pedicel slender, about 0-3 mm wide, upturned towards the tip, longer than sepals. Ripe seeds c. 0-4 X 0:4 mm. 2” = 20. South, west and central Europe extending locally north and east. subsp. SERPYLLIFOLIA A. serpyllifolia L., Sp. Pl., 423 (1753). Inflorescence lax. Capsule distinctly swollen at base, exceeding 3 mm long, 1-5-2 mm wide, possible to fracture when pressed at maturity. Pedicels stout, about 0-5 mm wide, straight, longer than sepals. Ripe seeds reniform, 0-5 x 0:4 mm. 2” = 40. Almost throughout Europe. subsp. macrocarpa (Lloyd) Perring & Sell, stat. nov. A. serpyllifolia var. macrocarpa Lloyd, Fl. Loire-Inf., 42 (1844). Inflorescence usually very dense. Capsule slightly swollen at base, exceeding 3 mm long and 2 mm wide, possible to fracture at maturity when pressed. Pedicels stout, about 0-5 mm wide, straight, often shorter than sepals. Ripe seeds reniform, exceeding 0-6 x 0-4 mm. Dunes on coasts of west Europe. The above three taxa have been variously treated. Sometimes A. serpyllifolia and A. leptoclados are regarded as species, with macrocarpa as a variety of serpyllifolia. Sometimes leptoclados is made a subspecies of serpyllifolia with macrocarpa a variety of subsp. serpyl- lifolia. Sometimes capsule shape is regarded as the most important, sometimes the seed size. When all characters are considered, as in F/. Eur. 1, 121 (1964), the descriptions make it very difficult to place many of the specimens. For this reason we think that the taxa are best regarded as subspecies, with the three main seed types as the primary character and the remaining characters subsidiary. A. marschlinsii Koch, Flora (Regensb.) 24, 509 (1841), (A. serpyllifolia subsp. marschlinsii (Koch) Nyman, Consp., 115 (1878)), from above 1,900 m in the Alps, is closely allied to subsp. /eptoclados, and is probably best placed as a montane subspecies of A. serpyllifolia. It differs from subsp. l/eptoclados in being often perennial, with fewer-flowered, dense inflorescences, and narrow scarious margins to the inner sepals. F. H. PERRING P. Dv SEEL Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 295 GENISTA TINCTORIA L. subsp. TINCTORIA Gy nnctoria L., Sp. Pl., 710 (1753). Plant more or less erect. Leaves 4-6 times as long as broad. Pods glabrous. subsp. LITTORALIS (Corb.) Rothm., Feddes Repert. 67, 5 (1963). G. tinctoria var. prostrata Bab., Man. Brit. Bot., 70 (1843) (Lectotype: Between Kynance and the Soap rock, Lizard, Cornwall, 23 July 1839 (CGE)). [G. humifusa Dickson ex Bab., loc. cit. (1843), nom. in syn.] G. tinctoria var. humifusa Bab., Man. Brit. Bot. ed 2, 73 (1847), nom. superfi. illegit. pro G. tinctoria var. prostrata Bab. G. tinctoria var. littoralis Corb., Nouv. Fl. Normand., 144 (1893). Plant prostrate. Leaves 2-4 times as long as broad. Flowers more handsome. Pods hairy on the back of the valves. Coasts of Cornwall (England), Pembroke (Wales) and Normandy (France). The British populations of subsp. Jittoralis, at least, are geographically isolated from those of subsp. tinctoria. Prostrate forms with glabrous pods and broad leaves are recorded from the coasts of Cornwall and north Devon and may show hybridization or gradation between the two subspecies. No specimens of subsp. /ittoralis have been seen from Nor- mandy, but Corbiére’s description seems to fit the British plants and it is a likely habitat and locality for it. The exact distribution of subsp. tinctoria is not known, but as several other subspecies have been described in central and east Europe, it may be confined to western Europe. The nomenclature of subsp. /ittoralis has been somewhat confused in that Druce, Brit. Pl. List, ed. 2, 2, 24 (1928), recognized each of the three synonyms given above as distinct varieties. G. humifusa Dickson (non L.) seems only to have been used as a name on herbarium sheets and never to have been validly published. Babington when describing var. prostrata presumably used a specimen in Herb. Dickson (not seen) and one in his own herbarium at Cambridge which is selected as the lectotype. In the second edition of his Manual Babington replaces the epithet prostrata with humifusa, using exactly the same diagnosis and citing the same manuscript name. Var. humifusa is thus a superfluous illegiti- mate name for var. prostrata. A plant of subsp. littoralis collected above Pentreath Beach, Lizard, Cornwall, by L. C. Frost in 1956, and grown in the Botanic Garden at Cambridge, has retained all its characters. F. H. PERRING P. D. SELL _ ANTHYLLIS VULNERARIA L. subsp. corbierei (Salmon & Travis) Cullen, comb. et stat. nov. A. maritima var. corbierei Salmon & Travis, J. Bot., Lond. 55, 320 (1917). J. CULLEN RUBUS ECHINATUS Lindley, Syn. Brit. Fl., 94 (1829) (Holotype: Hale End, no date, Forster in Herb. Lindley no. 43 (CGE)); Focke, J. Bot., Lond. 28, 132 (1890); Rogers, J. Bot., Lond. 30, 301 (1892); Rogers, Handb. Brit. Rubi, 64 (1900); E. S. Marshall, J. Bot., Lond. 43, 76 (1905); Rogers & E. F. Linton, J. Bot., Lond. 43, 203 (1905); Riddelsdell, J. Bot., Lond. 76, 311 (1938); non R. echinatus P. J. Muell., Flora (Regensb.) 41, 171 (1858). R. rudis var. attenuatus Bab., Ann. Mag. Nat. Hist., ser. 1, 17, 245 (1846) (Lectotype: Twycross, Leicestershire, 1845, A. Bloxham in Herb. C. C. Babington no. 792 (CGE)). R. fuscoater var. echinatus (Lindley) Bab., Ann. Mag. Nat. Hist., ser. 1, 17, 314 (1846). R. discerptus P. J. Muell., Pollichia, 16, 146 (1859); Focke in Ascherson & Graebner, Syn. Mitteleur. Fl. 6 (1), 565 (1903); Sudre, Bull. Soc. Etud. Sci. Angers 33, 115 (1904); Sudre, Rubi Eur. 132 (1908-13); W. C. R. Watson, Rep. Bot. Exch. Club Brit. Is. 8, 86 (1929 for 1928); W. C. R. Watson, Rep. Bot. Exch. Club Brit. Is. 11, 568 (1938 for 1937); W. C. R. Watson, Handb. Rubi Brit. Ireland, 132 (1958). Watsonia 6 (5), 1967. 296 NOTES ON THE BRITISH FLORA [R. rudis sensu Bab., Brit. Rubi, 189 (1869), et auct. angl. plur. ante 1890; non Weihe & Nees in Bluff & Fingerh., Comp. Fl. Germ. 1, 687 (1825).] Watson (1938, p. 566) states that R. echinatus Lindl. = R. histrix Weihe & Nees, basing this judgement on specimens identified for Leighton by Lindley several years after the publication of R. echinatus. Watson also states that there is no specimen of R. echinatus in Herb. Lindley (CGE). As Riddelsdell (1938, p. 311) points out, there is an excellent specimen of R. echinatus in Herb. Lindley, and this must be the holotype of the species. The holotype shows that R. echinatus Lindl. = R. discerptus P. J. Muell. It is very distinct from both R. histrix and R. rudis, the latter being the name given to it by most British botanists before 1890. Focke (1902, p. 566) rejects R. echinatus in favour of R. discerptus as he regards R. echinatus as a “nomen seminudum’. The name R. echinatus Lindl. is the oldest legitimate name of the species and must replace the later synonym R. discerptus P. J. Muell. B. A. MILES Sorbus wilmottiana E. F. Warburg, sp. nov. [S. wilmottiana E. F. Warburg in Clapham, Tutin & Warburg, Fi. Brit. Is. ed. 2, 432 (1962), anglice, sine diagn. lat.] Holotype: Clifton, edge of the greensward at the top of the Avon Gorge cliffs, v.c. 34 W. Glos., 4 October 1958, R. A. Graham, R. M. Harley, D. H. Lewis & E. F. Warburg, 5841 B (OXF). Arbuscula vel arbor parva. Rami hornotini brunneo-rubri, annotini nigri, glabri; gemmae 5-6 mm longae, subglabrae, apice tantum pilis lanatis instructae. Folia petiolata; lamina elliptica vel obovata, 7-12 cm longa, 1-6-2 (—2-7)-plo longiora quam latiora, apice acuta basi cuneata, duplicato-serrata marginibus per partem quintam basalem folii integris, dentibus acutiusculis plerumque ascendentibus apicale eisque nervos primarios terminanti- bus vix prominentibus, supra laete viridia nitidiuscula, subtus cinereo-tomentosa, nervis utrinque 6-9 supra vix elevatis; petiolus 1-2 cm longus. Inflorescentia pauciflora, glabra, c. 5 cm longa, 4 cm lata. Fructus subglobosus, 10-13 mm longus, latior quam longior, coccineus, lenticellis paucis magnis praeditus. This member of the Sorbus aria aggregate is named in memory of Alfred James Wilmott (1888-1950). E. F. WARBURGT LYTHRUM PORTULA (L.) D. A. Webb, Feddes Repert. 74, 13 (1967). subsp. PORTULA Peplis portula L. Sp. Pl., 332 (1753). subsp. longidentata (Gay) Sell comb. nov. P. portula var. longidentata Gay, Notic. Endress, 38 (1832). P. portula subsp. longidentata (Gay) Nyman, Consp., 252 (1879). P. fradini Pomel, Mat. Fl. Atl., 156 (1860). P. longidentata (Gay) Batt. in Batt. & Trabut, F/. Alger. (Dicot.), 320 (1889). Subsp. portula with the epicalyx segments about 0-5 mm long occurs throughout Europe. Subsp. Jongidentata with the epicalyx segments 1-5-2 mm long is found only in France, Spain, Portugal, Algeria and the Azores. Intermediates occur where the two subspecies meet and in the British Isles there is an east-west cline (cf. D. E. Allen, Watsonia 3, 85-91 (1954)). P. D. SELL ERICA PRAEGERI Ostenfeld and E. sruartTi E. F. Linton The earliest binomial deliberately given to a plant believed to be a hybrid between E. mackaiana and E. tetralix is E. < praegeri Ostenfeld, New Phytol. 11, 120 (1912). Dandy (List. Brit. Vasc. Pl., 85 (1958)) listed this hybrid as E. x stuartii E. F. Linton, citing E. praegeri Ostenfeld as a later synonym. The name E. stuartii was given by Linton Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA ae \)) (Ann. Scot. Nat. Hist. 1902, 177 (1902)) to a curious plant collected in Connemara by Stuart in 1890 and never found again, though it has been propagated by cuttings; all stocks of it in cultivation and all herbarium specimens are fragments of the same clone. For long it was regarded as E. mediterranea auct. (more correctly E. hibernica (Hook. & Arn.) Syme) x E. tetralix, but some years ago (Webb, Irish Nat. Jour. 11, 190 (1954)) I pointed out that there were serious objections to this explanation of it and concluded by saying that ‘if I am forced to choose, I prefer to regard it as a very peculiar mutant of E. mackaiana’. A little later I came to the conclusion that it was more like a mutant of E. xX praegeri, and I mentioned this to Mr. Dandy. Unfortunately he took this rather casual obiter dictum as a firm taxonomic opinion, with the result that this unique plant and the common and familiar hybrid are now regarded as two nothomorphs which must share the name E. X stuartii. I now wish to retract my opinion to the extent of saying that I do not know what E. stuartii is, and that I do not believe that anyone else does either. I think that it has some claim to be regarded as a ‘monstrosity’, in view of its very curious corolla, in which case it cannot stand as a nomenclatural type. Unfortunately our grandfathers had a much clearer concept of ‘monstrosity’ than their more sophisticated descents, and a recent attempt to give more precision to the term by defining it as a non-hereditable structural abnormality is not very helpful in the numerous cases of monstrosities or near-monstrosities which are (like E. stuartii) completely sterile. But even if this plant is not a monstrosity it is likely, in my opinion, to remain permanently incertae sedis. That it is a mutant of FE. x praegeri is a plausible guess and no more, for the evidence is conflicting. I do not think that a name whose application represents a rather desperate guess should, even if it is published earlier with a different application, take priority over one whose application was based originally on strong circumstantial evidence and has since been confirmed by experiment. A name whose type is inexplicable can be rejected as a nomen dubium (in so far as it desig- nates a taxon of definite rank or a hybrid of specified parentage) just as legitimately as one whose type cannot be ascertained; and I believe that the continued application of the name E. stuartii to the hybrid is likely to be a source of permanent confusion. In my opinion the hybrid E. mackaiana x tetralix should be called E. x praegeri Ostenfeld. D. A. WEBB ANAGALLIS ARVENSIS L. subsp. ARVENSIS forma ARVENSIS . arvensis L., Sp. Pl., 148 (1753). . phoenicea Scop., Fl. Carn. ed. 2, 1, 139 (1772). . arvensis subsp. phoenicea (Scop.) Vollman, Ber. Bayer. Bot. Ges. 9, 44 (1904). forma CARNEA (Schrank) Hyland., Uppsala Univ. Arsskr. 7, 256 (1945). A. coerulea var. carnea Schrank, Baier. Fi. 1, 461 (1789). forma PALLIDA (Hook. fil.) Hyland., Uppsala Univ. Arsskr. 7, 256, (1945). A. arvensis vat. pallida Hook. fil., Stud. Fl. Brit. Islands, 265 (1870). forma AZUREA Hyland., Uppsala Univ. Arsskr. 7, 256 (1945). A. latifolia L., Sp. Pl., 149 (1753). A. arvensis var. azurea Wilmott, Rep. Bot. Exch. Club Brit. Is. 14, 664 (1946). forma LILACINA Alefeld (quando ?) Hylander, Uppsala Univ. Arsskr.'71, 256 (1945). A. arvensis var. lilacina (Alefeld) Marsden-Jones & Weiss, Proc. Linn. Soc. Lond. 150, 3 147 (1938). A. arvensis subsp. phoenicea forma lilacina (Alefeld) Schinz & R. Keller, F/. Schweiz, ed. 2, 2, 173 (1905). subsp. FOEMINA (Mill.) Schinz & Thell. in Schinz & R. Keller, F/. Schweiz, ed. 3, 1, 419 (1909). A. foemina Mill., Gard. Dict. ed. 8, (1768). SENSE Watsonia 6 (5), 1967. 298 NOTES ON THE BRITISH FLORA The taxonomy of this species is that recommended by E. M. Marsden-Jones and F. E. Weiss, Proc. Linn. Soc. London 150, 146-155 (1938). The five flower-colour forms of subsp. arvensis breed true for colour. Any one of them can be crossed with any of the remainder and the offspring produces fertile seed. Subsp. foemina can be crossed with any of the colour forms of subsp. arvensis, but only the cross with forma carnea produces fertile offspring. PD Soe CALYSTEGIA SEPIUM (L.) R. Br. subsp. roseata Brummitt, subsp. nov. A subspecie typica corolla rosea, caulibus saepius pubescentibus et foliis apice acutiori- bus differt; a subsp. americana (Sims) Brummitt floribus minoribus (corolla plerumque 19-25 mm longa, staminibus plerumque 19-25 mm longis) differt. Holotype: Wales, Merionethshire, near Towyn, overlooking Afon Dysynni estuary, Grid Ref. 23/5703, 24 September 1959, J. A. Ratter, J. Cullen & R. K. Brummitt 59.815 (LIVU; isotypes in GB, K, NY). This subspecies is found on the Atlantic coasts of Europe (Norway, Sweden, Nether- lands, British Isles, probably also Spain and Portugal, perhaps also France), and the Pacific and perhaps Atlantic coasts of temperate South America, Easter Island, New Zealand and south-eastern Australia. The type specimen of C. sepium subsp. baltica Rothmaler in Feddes Repert. 67, 10 (1963), from the Baltic coast of northern Germany, has been examined by the author through the courtesy of the curator of the herbarium of the University of Greifswald. It is best referred to C. sepium subsp. sepium forma colorata (Lange) Dé6rfler, a pink-flowered variant of the typical subspecies occurring frequently in lowland areas of many parts of Europe. R. K. BRUMMITT RHINANTHUS L., Sp. Pl., 603 (1753); Gen. Pl., ed. 2, 263 (1754) (Type: R. crista-galli L. nom. ambig. = R. serotinus (Schénh.) Oborny). Alectorolophus Zinn, Cat. Pl. 288 (1757), nom. superfl. illegit. pro Rhinanthus L. RHINANTHUS SEROTINUS (Schénh.) Oborny, Verhandh. Naturforsch. Ver. Briinn. 22 (2), 435 (1884). R. crista-galli L., Sp. Pl., 603 (1753), nom. ambig. R. major Ehrh., Beitr. Naturk. 6, 144 (1791), non L., Amoen. Acad. 3, 53 (1756). Alectorolophus serotinus Schonh., Flora (Regensb.) 15, 601 (1832). subsp. APTERUS (Fries) Hyland., Uppsala Univ. Arsskr. 1945, 292 (1945). . Major var. apterus Fries, Noy. FI. Suec., Mantissa 3, 60 (1842). . apterus (Fries) Ostenf., Bot. Not. 1904, 83 (1904). . major subsp. apterus (Fries) Schinz & Thell. in Schinz & R. Keller, Fl. Schweiz, ed. 3, 2, 314 (1914). . major subsp. vernalis Zinger, Herb. Fl. Ross., no. 2530a. . serotinus subsp. vernalis (Zinger) Hyland., Uppsala Univ. Arsskr. 1945, 292 (1945). [R. borbasii et R. polycladus auct. brit.] I agree with E. F. Warburg (Clapham, Tutin & E. F. Warburg, Fi. Brit. Isles, ed. 2, 699 (1962)) that all British material of this species is referable to subsp. apterus. AA DAD RHINANTHUS MINOR L. subsp. MINOR R. minor L., Amoen. Acad. 3, 54 (1756). R. drummond-Hayi subsp. salmonii Soé, Feddes Rep. 26, 182 (1929). (Holotype: rocks foot of Craig Maid, Clova, Forfar v.c. 90, 16 July 1912, C. E. Salmon (BM)). Usually with only very short flowerless branches, but sometimes with longer suberect flowering branches from the middle and upper part of the stem. Internodes (except the lowest) more or less equal. Leaves of main stem (10-) 20-40 (-50) x (3-) 5-7 mm, usually Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 3) oblong and parallel-sided for the greater part of their length. Intercalary leaves 0 (-1) pairs. Calyx hairy only on the margins. Lowest flowers usually from the 6th—9th node. Corolla yellow with a violet (or rarely white) tooth. Flowers May to July. Grassy places on dry basic soils, common in England, rarer in Scotland where it often grades into subsp. stenophyllus. Throughout Europe extending to Caucasus and W. Siberia; S. Greenland, Newfoundland. subsp. calcareus (Wilmott) E. F. Warburg, comb. et stat. nov. [R. minor subsp. calcareus (Wilmott) E. F. Warburg in Clapham, Tutin & E. F. Warburg, Fl. Brit. Is., ed. 2, 700 (1962) sine ref. ad basionymum. | R. calcareus Wilmott, J. Bot., Lond. 78, 202 (1940) (Holotype: Grassy chalk down above Sutton, W. Sussex v.c. 13, 3 August 1901, E. S. Marshall, no. 2584 (BM)). Usually with long arcuate-ascending flowering branches from about the middle of the stem. Lower internodes short, upper very long. Leaves of main stems 10-25 x 1-5-3 mm, linear, more or less spreading. Intercalary leaves usually (2—) 3-6 pairs. Calyx hairy only on the margins. Lowest flowers usually from 14th-19th node. Corolla yellow; tooth violet. Local on chalk and limestone downs from Dorset and Sussex to Gloucester and Northants. ? Endemic. Plants from Ireland which somewhat resemble it are probably best placed in subsp. stenophyllus. subsp. STENOPHYLLUS (Schur) O. Schwarz, Feddes Repert. 46, 56 (1939). R. minus var. stenophyllus Schur, Enum. Pl. Transs., 511 (1866). R. crista-galli subsp. stenophyllus (Schur) So6, Feddes Repert. 26, 186 (1929). R. stenophyllus (Schur) Druce, Ann. Scot. Nat. Hist. 1901, 178 (1901); J. Bot., Lond. 41, 360 (1903). Usually with long arcuate-ascending flowering branches and shorter flowerless branches from the lower and middle part of the stem. Lower internodes usually short, equalling or shorter than the leaves, upper much longer. Leaves of main stems 15-45 x 2-5 (-7) mm, narrowly lanceolate or linear-lanceolate, more or less tapering from near the base. Inter- calary leaves (O—) 1-2 (-4) pairs. Calyx hairy only on the margins. Lowest flowers usually from (8th—) 10th—13th (-15th) node. Corolla yellow with violet tooth, sometimes becoming brown. Flowers July to September. Damp grassland and fens. Common in Scotland, north and west England, Wales and probably Ireland, rare (mainly in fens) in south-east England. Throughout much of Europe particularly in upland regions. subsp. MONTICOLA (Sterneck) O. Schwarz, Feddes Repert. 46, 56 (1939). R. minor var. monticola Lamotte, Mém. Acad. Clermont. 22, 880 et 166 (1881), nom. superfl. illegit. pro R. minor var. angustifolius Gren. & Godron. Alectorolophus monticola Sterneck, Abh. Zool.-Bot. Ges. Wien 1 (2), 111 (1901). R. monticola (Sterneck) Druce, Ann. Scot. Nat. Hist. 1901, 178 (1901); Ann. Scot. Nat. Hist. 1903, 171 (1903); J. Bot., Lond. 41, 360 (1903). R. crista-galli subsp. monticola (Sterneck) So6, Feddes Repert. 26, 187 (1929). R. spadiceus Wilmott, J. Bot., Lond. 78, 203 (1940). (Holotype: Glen Dee near Braemar, S. Aberdeen v.c. 92, 30 July 1906, E. S. Marshall no. 2947 (BM)). _ R. spadiceus subsp. orcadensis Wilmott, J. Bot., Lond. 78, 207 (1940) (Holotype: Grassy and healthy pasture, Point of Onston, Stenness, Mainland, Orkney, v.c. 111, 21 July 1922, H. H. Johnston no. 1879 (BM)). Usually with short or moderate flowerless branches from near the base, sometimes with 1-3 longer flowering branches in addition. Lower internodes usually very short, upper much longer. Leaves of main stems 10-20 (-25) x 2-4 mm, linear-lanceolate, more or less tapering from near the base, tending to be more erect than in subsp. stenophyllus. Intercalary leaves usually 1-2 (3) pairs. Calyx hairy only on the margins. Lowest flowers usually from the (7th-) 8th-llth (-12th) node. Corolla dull yellow becoming treacle brown or treacle-brown throughout, with a violet tooth. Flowers July and August. Watsonia 6 (5), 1967. 300 NOTES ON THE BRITISH FLORA Grassy places in mountainous regions from Yorks to Shetland, rare in Ireland. Also in the Alps. subsp. lintonii (Wilmott) Sell, comb. et stat. nov. R. drummond-Hayi sensu Sterneck, Abh. Zool.-Bot. Ges. Wien 1 (2), 115 (1901) quoad descript.; et sensu auct. brit. R. gardineri Druce, Rep. Bot. Exch. Club Brit. Is. 4, 205 (1915). R. lintonii Wilmott, J. Bot., Lond. 78, 209 (1940) (Holotype: Glen Fiagh, Forfar, 13 July 1889, E. F. Linton (BM)). R. lochabrensis Wilmott, J. Bot., Lond. 78, 211 (1904). (Holotype: near Meall Cumhaun, Glen Nevis, Inverness, 19 August 1902, C. E. Salmon (BM)). R. vachelliae Wilmott, Watsonia (1948), 1, 85 (Holotype: Glen Cannich, near the north-east end of Loch Mullardoch, E. Inverness, v.c. 96, 16 July 1947, E. Vachell (BM)). Simple or with 1-2 pairs of branches from the lower or middle part of the stem. Internodes more or less equal in length, or the lower shorter. Leaves of main stem (8-) 10-20 (-30) x 1-5-3 (-4) mm, very narrow linear-lanceolate, more or less tapering from the base. Intercalary leaves 0-3 pairs. Calyx hairy all over. Flowering from the 7th—-10th node. Corolla ? yellow or orange-yellow. Flowers July (or August). Grassy places in Scottish mountains. Apparently originating from hybridization between subspecies borealis, monticola and stenophyllus, but now often occupying a zone (where it is uniform) between borealis (higher up) and monticola and stenophyllus (lower down). ?Endemic. subsp. borealis (Sterneck) Sell, comb. et stat. nov. R. crista-galli var. drummond-hayi F. B. White, Ann. Scot. Nat. Hist. 8, 324 (1886). (Type: Ben Lawers, Breadalbane, F. Buchanan White (BM)). Alectorolophus borealis Sterneck, Annu. Cons. Jard. Bot. Genéve 1889, 25 (1889). Alectorolophus drummond-hayi (F. B. White) Sterneck, Abh. Zool.-Bot. Ges. Wien 1 (2), 115 (1901) quoad basionymum exclud. descript. R. drummond-hayi (F. B. White) Druce, Ann. Scot. Nat. Hist. 1903, 171 (1903); J. Bot., Lond. 41, 360 (1903). R. borealis (Sterneck) Druce, Ann. Scot. Nat. Hist. 1901, 178 (1901); J. Bot., Lond. 41, 361 (1903). R. borealis var. calvescens Wilmott, J. Bot., Lond. 78, 209 (1940). (Holotype: Ben Nevis, v.c. 97, July 1903, H. J. Riddelsdell (BM)). Usually unbranched, but occasionally with short flowerless (very rarely flowering) branches, and then intercalary leaves 0. Internodes more or less equal. Leaves 10-30 x 3-7 mm, oblong or oblong-linear, more or less parallel-sided. Calyx hairy all over. Lowest flowers from the 5th—-7th (-8th )node. Corolla bright yellow with violet or white teeth. Flowers July and August. Local in grassy places on mountains up to 3,200 ft in Scotland, Caernarvon and Kerry. Also in Alaska, Greenland, Iceland and ?Scandinavia. Extensive work on Rhinanthus has been carried out on the Continent by A. Chabert, Bull. Herb. Boiss. 7, 497-517 (1899), R. V. Sood, Feddes Repert. 26, 179-219 (1929), and J. V. Sterneck, Abh. Zool.-Bot. Ges. Wien, 1 (2), 1-150 (1901). E. S. Marshall, J. Bot., Lond. 41, 291-300 (1901) and G. C. Druce, J. Bot., Lond. 41, 359-361 (1903) published the identifications of their material by Sterneck, but not until A. J. Wilmott’s paper in J. Bot., Lond. 78, 201-213 (1940) was any serious work published on the British species of the genus. D. J. Hambler, Watsonia, 4, 101-116 (1958) reports on a series of investigations he carried out on the genus Rhinanthus, in which he concludes that only two species, R. minor and R. serotinus (Schénh.) Oborny, distinguished on corolla characters, are found in the British Isles. With this conclusion I agree. Although he points out that populations of R. minor are often distinct he gives no infraspecific categories, a situation hardly tolerable to a botanist who has examined the species in the field. The account by E. F. Warburg in Clapham, Tutin & E. F. Warburg, Fi. Brit. Isles, ed. 2, 699-701 Watsonia 6 (5), 1967. F ee ee ee ee me, ee oe ee NOTES ON THE BRITISH FLORA 301 (1962) is much more realistic, though I do not think that R. borealis should be distin- guished as a distinct species, and I think that the supposed hybrid, R. * gardineri, which produces uniform populations, is better regarded as a separate taxon. Examples comparable to this last plant are seen in Euphrasia vigursii Davy (formed from a cross between E. micrantha Reichb. and E. anglica Pugsl.) and Montia fontana subsp. variabilis Walters (a variable intermediate between subsp. fontana and subsp. amporitana Sennen which may be of hybrid origin). The nomenclature of the group is not difficult once the mass of synonyms have been sorted out. R. crista-galli L. is rejected as a nomen ambiguum. Sterneck was the first author to use most of the names of our taxa at the rank of species, but under the generic name Alectorolophus. They were transferred to Rhinanthus by Druce in Ann. Scot. Nat. Hist. (1901 and 1903), and J. Bot., Lond. (1903). Wilmott’s plants have good descriptions and holotypes. E. F. Warburg supplied the data for subsp. calcareus and is responsible for that combination. P. D. SELL EUPHRASIA L. EUPHRASIA MICRANTHA Reichb., F/. Germ. Excurs., 358 (1831) (Lectotype: In via qui iter verram Lauram, legi Jul. 1830 [Reich. p.] (W)). The type specimen is white-flowered. EUPHRASIA FRIGIDA Pugsley, J. Linn. Soc. (Bot.) 48, 490 (1930) (Lectotype: Ujaragsuit, Godthaabs Fjord, W. Greenland, 13 August 1885, S. Hansen (C)). [E. latifolia auct., non L., Sp. Pl., 604 (1753).] |E. arctica auct. amer., non Lange ex Rostrup, Bot. Tidsskr. 1 (4), 47 (1870-71).] EUPHRASIA TETRAQUETRA (Brébisson) Arrondeau, Bull. Soc. Polymath. Morbihan 1862, 96 (1863). E. officinalis var. tetraquetra Brébisson, F/. Normand. ed. 2, 183 (1849). E. occidentalis Wettst., Mon. Gatt. Euphrasia, 135 (1896). E. canadensis Towns., J. Bot., Lond. 36, 1 (1898). The types of none of these names have been traced, but in each case a good original diagnosis was given. EUPHRASIA NEMOROSA (Pers.) Wallr., Ann. Bot., 82 (1815). E. officinalis var. nemorosa Pers., Syn. P!. 2, 149 (1807). As there are no specimens in the Persoon herbarium at Leiden, we have typified the name by Plate 233 in Bulliard, Herbier de la France, 4 (1785), which was cited with the original diagnosis. EUPHRASIA ARCTICA Lange ex Rostrup, Bot. Tidsskr. 1 (4), 47 (1870-71) (Lectotype: Sandyford, Stromo, 31 August 1867, E. Rostrup (C)). [E. borealis auct. mult. incl. Wettst. pro parte.] E. borealis var. zetlandica Pugsl. et var. speciosa Pugsl., J. Linn. Soc. (Bot.) 48, 516 (1930). P. D. SELL PES YEO ODONTITES VERNA (Bellardi) Dumort. subsp. PUMILA (Nordst.) A. Pedersen, Bot. Tidsskr. 58, 291 (1963). Odontites serotina forma pumila Nordst., Bot. Not. 1898, 220 (1898). [Odontites litoralis sensu F. N. Williams, Prod. Fl. Brit., 316 (1909).] [Bartsia odontites var. litoralis auct. brit.] Plant up to 17 (-21) cm high, often unbranched but sometimes with one or two pairs Watsonia 6 (5), 1967. 302 NOTES ON THE BRITISH FLORA of short branches. Internodes short, giving the plant a condensed habit. Leaves lanceolate to ovate-lanceolate and distinctly toothed. Bracts longer than the flowers. Flowers apparent- ly darker red than in other subspecies. Flowers July-August (September). Grassy places by the sea. N. & W. Scotland (Sutherland, Caithness and Hebrides), Netherlands, Denmark and S. Sweden. subsp. VERNA Euphrasia verna Bellardi, Mem. Acad. Sci. Turin, ser. 1, 1790-91, 239 (1793). Odontites verna (Bellardi) Dumort., F/. Belg., 32 (1827). Odontites verna subsp. longifolia Corbiere, Nouy. Fl. Normand., 436 (1894). Bartsia odontites var. verna (Bellardi) Druce, List Br. Pl., 54 (1908). Bartsia odontites var. longifolia (Corbiére) Druce, List Br. Pl., 54 (1908). [| Bartsia odontites var. rotundata auct. brit.] Plant 10-30 cm high, with branches coming off at an angle of less than 45° and more or less straight. Internodes longer giving the plant a more open habit. Leaves lanceolate, distinctly toothed. Bracts longer than the flowers. Flowers June and July. 2n = 40. Cultivated, waste and grassy places. In the British Isles commoner in the north than the south. In Europe as a whole more common in the north and mountainous regions, perhaps absent from some areas in the south. subsp. SEROTINA Corbiére, Nouv. Fl. Normand., 437 (1894). Euphrasia odontites L., Sp. Pl., 604 (1753). Euphrasia serotina Lam., Fl. Fr. 2, 350 (1778) nom. superfi. illegit. pro E. odontites L. [Odontites rubra Gilib., Fl. Lith. 1, 126 (1781) nom. non rite publ.] Euphrasia rubra Baumg. in Hoffm., Deutschl. Fl., 215 (1791). Odontites vulgaris Moench, Meth., 439 (1794) nom. nov. pro Euphrasia odontites L. Odontites serotina Dumott., FI. Belg., 32 (1827), nom. superfi. illegit. pro Euphrasia odontites EP Euphrasia divergens Jord. in F. W. Schultz, Arch. Fl. Fr. Allem., 191 (1851). Odontites rubra subsp. serotina Wettst., Denkschr. Akad. Wiss. Math.-Nat. Kl. (Wien) 70, 321 (1901). Bartsia odontites var. serotina auct. mult. includ. Druce, List Brit. Pl., 54 (1908). Bartsia odontites var. divergens (Jord.) Druce, Brit. Pl. List, 54 (1908). Plant 20-50 cm high, with branches spreading at a wide angle, sometimes nearly at right angles, their tips often upcurved. Internodes long, giving the plant an open habit. Leaves linear-lanceolate, somewhat narrowed at the base, obscurely toothed. Bracts shorter than or equalling the flowers. Flowers July and August. 2n = 20. Similar habitats to subsp. verna. Common in S. England, rarer farther north and in hilly regions. Common in the lowlands of continental Europe, but rare in the north and sometimes the only subspecies in the south. ? East Europe. The genus Odontites seems to be represented in the British Isles by the above three subspecies. Subsp. verna and subsp. serotina are said to have different chromosome numbers, but their morphological characters (particularly in the British Isles) overlap so greatly, and there are so many intermediates, that it seems impossible to distinguish them as species. In continental Europe subsp. verna seems to be predominantly a northern or montane plant, while subsp. serotina is more lowland, with a distribution which goes much farther south. In the south and the north they seem to be much more distinct than in a zone running across central Europe and including the British Isles. Subsp. pumila is a dwarf coastal race most nearly allied to subsp. verna, with which it intergrades in the Hebrides. Round the Baltic occurs another dwarf coastal race which is usually unbranched, has long lower internodes, much longer than the leaves, and bracts longer than the flowers. It would seem to be most nearly allied to subsp. pumila. Its correct name seems to be O. VERNA subsp. LITORALIS (Fries) Nyman, Consp., 551 (1881) (Euphrasia odontites var. litoralis Fries, Fl. Scan., 40 (1835)). G. Marklund (Acta Soc. Fauna FI. Fenn. 72, 1-18 (1955)) when dealing with the Finnish plants of the genus, has described O. litoralis subsp. fennica Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 303 and O. litoralis subsp. littoralis var. baltica. | have not seen material of these plants, but from their descriptions they seem to be very close to O. verna subsp. litoralis, and may form connecting links between it and subsp. verna. Subsp. verna, subsp. pumila, subsp. litoralis and O. litoralis subsp. fennica thus form the northern elements of the O. verna complex. In the south, subsp. serotina is the dominant plant, but in Sicily plants with linear or linear- lanceolate, markedly dentate leaves and yellowish-pink flowers seem to be distinct. I call them QO. verna subsp. sicula (Guss.) Sell, comb. et stat. nov. (Euphrasia serotina var. sicula Guss., Fl. Sic. Prodr. 2, 149 (1828)). In eastern Europe the plants most resemble subsp. serotina, but they become increasingly pubescent eastwards and another subspecies may have to be distinguished there. The large number of small taxa of various ranks, and their combinations in Euphrasia, Bartsia and Odontites, plus a superfluity of names, have made the nomenclature of the group very difficult. I have done my best here to establish the correct names of the British taxa, and to show where synonyms or misapplied names belong, but only an exhaustive study of all the European literature could possibly establish the correct authors of the various combinations. P. D. SELL VALERIANELLA LOCUSTA (L.) Betcke subsp. dunensis (D. E. Allen) Sell, stat. nov. V. locusta var. dunensis D. E. Allen, Watsonia 5, 45 (1961). I have discussed this taxon with Allen. He tells me that he uses the rank of subspecies for geographical variation and that of variety for ecological variation. I see no clear difference between geographical, ecological or even altitudinal isolation of populations, all three factors sometimes applying to the same taxon. I have thus elevated Allen’s variety to the rank of subspecies. P. D. SELL SENECIO VULGARIS L. subsp. denticulatus (O. F. Muell.) Sell, comb. et stat. nov. S. denticulatus O. F. Muell., Fl. Dan. 5 (14), 4 et t. 791 (1780). This taxon appears to constitute a distinct ecological race, which in view of its wide geographical range seems to me to merit subspecific rank. For a discussion of the radiate forms of S. vulgaris see D. E. Allen, Watsonia 6, 280-282 (1967). Ps De SELL HIERACIUM L. We are here publishing only what is necessary to validate the new names and combina- tions of species being mapped in the Appendix to the Atlas of British Plants. A more detailed paper, bringing up to date the revision by Pugsley, J. Linn. Soc. (Bot.) 54 (1948), will follow. In Sectione ALPINA F. N. Williams Hieracium tenuifrons Sell & C. West, sp. nov. Ab H. alpino L. squamis involucri angustioribus, stylis obscuris et ligulorum pilis paucioribus differt, praeterea ab H. pseudopetiolato (Zahn) Roffey squamis longioribus angustioribus et capitulis latioribus distinguitur. Holotype: Fionn Ghleann, Glen Etive, v.c. 98, Argyll, July 1954, C. West (CGE). Planta phyllopoda. Caulis 13-25 (—36) cm altus, valde striatus, flexuosus, pilis aliis simplicibus pallidis nonnunquam basi nigris longis vel longissimis (ad 5 mm), aliis stellatis numerosis, aliis glanduliferis numerosis brevissimis flaviusculo-brunneis vel atriusculis vestitus. Folia medio-viridia; basalia numerosa; primigenia obovata, obtusa, denticulata, basi cuneata; cetera oblanceolata vel obovata; obtusa, denticulata vel leviter dentata, basi attenuata, petiolis mediocribus vel longis; caulina 1-3 (—7), anguste oblanceolata Watsonia 6 (5), 1967. 304 NOTES ON THE BRITISH FLORA vel linearia, obtusa vel subacuta, integra vel raro denticulata, basi attenuata, sessilia vel breviter petiolata; omnia utrinque et margine pilis simplicibus longis pallidis basi nigris, et margine pilis glanduliferis numerosis brevissimis vestita. Anthela capitulo unico. Involucri squamae 14-16 mm longae, 1-0-1-25 mm latae, atriusculo-virides, lineares, plus minusve acutae, pilis aliis simplicibus densissimis longis vel longissimis (ad 5 mm) pallidis, aliis glanduliferis numerosis brevissimis obscuris, aliis stellatis nullis obsitae. Ligulae saturate flavae, apice pilis brevibus vel brevissimis vestitae. Styli obscuri. Receptaculi alveoli margine breviter dentati. Cypsela obscura, 3-3-5 mm longa. Hieracium notabile Sell & C. West, sp. nov. H. eximium Backh., H. marginatum Sell & C. West, H. memorabile Sell & C. West et H. backhousei F. J. Hanb. quoad faciem simulat. Ab H. memorabile et H. backhousei foliis obtusis integris vel denticulatis differt, praeterea ab H. eximio et H. marginato squamis involucris latioribus obtusioribus distinguitur. Holotype: Ben More, v.c. 88, Mid-Perth, 16 July 1910, E. S. Marshall no. 3489 (CGE). Planta phyllopoda. Caulis 17-26 cm altus, valde striatus, flexuosus, parte basali interdum purpureo-tincta, pilis aliis stellatis paucis, aliis simplicibus numerosis inaequalibus brevibus mediocribus longisque basi nigris patentibus, aliis glanduliferis brevissimis paucis obscuris vestitus. Folia pallide viridia; basalia pauca; primigena late elliptica, mucronulata, integra, basi attenuata; cetera anguste elliptica vel elliptica, raro late elliptica, obtuso- mucronata vel subacuta, integra vel dentibus parvis paucis vel interdum dentibus tenuibus cuspidatis muniter, basi gradatim attenuata, petiolis alatis; caulina 2-3, inferioribus basali- bus similibus, superioribus bracteiformibus; omnia utrinque et margine pilis simplicibus numerosis longis saepe fere setiformibus, et margine pilis glanduliferis numerosis brevissimis vestita. Anthela capitulo unico. Involucri squamae 12-16 mm longae, 1-1-5 mm latae, atriusculo-virides, lineares, obtusae vel subacutae, raro acutae, pilis simplicibus densis inaequalibus longis pallidis, pilis glanduliferis nullis, apice pilis stellatis paucis obsitae. Ligulae saturate flavae, apice et nonnunquam in parte aversa pilis simplicibus brevibus vel brevissimis vestitae. Styli obscuri. Cypselae obscurae, circa 4 mm longae. Hieracium marginatum Sell & C. West, stat. et nom. nov. Basionym: H. globosiflorum var. lancifolium Pugsl., J. Linn. Soc. (Bot.) 54, 47 (1948) (Holotype: Ben Laoghal, v.c. 108, W. Sutherland, 2,300 ft, 2 August 1897, E. S. Marshall no. 2042 (BM)). Hieracium memorabile Sell & C. West, sp. nov. H. notabile Sell & C. West, H. eximium Backh., H. marginatum Sell & C. West et H. backhousei F. J. Hanb. quoad faciem simulat. Ab H. notabile squamis involucri angustioribus acutis, ab H. eximio squamis ligulisque pilis brevioribus, ab H. backhousei foliis pilosioribus stylisque obscuris distinguitur. Holotype: Beinn an Sochaich, 2,500—3,000 ft, v.c. 97, W. Inverness, July 1896, E. S. Marshall and W. A. Shoolbred in Linton, Set of Brit. Hierac. no. 55 (CGE). Planta phyllopoda. Caulis 15-24 cm altus, valde striatus, in parte basali nonnumquam purpureus, pilis aliis simplicibus longis vel longissimis (ad 6 mm longis) pallidis basi nigris, aliis stellatis inferne sparsis superne numerosioribus, aliis glanduliferis paucis vel numerosis brevissimis brevibus vel raro mediocribus obscuris vestitus. Folia saturate viridia; basalia plerumque numerosa; primigena suborbicularia, mucronulata, leviter dentata, basi abrupte contracta, petiolis brevibus; cetera dimorpha, exteriora late elliptica, mucronulata, dentata vel inciso-dentata, dentibus nonnunquam cuspidatis vel mammiformibus, basi cuneata, interiora anguste elliptica vel elliptica, plus minusve acuta, dentata vel inciso-dentata, dentibus saepe inaequalibus nonnunquam cuspidatis, basi attenuata; caulina 1-2 bractei- formia; omnia utrinque et margine pilis simplicibus plus minusve numerosis mediocribus et longis gracilibus, et margine pilis glanduliferis plus minusve numerosis brevissimis flaviusculis vestita. Anthela capitulo unico, nonnunquam furcata, utroque ramo capitulo Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 305 unico. Involucri squamae 13-16 mm longae, 1-1-5 mm latae, ante anthesin incumbentes, atriusculo-virides, lineari-lanceolatae, acutae, pilis aliis simplicibus inaequalibus longis pallidis basi nigris, aliis glanduliferis paucis brevissimis vel brevibus obscuris gracilibus obsitae. Ligulae saturate flavae, apice et in parte aversa pilis simplicibus numerosis brevis- simis vel brevibus vestitae. Styli obscuri. Cypselae obscurae, circa 3-5 mm longae. Hieracium atraticeps (Pugsl.) Sell & C. West, stat. sp. nov. H. chrysanthum var. microcephalum Backh., Mon. Brit. Hieracia, 34 (1856) (Lectotype: Lochnagar, 1855, J. Backhouse, jun. (BM)). H. hanburyi var. microcephalum (Backh.) Pugsl., J. Linn. Soc. (Bot.) 54, 51 (1948) quoad basionym exclud. descript. H. hanburyi vat. atraticeps Pugsl., J. Linn. Soc. (Bot.) 54, 51 (1948) (Holotype: Lochnagar, 1887, E. F. Linton (BM)). H. chrysanthum var. microcephalum Backh. was based on specimens collected from Lochnagar (Aberdeen), Clova (Angus), and Striding Edge and Glaramara (Cumberland). Plants in the Backhouse herbarium from Lochnagar are H. atraticeps and those from Cumberland H. subgracilentipes (Zahn) Roffey. No Backhouse specimens from Clova have been traced, but H. atraticeps occurs there. It is thus safe to assume that Backhouse included within his taxon two of ours. The original diagnosis applies almost equally well to both, with perhaps a slight tendency to fit H. atraticeps better. In the discussion that follows the diagnosis, there is also a slight tendency to suggest that the Cumberland form is slightly aberrant. We have therefore chosen a Lochnagar plant as the lectotype. Pugsley (1948) cited as a type of H. chrysanthum var. microcephalum an E. F. Linton specimen collected at Clova in 1890. This of course is impossible by the International Code. The plant he described is a slight variation of H. hanburyi Pugsl. On the same page he described H. hanburyi var. atraticeps giving as its type an E. F. Linton specimen collected at Lochnagar in 1887, and citing also the specimen we have now selected as the lectotype of H. chrysan- thum var. microcephalum. In Sectione SUBALPINA Pugsl. Hieracium diversidens Sell & C. West, sp. nov. Ab H. submuroro Lindeb. squamis capitulorum brevioribus pilis glanduliferis pauciori- bus distinguitur. Holotype: rocks at 1,700 ft, Allt Odhar, near Fortingal, v.c. 88, Mid-Perth, 18 July 1913, E. S. Marshall no. 3831 (CGE). Planta phyllopoda. Caulis ad 50 cm altus, valde striatus, sparse stellato-pilosus, inferne pilis simplicibus sparsis pallidis, superne raro pilis glanduliferis paucis brevissimis vel brevibus obscuris vestitus. Folia pallide viridia; basalia pauca; primigenia suborbicularia, mucronulata, denticulata, basi abrupte contracta; cetera dimorpha, exteriora late elliptica, mucronulata, denticulata vel leviter dentata, basi abrupte contracta, interiora elliptica vel late elliptica, obtusa vel acuta, denticulata vel irregulariter dentata, dentibus nonnunquam anguste mammiformibus, basi breve cuneata vel subtruncata; caulina O-1, bracteiformia; omnia utrinque glabra vel pilis simplicibus paucis mediocribus, margine pilis aliis simplicibus numerosis brevibus et interdum aliis glanduliferis brevissimis flaviusculis vestita; petioli pilis simplicibus densis longis flexuosis obsiti. Anthela laxa cymosa, capitulis 1-4; pedunculi suberecti, mediocres, pilis aliis stellatis numerosis, aliis simplicibus paucis brevibus vel mediocribus obscuris, allis glanduliferis sparsis brevissimis vel brevibus obscuris obsiti. Involucri squamae 9-10 mm longae, 1 (-1-25) mm latae, ante anthesin porrectae, atriusculo-virides, interiores margini- bus pallidioribus, lineari-lanceolatae, plus minusve acutae, pilis aliis stellatis sparsis, aliis simplicibus numerosis brevibus vel mediocribus obscuris, aliis glanduliferis paucioribus brevissimis vel brevibus obscuris vestitae. Ligulae saturate flavae, apice pilis simplicibus numerosis brevissimis vestitae. Styli obscuri. Cypselae obscurae, circa 3 mm longae. Watsonia 6 (5), 1967. 306 NOTES ON THE BRITISH FLORA Hieracium hastiforme Sell & C. West, sp. nov. Holotype: Cliffs, Glas Carnoch Burn [near Altguish Inn], v.c. 106, E. Ross, 24 July 1909, W. A. Shoolbred no. 716 (CGE). Planta phyllopoda. Caulis ad 45 cm altus, gracilis, valde striatus, e basi nonnunquam purpureus, pilosus, pilis simplicibus plus minusve numerosis longis pallidis (superne basi nigris), pilis stellatis paucis, et superne pilis glanduliferis sparsis brevibus vel brevissimis obscuris vestitus. Folia pallide viridia; basalia pauca; primigena nobis ignota; plurima lanceolata, raro elliptica, plurima acuta, raro obtusa, denticulata vel dentata, dentibus saepe anguste mammiformibus, basi plus minusve cuneata, petiolis longis; caulina 1-2, inferius basalibus simile, superius bracteiforme; omnia superne glabra vel pilis simplicibus paucis mediocribus, inferne pilis simplicibus paucis mediocribus, caulina inferne praecipuae pilis stellatis plus minusve numerosis omnia in marginibus scaberulis vel pilis simplicibus numerosis brevibus vel mediocribus vestita; petioli pilis simplicibus numerosis mediocribus longisque obsiti. Anthela cymosa, capitulis 1-3; pedunculi longi, suberecti, pilis stellatis numerosis, pilis glanduliferis numerosis brevibus brevissimisque obscuris, pilis simplicibus paucioribus brevibus mediocribusque obscuris vel basi nigris obsiti. Involucri squamae 12-13 mm longae, 1-1-5 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores margine pallidae, lineares, obtusae vel subacutae, pilis aliis stellatis sparsis, aliis simplicibus valde numerosis brevissimis obscuris vestitae. Ligulae saturate luteae (fide E. S. Marshall), apice glabrae vel puberulae. Styli obscuri. Cypselae non visae. Hieracium cuspidens Sell & C. West, sp. nov. Ab omnibus speciebus affinibus foliis alte irregulariter dentatis distinguitur. Holotype: Rocky gorge above a subalpine streamlet (1,400 ft), west of the road from Fortingal to Fearnan, v.c. 88, Mid-Perth, 3 July 1913, E. S. Marshall no. 3867 (CGE). Planta phyllopoda. Caulis ad 50 (ad 60) cm altus, valde striatus, parte basali plerumque purpure, pilis aliis simplicibus numerosis mediocribus longisque pallidis (superioribus basi nigris), aliis glanduliferis subnumerosis brevissimis obscuris, superne pilis stellatis sub- numerosis vestitus. Folia pallide viridia, nonnunquam purpureo-tincta; basalia pauca; primigenia suborbicularia, mucronulata, subintegra, basi rotundata; cetera plurima lanceolata, nonnunquam elliptica vel anguste elliptica, subacuta vel acuta, dentata vel inciso-dentata, dentibus inaequalibus plurimis mammiformibus vel anguste mammiformibus et saepe cuspidatis, nonnunquam ad basim dentibus longis angustis et interdum cum eis in petiolis longis descendentibus, basi cuneata; caulina 0-1, basalibus similes vel bracteiformia; omnia superne glabra vel pilis simplicibus paucis brevibus et mediocribus, inferne pilis simplicibus paucis vel subnumerosis brevibus mediocribusque, margine puberula vel pilis simplicibus plus minusve numerosis brevibus mediocribusque vestita; petioli pilis simplici- bus densis mediocribus longisque obsiti. Anthela compacte cymosa, capitulis 1-3 (—9 in specim. cult.); pedunculi mediocres, suberecti, pilis aliis stellatis numerosis, aliis simplicibus plus minusve numerosis brevibus mediocribusque basi nigris, aliis glanduliferis brevissimis vel brevibus obscuris obsiti. Involucri squamae 11-13 mm longae, 1-1-5 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari- lanceolatae, obtusae, pilis stellatis numerosis margine et apice, pilis aliis simplicibus plus minusve numerosis brevibus mediocribusque basi nigris et aliis glanduliferis brevissimis brevibusque vestitae. Ligulae saturate flavae (fide E. S. Marshall), apice puberulae. Styli obscuri. Cypselae obscurae, circa 3 mm longae. Hieracium glandulidens Sell & C. West, sp. nov. Ab H. hyparcticoide Pugsl. squamis involucri brevioribus, plerumque foliis latioribus cypselisque brevioribus differt. Holotype: near Inchnadamph, v.c. 108, W. Sutherland, 18 July 1908, E. S. Marshall no. 3295 (CGE). Planta phyllopoda. Caulis 30-60 cm altus, valde striatus, parte basali nonnunquam leviter purpureo, pilis simplicibus paucis pallidis mediocribus longisque, superne pilis Watsonia 6 (5), 1967. ee a ene ae ee he NOTES ON THE BRITISH FLORA 307 stellatis paucis et pilis glanduliferis subnumerosis brevissimis brevibusque obscuris vestitus. Folia saturate viridia; basalia plerumque pauca; primigena suborbicularia, mucronulata, subintegra, basi rotundata; cetera variabilia, late elliptica, ovata vel lanceolata, obtusa, acuta vel acuminata, glanduloso-denticulata, nonnunquam dentibus paucis mammiformibus cuspidatis, basi cuneata; caulinum 0-1, basalibus simile; omnia superne glabra vel pilis aliis simplicibus paucis mediocribus, inferne margineque et in petiolis longis pilis sub- numerosis brevibus mediocribus longisque vestita. Anthela compacte cymosa, capitulis 1-5; pedunculi longi, suberecti, pilis aliis stellatis numerosis, aliis glanduliferis numerosis brevissimis brevibusque obscuris, aliis simplicibus paucis vel numerosis, brevibus basi nigris obsiti. Involucri squamae 10-12 mm longae, 1-1-5 mm latae, ante anthesin incum- bentes, atriusculo-virides, interiores marginibus pallidioribus, lineari-lanceolatae, subacutae vel acutae, pilis stellatis margine sparsis et e apice numerosis, pilis aliis simplicibus numerosis, brevibus mediocribusque, basi nigris et aliis glanduliferis numerosis brevissimis brevibusque obscuris vestitae. Ligulae saturate flavae, apice pilis simplicibus brevissimis paucis obsitae. Styli obscuri. Alveolae receptaculi margine longidentatae. Cypselae obscurae, circa 3 mm longae. Hieracium chrysolorum Sell & C. West, sp. nov. Ab H. clovense E. F. Linton foliis majoribus saepe ellipticioribus obtusioribus emacu- latis et capitulis majoribus differt. Holotype: Midlaw Burn, Moffat, v.c. 72, Dumfries, July 1951, C. West (CGE). Planta phyllopoda. Caulis 18-30 cm altus, valde striatus, parte basali nonnunquam leviter purpureo-tincto, pilis aliis simplicibus subnumerosis longis pallidis (Superne basi nigris), aliis stellatis paucis, superne aliis glanduliferis paucis vel numerosis brevissimis brevibusque obscuris vestitus. Folia caesio-viridia; basalia plerumque numerosa; primigena suborbicularia, mucronulata, subintegra, basi rotundata; cetera suborbicularia, ovato- elliptica vel elliptica, plurima obtusa vel subacuta, raro acuta, denticulata vel grosse dentata, dentibus obtusis vel acutis mammiformibus, basi cuneata vel abrupte contracta, raro rotundata; caulinum 0-1, lanceolatum vel anguste ellipticum, acuminatum, acute denta- tum, basi abrupte contractum vel cuneatum, petiolo brevi; omnia utrinque glabra vel pilis simplicibus paucis mediocribus longisque, in marginibus petiolisque pilis simplicibus numerosis mediocribus longisque, caulina subtus pilis stellatis numerosis vestita. Anthela corymboso-cymosa, capitulis 2-6; pedunculi mediocres vel longi, suberecti, pilis aliis stellatis numerosis, aliis glanduliferis numerosis brevissimis vel brevibus obscuris, nonnunquam pilo simplice interdum mediocre basi nigro obsiti. Involucri squamae 11-13 mm longae, 1-1°5 mm latae, ante anthesin incumbentes, interiora marginibus pallidioribus, lineari-lanceolatae, obtusae vel subacutae, pilis aliis stellatis paucis, aliis glanduliferis numerosis inaequalibus brevissimis brevibusque mediocribusque obscuris, pilis simplicibus aliquando mediocribus basi nigris vestitae. Ligulae flavae, glabrae. Styli flavi. Cypselae non visae. Hieracium nigrisquamum Sell & C. West, sp. nov. Ab H. dasythrice (E. F. Linton) Pugsl. squamis involucri longioribus pilis aliis simplici- bus paucioribus obscurioribus aliis stellatis paucioribus, aliis glanduliferis numerosioribus vestitis differt. Holotype: On rocks, at about 2,500 ft, Coire nan Gall, north of Loch Laggan, v.c. 97, W. Inverness, 7 August 1916, E. S. Marshall no. 4308 (CGE). Planta phyllopoda. Caulis 30-55 cm altus, valde striatus, parte basali plerumque leviter purpureo, pilis aliis stellatis sparsis, aliis simplicibus paucis mediocribus vel longis pallidis (superne basi nigris) et superne aliis glanduliferis paucissimis brevibus vel brevis- simis obscuris vestitus. Folia medio-viridia, nonnunquam purpureo-tincta; basalia pauca vel numerosa; primigena suborbicularia, mucronulata, subintegra, basi rotunda; cetera plerumque elliptica vel late elliptica, raro lanceolata, plurima subacuta, interdum obtuso- mucronulata vel acuta; plerumque denticulata, nonnunquam integra vel ad basim dentibus paucis parvis, basi cuneata vel abrupte contracta, petiolis longis; caulina 0-1 (—2), plerumque Watsonia 6 (5), 1967. 308 NOTES ON THE BRITISH FLORA bracteiformia, sed nonnunquam basalibus similia; omnia superne glabra vel pilis simplici- bus paucis brevibus vel mediocribus, inferne et margine pilis simplicibus numerosioribus mediocribus vel longis, margine pilis glanduliferis paucis brevissimis flaviusculis, nonnun- quam utrinque pilis stellatis paucis vestita; petioli pilis simplicibus longis obsiti. Anthela laxe cymosa, capitulis 2-8, nonnunquam ramum ex axilla folii caulini superioris emittens; pedunculi plerumque longi, pilis aliis stellatis numerosis vel densis, aliis simplicibus numerosis brevibus vel mediocribus obscuris et aliis glanduliferis paucis vel numerosis brevibus vel brevissimis obscuris obsiti. Involucri squamae 10-12 mm longae, 1-1-:25 mm latae, ante anthesin porrectae vel incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari-lanceolatae, obtusae vel acutae, pilis altis stellatis paucis, altis simplicibus numerosis mediocribus vel longis obscuris, aliis glanduliferis plus minusve numerosis brevibus vel mediocribus obscuris vestitae. Ligulae pallide flavae, apice pilis simplicibus numerosis brevissimis obsitae. Styli obscuri. Cypselae obscurae, 4-4:25 mm longae. Hieracium pseudanglicoides J. E. Raven, Sell & C. West, sp. nov. Ab H. pseudanglico Pugsl. squamis involucri brevioribus foliisque angustioribus pilis stellatis vestitis differt, praeterea ab plantis parvis H. flocculosi Backh. squamis involucri plerumque brevioribus pilis aliis glanduliferis numerosioribus et aliis simplicibus pallidiori- bus vestitis, plerumque foliis angustioribus cypselisque longioribus distinguitur. Holotype: Rock ledges, corry of Ben Dourain, v.c. 98, Argyll, G.R. 27/330382, July 1953, M. C. F. Proctor & K. M. Goodway no. G3/94 (CGE). Planta phyllopoda. Caulis 10-50 cm altus, valde striatus, plerumque parte basali leviter purpureo, pilis aliis simplicibus paucis mediocribus vel longis pallidis, altis stellatis paucis, et superne nonnunquam aliis glanduliferis brevibus vel brevissimis vestitus. Folia pallide viridia, raro purpureo-tincta; basalia pauca vel numerosa; primigena late elliptica, obtuso-mucronulata, integra, basi rotundata; cetera plerumque anguste elliptica, nonnun- quam elliptica, obtusa, subacuta vel nonnunquam acuta, plerumque integra, raro dentibus parvis paucis, basi cuneata; caulina 1-2 (—3), basalibus similia vel bracteiformia; omnia superne pilis simplicibus nullis vel utrinque pilis simplicibus plus minusve numerosis mediocribus vel longis et pilis stellatis numerosis, in marginibus petiolisque pilis simplicibus numerosis mediocribus vel longis vestita. Anthela laxe cymosa, capitulis 1-2 (—10), raro ramum longum ex axillo folii caulini superioris emittens; pedunculi longi, erecti, pilis aliis stellatis subnumerosis, aliis glanduliferis numerosis brevissimis vel brevibus obscuris, aliis simplicibus subnumerosis brevibus mediocribus longisque obscuris obsiti. Squamae 10-13 mm longae, circa 1 mm latae, ante anthesin porrectae, atriusculo-virides, nonnunquam interiores marginibus pallidioribus, apice pilis aliis stellatis densis et ad basim aliis stellatis paucis, pilis altis simplicibus numerosis brevibus mediocribusque obscuris, aliis glanduliferis numerosis brevissimis vel brevibus obscuris vestitae. Ligulae pallide flavae, apice glabro vel puberulo. Styli obscuro. Cypselae obscurae, circa 5 mm longae. This species was first recognized as new, and the epithet suggested by J. E. Raven, who has joined us as a co-author in describing it. In Sectione CERINTHOIDEA Koch Hieracium magniceps Sell & C. West, sp. nov. Ab H. irico Fries pilis paucioribus foliis late illipticis differt; preaterea ab H. langwel- lense F. J. Hanb. in patibus omnibus majoribus distinguitur. Holotype: Grassy bank of the Shee Water, near the Spittal of Glen Shee, E. Perth, v.c. 89, 19 July 1959, N. D. Simpson, P. D. Sell & C. West (CGE). Planta phyllopoda. Caulis ad 60 cm altus, robustus, plerumque parte basali purpureo, pilis aliis simplicibus paucis vel subnumerosis longis pallidis nonnunquam basi nigris, aliis stellatis raris vestitus; pili glanduliferi nulli. Folia medio-viridia; basalia pauca; primigena non visa; cetera magna, late elliptica, obtusa vel acuta, leviter undulato-dentata, basi Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 309 attenuata; caulina 2-4, magna, basalibus similia; omnia utrinque et margine pilis simplici- bus dispersis nonnunquam fere setiformibus pallidis mediocribus vestita; petioli longi alati pilis simplicibus pallidis mediocribus vel longis obsiti. Anthela furcato-cymosa, capitulis 5-6; pedunculi plerumque longi, suberecti, pilis aliis stellatis numerosis, aliis simplicibus brevibus mediocribus vel longis pallidis basi nigris, aliis glanduliferis paucis brevissimis vel brevibus obscuris obsiti. Involucri squamae 12-15 mm longae, circa 1-5 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari-lanceolatae, subacutae vel acutae, pilis aliis stellatis numerosis praecipue in marginibus, aliis simplicibus numerosis pallidis basi nigris brevibus mediocribus vel longis, aliis glanduliferis paucis brevissimis obscuris vestitae. Ligulae saturate flavae, apice pilis simplicibus numerosis brevissimis obsitae. Styli obscuri. Cypselae non visae. In Sectione OREADEA Zahn Hieracium praetermissum Sell & C. West, sp. nov. Ab H. ebudico Pugsl., H. sarcophylloide Dahlst. et H. caledonico F. J. Hanb. facie simile squamis capitulorum pilis stellatis numerosioribus vestitis differt. Holotype: Limestone cliffs, 3-4 miles east of Durness, v.c. 108, W. Sutherland, 7 June 1900, E. S. Marshall no. 2336 (CGE). Planta phyllopoda. Caulis ad 40 cm altus, plerumque parte basale purpureo, pilis aliis simplicibus dispersis (raro numerosis) pallidis longis, aliis stellatis dispersis vestitus. Folia medio-caesio-viridia, plerumque superne maculis brunneo-purpureis magnis et parvis ornata, inferne purpureo-tincta; basalia numerosa; primigena suborbicularia, mucronulata, subintegra, basi rotundata, cetera ovata vel elliptica, nonnunquam lanceolata, plerumque obtuso-mucronata, nonnunquam acuta, raro cuspidata, plerumque dentata, raro denticulata, dentibus inaequalibus ascendentibus acutis apiculatis plerumque basi cuneata, sed in paucis locis basi subtruncata, nonnunquam dentibus in petiolis descenden- tibus; caulinum plerumque —1, in caule inferno basalibus simile petiolatum, raro 0; omnia superne glabra vel pilis simplicibus dispersis pallidis mediocribus raro fere seti- formibus, inferne et margine pilis simplicibus plus minusve numerosis pallidis mediocribus vestita, petioli pilis simplicibus densis longis obsiti. Anthela anguste cymosa, capitulis 2-9; pedunculi mediocres vel longi, suberecti, pilis aliis stellatis numerosis, aliis glanduliferis paucis brevissimis obscuris, aliis simplicibus paucis pallidis interdum basi nigris brevibus, pilis glanduliferis simplicibusque raro numerosioribus. Involucri squamae 12-14 mm longae, 1-1-5 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari-lanceolatae, plerumque obtusae, raro subacutae, pilis aliis stellatis numerosis praecipue ad margines et apicem, aliis simplicibus plus minusve densis pallidis basi nigris mediocribus vel longis, aliis glanduliferis paucis brevissimis vel brevibus obscuris vestitae. Ligulae saturate flavae, apice pilis simplicibus brevissimis obsitae. Styli obscuri. Receptaculi alveoli margine longe dentati. Cypselae obscurae, circa 4 mm longae. In Sectione VULGATA F.N. Williams _ Hieracium pruinale (Zahn) Sell & C. West, stat. sp. nov. _ H. murorum subsp. pruinale Zahn in Engler, Pflanzenreich 76 (IV, 280), 328 (1921). H. euprepes var. pruiniferum W. R. Linton, Brit. Hieracia, 57 (1905). Hieracium discophyllum Sell & C. West, sp. nov. Ab H. asteridiophyllo Sell & C. West foliis suborbiculariis pilis stellatis nullis differt, praeterea ab H. pellucido Laest. simili squamis capitulorum longioribus pilis simplicibus glanduliferisque vestitis distinguitur. Holotype: Pwll Byfre, v.c. 42, Brecon, 10 July 1929, H. J. Riddelsdell (CGE). Planta phyllopoda. Caulis ad 40 cm altus, robustus, plerumque parte basale purpureo, pilis aliis simplicibus interdum pallidis mediocribus vel longis, aliis stellatis dispersis Watsonia 6 (5), 1967. 310 NOTES ON THE BRITISH FLORA vestitus. Folia flaviusculo-viridia, supra maculis minutis brunneo-purpureis ornata, subter purpureo-tincta; basalia pauca vel numerosa; primigena suborbicularia, mucronulata, integra vel denticulata, basi rotundata; cetera suborbicularia, ovata vel late elliptica, obtuse-mucronulata, regulariter denticulata e basi dentibus leviter altis apiculatis, dentibus nonnunquam in petiolis descendentibus; basi rotundata vel subtruncata; caulina 1 (2), basalibus similes, petiolata; omnia superne glabra, inferne pilis simplicibus dispersis brevibus vel mediocribus pallidis vestita; petioli pilis simplicibus numerosis longis pallidis obsiti. Anthela late cymoso-corymbosa, capitulis 3-23; pedunculi mediocres, suberecti, pilis aliis stellatis densis, aliis glanduliferis dispersis brevissimis obscuris, aliis simplicibus paucis brevibus pallidis obsiti. Involucri squamae 9-11 mm longae, 1-25-1:5 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari- lanceolatae, abrupte angustatae ad apices obtusae vel subacutae, pilis aliis stellatis numerosis praecipue in marginibus apicibusque, aliis glanduliferis numerosis brevibus vel brevissimis obscuris, aliis simplicibus paucioribus brevibus vel mediocribus vestitae. Ligulae saturate flavae, glabrae. Styli obscuri. Receptaculi alveoli margine longe dentati. Cypselae obscurae, 3-5-4 mm longae. Hieracium piligerum (Pugsl.) Sell & C. West, stat. sp. nov. H. yariicolor var. piligerum Pugsl., J. Linn. Soc. (Bot.) 54, 156 (1948). Hieracium mucronellum Sell & C. West, sp. nov. Ab H. maculoso (Stenstr6m) Omang caulibus pilis simplicibus destitutis, foliis pedunculisque pilis paucioribus vestitis, et foliis praecipue obtusioribus distinguitur. Holotype: Near Bettyhill, v.c. 108, W. Sutherland, 6 July 1951, J. E. Raven (CGE). Planta phyllopoda. Caulis ad 50 cm altus, gracilis, flexuosus, valde striatus, parte basali saepe leviter purpureo, superne pilis stellatis dispersis, pilis simplicibus glanduliferis- que nullus. Folia medio- vel obscure viridia, superne maculis minutis brunneo-purpureis conspersa, inferne purpureo-tincta; basalia numerosa; primigena ovata, obtuso-mucronata, subintegra, basi rotundata; cetera variabilia, elliptica, late elliptica, ovata, oblonga vel lanceolata, obtuso-mucronulata vel acuta, leviter apiculato-dentata vel mammiformi- dentata, basi subtruncata, dentibus in petiolis nonnunquam descendentibus; caulinum 0-1, in parte basali caulis divaricatum, lanceolatum, longe acutum, petiolatum, dentatum omnia plerumque superne glabra, inferne et in marginibus pilis simplicibus dispersis mediocribus vel longis pallidis vestita; petioli pilis simplicibus densis longis obsiti. Anthela compacte cymosa, capitulis 2-8; pedunculi mediocres vel longi, suberecti, pilis aliis stellatis densis, aliis glanduliferis dispersis brevissimis obscuris, aliis simplicibus nullis. Involucri squamae 9-10 mm longae, 0-75-1 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, anguste lineari-lanceolatae, subacutae vel acutae, pilis aliis stellatis dispersis, aliis glanduliferis numerosis brevissimis obscuris, aliis simplicibus numerosis brevibus obscuris vestitae. Ligulae pallide flavae, glabrae. Styli obscuri. Recep- taculi alveoli margine breviter dentati. Cypsela obscura, 2-5-3 mm longa. Hieracium pauculidens Sell & C. West, sp. nov. Holotype: Abundant on grassy slopes of the coast, Skerray, W. Sutherland, v.c. 108, 23 June 1900, E. S. Marshall, no. 2327 (CGE). Planta phyllopoda. Caulis ad 55 cm altus, valde striatus, parte basali leviter purpureo, pilis aliis simplicibus paucis pallidis brevibus vel mediocribus, aliis stellatis paucis vel numerosis, superne aliis glanduliferis paucis brevibus obscuris vestitus. Folia medio-viridia, saepe purpureo-tincta; basalia plerumque pauca; primigena suborbicularia, mucronulata, subintegra, basi plus minusve rotundata; cetera ovata vel late elliptica, obtuso-mucronulata vel cuspidata, subintegra, minute denticulata vel raro dentibus paucis, basi inaequaliter rotundata vel subtruncata; caulina 0 vel raro 1, lanceolata, longe acuta, subintegra, basi cuneata, petiolata; omnia utrinque et in marginibus pilis simplicibus plus minusve numerosis mediocribus pallidis vestita, petioli pilis simplicibus densis mediocribus vel longis obsiti; Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 311 caulina inferne pilis stellatis dispersis vestita. Anthela compacte cymosa, capitulis 2-6; pedunculi breves vel mediocres, pilis aliis stellatis numerosis, aliis glanduliferis numerosis brevibus obscuris, aliis simplicibus paucioribus basi nigris obsiti. Involucri squamae 9-10-5 mm longae, circa 1 mm latae, ante anthesin plus minusve incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari-lanceolatae, obtusae vel acutae, pilis aliis stellatis dispersis praecipue ad marginem, aliis glanduliferis numerosis brevibus vel brevissimis, aliis simplicibus numerosis pallidis basi nigris brevibus vel mediocribus vestitae. Ligulae pallide luteae, glabrae. Styli flavi. Receptaculi alveoli margine breviter dentati. Cypselae obscurae, circa 3:5 mm longae. Hieracium variifolium Sell & C. West, sp. nov. Ab H. anguino (W. R. Linton) Roffey foliis latioribus magis dentatis capitulis numero- sioribus squamis involucri vestitis differt; praeterea ab H. prolixo Norrl. (facie simili) pedunculis pilis simplicibus numerosioribus vestitis, squamis involucri longioribus et capitulis angustioribus distinguitur. Holotype: Rocks just above the road at Coldbackie, v.c. 108, W. Sutherland, G.R. 29/ 612599, 8 June 1965, J. N. Mills & C. West (CGE). Planta phyllopoda. Caulis ad 40 cm altus, inferne nonnunquam leviter purpureus pilis aliis simplicibus paucis mediocribus vel longis, pallidis, aliis stellatis paucis, aliis glanduliferis paucis brevissimis obscuris vestitus. Folia medio-viridia, superne plerumque maculis magnis et parvis purpureo-brunneis ornata, inferne purpureo-tincta; basalia numerosa; primigena subrotunda, mucronulata, minute denticulata, basi rotundata; cetera variabilia, ovata vel late elliptica, obtuso-mucronulata vel acuta, saepe cuspidata, denticulata vel dentata, dentibus saepe anguste vel late mammiformibus, inferioribus saepe unguiculato-mammiformibus, basi inaequaliter subtruncata vel rotundata, dentibus nonnunquam in petiolis descendentibus; caulinum 0 (vel 1), lanceolatum, longe acutum, leviter dentatum, petiolatum; omnia superne glabra, inferne et margine pilis simplicibus pallidis mediocribus vestita; petioli pilis simplicibus densis longis obsiti, caulina pilis stellatis paucis vestita. Anthela anguste cymosa, capitulis 2-5 (-8); pedunculi mediocres vel longi, suberecti, praecipue acladio brevi, pilis aliis stellatis plus minusve numerosis, aliis glanduliferis numerosis brevibus obscuris, aliis simplicibus numerosis pallidibus nigri- basibus brevibus vel mediocribus obsiti. Involucri squamae 12-15 mm longae, circa 1 mm latae, ante anthesin porrectae ad apicem incurvatae, atriusculo-virides, interiores marginibus pallidioribus vel viridiusculae, anguste lineari-lanceolatae, subacutae vel acutae, pilis aliis stellatis sparsis, aliis glanduliferis numerosis brevissimis vel brevibus obscuris, altis simplicibus numerosis obscuris vel basi nigris brevibus vel mediocribus vestitae. Ligulae saturate flavae, apice puberulae. Styli obscuri. Receptaculi alveoli breviter dentati. Cypselae obscurae, circa 4 mm longae. Hieracium maculoides Sell & C. West, sp. nov. H. maculosum (Stenstrom) Omang, H. maculatum Sm. et H. glanduliceps Sell & C. West quoad faciem simulat. Ab H. maculoso ambitu et dentibus foliorum et squamis longioribus differt; praeterea ab H. maculato et H. glandulicipite squamis capitulorum pilis simplicibus numerosis vestitis distinguitur. Holotype: Scaurs above Dale Beck, Ingleborough, v.c. 64, M.W. Yorks, 7 July 1902, A. Ley (CGE). Planta phyllopoda. Caulis ad 35 cm altus, nonnunquam parte basali leviter purpureo, pilis aliis simplicibus paucis vel numerosis mediocribus vel longis pallidis, aliis stellatis dispersis vestitus; pili glanduliferi nulli. Folia pallide viridia, superne maculis magnis parvisque purpureo-brunneis ornata, inferne saepe purpureo-tincta; basalia plerumque pauca; primigena suborbicularia, mucronulata, subintegra, basi rotundata; cetera ovata, lanceolata, elliptica vel late elliptica, obtusa vel acuta, basi cuneata, denticulata vel dentata, dentibus saepe apiculatis; caulina 0-2 (—3), inferiora basalibus similia, superiora linearia vel bracteiformia; omnia utrinque et margine pilis simplicibus paucis vel numerosis pallidis Watsonia 6 (5), 1967. 3 SH NOTES ON THE BRITISH FLORA mediocribus vestita, petioli mediocres pilis simplicibus densis pallidis longis obsiti. Anthela anguste compacteque cymosa, capitulis 1-6; pedunculi mediocres vel longi, suberecti, pilis aliis stellatis densis, aliis simplicibus paucis vel numerosis pallidis nonnunquam basi nigris brevibus vel mediocribus, aliis glanduliferis paucioribus brevissimis vel brevibus obscuris obsiti. Involucri squamae 10-12 mm longae, circa 1 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, lineari-lanceolatae, subacutae vel acutae, pilis alts stellatis dispersis, aliis simplicibus numerosis mediocribus vel longis pallidis basi nigris raro aliis glanduliferis paucis brevissimis obscuris vestitae. Ligulae flavae, apice puberulae. Styli obscuri. Receptaculi alveoli margine longe dentati. Cypselae obscurae, circa 4 mm longae. Hieracium glanduliceps Sell & C. West, sp. nov. Ab H. maculato Sm. foliis basi abrupte contractis vel subtruncatis et squamis involucri pilis glanduliferis inaequalibus differt, praeterea ab H. maculoso (Stenstr6m) Omang et H. maculoides Sell & C. West facie simile praecipue foliis caulinis et squamis capitulorum pilis glanduliferis inaequalibus vestitis distinguitur. Holotype: Keld Head Scaur, Ingleton, v.c. 64, M.W. Yorks, 2 July 1903, A Ley (CGE). Planta phyllopoda. Caulis ad 40 cm altus, valde striatus, flexuous, pilis aliis simplicibus numerosis mediocribus vel longis pallidis, aliis stellatis dispersis, aliis glanduliferis nullis. Folia obscure viridia, superne maculis magnis parvisque purpureis ornata, utrinque purpureo-tincta; basalia pauca vel numerosa; primigena subrotundata, mucronulata, basi rotunda; cetera ovata, ovato-lanceolata, vel raro oblonga, obtuse mucronulata vel acuta, denticulata vel dentata, dentibus unguiculato-mammiformibus vel mammiformi- bus, basi abrupte contracta vel subtruncata; caulina 1-3, infima basalibus similia petio- lataque, superiora lanceolata vel linearia, acuta, dentata ut in foliis basalibus, sessilia; omnia superne glabra vel utrinque et margine pilis simplicibus paucis vel numerosis pallidis mediocribus vestita; petioli pilis simplicibus densis mediocribus vel longis pallidis obsiti. Anthela anguste cymosa, nonnunquam ramum longum ex axilla folii caulini superioris emittens; capitulis 1-8 (—10); pedunculi mediocres vel longi, suberecti, pilis aliis stellatis numerosis, aliis glanduliferis numerosis brevissimis obscuris, aliis simplicibus paucis vel numerosis brevibus obscuris vestiti. Involucri squamae 10-11 mm longae, circa 1 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, anguste lineari-lanceolatae, subacutae vel acutae, pilis aliis stellatis dispersis, aliis glanduliferis numerosis brevissimis brevibus et mediocribus obscuris, aliis simplicibus paucis mediocribus obscuris vestitae. Ligulae flavae, glabrae. Styli obscuri. Receptaculi alveoli margine breviter dentati. Cypselae obscurae, 3-3-5 mm longae. In Sectione TRIDENTATA F. N. Williams Hieracium sparsifrons Sell & C. West, nom. nov. H. sparsifolium var. oligodon E. F. Linton ex W. R. Linton, Brit. Hieracia, 78 (1905). H. oligodon (E. F. Linton ex W. R. Linton) Pugsl., J. Ecol. 33, 346 (1946), non Naeg. & Peter, Hier. Mittel-Eur. 2, 51 (1886). Hieracium ornatilorum Sell & C. West, sp. nov. Ab H. eboracense Pugsl. squamis involucri alio modo vestitis distinguitur. Holotype: Great Scar Limestone, Grassington, v.c. 64, M.W. Yorks, 570 ft alt., 11 August 1909, J. Cryer (CGE). Planta aphyllopoda. Caulis ad 110 cm altus, nonnunquam leviter purpureus, pilis aliis simplicibus sparsis brevibus mediocribus vel longis pallidis, aliis stellatis paucis vestitus. Folia ?obscure viridia; omnia caulina, 10-17, inferiora mediaque elliptica, obtusa vel subacuta, dentata, basi cuneata, plerumque sessilia, raro breviter petiolata; superiora ovata vel lanceolata, acuta, acute alteque dentata, basi rotundata, sessilia; omnia Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA 313 utrinque glabra vel pilis simplicibus paucis pallidis mediocribus, margine pilis simplicibus subnumerosis pallidis mediocribus vestita. Anthela divaricate cymosa, capitulis 2-12; pedunculi mediocres vel raro longi, divaricati, pilis aliis stellatis densis, aliis simplicibus paucis brevibus vel mediocribus pallidis, raro altis glanduliferis brevissimis obscuris obsiti. Involucri squamae 11-13 mm longae, circa 1-5 mm latae, ante anthesin incumbentes, atriusculo-virides, interiores marginibus pallidioribus, sublineares, obtusae, pilis aliis simplicibus paucis brevibus vel mediocribus obscuris, aliis glanduliferis paucioribus brevissimis bel brevibus obscuris vestitae. Ligulae ? saturate flavae, glabrae. Styli obscuri. Receptaculi alveoli margine ? breviter dentati. Cypselae obscurae, circa 3-5 mm longae. In Sectione UMBELLATA F.N. Williams HIERACIUM UMBELLATUM L., Sp. P/., 804 (1753). subsp. bichlorophyllum (Druce & Zahn) Sell & C. West, comb. nov. H. laevigatum subsp. bichlorophyllum Druce & Zahn, Rep. Bot. Exch. Club Brit. Isl. 9, 561 (1932 for 1931). P. D. SELL C. WEST PILOSELLA Hill, Brit. Herb., 441 (1756). The reasons for recognizing Pilosella as a distinct genus will be given in a later paper. PILOSELLA PELETERANA (Merat) C. H. & F. W. Schultz subsp. PELETERANA Hieracium peleterianum Mérat, Nouv. FI. Eny. Par., 305 (1812). P. peleteriana (Mérat) C. H. & F. W. Schultz, Flora (Regensb.) 45, 421 (1862). subsp. tenuiscapa (Pugsl.) Sell & C. West, comb. et stat. nov. Hieracium peleterianum var. tenuiscapum Pugsl., J. Linn. Soc. (Bot.) 54, 313 (1948). PILOSELLA OFFICINARUM C. H. & F. W. Schultz subsp. concinnata (F. J. Hanb.) Sell & C. West, comb. et stat. nov. Hieracium pilosella var. concinnatum F. J. Hanb., Ill. Mon. Brit. Hier., 6 (1889). subsp. OFFICINARUM Hieracium pilosella L., Sp. Pl., 800 (1753). Hieracium repens S. F. Gray, Nat. Arr. Brit. Pl. 2, 421 (1821) nom. superfi. illegit. pro H. pilosella L., non H. repens Willd., Sp. Pl. 3 (3), 1564 (1803). H. canum Vuk., Hierac. Croatica, 6 (1858), nom. superfi. illegit. pro H. pilosella L. Pilosella officinarum C. H. & F. W. Schultz, Flora (Regensb.) 45, 421 (1862), nom. nov. pro Hieracium pilosella L. subsp. nigrescens (Fries) Sell & C. West, comb. nov. H. pilosella var. nigrescens Fries, Symb. Hist. Hierac., 2 (1848). subsp."tricholepia (Naegeli & Peter) Sell & C. West, comb. nov. H. pilosella subsp. tricholepium Naegeli & Peter, Hierac. Mitt.-Eur. 1, 138 (1885). subsp. trichoscapa (Naegeli & Peter) Sell & C. West, comb. nov. H. pilosella subsp. trichoscapum Naegeli & Peter, Hierac. Mitt.-Eur. 1, 133 (1885). Pilosella flagellaris (Willd.) Sell & C. West, comb. nov., subsp. flagellaris Hieracium flagellare Willd., Enum. Pl. Hort. Berol. Suppl., 54 (1814, dated 1813). Planta ad 35 cm alta, capitulis 2-4 (7); pedunculi pilis simplicibus plus minusve numerosis ad 2 raro 3 mm longis obsiti; squamae involucri pilis simplicibus paucis vel subnumerosis ad 1:5 mm longis vestitae. Naturalized in a few localities in great Britain, widespread in Continental Europe. Watsonia 6 (5), 1967. 314 NOTES ON THE BRITISH FLORA subsp. bicapitata Sell & C. West, subsp. nov. Holotype: North-west of Hoove, Zetland, v.c. 112, 27 June 1963, W. Scott (CGE). Planta ad 18 cm alta, saepe capitulis 2; pedunculi pilis simplicibus, numerosis ad 7-5 mm longis obsiti; squamae capitulorum pilis simplicibus densis ad 2-5 mm longis vestitae. Known only from Zetland where it is apparently native. If native, the only taxon in the genus indigenous to the Islands. Pilosella lactucella (Wallr.) Sell & C. West, comb. nov. subsp. lactucella Hieracium lactucella Wallr., Sched. Crit. 1, 408 (1822). subsp. helveola (Dahlst.) Sell & C. West, comb. nov. H. floribundum subsp. helveolum Dahlst., Hierac. Exsicc. 4, nos. 14 & 15 (1891); Acta Hort. Berg. 2 (4), 13 (1894). Pilosella caespitosa (Dumort.) Sell & C. West, comb. nov. Hieracium caespitosum Dumortt., Fl. Belg., 62 (1827). Hieracium pratense Tausch., Flora (Regensb.) 11 Erg., 56 (1828). subsp. colliniformis (Naegeli & Peter) Sell & C. West, comb. nov. Hieracium collinum subsp. colliniforme Naegeli & Peter, Hierac. Mitt.-Eur. 1, 308 (1885). PILOSELLA AURANTIACA (L.) C. H. & F. W. Schultz subsp. AURANTIACA Hieracium aurantiacum L., Sp. Pl., 801 (1753). P. aurantiaca (L.) C. H. & F. W. Schultz, Flora (Regensb.) 45, 426 (1862). subsp. brunneocrocea (Pugsl.) Sell & C. West, comb. et stat. nov. Hieracium brunneocroceum Pugsl., J. Bot., Lond. 59, 67 (1921). PILOSELLA PRAEALTA (Vill. ex Gochnat) C. H. & F. W. Schultz subsp. PRAEALTA Hieracium praealtum Vill. ex Gochnat, Tent. Pl. Cich., 17 (1808). P. praealta (Vill. ex Gochnat) C. H. & F. W. Schultz, Flora (Regensb.) 45, 429 (1862). subsp. arvorum (Naegeli & Peter) Sell & C. West, comb. nov. Hieracium magyaricum subsp. arvorum Naegeli & Peter, Hierac. Mitt.-Eur. 1, 586 (1885). subsp. spraguei (Pugsl.) Sell & C. West, comb. et stat. nov. Hieracium spraguei Pugsl., J. Linn. Soc. London (Bot.) 54, 328 (1948). P. D. SELL C. WEST POTAMOGETON BERCHTOLDII X NATANS = P. X VARIIFOLIUS Thore, Essai Chlor. Dép. Landes, 47 (1803) (‘varii-folius’) (pro sp.). P. gramineus subsp. variifolius (Thore) Nyman, Consp. Fl. Eur., 682 (1882). P. natans x trichoides Hagstr., K. Svenska Vetensk. Handl., N.F. 55 (5), 193 (1916). P. javanicus subsp. variifolius (Thore) P. Fourn., Quatre Fl. Fr., 140 (1935). This remarkable hybrid was first described (as a species) by Thore from Dép. Landes in south-west France; it occurs also in the adjoining Dép. Gironde. It was unknown from outside this area until a plant apparently identical with it in all respects was discovered in Ireland by D. McClintock in August 1957, in the Glenamoy River below Glenamoy Bridge, West Mayo (v.c. H 27) (BM). The same plant was again found in the Glenamoy River by A. C. Jermy on 18 July 1962 (BM), growing in company with P. natans in a side eddy away from the fast-flowing stream. Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA a5 P. X variifolius has always been a botanical puzzle. Shortly after its original description De Candolle, in Lam. & DC., Fl. Fr. ed. 3, 3, 184 (1805), referred it to P. fluitans, as also did A. Bennett, J. Bot., Lond. 29, 75 (1891). Nyman (1882), however, treated it as a subspecies of P. gramineus, and it was placed in the synonymy of that species by K. Richter, Pl. Eur. 1, 13 (1890). Graebner, in his monograph of Potamogeton in Engler, Pflanzenr. 4 (11), 48 (1907), treated it as a ‘Planta dubia apparently related to the tropical and subtropical P. octandrus (P. javanicus); and P. Fournier (1935), following this up, went so far as to recognize it formally as a subspecies of P. javanicus. The idea that P. x variifolius is in any way closely allied to P. octandrus is quite fanciful. In fact, as Hagstrém (1916) pointed out, the plant is sterile and must be of hybrid origin. Such an origin had been suggested by Rouy, FI. Fr. 13, 304 (1912), with P. polygoni- folius and P. gramineus as putative parent species. Hagstr6m, however, rejected this sug- gested parentage, stating that the leaf characters “point in the direction of P. natans and some narrow-leaved species’. After discussing the plant further he decided that the narrow- leaved parent must be P. trichoides and he accordingly treated P. x variifolius as P. natans x trichoides. We are in full agreement with Hagstrom that P. x variifolius must be a hybrid between P. natans and one of the narrow-leaved (‘pusilloid’) species. It has floating leaves which, though much smaller, resemble those of P. natans and often show an indication of the discoloured joint at the top of the petiole which is characteristic of that species, and it has numerous narrowly linear submerged leaves which resemble the phyllodes of P. natans though more delicate in texture and with a nervation suggesting the influence of one of the ‘pusilloid’ species. Surprising though such a combination of parents may be, we can think of no other explanation for the origin of P. x variifolius; but while in accord with Hagstro6m on this point we have always had doubts about his choice of P. trichoides as the ‘pusilloid’ parent. P. x variifolius, like P. natans, has four-carpellate flowers, whereas in P. trichoides the number of carpels is normally reduced to one (rarely two or three). Hagstré6m mentioned this point, but regarded it as an instance where ‘P. natans is prevailing over P. trichoides’. Our own investigations have shown, however, that in hybrids between P. trichoides and four-carpellate species the number of carpels is reduced to 1-3, as in P. crispus X trichoides (P. x bennettii) and P. pusillus x trichoides (P. X grovesii; see below), and if P. x variifolius were P. natans x trichoides we should expect a similar reduction. The discovery of P. x variifolius in West Mayo confirms our doubts about the par- entage P. natans x trichoides, for P. trichoides is not known from Ireland, and we have no reason to suppose that the hybrid has not arisen in situ in West Mayo where it occurs along with P. natans. We conclude, therefore, that the ‘pusilloid’ parent must be a species other than P. trichoides, and the only possibilities are P. berchtoldii and P. obtusifolius, both of which have four-carpellate flowers and open stipules as in P. x variifolius. We can see no sign of any influence of P. obtusifolius in the hybrid, but the venation of the submerged leaves strongly suggests P. berchtoldii, a common and widely distributed species which occurs in West Mayo. The formula P. berchtoldii x natans is therefore here proposed for P. x variifolius. POTAMOGETON ALPINUS X LUCENS = P. X NERVIGER Wolfg. in J. A. & J. H. Schult., Mant. 3, 359 (1827) (pro sp.). P. rufescens var. nerviger (Wolfg.) Nyman, Consp. Fl. Eur., 681 (1882). P. rufescens subsp. nerviger (Wolfg.) K. Richter, P/. Eur. 1, 12 (1890). P. alpinus subvar. nerviger (Wolfg.) Ascherson & Graebner, Syn. Mitteleur. Fl., 1, 311 (1897). P. alpinus x praelongus Hagstr., K. Svenska Vetensk. Handl., N. F. 55 (5), 149 (1916) pro parte, excl. pl. angl. P. alpinus x heterophyllus Galenis, Lietuvos TSR Fl. 2, 63 (1963). The hybrid P. alpinus x lucens is known in the British Isles only from the River Fergus in Clare (v.c. H 9), where it was discovered above Inchiquin Lough by J. G. Dony Watsonia 6 (5), 1967. 316 NOTES ON THE BRITISH FLORA on 3 August 1947 (BM) and re-collected by him on 30 July 1948 (BM); it was also collected above Poplar Bridge, Killinaboy, by G. Taylor on 31 August 1962 (BM). P. X nerviger Wolfg. was originally described (as a species) from Lithuania. During the last century it was variously treated as a variety, subspecies or subvariety of P. alpinus (P. rufescens), and its affinity with that species is obvious. In 1907 G. Fischer, Ber. Bayer. Bot. Ges. 11, 46, suggested that it might be the hybrid P. alpinus x lucens though he could find no anatomical confirmation for this. Hagstrém (1916) made no comment on Fischer’s suggestion but identified Wolfgang’s Lithuanian plant with the British P. x griffithii under the formula P. alpinus x praelongus. As already explained in J. Bot., Lond. 77, 280 (1939), we believe this formula to be correct for P. x griffithii, but not for P. x nerviger, which, following Fischer’s suggestion, we accept as P. alpinus x lucens. Recently Galenis (1963) has treated P. x nerviger as a hybrid between P. alpinus and P. gramineus (P. heterophyllus), but we do not know the justification for this. POTAMOGETON CRISPUS X PRAELONGUS = P. X UNDULATUS Wolfg. in J. A. & J. H. Schult., Mant. 3, 360 (1827) (pro sp.). Dandy & Taylor, Bot. Soc. & Exch. Club Brit. Is. 12, 425 (1942); J. Bot., Lond. 80, 118 (1943). P. crispus X praelongus Casp., Schr. Phys.-6k. Ges. K6nigsberg 18, 98 (1877). P. praelongus xX crispus Ascherson & Graebner, Syn. Mitteleur. Fl. 1, 338 (1897). The hybrid P. crispus x praelongus was first recognized by Caspary (1877) from localities now in Poland. P. x undulatus Wolfg., which had already been described (as a species) from Lithuania, was later identified correctly as the same hybrid. There are authen- tic records of it also from Germany and Denmark. In the British Isles P. x undulatus was first found in 1906 in the River Lagan at Belfast, Antrim (v.c. H 39), where it was re-collected in September 1952 by G. Taylor (BM); it also occurs in the Six Mile Water at Antrim in the same vice-county. In addition to these Irish localities a Welsh station for the hybrid is now known. It was found in Llyn Hilyn, Radnor (v.c. 43), by J. A. Webb in 1938 (NMW), in company with P. crispus and P. praelongus; and it was collected there again by C. I. & N. Y. Sandwith in July 1945 (BM), along with P. crispus and P. praelongus. Potamogeton pusillus x trichoides = P. x grovesii Dandy & Taylor, hybr. nov.; inter parentes manifeste media, P. trichoidi facie similior sed stipulis basin versus tubulosis differt; a P. pusillo floribus 1—3-carpellatis praecipue distinguitur. Holotype: England, East Norfolk (v.c. 27), dike between Ingham and Palling, 11 August 1897, J. Groves (BM). P. trinervius sensu Dandy, List Brit. Vasc. Pl. 134 (1958), non G. Fisch. This apparently very rare hybrid is known to us only from a single locality between Ingham and Palling in East Norfolk (v.c. 27), where on 11 August 1897 James Groves made a voluminous gathering of ‘pusilloid’ pondweeds which he distributed extensively through the Botanical Exchange Club under the name P. trichoides, and which he recorded as P. trichoides in Bot. Exch. Club Rep. 1897, 569 (1898). The gathering does indeed contain specimens of P. trichoides, but examination of many sheets of it shows that altogether it comprises three different plants: P. trichoides, P. pusillus and their hybrid P. pusillus x trichoides. Thus every sheet of this gathering requires careful scrutiny to decide which of the three taxa are represented. P. pusillus and P. trichoides by themselves are very easily distinguishable, differing in leaf nervation, stipules, winter-buds, flowers and fruit. The hybrid, being intermediate in its characters, is of course more difficult to distinguish: in general appearance it more closely resembles P. trichoides, but it differs from that species, which has open stipules, in having the stipules tubular towards the base as in P. pusillus; it differs from P. pusillus, which has four-carpellate flowers, in having the carpels reduced in number to 1-3 as in P. trichoides. The flower spikes in the hybrid are mostly quite abortive; occasionally a drupelet is developed, but it may not have been viable. Watsonia 6 (5), 1967. NOTES ON THE BRITISH FLORA S119) An identical hybrid was later collected (with P. trichoides) on 1 August 1900 by A. and J. Bennett and C. E. Salmon in the New Cut between Stalham and Palling (GL), this presumably being the same as Groves’s original locality. The name P. trinervius G. Fisch., Ber. Bayer. Bot. Ges. 11, 29, 123 (1907), has been erroneously used for this hybrid. It was published by G. Fischer for Bavarian plants to which he attributed the parentage P. panormitanus x trichoides, i.e. P. pusillus < trichoides. He stated that there were two forms of the hybrid, one like P. pusillus (P. panormitanus) but more delicate, the other resembling P. trichoides but with the broader three-nerved leaves of P. pusillus (the leaves of P. trichoides are infact three-nerved but only the midrib is readily discernible). By the courtesy of Professor H. Merxmiiller we have been able to examine Fischer’s original material, now at Munich (M), and we find that all his specimens of the first (slender) form are in fact P. trichoides, while all those of the second (three- nerved) form are P. pusillus; one gathering, from Gaustadter Weiher, July 1905, is a mixture of the two species (the supposed parents), but we can find no trace of a hybrid. A sheet of this mixed gathering from Gaustadter Weiher and one of P. pusillus from Godeldorfer Weiher are the only sheets actually labelled with the binomial P. trinervius, and as the epithet trinervius clearly refers to the form with obviously three-nerved leaves (i.e. P. pusillus) we choose the specimen on this latter sheet as lectotype; it was collected in Godeldorfer Weiher in July 1902 by G. Fischer, and labelled ‘Potamogeton trinervius Fischer = P. panormitanus x trichoides’. It appears that Hagstr6m must have suggested the formula (though not the binomial) as there is a note ‘dt. Hagstrém’ at the bottom of the label. The name P. trinervius thus becomes a synonym of P. pusillus, and the hybrid P. pusillus x trichoides from East Norfolk is left without a binomial. We therefore name it after its discoverer, James Groves (1858-1933). J. E. DANDY G. TAYLOR DACTYLORHIZA INCARNATA (L.) S06 subsp. gemmana (Pugsl.) Sell, comb. nov. Orchis latifolia var. gemmana Pugsl., J. Linn. Soc. (Bot.) 49, 578 (1935). subsp. cruenta (O. F. Muell.) Sell, comb. nov. Orchis cruenta O. F. Muell., Fl. Dan. 5 (15), t. 876 (1782). I agree with J. Heslop-Harrison (Ber. Geobot. Forsch. Inst. Riibel 1953, 53-82 (1954)) and A. R. Clapham (Clapham, Tutin & E. F. Warburg, Fi. Brit. Is. ed. 2, 1044 (1962)) in considering the above taxa to be subspecies of Dactylorchis incarnata (L.) Vermeul. and now make the necessary transferences to the correct generic name Dactylorhiza Nevski. P. D. SELL CATAPODIUM RIGIDUM (L.) C. E. Hubbard subsp. majus (C. Presl) Perring & Sell, comb. et stat. nov. Sclerochloa rigida var. major C. Presl, Fl. Sic., xlv (1826) (based on Sclerochloa patens C. Presl). Sclerochloa patens C. Presl, Cyp. Gram. Sic., 45 (1820). Subsp. majus can be distinguished from subsp. rigidum by its taller habit, wider leaves and open pyramidal inflorescence. It occurs in the south and west of the British Isles, often near the sea, and abroad in south-west Europe and much of the Mediterranean region. F. H. PERRING P. D. SELL CATABROSA AQUATICA (L.) Beauv. subsp. minor (Bab.) Perring & Sell, stat. nov. C. aquatica var. uniflora S. F. Gray, Nat. Arr. Brit. Pl. 2, 133 (1821). Watsonia 6 (5), 1967. 318 NOTES ON THE BRITISH FLORA C. aquatica var. littoralis Parnell, Brit. Grass., 228, t. 102 (1842). C. aquatica var. minor Bab., Man. Brit. Bot., 366 (1843) (Lectotype: Sand of the sea shore, Hoylake, 24 August 1804, T. Gisborne (CGE)). Subsp. minor differs from subsp. aquatica in having shorter culms, leaves and panicles, and in the spikelets being only 1-flowered. It occurs in poor wet sandy soils near the sea at various places on the north and west coasts of Britain. No specimens have been seen from outside Britain. Watsonia 6 (5), 1967. BOOK REVIEWS Flora Europaea. Volume I. Lycopodiaceae to Platanaceae. Edited by T. G. Tutin, V. H. Heywood, N. A. Burges, D. H. Valentine, S. M. Walters and D. A. Webb, with the assistance of P. W. Ball and A. O. Chater. Pp. xxxii+-464 with 5 maps. Cambridge, University Press, 1964. Price £4 4s. Od. The publication of a European Flora was discussed for the first time in 1954 during the 8th International Botanical Congress in Paris. It is amazing that this British initiative materialized so quickly. This is no doubt due to the activities of the very energetic Executive Committee of which T. G. Tutin is president and V. H. Heywood secretary. The difficulties they have encountered during their gigantic undertaking must have been enormous. Not only did they have to find skilled collaborators, but they also had to strive for uniformity of treatment. Even before the first rough draft could be made, a number of conferences and progress reports proved to be necessary, which eventually culminated in the publication of a booklet entitled: The presentation of taxonomic information: A short guide for contributors to Flora Europaea (1958). The all-British composition of the Editorial Committee and the Organizing Committee must have been advantageous, because a frequent and close contact between their members was possible which must have contributed substantially towards the early publication and the reasonable uniformity. It is true that the Committees are supported in their task by a number of ‘advisory Editors’ and that a considerable number of prominent botanists all over Europe act as Regional Advisers or supplied geographical data and technical information, but it remains an achievement of the first order just the same. The Flora Europaea certainly meets a long-felt want in that improved methods of communication enable many European botanists to visit foreign countries more easily. There are of course many Floras, and some quite good ones among them, but several of these are not exactly up-to-date or are written in an alien tongue. Judged by contemporary standards, only very few Floras deal adequately with ‘critical’ groups, follow a sufficiently modern species concept and use a nomenclature which is not hopelessly out of date. The drawbacks are more particularly noticeable in Floras covering parts of S.E. Europe and the Mediter- ranean countries. The new Flora is arranged roughly according to the Englerian system (as used in Engler & Diels, Syllabus der Pflanzenfamilien, 11th ed., 1936). In a general preliminary survey the various families of the Spermatophyta are classified in a number of orders, but this kind of arrangement was relinquished in the case of the Pteridophyta, presumably because there is but little unanimity concerning the taxonomic significance of the various hierarchic subdivisions proposed for this group. Twenty-one families of ferns are treated under the heading Filicopsida, without the distinction of, e.g., Ophioglossales, Osmundales and Salvinales although these orders at least are recognized by several leading pteridologists. Still, the subdivision of the ferns into such a large number of families, several of which are included in ‘Polypodiaceae’ even in some recent floras, will delight many a fern specialist. Conceivably, the Flora of the British Isles by Clapham, Tutin & Warburg has served as a model for the set-up of this Flora Europaea. This is certainly no objection, because this flora is, in the reviewer’s opinion, the best European one, with very clear dichotomous keys and pithy descriptions of the species. In the Flora Europaea also a morphological description of every species is given which, though concise, is to the point. Taxa below the rank of subspecies are wisely left out of consideration. Not only would this have made the Flora much too unwieldy, but at the present state of our knowledge this might easily have given rise to much disagreement. Moreover, the relative uniformity of treatment could not possibly have been attained, because the infraspecific categories of some forms have been much more extensively studied than those of other taxa. In a book with so many co-authors, fifty-one collaborators having contributed to this first volume alone, a consistent uniformity of treatment could not possibly have been achieved, some systematists being born ‘splitters’ and others tending towards ‘lumping’. This does not only apply to the possible differences in the species concept, but also to their different ideas concerning generic delimitations which are sometimes rather broad and in other cases (and this seems to be the modern trend) are drawn much more narrowly. As an example, many botanists may find it difficult to become accustomed to the subdivision of the old ‘established’ genus Lycopodium into the four generic Huperzia, Lepidotis, Lycopodium (sensu stricto) and Diphasium. (Recently the genus Lycopodiella has been described, to which L. inundatum belongs.) Likewise, some workers may not feel too happy about the segregation of some species, formerly generally treated under Sempervivum, under the generic name Jovibarba. Consolida is considered as a separate genus apart from Delphinium but, on the contrary, Batrachium has not been separated from Ranunculus. On the whole the species concept is rather narrow. It may be a consolation to those who may not so willingly accept ‘small’ species that such a ‘modern’ classification has at least the advantage 319 Watsonia 6 (5), 1967. 320 BOOK REVIEWS of more readily becoming an incentive for new research. Most of the recent opinions are based on pains- taking and specialized biotaxonomical studies, often including an experimental approach. Nevertheless one may wonder why sometimes the specific variability is so summarily treated, e.g. under Capparis spinosa, of which species is said ‘incl. C. rupestris’, and of which the var. inermis is briefly discussed. In the author’s opinion, Capparis rupestris, though treated in several Floras as a distinct species, does not even qualify as a subspecies of C. spinosa; however, this remains a matter of personal opinion. New taxa, new names and new combinations are described previous to their publication in the Flora in the form of a series of notes in Fedde’s Repertorium (Notulae Systematicae ad Floram Europaeam spectantes), of which the first appeared in 1961. As far as can be ascertained, so much attention has been paid to the nomenclatural details, that in most cases the names have truly been brought up to date. A notable exception is the name Parietaria diffusa Mert. & Koch which is now generally thought to be conspecific with the older P. judaica L. (the ‘current’ name of many floras is P. ramiflora Moench). Such an omission can of course gladly be forgiven. Apart from the phytographical specific descriptions, the specific chromosome number, if available, is always indicated. We may expect that this information will have to be thoroughly revised after some decades when our knowledge will have considerably in- creased, not only because the hitherto unknown chromosome numbers of many plants will have become known, but also because the infraspecific variation of the chromosome number may prove to be more common than one is wont to admit at present. Cases of apogamy are not always mentioned. Among the ferns, apogamy is reported as occurring in Preris cretica, but nothing is said about the reproduction of Dryopteris borreri which most probably also propagates as an obligate apomict. Ecological data are conspicuously scarce and only supplied if the information is valid throughout the European range of the taxon concerned (and apparently only if it can be given in a concise and unambiguous form). The practical result of this editorial policy is that in the majority of the cases ecological characteristics are entirely lacking. In this respect, too, there is much room for improvement in future editions. The distribution of every species within the geographical boundaries of Europe is always mentioned, the countries generally being indicated by standard abbreviations. In this respect the data of the Flora are sometimes incomplete and there are certain omissions (e.g. concerning species of Asplenium). It is to be regretted that the distribution outside the European continent is only given in exceptional cases. In their introduction the Editors state that the extra-European areas of distribution are only mentioned in those cases in which the European range of the taxon in question is small and forms only a fraction of the total area of distribution. It must be admitted that the information concerning the extra-European distribution of European taxa is not always reliable and that in several cases taxonomic problems are immediately involved such as the status of supposedly conspecific forms and of a number of (for instance American) reputed vicariads. However, in many cases the distributional area is sufficiently known and worth mentioning. Apart from a list of basic and standard Floras, an extensive and very useful bibliography of the titles of books and periodicals cited in the text, and mainly referring to the original descriptions of the various taxa, is included. The glossary of technical terms will undoubtedly prove to be a good ‘service’ to ‘continen- tals’ and other alien botanists. In conclusion, it may be said that one must admire the achievements of the Editorial Board. The Flora will be a source of information and of inspiration. The minor errors and omissions were perhaps unavoidable and do, in any event, not detract much from the outstanding merits of this book. The present volume is of an excellent typographical quality. It is a good thing that no expenses were spared, so that the book stands a good chance of surviving the excessive handling it is surely to be submitted to. There can be very little doubt that this Flora will be gratefully welcomed by many botanists, professionals and amateurs alike, and that they will start looking forward to the appearance of the next volume. S. SEGAL Principles of Angiosperm Taxonomy. P. H. Davis and V. H. Heywood. Pp. xx-+-556, with 42 figures in the text. Oliver & Boyd, Edinburgh. 1963. Reprinted with corrections and supplementary bibliography, 1965. Price (hard covers) £4. 15s., (paperback) £3. 15s. The first thing that must be said about this book is that (as the cliché has it) it fills a long felt want. I remember clearly from my own student days feeling that, while it was (in principle at least) easy enough to learn about the results of the taxonomic activity of the last two centuries, it seemed to be nearly impossible to learn anything about the principles and methods by which these results had been achieved. Taxonomists, evidently, were born, not made. The authors of Principles of Angiosperm Taxonomy are both practising taxonomists whose work is widely known and respected. Starting from a consideration of basic taxonomic ideas, they trace the develop- ment of plant taxonomy from Theophrastus to the present day, giving an admirable outline of the contribution of pre-Darwinian taxonomists, and of the influence of evolutionary ideas on taxonomy in Watsonia 6 (5), 1967. BOOK REVIEWS S21 general, and the development of classificatory ‘systems’ in particular. They consider the taxonomic hierarchy and the units of plant classification as we have inherited them, and then return to examine the concept of characters, the construction and definition of taxonomic groups, and the evolutionary interpretation of comparative data. They give a brief and critical (but sympathetic) discussion of numerical taxonomy. Chapters 5—7 are devoted to consideration of the particular properties and problems of morpho- logical, cytological and chemical characters respectively. Chapter 8 deals with field and herbarium techniques, including a short section on typification and nomenclature, and the following chapter contains a great deal of useful information and advice on the presentation of taxonomic results. The remaining five chapters— ‘Modification of the phenotype’, ‘Variation within populations’, ‘Populations and the environment’, ‘Evolution and the differentiation of species’ and ‘Hybridization and taxonomy’—are essentially ‘biosys- tematics’ or “experimental taxonomy’, and deal with the biological background of the taxonomic situations that may be encountered in the field or the herbarium. The short section of ‘Conclusions’ is worth repeated reading. The authors take an essentially phenetic standpoint, and consider that ‘In the Angiosperms classifications based on phylogeny are unattainable except for certain special cases and then mostly at or near species level.’ I think this reflects the general consensus among present-day plant taxonomists. Indeed, I am not at all clear in just what sense any taxonomists have ever believed that their classifications were ‘based on phylogeny’. This is a catchphrase one occasionally hears vehemently defended; do its users think out its implications to their conclusions? The authors’ attitude to biosystematic evidence may be found more controversial, but here again I think we have been moving towards a consensus in the last decade or so, and that a majority of plant taxonomists will accept the authors’ propositions that “Taxonomic species represent different evolutionary situations. No assumptions about their breeding relationships are implicit in their definition, and it must not be assumed that sexual species comprise only interbreeding individuals, unless this can be demonstrated.’ A book about as large a subject as Angiosperm taxonomy is inevitably selective. I have found no serious omissions, and in general the examples are well chosen. The authors have delved widely and deep in the literature bearing on their subject, and their material is critically handled and well integrated. Other authors might well have arranged the book differently; it could be argued that a consideration of practical taxonomic units and the material of Chapters 8 and 9 would have been better placed after the biosystematic chapters. However, the logic of a book designed to be read and re-read is not necessarily the logic of an essay or a course of lectures, and the usefulness of the book is unaffected. I would recommend ‘Davis & Heywood’ unreservedly to students as a firmly based exposition of the ideas and practices of mainstream plant taxonomy in the 1960s. But obviously not all students will have the ability and willingness to tackle a long book which does not shirk the complexities or difficulties of a complex and difficult subject. There is still room for a much shorter, cheaper, tautly argued book covering some of the more essential ground of the present one. The book is well produced; one wishes it were less highly priced, though the price is not exorbitant by present-day standards. The paperback edition is hardly robust enough for its weight, and the amount of use it is likely to get, and is still expensive. Principles of Angiosperm Taxonomy will no doubt remain the standard work for a very long time to come. The authors have placed us greatly in their debt. MuGUE PROGTOR Principles of Numerical Taxonomy. R. R. Sokal and P. H. A. Sneath. Pp. 359, with 39 figures in the text. W. H. Freeman & Co., Ltd., San Francisco and London. 1963. Price £3. ‘The ideas on which numerical taxonomy rests go back to Adanson. . . . Numerical taxonomy aims to develop methods which are objective and repeatable both in the evaluation of taxonomic affinity and in the erection of taxa.’ This book sets out to provide a theoretical basis for numerical taxonomy, to review what has so far been achieved, and to provide the newcomer with a detailed step-by-step description of the procedures employed. The authors have kept their treatment broad ‘so as to be applicable to zoology, botany, microbiology, and palaeontology and other related sciences’ and hope that it will invite practitioners in these sciences to look at their material in a new way. The chapters dealing with the numerical methods assume a knowledge of elementary statistical methods; the rest of the book demands little or no mathematical background of the reader. The core of the book is an attempt to set up procedures for a maximum-attribute ‘natural’ taxonomy - on the lines suggested by Gilmour and others. The necessary stages are (a) selection of the individuals or taxa to be classified (operational taxonomic units, or OTUs), (b) selection of characters and evaluation of character-states for all the OTUs, (c) estimation of the taxonomic resemblance between all possible pairs of OTUs, and (d) construction of a taxonomic system from the information so obtained. There remain problems of identification and nomenclature, and of course the evidence may be used for other purposes such as the study of phylogenetic problems. Watsonia 6 (5), 1967. 322 BOOK REVIEWS The authors give a lucid discussion of the character concept, including the problems of dealing with various kinds of multi-state characters, of recognizing homology, and those posed by allometric growth (which impinge on the general problem of reducing complex shapes to minimal sets of independent para- meters; an important desideratum of numerical taxonomy which merits consideration at greater length). They emphasize that a reasonably large number of characters (say 50-100) must be used if the results are to be repeatable—and that even 100 characters will be only a small sample of the genotype. They review a large number of possible measures of taxonomic similarity, most of which have at one time or another been used for either ecological or taxonomic data. They conclude that the most useful are the simple matching coefficient (proportion of characters showing the same character-state for two OTUs), or correlation coefficients or Euclidian distances between OTUs after standardization of characters. The relationships between the OTUs are then examined by attractively simple cluster-analysis methods. Sneath’s single-linkage method is the simplest; OTUs are combined into clusters in descending order of their similarities with one another. In the average-linkage methods of Sokal and Michener, once the initial clusters are formed a measure of the similarity of the whole clusters to one another (and to unclustered OTUs) is used. In the successive re-computations of these measures of similarity, clusters may receive the same weight as individual OTUs, or they may be weighted proportionately to the number of OTUs they contain. Single-linkage will tend to give less clear cut answers than the average-linkage methods (but it may perhaps provide a more appropriate model of particular taxonomic situations). The results are presented as ‘dendrograms’—hierarchical schemes with the OTUs ranged along one axis and the levels of similarity (or taxonomic distance) at which these join into clusters along the other. ‘Numerical taxonomy’ can be applied not only to the ordinary purposes of taxonomy, but also to various other problems involving similarity between organisms. As the authors suggest, it could be used to examine environmental effects on a particular species, or to study phylogenetic problems, a subject to which they devote a thoughtful and stimulating chapter. The book concludes with a discussion of nomenclatural problems, a review of applications of numerical taxonomy up to 1962, a discussion of some objections to numerical taxonomy, a very brief consideration of the extraction of discriminatory features and the preparation of keys, and some speculation on the future of systematics. An appendix sets out computational methods for the procedures discussed in the book. The problem of homology has obviously given the authors much difficulty, and one feels that consider- able philosophical difficulties remain. Logically, it would be possible to reconstruct the organism completely from the full set of unit character states; for taxonomic purposes, we are replacing the organism by a model susceptible to our particular scheme of analysis. But the model may be constructed in different ways. The ‘bits’ specify the character-states, not the characters, and there seems to be no way of knowing whether we are constructing our models of different organisms in ways which are essentially equivalent for our purpose. The basic problems of correlations and discontinuity are only briefly and incidentally discussed. Characters may be correlated because they are linked to the same developmental process, because they are affected by the same environmental factor, or because they reflect different genes which tend to be associated together within or between populations. Redundancy of the first two kinds should, as far as possible, be eliminated from taxonomic data; redundancy of the third kind is fundamental to the idea of ‘natural’ classification. So too is the idea of discontinuity, and it is surely more important that (say) species- limits should follow natural discontinuities than that they should fall at a particular phenon level. In a subject developing as fast as numerical taxonomy, technical criticism can easily be unfair. Many people will probably find Williams & Dale’s (1965) discussion of a theoretical basis for numerical taxonomy more stimulating and intellectually satisfying than Sokal & Sneath’s; those authors had the advantage of reading the present book in proof. Others may be encouraged by the fact that the present authors often seem more concerned with empirical considerations than with mathematical niceties. The very powerful R-methods are scarcely considered; the authors do not pursue R-type factor or component analyses to the point of considering the distribution of OTUs in the common-factor or component space. The methods were freely available before this book was written. The Q-methods of Sokal & Sneath are appropriate for assessing levels of similarity between large numbers of diverse OTUs showing many independent directions of variation; they provide no information on discrimination. R-type principal component analysis and allied methods are appropriate where the OTUs are likely to fall into a relatively small number of rather discrete clusters, and where it is of interest to preserve something of the structure of the relationships as well as their magnitude; these methods provide information for discrimination of the clusters, but may lose information unrelated to the major directions of variation between the clusters. Discriminant analysis should surely receive more than a passing mention. The authors criticize the choice of title of The New Systematics, so they may be fairly criticized for their scanty treatment of the numerical methods most appropriate at specific and infraspecific levels in a book with the all-embracing title of their own! Watsonia 6 (5), 1967. BOOK REVIEWS 323 This would have been a far better book if it had included many more concrete examples of the application of numerical methods to actual taxonomic situations, and much less criticism of what the authors hold out as the practices of ‘conventional taxonomy’. Their criticism largely reflects one of those arguments which (like much political and religious controversy) have more to do with emotionally loaded labels than with reality, and it clouds more important issues. I suspect that a coolly rational appraisal would show rather little difference between the real working beliefs of many ‘conventional’ taxonomists and those of the authors. The main effect of phylogenetic ideas has probably been to encourage ‘natural’ (i.e. phenetic) as against artificial classification. So far as weighting is concerned, it seems to me that, in their quite proper anxiety to avoid appearing to condone any form of a priori weighting, the authors, lay altogether too little stress on the fact that a posteriori weighting is an essential goal of taxonomic activity. There is too ready a generalization from the fact that the present numerical methods give acceptable results in a few selected instances to the implied proposition that all biological taxonomy could be replaced by essentially similar methods. This book by-passes the really difficult problems that beset the plant taxonomist; adequate sampling of widely ranging taxa, plasticity, and the problem of extracting enough characters to describe the plant adequately without excessive and obvious redundancy. A group of students left to their own devices with numerical methods may well (as we have found in practice) produce a classi- fication which faithfully reflects superficial resemblance, with an emphasis on size. Numerical taxonomy does not guarantee good taxonomy. There is little doubt that (for different reasons) insects with complete metamorphosis, or micro-organisms, or even inanimate objects provide easier material for numerical taxonomy than most plants. These, to my mind, are serious flaws in an otherwise praiseworthy pioneering effort. I fear readers may too easily be provoked to one of two reactions—that numerical taxonomy as it is expounded here has all the answers, or that numerical methods are valueless. In my view both are dangerously wrong. Numerical taxonomy is not in a position to supplant ‘conventional’ taxonomy (the authors would not claim that it is), but it can contribute much to taxonomy in theoretical insights and in working tools. In appropriate circumstances numerical methods can save time and effort, and the authors are justified in their hope that numerical taxonomy can help us ‘to arrive at judgments of affinity ... without the time and controversy which seem necessary at present for the maturation of taxonomic judgments’. The taxonomist’s time is expensive as well as the computer’s! No doubt, with characteristic resilience, taxonomy will absorb into everyday use those parts of numerical taxonomy which are found to be of general value—as in the past it has absorbed and been enriched by the ideas of Darwinism or the ‘New Systematics’. To sum up, here is a book about which one has reservations, but which is nevertheless required reading, and will inevitably stand as something of a classic. One greatly hopes for a second edition in which the reservations can be dispelled. REFERENCE WILLIAMS, W. T. & DALE, M. B. (1965). Fundamental problems in numerical taxonomy. Ady. Bot. Res. 2, 35-68. M. C. F. PROCTOR Phenetic and Phylogenetic Classification. Systematics Association Publication No. 6. Edited by V. H. Heywood and J. McNeill. Pp. xi-+164. The Systematics Association, London. 1964. Price 15s. Phenetic and Phylogenetic Classification is a teport of the symposium held by the Systematics Association under this title in Liverpool in April 1964. It includes the text of the papers presented, together with extracts from the discussions which concluded the sessions. The contributors, well chosen, of very varied backgrounds and views, include some distinguished names. A controversial subject, the exigencies of time and space, and an obvious atmosphere of enthusiasm probably all contributed to the clarity and directness with which authors have presented their views. Every taxonomist will find plenty here to disagree with or argue about; this book would make an excellent basis for a series of discussions or seminars in a university taxonomy course. More or less ‘conventional’ views are put forward by Burtt and Blackwelder. Burtt produces some provocative examples of classificatory problems in flowering plants; are Astragalus and Oxytropis, separated by a single character, ‘good’ genera? Blackwelder talks some robust common sense about taxonomic principles and methods; I particularly liked his list of ‘acceptances’. Delevoryas is avowedly phylogenetic in his approach, and discusses the influence of palaeobotanical evidence on classifications of vascular plants. Introducing the session on numerical taxonomy, Sneath outlines succinctly the basic steps in his own numerical approach. Boyce presents a useful comparison of the results of a (Q) principal component Watsonia 6 (5), 1967. 324 BOOK REVIEWS analysis and cluster analyses of data from recent and fossil primate skulls. Edwards & Cavalli-Sforza discuss taxonomic aims, and go on to give a very interesting account of the interpretation of modern taxonomic evidence (on blood groups) in terms of a minimal phylogenetic model. Silvestri & Hill give a valuable discussion of basic general problems (ranging from the taxonomy of DNA chains to that of higher plants and animals); in considering the weighting of characters they outline briefly some conceptually interesting numerical techniques. Kendrick proposes a solution to a tiresome scoring problem. Meeuse, in his critique of numerical taxonomy, rightly emphasizes practical considerations; but these could be pursued further, and one is left feeling that the basic questions remain unanswered. Olson includes some assertions which will provide ready material for argument, and some valuable results from a study of character correlation groups in certain fossil vertebrate skulls, whose taxonomic implications will repay careful thought. The discussions are thoroughly worth reading, and have an immediacy and cogency often lacking in reports of this kind. The editors are to be congratulated on a well- (and promptly-) produced book. M. C. F, PRocTOR Botany: A Laboratory Manual. T. E. Weier, C. R. Stocking and J. M. Tucker. Ed. 3. Pp. viii+165. John Wiley & Sons Ltd., New York and London. 1964. Price £1 3s. Od. This book is designed for use in colleges in the United States. It is probable that it is satisfactory there where botany is taught rather differently. One wonders, however, why the publishers have bothered to bring it to the United Kingdom where teaching methods and ideas are such that most teachers will not readily find a use for it. Much of the material is of about ‘A’ level standard; a few exercises might be used in first-year university courses. Some of the illustrations are of a reasonable standard; others would be considered inadequate at any level. Fig. 14.3 (p. 39) ‘Leaf morphology’ is particularly poor. Here, leaf venation is illustrated by drawings of three leaves all of which are unnecessarily small—so small that the one intended to show parallel venation has completely indistinct veins. On page 119 appears: ‘Diagram the life cycle of ... as directed by the instructor.’ This sums up the whole book. With a competent instructor it is unnecessary. J. TIMSON The Arlington Practical Botany. Book I. Plant Anatomy. Mary-Anne Burns. Pp. 93, 84 x 11 in. 230 drawings. Arlington Books, London. 1964. Price 15s. Od. This book which is the first of a projected series of six is intended for students studying for ‘A’ level G.C.E. It is a laboratory guide and should also be useful for revision before examinations. It serves its purpose well and can be recommended for use in schools. It might also be of interest to amateur botanists who would like some knowledge of the internal structure of plants. The drawings are the main feature of the book and of quite a high standard although it is a pity that the ‘two groups of chromosomes’ on p. 9 are not more clearly drawn as there is always a tendency for students to draw these structures badly. The two drawings of leaf structure on p. 71 would have been much improved by being on a larger scale. It is un- fortunate that the Latin names of plants are not in italics and that in discussing the various theories of shoot apex structure no dates are given with the authors’ names. Theseare, however, relatively minor criticisms which could easily be avoided in any future edition and in the other volumes of this series. If these volumes main- tain the standard of Book I, the series will be a valuable addition to the books available for students at this level. J. TIMSON Flowers of the Mediterranean. By Oleg Polunin and Anthony Huxley. Pp. xii+257 with 32 plates in colour. Chatto & Windus Ltd., London. 1965. Price £2 2s. Od. There can be no doubt of the need for a good popular guide to the flora of the Mediterranean for the use of British tourists, and this is the public for which this book is intended. It provides pictures and descrip- tions to help them to identify some of the more conspicuous plants they find, and it should also assist them to take a real interest in the flora. The introductory chapter on the vegetation of the Mediterranean, though far too short, is just what the intelligent visitor needs, and the notes about uses, history, biblical and literary references, etc. added to the descriptions, and especially those on plants of economic importance, provide one of the most valuable features of the book. The descriptions cover over 700 species in 193 pages. The wording is appropriate for the reader for whom it is intended, but there is little attempt to emphasize diagnostic characters of special importance, and it is unlikely that users will wade through all the characters to enable them to distinguish the plants described from the hundreds of others which find no place in the book. Watsonia 6 (5), 1967. BOOK REVIEWS 525 Over 300 species are illustrated in colour with some of the best reproductions of flower transparencies which have yet appeared. Many of these are exceedingly beautiful and they include pictures which cannot be praised too highly. They show the plants growing in their natural habitat, and to people who already know them they will bring back many happy memories, but a as guide to identification they are far inferior to a good series of drawings in colour. The drawback to colour transparencies, as every experienced lecturer knows, is that several are needed to illustrate each species. A good picture of the flower will seldom show fruit; separate pictures have to be taken to show flower, fruit, leaves, habit, and various details and to reproduce all these would be too expensive. The identification of some of the colour pictures in Flowers of the Mediterranean (the Genistas for example) would defeat even the most experienced botanist. A further 127 species are illustrated on 28 pages of line drawings, and there are also sketches of fruits and other details in the text. These are by Mrs. Barbara Everard and are outstanding for their accuracy and success in conveying a true impression of the plants depicted. In attempting to include a representative selection of the common conspicuous plants to be found in all parts of the Mediterranean coast the authors have set themselves an impossible task. They have tried to solve it by exercising a strong bias towards flowers found in Greece and the Aegean, Turkey, Lebanon and Palestine, and adding some species found only elsewhere. The book is likely to prove a very much more useful guide for the relatively small number of tourists visiting the eastern part of the region, than for the vast numbers taking the cheaper holidays in Spain, France, Corsica, Sardinia, Italy and Yugoslavia. These will find many abundant conspicuous flowers not mentioned in the book at all. To give reasonable coverage for the ambitious title at least three books of this size are needed, and these should be on a regional basis so that the tourist needs to carry only one at a time. It is to be hoped that Messrs. Polunin and Huxley will follow this up with similar works for parts of the western Mediterranean, but in the meanwhile they have earned our gratitude by providing this inexpensive and well illustrated book which is a joy to own. J. E. LOUSLEY Illustrations to the Flora of Delhi. J. K. Maheshwari. Pp. xx-+282, with 278 line-drawings. Council of Scientific and Industrial Research, New Delhi. 1966. Price Rs. 28.00 (£2 16s. Od.; $8). Of the 942 species described in the Flora of Delhi, by the same author, which was published in 1963, only 278 are illustrated in the present work. However, when one considers that 411 of the species in the Flora are cultivated, and that the illustrations are all of indigenous or naturalized species, it is seen that the ratio is not as one-sided as it at first appears. The figures are set out one species to a page, and each is cross- referenced to the-main work. A scale is provided for each figure, and in many cases floral dissections and/or enlargements are also displayed. The family to which each species belongs is indicated at the top of the page, and the number of each figure corresponds to the page number. The illustrations are, on the whole, quite clear and sufficiently diagnostic to enable identifications to be made, although they do vary in quality to a certain extent; not only this, but they also vary somewhat in style which is rather singular since in his preface the author clearly states that the drawings are the work of but one artist, Shri D. M. Sonak. It is, for example, not easy to imagine that the carefully delineated and well-proportioned figure of Vicoa vestita on p. 102 proceeded from the same pen as the rather slapdash figure of Gomphrena celosioides which appears on p. 179. Several plants well known to British botanists appear in this work, so that even in the first twenty pages we come upon such familiar species as Ranunculus sceleratus, Coronopus didymus, Sisymbrium irio, Stellaria media, Arenaria serpyllifolia and Spergula arvensis. It thus appears that any botanist from these shores visiting the Delhi area for the first time would have quite a good nucleus to work from as a basis for getting to know the local flora. In his introduction the author enumerates some nomenclatural changes which have taken place since the publication of the Flora, and he also lists thirty-seven additional species recorded from Delhi State in the intervening three years—which would suggest that a certain amount of work remains to be done on the plants from the environs of India’s capital. Although somewhat limited in scope, this set of illustrations may be said to constitute a useful supplement to the Flora, and a worthwhile contribution to the steadily-growing volume of literature on the plants of the Indian subcontinent as a whole. A. R. SMITH Watsonia 6 (5), 1967. AS An) ex ann Peak Vee) 7 RRS j agy i) PROCEEDINGS OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Published Twice Yearly Price [5/- per part Obtainable from D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANDy for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf of a sub-committee of the Society appointed for this purpose. Sir George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and have become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8 vo., 176+xvi pages. Bound in cloth. Price : 10/- (postage 1]-). Please send cash with order Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. hes \ CAMBRIDGE A Dictionary of the Flowering Plants and Ferns J. C. WILLIS Seventh edition This edition has been completely rewritten and considerably enlarged. The 40,000 entries cover all generic names, both valid and invalid, published since Linnaeus (1753), and published family names from 1789. Each accepted genus is referred to its family. The geographical distribution is indicated together with the number of species. Each accepted family is described; its principal genera and, frequently, an outline of its classification are given. For the botanist and horticulturist this book is a necessity as the most authoritative, up-to-date reference work of its kind. CAMBRIDGE UNIVERSITY PRESS IRISH NATURALISTS’ JOURNAL A Magazine of Natural History Published Every Quarter by the I.N.J. Committee Edited by Miss M. P. H. KERTLAND, M.Sc. with the assistance of Sectional Editors Annual Subscription, 10/- post free Single Parts, 3/6 All communications to be addressed to:— The Editor, Department of Botany, Queen’s University, Belfast W384 &¢ WATSONIA JOURNAL OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Editors: C. D. K. Cook M. C. F. Proctor N. K. B. Robson Vol. 6 JUNE, 1968 Pt. 6 CONTENTS THE BROMUS MOLLIS AGGREGATE IN BRITAIN. By PHILIP SMITH a e 327-344 CHROMOSOME NUMBERS OF ORNITHOGALUM UMBELLATUM L. FROM THREE LOCALITIES INENGLAND. By R.CZAPIK .. ee Ais ee 345-349 COMPUTER MAPPING OF SPECIES DISTRIBUTION IN A COUNTY FLORA. By J. G. HAWKES, B. L. KERSHAW and R. C. READETT.. eh e ie 350-364 PARIETARIA OFFICINALIS AND P. JUDAICA. By C. C. TOWNSEND be: os 365-370 THE HYBRIDS OF MIMULUS CUPREUS. By R. H. ROBERTS oh Ha ae 371-376 THE INTERSPECIFIC RELATIONSHIPS OF JUNCUS EFFUSUS AND J. CONGLOMERATUS IN BRITAIN. By A.D.Q. AGNEW .. te ee a Pa ae 377-388 NOMENCLATURAL NOTES ON THE GENUS ANTHYLLIS L. By J. CULLEN .. ee 389 Book REVIEWS a a es ne a ne os a ve 390-396 ERRATA a A a, cea ae cue ue ae Be ae nie 397 INDEX TO VOLUME 6 ee ue ee 300-409 PUBLISHED AND SOLD BY THE ; BOTANICAL SOCIETY OF THE BRITISH ISLES Bs c/o DEPARTMENT OF BOTANY, BriTIsH MusEUM (NATURAL History), LONDON PRICE: TWENTY-FIVE SHILLINGS BOTANICAL SOCIETY OF THE BRITISH ISLES PATRON HER MAJESTY QUEEN ELIZABETH THE QUEEN MOTHER OFFICERS President: Dr. J. G. DONY Vice-Presidents: Miss U. K. DUNCAN, Prof. J. G. HAWKES, R. MACKECHNIE and KE. MILNE-REDHEAD Hon. General Secretary: D. E. ALLEN, c/o Dept. of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. Hon. Treasurer: J. C. GARDINER, Thrift House, 12 & 14 Wigmore Street, London, W.1. Hon. Editors: (Watsonia) Dr. M. C. F. PROCTOR, Hatherly Biclogical Laboratories, Prince of Wales Road, Exeter. Dr. C. D. K. COOK, Hartley Botanical Laboratories, The University, Liverpool 3. Dr. N. K. B. ROBSON, Dept. of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. (Proceedings) E. F. GREENWOOD, M.Sc., c/o City of Liverpool Museums, William Brown Street, Liverpool, 3. Hon. Meetings Secretary: Mrs. MARY BRIGGS, White Cottage, Slinfold, Sussex. Hon. Field Secretary: (vacant) Hon. Membership Secretary: Mrs. J. G. DONY, 9 Stanton Road, Luton, Bedfordshire. Applications for membership should be addressed to the Hon. Membership Secretary, from whom copies of the Society’s Prospectus may also be obtained. WATSONIA Price to non-members, 25/— per part. To be obtained from D. H. Kent, 75 Adelaide Road, West Ealing, London, W.13. SUBSCRIPTIONS The present rate of subscription is £1 10s. per annum for Ordinary Members, 15/— for Junior Members, and 10/— for Family Members, and the Society’s year runs from January Ist to December 3ist. All subscriptions should be paid te the Hon. Treasurer. ADVERTISEMENTS All enquiries for advertising space in the Society’s publications should be addressed to D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. THE BROMUS MOLLIS AGGREGATE IN BRITAIN By PHILIP SMITH Department of Botany, University of Edinburgh ABSTRACT Experimental studies on taxa within the Bromus mollis aggregate (sensu Tutin) have revealed that plants referable to B. thominii sensu Tutin are hybrid derivatives of crosses between B. mollis L. and B. lepidus Holmberg. The name B. x pseudothominii Philip Smith (hybr. nov.) is given to this material. It is demonstrated that the selection pressures favouring the persistence of these hybrid forms are associated with their success in contaminating sown pasture grass seed. The name B. thominii Hardouin should be applied to a dune species which is shown to be genetically and serologically distinct from inland forms of B. thominii sensu Tutin. It is proposed to include the dune material, together with B. mollis sensu stricto and B. ferronii Mabille, in a single species for which the correct name is B. hordeaceus L. Subspecific rank is given to these component taxa of B. hordeaceus, which therefore become: B. hordeaceus subsp. hordeaceus (syn. B. mollis L.); B. hordeaceus subsp. thominii (Hardouin) Hylander; and B. hordeaceus subsp. ferronii (Mabille) Philip Smith. INTRODUCTION Tutin (1962) includes Bromus mollis L., B. thominii Hardouin and B. ferronii Mabille in his conception of the B. mollis aggregate. These plants are closely similar morphologically, and seem to have strong affinities with at least three other taxa, namely B. molliformis Lloyd, B. interruptus (Hackel) Druce and B. lepidus Holmberg. Variation in the field in the B. mollis agg. is frequently puzzling since all variants of B. mollis and ‘B. thominii’ often grow together in the same sward, and their taxonomic separation is sometimes apparently arbitrary. Though several features are used to distinguish them (Tutin 1962), in practice the variation in lemma and spikelet sizes is the only character which can be used to dis- criminate between the two taxa. Both B. mollis and ‘B. thominii’ are ruderal plants of inland areas, sometimes found in maritime habitats, but B. ferronii is a cliff-top plant of very restricted distribution. It is known only from a few localities in the south and west of Britain, and western France. B. mollis and ‘B. thominii’ appear to be widely distributed in Europe and western Asia and are introduced in most other parts of the world, often becoming well established. An enquiry into the treatment of these taxa by different authors reveals a wide variety of opinions as to their taxonomic rank and nomenclature. Difficulties also arise because the application of two Linnaean names, B. mollis and B. hordeaceus, is uncertain. Discussion of Linnaeus’ interpretation in the literature (e.g. Munro 1861; Hackel 1884; Holmberg 1924) has further confused the position. My investigations into these problems, as part of a large-scale experimental treatment of the genus Bromus, were begun in 1961, and the present account relates to my first results bearing on the B. mollis aggregate. Experimental work with B. interruptus is not yet com- plete, and the relationship of this curious species to the rest of the British bromes will be discussed in a subsequent paper. The results of my work require some nomenclatural changes which are presented first so that the correct names can be used through the rest of the paper. FORMAL TAXONOMY OF THE MATERIAL Nomenclature and synonymy Results of examining the bromes of the Linnaean herbarium in London (LINN) (Smith 1965) indicate that while the modern conception of the species B. mollis L. (e.g. in Hubbard 1954, and Tutin 1962) is correct, it is not certain what Linnaeus meant by his name B. hordeaceus (Species Plantarum ed. 1, 1753). Munro (1861), Hackel (1884) and Holmberg 327 Watsonia 6 (6), 1968. 1 328 PHILIP SMITH (1924) have each provided partial explanations of Linnaeus’ meaning, which are unfor- tunately based on mutual misunderstandings and are incorrect. By comparing the Linnaean treatments of Bromus in the Flora Suecica ed. 1 (1745), Species Plantarum eds. 1 and 2 (1753 and 1762) and Flora Suecica ed. 2 (1755), one can trace a sequence of change in Linnaeus’ ideas about these plants (Smith 1965). Comparison with the Linnaean herbarium provides further information. Linnaeus’ criteria for distinguishing species of Bromus are restricted to laxness of panicle, gross spikelet shape, spikelet indumentum and whether the awns are straight or reflexed. Using them he had no difficulty in distinguishing a lax-panicled plant which he named B. secalinus from a plant with an erect, dense panicle which he named B. hordeaceus L. (Flora Suecica ed. 1, 1745). This treatment is repeated in the first edition of the Species Plantarum (1753), which is the earliest place of valid publication. In Flora Suecica ed. 2 (1755) Linnaeus transferred B. hordeaceus to the status of variety 8 of B. secalinus, retaining the epithet ‘hordeaceus’. (A printer’s error makes it appear as $8; since B. secalinus already had one variety, it should correctly have been var. y.) Linnaeus explains this change: ‘Bromus hordeaceus in hortis satus transit in Bromum secalinum, cujus varietas est, ex solo arido et duro pygmaeus.’ Clearly cultivation of a depauperate accession of B. hor- deaceus L. in the richer soil of a garden had produced a specimen approaching the vigour of B. secalinus. A further evolution of Linnaeus’ treatment of these bromes is marked in Species Plantarum ed. 2 (1762) which is the last evidence we have of Linnaeus’ intentions. ‘Var. 8 hordeaceus’ still appears under Bromus secalinus, and there is a new entry—B. mollis. From the Linnaean phrase name: ‘Bromus panicula erectiuscula, spicis ovatis aristis rectis, foliis mollissime villosis’ it is highly likely that the plant referred to is a soft brome, at present called B. mollis L. His description is expanded by a footnote: ‘Similis B. secalino sed albido-pubescens. Genicula crassiora, cylindrica, subtus inprimis valde hirsuta. Folia utrinque mollis sima vaginis striatis. Panicula brevius pedicellata.’ Of the six sheets of Bromus in the Linnaean herbarium referable to the B. mollis agg. sensu Tutin (Smith 1965), sheet 93.5 bears two specimens labelled B. mollis by Linnaeus (fide Spencer Savage in the Linnean Society’s Catalogue of the Linnaean Herbarium). Sheet 93.7 bears a depauperate tussock of a soft brome which though totally unlike B. secalinus has been so labelled by Linnaeus. This plant has been named ‘B. nanus’ by Weigel. Together with evidence from his published accounts these specimens indicate that both B. mollis L. and B. hordeaceus L. applied to a plant within the B. mollis agg. sensu Tutin. Linnaeus transferred B. hordeaceus to varietal rank beneath B. secalinus after culti- vation of a specimen resembling that on sheet 93.7. It is possible that this plant was his original collection. This explanation is supported by all the alterations of synonymy throughout the four Linnaean works cited (Smith 1965). If B. hordeaceus L. and B. mollis L. are conspecific, the former is the earlier and thus the correct name of the species. The genotypically depauperate material from dunes described later in this paper appears to be correctly named B. thominii Hardouin, judging from material collected and named by Hardouin in the herbaria at Kew (K) and Paris (P). Some of this material is illustrated in Plate 14a. B. thominii Hardouin replaced the earlier hononym B. arenarius Thomine-Desmazures (1824) non Labillardiére (1804). No type of either B. thominii or B. arenarius Thomine-Desmazures was cited in the original descriptions. Thomine’s herb- arium material seems to have been lost. His figure (Thomine-Desmazures 1824) plainly indicates material similar to Hardouin’s B. thominii. Wendelbo (1956) indicated that he thought B. thominii sensu Tutin (Tutin 1962) differed from French material collected and named by Hardouin. The depauperacy of Linnaeus’ B. hordeaceus has been used by many taxonomists as an indication that the name should apply to small plants like B. thominii Hardouin, and not to B. mollis L. If the depauperacy was only phenotypic and resulted from growth in hard, dry soils as Linnaeus reported (Flora Suecica ed. 2, 1755), then genetically dwarfed Watsonia 6 (6), 1968. | ; BROMUS MOLLIS AGG. IN BRITAIN 329 material should not be included in B. hordeaceus L., if it is regarded as specifically distinct from more robust specimens (i.e. B. mollis L.). The evidence in this paper that plants from sand dunes are genetically depauperate is reinforced by the work of Holmberg (1924), Nilsson (1931, 1937) and Wilson (1956). The two earlier workers regarded this material as specifically distinct and mistakenly named it B. hordeaceus L. The evidence above requires that, if such distinct status is accorded to these plants, then the name must be B. thominii Hardouin. Weigel published the name B. nanus in 1772 (Observationes Botanicae) and refers Linnaeus’ sheet 93.7 (LINN) to it. Weigel’s conception of this species probably included depauperate material of a number of related species. Sheet 93.7 material probably grew in an exposed place, in black, well-drained soil. It has rather short lemmas and short, weakly divaricate awns, and would be referred to B. hordeaceus subsp. thominii in the key given below. As explained above, if all these bromes are included in one species, then the name B. hordeaceus has priority over B. mollis. In view of the differences between the taxa, discussed in the rest of the paper, I believe that B. mollis L. (as in Hubbard 1954, and Tutin 1962), B. thominii Hardouin non sensu Tutin, and B. ferronii Mabille are best regarded as sub- species of a single species. Hylander (1945) has also concluded that the dwarfed form from dunes should be regarded as a subspecies in this group, basing his conclusion on Nilsson’s (1931, 1937) genetic work, which is discussed below. He has already made the necessary combination: B. hordeaceus L. subsp. thominii (Hardouin) Hylander. My conclusions from the evidence presented below are that B. lepidus is a distinct species, and that B. thominii sensu Tutin is a collection of hybrid derivatives (nothomorphs) which are, however, so common that a suitable binomial is needed for reference. The name B. xX pseudothominii Philip Smith is therefore proposed. Clearly it is very difficult to describe the limits of what is by definition a segregating hybrid complex. Some variants of B. xX pseudothominii are very close to pure B. lepidus, others resemble pure B. hordeaceus subsp. hordeaceus. The evidence of the artificial hybridizations reported in this paper, and of field collections, indicate that the commonest variants resemble B. hordeaceus more closely than they do B. lepidus. The key is written with this fact in mind. In most bromes there is variation in the development of spikelet indumentum from pubescence to glabrosity and this is acknowledged by the recognition of taxonomic forms and varieties based on the indumentum variation alone. It is an equivocal point as to whether the hairy spikelet form of the typically glabrous B. Jepidus is a genuine expression of intraspecific variation, or whether it may result from a flow of genes for hairiness from B. hordeaceus. The same problem exists for glabrous spikelet varieties of B. hordeaceus. It is not possible to know whether the types of these varieties are pure intraspecific variants or the products of intro- gressive hybridization and therefore synonyms of B. x pseudothominii. It is likely that such varieties contain genetic material from both sources and therefore it seems wiser to regard them as only pro parte synonyms of B. x pseudothominii. The correct nomenclature of the material considered in this paper, together with a selected synonymy, is listed below. B. HORDEACEUS L., Sp. Pl., 77 (1753). subsp. HORDEACEUS Be molis ©., Sp. PI. ed. 2, 112 (1762). subsp. THOMINI (Hardouin) Hylander, Upps. Univ. Arsskr, 1, (7) 84 (1945). 2B. nanus Weigel, Obs. Botan. 8 (1772). B. arenarius Thomine-Desmazures, Mém. Soc. Linn. Calvados 1824, 40 (1824) non Labill., Novy. Holl. Pfi. 1, 23 (1804). B. thominii Hardouin, Congrés Sc. Fr. 1, 56 (1833) non sensu Tutin in Clapham et al., Flora of the British Isles, ed. 2, 1152 (1962). Watsonia 6 (6), 1968. 330 PHILIP SMITH subsp. ferronii (Mabille) Philip Smith, comb nov. B. ferronii Mabille, Act. Soc. Linn. Bord., 25, 612 (1864). B. LEPIDUS Holmberg, Bot. Notiser, 1924, 326 (1924). B. gracilis Krosche, Feddes Rep. 19, 329 (1924) non Leysser, F/. Hal. 16 (1761). B. britannicus I. A. Williams, J. Bot. Lond., 67, 65 (1929). Bromus x pseudothominii Philip Smith, hybr. nov. B. mollis L. var. leiostachys Hartman, Handb. Skand. Fl. ed. 2, 33 (1832) pro parte. B. gracilis Kroésche, var. micromollis Krosche, Feddes Rep. 19, 329 (1924) pro parte. B. lepidus Holmberg, f. lasiolepis Holmberg, Bot. Notiser 1924, 326 (1924) pro parte. B. thominii sensu Tutin in Clapham et al., Flora of the British Isles ed. 2, 1152 (1962) non Hardouin, Congres Sc. Fr., 1, 56 (1833). Annuum vel bienne. Culmi graciles, erecti, 20-70 cm alti, vaginis foliisque pubescentibus. Panicula laxa vel contracta, erecta, 1-10 cm longa. Spiculae lanceolatae, glabrae vel pubescentes, 10-15 mm longae, breviter pedicellatae. Glumae inaequales, prominenter nervosae, inferior lanceolata, superior ovato-lanceolata. Lemma chartaceum, ovato- lanceolatum, 6-5-8 mm longum, palea longius, prominenter nervosum, marginibus angustis hyalinis plerumque obtuse angulosum. Arista recta, 3-7 mm longa. Palea bicarinata, nervis ciliatis. Antherae 0-2-1-5 mm longae. Caryopsis palea aequilonga. A B. hordeaceo subsp. hordeaceo lemmate breviore saepe marginibus acutius angulosis differt. A B. hordeaceo subsp. thominii culmo altiore erecto panicula majore laxiore, pluri- spiculataque differt. A B. lepido \lemmate longiore marginibus obtusius angulosis, parte hyalina angustiore, palea caryopsim aequante differt. Holotypus: Roadside near Queen Elizabeth Hospital, Birmingham, England. Coll. 1962, P. Smith 458. In Herbarium of the Royal Botanic Garden, Edinburgh (E). Key to taxa 1. Awns stout, recurved in fruit. Lemmas 6-5-8-5 mm long, bluntly angled. Spikelets densely hairy, in a dense panicle. (Coastal cliffs). 2. B. hordeaceus subsp. ferronii Awns straight, or if weakly divaricate then the culms prostrate or procumbent. Lemmas 4-5—11 mm long, sharply or bluntly angled. Spikelets hairy or glabrous 2 2. Lemmas 4:5-6:5 mm long, margins sharply angled with a broad hyaline margin. Awns weak. Grain exceeding palea. 4. B. lepidus Lemmas 6:5—11 mm long. 3 3. Culms 1-8 (-12) cm, prostrate or procumbent to ascending. Lemmas 6-5-7-5 mm long. Awns sometimes divaricate in fruit. Grain shorter than palea. (Sand dunes). 3. B. hordeaceus subsp. thominii Culms 3-80 cm, normally erect. Lemmas 6-5-11 mm long. Awns straight, erect. Grain shorter than or equalling the palea. 4 4. Lemmas 6-5-8 mm long, usually glabrous, with broad or narrow, normally bluntly angled, hyaline margins. Grain usually equalling the palea. 5. B. x pseudothominii Lemmas 8-11 mm long, usually hairy, with narrow, bluntly angled hyaline margins Grain shorter than the palea. 1. B. hordeaceus subsp. hordeaceus Descriptions 1. B. HORDEACEUS L. subsp. HORDEACEUS Annual or biennial, 3-80 cm. Culms erect or rarely ascending, usually stout, but Watsonia 6 (6), 1968. BROMUS MOLLIS AGG. IN BRITAIN 573) slender in depauperate material. Panicles 5-10 cm, erect, lax or rather dense, or reduced to a single spikelet. Spikelets 12-25 x 4-6 mm, usually hairy, lanceolate. Glumes unequal, lower lanceolate, upper ovate. Lemmas 8-11 mm, ovate, bluntly angled, with a narrow hyaline margin. Awn stout, straight, erect, 4-11 mm long. Palea shorter than lemma. Anthers 0-:2—2-0 mm. Grain shorter than palea. (2n=28). Europe, W. Asia, N. Africa. Introduced and naturalized in many parts of the world. 2. B. HORDEACEUS L. subsp. ferronii (Mabille) Philip Smith Annual or biennial, 2-15 (-20) cm. Culms erect or ascending, usually stout. Panicles 2-5 cm, erect, dense, often simple, sometimes reduced to a single spikelet. Spikelets 8-18 x 3-6 mm, densely hairy, ovate-oblong. Glumes unequal, lower lanceolate, upper broadly ovate. Lemmas 6:5-8:5 mm, rather broad, bluntly angled, with a narrow hyaline margin. Awn stout, flattened at base, recurved in fruit, 2-5-5 mm long. Palea shorter than lemma. Anthers 0:2—1-5 mm. Grain shorter than palea. (2n=28). W. Europe. Cliff-tops. 3. B. HORDEACEUS L. subsp. THOMINII (Hardouin) Hylander. Annual, 1-8 (-12) cm. Culms prostrate, procumbent or ascending. Panicles 1-3 cm, erect, dense, often simple, or reduced to a single spikelet. Spikelets 8-12 x 2-3 mm, hairy or glabrous, lanceolate. Glumes unequal, lower lanceolate, upper ovate. Lemmas 6:5—7:5 mm, ovate, bluntly angled, with a narrow hyaline margin. Awn weakly divaricate in fruit, 3-7 mm long. Palea shorter than lemma. Anthers ca. 1:0 mm. Grain shorter than palea. (Qn=28). W. Europe, probably elsewhere. On coastal dunes and similar habitats inland. 4. B. LEPIDUS Holmberg. Annual or biennial, 4-70 cm. Culms erect, often rather slender. Panicles 2-10 cm, erect, rather narrow, lax or dense, or reduced to a single spikelet. Spikelets 5-15 « 2-4 mm, usually glabrous and glossy, lanceolate. Florets tightly imbricate at first and giving a green and white striped appearance to the spikelet, later slightly divaricate. Glumes unequal, ovate with broad hyaline margins. Lemmas (4-5-) 5-5-6-5 mm, ovate or broadly ovate, sharply angled, with a broad hyaline margin. Awn straight, erect, weak, setaceous, 2-5-5 mm long. Palea shorter than lemma. Anthers 0-5-2 mm. Grain exceeding palea, rarely exceeding lemma. (2n=28). Introduced in Britain in ruderal habitats, perhaps also introduced in W. Europe. Native distribution unknown. 5. B. X pseudothominii Philip Smith (Bromus hordeaceus L. x Bromus lepidus Holmberg) Annual or biennial, 20-70 cm. Culms erect. Panicles 1-10 cm, erect, lax or dense, sometimes reduced to a single spikelet. Spikelets 10-15 x 3-4 mm, hairy or glabrous, lanceolate. Glumes unequal, lower lanceolate, upper ovate-lanceolate. Lemmas 6-5-8 mm, usually bluntly angled, with a narrow hyaline margin. Awn straight, erect, 3-7 mm long. Anthers 0:2-1-:5 mm. Grain usually equalling the palea. (2n=28). Europe, probably elsewhere. MATERIALS AND METHODS (i) Seed sources The accessions of Bromus species used in this work originate from a wide geographical area. British material used in serological and genetic experiments is listed in an Appendix. Morphological studies were made on plants from many parts of the world, but especially Watsonia 6 (6), 1968. 352 PHILIP SMITH from Britain. The Flora of Warwickshire Bromus collection was placed at my disposal, and I am indebted to Messrs. Anton Nielsens, of Vejle, Denmark, for samples of Danish- produced forage grass seeds. All accessions used in experiments are being maintained in a living condition, together with voucher specimens, in the author’s Bromus collection at the Department of Botany, University of Edinburgh. (ii) Morphological studies The herbaria of the Department of Botany, University of Birmingham (BIRM), of the Birmingham Natural History Society (Midland Institute, Birmingham), and of the Royal Botanic Gardens, Kew (K), have been extensively consulted to determine the limits of natural variation. The Birmingham collections included the recent Flora of Warwickshire material, so that an up-to-date idea of field variation was obtained. Specimens have been received on loan from the herbaria of Oregon State University, Corvallis (OSC), the Naturhistorisches Museum, Vienna (W), and the Muséum National d’Histoire Naturelle, Paris (P). I am indebted to the Curators of these institutions for their co-operation. Data from herbarium material was used to construct scatter diagrams to demonstrate the extent of variation in some selected morphological features. Owing to the paucity of collections of B. ferronii, there has been a deficiency in the amount of B. ferronii material available for scoring. (iii) Karyological and Genetic Studies The chromosome numbers of cultivated accessions were determined. Attempts at hybridization were made using a variety of techniques. Crossing technique proved to be an obstacle to progress and most success was achieved using the laborious method of hand pollination described by Beddows & Davies (1938). Since the plants are all self-compatible, as was evident from good seed-setting in enclosed panicles, emasculation was essential. Maternal panicles were enclosed in pergamene bags. (iv) Ecological Studies Enquiries were restricted to observations of the various habitats in which the taxa were commonly found, and how they were associated with one another. Further informa- tion was supplied by analysis of the ecological records of the Flora of Warwickshire survey. (v) Serological Studies Serological techniques have been used in an attempt to determine the relationships between the taxa principally involved in this account. Although it has not yet been possible to apply these methods to B. hordeaceus L., B. lepidus and B. x pseudothominii as com- prehensively as is desirable, the preliminary results seem to contribute valuable corrobora- tion to my hypothesis of the evolution of this group of species developed from other criteria. Serological methods applied to many other bromes in this section of the genus (Smith 1965) have proved to be a valuable adjunct to morphological and cytological methods in determining patristic relationships. Hence the present data are presented with confidence. The serological techniques adopted were basically those of Gell, Hawkes & Wright (1960), involving the use of double diffusion and immuno-electrophoretic analysis of the plant proteins, in agar gels. In their work on Solanum, tuber proteins were used, but in the present work the seed proteins were employed. Storage proteins seem to offer the best material for serological comparison, since greater inter-specific differences may be expected in them than in proteins concerned with activities such as photosynthesis, which are perhaps fundamentally similar in all green plants. Kloz (1960) has demonstrated the greater inter- specific differentiation of the storage proteins in a study of Phaseolus species. Kloz surmises that the reasons for this may be that the storage proteins are phylogenetically older, but it is hard to find evidence for this view. It seems more likely that storage proteins in seeds Watsonia 6 (6), 1968. BROMUS MOLLIS AGG. IN BRITAIN 95 are under particularly severe selective pressure from the environment because during the critical establishment phase when the survival capacity of a plant is most searchingly tested, variation in the composition, size and shape of stored protein molecules might well affect the success or failure of a plant competing with others. Since storage protein acts largely as an amino-acid reserve, the precise sequence of the residues may not be so rigidly fixed as in enzymes carrying out specific tasks in green tissues. So it may be that selective forces have a wider range of viable mutant molecules to operate on, and hence produce a com- paratively rapid divergence between species. Storage proteins seem therefore to be the plant serologists’ material of choice, and have the merit of being available in high concentration at a particular stage in the life of the organism. Kloz (1960) was working with globulins, whereas the storage protein of grasses comprises the prolamine and glutelin fractions of the seeds. Unfortunately, the solvents of prolamine and glutelin proteins (ethanol and alkaline solutions respectively) cause non-specific precipitation of the serum proteins in the antisera used to determine interspecific differences in protein precipitation. Therefore in the present work it was necessary to restrict the serological analysis to the water and saline soluble fractions, i.e. the albumins and globulins. It must be borne in mind that serological differ- ences which may be demonstrated in these are likely to be smaller than the differences which exist in the prolamine and glutelin fractions. 0-8 per cent. saline extracts of finely ground, dehusked seeds of Bromus species were compared by raising antisera to them in rabbits. Courses of up to nine injections of the extracts were administered over a period of about six weeks. The earlier injections were of a 1:1 mixture of a 1 per cent. protein solution (the antigen solution) and Freund’s adjuvant (Freund & Bonanto 1944) and were intramuscular. Later injections were of unadulterated saline extract into the marginal veins of the ears, or into the muscle of the thighs. 100-150 mg. of plant protein was injected during the course of the immunization. The blood serum, containing antibodies to Bromus proteins, was used to analyse the antigen content of extracts of the seeds. The antiserum and the antigens were allowed to diffuse towards one another in 0-6 per cent. agar gel, buffered at pH 8-6 by barbital-HC1 buffer (Michaelis 1930). Lines of precipitated protein (“‘precipitin’) representing the reaction of the antibodies with antigens were developed after 24 hour incubation at 30°C. The differences in pattern produced by the reaction of antigen preparations from different species were regarded as taxonomic criteria. For immuno-electrophoresis, the seed ex- tracts were electrophoresed in agar buffered at pH 9-4 (barbital-HC1) for eight hours (1-2V/cm, 20mA). After this time the antiserum was allowed to diffuse towards the now separated antigenic proteins, to produce an immuno-electropherogram. After 24 hour incubation, the precipitin lines, each representing an antigen-antibody complex, were sufficiently well-developed to be observed and drawn. After a further three days’ develop- ment, the buffer salts and soluble proteins were washed out of the agar, and the agar itself was dried on a glass plate, protected by a piece of filter paper. When drying was complete the paper was removed and the pattern of lines was stained in 0-2 per cent. Ponceau S in 30 per cent. trichloracetic acid. After further drying, contact prints or photographs could be made from the finished preparations. RESULTS (i) Morphological The distinguishing features of the taxa in the B. mollis agg. sensu Tutin (1962) and of the related species B. Jepidus are listed in Table 1. Plate 145 shows the gross differences in floret size and shape between B. hordeaceus, B. lepidus and ‘B. thominii’ sensu Tutin (B. X pseudothominii). Herbarium material was scored for these features and the results are presented on a scatter diagram (Fig. 1). Recent Warwickshire material is plotted separately in Fig. 2. Floret characters were measured on the lowest floret in the uppermost spikelet of the panicle. Some of the features showed no consistent pattern of variation, and have Watsonia 6 (6), 1968. 334 PHILIP SMITH thus been omitted from the scatter diagrams. Anther-length varied widely within each taxon, as did stem-width. Spikelet-width and -length appeared very strongly correlated with lemma-length, while awn-length showed a lower correlation. As suggested by Tutin (1962) B. hordeaceus subsp. ferronii has quite wide spikelets, but my measurements showed that they were no wider than the spikelets of some specimens of B. hordeaceus subsp. hordeaceus. TABLE 1. Distinguishing features of B. hordeaceus subsp. hordeaceus, subsp. thominii, subsp. ferronii, B. X pseudothominii and B. lepidus (All measurements are in millimetres. Stem width was measured 1:5 cm below the first node of the inflorescence.) B. hordeaceus Character subsp. subsp. subsp. B. X pseudo- hordeaceus thominii ferronii thominii B. lepidus 1. lemma-length 8-11 6-5-7°5 6:5-8-5 6-5-8 (4-5-)5-5-6-5 2. awn-length 4-11 3-7 weakly 255 3-7 2-5:5 straight divaricate recurved straight straight 3. spikelet-length 12-25 8-12 8-18 10-15 5-15 4. spikelet-width 4-6 2-3 3-6 3—4 2-4 5. anther-length 0-2-2 ca. 1-0 0:-2-1°5 0-2-1-5 0-5-2 6. stem-width 0-5-1-5 0-5-1 0-5-1-25 0-5-1-25 0-5-1 7. spikelet usually hairy hairy or densely hairy hairy or hairy or indumentum glabrous glabrous glabrous 8. hyaline lemma- narrow narrow narrow usually broad margin narrow 9. lemma shoulder bluntly angled bluntly angled bluntly angled bluntly angled sharply angled 10. length of grain grain shorter grain shorter grainshorter grain about grain longer relative to palea equal to palea In both scatter diagrams it will be seen that the spread of points exceeds the variation range quoted by Tutin (1962) and by Hubbard (1954) for these features, emphasizing the extreme morphological variability of the group as a whole. Though few plants of subsp. ferronii were scored (none, of course, from Warwickshire) they probably fall into quite a distinct group with dense panicles, recurved awns and densely hairy lemmas. They do not appear to be as distinct from B. hordeaceus subsp. hordeaceus as is B. lepidus, though both subsp. hordeaceus and B. lepidus are connected by a whole range of variants which approach the Tutin definition of ‘B. thominii’, but which intergrade into both the former taxa. The indicated distribution of spikelet pubescence is consistent—a gradual increase of hairiness with increasing lemma size. The existence of names for hairy and glabrous variants of all the taxa concerned, except subsp. ferronii, accords with this finding. It seems reasonable to postulate from these scatter diagrams that some hybridization may be occurring between B. hordeaceus subsp. hordeaceus and B. lepidus, which is either obscuring a discrete taxon ‘B. thominii’, or which may be the sole source of bromes so named. The consistent spread of spikelet hairiness from B. hordeaceus subsp. hordeaceus to B. lepidus may indicate introgression of genes for hairiness from subsp. hordeaceus into B. lepidus, and factors for glabrosity in the reverse direction. (ii) Ecological Field collecting has indicated that pure stands of both B. hordeaceus subsp. hordeaceus and B. lepidus can be found but that mixed populations of both these taxa and ‘B. thominii’ are more frequent. Often the herbarium sheets examined proved to bear mixed gatherings Watsonia 6 (6), 1968. BROMUS MOLLIS AGG. IN BRITAIN 585) of two or three of these taxa. B. hordeaceus subsp. hordeaceus collections often included material with shorter lemmas, referable to ‘B. thominii’ sensu Tutin. Pure stands of B. hordeaceus subsp. hordeaceus were found in woods, in hayfields and occasionally on road- sides. Pure stands of B. /epidus were almost invariably associated with sown grassland of one sort or another, for example, sown sheep pasture, lawns, and newly sown roadside verges and embankments. In these latter habitats, B. Jepidus was quite commonly found mixed with ‘B. thominii’ and sometimes with B. hordeaceus subsp. hordeaceus as a third component of the brome flora. A mixed population of B. lepidus and ‘B. thominii’-type variants grows on an area of the Knocking Hoe National Nature Reserve in Bedfordshire, parts of which were sown after the war (J. G. Dony, personal communication). A curious population growing in disturbed soil on a cereal field margin at Ufton, Warwickshire, was scored for various morphological features, and the results are plotted in Fig. 3. Variation in lemma-size, sharpness of lemma shoulder, and spikelet hairiness place the plants of this population in a position intermediate between B. hordeaceus subsp. hordeaceus and B. lepidus. No B. lepidus could be found nearby at the time of sampling, but several large patches of B. hordeaceus subsp. hordeaceus grew within thirty metres of this population. According to the key in Tutin (1962) such plants would be referable to ‘B. thominii’, but the recognition of such intermediate populations as a separate species requires some experimental validation. Material of ‘B. thominii’ accession 121 was found to segregate slightly in experimental cultivation, producing some plants with shorter lemmas in the second experimental genera- tion. Such segregation supports the idea that ‘B. thominii’ populations are composed of plants of hybrid origin. Most samples of ‘B. thominii’ came true from seed, however, and showed no variation between generations which could be attributed to segregation. Further insight into the ecological position of the taxa in Britain comes from an analysis of the copious ecological records of the new Flora of Warwickshire survey (Hawkes & Readett 1963). Fig. 4 represents the distribution of the three taxa, B. hordeaceus subsp. hordeaceus, B. lepidus and ‘B. thominii’ sensu Tutin, in the sixteen habitats in which they have so far been recorded. It can be seen that while all three taxa are commonest on road- sides, in the other habitats their distribution appears most interesting. In cultivated land B. lepidus is commonest, followed by ‘B. thominii’, with B. hordeaceus subsp. hordeaceus a poor third. B. lepidus and ‘B. thominii’ are much commoner in pasture land than is B. hordeaceus subsp. hordeaceus, while in rough grassland B. lepidus is much less common than ‘B. thominii’ and B. hordeaceus. B. lepidus is commonest on waste ground, and, where B. lepidus is not recorded from a ruderal habitat, ‘B. thominii’ is commoner than B. hor- deaceus subsp. hordeaceus. It is perhaps worth emphasizing that the identifications on which these ecological records are founded were all supported by specimens from the various collectors, and were confirmed by the author, in collaboration with the Flora Identification Panel. From these data it appears that both B. lepidus and ‘B. thominii’ are associated with artificially sown grassland. Both are poor pasture grasses which are never deliberately sown in Britain, so that the most likely reason for their distribution seems to be that they occur as contaminants in the seeds which are deliberately sown. To investigate this hypo- thesis, samples of commercial grass seeds have been inspected for contamination by bromes. These samples were purchased from seed firms in Britain, and a valuable collection of Danish strains, both cleaned and uncleaned samples, was made available by the firm of Anton Nielsens, Vejle. Bromus seeds found in imported cleaned seed of Danish Festuca pratensis Huds. were referable to B. Jepidus and ‘B. thominii’, while in an uncleaned sample of F. pratensis from another producer, B. hordeaceus subsp. hordeaceus seed was seen. Clearly this shows that the seed size and shape of B. /epidus and ‘B. thominii’ are so similar to those of meadow fescue that seed cleaning methods do not remove them, though the larger seeds of B. hordeaceus subsp. hordeaceus are removed with less difficulty. B. lepidus has also been found in uncleaned seed of Chewings Fescue (Festuca rubra subsp. commutata Gaud.), and in cheap, cleaned lawn grass seed consisting largely of Lolium perenne L. One Watsonia 6 (6), 1968. PHILIP SMITH 336 ‘snpida ‘g pue ‘un, nsuas “33¢e SI]JOW “{ IY} JO SUOIJSOTJOO OATYSYOIMIVAA JUSIII JO WIZLIP 19}}k9S °7 “BI (ww) y}Hue|-ewwe| I 0! 6 8 L 9 S y 0 (uw) yjHuaj-ume *[elIoyeU WINTIegJoYy WOIJ poyidui0s ‘snpida] ‘gq pue “UN, Nsuos “33e s1jou “g OY} JO UIeISEIP 10}}89G "| “3Ly (Ww) yyHuej-euwa; ll ol 6 8 L 9 S v0 6 | d} o Ww 2 fe) ¥ (ate ae Cee ¢ o w d onmege earitre b ° od e ¥ Y¥ 2 O° Sirs. wy i2= 00 CO ede 9 a @ °o Oo ry 9 5 (oe) Bo (6) 2 fo) 0d 00 @ L O O fe) fe) g 00 O O 0 6 Ol Watsonia 6 (6), 1968. Sai) BROMUS MOLLIS AGG. IN BRITAIN “OTYSYOIMIVAA ‘UO, Wor UONLNdod “se syjjou “_ Jo UILISIP 19}}89G “¢ “SIA (ww) y}Guej-ewwe| 61 6 8 L 9 S y 0 ¢ ¢ Japjnoys Dwway pajbup Aydsoys------- 6d » S ‘ ¥ SUMD paxa}jas-*°-*: d b VW cot dw S 5 5 wibsow Durway euijDAy‘poosq:---- ‘eo BSS00 Wy C aS a es eS 0d Sou a yajaxtds snojqo\b OO Cave OG @ oho) (ox: Moka: } 9 = Che (ef =} yayayids Autoy Aiyysis<---- o a yajayids. Autpy we te wee O L Watsonia 6 (6), 1968. Percentage of species records B. lepidus M-ro M-hr MwaM-ry C-f C-g G-m'G-p G-r RU RUW RUG AUb WS WOscWO-m Warwickshire habitats Fig. 4. Histogram of distribution of B. hordeaceus subsp. hordeaceus, B. thominii sensu Tutin and B. lepidus in all Warwickshire habitats in which they have been recorded. The percentages are of the total number of records of each taxon in the Flora of Warwickshire Survey. Key to abbreviations: M-ro—roadside G-m—meadow RU-q—quarry M-hr—hedgerow G-p—pasture RU-b—brick or marl pit M-wa—wall G-r—rough grassland WS-p—pondside M-ry—railway embankment or cutting RU-f—farmyard WO-sc—scrub or derelict wood C-f—cultivated field RU-w—waste ground WO-m—mixed wood C-g—garden BROMUS MOLLIS AGG. IN BRITAIN Bev landscaped site in the University of Birmingham grounds, which bore a rich brome flora of three taxa, B. lepidus, B. hordeaceus subsp. hordeaceus and ‘B. thominii’, with all spikelet indumentum variants, was found to have originated from a sown mixture including an impure source of Festuca longifolia Thuill. Investigation of other samples of cleaned F. longifolia revealed B. lepidus as a contaminant. Beddows (1953) shows that large-scale importations of rye-grass and other grass seed into Britain date from about 1831, so that B. lepidus and ‘B. thominii’ may have entered Britain in quantity from about this time. The earliest British collection of B. lepidus is dated 1836. Early agricultural literature on grass husbandry (e.g. Sturrock 1865; Everitt 1897) indicates that imported grass seed was very impure and needed considerable cleaning. Britain is now a major producer of grass seed, and advice on this very specialized trade is given to farmers by the National Institute of Agricultural Botany, Cambridge. A series of Herbage Seed Growers’ Leaflets has been produced and additionally a series of Seed Notes. Seed Notes No. 72 shows that Bromus is a recognized weed in rye-grass and fescue crops, and the Herbage Seed Growers’ Leaflet No. 2 advises farmers to remove from the field and field margins any rogue grasses whose seed may be difficult to remove from the crop. It seems clear that B. lepidus and ‘B. thominii’ owe much of their British distribution to the practice of sowing grass seeds in pastures and on roadsides. British Certified Seed, or its foreign equivalent, of very high purity is commonly used, and in this brome contam- ination is at a minimum, though still present. However, much cheap seed is used by road- side contractors and local councils, and this is often impure. Methods of seed cleaning by screening for seed size will explain the strong tendency for small-seeded bromes to occur on cultivated ground among sown grasses of similar seed size. B. hordeaceus subsp. hordeaceus can more easily be removed, and this species seems more commonly established on old roadsides and hayfields than either B. lepidus or ‘B. thominii’. Some seed contamination by B. hordeaceus subsp. hordeaceus is rendered possible by the occasional maturation of smaller seeds in the upper part of the spikelet. The above discussion relates only to the inland members of this aggregate of species, but it is a common observation that ‘B. thominii’ occurs, often abundantly, on coastal dunes where there is no likelihood of artificial sowing. B. /epidus is uncommon here, moreover. B. hordeaceus subsp. hordeaceus occurs on coastal cliffs and on some dunes. It is sometimes depauperate and occasionally has procumbent culms, but these characters have not been found to persist in cultivation (Wilson 1956). ‘B. thominii’ from sand dunes has a rather distinct morphology not manifested in the scatter diagrams owing to lack of material. These plants, which are here regarded as B. hordeaceus subsp. thominii, are usually very small and often prostrate or procumbent, with small, dense paucispiculate panicles. The awns often become divaricate. Experiments revealed that these features were maintained in cultivation, recalling the similar experience of Nilsson (1937) and Wilson (1956). The dune material used came from the Ynys-Las dune system in Cardiganshire and from Burry Port, Carmarthenshire. (iii) Serological Using antisera raised to two accessions of B. hordeaceus subsp. hordeaceus (Nos. 13 and 21 in appendix), the albumin and globulin fractions of the seed proteins of taxa within the B. mollis aggregate (sensu Tutin) and B. Jepidus were studied and compared. Both double diffusion and immunoelectrophoretic analysis of the antigen components in the saline extracts gave compatible results. The precipitin spectra obtained are illustrated in Pl. 155 and c. The homologous reaction of B. hordeaceus subsp. hordeaceus with its own antiserum cannot be distinguished from that of B. lepidus nor from inland accessions of ‘B. thominii’. Other Bromus species tested, most of which do not form part of the present investigation, gave reactions varying from similar to very different from the homologous reaction, with the exception of B. interruptus which resembled B. hordeaceus subsp. hordeaceus completely in its serological behaviour. B. hordeaceus subsp. ferronii and subsp. thominii from dunes Watsonia 6 (6), 1968. 340 PHILIP SMITH both showed slight though persistent differences from the homologous reaction. In both these taxa, one antigen line is lacking and the main antigen moves more slowly in electro- phoresis than that of the homologous reaction, and that of B. Jepidus and inland ‘B. thominii’. These immunological differences were detected by the antibody-forming mechanisms of four adult rabbits. (iv) Karyological and Genetic B. hordeaceus (all subspecies), B. lepidus and ‘B. thominii’ were found to have a somatic chromosome number of 2n=28 and are tetraploid. Using the technique described by Beddows & Davies (1938) it was possible to establish a number of hybrids of B. hordeaceus subsp. hordeaceus and B. lepidus. These hybrids resembled ‘B. thominii’ sensu Tutin in all their morphological features, and were almost completely fertile. A typical hybrid panicle is shown in Pl. 15a. The F, generation, raised from the seeds of the hybrids, proved to be highly variable and to include types which would be referable to B. hordeaceus subsp. hordeaceus, B. lepidus and ‘B. thominii’. Plate 14c illustrates some of the variations in floret size which were produced. Hairy and glabrous spikelets occurred throughout the range of variation. A fully fertile hybrid of B. hordeaceus subsp. hordeaceus and subsp. ferronii was also obtained, which was morphologically intermediate between the parents, DISCUSSION AND CONCLUSIONS Morphological studies suggested that plants referable to B. thominii sensu Tutin might be the result of hybridization between B. hordeaceus subsp. hordeaceus and B. lepidus. Field observations showed the common occurrence of mixed populations of all three taxa in fields and on roadsides. B. hordeaceus subsp. ferronii and ‘B. thominii’ from coastal dunes seemed rather more distinct, but inland ‘B. thominii’ and B. lepidus were shown to be associated with sown grassland. Examination of grass seed samples revealed that both occurred as contaminants in the sown seed of various species of pasture grass. Their seed- size so closely matches that of a number of rye-grasses and fescues that they cannot readily be separated by normal mechanical methods. A similar situation has been reported in the weed Camelina by Tedin (1925). Contamination of the seed must occur in the production field, though examination of the uncleaned seed purity records at the National Institute of Agricultural Botany failed to demonstrate the expected high contamination. This may readily be explained. Firstly, the sample sizes used by the N.I.A.B. testers may be too small to reveal the comparatively small number of brome seeds required to initiate substantial contaminating populations. Secondly, it seems likely that the records of contamination by ‘Bromus mollis’ (B. hordeaceus subsp. hordeaceus), of which there are a good many, actually refer also to the smaller seeded bromes. Thirdly, the incidence of contamination may be a regional or climatic variable depending on methods of cultivation and cleaning and on the amount of effort expended in certifying the seed. In the present work it was obvious that seed of Danish origin and seed from southern Britain were frequently significantly contam- inated with small-seeded bromes. Gooch (1963) shows from an analysis of purity records of imported Danish meadow fescue grass that 69-8 per cent. of samples were contaminated with “Bromus mollis’ (B. hordeaceus subsp. hordeaceus), while 40-7 per cent. of creeping red fescue samples were contaminated. A contaminated sample was one which contained at least one contaminant seed in the portion analysed (though the size of this sample is not defined). Contamination of rye-grass by B. hordeaceus subsp. hordeaceus was found in only 7-8 per cent. of Danish samples. In samples of English origin, Gooch found B. hordeaceus subsp. hordeaceus contamination in 17-1 per cent. of samples of Lolium perenne, and 20-4 per cent. of samples of Festuca pratensis. B. hordeaceus contamination was almost twice as frequent in samples from western counties of England as in those from eastern areas. Astonishingly high incidence of contamination—76 4 per cent. and 96-5 per cent. respect- tively—was recorded in Irish samples of Italian and perennial rye-grass. Watsonia 6 (6), 1968. ee ee eee eects Pas ele ‘ Oo ee bac big prenville el Cabwo wt - ts ™"( eghoades ! 2 Fe = Srer’d. Meoguin - Tendon S ~ 2 / a PLATE 14a—Authentic material of B. thominii Hardouin (collected and named by Hardouin), in herb. Muséum National d’Histoire Naturelle, Paris (P). r ttt PLATE 14b—Florets of B. hordeaceus subsp. hordeaceus (left); B. thominii sensu Tutin (centre); and B. lepidus (right). pypeernene PLATE 14c—Range of variation in florets of the F, generation of the cross B. hordeaceus subsp. hordeaceus x B. lepidus. [facing page 340 PLATE 15a—Panicle of artificial F; hybrid of B. hordeaceus subsp. hordeaceus and B. lepidus. WUE ELL Cc d a PLATE 15b—Immunoelectropherograms of seed protein extracts of: (a) B. hordeaceus subsp. Hordeaceus: (b) B. thominii sensu Tutin; (c) B. lepidus; (d) B. hordeaceus subsp. thominii (Hard.) Hylander. PLATE 15c—Double diffusion spectra of seed protein extracts of: TOP (left to right) B. thominii sensu Tutin; B. hordeaceus subsp. hordeaceus; B. hordeaceus subsp. thominii (Hard.) Hylander; B. thominii sensu Tutin; B. interruptus; B. hordeaceus subsp. thominii. BOTTOM (left to right) B. thominii sensu Tutin; B. lepidus; B. hordeaceus subsp. thominii; B. thominii sensu Tutin; B. hordeaceus subsp. ferronii; B. hor- deaceus subsp. thominii. Note missing inner lines from B. hordeaceus subspp. thominii and ferronii. BROMUS MOLLIS AGG. IN BRITAIN 341 Thus there seems to be abundant evidence of the origin of the small-seeded annual brome populations from contaminated seed in sown grassland. Serological evidence showed that B. hordeaceus subsp. hordeaceus and B. lepidus were indistinguishable in terms of their seed albumins and globulins, implying therefore a very close patristic relationship. Inland *B. thominii’ was also indistinguishable, thus sup- porting the idea that it may result from hybridization, and possibly introgression, between B. hordeaceus subsp. hordeaceus and B. lepidus. B. hordeaceus subsp. ferronii and subsp. thominii sensu stricto were slightly different, which agrees with their morphological and ecological distinctness from the typically inland forms. Genetic evidence showed that B. hordeaceus subsp. hordeaceus and B. lepidus were interfertile tetraploids, thus genetically closely related, and that they could hybridize to produce highly fertile progeny indistinguishable in morphology from inland material of ‘B. thominii’. | It is therefore suggested that inland ‘B. thominii’ in Britain results from the hybridiza- tion of B. hordeaceus subsp. hordeaceus with B. lepidus. It seems highly likely that such hybridization has been followed by back-crossing of the hybrids to the parents to some extent, a process which would contribute a further range of intermediates to the initial hybrid populations. This introgression hypothesis is supported by the distribution of hairy- spikelet variants in B. lepidus, and of glabrous-spikelet variants in B. hordeaceus subsp. hordeaceus. In each case the spikelet- and lemma-size of these variants is at the extreme end of the variation range which approaches the other taxon. Only rarely are glabrous-spikelet variants of B. hordeaceus subsp. hordeaceus found with lemma-length in the upper end of the species range. Small-lemma, hairy-spikelet variants of B. lepidus have not been seen during this investigation. The hybridization may not have taken place originally in Britain, but wherever in Eurasia B. Jepidus is native and is sympatric with B. hordeaceus. Certainly neither B. lepidus nor ‘B. thominii’ give the appearance of being native in Britain, since they are not well-established in any natural community. Their populations are ephemeral and they are quickly ousted from the sown grassland in which they are most commonly found by the intended components of the seed mixtures applied. B. hordeaceus subsp. hordeaceus on the other hand, while also a ruderal plant, seems much more persistent wherever it grows, and tends to spread rather than to disappear. In old hayfields it often seems to be a permanent feature of the flora. Although both B. hordeaceus subsp. hordeaceus and B. lepidus are predominantly self-fertilizing in Britain, Beddows (1936) concluded that the wide variation exhibited by the former could be explained only on the basis of some degree of outbreeding. Observations during the present work indicated that both species sometimes shed pollen, and it is possible that elsewhere both species may normally be outcrossing. The variation generated by hybridization would produce a range of types which would eventually become stabilized by a resumption of selfing (Stebbins 1950). This would explain the wide variation range which exists between B. hordeaceus subsp. hordeaceus and B. lepidus. The success of the hybrid derivatives, at least in Britain and probably elsewhere where agricultural practices are similar, appears to be a result of the ease with which they contaminate grass seed crops, and so persist with the intended harvest. This type of adapta- tion in response to ‘artificial natural selection’ has been frequent in Bromus, as witnessed by the cereal field contaminants B. secalinus and B. arduennensis Dumort., as well as by other species yet to be considered. These are all examples of the effectiveness of the facul- tative inbreeding system in changing environments, and of the unconscious selection by Man of weed as well as crop plants. With the coastal populations of B. hordeaceus subsp. thominii, the explanation of origin by hybridization and possible introgression from B. lepidus is not satisfactory. They are well-established in natural communities, and have been, in many cases, observed for over 150 years, and are clearly not established by artificial sowing. The persistence of their morphological peculiarities in cultivation and their serological distinctness from the rest of the B. hordeaceus subsp. hordeaceus-B. lepidus-‘B. thominii’ complex indicates that they have a different evolutionary history. Genetic evidence of their distinctness has been Watsonia 6 (6), 1968. 342 PHILIP SMITH reported by Holmberg (1924) who described a naturally occurring plant as a hybrid of B. hordeaceus subsp. thominii and B. hordeaceus subsp. hordeaceus. The plant was only 50 per cent. fertile. Nilsson (1931; 1937) produced artificial hybrids of the same parentage and found that fertility increased very markedly after comparatively few generations. Presumed hybrids of these taxa appear regularly in the literature and they have been named B. X jansenii A. Camus (Camus 1957) in another report indicating partial sterility. Wilson (1956), in an attempted artificial hybridization, failed to get any seed in a small trial. Spatial isolation, and possibly disruptive and stabilizing selection, are the presumed causes of the ecological and genetical peculiarities of these dune plants. Because of these distinctions it seems best to regard them as ecotypes of B. hordeaceus. On dunes, the course of genecotypic adaptation, if this is in fact the explanation, has been in the direction of small size of plants, with prostrate or procumbent culms, a tendency to broad short lemmas and small seeds, and rather dense small panicles. The prostrate or procumbent habit is a familiar develop- ment in seaside plants, while the broad lemmas may perhaps be a photosynthetic compen- sating mechanism since the leaves are short-lived. It is the small lemma and seed size which is of particular systematic interest. Perhaps the genetically fixed ‘depauperation’ was initially disadvantageous in terms of the number of seeds maturing, but was followed by selection of variants with a larger number of smaller seeds. As well as making most use of possibly limiting nutrient supplies, biotypes with smaller more numerous seeds might have the developmental advantage of faster maturation in conditions of imminent drought. The isolation of these gene-pools on dunes would naturally offer the opportunity for divergence in seed protein complement such as has been demonstrated. Some ecotypic adaptation may have occurred in the bromes inhabitating areas used for grass seed production, but owing to the rarity of variants of B. hordeaceus subsp. hordeaceus with sufficiently small seeds, and the inconstancy of the selection pressure on any particular gene complement, it has probably been insignificant. Where the crop and weed seed sizes are more closely similar at the outset, genecotypic adaptation is probably the major means of evolutionary change towards the size and shape of the crop seed. Such adaptation is presumably the cause underlying the contamination of rye fields by B. secalinus and of Triticum spelta L. by B. arduennensis. In the present instance, where the seed sizes of B. hordeaceus subsp. hordeaceus and the crop grasses are rather different, only hybrids com- bining the characters of a very small-seeded brome could have exploited this niche. A further factor promoting genecotypic adaptation in cereal field contaminants is the regular harvesting of seed with resowing of some fraction of it. The cereal field ecotypes were common enough to persist through these generations of crops and so progressively become better and better adapted. The incidence of subsp. hordeaceus variants with seeds small enough to escape removal in the normal seed cleaning techniques is very low, if they exist at all, and so the chance of their reappearing in another generation of the crop plant grown for seed multiplication is minimal. Probably they would be dispersed in the part of the crop sold commercially, and so not be subject to further selection. Bromus hordeaceus subsp. ferronii appears to have serological and ecological peculi- arities and its morphological features were found to persist in cultivation. It is interfertile with subsp. hordeaceus and most of its characteristics—dense hairiness, short, stout culms and recurved awns—can be interpreted as ecotypic adaptations to the cliff-top environment. Awn recurvature in bromes seems to be xeromorphic feature, and the hairiness may also be so explained, or else be of some protection against grazing. For an ecotype it may seem to have a rather restricted geographical distribution. There is no reason, however, why an ecotype need be present in all the possible places where its particular environment occurs. Baker (1953) has cited an excellent example of the restriction of morphologically striking apomicts of Limonium binervosum (G.E.Sm.) C. E. Salmon to particular localities in south-west England and West Wales. The habitat is similar but the particular apomict in each one differs in morphology sufficiently to have been recognized as distinct species by some taxonomists. B. hordeaceus subsp. ferronii may illustrate a similar phenomenon in an autogamously reproducing group. Watsonia 6 (6), 1968. BROMUS MOLLIS AGG. IN BRITAIN 343 The close morphological resemblance of material of B. hordeaceus subspp. hordeaceus, thominii and ferronii, supported by the nature of their serological and genetical resemblances appears to expose them as subspecific variants within a single species. Accordingly this is the course which has been followed in the taxonomic treatment given earlier. The material formerly named ‘B. thominii’, which has been shown above to be a hybrid complex between B. hordeaceus subsp. hordeaceus and B. lepidus, is so common and conspicuous that some reference name seems desirable. The name B. Xx pseudothominii Philip Smith has therefore been proposed for it. Though B. /epidus is shown to be interfertile with B. hordeaceus, and to have an indistinguishable serological spectrum, its specific status is maintained since it contributes to the hybridization as an independent and distinct source of genes, and has a very characteristic morphology, which suggests that at least one of its diploid ancestors is not shared by B. hordeaceus. ACKNOWLEDGMENTS I wish to thank Professor J. G. Hawkes, who supervised the Ph.D. project of which this paper gives a partial account, for his advice, encouragement and help. The work reported here was almost all completed in the Department of Botany, University of Bir- mingham. Professor P. G. H. Gell generously provided some of the antisera. I am indebted to Professor Hawkes, Professor G. L. Stebbins, Professor T. G. Tutin, Dr. C. D. K. Cook, Dr. P. H. Davis and Dr. A. Melderis for their criticism of this manuscript. Dr. C. E. Hubbard provided valuable facilities and encouragement, and I am grateful also to Mr. J. P. Shildrick of the National Institute of Agricultural Botany for most helpful co-opera- tion. I acknowledge with thanks the advice on nomenclatural matters given me by Mr. J. E. Dandy. Thanks are due to the Flora of Warwickshire Committee, and other institutions and individuals who provided seed or herbarium material, to the Linnean Society for per- mission to consult the Linnaean material, and finally to the Science Research Council for the Research Studentship during the tenure of which most of this work was carried out. REFERENCES BAKER, H. G. (1953). Race formation and reproductive method in flowering plants. S.E.B. Symposia 7, Evolution, 114-145. Beppows, A. R. (1936). Vegetative vigour and possibility of natural cross-fertilization in Soft Brome (B. hordeaceus). Welsh J. Agric. 12, 174-182. Beppows, A. R. (1953). The rye-grass in British agriculture: a survey. Bull. Welsh PI. Breed. Stn. H17, 1-81. Beppows, A. R. & Davies, A. G. (1938). Illustrated notes on the technique of grass breeding at Aberystwyth. Herb. Rev. 6, 22—227. Camus, AIMEE (1957). Une Bromus hybride des dunes du Cotentin. Bull. Mus. natn. Hist. Nat., Paris, Sér. 2, 29, 184. Everitt, W. S. (1897). Practical Notes on grasses and grass growing in East Anglia. Ed. N. Everitt, London. FREUND, J. & BONANTO, Mary F. (1944). The effect of paraffin oil, lanolin-like substances and killed tubercle bacilli on immunisation with diphtheric toxoid and Bact. typhosum. J. Immun. 48, 325-334. GELL, P. G. H., HAWKES, J. G. & WRIGHT, S. T. C. (1960). The application of immunological methods to the taxonomy of species within the genus Solanum. Proc. R. Soc. B 151, 364-383. Goocu, S. M. S. (1963). The occurrence of weed seeds in samples tested by the Official Seed Testing Station, 1960/61. I. Grass seed of United Kingdom origin. II. Grass seed of Scandinavian origin. J. natn. Inst. agric. Bot., 9, 353-360. HACKEL, E. (1884). in Kerner, Schedae ad Floram exs. Austro-Hungaricam, 142. Mus. Nat. Hungarici, Budapest. Hawkes, J. G. & READETT, R. C. (1963). Collecting the data, in Wanstall, P. J. (ed.), Local Floras. B.S.B.1. London. HoimserG, O. (1924). Bromi molles—eine nomenklaturische und systematische Untersuchung. Bot. Notiser, 1924, 313-328. HUBBARD, C. E. (1954). Grasses. Penguin Books, Harmondsworth. Hy LAnpber, N. (1945). Nomenklatorische und systematische Studien tiber nordische Gefadsspflanzen. Uppsala Univ. Arsskr. 1 (7), 1-337. Watsonia 6 (6), 1968. 2 344 PHILIP SMITH Koz, J., TurRKovA, V., & KtLozovA, E. (1960). Serological investigations of taxonomic specificity of proteins in various plant organs in some taxons of the family Viciaceae. Biologia PI. 2, 126-137. MICHAELIS, L. (1930). Diethylbarbiturate buffer. J. Biol. Chem. 87, 33. Munro, W. (1861). On the grasses of Linnaeus’ Herbarium. Proc. Linn. Soc. Lond. (Bot.) 6, 46. Niisson, F. (1931). Die Hybride Bromus hordeaceus L. X B. mollis L. experimentell dargestellt. Bot. Notiser 1931, 1-9. NILSSON, F. (1937). Undersokningar 6ver hybriden Bromus hordeaceus < Bromus mollis och des avkomma. Bot. Notiser 1937, 463-486. SmiTH, P. (1965). Experimental Taxonomy of Bromus. Ph.D. thesis, University of Birmingham. STEBBINS, G. L. (1950). Variation and Evolution in Plants. Columbia University Press. STuRROCK, A. (1865). Comparative value of the different grasses. Trans. R. Highld. Agric. Soc. Scotl. 1863— 1865, 251-280. TEDIN, O. (1925). Vererbung, Variation, und Systematik in der Gattung Camelina. Hereditas 6, 275-386. THOMINE-DESMAZURES, C. (1824). “Extrait d’une mémoire.’ Mém. Soc. Linn. Calvados 1824, 40-41. TuTIN, T. G. (1962). in Clapham, A. R., Tutin, T. G. & Warburg, E. F., Flora of the British Isles, ed. 2, Cambridge. WENDELEO, P. (1956). Anthropochore Bromus—arten i Norge. Blyttia 14, 1-3. WILSon, D. (1956). Cytogenetic Studies in the genus Bromus. Ph.D. thesis, Aberystwyth. APPENDIX Accessions of Bromus mentioned in the account. Voucher material at Dept. of Botany, University of Edinburgh. Flora of Warwickshire material at BIRM. Species Acc. no. Origin Bromus hordeaceus subsp. hordeaceus 13 England, Warwickshire; Edgbaston, Birmingham 15. Vincent Drive roadside. P. Smith, 457, 1962. Di England, Worcestershire; Queensway, Halesowen, Old allotment. P. Smith 503, 1962. 26 England, Warwickshire; Quinton, Birmingham. Road- side. P. Smith 494, 1962. 474 Flora of Warwickshire accession. Myton Farm, near Warwick. On a roadside. 1963. Bromus hordeaceus subsp. thominii 429 Wales, Cardiganshire; Ynys-Las, sand dunes on the Dyfi estuary. P. Smith 671, 1963. 510 Wales, Carmarthenshire; Burry Port. Sand dunes near the Old Harbour. P. Smith 813, 1965. 99 Bromus hordeaceus subsp. ferronii 462b Wales, Anglesey; South Stack, on a wall top. Coll. J. B. Phipps. 1963. B. lepidus 31 England, Worcestershire; Halesowen by-pass. Roadside embankment. P. Smith 502, 1962. 92 Flora of Warwickshire accession. S.E. of Willoughby. Railway embankment. 1959. B. X pseudothominii 116 Flora of Warwickshire accession. S.E. of Barford. Farmyard. 1962. 121 Flora of Warwickshire accession. Green Lane, E. of Cloudesley Bush. Farmyard. 1962. The panicle illustrated (Plate 15a) is from a hybrid between B. hordeaceus subsp. hordeaceus 26 and B. lepidus 31. Watsonia 6 (6), 1968. CHROMOSOME NUMBERS OF ORNITHOGALUM UMBELLATUM L. FROM THREE LOCALITIES IN ENGLAND By R. CZAPIK Institute of Plant Anatomy and Cytology, Jagellonian University, Cracow, Poland* ABSTRACT The karyological examination of 49 specimens of Ornithogalum umbellatum L. from three localities in England showed the occurrence of two euploid cytotypes, a triploid (2n = 27) and a hexaploid (2n = 54). In one locality two aneuploid plants (2n = 28) were found among the triploids. A few cases of mixoploidy (polysomaty and aneusomaty) and some karyotype aberrations are described. The origin of cytological diversity within the population is briefly discussed. INTRODUCTION Ornithogalum umbellatum L. shows a high degree of cytological differentiation. Euploid (Qn = 18, 27, 36, 45, 54, 72) and aneuploid cytotypes (2n = 19, 20, 21, 22, 23, 28, 43, 52) as well as plants with B-chromosomes occur within this species according to the data given by several research workers (Table 1). In addition some other aneuploid chromo- some numbers unknown for adult plants were found by the author in young seedlings in the progeny of triploids (Czapik 1966). It should also be emphasized that the cytotypes with 54 and 72 chromosomes have been reported chiefly for cultivated plants and that aneuploid specimens have been found more frequently in wild populations than in garden samples. In view of this, it seemed advisable to study a large amount of material of various origins. In this way the rdéle of wild populations and cultivated material in the process of cytological differentiation within the species could be investigated. MATERIAL AND METHODS Origin of Plants O. umbellatum is native in some parts of the British Isles (Warburg 1962) and occurs in natural habitats. It is also in cultivation and it may be a garden escape in many places (Perring, Sell & Walters 1964). The bulbs examined originated from three localities in England (Table 2). The plants growing on the heath near Icklingham (Suffolk, Grid. ref. TL 7673) were probably native. The specimens from two other colonies, viz. a very small one on the roadside near Fleam Dyke (Cambridgeshire, Grid ref. TL 549542) and a large one in the pasture near the church in Hitcham (Suffolk, Grid ref. TL 983512) might be considered as garden escapes of unknown age. The plants from Icklingham were collected by Mrs. M. Southwell. Since 1951 they have been cultivated in the University Botanic Garden, Cambridge. Six bulbs were dug out from various places of the bed but their clonal origin cannot be excluded. Two adult plants and 6 young bulbs which belonged probably to two clones were collected by the present author in the locality near Fleam Dyke. The material collected in Hitcham was more ample. The bulbs were taken from 7 well-separated groups of plants. According to the number of plants in each group only 2-10 bulbs of various size were taken for the karyological examination, and the remaining plants were left in situ. Cytological Technique This was the same as used in the author’s previous studies on Ornithogalum (Czapik 1965, 1966). Single bulbs were grown in jars with tap water. Root growth began after * The material was collected during the stay of the author at the Botany School, University of Cambridge, in May 1966. 345 Watsonia 6 (6), 1968. 346 R. CZAPIK TABLE 1. Karyological differentiation within Ornithogalum umbellatum L. 2n Authors Euploid chromosome numbers: 18* Neves (1952, 1956), Giménez-Martin (1958), Czapik (1965) ALife Heitz (1926; n=12-14), Sprumont (1928), Nakajima (1936), Satd (1942), Polya (1950), Neves (1952), Gadella & Kliphuis (1963), Czapik (1961, 1965) 36* Neves (1952), Czapik (1965) 45 Sprumont (1928), Neves (1952) 54 Matsuura & Sut6 (1935), Holzer (1952), Neves (1952) WD. Neves (1952) Aneuploid chromosome numbers in adult plants: 19* Neves (1952), Giménez-Martin (1958), Czapik (1965) 20* Neves (1952), Giménez-Martin (1958), Mesquita (1964) D1 Neves (1952), Giménez-Martin (1958) 22 23 Neves (1952) 23s Neves (1952), Czapik (1965) 43, 52 Neves (1952) Aneuploid chromosome numbers in young seedlings and embryos only:* 24, 25, 26, 29, 30, 32 Czapik (1966) B-chromosomes (adult plants): 18+1 Neves (1952), Mesquita (1964-65) 18+3, 19+1, 19+2, 20+1 Neves (1952) 21+1 Neves (1952), Giménez-Martin (1958) 23+1, 27+1 Neves (1952) * Found in the progeny of triploid plants (Czapik 1966) approximately two months and was more abundant in darkness. Root-tips were fixed in a mixture of ethyl alcohol, chloroform and acetic acid (6:3:1) after pre-treatment with 8-hydroxyquinoline (0-002 mol./l. aqueous solution) and stained in bulk in alcoholic hydrochloric acid carmine (Snow 1963). Squashes were prepared separately from each root- tip in order to avoid the difficulties of interpretation in case of mixoploidy within a bulb or particular roots. The cellophane method was used to make the squashes permanent (Murin 1960). RESULTS The main results of cytological examination are summarized in Table 2. Three cytotypes were found in the material studied (2n = 27, 28, 54). In two localities triploid plants were prevalent (2n = 27, Plate 16a); in the third one near Fleam Dyke hexaploid cytotypes occurred (2n = 54, Plate 16d). The population in Hitcham showed some degree of cytological differentiation. The examination of 35 bulbs from 7 well separated groups of TABLE 2. Chromosome numbers of investigated plants Number Localities 2n of bulbs 1. Heath of Seven Tree Road, Icklingham, Suffolk 27 6 2. Pasture, behind the churchyard, Hitcham, Suffolk 27 33 28 2 3. Roadside near Fleam Dyke, Cambridgeshire 54 8 Watsonia 6 (6), 1968. CHROMOSOMES OF ORNITHOGALUM UMBELLATUM 347 plants revealed that 33 bulbs belonged to triploid plants, while two others had 28 chromo- somes (Plate 16b). The aneuploid number was found in single plants in two different groups. No group of plants consisted of aneuploids exclusively. The same types of chromosomes were easy to identify in the karyotypes of the British plants as those of the plants originating from Poland (Czapik 1965). The haploid set con- sisted of one SAT-chromosome of medium size (chromosome VI), three large chromosomes with submedian centromeres characterized by various proportions of their arms (chromo- somes I, I, III), two chromosomes of medium size with submedian centromeres (chromo- somes IV and V) and three small chromosomes with submedian centromeres (chromosomes VII, VIII and IX). The hexaploid and most of the triploid plants had normal karyotypes, and no distinct structural changes of chromosomes could be detected. It should be noted, however, that in the triploid plants only one of the chromosomes VI had a visible satellite. Two satellites could be discerned only in root-tip squashes of two plants from Hitcham, The numbers of satellites in the metaphase plates and their size, as well as the length of the constriction between the satellite and the arm, also showed some degree of variability in plants of O. umbellatum studied previously (Sprumont 1928, Czapik 1965). No special examination was done, however, to test the cause of this variation. Only one triploid plant originating from Hitcham showed a change in the composition of its chromosome complement. It had two chromosomes VII while the chromosome IX was represented four times; the other types of chromosomes were present in the normal number. In plants with 28 chromosomes a strictly triploid complement was increased by the additional occurrence of the chromosome IX, the smallest one in the karyotype. It is very remarkable that variation in the number of these particular chromosomes seems repeatedly to be responsible for the aneuploid differentiation in O. umbellatum (Neves 1952, Giménez- Martin 1958, Mesquita 1964, Czapik 1966). Mixoploid root-tips were detected in one aneuploid and in two triploid plants from Hitcham. Polysomaty was observed in one triploid plant which had a hexaploid sector in one root-tip. One metaphase (Plate 16c), one anaphase and three prophases with 54 chromo- somes formed an island among the triploid cells of the root meristem. This type of mixo- ploidy in O. umbellatum was described also by Neves (1952). One of the aneuploid plants appeared to show a tendency to reversion to triploidy. Eight root-tips were examined: in six roots 28 chromosomes were counted consistently; by contrast in a single root triploid metaphases (2n = 27) occurred side by side with plates having 28 chromosomes and in the last root only triploid plates were found. In a further plant three metaphases with 26 chromosomes were found among the normal triploid plates. In the same root-tip early telophases with bridges were visible. Such abnormalities might lead to numerical differences of the chromosome complements in adjacent cells (Rychlewski 1967). Bridges at anaphase and telophase occurred also in three other roots where no change in karyotypes could be detected. The case of aneusomaty described above could be interpreted as a result of the in- fluence of external factors on the mitosis; abnormal conditions created by water culture of bulbs before fixation must be also taken into consideration. In the majority of bulbs, however, these disturbances did not greatly affect the normal course of mitosis. The deviat- ing chromosome numbers were also found in some young seedlings of O. umbellatum (the progeny of triploids) as well as in some root-tips of adult plants (Czapik 1965, 1966). In seedlings, however, such facts might point to some internal instability caused by aneuploidy. DISCUSSION Embryological investigations have shown that triploidy may play a role in the cyto- logical evolution of O. umbellatum (Czapik 1966). The progeny of triploids show a wide range of karyological differentiation. Chromosome numbers from 2n = 18 to 30 were found in seedlings while in embryos 2m = 32 and 36 also occurred (Table 1). It seems that the Watsonia 6 (6), 1968. 348 R. CZAPIK type of karyological differentiation revealed in O. umbellatum is similar to that known for Hyacinthus orientalis, where the full range of chromosome numbers between 2x and 4x exists in seedlings and adult specimens of the progeny of triploid plants (Darlington, Hair & Hurcombe 1951; Darlington 1956). However, some of hypo- and hypertriploid numbers (2n = 24, 25, 26, 29, 30 and 32) occurring in young plants of O. umbellatum have not yet been found in adult plants. The present author supposes that in Polish material the con- ditions of life in crop-field associations favour vegetative propagation and are disadvant- ageous for germination of seeds and undisturbed growth of seedlings. In this connection a careful examination of specimens from triploid populations of O. umbellatum having more convenient conditions for generative reproduction is needed. The population of O. umbellatum in Hitcham seemed to grow in such a habitat. The cytological diversity observed there was, however, slight. Triploid specimens prevailed and only two from among 33 plants had 28 chromosomes. The aneuploids grew in two separate groups of triploid plants; no group consisted of aneuploids exclusively. According to Lévkvist (1963), in sexually reproducing plants that also reproduce vegetatively e.g. in Cardamine, Sagina nodosa and Saxifraga granulata, aneuploids may have a more pro- nounced ability to form vegetative diaspores than the euploids. Aneuploid plants of O. umbellatum do not seem to belong to this group. They were not found in great numbers nor was the existence of any exclusively aneuploid population reported. As far as the rarity of aneuploid cytotypes within the population is concerned the possibility of the elimination of the additional chromosomes in somatic tissue should be taken into consideration. It is possible that this process is sufficient to prevent the accumulation of aneuploid plants. The case of aneusomaty observed in one bulb of the plant with 28 chromosomes may be con- sidered as an indication of a tendency towards such an elimination. The vegetative origin of 28-chromosome plants also cannot be excluded. They may be bud sports like some of the garden varieties of Hyacinthus whose vegetative origin in cultivation is well documented (Darlington, Hair & Hurcombe 1951). Mixoploidy observed in a few plants from Hitcham seems to be rather common in O. umbellatum but the sup- position of the vegetative origin of some aneuploids requires further examination. The cytological differentiation established for O. umbellatum in the course of the present investigation does not exceed the limits of the variability hitherto known for this species. Still more material is necessary in order to get a more exact view of the processes contributing to the karyological evolution of the species. ACKNOWLEDGMENTS The author wishes to express her gratitude to Dr. S. M. Walters, Botany School, Cambridge, for all facilities provided at the Herbarium and Laboratory as well as for the plant specimens from Icklingham. Thanks are due also to the British botanists who kindly gave the necessary information concerning the localities of O. umbellatum. The support of the Botany School, Cambridge University, and of the Jagellonian University, Poland, is also gratefully acknowledged. REFERENCES CZAPIK, R. (1961). in Skalifska, M., Piotrowicz, M., Sokotowska-Kulczycka, A. et al. Further additions to chromosome numbers of Polish Angiosperms. Acta Soc. Bot. Pol. 30, 463-489. CZAPIK, R. (1965). Karyotype analysis of Ornithogalum umbellatum L. and O. Gussoni Ten. Acta Biol. Cracoy., ser. Bot. 8, 21-34. CZAPIK, R. (1966). The mechanism of cytological differentiation in triploid populations of Ornithogalum umbellatum L. Acta Biol. Cracov., ser. Bot. 9, 65—86. DARLINGTON, C. D. (1956). Chromosome Botany. London. DARLINGTON, C. D., Hair, J. B. & Hurcomse, R. (1951). The history of the garden Hyacinths. Heredity, Lond. 5, 233-252. GADELLA, Th. W. J. & Kirpuuls, E. (1963). Chromosome numbers of flowering plants in the Netherlands. Acta Bot. Neerl. 12, 195-230. Watsonia 6 (6), 1968. PLATE 16 Somatic chromosomes of Ornithogalum umbellatum L. a—triploid metaphase (2n=27, Icklingham); b— prometaphase of an aneuploid plant (2n=28, Hitcham); c—metaphase plate from the hexaploid sector of a triploid root tip (54 chromosomes, Hitcham); d—hexaploid metaphase (2n—54, roadside near Fleam Dyke). 1000 [facing page 348 CHROMOSOMES OF ORNITHOGALUM UMBELLATUM 349 GIMENEZ-MARTIN, G. (1958). Mutaciones espontaneas en Ornithogalum umbellatum L. Phyton, B. Aires 10, 51-58. HEITZ, E. (1926). Der Nachweis der Chromosomes. Vergleichende Studien tiber ihre Zahl, Grosse und Form im Pflanzenreich. I. Z. Bot. 18, 625-681. Houzer, K. (1962). Untersuchungen zur karyologischen Anatomie der Wurzel. Ost. Bot. Z. 99, 118-155. LOvKVIST, B. (1963). Taxonomic problems in aneuploid complexes. Regnum Vegetabile 27, 51-57. MatsuurA, H. & Suto, T. (1935). Contributions to the Idiogram Study in Phanerogamous Plants. I. J. Fac. Sc. Hokkaido Uniy., Ser. 5, Botany 5, 33-75. MesquiTa, J. F. (1964). Natureza e comportamento dos cromosomas supranumerdarios isobraquiais em Ornithogalum umbellatum L. Bolm. Soc. Broteriana sér. 2, 38, 119-131. Muri, A. (1960). Substitution of cellophane for glass covers to facilitate preparation of permanent squashes and smears. Stain Technol. 35, 351-353. NAKAJIMA, G. (1936). Chromosome numbers in some crops and wild Angiosperms. Jap. J. Genet. 12, 211-218. Neves, J. B. (1952). Estudos cariolégicos no género Ornithogalum L. Bolm. Soc. Broteriana sér. 2, 26, 5-192. NEVES, J. B. (1956). Sur la caryosystématique d’ Ornithogalum paterfamilias Godr. Bolm. Soc. Broteriana sér. 2, 30, 141-154. PERRING, F. H., SELL, P. D. & WALTERS, S. M. (1964) in Perring, F. H., Sell, P. D., Walters, S. M. & Whitehouse, H. L. K. A Flora of Cambridgeshire. Cambridge. PO6LyYA, L. (1950). Chromosome numbers of Hungarian plants II. Annls. Biol. Univ. Debrecen. 1 (VII), 46-56. RYCHLEWSKI, J. (1967). Karyological studies on Nardus stricta L. Acta Biol. Cracov., ser. Bot. 10, 55—72. SATO, D. (1942). Karyotype alteration and phylogeny in Liliaceae and allied families. Jap. J. Bot. 12, 57-161. SNow, R. (1963). Alcoholic hydrochloric acid-carmine as a stain for chromosomes in squash preparations. Stain Technol. 38, 9-13. SPRUMONT, G. (1928). Chromosomes et satellites dans quelques espéces d’Ornithogalum. Cellule 38, 271-292. WARBURG, E. F. (1962) in Clapham, A. R., Tutin, T. G. & Warburg, E. F. Flora of the British Isles. Ed. 2. Cambridge. Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION IN A COUNTY FLORA By J. G. HAWKES Department of Botany, University of Birmingham B. L. KERSHAW Computer Services, University of Birmingham and R. C. READETT Birmingham Natural History Society ABSTRACT Methods of computer mapping of distribution, habitat and frequency data of vascular plants in Warwickshire are described. A simplified mapping scheme for bryophytes is also outlined. The primary data for vascular plants are transferred in the form of coded information to punched tape by means of a teleprinter and checked for errors. The data are then processed in the Birmingham University KDF9 computer by means of three separate programs, each taking the data forward to the next stage by means of magnetic tape. Intermediate print-outs give a preliminary statistical treatment of the data, as well as a simple sketch map, which can be printed on the teleprinter or line printer. The end product of the final program is a tape designed to give the necessary instructions to an incremental graph plotter. This draws symbols in the correct positions indicative of the habitat or habitats in which the species has been recorded. Two grades of frequency are shown by means of different line thicknesses. The plotted map will be reproduced with an overprinting in another colour, indicating the county boundary and various other features, as well as grid lines and border. A series of bryophyte maps without habitat and frequency information but with data on fruiting will be printed on the teleprinter. The wider applications of these mapping techniques are briefly discussed. INTRODUCTION The hand plotting of distribution data, whether of plants or of any other class of organisms, can be both tedious and expensive. If the data to be mapped are at all complex, errors may easily occur, which cannot entirely be eliminated, even after the expenditure of even more time and effort. Within the last decade high speed electronic computers and associated equipment have become generally available and have made it possible for data to be stored with accuracy, and sorted and retrieved in a wide variety of ways. It would seem desirable, therefore, to use such equipment to sort and store distributional data, provided they can be presented to the computer in an appropriate way, and to devise systems by which they can be printed out in map form. An excellent example of the use of electronic equipment in the production of distri- bution maps in this country is the B.S.B.I. Distribution Atlas (Perring & Walters 1962). The original maps for this publication were printed on a tabulator from data supplied to it from J.C.T. punched cards. The symbol on the map, which was a filled-in circle (except for some other symbols put in by hand afterwards), was positioned by making use of numerical co-ordinates taken from the National Grid at 10 km intervals. Soper (1964, 1966) has discussed several methods of mapping the distribution of the vascular plants of Ontario by tabulating machines and digital plotters. The primary data, consisting of a code reference for the species together with geographical co-ordinates and some other information, were put on to punched tape and later transferred to 80- column I.B.M. punched cards. Soper pointed out that both tabulators and digital plotters with a printing head were capable of printing a limited number of symbols in a reasonably 350 Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION 351 satisfactory manner. On the other hand, a digital plotter with a drawing pen could be made to draw symbols of any shape and size, though it would, of course, need to be ‘programmed’ for each symbol. He pointed out that the method was not very satisfactory for drawing completely blacked-in symbols, and that smooth curves could not be formed very well because the pen could move only in successive rectilinear steps. THE FLORA OF WARWICKSHIRE PROJECT A survey of plant distribution in vice-county 38, Warwickshire, was begun in 1950 as a joint project of the Birmingham Natural History and Philosophical Society and the Birmingham University Department of Botany. This has been a co-operative study, involving a team of amateur and professional botanists in many hours of field work over some 16 years. It was begun by Mr. R. C. Readett and Mr. Peter Green. After the latter’s move from Birmingham one of the present authors (J.G.H.) took his place, and the organ- izing panel now also includes Miss D. A. Cadbury and Mr. M. C. Clark. The scheme was modified in 1954, when the “basic square’ technique of area recording was introduced (Hawkes & Readett 1954, 1963). This involved the detailed survey of 1 km squares as the unit of recording, one square at random being selected from each block of four, or ‘tetrad’. An even survey over the whole county was thus feasible with only one quarter of the effort that would have been involved in recording from every square. There are thus 25 tetrads in each ‘major square’ of 10 x 10 km, which was the unit of recording for the B.S.B.I. scheme referred to above. In addition to this, a more conventional method of recording rare plants was carried out, since they would have been likely to have been missed altogether on the random square method. The records obtained from the basic square surveys were written on to species cards representing a 10 x 10 km block of 1 km squares, printed at the scale of 1 in. to the mile and hence directly comparable with the Ordnance Survey map of that scale (Hawkes, Readett & Skelding 1955). This method also helped to give an idea of the distribution of the species at any time during the survey, since the records were displayed on a grid system in their correct spatial relationship. Habitat data and frequency were also recorded on a pre-determined scheme using a system of major habitat groupings (e.g. woodland, waterside, water, cultivated land, etc.); each of these was divided into a number of minor habitats for greater precision (e.g. mixed woodland, oak woodland, conifer woodland, etc.) and suitable lexicographic symbols were devised for each (e.g. WO-m; WO-0; WO-<; for the habitats just given). Frequencies were recorded on the usual simplified system: abundant (a), locally abundant (1a), frequent (f), locally frequent (1f), occasional (0) and rare (r). It was emphasized that the frequencies should refer to each habitat for which the species was recorded, and not to the square as a whole. From the inception of the scheme it had been hoped that it would be possible to transfer the habitat data in some way on to species distribution maps by means of symbols that would give a general impression of the community or communities in which a species was to be found throughout the county. A number of problems had to be solved here; in the first place, none of the symbols could be of the blacked-in type, since they would need to be presented in various combinations in the same square. Each must therefore be clearly visible, even when combined with several of the others. It was also necessary to devise symbols which would give easily recognizable patterns over the map as a whole. Thus, symbols which relied for their interpretation on their position within a square would clearly be valueless, since such a scheme would not enable the reader to obtain a general picture at a glance; each square would have to be examined individually. It was further decided to represent the record from each random square in the complete tetrad in which the random square occurred. This would prevent a false impression of distribution being produced as a result of ‘clustering’ of the random squares. A reasonably satisfactory series of habitat symbols was devised which conformed to the needs outlined above (see Cadbury, Hawkes & Readett 1957). These have been incorporated into the present scheme Watsonia 6 (6), 1968. Sy) J. G. HAWKES, B. L. KERSHAW and R. C. READETT with some modifications. At that stage (prior to 1964) it seemed impossible to devise a scheme for habitat recording which at the same time showed frequencies within habitats, though a general idea of frequencies could of course be obtained by looking at the spread of the species over the county as a whole. The major habitat categories (woodland, waterside, water, heath, marginal, grassland, cultivated land and ruderal), eight in number, were, as we have already mentioned, divided into minor ones; there were 34 of these minor habitat categories for Warwickshire, and it was clearly impossible to devise a system of symbols to denote all these when each symbol needed at the same time to be distinguishable in a large number of different combinations. It was therefore considered that symbols showing the eight major categories would be all that could be reasonably expected for maps of this sort. ELECTRONIC DATA PROCESSING By 1964 a number of problems urgently needed attention. Since it was hoped to draw maps of some 700 species of vascular plants and 200 species of bryophytes a long and expensive period of map drawing seemed to be ahead of us. Although the bryophyte data were fairly simple, the prospect of drawing by hand the habitat symbols on 700 vascular plant distribution maps was one to which we did not look forward with much pleasure. There were somewhere in the region of 1,000 records a species for the more common species, each of which would need to be mentally interpreted into a symbol by the draughts- man and placed accurately on to the base map. A further difficulty became apparent about that time. The habitat coding scheme, which was devised in 1950 and came into general use in its present form in 1954, was later felt to contain some errors of arrangement of minor into major habitat categories. Thus, the marginal category was found to be rather artificial from an ecological point of view. Whilst retaining the system for recording purposes, it was decided to keep ‘hedgerow’ in the marginal category and add ‘scrub woodland’ to it, at the same time removing all the others. ‘Roadside’ and ‘railway’ would fit best with the ruderal category (enlarged slightly to include ‘viatical’ elements); ‘walls’ and ‘quarries’ were removed from marginal and ruderal, respectively, and a ninth major category of ‘stones, etc.’ was created for them. When this was done, clearly an even more complex mental process would be needed to convert the original code letters on the species card into the correct map symbols. A final point might be considered here. It was planned at a fairly early stage (see Cadbury, Hawkes & Readett 1957) to include habitat analyses in the Flora; these would show for each of the commoner species the range of habitats in which it had been recorded, expressed as the number of records for each habitat and the percentage figure, as well as the apparent ‘preference’ of the species for a particular habitat. For the second time in a county flora numerical data would be available for statistical analysis, the first being Good’s Geographical Handbook of the Dorset Flora, 1948. The information available was sorted in one or two cases by counting occurrences in the different habitats and calcu- lating percentages, as the first stage of the process. The data were extremely interesting, but the labour involved was quite considerable and carried with it as many possibilities of error as the drawing of the maps themselves. Fortunately, at about this time Birmingham University bought an English Electric KDF9 computer, and it was therefore decided to process the whole of our data on this machine and at the same time to explore the possibilities of computer mapping of the distribution data. Clearly, some kind of graph plotter seemed to be the answer to the mapping problem, where X and Y co-ordinates could be used to position the symbols in the correct places on a previously printed map on continuous stationery. Unfortunately, most graph plotters incorporate a printing head with very small standard symbols. These completely obscure each other if overprinted, only one symbol can generally be printed in each position, and the symbols can often only be printed in a fixed sequence. Furthermore, the symbols available seemed to be far too small for our purposes. Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION 338, The machine we finally adopted, after many exploratory discussions with a variety of firms, is an incremental plotter with a drawing pen, marketed by Benson-Lehner Ltd., of Southampton. Since these instruments are relatively expensive and we should only be needing one for this particular project, we agreed in April, 1965 to buy time on one of Benson-Lehner’s own machines. We should like at this point to record our thanks to the representatives of this firm for their valuable help and co-operation in the project. PREPARING THE DATA FOR THE COMPUTER Apart from the very necessary programming work,* the first stage in preparing the data for plotting is to put it into a form which can be used by the computer. We had no punched card facilities at Birmingham when this phase of the work started and we therefore decided to use 5-hole punched tape. To this end we bought a Westrex teleprinter, which is in effect an electric typewriter with tape punching and reading facilities. Punching began in December 1965, and we were fortunate in having the services of a skilled teleprinter operator (Mrs. M. Kershaw) for the whole period, working on a part-time basis, until Easter, 1967. This included not only the punching of the tapes, but also editing them; it must be admitted, in all fairness to the operator, that a lot of the ‘editing’ was nothing more than adding recently received records to a tape which was in fact a perfect copy of the original data. The data were taken directly from the species cards, species by species, and in sequence of major square cards, the data being copied exactly and a ‘print-out’ produced as a by- product of the punching operation (Fig. 1). The print-out was used for checking against the original cards, a tedious exercise carried out by Miss D. Cadbury, Mrs. E. Pickvance, Mrs. Kershaw and others. Thanks are due to them for the dogged way in which they carried out this monotonous but all-important stage of the work. After punching and checking, the data are ready to be passed to the computer. Three programs are used, but only the first reads the paper tape data. Each in turn operates on the data, taking the work a stage further and passing on partly processed intermediate results to the next stage by means of magnetic tape. The last program produces punched paper tape which is sent to Benson-Lehner and used to control the plotter to draw the final map. ENTERING THE DATA INTO THE COMPUTER The computer first checks the format of the data for four types of error: (1) Major square number unacceptable. (2) Minor square number unacceptable. (3) The same | km square mentioned twice. (4) Observation code unacceptable. If an error is detected, the computer prints a message to guide the punch operator in correcting the mistake, and switches into a mode of operation in which it does not prepare intermediate results for further processing, but continues to read the data and print messages for any errors detected. In this way all errors in a particular data tape are found in one computer run and can all be corrected before the tape is resubmitted. If, however, no errors are detected, the computer assembles within its memory an “electronic image’ of the required distribution map. Once this image is obtained, the rest of the computer process is directed to translating it into suitable control signals for the plotter. The first level map image is placed on magnetic tape ready to be further treated by the second program. The first program, however, does rather more than this. Observations which are accepted and entered on the map image are counted, giving 204 totals, one for each minor habitat/frequency combination. These totals are printed to form the first table in the * See Appendix, p. 360. Watsonia 6 (6), 1968. 354 J. G. HAWKES, B. L. KERSHAW and R. C. READETT 3 acer campestre le 3 Q4 69emhr£. / 0S 41-0006 43-mhfo. 44.-MTOTWOS8CEe SO-embhloe 54-40 SCOemMbhrOe SS Bbhro.WORmOe 60.mhroe 62-@Qhtle 65-womlft. 66emhrf. 68-eBkTO6 75-HWOMO oMbTEe 7T6-WOSCCeMATO>” 62. Bhroe &9.cEr. 92.mhroe 94 mbrlft. 96 «WO BOs 98emhberre 7 06 43-MhC06 44-480 8CO. Sieahret. 54-vomoemhrfe S6emhbrere 60-M@brOe” 64 mbr£. 6Semhbt£e 67> MhEOe TO-mhto. 73-mherEe 7TSeMhlOe T6-mhroe 77 -mhret. SCeMhAFOe B2omhroe S4.Wwome£ 87 - MALO” 92-mhro. 93.mhro. 94-MhroeWOSCre G6. mhro. 7 O7 7T2 @ahbhre. 713e¥0 3COe G2.usd cs 90-mgbtoe 4 Fig. 1. Part of the data for Acer campestre, Field Maple, as typed on the teleprinter, showing major square numbers heading blocks of data, minor square numbers at the start of lines, and letter strings representing individual observations (e.g.: mhrf = hedgerow, frequent; wooo = oak wood, occasional; wosca = scrub wood, abundant). The figure 3 at the top represents the B.S.B.I. code number for this species. Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION 355 statistics (Fig. 2). The second table is simply the first recast in terms of nine major habitats and two frequencies, corresponding to the symbols drawn on the maps. It should be noted that this first program also incorporates a ‘random square sieve’ which allows habitat and frequency data through for the simple statistics only if they have been recorded for random squares. Data for non-random squares are therefore suppressed, thus rendering what remains more statistically reliable. Many records have been submitted by collectors for non-random squares as well as for random ones and all these have been used for mapping. However, for showing habitat spread and preference of the species to be mapped it was considered essential not to bias the data in any way towards a preference of a collector for recording in a particular habitat, rather than a preference of a plant for growing in it! This preliminary statistical treatment of the data will not only help us to make useful and meaningful statements on habitat spread and preference in the general text of the Flora; it is also intended to provide a comparative table of habitat records for each of the 700 species that are mapped, giving both actual and percentage records in each of the nine major habitat categories. Acer campestre L. 3 A LA F LF 0 R TOTAL MHR 2 3 134 24 255 7A 489 WOM j 7 1 43 14 66 MSC 3 1 15 6 25 RURO 5 t 8 | 15 RURY \ 7 3 11 WSR | 8 j 10 WSC 3 1 4 WSD j | 1 ! 4 woo 3 | 4 GP } | | 3 GM 1 ] 2 GR 2 2 WSP 2 2 RUW 1 1 STQ 1 1 TOTAL 2 4 154 248 344 107 639 HABITAT ABUNDANT OCCASIONAL TOTAL OBS. PERCENT OBS. PERCENT ORS PERCENT MARGINAL 167 260} 347 54,3 514 80,4 WOODLAND 9 1e4 41 9e5 70 11,0 RUDERAL 7 Trek 20 33! 27 4e2 WATERSIDE 3 005 17 Dig th 20 3a i GRASSLAND 2 003 5 0.8 eataie iy, STONES ETCe l 002 | 02 TOTAL 188 2904 45} 7006 639 100.0 Fig. 2. Data format check and simple statistics for Acer campestre. The input tape for this species contained no errors, so the computer has printed no error messages. In the first block of data the records are listed by minor habitats in one direction and frequencies in the other; general totals are given in the right-hand column and bottom line. In the second block of data the records are listed in major habitat groups on two grades of frequency (abundant and occasional). Totals and percentages are also shown. Note: these figures are based on random square data only. Watsonia 6 (6), 1968. 356 J. G. HAWKES, B. L. KERSHAW and R. C. READETT The ‘sketch map’ (Fig. 3) uses each print position to represent one tetrad. If the species has not been observed in that tetrad, ‘space’ (blank) is printed; if the only observation(s) are occasional or rare, ‘+’; and if at least one abundant to frequent observation has been recorded, ‘X’. There are 10 print positions to the inch across the page, but only 6 lines to the inch vertically; by inserting ‘space’ in alternate positions we have 5 symbols to the inch horizontally and 6 to the inch vertically, which is as near as it is possible to get to equalizing the horizontal and vertical scales. The maps are therefore slightly wider than they should be, but are nevertheless most useful in giving a general impression of the distribution of each species throughout the county. They will not appear in the Flora (but see bryophyte maps, mentioned below) and will be used merely to provide a general indication of species distribution when the textual part is being drafted. Those tetrads through which the county boundary passes are ‘marked’ in the computer’s memory. When such a tetrad is empty, full stop is printed for it instead of space. These full stops build up a suggestion of the county boundary, assisting the interpretation of the sketch map. FURTHER TREATMENT OF THE MAP IMAGES The map images are taken from magnetic tape by the second program and converted into suitable control signals for the plotter, which are placed on a further magnetic tape. The third program transfers from this magnetic tape to paper tape; this is done for the convenience of the computer room staff and need not be discussed further here. The plotter consists, effectively, of a pen moving over a sheet of paper, capable of drawing any shape required by incremental movements of 0-1 mm (or 0-005 in. on the non-metric model) in any one of the four cardinal directions. Left to right movements are Acer campestre L. 3 + 4 e+ ¢@ X @ @ ++ @ e+e ¢ X « o> + + + + e es ©X + + + + -% Att +X ¢ + ¢€ + X + X + e+ ¢ x ¢ no ESD > xX « ° +>~+? EKE XK O 'X SD, oS fy e+ € Ke OK ee) a X + + ee * «e+ X XX + X X @ + OX OK Ke oe ee ¢ OK KK Ke eae Ke OX x eo t-e ee + CPO be KEK Fh oa! ER PXEE ROOK: OX * e Mote) Sd.) 4 4 S S22SSN OKO KK wwe Pe 2 e¢ + + © H+ © + SOK Kt Ke Xe X » det se th Gea Ge SAG Pa De Mat TC Gand, Gee dh Vimy. Gah Te, Shy en: Koo @ XU ew Ke 2 2 4 4. se 974 KONE eh ee cine £0 XX) Sete & OK [e O5K MX SEs e EE xX ¢ ¢ + + + © eo + & & KX *+*-e xX X HIKER SP) OREO OK Ke ae Te Rw Ke ree eX OK S°X 1X Sere KK 2: SEM Se 4 Se SO XPS eK -o X © © XX XXX @ X XX + ¢¢X ¢@ Ke KEK KS em eee Or Pi hea a aX CMO py Sa At OD eM we OM OM ROK eX 4 es + 4 eS eS OX! KM KO Wee SS XoeeXeeXve X Cie) Gu Vea oat eer Bm Cun) uur Teer Ur ya) Gap < © 2 2 1 XK KX SK SK MK XK XK ee SK OX OX ee OX OK XX PPK IK OK OK COOK Kee OX ee Sey Mwy Ai ie Patan htneas ed wind BP BODY DS aDS PY RU PS Che Re 24 ROS @ 2 Fe S*XP SSK S SEK) Se OX XT KKK K oat + ¢ ¢ a Ne mae OE oe - +e + X Xe X PNG SEK Ki Re Keay ae ok 2 eK XS ee OK eK OK eh KR OK RO Kaye Xoce x CS ee) eee fuer. Cen e SMe a KOK erie e+ + + ¢ xX +++ X + X 2+) bb XK So? Gem Ship Cone ¢ + + Fig. 3. Interim ‘Sketch Map’ for Acer campestre L. Printed on the computer’s fast output line printer, this map took only three seconds to produce. + Occasional to rare in tetrad. x Frequent to abundant in tetrad. Indicates position of county boundary when no record exists in this tetrad. Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION Boil accomplished by the actual movement of the pen, whilst up and down movements are accomplished by corresponding movements of the paper. Diagonal lines are drawn by combined pen and paper movements, whilst further control signals instruct the pen to begin or to cease marking. A reasonable solution to the problem of showing frequencies, as well as habitats, on the same map has also been devised, though admittedly not all grades of frequency can be given. This was done by indicating ‘rare’ to ‘occasional’ by means of a thin line and ‘fre- quent’ to ‘abundant’ (including ‘locally frequent’ and ‘locally abundant’) by a thick one. If these ideas were to be translated into a completely automated system it would clearly be impossible to change pens every time a change of line thickness was required. We therefore decided to use a thin pen throughout and to program it to re-draw the symbol after displacing the pen by two increments if a thick line was required. Since the pen thick- ness (0-2 mm) is the same as this displacement, the two single thin lines will ideally produce a line twice the thickness of the thin one. In theory there should not be a perfect coincidence of the two lines when diagonals and circles are drawn, but in practice they seem to coincide quite well. All the symbols (except the circle, which is a special case) are made up of vertical, horizontal or diagonal straight lines (Fig. 4). Fig. 5 shows how the circles, thick and thin, are built up from short horizontal, vertical and diagonal straight lines. The exact arrangement of these lines was decided by a preli- minary computer run to design the best possible circle within the limitations of the plotter. Two problems arise in preparing the plotter control tape: moving the pen to the required map location for a symbol, and drawing the symbol once the pen is correctly located. The second is easier of solution, since the drawing of a symbol is identical, regardless of its position on the map: 18 standard sets of drawing instructions, one for each symbol, are held in the computer and the correct one is selected after the pen has been positioned. Positioning the pen to draw a symbol is complicated by the desire to keep pen move- ment, and therefore plotter time and cost, to a minimum. As each symbol is drawn it is erased from the computer’s map image, and the machine then chooses the nearest remaining syinbol to the plotter pen’s current position to be drawn next. Besides drawing the symbols, the plotter writes a reference number on each map so that it can be identified and given the correct title in printing; two register or ‘fiducial’ marks, one at the top left of the map and the other at the bottom right, are made so that the printers can correctly superimpose the plotted symbols on the standard outline of the county which is overprinted on every map (Fig. 6). PUBLICATION OF THE MAPS In this paper we cannot go into the plans for printing the Flora as a whole, but shall confine ourselves to the mapping process. Pergamon Press, Ltd., have agreed to publish the Flora, and have, of course, been consulted over the presentation of the maps. They plan to make a map of two printings. One of these, printed in black, will be reproduced from the plot of the symbols for each species, and will have keys and species names added at the time of photographing. The other will be printed from a standard block, in some other colour (probably blue, red or brown) and will comprise the county boundary, the grid lines, some physical features, such as conurbations, roads, railways, etc., and the border and panels. These will be combined, using the fiducial points for exact registration of the two printings. Transparent overlays will also be provided, giving geological and other information of possible interest for an interpretation of distribution patterns. BRYOPHYTE DISTRIBUTION MAPS The bryophyte survey, which has been planned and largely carried out by Mr. T. Laflin, is conceived on a slightly different basis from the vascular plant survey. In the bryophyte study, species lists without habitats but with data on ‘fruiting’ or ‘sterile’ will be punched Watsonia 6 (6), 1968. 358 J. G. HAWKES, B. L. KERSHAW and R. C. READETT awe ) Woodland (2) Waterside (3) Water ) Acid heath, bog bil (5) Hedgerow, scrub ) Grassland 7) Arable, gardens ~ (8) Quarries, walls en Ce (9) Woodland (10) Waterside ) Water ) Acid heath, bog ee E (13) Hedgerow, scrub | j (14) Grassland . (15) Arable, gardens (16) Quarries, walls (1) — (8) Rare to occasional — (16) Frequent to abundant Fig. 4. Enlarged drawings of habitat symbols (other than ‘ruderal’) showing the track of the pen centre. At the actual size at which these are plotted (in a square of 6 mm sides) the line drawn (0:2 mm) is wide enough to fill completely the space between the lines of a thick symbol. on the teleprinter. Thus, each record will consist of a species code reference, based on the bryophyte census catalogue number, together with a grid reference and a symbol denoting fruiting or sterile. The tapes will be passed through the computer with a different pro- gram, and an output tape produced which will give a map on the teleprinter. This device is used rather than the fast printer, since it is proposed to publish these maps direct, and the print quality of the fast printer is not really good enough for publication. Pilot maps have been produced already but at the moment of writing the primary bryophyte tapes have yet to be punched. The teleprinter maps will bear two distinct kinds of symbol, for fruiting or sterile, and the records for these will be added up and percentaged as a by-product of the computer program (Fig. 7). These maps will also be overprinted with a standard map in some other colour, as with the flowering plant maps, though they will be reduced in size so that Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION Sy, Fig. 5. The circle (Ruderal symbol), not shown in Fig. 4. On the left is the pen track to draw a thin circle. The two tracks on the right, separated by about two steps throughout, combine to form a thick circle. four maps, instead of two, can be printed on each page. These maps, like those produced on the fast printer, suffer from inequality of vertical and horizontal scales, so the standard overprint map will have to be correspondingly distorted. FURTHER APPLICATIONS At the time the Warwickshire Flora survey was begun there was no prospect of a computer becoming available at Birmingham and for this reason the storage of data on punched cards or tape was not envisaged. Our data were therefore rather laboriously written on to species cards and only much later transferred to tape. It would almost certainly have been easier to have put our records on to punched cards in the first instance, as was done for the B.S.B.I. Atlas of the British Isles. A Fungus Survey of Warwickshire has now been started and a flowering plant survey of the Wyre Forest, near Bewdley, is just beginning. With the card punching, sorting and reading facilities now available to the biology departments at Birmingham it is hoped that the data can be stored directly on punched cards or on tape. In both of these schemes mapping of some kind is envisaged, though the exact details have not yet been decided. In County Flora work or in recording distribution of plants for even smaller areas one has limited and often specific objectives. For surveys of a wider nature on a national scale, on the other hand, the perspectives are remarkably wide. Such mapping techniques could be used to portray the distribution of any kind of organism, of fungal or other diseases, of genetic variability in cultivated plants or the results of demographic surveys, to name only a few applications. Simple and inexpensive maps of the ‘sketch map’ type can be produced with great speed, once the basic programs are available, and provided the data are collected in a suitable form. The more sophisticated type of map which is drawn on a graph plotter has an even wider application, limited perhaps only by the capacity of the data collectors to obtain the information desired. What seems quite clear is that computer mapping has an exciting future whose outlines at present we can only indistinctly perceive. Watsonia 6 (6), 1968. 3 360 J. G. HAWKES, B. L. KERSHAW and R. C. READETT WARWICKSHIRE VICE-COUNTY 38 Acer Campestre L. Field Maple FREQUENCIES r-o f-a | Hedgerow, Scrub Acid heath,Bog [_ Grassland Waterside, Marsh —— Water We i Waysides Waste. Miles Quarries, Walls I oes sie 3 | Arable, Gardens ~ Kilometres 07 129% 6) ce 40 10§ INF /— Fig. 6. Plotted map for Acer campestre. The county boundary, title, grid references, etc., have been added by hand. In the final publication these and other data will be overprinted in another colour. (Note: This is only a pilot map; symbol connectivity will be improved in the Flora itself.) APPENDIX FURTHER DETAILS OF THE COMPUTER PROGRAMS Some readers may be interested in a rather more detailed description of the computer programs than was thought admissible in the body of the paper The language used was K-Code, the KDF9 dialect of Mercury Autocode, except for the final program which was written directly in symbolic assembler. K-Code possesses many sophisticated and useful bit handling and logical operations and allows the easy introduction of patches of assembler in areas of high-level language. The unit of recording, the minor square, is equated with the computer word. The KDF9 word has 48 bits, of which the least significant 18 are used to represent the presence or absence of the species under consideration in the 18 possible habitats, 9 ‘abundant’ and 9 ‘occasional’. One other bit indicates whether or not the square is random; this is the sign bit, giving fast discrimination of random/non-random squares by the sign detection hardware. Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION 361 2129011 amblysteGium Serpens (hedWe) bey Se and Go + * * Ok Ok + + r * x + * + * ++ * rs * + + + * * * xO + * % * FORK + + + * Ke KK + * + * * + * xe KK + + * * x * * * kK OK + * Ok ++ * x + x * * Ox * * * * Ok * + * * * ke 5 Pa + * * * kK + kK OK + Ok * * * ke OK * +k Kk K OX * * + x * * x kOe KOK * * + * OK * *x* + + + kK Kk OK & * ok | + kK * * + k kok Ok Ok * | * Ok kok OK KOK OR * x +k OK OK + * + KK OK KK OK OK * * ok x * * | ee kk KK + OR OK + OK OK OR * ++ kh * * * ok * Ok x of x + ak | + + kK Kk KOK K ek Kk + X x ok Ok * * * * * Ok * Pa | * * ek Ok OK * * oe Ok * | KK OK OK # x * Ok | ek * * * * Ok | Ok Ok * Kk OK KOK OK OK + | * * * xk ok Ok * | kk ke * + * * 471 ebservationS, 375 (79-62 percent) Eruiting. _ Fig. 7.7A simple moss distribution map (Amblystegium serpens), produced on the teleprinter. Borders, grid references, etc., will be printed in another colour in the final publication. + Not fruiting, sterile. * Fruiting. Records of sterile and fruiting are totalled and percentaged. If K-Code had facilities for 2-dimensional arrays, one would have been used to represent the map plotted by squares. This facility, however, is missing, so it is necessary to use a I-dimensional array and calculate unique subscripts from the X- and Y-positions of the Square under consideration, themselves calculated from the major and minor square Watsonia 6 (6), 1968. 362 J. G. HAWKES, B. L. KERSHAW and R. C. READETT numbers. If, when first mentioned, a square has any bit other than the sign bit non-zero, an input error, minor square mentioned twice, is signalled. Whenever an error is detected, an appropriate message is output and an error marker is set. The input letter strings are packed into one word each. The input symbols read by K-Code are of 8 bits each, so the longest possible string, 6 letters (e.g. WOSCLA) just fits into one 48-bit word. Had there been 7-letter input strings the programming would have been considerably more difficult. An assembled input string is compared word by word with a permanent list of 204 permissible strings. The subscript of the list word which matches the string gives a simple encoding, a number from 0 to 203 inclusive, of the input word. If no match is found the ‘code unacceptable’ error occurs. After a successful search of the list, the encoded form of the input string is used in three ways: 1. Placed directly onto magnetic tape, so that the input data, encoded, will be available for future work without the necessity of reading the paper tapes again. 2. As subscript to control the adding of | to one of 204 counters, thus accumulating the raw material for the statistics. 3. Also as subscript, to select a shift length. A 1 bit, shifted by this amount, becomes a marker for an occurrence in that major habitat/frequency and is combined with the previous contents of the correct work in the map image by the logical ‘OR’ operation, thus avoiding disturbing any markers already present in that word. When the map is complete, known by the detection of the character ‘>’ on the input (an arbitrary choice from the unused symbols on the teleprinter keyboard) the error marker is consulted. As messages are output as soon as an error is detected, if the marker is set no further action for that species is needed. Assuming the error marker clear, however, the next action is to convert the map plotted by squares into the map plotted by tetrads. Suitable indexing selects the four squares forming one tetrad, and they are combined into one word. A slight complication arises here. A species may be both abundant and occasional in the same major habitat—e.g. abundant in oak woodland, occasional in mixed woodland. An abundant symbol blots out an occasional symbol in the plot, but to draw the thin symbol would be a waste of time and money. The routine which combines squares into tetrads therefore erases occasional observations wherever there is an abundant observation in the same tetrad and major habitat. This is achieved by a combination of shifting and logical oper- ations. The tetrad record is placed into the appropriate location of a tetrad array, which is filled up by a suitable scanning process and becomes the final map image. This image is placed on magnetic tape ready to be handed on to program 2. The statistics are assembled from the 204 counts, suitable totals being taken and the information re-ordered with the most favoured habitat being printed first, then the next most favoured and so on. Zero counts are suppressed. The sketch map is formed by printing ‘space’ if a zero word is found during scanning of the tetrad array, ‘X’ if the value is greater than 511 and ‘+’ if the value is non-zero but not greater than 511. Details of the county boundary are held in preset (‘V’) stores in which each bit represents a tetrad, being 1 if the boundary passes through that tetrad, zero if not. If the tetrad is empty, the appropriate bit is consulted to decide whether or not to print full stop. The second program takes the first program’s map images from magnetic tape and prepares a new magnetic tape which the final program converts into paper tape for the plotter. Having obtained an image from magnetic tape, program 2 scans it row by row. Each row is ‘pre-scanned’ from left to right and the positions of the left-most and right- most occupied tetrads compared with the current pen position. The latter is held in two words, one for X and the other for Y, and is assumed to start at the top left corner of every map. If the pen is nearer the left-most symbol the row is plotted with the pen moving generally from left to right, and vice versa. This is not a perfect way of choosing the shortest Watsonia 6 (6), 1968. COMPUTER MAPPING OF SPECIES DISTRIBUTION 363 path to visit every symbol, but it is a reasonable compromise between waste pen movement and program complexity. Having selected an occupied tetrad, the computer recalculates the current pen position in terms of an origin at the bottom left corner of that tetrad. All plotting uses this ‘floating origin’ technique; its advantage is that the symbol-drawing routines need not keep track of the varying positions of the symbols relative to a fixed origin. Instead they work relative to an origin which is always in the current tetrad, and a short and fairly simple routine moves the origin from tetrad to tetrad. The computer now examines all entry points to all symbols in the selected tetrad and picks the nearest one to the current pen position. An entry point is a point at which drawing of a symbol may start; for example, the ‘woodland occasional’ symbol, a vertical line, may be drawn either by taking the pen to the top centre of the tetrad and moving it down, or by taking it to the bottom centre and moving it up. Top and bottom centre are the two entry points for this symbol. For the thick symbol, a narrow vertical rectangle, there are four entry points, one at each corner. Similar entries exist for all symbols, and having chosen the handiest the computer raises the pen and takes it to that point. No movement may be necessary, as with ‘waterside’ and ‘water’ symbols in the same or horizontally adjacent tetrads, in which case the pen is not raised. Each symbol, including the circle, is built up from straight lines, and each straight line is defined by two items. It is rather easy to split KDF9 words into three 16-bit parts, known as the C, I and M parts. (Exactly why they are given these names would take us too far from our present subject.) Having selected a symbol entry point, the computer refers to a word whose C part specifies how many steps are needed to draw the first line of the symbol. The M part holds a bit pattern which can be interpreted as a step instruction by the plotter, so that between them the C and M parts completely specify the line. The I part is zero if that line completes the symbol, non-zero if at least one more line is needed to complete it. In the latter case, similar details for the next line are in the next word on the list. This can best be explained by an example. Three lines are needed to draw the ‘wood- land abundant’ symbol starting at its top left corner. The symbol is a rectangle 60 steps high and 2 steps wide, and the relevant control words are: Word 1 C part 60 (Steps down) Word 1 I part Non-zero (Another line follows) Word | M part ‘Down’ (Expressed as a plotter control) Word 2 C part 2 (Steps across) Word 2 I part Non-zero (Another line follows) Word 2 M part ‘Right’ (As a plotter control) Word 3 C part 60 (Steps up) Word 3 I part Zero (Symbol complete) Word 3 M part ‘Up’ (As a plotter control) Compare with this the single word used to draw the ‘woodland occasional’ symbol (vertical line) from the top: C part 60 (Steps down) I part Zero (Symbol complete) M part ‘Down’ (As a plotter control) Similar sets of words are provided for every entry point of every symbol, together with a ‘symbol directory’ giving details of where they begin in the control word array and of the pen positions before and after drawing them. Having plotted one symbol the computer removes its marker from the map image and checks if the current tetrad is now empty, i.e. all symbols in it have been plotted and their markers removed. If not, the process of finding the nearest entry point to the (new) pen position, going to it, plotting the symbol and removing its marker is repeated until the tetrad is clear, when the next occupied tetrad is found, dealt with and cleared in the same Watsonia 6 (6), 1968. 364 J. G. HAWKES, B. L. KERSHAW and R. C. READETT way. After one row has been plotted the next is examined, the decision to plot from left to right or right to left made afresh and the whole process repeated until the map is complete. The map area is extended by four rows above and below the county, and the fiducial marks are created by putting dummy ‘woodland occasional’ and ‘waterside occasional’ observations into the top left and bottom right tetrads. These plot as vertical and hori- zontal lines, which combine to form ‘+’ crosses whose centres, always in fixed positions relative to the map grid, are the fiducial points. On magnetic tape a line is represented by two or four 6-bit characters. In the two character case, the first is the actual control symbol for the plotter, which occupies 5 of the 6 bits. The next 6 bits are a binary number giving the length of the line in terms of plotter steps required to draw it. If the line is longer than 63 steps, the sixth (32) bit of the first character is made 1 and the second character is the quotient on dividing the number of steps required by 64. Any remainder is dealt with by another two characters, the first with its 32 bit zero and the second directly specifying the remainder. This system cannot draw lines longer than (63 x 64) + 63 = 4,095 steps. The longest possible line on a map, right from top to bottom, is 3,180 steps, so this limitation is not important. The final program simply transfers the information from magnetic tape to paper tape. It takes two 6-bit characters from magnetic tape, examines the 32 bit of the first and if it is 1, clears it and multiplies the second number by 64. The first character is then output to paper tape the specified number of times. This system, which effects a considerable economy of magnetic tape, is to be adapted as part of the on-line plotting system when Birmingham University’s plotter is delivered later this year. REFERENCES CADBURY, Miss D. A., HAWKES, J. G. & READETT, R. C. (1957). Flora of Warwickshire: Species distribution maps and habitat analyses. Proc. Birmingham Nat. Hist. Phil. Soc., 18, 135-145. Hawkes, J. G. & READETT, R. C. (1954). The Warwickshire County Flora revision: a new method of recording. Proc. Birmingham Nat. Hist. Phil. Soc., 18, 61-74. HAWKES, J. G. & READETT, R. C. (1963). Collecting the data. A description of the methods used in the current revision of the Flora of Warwickshire. Local Floras (Ed. P. J. Wanstall), B.S.B.I., London. Hawkes, J. G., READETT, R. C. & SKELDING, A. D. (1955). The Warwickshire County Flora Revision. Mapping distribution. Proc. Birmingham Nat. Hist. Phil. Soc., 18, 109-112. PERRING, F. H., & WALTERS, S. M. (1962). Atlas of the British Flora. B.S.B.I. and Nelson, London. Soper, J. H. (1964). Mapping the distribution of plants by machine. Canad. J. Bot., 42, 1087-1100. Soper, J. H. (1966). Machine-plotting of phytogeographical data. Canad. Geogr., 10, 1, 15-26. Watsonia 6 (6), 1968. PARIETARIA OFFICINALIS AND P. JUDAICA By C. C. TOWNSEND The Royal Botanic Gardens, Kew ABSTRACT A full account is given of the nomenclature and typification of the British species of Parietaria. It is shown that the type of P. officinalis L. is specimen No. 1220.2, and that of P. judaica L. specimen No. 1220.3, in the Linnaean Herbarium (LINN). P. officinalis is a species found in the Orient and central and northern Europe, but absent from the British Isles. The British species should be called P. judaica L., of which P. rami- flora Moench and P. diffusa Mertens & Koch are synonyms. The diagnostic features of P. officinalis and P. judaica are given. Although name-changes in the British flora have been not infrequent in recent years, comparatively few have been based purely on taxonomic conclusions; rather have the majority resulted from the exploration of the very necessary, but to most botanists rather arid, fields of bibliography and nomenclature. It would appear, however, that the name by which British botanists know the common Pellitory-of-the-wall must be changed on purely taxonomic grounds. In the current authoritative check-list of British plants (Dandy 1958), the Pellitory is listed as Parietaria diffusa Mertens & Koch, with P. officinalis auctt. and P. ramiflora auctt. in synonymy. These synonyms reflect two taxonomic conclusions; firstly, that the names P. officinalis L. (1753) and P. ramiflora Moench (1794) have been misapplied to our British species, and secondly, that taxa described under these names represent one or more species distinct from P. diffusa Mertens & Koch. The existence of at least two closely related species of Parietaria—one erect in habit with larger leaves, the other more decumbent with smaller foliage (to which the degree of accrescence and form of the perianth of the hermaphrodite flowers and various other characters have been added)—has been indicated by three principal authors, viz. Linnaeus (1756), Moench (1794) and Mertens & Koch (1823). Linnaeus, in the dissertation ‘Flora Palaestina’, described Parietaria judaica L. in the following terms: Parietaria Judaica accedit ad Parietariam officinalem, sed caule glabro, petiolis brevioribus, florum aceruis ad ramos, vix vero ulli ad caulem. This diagnosis was based on material collected by his student Hasselquist in Palestine, as the entry here ‘582 judaica H. Palaestina’ indicates. Moench contrasts his new species P. ramiflora with P. officinalis in the following manner: officinalis L., caule erecto: foliis lanceolato-ovatis: subhirsutis: florum glomerulis caulinis pluribus pedunculatis. [con Oederi, Tab. 521. ramiflora, caule decumbenti: ramis non caule floriferis: glomerulis sessilibus: folliis hirsutis ovatis. Parietaria judaica Linn. Parietaria officinalis Pollich. Pal. n. 940 huc pertinet. Folia minora atro-virentia. Mertens & Koch simplified the matter (for themselves!) by distinguishing two species of Parietaria and, feeling that other names had been confused by previous authors (as reflected in their synonymy), they renamed these species P. erecta and P. diffusa. The single name P. officinalis (of various authors) figures in the synonymy of P. erecta, while in the synonymy of P. diffusa the names P. officinalis ‘nach Smith’, P. judaica Wild. and P. rami- flora appear. From the above facts, and from Mertens & Koch’s further observations, it is clear that: 1. In describing P. ramiflora, Moench was also superfluously renaming P. judaica L. 2. In describing P. diffusa, Mertens & Koch believed themselves to be renaming 365 Watsonia 6 (6), 1968. 366 C. C. TOWNSEND P. ramiflora Moench, but did not consider that this was in fact true P. judaica L. Neither did they know for certain whether their P. erecta was true P. officinalis. The first requirement is thus to typify both Parietaria officinalis and P. judaica. Paclt (1952) has stated that the type of P. officinalis is in the Hortus Cliffortianus Herbarium at the British Museum, where the species is represented by two specimens, of which Paclt gives photographs. One of these is ‘P. erecta’ and is described by Paclt as the holotype; the other is ‘P. diffusa’, and is described by Paclt as the paratype. In point of fact, Linnaeus did not cite a holotype and therefore, by definition, there could be no paratype. If the Hortus Cliffortianus specimens could be regarded as types at all they would be syntypes, because Paclt selected the specimen resembling ‘P. erecta’ as the lectotype of P. officinalis L. Clearly, however, neither of these specimens can be regarded as the type of P. officinalis. Linnaeus’ concept of this species had altered from that which he held when writing the Hortus Cliffortianus by the time he wrote the Species Plantarum. Two points alone will illustrate this: 1. In the Hortus Cliffortianus, ‘Parietaria sicula, alsines folio. Bocc. sic. 47, t. 24 is cited under P. officinalis. In the Species Plantarum this has been transferred to P. Jusitanica. 2. In the Hortus Cliffortianus the main diagnostic phrase given for P. officinalis is ‘Parietaria foliis lanceolato-ovatis’, while in the Species Plantarum this reads ‘Parietaria foliis lanceolato-ovatis alternis’. From point (2) it is obvious that, in addition to his changed concept of P. officinalis, Linnaeus had before him at the time of writing the Species Plantarum (when the Hortus Cliffortianus Herbarium was not available to him) a further specimen enabling him to add the word ‘alternis’ to his diagnosis. Thus there can be no doubt that the true type of Parietaria officinalis is the specimen No. 1220.2, in the Linnaean Herbarium housed at the rooms of the Linnean Society of London (LINN). The statement by Scholz (1955) that P. officinalis must be rejected as a nomen ambiguum is quite incorrect. At this point, the name P. officinalis departs from consideration for the British Pellitory- of-the-Wall. Moench typified this species correctly, for the Linnaean specimen is the plant also known as P. erecta, and is not a British species. Dandy is thus correct in calling the British Pellitory ‘P. officinalis auct.’ (non L.) There is in the Linnaean Herbarium one specimen (1220.3) labelled ‘judaica’ by Linnaeus in his own handwriting and bearing a sign indicating a western Asiatic origin. This is evidently the plant of the Flora Palaestina dissertation, and can be accepted as the holotype of P. judaica. Another specimen in the Linnaean Herbarium (1220.4) bears the abbreviation ‘H.U.’, indicating that it was cultivated in the Uppsala University Botanic Garden, which was under Linnaeus’ control; but the epithet judaica was added by Smith, the subsequent owner of the Linnaean Herbarium, and not by Linnaeus himself. The real problem is—are P. judaica L. (P. ramiflora Moench) and P. diffusa Mertens & Koch conspecific, or are they not? Opinions have differed widely, and during the course of naming general collections of plants from the Balkans and south-west Asia the writer has had an opportunity to test the various treatments, and to resolve the matter while preparing an account of the Urticaceae for the Flora of Iraq. All that can be said is that from the existing accounts one is quite unable to distinguish confidently between P. diffusa as represented by British material and P. judaica as represented by Oriental material. The greatest authority on the flora of eastern Europe and western Asia, Edmond Boissier, had no doubts in this matter. In 1875 he reduced P. diffusa to synonymy under P. judaica, describing at the same time a var. brevipetiolata, into which he sank P. multicaulis Boiss. & Heldr. Hayek (1924) took up the name P. vulgaris Hill, Brit. Herb. 491 (1757), for the decumbent Pellitory, and cited P. diffusa, P. ramiflora and P. judaica Hal. (non. L.) in his synonymy, with reference to a paper by Vierhapper (1919). P. vulgaris Hill, however, is a name which must be rejected under Article 23 of the International Code of Botanical Nomenclature, as an incidental binomial published in a work which did not consistently employ binomial nomenclature. Vierhapper appears to have attempted to separate P. judaica and P. ‘ramiflora’ by means of vegetative characters which become less impressive as one Watsonia 6 (6), 1968. PARIETARIA OFFICINALIS AND P. JUDAICA 367 studies material from various points in the geographical range of the aggregate as a whole. Hegi (1912) adopted the strange course of using the name P. ramiflora Moench, while placing P. judaica L. in synonymy, together with P. diffusa. This has been rectified by Schreiber (1958) in the second edition of Hegi’s Flora, who omits P. judaica from the synonymy of P. diffusa without giving reasons for so doing. Finally, the two European Floras which use, probably, the broadest and narrowest views of species were consulted. Fiori (1923) relegated both P. erecta and P. judaica to varietal rank under P. officinalis with P. ramiflora and P. diffusa in synonymy under P. officinalis var. judaica. In the Flora URSS, Jarmolenko (1936) separated P. judaica from ‘P. officinalis’ (i.e. P. ‘ramiflora’/ diffusa) and ‘P. erecta’ (P. officinalis), on the grounds that in P. judaica all the flowers are hermaphrodite, while in the other two species the lower flowers of the inflorescence are pistillate. Examination of the Linnaean type, and general material of P. judaica in herbaria, did not bear out this distinction; though the female flowers of P. judaica are always harder to see than those of P. officinalis L., at least one can be seen on the type specimen without dissection of the cymes. While examining the Kew herbarium (K) material of Parietaria an interesting feature arose from this distinction brought to light by Jarmolenko. This is the gradual reduction eastwards in the number of female flowers to be found in the decumbent Parietaria. In the extreme west of its range (e.g., in Great Britain), female flowers are plentiful. On reaching Greece these are still readily found but are considerably fewer in number. A specimen illustrating this is Heldreich, Herb. Fl. Hellen. 94, gathered on the Athens Acropolis. In western Turkey (e.g., Sintenis, Iter Trojanum 486), female flowers are still frequent, but as one moves eastwards they become progressively harder to find and are often absent. The writer could detect no female flowers, for example, on isotype material of P. multicaulis Boiss. & Heldr. (from Pisidia), or on Davis, Dodds & Cetik 19045, gathered at Nigde, C. Anatolia, or on several other recent specimens gathered by Dr. P. H. Davis and his collaborators. On the following specimens: McNeill 502 (Bitlis), Davis 16616 (Bozkir) and Aucher 9794 (Aintab), however, female flowers were scarce but certainly present. Sometimes in the Orient the female flowers are solitary and are found in the axils of the lower leaves—e.g., Maitland 100 (Baalbek, Syria) and Post 1015 (Beirut, Lebanon). When female flowers are present, their morphology is similar to those of the British and western European P. diffusa. Somewhat strangely, all the Egyptian material seen has plentiful female flowers. In Iraq, female flowers may be present or not in gatherings otherwise taxonomically indistinguishable, and in Iran female flowers are rarely found at all. The pattern of gradual decline makes specific or even infraspecific separation by means of this character impossible. Weddell (1856) attempted to separate P. judaica and P. officinalis var. diffusa (Mertens & Koch) Weddell by stressing the latter part of Linnaeus’ diagnosis— *Florum acerui ad ramos, vix vero ulli ad caulem’, observing: “Par ces mots: florum acerui, etc., l’ auteur de l’espéce a en effet mis en relief un des traits les plus essentiels de la plante, et celui qui la distingue le plus nettement du P. officinalis. En d’autres termes, les tiges cespiteuses du P. judaica sont sous-frutescentes et vivaces, tandis que celles du notre Pariétaire (P. officinalis) commune sont ordinairement herbacées et annuelles’. A footnote guards us against using plants of P. judaica flowering in the first year for this comparison. Weddell also observed in his description of P. judaica: ‘floribus hermaphroditis quam feminei multo numerosioribus’, but does not go so far as to say that female flowers may be absent. Béguinot (1908) did not find this growth character at all satisfactory, nor does it appear to be so; in the Orient many species become more woody than in Europe, often due to constant grazing by the ubiquitous goats, and all manner of transitional forms occur. Two specimens from the Rowanduz Gorge of Iraq, Guest 13094 and Guest 2091, show respectively a much-branched plant with almost all the flower-clusters on the branches and a less-branched plant with most of the cymes disposed along the stem. The remarks of Mertens & Koch follow much the same pattern as those of Weddell, but it is clear that none of these authors had seen the type specimen of P. judaica. No character has come to light by which P. judaica Watsonia 6 (6), 1968. 368 C. C. TOWNSEND may be separated from P. diffusa and P. ‘ramiflora’ or by which two distinct taxa representing the latter two species may be recognized. Thus, the earliest name available for the British Pellitory-of-the-Wall is P. judaica L. It is remarkable that this name has not hitherto been adopted in this country in view of the fact that Villars (1787) used the name P. judaica L. for the Western species before either P. ramiflora Moench or P. diffusa Mertens & Koch was published for which correct decision he was taken to task by Weddell (1856). Various authors have included the decumbent Pellitory (under one name or another) in P. officinalis L. at varietal or subspecific rank, but it is in fact a distinct species. P. officinalis has a much more restricted range, being chiefly found in central and northern Europe, reaching westwards to the Iberian Peninsula (where it is rare) and eastwards to northern Iran. P. judaica, on the other hand, extends further west to the British Isles and Macaronesia, east to Tian Shan and Pakistan and south to N.W. Africa, Egypt, Palestine and Iraq. The differences between the two species are dealt with in detail by Mertens & Koch; much of the difficulty in separating them has been caused by undue emphasis being placed on characters of foliage and habit, in both of which P. judaica in particular is very variable. Table 1 summarizes the characters found most useful by the present writer. TABLE 1. Comparative characters of Parietaria officinalis L. and P. judaica L. P. officinalis L. P. judaica L. Female flower Length in fruit 2-75-3 mm, much Length in fruit 2-0-2-:25 mm, not exceeding the bracts, very con- much exceeding the bracts, rather spicuous. Segments more broadly inconspicuous. Segments narrowly lanceolate, midrib and margins con- lanceolate, subaristate with the con- fluent only at the tip. fluence of the midrib and margins. Hermaphrodite flower Petrianth remaining campanulate in Perianth in fruit tubular, accrescent fruit, not exceeding 3 mm. to 3-3-5 mm. Achene More narrowly ovoid, 1-5-1-8 mm Shortly ovoid, 1-0-1-:2 mm long. long, frequently distinctly asymetri- cal. The following is a limited synonymy, omitting the numerous infraspecific taxa based mostly on leaf characters. PARIETARIA OFFICINALIS L., Sp. P/. 1052 (1753). P. erecta Mertens & Koch, Réhlings Deutschlands Flora ed. 3, 1, 825 (1823). P. officinalis L. var. erecta (Mertens & Koch) Weddell, Archs. Mus. natn. Hist. nat., Paris 9, 507 (1857). P. officinalis subsp. erecta (Mertens & Koch) Béguinot, Nuoy. G. bot. ital. 15, 341 (1908). PARIETARIA JUDAICA L., Fl. Palaest. 32 (1756); Amoen. Acad. 4, 466 (1759). P. vulgaris Hill, British Herbal 491 (1757), nomen nonrite binom. P. ramiflora Moench, Meth. 327 (1794), nomen superfluum illegitimum. P. assurgens Poir., Encycl. Méth. 5, 15 (1804). P. punctata Willd., Sp. Pl. ed. 4, 4, 953 (1805). P. diffusa Mertens & Koch, Réhlings Deutschlands Flora ed. 3, 1, 827 (1823). P. maderensis Reichb., Flora (Regensburg) 13, 131 (1830). P. officinalis L. var. diffusa (Mertens & Koch) Weddell, Archs. Mus. natn. Hist. nat. Paris 9, 507 (1857). P. officinalis L. var. ramiflora Aschers., Fl. Brandenb. 610 (1864). P. officinalis L. var. judaica (L.) Gort., Fl. Friul. 2, 140 (1906). Watsonia 6 (6), 1968. 369 PARIETARIA OFFICINALIS AND P. JUDAICA ‘bl X IV “(etoyeur rbeay “9g9€7 beay jo wmnireqioyy [eUOeN ul IMeY ody) susyoe “p Syjueliod syewoy “¢ SyyuelIod oyposyd “euloy *Z :pouodo “yyuelod sews ‘] “YT volwpnl -g ‘g ‘([elroyew uvliesuny ‘goge joury WIOJJ) OUDYR “Pp -Yjueliod oyposydeunosy ‘¢ ‘yjuersod syeursy ‘z < pousdo Yiueltod seulas ‘[ ‘Ty syousoifo viavjaiavg “Wy [Sly ekSy y OW Cain ey | | | | | | | | | | | | | | | | | | | | eS Watsonia 6 (6), 1968. 370 C. C. TOWNSEND P. officinalis L. subsp. judaica (L.) Béguinot, Nuoy. G. bot. ital. 15, 342 (1908). P. officinalis L. “Rasse’ ramiflora Aschers. & Graebn., Syn. Mitteleurop. Fl. 4, 623 (1911). I am grateful to Dr. W. T. Stearn and Mr. A. A. Bullock for kindly confirming the Linnaean typification set out above and raising interesting additional points in connection with this, and to the latter and Mr. J. E. Dandy for correcting certain terminological aberrations; also to Miss Mary Grierson for preparing the illustration. REFERENCES BEGUINOT, A. (1908). In Fiori, Béguinot & Pampanini, Schedae ad Floram Italicam Exsiccatam, Nuov. G. bot. ital., 15, 341-344. BolssigER, E. (1875). Flora Orientalis, 4, 1149-1150. Geneva, Basle & Lugano. Danpy, J. E. (1958). List of British Vascular Plants. London. Frori, A. (1923). Nuova Flora Analitica d’Italia, 1, 378-379. Firenze. Hayek, A. (1924). Prodromus Florae Peninsulae Balcanicae, 1, 94-95. (Reprium nov. Spec. Regni. veg. Beih., 30). Heat, G. (1912). I/lustrierte Flora von Mittel-Europa, 3, 142-144. Miinchen. JARMOLENKO, A. V. (1956). Urticaceae, in Komarov, V. L. (ed.), Flora URSS, 5, 399-401. Leningrad. LINNAEUuS, C. (1756). Flora Palaestina sistit B. J. Strand. Uppsala. LINNAEUS, C. (1759). Amoenitates Academicae, 4, 466. Uppsala. MERTENS, F. C. & Kocu, W. D. J. (1823). Réhlings Deutschlands Flora, ed. 3, 1, 825-828. Frankfurt am Main. MoeEncH, C. (1794). Methodus Plantas Horti Botanici et Agri Marburgensis, 327. Marburg. PAcLT, J. (1952). Uber die Identitét von Parietaria ramiflora Moench mit Parietaria erecta Mertens et Koch (= P. officinalis L.). Phyton, 4, 46-50. ScHOLZ, H. (1955). Parietaria erecta Mertens & Koch und Parietaria ramiflora Moench. Phyton, 6, 31-32. SCHREIBER, A. (1958). In Hegi, J//ustrierte Flora von Mittel-Europa, ed. 2, 3, (1), 304-307. Miinchen. VILLARS, D. (1787). Histoire des Plantes de Dauphiné, 2, 346. Grenoble, Lyon & Paris. VIERHAPPER, F. (1919). Beitrage zur Kenntnis der Flora Griechenlands. A. Anthophyta und Pteridophyta. Verh. Zool.—bot. Ges. Wien, 69, 292. WEDDELL, H. A. (1856). Monographie de la Famille des Urticacées, 506-511. (Archs. Mus. nat. Hist. nat., Paris, 9). Since this paper was submitted for publication, J. Mennema (Leiden) & S. Segal (Amsterdam) have published two papers on ‘Het geslacht Parietaria in Nederland’, in Gorteria 3 (7), 96-102 (1967) and 3. (8) : 109-118 (1967). Working primarily from a Western European viewpoint, they reach the same con- clusions as the present author; unfortunately, Dr. Mennema and I each heard of our interest in this. problem at too late a stage for a co-operative effort to be produced. Watsonia 6 (6), 1968. THE HYBRIDS OF MIMULUS CUPREUS By R. H. ROBERTS 51 Belmont Road, Bangor ABSTRACT The hybrid M. 2 guttatus x 3 cupreus has been synthesized and found to be completely sterile. All attempts to obtain the reciprocal cross have failed. Naturalized plants from several northern British localities are identical with plants of the synthesized hybrid, with one exception. This is a double-flowered plant from Scalloway, Shetland, v.c. 112. The results presented here suggest that this plant is also a hybrid of the same parentage. M. cupreus and M. luteus have been found to be completely interfertile, and it is suggested that M. cupreus may not merit specific rank. The hybrid M. 9 guttatus x 3 (M. luteus x cupreus) has also been produced and is also completely sterile. INTRODUCTION In a previous note (Roberts 1964) it was suggested that certain plants of naturalized Mimulus from Scotland and Shetland might be the hybrid of the North American species M. guttatus DC. and the Chilean M. cupreus Dombrain.* A similar plant has also been seen from Cumberland, v.c. 70. Like the hybrid between M. guttatus and M. luteus L., all of these have highly sterile pollen and a short pubescence on the calyx, pedicels and upper parts of the flowering stem. They differ from M. guttatus < luteus mainly in their smaller size and in having a more or less uniformly copper-coloured corolla, marked only with small spots in the throat. In the M. guttatus < luteus hybrid the corolla is yellow and usually has some larger reddish spots on the lobes, in addition to numerous small dots down the throat. An additional feature of the Shetland plant is that it has a petaloid calyx, coloured like the corolla, but spotted with somewhat larger and more deeply coloured spots. The pollen of this plant also has a higher percentage of fully-formed grains (Table 1), but in all other respects it is indistinguishable from the single-flowered putative hybrid. TABLE 1. Percentage pollen fertility in the naturalized hybrid M. guttatus x cupreus. Flower no. Locality 1 2 3 4 Single-flowered plants: Vale of St. John 3395) 7 15 10-5 Rhunahaorine 3 0 Voss) 2 Double-flowered plants: Scalloway 24 35 30 26 River Whiteadder 24 The artificial synthesis of this hybrid was undertaken and the results are summarized below. Although the differences between M. guttatus and M. luteus were first pointed out as early as 1813 by de Candolle (quoted by Grant 1924), other botanists believed them to be conspecific. It was not until some years later, when seeds had been sent from South America by various collectors, that it became possible to distinguish M. /uteus from its North Ameri- can relative. M. guttatus is a more or less pubescent perennial, often glandular-pubescent above and glabrous below. It is highly variable; in North America some local races of the aggregate are said to be annual, and a few races are entirely glabrous. It can generally be separated * M. cupreus Dombrain, Floral Mag., 2, t.70 (1862). 371 Watsonia 6 (6), 1968. S72 R. H. ROBERTS from other members of the genus by its stout, fistulous stems, and ovate or orbicular leaf-blades, which grade into small bracts at the top of the racemose inflorescence. It has a bilabiate, yellow corolla, with the lower lip much longer than the upper, and the throat nearly closed by two densely hairy ridges, marked with small red or brown spots. It is a native of western North America, where it is the commonest and also the most poly- morphic species of the genus, occurring along streams and in wet places from Alaska to northern Mexico (Grant 1924; Pennell 1951; Vickery 1959). M. luteus, on the other hand, is a completely glabrous, more slender plant, with de- cumbent flowering stems and smaller, acute leaves. The inflorescence also has fewer flowers and longer pedicels than M. guttatus. The corolla tube is relatively narrow, but the throat is open and red-spotted, while the lips of the corolla are nearly equal, with spreading lobes. Corolla colour is more variable than in M. guttatus, ranging from pale to deep yellow, and either unspotted, as in M. guttatus, or with large reddish-brown spots on one or more of the lobes. It is native only in Chile, where it is common along streams and in moist places (Grant 1924). M. cupreus is similar in many respects to M. luteus, but differs in being usually annual and sometimes puberulent. It is a smaller plant, with a compact habit and smaller, sub- rhombic-ovate or elliptical leaves. The corolla is also like that of M. Juteus, with a wide throat, nearly equal lips and spreading lobes; it differs only in its smaller size and in its colour: golden-yellow at first, changing to a brilliant copper colour at maturity. This species is restricted as a native plant to southern Chile (Grant 1924). Since M. luteus and M. cupreus occupy the same geographical region, their crossability has also been investigated and the results are discussed below. RESULTS M. cupreus has been reported as an escape in the Lake District by McClintock (1957). However, all the specimens of naturalized Mimulus with copper-coloured flowers which have been examined, including one from the Vale of St. John, Cumberland (collected by Mrs. E. M. Satow), were found to have highly sterile pollen. Seeds of two horticultural strains of M. cupreus were obtained, and the plants raised from them were used for crossing with M. guttatus. The same material of M. guttatus was used as in previous hybridizations with M. luteus: one plant from the River Ogwen, Caernarvonshire, v.c. 49, and another from Dolgellau, Merioneth, v.c. 48. Thirty cross-pollinations were made in both directions, the usual precautions being taken of emasculating the flowers in bud, and covering with a cellophane bag after pollen transfer had been effected. In the following account the name of the female parent is given first. The cross M. guttatus < cupreus produced a fairly good seed set. Eight of the resulting capsules were opened before dehiscence and gave the following seed counts: 315, 430, 286, 210, 324, 184, 452 and 416. But these seeds were smaller and more variable in size than those of the parent species, and their germination was much lower: around 20 per cent to 35 per cent compared with 87 per cent to 100 per cent in the parents. The reciprocal cross M. cupreus X guttatus produced an average of approximately 340 empty testae per capsule. Since many of these looked deceptively like good seeds, the contents of all the capsules were kept on damp blotting paper for ten weeks to test for any germination. Only one fully-developed and two poorly-developed seeds were seen, and even these failed to germinate. These results were almost identical with those obtained when M. guttatus was crossed with M. luteus, the only difference being that one seed of the cross M. luteus X guttatus germinated and grew into a robust plant (Roberts 1964). As might be expected from the poor quality of the seeds, the seedlings of M. guttatus Xx cupreus were generally smaller and weaker than those of the parent species; several had only one cotyledon and died 3 or 4 days after germinating. Fifty seedlings were transplanted Watsonia 6 (6), 1968. THE HYBRIDS OF MIMULUS CUPREUS 59/2. into sterilized soil, but three of them proved to be severely stunted and malformed. In one of these the leaves bore hardly any resemblance to those of normal Mimulus; they were smoothly elliptical with no trace of serrations, and the mid-vein was continued as a pro- jection from the lower surface of the leaf a short distance from the tip. The second had minute, alternate leaves, and produced root initials from the leaf axils, as well as from various points between the nodes. These two survived for over 3 months, but remained extremely dwarfed and slender. They were both under 2 cm high when they became chlorotic and died. The third showed similar anomalies, but was so minute and weak that it only reached a stature of 0-7 cm before dying. The remainder of the transplanted seedlings soon overcame their initial disadvantage and grew into vigorous plants. The majority of them were taller and more robust, and produced more flowers per plant than either of the parent species. Hybrid vigour was also shown in the larger size of the flowers, which were uniformly copper coloured with small spots only in the throat of the corolla. All of these plants were puberulent, mostly on the calyces, pedicels and upper parts of the flowering stems. There was, however, considerable variation in the density of the puberulence, and those plants in which it was very sparse could easily be mistaken in the field for M. cupreus. Pollen from a few flowers of each of these plants was examined in a few drops of aceto-carmine, when the fully-formed grains took up the stain and were easily counted. As the results from all the Fl progeny were so similar, those from only ten plants are given (Table 2). Thirty-five flowers of these hybrids were either selfed or intercrossed with other TABLE 2. Percentage pollen-fertility in ten plants of the synthesized hybrid M. guttatus X cupreus. Plant Flower no. reference no. 1 2 3 4 Cl 3 25 3 2 C2 0 4-5 2 0:5 C3 3-5 5°5 5 a C4 10 9-5 4-5 13 C5 2 4 3 4 C6 0:5 5 2-5 Z C7 6)95) 6 3°5 6 C8 3 4:5 7 —— C9 8-5 6 11 TSS C10 0:5 3 2 1:5 F1 individuals; twenty-five flowers were pollinated with M. guttatus and another twenty- five with M. cupreus pollen. Pollen of the Fl plants was aiso used on the stigmas of ten flowers of each of the parents. In no case was any seed obtained. Selfing and back-crossing were also tried with two plants of the naturalized putative hybrid which had been brought into cultivation. These were the single-flowered plant from the Vale of St. John, Cumberland, and the double-flowered plant from Scalloway, Shetland. In the case of the Cumberland plant the results were exactly the same as with the synthesized hybrid: six of its flowers were selfed, and another six pollinated with M. guttatus and six with M. cupreus pollen. Again not a single seed was obtained. This plant was morphologically identical with the synthesized Fl, and its pollen was similarly highly sterile (Table 1). It is therefore presumed to be a hybrid of the same parentage. In addition to having a double flower, the Shetland plant differed in two other respects: its pollen had a higher percentage of well-formed grains, and it produced a very low number of seeds, both after selfing and after pollination with M. guttatus. During 1963 and 1964 a total of eighteen flowers were selfed, ten others were pollinated with M. guttatus and another ten with M. cupreus pollen. The selfed flowers produced from 0 to 3 seeds per Watsonia 6 (6), 1968. 374 R. H. ROBERTS capsule; only twenty-three seeds were obtained in the course of two seasons. Seventeen of these germinated, but several of them had small, malformed cotyledons, and others lacked the normal green colour. Although they were transplanted into sterilized soil none of them made any growth beyond the cotyledon stage, and all of them died within a fortnight of germination. From the capsules which had been fertilized with pollen of M. guttatus a total of twelve seeds was obtained. Seven of these germinated readily and the resulting progeny are discussed below. Pollination with the pollen of M/. cupreus gave no seeds at all, neither were any seeds set when pollen of the Shetland plant was placed on the stigmas of M. guttatus or M. cupreus. The back-cross progeny of the Shetland plant with M. guttatus showed many peculiar- ities. They were all densely covered with long, viscid, glandular hairs, and exhibited various flower abnormalities, the commonest being severe malformation of the stigma, style (often trifid) and ovary. In some of these plants the flowers were small (corolla less than 1-5 cm long), malformed, and failed to open; later flowers on the same plants, however, were less severely stunted and opened normally. Curiously, all of them had petaloid calyces like the Shetland plant, and flower colour was more or less intermediate between that of the parents. The pollen of these back-crossed plants had a higher fertility than that of the seed parent (Table 3), and it was anticipated that seed production would show a corresponding TABLE 3. Percentage pollen fertility in three plants each of the first and second back-cross of the Shetland plant with M. guttatus. Bl, B2, B3: (MV. guttatus < cupreus) Xx guttatus BGI, BG2, BG3: [(M. gutt. x cupr.) < gutt.] < guttatus. Plant Flower no. reference no. 1 D 3 4 Bl 64°5 72 54 69 B2 66 58 60 32 B3 42 30 34 34 BGI 19-5 22 26 20 BG2 62 55) 58 64 BG3 23 32 28 25 increase. Selfing and inter-crossing of the back-crossed individuals gave only a slightly higher seed-set than was seen in the Shetland plant (from three to seven seeds per capsule in eight capsules), but none of them germinated. Ten capsules which had been fertilized with pollen of M. guttatus gave a closely similar result: only thirty-four seeds were obtained altogether. But these were better developed and germinated in 3 to 5 days. The plants of this second back-cross to M. guttatus, i.e. (Shetland plant x guttatus) < guttatus, grown in 1966, showed much the same abnormalities as the first back-cross. Some of them had greatly reduced vigour and minute corollas which did not open, but withered on the plant without dropping. One or two of the more vigorous plants had flowers which were closer in shape and colour to those of M. guttatus. But it was remarkable that all of them possessed petaloid calyces, even when the corolla itself was severely reduced. Seed production in these plants was no greater than in the seed parents, while pollen fertility showed a slight downward trend (Table 3). The plants of both back-cross generations appeared to be less vigorous than the putative Fl Shetland plant, and it seems extremely doubtful whether any of them would have survived in the wild. All the available evidence therefore supports the view that the Shetland plant is the hybrid M. guttatus < cupreus, and that its petaloid calyx is the result of a simple mutation which may also be responsible for its higher pollen fertility and its ability to set a few viable seeds. It is difficult otherwise to account for the appearance of Watsonia 6 (6), 1968. | THE HYBRIDS OF MIMULUS CUPREUS SM) the double-flowered character in every individual of successive back-cross generations. Among the plants of M. cupreus raised from seed, several were found with one or more of the calyx-lobes enlarged and coloured like the corolla. It seems possible therefore that the petaloid-calyx character in the Shetland plant has been inherited from its M. cupreus parent. Nevertheless, the appearance of gross morphological abnormalities in the first and second back-cross generations indicates a severe genetic disharmony between M. guttatus and M. cupreus despite the ease with which the F1 is formed. The plants of the two horticultural strains of M. cupreus were crossed in both directions with M. luteus (progeny of a naturalized plant found in Caithness by D. P. Young in 1959). The hybrids were easily obtained, and there was a full set of well-developed seed, equal in every respect to those of the parents. There was no difference in their quality whichever species was the seed parent, and germination in both crosses was very high: from 94 per cent to 100 per cent. The resulting seedlings were very vigorous, and showed no morpho- logical abnormalities. Hybrid vigour in these crosses was displayed in their larger size, bigger flowers (with various combinations of the corolla colour and markings of the parent species), and greater number of flowers and capsules. Pollen fertility was unimpaired, and seed production showed no diminution either on selfing or back-crossing to the parents. The F2 progeny similarly suffered no reduction in either pollen fertility or seed production, and it is evident that M. cupreus and M. luteus are completely interfertile. These fertile hybrids were cross-pollinated in both directions with M. guttatus. The results were identical with those obtained when M. luteus and M. cupreus were crossed with M. guttatus: hybrids were only produced when the latter was used as the female parent, and they all proved to be completely sterile. DISCUSSION These observations strongly suggest that the coppery-flowered Mimulus naturalized in several northern localities in Britain are mostly hybrids of M. guttatus and M. cupreus, the former being the seed parent. They are easily mistaken in the field for M. cupreus. They can, however, be distinguished from it by the presence of a minute pubescence on the calyx, pedicels and upper parts of the flowering stems, particularly the leaf-bases and nodes, and by the complete absence of normally-developed capsules. Confirmation of their hybrid nature can be obtained from microscopic examination of the pollen, which, as has been seen, consists mostly of empty and shrivelled grains. Even in the double- flowered Shetland plant, whose pollen fertility is somewhat higher than that of the single- flowered hybrid, the full grains constitute only 20 per cent to 35 per cent of the total. These figures, based on larger samples of pollen, are somewhat lower than the provisional estimates formerly given (Roberts 1964). A specimen recently collected in Berwickshire, where it has been ‘known for about 40 years in the River Whiteadder and its tributaries’ (D. McClintock, personal communication, 1966), is practically identical with the Shetland plant. The only flower available for examination had a pollen fertility of 24 per cent, and there is little doubt that this plant is also the same hybrid. The ability of the double-flowered hybrid M. guttatus < cupreus to produce a few seeds after selfing implies that some of its pollen is functional. The production of viable seeds after fertilization with the pollen of M. guttatus lends further support to the view that the latter is one of the parents of the double-flowered hybrid. This low seed set also seems to contrast sharply with the complete failure to obtain seed from the single-flowered hybrid—both naturalized and synthesized—and to conflict with the conclusion that the double-flowered hybrid is an Fl of the same parentage. But this difference between the two forms of the hybrid may not be as significant as it appears, and, in fact, may be only a matter of degree. Stebbins (1950) quotes an instance of hybrids in the genus Paeonia which, though almost completely sterile, produce one seed in perhaps one thousand or ten thousand, or even one in a million. There is, so far, no authentic record of M. cupreus as a naturalized plant in the British Watsonia 6 (6), 1968. 4 376 R. H. ROBERTS Isles. The most likely origin of its hybrid with M. guttatus is from a deliberate garden cross, at the time when Mimulus was popular in gardens around the middle of the 19th century. The persistence of the hybrid as a naturalized alien has no doubt been facilitated by the ease with which it propagates and spreads vegetatively, as well as by its hardiness which enables it to survive the winter in Britain. In this respect it is like M. guttatus. The complete interfertility of /. cupreus and M. luteus seems to lend support to the view (e.g. Bailey 1927) that their morphological differences are hardly sufficient to maintain them as separate species. The analogous behaviour of both species when crossed with M. guttatus also implies that the genetic differences between them are small. Reference to the horticultural literature indicates that it is these two species which have been crossed to produce the vigorous, large-flowered, fertile hybrids which were once popular as garden plants. It is probable that some of the sterile naturalized plants, recently shown to be widespread in Britain, are the trihybrid M. guttatus x (luteus < cupreus). Artificially produced plants of this hybrid were so variable that it was found impossible to distinguish many of them from the hybrid M. guttatus < luteus. LOCALITIES FROM WHICH THE HYBRID M. GUTTATUS X CUPREUS HAS BEEN SEEN v.c. 70, Cumberland: Vale of St. John, Mrs. E. M. Satow. v.c. 81, Berwick: River Whiteadder and its tributaries, 1966, Miss E. Logan Home. v.c. 89, North Perth: River Tay, near Dunkeld, 1958, D. P. Young (herb. D. P. Young). v.c. 90, Angus: Tarfside, Glen Esk, 1960, Miss U. K. Duncan (herb. U. K. Duncan). v.c. 91, Kincardine: near Fettercairn, 1960, Miss U. K. Duncan (herb. U. K. Duncan). v.c. 101, Kintyre: Rhunahaorine area, south of Gigha Ferry, 1964, A. G. Kenneth (herb. R. H. Roberts). v.c. 103, Mid Ebudes: Coll, 1966, Miss U. K. Duncan (herb. R. H. Roberts). v.c. 108, West Sutherland: near Inverkirkaig, 1961, Miss V. Gordon (herb. V. Gordon). v.c. 112, Shetland: near Scalloway, 1962, R. C. Palmer & Walter Scott (herb. R. H. Roberts). Specimens of the synthesized hybrid have been deposited in the herbarium of the National Museum of Wales, Cardiff (NMW). ACKNOWLEDGMENTS The living material of the naturalized hybrid was kindly supplied by Mrs. E. M. Satow, of Great Langdale, and Mr. Walter Scott, of Scalloway, to both of whom I wish to express my thanks. I am also indebted to all those who sent the herbarium specimens without which these observations could not have been made. REFERENCES BAILEY, L. H. (1927). The Standard Cyclopedia of Horticulture. New York. GRANT, A. L. (1924). A monograph of the genus Mimulus. Ann. Missouri Bot. Gard., 11, 99-388. McCtiintTock, D. & Fitter, R. S. R. (1956). The Pocket Guide to Wild Flowers. London. Supplement by D. McClintock (1957). PENNELL, F. W. (1951). In Abrams, L.R., I/lustrated Flora of the Pacific States, 3, 688-731, Stanford, California. RoseErTS, R. H. (1964). Mimulus hybrids in Britain. Watsonia, 6, (1), 70-75. STEBBINS, G. L. (1950). Variation and Evolution in Plants. New York. VICKERY, R. K. (1959). Barriers to gene exchange within Mimulus guttatus (Scrophulariaceae). Evolution, 13, 300-310. Watsonia 6 (6), 1968. THE INTERSPECIFIC RELATIONSHIPS OF JUNCUS EFFUSUS AND J. CONGLOMERATUS IN BRITAIN By A. D. Q. AGNEW Botany Department, University College Nairobi, P.O. Box 30197, Nairobi, Kenya ABSTRACT Juncus effusus and J. conglomeratus appear to be specifically distinct in the British Isles; J. conglomeratus is the rarer species. Their principal diagnostic features are given. Evidence is presented to show that the two species hybridize at higher altitudes in Britain, probably because the overlap of their flowering periods is longer in these regions than at low altitudes. Evidence is also found for gene flow within J. effusus populations from J. conglomeratus and the reasons for the continuing specific distinctness of these two species is discussed. INTRODUCTION Juncus effusus has long been known as an extremely variable species. This is reflected at once both in its wide range of ecological tolerance (Agnew 1961) and in its morpho- logical variability, and it is clear that, since both these features are simultaneous expressions of the inherent variability, no full appreciation of the species can be made without consider- ing its variation with respect to its taxonomy. From the start the species in its relation to J. conglomeratus has been a puzzle to taxonomists. Thus, although the species is known under its Linnaean name, Linnaeus himself seems to have had no clear idea of the differences delimiting it from J. conglomeratus or J. inflexus, two closely related species, although he evinced some interest in them, as his annotated copy of the Species Plantarum (1753) shows. The position may be summarized thus: in the Linnaean Herbarium (LINN) the specimens annotated by Linnaeus as J. conglomeratus (Savage 1945, no. 449.2) are in fact J. effusus var. compactus as we know it, while those annotated as J. effusus (Nos. 449.4 and 449.5) are J. inflexus as we know it. Due to the extraction of some parts of the original description of J. effusus for inclusion in other later species (this has been summarized by Tweed and Woodhead 1947) and subse- quent further doubt concerning the delimitations of the two species, they were both in- cluded under the one specific name J. communis by Meyer (1819). Krisa (1962) has drawn attention to this confusion in Czechoslovakia, and has con- cluded that the two species are only recognizable as end points in a continuously variable series of forms, which are regarded as introgressive hybrid populations. The whole series of forms including both species is regarded by Krisa as a ‘conspecies’. It will be noted that the conclusions of the present author, who upholds the complete validity of the two species, based on examination of material from the British Isles, differ from those of Krisa. It is due to this discrepancy in views that the following account of work under- taken between 1952 and 1955 is presented. The present author has examined some stands of the two species in Czechoslovakia and feels that the situation there may not be very different from that in Britain, once the restricted nature of the much rarer J. conglomeratus is recognized. Table 1 is a list of the specific differences which have been found most constant in Britain. CHARACTERS FOR BIOMETRIC ANALYSIS Work on the variation of both species was done at Edinburgh (Scotland) in 1952 and in 1953-55 at Bangor (North Wales). To make comparisons between J. effusus and J. conglomeratus, characters must be used which show a difference between the two species; the form of the inflorescence is such 377 Watsonia 6 (6), 1968. 378 A. D. Q. AGNEW TABLE 1. Diagnostic features of J. effusus and J. conglomeratus. Character Stem sheathing scales Stem-ridges at inflorescence Spathe-length Bract of second cyme Inflorescence Perianth Stamens Style Fruit J. effusus L. Black to red-brown at base 30-60 Usually 15-30 cm Not as long or as long as split in spathe Diffuse or compact Outer whorl of three segments about 0-5 mm longer than inner whorl Perianth segments olive with green or light brown centres Anthers oval, shorter than filaments Very short or absent Retuse or truncate, not apiculate J. conglomeratus L. Red-brown to olive at base 15-24 Usually 5-15 cm Longer than split in spathe Almost invariably compacted into globose head Outer whorl as long as inner whorl Perianth segments with dark red— brown edges, lighter centres Anthers strap-shaped, longer than filaments One third as long as ovary Truncate, apiculate a character. The other two chosen were ridge-number and the length of the spathe above the inflorescence. These will be dealt with in turn. Since the varietal classification of both J. effusus and J. conglomeratus depends on the condensed or diffuse nature of the inflorescence, this is an obvious character to choose; however there are some difficulties in its measurement. The method chosen has been to measure three axes of the inflorescence at right angles to one another, taking the first measurement parallel to the stem and spathe. The product of the three lengths is then calculated and from this and a count of the total number of flowers in the inflorescence a figure is obtained of the number of flowers per unit volume (ml) of inflorescence (‘conden- sation’). It will be seen that this is in many ways an arbitary figure since the product of the axes does not give the true volume of inflorescence. However, sufficiently comparable figures can be obtained in this way. The axes were measured to the nearest millimetre and commonly were less than 10 mm in length. Clearly, this introduces an error in the figure obtained for ‘condensation’ which increases as length decreases; thus the variance of figures obtained for condensation should increase as the mean condensation increases. This has proved a drawback to the treatment of J. conglomeratus by this character, and it has been necessary to transform all figures of condensation logarithmically before statistical treatment. Spathe length is short (5-15 cm) in J. conglomeratus and long (15-25 cm) in J. effusus. However it is not a very good character since there are many intermediates. It is very quick to measure and does give a separation of the two species on the scatter diagrams mentioned below. Ridge-number on the stem below the inflorescence has been found to be very useful indeed. Not only is it easy to observe, especially when the stems have been left to dry for a period, but it gives effective separation between J. effusus and J. conglomeratus with no overlap and takes an intermediate value for hybrids. An experiment was carried out to test the plasticity of these characters. Two adjacent populations of J. effusus differed in their growth form, i.e. clumped and unclumped (diffusely growing), and occupied distal ends of a soil and vegetation catena from very wet conditions (the unclumped form) to drier grassland (the clumped form). Twenty individuals of each population were transplanted and grown under similar conditions; the results are sum- marized in Table 2. The previously significant difference in the spathe-length disappeared Watsonia 6 (6), 1968. Number of stem ridges JUNCUS EFFUSUS AND J. CONGLOMERATUS S79. TABLE 2. Transplant experiment of J. effusus plants from Braint, Anglesey. Mean Significance Mean Significance Meanno. Significance spathe- of ridge of ‘flowers of length difference no. difference per ml’ difference Before cultivation: Clumped population 15-48 +--+ 43-9 N.S. 43-05 a aigaty Unclumped population 19-90 42-9 20-28 After cultivation: Clumped population 16-12 N:S. 34-60 N.S. 72:53 IP SF Unclumped population 14-13 36-20 44-88 ‘t’ test probabilities given as: + = 5 per cent. >p>1 per cent., +-+ = 1 per cent. >p>0-1 per cent., +++ =p<0°1 per cent. under cultivation, but the difference in inflorescence condensation remained, indicating that this character at least is genotypically controlled. The value of this experiment is enhanced by the seed-size measurements. The mean seed-size of these two populations was almost identical so that it was not expected that a significant difference in any other character would remain after cultivation. At the same time these results do not invalidate the continued use of ridge-number and spathe-length as population parameters when comparing J. effusus and J. conglomeratus, although spathe-length is of doubtful use when J. effusus alone is dealt with. POPULATION ANALYSES The population analyses initially carried out were those of mixed J. effusus and J. con- glomeratus populations. A scatter diagram of such an analysis is shown in Fig. 1. It can be seen that there is a good separation of the two species on ridge-number, spathe-length, 60° =O - light perianth (J.effusus) @ - dark perianth (J. conglomeratus ) 0 30 40 30 ) 3 10 15 20 25 Spathe length (cm) Fig. 1. Scatter diagram of a community containing J. effusus and J. conglomeratus. The scatters of the two species are easily separated. Watsonia 6 (6), 1968. 380 A. D. Q. AGNEW and a qualitative character, that of perianth-colour. There are no complications in this diagram. Two populations have been found, one near Edinburgh (Fig. 2) and one near Glasgow, which do not show such a clear separation of the two species. Here there are individuals which have some characteristics of each species. These plants have been found to be easily separable from both J. effusus and J. conglomeratus by their intermediate number of stem ridges, the dark perianth segments (the latter typical of J. conglomeratus) and the intermediate character of the inflorescence. 60 O - light perianth @ - dark perianth Number of stem ridges 5 10 15 20 25 Spathe length (cm) Fig. 2. Scatter diagram showing an intermediate between J. effusus and J. conglomeratus, from a popu- lation at Penicuik, Midlothian. In North Wales this type of complex has been found on two occasions. The analysis of one of these populations (Fig. 3) shows as before a group of plants intermediate between J. effusus and J. conglomeratus. In this case there were enough plants of all groups to attempt a statistical analysis, when a positive correlation was found between ridge-number and spathe-length only in the intermediate group. This shows, in my opinion, that the inter- mediate type is a hybrid, since one would expect a hybrid to show correlations between the characters which distinguish its parents. It has been supposed in this case that these intermediates are groups of F, or subsequent generation hybrids, with a little back-crossing. Populations showing introgression have however been found at Edinburgh and on the Snowdon massif. Scatter diagrams of the former are shown in Figs. 4 and 5. They can be seen to include characters of both species in a complex which shows recombinations of many types. These populations show also introgression only with J. effusus, the scatter being between the position of the supposed hybrid and that of J. effusus. The correlation of these two characters, ridge-number and spathe-length, is significantly positive in both cases. That hybrids can occur between the two species has been recognized for some time (Buchenau 1906; Richards & Clapham 1941). The two species were artifically crossed in the garden in 1955. Seeds set in normal numbers in all the flowers crossed and germinated Watsonia 6 (6), 1968. Number of stem ridges JUNCUS EFFUSUS AND J. CONGLOMERATUS 381 O- light perianth (J. effusus) ©- dark perianth (putative hybrid) @- dark perianth (J. conglomeratus) 30 > (=) ul © 20 0 P) 10 IS 20 25 30 Spathe length (cm) Fig. 3. Combined scatter diagram of three populations, distinguished on this diagram by the use of three symbols. The intermediate grew spatially between the others and in an intermediate habitat. All were from within 100 m of each other at Llyn Idwal, Caernarvonshire. O - light perianth @ - dark perianth 30 E = 40 : Sats oo O : 8 30 (= =) Zz 20 10 , 0 5 10 15 20 25 Spathe length (cm) Fig. 4. Scatter diagram of a population from West Linton, Midlothian. The intermediate introgressive nature of the population can be seen by comparison with Fig. 1. Watsonia 6 (6), 1968. Number of stem ridges 382 A. D. Q. AGNEW O- light periantk © - diffuse inflorescence @ - dork perianth © - intermediate inflorescence 30 b- condensed inflorescence Oo 40 oS , awh. 30 ® 4° 20 be) l0 IS 20 25 Spathe length (cm) Fig. 5. A further population from West Linton, Midlothian, where inflorescence condensation is also represented. normally, the seedlings unfortunately being killed by damping off before reaching the 4-leaf stage. There seems no reason to suppose that they could not have developed into mature plants. The hybrids observed in the field moreover appeared fully fertile. This is in contrast to the statements made by Buchenau (1906, p. 137) that he considers the hybrids to be highly sterile and rare. Richards & Clapham (1941, p. 378) state that the presence of hybrids needs confirmation in Britain. It seems, then, that hybrids exist both as populations of hybrids and as populations introgressing with J. effusus. No signs of introgression with J. conglomeratus have been found in any population. This may be due to the type of crossing available to the species. As has been repeatedly noted (Richards & Clapham 1941; Buchenau 1906; Krisa 1962) the flowering times of the two species differ and are perhaps responsible for the lack of hybrids. J. conglomeratus flowers before J. effusus and the flowers of both are protogynous. This means that most crossing between species will be between the pollen of the earlier and the ovules of the later flowering species and this should hold true for hybrids as well. Thus all crossing, introgressive or otherwise, would be in the direction of J. effusus as seed parent. This will lead to the seeds being deposited around the plant genetically nearest to J. effusus giving an increased likelihood of the progeny being themselves crossed with J. effusus. Since the two species usually grow in slightly different habitats, J. conglomeratus preferring the drier areas, it can be seen that this effect could have great significance. Having shown the evidence for hybridization, I propose now to deal with the analyses of the variation within J. effusus itself. VARIATION OF J. EFFUSUS The figures for the variation of populations of J. effusus were obtained at Edinburgh with the exception of those analysed for the experimental investigation into the amount of phenotypic variation to be expected. A scatter diagram, made up of four populations, of condensation of inflorescence against ridge-number is shown in Fig. 6. The means of the Watsonia 6 (6), 1968. a. er ee a Number of stem ridges JUNCUS EFFUSUS AND J. CONGLOMERATUS 383 ® 0 20 P40) 60 80 OO 120 Inflorescence condensation (number of flowers per ml) Fig. 6. Scatter diagram of four populations of J. effusus where the individuals are numbered according to their population and the mean for each population is plotted as a large, uncircled figure. A population of J. conglomeratus (solid dots) is given so that it can be seen to lie on the extrapolated regression line of the four J. effusus populations. populations are shown within the diagram. It is evident that there is some correlation between these two characters, and that the correlation is in the general direction of the position that J. conglomeratus occupies in this diagram. With this in mind the four populations were analysed for correlations between the three parameters measured. The significance of any correlation observed is shown in Table 3. It can be seen that only one population shows significant correlations within TABLE 3. Correlations between and within four populations of J. effusus. Correlations determined between spathe length (A), number of ridges (B), condensation of inflorescence (C). Significance of correlation coefficient gives as: N.S., not significant; + = 5 per cent. >p>2 per cent.; +-+ = 2 per cent.>p> 1 per cent.; +++ = p<0-1 per cent. Analysis of each population: A/B A/C C/B 1 N.S. N.S. N.S. 2 N.S. N.S. N.S. 3 ste aie N.S. 4 N.S. N.S. N.S. Combined analysis of all populations: SP AP N.S. +++ Watsonia 6 (6), 1968. 384 A. D. Q. AGNEW itself and that both of these are positive correlations. The populations taken together show two significant correlations—between spathe-length and ridge-number (positive) and between ridge-number and condensation (negative). Anderson (1953, p. 287) has suggested that when a scatter diagram shows a correlation of characters there is evidence for gene flow from some other species, and that the characters of the species responsible for this may be found roughly by extrapolation—that is, by producing the line of the correlation until it enters the region in the diagram which would be occupied by a known species were it present. That species is suggested as the one responsible, by hybridization and subsequent backcrossing, for the gene-flow responsible for the correlation. In the present instance, it can be seen from Fig. 6 that the correlation line would in fact produce through the area representing J. conglomeratus on the diagram; that is, the area of high inflorescence-condensation and low ridge-number. In the same way the correlation of spathe-length against ridge-number also is in the direction of J. conglomeratus. The lack of correlations within populations shows that the correlation between these characters is not due to genetic linkage. The correlation between spathe-length and ridge- number is possibly not very important; as noted above spathe length is very plastic, and there is a correlation within one of the populations of the group. The correlation between ridge-number and condensation however appears to be very important. Not only are no correlations between these two characters shown within the population, but also these two characters seem to be the most environmentally stable of those used. The explanation, in the author’s opinion, is that gene flow is occurring from J. conglo- meratus into J. effusus populations. This would fit all the facts of variation within and between populations and explain why the correlations between populations are in the direction of the J. effusus—J. conglomeratus regression. Another line of evidence, from the mean seed-size of populations of J. effusus, confirms this hypothesis. Seed-size was taken as the product of length and breadth. Fifty seeds were taken from each population by taking at random 50 stems from each, shaking these over paper, and taking a small sample of the seeds so gathered. The size of samples was checked several times by measuring a further 50 seeds when it was invariably found that the means approached each other very closely. The means observed differed significantly from popu- lation to population and this was found to correlate with the altitude at which the sample was taken, the mean seed-size increasing with altitude. The means of seed-size are shown in Fig. 7. Analysis gives the regression as highly significant (p<0-1 per cent). Now J. conglomeratus has a larger mean seed size than J. effusus (0:13-0:15 sq. mm in the scale given in Fig. 7). Gene flow from hybridization at high altitudes could give a series of this sort with the genes of J. conglomeratus becoming progressively diluted with those of J. effusus as lower altitudes are reached. This hypothesis implies greater hybridi- zation at high altitudes; there is evidence for this and for its cause. The hybrids and introgressive hybrid swarms so far found in Scotland and North Wales have been at altitudes exceeding 150 m although both species occur, in non- hybridizing populations, down to sea level. There follows a list of the hybrids found, with their altitudes. Type of hybrid Place Altitude Probable F, Penicuik, Midlothian (Scotland) 183 m Probable F, Rutherglen Park, Glasgow 152m Probable F, Llyn Idwal, Caerns. (Wales) 366 m Probable F, Cwm Glas, Caerns. (Wales) 381 m Introgressive hybrid West Linton, Midlothian (Scotland) 305 m Introgressive hybrid West Linton, Midlothian (Scotland) 305 m Introgressive hybrid = Llyn Llydau, Caerns. (Wales) 457 m Taking the latitude into account, it can be seen that the hybrids so far found have been at the higher altitudes. This was realized some time ago and a search has been made in Watsonia 6 (6), 1968. JUNCUS EFFUSUS AND J. CONGLOMERATUS 385 Altitude (metres a.s.I.) 0.10 0.11 0-12 0-13 0-14 Mean seed size given as length x breadth (sq mm) Fig. 7. Mean seed size of twelve J. effusus populations from North Wales plotted against altitude of occurrence. lowland Junceta for the hybrids, as yet with no result. It is not intended to show that there are no hybrids at low altitudes, but that they are much more frequent on high ground. Since J. effusus is tolerant of the acid conditions so prevalent on British mountains of high rainfall, whereas J. conglomeratus is not, it follows that the number of habitats available to J. conglomeratus in the mountains are few. J. conglomeratus does however occur to 450 m on the Pentland Hills, Midlothian, and to 630 m on the Tarmachan crags above Killin (Scotland). It can be seen therefore that altitude is no barrier to this species in Britain. In view of the sympatry of the species it is surprising that hybrids do not occur more often. Anderson (1948) has postulated the existence of ‘hybrid habitats’ as accounting for many of the hybrids observed between sympatric but not concomitant species. In the present case, of course, the species are almost always concomitant. Possible ‘hybridity’ of the habitat occurs often through human interference, but although search has been made for hybrids in those types of community with agricultural disturbance, none has been found. The hybrids observed were only in one case found to occupy an intermediate type of habitat which was recognizable as such. This was at Llyn Idwal (Wales). FLOWERING PHENOLOGY One difficulty attendant on a comparative study of flowering times is that a large distance must be covered during the flowering season to observe sample localities. It has thus been found possible to sample only four localities, ranging from sea level to 381 m. It is un- fortunate that the information provided by the analysis of seed size was obtained, initially, somewhat fortuitously and no effort was made in previous years to study the phenology of the two species. Thus a record of flowering times has been made for only one year in North Wales. Watsonia 6 (6), 1968. 386 A. D. Q. AGNEW Difficulties have also been experienced in the measurement of flowering. The method finally employed was to take five random stems from the stand under consideration and dissect flowers from them. If both dehisced and undehisced anthers were present on any stem the stand was considered to be in flower. The data obtained are represented in Fig. 8. The altitude has been placed on a logarith- mic scale for convenience only. With the few samples taken it cannot be said to be significant - J.effusus ----—- J. conglomeratus 38! pea Altitude (metres a.s.I.) oO nm JUNE JULY AUGUST Fig. 8. Variation in flowering periods of J. effusus and J. conglomeratus with altitude in North Wales, in 1956. that this allows a straight line to join the starting points of flowering for each species. Due to the preference of J. conglomeratus for the more base-rich habitats there are no intermediate stations between 121 and 366 m at which both species may be found close enough together to be influenced by the same environmental conditions. Thus in this area of North Wales, at least, it is unavoidable that there should be a gap here. For reasons of transport and time it was impossible to visit continuously more than one site above 366 m. Fig. 8 shows that there is some overlap in flowering times at all altitudes. This has also been found in Midlothian. The high altitude population however shows the same difference in the start of flowering times between the two species but a much greater overlap between them. These few results are somewhat inadequate, but indicate that flowering times of the two species may show greater overlap and thus allow a greater chance of hybridization at high altitudes. No cause, phenological or otherwise, has been found for the apparent lack of hybrids over part of the range of the two species in Britain. DISCUSSION It has been shown that hybrids and particularly introgressive hybrids do occur between J. effusus and J. conglomeratus. Back-crossing does appear to take place, but only with J. effusus, and there is thus a possibility of gene flow from J. conglomeratus into J. effusus. That this occurs is shown by two lines of evidence: (i) correlations between characters which separate the two species in four populations of J. effusus taken together, but not within each population, and (ii) increasing seed-size with altitude. The latter corresponds Watsonia 6 (6), 1968. JUNCUS EFFUSUS AND J. CONGLOMERATUS 387 to the increased hybridization at high altitudes, allowing, so to speak, J. conglomeratus genes into the populations there, from whence they ‘percolate’ in increasing dilution to sea level. The greater overlap of flowering times observed at high altitudes is responsible for greater frequency of hybrids there. From the foregoing some tentative conclusions may be drawn as to the nature of J. effusus variation. It has already been stated that the species is extremely variable both ecologically and morphologically. This variation may be due in part to the possible genetic amplitude available to the species due to gene flow with J. conglomeratus. Thus J. conglomeratus has the more condensed inflorescence of the two species. If gene flow is occurring pre- dominantly in the direction of high to low altitude, one would expect the more compact forms of J. effusus at higher altitudes. This is found to be generally the case in the writer’s experience, and in that of Tweed and Woodhead (1946) who state that the variety (J/. effusus var. compactus) is the dominant one on high ground. In fact no truly effuse forms have been found by the writer in high or exposed habitats. However, this does not altogether fit the pattern of the ecological variation of J. effusus (Agnew 1961). J. conglomeratus is associated with the more basic type of habitat. Thus in North Wales at least it is associated with the basic rock outcrops of the mountains, and with the boulder clay and limestone regions of Caernarvonshire and Anglesey. The only site in my experience where it is dominant to the exclusion of J. effusus is on the Carboniferous Limestone at Penmon Point, Anglesey. No pH values of lower than 4-3 have been recorded for it by the present author, and Pearsall is quoted by Richards & Clapham (1941) as considering it more basicole than J. effusus. The flow of J. conglomeratus genes through J. effusus should confer a basicole tendency and, since the compact form of J. effusus presumably contains more of these genes than the effuse form, one would expect the latter to be the more acidicole variety. In fact the reverse is the case in North Wales at least. This is one of the problems that might be solved by further transplant experiments. There is also the problem of interspecific barriers between J. effusus and J. conglomer- atus. A hypothesis has been put forward in this paper to explain the apparent lack of introgression and gene-flow into J. conglomeratus, the latter species acting as a ‘gene reservoir’ to J. effusus. It is curious that J. effusus has remained distinct from J. conglomer- atus, since it cannot be envisaged that this gene flow is a recent process; the factors acting in favour of J. effusus cannot include a sterility barrier, and the effect of phenology has been discussed. There remains therefore selection pressure. Only one possible example of this has been found. It has been noted that the J. effusus plants of North Wales are prone to a fungal infection of the capsules. This was first inves- tigated during the comparatively wet summer of 1954 when many samples taken from a marsh area showed no uninfected capsules at all. The infection allows the carpel wall to reach maturity but it contains no seed, only a mass of fungal hyphae. This fungus was not identi- fied. It was noted that the compact variety of J. effusus was very much more heavily attacked than the effuse form, probably owing to contagion between capsules. In one pair of popu- lations, one compact, the other effuse, of 5,000 capsules examined from each, 2,560 in the former and 606 in the latter were infected. This is the only quantitative evidence obtained for selection against J. conglomeratus characters in J. effusus. It is interesting to note that this fungal infection was never observed by the present author on J. conglomeratus. In any case it seems likely that this is one of the factors operating in favour of the effuse J. effusus form, thus preventing the species from drifting wholly towards J. conglomeratus. ACKNOWLEDGMENTS Dr. P. H. Davis initiated and guided this study while I was an undergraduate, and it gives me pleasure to acknowledge my debt to him. I am also grateful to Professor P. W. Richards under whose supervision this work was continued in North Wales. Watsonia 6 (6), 1968. 388 A. D. Q. AGNEW APPENDIX Locations and habitats of populations of J. effusus and J. conglomeratus cited in text figures. Population Location and Habitat Fig. 1 All Marshy grassland at edge of Threipmuir reservoir, Balerno, Midlothian, with Deschampsia caespitosa and Molinia caerulea. Fig. 2 All Flush area on hillside, with Agrostis spp. at Penicuik, Midlothian. Fig. 3 All Flush area and adjacent hillside, with consequently varying associated species, at Llyn Idwal, Caernarvonshire. Fig. 4 All Grassland (rough pasture) with Nardus stricta and Des- champsia flexuosa adjacent to streamsides with Agrostis tenuis and Holcus lanatus ? km west of West Linton, Midlothian. Fig. 5 All As for Fig. 4. Fig. 6 1 Threipmuir reservoir, Balerno, Midlothian, as in Fig. 1. 2 Penicuik, Midlothian, as in Fig. 2. 3 Wet open woodland with Holcus lanatus at Penicuik, Midlothian. 4 In Salix carr bordering Threipmuir reservoir, Balerno, Midlothian. Fig. 7 All These were spot samples and no record remains of their detailed provenance. Fig. 8 Lowest Agrostis tenuis grassland at Bangor, Caernarvonshire. next Agrostis tenuis grassland 1 km south-west of Bangor. next Coarse wet grassland with Deschampsia caespitosa at Llandegfan, Anglesey. highest As for Fig. 3 at Llyn Idwal, Caernarvonshire. REFERENCES AGneEw, A. D. Q. (1961). The ecology of Juncus effusus L. in North Wales. J. Ecol., 49, 83-102. ANDERSON, E. (1948). Hybridization of the habitat. Evolution, 2, 1-9. ANDERSON, E. (1953). Introgressive hybridization. Biol. Rev., 28, 280-307. BUCHENAU, FR. G. P. (1906). Juncaceae in Das Pflanzenreich, 25. Leipzig. KrisA, B. (1962). Relations of the ecologico-phenological observations to the taxonomy of the species Juncus effusus L. s.l. Preslia, 34, 114-126. LINNAEuS, C. (1753). Species Plantarum. Uppsala. Meyer, E. (1819). Junci generis monographiae specimen. Gottingen. RICHARDS, P. W. & CLAPHAM, A. R. (1941). Biological Flora of the British Isles. Juncus effusus L.; Juncus conglomeratus L. J. Ecol., 29, 375, 381. SAVAGE, S. (1945). A catalogue of the Linnaean Herbarium. London. TWEED, R. D. & WOODHEAD, N. (1946). A consideration of Juncus effusus L. and Juncus conglomeratus L. J. Ecol., 33, 210-213. TWEED, R. D. & WooDHEAD, N. (1947). The taxonomy of Juncus effusus L. and Juncus conglomeratus L. North Western Naturalist, 21, 216-222. Watsonia 6 (6), 1968. NOMENCLATURAL NOTES ON THE GENUS ANTHYLLIS L. By J. CULLEN University of Liverpool Botanic Gardens The following new combinations, etc., have been made necessary by the preparation of an account of the genus for Flora Europaea. ANTHYLLIS MONTANA L. subsp. hispanica (Degen & Hervier) Cullen, comb. nov. A. montana var. hispanica Degen & Hervier, Bull. Acad. Int. Géogr. Bot. 15: 68 (1905). ANTHYLLIS VULNERARIA L. subsp. bulgarica (Sagorski) Cullen, comb. nov. A. dillenii Schult. ‘Unterrasse’ A. spruneri Boiss. var. bulgarica Sagorski, Allg. Bot. Zeitschr. 14: 173 (1908). subsp. argyrophylla (Rothm.) Cullen, comb. nov. A. argyrophylla Rothm., Feddes Repert. 50: 190 (1941). A. webbiana auct. mult. non Hooker, Bot. Mag. t.3284 (1833). subsp. reuteri Cullen, nom. nov. A. hispida Boiss. & Reuter, Pug. Pl. Nov. 36 (1852) non A. vulneraria var. hispida Boiss., Pl Or: 27158 (1872). subsp. forondae (Sennen) Cullen, comb. nov. A. forondae Sennen, Sched, ad Pl. d’Esp. 5695 (1926). subsp. pindicola Cullen, subsp. nov. A subsp. maura (G. Beck) Lindb. foliis solum in parte inferiore caulis insertis, foliis inferioribus equifoliatis, foliolis (7—) 9-13 (—15), ellipticis vel orbiculatis, calyce concolore, sparse hirsuta differt. Holotype: Greece, Epiros, inter Vriskos et Joannina, 8. vii. 1958, Rechinger 20721 (W—Herb. Rechinger). subsp. weldeniana (Reichenb.) Cullen, comb. nov. A. weldeniana Reichenb., Fl. Germ. Excurs. 515 (1832). subsp. hispidissima (Sagorski) Cullen, comb. nov. A. hispidissma Sagorski, Allg. Bot. Zeitschr. 14: 174 (1908). 389 Watsonia 6 (6), 1968. BOOK REVIEWS Flora of Iraq, Volume 1. Evan Guest (editor). Pp. xvi +207, with 9 plates (2 in colour), 36 figures (including 2 coloured maps) and 6 tables in the text. Ministry of Agriculture, Baghdad, Iraq, 1966. Obtain- able from R. Maclehose & Co., Glasgow. Price £1 10s. Turkey and Iran border Iraq, and form with it one of the most fascinating botanical areas in the world, particularly with regard to plant geography and plant domestication. The floras of all three countries are at present being covered by exhaustive treatments, Iran’s by K. H. Rechinger, Turkey’s by P. H. Davis and Iraq’s in the work being reviewed here. The area has not been covered since Boissier’s Flora Orientalis (1867-1888) and so this is a timely renewal of interest in it. The flora of Iraq has many of the easterly extensions of Mediterranean weeds, as well as many British and European casuals; this Flora will therefore be of value to British botanists. Volume 1, however, requires special consideration if its context is to be understood. It is a general treatise on vegetation and environment in Iraq, and includes diverse matter such as a gazetteer which will be of real value only to those attempting to elucidate the labels and field notes of herbarium sheets, and a 49-page illustrated glossary aimed at Arabic-speaking users of the Flora. There are only 55 pages on vegetation. The book is written as an adjunct to school and University texts within Iraq as well as a scientific work. Thus its scope is wide, and criticism comes easily to the specialist in any field, but this should not detract from the achievement that it represents. No one who has not worked in Iraq can envisage the delays and red tape attached to any project there. The gaps between impulse, action and completion of any work are wide, and Mr. Guest is to be looked on with awe as someone who has surmounted innumerable obstacles. He is an Arabist whose contribution to the proper understanding of local plant names will shine throughout all the volumes of this work. But it is due to the special conditions in Iraq that this volume is dated 1966 while little new material has been incorporated since 1961. My criticisms are confined to the treatment of vegetation except for two items: the useful rainfall records of the Hydrological Survey of Iraq (1959) are not mentioned in the section on the climate of Iraq, and in the bibliography (which incorporates material to 1964) it would have been better to use the World List for abbreviations of periodicals—‘Journ. Ecolog.’ is clumsy. Except for the list of plant ‘associations’, apparently no new material has been incorporated into the section on the vegetation of the country. This is a pity, for Mr. Guest must have made many original observations which, backed by research at Kew where this volume was written, could have added to the value of the section. Thus in the treatment of the Irano-Turanian region he quotes Drude (1890) for endemism rates, as cited by Eig (1931). Surely more up-to-date estimates would have been possible? This comment also applies to the uncritical acceptance of Zohary’s (1950) figures for the proportions of Iraq species in each biogeographic region. The special nature of Saharo-Sindian vegetation is dismissed in a few words, without considering Murbeck’s (1920) old but thought-provoking work; and surely some groups of plants do show special development in this area, such as Neurada procumbens (in its own subfamily of the Rosaceae), Gymnarrhena micrantha, and Launaea spp. The list of plant communities in Iraq is given in terms of ‘associations’ which are assigned binomials without published relevés and without any mention of the availability of the data for future phytosociolo- gists. The desert area is covered as far as the major vegetation types are concerned (though I would not imagine Pulicarietum guestii to be anything but a local facies of Artemisietum herbae-albae, not necessarily degraded), but the other areas of Iraq are very cursorily covered. The quality of the printing is good, and the work as a whole is relatively free from errors, but the standard of photographic reproduction is poor in the examples that I have seen. Flora of Iraq, Volume 2. C. C. Townsend and Evan Guest (editors). Pp. viii+ 184, with 32 plates, 1 coloured frontispiece and 3 text figures, 1966. Price £1 10s. This Flora will, I suspect, remain as one of the curiosities of botanical literature since it is arranged on the basis of Hutchinson’s 1959 treatment of the Families of Flowering Plants, a fact for which neither Mr. Townsend nor Mr. Guest can be held responsible. In this volume only Equisetum, ferns, gymnosperms and Rosaceae are covered. There is a key to the families of Iraqi Angiosperms which is stated to have been modified from Hutchinson’s key. A little more modification with the Iraq flora in mind would have made the key much shorter. Most families are keyed out in the present work just as many times as in Hutchinson’s original, although the Iraq flora is poor in genera of some families. For instance the Sterculiaceae is keyed out eight times but, as far as I know, apart from ornamental cultivands only one species (Glossostemon bruguieri DC.) grows in Iraq. 390 Watsonia 6 (6), 1968. BOOK REVIEWS 391 Throughout the work little reference has been made to collections in the Baghdad University Herbarium. I appreciate that in the past it has been difficult to procure specimens on loan from the University autho rities, but I hope that renewed efforts will be made to do so in the future. In the present volume Agri monia eupatoria L. is not recorded for the Baghdad district although it grows there near the site of the old R ustam Herbarium and is represented by specimens in BUH. R. D. Meikle has given a fine treatment of the Rosaceae, and I enjoy the broad view of the genus Prunus taken here, sinking Amygdalus and Cerasus. J.B. Gillett has revised the late R. A. Blakelock’s account of Ephedra, which has left the genus in an easily understandable order. The plates are excellent and show the high standard of draughtsmanship that we associate with Kew productions. REFERENCES Eric, A. (1931). Les éléments et les groupes phytogéographiques auxiliaires dans la flore palestinienne. Feddes. Rep. Beih., 63, 1-201. HYDROLOGICAL SURVEY OF IRAQ. (1959). Summary of Monthly Precipitation at stations in Iraq. Republic of Iraq Development Board. MursBeEckK, Sv. (1920). Beitrage von Biologie der Wiisten Pflanzen. II. Die Synaptospermie. Lunds Univ. Arsskrift. N.F., 15, 3-53. A. D. Q. AGNEW Botanical Latin. W. T. Stearn. Pp. xiv+566 and 42 illustrations. Nelson & Sons, Edinburgh. 1966. Price £5 5s. Od. Systematic botany is one of the very few surviving strongholds of the Latin language as a /Jingua franca. This may seem to many people a curiosity, worth preserving, perhaps, for its historical associations, but of no significance in the modern world of science-based technology. Yet to any systematist who has experienced the very real relief of finding, for example, a new species in the Bulgarian Flora described in Latin in a special appendix—or even to the forester who uses Latin binomials with confidence as an international nomenclature—botanical Latin is very much more than a ‘freak’. It was, therefore, all the more surprising that no comprehensive book in English existed which attempted to teach this language before William Stearn’s masterly opus. Reviewers often say, with varying degrees of conviction, that a book ‘fills a long-felt want’. The phrase can be sincerely used to commend this remarkable work. ‘Botanical Latin’, in the words of the author (in a preface entitled ‘Apologia pro Libro meo’), ‘aims to provide a working guide to the special kind of Latin internationally used by botanists for the description and naming of plants.’ It is a reasonable claim; after a fascinating historical and introductory Part I, the author devotes Part II to basic grammar, and some old friends appear (‘amo, amare’) as well as some less familiar, more technically useful ones (‘florere’ to flower). The idea is excellent, and the book is completely self-contained. Part III, somewhat uninformatively entitled “Syntax and Other Matters’, provides the systematic botanist with everything he needs to write his own Latin diagnosis or description. Part IV is an extraordinarily complete vocabulary working both ways. Characteristically the author says ‘it does not claim to list every word which has ever been used in descriptive botany’, but few will criticise it on the score of inadequacy. An excellent work, complete and attractively produced, ideal for the student of botanical taxonomy in every respect except one—the price. Perhaps nothing can be done about this problem; but 5 guineas is too much to expect the student to pay for a book which, to be honest, is only one of many which he is going to have drawn to his attention during a three-year University course. How many will persuade a rich parent or relative that this is the ideal Christmas present? It is sad to think that, to very many students forgwhom this book could be really useful, it will remain yet another work to be consulted hurriedly and inadequately in a Departmental library. Perhaps a shortened version, incorporating the excellent Part II, might be produced at a more modest price? I sincerely hope so. S. M. WALTERS A Dictionary of the Flowering Plants and Ferns. J. C. Willis. Seventh Edition, revised by H. K. Airy Shaw. Pp. xxii-+1214+liii. Cambridge University Press. 1966. Price £5. It is good to be able to welcome a new edition of this familiar work which, in one or other of its previous manifestations, has occupied a prominent position on the desks of many botanists. For, indeed, it provides in a handy concise form much information that is otherwise only to be found widely scattered through the literature. Willis’s Dictionary first appeared in 1897 in two volumes, of which the first contained articles on Watsonia 6 (6), 1968 5 392 BOOK REVIEWS various botanical topics such as Morphology, Classification and Economic Botany, while the second was the dictionary proper. The second (1904) and third (1908) editions were in one volume with, in the latter, an additional ‘glossarial index’. In 1919, Willis completely revised the work, omitting the general topics of Part 1. In two further editions (1925, 1931), he rewrote certain pages and added a supplement. The sixth edition was reprinted four times between 1948 and 1960, but no additional material appears to have been added since 1931. In producing this new edition, Mr. Shaw has adhered mostly to the previous format; but, in order to provide adequate discussion of certain topics and still keep the size of the book within a reasonable and handy compass, he has had to omit certain information that appeared in previous editions. Thus all headings concerned with popular and economic names have been omitted, as well as the articles on morpho- logical topics (e.g. perianth, stipule), so that the alphabetical entries are largely confined to names of genera and families. The reviser’s aim has been “to include every published generic name (whether validly published or not) from 1753 onward, and every published family name from the appearance of the Genera Plantarum of Jussieu in 1789’. The only other entries other than genera or families are some supra- and infra-familial taxa and the uninomials of Ehrhardt and Du Petit Thouars. In addition, the family entries have been some- what altered. Instead of Engler & Prantl and Bentham & Hooker equivalents of each family, the author of the family name is given. This is a very valuable piece of information which could not have been included in previous editions as much of the relevant research had not been done; but it is a pity that space was not available for reference to the two most widely used systems of classification. Another feature of previous editions that is here conspicuous by its absence is the reference, under each genus of the larger families, to the position of the genus according to the synopsis of ‘Classification and Chief Genera’ given at the end of the account of the family. At first glance the number of such synopses appears to have been reduced; but this is due to the elevation of several subfamilial taxa to family rank (e.g. in Gentianaceae and Onagraceae). Many of the synopses which remain, however, have been elaborated and revised according to modern treatments (e.g. Ranunculaceae, Ericaceae). One of the notable features of this edition is the introduction of a completely revised family concept. At a time when workers continue to produce evidence that many of the well-known families are to some extent heterogeneous and that the classification of Angiosperm families into orders has not yet been success- fully achieved, it appears wise to avoid strict adherence to any of the published family treatments. Instead, Mr. Shaw has attempted to ‘purify’ many families by the exclusion of the heterogeneous elements, while being apparently in no hurry to re-allocate these elements to other existing families. This policy has resulted in the recognition of many unfamiliar families, often monogeneric e.g. Sladeniaceae (S/adenia; ex Theaceae), Uapacaceae (Uapaca; ex Euphorbiaceae), Kingdoniaceae (Kingdonia; ex Ranunculaceae), but sometimes quite large (e.g. Alliaceae). These families, including as they do many of the anomalous genera of existing families, usually show features indicative of relationships in several directions. It is therefore very helpful to have, at the end of each family entry, Mr. Shaw’s ideas regarding the affinities of that family. Even though not everyone will agree with his suggestions—and some of them are most unexpected—it would be difficult, if not impossible, to find another modern work of comparable scope in which the relationships of families and genera have been considered on such a wide scale. The treatment of genera is necessarily dependent to an even larger extent on the availability of recent revisions and the author’s own opinions. It is not surprising, therefore, to find some large genera sub- divided when, in the reviewer’s opinion at least, it would be preferable to take a broad view (e.g. Hippo- cratea L.). Irrespective of these considerations of rank and taxonomic recognition, it is good to note the addition of such a large number of genera omitted from previous editions and the inclusion, under both generic and familial heads, of many references to recently published work. The value of the volume as a research tool is considerably enhanced thereby. This new edition of ‘Willis’ has obviously entailed much careful and detailed work, but two points of criticism come to mind. Firstly, the synonymy signs are confined to= (i.e. ‘a synonym of .. .’) and ~ (i.e. ‘sometimes included in ...’). It would have been helpful to distinguish in addition between nomenclatural and taxonomic synonyms, but the extra work involved would no doubt have been enormous. A more practical criticism concerns the summary of the classification of Bentham & Hooker at the end of the work. This, like the summary of Engler & Prantl’s system, has been reprinted unchanged from the previous edition. Willis, however, had added two small families to the numerical sequence adopted by Bentham & Hooker. Since these are both near the beginning (21. Vochysiaceae and 48. Cyrilleae), most of the numbers of the families differ by one or two from those in the original system as used in Kew, the British Museum and other herbaria. This difference has already proved confusing to herbarium workers and it is a pity that the oppor- tunity to revert to the original enumeration was not taken. Although now somewhat larger than before, ‘Willis’ is still convenient for use as a desk-top reference Watsonia 6 (6), 1968. ee BOOK REVIEWS 393 volume. The differences in type-size and page-headings, small though they are, will also make it easier to use. The only inconvenient feature in this connection is the price which, at £5, renders ‘Willis’ less easily available than before; but the additional information in this edition makes it well worth the increased cost. N. K. B. RosBson Primary Productivity in Aquatic Environments. C. R. Goldman (editor). Pp. 1-464. University of California Press. 1966. Price £3 10s. This volume records the proceedings of the first symposium to be held in the International Biological Programme. A total of 27 papers were presented in six sections dealing with various aspects of primary productivity in both freshwaters and the sea. As with all symposium volumes there is considerable variation in the standard of individual papers. Some are useful surveys of an area of investigation, and others are mere abstracts of work published or about to be published elsewhere. The editor admits that the papers did not lend themselves to rigid organi- zation, so that the book lacks the form that might be achieved by a single author. This is a feature of many symposia, and should not be allowed to detract from the general value of the volume, which covers many topics with greater expertise than could be expected from a single author. Some of the authors also acknow- ledge that their first drafts have been modified by the discussions which took place at the symposium. For the non-specialist in primary production this book provides an introduction to the people who are studying this topic in aquatic habitats and provides an idea of the directions in which their lines of research are travelling. Most of the papers deal with primary production by planktonic algae, but some of the papers deal with higher plants; of these the paper by Westlake on the productivity of aquatic macrophytes is the most likely to be of direct interest to readers of this journal. The volume is well produced, with few printer’s errors, and the general usefulness is increased by the addition of indices of water bodies, organisms and subjects. J. GREEN Plant Taxonomy. V. H. Heywood. Pp. 60. Studies in Biology, No. 5. Edward Arnold. 1967. Price 12s. 6d. hardback; 7s. 6d. paperback. This is a clear and closely reasoned exposition of the main aspects of modern taxonomy, and the author is to be congratulated in condensing so much information into some 60 pages of text and diagrams. It touches on most of the newest lines of advance in the subject, relating them where appropriate to the more traditional methods. The author starts with the need for the classification of organisms and the historical development of taxonomic thought from the classical to the experimental approach. He then sets out to examine the raw materials of taxonomy, populations. This is perhaps arguable: populations may form the raw material of evolutionary development, but individuals could be considered as the raw material of any classificatory system. The various breeding systems which influence the pattern of variability and hence of evolutionary change are next dealt with, followed by a section on patristic similarity, cladistics and homoplasy, which might appeal much less to the average student, but certainly ought to be understood by him. Then comes a chapter on phenotype, genotype and ecotypic differentiation which should, in the reviewer’s opinion, have come earlier in the book when the raw material was being discussed. Taxonomic characters are treated in chapter 6, but seem to include only morphological and anatomical ones, though it must have been the author’s intention to lead from these through to the biochemical ones, which are the subject of the following chapter. Cytological characters, however, are not mentioned until much later, in the last chapter. The chapter on biochemical systematics is easily the best in the book, and in the space of a few pages opens up new and exciting vistas to the reader. The concept of character weighting is discussed in the next chapter, followed by a good, but necessarily simplified exposition of numerical taxonomy in chapter 9. The final chapter, on chromosomes, taxonomy and evolution, contains a section on karyotypes and chromosome number which should surely have been placed in chapter 6. However, the evolutionary aspects of polyploidy and the significance of isolating mechanisms clearly find their correct place in this final chapter. One would like to have seen the section on isolating mechanisms as well as that on hybridization considerably expanded, and more mention made of Anderson’s introgressive hybridization concepts. No doubt, however, the author was strictly limited as to space, so that the extension of one section would have forced him to exclude or shorten another. This book presents modern taxonomy in a clear and exciting manner in a form likely to appeal to the: average modern student who is so often nurtured on a pure diet of molecular biology. The biochemical, numerical and evolutionary sections are particularly interesting and should stimulate the student to read more widely. Here again, the author gives a useful select list for further reading, referring, amongst other Watsonia 6 (6), 1968. 394 BOOK REVIEWS items, to the excellent work by Dr. P. H. Davis and himself, The Principles of Angiosperm Taxonomy. The provision of an index would greatly have enhanced the value of the book, but may have been omitted for reasons of space or because it would not have been consistent with editorial policy. The book is remarkably good value and includes a number of good diagrams and some excellent plates. It should be of interest to many sixth-form pupils and should be required reading for all first and/or second year University students reading Botany and Biology. J. G. HAWKES Vergleichende Chorologie der zentraleuropdischen Flora. Pteridophyten—Gymnospermen—Monokotylen— Dikotylen (Salicales—Fabales). H. Meusel, E. Jager and E. Weinert. Text band: Pp. 583 with 9 plates. Kartenband: Pp. 258 with 992 distribution maps. Format: 34 x 30 cm. VEB Gustav Fischer Verlag, Jena. 1965. Price: MDN 295,00 (approx. £30). This work is published in two parts and describes the world-wide distribution of selected plants from the Pteridophytes, Gymnosperms, Monocotyledons and Dicotyledons (Salicales to Fabales). Further parts are to be published later. This is not just a revised version of Professor Meusel’s Vergleichende Arealkunde (1943) but is completely rewritten and very much enlarged. Stress is placed on the comparative distributions of Central European species but attempts are made, where possible, to plot patristically related species that are not found in Central Europe; in some genera such as Colchicum, Glyceria, Larix, Oxytropis, Picea and Pseudostellaria this is well done but in many other genera no information is given on extra European species. The first part or ‘Textband’ is particularly valuable and, at the beginning, reviews the main systems for the classification of distribution patterns and discusses, in detail, the floristic elements that make up the Central European flora. The next chapter deals with the distribution patterns of whole families and selected sub-families, genera and sections. The amount of detail given for each group varies; some groups are well covered while others merely have short descriptions. It must be borne in mind that gathering information of this nature is very tedious and we must be thankful for any data on this subject. The next part of this volume of text contains a table (running to 179 pages) of the selected species with their distribution in abbreviated form, the floristic element to which they belong and selected references. The next 272 pages contain full descriptions of the dis- tribution of the mapped species and with each species is given a selected synonymy, some information on re- lated species and some relevant references. This volume ends with a bibliography which contains well over 1,000 references but, unfortunately, some of the references cited in the text are not to be found in the bibliog- raphy—it is to be hoped that this will be rectified in later volumes. The second part, which is bound as a separate book, contains 992 distribution maps. The maps are printed four to each page and they are clear and well reproduced. Most distributions are shown by shading but simple outlines and point distributions are also used. Where there is suitable information on the recent spread of species this is often given; good examples are: Camelina microcarpa Andrz. ex DC., Sisymbrium altissimum L. and S. loeselii Juslen. Well documented extinctions or contractions of range are indicated and good examples are: Aceras anthropophora (L.) Ait., Ophrys apifera Huds. and Batrachium (Ranunculus) hederaceum (L.) S. F. Gray. Obvious introductions, doubtful records and a few fossil distributions are also given. Most maps illustrate the distribution of two or more species, subspecies or varieties; map 71b, for example, shows the separate distributions of 15 taxa included within the Carex angarae Steud. group. Many maps illustrate the distribution pattern of genera and sections. The species densities within these genera and sections are usually given by printing the number of species to be found in selected areas. It must be admitted that this book does contain some mistakes, but mistakes in a work like this are inevitable where the authors have relied on published literature for nearly all of their records. The mistakes often arise from differences in taxonomic opinion and not just simple misidentifications of plant material. However, these mistakes should not detract from the book as a whole. Professor H. Meusel and his colleagues Diplombiologen E.-Jager and E. Weinert have collected and handled an immense amount of material which has been sorted and published in a readily accessible form. The price of this book is prohibitive and it is to be hoped that libraries, at least, can be persuaded to make the knowledge contained in this book ‘readily accessible’. C. D. K. Cook Floral Biology. Mary S. Percival. Pp. xv+243, with 67 text figures. Pergamon Press, Oxford, London, Edinburgh, New York, Paris and Frankfurt. 1965. Price 20s. Floral biology is an attractive subject, dealing as it does with the form, structure and mechanisms of the flower and with associated pollinating agents. It has been out of favour for much of this century, largely Watsonia 6 (6), 1968. BOOK REVIEWS 395 as a reaction against the ‘excesses of frequently teleological “‘adaptationists” ’ in the post-Darwinian period who saw function at every turn on the most slender evidence. It is thanks to the work of people such as Baker, Grant, Jaeger, Leppik, van der Pijl and Stebbins that floral biological studies are now on a sound and scientific footing again. Dr. Percival’s account is a brave attempt to present the subject and outline the modern techniques that can be employed, in the hope of attracting new workers into this field. There can be no doubting her enthusiasm, but her book can only be recommended with grave reservations. It is, in fact, riddled with errors, inaccuracies, inconsistencies and stylistic infelicities, and appears to have been badly edited. After a brief introduction which includes a potted history of floral biology, the first chapter is devoted to “Sex in flowers’. There is a useful section on terminology in which we are told that ‘Species, all individuals of which have the same kinds of flowers, and possessing hermaphrodite and female flowers are gynomonoe- cious’ —a baffling statement unless one knows beforehand what the term means. The frequently used term ‘polygamous’ is not given. Fig. 2 gives an extraordinary illustration of Begonia in which the label pointing to ‘anthers with line of dehiscence’ in fact leads to nothing comprehensible even under a x 10 lens. In figures 4 and 5 the spurious accuracy of the width of the lower lip of the corolla and the length of the tube (3-7 mm, 4 mm, 6:7 mm and 8:5 mm respectively) is misleading. On p. 7 the statement that gynodioecism in Glechoma hederacea ‘appears to be due to a virus-like particle in the cytoplasm of the female plant, for if the female is grafted on to the hermaphrodite plant, the latter becomes female’ surely demands some further explanation. Much more serious is the unpardonable confusion between pollination and fertilization in the definitions of autogamy, allogamy and hybridization. In Fig. 8, Helleborus corsicus, the stigmas are described as ‘proterogynous’ whereas on p. 8 the term is correctly applied (and differently, although equally correctly spelled) to the flower. The pictures of Fuchsia (Figs. 9(a) and (b)) and of Asclepias (Fig. 14) are far below the quality one expects in a published work, while Fig. 21 which is said to be ‘Momordica balsamina L. S. flower centre’ is a baffling black and white silhouette resembling an ink blot. Both the description and illustrations of the stigma of Grevillea are difficult to follow and appear in any case to be contradictory. And so one could go on. Whole pages are undisciplined and rambling such as the account of the features of anemophilous flowers on pp. 55 seq. when the author jumps from group to group with scant regard for subjects, antecedents, paragraphs and so on, not to mention factual errors, misleading generalizations and bizarre notions of taxonomy. Chapter 3 is concerned with the agencies of pollination. Bat pollination is included but there is no reference to or pictures of bat flowers or the extensive work of Baker and Harris on this topic. Myrmecophily receives no mention. It is interesting to learn that maize pollen is so large (700,000?) and that it falls almost vertically to the ground. Unfortunately, the dimensions, when one has worked out what they mean, are of little help without those of other pollen grains for comparison. The section on hydrophily is difficult to follow and with regard to flowering in the Podostemaceae (p. 61), self-contradictory. Fig. 33 is not of Zannichellia palustris as stated, which has quite a different structure from that figured. Chapters 4 and 5 deal with animal food in flowers and are full of interesting information—‘in Helleborus orientalis stamens dehisce all round the clock’—and Table 2 gives a useful survey of the rhythm of pollen presentation in a wide range of species though one might be forgiven for wondering how accurate it is. It would be easy to continue this sad commentary. A last quotation illustrates the kind of passage that induces a feeling of helplessness in this reviewer—‘members of the Lobelioid group of the Lobeliaceae lost their only pollinators in areas where the Hawaiian Honey-creepers, Drepanididae, became extinct. The birds had bills of the right proportions to probe the flower tubes. These plants, one is Cyanea hirtella Rock, are persisting and setting seed although the birds have gone and have successfully reverted to autogamy.’ And who can blame them? The number of omissions is remarkable—no mention of H. G. Baker’s outstanding series of papers and reviews of fioral biology and breeding systems, nor of E. J. H. Corner (not even a mention of figs); the fundamental review of reproductive modes in higher plants by Fryxell is not mentioned; van der Pijl’s review (1960-61) of flower evolution finds no place (although there is a reference to earlier of his works); Meeuse’s The Story of Pollination is omitted as is the much earlier volume by Clements and Long, which said much of it all before. Other important omissions are Verne Grant’s Natural History of the Phlox Family and The Origin of Adaptations, both highly relevant works. Other works are misquoted such as “Principles of Evolution of Genetic Systems by C. A. Darlington’, and ‘Dr. Hans Kugler’s Introduction to Floral Biology’ mentioned in the Introduction which is in fact a German work as one can deduce from the Bibliography, and ‘Dr. Stefan Vogel’s Floral Biological Types as a Factor in Classification (1954)’ which is not only not in English, but is not even a book as one is led to expect: it is a review (given in the References, not the Bibliography) published in Bot. Studien 1, Jena. The references (apparently papers cited in the text) and the Bibliography (the books cited) are in thorough need of revision. Both the Darwin references in the Bibliography have the wrong dates and the wrong titles Watsonia 6 (6), 1968. 396 BOOK REVIEWS (correct, however, in the Introduction). The references to Kerner and Knuth contain errors regarding the dates of publication and the number of volumes. There is an index to animals, one for plants and a very brief and incomplete subject index. The use of common and scientific names is quite haphazard—see pp. 158-59, paragraphs 3—4 for a concentrated example. There are numerous mis-spellings such as Robinia pseud-acacia, Lithospermum purpuro-caeruleum, and L. purpurpocaeruleum, Myosotis discolor Pers, secretary (secretory), D.C. (DC.), rachices, data is, fasicles, Plantage, Rhododendrom, Escholtzia and Eschscholzia, Rubus fruticosus age, Manilla, Helianthemum chamaeocistus, C. sinensis, v: Valencia, apomyxis, Coleus thryrsoideus, C. thrysoides and C. thrysoideus (none of them correct!), Cyania hitrella, Echium vulare, Heracleum spondylium and Ranunculus puberculus. Other errors noted included ‘Royal Horticultural Society’s Dictionary of Horti- culture’ and Larsen & Tung (1952) in text, which appears correctly as (1950) in references where the volume number of the Botanical Gazette is wrongly given as 3 instead of 111. The generally slipshod approach employed in writing the book leaves one apprehensive about its overall value. With so many obvious inaccuracies one simply does not know what pieces of information to trust. And yet there is clearly much of value in the book. I can only suggest to Dr. Percival, her Editors and Publishers that the volume be meticulously revised and republished. There are few enough books in this field which one can recommend for student use and it would be a service to botany if this one could be revised and added to the list. V. H. HEyYwoop Hortulus. Walahfrid Strabo. Translated by Raef Payne. Commentary by Wilfrid Blunt. Pp. xii+ 92. Number 2 of the Hunt Facsimile Series. The Hunt Botanical Library, Pittsburgh, Pa. 1966. Price $12. A book about Walahfrid Strabo with a commentary by Wilfrid Blunt, text and translation by Raef Payne, and published by the Hunt Botanical Library, must be excellent. This book, both in substance and form, is, in fact, superexcellent. Wilfrid Blunt’s account of Strabo—itself a piece of literature—shows clearly that the widely held opinions about the cross-eyed are not always true. Among other treasures we have here a facsimile of the best-known manuscript Hortulus in fine Caroline minuscules. Of those who contributed to the design of the volume, special mention should be made of Wilfrid Blunt, at whose suggestion the pretty green linoleum cuts were spread over the crystal-clear type of the Latin and English texts, which they by no means obscure. When these days are past and gone, this work will remain—an emblem of the civilization we now enjoy. H. GILBERT-CARTER Flora Slovenska. Edited by J. Futak. Vol. 1, Pp. 602, 93 x7 in. Bratislava, 1966. Price 40 Czech crowns. Vol. 2, Pp. 345, 94 x7 in. Bratislava, 1966. Price 28 Czech crowns. These are the first two volumes of a new Flora of Slovakia, which is being prepared under the general editorship of Doz. J. Futak, of the Slovak Academy of Sciences. The authors of Volume 1 are J. Dostal, J. Futak & F. A. Novak, and of Volume 2, J. Futak, M. Jasi¢ova & E. Schidlay. Volume 1 has an introduction in Slovak, Russian and German which sets out the history of the project, the methods used in preparing the flora, and the system on which it is arranged. The whole of the remainder of the volume is devoted to a survey of plant morphology, item by item; every term is defined (its Latin equivalent being given) and most of them are illustrated by large, clear diagrams. There is an excellent index to the Latin terms, with 1200 entries. Volume 2 is the first volume of the Flora proper, and it covers the Pteridophyta and Coniferophytina. The treatment of both groups is up to date. Families, genera and species are keyed and concisely described. Place of publication is given for the species and basionym and synonyms are cited. There are notes on ecology and chorology, and chromosome number (usually from the literature) is given; localities are often given in detail, herbarium specimens frequently being cited. There are good habit drawings (3 or 4 to apage) of every species; and dot distribution maps, on a reasonably large scale, of about half the species. There are also, at intervals in the text, short lists of literature references; this is an excellent feature. We can heartily recommend this new arrival to the list of European Floras; Dr. Futak and his colleagues are to be congratulated on a well produced and well organized piece of work. The Flora will be complete in 7 volumes. D. H. VALENTINE Watsonia 6 (6), 1968. ERRATA p. 86, lines 5-6 from bottom should read: 1. Leaves up to four times as long as broad Leaves more than four times as long as broad p. 87, lines 23-24 should read. 15. Leaves dentate Leaves finely denticulate or subentire p. 87, lines 29-30 should read: 18. Leaves over 20 mm. broad 10 Leaves under 20 mm. broad 397 Watsonia 6 (6), 1968. 7. dovrense 16 . attenuatifolium 19 HO. Belle ee ee . - # 1 ay { Sahing ‘ uu h ) or ¥ H ae) = i ‘. < * ey Vite hs « ' , iy j a C 4 meal i Fy i y ‘ } . t INDEX TO VOLUME 6 New taxa, names and combinations are given in heavy type Agnew, A. D. Q., The interspecific relationships of Juncus effusus and J. conglomeratus in Britain, B77 Alchemilla filicaulis, 76, Plate 2 filicaulis subsp. vestita, 78 minima, 76, Plate 2 Allen, D. E., A list of infraspecific taxa of British Phanerogams tested in cultivation, 205 Taxonomy and nomenclature of the radiate variants of Senecio vulgaris L., 280 Allieae, distribution of, 118 Anagallis arvensis subsp. arvensis, formae, 297 arvensis subsp. foemina, 297 Anthyllis montana subsp. hispanica, 389 vulneraria subsp. argyrophylla, 389 subsp. bulgarica, 389 subsp. corbieri, 295 subsp. forondae, 389 subsp. hispidissima, 389 subsp. pindicola, 389 subsp. reuteri, 389 subsp. weldeniana, 389 Apomixis in Euphrasia, 217 Arenaria serpyllifolia subsp. leptoclados, 294 Arenaria_ serpyllifolia subsp. macrocarpa, 294 The Arlington Practical Botany. Book I. Plant Anatomy by Mary-Anne Burns (Review by J. Timson), 324 Arthrocnemum, 25 Assessment of the taxonomic status of mixed oak (Quercus spp.) populations, by A. Carlisle and A. H. F. Brown, 120 Athyrium distentifolium var. flexile, 292 Barling, D. M., Leaf measurements and epi- dermis in Poa angustifolia, 109 Botanical Latin, by W. T. Stearn (Review by S. M. Walters), 391 Botany: A Laboratory Manual, by T. E. Weier, C. R. Stocking and J. M. Tucker, ed. 3 (Review by J. Timson), 324 Bowen, H. J. M., Sulphur and the distribution of British plants, 114 Bradshaw, Margaret E., Studies on Alchemilla filicaulis, Bus., sensu lato, and A. minima Walters. II. Alchemilla minima, 76 Breeding relationships of some European Euph- rasiae, by P. F. Yeo, 216 British Phanerogams, intraspecific taxa tested in cultivation, 205 Brees hordeaceus, 329, 330, Plates 14b, 14c, 15a, 15 Bromus hordeaceus subsp. ferronii, 330, 331, Plate 15c subsp. thominii, 329, 331, Plates 14a, 15b, 15c interruptus, Plate 15c lepidus, 330, Plates 14b, 14c, 15a, 15b mollis agg., 327 xX pseudothominii, 330, 331, Plate 15a thominii sensu Tutin, 330, Plates, 14b, 15a, 15b Bromus mollis aggregate in Britain, by Philip Smith, 327 Brown, A. H. F., see Carlisle, A. Calystegia sepium subsp. roseata, 298 Cannon, J. F. M., Infraspecific variation in Lathyrus nissolia L., 28 Watsonia 6 (6), 1968. 399 Carlisle, A. & Brown, A. H. F., The assessment of the taxonomic status of mixed oak (Quercus spp.) populations, 120 Catabrosa aquatica subsp. minor, 317 Catapodium rigidum subsp. majus, 317 Cerastium fontanum subsp. scoticum, 293 fontanum subsp. triviale, 293 Chromosome Botany and the Origins of Cultivated Plants, by C. D. Darlington (Review by D. A. Webb), 134 Chromosome morphology in Trifolium, 274 Chromosome number, Alchemilla filicaulis, 80 minima, 78 Bromus mollis agg., 340 Cochlearia aestuaria, 188, Pl. 12 alpina, 188, Pl. 12 anglica, 188 arctica, 188, Pl. 12 danica, 188 groenlandica, 188 micacea, 188 oblongifolia, 188 officinalis, 188 polonica, 188 pyrenaica, 188, Pl. 12 scotica, 188 tatrae, 188 Epilobium adenocaulon, 36 alsinifolium, 36 anagallidifolium, 36 angustifolium, 36 hirsutum, 36 lanceolatum, 36 montanum, 36 nerterioides, 36 obscurum, palustre, 36 parviflorum, 36 roseum, 36 tetragonum, 36 subsp. lamyi, 36 Euphrasia brevipila, 216 curta, 216 davidssonii, 216 rotundifolia, 216 salisburgensis var. hibernica, 216 stricta, 216 Myosotis brevifolia, 276 ,278 stolonifera, 276, 278 secunda, 278 scorpioides, 279 caespitosa, 279 Ornithogalum umbellatum, 345 Ranunculus hederaceus, 248 omiophyllus, 248 Sisyrinchium bermudianum, 286 montanum var. crebrum, 287 mucronatum, 287 Spartina x townsendii sensu lato, 290 Trifolium occidentale, 271, 273 repens, 273 thalii, 273 Viola arvensis, 40, 41 curtisii, 40, 41 lutea, 40, 41 nana, 40, 41 tricolor, 40, 41 Chromosome numbers of Epilobium in Britain, by Peter H. Raven & D. M. Moore, 36 400 INDEX TO VOLUME 6 Chromosome numbers of Ornithogalum umbel- latum L. from three localities in England, by R. Czapik, 345 Cleistogamy in Spartina, by J. C. E. Hubbard, 290 Cochlearia, diploid species, 188 Computer mapping of species distribution in a county flora, by J. G. Hawkes, B. L. Kershaw & R. C. Readett, 350 Cook, C. D. K., Studies in Ranunculus subgenus Batrachium (DC.) A. Gray. III. Ranunculus hederaceus L. and R. omiophyllus Ten., 246 Coombe, D. E. & Morisset, P., Further obser- vations on Trifolium occidentale, 271 Cousens, J. E., The status of the Pedunculate and Sessile Oaks in Britain, 161 Cross-fertility in Euphrasia, 218 Cruciferae, distribution of, 118 Cullen, J., Nomenclatural notes on the genus Anthyllis L., 389 Czapik, R., Chromosome numbers of Orni- thogalum umbellatum L. from three localities in England, 345 x Dactylitella, 132 x Dactylocamptis, 132 xX Dactyloglossum, 132 Dactylorhiza Nevski, the correct generic name of the dactylorchids, by P. F. Hunt & V. S. Summerhayes, 128 Dactylorhiza aristata, 131 cataonica, 131 cilicica, 130 cordigera, 131 cruenta, 130 elata, 131 foliosa, 131 fuchsii, 132 subsp. hebridensis, 132 subsp. okellyi, 132 graggeriana, 130 hataginea, 130 iberica, 129 incarnata subsp. cruenta, 317 incarnata subsp. gemmana, 317 incarnata, 130, 317 subsp. coccinea, 130 subsp. ochroleuca, 130 subsp. pulchella, 130 kerryensis, 131 kotschyi, 130 lapponica, 131 maculata, 132 subsp. ericetorum, 132 subsp. rhoumensis, 132 majalis, 130, 260 majalis subsp. cambrensis, 261 osmanica, 130 persica, 130 praetermissa, 131, 260 pseudocordigera, 130 purpurella, 131, 260 romana, 129 saccifera, 132 salina, 130 sambucina, 129 sanasunitensis, 130 traunsteineri, 131, 260 umbrosa, 130 Dictionary of the Flowering Plants and Ferns, by J. C. Willis, 7th Edition revised by H. K. Airy Shaw (Review by N. K. B. Robson), 391 Pe in the genus Cochlearia, by J. J. B. Gill, Watsonia 6 (6), 1968. Distinguishing characters and geographical dis- tribution of Ulex minor and Ulex gallii, by M. C. F. Proctor, 177 Distribution of British plants in relation to sulphur, 114 Distribution of Ranunculus hederaceus and R. omiophyllus, 246 Distribution of species, computer mapping of, 350 Distribution of Ulex minor and U. gallii, 177 Drawings of British Plants. Part XVIII, Comp- ositae (4), by Stella Ross-Craig (Review by P. D. Sell), 83 Ecology of Euphrasia species, 243 Epilobium, 36 Erica praegeri, 296 stuartii, 296 Errata, 397 Euphrasia, 1 Euphrasia, breeding relationships of some Euro- pean species, 216 Euphrasia, ecology of British species, 243 Euphrasia anglica, 216 arctica, 301 borealis, 216 brevipila, 216 confusa, 216 curta, 216 davidssonii, 216 frigida, 301 micrantha, 301 nemorosa, 301 pseudokerneri, 216 rotundifolia, 216 salisburgensis, 217 salisburgensis var. hibernica, 216 tetraquetra, 301 Ferguson, I. K., Notes on the stigma morpho- logy and flowering behaviour in British Sali- corniae, 25 Flora of the British Isles, by A. R. Clapham, T. G. Tutin and E. F. Warburg. J/lustrations, Part II, Boraginaceae—Compositae. Drawings by Sybil J. Roles (Review by M. C. F. Proctor), 82 Flora Europaea, Volume 1 Lycopodiaceae to Platanaceae, edited by T. G. Tutin, V. H. Heywood et al. (Review by S. Segal), 319 Flora of Iraq, Volume 1, edited by Evan Guest (Review by A. D. Q. Agnew), 390 Flora of Iraq, Volume 2, edited by C. C. Townsend & E. Guest (Review by A. D. Q. Agnew), 390 Fléra Slovenka, edited by J. Futak, Vols. 1-2 (Review by D. H. Valentine), 396 Floral Biology, by Mary S. Percival (Review by V. H. Heywood), 394 Floral biology in Euphrasia, 238 Floral biology of Ranunculus hederaceus and R. omiophyllus, 248 The Flowering Process, by Frank B. Salisbury (Review by B. Frankland), 137 Flowers of the Mediterranean, by Oleg Polunin and Anthony Huxley (Review by J. E. Lousley), 324 Fruit variation in Polygonum persicaria L., by J. Timson, 106 Futher observations on Trifolium occidentale, by D. E. Coombe and P. Morisset, 271 Genista tinctoria subsp. littoralis, 295 Gill, J. J. B., Diploids in the genus Cochlearia, 188 Growth of Evphrasia in cultivation, by P. F. Yeo, 1 INDEX TO VOLUME 6 401 Hawkes, J. G., Kershaw, B. L., & Readett, R. C., Computer mapping of species distribution in a county flora, 351 Hieracium section Alpestria, 85 Hieracium atraticeps, 305 attenuatifolium, 95, Plate 7b australius, 92, Plate 5b breve, 89, Plate 3b carpathicum, 103, Plate 11b chrysolorum, 307 cuspidens, 306 dewarii, 102, Plate lla difficile, 91, Plate 5a dilectum, 97, Plate 8b discophyllum, 309 diversidens, 305 dovrense, 93, Plate 6a glanduliceps, 312 glandulidens, 306 gratum, 90, Plate 4b hastiforme, 306 hethlandiae, 94, Plate 7a maculoides, 311 magniceps, 308 marginatum, 304 memorabile, 304 mirandum, 101, Plate 10b mucronellum, 310 nigrisquamum, 307 northroense, 100, Plate 9b notabile, 304 ornatilorum, 312 pauculidens, 310 piligerum, 310 praetermissum, 309 praethulense, 94, Plate 6b pruinale, 309 pseudanglicoides, 308 pugsleyi, 96, Plate 8a solum, 89, Plate 4a sparsifrons, 312 subtruncatum, 98, Plate 9a tenuifrons, 303 umbellatum subsp. bichlorophyllum, 313 variifolium, 311 vinicaule, 100, Plate 10a zetlandicum, 87, Plate 3a Hfortulus, by Walahfrid Strabo, translated by. Raef Payne, commentary by Wilfred Blunt (Review by H. Gilbert Carter), 396 Hubbard, J. C. E., Cleistogamy in Spartina, 290 Hunt, P. F., & Summerhayes, V. S., Dactylorhiza Nevski, the correct generic name “of the dacty- lorchids, 128 Huntia. A year book of botanical and horticultural bibliography, edited by G. H. M. Lawrence. Vol. I (Review by N. Douglas Simpson), 139 Hybridization, in Primula, 191 in Quercus, 125, 161 in Ranunculus hederaceus and R. omiophyllus, 252 in Ulex, 186 Hybrids, in Erica, 296 in Euphrasia, 218 in Mimulus, 70, 370 in Myosotis, 276 Hybrids of Mimulus cupreus, by R. H. Roberts, 371 Illustrations to the Flora of Delhi, by J. K. Mahesh- wari (Review by A. R. Smith), 325 Infraspecific variation in Lathryrus nissolia L., by J. F. M. Cannon, 28 Waisonia 6 (6), 1968. Ingram, Ruth, On the identity of the Irish popu- lations of Sisyrinchium, 283 Interspecific relationships of Juncus effusus and J. conglomeratus in Britain, by A. D. Q. Agnew, ST) Juncus effusus and J. conglomeratus, 377 Juncus, introgression in, 380 Kershaw, B. L., see Hawkes, J. G. Key to British species of MHieracium section Alpestria, 86 Lathryus nissolia, 28 Lathyrus nissolia var. pubescens, 32 Leaf measurements and epidermis in Poa angusti- folia L., by D. M. Barling, 109 List of infraspecific taxa of British Phanerogams tested in cultivation, by D. E. Allen, 205 Lythrum portula subsp. longidentata, 296 ee Orchids, coexistence of tetraploid species, 6 Mimulus cupreus, 70, 370 guttatus, 70, 370 guttatus X cupreus, 70, 371 guttatus X luteus, 70, 370 luteus, 70, 370 moschatus, 70 Mimulus hybrids in Britain, by R. H. Roberts, 70 Moore, D. M., see Raven, Peter H. Morisset, P., see Coombe, D. E. Morphology of Euphrasia hybrids, 228 Myosotis brevifolia, 276 caespitosa, 276 scorpioides, 276 scorpioides agg., 276 secunda, 276 sicula, 276 Natural hybridization between the cowslip (Primula veris L.) and the primrose (P. vulgaris Huds.) in Britain, by S. R. J. Woodell, 190 A New Illustrated British Flora, by Roger W. Butcher (Review by J. E. Lousley), 137 Nomenclatural notes on the genus Anthyllis L., by J. Cullen, 389 North Atlantic Biota and their History, by Askell Love and Doris Léve (Review by D. H. Kent), 82 Notes on Myosotis scorpioides agg., by D. Welch, 276 Notes on the stigma morphology and flowering behaviour in British Salicorniae, by I. K. Ferguson, 25 Odontites verna subsp. litoralis, 302 subsp. pumila, 301 subsp. serotina, 302 subsp. sicula, 303 subsp. verna, 302 On the identity of the Irish populations of Sisyrin- chium, by Ruth Graham, 283 x Orchidactyla, 133 Orchis comosa, 131 fistulosa, 131 latifolia, identity of, 130 palmata, 131 Ornithogalum umbellatum, chromosome numbers in, 345, Plate 16 402 INDEX TO VOLUME 6 Parietaria judaica, 365, 368 officinalis, 365, 368 Parietaria officinalis and P. judaica, by C. C. Townsend, 365 Pettet, A., Studies on British pansies. oO Chromosomenumbers and pollen assemblages, II. The status of some intermediates between Viola tricolor L. and V. arvensis Murr., 51 III. A factorial analysis of morphological variation, 141 Phenetic and Phylogenetic Classification, System- atics Association Publication No. 6, edited by V. H. Heywood and J. McNeill (Review by M. C. F. Proctor), 323 Pilosella aurantiaca subsp. brunneocrocea, 314 caespitosa, 314 caespitosa subsp. colliniformis, 314 flagellaris, 313 flagellaris subsp. bicapitata, 314 lactucella, 314 lactucella subsp. helveola, 314 officinarum subsp. concinnata, 313 subsp. nigrescens, 313 subsp. tricholepia, 313 subsp. trichoscapa, 313 peleterana subsp. tenuiscapa, 313 praealta subsp. arvorum, 314 subsp. spraguei, 314 Plant Taxonomy, by V. H. Heywood (Review by J. G. Hawkes), 393 Poa angustifolia, leaf measurements and epi- dermis, 109 subcaerulea, 109 Pollen fertility, in Euphrasia hybrids, 223 in Primula, 194 Pollen polymorphism, in Viola subgenus Melanium, 44 Polygonum persicaria, fruit variation, 106 Polyploidy and chromosome morphology in Ranunculus hederaceus and R. omiophyllus, 248 Polypodium vulgare subsp. prionodes, 292 vulgare subsp. serratulum, 292 Potamogeton alpinus X< lucens, 315 berchtoldii < natans, 314 crispus X praelongus, 316 Xx grovesii, 316 x nerviger, 315 pusillus x trichoides, 316 x undulatus, 316 x variifolius, 314 Primary Productivity in Aquatic Environments, ead by C. R. Goldman (Review by J. Green), Primula veris, 191 vulgaris, 191 Principles of Angiosperm Taxonomy, by P. H. Davis and V. H. Heywood (Review by M. C. F. Proctor), 320 Principles of Numerical Taxonomy, by R. R. Sokal and P. H. A. Sneath (Review by M. C. F. Proctor), 321 Proctor, M. C. F., The Distinguishing characters and geographical distributions of Ulex minor and U. gallii, 177 Quercus, introgression in, 161 Quercus, taxonomic status of mixed populations, 120 Quercus, taxonomic status of pedunculate and sessile oaks, 161 Quercus cerris, 169 conferta, 171 pedunculiflora, 171 Watsonia 6 (6), 1968. Quercus petraea, 120, 161 robur, 120, 161 Ranunculus subgenus Batrachium, 246 Ranunculus hederaceus, 246, 257 omiophyllus, 246, 258 Raven, Peter H. & Moore, D. M. Chromosome numbers of Epilobium in Britain, 36 Readett, R. C., see Hawkes, J. G. Revision of the British species of MHieracium Section Alpestria [Fries] F. N. Williams, by P. D. Sell and C. West, 85 Rhinanthus, nomenclature of, 298 Rhinanthus minor, 298 minor subsp. borealis, 300 subsp. calcareus, 299 subsp. lintonii, 300 subsp. monticola, 299 subsp. stenophyllus, 299 serotinus, subspecies, 298 x Rhizanthera, 133 Roberts, R. H., Mimulus hybrids in Britain, 70 Studies on Welsh Orchids. III. The coexistence of some of the tetraploid species of marsh orchids, 260 Roberts, R. H., The hybrids of Mimulus cupreus, 371 Rubus echinatus, 295 Salicornia, 25, Plate 1 Seed production and germination in Euphrasia hybrids, 226 Self-fertility in Euphrasia, 217 Sell, P. D. & West, C., A revision of the British species of Hieracium Section Alpestria [Fries] F. N. Williams, 85 Senecio vulgaris, taxonomy and nomenclature of radiate variants, 280 subsp. denticulatus, 303 var. denticulatus, 280 var. hibernicus, 280, 281 forma radiatus, 280, 282 Sisyrinchium, identity of Irish populations, 283 Sisyrinchium bermudianum, 284 montanum var. crebrum, 284 mucronatum, 284 Smith, Philip, The Bromus mollis aggregate in Britain, 327 Sorbus wilmottiana, 296 Spartina, cleistogamy in, 290 Spartina X townsendii, sensu lato, 290, Plate 13 Status of the Pedunculate and Sessile Oaks in Britain, by J. E. Cousens, 161 Studies on Alchemilla filicaulis Bus. sensu lato and A. minima Walters. III. Alchemilla minima, by Margaret E. Bradshaw, 76 Studies on British pansies, by A. Pettet. I. Chromosome numbers and pollen assem- blages, 39. II. The status of some intermediates between Viola tricolor L. and V. arvensis Murr., 51. Il. A factorial analysis of morphological variation, 141 Studies in Ranunculus subgenus Batrachium (DC.) A. Gray. IL. Ranunculus hederaceus L. and R. omiophyllus Ten., by C. D. K. Cook, 246 Studies on Welsh orchids. III. The coexistence of some of the tetraploid species of marsh orchids, by R. H. Roberts, 260 Sulphur and the distribution of British plants, by H. J. Bowen, 114 Summerhayes, V.S., see Hunt, P. F. INDEX TO VOLUME 6 403 Taxonomic and nomenclatural notes on the British flora, by various authors, 292 Taxonomy and nomenclature of the radiate parities of Senecio vulgaris L., by D. E. Allen, Timson, J., Fruit variation in Polygonum persi- caria L., 106 Townsend, C. C., Parietaria officinalis and P. judaica, 365 Trifolium occidentale, 271 Ulex minor and U. gallii, distinguishing characters and geographical distributions, 177 Valerianella locusta subsp. dunensis, 303 Variation, in British Pansies, 141 in Ranunculus hederaceus and R. omiophyllus, 251 Various authors, Taxonomic and nomenclatural notes on the British flora, 292 Vascular Plants of the Pacific Northwest. Part 2: Salicaceae to Saxifragaceae, by C. L. Hitchcock and A. Cronquist (Review by N. Y. Sandwith), 203 The Vegetation of Scotland, edited by J. H. Burnett (Review by J. J. Moore), 268 Vergleichende Chorologie der zentraleuropaischen Flora. Pteridophyten-Gymnospermen-Monokoty- len- Dikotylen (Salicales-Fabiales), by H. Meusel, E. Jager & E. Weinert (Review by C. D. K. Cook), 394 Watsonia 6 (6), 1968. Viola subgenus Melanium, 141 chromosome numbers and pollen assem- blages, 39 factorial analysis of morphological vari- ation, 141 arvensis, 39, 51, 141 carpatica, 142 contempta, 40, 52 curtisti, 40 kitaibeliana, 39, 142 lepida, 142 lutea, 42, 142 maritima, 45 nana, 39 tricolor, 39, 51, 141 tricolor < arvensis, 52 tricolor subsp. curtisii, 142 tricolor and V. arvensis, Intermediates between, 51 variata var. sulphurea, 40, 52 Welch, D., Notes on Myosotis scorpioides agg., 276 Welsh Orchids, 260 West, C., see Sell, P. D. Wild Flowers of Southern Africa—Natal, by Winifred Wright (Review by A. A. Bullock), 139 Woodell, S. R. J., Natural hybridization between the Cowslip (Primula veris L.) and the Primrose (P. vulgaris Huds.) in Britain, 190 Yeo, P. F., The growth of Euphrasia in culti- vation, 1 The breeding relationships of some European Euphrasiae, 216 . i dae Perk an a alte " vr i ti ahs Yaka VA) ee ie et aye Tht Le <7 w iy) er a PROCEEDINGS OF THE BOTANICAL SOCIETY OF THE BRITISH ISLES Published Twice Yearly Price 15/- per part Obtainable from D. H. KENT, 75 Adelaide Road, West Ealing, London, W.13. LIST OF BRITISH VASCULAR PLANTS prepared by J. E. DANpby for the British Museum (Natural History) and the Botanical Society of the British Isles. Many botanists have assisted in the preparation of this work, which incorporates the London Catalogue of British Plants and was undertaken by Mr. Dandy on behalf of a sub-committee of the Society appointed for this purpose. Sir George Taylor, Director of the Royal Botanic Gardens, Kew and a former President of the Society, writes in a foreword: ‘British botanists have been unusually fortunate in having Mr. Dandy’s unequalled knowledge of nomenclature and sure taxonomic insight placed so generously at their disposal. There has been a pressing need for a new British plant list and that deficiency has been most worthily met.’ The names given in the list are now used in the publications of the Society and have become adopted by British botanists generally. Synonyms relating to previous lists are given, and the list will prove invaluable to members as a work of reference. Genera and species are numbered, making the list useful for arrangement of herbaria and local lists of plants. Demy 8vo, 176+ xvi pages. Bound in cloth. Price : 10/- (postage 1|-). Please send cash with order Obtainable from E. B. Bangerter, c/o Department of Botany, British Museum (Natural History), Cromwell Road, London, S.W.7. ROYAL IRISH ACADEMY ROBERT LLOYD PRAEGER FUND FOR FIELD NATURAL HISTORY Grants, not normally exceeding £40 in any one year, will be awarded for field work relevant to the Natural History of Ireland. Application forms, which should be returned before February 15 of next year may be obtained from: The Secretary, Royal Irish Academy, 19 Dawson Street, Dublin, 2, who will also be glad to give further information. IRISH NATURALISTS’ JOURNAL ®@ A Magazine of Irish Natural History | Published Every Quarter by the I.N.J. Committee Edited by Miss M. P. H. KERTLAND, M.Sc. with the assistance of Sectional Editors Annual Subscription, 10/- post free Single Parts, 5/- All communications to be addressed to: The Editor, I.N.J., Science Library, David Keir Building, Queen’s University, Belfast, 9 INSTRUCTIONS TO CONTRIBUTORS PAPERS having a bearing on the systematics of British vascular plants are invited from both members of the Society and others. Careful attention to these notes before a paper is submitted can save the author and editor much work, and do much to speed publication. Papers must be submitted in final, fully corrected form. They should be typewritten, with wide margins, double-spaced, on one side of the paper only. If possible two copies should be sent, of which one must be a top copy. Authors should keep a fully corrected carbon copy of their typescripts for reference. Write clearly and concisely, rigorously excluding extraneous matter. It sometimes helps the clarity of an argument to place extensive experimental or tabular data in appendices, rather than in the main text. Technical terms are of value only as aids to clarity, precision and conciseness of expression. Take care that they are used consistently, and defined where necessary. Ill-defined or general terms often need definition in a particular context, and can often with advantage be replaced by something more precise, as by the ‘deme’ terminology of Gilmour and Gregor in genecology. Make certain that the main conclusions of the paper are clearly displayed to the reader. Avoid complicated hierarchies of headings, and check carefully the consistency of those that you use; a table of contents, setting out the full hierarchy of headings with the MS. page numbers, is often helpful even if it is not to be printed with the paper. Avoid footnotes as far as possible, and keep cross references by page number to a minimum. Tables, unless very small, should be typed on separate sheets and attached at the end of the typescript. They should be kept within a reasonable size, and as simple in structure as possible. Keys should be in one of the generally accepted forms; Flora of the British Isles by Clapham, Tutin and Warburg (ed. 2, Cambridge, 1962) and Flora Europaea (Vol. 1, Cambridge, 1964) provide suitable patterns. Names of genera and species should be underlined, but any other typographical indications should be inserted lightly in pencil. Names of British vascular plants should normally follow the List of British Vascular Plants by J. E. Dandy (British Museum (NH) and BSBI, London, 1958), and may then be cited without authorities. Otherwise, authors’ names must be cited, at least on the first occasion where they appear in the text. Authors must follow strictly the International Code of Botanical Nomenclature (Utrecht, 1966). Synonyms should be cited in chronological order; misapplied or illegitimate names should be placed in square brackets. Examples will be found in recent parts of Watsonia. Except for citations of ihe place of publication of plant names, which may be given in abbreviated form in the text, full references should be listed in alphabetical order of authors’ names at the end of the paper; the form used for references in a recent part of Watsonia should be followed carefully. Names of periodicals should be abbreviated as in the World List of Scientific Periodicals (ed. 4, London, 1963-1965). References to herbaria should follow the abbreviations given in British Herbaria (BSBI, 1958) and Index Herbariorum Part 1 (ed. 5, IAPT, Utrecht, 1964). Foreign language references should where necessary be transliterated into the roman alphabet according to a recognized convention; for Cyrillic script (Russian, etc.) the Royal Society (BS 2979: 1959) or UNESCO schemes are acceptable. Unless authors expressly indicate that they cannot get access to necessary literature the editor will assume that they have checked the correctness of all titles, abbreviations, transiiterations and references. Papers should begin with a short abstract, in the form of a piece of connected prose conveying briefly the content of the paper, and drawing attention to new information, new names and taxa, and the main conclusions. Line drawings should be boldly drawn in Indian ink on Bristol board, smooth surface white cartridge paper of good quality, or suitable draughtsman’s tracing materials, and should normally be suitable for reproduction at about one-half to two-thirds (linear) their original size. Very large originals should be avoided. Scale is best indicated by a rule on the drawing itself. If a magnification is quoted this should be checked carefully against the final block. Graphs can be reproduced from originals on graph paper with feint blue ruling, but all lines to appear on the finished block must be inserted in Indian ink. Lettering on line-drawings and graphs should be inserted lightly in pencil, and will be finished in uniform style. The following standard symbols are available for explanations of graphs, etc.: A, A, O, @, 0, +, x, ©. Avoid other symbols as far as possible. If an illustration includes plant-names or place-names these should be typed clearly on a separate sheet of paper. Photographs can be accepted only where essential. They must be of first-rate technical quality, of good but not excessive contrast, and of a size and character suitable for the necessary reduction. Remember that fine detail may be lost on even the best half-tone blocks. Legends to text-figures and plates should be typed on a separate sheet of paper and attached at the end of the typescript. If in doubt about the citation of names or references, or the presentation of illustrations or tabular matter, contributors are advised to consult the editors before submitting their typescripts. Further useful information and advice will be found in General Notes on the Preparation of Scientific Papers (ed. 2, Royal Society, London, 5s. 0d.). Proofs. Authors will normaliy receive a galley proof. Particular care should be given to the thorough checking of references and tables. It should be assumed that only actual errors of typography and fact can be corrected in proof; as far as possible any words or phrases deleted should be repiaced by others of equal length. Twenty-five offprints are given free to authors of papers. Further copies may be obtained at the Society’s current price, and must be ordered when the proofs are returned. The Society as a body takes no responsibility for views expressed by authors of papers. Papers should be sent to Dr. M. C. F. Proctor, Hatherly Biological Laboratories, Prince of Wales Road, I LT DEL IEE DLL TE TET EE Exeter. Books for review should be sent to Dr. N. K. B. Robson, Dept. of Botany, British Museum (Natural a EE ET a LS SI ELS EIEIO ELLA LD LE EE TO I EE I ED EA EDS History), Cromwell Road, London, S.W.7. A DARWIN CENTENARY Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the very successful Conference arranged by the Botanical Society of the British Isles in 1959. The contributors include Lady Barlow, G. D. H. Bell, Prof. T. W. Bocher, Prof. W. H. Camp, Prof. N. Hylander, Prof. I. H. Manton, Prof. T. G. Tutin and Prof. D. H. Valentine. Demy 8vo, 140 pages, with 7 half-tone plates and 11 maps and figures in the text. Bound in buckram. PRICE £1 (postage extra) LOCAL FLORAS Edited by P. J. WANSTALL The purpose of this volume is to make available to a wider public the papers read at the Conference held by the Botanical Society of the British Isles in 1961. The contributors include D. E. Allen, Dr. H. J. M. Bowen, Dr. J. G. Dony, J. S. L. Gilmour, Prof. J. G. Hawkes, D. H. Kent, J. E. Lousley, D. McClintock, R. D. Meikle, Dr. F. H. Perring, R. C. Readett and C. C. Townsend. Demy 8vo, 120 pages, with 2 half-tone plates, and figures in the text. Bound in buckram. PRICE £1 (postage extra) THE CONSERVATION OF THE BRITISH FLORA Edited by E. MILNE-REDHEAD The Report of an important and most successful Conference held by the Botanical Society of the British Isles at Durham in April 1963. This report includes most of the papers read at the Conference, together with a verbatim account of the discussions which followed. Demy 8vo, 80 pages. Bound in buckram. PRICE 15/- (postage extra) The above items are obtainable from D. H. Kent, 75 Adelaide Road, West Ealing, London, W.13. Printed by Headley Brothers Ltd 109 Kingsway London WC2 and Ashford Kent oT eee nS ee all apt ‘ Ego tet Pi tn Hahei oe ed a iP, ae a aoe He ete Y Vater iseti sci TF poset eae eas iF a N BS “4p ae Saas eet t/a aS, : wea ie Ma erin ei gt fis ibe Aol ¥ a “Sgn 3 Ms ee, en” 5% Ke msi | | ' fe) neon” AN Ra fy ~ WX, \ op “ty | | 5 Vey WO i i xe i | | i ve, 1p, Wt | oe 1 i H Hf ie; le Ws “1 | i | & Si mi Li es : _ 4% ; 25 | i be } ‘e a 4 A = - = : AN on “Dp Ss Te Mek Li Pps ae a Mul ale Mal! sa Tre, f. oa an Fae ee oe ry Seale ere eee ee tea - ™, ns, iv PY Oies 438 an] ah. Ape OE a m ae = i fax a ee gi i oa 2 BMOy at ass aa if eee my, err) AE, Jas Y dL ee es ae <* I o€ re *S} hee SSCA RE ei EEC STITUTION LIBRARIES NO ae $405 ‘ Pinte ¢ Wi: ; | : ise tae west } | | , i Aig fe, : SoS EG 1h es Letcira td | } Vea seed. he biter 4 «i : | : f 1 } Rats A Sieg . : | whe | Se eR | e sere ) : | | ‘ } \ STS OY 7 a ua SUS eee . +i BY a 2 | Y Wee | 7 i q $1 . P i Sg aL Y , ' : j is Ve 4 ' : ‘ t¢ q ‘ i 4 Ws Fea ! tse Baas | . | poate ale Mts i “tis \ Ay Ae La ; a ‘ Mii ae | t é ; : PES a he