АА RAA AA et FRE LA AREAS у Ам of amené à DI Pr < —-- р № =: ie 4 7 DL И ANA И ААА» | А RAN AAA Wan PT EEE ) Ana Annan Ramana hn AN AAA PAANARE DA A D ve ED r: . № ) 2 Dr > D ae 2 : ) Dy. i } » > я = ei > ars D», we pt > 32 DD 2:29 D >> 2 ) IE i D>» =D. BIS DIE 3335 № » >» py DD yy. >> ВР» рн II DIDI De >) YW 5 po Do XD DI DD » DD) yD 3% > »d 7 DI) 7 > DIDI D PM D “Yodo » 1» >>, VAN И И И fk A ØR rar a i U Bo il AA f NT des"! Meddelelser om Grønland. Meddelelser om Gronland. udgivne af Commissionen for Ledelsen af de geologiske og geographiske Undersagelser i Granland. Syv og tredivte Hefte. Kjebenhavn. C. A. Reitzel Boghandel. Bianeo Lunos Bogtrykkeri, 1921. The Structure and Biology of Arctic Flowering Plants. Il. Udgivet paa Carlsbergfondets Bekostning. Indhold. Side Nekrolog (H. P. Steensby) af C. F. Wandel...................... I—II bosser: By Knud Jessen. и cu... eee be 1 Doniaces.-By Carsten: Olsen ааа... 127 10. Caprifoliaceae. Linnaea boreartis L. Ву 0. Hagerup.......... 151 Be Primulaceae..By Fr. Je Mathiesen. .: .:...........:........... 165 12. A List of Arctic Caryophyllaceæ, with some synonyms. By С. Le DT I ae a ned ee 221 kam Caryophyilaceæ By Hug. Warming::.......:....:......,. 229 aies By Morten RB. Porsild.... .. ::..:.-:=:..12.: 1.1... 343 15. Scrophulariaceae. By Fr. J. Mathiesen ...................... 359 38344 i 4 ; ТИ EJ Bm Donau 7 + À ex р L 4 ? * 1 MR kr eleg ‚© ù af RME > så > & 4 у a a > mn x & | in » ” Avr sy ase af « ‚ Re à + ! à “ in ‘ x - . ‘ ` 4 ‘ . + » ‘ A H. P. Steensby. EN 19. Oktober ifjor meldte den traadlose Telegraf det D sørgelige Budskab fra Atlanterhavet, at Professor Steensby samme Dag var afgaaet ved Doden ombord i Ame- rikadamperen »Frederik den Ottende« paa Hjemrejsen fra Amerika. Professor, Dr. phil. H. P. Steensby var født 25. Marts 1875; han studerede Geografi og Naturhistorie, blev cand. mag. 1 Aaret 1900, Dr. phil. 1 1905 paa en Afhandling om Eskimo-Kulturens Oprindelse og i 1911 Professor i Geografi. Længe for det sidstnævnte Aar havde Steensby udfoldet en betydelig Forfattervirksomhed om ethnografiske, geogra- fiske og anthropologiske Emner og foretaget lange Rejser, men iser skal nevnes hans Arbejde for Udviklingen af det geografiske Laboratorium, der nu kan tjene som et Monster. for andre Lande. QV GILASS AST. m +} Te Im iv Å IT Endelig skal nævnes hans Interesse for den historiske Geografi; hans forste Arbejde paa dette Omraade var hans Afhandling om den græske Kobmand Pytheas, 1 hvilken han mente at paavise, at Pytheas var naaet til Ringkobing-Fjord; hans næste Arbejde paa dette Felt var Afhandlinger om de gamle Nordboers Vinlandsrejser, hvor han folger dem langs Labradors Kyst til selve »Vinland« ved Mundingen af St. Lawrence-Flod. For paa Stedet at konstatere sine Slutningers Rigtighed, foretog han i 1919 en Rejse til de paagældende Egne, og det var paa Hjemrejsen derfra, at Doden ind- hentede ham. I 1909 udsendte Kommissionen ham tilligemed Muse- umsassistent Thomsen til Kap York paa Grønlands Vestkyst for at studere den derboende Eskimostamme, og Resultatet al denne Rejse foreligger for hans Vedkommende i to inter- essante Afhandlinger i »Meddelelser om Gronland« Bind XXXIV og L. I 1913 indtraadte Steensby som Medlem i Kommissionen, hvor han desværre altsaa kun virkede i seks Aar; men paa Grund af hans elskværdige, rolige og saglige Deltagelse i Kommissionens Arbejde og hans levende Interesse for Stu- diet af Gronland bliver hans Bortgang et stort Tab for Kommissionen. C. F. Wandel. 8. Rosacee. By Knud Jessen. 1913. XXXVII In the present paper Г have given an account of the growth- forms and of the anatomy of the root, stem and leaf, and also of the flower-biology, of some Arctic and Subarctic Rosaceæ in a similar manner to my work on the Arctic Ranunculaceæ (vol. 36). — In addition to the alcohol-material in the Bot. Museum in Copen- hagen and the herbarium in the same place I have also made use of the literature in order, as far as possible, to collect in one place what is known about, these plants. I wish to convey to Professor WARMING my thanks for the readi- ness with which he has placed at my disposal his notes from his jour- neys in Arctic Norway and Greenland. Some of the figures are drawn by him and a good many of these have not previously been published. The rest of the figures, with the exception of figs. 8 and 26, have been drawn by myself. The following species have been investigated: — Potentilla palustris ............ р ©. — РЕНИ FRERE р. 15. Se ее 0... MES pe 22. = MIDER ae a NU: ВЕРЬ Daven. — Vahliana.'........... р. 38. — emarginata........... p. 36. — аи. Eee: p. 42. -— ANSE Ai halt nee p. 47. Sibbaldia procumbens .......... p. 54. Dryas octopetala and D. integrifoha р. 60. RUDS ВОТ EEE. NS p. 74. — chamæmorus........... p. 82. NO ее, Het een eee a р: "91: Alchimilla alpina and A.feröensis р. 99. — DUMAS 2 2 safe ue see p. 105. Sorbüs ‚americama с... ....:....., р. 108. 1* 4 Кмор JESSEN. Principal literature. ABROMEIT, 1899: Bot. Ergebnisse der ... unter Leitung von Dr. у. Dry- galski ausges. Grönlandexp. B. Samenpfl. (Phanerog.). Biblioth. Bot. 42. Stuttgart. ANDERSSON & HESsELMAN, 1900: Spetsb. och Bären Eilands kärl- växtflora. Bih. Kgl. Sv. Vet. Akad. Handl. Bd. 26. Afd. 11. Bonnier, 1894: Les plantes arctiques comparées aux mêmes espèces des Alpes etc. Rev. gén. d. bot. Tom. VI. Воснемло & Focke, 1872: Gefässpflanzen Ost-Grönlands. Zweite Deutsche Nordpolfart. Ш, 1. Botanik. CLEVE, Astrip, 1901: Zum Pflanzenleben in nordschwedischen Hoch- gebirgen. Bih. Kgl. Sv. Vet. Akad. Handl. Bd. 26. Afd. III. CovILLE, 1896: Bot. of Yakutat Bay, Alaska. Contrib. f. the U.S. National herbar. Vol. III. No. 6. | Dusen, 1901: Zur Kenntniss der Gefässpflanzen Ost-Grönlands. В. Kgl. Sv. Vet. Akad. Напа]. Bd. 27. А. III. Exstam, 1897: Einige blütenbiol. Beobacht. auf Novaja Zemlja. Tromsø Museums Aarshefter. У. 18. — 1899: Einige blütenbiol. Beobacht. auf Spitsbergen. Ibid. 20. FREIDENFELT, 1904: Der anat. Bau d. Wurzel in seinem Zusammen- hange mit dem Wassergehalts des Bodens. Bibl. bot. H. 61. Stuttgart. Fritsch, 1887: Anatom. system. Studien über 4. Gatt. Rubus. Sitzungs- рег. 4. К. К. Academie 4. Wiss. in Wien. Math.-naturw. Classe. XCV. HAGLUND, 1905: Ur de högnordiska vedvaxternas ekologi. Upsala, 1905. Hartz, 1894: Bot. Rejseber. fra Vest-Grenland. Medd. om Grønland. XV. — 1895. a: Ost-Gronlands Vegetationsforhold. Ibid. XVIII. — 1895, b: Fanerog. og Karkryptogamer fra N. 0. Grønland, etc. Ibid. XVIII. Hoım, Тн., 1885: Novaja Zemljas Vegetation .... Dijmphna Togtets 2001. bot. Udbytte. Kobenhavn, 1877. Ногтзтегх, 1907: Vergl. Anatomie 4. Stängel und Rhizom von dicotyl. Alpenpflanzen. Diss. Göttingen. Inmiscu, Тн., 1861: Über ... Comarum palustre. Bot. Zeit. Jonsson, Heıcı, 1895: Optegnelser fra Vaar- og Vinterexcursioner i Ost-Island. Bot. Tidsskr. Bd. 19. København. Kınıman, 1890: Pflanzenbiolog. Studien aus Russisch Lappland. Act. Soc. Faun. Flor. Fenn. VI. KJELLMAN, 1901: Om arten och omfat. af det uppbygg. arbet. Bot. Not. Lund. Кмотне, 1902: Vergl. Anatomie 4. unbenetzbaren Blätter. Diss. Berlin. Knurn, 1894: Blumen und Insekten auf d. Nordfriesischen Inselen. Rosaceæ. о Квосн, 1897: L’epiderme musilaginosa nelle foglia delle Dicotiledoni. Ann. del R. Istituto Botanico de Roma VI. Milano. LANGE, JoH., 1890: Conspectus flore Grünlandicæ. Meddel. om Gron- land. H.3. Kobenhavn. Leıst, 1889: Ueber den Einfluss des Alp. Standortes auf die Ausbild. d. Laubbl. Mittheil. d. Naturf. Gesellsch. v. Bern. LınpMman, 1884: Om postflorationen och dess betydelse etc. Kgl. Sv. Vet. Akad. Handl. Bd. 21. — 1887: Skandinav. fjällväxternas blomning och befruktning. Bih. Kgl. Sv. Vet. Akad. Handl. Bd. 12. Afd. III. LuUNDAGER, 1912: The Vegetation of Germania Land, N. Е. Greenland. Danmark-Eksped. til Grønlands Nordestkyst, 1906—1908. Bd. III. No.13. Medd. om Grønland XLIII. København. Маттев, HERMANN, 1881: Alpenblumen. Leipzig. Norman, 1895: Norges arkt. Flora. II. Kristiania. Porrıus, 1903: Blombiologiska Iakttagelser. Acta Soc. pro Fauna et Flora Fennica, 25. Vol. 1. Ravn, Е. Korpın, 1894: Om Flydeevnen hos Frøene af vore Vand- og Sumpplanter. Bot. Tidsskrift. 19. Kobenhavn. Res. francaise. Rosenvince, KoLDERUP, 1892: Andet Tillæg til Grønlands Fanerog. og Karsporeplanter. Meddel. om Gronland, III. — 1896, a: Nye Bidrag til Vest-Gronlands Flora. Ibid. H. XV. — 1896, b: Det sydlige Gronlands Vegetation. Ibid. SCHROETER: Das Pflanzenleben d. Alpen. Zürich, 1904—1908. SIMMONS, H. G., 1906: The vascular plants in the flora of Ellesmere- land. Published by Videnskabs-Selskabet, Kristiania. SYLVÉN, 1906: Om de sv. dicotyledonernas förste förstärknings sta- dium, m. 25 Tafl. Kgl. Sv. Vet. Akad. Handl. Bd. 40. TrECUL, 1865: Du Tanin dans les Rosacées. Compt. rendus de l’acad. dese: ара" T.LX: WAGNER, 1892: Zur Kenntniss des Blattbaues d. Alpenpflanzen und dessen biolog. Bedeutung. Sitzungsber. d. K. K. Academie d. Wiss. in Wien. Math.-naturw. Classe. Bd. CI, АБВ. I. WARMING, Euc., 1884: Om Skudbygning, Overvintring og Foryngelse. Naturhistorisk Forenings Festskr. København. — 1886, a: Om Bygningen og den formodede Bestøvningsmaade af nogle grønlandske Blomster. Oversigt over d. Kgl. D. Vidensk. Selsk. Forhandl. København. Resumé francais. — 1886, b: Om nogle arkt. Våxters Biologi. Bih. K. Sv. Vet. Akad. Handl. Bd. 12. Afd. III. Stockholm. — 1887: Om Grønlands Vegetation. Meddel. om Grønland. Vol. ХИ. København. Worr, 1908: Monographie d. Gatt. Potentilla. Bibliotheca bot. H.71. Stuttgart. 6 Кмор JESSEN. Potentilla palustris (L.) Scop. Lit. Irmisch, 1861. LINDMAN, 1884. WARMING, 1884. KNUTH, 1894. KöLrın Ravn, 1894. Norman, 1895. ASTRID CLEVE, 1901. FREIDENFELT, 1904. SyLVEN, 1906. Wozr, 1908. Potentilla palustris 13 widely distributed over a great part of the northern hemisphere. It lives in Arctic and Sub- arctic regions and extends southwards as far as 40° N. lat.: and the monographer of the Potentillas, Tu. Wozr (1. с.), d FR, Е: Fig. 1. Potentilla palustris. (Denmark, Oct. 20, 1912; about 1/2). The leaves are dead. я, n, Scale-leaves of the winter bud, 1911—1912; its internodes are slightly elongated; а, 6, с, 4, foliage-leaves оп axis I unfolded during early summer; 1, floral axis; с subtends the principal shoot II with the leaves 1—5 which have unfolded during summer; 6 subtends a)supplemental shoot : (“ Bereicherungsschuss’’). states that it is highly probable that the species is a remnant of the Arctic tertiary flora and that it has doubtless only recently — perhaps during the Glacial period — advanced south- wards. It is everywhere confined to boggy and peaty soil and thrives best at the edge of swamps and high-moors among the peat-forming mosses. Norman (l. с.) writes that in Arctic Norway it frequently occurs on the strand and he believes that it is in part distributed by ocean-currents. The alcohol material on which the following description is based was collected in Greenland and in Denmark. Rosaceæ. 7 The figure given of the structure of the shoots of Poten- tilla palust., illustrates a common feature as regards the shoot- succession when flowering begins. After the winter-bud, which is shown by the short internodes, the first 2—3 elon- gated internodes of the floral shoot form a continuation of the older axis, while the following internode constitutes the basal part of the ascending branched floral axis. Often even be- fore flowering begins the principal bud grows out, subtended by the leaf (Fig. 1,c) which is seated at the base of the flower-stalk, and forms a sympodial continuation of the axis. But different specimens from the same locality, at any rate in Denmark, may differ greatly as regards the degree of development reached by the principal shoot during the year in which the parent-shoot flowers, in that the length may vary from 2 cm. to about 10 cm., and this proleptic development is by no means a fixed rule even though it is certainly typical in tem- perate regions. By comparison of herbarıum specimens from Denmark, Iceland and Greenland in the Botanical Museum in Copenhagen, it appears, however, to be the rule that the pro- leptic development of the principal shoot is rarer in the two latter countries than in Denmark, and rarer in Greenland than in Iceland. It was found to be the rule among the Green- land specimens that the principal bud did not emerge until next spring. In vigorous specimens from Denmark, the leaves a and b in Fig. 1 also often subtend proleptically developed supplemental shoots (‘‘Bereicherungsschiisse’’) while such were not found in the Greenland specimens in the herbarium, and in only a few of those from Iceland did the leaf b subtend a slightly precocious shoot. This proleptic development may perhaps be regarded, for instance, as a weapon in the struggle against the mosses among which Potentilla palustris commonly grows; since if the growth in length of the obliquely placed stem was arrested in the middle of summer by the formation of the floral shoot it might easily be overgrown by the mosses. 8 Knup JESSEN. Furthermore, a more extensive assimilation is attained by the expansion of the new leaves. Other leaves also may subtend shoots which, however, usually do not grow out until next summer, and it should be mentioned that such shoots which pass their first winter in bud differ in certain-respects from vigorous precocious shoots; the latter begin with a more-or-less elongated internode, and the first leaf is a foliage-leaf, so that they can begin to assi- milate immediately, while the other shoots begin with 1—5 short internodes and the first 1—3 leaves are scale-leaves. After germination — probably in the spring — Potentilla palustris begins a stage of vegetative growth which lasts for several years. The first year it may form a rosette, but even then may also have elongated internodes (IRMISCH, SYLVEN, WARMING). The primary root is slender, but does not die during the first year (WARMING); afterwards the plant is entirely depen- dent on its adventitious roots which according to IRMISCH arise upon the epicotyledonary axis even during the first year. So far as I have observed they do not arise upon the prolep- tically developed shoots until during their second summer. The brown, somewhat compressed stem may attain a con- siderable length; Г have dug up sympodia above two metres in length; they may live for at least seven years. As the stem branches freely, vegetative propagation takes place abun- dantly. The two-rowed leaves usually live for one summer only, but SYLVEN says that a single leaf may pass the winter in a green condition. After the leaves have died the large sheaths persist and those seated at the apex of the year’s shoot envelop the winter-bud; in the latter a few scale-leaves may also be developed (see Fig. 1). In Denmark there is no fixed rule as to the height above the surface of the bog at which the shoot-apex lives through Rosaceæ. 9 the winter. Some shoot-apices are found raised as much as 10 cm. above the surface, while others are at the surface, and others again are overgrown by the sphagnum of the bog. To judge from specimens in the Copenhagen herbarium 1 1$ not a fixed rule for the shoot-apices to die away during winter _ in Greenland and Iceland, though they no doubt often do so; normally the plant forms а monopodium till flowering begins. In the event of the shoot-apex dying during winter there will always be found a lateral bud upon the stem, situated at the surface of the bog or below it, which can continue the life of the plant. [вмтзсн does not understand why this plant ıs not reckoned among woody plants as much, for instance, as Vac- cinium oxycoccus, and WARMING (1884) very properly calls it a shrub. When Potentilla palustris is a chamæphyte it cer- tainly resembles most closely a dwarf shrub, but the case becomes more doubtful when the plant is a hemicryptophyte or helophyte. SyLvEn, TH. WoLr and ASCHERSON and GRAEB- NER! mention it as an undershrub. Anatomy. The adventitious roots proceed from the nodes of the stem especially during the second period of vege- tation; they branch abundantly, but in each system there is always a well-marked main axis. The anatomy has been treated by FREIDENFELT (I. с.). I have, on the whole, found the structure to be the same as that which he describes, but in my material there were no roots with periderm. The epi- dermis was characterized by being exceedingly small (Fig.4,A); its contents were always brown in colour. In the periphery of the cortex of the adventitious roots of the first order there occur a few layers of cells with shghtly thickened walls; inter- cellular spaces are absent from this part; in the inner part of the cortex the latter are large, and lysigenous lacune occur. The endodermis is rather thick-walled. FREIDENFELT records that the primary cortex ultimately dies and falls off, a peri- — 27 Synop. d. mitteleurop. Flora. Bd. VI, 1, p. 663. 10 Кмор JESSEN. derm being formed — the latter contains small intercellular spaces — and a secondary woody part. The most vigorous roots in my Greenland material had just begun the develop- ment of these structures, but the cortex was living and con- tained a large amount of starch especially in the part within the exodermis. In the cortex of the absorbent roots from Greenland fungal hyphæ occurred; these are also recorded by FREIDEN- FELT in plants from Sweden. The creeping stem is distinguished anatomically from the floral shoot, partly by its secondary formations and partly by the fact that normally bast is absent from it. The epidermis of the year’s shoot is rather small-celled, the outer wall is about 4.5 thick; and the outermost layer of the cortex is slightly collenchymatous. The cortex, in the middle, consists of especially large cells, some of which die at an early period, forming large lacunæ separated by radiat- ing trabeculæ. The cortex of the floral stem behaves in the same manner. During the first period of vegetation the cortex of the shoot dies, but it is found attached to the shoot even during the third year. The central cylinder is surround- ed by an endodermis. In the floral shoot the several-layered pericycle is transformed into bast tissue, and only in one particular case did I find bast in the rest of the shoot also, viz. in a plant which grew among shrubs and whose annual shoots of the three last years protruded about % metre verti- cally above the surface of the bog. The periderm is formed in the outermost layer of the pericycle, and even in the first period of vegetation it reaches a thickness of several layers. Moreover it appears to be stronger in Arctic specimens, and while in Denmark, at any rate when young, it probably does not contain cork-cells, in the material from Greenland 1—2 layers of phelloid-cells alternated with one layer of cork-cells, and as many as 6 Rosaceæ. 11 layers of the latter may occur. The periderm contains fairly large intercellular spaces, which however are always absent from the inner side of the cork-layers. On older parts of the stem the periderm peels off as brown concentric flakes. During the first period of vegetation of the shoot is form- ed the greatest amount of the total secondary wood, since often the addition during following years is only a few layers thick. The annual rings are, however, fairly distinct. The structure of the wood is regular, one-layered medullary rays being separated by vessels and tra- cheids in one to a few layers. Some wood-parenchyma is, however, also present. The pith is heterogeneous. It con- sists of a broad peripheral part of homogeneous living cells which lasts during the life of the shoot, and ın addition of a central part which dies away early. The pith, the medullary rays, the pericambium and the phel- loid tissue are the chief depositories Fig. 2. Potentilla palustris. of food-material, and during autumn + % (anchor nas tom ene they are filled with starch, but on arian (Dae LENS investigating the shoots in November it will be found that almost all the starch has disappeared and has been replaced by fat the presence of which is indicated by use of Sudan III. In Denmark this change takes place in October. The imparipinnate leaves are covered on both sides to a greater or less degree with hairs of the usual unicellular form; older leaves are often almost glabrous. Швмизсн (1. с.) writes that in the young plant the leaf-stalk and the margin of the leaves bear stalked glands, and on older plants he found such glands also on the leaves which had developed during spring, while they were absent from the leaves which 12 Кмор JESSEN. had unfolded later in the summer. In the neighbourhood of Copenhagen 1 found in spring numerous glands upon both sides of the leaf (Fig. 2) and upon the leaf-stalk of Potentilla palu- stris, but they were wanting later in the summer, probably they had then fallen off. With the exception of a few speci- mens, leaf-glands were entirely absent from the Greenland plants which have been investigated. Fig. 3. Potentilla palustris. Transverse section of leaf from Ivigtut in South Greenland (*6°/1), The inner wall of the epidermis of the upper surface is highly mucilaginous. The under side usually presents a greyish blue colour; this ıs perhaps due to a thin, granular covering of wax, which must, in addition to the hairs, serve to prevent the under side, which is rich in stomata, from getting wetted. Stomata may also be found on the upper side; they then occur either singly or in groups (Fig. 4, С). The stomata are sunken (Fig. 3). The upper side is slightly reticulated and the outer wall of the stout epidermis is not specially thick, about 2.5 и; the inner wall is highly mucilaginous (Fig. 3). The Rosaceæ. 13 palisade parenchyma usually consists of two layers which may be more or less strongly developed; the leaf figured has a relatively compact structure. The spongy parenchyma 1$ rather loose in structure (Fig. 4, B), and a slit-like lacuna is often found within the subepidermal layer. No stereom pro- per is present in the leaf, only along the main bundle some bast occurs. Fig. 4. Potentilla palustris. (From Ivigtut in South Greenland). A, Epidermis and the outermost layer of cortex of an adventitious root of the first order; the epidermis is about 6.5 / high. В, Section of the spongy parenchyma of the leaf. С, Epidermis af the upper and D of the lower surface of the leaf. (A, В, 2/1; С, D, 139/1). At the apex of the leaf-teeth there are hydathodes with water-stomata on the upper side of the teeth. The dark brown flower is about 25 mm. in diameter (KNUTH); the petals are small, about three times as short as the sepals. It is usually stated that the petals persist while the fruits are ripening (e.g. Бу IRMISCH, LINDMAN, Кмотн and NEUMAN in “Sverige’s Flora”), the sepals after pollination closing round the inner parts of the flower, but 14 Кмор JESSEN. Tu. Worr, on the other hand, states that they fall off after flowering. The flower-biology has been exhaustively des- cribed by Heınsıus! and Кмотн (1. c.) from Holland and from the North Frisian Islands respectively; unfortunately my Greenland material contained only a very few flowers, but they corresponded with the descriptions of the above authors in being decidedly proterandrous. According to Кмотн Po- tentilla palustris behaves as follows: Shortly after the young flower has expanded, the extrose anthers on the numerous erect stamens open; when the anthers have fallen off the filaments bend outward toward the calyx and the corolla so that room is made for the stigmas which are now furnished with small, yellow stigmatic papille. The flowers are eagerly visited by bees. Honey is secreted by a green disk between the carpels and the stamens; this had already been observed by KONRAD SPRENGEL. NORMAN writes that the great amount of honey contained in the flowers attracts numerous insects, flies, butterflies and bees, which continue to visit the flower a long time after fertilization has taken place. The time of flowering is in Central Europe in June and July; in Arctic Norway it begins about July 20; in the moun- tains of northern Sweden it is recorded to be in August, in Greenland in July and August. In specimens from Denmark I found young flower-buds in the beginning of May, while such were not found in the middle of March. So the flower- development probably begins at the same time as the closely folded winter-bud begins to open. In some of the northern- most localities of Arctic Norway it may happen that flowers are not developed. Miss CLEVE found no fruit in the mountains of northern Sweden. The theory advanced by Norman that in the coastal districts of Norway Potentilla palustris is dispersed by marine currents is borne out by the fact that the fruit is "1 Bot. Jaarboek, 1892, Tab. II, figs. 7—9. Rosaceæ. 15 certainly able to float. This point has been discussed by Котртх Ravn (I. с.) among others. He found that the fruits keep afloat even while they are germinating, and anatomical investigation showed that the thick testa consists of several layers of cells, which contain air. The inner part of the peri- carp consists of protective stone-cells. Potentilla tridentata Soland. Lit. Hartz, 1894; Wozr, 1908. This plant grows in Greenland, Arctic North America, Labrador, Newfoundland and Canada. It occurs in clefts of rocks, on lichen-heaths and in other dry localities. The alcohol material was collected in the following loca- lities in Greenland: Christianshaab, Ivigtut, Sukkertoppen and Holstensborg. LANGE and Worr call Pot. tridentata an undershrub, and as is the case in Potentilla palustris the lignified axis lives several years. It is a typical wandering plant with a horizontal, richly branched rhizome from which slender adventitious roots arise. Usually the runners, which may be above 30cm. long, are etiolated and bear scale-leaves, but creeping shoots with elongated internodes, bearing only foliage-leaves, are also met with; such epiterranean shoots may also be ascending. The scale-leaves are distichous. The shoots which arise from the runners and which often grow out from older stems have probably a life-cycle of several years. In the first period of vegetation they are wandering and pass the winter with straight apex, and not until the second period or perhaps even later do they enter the rosette- stage and form foliage-leaves. It appears to be necessary for the young rosettes to pass through a stage of vegetative growth before they attain the flowering stage. The apex . of the rosette-shoot is protected during winter by the large, closely folding sheaths of the older leaves; scale-leaves are 16 Knup JESSEN. probably not developed (Fig. 5,A, В ,C). The foliage-leaves probably remain green during winter; when they die the three leaflets fall off separately, and the stalk persists (Fig. A). The floral shoot is terminal and the principal bud arises in the axil of the uppermost rosette-leaf; it begins with a small transversly placed scale-leaf, and opens its first foliage-leaf in the year in which the parent-shoot flowers. As other lateral shoots may be developed in the rosette, the plant (gg Fig. 5. Potentilla tridendata. A, Flowering rosette-shoot; 1, floral axis; the principal bud II is borne in the axil of the uppermost rosette-leaf, the dead leaf-stalks are seen below the fresh leaves. B, Drawn from a herbarium specimen; a scale-leaf upon an older and now flowering shoot subtends a new runner. С, Apex of a runner. D, Carpel; the ovaries are hidden by long hairs. A, C are from Ivigtut, 20.8.1883; B from Ivigtut, 1868; D from Præstefjæld (4, B and С about natural size: D 32/1). may form loose tufts. The floral axis bears a few foliage- leaves below the often profuse inflorescence. Anatomy. Only older adventitious roots have been investigated. The phelloid of the periderm contains, as in the rhizome, scattered thick-walled wood-cells. The cells of the cork-layers are rather thick-walled. The intercellular spaces are fairly large. The young runner is protected by a strong epidermis the outer and inner walls of which measured 4.4 and 3.5 м respectively. The outermost layer of the cortex is also fairly Rosaceæ. 17 strong, but otherwise the thick cortex is rather loose and perishes early. The central cylinder which is surrounded by an endodermis contains only a few primary bundles and a homo- geneous pith of long duration. A continuous cambium de- velops during the first period of vegetation a considerable amount of wood as in Pot. palustris. This — the secondary wood — consists, in addition to large vessels, of one-layered medullary rays usually one layer high, and also of about equal quantities of thick-walled stereom-cells and woody parenchyma. Some indistinct annual rings may be distin- guished in the older rhizomes which may attain an age of at least about ten years. The periderm is thick and consists of series of layers of usually 3-layered phelloid separated by one layer of cork. The outer tangential wall of the cork- cells is especially very thick. Some of the older phelloid- cells in rhizomes from several localities had very thick walls which were lignified (Fig. 6, H). These strengthening cells form connecting bands between the vigorous cork-cells in the periderm and probably thereby delay the peeling-off of the latter. There are large intercellular spaces in the peri- derm; these are, however, absent from the inner side of the corky layer. The often rather tall flowering shoot is stiffened partly by a strong bast cylinder situated outside the sieve-tissue, and partly by the secondary wood formed by a continuous cambium the peripheral part of which consists chiefly of thick-walled tracheids. The tangential walls of the epidermis are strong, the outer one measured 4.5 и and the inner 3.3 y. The outer layers of the cortex are somewhat collenchymatous and as the otherwise loose cortex is rather few-layered, the outer stereom-ring is close to the periphery of the stem. An endodermis is present. The pith consists of a narrow, small-celled peripheral part formed of living cells only, and a larger central part consist- XXXVIII. 2 7 18 Кмор JESSEN. ing of larger cells of which a greater or smaller number often dies, but nevertheless the shoot cannot be called hollow. The leathery leaf, which is 238—270 м thick, has a deci- dedly xeromorphic structure in correlation with the fact that the plant grows in dry localities; perhaps it also may live through the winter. It is glabrous upon the upper surface, but upon the lower surface it bears in addition to scattered adpressed hairs of the common type also numerous glandular hairs of two other forms (Fig. 6, E, F.). In the smaller form the apical cell is almost spherical and in the larger ovate to cylindrical. The stalk is often several-celled. These hairs spring from epidermal cells which are smaller than the rest and have a thin outer wall. Perhaps these hairs with their thin-walled and plasma-filled apical cells are able to absorb water which may then be carried into the leaf. Both the upper and lower epidermis have very strong outer walls; thus that of the upper surface measured between 8 and 11 и and that of the lower about 11 и. The lower surface of the leaf is moreover highly reticulated. The lower epidermis in particular has a strong cuticle which is somewhat rough. The epidermis of the upper surface is characterized by its tangential walls being mucilaginous to a very high degree, especially the inner; the mucilaginous covering is always wanting above the larger bundles. (Fig. 6, A.). Both leaf-surfaces bear stomata, but upon the upper surface they are very scattered and are surrounded by a group of smaller cells (Fig. 6, В). Upon the lower surface they are placed closely and are not definitely arranged. They are sunk below the level of the surface (Fig. 6, A). The radial walls on both the upper and lower surface are only slightly wavy; they are highly perforated on the lower sur- face (Fig. 6, C); on the upper surface the pores occur in larger quantities only near the bundles. The palisade parenchyma occurs in 2—3 layers of closely Rosaceæ. 19 Fig. 6. Potentilla tridendata (From Christianshaab in Greenland). A, Transverse section of leaf; both the tangential walls of the upper epidermis are mucila- ginous. B, Epidermis of the upper leaf-surface, a stoma is seen surrounded by small cells. C, Epidermis of the lower leaf-surface. D, Section parallel to the surface of spongy parenchyma from the middle of the leaf between a pair of bundles. Е and Е, Two glandular hairs from the lower leaf-surface. @, Longitudinal section of leaf-stalk through the zone where the leaf- blade is cast off; 6, bast; ch, collenchyma; J, the basal portion of the leaf; st, stalk ; z, xylem. H, Fragment of a transverse section through the periderm of a rhizome; №, cork- layer, the striation of the walls is not so distinct as shown in the figure; p, phelloid-cells ; the upper side is the outer. (A, В, C, D, Е. Е and H 22/1; @ 18/4), РА 20 Knup JESSEN. placed narrow cells, which to a certain extent are grouped around the bundles. The proportion between the thickness of the leaf and the thickness of the palisade parenchyma is about = The spongy parenchyma is built in such a manner, that between the bundles in the middle of the leaf areas occur where the intercellular spaces are very large. This structure has this effect that the thick, solid leaf suc- ceeds in getting a well-developed system of larger canals in the respiratory tissue!. For the rest the spongy paren- chyma consists of rather small cells of simple form (Fig. 6,D). The larger vascular bundles are accompanied by bast both upon the upper and the lower surface. The leaves are very rich in sphaero-erystals which occur especially along the bundles. When the trifoliolate leaves die, the leaflets fall off sepa- rately, as mentioned above, and leave a scar upon the leaf- stalk which persists upon the rhizome for a year or more. Fig. 6, С shows a median, longitudinal section through a leaf-stalk just below the blade. The leaf-fall is in several ways prepared for through the anatomical structure. The three bundles of the leaf-stalk divide, so that three enter each leaflet. The bundles of the median leaflet are all of almost the same strength, and lie in almost the same plane; in the lateral leaflets the two lateral bundles are very small. In the zone where the leaflet is to be thrown off there is no bast along the bundles, as is the case both in the leaf-stalk proper and along the main bundles of the blade, and the zone in question is further distinguished by a collenchy- matous area which extends straight through the stalk. In the peripheral part of the collenchymatous area the direc- tion of the line of separation is indicated by the longitudinal direction of the cells; thin-walled areas further facilitate the falling process. 1 Cf. Coptis trifolia. Arct. Flow. Pl., I, 6. Ranunculaceæ, by KNUD JESSEN; Fig. 58, В. Meddel. om Grönland, Bd. 36. Rosacee. 21 The leaf-stalk is also stiffened, in addition to the fibrous tissue of the bundles, by some collenchyma which at the angles attains a thickness of several layers. The outer wall of the epidermis is also strong. The open, cymose inflorescence is few- to many-flowered. The entire petals are white, and the stamens — 20 in num- ber — have very long reddish filaments. The diameter of the flower is 10—15 mm. (Wo tr). The flower is scentless; there is a greenish yellow disk around the base of the stamens; the intensely flesh-coloured anthers appear to dehisce before the stigma is functional (WARMING's notes). Investigation of about 10 flowers preserved in spirit and collected from different districts in Greenland showed that the minute stigma of the long styles does not bear papillæ even in flowers which were so old that the anthers had fallen from the fila- ments. In spite of a close search I found only one pollen- grain upon a stigma; among the crowded long hairs of the gynophore much pollen is usually found in older flowers. The anthers form pollen abundantly, which is, apparently, capable of germination. WARMING states in his notes that in some inflorescences from the previous year from different localities in Greenland he found a few fruits which appeared to be ripe. Harrz (1894) found Pot. tridentata in flower on Præstefjæld on June 15, 1890; the flowering is continued into September. Potentilla pulchella В. Br. Lit. ANDERSSON and HESSELMAN, 1900. Simmons, 1906. Wozr, 1908. LUNDAGER, 1912. This is а purely Arctic species; it grows in Spitzbergen, Greenland, Arctic North America and оп Wrangel Island off N. E. Siberia. Its habitats are rocky and grassy flats. — The alcohol material I had for investigation was from Snenæs in N. Е. Greenland and Tempel Bay in Spitzbergen. 22 Knup JESSEN. Pot.pulchella has a vigorous multicipital primary root,which may attain a considerable length — about 30 cm. or perhaps more — and probably lives as long as the plant; adven- A a Fig. 7. Potentilla pulchella (From Snenes in Greenland). A, Shoot with internodes elongated to an unusual degree ; J—J, the monopodial main axis bearing in its basal part dead fragments of leaves from the previous year; 11, flowering lateral axes (12. 7. 1908; the upper floral axis is wrongly marked Г); about natural size. B, f. humilis; d, closely packed dead fragments of leaves; at the top the living leaves form a layer a few mm. in height (22. 6. 1908); about natural size. C, Pistil which has not as yet received any pollen; 25/1. D, Fragment of a style with stigma upon which are germinating pollen grains ; 24/1. titious roots are probably not developed. The often numer- ous, rosette-leaved shoots are usually very close-set, and solid tufts of considerable size (about 20cm. in diameter) may be formed. The new leaves project between the crowded Rosaceæ. 23 old leaves and their remains. The older portions of the stems appear ultimately to become bared (Warmıng’s notes). The prostrate floral shoots (Tu. Wozr) are lateral (Fig. 7, A.); their subtending leaves are often dead when they flower, so that they arise from the stem below the terminal rosette. The leaf-sheaths are large and closely folded around the young leaves at the apex of the stem. I found no scale- leaves. The much divided leaves are densely hairy especially in LANGE’s f. humilis, the leaves of which bear long close-set hairs on both sides, while his f. elatior is only slightly hairy upon the upper surface. LUNDAGER (1. с. р. 406) writes regarding 7. elatior that on Lille Snenæs in М. Е. Greenland it was found along river banks, and had very long roots which might reach a length of 1.15 metre “and to a great extent were lying so high that a portion of them lay bare.” On the other hand, he found f. humilis on gravel-banks exposed to the wind and so greatly influenced by the sand-drift that they ac- quired columnar forms, “pillars,” as shown in figures 7, В and 8. Anatomy. The very slender absorbent roots con- tain fungal hyphe in the large-celled, few-layered cortex. The epidermis dies early and collapses; the same happens to large portions of the cortex with the exception of the exodermis and the cell-layer outside the strong endoder- mis. The structure of the exodermis is specially strong, its radial walls, especially on the outside, being highly thickened in the same manner as in Pot. emarginata. Secondary growth in thickness begins early in the roots and also the formation of periderm. Usually I found in the lateral roots a conti- nuous woody portion with scattered vessels, but in these roots the woody portion may also be divided into radiating parts separated by broad non-lignified medullary rays. This is what happens in the primary root, and owing to this struc- ture and also on account of its considerable thitkness (5—6 | Fig. 8 Potentilla pulchella; “рШагз” from М. В. Greenland (Nat. size). (From С. Н. Ostenfeld and Andr. Lundager: List of vascular plants from N.E. Greenland. ‘‘Meddel. om Grønland”, XLIII, København, 1910). Rosaceæ. 25 mm.) it becomes а spacious reservoir of food-material. А plant collected on Snenæs in the middle of June contained in the primary root in addition to some starch large quantities of fat. The walls in the older secondary cortex are collen- chymatous and the tissue has the appearance of being highly compressed. — The periderm around the roots does not attain any considerable thickness (about 10 layers) and consists of alternating layers of cork and phelloid. Small intercellular spaces occur between the phelloid-cells, but there are pro- bably none upon the inner side of the cork-layers. The structure of the rhizome is similar to that of the primary root, the vessels being gathered into rays rich in parenchyma and separated by broad non-lignified medullary rays (Fig. 9,G.). Distinct annual rings were not observed either in the rhizome or in the primary root. The pith is large and persistent; it contains large intercellular spaces. ‚ The secondary cortex becomes compressed by the growth in thickness; the walls are somewhat collenchymatous. In the young rhizome there is an endodermis. The periderm con- sists of layers of cork and phelloid-cells which alternate; it does not attain any considerable size (about 8 layers). Inter- cellular spaces occur as in the root. In a rhizome collected on July 12 on Snenæs (N. E. Greenland) in addition to some small-grained starch a considerable amount of fat was found in all the living tissues. The flower-stalk contains about 10 vascular bundles ; no continuous cambium is developed. There is no endo- dermis proper. In the older stalk the several-layered peri- cycle is probably transformed into bast; but my material contained young stalks only. The few-layered cortex con- sists of an inner part with large cells without chlorophyll and an outer part with chlorophyll; in the latter the outer- most cell-layer is slightly collenchymatous. The outer wall of the epidermis is about 3 м thick. The pith is homogeneous. a es 26 Кмор JESSEN. Fig. 9, А shows that the leaf-margins may be revolute, which is especially the case in f. humilis; and the tomentose hair-covering of the lower surface is a good protection for the leaf against the danger of excessive transpiration. Both Fig. 9. Potentilla pulchella. A, Transverse section of a leaf-segment. B, Fragment of a transverse section of a leaf of f. elatior, the inner wall of the upper epidermis is highly mucilagi- nous. С, Stoma from the lower surface of the leaf. D, Epidermis of the upper and Е of the lower surface of the leaf; Ah, the base of a hair. И, Section of spongy parenchyma parallel to the surface. G, Fragment of transverse section of rhizome ; с, cambium; m, pith; р, periderm; r, medul- lary rays. x xylem. A, From Tempel Bay in Spitzbergen: В, С, D, Е, Е and С from Snenes in Greenland (A and G about 16/1; В, С, D, E and Е 1/1), leaf-surfaces have scattered glandular hairs. Upon the upper surface only a few stomata occur which are situated on a level with the leaf-surface; the numerous stomata upon the lower surface are on the other hand usually slightly projecting Rosaceæ. 27 (Fig. 9, С.). The outer wall of the upper epidermis in f. elatior is 2.5—3.5 и thick; in f. humilis from Lille Snenæs it was found to be about 6.5 и; the inner wall is mucilaginous (Fig. 9, B); the radial walls are straight (Fig. 9, D); in the lower epidermis they are wavy (Fig. 9, Е). In the palisade paren- chyma also a difference was found between the two forms, there being 2 layers in f. elatior (Fig. 9, В), but 3 layers in f. humilis. which is on the whole somewhat more xeromorphic. The spongy parenchyma is rather loosely built and consists of somewhat branched cells (Fig. 9, F). Bast either does not occur or is slightly developed only along the strongest bundles. The yellow petals are slightly longer or according to SIMMONS usually only just as long as the sepals, and the diameter of the flower is about 10 mm. (Worr); there are 20 stamens with short filaments. According to the alcohol material from Snenæs in Greenland the species appears to be homogamous. In the young flower which was just opening the anthers were found to be closed, and the short, slightly lobed stigma had not as yet received pollen (Fig. 7, C). In an older flower some of the stamens were bending with open anthers inwards towards the pistils, and the appearance of the stigmas was as shown in Fig. 7, D; they were densely covered with germinating pollen. Spontaneous self-pollina- tion probably occurs. Pot. pulchella flowers in July; in Spitz- bergen flowering begins at the end of June and is continued into September. Fruit is set regularly and abundantly (A. and H.; Stmmowns). Potentilla nivea L. Lit. Bonnier, 1894. Hartz, 1894, р. 6; 1895 a, рр. 287, 300; Ъ, р. 322. Norman, 1895. Квосн, 1897. ABROMEIT, 1899. ANDERSSON and HESsELMAN, 1900. CLEvE, 1901. SyLvEÉN, 1906. HozzsTEIN, 1907. Wozr, 1808. This species grows on rocks, and on rocky and grassy flats; it avoids well-aerated moist soils (CLEVE). It is circum- 28 | Кмор JESSEN. polar and is found in addition in Alpine regions in the tempe- rate parts of Asia, America and Europe. The alcohol material is from northern Norway and from Greenland. Pot. nivea has a multicipital main root, which may attain a considerable length, at least 60 cm. (WARMING's notes) and may also become very thick (6 or even 9 mm.). The often numerous shoots are usually short and close-set, so that they form compact tufts which may attain а diameter of at least 20 ст. The shoots are covered by the old, brown leaf-sheaths which, however, disappear in the course of the following years. Sometimes looser tufts are also found; WARMING, in his notes, even mentions short, horizontal shoots. In the Botanical Museum in Copenhagen I have measured obliquely ascending shoots which were about 10 cm. in length. Ad- ventitious roots, which are even fairly strong (about 2 mm. in diameter), are frequently found, and it must undoubtedly be presumed, as WARMING also is of opinion, that new indi- viduals may be formed by vegetative propagation. The floral shoots are lateral; in a small tuft gathered on Dan- marks © in East Greenland at the end of the winter the young floral shoots were found to be subtended by dead leaves. The leaves wither in the autumn, but in favourable localities in southern regions a few small new leaves may protrude a little before the winter owing to the fact that the plant has no real winter-bud with scale-leaves. According to Harrz, in Scoresby Sound in East Greenland it lives through the winter in places free from snow. Anatomy. The absorbent roots contain fungal hyphæ in balls in the cells of the few-layered, loosely built cortex. The structure of the epidermis and the exodermis is as in Pot. emarginata; but the walls of the exodermis are some- what weaker than in the latter species. In the primary root the secondary woody portion is at first continuous, and not Rosaceæ. 29 until the roots get somewhat older are the broad non-lignified medullary rays developed. The periderm is not very thick; in it one layer of cork-cells alternates with one layer of phel- loid-cells. The older secondary cortex becomes highly com- pressed and is somewhat collenchymatous. The rhizome, with the exception of the periderm, has a structure similar to that in Pot. pulchella. In the periderm one layer of phelloid-cells alternates with one layer of cork- cells, as in the root. The inner cellulose-layer of the phelloid- cells is fairly strong. An endodermis is present in the young rhizome. The structure of the rhizome from Dauphine, as described by HoLLSTEIN, is similar to that of the present one. HOLLSTEIN writes that the outermost layers of the phloem is somewhat collenchymatous, and I find this gene- rally to be the case in the greater part of the area between the cambium and the phellogen, but the peculiar appear- ance cannot be explained solely by this fact; it is no doubt chiefly due to a strong radial pressure. The rhizome together with the primary root is the chief depository for food-material; and I find the alternation in the abundance of starch and fat according to the season of the year to be the same as is indicated in Pot. emarginata. The materials contained in the rhizome in five different months from some- what different localities in Greenland varied as follows: — 18.11.92. 10. IV. 92. 31.V. 08. DNA OR 4 XI. 91 Starch. | a little. | 0. 0 to rather much. | much. | 0. Fat. | much. | rather much. a little. | very little. | rather much. The plants gathered in November, February and April were from Danmarks © in East Greenland, while the plants gathered in May and August were from other localities in East Greenland. The quantity of the starch increases during summer till in August all the tissues are crowded; in November the starch has disappeared entirely; in February a small 30 Кмор JESSEN. quantity was, however, present. During summer only a very small amount of fat occurs; during autumn the quantity increases; а maximum is reached about the middle of Fe- bruary, then the amount begins to decrease. For the structure of the flowering axis the reader is referred to HOLLSTEIN. The leaves, which are usually tripartite, are covered with а dense felt of white hairs on the lower surface; the Fig. 10. Potentilla nivea. A, Transverse section of а leaf, a few of the cells of the upper epidermis have mucilaginous inner walls. B, Stoma on the lower surface of the leaf; it projects slightly. С, Epidermis of the upper surface of the leaf and D, epidermis of the lower surface. Е, Section of the spongy parenchyma of the leaf parallel to the surface. F, Stalkedfgland from the upper sur- face of the leaf. Kongsvold in Norway. (А, В, С, D, Е, Е 5). upper surface is glabrous or may sometimes be more or less hairy. Fig. 10, F shows a glandular hair. Such hairs were found, both on the leaves from Norway, and on leaves of plants in the Herbarium of the Botanical Museum in Copen- hagen gathered on Danmarks Ø in East Greenland (HArTz), at King Point in Arctic America (determined by Dr. OSTEN- FELD), in the Tyrol (Huter) and in the Altai. The glandular hairs are found on both leaf-surfaces though not in great quantity. There is the more reason for emphasizing this fact as WOLF Возасег. 31 (1. с.) in his key of the species in the group Nivea states that among others the present species and Р. Vahliana (see p. 34) are destitute of such hairs. The method for discover- ing hairs described by Worr (loc. cit. pp. 32 and 38) cannot be employed with certainty in the case of leaves with a dense layer of felt as in these two species, when the object is to prove the possible presence of glandular hairs; I have de- monstrated the hairs in transverse sections. The leaf-margins are often more or less revolute. The stomata do not occur exclusively upon the lower surface; the few scattered stomata upon the upper surface are situated on a level with the surface, while those upon the lower surface often project slightly (Fig. 10, ВБ). The outer wall of the upper epidermis varies in thickness from about 3 to about 4.5 и, in that of the lower it is about 2.5 и thick. The radial walls of the epidermis are straight and undulating on the upper and lower surfaces respectively (Fig. 10, С, D). Several of the cells of the epidermis on the lower surface of the leaf have mucilaginous inner walls (see also Квосн). In the leaves investigated the palisade parenchyma consisted of 2—3 layers of cells. The thickness of the leaf varied between 132u (М. Е. Greenland) and 192 м (Norway) and the proportion between the leaf-thickness and the thickness of the palisade parenchyma was 7. In the two upper layers of the palisade parenchyma the cells were well-formed; they inclosed large intercellular spaces. The spongy parenchyma was loose in structure and consisted of slightly branching cells (Fig. 10, E). Here reference should be made to BonnIEr (1. с.) who in leaves from Spitzbergen found two layers of palisade-cells which were scarcely longer than they were broad. In speci- mens from the Alps he found 3—5 layers of palisade-tissue: the cells were narrow and well-formed, and intercellular spaces were either wanting or few in number. WARMING says in his notes regarding the flowers from Greenland: “The flowers are highly conspicuous on account 32 Кмор JESSEN. of the intensely yellow petals which are either of one colour or have an orange-coloured spot at the base. Honey is secre- ted abundantly.” According to Worr the diameter of the flower is 10—15 to 18mm. The flowers in my material from northern Norway and from Greenland were homogamous, and appeared to behave like those of P. emarginata. The _ five innermost stamens which stand 27 before the sepals are longer than the others and rise upwards almost to the level of the pistils. In flow- ers from Kongsvold and Alten in Norway it could be seen how the somewhat outwardly bent styles touched with their stigmas the open anthers of these stamens. А, Longitudinal section through a WARMING (1. с.) also finds that flower from Spitzbergen (about */). the anthers and stigmas are deve- Drawn by Eug. Warming. В, Carpel showing the peg-shaped cells [oped simultaneously “and the sta- at the base of the style (about 23/1). Fig. 11. Potentilla nivea. mens stand so erect and close to the pistils that self-pollination must undoubtedly easily take place (Fig. 11, A).” In Scoresby Sound Hartz (1. с. 1895 a) observed Syrphids to visit the flowers of P. nivea. Norman records that the flowering period is in the latter half of July and in the first half of August; CLEVE: from June 30th to July 10th; LANGE: July—August in Greenland where Harrz, however, in 1889, 1890 and 1892 found it flowering in the middle of June. In tke botanical museum of Copen- hagen there are flowering specimens from different places in N. Greenland gathered at the end of June. P.nivea sets fruit abundantly probably everywhere. Fruit-dispersal is as in Р. emarginata. Greenland specimens gathered on the 4th of February on ground free from snow had ripe fruit enclosed between the contracted sepals, and "WARMING (notes) found (July 15) fruit from the preceding year hidden in the same manner. Rosaceæ. 33 Potentilla Vahliana Lehm. Lit. Wozr, 1908. This is an Arctic species which lives in North-West Green- SIMMONS, 1906. land, Arctic. America and adjacent islands (Unalaschka) and perhaps the Rocky Mountains. It occurs in dry localities exposed to the sun, especially when they are gravelly, and . extends to the snow-line (WoLr, LANGE). terial was collected near Godhavn on Disco. The alcohol ma- Pot. Vahliana is a plant with a tufted growth; it has a strong, long, multicipital primary root. An almost unbranch- ed primary root on a small tuft from Godhavn, which WARMING had measured, was 60 cm. long. The numerous shoots are covered, almost for their whole length, with dead leaves and remains of leaves, pressed closely together, and upon the surface of the tuft the living leaves hardly protrude above the old dead ones (Fig. 12, А). Adventi- tious roots, which may be fairly vigorous, may arise from the shoots, but vegetative propagation 15 undoubtedly only а rare occurrence. The shoot-structure is monopo- dial. small leaves and 1—2 large flowers. The The lateral flowering stems bear 1—2 whole plant is covered with long, yellow hairs. Anatomy. I have not examined the roots. The structure of the rhizome agrees in its chief points with that of Pot. pulchella. The pith is however considerably smaller; the greater part of it was found to be com- pressed; it contained starch, as did the other XXXVII. Fig. 12. Potentilla Vahliana (God- havn on Disco, July 7). A. А shoot-complex about 10 years old; about 1/2. Above are seen two flowering stems and one which has finished flowering pro- truding from among the living leaves; for the rest, the greater part of the shoots are covered with dead masses of leaves, through which the re- mains of the flowering stems of previous years are protruding. В, Leaf; about 1/2. Drawn by Eug. Warming. 3 34 Кмор JESSEN. depository tissues of the rhizome, and, in addition, large sphaero-crystals. The shoot may attain a considerable age. In the oldest part of the shoot in my material about 15 partially indistinct annual rings could be counted with an average breadth of about 40 и. Even at this age the rhi- zome is closely covered with dead fragments of leaves. An endodermis is present. In the flowering stem there is no distinct endodermis. Outside the vascular bundles, about 10 in number and sepa- rated by broad medullary rays, there is a ring of stereom consisting of 3—7 layers of shghtly lignified bast. The outer part of the medullary rays also becomes woody. The struc- ture of the cortex 13 similar to that of Pot. pulchella. The epidermis consists of rather small cells; the tangential walls are about 6.5 thick. The pith is homogeneous. The leaf is hairy on both surfaces, but more so upon the lower (Fig. 13, A), where there also occurs a fair number of glandular hairs of the type figured (Fig. 13, Е); the stalk is 2—4 celled. Besides in the alcohol material, fairly large numbers of glandular hairs are shown also on leaves from Ellesmereland (determ. Simmons) and from King Point in Arctic America (determ. OSTENFELD), contained in the her- barium of the Botanical Museum in Copenhagen; (see Pot. nivea р. 31). The glandular hairs may also occur upon the upper surface of the leaf along the margin. The outer wall of the upper epidermis is about 2.5 « thick; the inner wall is not mucilagmous. The radial walls of the upper epidermis are almost straight (Fig. 13, С), of the lower undulating. Stomata occur only upon the lower surface, and as seen in the figure, they often project some- what, a fact which should doubtless be correlated with the dense hairy covering (compare Pot. pulchella and Pot. nivea). There are two layers of coarse palisade-cells and a rather loose spongy parenchyma the cells of which are slightly Rosaceæ. 35 branched (Fig. 13, В, D). The thickness of the leaf was about 140 and the proportion between the leaf-thickness and the thickness of the palisade parenchyma was dd Some weak bast occurs upon the under side of the main bundle of the leaf and also at the leaf-margin when a bundle runs along the latter. The diameter of the flower is somewhat large, WoLF records 18—20 mm.; the petals are broader than long and yellow in colour, and these large, conspicuous bowls doubtless Fig. 13. Potentilla Vahliana (Godhavn on Disco). А, Transverse section of a leaf-lobe, the edges are revolute, the lower surface is tomentose. В, Transverse section of leaf; st., stoma. С, Epidermis of the upper surface ; h, cross section of а hair which is standing on a small stalk. D, Section of spongy parenchyma parallel to the surface. E, Glandular hair from the lower surface of the leaf. Е, A highly projecting stoma which was found behind the revolute leaf-edge. (A, about 25/1; В, С, D, Hand F %%5/;). readily attract insects. WARMING states in his notes that honey is secreted abundantly. According to WARMING’s notes, of the 20 stamens the episepalous are longer than the epipetalous and stand more erect against the pistils. To judge from my alcohol material Pot. Vahliana appears to be homogamous, and self-pollination probably takes place. Flowering begins in July; according to Simmons it lasts only a short time. Fruit is set abundantly (Stmmons). 3% 36 Кмор JESSEN. Potentilla emarginata Pursh. Lit. ABROMEIT, 1899. Hartz, 1894, р. 32; 1895a., р. 288; 1895 b., р. 322. Exstam, 1897. Simmons, 1906. Wozr, 1908. This plant is high northern and cireumpolar. In Green- land it lives on rocky and grassy flats; on Nova Zembla it prefers dry slopes exposed to the sun (ЕкзтАм).1 The alcohol material was collected in Greenland in dif- ferent places. Pot. emarginata has a strong, multicipital primary root which may attain a length of at least 25 ст. and a thickness of 6mm. During the first years the seedling is in the form of a rosette without lateral shoots; the specimen figured has just entered on its second- or perhaps its third summer (Fig. 14, A). Later on it branches freely and very compact tufts rich in shoots, measuring as much as 20 cm. in diameter, may be formed. Looser tufts are however often found; these plants have probably lived in more favourable localities. The shoots in such tufts are ascending or may at first have a horizontal part several cm. in length. Often adventitious roots, which may even be somewhat vigorous, arise from older shoots, though this is hardly a normal occurrence, and then vegetative propagation may take place or, rarely, even а slight vegetative wandering may occur. The large brown sheaths of the spirally arranged leaves persist for several years, but the shoots may ultimately get rid of them. The floral shoots are lateral. They usually bear 1—2 leaves and are ascending or the lower portion is prostrate. They are subtended by those leaves which in the autumn occur uppermost in the rosette upon the monopodial main 1 Exstam (1. с.) records Р. fragiformis Willd. from Nova Zembla, but as this species, according to Wo tr, does not occur on Nova Zembla, he has probably had specimens of P. emarginata before him, as these two species are commonly confused (cf. Wozr, р. 510). Rosaceæ. 37 shoot. Plants from Kekertak in Greenland collected Sept. 14 had formed flower-buds for the next summer, which had developed so far that the exterior of the stamens was fully formed and in the interior the mother-cells of the pollen- grains could be seen; the carpels were formed, but no indication of the formation of ovules could be observed. In the specimens from Upernivik (May 17) the young flowers were somewhat larger; as yet no pollen was to be seen, but in the carpels the style was formed, and in the ovules the embryo-sac could be seen. The plants probably pass the winter with dead leaves; in Kekertak they had withered in the middle of September 1886, and in Upernivik they began to unfold in the beginning of June 1887. The shoot-apices are protected by the leaf-sheaths; scale-leaves are not developed. Fig. 14. Potentilla emarginata. A, Seedling from Lille Snenæs in N. Е. Greenland (25. 6. 1908; about */s); the cotyledons and the lower darkly shaded leaves were dead and date probably from 1906 and 1907; there are two fresh leaves at the shoot apex; the primordial leaves have a simpler form than the leaves of the full-grown plant. B, Leaf of a full-grown plant from Spitzbergen ; somewhat reduced. (Drawn by Eug. Warming.) Anatomy. The struc- ture of the slender absorbent roots is shown in Fig. 15, F. The epidermis is very weak and collapses, on the other hand the structure of the exoder- The rest of the cortex consists of a few layers of cells in which balls mis is very strong especially in its radial walls. of fungal hyphæ occur; in the intercellular spaces fairly vigorous hyphæ are found. In the secondary woody portion, which begins to develop early, the medullary rays are absent from the slender roots, but in the older roots broad, non- 38 Кмор JESSEN. lignified medullary rays are gradually formed, similar to those which are found in the primary root, of which they may constitute the main portion. The rays of wood contain much parenchyma. The periderm consists of alternating layers of cork-cells and phelloid-cells; in the former the walls of the lateral roots are specially strong. The intercellular spaces are fairly large in the periderm; they are absent from the inner side of the cork-layers. The outer part of the se- condary cortex becomes gradually compressed owing to growth in thickness. The primary root which, as already mentioned, may attain a thickness of at least 6 mm. is doubt- less the chief depository of food-material. The rhizome is similar in structure to that of Pot. pul- chella. In the periderm 1—2 layers of phelloid-cells alter- nate with a layer of cork-cells. The intercellular spaces are rather large; they are absent from the inner side of the cork- layers. The rhizome may attain an age of at least about15 years. The material at my disposal allowed an investigation of the amount of the food-materials — starch and fat — con- tained im the rhizome in the months of May, June, July and September, with the following result:— Ws NI 4. VI medio. VII TÆER Starch. | a little | rather much much | much Fat. | much | rather much | a little | a little (СГ. Pot. nivea). Though the rhizomes investigated were collected in different localities and in different years yet the investigation is of some interest particularly as regards this plant which is hardly green during winter. ‘Plants from the second column were collected on June 4, and the leaves were not unfolded; consequently, the very decided increase in the amount of the starch as compared with the plants in the first column cannot be due to new acquisition, but must have been produced by the transformation of other materials which Rosaceæ. 39 were already present. The simultaneous decrease of the amount of the fat may indicate that some of this material has been used in the starch-formation, though naturally nothing can be stated with certainty on this basis. Besides, there are other materials than fat in herbaceous plants which, like the latter, decrease in quantity during spring. The vascular bundles in the floral shoot vary greatly in number (7—17) and are surrounded by a sheath of several- layered stereom which is well-developed especially opposite the vascular bundles; the outer part of the woody portion of these bundles is connected with the bast-sheath, so that the groups of sieve-tissue are surrounded by stereom. The cortex is similar to that in Pot. pulchella. The outer wall of the epidermis is about 5 in thickness. The inner portion of the pith breaks down. An endodermis is not present. The leaf of Pot. emarginata is most nearly mesophilous; the thickness is about 170. Along the margin and along the bundles of the leaf a fair number of long hairs occur; upon the surface of the blades these hairs are fewer, but short-stalked glands are rather numerous especially upon the lower surface (Fig. 15, A B), a fact which has also been observed by Abromeit (1. с.). The outer walls of both upper and lower epidermis are about 24 in thickness; the inner walls of the upper epi- dermis are not mucilaginous. As regards the radial walls of the upper and lower epidermis the reader is referred to figure 15,C and D. Stomata occur on both surfaces, but they are more numerous upon the lower surface where they may project slightly; upon the upper side they are on a level with the surface. The coarse, somewhat barrel-shaped pali- sade-cells occur in 2—3 layers. The proportion between the leaf-thickness and the thickness of the palisade parenchyma is found varying from Е to 1° 1 Ernst MICHEL, Zur Kenntniss der Jahresperiode unserer Stauden. Inaug. Dissertat. Göttingen, 1909. 40 Кмор JESSEN. The spongy parenchyma consists of shortly branched cells and has a loose structure (Fig. 15, A, Е). Slit-like lacune are often found upon the dorsal side of the leaf. Stereom 1s absent. Fig. 15. Potentilla emarginata. A, Fragment of а transverse section of a leaf; К, a glandular hair. В, Epidermis from the lower leaf-surface with 2 stomata and a glandular hair К. С Epidermis of the upper leaf-surface. D, Epidermis of the lower leaf-surface. Е, Section of the spongy paren- chyma of the leaf. Е, Transverse section of an absorbent root; ep, the collapsed epider- mis; ex, exodermis; A, fungal hyphæ, which partly forms balls in the cortical cells and partly fills out the intercellular spaces; $, the leptome-groups of the central cylinder. A, From Lille Snenes; В, С, D, Е and F from Upernivik (A, В, С, D, Е and Е about 72/3). Возасег. 41 According to LANGE the flowering period in Green- land is in July—August, and on Nova Zembla the plant flowers in the middle of summer and late in autumn (Ек- sTam)!; but Hartz found it in flower on Præstefjæld in Greenland on June 15, and at Scoresby Sound on June 19; in Ellesmereland the flowering begins in the middle of June and is continued into August (Stmmons). In my alcohol material there were numerous flowers from Upernivik gathered on June 21. Pot. fragiformis Willd. f. parviflora Trautw., which is perhaps a synonym of Pot. emarginata (cf. WoLr, р. 533), according to ANDERSSON and HESSELMAN (|. c.), is a spring plant in Spitzbergen; it flowers in the middle of June and in the beginning of July. Pot. emarginata sets fruit abun- dantly (EKSTAM, SIMMONS). The large, pale yellow flowers occur either singly or 2—3 together оп the floral shoot. Wozr states the diameter to be 15—18 (—20) mm., Exstam records 20 mm. from Nova Zembla, while Pot. fragiformis f. parviflora from Spitzbergen has flowers only 10—14 mm. broad (A. and Н.). Exsram found no scent, but he found honey which is secreted between the stamens and on the base of the filaments. He was of opinion that homogamy and self-pollination were possible as the stamens were bending over the stigmas. Of the stamens the 5 inner- most are the longest and in the spirit-material they were only very slightly separated from the stigmas. WARMING writes in his notes regarding specimens from Spitzbergen that these 5 antesepalous stamens which he also found to be the longest, are without doubt easily able to shed pollen upon the stigmas. They open later than do the others. I also find that Pot. emarginata is homogamous (Upernivik and Lille Snenæs), perhaps with a short proterogynous stage. In case = pollination fails for some reason or other, the 5 innermost 1 See note, р. 36. 42 Кмор JESSEN. and longest stamens remain as a reserve which ensures self- pollination (Fig. 16). Of a large collection of flowers preserved in spirit and gathered at Upernivik on June 21, 1887, a great number were found with more or less abortive stamens. АП transi- tions from normal to purely female flowers were found. The abortive stamens were shorter than the head of pistils and the anthers were small and shrunken, even in quite young flowers; they con- tained no pollen. The carpels were appa- rently normal though the styles were some- what short, but no pollen was found on the stigmas which were furnished with large papille, and Fig. 16. Potentilla emarginata. no fertilization had A, Longitudinal section through a flower; 1/1. В, taken place in the Pistil with hairs at the base; stigma with many small conical protuberances ; 18/1. Spitzbergen. (Drawn by Eug. older flowers. en According to Ек- STAM on Nova Zembla “Pot. fragiformis Willd.” is visited by flies and other diptera. As is usual in the Potentillas the calyx is persistent and closes around the fruits, whereby their dispersal is regulated. The fruit-dispersal is anemophilous (EKsTam). P. maculata Е. M. vix Pour. Syn. Potentilla alpestris Hall f. Lit. Hartz, 1894, p. 4; 1895 a., р. 288. Norman, 1895. ABROMEIT, 1899. ANDERSSON and HESSELMAN, 1900. CLEVE, 1901. Syıven, 1906. Worr, 1908. The alcohol material is from Spitzbergen and Greenland Rosaceæ. 43 Wozr is of opinion that, judging from its present distri- bution, this plant is an old species which was probably origi- nally circumpolar and has only recently advanced southwards to about 40° N. lat. In Arctic and Subarctic regions it is found both in the mountains and near the sea, in southern regions (Europe and western Asia) only in Alpine and Sub- alpine mountain regions. It grows on rocks, on heaths оп grassy slopes and in copses. The shoot-structure is monopodial and caespitose, but the tuft is rather loose in form, and ascending or almost horizontal branches, 10—15 cm. in length, commonly occur. There is some individualization of the older branches. The branches are for many years covered with the remains of brown leaf-sheaths. SYLVEN writes that in Sweden Pot. maculata is evergreen even in mountainous regions and in the Copenhagen her- barium there are specimens with green leaves from Disco in Greenland gathered in September and in the beginning of April. In a garden in Copenhagen it remains green through the winter. Anatomy. My material of Pot. maculata was extremely poor with regard to roots; but it appeared that the exodermis of the absorbent roots has thin walls very much as in Pot. Anserina. The epidermis of the absorbent roots proved to have collapsed. No mycorrhiza was seen. The structure of the rhizome resembles in its main points that of Pot. pulchella (see Fig. 9,G), the medullary rays are however fewer in number—about 4. Fairly distinct annual rings were seen in the wood. A shoot, 7 cm. long, was about 6 years old, and another which measured about 3 cm. was of the same age. The pith in these shoots was still alive. The secondary cortex had somewhat thickened walls and its outermost part was found to be somewhat com- pressed. The periderm is thick, there are about 6 layers of A4 Knup JESSEN. cork each separated from the other by a layer of phelloid. In this species also, intercellular spaces are absent from the inner side of the cork-layers. Fig. 17 shows a portion of the transverse section of a floral shoot which may serve as an example of the stem- structure of the monopodial Potentillas mentioned in the pre- sent work. The vascular bundles may vary somewhat in number in the different species. The primary groups of vessels always remain distinct: the function of the cambium is slight. The pericycle is transformed into а rather strong bast-ring, about 3—6 layers thick. No distinct endodermis is present. The structure of the cor- tex ıs similar to that described for Pot. pulchella. The leaves are almost glabrous or slightly Вашу upon the upper Fig. 17. Potentilla maculata. surface, and upon the lower surface Portion of a transverse section of à а floral shoot (about 1h); 6, bast; the hairs are sparsely scattered. The ne PO т Pb hairs are projecting. LANGE maintains what disorganized; 2, xylem. that Pot. FrieseanaLge.isdistinguished from Pot. maculata Pour. among other things by the fact of the leaves being glandular, but Abromeit states that in Sweden he has found numerous yellow glands upon the lower surface of the leaves of the latter species. In the key given by Wo tr of the determination of the forms of Pot. maculata, among the northern forms leaf-glands are assigned only to Pot. Frieseana (found only on Disco). I have, however, usually found numerous short-stalked glands on both leaf-surfaces of Pot. maculata from the majority of the localities represented in the Arctic her- barium in Copenhagen. The glandular hairs occur in unequal quantities in different plants and on the different leaves of the same plant; upon the upper surface they often occur especially in the furrows above the larger bundles. The form of the glands is shown in Fig. 18, D. Rosaceæ. 45 The outer wall of the epidermis of both leaf-surfaces was 3—4 u thick. In the epidermis of the upper surface a muci- laginous inner wall was found in а great many of the cells, and in specimens from a single locality in Greenland (Kang- erdlugsuatsiak) а few cells in the lower epidermis had also mucilaginous inner walls. Such cells were often larger than the other cortical cells (Fig. 18, А, В, С; cf. р. 57). The stomata are far more numerous upon the lower than upon the upper surface; they are situated on a level with the sur- face. The palisade parenchyma consist of 2—3 layers of cells and contains large intercellular spaces. The spongy paren- chyma is often composed of short, stellately branched cells and is more or less loose in structure. In the subepidermal layer of the present species more branched cells occur than in the other Potentillas mentioned in this work. As may be seen in the two figures of the transverse section of the leaf, this varies considerably in thickness, from 160 и to 240%. In the thick leaf from Spitzbergen Fig. 18, A, which was somewhat fleshy and almost smooth there were three layers of well-marked palisade-cells, and the propor- tion between the leaf-thickness and the thickness of the pali- sade parenchyma was about a The spongy parenchyma was specially loose in structure owing to its large intercellular spaces. In Fig. 18, B are shown 2 layers of less well-marked palisade-cells and the proportion mentioned above was in this leaf about =; here the spongy parenchyma was somewhat more compact. Unfortunately nothing is known with regard to the conditions in which the different plants were living. It is however of some interest that this difference in the leaf-structure in a single species is also found, though in a more marked degree, in other Arctic plants, viz. among the land-species of Ranunculus. The thin leaf of Pot. maculata (Fig. 18, В) resembles in the structure of the mesophyll the type 46 Кмор JESSEN. exhibited by the majority of the land-species of Ranunculus, while a further development in the direction indicated in Fig. 18, А would lead to а leaf of the А. glacialis type!. The diameter of the flower varies greatly; WOLF records Fig. 18. Potentilla. maculata. ` A, Transverse section of leaf. В, Transverse section of leaf. С, Cortical cell with mucilaginous inner wall from the lower surface of the leaf. D, Stalked gland from the upper leaf-surface. Е, Epidermis of the upper leaf- surface. F, Epidermis of the lower leaf-surface. G, Section of the outer- most layer of the spongy parenchyma. А and G are of leaves from Spitz- bergen, the others are of leaves from Greenland (4, B C, D, F and @ about 15/1; Е, about %/ı). from 10 to 25mm. I measured flowers from Greenland which were about 15 mm. in diameter. The petals are usually golden yellow in colour, sometimes they are of one colour, sometimes there is a saffron coloured spot at the base (Worr). 1 Ranunculaceæ by Кмор Jessen: Arctic Flow. Pl. I. 6. Meddel. om Grönland, Bd. 36. Rosaceæ. 47 The flowering period is June-September. Flowers from Bose- kop (Arct. Norway) were found to be distinctly proterandrous; the next stage is homogamous, and it is probable that self- pollination takes place in the latter, as the outermost of the somewhat outwardly directed stigmas are either in direct contact with the inner erect or slightly inwardly bent anthers, or they are only slightly separated from these inner anthers which occur either at the same level as the stigmas ог at a somewhat higher one. Herman Müller investigated а Pot. verna L! in the Alps in which he found the anthers developed simultaneously with the stigmas. He writes that as the flowers in inclement weather remain half-closed and during the night entirely closed, self-pollination is ensured И cross- pollination fails. The fruit does not ripen in the mountains of northern Sweden (CLEVE). Potentilla anserina L. Lit. WARMING, 1884. Кмотн, 1894. Norman, 1895. CovILLE, 1896. Porsizp, 1902. Poppius, 1903. SYLVEN, 1906. Wozr, 1908. As regards the distribution of this species in Arctic Nor- way Norman writes that the strand and the fore-shore were most likely its original habitats and as the result of civi- lization it has been brought inland where it occurs especially along road-sides and near houses. This supposition is no doubt true for other regions also; thus CoviLLe (I. с.) records that in Yakutat-bay in Alaska where Pot. anserina is common it is confined to the beaches. On Disco Ровзил found var. Egedii in such places only, where it would be washed by the sea. Both Norman and Wotr think that P. anserina is dispersed by the agency of the sea, and the latter author writes that the fruit or even whole shoots (“ganze Stöcke”) 1 It is uncertain whether this was a P. maculata or a P. minor Gil. 48 Кмор JESSEN. can withstand being carried a long distance in salt water without getting damaged. Fig. 21, Е shows a transverse section through a ripe fruit. In the wall, within a few layers of cells which keep alive a long time, there occurs a consider- able area of dead cells containing air. The shaded part is sclerenchymatous. I have found no difference in the degree of development of this floating tissue in the fruit of var. Egedii which Porsitp gathered on the strand on Disco and in the fruit of a specimen of the typical form which had been ‚ growing in Denmark at a distance from the sea. Pot. anserina is almost cosmopolitan; for further notes regarding its distribution the reader is referred to Wozr. The alcohol material is from Greenland and Denmark. SYLVÉN has observed the germination during spring in nature. During the second year the slightly branched pri- mary root was found to be about 3 mm. thick and spindle- shaped; sometimes, however, it is rather weak and then an adventitious root takes on its function. In older plants the primary root is absent. The short, vertical rhizome is monopodial and the annual growth-increment hardly lives more than one year; the leaves die away every autumn and the shoot-apices are protected by scale-leaves. The rosette-leaves subtend from one to several runners, which may attain a length of about one metre; but the longest runner I have seen from Greenland was only about 4% metre long. According to WARMING, who mentions Pot. anserina in his paper of 1884, these sym- podial runners are to be compared with the lateral floral- shoots in Pot. erecta (L.) Dalla Torre. The single shoot-gene- rations consist normally of a long and a quite short inter- node, after which they terminate in a flower. At the base of the flower-stalk are two small connate bracts of which the lower subtends an assimilatory shoot which has leaves in a rosette and as a rule soon produces 1—2 strong adven- Rosacex. 49 titious roots, while the upper bract subtends a new runner of the same appearance as the parent-shoot. “The leaf- ‘shoots lie alternately to the right and left on the runner- sympodium, from which it is evident that the branching is as in a true scorpioid суше” (WARMING 1. с.). From the new adventive plants new runners may arise even during the first summer (WOLF). As every runner may produce several new rosettes, when circumstances are favourable, a parent-rosette of this plant may give rise to a considerable number of new individuals all around it. The runners die during autumn and by this means the young shoots become independent. On the other hand, in less favourable localities, where the runners are short, the number of off-shoots is considerably reduced, and it is espe- cially those shoots which are nearest to the parent-plant which succeed in striking root and becoming independent. It appears, especially in dry soil, to be a critical moment for the young shoot when its roots emerge, because these easily get dried up before they succeed in penetrating deep enough into the soil. Further, it is only the inner shoots or only the innermost one of the shoot-generations of the runners which attain the flowering stage in unfavourable localities. In cases where the flower is not immediately visible, an ex- tremely small dead flower-bud may often be found hidden among the bracts. Besides these creeping floral shoots the leaves of the parent-rosette may also subtend new rosette-shoots which may sooner or later become independent. As already mentioned the young adventive shoots are provided with 1—2 strong adventitious roots. The latter die in the course of the followmg summer while new ones are developed. In older plants these one-year-old roots often take the character of a food-storing root as shown in Fig. 19. They are usually spindle- or club-shaped and may become XXXVII. 4 50 | Кмор JESSEN. about 8 mm. thick and about 10 em. long; they do not branch, but bear thin absorbent roots. During winter they are found crowded with starch which appears to be transformed to fat only in a very slight degree. Pot. anserina, like many other plants, has the power of keeping its growing point on a level with the surface of the ground, at any rate if the shoot-apex has been covered with a layer of earth and in such a case the fresh food-storing roots are in the autumn often found to be raised as much as 2 ст. above the re- mains of the shoot-base of the previous year upon а portion of the axis with elongated internodes. Anatomy. In the slender absorbent roots of the first and second order considerable quantities of strong fungal hyphæ occur which form balls in the cells of the loosely built cortex. The epidermis, especially of the ab- sorbent roots of the second order, , | . 13 very small-celled, very much as Fig. 19. Potentilla anserina . } я (Denmark; 28. 9. 1912.. in Potentilla palustris, but the Autumn stage; the runners dead or Outer wall 15 thinner, and the dying; 2, the leaf-bases; а, the dead 3 = food-storing root of 1911—12; f,two Skin is collapsed; nor are the dark fresh food-storing roots which have ` : . been developed in the summer of 1912 contents found ın Pot. palusiris (somewhat reduced). present here. Fig. 20, A shows the structure of the fully developed food-storing root. On the outside is seen the thin cork with a few fragments of the primary cortex, then follows the thick secondary cortex and lastly a few groups of vessels around the small pith. Not until during the beginning of summer is a continuous woody part developed in the middle Rosaceæ. 51 of the root. The secondary cortex consists chiefly of paren- chyma. It shows concentric layers which result from circles of sieve-tube-groups, the cambium for each circle of such groups producing about 10 layers of parenchyma. Fig. 20, B shows a group of sieve-tubes just outside the cambium, while Fig. 20, C illustrates two such tubes from about the middle of the cortex. In the rhizome, as in the food-storing root, a conti- nuous cambium develops chiefly parenchyma. The periderm in correlation with the short duration of life is not very thick. An endodermis is present. The horizontal portions of the floral shoots are some- what flattened. The activity of the cam- bium is slight, and the few groups of vessels always re- main separated by primary medullary rays. The two- layered pericycle is Fig. 20. Potentilla anserina. A,Fragment of a transverse section of afood-storingroot taken in the autumn (15/1); сс, central cylinder with groups of vessels ; which becomes some- cb, cambium; se, secondary cortex; pe, fragments of the dead £ primary cortex; k, cork; [, concentric circles in the secon- what stronger in the dary cortex with groups of sieve-tubes. В, A group of sieve- : tubes just outside the cambium. C, Groups of sieve-tubes peduncle than т the from about the middle of the secondary cortex. B, C horizontal part. An about =). transformed to bast endodermis-like layer is present! Within the epidermis there is a layer of weak collenchyma. The stomata project above the level of the surface. The leaf varies considerably in size being from about one-third of а metre (f. grandis Lehm.) to 3—4 cm. in length, the hair-covering is also very different; in f. sericea Hayne, 1 Cf. Ермомь Овтн: Beitr. 2. Anatomie d. Gattung. Potentilla. Dissert. Hamburg. 1893, p. 27. 4* 52 Кмор JESSEN. which occurs especially in dry localities, both surfaces of the leaf are white with silky hairs, in the typical form this is the case only on the lower surface, and in the Arctic and Sub- arctic var. Zgedii Torr. and Gr. and var. gronlandica Tratt. the leaves are entirely glabrous or somewhat hairy, on the lower surface only, respectively. Fig. 21, D shows a stalked BR RTE Sor EEE @. Il DREI у В МАИ | | N an | Sue Fig. 21. Potentilla anserina (Greenland). A, Transverse section of leaf. B, Surface view of the upper epidermis of the leaf. С, Surface view of the lower epidermis of the leaf; Ah, а hair- bearing cell. D, Stalked gland from the lower surface of the leaf. Е, Sur- face view of spongy parenchyma. F, Transverse section of an almost ripe pistil; sm, the seed cavity; %, outer layer of the fruit-wall; ae, floating tissue; sc, selerenchyma (A, С, D, Е about 1/1; В about 65; F about 15/1). gland; such occur scattered along the bundles and are more common on the lower surface. The anatomical structure of the leaf, which only lives through one summer, is mesomorphic. The outer wall in the epidermis of the upper surface is about 2.5 м thick; the radial walls are straight or almost straight (Fig. 21, B), but in that of the lower surface somewhat undulating (Fig. 21, С). Rosacee. 53 The palisade parenchyma consists of 2—3 layers of cells, of which the two upper layers are fairly well-marked. The proportion between the thickness of the leaves from Green- land and the thickness of the palisade parenchyma is about a The spongy parenchyma is rather loose and consists of slightly branching cells (Fig. 21, A, В). It is normal for the yellow flowers to occur singly on the often long flower-stalks. Worr records the diameter of the flower as 18—20 mm., and KNUTH as 20 mm. The flowering period occurs in Arctic Norway between June 25 and Sept. 21 (Norman), LANGE records as regards Greenland July—August; in Central Europe the flowering begins in May. ' My material contained flowers from different localities in Greenland, and those which had not entered the homo- gamous stage showed a somewhat considerable degree of proterogyny. Anthers in fully expanded young flowers were still closed while the stigmas were found to be highly papil- lose even before the petals had opened. It was, however, impossible to find a single pollen-grain on these highly papil- lose stigmas, but, as soon as the anthers had opened the stigmas were found to be densely covered with pollen. This fact does not seem to indicate that insect-invisitors have been frequent. The stamens are longer than the styles; they are erect or somewhat bent inwards when the anthers are open. Self-pollination is consequently possible, which opinion is also maintained by Knuru. In the Friesian Islands he found Pot. anserina to be purely homogamous, and he writes that spontaneous self-pollination must take place as soon as the pollen-grains no longer adhere to the drying pollen-sac; he found that there was a possibility for insect-pollination in the fact that at first the pollen was rather firmly attached to the open pollen sacs. According to KNUTH in inclement weather the flowers are half-closed during the day and entirely 54 Кмор JESSEN. closed at night, and this makes self-pollination possible, Porrrus found the flowers to be slightly protandrous at Esbo in Finland; they were very sparsely visited by insects. Concerning the fruits see p. 48. Sibbaldia procumbens L. Lit. WARMING, 1886, а. LinpmMan, 1887. Hartz, 1895, a. Norman, 1895. А. Creve, 1901. KJELLMAN, 1901. SYLvÉN, 1906. Нотлятетх, 1907. This species grows in grassy and often somewhat damp localities, not, however, in bogs (CLEVE). It is circumpolar, and occurs also in Great Britain, Iceland, the Færües and the Alps. The alcohol material was collected in Spitzbergen and in Greenland, in different places. According to SYLVÉN the seedling has a slender, slightly branched main root which together with the hypocotyl quickly becomes woody. From the leaf-axils of the first young rosette monopodial lateral shoots are developed sooner or later, which in the course of time may attain a considerable length; they become prostrate and have a somewhat up- wardly turned apex. The primary root may live for a long time, and in the museum of the Botanic Garden in Copen- hagen tufts are found, 20—30 cm. in diameter, the numerous, closely placed shoots of which are still in connection with the vigorous main root. KJELLMAN (1. с.) writes that Sibbaldia either does not develop adventitious roots or does so only very slightly; but this statement does not quite agree with the specimens I have seen in the herbarium mentioned above; to judge from the latter, adventitious roots — even vigorous ones — may be fairly often developed from older branches, and the plant is certainly not dependent on its primary root, but has some power of vegetative propagation. The shoots are 2 Rosaceæ. 55 more or less covered by the dead shoot- and leaf-masses of the tuft; they branch freely and often have a characteristic appearance on account of the lateral shoots often arising al- most at right angles. The internodes are usually short, and KJELLMAN refers Sibbaldia procumbens to one of the charac- teristic types of rosette-plants; in vigorous specimens from the Botanic Garden in Copenhagen the internodes are, how- ever, as much as about one cm. long. The leaf-arrange- ment is two-rowed. The dead parts of the leaves persist for a Jong time on the shoot and may impart to the latter a peculiarly scaly appearance especially in its older parts; or the shoot becomes quite smooth owing to formation of periderm. The floral shoots with the compact inflorescence are found during summer subtended by dead leaves which are seated below the living terminal rosette. Often 2—3 floral shoots are developed on the same branch. As is the case with the vegetative shoots, they vary greatly in size accor- ding to the nature of the habitat; in much exposed places they attain a length of 1—2 cm. only. — Plants which Harrz collected at Scoresby Sound in the autumn had vegetative lateral shoots of the same year in which a small foliage-leaf had unfolded; in the Botanic Garden in Copenhagen this proleptic development is more pronounced. The new shoots begin either with a scale-leaf or with a weak foliage-leaf. Afterwards no scale-leaves are developed, and the shoot- apices are protected by the older, closely folding leaf-sheaths. The leaves wither in the autumn — at least in the Botanic Garden in Copenhagen. | Anatomy. The primary structure of the adventiti- ous roots of the first and second order is practically the same, with the exception of the central cylinder which in the former is triarch-tetrarch and in the latter diarch. But the outer walls of the epidermis of the lateral roots are somewhat 56 Knup JESSEN. stronger than those in the roots of the first order. The lateral roots have endotrophie mychorrhiza. The older roots are surrounded by а periderm which is formed in the outermost layer of the pericycle. The phelloid-cells occur in a few layers, and each series is separated by a layer of cork. Large intercellular spaces occur between the phelloid-cells them- selves and between the latter and the next inner layer of cork. In the rhizome there is formed during the first period of vegetation a complete ring of cambium which, however, does not produce a particularly large amount of secondary, wood during the first year. With increasing age the latter reaches a considerable thickness and the continuous woody portion is very large; the annual rings in older rhizomes are not dis- tinct. When older the rhizome becomes highly hyponastic. An endodermis is present. Periderm is formed during the first or second period of vegetation, alter which the primary bark with epidermis dies and falls off. In the periderm one layer of cork alternates with about three layers of phelloid- cells. Specimens investigated from the Botanic Garden in Copenhagen had a smaller amount of cork in the periderm than the specimens from the other localities. Large inter- cellular spaces occur which are distributed as in the root. The outer, dead portion of the periderm persists as a very thick protective layer. Normally, no bast is produced in the rhizome, only in very vigorous specimens from the Bota- nic Garden in Copenhagen did I observe scattered groups of bast-cells. The pith is large and lives for a long time; it is, together with the periderm and the medullary rays, the chief depository of food-material. The starch collected during summer disappears, at any rate partially, during winter and is replaced by fat (Bot. Gard. in Copenhagen, January). HOLLSTEIN has described the floral stem from ma- terial from the Dauphiné Alps. There occurred in the stems from these regions considerably more stereom than Rosaceæ, 57 in the stems from Spitzbergen and Greenland investigated by me. Within the epidermis, which had fairly strong outer walls, I found 1—2 layers of weak collenchyma; this was practically similar in the floral region and the rest of the stem, but the degree of development reached by the stereom in these two parts was different. In the floral region there were a great number of primary bundles within a com- mon endodermis; secondary formations were of slight thick- ness only, as no interfasicular cambium occurred. The whole of the pericycle was transformed into bast of medium strength or weaker. In the rest of the stem a complete cambium had formed a considerable amount of secondary wood consisting chiefly of stereom, and in the pericycle isolated bundles of bast occurred. Here, also, there was a common endodermis. Both leaf-surfaces are slightly hairy with compressed, unicellular, thick-walled hairs. Moreover, there occurred on both sides, but in greater abundance on the under side, small stalked glandular hairs of the form shown in figure 22, B. Stomata occur on both surfaces of the leaf in almost equal number; they are sunk below the level of the surface (Fig. 22, А). The outer wall of the upper epidermis is 3.3 to about 4 thick and that of the lower about 3 и. The cuticle is uneven. The surface views of both the upper and lower epidermis have almost the same appearance, the radial walls are slightly undulating (Fig. 22, Е, Е). The inner wall of the epidermis of the upper surface is sometimes mucilaginous (Fig. 22, C). Cells with such mucilaginous walls are often larger than the surrounding cells (this fact is also mentioned by WARMING in his “Halofyt Studier’’)! and they appear to occur in varying quantities in the different localities; in the leaves from the Botanic Garden in Copenhagen more epidermal cells were found with than without mucilaginous walls. The thickness of the leaves showed some variation: 1 See p. 59. 58 Кмор JESSEN. 123 (Nunatsuk in Greenland), 154 м (Gaaseland in Spitz- bergen) and 1744 (Botanic Garden in Copenhagen), and the proportion between the thickness of the leaf and the >=. т and = respec- tively. The palisade parenchyma consists of 2—3 layers of cells. thickness of the palisade tissue was Fig. 22. Sibbaldia procumbens. A, Transverse section of leaf; В, glandular hair from the upper surface of a leaf; С, epi- dermal cell from the upper surface of the leaf with mucilaginous inner wall; D, air-cavity below a stoma upon the upper surface of the leaf produced in part by the formation of a ‘Trichtercelle” (cf. text). Z, Epidermis of the upper surface of the leaf; Е, epider- mis of the lower surface of the leaf, and G, section of spongy parenchyma. А, В, С, D, Е, Е and @ 25|]. A, D from Nunatsuk in Greenland, С from Godhavn in Greenland, D, Е, F, G Gaaseland in Spitzbergen. Below the stomata there is a large air-cavity which is formed either by the surrounding palisade-cells separating from each other at their upper ends, or by the adjacent cells taper- ing in the form of a funnel to the outside (Fig. 22, D). Sibbaldia procumbens may consequently be placed side by side with those specimens which Rywoscu! mentions in support of his 1 Beitr. zur Anat. d. Chlorophyligewebes, Zeitschr. f. Bot., 1912. Rosaceæ. 59 theory as against HABERLANDT’S, as regards the signification of the so-called “Trichtercellen’ in the foliage-leaves; a theory which should more properly be called the WarmınG-RywoschH’s theory, as it corresponds with the views of the former investi- gator’) (see also Alchimilla vulgaris р. 108). The polygonal or slightly branching cells of the spongy parenchyma are arranged in about four layers and include inter- cellular spaces which are usually smaller than the adjacent cells (Fig. 22, А, С). The bundles are not accompanied by stereom; the smaller branches are surrounded with a sheath while the larger ones are in addition accompanied by the usual vein-parenchyma, which compensates for the assimilatory tissue, up to both the upper and lower epidermis. In fresh leaves from the Botanic Garden in Copenhagen tannic acid occurred especially in the tissue of the veins and in both the upper and lower epidermis. The insignificant flowers are collected in 6—12—flow- ered heads which consist of several small cymes. The petals are pale yellow and smaller than the inner sepals. The flowers are 7—8 mm. in diameter (WARMING’s notes). The flowering period begins in July; Norman records July 9 for Arctic Norway, and at Scoresby Sound Hartz observed the first flower on July 7. It no doubt usually sets ripe fruit. MÜLLER is of opinion that it has lost the power of spontaneous self-fertilization (Alpenbl., р. 222); WARMING, on the other hand, writes in 1886, that this cannot apply to the Green- land specimens because he found the anthers placed so close to the stigmas that self-pollination must no doubt easily take place (Fig. 23, B), and in his notes he remarks that the species is homogamous and that he has seen the anthers in contact with the stigmas. Similar observations are recorded by Lınpman (1887) who investigated the species on Dovre ! Halofyt-Studier. Kgl. Danske Vidensk. Selsk. Skr., 6. Række, naturvidensk. og math. Afd. VIII, 4, p. 245. 1897. 60 Кмор JESSEN. (Scandinavia). He writes that the anthers are introrse and the filaments are bent inwards so that the anthers are sepa- rated by only 1% mm. or even less from the stigmas, and stand at а somewhat higher level than the latter; therefore Fig. 23. Sibbaldia procumbens. A, Flower seen from above. The disk (d) curves outwards between the stamens towards the petals (about 5/1). B, Longitudinal section of a flower; the anthers are almost in con- tact with the stigmas (6/1). С, Carpel with germinating pollen and an anther with open pollen-sacs from fiower from Kongsvold in Norway (about 20/1). Drawn by Eug. Warming. he thinks it is probable that self-pollination takes place. LINDMAN says that the flower agrees with MÜLLERr’s Fig. 87. As Fig. 23, A shows, there is a disk that curves towards the petals. Dryas octopetala L. and D. integrifolia Vahl. Lit. BucHEnAu and Focke, 1872. Ногм, 1885. WARMING, 1886, а; 1887. Шимомлдм, 1887. Клнимам, 1890, р. 229. WAGNER, 1892. Hartz, 1894, р. 48; 1895, a, pp. 288, 300 and 311; 1895, b. Jönsson, 1895. Norman, 1895. ЕкзтАм, 1897 and 1898. Квосн, 1897. ABROMEIT, 1899. ANDERSSON and HEssEeLMAN, 1900. HESSELMAN, 1900. А. CLEvE, 1901, р. 45. Dusen, 1901. Кмотне, 1902. ScxroTER, 1904 Hac- LUND, 1905. Simmons, 1906. ЗутмЕм, 1906. HozLsTEIN, 1907. These two species are treated here jointly for brevity’s sake, as they only differ in a few morphological and ana- tomical points. As regards their mutual systematie relation Rosaceæ. 61 the reader is referred to DRUDE!, ABROMEIT (1. с.) and SIM- MONS (I. c.)?. Dryas octopetala is circumpolar and is moreover found on the mountains of the temperate regions of Europe, in Asia and North America, in Iceland and the Færües. D. in- tegrifolia is found in Greenland, in Arctic America, where it has extended somewhat southwards, and in N. E. Siberia (cf. Simmons, where the distribution of both the forms is exhaustively discussed). D. octopetala and D. integrifolia occur chiefly in dry localities, and Norman and A. CLEVE emphasize the fact that D. octopetala avoids damp localities m northern Scandi- navia, and according to Simmons D. integrifolia does not thrive in damp soil in Ellesmereland. Нотм, however, ob- served D. octopetala growing in the Tundras of Nova Zembla. Heaths and rocky flats are their home proper. The alcohol material was collected in Greenland, Iceland, the Færües, northern Scandinavia, Spitzbergen and Nova Zembla. DruDE has included D. octopetala among his “Holzstau- den” (Halbsträucher),” but both HAGLUND and SCHROTER (1. с.) refer it to the dwarf shrubs proper, and their reason for doing so is that the destruction by frost during winter is not great and that the plant is woody and attains a very considerable age. Thus Kınıman (loc. cit. p.229) records that he has found a plant which was upwards of 108 years old. And ScHRØTER figures (Fig. 65) a shrub about 50- years-old which attained a length of about 1.60 metres. The growth is espalier-like* and the branches may be fixed to the soil by adventitious roots. WARMING remarks in his ' Die system. und geograph. Anordnung der Phanerog. in SCHENK: Handbuch der Botanik, III, 2, pp. 212 and 256, fig. 1. 2 Fig. 24 represents different leaves of an intermediate form. 3 Handb. 4. Pflanzengeographie, 1890. + WARMING'S term for prostrate, outspread growth. 62 Кмор JESSEN. notes as regards D. integrifolia in Greenland, that it grows in flat, low tufts, usually almost circular in form and about a foot or so in diameter. Each tuft has a strong main root, and a few slender, weak adventitious roots arise from the branches. From the upper end of the main root the branches proceed laterally in a horizontal direction, and they may become fairly long. It hardly propagates vegetatively. — In D. octopetala from Russian Lappmark Ктнымлм (1. се.) measured branches which were about two metres in length. HAGLUND ascribes to D. octopitala a rich formation of adven- titious roots from the prostrate parts of the shoots, according Pr Fig. 24. SCHROTER records a rich Leaves of a form intermediate between Dryas rc octopetala and D. integrifolia collected at Kin- supply of adventitious roots gigtok in Vaigat by Hartz. АП the sixleaves from the under side of the were found on the same plant, and the dif- ferent forms were represented almost to an branches of D. octopetala in equal degree (about natural size). to him it is almost exclu- sively the adventitious roots, which provide the older in- dividuals with nourishment; he has not been able to find a persistent main root. the Alps. InWARMING’S notes we read further as regards D. integrifolia: “The shoots are close-set and among them masses of dead leaves are found which persist several years; it grows in its own remains; the fresh shoots are seen in the midst of the blackish brown re- mains of old leaves. Large patches may be found dead in a tuft, but the leaves are still densely crowded.” The young seedling of D. octopetala has been described by HAGLUND and Sytven; it forms a rosette which is provided with a main root and which sooner or later develops lateral shoots. The structure of the shoot of the full-grown plant has been treated by Hacıunn (loc. cit. p. 7) who, how- ever, doubtfully refers Dryas octopetala to dwarf shrubs Rosaceæ. 63 with racemose shoot- structure, citing as additional examples Cassiope tetragona and Empetrum; and on р. 33 he says that there is a clear distinction between vegetative or vegetative- floral short shoots and vegetative long shoots. In Lapp- mark where he was staying the long shoots measured 2—4 em., and had spirally arranged leaves in the axils of which short shoots developed even during the same year. ASCHERSON and GRAEBNER! ascribe to D. octopetala а long and much branching stem with prostrate long shoots and numerous erect short shoots. — The results of my investigations of Dryas octopetala and D. integrifolia are in several respects different from those arrived at by HAGLUND. The same year in which a shoot flowers it produces a small number of usually close-set lateral shoots, as generally only the upper leaves subtend shoots. At first the shoots have a scale-leaf, afterwards foliage-leaves follow. In vigo- rous and luxuriant plants the shoots are usually distinguished into short shoots and long shoots, which otherwise behave as described by ASCHERSON and GRAEBNER. In investi- gating shoot-apices in plants both in the Botanic Garden in Copenhagen and in the alcohol material, small flower-buds were as а rule found on all shoots gathered in the autumn; when such were absent the shoots were frequently weak short shoots. Fig. 25 shows a longitudinal section through a flower-bud gathered in August in Greenland; the stamens have just been developed and there are indications of the outermost carpels. The young flowers in the Botanic Garden in Copenhagen had in January developed so far that the stamens were externally fully developed, but in the carpels no ovules were formed. Probably the development is not so advanced at the same time in colder regions where the flowering period also occurs later than in Copenhagen (see ge dé 1 Synopsis 4. mitteleurop. Flora. VI, 1, рр. 889—90. Fr É 64 Кмор JESSEN. p. 73). It is not, however, by any means all the flower- buds that are formed which reach maturity. In Copenhagen Fig. 25. Dryas octopetala. Median section through a young flower-bud which would not have attained the flowering stage until the following summer (Kingigtok in Vaigat; Ар 85) ©, in- dication of carpels; st., young stamens ; р, petal; sp, sepal. numerous dead buds were found during autumn and winter. Otherwise such shoots with a dead apical flower-bud behave as do those that flower. The above-mentioned weak shoots which do not at once produce buds, may do so after having lived through one or perhaps several years of growth in vigour, but I also find that such shoots may after the lapse of a few years die in a purely vegetative stage, probably suffocated by the more vigorous shoots overshadowing them. HAGLUND writes that the short shoots im- mediately put forth new short shoots, and this is undoubtedly often true. I must regard the short shoots in Dryas octopetala and D. inte- grifolia simply as shoots which have become weakened by competition among themselves, and it is undoubtedly normal for the shoot-develop-. ment to end flower-formation. cording to the above the branch-system in Dryas! is sympodial, a few new shoots being produced every of these usually one or two are more vigorous than the others. with Ac- year; Fig. 26. Dryas octopetala (Iceland). 16. 1. 1894. а А shoot-apex which shows how the rolling back of the leaves increases towards the apex of the shoot; the uppermost leaves are living the older are dead ; The length of the b, a leaf seen from the lower surface; €, а highly re- volute leaf seen from the lower surface (HelgiJöns- long shoots naturally son’s figures 1. c.). ı The North American and the North-east Asiatic Dryas Drumondii Richards. behaves in a similar manner. Возасезе. 65 varies greatly; I have found them to be from one to about 10 em. long; the longest are often found at the periphery of the tufts. HAGLUND (I. с.) records that the leaves of D. octopetala die away during winter; others find that the leaves remain green during winter, viz. SCHROTER, Hartz (1895a, р. 182) and Jonsson (cf. Fig. 26); in my material there were specimens collected by Jonsson in January, and they agreed exactly with his figure and in this manner D. octopetala also behaves in the Botanic Garden in Copenhagen. Finally, WARMING remarks in his notes that D. integrifolia has leaves which live through the winter. The oldest leaves on the shoot wither in the autumn but as figure 26 shows there are а few half expanded young leaves at the apex of the shoot and these live through the winter. Scale-leaves are not developed (cf. HAGLUND); the shoot-apices are protected by the leaf- sheaths. | Anatomy. The root-system as mentioned above is characterized by a vigorous main root which bears vigorous branches. The absorbent roots are richly branched and their apices are encompassed with a thick mantle of fungal hyphæ. HESSELMAN figures mycorrhiza of D. octopetala which resemble those of the Cupuliferae; D. integrifolia also has mycorrhiza. HESSELMAN also illustrates the structure of the root by several figures. The central cylinder in the primary root is surrounded by a vigorous endodermis and in the innermost layer of the cortex the radial walls are highly thickened in the middle, so that the transverse section of the wall shows an almost circular figure. Periderm is developed in the outermost layer of the pericycle, and gradually as the periderm and the secondary wood grow the cortex dies and is thrown off together with the hyphe. In older roots it attains a consider- able thickness; it consists exclusively of cork-cells which XXXVII. 5 66 Knup JESSEN. often have somewhat curving tangential walls. There are no intercellular spaces in the periderm. The older roots consist mainly of secondary wood with medullary rays of one layer and several cells high which are united with one another by tangential bands of wood-parenchyma containing starch. In the young shoot there are 3—5 primary vascular bundles. The cambium begins its activity quickly and in the first period of vegetation it forms a considerable quantity of wood. There is no endodermis, as HOLLSTEIN also has noticed. Outside the primary sieve-tissue, and afterwards also outside the secondary, bast is developed, usually in isolated groups, rarely as a continuous ring. The secondary bast is often present in less quantity than the primary, naturally enough, as the woody part affords sufficient support. In the Botanic Garden in Copenhagen there always occurred a more strongly developed bast than in plants from more northern regions. The cortex is of simple parenchyma and the epi- dermis is small-celled and has strong outer walls, about 3 y thick, with a well-developed smooth cuticle. The pith is heterogenous. There is a peripheral part of active, starch- containing cells and in the central part similar cells form a network between larger and dead cells. The active tannin- containing pith-cells have strong walls. Shortly after the cambium has begun to develop wood a phellogen arises in the cortex outside the bast. The first phellogen forms a cylinder which may be closed, but after- wards a typical crust-bark, “Schuppenborke” (cf. HOLLSTEIN) is formed. Each phellogen produces 3—4 layers of cork, and each new phellogen-layer cuts off a portion of the secon- dary cortex with the bast-groups or bast-cells enclosed in it. The primary, but dead cortex may be found on the branches even during the third year, and 5—6 cork-plates may be counted, separated by dead cortical tissue. Rosacez. 67 The secondary wood often, but by no means always, shows distinct annual rings. The growth-increment of the first year is equally vigorous on all sides, but afterwards the branch becomes highly hyponastic. The yearly growth on the hyponastic side is naturally very different in different localities; but it was always found to be less in the Arctic and Subarctic material of D.octopetala, on an average 0.2 mm. in younger branches, than in plants from the Botanic Garden in Copenhagen where the average growth, likewise for younger branches, was found to be 0.9mm. Kıkrman found an average yearly growth of 0.07mm. in the 108-years-old branch mentioned above, and SCHROTER measured a yearly mean growth of 0.22 mm. in a 50-years-old stem from the Alps. The maximum of the radial vessel-diameter was also greatest in specimens from Copenhagen. — The chief deposi- tory for reserve food-material in the two Dryas forms is naturally the branches, and the tissues containing starch are, in addition to the active part of the pith, the medullary rays, which may be as much as 3 cell-layers thick, and several cells in height, and here as in the root are connected by tan- gential bands of wood-parenchyma. During winter the starch disappears from the branches almost entirely and is replaced by fat. In Copenhagen this change took place in November. In the peduncle there are no secondary formations. Outside the numerous (about 20) primary vascular bundles there is gradually formed a strong bast-cylinder which is correlated with the fruit-bearing function of the peduncle; there is no endodermis. The active cells in the pith have weaker walls than have those in the stem. The leaf in D.octopetala and D. integrifolia is leathery and especially in the former is highly reticulated upon the upper surface. The lower surface is silver-white with a dense felt of hairs; the upper surface is usually glabrous, but in several forms it may be more or less hairy; for instance in Dryas 5% 68 Кмор JESSEN. octopetala у, hirsuta Hartz and 0, argentea A Blytt, in the latter the upper surface is densely woolly, and Dryas integri- folia var. canescens Simm., in which the upper surface is - densely tomentose. The leaves are revolute in the bud- condition and also as a rule, to a greater or less degree, when full-grown, especially in D. integrifolia (Fig. 27, A). In this respect it is of importance whether the plant grows in favour- VE TA) $ ESS Fig. 27. Dryas integrifolia. A, Transverse section of leaf; >/, (Eug. Warming, 1887). В, Frag- ment of transverse section of leaf 2*°/,). С, Surface view of epidermis of the upper surface (4/,). D, Section of the spongy parenchyma of the leaf (1%/,). — Dryas octopetala. Е. Hairs from the bundles on the upper surface of the leaf; a, glandular hair (2%/,); b, branched hair (°°/,). able or unfavourable localities. Thus the most revolute leaves are met with in dry rocky flats or on downs (Dove Bay in N. E. Greenland). Even in one and the same tuft there may be differences as regards the degree to which the leaf is revolute. Hartz (1. с. 1895 а, р. 311) figures a tuft of D. octopetala 8 minor from East Greenland, and writes regarding it that “to windward the branches are stunted and decorticated, Rosaceæ. 69 the leaves are torn off or extremely small, and linear with involute leaf-edges; the leaves to the leeward of the shrub are larger and their leaf-edges are involute in a less degree or not at all.” Hartz gives in the same place some information regarding the habitats of the different forms in East Greenland. Thus he found D. octopetala a genuina, which has large broad leaves, in damper localities in sheltered spots, and it was covered with snow during winter, while В minor, which had small narrow leaves, у hirsuta, д argentea and D. integrifolia were found in dry localities, and were usually not snow-covered during winter. The young leaves which live through the winter are so highly revolute that the lower surface is almost entirely covered. WARMING remarks in his notes that the leaves when dry roll back towards the condition of the bud-stage, but on getting wetted they spread out again. As the stomata only occur on the lower surface the importance of the revolute condition is easily understood. The epidermis of the upper surface has a very strong outer wall which is somewhat thicker in the winter leaves than in those of the summer. Thus in the latter it was found to be 2.5 » (Iceland and Copenhagen), 4.4 » (the most common) and 6 » (D. octopetala from downs at Dove Bay in N. Е. Green- land) and in the winter leaves 6—7 y (D. octopetala from Dan- marks Ø off East Greenland) and 8 и (Iceland). The cuticle is considerable; it is smooth and the cutinized portion con- stitutes above one-half of the wall; its inner limit is distinetly toothed (Fig. 27, В). Квосн (1. с.) records that т D. octopetala the walls of the upper epidermis are mucilaginous; this I also have found to be the case, though not in all the cells; on the other hand, he denies the presence of mucilage in D. integrifolia, but this difference I have not been able to verify (cf. Fig. 27, P). The palisade parenchyma is very thick and is composed of about three layers of cells. The spongy parenchyma is rather loose and consists of slightly branching cells (Fig. 27, D). 70 Кмор JESSEN. The epidermis of the lower surface together with the sub- epidermal layer is as а rule separated from the rest of the leaf-tissue by large slit-like lacunæ occurring between the larger bundles (see Fig. 27, А, В). Нотм (1. c.), who has des- cribed and figured the anatomy of leaves from Nova Zembla, mentions similar lacunæ. According to Hooker (Fl. Bor. Amer.) the veins in the leaves of D. octopetala are distinctly visible on both sides of the leaf, and the upper surface acquires thereby a somewhat wrinkled appearance, while in the leaves of D. integrifolia they are only slightly or not at all visible, and therefore the upper surface 15 smooth. Also as regards the anatomy there is a difference in the leaves of the two forms in connection with the bundles, there being in the leaf of D. integrifolia less chloro- phyll above and below the bundles than in the leaf of D. octo- petala, so that the transverse section of the former shows as many bright transverse bands as there are larger bundles, while this appearance is far less distinct in D. octopetala. In other words, the vein-parenchyma is more strongly developed in D. integrifolia than in D. octopetala. The stomata are situated on a level with the surface and occur especially in the spaces between the bundles. — The majority of the hairs are unicellular, they are erect and intertwining and form thereby the dense, thick layer of felt (Fig. 27, A). Upon the larger bundles which project as ribs the felt-hairs are partially wanting, but here large, several- celled, branched hairs and several-celled glandular hairs oceur (Fig. 27, E). During winter the chlorophyll takes up in a marked degree the winter-position described by Krause. In the palisade-cells it gathers in the inner part, and in the cells of the spongy parenchyma it retires from the parts of the walls which border on the intercellular spaces. The two species of Dryas flower generally very abundantly; Rosaceæ. 71 they appear to agree as regards the structure and the biology of the flower. WARMING (1. с. 1886a, pp. 127—28) has described and figured the flower of D. integrifolia. He writes: “that as D. octopetala (in Norway, Spitzbergen and the Alps) this also has hermaphrodite and male flowers and from a biological point of view they appear on the whole to resemble each other exactly.” — Fig. 28 is a reproduction of his figure; А shows а branch with the terminal flower; В, а herma- phrodite flower; С, а male flower, at the base of the hypan- thium of the latter are seen some pistil-rudiments. — “As in D. octopetala the hermaphrodite flowers are larger than the male flowers, but the smallest forms of the former are some- what smaller than the largest of the latter”’!. — According to Simmons (1. с.) the flower of D. integrifolia is оп an average somewhat smaller than the average size of the flower of D. octopetala. — “The petals (8—11) are white or yellowish-white while young, but with increasing age become darker yellow or brownish yellow.” — In D. octopetala they are almost pure white (Simmons). — “It appears that the hermaphrodite flowers may be both homogamous and slightly proterogynous or slightly proterandrous; on July 15, 1884, I gathered near Amerdlok- fjord (W. Greenland) flowers, the anthers of which were open while the styles were still short, and the stigmas appeared to be quite immature; in other cases the pistils are very long and have ripe stigmas while the stamens are still lying bent inwards with closed anthers (Fig. 28, D); lastly there are cases where the anthers and stigmas begin to function simul- taneously and as they are at the same level or the anthers are slightly higher than the stigmas and very close to the latter or rise above them, self-pollination will easily be able to take place; I have even repeatedly observed the anthers to be carried in between the stigmas.” “Honey is secreted by the inner side of the hypanthium (the shaded part of the 1 A similar case is described by Miller in Alpenbl., p. 227. 12 Кмор JESSEN. figures). The stigmas are not very conspicuous or delimited from the rest of the smooth upper part of the style (Fig. 28, F).’ LinpMAN has investigated D. octopetala on the Dovre (Scandinavia) (1. с. Tab. III, Fig. 30). AXELL! mentions polygamy, writes LINDMAN, but he himself found hermaphro- Fig. 28. Dryas integrifolia. А, В, С, D, See text. Е, Pistil. Е, Apex of style with germinating pollen. С, Anther (Warming, 1886 a). — Dryas octopetala. H, Rudimentary carpel of a male flower from Bosekop (Arctic Norway: 1/;). dite flowers only. In the flowers he investigated he found proterogyny to be more decidedly pronounced than is the ° case in Central Europe according to Ricca (Att., XIV, p. 253) and Mi ier (Alpenbl., р. 227). LiNDMAN does not mention proterandry. Exsram (1897) records D. octopetala from Nova Zembla; the diameter of the flower was 10—25 mm., the 1 Ош Anordningarna for de fanerog. vaxternas befrugtning, 1869. Rosaceæ. 13 flowers had по perfume, and he found homogamy only. Self- pollination is made possible by the fact that the anthers bend over the stigmas or towards them. In 1898 the same author writes regarding the flowers from Spitzbergen; here the dia- meter was found to be 20, 25, 27 mm., exceptionally 28, 32, and even 38 mm. (MÜLLER records the diameter of the flower in the Alps to be 27—40 mm.; 1. с.). Self-pollination may easily take place as the anthers shed the pollen immediately upon the stigmas. Honey is secreted between the androecium and the gynoecium. Lastly, РозЕм describes the flower-bio- logy in East Greenland between 70° and 75°. He found D. octopetala to be homogamous; regarding D. integrifolia he writes that it is “dioecious ($).” In my material were 18 flowers (preserved in spirit) from Bosekop in Arctic Norway (collected by WARMING, 1885), of these four were male, the others hermaphrodite; one of the latter showed a distinctly proterandrous stage: almost all the sta- mens had raised themselves into the air and had opened their anthers, while the pistils were yet without papillæ. The others were in a homogamous stage. Fig. 28, Н shows a rudimentary carpel from Bosekop; the ovule with the em- bryo-sac is formed. Moreover, the material contained four flowers gathered by Нотм at Mejduscharskji(?) and probably from Nova Zembla; three of these were distinctly male, the rudimentary carpels did not rise above the hypanthium. In the material was also found a male flower from Lille Sne- паз in East Greenland. The insect-visitors in northern regions appear chiefly to be flies and similar diptera (LINDMAN, Hartz and Exstam). D. octopetala and D. integrifolia flower in early and middle summer (June and August; LANGE, Norman, EKSTAM, ANDERSSON and HESSELMAN, and SIMMONS), and ASCHERSON and GRAEBNER (I. с.) record, as regards Central Europe, (May) June and August. In Ellesmereland D. integrifolia is one of 74 Кмор JESSEN. the earliest flowering species (Simmons). The Scandinavian Floras, on the other hand, record July, August. Usually fruit is set abundantly and regularly; as regards Spitzbergen ANDERSSON and HESSELMAN write, however, that fruitsetting is sometimes irregular, and in northern Sweden the fruit does not ripen at a height of 1000 metres above sea-level (A. CLEVE). Norman (1. с.) often finds empty fruit without kernel. The fruit dispersal is decidedly anemophilous. The highly elongated and densely hairy styles in the ripe fruit spread themselves out in dry weather, but in damp weather they again unite into a single cone usually twisted to the right. The peduncle elongates considerably during fruit- setting and attains double its usual height (Exstam, 1898). Rubus saxatilis L. Lit. WARMING, 1884 and 1886, b. Norman, 1895. Popptus, 1905. This species 15 not really Arctic; it is found in Greenland, but south of the Polar Circle, in Iceland, the Færûes, in Cen- tral and North Europe (also north of the Polar Circle), Siberia and in Caucasia (LANGE). The alcohol material was collected in Greenland, Iceland and in several places in Denmark. The shoots have a two-years period of development. In the first year a shorter or longer scale-leaf-bearing shoot is formed, which is erectly ascending or has an oblique lateral growth; it lives through the winter with its apex at the sur- face of the ground; next year the aerial shoot 1s formed. The length of the shoot of the first year 15 dependent on the distance of the subtending leaf from the surface of the ground, and also on the nature of the soil; as may be expected, in the loose leaf-mould of woods longer shoots are developed (Fig. 29, А), while in harder soil the shoots are often quite Rosaceæ. 75 short. In conformity to this the rhizomes which persist for several years are shaped very differently. The scale-leaves are two-rowed as are the leaves of this species generally. There appears to be a difference be- tween the upper and lower rejuv- enating shoots, only the upper having been found to be floral, and those of them, which had not been proleptically developed in the same year as the parent-shoot, de- velop at the end of summer flower- buds for the next year (Fig. 29, A, C). I have not found flower-buds on the lower and longer shoots. Next year the flower-buds pro- duce first 2—6 foliage-leaves and then the flowers expand, but the vegetative pagating produce first an pro- shoots Fig. 29. Rubus sazatilis. A, Shoot-base from Denmark (Oct. 22, 1912). The plant has been growing in loose leaf-mould and has therefore especially elongated internodes. The lateral shoot a was floral and would have flowered in the summer of 1913, while the two other shoots, at any rate the strongest, b, would have produced assimilatory and wandering shoots. The buds a and b reached the surface of the ground. Lowermost in the figure is seen the upper part of the shoot from 1911; there are dormant buds in the axils of the scale-leaves (slightly above natural size). B, Stolon-apex from Denmark (Oct. 1912). Just at the point where the stolon bends the axis is somewhat thickened, and from here numerous vigorous adventitious roots arise (slightly above natural size). С, Longitudinal section through the bud а in Fig. A; the exterior of the stamens well-developed, but the pollen was found not to have developed yet; nor had the ovules begun to develop in the carpels (°°/,). 76 Кмор JESSEN. erect lower portion like the floral shoot and then creeping runners. The runners have elongated internodes, small leaves, and develop no roots; from the leaf-axils other runners arise. Towards the end of summer the apex of the stolon slightly penetrates the ground; it then becomes negatively geotropic, the internodes shorten and scale-leaves are developed. The stolon is somewhat thickened at the point where it bends and thence long, vigorous and richly branching roots arise (Fig.29, B). These shoots also pass the winter at the surface of the ground. According to WARMING (1884) the stolons may attain a length of 144—2 metres. In Arctic Norway they are, according to Norman, often above 2 metres, and he places the maximum at about 4 metres. During autumn the aerial shoots die away, at any rate partially, and the terminal buds of the stolons are set free and become independent individuals. The structure of the shoot in Rub. saxatilis may be pro- duced through a slight modification of the type common among the Eubatus. As the stolons live through one sum- mer only, the flower-buds are placed near the surface of the ground, and new shoots arise from the base of the floral shoots as well as from the wandering shoots. Moreover, e. g. in Denmark, specimens are met with which form a transition between the two types, as they enter the winter with living shoot-bases, a few cm. long, which protrude above the ground and bear flower-buds. Anatomy. Of roots only those adventitious roots have been investigated which arise from the apex of the stolons. The epidermis in the roots of the first order is thin- walled, while the subepidermal layer has stronger, brown walls. In the roots of the second order the epidermis has thicker outer walls, and the exodermis is not so well-marked. The central cylinder is in the former tetrarch and in the latter diarch. The roots of the second order are particularly characteristic owing to the structure of the cortical layer Rosacee. 77 which is situated just outside the endodermis, the radial walls appearing as thick, refractive, non-stratified cushions (Fig. 30, Е). Also in roots of the first order a similar struc- ture has been found, but it is less common there. The peri- derm is formed during the first period of growth in the outer- most layer of the pericycle at the same time as the primary cortex dies. The stem is covered with a fairly dense layer of uni- cellular, thick-walled hairs; sometimes a few weak spines occur. The axis-system in this species illustrates fairly well how highly the development of the different tissues is de- pendent on the length of life and the function of the different parts of the stem. In the erect part of the stem no periderm occurs, with the exception, however, of the lower scale-leaf- bearing part which as already mentioned becomes at any rate partially perennial. In the first year it consists here of only about three layers, which develop immediately out- side the bast in the cortex as is the case in the species of various other groups of Rubi investigated by Евтзсн (1. с.). In the stolon the same features occur as in the stem of Rub. arcticus; a phellogen is developed which probably does not become functional. On the other hand, in the some- what thickened apex of the stolon a fairly well-developed periderm is found towards the end of summer, in correlation with the altered conditions. In Denmark it consists chiefly of phelloid cells which at that time when the bud begins to become independent, contain a large quantity of starch. Also in the perennial subterranean axis the periderm, which is 10—15 layers thick, consists chiefly of phelloid, at any rate in Denmark in humus soil, while the periderm in the Iceland specimens was richer in cork. The periderm rich in phelloid makes a spacious reservoir of food-material. It con- tains small intercellular spaces. I have seen no endodermis either in the rhizome or in the aerial stem. 78 Knup JESSEN. Bast-tissue is absent from the rhizome; it is most vigor- ously developed in the stem where it forms a continuous cylinder outside the sieve-tissue; the bast is connected with the secondarily formed portions of stereom in the outer part of the groups of wood by the lignified medullary rays. The stolon also contains bast abundantly; it is more slightly developed in the thickened apex. In the rhizome a continuous cambium forms a conti- nuous woody portion which shows distinct annual rings. The numerous medullary rays are 1—3-layered and upwards of 20 cells in height. In the stem there is no continuous cambium except in its basal part; on the other hand a con- tinuous cambium begins to function early in the thickened apex of the stolon. While the cortex in the stem is distinguished into an outer small-celled part filled with chlorophyll and an inner larger-celled part poor in contents, the whole cortex in the apex of the stolon functions as food-storing tissue, and is crowded with starch in the autumn. The 1—8 outer cor- tical layers are collenchymatous especially in the erect stem. The pith is heterogeneous as in all the Rubus-species hitherto investigated (cf. Frirscx). The active cells, espe- cially the starch-containing ones in the rhizome and the apex of the stolon, are found chiefly in the periphery of the pith, but also in addition dispersed longitudinal rows of living cells occur in the central part which for the rest con- sists of larger dead cells. Anastomoses are rare. During autumn the quantity of starch diminishes in the different tissues of the rhizome and at the same time large quantities of fat may be demonstrated. The leaves are slightly hairy on both sides and are somewhat folded. No glandular hairs were found. On the lower surface the bundles are very prominent. The structure is mesophilous in correlation with the fact that the leaf lives St 2 Rosaceæ. 79 one summer only, and is moreover strongly influenced by light The leaves from Greenland (Tasermiut-Fjord) had the following dimensions: Thickness of leaf 110 », thick- ness of the outer wall in the upper epidermis about 5 y, Fig. 30. Rubus saxatilis. A, Transverse section of leaf. В, Epidermis of the upper surface of the leaf. С, Epidermis of the lower surface of of the leaf. D, Section of the spongy parenchyma of thelleaf parallel to the surface. Е, Fragment of an absorbent root of second order; с, the cortex, the radial and outer tangential walls of its innermost layer have highly thickened walls; end, endo- dermis; p, outermost layer of the pericycle. A, B, C, D from Greenland, E from Denmark (А, С, D, Е =; В TEL): in lower epidermis about 2, the corresponding dimensions from leaves collected in Iceland (Adalvik in N.W. Ice- land) were : 130 м, about 4.44 and about 3.3y. An explanation of these differences is suggested by the vague 80 Кмор JESSEN. records on the labels: The Greenland specimens are said to be collected “under and between birch copses,” consequently, most likely on shady ground, while those from Iceland grew on ‘Lien’: (grassy slopes), consequently, probably in more open ground. Leaves collected in very shady localities in Denmark were about 904 thick and had in a high degree the character of “shade leaves,” with only one layer of palisade-cells, which were funnel-shaped, and a few layers of spongy parenchyma. The proportion between the leaf- thickness and the thickness of the palisade tissue was about i in leaves both from Greenland and Iceland. The figures show the leaf-structure of the common type; a loose palisade parenchyma consisting of 2 layers of cells ,about twice as high as broad, and a likewise very loosely built spongy paren- chyma. The stomata occur only on the lower surface of the leaf; they are situated on a level with the surface. (Fig. 30, 4, B, C, D.) — Bast-tissue is absent from the leaf. Three distinct bundles enter the leaf-stalk (see p. 87); shortly afterwards the two upper lateral ones divide each into two bundles of which the lower one on both sides gra- dually approaches the median lower bundle and ultimately fuses with the latter the transverse section of which then becomes crescent-shaped. The fusion takes place a short distance above the point where the sheath ends, and in the rest of the stalk there are 3 bundles. The two upper are round in transverse section and lie in the ribs which form the furrow on the upper surface of the stalk; the two latter especially are accompanied by a strong fibrous tissue. The flower-biology has been investigated by WAR- MING (1886, 65). He writes (loc. cit. р. 40) “The petals in the latter (Rub. saxatilis) stand erect, and the stamens bend somewhat toward the centre without however uniting so closely together above it as in Rub. arcticus; there is, at any rate in a somewhat older flower, an entrance between them Rosaceæ. 81 down to the pistils; they have similar finely warted filaments to those of Rub.arcticus (Fig.37), which likewise terminate abruptly above in a smooth and much thinner portion, and in my opinion the pollen from them must certainly fall upon the stigmas.” H. Mütter (Alpenblumen, р. 216) arrives at the same conclusion. (Compare Fig. 31). WARMING states in his notes that the stigmas become functional very early; even in the hardly expanded flowers from Talvik in northern Nor- way they are highly papillose; the outermost stamens are the next to become functional and afterwards the others. But Fig. 31. Rubus saxatilis (Talvik in Arctic Norway). A and В, Flowers (about ®/,); В, in longitudinal section. С, A stamen (about %/,); ef. text. Drawn by Eug. Warming. even if the flowers have perhaps at that time a short prote- rogynous stage, they become homogamous at any rate early. As regards the supposed importance of the warts on the filaments see Rub. arcticus, р. 97. Poppius (1. с.) records that in Finland Rub. saxatilis is eagerly visited by Vespa vul- garis L. The flowering begins in Scandinavia in June; Norman records June 21. In some years it does not set fruit in the northern districts of Arctic Norway. The fruits may be ripe on July 28 (Norman). In Greenland it is usually found without fruit, and in the Feeröes it sets fruit sparsely. 1 Ostenfeld, Plantevexten paa Færôerne, København, 1906. XXXVII. 6 82 Кмор JESSEN. Rubus chamæmorus L. Lit. WARMING, 1886, a. LINDMAN, 1887. Клнимам, 1890, pp. 55, 113. Hartz, 1894, р. 7. Norman, 1895. KOLDERUP ROSENVINGE, 1896, а. Квосн, 1897. ANDERSSON and HESSEL- MAN, 1900. А. OLEVE, 1900, р. 47. Poprius, 1908. SyYLvEn, 1906. Rubus chamemorus is circumpolar, but thrives also in temperate regions in Asia, America and Europe, where in many places it must most properly be regarded as a relict. It occurs almost everywhere in damp localities, and grows by preference on the loose surface of bogs and marshes. A few exceptions from this rule are however known, the plant in Arctic Norway, according to NORDMAN, growing in dry _localities furthest out towards the sea and especially at some height above it; an explanation of this phenomenon has been sought in the fact that the аш from the sea is extremely damp. In Greenland also, where it only occurs in the coas- tal districts and on islands in the skerries, it grows on rela- tively dry and heath-like tracts. The aerial shoots are vegetative or vegetative-floral. They usually bear 1—5 foliage-leaves and are terminated by a solitary flower or by a dead bud enclosed in the sheath of the uppermost foliage-leaf. The aerial shoot usually has at its base a few scale-leaves with intervening short inter- nodes (Fig. 32, С). The uppermost of these scale-leaves are doubtless as a rule situated above the surface of the bog, and often from the axils of the lower arise during summer scale-leaf-bearing shoots which will next year produce aerial shoots. Consequently, the plant has a two-years develop- ment. During autumn the scale-leaf-bearing shoots form a distinct resting-bud which is situated at the surface of the bog. Ifthe shoot arises from a point which is situated at some depth in the bog it assumes a more or less decided wandering stage before it enters into the assimilatory or ultimately floral stage; the shoot to the right im Fig. 32, D Rosaceæ. . 83 exemplifies this; but the development is nevertheless of two years duration. The plant has, however, a much greater power of wandering, because it gives off horizontal runners which may attain a length of. above 30cm. Fig. 32, A, B are apices of the latter in the wandering and resting stages respectively. These well-marked wandering shoots arise from deeper-lying parts of the plants, and their two-rowed sheath- 1912 1912 Fig. 32. Rubus chamzmorus. A, Apex of а runner (Norway; July 22); k, bud. В, Runner with resting bud in which the ifirst foliage-leaf is formed (Denmark; Oct. 6). C and D are branch-systems in the winter-stage (Denmark; Oct. 6). The shoots marked 1912 have that year been assimilatory shoots, from their base new shoots arise bearing winter-buds. Portions of axes marked 1911 and 1910 have been assimilatory during these years. All the figures are about natural size. like scale-leaves may subtend new runners. In the begin- ning of autumn they may either develop a resting bud which is situated almost at the surface of the bog and which in Denmark in October contains the first foliage-leaf of the assimilatory shoot of the next year, or the runner may live through the winter deeper down in the bog without devel- oping a resting bud proper and next summer resume its 6* 84 : KNUD JESSEN. wandering life. The shoot then takes more than two years to develop. When a runner is entering on its last stage it grows upwards toward the surface of the bog and at the point where it bends the internodes usually become some- what shorter, but by no means always, as easy transitions from runners with 3—4 short internodes at the bending point to runners which form aerial shoots without any shortening of their internodes are met with. This difference expresses the state of dependence in which plants growing in a bog stand to the moss-vegetation of the latter. The shoots, without the shortening of the internodes, were probably in danger of being overgrown by the moss, consequently, no short internodes have been formed but only elongated ones to carry the terminal bud upwards to reach the light. The older plant probably develops only adventitious roots; the latter arise from the nodes of the runners during the second period of growth. Among related species this Rubus has no doubt an unique shoot-structure which appears to make it specially adapted for life in the habitats it has chosen for itself. Anatomy. The adventitious roots are diarch—tetrarch. The endodermis has rather thick walls which have become corky. Even before the periderm, which 1s developed in the outermost layer of the pericycle, is formed, consequently, in the absorb- ent-root stage, the epidermis is destroyed and the two outer layers of the cortex assume the appearance shown in Fig. 33,G. At a distance of a few mm. only from the root-apex we find the epidermis dead, while the cortical layer next to the outermost one only gradually attains the wall-thickness figured, viz. about 2.5 и. The dead epidermis does not however persist everywhere, as the figure indicates, often only frag- ments of the radial walls are found. There are somewhat considerable differences in the ana- tomical structure of the aerial and the under-ground stem Rosaceæ. 85 in Rub.chamaæmorus, differences which are connected with differ- ences of life-duration and function. The conducting tissue ex- hibits the greatest differences. In the aerial shoot there are about 10 primary bundles separated by narrow medullary rays which remain in their primary condition; the intrafas- cicular cambium forms only a small amount of wood. I have not been able to see any endodermis in the aerial shoots, though such is distinetly seen in the wandering shoots. In the latter a continuous cambium begins to function very early and a considerable amount of wood is formed contain- ing numerous medullary rays which are usually one cell- layer broad. In older rhizomes more or less distinct annual rings occur; the oldest I have seen were, however, only three years old. As regards the proportion between the diameter of the pith and that of the whole organ, the usual difference obtains between aerial and under-ground shoots; the propor- tion in the aerial shoot is larger (about 1) than in the runners measured in their primary condition (i—2) In addition there is found in the aerial shoot, but not in the runners, a fairly thick fibrous tissue outside the phloem, and the outermost cortical layer is somewhat collenchymat- ous. The cortex in the runners is considerably thicker than in the aerial shoot, but more space is not thereby gained for reserve food-material, because even in the first vegeta- tion-period of the shoot a periderm is formed in the outer- most layer of the pericycle. Then the cortex collapses and persists during the following years as a loose brown mantle around the rhizome. But even in the third year the phel- logen has not been able to form more than 3—4 layers of which the outer have become corky. In the aerial stem no periderm is developed, not even so much as in Rub. arcticus (Fig. 36, К); consequently, the structure of the assimilatory shoot of Rub. chamæmorus is the simplest among the species of Sect. Cylactis here investigated. — The epidermis around the young runner 86 Knup JESSEN. has thin walls, and the protective layer described for the roots is altogether wanting. The pith is heterogenous as in the two other species; Fig. 33, F shows a somewhat tangential longitudinal section through the central part of the pith of a runner. The active cells, which have somewhat stronger walls in the aerial stem than in the runners, form in addition to a continuous mar- ginal part (of which mp in the figure is the innermost layer) also slightly anastomosing rows in the central part (mc) of the pith where the majority of the cells are dead. But not- withstanding this structure the pith is the chief depository of food-material, and during autumn is found crowded with starch; in addition, the medullary rays and the pericycle in the runners serve as reservoirs of starch. Though Rub. chamæmorus is not an evergreen yet its leaves have in one respect a xeromorphous structure: the upper epidermis has mucilaginous inner walls (Fig. 33, A), a fact which has already been mentioned by Квосн (1. c.). Клит- MAN (l.c., p.113) was probably ignorant of this since he mentions, among others, Rub. chamæmorus and Potentilla palu- stris (which has also mucilaginous walls in the upper epidermis of the leaf) as striking examples of bog-plants without special protective devices against excessive transpiration (see p. 119). Otherwise the leaf has the usual mesophyllous structure; it is rather thin (170—190 ») and the proportion between the thickness of the leaf and the thickness of the palisade tissue varied from about 2 (Godthaab in Greenland) to = (Bosekop in Arctic Norway). The outer wall of the upper epidermis is 2—2.5 и thick and that of the lower somewhat thinner. Stomata occur only upon the lower surface; they are not definitely arranged. The radial walls in the epidermis are somewhat wavy, usually somewhat more so upon the lower than upon the upper surface. The palisade parenchyma is present in 2—3 layers and the spongy parenchyma is loose in structure. (Fig. 33, А, В, С, D). Rosaceæ. 87 The larger bundles are very prominent upon the lower sur- face of the leaf which helps to support the large reniform blade. These ribs consist not only of conducting tissue proper, but also of rather thick-walled vein-parenchyma. True stereom was not observed in the leaf. Along the veins upon both sides, but especially upon the under side, such glandular Fig. 33. Rubus chamæmorus. A, Transverse section of leaf, the inner wall of the upper epidermis is mucilaginous. B, Epidermis of the upper surface of the leaf. C, Epidermis of the lower surface of the leaf. D, Section of the spongy parenchyma of the leaf parallel to the surface. (4, B, C and D from Greenland.) Е, Glandular hair from the lower surface of the leaf (Norway). F, Longitudinal section through the central part of the pith in a runner (Denmark). G, The peripheral part of an absorbent root in transverse section, the epidermis has collapsed, the layer within the exodermis has strong walls (Denmark). А, ВБ, С, D, Fand С ®®/,; E,"/1. hairs are found as shown in Fig. 33, Е; scattered unicellular hairs of the usual appearance are also met with. In the anatomy of the leaf-stalk Rub. chamemorus differs considerably from Rub. saxatilis and Rub. arcticus, which is of interest as Fritsch has shown that a certain systematic 88 Кмор JESSEN. importance must be attached to the number of the vascular bundles and their course in the leaf-stalk of Rubus. And as the group Cylactis (Raf.) Focke in this as well as in most other points has not previously been anatomically investigated I shall give a description of the present case. With the ex- ception of some individual variations the facts are as follows: As in all the species investigated by Евттзсн, and as in the two other species mentioned here, so also in the present spe- cies three vascular bundles enter the leaf-stalk; after a course of about one ст. the two upper divide each into two bundles, and in the whole length of the stalk up to a few mm. from the leaf-blade there are five bundles; at this level one lateral bundle on each side unites with the median one, but the latter divides again into three before the stalk passes into the leaf-blade. The bundles are stiffened by a fibrous tissue along the sieve-tissue. As regards the structure of the flower of Rub. Chame- morus WARMING says in his notes: “Both in northernmost Norway and in Greenland I found the cloud-berry to be purely dioecious and the degree of abortion of the one sex was almost the same in both places. If I may point out any differences it would be that: (1) The Greenlandic flowers are perhaps slightly smaller than the Norwegian; the diameter of the latter I have found to vary between 21% and 31% cm. as regards the male flowers and to be about 21% cm. for the female flowers; the Greenlandic were 2—2 и, cm. for both kinds of flowers, but the material in hand was extremely sparse. (2) The Greenlandic appear to have slightly less rudimentary stamens in the female flowers and likewise slightly less rudimentary pistils m the male flowers than have the Norwegian. In the former the stamens had a some- | what larger anther, in the latter the carpel was more highly developed and the ovule was even indicated, as the figures show.” .... “As regards the flower it may be added, Rosacee. 89 that it is 4—6-merous in calyx and corolla, sometimes purely so, but sometimes with one or other sepal broader than the rest and cleft, a transitional stage to becoming two. In some flowers the hypanthium is quite smooth in the interior, in others it is, over the greater part of the surface or only at the bottom, covered with long, stiff, thick-walled hairs which are quite smooth exteriorly. In the male flowers many pollen-grains were found between these hairs and between the pistil-rudiments.”? In 1887 Linpman recorded Rub. chamæmorus from Fig. 34. Rubus chamæmorus. A, Male flower from Arctic Norway; June 25 (2/,). В, Female Hower from Greenland (*/,). С, Carpel of a male flower from Arctic Norway, in the hollow a slight indication of the ovule is seen (#/,). D, The somewhat folded and ‘slightly lobed stigma of a female flower from Arctic Norway (about 5/,). Drawn by Eug. Warming. Dovre (Scandinavia). He found only unisexual flowers with very small rudiments of the other sex. In the female flowers the petals were 8—10 mm., in the male flowers about 15 mm. long. AXELL! (pp. 46—48, Fig. 11, a, b) records Rub. chamem. as trioecious-polygamous (quoted after LINDMAN). In 1903 Poprrus mentions Rub. chamæm. In the female flowers he finds slight or no secretion of honey; in the male flowers honey 1 Om Anordningarne för de fanerog. växternas befruktning. 1869. 90 Кмор JESSEN. is secreted more abundantly, the stamens are well-developed, with brown anthers, the carpels are very rudimentary and scarcely visible in the hypanthium. Insect-visitors to the flowers are fairly frequent, especially flies. In the male flowers they devour partly honey and partly pollen, in the female flowers it is probably the white stamen-rudiments which attract them, because their attention is always drawn to these staminodes. In Spitzbergen the plant has been collected with flowers twice only; the second time numerous flowers were found at Kol Вау (23. 7.1898); they were smaller than those from the district of Tromsö (Arctic Norway) and were all female (ANDERSS. and HESSELM.). In Greenland it is peculiar that the two sexes have not been found in the same locality. Thus, only male flowers have been found at Julianehaab and on the large island to the west of Ameralikfjord, while only female flowers have been gathered in different localities near Godthaab (VAHL [see WARMING, 1886], LANGE, RosENvINGE and Hartz). WARMING (1. с.) mentions that WORMSKJOLD records fruit of Rub. Chamemorus from Greenland, and ROSENVINGE (1. с.) says that it appears to set ripe fruit at Kasigi- änguit near Godthaab. Otherwise it is not known to bear fruit in Greenland. As regards this point NorMAN records from Arctic Norway that it sets fruit there in favourable years, even in the northernmost districts; in the highest habitats it never flowers, and A. CLEVE (loc. cit., p. 47) records that it is often sterile in the mountains of northern Sweden. In places on the coast of Arctic Norway, at higher levels above the sea, it bears almost exclusively male flowers and probably every year; and NORMAN says that in certain years, only or almost only male flowers occur over very large areas so that the harvest fails entirely. Lastly, Porrivs (1. с.) mentions the distribution of the two sexes; Rosaceæ. Эт he found that at Esbo in Finland the female flowers which were somewhat more numerous than the male, usually oc- curred in more open places in the marshes and bogs, while on the other hand the male flowers were somewhat more numerous in shady localities. Flowering begins in June (LANGE, Norman). In Arctic Norway birds are important agents in fruit-dispersal; and bears probably also play a part in this connection (NoRMAN). Rubus arctieus L. Lit. Твесог, 1865. WARMING, 1886, b. Norman 1895. Poprius, 1903. SYLvÉN, 1906. This species 15 found in the whole of the Subarctic zone, and extends but rarely into the Arctic, namely in Norway, where its northern limit lies somewhat to the south of 71°. Norman writes that it thrives both in dry and damp localities and even in bogs, where it occurs especially on knolls. The alcohol material was collected in northern Scandi- navia. Rubus arcticus does not develop wandering shoots, but vegetative propagation is effected by bud-producing roots; WARMING (1886) records this from Bosekop in Arctic Norway: “Its (Rub. arct.) unbranched or slightly branching shoot, which is a few inches high, occurred always singly but socially over larger or smaller patches of the field” and “it was proved that this species was chiefly propagated by root-shoots, con- sequently it has shoots of the same structure as have those of a great many similar plants, e. ©. Cirsium arvense, Sonchus arvensis, Pyrola uniflora, etc. Its shoots have elongated internodes without rosette-leaves at the base; remains of it may keep alive through the winter and next spring produce a new shoot as the figure (Fig. 35, A) shows.” A sympodium is thus formed which may include at least four shoot-genera- tions. These root-shoots sometimes go through a stage of 92 Кмор JESSEN. growth in vigour before they flower, but I have not been able to investigate such shoots in the material at my disposal, therefore I shall only deal with the propagative shoots formed by the stems. In correlation with the fact that vegetative propagation Fig. 35. Rubus arcticus. A, Portion of root with a shoot-complex arising from it. The first leaf of the lateral shoots is marked with a x (Arctic Norway, July, 1885; %,); drawn by Eug. Warming (l.c.). B, A plant which had been growing in a sphag- num-bog in the province of Quebec in Canada, Aug., 1907; I, the stalk of a foliage-leaf; s, propagating shoot; (slightly above nat. size); drawn from a specimen in the Botanical Museum in Copenhagen. is effected by the roots, the plant has one kind of shoot only, viz. assimilatory floral-shoots which have a two years devel- opment. In the first year an erectly ascending, and when the plant is growing in dry soil, short propagating shoot is Rosaceæ. 93 formed, which usually bears 4—6 scale-leaves in two rows; it probably does not elongate until the end of summer. The specimens figured in Fig. 35, B had been growing in a sphag- num-bog, and the height of the latter has compelled the young shoot to elongate to an unusual extent. The propa- gative shoots pass through the winter with their apices at the surface of the ground. Next year the assimilatory shoot is formed, which also has elongated internodes; it is termi- nated by the solitary flower or by the dead apex of the stem which is hidden in the upper leaf-sheath. The foliage-leaves, which are usually few in number, are two-rowed as are the scale-leaves. Often, the foliage-leaves, especially the upper, also subtend solitary flowers, or slender vegetative shoots occur. The propagating shoots, two or more of which may occur, are as a rule subtended by the lower scale-leaves; buds subtended by other scale-leaves remain dormant or the upper ones may develop proleptically during the same year as the parent-shoot, which results in the individuals becoming fairly rich in shoots. — According to SyYLvEn Rub. arcticus appears to pass through the winter in a partially green condition. Anatomy. The horizontal shoot-bearmg roots are protected by a cork-periderm of about 6 layers without inter- cellular spaces. They are in connection with the shoot-bases, the plant’s reservoir of food-material, and in them the several- layered pericycle and the one-layered, but several cells high, medullary rays in the secondary wood become filled with starch. From the rhizome adventitious roots arise. The absorbent roots are of the nature of mycorrhiza and have a peculiar structure. The epidermis consists of low cells with very strong outer walls (4x thick). The structure of the inner layer of the cortex, which is about 4 layers thick, is as in Rub. saxatilis, but the thickenings of the walls are less and there is an unthickened part on the outer tangential wall. The rather thick endodermis, the thickenings of the 94 Кмор JESSEN. walls mentioned above, and the epidermis turned yellow when treated with chloride-zinc-iodine. (Fig. 36, Æ). The anatomical structure of the annual aerial shoot is naturally very different from that of the perennial rhizome. In the latter secondary wood is developed which forms annual rings, and the primary cortex is thrown off, a periderm of about 6 layers without intercellular spaces and consisting chiefly of cork being formed in the pericycle. The latter, a few layers of phelloid-cells, the medullary rays, which are of the same structure as those of the root, and especially the peripheral part of the pith are depositories for starch. The aerial shoot remains in a far higher degree in the primary condition. A continuous cambium in the lower part forms only a small quantity of wood, and it is hkewise only in the lower portion that a periderm is found like that shown in Fig. 36, F. In the cortical layer bordering on the bast-ring tangential walls are developed, the walls in the outer layer thereby formed become corky, but the inner layer which corresponds to the phellogen remains unaltered. This rudi- mentary periderm distinguishes with certainty the stem of Rub. arcticus from that of Rub. Chamæmorus. — The bast, which is absent from the rhizome, is about 3 layers thick in the lower part of the stem and forms a continuous ring, becoming weaker upwards. Endodermis is absent, but is present in the perennial part of the stem. The outer wall of the epidermis is rather thick and has а well-developed cuticle. Within the epidermis there are two layers of slightly developed collenchyma. — The pith has the usual peripheral layer of starch-containing cells, but in the central part I have seen no such cells. Твесог (1. с.) who has described the pith records that in the central part a few isolated cells were found containing tannin. The leaf, which probably as a rule lives one summer only (see p. 93), has the usual mesophyllous structure; Возасезе. 95 it bears on both sides, especially on the veins, scattered uni- cellular hairs. The outer wall of the upper epidermis is about 2.5 thick. The stomata occur only upon the lower surface and are on a level with the surface. Palisade parenchyma occurs in 1—8 layers. The proportion between the leaf- thickness (about: 155 ») and the thickness of the palisade tissue was about 2. The spongy parenchyma is loose. Fig. 36, A, ЕС... < The structure of the leaf-stalk is as in Rub. saxatilis ; Fig. 36. Rubus arcticus. А, Transverse section of leaf. В, Epidermis of upper surface. С, Epidermis of lower surface. D, Section of spongy parenchyma parallel to the surface. Е, Portion of a transverse section through an absorbent root; end, endodermis; Ay, fungal hyphae, the innermost layer of the cortex has highly thickened radical walls. F, Portion of a transverse section through an aerial stem; b, the bast outside the leptome; с, cortical cells; %, the cork-layer in the rudimentary periderm p. A, B, C, D, E from Bosekop and F from Ängermanland ın Sweden. (A, D1/,; В, С 14]; Е, F SST): along the greater part of the length of the stalk there are three bundles, and only along а very short distance, imme- diately after the three bundles have emerged from the stem there are 5 bundles. The stereom is very much as in Rub. saxatilis. The flower-biology was first treated by WARMING 96 Кмор JESSEN. (1886), who says: “The hermaphrodite flowers are 6- or 7- merous, more rarely 8-merous and have a varying diameter (about 18—30 mm.). The dark-crimson, stellately expanded or even somewhat reflexed petals are highly conspicuous with the green field as а background; therefore it is inter- esting that the flower is nevertheless well adapted for self- pollination.” “On regarding the flower from above one does not see, as in the other Rubus-species, that the styles with the stigmas Fig. 37. Rubus arcticus. A, A flower seen from above. В, A longitudinal section through a ffower the anthers of the outermost stamens have opened. D and G, Two stamens from the innermost circle. Е, A stamen from the outermost circle. F, А pistil. — Cf. text. D, Е, Е С !/,; Arctic Norway. (Bug. Warming 1. с.) are situated in the middle and are closely surrounded by the stamens, which are of about the same length; here, only the stamens are seen which from all sides bend inwards toward the centre of the flower so that the central part is quite hidden. A longitudinal section shows best how they unite closely in a conical fashion with the upper ends more or less bent in- wards and with all the anthers gathered above the centre of the flower. (Fig. 37, А, В). The outermost stamens are the longest, their filaments become, from a thinner, entirely smooth base, gradually thicker and also covered with small Rosaceæ. 97 warts. Immediately below the anther the filament becomes suddenly thinner. The nearer the stamens approach the centre of the flower, the smaller they are and the more bent inwards; the smallest of all have short much bent filaments the upper thinner part of which is however аз long or even longer than that of the outermost stamens (Fig. 37, D, С). As the stamens are so densely and closely united above the centre of the flower it appears as И all smaller and weaker insects must be entirely excluded from access to the honey secreted by the inner side of the hypanthium. I think that only strong insects such as humble-bees and bees will be able to force a way, but perhaps the slender proboscis of the butterfly may discover a way down between the filaments.” “Longitudinal sections show moreover that the pistils are lying hidden just below the stamens. The styles, which are slightly thickened towards the apices, are erect, and the irregularly lobed stigmas (Fig. 37, К) are thereby placed immediately below the anthers, and partially even in direct contact with those of the innermost, short stamens. The flowers are slightly -protandrous; the outermost stamens open their anthers very early even at the time when the petals still stand almost erect, and before the stigmas are quite mature, though large. But the stigmas evidently be- come functional very quickly after the outermost stamens, because I have seen the stigmas covered with quantities of germinated pollen before the anthers of the innermost sta- mens have opened. Cross-pollination must be possible; but F also believe that pollen must almost certainly fall from the anthers of a flower upon its own stigmas, when an insect thrusts its proboscis down between the stamens, and this evidently happens the more easily because the filaments are rough and must therefore offer more resistance to the intrud- ing proboscis of an insect, and thus be shaken so much more effectively. It will then depend on whether the pollen XXXVIL. 7 98 Кмор JESSEN. from the flower of another individual germinates or deve- lops more quickly than does the pollen of the same flower. But in case cross-pollination does not take place, self-pollina- tion is inevitable ultimately when the short stamens dehisce.” “Т know no other Rubus-species in which self-pollination is so well secured as in the present species, and this must undoubtedly be regarded as an adaptation of the species to the inclemency of the weather, to which it is probably often exposed during its flowering period. The species which comes nearest to it is, according to my observations, Rubus saxatılıs, which as is well-known extends even to the North Cape and East Finmark and above the wood-limit in the mountains of Norway and in the Alps.” Poprtus (1. с.) finds that the flowers are without scent and secrete very little honey; he also records a proterandrous stage before the homogamous. He writes that in foggy or overcast weather the stamens bend inwards so that the an- thers and filaments entirely cover the pistils. Numerous in- sect-visitors are mentioned. Norman records that in the districts of Arctic Norway where the plant sets fruit rarely or only sparingly, the corolla is largest and most showy, frequently with 6—8 petals which are 13—14 mm. long and 8—10mm. broad, while in districts where the plant bears fruit frequently the corolla is often much smaller, with petals the length of which may be reduced to 9mm. and breadth to 5.5 mm. In Arctic Norway Rub. arcticus flowers in the end of June and has ripe fruit in August (Norman). — Fruit- dispersal probably by birds. According to the microscopical investigation of Rubus chamemorus, Rub. saxatilis and Rub. arcticus mentioned above, the first-named differs in several points from the other two, viz. (1) in the structure of the absorbent roots, (2) in the structure of the aerial shoot, (3) by the fact of its leaves bearing stalked Rosaceæ. 99 glands and having mucilaginous inner-walls in the epidermis of the upper side, and lastly (4) in the number of the vascular bundles in the leaf-stalk, where Rub. chamemorus has five and the others three. According to Fritscu (1. с.) special import- ance should be attached to the last feature when the question is to distinguish anatomically species and groups of species among the Rubi. Thus, the anatomical investigation together with the morphological appears to support Еоске! who sepa- rates Rub. chamæmorus from the other herbaceous Rubi and places it in a distinct subgenus. Alchimilla alpina L and A. færoensis (Lge) Buser. Lit. MÜLLER, 1881. Jonsson, 1895. Norman, 1895. Кмотне, 1902. SCHROTER, 1904. HoLLsTtein, 1907. As these two species agree in many of the points which have been investigated they are here treated jointly for bre- vity’s sake. A. alpina extends north of the Polar circle in Scandinavia only; it is found also in Greenland, Iceland, the Færåes, the mountains of Central Europe and in the Urals. А. feroensis has been found only in the Fzeröes and in Iceland. The habitats are more or less dry, but A. alpina at any rate can live in widely different conditions: in grassy localities, on cliffs, on heaths, rocky flats, in willow copses, etc. The alcohol material was gathered in Greenland, Iceland and the Feerges. A. alpina has a primary root of rather long duration from which the shoots spread out upon the ground with apices turned upwards. Usually the shoots do not reach. any con- siderable length; they branch freely and are fixed to the ground by adventitious roots. Looser or denser tufts are formed. The rhizome in A. fergensis is thicker than that in A. alpina; 1 Abh. d. Naturw. Ver. Bremen, IV, 1874 and Synopsis Ruborum Germaniæ. Bremen, 1877. Tr 100 Knup JESSEN. it is vertical or obliquely ascending, branches less freely, and the branches are considerably shorter, but they develop ad- ventitious roots as in A.alpina. Undoubtedly, neither species is entirely dependent on the primary root for the continu- ation of life; vegetative propagation takes place. Fragments of the dead leaves persist for a longer time, but in A.alpina at least the shoots may ultimately get rid of them. The leaves are placed in a rosette, but in luxuriant spe- cimens of A. alpina the internodes may be somewhat elon- gated. The first leaf of the shoots — both of the floral and the vegetative — is a scale-leaf in the form of a closed sheath; also the other leaves have closed sheaths which are ruptured on the vegetative shoots owing to growth in thickness. In Ice- land Jonsson found A. alpina with green leaves in the winter. The old leaves had not fallen off, but formed a sort of water- proof roof over the shoots (1. c.). In the Botanic Garden in Copenhagen the leaves which remain green through the winter are, in both the species, very small, few in number and not quite expanded. The summer-leaves wither in the autumn and a few new leaves may grow out, but strictly speaking the plant is scarcely winter-green. The shoot-apices are pro- tected by the closely folding leaf-sheaths; scale-leaves are not developed. In the Botanic Garden in Copenhagen there is no fixed flowering-period, but new floral shoots are constantly being developed as in the A. vulgaris-forms, though not to the same extent as in the latter; and not until the low temperature of winter sets in does the development cease. In December all transitional forms are found, from flowering shoots to quite young buds which are hidden by the leaf-sheaths of the terminal bud. I am not prepared to state whether a similar proleptic development takes place in Arctic Norway and in Greenland. Norman and LANGE record that the flower- ing in these regions ends on September 20 and in August Rosaceæ. 101 respectively. In Iceland Jonsson (I. се.) found no flower- buds on March 26, but, as Fig. 38, А shows, young inflores- cences are at any rate to be found in Greenland in August. Anatomy. The adventitious roots are triarch-tetrarch. The central cylinder is surrounded by a suberized endodermis. The outer wall of the epidermis, especially in А. alpina, is some- what thick (2.5), but the skin is functional for a certain period only, after which it is thrown off; this is due to a struc- tural feature which is shown in Fig. 38, В and which was especially pronounced in A. fergensis. The exodermis consists of large cells, the thin radial walls of which are ruptured sooner Fig. 38. Alchimilla alpina: A, Longitudinal section through a young bud containing the primordium of an inflorescence (Greenland; 1. 8. 1885; 1/4). С, Portion of a transverse section through the periderm of a rhizome (Greenland; 15/,); k, layer of cork; р, phelloid-cells. — Alchimilla feroensis : В, Portion of a transverse section of a young adventitious root of the first order; lp, epidermis (*4°/;). or later, and the following layer of the cortex which has fairly thick walls and gradually becomes corky, functions then as epidermis till the periderm is formed. The primary epidermis is thrown off before the root in the zone in question has begun its secondary growth in thickness. In the primary cortex in the roots of A. alpina fungal hyphæ have been found. Periderm is probably formed in the pericycle; it is not developed until a continuous woody portion is produced in the central cylinder; it consists of alternating series of layers of 1—3 layers of cork-cells and phelloid-cells; it contains rather large intercellular spaces which are however often absent 102 Кмор JESSEN. from the inner side of the cork-layers. The phelloid-cells contain starch-grains which are found especially along the inner walls. The periderm may constitute as much as + of the radius of the root. Medullary rays are absent from the wood. There are distinct annual rings; the oldest roots I have seen were 5 years. | The central cylinder in the rhizome is surrounded by a distinct endodermis. The primary bundles are considerable in number and are more or less confluent. The pith which is of simple parenchyma lives in A. alpina as long as the shoot, but in A. feeroensis it apparently dies at an earlier period. In the latter species the pith constitutes a larger portion of the diameter of the shoot than in the former; it is the chief reser- voir of food-material and in the autumn is found crowded with starch. Rhizomes of A. alpina collected in Iceland in March contained, on the other hand, almost no starch, and shoots of both species collected in the Botanic Garden in Copenhagen on Jan. 5 contained only a small quantity; on the other hand, they contained large quantities of fat. In the first period of vegetation of the shoot a continuous cambium produces a small quantity of wood. Medullary rays are absent from the latter (cf. HOoLLSTEIN who has stu- died A.alpina). In A.alpina the diameter of the vessels was larger in specimens from Copenhagen than in those from more northern regions. The periderm is formed by the division of the outermost layer of the pericycle and is developed even in the first period of vegetation of the shoot. 2—8 layers of phelloid-cells alternate with one layer of cork-cells. But in the Botanic Garden in Copenhagen the periderm in these two species contained only scattered cork-cells; in such plants № func- tions almost solely as a depository of food-material while the protective layer proper consists of dead masses of leaf- sheaths, primary cortex and the older layer of the periderm. Rosaceæ. 103 Fig. 38, C shows a transverse section of the periderm of А. alpina; it contains large intercellular spaces. During summer . the starch-grains are found especially along the inner walls of the phelloid-cells; in the autumn the cells are entirely filled by them. The primary cortex is ruptured owing to secondary formations, but even in the second summer it may be found alive. Bast does not occur. The structure of the floral-shoot is very much like that described by HoLLSTEIN from the Dauphiné-Alps. A collenchyma, 1—2 layers thick, occurs, and within the endo- dermis 2—3 layers of bast unite with the secondary wood, which consists of several-layered strong-walled tracheids, so that the groups of sieve-tissue are entirely surrounded by stereom. The pith dies away in the centre. The leaves are from 200 to 220% thick, and generally somewhat thicker in A. alpina than in A. feroensis. The epidermis of the upper surface is glabrous in the former and slightly hairy in the latter; it is high, and the outer wall in A. alpina is about 6y and in A. feroensis 4—5 y thick, while the outer wall in the epidermis of the lower surface is about 3.3 и and about 24 thick respectively. Stomata serving as respiratory organs are absent from the upper surface. The lower surface is densely hairy and here the stomata occur abundantly and are somewhat sunk below the level of the surface (Fig. 39, В). Кмотне (1. с.) illustrates by a figure of A. alpina how the hairs are appressed upon the lower surface of the leaf, and he states that this is not wetted. The palisade parenchyma consists of a well-marked layer which constitutes about one-half of the thickness of the mesophyll. The proportion between the leaf-thickness and the thickness of the palisade tissue is in A. alpina about 2 : i 2. : = and in À.færoensis about ~*. The intercellular spaces in the spongy parenchyma are not large, but sections 104 Кмор JESSEN. often show large slit-hke lacunæ on the dorsal side of the leaf. Stereom is absent from the leaves. Fig. 39 (A, В, С, D, Е). At the apex of the leaf-teeth there is a slightly developed epithema, below which tracheids terminate; water-stomata occur upon the upper surface of the leaf-teeth on a limited area. The three bundles in the leaf-stalks of these two Fig. 39. Alchimilla alpina from Greenland. A, Transverse section of leaf (1#/,). В, Stoma (?%/,). С, Epidermis of the upper surface of the leaf (!%/,). D, Epidermis of the lower surface of the leaf with hairs and stomata (?%/,). Е, Section of spongy parenchyma parallel to the surface (2%/,). — Alchimilla feroensis. Е, Transverse section of leaf-stalk (2/,); ch, collenchyma; e. endodermis ; 7, leptome ; x, xylem; m, ‘‘pith’’. species are concentrie in structure. This peculiar feature was first demonstrated by С. de CANDOLLE! in a great many Alchimilla-species, among others also in A. alpina and A. minor, and afterwards Bouyeurs?, without mentioning de CANDOLLE’S paper, has published a short note upon the same subject. Buser? has been studying de CANDOLLE’S 1 Bull. de l’herbier Boissier, T. I, 1893. 2 Act. soc. Linn. de Bordéaux, 1900. 8 Bull. de Vherbier Boissier, T. II, 1894. Rosaceæ. 105 results and shows that the species which have the concentric structure of the bundles of the leaf-stalk are found in the “region montagneuse” of the Alps, while the subnival Alpine species have the common collateral structure of the bundles. Fig. 39, F shows a transverse section of the stalk of A. fæ- roensis. As is well-known 5. Моввеск* has demonstrated par- thenogeny in a whole series of Alchimilla-species including A. alpina in which development of the embryo frequently takes place even while the flower 15 in the bud-condition. Murgecx did not find а single pollen-grain capable of germin- ation in А. alpina. Nor does А. færoensis appear to develop pollen-grains capable of germination. In the insignificant yellowish-green, apetalous flowers there is around the gynaeceum a broad fleshy disk which secretes honey. H. MüLter (1. с.) has described the structure of the flower of A. alpina. Alchimilla vulgaris L. Lit. H. Murer, 1881. Leıst, 1889. Jönsson, 1895. Nor- MAN, 1895. SyLvéN, 1906. In Arctic Norway this collective species grows in widely different localities, having been found both in dry and in damp places, in open spots and in birch-woods (NORMAN). In Greenland it grows in damp grassy fields and in willow- copses. In my material the following elementary species occurred: А. acutidens Bus., glomerulans Bus., and minor Huds. АП these species extend to the north of the Polar cirele in Green- land and Europe; they are also found т Iceland, and А. acut- idens and А. minor occur in the Færøes; they are found, in 1 Parthenogen. Embryobild. in d. Gattung Alchimilla. Act. Univer. Lundensis Т. 36, 2. LA) 106 Кмор JESSEN. addition, in northern temperate Europe”. — The alcohol material is from Greenland, Iceland and the Færges. The young seedling of A. vulgaris is according to SYLVEN a rosette, and in older plants also the leaves are always placed in a rosette even if the internodes are often somewhat elon- gated. The rhizome is vertical or oblique or almost horizon- tal, and in the last case there may be a slight wandering. It does not branch so freely as in A. alpina. Even during the year in which it germinates adventitious roots arise, and these are afterwards developed abundantly. The floral axes are lateral, and there occurs, at any rate in several of the elementary species growing in Denmark, a similar proleptic development of the floral shoots as in A.alpina in Copenhagen, only the development is more vigorous. It appears to be external factors (e. g. fall of temperature) rather than internal which limit the flowering period; we may be justified in presuming that something similar happens also in the Arctic regions; NorMAN records that in Arctic Norway A. vulgaris may flower till September 20. The plant lives through the winter in a more or less green condition (WARMING, 1884; J6NSSON and SyLvEN); but we may also find it without any green leaves (in Denmark) and as in A. alpina the summer leaves wither in the autumn. The shoot-apex is protected by the sheaths of the foliage- leaves; scale-leaves are not developed. Anatomy. — Roots of A. minor and A. acutidens have been investigated; in structure they resemble essentially those of A. alpina and A. færoensis (cf. Fig. 38, В). Mycorrhiza has been found in A. acutidens from Iceland. The rhizomes are on an average thicker than those of A. alpina; this is especially due to the fact that the pith is 1 See Harald Lindberg. Die nordischen Alchimilla vulgaris-For- men und ihre Verbreitung. Acta Soc. Scient. Fenn. XXXVII. Helsingfors, 1909. Rosaceæ. ‚ 107 largely developed. Otherwise, the structure of both the rhizomes and the stems is practically similar to that in А. alpina and A. færoensis. The anatomy of the leaves (Fig. 40) is the same in all these three species, and closely resembles that of the two species mentioned above; the leaves are, however, somewhat more mesomorphic than those in A. alpina. The upper surface of the leaf is glabrous, or in A. glomerulans slightly hairy, the lower surface is slightly hairy. The outer wall of the epidermis Fig. 40. Alchimilla vulgaris. A. glomerulans. A, Transverse section of leaf. С, Epidermis of the lower surface of the leaf. E, Section of spongy parenchyma parallel to the surface. — A. minor. B. Epidermis of the upper surface of the leaf. D, Section of spongy parenchyma parallel to the surface. — A. acutidens. F, Stoma from the epidermis of the lower surface of the leaf (4, about 220/,; В, С. D and Е, about !%/,; Е, about 15]; A, С and Е from West Greenland, В and D from Dr. Louises © in East Greenland, F from Eskefjord in East Iceland). of the upper surface is about 2y thick; the radial walls are slightly undulating. Stomata occur scattered upon the upper surface and, as upon the lower, they are situated below the level of the surface and on an average somewhat more so than is the case in the two species of the Alpina-group (Fig. 40 A). The radial walls of the lower epidermis are somewhat more strongly undulating than those of the upper. A layer of well-marked palisade-cells occur and sometimes another layer consisting of more irregularly formed cells. LEIST records only one layer of palisade-cells. The proportion between the leai- 108 Кмор JESSEN. thickness (about 190) and the thickness of the palisade-tissue is about 3. Below the stomata of the upper surface air-spaces are produced in the same manner as described for Sibbaldia procumbens (p. 58). The spongy parenchyma is rather loose in structure and contains, as shown in Fig. 40, D, besides large intercellular spaces numerous very small ones. The cells are branched. The leaf-teeth contain a slightly developed epithema; the water-pores occur on a limited space upon the upper surface; these structures have been mentioned by REINKE}. The structure of the leaf-stalk is as in A. feroensis and A. alpina (see Fig. 39, F). Моввеск has demonstrated parthenogeny in A. minor, which produces only a very small quantity of imperfect. pollen. — In the flowers of A. glomerulans from Greenland a great many of the anthers had fallen off; in those which were still attached I found no fully developed pollen. The flowers of A. vulgaris, which in structure almost agree with those of A. alpina, have been described and figured by H. Miter 1873” and 1881 (1. c.). Sorbus americana Willd. Lit. WARMING, 1887. RosENvINGE, 1896. The alcohol material is from Ivigtut in Greenland. This Sorbus which bears much resemblance to S. aucu- paria is a small tree with an erect, straight growth; it does not branch much, but several stems may proceed from the same root. The largest trees hitherto found in Greenland measured about 4 metres in height and the stems were about 9 cm. thick. These trees were 56 years old (ROSENVINGE). Generally, however, it is lower in growth. It occurs in south- 1 Secretionsorg. Pringsheim Jahrb. Bd. X. 1876. 2 Die Befruchtung der Blumen durch Insekten. Leipzig. Rosacex. 109 ern Greenland, in the valleys within the fjords, dispersed through Betula and Salix copses; it grows also in N. America. The leaves are smooth. The epidermis shows strong cuticular stripes, strongest upon the upper surface, and upon the lower surface stronger along the margins (Fig. 41, А, С, D). The outer wall of the upper epidermis is about 4.5 и Fig. 41. Sorbus americana. A, Transverse section of the leaf. B, Transverse section of the epidermis of the upper surface showing a cell with highly mucilaginous inner wall. C, Surface view of the upper epidermis of the leaf. D, Section parallel to the surface of the lower epidermis of the leaf; the cuticular folds of some of the cells are shown as in C, usually each epidermal cell has its own sys- tem of folds. Æ, £ection of the spongy parenchyma of the leaf parallel to the surface. — A, В. С, D, Е “°/,. Ivigtut in Greenland. thick, of the lower about 2.5 и. Some of the cells in the upper epidermis have highly mucilaginous inner walls, and such cells are often particularly large (Fig. 41, B), compare p. 57. Stomata occur, with the exception of the water-stomata on the leaf-teeth, only on the lower surface: they are on a level with the surface. There are two layers of palisade-tissue of 110 Кмор JESSEN. which the upper consists of high, narrow cells which enclose large intercellular spaces. The cells in the second layer are often very irregular; this layer is absent above the smaller bundles in the marginal portion of the leaf. The proportion between the leaf-thickness (about 220 4) and the thickness of the palisade tissue is about 2. Above the larger bundles the assimilatory tissue is replaced by vein-parenchyma. The spongy parenchyma is very loose in structure, the cells are slightly or shortly stellately branched (Fig. 41, В). In the flower of Sorbus americana 1 have found about 15 stamens, of which the ten episepalous were only a little taller than the five epipetalous. In Sorbus aucuparia there are twenty stamens!. — In flower-buds which are on the point of open- ing the stamens are found to be bent inwards over the 2—3 styles. The anthers are closed, but the stigmas are highly papillose and apparently ready to receive the pollen. When the flowers expand the stamens gradually straighten and bend outwards towards the petals. I have not found open anthers which were in contact with the stigmas, but they are seen to be bending over the latter some time after they have dehisced, and the pollen must almost certainly fall upon the stigmas. In this homogamous stage the stigmas were found to be covered with germinating pollen. In Greenland the flowering period occurs in July and August, but according to LANGE the plant does not flower at higher altitudes. RosEnvInGE thinks that the rowan sets fruit regularly in favourable localities, and he gives some instances of its having probably been dispersed by the ageney of birds. 1 Cf. fig. 422 in Eug. Warming: Froplanterne (Spermatofyter), København, 1912. Rosaceæ. 111 Summary. A. Account of the growth-form of the species. Г FE IT. Microphanerophyte: Sorbus americana. Tree, low in growth. Lives through the winter without green leaves; the shoot-apex is covered by smooth bud-scales. The leaf-structure is mesomor- phic. Chamephytes: Dryas octopetala and D. integrifolia. Dwarf shrubs with a primary root of long dura- tion. The prostrate long shoots can be somewhat indi- vidualized. In autumn the leaves die with the excep- tion of a few half-unfolded ones at the shoot-apices. The apex of the shoot is covered by the sheaths of the young and dead foliage-leaves. The leaf-structure is xeromorphic. Hemicryptophytes: a. Dwarf shrubs or undershrubs. Potentilla palustris is а dwarf shrub (see р. 9) which may be a hemicryptophyte, а chamæphyte! or a helophyte. The shoots with elongated internodes live through the winter without green leaves and with their apices covered by the sheaths of the foliage- leaves and by scale-leaves. Much wandering takes place, as also vegetative propagation. The leaf-struc- ture is slightly xeromorphic. Pot. tridentata is а creeping rosette-forming under- shrub. Some of the summer-leaves may live through the winter: there is no closed winter-bud. The leaf- structure is xeromorphic. 1 See also Vahl, Growth-forms of some Plant Formations of Swed- ish Lapland. Dansk bot. Arkiv, Bd.I, No. 2, Köbenhavn, 1913. 112 Кмор JESSEN. These two species have terminal floral-shoots while all the other Potentillas mentioned in this paper are monopodial. b. Perennial herbs. 1. Spotbound are Potentilla pulchella, Pot.nivea, Pot. Vahliana, Pot. emarginata, Pot. maculata, Sibbaldia pro- cumbens, Alchimilla alpina, А. færoensis and A. vulgaris. Features common to them all are monopodial, rosette- formed assimilatory-shoots and lateral floral-shoots. The Potentillas, Sibbaldia and Alchimilla alpina have primary roots of long duration, and they distinctly belong to the type “Radix vel rhizoma multiceps”; the other Alchi- millas branch less freely, and their primary root un- doubtedly dies rather early. They live through the winter with their shoot-apices protected by the sheaths . of dead and living leaves. To evergreens proper only Pot. maculata can be referred. It is however probable that other species also may live through the winter with a few green leaves, and as regards this point local differences no doubt take place. In species such as these which have no true winter-buds it is difficult to draw a sharp limit between the winter- green and the not winter-green habit, since the leaves of the following summer may grow out slightly in the preceding autumn. This uncertainty is no doubt the reason of the often contradietory records. — The leaf- structure would most nearly be characterized as mezo- morphic in Pot. emarginata and Alch. vulgaris and as more or less xeromorphic in the others. 2. Wandering perennial herbs are Pot. anserina and the Rubus-species. Pot. anserina has a rosette-formed, monopodial main shoot and lateral floral axes which behave as runners. Vegetative propagation takes place by the Rosaceæ. he individualization of the rosette-formed lateral shoots of the runners. Food-storing roots often occur. It lives through the winter without green leaves and with its shoot-apex covered by scale-leaves. It is typical for the leaf-structure to be slightly xeromor- phic. As features common to all the Rubus-species may be mentioned that they do not form rosette-shoots, that the floral shoots are terminal; and that they live through the winter without green leaves! and with their winter- buds covered by scale-leaves. Rubus arcticus wanders by means of its roots, and Rub. saxatilis by the develop- ment of certain shoots into aerial runners the apex of which is developed into a winter-bud situated at the surface of the ground. Rub. chamæmorus has scale-leaf- bearing underground runners. The leaf-structure in the two first-named is mesomorphic and in the last-named slightly xeromorphic. There are at least two points connected with the shoot-structure of several of these Rosacee which appear to be favourable to plants living in Arctic climates, as regards the way the latter influence the plants, e. ©. by desiccation (physiological desiccation) and by the shortness of the growth-period. A great many of the species, and more particularly those that are true Arctic (see p. 121), are more or less highly cæspitose, and may have in addition very closely packed shoot-masses (closest in Pot. Vahliana); it appears certain that by this method of growth the amount of transpiration is lessened, in that the outer shoots of the tuft protect the inner from the wind, and the shoot-apices are surrounded and protected by the dead masses of leaves. 1 See р: 93. XXXVI. 8 114 KNUD JESSEN. From Kyeııman’s classical work “Ur polarväxt- ernas ПР” it has long been well known that the Arctic summer begins very suddenly and that a great many Arctic plants live through the winter with leaves which need only elongate slightly in order to attain full power of assimilation. The majority of the true Arctic species mentioned here live through the winter in this way; some even have small, green, half-opened leaves (Dryas). They have moreover no closed winter-bud sur- rounded by scale-leaves (except, е. g. Rubus chamæmorus) and with but very little room for the new foliage- leaves; but the “buds” are open, and the new organs — leaves and flowers — have ample room and grow out quickly, as soon as ever external conditions permit their doing so, and begin work in the short summert. . In the following species the flowers are formed during the = As regards light-conditions in Arctic regions I would remark, that from Porsild’s “Actinometrical observations from Greenland” (Arbejder fra den danske arktiske Station paa Disco, No. 4, Medd. om Grönland, XLVII, Köbenhavn, 1911) it is seen (from Tables II and III) that the mean light-total for the months May—August is about the same on Disco as in Denmark (An- holt). On comparing the assimilation-conditions of the plants on Disco and Anholt it is seen that the Greenlandic plants must, besides a lower temperature, also suffice with less light for their summer work than the Danish, the latter having longer time at their disposal. So far as I know it is an open question what use the plants make — if any — of the light of the midnight sun. The investigations made by Curtel (Rev. générale de botanique, Tome II, 1890) and Porsild (Meddel. om Grönland, ХХУ, 1902, pp. 25 and 26) regarding this point gave contradictory results. Besides Porsild, Wiesner has also studied the light-conditions in Arctic regions (Norway and Spitzbergen) and he finds by com- parison with the Alpine regions of Central Europe that these are characterized not only by a greater light-intensity, but also by a greater amount of light in the period of vegetation. (Beitr. z. Kenntniss des photochem. Klimas im arct. Gebiet. Denkschr. math. nat. CL 4. К. К. Ak. Wien. LXVII, 1908, and Beitr. 2. Kenntniss des Lichtklimas von Tromse, etc. Tromso Museums Aarshefter, 24. Tromso, 1901). Rosaceæ. 115 summer previous to the year in which they expand, viz. Alchimilla alpına, Alch. vulgarıs, Dryas octopetala, Potentilla emarginata and Rubus saxatılis, all of which flower in June. It is probable that the great majority of the species behave as these, with the exception however of Pot. palustris in which the young flower-buds have not been observed until the spring of the year in which they expanded. . The absorbent roots often present anatomical peculi- arities, for instance, a thin epidermis which is usually more or less collapsed is found in the majority of the Potentillas (anserina, emarginata, maculata, nivea and pulchella), and in addition in Sibbaldia and Rubus saxatilis. In Rubus chamemorus it is destroyed early. A stronger epidermis occurs in Pot. palustris, where it is very small- celled, and in the Alchimilla-species where there is a ten- dency to throw it off entirely, because the thin-walled exodermis ruptures. In Pot. emarginata, P. nivea and P. pulchella the radial and the outer walls of the exodermis are very highly thickened, and this cell-layer is also very strong in Rubus chamæmorus. In the Alchimilla-species the cortical layer within the thin exodermis is fairly strong. In Dryas, Rubus arcticus and В. saxatilis the radial walls — and in the two last species also the outer walls — of the cortical layer outside the endodermis are very strong. Ectotrophic mycorrhiza occurs in Dryas; and endo- trophic mycorrhiza has been found in Pot. anserina, Pot. emarginata, Pot. nivea, Pot. palustris, Pot. pulchella, Sibbaldia, Alchimilla alpina, Alch. acutidens and Rubus arcticus. All the Rosacee mentioned here have probably mycorrhiza, except Rub. chamæmorus. . A continuous cambium is found in the perennial portion of the shoots in all the species. According to the nature of the tissue produced by the cambium the species may 8% 116 Кмор JESSEN. be divided into two main groups. One, more herbaceous, in which the secondary formations consist of groups of vessels rich in parenchyma and separated by broad medul- lary rays; the Potentillas with the exception of Pot. palustris and Pot. tridentata belong to this group. Another, more lignified, in which the cambium produces a ring- shaped woody portion; to this belong the rest of the species. In Dryas, Potentilla palustris., Р. tridentata, Sibbaldia and Rubus are found medullary rays, one to a few cells broad; medullary rays are absent from Alchimilla except above the leaf-trace bundles (“Blattspurstränge”). In Alch. al- pina and especially in Alch. vulgaris and Alch. feroensis the secondary wood-formations constitute only a compara- tively small portion of the shoots, and these are then more herbaceous. Bast occurs regularly only in Dryas. In Dryas, from which, as already mentioned, an endodermis is absent, cork is formed in the inner part of the cortex; in all the others which have an endo- dermis, a periderm is probably always formed in the outermost layer of the pericycle (cf. SOLEREDER!). Dryas, only, has a scaly bark (“Schuppenborke”’) formed of cen- tripetally developed phellogens which develop only cork- cells without intercellular spaces. In the other species, on the other hand, the periderm consists of one layer of cork-cells alternating with usually three layers of phelloid- cells. It contains large intercellular spaces, which are however absent from the inner side of the cork-layers in the great majority of the species. SOLEREDER (I. с.) who also describes the periderm has found intercellular spaces in the phelloid in Pot. palustris and Poterium sanguisorba; in his opinion this suggests a resemblance to aérenchyma and he regards it as an adap- 1 Anatomie d. Dicotyled. Stuttgart, 1899, p. 348. Rosacee. 117 tation to life in bogs in these species. It is true that respiration in the inner part of the shoots is facilitated by the presence of the intercellular spaces in the peri- derm, but the comparison with aërenchyma does not appear appropriate when we consider that species from extremely dry habitats, such as Pot. tridentata, Pot. nivea, Alch. alpina and others have a quite similar structure of the periderm. A contribution to the explanation of the structure of the periderm in these Rosaceae may perhaps be found in the fact that it no longer functions only as a protection from excessive transpiration, as for instance in Dryas, but that it has also developed into a very important starch-storing organ; at any rate, a peri- derm with this function requires an ample supply of oxygen so that the necessary condensations and hydro- lyses can take place in it. — It may be mentioned that a similar periderm also occurs in the roots of the species which have been investigated, with the exception of Dryas and Rubus arcticus. As regards this point the reader is referred to MARTHA BUNDING! who has investigated the root-cork in many Rosaceae and finds as а rule intercellular spaces in the herbaceous species, but not in the woody or shrubby (cited from FREIDENFELT). . Special food-storing organs are absent except in Pot. an- serina which frequently has spindle or club-shaped food- storing roots of a peculiar anatomical structure. Other- wise reserve food-material is found in the root (many of the species have primary roots of long duration and con- siderable size) and in the perennial portions of the shoots. The tissues which contain reserve food-material are the = Structure of the Cork Tissues in Roots of some Rosaceous genera. Publ. of the Univer. of Pennsylv. New Ser. No. 5. Contrib. fr. the Bot. Laborat., Vol. 2, 1898. (Unfortunately, the paper has not been available to me). 118 1 Кмор JESSEN. phelloid, the perieycle, the medullary rays and the wood- parenchyma; and in the pith, especially the peri- pheral part. The reserve food-material consisted especially of starch in all the species which have been investigated in regard to this point; but in several of these species (Dryas, Alchimilla, Pot. emarginata, Pot. nivea, Pot. palustris, Sib- baldia, Rubus saxatilis) it has been demonstrated that during winter the starch disappears more or less com- pletely and is replaced by fat.! Probably during autumn the starch is converted into fat, and then during spring the fat is again converted into starch. Similar conditions are well-known as regards several trees. It is apparently a common occurrence among perennial herbs and dwarf- shrubs that during winter the starch is replaced by fat, the amount of fat increasing from October or November and reaching its maximum in January or February, then during spring it disappears entirely in temperate regions; this is at least the case as regards a great many Danish species. On the other hand, there are a number of species in which no fat is visible during winter; as an example, the food-storing roots in Pot. anserina may be mentioned. While, as mentioned above, fat is absent from the deposi- tory organs of the herbs during summer in Denmark, I have, in P. nivea, P. emarginata and especially in P. pul- chella (from N. E. Greenland) found fat in rhizomes and roots throughout the whole summer; this is probably due to the comparatively low temperature during summer in Arctic regions. . A great many of the species have leaves which in one way or other illustrate xeromorphy. Transpiration is lessened (a) by a hairy covering: in several of the species the under side of the leaf is densely matted or hairy, espe- Indicated by use of Sudan 111. Rosaceæ. 119 cially in Dryas, Pot. nivea, Alch. alpina and others; (b) by a revolute leaf-edge, which is found especially in several Dryas-forms and in Pot. pulchella {. humilis; (с) by sunken stomata, which are found in Alchimilla, Pot. palustris, Pot. tridentata, and Sibbaldia, while in several of the spe- cies, which have a dense hairy covering upon the under side of the leaf, the stomata there project slightly above the leaf-surface. For the rest, stomata occur on both surfaces of the leaf in almost equal number in Sibbaldia only; of the other.species, stomata are entirely wanting upon the upper side in the following: Alch. alpina, Alch. feroensis, Dryas, Pot. Vahliana, Rubus and Sorbus. Lastly, transpiration must also be lessened by (4) a thick outer wall in the epidermis, and here Dryas and especially Pot. tri- dentata may be mentioned; and (e) by the mucilaginous covering of the inner wall in the upper epidermis (pectin- mucilage). This is found in Dryas, Pot. maculata, Pot. nivea, Pot. palustris, Pot. pulchella and Pot. tridentata, Sib- baldia and Sorbus. Very often not all the cells are muci- laginous; in Pot. tridentata besides the inner wall the outer wall is also somewhat mucilaginous and in Pot. maculata a few mucus-cells were also found in the lower epidermis. The importance which should at any rate be attached to a cellulose-mucilaginous covering of the inner walls of a leaf-epidermis is among others mentioned by VOLKENS?, who is of the opinion that such mucilage in the cells does not function as a water-reservoir, but checks transpira- tion in the same manner as does a gelatinous layer which is spread over a slightly evaporating surface of water (compare also WARMING, 1887). The mesophyll in the leaves of the species which have been investigated has a very ordinary structure, and it shows only to a slight degree such adaptations to 1 Die Flora der Aegypt.-Arab. Wüste. Berlin, 1887, pp. 48—45. 120 Knup JESSEN. external conditions as are evident-in the structure and hairy covering of the epidermis of at least several species. The palisade-tissue in almost all the species consists of 2—3 layers of cells; in the Alchimilla- species only, there is usually a single layer of very long palisade-cells. In these the relation between length and breadth is as 4 or 5 to 1; the same relation is also found in the upper layer of palisade-cells in Sorbus; but in the remaining species the palisade-cells are only 2—3 times as long as they are broad; they are often somewhat barrel- shaped. In all the species the relation between the thick- ness of the leaves and the thickness of the palisade-tissue is about 5; as regards this point some variations have been found to occur in various species, but such special records of the nature of the habitats as are necessary in order to be able to understand such variations have unfortunately not been at my disposal. In the majority of the species there are large intercellular spaces in the palisade-tissue. In Pot. tridentata these are however small, and as regards the mesophyll, this species has the most xeromorphic leaf, while the least xeromorphic is that found in Rubus saxatilis and R. arcticus. The transpiration-parenchyma is throughout of loose structure and consists of short, branched cells. Large slit-like lacunæ occur commonly within the subepidermal layer on the lower surface of the leaf (e.g. in Dryas, Pot. pulchella, Alch. alpina). The species may be divided into two groups accord- ing to the anatomy of the leaves; viz. into those species which have a mesomorphic leaf-structure and those which have more or less well-marked xeromorphic leaves. To the first group belong Rubus saxatilis and Rub. arcticus, Alch. vulgaris and Alch. færoensis, and perhaps also Pot. emarginata and Sorbus americana. The remaining species Rosaceæ. 121 are referable to the other group; particularly, Dryas, Pot. tridentata, Pot. nivea and Pot. pulchella (especially f. hu- milis). Some understanding of the leaf-structure in the different species may be gained by а consideration of the habitats. Species of the first group, with the exception of Pot. emarginata, are generally associated with relatively favourable soil, moreover, none of them — again with the exception of Pot. emarginata — are true Arctic species. The majority of the species in the other group are natives of; dry localities such as fell-fields, rocks and heaths, and a great many of them are High Northern species, while two of them are bog-plants. The few species which may be regarded as more or less evergreen belong to this group. It does not appear, however, that the leaf-structure is governed thereby to any special degree, perhaps, how- ever, with the exception of Pot. tridentata; it is tempting, more particularly on account of the leaf-structure in this species, to conclude that it is an evergreen; but beyond the surmise that a few of the summer leaves may un- doubtedly live through the winter I am prepared to state nothing. In Dryas the outer wall in the upper epidermis of the leaf is somewhat thicker in the leaves which live through the winter than in the summer leaves which die away during autumn; but the common xeromorphy in the leaf-structure of this species must upon the whole be due to other reasons, because the number of the sum- mer leaves is far greater than that of the leaves at the shoot-apex which live through the winter. One of the reasons for the xeromorphy in the Dryas-leaves is un- doubtedly that already mentioned: the dryness of the soil, or, at any rate, the physiological dryness of the heaths and rocks where the plants grow. A similar con- sideration to that which holds good for Dryas may be applied to the other species which may occasionally be Тая Кмор JESSEN. = | found with green leaves during winter. Pot. maculata 15 however an exception, as it must undoubtedly be called a true evergreen; but in comparison with the other herba- ceous Potentillas its leaf-structure does not exhibit any special xeromorphic characters. Among the 14 species of Arctic Ranunculaceæ, which I have investigated!, there are only two, at most three species, the leaves of which have a xeromorphic character, viz. Thalietrum alpinum and Coptis trifolia which have leathery leaves, which in the latter live at least two years, and Ran. glacialis with a somewhat succulent leaf which has a well-developed palisade-tissue. Of the remain- ing species all the land-plants have a decidedly meso- morphic leaf-structure. By comparison of the individual leaf-tissues in the species of Ranunculaceæ and Rosaceæ which have been investigated the difference between the two families is distinctly seen. The epidermis in all the Ranunculaceæ, with the excep- tion of certain forms of Ran. acer., is glabrous or only slightly hairy. The outer wall of the epidermis of the upper leaf- surface, except in Thalictrum alpinum and Coptis trifolia, is thin, not above 24 thick, and the walls are not mucilaginous as in the Rosaceæ. In the paper cited it is recorded that chlorophyll was usually found in the epidermis of the leaf of the Arctic Ranunculaceæ which had been investigated, but I did not observe this in the Rosaceæ which I have had for investigation. The stomata in our Ranunculaceæ are always situated on a level with both leaf-surfaces, and as in the Rosaceæ, they occur especially upon the lower surface of the leaf; three of the species have, how- ever, the majority of their stomata upon the upper The Structure and Biology of Arctic Flowering Plants, I. 6. Ranunculaceae by Knud Jessen. Medd. om Grönland, XXXVI, Kobenhavn, 1911. Rosaceæ. 123 surface of the leaf, viz. Ran. glacialis, R. hyperboreus and В. reptans. The palisade-tissue is on an average some- what more strongly developed in the Rosaceæ than in the Ranunculacee, and the intercellular spaces are usually smaller. It is recorded that the palisade-tissue in the Ranunculaceæ constitutes about one-half of the thickness of the mesophyll, but in several of the species it is, how- ever, considerably less thick, е. ©. Anemone Richardsoni, Ran. РаЙази, Ran. pygmæus and Ran. nivalis. The respi- ratory tissue is on an average looser in structure in the Ranunculaceæ. Several of the Ranunculi have leaves which in their anatomical structure recall the type which is common among woodland plants in temperate regions, thus especially А. pygmeus and R. nivalis, and a similar structure is found in Rubus arcticus and Rub. saxatilis. The fact that the leaves of the Ranunculaceæ which have been investigated, are, as already mentioned, on the whole, to a greater extent mesomorphic than are those of the Arctic Rosacex is I find, partially explained — as in the Rosaceæ — by a consideration of the habitats. It is commonly stated that the Arctic Ranunculaceæ prefer damp localities, often bogs; thus Ranunculus reptans, Ran. hyperboreus, Ran. lapponicus, Ran. Pallasu and also Ran. affinis, Ran. nivalis, Ran. pygmeus, Ran. sulphureus and Ran. glacialis are mountain plants that live near the snow line in soil saturated with melting snow. Anemone Richardsoni lives by preference in copses, and Thalictrum alpinum and Coptis trifolia are met with in similar localities, or at any rate on luxuriant heaths. Ran. acer, on the other hand, often occurs on rather dry and stony soil, and in such localities its leaves are more hairy. We have, however, a peculiar instance of the fact that different plants react differently under appa- rently the same conditions. Thus, both Potentilla palu- 124 KNUD JESSEN. 51715 and Rubus chamæmorus show characters in their leaf- structure which undoubtedly should be regarded as agents serving to prevent excessive transpiration, and yet they grow in sphagnum-bogs as do Ran. Pallasii and Ran. lapo- nicus, the leaves of which I should characterize more closely as mesomorphic-hydrophilous — a fact which indicates that the adaptation of plants is not evidenced solely by morphological conditions. . In the majority of the species the flowers are fairly con- spicuous, least so in the Alchimilla-species and in Sibbal- dia, but here they are gathered into more or less con- siderable inflorescences; Sorbus has richly flowering co- rymbs, but in the remaining species the flowers are either solitary or gathered into usually few-flowered cymes. Two species have red flowers, viz. Pot. palustris, and Rub. arcticus; the following have white flowers: Pot. tridentata, Dryas, Rub. chamemorus, Rub. saxatilis and Sorbus; the rest have yellow petals. All the species undoubtedly produce honey, which is secreted by the inner side of the hypan- thium; in several species a disk is developed (Pot. triden- tata, Sibbaldia and Alchimilla). A slight perfume has been noted in a few species only. Well-marked hercogamy occurs in only a few of the species; Rubus chamemorus is dioecious, Dryas is andro-monoecious and Pot. palustris is highly protandrous. The others. are homogamous or have, at the most, a short preliminary unisexual stage. Slight protandry has thus been observed in Pot. tridentata, Pot. maculata, Dryas (some of the hermaphrodite flowers) and Rubus arcticus, while Pot. anserina, Pot. emarginata(?), Dryas (some of the hermaphrodite flowers) and Sorbus are slightly proterogynous. It must be assumed that self-pollination can take place in all species except in the Alchimillas, Rub. chamemorus and Pot. palustris. There are only very sparing notes to hand as regards direct Rosaceæ. 125 observations of insect-visitors, in true Arctic regions, to the species in question, viz. only in the case of Pot. emarginata, Pot. nivea, Pot.emarginata and Dryas, which were visited by flies and other diptera. . Many of the Arctic Rosaceæ mentioned here have a very considerable area of distribution. Eight of the species are circumpolar or almost circumpolar — of these P. pulchella and emarginata are distributed only in the Arctic zone; the others extend far southwards, and of these the two bog-plants P. palustris and Rubus chamæmorus are espe- cially confined to the low lands, while the others (2. nivea, maculata, Sibbaldia and Dryas octopetala) are confined to mountainous districts in temperate regions. Pot. tridentata and Pot. Vahliana are also true Arctic species, but their area of distribution is smaller; this is also the case with Dryas integrifolia, which in America, however, has ad- vanced somewhat southwards. Pot. anserina is almost cosmopolitan. According to Wo xr the Tertiary circum- polar land was probably the home of the original Poten- tıllas, whence they have migrated southwards. He ascribes to those groups, among others to which our Potentillas belongs, avery high age, and regards several as palaeotypic ; the majority of the species must be regarded as very old. Dryas octopetala and Sibbaldia procumbens are probably also originally true Arctic forms, which have afterwards ex- tended towards the south. The great Glacial period may have helped them, together with P. nivea and maculata, to cross the Central European low land; they are also very old species. Of these species only the following appear to have qualifications for wide fruit-dispersal: P. palustris and P. anserina by the agency of the water, Dryas by the wind and Rubus chamæmorus endozoically. Another group of the species in question comprises 126 Knup JESSEN: Rosaceæ. temperate forms, viz. Rubus saxatilis, R. arcticus and Alchi- milla vulgaris; they have extended more or less further to- wards the north, and have in several places passed beyond the Polar Circle. To these Sorbus americana may also be referred. This and the two Rubus-species have endozoic or at least synzoic fruit-dispersal, and the Rubus-species (especially В. arcticus) have attained a very wide distri- bution. Alchimilla alpina is presumably an Alpine form, which has probably during the Glacial period migrated north- wards and has in Scandinavia passed over the Polar boundary (BUSER in SCHROTER; |. с.). 12.—3.— 1914. AT oe ‘à Ds + ’ Carsten Olsen. 1914. Di IE LU = - Preface. The material used for the following investigation is the pro- perty of the Botanical Museum in Copenhagen. A large material of Cornus suecica, preserved in spirit and collected at several places in Greenland has been placed at my disposal, also a smaller quantity of similar material from Finmark, other places in Norway, from Sweden and Denmark. Besides this, I have used dried material of Cornus suecica from Greenland, Iceland, the Ferées, Denmark, Norway, Sweden, Finland and North America. Amongst the dried specimens from Greenland I found three specimens of Cornus canadensis, determined as Cornus suecica ; these specimens were collected by A. Jessen, July the 28th 1894, on a small, remotely situated island, Inuarudgligak, near the southern point of Greenland (60° 23"). Cornus canadensis has not hitherto been known to exist in Greenland; certainly У. МеЕНАХ mentions it as found here, but Tu. Hozm, who has gone through his material from Greenland, has shown, that Meehan was probably not in position to determine his plants correctly, especially as he has not even collected the plant, but simply made a note of its occurrence. I have also had ample material of Cornus canadensis from North America and Japan. Fig. 2 is drawn by Professor Warmine; the photograph fig. 10 is taken by Н. Е. Petersen, mag. sc., while the remaining figures are drawn by myself. XXXVI. 9 130 É CARSTEN OLSEN. Principal literature. BucHENAU, Fr., 1859: Bemerkungen über Cornus suecica. (Flora XLII). Hartz, N, 1894: Botanisk Reiseberetning fra Vest-Gronland. (Med- delelser om Grønland. Bd. 15). HesseLMAN, H, 1897: Nagra iakttagelser öfver växternas spridning (Bot. Notiser). Hom, Tu., 1895: Contributions to the Flora of Greenland (Pro- ceedings of the Academy of Natural Sciences of Philadelphia). LeisT, К, 1889: Ueber den Einfluss des alpinen Standortes auf die Ausbildung der Laubblätter (Separat Abdruck aus den Mittheil- ungen der Naturforschenden Gesellschaft von Bern). LovELL, Joun, 1898: The Insect-visitors of Flowers. (Bulletin of the Torrey Botanical Club. Vol. XXV). MEEHAN, W, 1393: A Contribution to the Flora of Greenland (Pro- ceedings of the Academy of Natural Sciences of Philadelphia). MEEHAN, THOMAS, 1892: Contributions to the Life Histories of Plants, No. 8 (Proceedings of the Academy of Natural Sciences of Phi- ladelphia). Norman, J, 1895—1901: Norges arktiske Flora. ROSENVINGE, L. KoLpErup, 1896: Det sydligste Grønlands Vegeta- tion. (Meddelelser om Gronland, Bd. XV). RosENVINGE, L. Kotperup, 1896: Nye Bidrag til Vest-Gronlands Flora (Meddelelser om Gronland, Bd. XV). SCHAFFNER, JOHN, 1905: Key to the Ohios Dogwoods in the Winter- condition. (Ohio Nat. VI). SERNANDER, R, 1901: Den skandinaviske vegetationens spridnings- biologi. SERTORIUS, A, 1993: Beitrage zur Kenntniss der Anatomie der Cor- naceae. Diss. München (Bulletin de l’Herbier Boissier). Siren, Е, 1905: Blomsterbiologiska iakttagelser i Kittila Lappmark. (Meddelanden af Soc pro Fauna et Flora Fennica. 31). SYLVÉN, N, 1906: Om de svenska Dikotyledonernas forsta forstark- ningsstadium eller utveckling fran frö til blomming. 1 (Kungl. Sv. Vet. Akad. Handl. XL Nr. 2). WARMING, Euc., 1887: Om Grønlands Vegetation (Meddelelser om Gronland XII). Wırrıs and Вовкил, 1903: Flowers and Insects in Great Britain (Ann. of Bot. XVII). Cornaceæ. 131 Cornus suecica L. Norman, 1901, р. 322. WARMING, 1887, р. 18, 34, 40, 75, 138. RosENVINGE, 1896, р. 67, 128, 161, 197, 198. Harrz, 1894, р. 57. `ВоснемАю, 1859, р. 87. ЗутмЕм, 1906, р. 135. SILÉN, 1905, р. 87. Wircis & Вовкил, 1903, р. 343. SER- NANDER, 1901, р. 141, 142, 164. SErrorius, 1893, р. 622. Кмотн, 1898, vol. 2, part 1, р. 51). Cornus suecica is mostly found in the subarctic regions, and most frequently in the northern part of the Scandi- navian peninsula, but is fairly common everywhere in Fin- land, Sweden, Norway and in the northern parts of Den- mark. Elsewhere in Europe it is found near Reval, in Hol- stein and Pomerania, and the Alps, also in the Ferées and Iceland. Outside Europe it grows in West Greenland, in Labrador, New Foundland, Alaska and Baffinsland. In Greenland it appears most frequently in “Urtelierne” and in thickets; Hartz says it is nearly always found in birch thickets and RosENVINGE, that it is also present in the willow thickets and on the heaths in the southern part. Its northern boundary in Greenland is according to Rosen- vinge, at 69°14. According to Norman it occurs in Finmarken in the birch woods, where it often by itself forms the undergrowth, growing closely like corn in a field. The development of the shoots of Cornus suecica has been described by BucHEnau (1859), and my own obser- vations quite agree with his description. The rhizome is horizontal and bears opposite triangular scale-leaves, two to three centimeters apart. It is of several years duration, while the scale-leaves die already the first year; and from their base adventitious roots appear, with a rich rami- fication. In the axis of the scale-leaves buds develop, some becoming runners, resembling in all particulars the mother- 9* 132 i CARSTEN OLSEN. shoot, others remain dormant. The runners, which attain a length of up to 30 centimeters, finally turn their apex upwards, forming a stem with foliage-leaves, terminating with an inflorescence or else remaining vegetative. The perpen- Fig. 1. Cornus suecica. A, Flowering Plant; В, Apex of Runner (Godthaab, Greenland); A. about 5/4, В, 2/1. dicular part of the shoot commences with a series of pairs of scale-leaves, red while fresh, but withering quickly; their number varies from 3 to 9 pairs, and is as a rule 5. The foliage leaves develop next, Danish specimens showing 4 to 8 pairs, those from the arctic 3 to 4 only. Cornaceæ. 133 BUcHENAU states, that the two uppermost pairs are closely placed, resembling а whorl; all the specimens from one Danish locality (Maglemose in Grib Skov) examined by me showed the same, while plants from other Danish loca- lities did not. On arctic material I never found it. The aérial part of the shoot attains a height of from 6 to 30 centimeters and is annual. The stem is quadrang- ular and has two opposite grooves (fig. 12), which change sides at each internode, the leaf pairs being placed at their lower ends. The scale-leaves, with which the perpendicular part of the shoot commences, support buds, the first pair excepted; and some of the buds develop next year into shoots with foliage leaves, growing upwards immediately, and showing no inclination to become runners. These shoots also commence with a number of red scale-leaves, after which follow foliage-leaves. The same ramification takes place the following year from the scale-leaf buds of these shoots, and so on, thus forming a tufty growth. According to BUCHENAU, it is always the buds ap- pearing in the axis of the second pair of scale-leaves which develop, so that all the shoots are placed at the same level. This is however far from always being the case, often it is from the uppermost pair of scale-leaves or the one next to this, from which next year’s shoot develops, and often only one bud from a pair of scale-leaves develops, although the second one can very well develop later during one of the following years. The buds which are to develop into shoots next year, are in the autumn of a considerable size and the first two pairs of leaves are brown and formed like bud-scales. The whole shoot completely formed is inelosed in the winter- bud, also the inflorescence and the first indications of the buds to develop the 2nd year. The leaves of Cornus suecica fall in the autumn, while 134 CARSTEN OLSEN. the stem is yet fresh; towards the winter it dies, but can very well remain standing in a withered state for se- veral years. The winter buds are located near the surface of the ground, and the plant is thus, as regards its life habit, a hemikryptophyte. Buds are formed in the axis of the foliage leaves, which however are very small, and as a rule only the buds of the uppermost pair of foliage leaves de- velop; at the time of flowering they are only small, but during the summer they develop into large side branches. I have noticed a Danish specimen, not flowering, where all the buds of.the foliage leaves were developed, with the excep- tion of the lowest pair. The leaf is ovate and sessile, with two or three veins diverging from each side of the midrib, near its base. On arctic specimens the uppermost pair of leaves (100 specimens were measured), had an average length of 19 mm, the average width being 12 шт. On Danish specimens, however, the average length was 23mm with an average width of 17mm, 100 specimens also being measured in this case. The leaves of the Danish specimens thus distinguish them- selves by being considerably larger than those from the Arctic. The seedlings have been described by SYLVÉN; the cotyledons are shortstalked, ovate-lanceolate, about 7 mm long. The primary root is strongly ramificated, and the devel- opment of the adventitious roots seems only to take place during the second period of growth; only few internodes are formed on the aerial shoot during the first year. The hibernation takes place by means of buds in the axis of the cotyledons, which next year develop into new aérial shoots, and the formation of runners only commences several years later. In Greenland the flowering takes place in July and August. The inflorescence is a head-like or umbelliforme Cornaceæ. 135 cyme, surrounded by four yellowish-white involucrals, by which, as is well known, the whole inflorescence gets the appearance of one single big flower, with four petals. Each flower has a short stalk, about 2 mm long; the sepals are very small, and the petals, of which one is often elongated, have a dark red-violet colour, and are turned sideways. The style is surrounded at its base by an annular brown disc; the stamens are turned to the side in the spaces between the petals. Fig. 2. Cornus suecica. A—E;from West Greenland, (Sukkertoppen, 16th August); A. Flower in bud (2%/,), В, Flower in full bloom (just opened); С, Anther, not open (*/1); D, Anther, open {(%/,); Е, Discus, seen from above (*/,); F—K from Northern Norway (Tromsø, 24th July) (11); Е, Flower with one anther (to the left) open (!%/,); С, The elongated petal (see fig. A); H, Part of style with stigma; J, Empty anthers, pollen apparently on stigma; K, Same, but filaments more curved; Г, Two pollen grains. from Northern Norway (Bosekop) (#°/,) ; (A—K, Eug. Warming del.). Concerning specimens from “Sukkertoppen’’ WARMING writes in his notes from Greenland: ‘“‘Shght protandry is noticed. 4 Simultaneous with the opening of the petals, the stamens extend and the anthers open, but the stigma is not yet fit to retain the pollen; when this shortly after- wards happens, the anthers are still full of pollen and self- pollination possible, though I think, not easily, by reason of the spread of the filaments (Fig. 2 В, Е); as the flowers 136 CARSTEN OLSEN. stand so very close to each other, it is more likely they could reach the nearest ones and deposit pollen on their stigmas, and, аз а matter of fact, [ have seen the anthers of neighbouring flowers in close proximity to a flowers stigma””. WARMING found the same form of development in arctic Norway (Tromse, Alten) and pure homogamy was only found once. The pollination takes place by insects, which are guided to the flowers by the four big and petal-like involucrals; according to Wırrıs and Вовкил, the visitors are few, mainly of the order of Diptera. The pollen grains, which are only few in number, have three strongly projecting pores (fig. 2). Deformity of the inflorenscence is not quite uncommon; sometimes the involucral is partly and even completely green. WARMING mentions such deformities in his notes from “Sukkertoppen”; in the same locality he found spec- imens with 6 or 7 involucral leaves. The fruit is a drupe; the stone is two-celled, one of which nearly always remains barren; the ripe fruits, red in colour, are found in Greenland in August and September. The dispersal takes place by animals (HESSELMAN found 40 stones in the excrements of a Larus marinus), and partly by means of water; the plant often grows near water- courses, and the stones being lighter than water, will often be carried far away; SERNANDER found them thrown up by the sea near a small island in the Baltic, more than 400 km from its nearest habitat. Anatomy. The leaf is dorsiventral, with a pallissade tissue of one layer of cells and a pronounced spongy tissue (fig. 4). The epidermis of the upper surface consists of very large cells, while those of the epidermis of the lower surface are Cornaceæ. 157 smaller and with undulating sidewalls. This undulation is more pronounced in leaves of Danish specimens than in those from the Arctic (fig. 3). The sidewalls of the upper surface are also undulating, but in a lesser degree, and also in this case is found the same discrepancy between Danish Fig. 3. Cornus suecica. A, Epidermis, upper surface, with a two-armed hair (Sukkertoppen, Greenland); B, Epi dermis, upper surface (Maglemose, Denmark); C, Epidermis, lower surface (Sukkertoppen, Greenland); D, Epidermis, lower surface (Maglemose, Denmark); A, В. С and D about :36]1. and Arctic leaves, though less pronounced. The cuticula is thin on both sides, and shows opposite the larger veins fine stripes, running parallel to the longitudinal axis of the leaf. Only the lower surface has stomata, these are placed 138 CARSTEN OLSEN. over the level of the other epidermal cells, more pronounced on Arctic than Danish leaves. As to the number of stomata on а given area, they are perhaps more numerous in the Arctic specimens than in Danish. The table below gives the number of stomata pr. Sq. m. m. of a number of Da- nish and arctic leaves: Danish Leaves | Arctic Leaves Number of | | | Number of Locality Stomata | Locality Stomata pr.omm | | pr.omm Maglemose(Grib Skov) | 70 | Egedesminde......| 100 Waldiniosen wann... 100 Tromse...... es 84 Hssede ле | 78 Nunatsuk rer, 82 ВИО | 85 | Julianehaab....... 120 Skibsted” 2.7. 0 | 665, Tuy) flav sce ee 134 Мое 3e 127 | Godtbhaabr. ee 75 Mariager she 82 Sukkertoppen ..... 109 Hillerad 2. ae 70 | Sukkertoppen..... 79 Among other differences between leaves from the Arctic and those from temperate localities, the following is particularly pronounced and has shown itself to be very constant: The Arctic leaf is thicker than the Danish (fig. 4). This substantial difference is chiefly due to the strongly developed spongy tissue of the Arctic leaf, while the length of the cells of the palisade tissue is rather smaller than in Danish leaves. Besides this, the Danish leaves have a more loosely built mesophyll with the palisade cells only loosely connected, diverging below, and the spongy tissue has large intercellulars and more slender cells than is the case with arctic leaves (fig. 4, C—D). The relation between the size of the spongy tissue and that of the palisade tissue of Arctic and Danish leaves is plainly given by the following measurements of leaves from eight different Danish locali- ties and a similar number from the Arctic. em ne Cornaceæ. 139 Danish Leaves - Arctic Leaves Thickness of | Thickness of Palisade | Spongy Locality | | Palisade | Spongy a alisade Leaf | tissue tissue | Leaf | tissue | tissue Locality Maglemose | Godthaab ... | 231и 51 | 138 a (GribSkov) | 1744, 66 » | 90 y | Sukkertoppen | 284- 54 - | 135 - Нэзеде.... | 174- 57 - | 78 - | Egedesminde. | 210-| 57 - | 120 - Mariager ..|168-| 57 - | 81 - | Julianehaab . |246- 60 - | 150 - Helsingør. |174-| 66 - | 75 - | Nunatsuk … |246-| 72 - | 144 - Hillerad ... |177-) 66 - | 78 - | Ilua......... 1210-| 54 - | 120 - Skibsted...|174-| 66 - | 81 - | Tromse...... 189- 48 - | 114 - Vildmosen . |180- 75 - | 78 - | Sukkertoppen | 240- 60 - | 138 - Silkeborg .. | 156-| 60 - | 66 Fig. 4 Согпиз suecica. A, Transverse section of leaf, р, two-armed hair (Godthaab. Greenland); В, Transverse section of leaf (Maglemose, Denmark); C, Section through the spongy tissue, parallel to the surface (Godthaab, Greenland); D, Section through the spongy tissue (Maglemose, : Denmark); A, В, С and D about #5/.. 140 CARSTEN OLSEN. The upper as well as the under surface of the leaf is provided with large two-armed hairs (Malpighia-hairs), fig. 3 А, shaped like a pickaxe, the upper-surface having, on the same area, about three times as many as the under surface; they are one-celled and the stalk is exceedingly short (fig. 4 A). The иле Was oa SK CRT ph Е == > Fig. 5. Cornus suecica. Transverse section of stem; ph, Leptom ; had, Hadrome (Egedesminde, Greenland); about 2/,. surface of the hair (the short stalk excepted) has numerous projections, thickenings of the walls, and the latter plainly show several layers which are much thickened; the lumen is extraordinarily small and seldom reaches the ends of the arms, which are consequently solid (fig. 3, 4). The main veins of the leaves have on their under Cornaceæ, 141 surface, below the phloëm, a slight sclerenchyma, which is not present in the smaller ones. The Stem (fig.5). The aerial shoot is annual. The epid- ermis has thickened outerwalls and is covered by a thin cuticula, striped parallel to the stem’s longitudinal axis; PZ Re X № er Fig. 6. Cornus suecica. Transverse section of rhizome; 5, Cork; cb, Cambium; had, Hadrome (Egedesminde, Greenland), about ?’°/,. stomata are present. Five layers of large chlorophyllaceous bark-cells follow inside the epidermis, and after these a more or less developed annular schlerenchyma, consisting of elongated parenchymatic cells, with much thickened walls (fig. 5, sc). The phloëm appears inside this annular scleren- chyma and, like the woody tissue, forms a coherent ring 142 CARSTEN OLSEN. round the stem. The cambium is rudimental in the fully developed stem, its activity soon ceasing as the whole of the woody tissue is already formed while the stem is quite young. The wood consists chiefly of spiral and pitted vessels, tracheids and wood-parenchyma. The vessels are fairly narrow, and in transverse sections, angular; the pitted ves- sels have ringpores, while the wood-parenchyma has single Fig. 7. Cornus suecica. Transverse section, showing primary structure of root; the epidermis is dead and partly fallen off; exod, Exodermis; end, Endodermis (Sukkertoppen, Greenland) about 339/.. pores; medullary rays are not present, but starch is found in the wood-parenchyma, and inside the latter a medulla with large cells runs in the longitudinal direction of the stem. The rhizome (fig. 6) is covered with cork, consisting of rather thickwalled cells and formed sub-epidermally. The bark-cells are arranged rather irregularly inside the cork, they are somewhat thickened collenchymatically and contain starch, which as far as Cornus suecica is concerned, always Cornaceæ. 143 consists of composite granules. On the innerside an even transition from the bark-cells to the phloëm is to be found. As in the stem, the wood consists of a coherent ring, and has rather indistinct annual rings. The secondary wood consists of trabecular and ringpored vessels, together with tracheids and wood-parenchyma, this latter being rich Fig. 8 Cornus suecica. Transverse section, showing secondary structure of root; S, cork (Greenland); about %?/,. in starch. Medullary rays are not present, the medulla con- sisting of large cells filled with starch. The bark-cells of the rhizome contain large quantities of tannin, which is also found in the bark-cells of the stem and in the leaf. The Root: the primary structure is diarch, it has a suberized rather thickwalled endodermis and а distinct pericycle (Fig. 7). 144 CARSTEN OLSEN. The bark consists of rather large cells; the epidermis dies away very quickly, and remains as a loose sheath round the very young root, while the exodermis, the cells being yet well filled with protoplasm, soon produces а layer of corky tissue with thickened cells, and resembles otherwise the cork of the rhizome. The secondary root (fig. 8) is on the whole of the same Fig. 9. Cornus suecica. Tangential, longitudinal section through the bark of a young root, containing mycorrhizae; a, destroyed hyphae (Greenland) about */,. structure as the rhizome, the exception being that no me- dulla is found. Endotrophic mycorrhiza is always found in the young root, in Arctic as well as in Danish specimens (fig. 9); the fungal hyphae are rather thick and not very closely knotted together, as is most often the case with roots having endo- trophic hyphae. The hyphae are both inter- and intracel- lular, and in the cells are also found irregular or coral-shaped clumps, probably destroyed hyphae, which like the former Cornaceæ. 145 colour strongly when treated with Hansen’s haematoxylin. The cells containing these lumps or living hyphae contain no starch, other cells are rich in this. Cornus canadensis L. SERTORIUS, 1893, р. 619; Lover, 1898, р. 386; Hom, 1895, р. 544; MEEHAN, Тн., 1892, р. 376; МЕЕнАм, W., зв poll. Cornus canadensis (fig. 10) is found most frequently in North America; № is met with throughout Canada, with the exception of the most northern parts, from New Foundland and Labrador to Southern Alaska. Its south- ern boundary passes through the states of West Virginia, Ohio, Indiana, Wisconsin, Fig. 10. Cornus canadensis Minnesota, N ebraska, Colo- from Inuarudgligak in Greenland; about */, (H. t rado and California. It is u Be Blau). found also in the northern Islands of Japan, and now, as mentioned in the preface, in the most southerly part of Greenland, in the small, remotely situated island, Inuarudgligak. It does not reach as far North as Cornus suecica, which е. g. is met with at Baffinsland, where С. canadensis is not found, whereas it is found consider- ably further South. In North America C. canadensis has its habitat in damp, cold woods (Pinus, Larix) and in thickets, and evidently grows in more shady places than C. suecica; in Japan it grows in the mountains. As to the structure of the shoot, this corresponds in the main to that of С. suecica; all the foliage leaves are XXXVII. 10 146 CARSTEN OLSEN. placed in а whorl at the top of the stem, though occasion- ally one or two leaf pairs are found lower down. The leaf- whorl is formed by the suppression of the uppermost inter- nodes on the main as well as on the secondary axis. In contrast to C. suecica the leaves are said to be everlasting (ASCHERSON and GRAEBNER); they are considerably larger and tapering at the base; and also differ as regards the venation; in C. suecica the veins all diverge from the base of the leaf, while in C. canadensis the two uppermost side- veins diverge from the midrib somewhat higher up. The involucral leaves are as a rule larger than those of C. suecica. The petals are green, one of them often having an awl-shaped elongated point. In material from North America I have found specimens with red petals edged with green, which together with other characteristics made them form a transition in appearance to С. suecica!). The three specimens from Inuarudgligak which constitute the whole material from Greenland had green petals. The biology of the flower has been treated by LOWELL, who writes: “Аз the flower-bud approaches maturity the filaments grow rapidly, breaking apart the petals at the base, and protruding as four V-shaped arches. If at this stage the awl-shaped appendage is touched by the body of an insect, or by a needle, the petals are instantly reflexed, the elastic filaments straighten, and a minute shower of pollen is projected upward. In fully matured flowers a faint snap may be heard. Self-fertilization is prevented by the immaturity of the stigma. Both stamens and petals soon fall rot and the capitate stigma, composed of white i I var seen a similar intermediate specimen from Neriak on the East coast of Greenland (61° 3”); the involucral leaves were very large, and the petals red, edged with green; it had 2 pairs of large, opposite leaves of which the uppermost pair supported short-jointed shoots; the leaves tapered at the base and the venation and anatomy were quite similar to those of С. canadensis. — Cornaceæ. 147 papillae, subsequently matures. Though self-fertilization is prevented by proterandry there may occasionally be ob- served, resting against the stigma, the anthers of neighbouring flowers. The honey is secreted in a very thin layer by a dark brown ring at the base of the style!. eaves Fig. 11. Cornus canadensis. A, Transverse section of leaf (Inuarudgligak, Greenland), B, Epidermis, upper surface (Greenland); С, Transverse section of leaf (Minnesota); D, Epidermis, upper surface (Minnesota); А, В, С and D, about 1*/,. 1 THOMAS МЕЕНАМ reports having observed, that С. canadensis is partly monoecious, partly dioecious. He has only seen plants with ripe fruit, not any flowering specimens, and found, that the inflorescences of some plants were quite sterile, others had a single berry in the middle of the cyme, while others again were rich in fruits; from this he draws his conclusion. All the individuals I have examined have been hermaphrodite, and as Lower, who has studied the biology of the flower, does not 10* 148 CARSTEN OLSEN. Anatomy. The Leaf. The leaf is dorsiventral as in C. suecica, but differs in having the epidermal cells of about the same size on both surfaces; also by the cells having strongly thickened outer walls, this latter being caused by the cellulose-mem- brane; the cuticula is quite thin. The palisade-cells are rather thick and short. Apparently there is the same dif- ference between leaf specimens from Greenland and those from more southerly latitudes, as is the case with those of С. suecica. The leaves from Greenland are thicker than those from further South, and the extensiveness of the spongy-tissue is much greater in comparison to the palisade- tissue. The following table gives a few measurements of leaves from different latitudes: | | | ‚ Exten- | ] Exten- Locality ns [den - | tissue | tissue Greenland Ay 90 м | би | 56 м Vancouver, Islands "1 #77 60 - | 18 - 30 - Minnesota. о. | 55 - | 20 - 26 - New. Ноно ао. 70 - | 16 - | 37 - Turnavik (Labrador)......... OS 2 - Specimen from the Botanical | | Garden.at:Berlin........... | 77 - | 27 - 40 - On the American specimens the epidermis of both surfaces have strongly undulating sidewalls, while this is much less pro- nounced in those from Greenland (Fig. 11). Stomata are only present оп the under surface; the specimens of С. canadensis mention diclinous flowers, these must consequently be very rare, if existing at all. Cornaceæ. 149 from Greenland having 160 of these pr. square: mm, while specimens from Canada had only с. 140 pr. sq. mm. The Stem. The stem is very much like that of A EN С SH IE transverse | + section being however es co) ) somewhat different. As “> / A Fig. 12 shows, the stem of C. suecica seen in trans- 5 > i 2: verse section is more bar- Fig. 1 Transverse sections of stems. A, Cornus cana- rel-shaped, and the groove densis from Vancouver Island; B, Cornus a hal h suecica (Greenland); C, Cornus canadensis from not very deep, while that Greenland; A, В and С about %/,. of С. canadensis is con- siderably deeper, making the stem ‘“‘four-winged”’; besides this, the circumference of the medulla of C. canadensis is quadrangular, while that of С. suecica is round. In С. cana- Fig. 13. Transverse section of the winged part of the stem. A, Cornus canadensis from Vancouver Island; В, Cornus suecica from Greenland; С, Cornus canadensis (Greenland); A, В and С about 15/,. densis the distal part of the wings of the stem consists of collen- chyma while such is practically missing at this place in C. suecica (Fig. 13). The sclerenchyma in the bark of C. canadensis is often greatly developed in vigorous specimens, while it is about 150 CARSTEN OLSEN: Cornacex. missing in poorly developed plants, e. g. in the specimens from Greenland. The anatomic structure of the rhizome and the root is exactly like that of С. suecica. Tannin is present throughout the plant. The mycorrhizas of C. canadensis and C. suecica are alike. Summary. Cornus suecica is a distinct mesomorphic plant which completely lacks xeromorphic characters; such are however found in C. canadensis: the epidermal cells of the leaves of this species have thickened outer walls, which is correlated to the fact of the leaves being perennial. The more loose construction of the mesophyll of Arctic plants mentioned by BorGEsENn! does not apply to the species here dealt with; on the contrary, the Arctic specimens of С. suecica have а mesophyll of а much closer structure than the spec- imens from more southerly localities. Besides this, the leaves of С. suecica, grown in Arctic localities, are disting- uished by being smaller, but decidedly thicker than those of specimens from Denmark. The spongy-tissue is very strongly developed in Arctic specimens. 7 Bidrag til Kundskaben om arktiske Planters Bladbygning (Bo- tanisk Tidsskrift, Bd. 19, 1895; also in “Journal de Botanique”, vol. 9, Paris 1895). _ 10. _ Caprifoliaceae. % Linnaea borealis Г... >. р By Г 0. Hagerup. р 1915. kr 1 A u № . 7 MU te fé ты NE wo) auf The present investigation is carried out with the aid of material from the herbaria belonging to the Botanical Museum at the University of Copenhagen; it is brought from many different localities by several investigators. Besides this, material from Denmark, preserved in spirit has also been used. All material from Greenland was dried, which has caused difficulties, especially as regards the leaf ana- tomy; some material was also collected by WARMING in Northern Scandinavia. Linnaea being a plant about which many investigations have been made and many papers written, especially GIGER's monograph, a more complete treatment is superfluous, and in this general statement it has only been necessary to mention the main points. As regards the distribution of the plant the above named paper may be referred to. As to the localities in Greenland in which Linnaea is found, the following are represented in the herbaria: 1. Prestefjeldet near Holsteinsborg, in a thicket of Salix glauca, Empetrum and Vaccinium uliginosum, collected by WARMING and TH. Horn. 2. Ivnarsuak in Tasermiut Fjord, leg. F. BENJ. PETERSEN. 3. Ivigtut, leg. Aue. BERLIN. 4. Kvannit Sarok in Diskofjord (Lat. 69° 33’N.), copiously flowering in the outskirts of willow thickets; leg. Мовтем P. Porsttp; this being the most northern locality in which Linnaea has been found in Greenland. 154 О. НАСЕВУР. The following literature has been used (complete list of literature available is given in GIGERS monograph). Brenner, M., 1909: Nagra Linnaea-former i Finland (Medd. af Soc. pro fauna et flora Fennica, h. 34, p. 89—97 og h. 35, pag. 56). Giger, EMIL, 1913: Linnaea borealis L., eine monographische Studie (Beihefte zum Bot. Centralblatt, Band XXX, Zweite Abt.). GUNDERSEN, 1910: Recherches anatomique sur les Caprifoliacées (These; Paris). Leist, K., 1889: Uber den Einfluss des alpinen Standortes auf die Ausbildung der Laubblatter. (Mittheil. d. Naturf. Gesellsch. v. Bern.) Norman, J. M. 1895: Norges arktiske Flora, II, pag. 324. Kristiania. PETERSEN, О. G., 1901: Diagnostisk Vedanatomi, pag. 93. København. Porprius, B. R., 1903: Blombiologiska iakttagelser (Acta Soc. pro fauna et flora Fenn. 25, No. 1). ScHRÔTER, C., 1904: Das Pflanzenleben d. Alpen, I. р. 235. SERNANDER, R., 1891: Studier öfver skottbyggnaden hos Linnaea borealis L. (Bot. Notiser 1891, p. 225). SILÉN, F., 1905: Blombiologiska iakttagelser i Kittilä Lappmark (Medd. af Soc. pro fauna et flora Fenn., h. 31, p. 88). SyLVÉN, N., 1906: Om de svenska Dikotyledonernas första förstärk- ningsstadium; 1 (Kungl. Svenska Vetenskapsakademiens Hand- lingar, Band 40, No. 2, p. 65). WAGNER, А., 1892: Zur Kenntniss des Blattbaues 4. Alpenpflanzen und dessen biol. Bedeutung. (Sitz. Ber. der Wien. Akad., math.- nat. Cl р. 615, Baye 1, Abt. D} Wittrock, У. B., 1879: Om Linnaea borealis (Bot. Not. 1878, р. 17, 49, 83, 122:and 1879, р. 9, 137). — 1907: Linnaea borealis (Acta Ноги Bergiani, Bd. 4, No. 7). Wyo ter, H., 1859: Morphologische Mittheilungen (Flora N. В. 17, р. 1). Germination. According to the literature Linnaea seldom produces germinable seeds in Central Europe, while seeds from the North germinate within a few weeks. WITTROCK has found a seedling grown under natural conditions. The further development of the seedling has been examined through cul- tures by GIGER; the results are briefly as follows: the germinable power of the seeds lasts for at least three years, and the deve- lopment of the seedling proceeds slowly, especially at first ; during the second year the main axis, consisting of 2—4 inter- nodes, inclines horizontally towards the ground, and becomes Caprifoliaceae. 155 therejuvenatingrunner, bearing buds, which, as described later, become fructification or assimilation shoots, respectively. In its third year the plant has reached its full development, con- sisting of a main root (which can live a few years), some adventitious roots, and two succesive long-shoots. Morphology of the Stem. Linnaea borealis L. is an ever- green woody plant (not an undershrub) with a long, creeping main axis, sending off rootlets, and bearing two kinds of erect, short shoots, which according to their function may be called, respectively: assimilation and fructification shoots (WITTROCK). The Rejuvenating Shoot. The rejuvenescence takes place by the terminal bud of the main axis yearly producing a creeping shoot, 6—45 centimeters long, consisting of 10—16 internodes, the individual length of which is lessened towards both ends of the shoot. WITTROCK mentions, eo that these shoots can reach an age Fe À of up to 13 years. The plant can quickly spread over large ie areas. The terminal bud is covered NG о by rudimentary green leaves, which =< в. do not wither during the next period of growth, but on the con- Fig. 1. trary lengthen somewhat in their Disram, showing the more vigorous buds (the larger circles) supported outer half, and act as organs of by the leaves placed on the under- ser Ä side of the rejuvenating shoot. assimilation. As a rule all the leaves of these rejuvenating shoots support buds, which during the next period of growth develop into one of the three kinds of shoots. Generally however one of the two buds of a leaf-pair becomes a “dormant bud”, while the other, distinguished by its greater size, develops into an 156 О. HAGERUP. assimilation or fructification shoot. Through the twisting and turning of the petioles, the leaves apparently stand in two rows, whereas in reality they are placed crosswise oppo- site. It is always the leaves nearest the ground, which support the most vigorous bud, and as а consequence of this, the short-shoots are placed in a zig-zag line with an ascending base. Fig. 1 illustrates this, which disagrees with the general view, that the leaves have been placed in two rows by the twisting of the stem. New rejuvenating shoots may appear by the development of older and younger buds from all three kinds of shoots, which enables the plant to spread in all directions (see GIGER). The Assimilation Shoot develops, as already mentioned, from a side bud on the horizontal rejuvenating shoot. It is characterised by a short axis (4—6 cm high) with 4 or 6 internodes, and especially by a relatively great number of folıage leaves, which gain for it the name of assimilation shoot; besides these leaves, one or two pairs of scale leaves are present, which quickly wither. The axillary buds seldom develop: the rejuvenescence is, as in the rejuvenating shoot, monopodial; the terminal bud, covered by 1 or 2 pairs of scale leaves, develops during the following period of growth into a shoot similar to that of the previous year; this is continued for 3 or 4 years, when at last the shoot ceases its function. New assimilation shoots may also develop from older or younger side shoots or end shoots, of any of the other kinds of shoots (see GIGER). The Fructifieation Shoot consists of two different parts, viz. a lower vegetative part, which is perennial, and an upper floral one, which is annual. Of these the lower one is much like the just mentioned assimilation shoot, but differs by nearly always sending out adventitious roots from its ascending base. The shoot ter- Caprifoliaceae. 157 minates in а floral part, which withers every year, and the rejuvenescence takes place by a vigorous bud, situated in the axis of one of the upper leaves, so that the fructi- fication shoot, contrary to the two other forms of shoots, becomes a sympodium. The floral part consists of а long thin internode, the terminal bud of which does not reach development under natural conditions. Below this point is а pair of bracts, each supporting a flower, the inflorescence thus being a two- flowered cyme. Just below each flower, three alternating pairs of bracts are found; the uppermost pair grows bigger as the fruit ripens, finally quite enclosing it. The whole of the floral part of the shoot is covered by glandular hairs, which have а many-celled head, and а long, two- or three- celled stalk. These hairs increase in numbers towards the upper end of the shoot; they are especially numerous on the uppermost pair of bracts, which (according to GIGER) produces additional glandular hairs towards the time of maturity; by means of their sticky secretion they are said to further the dispersal of the fruits by causing it to become attached to passing anımals. The Flower. A zygomorphy is often present both in calyx and corolla ; it varies however much, as does also the colour of the corolla; both of these peculiarities have been closely studied by Wittrock (1907), to whose work reference is made. The flowers have a pleasant, strong scent of vanilla, and on the innerside of the corolla, at the base of the lowermost petal, is a nectarium partly covered by long hairs. The secreted fluid shows distinct sugar reaction. The anthers are didynamous and develop simultaneously with the stigma (according to WARMING). The list of the visiting insects (GIGER) shows a comparatively great number of Diptera, honeyseekers with short proboscis, which 158 О. HAGERUP. without any difficulty can enter the corolla by reason of its large opening. Besides, visits by Lepidoptera and Hymenoptera have been observed. The ovary is trilocular, two compartments containing several ovules which do not develop, while the third contains one only, which however is fit for development. Fig. 2. Transverse section of the rejuvenating shoot (material from Greenland). Compared with material from Denmark, the Arctic Lin- naea has a compact growth, owing to shorter internodes; this mode of growth causes the deviating habitus, peculiar to many Arctic plants. Anatomy of the Stem. The bark and the wood together only amount to half the diameter of the stem, the pith being relatively strongly developed. The bark passes through Caprifoliaceae. 159 peculiar phases of development. The epidermis and hypodermis (Fig. 2a) start to fall off in the 3rd year, through the destruction of the underlying 2—3 layers of cells (Fig. 25); the protective function of the former is then taken up by the 1—(5) layers of large bark cells (Fig. 2c) underneath, the walls of which become woody. During the 4th year however, this protecting layer is also partly thrown off, owing to the cells immediately underneath being similarly destroyed; the outer cells then form a third protective layer by a strongly developed forma- tion of cork. The diameter of the stem is about the same all over, as the growth in thickness keeps pace with the decrease in dia- meter, caused by the falling off of the several layers of bark. Under the bark is the cambium (Fig. 2d) and under this the wood (Fig. 2e), which yearly grows a little in thickness and shows annual rings. The wood consists of vascular tissues, trach- elds, libriform-cells, wood-parenchyma and medullary rays. The original transverse walls have the appearance, even in the fully developed vessels, of a grate, consisting of very thin, parallel staffs. The vessels lie scattered in the wood, but not in such great numbers as the tracheids; all kinds of transitions to these latter are however found, and the circular or oblong pores have the same trabeculate arrangement. . The greater part of the wood consists of tracheids ; hbriform-cells and wood-parenchyma are rarely met with, and are of quite subordinate importance. Between the pith (Fig. 22) and the wood are found a few layers of cells, which both in shape and position form the transition between the medullary cells and the tracheids; as the figure shows, their walls have circular pores (Fig. 2 f). The numerous medullary rays consist of 4—8 radiant rows of cells placed vertically above each other; they are always well filled with starch, which can however also be the case with the cells of the living bark. 160 О. HAGERUP. Short- and long-shoots have in the main the same anatomic structure. The Leaf. Its duration is from 2 to 3 years; in shape it varies somewhat, most often it is ovate, but elliptical-round or nearly kidney-shaped leaves are also found. As a rule 2 faint incisions are to be seen on each side of the leaf, but some variations are also found in this respect; it is Fig. 3. The leaf. Epidermis: g, lower surface (material from Greenland); d, lower surface, 0, upper surface. with palisade cells added (material from Denmark). pinnately veined. Scattered hairs are present, of the same type as those on the stem; on the lower surface of the leaves are some glandular hairs, with a short one-celled stalk and a many-celled head, which secretes a sticky sub- stance in which fungus spores are often caught and germinate. The mesophyll shows that the leaf is a typical shade- leaf, in conformity with its usual habitat (Fig. 4). The epidermis contains chorophyll; the cells are largest on the upper surface (Fig. 3). The perpendicular walls are undulating with thickened outer ends (Fig. 4). This was especially pronounced on the Caprifoliaceae. 161 material from Greenland, where the walls also were thicker. The stomata, which are only found on the lower surface, are somewhat projecting, oval, and of the usual structure (Fig. 3). The following table shows the average number of sto- Fig. 4. Transverse section of leaf (from Denmark). mata on '10 mm? of 6 leaves from Denmark compared with 6 from Greenland: ie ANE I 2 3 4 5 6 pr. mm? Ванн: > 49:75 397 7207, "abe 789 >50 463 Arctic... 40 44 47 440 48. 45 440 On 1 mm? the average was about 450 stomata. The undulation of the perpendicular cell walls of the epidermis is still more pronounced on the lower surface of the leaf than on the upper surface. Besides, the material from Greenland also showed a thickening, and larger undu- XXXVI. au: 162 О. HAGERUP. lations of the perpendicular epidermis walls, on the lower surface, than was the case with leaves from Denmark. The cell (see Fig. 32), which carries a glandular hair, has however straight cell walls, and the spongy tissue underneath is without intercellular spaces. The mesophyll has a pretty, regular structure rich in intercellular spaces, especially so in the spongy tissue, but also in the palisade tissue (Fig. 4). Fig. 5. The lowermost layer of spongy tissue, viewed from above; g, from Greenland; d, from Denmark. The leaves from Denmark had, with scarcely any excep- tion, one layer of short, broad palisade cells only, while those from Greenland had two; further, the spongy tissue of the Danish leaves had throughout a looser structure than those from the Arctic. These relations are the opposite to what has been found by examination of other Arctic plants with regard to leaf-structure; the explanation is probably, that Linnaea in Greenland grows on open heaths, and consequently is not, as the Danish plant, a typical shade-plant. - - - Caprifoliaceae. 163 The spongy tissue, generally consisting of three layers, is connected with the palisade cells by a layer of collecting cells. The leaves from Greenland are, as is generally the case, of a smaller size than those from outside the Arctic zone. Fig. 6. Transverse section of triarch root (from Greenland). The Root (Fig. 6). The primary root dies quickly and is replaced by the fairly numerous adventitious roots, which sprout from ‘nodes on the rejuvenating and fructificating shoots, and ramify freely in the ground; this latter generally contains many decaying plant tissues, why mycorrhiza formations might be expected on the roots, but such have not yet been found. The epidermis of the root has thickened inner walls; the outer walls however are thin and, together with the root-hairs, decay during the second ls 164 О. Hacerup: Caprifoliaceae. year. The exodermis, with the remainder of the epidermis, thus becomes the outermost layer; its cell walls are slightly thickened. Underneath is the starch-carrying bark, which occasionally exceeds several times in thickness the diameter of the central cylinder. Towards the centre, the bark adjoins an endodermis, the cells of which has thickened walls, those of the passage cells excepted. During the fourth year the whole of the bark is forced off, and the pericambium beco- mes phellogen, which forms a cork, consisting of about 10 cell layers (GicER). The central cylinder is diarch to pentarch, and outermost is a distinctly bounded pericycle. Wood, to fill the central cylinder, is quickly formed; its anatomic structure is practically identical with that of the stem, and shows distinctly annual rings. Summary. When compared with material from Denmark, Linnaea from Greenland is distinguished by the following characteristics: The stem has shorter internodes, giving the plant a more compact appearance. The leaves are smaller with slighter incisions, and the perpendicular walls of the epidermis cells are thicker and have more pronounced undulations. The mesophyll has smaller intercellular spaces and larger palisade tissues. И. Primulaceae. By: Fr. J. Mathiesen. 1916. aA ON i | ane ax ‘i a RTS ET de ALT RON ! . ids aah ‚ А ый Cod: би, г Vey МГ Ni ER LA ++ AMIE 4 # * и Es à A) ARTE + LE Zu N Suhl F or this work, entrusted to me, the necessary material has kindly been placed at my disposal by the Botanical Museum of the University in Copenhagen and by the Riksmuseum in Stockholm. Т have throughout had the kind advice and guidance of Professor Е. WARMING, and wish to tender my hearty thanks for his great kindness and courtesy. Г am also greatly indebted to Dr. С. Н. OSTENFELD, Inspector at the Botanical Museum, for the willing help extended to me on several occasions. With the exception of figs. 6 A—T and 10, which were placed at my disposal by Prof. WARMING, all the others have been prepared by me. The anatomical sketches were drawn with the help of Abbe’s drawing-apparatus and reduced to half-size when reproduced. The lens-combination used in drawing is given in the text for each figure, one index is the number of the Zeiss apochromatic, the other that of the compensation-eyepiece. The following species have been examined: — Pag. OO ES OSES Le. Du à aaa de À SRE TAN 0 lege Sa 169 re БОИ A SARA Re 176 а de Ci 182 u jorınasa. var, LKO ION ALC AN. u... Ze ern 187 SMM Team SH A, PII,» GIS) APN ия HELE) EN, à 190 оО... ее 194 == RANGE AS MEL. J sib thie) Л FAIR, calles biog: 201 LEON rin: ne de ne 208 168 Ев. J. MATHIESEN. Principal literature. ABROMEIT, J., 1899: Bot. Ergebnisse der... unter Leitung von Dr. у. Drygalski ausgesandten Gronlandexpedition. В. Samenpflanzen (Phanerogamen). (Bibliotheca botanica. 42. Stuttgart.) Вьутт, A., 1906: Haandbog 1 Norges Flora. (Kristiania.) Brunvın, J. A. Z., 1898: Bidrag till kännedomen om de svenska = fanerogama örternas skottutveckling och öfvervintring. (Akad. Afh. Uppsala:) | BôüRrGESEN, F., 1895: Bidrag til Kundskaben om arktiske Planters Bladbygning. (Botanisk Tidsskrift, Bd. 19. — Resumé i Jour- nal d. Bot. IX, 1895.) ER Decrock, E., 1901: Anatomie des Primulacées. (Ann. des sciences nat., Botanique, 8 série, tom. XIII.) Eastwoop, ÅLICE, 1902: A descriptive list of the plants collected by _ Dr. Е. С. Blainsdell at Nome City, Alaska. (Bot. Gazette, 38.) à Exstam, O., 1897: Einige blütenbiolog. Beobachtungen auf Novaja Semlja. (Tromsö Museums Aarshefter, 18.) FLORA DANICA. .3 Hozm, Тн., 1885: Novaja Zemljas Vegetation, særligt dens Phanero- gamer. (Dijmphna Togtets zool. bot. Udbytte, pp. 1—71 Kjo- benhavn, 1887.) + KseELıman, Е. R., 1882: Asiatiska Beringssundskustens Fanerogam- flora. (Veg. Exp. Vetenskap. lagtt. I.) Клегтмлм, Е. R., 1883: Ur polarvaxternas lif. (Nordenskjölds Stu- dier och Forskninger.) i Кмотн, Р., 1898—1905: Handbuch der Blütenbiologie. LANGE, J., 1880: Conspectus florae Grénlandice. (Meddelelser om Grønland, H. III.) LANGE, J., 1887: Tillæg til Grønlands Fanerogamer og Karspore- planter. (ibid. H. III, 2.) Lınpman, С. A., 1887: Bidrag till kännedomen om skandinaviska fjallvaxternas blomning och befruktning. (Bihang till Kgl. Sv. Vet. Akad. Handl., Bd. 12, Afd. III.) ; Murer, H., 1881: Alpenblumen, ihre Befruchtung durch Insekten und ihre Anpassungen an dieselben. Leip zig. ÖsTENFELD, С. H., 1910: Vascular plants collected in arctic North America (King William Land, King Point and Herschell Island) by the Gjöa Expedition under Capt. Roald Amundsen 1904—1906. (Kristiania. Vidensk. Selsk. Skrifter.) Pax, F., 1889: Monographische Ubersicht über die Arten der Gattung Primula. (Engler’s Bot. Jahrbücher, X, pp. 75—241.) Pax, Е. og Кмотн, R., 1905: Primulaceae. (A. Engler: Das Pflanzen- reich, IV, 237. Leipzig.) RosEnvinGE, L. KoLderup, 1892: Andet Tillæg til Grønlands Fane- rogamer og Karsporeplanter. (Meddelelser om Grønland. H. МТ, 2.) Primulaceae. 169 SYLVÉN, N., 1906: Om de svenska Dicotyledonernas första Förstärk- ningsstadium. (Kgl Sv. Vet. Akad. Handl., Bd. 40.) WAGNER, A., 1892: Zur Kenntniss des Blattbaues der Alpenpflanzen und dessen biologischer Bedeutung. (Sitzungsber. der kaiserl. Akad. der Wissenschaften in Wien. Matem.-naturw. Classe; Bd. GT" Abth. 1.) WARMING, Euc., 1884: Om Skudbygning, Overvintring og Foryngelse. (Naturhistorisk Forenings Festskrift, Kjøbenhavn.) — 1886 a: Om nogle arktiske Væxters. Biologi. (Bihang till Kgl. Sv. Vet. Akad. Handl., Bd. 12, Afd. III.) — 1886 b: Om Bygningen og den formodede Bestøvningsmaade af nogle grønlandske Blomster. (Oversigt over D. К. Danske Viden- skabernes Selskabs Forhandlinger.) WIDMER, 1891: Europ. Arten der Gattung Primula. München. WypLer, H., 1859: Morphologische Mittheilungen. (Flora, pp. 20—23.) Primula egaliksensis Wormskj. Lit.: Flora Danica, $. 1511. Lance, 1880, p.71; 1887, р. 260... WARMING, 1886 a, р. 24. WiDMER, 1891, р. 119. RosEn- VINGE, 1892, р. 684. BorGEsEN, 1895, pp. 223, 225, 236, 237. Pax and Кмотн, 1905, р. 77. Herbarium and alcohol material from South-West Greenland collected by J.Vanı (1828); RosenvinGe (1888), MELdorPH (1900), and Deicumann (1906). Also, as alcohol material, seedlings from seeds collected in Greenland, and grown in the Botanical Garden at Copenhagen (one of these is illustrated in fig 1, a). A perennial herb of the common Primula type (War- MING 1884, pp. 33, 37 and 38; Pax in Englers bot. Jahr- biicher, 1889). The vertical rhizome is very short, since the lower end dies away from year to year; it is richly covered with adventitious roots, the duration of which must also be about a year. New adventitious roots had already appeared on specimens (gathered in July) where the buds, which continue the growth of the rhizome, had as yet only developed one pair of foliage leaves. The young roots were comparatively thick at their base (1.5—2 mm), and the cortical cells were full of starch. On the other hand, the 170 FR. J. MATHIESEN. roots of the preceding summer had already begun to decay and were completely emptied of starch. The leaves are arranged in а 2 spiral. They vary some- what in form and size, but with no apparent regularity in the variations, either as to arrangement or stage ol development. The principal bud is placed in the axil of the uppermost foliage leaf, and, as mentioned, it had already developed several foliage leaves by the middle of the summer; as the result of further increases during the sum- mer, a rosette of 6—8 leaves is formed. The growth of the bud ceases temporarily during the autumn, some of the leaves remaining as scale-leaves, to form a protective covering for NP the younger growth, which next spring will develop into ordi- nary foliage-leaves. These scale- leaves may be found at the base of the spring-rosette as late as July, but only in a very Fig. 1. Primula egaliksensis. damaged state. Unfortunately, a, Seedling from the Botanical Garden, J had at my disposal no Copenhagen, @/ı) 6, Young plant from - Igaliko (*/, 1848); 2, the two first leaves buds gathered during the win- NT GOON WINE ea eek ter, but it is probable that the bud-scales contain starch, as is the case with kindred species. The principal bud has sometimes been found to develop its peduncle during the same year as the parent shoot, and this even on plants showing relatively poor development, so that the plant may have two inflorescences flowering about the same time. Vegetative augmentation-shoots may be formed by small leaf-rosettes developed from some of the Primulaceae. 171 microscopic buds, found in the axils of many of the foliage leaves (WIDMER, p. 10). I have not been able to prove the presence of dimorphic flowers in this species (see E. WARMING 1886 a, p. 24). The length of the style varies only slightly, the majority of the flowers examined showing the stigma placed as in fig. 2, a. Self-pollination may occur, as the pollen and the stigma in the same flower seem to be fully developed at the same time. In Flora Danica the stigma is figured as ex- tending a trifle above the The corolla is 5—8 mm. in diameter, the colour anthers. being white, with a yellowish (Fig. 2.) The peduncle increases ring in the throat. in length after flowering, as in the case of Pr. stricta. Anatomy. A transverse section of the young primary root (fig. 3, А) shows the cen- tral cylinder to be very thin, only + of the diameter of the root; the fourth layer of the cortex seems to be in process of formation. root is diarch. The primary оз В: Fig. 2. Primula egaliksensis. a, Longitudinal section of a flower, the stigma is seen level with the base of the anthers, (5/,). 6, The corolla-limb of a somewhat larger flower : in the throat the anthers are seen, (5/,). с, Longitudinal section of a пре fruit, (about °/,). (Igaliko, ”/, 1888.) shows the central cylinder with the endodermis of а very young adventitious root. The cells of the cortex are filled with starch, except the outermost layer, which functions as an exodermis; this is 172 Fr.:J. MATHIESEN. also the case in the primary root. As the root gets older, the vessels increase in number, the epidermis dies away, after which the cortical cells become emptied of starch and finally perish. Crystals of а calcium salt, single as well as aggregate, are found in the bark, especially in the exodermis of the older roots. The Rhizome. A transverse section through the lower part of this shows the severed adventitious roots and the Fig. 3. Primula egaliksensis.. A. Tranverse section of the primary root (Seedling from the Botanical Garden, Copen- hagen); (Obj. 4, Ос. 4). В, Transverse section of a young adventitious root, showing the central eylinder and the inner layers of the primary cortex; E, Endodermis with Casparian spots; (Obj. 4, Oc. 6). (Igaliko.) vascular strands of the leaves, also a central group of 2—4 vascular bundles. Towards the upper end of the rhizome, where this passes over into the terminal peduncle, the number of the vascular bundles increases, but with a reduction in the size of each bundle, and at the same time several layers of the surrounding parenchyma develop into a continous sclerenchymaring. The vascular bundles differ from the nor- mal and approximate to the hadro-centrie type with sieve- tissue closing more or less distinct'y round the inner side of the bundle; exceptions to this are seen in the bundles at the modifié Primulaceae. 173 apex of the rhizome which are always distinctly collateral. These bundles (,,steles”, у. Tieghem) are surrounded by sheaths having Casparian spots on the radial walls, and vary in shape from circular to oblong-arched. The walls of the parenchyma show pores, and its cells contained some starch (at end of July). In the proximity of the vascular bundles, the ele- ments of the parenchyma are often arranged in fairly regular radial rows. The Peduncle is provided with a 4—6-layered scleren- chyma-ring, which however is reduced to 3—4 layers opposite each of the vascular bundles; these latter are arranged ina regular circle. In a longitudinal section this sclerenchyma- ring appears to consist of elongated, cylindrical parenchyma- tous cells, densely and finely pored and with lignified walls. On the inner side of the ring, the sclerenchymatous tissue closely surrounds the leptome of the vascular bundles, thus affording some degree of protection. In the stereom, as well as in the medulla, near the vascular bundles, there are cells containing a brown substance, the nature of which, however, could not be further determined. Outside the sclerenchyma- ring the peduncle shows 4—6 layers of parenchymatous cells, lengthened axially and with large intercellular spaces com- municating with the atmosphere through numerous stomates in the epidermis. Numerous chlorophyll granules are present in this parenchyma. The pedicels attain a thickness of about 0.75 mm, and have on the whole the same anatomical structure as the peduncle. The Leaf. The epidermal cells on the upper surface are larger and have thicker walls than those on the lower surface (fig. 4). The lateral walls are undulating, more so on the lower surface of the leaf. The outer walls show fine cuticular striations, especially near the margin of the leaf and above the midrib. The stomates are level with the surface 174 FR. J. MATHIESEN. or slightly protruding; they are surrounded by 4 to 6 cells, most often 5. BorGESEN states that the number of stomates is nearly equal on the upper and lower surfaces of the leaf, but such is not the case in the specimens of Pr. egaliksensis Fig. 4. Primula egaliksensis. a, The epidermis from the upper surface of the leaf. b, The epidermis from the lower surface of the leaf. (a and b, Obj. 4, Ос. 4). с, Transverse section of the leaf; on the lower surface a glandular hair; (Obj. 4, Ос. 6). (igaliko.) examined by me where the number of stomates was always greater on the lower surface. The mesophyll (fig. 4, c) has a very lacunar structure, and palisade cells are scarcely distinguishable, though the Primulaceae. 175 elements of the uppermost layer above the veins, are more elongated and cylindrical; the lower layers consist of slightly ramified cells. The whole of the mesophyll is rich in chloro- phyll granules, which are also found in the epidermis of the lower surface. Hydathodes are present on the leaf-margin, opposite the stronger veins. Small glandular hairs, with a one-celled Fig. 5. Primula egaliksensis. a, Transverse section of the calyx near its base. 6, The epidermis from the inner surface of the calyx. с, The epidermis from the outer surface of the calyx (the apex of the lobes). (a, b and с, Obj. 4, Oc. 4). d and e, Forms of hairs from the margin of the calyx-lobes; (Obi. 4, Oc. 6). f, The epidermis from the upper surface of the corolla-limb. 9, The epid- ermis from the lower surface of the corolla-limb ; (Obj. 3, Oc. 12). (Igaliko.) head on a one-celled stalk are found on both surfaces of the leaf, but only in small numbers. (Fig. 4, с.) The Flower. The epidermis on the inner surface of the calyx, near its base, shows strongly thickened and arched cupola-like outer walls, the lateral walls being undulating and with irregular pores (fig. 5, a and b). Towards the apex the epidermal cells have more resemblance to those of the outer surface (fig. 5, с). Stomates appear fairly frequently on the outer surface of the calyx, being most numerous in the 176 Ев. J. MATHIESEN. somewhat depressed areas between the 5 chief veins; here the outer epidermal walls show fine cuticular striations. The margins of the calyx-lobes are covered with glandular hairs, the form of which is illustrated (fig. 5, d and e); the globular or club-shaped head is always filled with a brown substance. The cells of the mesophyll are somewhat axially elongated and are amply provided with chlorophyll granu- les, which also occur in the epidermis of the outer surface. There is a distinet difference hetween the epidermal cells on the upper and lower surface of the limb of the corolla. The outer walls of the upper side are strongly arched and papillose, with cuticular striations radiating from the top of each papilla. In the throat these papillae become very, large, and their contents have here a strong brown colour. On the lower surface of the limb of the corolla, the epidermal cells are somewhat cupola-shaped, the lateral walls being provided with ridges at right angles to the outer walls (fig. _5,g). Similar ridges are also present in the epidermis of the upper surface, but much less pronounced in character, especially near the throat (fig. 5, f). Primula sibirica Jacq. Lit.: Flora Danica, $. 188 and $. 1809. "WARMING, 1886 a, pp. 25, 26, 27; 1886 b, р. 21. Wıpmer, 1891, р. 119. DECROCK, 1901, P.'98. "Pax and Кмотн, 1905, pp. 16, 7127 Beer 1206; р. 569. Alcohol material, collected by Е. Warmıng, from Norway, 1885 (Alten, Bosekop and Käfjord). Herbarium material from the same places and also from Hammersfest (VAHL), the shores of the White Sea (Ancsrrém 1843), Lapponia Imandrae (Brorurrus, 1885) and Kola (BROTHERUS, 1889). The following observations were made in addition to what is stated Бу WARMING and WiDMER as regards the morphology of Primula sibirica: The leaves in the rosette appear to be arranged in a à spiral, as is the case with Pr. egaliksensis, but 3—5 of the Primulaceae. 217 embryonic leaves of the rejuvenation bud are developed as foliage-leaves during the summer. The succeeding leaves remain as scale-leaves, forming a protection for the young tissues of the bud, which in the following spring will develop into foliage-leaves, and terminate in the formation of the pe- duncle. The duration of the adventitious roots is similar to that of the preceding species. Apparently Pr. sibirica never developes more than one peduncle during one period of growth; at least, I have seen no specimens, which show indications of the autumnal rosette developing its peduncle during the same year as the parent- shoot. The peduncle bears only a few flowers, and frequently only one of these attains to full development. But two subtending leaves at least are always present, and in the axil of the one which apparently bears no flower there is a microscopic rudi- ment of a flower-bud (fig. 6, J). Amongst the plants from Bosekop a few specimens were found to have four-partite flowers. The diameter of the limb of the corolla varies from 10—15 mm; the colour is a light lilac (Biyrr), but according to WARMING (1886 b) there is a white variety at Alten. The biology of flower has been fully dealt with by Е. Warmine (1886, а, р. 27), who states as follows: “As regards the specimens from Alten, the following observa- tions have been made: Normally developed, they are distinctly heterostylous, the figures showing the differences. According to my experience, the forms illustrated in fig. 8 (here re- produced as fig. 6) C and D, must be looked upon as the most common forms. In the long-styled flower the stigma is just in the mouth of the tube, level with the dark-yellow, wavy ring, which is present in this species as well as in Pr. stricta and other species. The very same differences known XXXVII. 12 178 Ев. J. MATHIESEN. in other heterostylous species are again met with here. The stigma of the short-styled flower is smaller, with smaller papillae, but the pollen-grains are larger (see fig. С"). On the D! Fig. 6. Primula sibirica. A, Base of plant in bloom, the foliage-leaves are marked in their order from below up- wards, 1, 2, 3, 4; the uppermost, 4, subtends the main hud, which has already developed 2 foliage-leaves (9). В, a foliage-leaf. С, short-styled flower. D, long-styled flower. C1 and D!, the corresponding stigmas, pollen-grains and stigma-papillae, in the same size respectively. Е, a long-styled flower, seen from behind. F, the stigma level with the distal parts of the anthers; the stigma papillae are illustrated to the left. @, a long-styled flower with much protruding style. Н, a similar one, with still Jonger style. Г, anther. J, longitudinal section through the distal part of the peduncie (2°/,); а and a’, 2 subtending leaves; b, pedicel; с, a rudiment of the second flower. (Alten Fjord, Norway). (А—1 from Warming 1886.) other hand the stigma is larger in the long-styled species, the papillae are also larger, but the pollen-grains smaller (DW. Primulaceae. 179 Flowers are, however, also found, which have a relatively much larger pistil, with a stigma reaching above the limb of the corolla, for instance fig. G; I found the most protruding style in the flower illustrated in fig. Н; the pistil is of about the same length as the one in fig. G, but the corolla tube is noticeably shorter, causing a larger portion of the style to protrude; the anthers in fig. Н were not yet open. I have, however, never met with a short-styled flower, the anthers Fig. 7. Primula sibirica. a, transverse section of a thin adventitious root. b, central cylinder with adjoining paren- chyma of а 3-rayed adventitious root. some cortical cells showing contents of starch (a and;b, obj. 8, oc. 6). с, the outer layers of the same root, greater enlargement (obj. 4,: oc. 4) ;}the section taken near the apex of the root. d, the epidermis and outer layers of the cortex of a root from the previous year (obj. 4, oc. 4). (Alten Fjord, Norway.) of which corresponded in height with a stigma protruding to such extent. Finally, I have seen a plant at Käfjord, with anthers and stigma at about the same level, as illus- trated in fig. F; the calyx was here unusually short; the stigma was level with the tip of the anthers, showing its greater part above these; the anthers were open and full of pollen, and germinated pollen was found on the stigma. This was large and its papillae were in size just between those in the common long- and short-styled flowers (see fig. F, to 12* 180 Ев. J. MATHIESEN. the left). This specimen was manifestly a self-pollinator, but this species is, on the whole, as eminently suited for cross-fertilization as our large flowered and dimorphic species.” Anatomy. Two-rayed adventitious roots occur (fig. 7, а), but the most common number of rays is 3 or 4. The epidermal cells perish quickly, but their outer, and especially their lateral walls are fairly resistant, so that well-preserved re- mains of the original outer layer may be seen even in older roots. On juvenile roots, the cortical cells have somewhat thickened walls (according to DECROCK, OI. a rather common feature of Pri- GE © XS mulaspecies). Strong thickenings te are especially noticeable on the С outer and, in some measure, also ue on the lateral walls of the layer SE of cells, directly under the epidermis (fig. 7, а and с). These thickenings, consisting of cellu- lose, had disappeared in the root of the preceding year (fig. 7, 4). AN } и, The cells in the outermost layer Fig. 8 Primula sibirica. Е Transverse section of a peduncle: the Of the primary cortex hitelosely ome and assimilation-tissue, together — together and in close contact with epidermis (obj. 4; 0с.4). (Alten “with the Epidermis, REA Fjord, Norway.) becoming somewhat cutinised with age. The young roots contain some starch in their cortical parenchyma. In the roots of the preceding year the starch was almost exhausted, probably during the spring growth, but I have found small single crystals of a calcium salt, especially in the exodermis (fig. d). The rhizome has in all essentials the same structure as in Pr. egaliksensis, but both Pr. sibirica and the following species are characterized by having sclereids, singly or in small groups, in the medulla of the rhizome. Primulaceae. 181 The transverse section of a peduncle (fig. 8) shows а structure somewhat less resistant than in the case of Pr. egalik- sensis and Pr.stricta. The assimilation tissue is more lacunar, and the stereome is provided with larger and less thick-walled elements. Chlorophyll granules are present also in the epidermis. AN ( ÇA AS YY ae wane Så i 7 J ous 7. — a pate mn ( KU LE SS JW \ Fig. 9. Primula sibirica. a, the epidermis of the upper surface of the leaf. b, the epidermis of the lower surface of the leaf. с, transverse section of the leaf. d, transverse section of the margin of the leaf. e, glandular hair from the upper surface of the leaf; the outer walls of the surrounding epi- dermis cells, with fine cuticular striations. (a, b, с, and а, obj. 4, oc. 4.) (Alten Fjord, Norway.) The Leaf. The lateral walls of the epidermis on the upper surface show faint undulations, those of the lower surface being somewhat more pronounced (fig. 9, a and b). The outer walls are provided with fine cuticular striations (fig. 9, e) which are more strongly marked towards the margin of the leaf where the outer epidermal walls become thicker 182 Ев. J. MATHIESEN. (fig. 9, d). Both surfaces of the leaf are provided with stom- ates, the lower having by far the greater number; these are level with, or slightly raised above the surface, and are sur- rounded by 4—6, most frequently 5, cells. Glandular hairs of the usual Primula type are distributed fairly regularly, though very scantily, over the whole surface. A palisade layer is hardly distinguishable, though the cells of the corresponding layer are almost barrel-shaped (fig. 9, с). The mesophyll is, on the whole, very loose in structure, its bottom layer consisting slightly branched cells. It is amply pro- vided with chlorophyll granules, and these are also found in the epidermis of both surfaces, although only scantily im that of the upper surface. Hydathodes are present on the leaf-margin opposite the stronger veins. As in the case of the vegetative parts, the floral organs of Pr. sibirica differ only slightly from the corresponding parts of Pr. egaliksensis. The difference is limited to the less thickened outer and, especially, lateral walls of the epidermis on the inner side of the calyx (compare Pr. egaliksensis in fig. 5, а, band с). The hairs, most often present on the margin of the sepals, are those with club-shaped heads. Primula strieta Hornem. Lit.: Flora Danica, $. 1385. WARMING, 1886 a, р. 21 and the following; 1886 b, p. 7. LinpMman, 1887, pp. 28, 77. WIDMER, 1891, р. 122. BOoRGESEN, 1895, pp. 223, 225, 229, 236, 237. Ек- STAM, 1897, pp. 166, 167. Pax and Кмотн, 1905, р. 86. SYLVÉN, 1905, р. 125. Buyrr, 1906, p. 568. Alcohol material from the north of Norway (Bosekop near Alten Fjord, Käfjord), collected by Е. Warminc, June 1885. Herbarium material from the same places, and also from Iceland (Restara, 1894, ST. STEFANSSON) and Lapponia Imandrae (А. О. KınLmAann, 1892). According to SYLVEN, Pr. stricta shows, during its first stage of vegetative development, a great resemblance to the closely allied species Pr. farinosa and Pr. scotica. These latter develop | Primulaceae. 183 during the first summer a fairly dense rosette, and numerous adventitious roots arise from the base of the leaves, whereas the ramification of the main root is poor. The hibernation is effected by a winter-bud. In the case of this species, it is not uncommon to find the same rhizome bearing several inflorescences during the summer. The rejuvenation shoot in the axil of the upper- most foliage leaf, as well as the augmentation shoots from buds in the axils of the lower leaves of the rhizome, are able to reach the flowering stage simultaneous with, or some- what later than, the parent shoot. The colour of the corolla is lilac with a yellowish throat- ring, the diameter of the limb being from 5—8 mm. Ex- STAM (p. 116) describes Pr. stricta as being scented, and states (р. 167): “während nur zwei, Polemonium pulchellum und Pr. stricta (mit starkem, unangenehmen Geruch, nach Kjell- man), einen unangenehmen Geruch haben”. LINDMAN ascribes to Pr. stricta “mycket stark, frän lukt, lik den af orchisknôlarna”. The morphology of the flower has been described by Warmine (1886 a), and from this the following is quoted. “The Norwegian specimens (the author is comparing the details of Pr. stricta Hornem. with those of Pr. stricta var. groenlandica Warm. — Pr. farinosa var. д. groenlandica (Warm.) Pax) are slightly protrandrous or perhaps some- times homogamous. The anthers open very early and, as far as I have been able to see, most often somewhat before the stigma is fully ripe. Amongst the very numerous Scan- dinavian flowers, examined by me, I have found, but only in very few cases, the stigma in the precise relative position as in those from Greenland; the rule is, that the stigma protrudes more than in the latter; in some instances the stigma is found level with the tip of the anthers (fig. 10, E), while in others it is as much as 1 mm. above them (fig. 1). 184 Ев. J. MATHIESEN. In the first case self-pollination may yet be possible and comparatively easy, while in the latter, it is rendered very difficult, or well nigh impossible, the flower being in an erect position all through. This greater length of the style 1s not caused by growth during flowering, as, in all the many spe- cimens of buds examined, I have always found the same Fig. 10 A—D, Primula farinosa, var. groenlandica, from Greenland (Itivnek near Holsteinsborg). A, the whole plant and an isolated inflorescence, natural size. B, longitudinal section of a flower. С, the corolla-limb seen from above; the stigma surrounded by the anthers is seen through the throat (/,). D, an anther. E-F, Primula stricta, from Alten Fjord, Finmark. Е and F, longitudinal section of two flowers. G, the limb of the corolla, seen from above, 6h). Н, stigma and anthers are at the same level at *. J, longitudinal section of a long-styled flower. (From Warming 1886.) relative lengths, the stigma being placed at a higher level, than in the specimens from Greenland..... Only once I have found an inflorescence, the two flowers of which might Primulaceae. 185 appear as being short-styled (fig.10,F), and the flowerillustrated in fig. /, when compared with this might be looked upon as the corresponding long-styled form! considering, however, that this fig. J depicts the longest styled flower I ever found, and that the length varies generally between that given in fig. I and in fig. Е, while no flowers are found with the stigma below the anthers, | must look upon the form, illu- strated in fig. F, as being accidental and of no great im- port.” To this I have only to add the following: Although the most common form amongst flowers from Norway is like that illustrated in fig. 10 Е, yet it seems that long-styled flowers (as in fig. 10, 7) are not so very rare. I have also found a few short-styled flowers (as in fig. 10, К). A diffe- rence was noted in the size of the stigma-papillae of the “long” and ‘“short”-styled types, though this was certainly small. Flowers from Iceland all resemble the type in fig. 10, E. The pedicel is lengthened a good deal after the flowering. Anatomy. The adventitious roots are 4-rayed. The epidermis dies away quickly, and the protective function is transferred to the outermost layer of the primary cortex which becomes an exodermis, the cells of which are slightly cutinised on the outer and, to some extent, also on the lateral walls. Single crystals are frequent in the exodermal cells. Distinct Casparian spots occur on the radial walls of the fairly well-marked endodermis. The more vigorous adventitious roots exhibit an incipient development towards the secondary type, in that the middle of the central cylinder is filled with hadromal elements, while the leptome groups increase in size; but no continuous cambium is formed, because the groups of meristematic cells situated between the protohadromal rays do not succeed in joining outside these. Young roots have the cortical cells filled with starch. 186 Ев. J. MATHIESEN. The rhizome quickly dies away from behind. A transverse section discloses the vascular bundles (“steles’’, у. TIEGHEM), circular to oblong-arched, surrounded by an endodermis, and resembling those of Pr. egaliksensis already described. Thick- walled and finely pitted sclereids appear both as groups in the medulla and isolated in the steles. The surrounding parenchyma Fig. 11. Primula stricta. a, the epidermis of the lower surface of the leaf. 6, the epidermis of the upper surface. с, transverse section of the leaf, (а, b, and с, obj. 4, oc. 4). (Alten Fjord, Norway.) has its adjacent cells arranged in radial rows, radiating from the vascular bundle; the cells are pitted and were found empty of starch (June, July). The peduncle has the same structure as that of Pr. egaliksensis. The Leaf (fig. 11). The lateral epidermal walls of the upper surface may be nearly straight or slightly undulating, while those of the lower surface are more strongly undulated. Primulaceae. 187 The epidermal cells of both surfaces have their lateral walls strengthened by thickened bands at right angles to the surface; these originate fairly regularly from the convex side of the wall-foldings (fig. 11, а and 6). Numerous cells with brown contents are found in the epidermis. BoRGESEN states that the number of stomates is somewhat larger on the upper than on the lower surface, but the opposite is the case in the leaves examined by me, where the larger number of stomates is on the lower surface. The stomates both on the upper and lower surfaces of the leaf, are most numerous towards the apex and near the margin. Glandular hairs occur evenly distributed, though rather scantily, over the whole surface of the leaf. As was the case with the two pre- ceding species, the mesophyll is loose and little differentiated in structure (fig. 11,c). The uppermost layer consists of short, thick and often almost isodiametric cells; in the lowest layer the elements of the mesophyll become somewhat ramose and more horizontally elongated. Chlorophyll granules are plentiful throughout the mesophyll, and also in the epidermis of the lower surface, but they are few or entirely absent in that of the upper surface. Hydathodes are present on the margin of the leaf. The floral parts are anatomically very similar to the corres- ponding parts of Pr. sibirica. The hair-covering of the calyx lobes consists principally of the type of glandular hairs illustrated in fig. 5, e. Primula farinosa L., var. groenlandica (Warming) Pax. Lit.: LANGE, 1880, p. 70 (Pr. stricta); 1887, p. 260 (Pr. stricta var. groenlandica, Warming). Е. WARMING, 1886 a, р. 21 etc. (Pr. stricta var. groenland.). RoSENVINGE, 1892, р. 683 (Pr. farin. var. mistassinica (Mich.) Pax). ABRoMEIT, 1899, pp. 37—40 (Pr. farin. var. mistassin.). Pax and Кмотн, 1905, р. 84. 188 Ев. J. MATHIESEN. Alcohol material from Itivnek near Holsteinborg (Е. WARMING, A4]; 1884). Herbarium material from Umanak, Stromfjord, Isortok and Itivnek in Greenland. Pax and Кмотн consider that the small Primula from Itivnek, described and illustrated by Е. WARMING (1886, a) belongs to the species Pr. farinosa, but as a special variety: 6, groenlandica (Warm.) Pax. The Greenland plants of this species differ from Pr. stricta Hornem., (which latter accord- ing to the above authors does not occur in Greenland; see also ROsENVINGE), by having deeper incised corolla lobes, and by a calyx about the same length as the corolla tube, whereas in Pr. stricta the latter is considerably longer than the calyx (fig. 10). The relative length of the calyx tube and that of the corolla seems, however, to vary somewhat, not only in the true Pr. stricta, but also in the Greenland form of Pr. farinosa. No doubt, the plants from Itivnek have the calyx as long as the corolla tube, but, amongst about half a hundred specimens from various other places in Greenland, I found only a few flowers exhibiting similar proportions, the calyx being generally somewhat shorter. The same has been observed by ABROMEIT, who writes “die gelbe Kronröhre ragt etwa 2 mm aus den Kelch hervor”. According to this author the pedicels increase greatly in length after flowering: “Die früher nur 2 mm langer Blütenstiele verlängen sich nach der Blütezeit bis zu 10 mm, und an einer alten abgestorbenen Fruchtdolde konn- ten gegen 24 mm lange Fruchtstiele gefunden werden”. The deeply incised corolla lobes were found in all the specimens of this form, examined by me. Besides in South-West Green- land it also occurs in Labrador, thus sharing the same geo- graphical area with Pr. egaliksensis (acc. to Pax and Knuth). In vigorous specimens, the lower surface of the leaves is mealy (ABROMEIT, ROSENVINGE). The shoot-structure closely resembles that of the nearly related species already described. | Primulaceae. 189 Е. WARMING (1886, a) gives a description of the flower biology, and states in this connection: “All flowers from Greenland, examined by me, showed homogamy, as far as could be ascertained; in a bud, near its expansion, the pollen grains were observed lying loose and fully developed in the anthers still closed, and the papillae of the stigma were as large as those in an older flower; the stigma was in all species level with the middle of the anthers (fig. 10, B); self-pollination is inevitable as the corolla tube is very nar- row, and the anthers lie close up to the stigma; the anthers incline somewhat towards the stigma (fig. 10, 5), and looking into the throat from above, the stigma is found to be sur- rounded by the five anthers (fig. 10, С). Cross-pollination is of course not impossible, but it seems to me, that to be effective, the tubes of the foreign pollen must develop more quickly, than those of its own pollen. I found germinating pollen grains on the stigma, while pollen yet remained in the anthers’. In all the flowers examined by me, I also observed, that the position of the stigma, in relation to the stamens, was as given in fig. 10, В. The limb of the corolla reaches a diameter of 8 mm. As to the anatomy, this form corresponds so closely with the species described earlier, that a few remarks will suffice. Crystals of a calcium salt were also found in the exoder- mis of the root of Pr. farinosa var. groenl. The sclerenchyma of the peduncle is only slightly developed in tiny speci- mens such as illustrated in fig. 10, A, but the plants may develop much more vigorously, and in these the structure of the peduncle is like that of Pr. egaliksensis and Pr. stricta. ABROMEIT states, that owing to the presence of anthocyanin in the epidermis, the peduncle, especially near its apex, also the pedicels and calyces, may be dark blue-black in colour. 190 Ев. J. MATHIESEN. The Leaf. The epidermis of the upper surface exhibits slightly undulated to nearly straight lateral walls, but those of the lower surface are strongly undulated. Thickenings of the lateral walls, similar to those in fig. Та and 6, are also present, though less pronounced. No proper palisade layer is developed, nor in Pr. farınosa var. groenl. does the corresponding layer consist of the barrel-shaped cells, already described in the previous species. The bottom layer of the mesophyll has shghtly branched cells. The mesophyll is liberally provided with chlorophyll granules, and these are also present in the epidermis of the lower surface. The number of stomates is, in this species, much greater on the lower than on the upper surface of the leaf. Primula nivalis Pallas, var. pumila Ledeb. Lit.: KJELLMAN 1882, р. 515; 1883, р. 492. Pax and KNUTH, 1905, p. 102. Herbarium material from arctic Siberia (Pitlekaj and the island of St. Lawrence) collected July 7th and August 156, 1879, during the Vega expedition by Fr. В. KJELLMAN, and by him determined as being the variety pumila Ledeb. The chief difference from the main type, according to this author, lies in the almost entire leaves (1882, p. 515: “Mest afvika de med hänsyn till bladen, hvilka hos hufvudformen dro tätt och hvasst sagade, hos den arctiska formen ater alldeles helbräddade. Emellertid finnas 6fvergangsformer’’). In size they vary greatly; and KJELLMAN records finding specimens (St. Lawrence and Konyam bay), which, as regards vigorous development, were quite equal to the main type, wherefore he does not think that the designation pumila is quite appropriate. My material consisted of specimens with peduncles measuring in height up to 10 cm, and the leaves attained to Primulaceae. 191 5 cm long and 7—10 mm broad. The inflorescences vary up to 10 flowers. In the monograph by Pax and Кмотн, Pr. nivalis var. pumila Ledeb. is considered to be identical with KJELLMAN’s species Pr. Tschuktschorum (1882, р. 516, +. IX), and the name Pr. pumila (Ledeb.) Pax, is consequently given to both. The description although it agrees closely with Pr. T. Kjell- man, scarcely applies to Pr. nivalis f. pumila Ledeb. ac- cording to the material at my disposal here. This latter differs distinctly from the former in having broader leaves, and a longer corolla tube in proportion to the calyx- It also has, all over, a more vigorous habit, and the inflorescence is generally fuller. The two forms have the same area of distri- A B bution VIZ. north- Fig. 12. Primula nivalis var. pumila. eastern Siberia and the A, longitudinal section of а flower; the stigma is А the Behring. |) jum section of à chore styisa dower, 0) (its) Strait. KJELLMAN gives (1883, р. 492, Вос. pp. 490 and 491) the following description of the hibernation bud of Pr. nivalis from Chuckesland: “Det år, såsom synes, en mycket kraftig bildning. Dess längd ar 35, dess tvärdiameter 15 mm. Dess hölja utgöras af ett antal kraftiga, tjocka lagblad, hvilka antagligen i viss man bidraga till de inre delarnas skydd, men som derhjämta, och kanske huvudsakligen, tjenst- göra säsom förvaringsrum för upplagsnäring. Atminstone äro deras parenchymcellar öfverfulla of stärkelse. Innanför dessa sitta knoppskottets assimileranda blad, — de äro stora 192 Ев. J. MATHIESEN. med tydligt urskiljbara delar, — och innerst sjelfa blom- ställningen, hvars hufvudaxel har en märkbar längd och hvars blomknopper äro fullt tydliga, ungefär 2,5 mm länga, och med alla blommans delar ganska långt utbildade . . .”. In this Primula, the main bud also developes foliage leaves during the same summer, in which it is formed. Flowers of type A (fig. 12) were found as frequently as of type В. The stigma in type A is level with the base of the anthers and self-pollination may thus readily be effected. Fig. 12, B shows a short-styled flower but corresponding long-styled form, which might be expected to be present, did not occur in the material examined by me. The stigma- papillae of A were slightly larger on the average, than those of B. The corolla is dark purple in colour. Anatomy. The adventitious roots vary in thickness and in number of rays, and roots with 4, 5 and 6-rays were found. The endodermis is rather prominent, and Casparian spots are very distinct (fig. 13, А). The cells of the primary cortex were devoid of reserve-materials (July), and their walls showed collenchyma-like thickening. The epidermis is thin-walled and perishes quickly. The outermost layer of the primary cortex is developed as an exodermis, with the cells filled with a dark-brown substance. The peduncle (fig. 13, B) differs from that of the spe- cies previously described by the absence of a sclerenchymatic ring during flowering. It appears, however, from a peduncle of the previous period of growth that the cells between, and just outside, the vascular bundles are somewhat thick- ened during the ripening of the fruits. Chlorophyll granules are present in the cells of the exceedingly lacunar cortex. The vessels in the strands are few and narrow (fig. 13, В), whereas the leptome is rather strongly developed. The paren- chyma is thin-walled. Primulaceae. 193 The leaf. The epidermal cells of the upper surface have straight lateral walls, or nearly so, which here and there are provided with flanged thickenings at right angles to the surface of the leaf (е. g. the two upper cells to the right, fig. 14, a). The outer-walls often have fine cuticular striations, especially on the epidermal cells over the midrib. The epi- dermis of the lower surface has undulating lateral walls, ie NEO) x | И DAR Fig. 13. Primula nivalis var. pumila. A, the central cylinder and the adjoining tissues from а 6-rayed adventitious root; (obj. 8, oc. 6). B, the peduncle; the fig. shows a sector of the transverse section with a vascular bundle; (obj. 4, oc. 4). (Pitlekaj). provided with thickenings similar to those already described for the upper surface; these are especially well-marked near the stomates (fig. 14, b). The stomates are exceedingly rare on the upper surface, whereas, on the lower surface, they are fairly numerous, from 90—110 per sq. mm; they are surrounded by 4 or 5 cells. Glandular hairs of the usual Primula type are present on both surfaces of the leaf, and they always originate from an epidermal cell much smaller than the surrounding ones (fig. 14, a and 6). The XXXVII. 13 194 Ев. J. MATHIESEN. stomates are on a level with the surface of the leaf. Numerous cells, filled with a yellowish-brown substance, are present in the epidermis of both surfaces. | In this Primula the upper 1 or 2 layers of the meso- phyll consist of elongated and closely placed cells, so that Fig. 14 Primula nivalis var. pumila. a, the epidermis of the upper surface. b, the epidermis of the lower surface. c, transverse section of the leaf. (a and 6. obj. 4, oc. 6; с, obj. 4, oc. 4). (Pitlekaj.) they may be described as palisades (fig. 14, с). Towards the lower surface, the mesophyll becomes unusually lacunar, and consists of much branched cells extended in the plane of the leaf surface. Androsaces septentrionalis L. Lit.: Flora Danica, t. 7. MÜLLER, 1881, р. 358. BRUNDIN, 1898, рр. 24, 25, 27. P. Кмотн, 1899, р. 307. Pax and Кмотн, Primulaceae. 195 1905, р. 214 SYLVÉN, 1905, р. 126. Bryrr, 1906, р. 566. OsTEN- FELD, 1910, р. 61 (var. Gormannii). Plants preserved in alcohol from Öland (Е. WARMING 315 1907). Herbarium material from Arctic Siberia (the mouth of the Lena River, 1883; Dr. Bunce), and Arctic America (King Point, А. Н. Linpstrém, June 17th and 28th, 1906). The plants from the latter habitat be- longed to А. Gormannii Greene, formerly considered a separate spe- cies, but the only difference from A. sept. is the shorter pedicels; OsTENFELD therefore ranks it with the latter species, but as a special variety (A. septentrionalis var. Gormannii (Greene) Ostf.). A rosette plant of the Draba verna type (BRUNDIN,SYLVEN), generally appearing as a hapaxanthic winter annual, but sometimes it succeeds in attaining to pollacanthic growth (SYLVEN, one specimen from Uppsala). According to these authors, the seeds germinate at once during the summer, and a great number of leaves are developed before winter sets in. These remain green throughout the winter, except the lowest in the rosette, hence they can re- sume their assimilative functions in spring, as soon as con- ditions permit. No special bud-scales are provided for pro- tecting the growing-point and the embryonic leaf-tissues against desiccation during the winter. The peculiar flat form of the rosettes is caused by the upper leaves growing more on the upper than on the lower surface, so that they are deflected downwards, pressing the lower and older leaves closely towards the substratum (Brunpin). During the winter the leaves assume a deep red-brown colour (ibid.). The primary root developes into a thin, slightly branched tap-root; adventitious roots do not occur. The growth of the shoot terminates with a 7—10 flow- ered peduncle; weak specimens develop only one peduncle, whereas stronger ones produce several, through development of the axillary buds of the upper foliage-leaves; on a spe- cimen from Boganida, I have found 19 peduncles, and A. septentrionalis is also depicted with several peduncles both 13* 196 Fr. J. MATHIESEN. in Flora Danica and by Втутт. These secondary axes have often some foliage-leaves developed at their base. Like other winter annual plants, А. septentrionalis is, according to BRUNDIN, sometimes found flowering in the same year as it germinated. The calyx is campanulate with 5 protruding ridges at its base, and its lobes reach somewhat above the corolla-tube. The diameter of the limb was found to be 6 mm on spe- cimens from Oland, 3—5 mm ‘on those from the Arctic; the limb and tube are white. In the throat of the corolla there are 5 projec- tions, alternating with the corolla- lobes; they are sac- like expansions of the throat (fig. 15, D, shows part of the corolla seen from the outside). The centre of each Fig. 15. Androsaces septentrionalıs. projection is some- A, longitudinal section of flower-bud, (%/,). В, a recently opened flower, (5/;). С, a young inflorescence, developed what depressed (fig. from а bud in one of the uppermost leaf-axils of the rosette, (/,). D, part of the corolla, seen from the outside; the 15, A and Е). Н. limb is bent somewhat upwards, to show the throat-projec- Nes tions. Е, the corolla, seen from above, (4/,). (Öland.) MÜLLER remarks, concerning these throat-projections in Androsaces species, that they are yel- low and often very conspicuous, and for this reason he thinks they are of importance as nectar contrivances; they are also useful in narrowing the throat, thus preventing drops of water from finding their way into the corolla tube, and moistening the reproductive organs. The uppermost flat part of the ovary functions as a nectary, and, according to the above author, a drop of honey is here secreted on dry and warm days. Primulaceae. 197 H. MÜLLER has found, by studying A. septentrionalis in the Alps, that visits by insects (various Diptera) are not fre- quent. The inflorescence, which consists of a few rather small and insignificant flowers, attracts but few visitors, hence he assumes, that the plant will often be dependent on self- pollination. The proximity of the stamens to the stigma (fig. 15, A) must greatly facilitate this. Anatomy. In its anatomical structure, the root of A. septentrionalis, according to Ресвоск’з description (р. 121), о ] > ig OM 9%) > О и Fig. 16. Androsaces septentrionalis. А, transverse section of a root-branch; (obj. 4, oc. 6). В, the peduncle; a sector of a transverse section with a vascular strand ; (obj. 4, oc. 4). (Öland.) seems to resemble closely that of A. maxima; no difference was found between the main and the secondary roots. The primary cortex is few layered and thin-walled, and perishes quickly, with the exception of the innermost layer, the endodermis. This consists of comparatively large cells, which, during the development of the root, divide by radial walls, so that the endodermis can keep pace with the considerable growth of the central cylinder (fig. 16, A); the outline of each original endodermal cell is marked by a cutinised lamella. The young root, provided with 2 hadromal rays, quickly attains the characteristics of the secondary type; the one 198 FR. J. MATHIESEN. depicted (fig. 16, А) 15 thus only 0.5 mm thick; the paren- chymatous ground tissue, surrounding the numerous ves- sels, is lignified at the periphery of the hadromal part. In the leptome the sieve tubes seem to be few, but their pre- cise determination is made difficult by the rapid thickening of their lateral walls, which in a transverse section makes them resemble the somewhat collenchyma-like parenchyma cells surrounding them. In the periphery of the root, the parenchyma cells appear to be tangentially extended and are frequently divided by thinner radial walls. The peduncle. A lacunar tissue of axially extended cells, containing chlorophyll granules, is surrounded by a thick, strongly cuticularised epidermis, provided with stom- ates (fig. 16, В); the peduncle is therefore, in this case, an organ with assimilative functions. Inside this tissue is seen a stereome including up to 10 layers of very elongated and thick-walled parenchymatous cells, with finely pitted lateral walls. The first-formed vessels (see fig. 16, B, below) are placed nearly in a radial line, and have only a few ill-defined annular thickenings, so that in the fully developed peduncle, they are found quite flattened by the turgid pressure of the surrounding tissue. The ground-tissue has very thin walls, and often perishes during the ripening of the fruit, when the peduncle becomes hollow. The pedicels, about 0.5 mm thick, are very similar in structure to the peduncle. The leaf. The epidermis of the upper surface (fig. 17, A) has straight, often pitted, lateral walls, those of the lower surface being rather strongly undulated (fig. 17, B). The outer walls show cuticular striations, especially distinct around the hairs. Of these two kinds are present: I. glandu- lar hairs with a one-celled stalk (less frequently two-celled) and with a unicellular head; these originate from a cell which is smaller than the surrounding epidermal cells. Primulaceae. Too II. multicellular, thickwalled and most frequently branched cover-hairs with a warty cuticle; various forms are depicted in fig. 17, D, Е and Г; D and Е showing the thickness of the walls. Stomates, surrounded by 4—5 cells, are present м АЕ еее & Fig. 17. Androsaces septentrionalis. A, the epidermis of the upper surface of the leaf. В, the epidermis of the lower surface of the leaf. С, transverse section of the leaf; the vascular strand to the left shows collen- | chyma-like thickened tissue оп the lower side. D, Е and Е, different forms of cover-hair; D and £ are seen in optical section. G, epidermal cells from the upper surface of the leaf, with contents of spheroidal formations. (С, obj. 4, oc. 4; the rest, obj. 4, oc. 6.) (Öland.) on both surfaces of the leaf, but there are about double as many on the lower than on the upper surface, the propor- tion varying only slightly in plants from different habitats. 200 Ев. J. MATHIESEN. N The total number of stomates per leaf is, however, on an average far greater on specimens from Öland, than on those from Arctic America and Siberia. The apex of the leaf and of each of the 2—4 teeth in its margin, are provided on the lower surface with a group of hydathodes, from which the distal portion of the veins extends fan-shaped as a well-developed epitheme. A transverse section (Fig. 17, С) discloses from 2—-4 layers (the thickness of the leaf is somewhat. variable) of short and broad palisade cells, which like the rest of the mesophyll are rich in chlorophyll granules. The undermost layer of the mesophyll consists of rather strongly branched cells. Frequently the epidermis of the lower surface is found loosened from the mesophyll. The vascular bundles are sometimes accompanied by cells with collenchymatous thicken- ing (fig. 17, C, the vascular bundle to the left). In the epi- dermis of both lower and upper surfaces, though most fre- quently in the latter, there are cells filled with a brown substance. In all the leaves examined by me, but only in the upper surface, this contains some peculiar spheroidal and deeply brown-coloured bodies (fig. 17,G), which were ex- ceedingly resistant towards solvents; they may perhaps be only a precipitate caused by evaporation, or by preserva- tion in alcohol. With regard to the floral parts of A. septentrionalis, the calyx presents a peculiar anatomical structure, so that it deserves further attention. A transverse section (fig. 18, A) of one of the 5 pro- jecting ridges of the calyx (see p. 196) shows a layer of thin- walled cells within the epidermis of the outer surface of the calyx; these cells are in close contact with each other, and also with the epidermis without any intercellular spaces. This layer (probably an aqueous tissue) is seen in fig. 18, B, in surface-view; its elements appear very elongated, with Primulaceae. 201 their ends wedged together, and they contain chlorophyll granules, like the mesophyll. The stomates are limited to the upper part of the calyx (i.e. the lobes and a narrow belt below these), and they are about equally numerous on both surfaces. The lateral walls of the epi- dermis are undulating, more especially on the inner surface of the calyx and also on the 5 concave depressions between the protruding parts of its outer sur- face;. they are less strongly undulated, or nearly straight, on the protruding parts. The limb of the corolla has papilla-like arched epidermal cells, which are most promi- nent on the upper surface, especially so round the throat-scales, but these, however, are not papillary. The flat polygonal cells, which cover the throat-scales, LME © ` 4 АК un, fe Ar = F ur Cy" хе Ил, == ——— as Fig. 18. Androsaces septentrionalis. A, transverse section of one of the protruding parts of the calyx; under the epidermis of the outer surface is seen a continuous layer of thin-walled cells. B, sur- face-view of this layer. C, the epidermis from the inner surface of the calyx. D, the epidermis from the outer surface, (A, В, С and D, obj. 8, oc. 6.) (Oland.) were filled with a brown substance. Androsaces chamaejasme Host. Lit.: Ууогев, 1859, р. 22. MÜLLER, 1881, рр. 358, 359. Horn, 1885, р. 44; tab. VIII. Easrwoop, 1902, p.211. Кмотн 1899, 202 Ев. J. MATHIESEN. р. 307. Pax and Кмотн, 1905, р. 188, etc. OSTENFELD, 1910, p. 61. Alcohol material from Jugor Schar (*ls 1882), collected by Tu. Horn. Herbarium material from Arctic America (var. arctica Knuth, Herschell Island #/; 1906, A. H. Linpstrém); Arctic Siberia (Chaba- rowa, August, 1894, H. Fısuer; Wajgatsch, July, 1875, Fr. В. KJELL- man and А. N. Lunpstrém); also from Central Europe (the Alps, the Carpathians). This plant developes long and short shoots alternately on the same axis (see fig. 19, A) and is similar to A. lactea, which is described by WYDLER, as having the densely leaved rosettes separated by long-jointed internodes. In the species here de- scribed, the peduncles are terminal (thus differing from A. lactea), the reju- venescence must therefore be effected by means of axillary buds (fig. 19, B). Fig. 19. Androsaces chamaejasme. During the summer in A, a small specimen from Herschell Island, (see which the parent-shoot the text); the plant shows the withered remains of 4 peduncles, the one furthest to the left is the forms its peduncle, some youngest, the middle one the oldest, (?/;). В, ro- sette with 3 rejuvenation buds, (2/,). О, те- of these buds develop juvenating shoot, where the change from long : etre internodes to the densely leaved rosettes has taken Into lateral shoots similar place twice, (?/,). (А from Herschell Island, В and in structure to that of С from Wajgatsch.) the parent-shoot. The length of the long-jointed internodes varies greatly, from 5 em down to a few mm, and sometimes even less, so that the plant becomes almost cushion-lke. The rosettes are generally bent towards the ground, or even rest on this, in which case adventitious roots may develop, and facilitate the vegetative extension of the plant; Primulaceae. 203 this development is however rare. The plant (from Herschell Island) depicted in fig. 19, A, has, however, a more erect growth, and this is the more characteristic form of speci- "mens from this locality. There was a considerable amount of clay and sand amongst the withered leaves of the three lower rosettes (fig. 19, A), and it seems probable, that the plant has grown on a spot, where, under certain conditions, its vegetative organs were periodically covered by deposits, thus causing the plant to assume a more erect manner of growth, necessary for maintaining life. The leaves do not seem to retain their assimilative function beyond one period of growth, and the leaves de- veloped during the previous summer were nearly always withered before the following spring. The withered leaves are fairly resistant, and they may remain on the plant for several years before crumbling away. Special bud-scales are not formed, but the young leaves, which are very hairy, especially on the margins, protect the growing point by bending across it. There are two fairly well differentiated types of axial shoots: one has a thicker elongated internode and a densely leaved rosette; the other has a thinner axis and fewer leaves in the rosette, and its growth never terminates by the development of a peduncle. As regards the anatomy, there is a marked difference be- tween the long-jointed internodes of these two types, which will be dealt with later. The flower. The size of the corolla seems to vary greatly (fig. 20, A and В). I found it largest on plants from Herschell Island and from Wajgatsch, a diameter of 12 mm being attained; it was distinctly smaller (6—8 mm) оп plants from other habitats. The colour is whitish-yellow, the 5 small projections narrowing the throat, having a more pro- nounced yellow colour. According to H. Мбмяв, who tho- roughly describes the biology, these throat-scales change their 204 FR. J. MATHIESEN. colour and become red when the flower has reached the stage of pollination. This is effected either by insects (Le- pidoptera, Hymenoptera and, especially, Diptera), or by self-pollination, in the absence of these. The limb of the corolla also assumes a faint pink colour. This difference in colour between younger and older flowers is presumed to be a signal to the more intelligent insects, that the older flowers have already been cleared of honey by earlier visitors, hence they avoid the red flowers, and the result for the plant is a more intensive utilisation of the visits by insects (H. Mütter). Fig. 20. Androsaces chamaejasme. A and B, flowers seen from above; A from Herschell Island, B from Chabarowa, (5/.). С, flower, transverse section, (5/.). Miss А. Eastwoop mentions the change of colour in the corolla, so that it must also occur in arctic localities (Alaska). Honey is secreted on the cup-shaped surface of the ovary (H. MULLER). Anatomy. An adventitious root, 0.5mm in thickness, examined by me, was found to be 3-rayed, but had otherwise the same anatomical structure, as the root of A. septentrionalis. The stem of А. chamaejasme, like that of other Androsaces species with a similar mode of growth [A. lactea (WYDLER, 1859), А. villosa (DEcrock, 1901)], shows а di- stinct difference as regards the structure of the naked long- shoots and the densely-leaved short-shoots. As to the first, fig. 21 А illustrates a transverse sec- tion of a stem, which has carried a rosette with few leaves; Primulaceae. 205 В and С show similar sections of the cortex and cen- tral cylinder, respectively, of a stem, where the rosette has been more densely leaved. Both have this in com- mon, that the inner layers of the cortical parenchyma perish fairly quickly, except the endodermis which by divi- sion of its radıal walls keeps pace wıth the growth of the central cylinder. The outer layers of the cortex acquire thickened and lignified walls, thus forming; а continous Fig. 21. Androsaces chamaejasme. A, tramsverse section of a stem, which has borne a few-leaved rosette. B and C, respec- tively, the outer layers of the cortex with epidermis attached, and the centralcylinder with the endodermis of a stem with many-leaved rosette. (A, B and С, obj.8, oc. 4.) (Jugor Schar.) sclerenchymatic ring. By comparing fig. 21, A with fig. 21, В it is apparent, that the thickness of this ring differs in the two types of shoots; in В, 5—6 layers take part in the forming of the stereome, while in A, this only consists of 1—2 layers. When the shoots have reached a certain age, the central cylinder will be found lying free within a tubular sheath, formed by the lignified and thickened cor- tical parenchyma, together with the persistant epidermis filled with a brown secretion. The cells of the cortical parenchyma are elongated axially, 206 Ев. J. MATHIESEN. and the elements of the external thickened layers are finely pitted, both in the end and the lateral walls. The central cylinder has а narrow medulla, surrounded by a continuous ring of vessels (fig. 21, A and С). А circular, continuous cambium commences its functions at a very early stage of the development, forming new vessels towards the centre. Outwards it forms sieve-tissue, the walls of which soon be- come somewhat thickened collenchyma-like, thus resemb- ling, in a transverse section, the thick-walled parenchyma of the many-layered pericycle; in this, thin radial walls were noticed in a few of the cells (fig. 21, С). The medulla becomes larger in the stem of the rosettes, and the continuous ring of vessels and sieve-tissue, present in the long internodes, is here interrupted by the leaf-traces inserted amongst them. The sclerenchyma in the periphery is here composed of much shorter elements. The peduncle differs from that of the previous species, by having, most frequently, a somewhat less-developed stereome. It also has, like the calyx, long cover-hairs consisting of up to 7 thin-walled cells; in appearance they are minutely warted by cutinization. Glandular hairs with a 1—2-celled head on a 1—2-celled stalk, are also present. The leaf. The epidermis has thin, undulating lateral walls; those of the upper surface are somewhat less undulated than those on the lower surface, though there is no great difference (fig.22, A and B). Glandular hairs, with a structure similar to those on the peduncle, are evenly distributed all over the surface. The cover-hairs are most numerous near the margin of the leaf; they are hyaline, 2—5 cells long, more or less thick-walled, and their surface is closely and minutely warted by cutinization ; on the lamina of the leaf these hairs are few, and the short (2-celled) form is most frequent. The variety arctica Knuth (Herschell Island) is, however, distinguished by being closely covered with long Primulaceae. 207 cover-hairs on both surfaces of the leaf. The epidermal cells from which the cover-hairs originate, are strongly arched. Stomates are present in great numbers on both surfaces. Where the leaves are pressed closely together in rosettes, the proportion between the number of stomates on the upper Fig. 22. Androsaces chamaejasme. A, the epidermis of the lower surface of the leaf. В, the epidermis of the upper surface of the leaf. С. a piece of the epidermis, more enlarged. D and Е, forms of glandular hairs. F, cover-hairs from the margin of the leaf, seen in optical section. С, transverse section of the leaf. H, a single stoma. (A, В and С, obj. 8, oc. 4; Е, obj. 8, oc. 6; С, D, Е and Н, obj. 4, oc. 4.) (Jugor Schar.) and lower surface is almost the same, but they are always somewhat more numerous on the lower surface. On shoots with few leaves (see fig. 19, А, furthest to the right), the stomates are much fewer on the upper than on the lower surface of the leaves: but, in this case the stomates are 208 Ев. J. MATHIESEN. more evenly distributed as compared with the densely leaved rosettes, where they are especially numerous towards the apex of the leaf. The total number of stomates per leaf seems generally greater on plants from Central Europe, than on those from the Arctic Zone. In fig. 22, G, a transverse section of the leaf is given. The outer walls of the epidermis are only slightly thickened, and the stomates are either level with the surface of the leaf, or slightly raised above this. The illustration shows only one layer of palisades, but 2 and even 3 layers may occur (see TH. Нотм, Tab. VIII, fig. 4). The aérenchyma is exceedingly lacunar in structure, änd its cells are strongly branched. Dodecatheon frigidum Cham. et Schlecht. Lit.: KJELLMAN, 1883, р. 517. Easrwoon, 1902, р. 211. Pax and Кмотн, 1905, р. 239. OSTENFELD, 1910, р. 62. Herbarium material from Arctic America (Port Clarence, 22nd and 26th July 1879, Fr. В. KJELLMAN; Herschell Island "Js 1906, А. H. LINDSTROM). The geographical area of Dodecatheon frigidum seems to be fairly limited. With the exception of a couple of records in Oregon and Utah, which Pax and Кмотн con- sider to be doubtful, it has only been found on the shores of Berings Straits and Alaska, and on adjacent islands. According to KJELLMAN, it grows chiefly amongst shrubs on mountain-slopes, but it may also be found on damp plains near the beach. The shoot-structure resembles that of the Primula species already described. The peduncle is terminal, and the bud, which is to continue the growth of the rhizome, developes in the axil of the uppermost foliage- leaf. In the material examined by me, were some specimens with an upward branching rhizome, which showed that the buds formed in the axils of the lower foliage-leaves may also attain to full development under favourable conditions. Primulaceae. 209 The rhizome is closely covered with slightly branched ad- ventitious roots, up to about 2 mm in thickness; both have a bright, brownish colour. The rhizome attains to a considerable length (up to 5 cm) by continuous annual growth, remaining alive for several years; on one rhizome I have counted up to 3 generations of shoots, plainly defined by the remains of the old peduncles. The withered bud- scales are found at the base of the spring rosettes, which consist of long-stalked ovate foliage-leaves 2—4 cm long, 0.7—2 cm broad. About the middle of July the hiber- nation buds are still small. The peduncle carries 2—3 reddish-violet flowers, droop- ing when in full bloom. The pollination process in species of Dodecatheon, according to KERNER (Pflan- zenleben II p. 303), is as follows: The pollen-bearing insect, in search of honey, Fig. 23. Dodecatheon frigidum. forces the cone-like column A, inflorescence (slightly reduced). B, an å 5 anther seen from the inside; the cavities of the of anthers apart with its anther are open. С, dorsal view of an anther, proboscis, and during this mia. en Se process part of the pollen falls out and becomes attached to the visitor. On the next flower visited by the insect, some of this pollen is left on the stigma and cross-fertilization is brought about. Should such fertilization not be effected, by the non-appearance of visiting insects, autogamy will take place. The short filaments become relaxed towards the end of flowering, and cause the anthers to separate somewhat, whereby the space inside the cone of anthers filled with pollen is opened, and the slightest shake causes the mealy pollen to fall on the stigma situated vertically below the anthers, because of the inverted position XXXVII. 14 210 Ев. J. MATHIESEN. of the flowers (fig. 23, А). This figure also shows that the style of D. frigidum only reaches а few mm beyond the anthers. The papillae of the stigma are fairly long (fig. 23, D). Anatomy. The adventitious roots are 3—5 rayed, with the central-cylinder surrounded by a strong endodermis (fig. 24,B). Some cells with thin walls occur scattered in this, especially opposite the hadromal rays. The primary cortex consists of a rathe r thickwalled parenchyma, of 16—20 layers. The outermost layer developes into an exodermis which in tan- gential section (fig. 24, D) shows both axially extended cells and short almost isodiametrical cells; these latter are comparatively thinwal- led. The middle lamella is undulated, and in older roots (these alone were at my disposal) Fig. 24. Dodecatheon frigidum. А, transverse section of а root-branch. В, the central- cylinder with endodermis and a layer of the surround- the cells, and this was ing cortical parenchyma of an adventitious root; the starch contents are shewn in some of the cells. C,the also the case with the outer layers of the adventitious root. D, a tangential ag section of the exodermis, (A, В, С and D,obj.4,oc.4. middle lamella of the it 15 cutinised all round ioral an) endodermis. À trans- verse section of the outer layers of the root is illustrated in fig. 24, С; the fourth cell in the exodermis, from below, is one of the above-mentioned short and thin-walled cells; these are frequently filled with a brown coloured substance, similar to that which characterises the epidermis and the Primulaceae. at root-hairs. All the wall thickenings in the cortex consist of pure cellulose. It is possible, that the unthickened elements mentioned in the exo- and endodermis, have served as passage-cells for water at an earlier stage of the devel- opment of the roots, before the cutinization of the middle lamella reduced their permeability. The fact that the epidermis keeps alive and active as a water absorbing organ for a long time (I have found well preserved root-hairs nearly up to the base of the root), seems to me, at least, to confirm this supposition. Fig. 24, A depicts a transverse section of a root-branch, about 0.5 mm thick; it is two-rayed and both endo- and exodermis are well developed. The vascular bundles of the rhizome are arranged in a ring. The number varies somewhat according to the thickness of the rhizome, and I have counted up to 7 in a transverse section. In shape they vary from round to much elongated tangentially; anastomosis is frequent. They are collateral, and each is surrounded by a strand-sheath, with its cells often divided by thin radial walls. Outside the vascular-bundles, a “réseau radicifère” is formed (see Decrock р. 183 and 196), from which the adventitious roots initiate. The ground-tissue consists of fairly thick- walled, many-pitted cells, containing some starch (July). There was an active cork-cambium directly under the epidermis, on the part of the rhizome formed the pre- vious year; the cells of the cork contained a brown substance. The peduncle. The stereome, so common in the Primulaceae, is very slightly developed here; it encloses a circle of 15—20 vascular bundles, and to some extent, it surrounds each single strand. A quickly perishing ground tissue is present in the middle. Outside the stereome, an assimilative parenchyma with large intercellular spaces is 14* 212 Ев. J. MATHIESEN. found. The epidermis is provided with stomates. Glandular hairs occur on the peduncle as well as on the pedicels (here particularly abundant), and on the calyx. The leaf (fig. 25). The epidermis on both surfaces has undulating lateral walls. Stomates only occur on the lower surface, and are fairly evenly distributed, an Fig. 25. Dodecatheon frigidum. A, the epidermis of the upper surface; to the left the underlying palisade cells are drawn with dotted lines. B, the epidermis of the lower surface. C, transverse section of the leaf. D, a hydathode. (A, В and D, cbj. 8, oc. 4; С, obj. 8, oc. 6.) (Herschell Island.) average about 55 per sq. mm. Glandular hairs, with a unicellular, globular head on a 1-celled stalk, were present on both surfaces. The hydathodes are charac- teristic, each single water-pore (fig. 25, D) resting on a small wart-like base; they are situated on the upper sur- face of the leaf, near the margin and above the ends of the veins. The transverse section (fig. 25, С) shows a single layer Primulaceae. 213 of palisades and an aérenchyma with an exceedingly loose structure consisting of much-branched cells. Chlorophyll granules are plentiful throughout the mesophyll. Summary. A. Growth forms! and shoot-formation. 1. Chamaephytes: Androsaces chamaejasme. A sedentary chamaephyte with a very persistent main root. When many leaf-rosettes are crowded together, Andro- saces chamaejasme shows at times distinct transitions to the cushion-plant form. The adventitious roots appear very sparingly, hence vegetative propagation by root-forming rosettes can scarcely be of much importance. The foliage- leaves die in the autumn, except the innermost, partly expanded leaves. The withered leaves form a protection for the embryonic tissues of the axis, while the hair covering of the young leaves is most probably useful in the same way. The structure of the leaves is mesomorphic. The peduncle is terminal. 2. Hemicryptophytes: All the Primula species described above, and also Dodecatheon frigidum. Perennial herbs with a quickly perishing main root, a vertical rhizome, and with leaves arranged in basal rosettes. The rhizome of Dodecatheon frigidum reaches a compara- tively considerable length, while the Primulae have only short ones, as they perish quickly from behind. The living part of the rhizome of the first four species described (viz. Pr. egaliksensis, Pr. sibirica, Pr. stricta and Pr. farinosa var. groen- landica) consists of a single year-growth only. 1 C. Raunkier: Planterigets Livsformer og deres Betydning for Geografien. København og Kristiania, 1907. 214 2 Fr. J. MATHIESEN. The main bud of the Primula species developes in the axil of the uppermost foliage-leaf, and generally produces a few leaved rosette, almost simultaneously with the flowering of the parent shoot. During the autumn some of the leaves remain in the scale-leaf stage, and as bud-scales form a protection for the embryonic tissues in the winter-bud. In spring these tissues develop into regular foliage-leaves (the spring-rosette) and the growth of the axis terminates with the formation of the peduncle. Buds can, however, sometimes develop in the axils of other foliage-leaves, thus forming a “rhizoma multiceps”. As adventitious roots develop in abundance, and (as mentioned above) the rhizome of the Primulas quickly perishes from behind, it is possible for vegetative propagation of these species to take place. A single individual may disintegrate into a larger or smaller group by the dissolution of the parts connecting the mother rhizome with its branches, so that the latter, by the development of adventitious roots, be- come independent plants. The leaves have a mesomorphic structure. 3. Androsaces septentrionalis appears most frequently as a hapaxanthie winter-annual. The seeds germinate in the late summer, and before the approach of winter the plant has formed a small leaf-rosette, closely pressed towards the ground, and with a strongly developed main root. Bud-scales are not present, but the tip of the axis is protected by the half-developed innermost leaves of the rosette. During the spring following, the axis completes its growth by the development of a peduncle, and in the case of robust individuals, other flowering shoots are developed in several of the axils of the foliage-leaves. The leaves developed during the autumn, hibernate and have a somewhat xeromorphic structure and appearance (the felted hair covering). BRUNDIN states, that Androsaces sep- Primulaceae. 215 tentrionalis, besides being a hapaxanthic winter-annual, may sometimes be seen to flower during the same year as germina- tion has taken place (this is also the case with such plants as Draba verna); according to SYLVEN it sometimes attains to pollacanthic growth. The plant seems thus to be somewhat variable as regards its growth-form. (Therophyte - Hemi- cryptophyte.) B. Flower Biology. The nature of my material has prevented me from ascertaining, whether the inflorescence is formed during the year previous to its development, which is the case with numerous Arctic plants. According to KJELLMAN this takes place with Pr. nivalis (Arctic Siberia) and the same is most probably the case with the other Primula species described here, and perhaps also with Dodecatheon frigidum. The Primula species, Dodecatheon frigidum and Andro- saces chamaejasme have quite conspicuous flowers, those of Androsaces septentrionalis being less so. As to the colour of the corolla, this is reddish-violet to purple in Pr. nivalis and Dode- catheon frigidum, while Pr. egaliksensis, Pr. sibirica, Pr. stricta and Pr. farinosa v. groenlandica have a light purple to almost white corolla-limb with yellow throat-scales. The Androsaces species have white corollas with yellow throat-scales, but in Andros. chamaej. a change of colour takes place after fertili- sation, the corolla then assuming a light purple colour. The uppermost flat part of the ovary serves as a nectary in the two species of Androsaces. The alpine Primulas, amongst others Pr. farin., secrete honey from the wall of the ovary according H. MüLLer, but whether this is the case with the Arctic species, must remain an open question for the present. Pr. stricta has a strong unpleasant scent. As to dimorphism of the flowers, it can be stated, that Pr. sibirica appears to have a characteristic dimorphic hetero- styled flower, with a very perceptible difference in size 216 Fr. J. MATHIESEN. between the stigma papillae and pollen granules of the short- and long-styled flowers. Pr. egaliksensis and Pr. farin. v. groenlandica have homomorphic flowers, the stigma being level with or slightly below the middle of the anthers. The flowers of Pr. stricta may perhaps generally be regarded as homomorphic, with the stigma placed a little above the anthers, but flowers are also found with the stigma raised considerably above the anthers, while others, have it situ- ated much lower than the base of the anthers. Although Г found a considerable number of flowers of Pr. nivalis with a very short style, the corresponding long-styled form seemed entirely lacking, and the greater part of the flowers examined had the stigma placed a trifle below the base of the anthers. АП species must be considered entomophilous; the Androsaces species are visited especially by Diptera (H. MULLER). Failing visits by insects, self fertilisation can probably take place without difficulty in the Androsaces species, Pr. egaliksensis, Pr. farin. v. groenl. and also in those flowers of Pr. nivalis, where the stigma is about level with the anthers. C. Anatomy. I. According to their anatomical struc- ture, the roots of the species examined can be classed into two groups, the first representing the mesomorphic, the second the more xeromorphic type (see DECROCK). 1. The root of all the Primula species has a thin-walled and short-lived epidermis. Its protective function is as- sumed by the outermost layer of the primary cortex, the elements of which fit closely together, forming an exodermis, with the outer walls often slightly cutinised. The primary cortex is relatively an extensive tissue and, in Pr. sibirica and Pr. nivalis, its cells have somewhat thickened walls. An endodermis with Casparian spots can always be distinctly observed. The number of hadromal rays in the central Primulaceae. Be cylinder varies from 2 in the primary, and thin adventitious roots of Pr. sibirica to 6 (Pr. nivalis); 4 seems the usual number in the group Farinosae. The primary structure is retained throughout the existence of the root, though, in very strong roots of Pr. stricta I have found a few hadromal elements apparently of secondary origin. The primary cortex of the young roots is filled with starch, which is consumed during the spring growth. The adventitious roots of Pr. egaliksensis, Pr. sibirica, Pr. stricta and Pr. farinosa v. groenl. developed during the preceding summer, were found to die away; the primary cortical cells collapse and become empty, and in the case of Pr. sibirica, the thickenings of the cell walls dis- appear. In these roots of the preceding year, I have noticed diminutive crystals of a calcium salt, especially in the outer layer of the cortex; these were most frequent in Pr. stricta. The anatomical structure of the root of Dodecatheon frigidum resembles most that of the Primulas. It differs in its characteristic exodermis and in the longevity of the epidermis. The number of hadromal rays in the adventitious roots varies from 3 to 5, but the finer lateral roots have only 2 rays. The cells of the primary cortex have somewhat thickened walls 2. In contrast to the type described above, the roots of Androsaces species pass quickly into the secondary stage. An active cambium ring is developed, and forms numerous vessels and wood-parenchyma toward the inside, and out- wards leptome elements with somewhat collenchymatous thickened walls. The epidermis and primary cortex lying outside the endodermis, are thrown off. Through division along the radial walls, the cells of the endodermis are enabled to keep pace with the increase in thickness of the central cylinder. II. All the species of the genus Primula described here are characterised by an anomalous structure of the rhiz- 218 Ев. J. MATHIESEN. ome (polystely and a “réseau radicifère”); polystely could not be proved in the case of Dodecatheon frigidum. These anomalies have already been thoroughly described by van ТтЕСНЕМ! and others for Pr. nivalis, Pr. stricta, Pr. sibirica and Pr.farin. Stone-cells, single or in small groups, are present in the medulla of Pr. sibirica and Pr. stricta, and the latter also has them enclosed in the individual steles. The cells of the ground-tissue are parenchymatic, coarsely pitted and contain starch. Dodecatheon frigidum has an active cork-cambium in the subepidermal layer of the cortex. III. The vascular bundles of the peduncle and flower stalks are arranged in a regular circle. In most of the species they are surrounded by a sclerenchymatic ring (especi- ally strong in Androsaces sept.); Pr. nivalis seems however to lack this stereome. A parenchymatic tissue is present outside the stereome; it consists of axially extended cells with large intercellular spaces, and is rich in chlorophyll granules. The epidermis is provided with stomates and covered with hairs of the same type as those present on the leaves of the respective species. The central ground- tissue consists of thin-walled cells, and is always homo- geneous. IV. The leaves of the Primula species, and those of Dodecatheon frigidum are provided on both surfaces, with glandular hairs of the usual Primula type (a unicellular head on a one-celled stalk). Androsaces sept. has in addition hairs with two-celled stalks, while in Androsaces chamaejasme the head of the glandular hairs may also be two-celled, the division being always effected by a central vertical wall. The dense mealy covering, so characteristic for the 1 van TIEGHEM: Struct. de la tige d. Primeveres & Groupement 4. Primevères, etc.; Bull. Soc. bot. de France, 1886. van TIEGHEM et Douuior: Polystélie; Ann. sc. nat. Ser. 7, T. III, 1886. For the literature of this subject see also SoLEREDER: System. Anat. der Dicotyl.; Stuttgart, 1899. Primulaceae. 219 lower surface of the leaves of many Primulas and which is known to be a resinous substance secreted from the head of the glandular hairs, is not met with, or only in а very limited degree, in the Arctic species closely related to Pr. farinosa (viz. Pr. egaliksensis, Pr. stricta and Pr. sibirica), but Pr. farin. у. groenlandica may have the leaves “mealy” on the lower surface, at least on vigorous specimens (ABROMEIT). The leaves of the Androsaces species are covered with cover hairs, on the margins and the upper surface, those of Androsaces sept. being short and branched, while on Andro- saces chamaej. they are long, hyaline and simple. The lateral walls of the epidermal cells of Pr. nivalis and Androsaces sept. are straight or nearly so on the upper surface of the leaf: on the lower surface, however, they are undulating. The other Primulae and also Dodecatheon frigi- dum and Androsaces chamaej. have undulating lateral walls on both surfaces of the leaf. Fine cuticular striations were present on the outer walls of the cells of the genus Primula and also in the case of Androsaces sept. Primula nivalis and Dodecatheon frigidum have stomates only on the lower surface of the leaves, but the other spe- cies are provided with stomates on both surfaces. Pr. egaliksensis, Pr. sibirica, Pr. stricta, Pr. farin. у. groenl. and Androsaces sept. have always by far the greater number on the lower surface. In Androsaces chamaej. the number of stomates on the upper and lower surfaces is almost the same, but only where the leaves are crowded together in dense rosettes, so that the stomates, on account of this, are mainly situated towards the apex of each leaf; where the leaves are arranged more favourably for development, as on the few- leaved shoots, then the stomates are much more numerous on the lower surface. The stomates are level with the surface of the leaf, at times even somewhat raised above this. 220 Fr. J. MATHIESEN. Primulaceae. The mesophyll is on the whole loose in structure. Dode- catheon frigidum has 1 layer, Pr. nivalis 1—2 and Androsaces chamaej. 1—3 layers of palisades. The leaves of Androsaces sept. vary somewhat in thickness, and show from 2—4 layers. The Primula species of the group Farinosae are characteri- sed by having the uppermost layer of the mesophyll com- posed of somewhat irregular, ovate or barrel-shaped cells, with large intercellular spaces, so that a true palisade layer can scarcely be said to have developed. Pr. farinosa in the Alps shows, according to WAGNER, the same peculiarity. In the case of Androsaces chamaej. and Dodecatheon frigidum the other layers of the mesophyll consist of richly branched cells, whereas in Primula species and Androsaces sept., it is only in the lowest layer that the elements show any stellate ramification. Chlorophyll is abundantly present throughout the meso- phyll, and in special cases it was also found in the epidermis. As is generally the case with Primulaceae, hydathodes are found at the tip of the leaves as well as on the margin opposite the stronger lateral veins. Androsaces sept. has a small group of hydathodes at the tip of the leaves and also on the lower surface of each leaf-tooth; on Dodecatheon frigidum each hydathode is placed on a small, raised, wart-like base. 12. А List of Arctic Caryophyllacee, with some synonyms. By C. H. Ostenfeld. 1920. XXXVII. 1. Moehringia lateriflora (L.) Fenzl, Verbr. Alsın. tab. zu S. 18 (1833); Агепата I. Г. Sp. pl. 423 (1753). 2. Merckia physodes Fisch. ap. Cham. & Schlecht. in Linnæa Г. 59 (1826); Атепата ph. Fisch. in D.C. Prodr. I 413 (1824). 3. Arenaria ciliata L. Sp. pl. 425 (1755). 1. subsp. norvegiea (Gunn.) Fries, Mant. II, 34 (1839); A. norvegica Gunnerus, Fl. Norv. II, 144 (1772), tab. 9, fig. 1—9; A. humifusa Wahlenb., Fl. Lapp. 129 (1812); A. ciliata, В, humifusa Lange, Consp. Fl. Groenl. 27 (1880). 2. subsp. pseudofrigida Ostf. & Dahl, Nyt Magaz. Naturv. LV, 217 (1918). 4. Stellaria media (L.) Vill., Hist. pl. Dauph. III, 615 (1789); Lange, Consp. Fl. Groenl. 27 (1880). 5. Stellaria longipes Goldie, Edinb. Philisoph. Journ. VI, 327 (1822); Lange, Consp. Fl. Groenl. 29 (1880); S. Ed- wardsit R. Br., Chlor. Melvill. 13 (1823). 6. Stellaria borealis Bigel., Fl. Boston. ed. 2, 182 (1824); Lange, Consp. Fl. Groenl. 28 (1880); S. alpestris Hartm., Handb. Scand. Fl. ed. 2, 132 (1832); 5. calycantha Bong., Veget. Ins. Sitcha 127 (1853). 7. Stellaria humifusa Rottb., Kiobenh. Selsk. Skrift. X, 447 (1770), tab. 4, fig. 14; Lange, Consp. FI. Groenl. 28 (1880). 8. Stellaria crassifolia Ehrh., Hannov. Magaz. VIII, 116 (1784). ? 15* 224 С.Н. OSTENFELD. 9. Stellaria longifolia Mühlenb. in Willd. Enum. Hort. Berol. 479 (1809). 10. Stellaria graminea L. Sp. pl. 422 (1753). 11. Cerastium trigynum Vill., Prosp. 48 (1779); Lange, Consp. Fl. Groenl. 30 (1880); Stellaria cerastioides L. Sp. pl. 422 (1753); C. cerastioides Britton, Mem. Torr. Bot. Club. V, 150 (1894); (?)C. lapponicum Crantz, Instit. II, 402 (1766); Lindman, Svensk Fanerog. Fl. 240 (1918). 12. Cerastium nigrescens Edmondston, Fl. of Shetland 29 (1845); С. Edmondstonu (Watson) Murb. & Ostf., Bot. Notis. 246 (1898); C. latifolium auctt., non L.; C. latifolium, 8, Edmondstonii Watson, in Edmondston, Fl. of Shetland 29 (1845); С. arcticum Lange, Fl. Dan. Fasc. 50 (1880), tab. 2962, ex parte, non Lange, Consp. Fl. Groenl. 31 (1880). 13. C. Regelii Ostf., Vidensk. Selsk. Skr. Kristiania 1909, No. 8, 10 (1910); C. alpinum, y, cæspitosum Malmgren, Spetsb. Fanerog. fl., Öfvers. Sv. Vet. Akad. Förh. Stockholm 242 (1862). 14. C. alpinum L. Sp. pl. 438 (1753); Lange, Consp. FI. Groenl. 31 (1880). ' var. lanatum (Lam.) Hegetschw., Reisen 154 (1825); С. lanatum Lam., Encyclop. I, 680 (1783—84). var. pulvinatum Simmons, Sec. Arct. Exp. Fram 1898—1902 No. 2, 122 (1906); C. alpinum, var. cæspito- sum auctt., non Malmgren. subsp. Fischerianum (Ser.) Torr. & Gray, Fl. North Am. I, 188 (1838); Ostenfeld, Vidensk. Selsk. Skr. Kri- stiania 1909, No. 8, 39 (1910); C. Fischerianum Seringe, in D.C. Prodr. I, 419 (1824). i 15. Cerastium arvense L. Sp. pl. 438 (1753); Lange, Consp. Fl. Groenl. II, 245 (1887). 16. Cerastium eæspitosum Gilib., Fl. Lithuan. У, 159 (1781)5.C. vulgatum L. Sp. pl. ed. 2, 267 (1762), non Fl. A List of Arctic Caryophyllaceæ. 225 Suecic. ed. 2 (1755), 158; С. vulgare Hartm., Handb. Scand. FI. ed. 2, 182 (1820). subsp. fontanum (Baumg.) Gürke, Pl. Eur. II, 223 (1899); С. fontanum Baumgarten, Stirp. Transilv. I, 425 (1816); С. vulgatum, *alpestre Lindbl. in Fries, Summ. Veg. Scand. I 37 (1845); Lange, Consp. Fl. Groenl. 30 (1880); C. vulgare, subsp. alpestre Murb., Botan. Not., 253 (1898). 17. Cerastium maximum L. Sp. pl. 439 (1753). 18. Honckenya peploides (L.) Ehrh., Beitr. II, 181 (1788); Arenaria р. L. Sp. pl. 423 (1753); Halianthus p. Fries, Fl. Halland 75 (1817); Ammodenia p. Rupr., Fl. Samojed. eisural. 25 (1845). var. diffusa (Hornem.); Arenaria peploides diffusa Hornem., Oecon. Plantel. ed. 3, 501 (1821); Halianthus peploides var. diffusa Lange, Comp. Fl. Groenl. 26 (1880). 19. Minuartia verna (L.) Hiern., Journ. of Bot. 37, 320 (1899); Arenaria v. L. Mant. I, 72 (1767); Alsine verna Wah- lenb., Fl. Lapp. 129 (1812). 1. var. hirta (Wormsk.); Alsine verna, var. hirta Fenzl, in Ledeb. Fl. Ross. I, 349 (1842); Arenaria hirta Worm- skiold, Fl. Dan. fasc. 28 (1819), tab. 1646; Arenaria pro- pinqua Richardson, in Frankl. Journ. App. 17 (1823); А. hirta, В, pubescens Cham. & Schlecht., Linnæa I, 56 (1826); Alsine verna, у, hirta et à, propinqua Lange, Сопзр. FI. Groenl. 24 (1880). 2. var. rubella (Wahlenb.); Alsıne rubella Wahlenb. Fl. lapp. 128 (1812); A. verna, var. rubella Hartm., Scand. Fl. ed. 6, 149 (1854); A. verna, В, rubella Lange, Consp. Fl. Groenl. 24 (1880); Arenaria hirta, a, glabrata Cham. & Schlecht., Linnæa I, 56 (1826). 20. Minuartia Rossii (R. Br.) Graebn., in Ascherson & Graebner, Synops. Mitteleurop. Fl., Bd. V, 772 (1918); Are- naria Rossii R. Br., Chloris Melvill. 14 (1823); Alsine R. 226 С. Н. OSTENFELD. Fenzl, Verbr. Alsin. tab. zu 5. 18 (1833); Lange, Consp. Fl. Groenl. 25 (1880). 21. Minuartia strieta (Sw.) Hiern., Journ. of Bot. 37, 320 (1899); Spergula stricta Swartz, Vetensk. Akad. Handl., Stockholm, 20, 227 (1799); Alsine $. Wahlenb., Fl. lappon. 127 (1812); Lange, Consp. Fl. Groenl. 25 (1880). 22. Minuartia arctiea (Stev.) Graebn., in Ascherson & Graebner, Synops. Mitteleurop. Fl., Bd. V, 772 (1918); Are- naria arctica Stev., in D.C. Prodr. I, 404 (1824). 23. Minuartia macrocarpa (Pursh) n. comb.; Arenaria macrocarpa Pursh, Fl. Am. Septentr. 1, 318 (1814); Alsine m. Fenzl, Verbr. Alsin., tab. zu S. 18 (1833); Ostenfeld, in Vid. Selsk. Skr. Kristiania, 1909, No. 8, 37 (1910). 24. Minuartia biflora (L.) Schinz & Thell., Bull. Herb. Boiss., 2 ser. VII, 403 (1907); Stellaria biflora L. Sp. pl. 422 (1753); Alsine b. Wahlenb., Fl. Lappon. 128 (1812); Lange, Consp. Fl. Groenl. 23 (1880). 25. Minuartia groenlandiea (Retz.) n. comb.; Stellaria groenlandica Retz., Fl. Scand. ed. 2, 107 (1795); Arenaria g. Spreng., Syst. II, 402 (1825); Alsine g. Gray, Manuel, ed. 2, 58 (1856); Lange, Consp. Fl. Groenl. 26 (1880). 26. Sagina eæspitosa (J. Vahl) Lange, Tillæg No. 6 ad Rink, Grønland Bd. II, 33 (1857); Consp. Fl. Groenl. 22 (1880); Spergula (Arenaria) c&spitosa J. Vahl, Fl. Dan. fase. 39 (1840), tab. 2289; ? Spergula saginoides, var. nivalis Lindbl., Physiogr. Sällsk. Skr. 328 (1837—38). 27. Sagina intermedia Fenzl, in Ledeb. Fl. Ross. I, 339 (1842); 5. nivalis Lange, Consp. Fl. Groenl. 22 (1880), et auctt.; Fries, Mant. III, 31 (1842) ex parte. .23. Sagina Linnæi Presl, Reliq. Haenk. II, 14 (1831); Lange, Consp. Fl. Groenl. 21 (1880); Spergula saginoides Г. Sp. pl. 441 (1753); Sagina saginoides Dalla Torre, Anl. Beob. A List of Arctic Caryophyllaceæ. 227 Alpenpfl. 189 (1882); 5. saxatilis Wimm., Fl. Schles. 193 (1832). 29. Sagina proeumbens L. Sp. pl. 128 (1753); Lange, Consp. Fl. Groenl. 21 (1880). 30. Sagina nodosa (L.) Fenzl, in Ledeb., Fl. Ross. I, 340 (1842); Lange, Consp. Fl. Groenl. 23 (1880); Spergula nodosa L. Sp. pl. 440 (1753). " 31. Dianthus alpinus L. Sp. pl., 412 (1753). 32. Dianthus superbus L. Sp. pl. ed. 2, 589 (1762). 33. Silene acaulis L. Sp. pl. ed. 2, 603 (1762); Lange, Consp. Fl. Groenl. 19 (1880). 34. Silene tenuis Willd., Enum. Hort. Berol. 474 (1809). 35. Viscaria alpina (L.) G. Don, Gen. Syst. I, 415 (1831); Lange, Consp. Fl. Groenl. 19 (1880); Lychnis a. L. Sp. pl. 436 (1753). 36. Melandrium apetalum (L.) Fenzl, in Ledeb., Fl. Ross. I, 526 (1842); Lange, Consp. Fl. Groenl. 19 (1880); Lychnis apetala L. Sp. pl. 437 (1753); Wahlbergella a. Fries, Summa Veg. Scand. 56 (1845). 37. Melandrium affine J. Vahl, Fl. Dan. fasc. 40, 5 (1843); Lychnis af. J. Vahl, in Fries, Mant. III (1842); M. involucratum (Cham. & Schlecht.), 8, affine Rohrbach, Linnæa 216 (1869); Lange, Consp. Fl. Groenl. 20 (1880); Wahlbergella aff. Fries, Summa Veget. Scand. 56 (1845). 38. Melandrium triflorum (В. Br.) J. Vahl, Fl. Dan. fasc. 40 (1843), tab. 2356; Lange, Consp. Fl. Groenl. 20 (1880); Lychnis triflora В. Br., in Ross, Voy. (1819). " i > i 6 LEER: NER re ur я k 2 в Ar ey Ц MATE DES 15 Caryophyllaceæ. By Eug. Warming. 1920: s an introduction to my description of the morphology, А anatomy and flower-biology of the Arctic Caryophylla- ceæ, Professor С. H. OSTENFELD has kindly prepared the preceding systematie list of the species. My description, which follows, is based: — (1) On the investigations made in the Arctic Herbarium of the Botanical Museum of the University of Copenhagen, and on the alcohol material belonging to the same Museum, which has, in the course of years, been collected by the nume- rous Danish scientific expeditions to Greenland, (P. EBEr- LIN, N. Hartz, TH. HoLm, Cur. KRUUSE, ANDR. LUNDAGER, С. H. OÖSTENFELD, L. KoLDERUP RosENVINGE, С. RYDER, Euc. WARMING, etc.). To this may be added some material which was col- lected in Iceland by Hercı Jonsson and С. H. OstENFELD, and in Scandinavia (Finmark, Dovre, Finse) by Euc. WARMING and others, and a small amount brought home by Swedish Expeditions to Spitzbergen. (2) On the investigations and notes which I have made on living material in Greenland, Scandinavia and the Færôes. (3) On the numerous notes found in the literature on the subject, especially in the Danish, Norwegian and Swe- dish. With regard to these notes the reader is referred to the list of literature at the end of this treatise. 232 Euc. WARMING. Г have divided the following description into four sec- tions, viz. I. Morphology and Vegetative Propagation. II. Leaf Anatomy. III. Adaptations to the Environments. IV. Flower Biology and Notes on Seed-production. I. Morphology and Vegetative Propagation. According to the duration of life, and the forms of the vegetative shoots and the combinations of them, ete., I di- vide the species — for the sake of a general survey — into the following groups: — A. The Melandrium type. Spot-bound,! pollacanthic, semi-rosette plants. B. The Silene acaulis type. Spot-bound species, of which the long-shoots have short internodes, and narrow leaves. C. The Sagina nodosa type. Spot-bound, pollacanthic species, with shoots which must be described as long-shoots with elongated internodes, although the basal leaf-pairs are close-set. Monopodial rosette shoot, at least in some of the species. No propagation by means of aérial runners. D. The Cerastium alpinum type. Spot-bound, pollacan- thic species, the shoots of which also usually have shorter internodes at the base, and must be called long-shoots, but which have terminal flowers (sympodia), and on which aérial runners are developed. Vegetative propagation can take place by means of layers. E. The Cerastium arvense type. Species with shoots like those of D, but which have both aérial and subterranean runners, and a more constant vegetative propagation by means of runners. ' Spot-bound, that is, they have no horizontal shoots or runners, with which they can form “layers” (see Warming, 1884). Caryophyllaceæ. 233 Е. The Stellaria crassifolia type. Shoots like those of E, but with hibernacula. G. The Stellaria media type. Hapaxanthic species. It must be remarked, that there is absolutely no distinct limit between these groups — nor anywhere else in the plant-world — and it is possible that more exhaustive in- 3 Fig. 1. Melandrium apetalum. vestigations will prove, that there are several more species with subterranean runners, than those quoted by me. A. The Melandrium type. Spot-bound, caespitose, pollacanthic species, with typical rosettes on semi-rosette shoots. The uppermost parts of the flowering shoots with elongated internodes usually protrude far above the rosettes (Fig. 1). No markedly pul- vinate forms. 234 Вос. WARMING. To this type belong the following Silenee: — Melan- drium affine, M. apetalum, М. triflorum, Viscaria alpina and Of non-Arctic: Lychnis flos cuculi, Silene otites, etc. Dianthus superbus. Melandrium diurnum, These are all “perennial, usually vigorous herbs with a tap-rootwhich remains during the whole life of the plant; but they are either devoid of or rarely provided with adventi- tious roots, which, when pre- sent, are insignificant. Vegeta- tive propagation does not take place. The larger the individ- the the tap-root, and the deeper | uals are, stronger is does it grow down into the ground. The shoots are semi- rosette-shoots, 1. e. they have a typical rosette at their have base, but otherwise Fig. 2. Melandrium apetalum B arcticum. (Spitzbergen). g, Budson the naked axis, either root-shoots or rather buds from the axils of rosette- leaves which have died and dis- appeared. a, Fragments of dead leaves. b—e, fresh leaves. Hairs elongated internodes (long- shoots). The leaves are large when compared with those of the Alsinee. Scale-leaves are absent. Lateral shoots, similar in structure to the omitted. (E. W.) main shoot, are developed from the axils of the rosette-leaves. All these shoots are erect or ascending. They are most frequently di-pleio- cyclic, 1. e. they remain 2—3 or more years in a vegetative stage, before their internodes elongate and flowers are developed; many undoubtedly remain throughout their Caryophyllaceæ. 235 whole life in the vegetative stage. On an older plant the basal part is therefore composed of barren shoots (rosette- shoots) and of flowering shoots or the basal, dead portions of such shoots (Fig. 2), since at the commencement of the cold season, the shoots die so far down, that only small basal portions remain alive, which bear the buds for the next year. This results in a complex of shoots and portions of shoots which are held together by the tap-root and which have formerly been called “Rhizoma multiceps”, “Radix multiceps” or, by Hj. Nırsson, “pseudorhizom”, “Skott- basis-Komplex’’. In 1918 I suggested the old name “meso- cormus” for it (“caudex”, “Mellemstok”, the French “souche’’). Plants, with the structure described above, are typi- cally cæspitose in habit (plantæ cæspitosæ), and occur singly, often growing scattered on the ground. The flowering shoots can, in the same species, e. g. in Melandrium triflorum, be either very short or long (for in- stance 25 cm.) — according to the prevailing conditions. The germination, in some of the species, has been investigated by SYLVÉN and WARMING. The cotyledons are epigeous and resemble, more or less, the rosette-leaves (cfr. Figs. 5, 8). The first rosette is found immediately above the cotyledons. On cultivating Viscaria alpina I found no develop- ment of lateral shoots in the year in which it germinated, but in the second year lateral rosettes occurred, and already in these there were buds for the lateral shoots of the 3rd order. Three-year-old plants of this species were already in flower; but otherwise the number of years necessary for a plant to spend in a vegetative stage before it can flower, undoubtedly varies greatly, according to the prevailing con- ditions. In nature the vegetative stage appears to last for several years (Sylvén). 236 Euc. WARMING. Melandrium apetalum agrees in its main points with Viscaria (Sylvén). Winter-stage. АП the buds are open, and the youngest foliage-leaves are protected during the winter by the older, withering leaves. How many of these remain green during the. winter, depends evidently on the prevailing conditions, for instance, if the plants are snow-covered during the winter, or are uncovered and perhaps exposed to cold and desic- cative winds. Statements made by botanists regarding the appearance of one and the same species during winter, there- fore vary somewhat. As regards Viscaria I made the follow- ing note in West Greenland, on June 28th, at the begin- ning of spring: “passes the winter in a green condition”. SYLVEN (р. 291) says the rosette of the young plant passes through the winter “in an open rosette-stage”. As regards Melandrium apetalum SYLVÉN says that the rosette-shoot of the young plant appears to pass through the winter in a more or less green condition, and regarding Wahlbergella angustiflora from Lapmark: the plants reared under cul- tivation remained green during the winter (p. 290). Foliage leaves which have remained green during the winter, undoubtedly die immediately upon the commence- ment of the following spring. The old, dead leaves may persist for a long time, until they gradually disappear, the cold Arctic climate having no great putrefactive power. Lateral shoots may be somewhat arching, or even slightly prostrate, at their base. This appears to be most decided in Dianthus superbus in which, from the mesocorm, there may proceed obliquely-placed branches with elongated internodes on which there may even occur, here and there, slender adventitious roots, but it appears to be absolutely certain that no vegetative propagation takes place. BLYTT (Norges Flora, p. 1073) mentions also these lateral shoots: Caryophyllaceæ. 237 “numerous, short, sterile, leaf-bearing shoots,” besides the “ascending or erect... stems.” I have observed similar fea- tures in Viscaria purpurea (in Norway near Finse and in Den- mark), and in Viscaria alpina there even appears to be an attempt towards the formation of scale-leaves on such shoots. In Denmark Lychnis flos cuculi, Silene nutans and Melandrium dioecum have similar shoots. Pulvinate habit. Pulvinate forms develop easily from the typically cæspitose forms when the external conditions are unfavourable. When cold or strong wind checks the growth, the shoots become short and very much crowded, the entire mesocorm becomes very compact and closely set with branches, more or less semi-globular, and the flower- bearing portions of the shoots protrude only slightly above the surface of it. The pulvinate form is only a modification; to the same species may belong individuals with a divergently branched and long-branched mesocorm, and also dwarf-in- dividuals with a pulvinate habit. The pulvinate form occurs at any rate in Melandrium affine and apetalum. Adventitious Buds. On the naked vertical axis below the leaf-bearing mesocorm of old plants, small buds _ are sometimes seen, the origin of which may be doubtful (Fig. 2, g). Perhaps they are adventitious buds on the primary root, but it is more probable that they are axillary buds from former foliage-leaves on the main axis which buds have not had favourable conditions for their development. I have seen such buds in Melandrium apetalum, M. affine and M. triflorum. B. The Silene acaulis type. To this I refer only this one species, which is decidedly cæspitose in habit, and has а tap-root growing far down into the ground, but differs from the Melandrium-type in not having semi-rosette shoots, but long-shoots with short inter- XXXVI. 16 238 Euc. WARMING. nodes. The primary root remains throughout the whole life of the plant, and can become very long. KRUUSE records the length as being 2 m, and SEIGNETTE almost 3 m. The shoots may most properly be called long-shoots, and have close-set leaf-pairs, which do not cross each other at right angles, but grow in such a manner that, for instance, the Ist and the 4th leaf-pair stand perpendicularly above each other. Only a few (usually 4—5) fresh leaf-pairs occur simultaneously on each shoot, and these appar- ently form a small ro- sette at the end of the stem which is usually closely covered with old leaves and fragments of leaves (Fig. 3). The fresh leaves are immediately Fig. 3. Silene acaulis. succeeded by the soli- A, A branch (Norway). B, The apex of tary, terminal flower a branch seen from above; the leaf-pairs, fi following each other in succession, are (Fig. 3, A). indicated by figures, and are, moreover, From the base of shaded in various ways. In the centre are seen two fresh, young foliage-leaves, not yet unfolded (Iceland; H. Jönsson, proceed in all directions, 26.3.1894). C, A branch from northern Sweden. (Е. W.) the tap-root, leaf-shoots which shoots are more or less branched, and are almost equal in height and are, moreover, almost always densely crowded, so that a flattened, semi-globular cushion is produced. SCHRÖTER (pp. 582 and 583) calls this form “Flachpolster.” The branching of the shoots in the interior of the cushion is well-illustrated in Mrs. THEKLA RESVOLL’s Fig. 36, p. 144; a compact tuft is shown “opened” and the branches are spread out. There are scarcely any of the Arctic Caryophyllaceæ which are so typically pulvinate Caryophyllaceæ. 239 in habit: as the present species. Many travellers have there- fore described and illustrated these fresh-green, flatly-arched cushions, which during the flowering-period are covered with pink stars of 5 rays. Figures are found for instance in BONNIER, LAZNIEWSKI, KERNER (Pflanzenleben, Pl. 193), BorGESEN (Botany of the Ferées), and in FEILDEN (Geog. Soc., April, 1898), С. Kruuse (1911, р. 358, Fig. 4) and others. The shoots with their numerous old leaves, are usually so densely crowded in the cushions, that they form a spongy mass, which by its capillarity can absorb a great quantity of water. It is therefore very natural that adventitious roots are developed in their in- terior, which must to a certain degree take their nourishment from the old fragments of the plant’s own leaves. But mention is also made of tolerably loose and divergently branched tufts, which may, for instance, have Fig. 4. Silene acaulis. A small portion of a tuft; about+/,. On some of the shoots due to the nature of the station; the green leaves are seen, sur- rounded by dead leaves. (From у ; Iceland; Helgi Jonsson; 31. 12. among damp moss, or in а parti- 1893.) stems as long as 20 cm.; this is as e. g. when the plant grows cularly shady, damp locality. Tufts, eroded by the wind, are described and figured by Hartz and КвоозЕ. Kruuse (1912, Fig. 41, р. 270) writes: “The tufts reach a diameter of 30—40 cm and а height of 7—10 cm, but the windward is often eroded and dead.” KIHLMANN also records such tufts from Kola. Vegetative propagation under natural conditions is not known to me, but as gardeners are able to divide the tufts, adventitious roots must be so vigorously developed, that propagation by layers can take place. 16* 240 Euc. WARMING. Winter-stage. Heıcı Jonsson (1895, pp. 279 and 285) has described and figured shoots in the winter-stage (1. с. Fig. 3,B, reproduced in this treatise as Fig. 4), and I have, in Fig. 3, B, given an illustration of the apex of a shoot in the winter-stage. Imbedded between the dead (shaded) leaves are seen two young, erect, conical, fresh-green leaves, which cover still younger leaves. But the features evidently vary, according to the prevailing conditions; on a speci- men from a spot recently bared of snow (Upernivik, May 10th; C. Ryder) I observed pale, flabby, dead leaves at the apex of the shoots, but far down between them fresh leaves occurred, some of which contained starch, and also flower- buds were found (Fig. 24). Негст Jonsson also found starch in the living leaves in winter time. KJELLMAN writes (ac- cording to citation in Sylvén, p. 287) regarding young plants under cultivation: “The plant passes through the winter in a green condition, and at the end of the winter, continues its development just where it was interrupted by the com- mencement of winter,” and SYLVÉN himself says: “The plant seems to pass the winter in an open rosette-stage; the outer leaves, however, appear to become brown and withered at an early stage.” The seedling has, according to KyELLMAN (1901), an epicotyl-shoot with elongated internodes. He says that“ 5. acau- lis under cultivation belongs to the type with elongated internodes.” SYLVEN mentions (р. 287) the young plants as remaining in nature in a rosette-stage during the three first years, but as distinctly, although only slightly, elongating their internodes during the following years. TH. REsvoLL (1917; p. 141, Fig. 36) also found that the seedling develops a leaf-rosette, which during the year of germination pro- duces “as many аз 3 shoot-generations of lateral branches.” Not until the 3rd year did the first individual flower. It Caryophyllaceæ. 241 consequently appears that the plant is on the point of transi- tion to the semi-rosette stage. Silene maritima, which so far belongs to the Arctic flora as it is found in Finmark and also in Iceland, appears to have exactly the same morphology as 5. venosa, viz. a numer- ously branched mesocorm with erect long-shoots. If any of the other species should be placed together with Silene acaulis, | believe it must be Arenaria ciliata. C. The Sagina nodosa type. To this group I refer Minuartia biflora, М. arctica, M. hirta, M. groenlandica, M. macrocarpa, М. Rossii, M. stricta, M. verna and the Sagina-species: 5. cæspitosa, S. inter- media, 5. Linnei, 5. nodosa, 5. procumbens, and, for in- stance, the non-Arctic species, S. saxatilis and S. subulata. Arctic species, which I have not been able to investigate more closely, are Alsine arctica (Arctic America) and А. та- стосагра. The above-mentioned species are generally low-growing plants, with the basal portions of the shoots close-set, and the flowering portions of the shoots protruding far outwards. Primary root. Here also the primary root is, almost without exception (5. procumbens), the only root which 1$ of any importance; № holds together the erect branches of the mesocorm. It remains during the whole life of the plant, and may be long, although it is most frequently slender, in accordance with the fact that the plants are small. Although adventitious roots may perhaps occur in all the species yet № appears that only in 5. procumbens do they promote layering. The characteristic feature in this group, when compared with all the others, as far as I have been able to see distinctly from the material to hand, is the fact that the primary 242 Euc. WARMING. shootismonopodial;indeed,it remains vegetative through- out the whole life of the plant, and the flowering shoots occur as lateral shoots of short duration, upon it and upon the other monopodial shoots developed as lateral branches. As the mesocorm is short and has very short internodes, it bears a rosette of foliage-leaves, and for lateral shoots it has partly flowering shoots, with elongated internodes, and partly new monopodial rosette-shoots. In the middle of the mesocorm of a plant of the above-mentioned species, we Fig. 5. Sagina nodosa. A, (about 1/,). In August; I, primary shoot. В, Bulbil in the axil of a subtending leaf. C, In the axil of one of the lower-most pair of leaves there is only a bulbil, in that of the other there is a small shoot and а bulbil as an accessory shoot; likewise in all the following leaf-axils. The branches are arranged according to the usual, peculiar Caryophyllaceæ-mode of branching. D, Germinating bulbils (October). Е, Seedling. F, А foliage-leaf. (Material from Denmark). (Е. W.) therefore always find a rosette, which is formed of narrow. linear leaves, and below this rosette there are often a great number of laterally placed, flowering shoots. These lateral flowering shoots may be most appropriately called semi- rosette shoots. À specimen of such a plant is seen in the accompanying figure (Fig. 5) of Sagina nodosa. А monopodial primary shoot and monopodial shoots of Caryophyllaceæ. 243 higher order are also found in Sagina procumbens and 5. subulata, but hardly in all the Sagina-species; Г do not think they are found in 5. apetala for instance, nor in every species of Minuartia. It is often difficult to decide the exact condition, especially if the material to hand is not good. It is very desirable that the above-named and other species should be thoroughly investigated. In monopodial species a terminal flower has excep- tionally been found on the main axis, but it expands later than does the first flower on the lateral axes (Irmisch, Wydler). The monopodial structure has been treated of by the follow- ing authors: Ar. BRAUN (Flora, 1843); [вмтзсн (1848, pp. 529 and 561; Bot. Ztg., 1850, р. 297); Wypıer (Flora, 1851, р. 328; 1859, р. 314); Gopron (Flore de France, vol. 1, p- 245). The leaf-pairs in the rosettes alternate with each other at acute angles, so that, for instance, first the 4th or 5th leaf-pair stands under the Ist. The flowering lateral shoots appear, as a rule, to last for two years only (this was observed in Minuartia biflora, groenlandica and verna), but cases in which they last a longer time may undoubtedly occur. In addition to the flowering lateral shoots, sterile mono- podial shoots, like the primary axis, may also be developed. Sometimes, some of these become prostrate, for instance in Sagina nodosa, 5. procumbens, S. Linnei, Minuartia biflora and M.verna, and may then strike roots. With this there occurs a possibility for the formation of layers, and this may happen, at any rate in the non-Arctic Sagina subulata. As regards Minuartia biflora and verna THEKLA RESVOLL states decidedly that she has not observed vegetative pro- pagation; but it is possible that this may take place once in a way in М. verna. In Sagina procumbens, on the other 244 Euc. WARMING. hand, vegetative propagation undoubtedly takes place by means of its rooting aërial shoots. To a certain degree it merits a special place, side by side with the Stellaria- type. | баста nodosa has a peculiar mode of propagation. In the axils of the foliage-leaves small, few-leaved shoots are developed (see Fig. 5): these shoots fall off and, on striking roots, form new individuals. For further details the reader is referred to the explanation of Fig. 5. The Seedlings of Sagina nodosa immediately develop a rosette-shoot which sometimes forms branches during the first year. Itsfirstinternodes, however, aresometimeselongated (Fig. 5, В). In plants of Minuartia stricta, cultivated by me, small tufted plants were developed in the course of two months, lateral shoots with short internodes being quickly developed in some of the leaf-axils. In M. biflora THEKLA RESvoLL also found lateral shoots (Fig. 32), but SYLVEN found none (loc. cit., p. 304). In other respects the seedlings of Sagina Linnæi, Minuartia verna and M. intermedia appear to correspond with those of М. stricta and Sagina nodosa (see Sylvén, p. 266 and Thekla Resvoll, Fig. 29). Аз regards the time preceding the first flowering period THEKLA RESVOLL states, that it is not probable that Sagina intermedia flowers until the third year, but that Minuartia verna, when cultivated, had flowered in the 3rd year, and even in the 2nd year though sparingly, and that seedlings of M.biflora, when cultivated, flowered in the 3rd year; but, according to SYLVÉN, it appears that in nature the first vegetative-stage (“Forstærknings-Stadium”) lasts for several years. It seems evident that the length of time is dependent on the prevailing conditions. Pulvinate forms. The above-mentioned species are plants with a tufted habit, which may sometimes form dense Caryophyllaceæ. 245 cushions. In the same species the branch-systems may either be long and divergently placed, with flowers raised high above the ground, or they may be so short and crowded together that dense semi-globular or globular cushions are formed with flowers, lying almost upon the surface of the cushion, or resting upon it. In Fig. 6 three specimens are shown which belong to the form-cycle of Minuartia verna, and A С В Fig. 6. Minuartia verna (reduced). A, var. hirta. W. Greenland, 74°18’ N. lat. (С. Ryder). В, var. hirta. Scoresby Sound (N. Hartz). C, var. rubella. Spitzbergen (Hanna Resvoll Holmsen). which show all these modifications within the same species. It is evident that the forms are dependent on the life-con- ditions prevailing in the stations. A pulvinate habit has been found to occur in Minu- artia verna (Fig. 6), M. biflora, М. Rossi (Fig. 7), Sagina Linnei, S.cæspitosa, S.intermedia and S. procumbens (as also in 5. subulata). See figures in Simmons, Pl. 6, pp. 4—6 and in OSTENFELD and LUNDAGER, p. 19. In his excellent 246 Euc. WARMING. treatise “Das Pflanzenleben der Alpen” С. ScHRÔTER (рр. 569—584) has given an account (accompanied by several figures) of plants with a pulvinate habit, and on p. 579 he has figured a cushion of Alsine (Minuartia) sedoides; he Fig. 7. Minuartia Rossti. (Dragon Point; 21. 1. 1916; Thorild Wulff). A cushion seen from above. At the bettom the primary root is seen to protrude. describes it as “halbkugelige Polster vom kahlem Azorella- Typus mit völlig glatter Oberfläche” . . . “Die Früchtchen sind in das Polster eingesenkt,” etc. The species “gehört zu den letzten Pionieren der Vegetation.” The Arctic cushions may be fairly flattened and carpet- m Caryophyllaceæ. 247 like, without, however, being in reality “wandering.’ Slen- der, adventitious roots may occur in the interior of the cushions. The Winter-stage. The same is the case here as in all the other Arctic Caryophyllaceæ: the buds are open; in the autumn there are fresh, green leaves, but during the winter many, or we may say, the majority of them, wither; nevertheless, the species may be called evergreen. At the commencement of spring, or of the time for the opening of the buds, the remaining leaves wither quickly. The degree of greenness of the plant is dependent on the station, espe- cially if this is snow-covered or not. D. The Cerastium alpinum type. To this may be referred: Cerastium alpinum, C. cæspi- tosum, С. nigrescens, С. Regelii and Arenaria ciliata. These also are typically spot-bound species, but there are no monopodial shoots, these being now if anything long-shoots with elongated internodes, even if the prevailing conditions may sometimes produce semi-rosette-like forms. Here also the internodes are shortest at the base of the shoots, and as branches arise more particularly from the base, meso- corms of the usual type are developed; the barren or flower- ing branches are held together by the primary root which remains during the whole life of the plant, but the branches of the mesocorm are not particularly close-set, and the basal shoots are often lying immediately upon the ground, spread out in all directions (cf. Fig. 9). This is probably an instance of psychrokliny, 1.е. the direction of the stems is induced by the low temperature in accordance with VOoECHTING’s view. Added to this, the ground may possibly have been damp and this may have produced root-development; the pressure of the snow covering may also have had a part in it. 248 Euc. WARMING. Adventitious roots are frequently developed on the prostrate shoots, but vegetative propagation does not take place, or rarely does so, through their agency. Here also the habit of the plant is properly speaking tufted; to what a high degree the length of the shoots is dependent on the prevailing conditions, is evident from SYL- vEn’s observations (Cerastium alpinum, pp. 299 and 300). Fig. 8. Cerastium alpinum. A, Young plant (from cultivation); about 1/1. В, Basal portion of the cotyledons; buds are seen in both axils. С, Older plant (from cul- tivation; the internodes are less elongated than in A; after the cotyle- dons, the typical Caryophyllaceæ-branching appears, first with a vigorous shoot in the axil of the one leaf and a less vigorous in that ot the opposite leaf, then with bud only in one of the axils of the two opposite leaves; in A—C the hairs are omitted. D, A leaf with its covering of hairs. Е, Young shoot in spring. F, From №. В. Greenland; winter-stage, 21. 5. 1908 (А. Lundager). A pair of leaves with axillary shoot; there are no winter-buds nor any special covering for the buds. G, Branch of Cerastium alpinum Y cæspitosum (from E. Greenland). Secondary roots are developed; the hairs are omitted. (Е. W.) In nature the young plants appear to pass the winter “in a rosette-like-stage,” he has, however, found the form glabra to develop in the first year‘ directly elongated internodes.” The young plants which had been reared under cultivation Caryophyllaceæ. 249 had, on the other hand, an elongated, richly-branching and 29 plagiotropic epicotyl-shoot.” This was also the case with the young plants reared by me under cultivation (Fig. 8). The commonly occurring Arctic species Cerastium alpi- num, which varies exceedingly in colour and hairiness, may serve as a type. It may in certain conditions form very Fig. 9. Cerastium alpinum. A, From W.Greenland, 29th July, 1884. B, f. pulvinata. From E. Greenland, 25th Aug., 1900. (Hartz and Kruuse). dense cushions, with short flowering shoots; but long, erect shoots, as much as 25 cm in length, may also be developed from the mesocorm (Fig. 9). Cerastium nigrescens appears to behave like С. alpinum. It may be divergently branched, with long, erect shoots — and this is the rule — or it may be densely branched, even 250 Еос. WARMING. — in rare Cases — so densely, that it becomes somewhat pulvinate. С. caespitosum with its variety alpestre varies in the same manner as C. alpinum (Sylven, p. 300). It has some- times very long stems. Cerastium Regelii I have only seen divergently branched. At the apex of the shoots a kind of rosette is often met with, in that the internodes here are short temporarily, and a kind of winter-bud is developed, as for instance in Cerastium alpinum; the youngest leaves and flower-buds of the latter must thereby and by means of the densely villous- haired leaves be provided with an excellent winter-covering (Fig. 8, F). It is not always that such a shoot-development appears. Similar features are met with in Arenaria ciliata. The stems with elongated internodes may be erect, but mayalso be more or less prostrate and entangled, covered with white flowers. Pulvinate forms may develop, although not so decidedly as in the Minuartia’s and т Silene. In other respects, it most resembles the latter in the form of its branches and the position of its flowers. It may have several (apparently lateral) flowers upwards along a long axis. I regard this as a sympodium, but have observed cases which were doubtful. Vegetative propagation it appears does not generally take place. Neither THEKLA REsvoLL nor I have observed such in С. alpinum, although this species may have adventi- ce tious roots (Fig. 8, (С). SYLVÉN, however, mentions “sub- terranean runners with scale-leaves (or scale-foliage-leaves)” in С. alpinum. Nor has THEKLA Везуотл, found vegetative : propagation to take place in С. nigrescens (С. Edmonstonit). The seedlings. SyLv£n found that seedlings of Cera- stium alpinum when growing in nature hardly flowered “until after the lapse of 2 years.” The seedlings of Arenaria Caryophyllaceæ. 251 ciliata 3 norvegia developed in Lappmark, in the year in which they germinated epicotyl-shoots with long inter- nodes which often branched in the first year, but not until after the lapse of some years did they flower. Also under cultivation by me, they branched very soon. Pulvinate forms. It is evident from what has been said regarding the form of the shoots, that there are not Fig. 10. Cerastium alpinum. The youngest part of the shoots are densely covered with woolly hairs. (From Iceland; Helgi Jönsson; 16. 1. 1894.) such good conditions for the development of a pulvinate habit as in the first three types. It does occur however (Fig. 9). Cerastium caespitosum has been found in East Green- and with globular cushions, and Ровзию (Medd. om Grönl., vol. 50, p.370, Fig. 12) has figured a form regarding which he says “it differs greatly from the numerous forms of the species;” “especially distinctive is the low, densely-tufted growth, and the glabrous, obtuse, rosette leaves, etc.” There is also mention of dense tufts in Cerastium nigrescens, and regarding Arenaria ciliata LUNDAGER records “dense and 252 Euc. WARMING. large tufts.” But these forms are hardly typically pulvinate in habit. Adventitious roots may be found in the interior of dense mesocorms. Winter-stage. Cerastium alpinum appears to be parti- cularly evergreen. Негст Jonsson examined it in East Iceland on January 2nd; he mentions the dense hairy-cover- ing of the protecting leaves (see Fig. 10). The living leaves contained starch. He found living leaves also on C. caespi- tosum, as SYLVÉN did in Central Sweden. Likewise, there evidently occur living leaves on С. alpinum, according to material preserved in alcohol, both at Upernivik on the west coast of Greenland (C. Ryder) and in N.E. Greenland (Lundager). But I found no starch in them. LUNDAGER and OsTENFELD (1917, р. 20) write about С. alpinum in N.E. Greenland: “When the flowering plant, in the autumn, is suddenly covered with snow, which remains during the winter, all the parts of it are so well preserved, that in the spring, when the snow has melted, they appear again and apparently are as fresh as if they had quite recently un- folded themselves.” The older branches, as in the other types, are apparently dead, because their bark is grey and wrinkled. This is most distinctly seen in the prostrate branches; but in the spring, fresh leaves develop from their terminal and lateral buds. It is true that the bark-layer is dead, but on loosening it, a fresh, green axis is found under it. In Denmark the same thing happens in nature. I have described and figured the wrinkled, older, subterranean runners which are also found in Honckenya. Perhaps this phenomenon should be regarded as a means of protecting the stem against transpiration due to cold winds. ” Caryophyllacee. 253 E. The Cerastium arvense type. This type differs from the preceding, by the fact that the prostrate shoots with elongated internodes, have not so short internodes at their base, that they may have there rosettes or rosette-like shoot-forms; also they develop roots so abundantly that they might justly be termed “creeping, and vegetative propagation (layering) must be able to take place both by means of aérial runners and — in some species — by subterranean runners. It is a further development of the Cerastium alpinum type. To this I refer: Cerastium trigynum, C. arvense, Moehrin- gia lateriflora, Stellaria longipes, S. borealis and Honckenya peploides. To these may be added of non-Arctic species, for instance Stellaria graminea, S.nemorum (Warming, 1918, Fig. 14), 5. holostea, S. uliginosa and 5. glauca, each with its special peculiarities. When, under the Cerastium alpinum type, mention was made of pale, subterranean, plagiotropic shoots, they were probably aérial shoots which had been accidentally covered over with wind-carried soil and leaves, or had been developed in the shade. Those we are now discussing are, on the other hand, typical subterranean runners with elongated inter- nodes, pale in colour, with scale-ieaves and furnished with an apex which can push its way into the soil; but cf course the boundary line between these two types is not sharp. Cerastium trigynum we will first mention. It has, on the whole, the same mode of branching as C. alpinum, but the stems have their internodes, also those at the base, as a rule, elongated. SyLvEN mentions, however, rosette-like shoots in the young plants in their first year, and THEKLA Везуотл, (loc. cit. р. 137, with Fig. 11) records that, in the snow-fields, the internodes of the primary shoot remain short even for а longer period. Branches, at any rate, are developed XXXVII 17 254 Euc. WARMING. most abundantly at the base, and therefore, numerous branches proceed from а central point, which is the upper end of the primary root. They may, in some cases — at any rate as long as they are still short — be more or less erect, with curved and rooting branches, but soon they lay themselves upon the ground, spreading out in all directions and rooting as aërial shoots, and in а comparatively short time they may attain a length of 10—25 cm, or even more. Ultimately many slender adventitious roots are thereby fastened to the ground, and by this means vegetative pro- (i x Wy LAS Fig. 11. Cerastium trigynum. А runner (about :/,) As far as to L, reckoned from the flower, the foliage-leaves are fresh, then come dead foliage-leaves which subtend small shoots. Two leaves have been magnified. (Iceland; C. H. Osten- feld). (E. W.) pagation may take place (Fig. 11). The apices may be curved slightly upwards and the leaves may be unilaterally curved (Fig. 11). From these runners a few or several flowering shoots rise into the air; their duration of life is two years (Thekla Resvoll). Sometimes the shoots are so richly branched and so closely interwoven that broad, but low cushions may almost be formed. Cerastium arvense has erect or ascending long-shoots, with short, sterile shoots at rts base. It also develops plagio- tropic shoots quickly, and the seedling may, under cultiva- tion, develop, in the year of germination, a richly branched shoot-system, with typical scale-leaf-bearmg runners. It may be densely branched at its base. Caryophyllaceæ. 255 Stellaria longipes. The primary root lives for a long time. This species has prostrate shoots with very long inter- nodes, with a few or several adventitious roots, and with very distinct, long and thin, scale-leaf-bearing runners, each with a straight apex. Stellaria borealis. The stems are thin, spreading and often very long. From the basal nodes there proceed typi- cally pale, very slender subterranean runners — shoots bearing scale-leaves. Moehringia lateri- flora has long, thin, prostrate and rooting shoots, with long in- ternodes, and, as far as: could be observed, thin subterranean run- ners, with slender roots proceeding from the Fig. 12. Stellaria crassifolia. Its winter-buds are densely filled with starch. (From Tromso). (EK. W.) leaf-axils. Honckenya peploi- des is the species which has the most decidedly typical, strong and far-spreading subterranean runners. The Arc- tic specimens appear to agree exactly with the Danish as regards the peculiarities of their runners, their serially-placed axillary buds, their dwarf-shoots with as many as 45 close-set leaf-pairs, etc. (described and figured by WARMING in 1877—79, see also 1891, 1906, 1918, Fig. 12). The germination is described by Jon. Erikson (1896). When growing on beaches, the cotyledons and the lowermost internodes of the epicotyl stem are bu- ried in the sand by the wind, and are thereby, and perhaps also by root-contraction, carried down into the ground. In a 256 Euc. WARMING. each of the axils of the cotyledons, most frequently three serial buds are developed, the uppermost being the largest. Pulvinate forms. In these species, in which the shoots with elongated internodes immediately become pro- strate and spread out in all directions, the typical develop- ment of semi-globular cushions must be very difficult, if not impossible. In places where there is active growth, flat cake-shaped forms can however occur, being formed by the interwoven stems. Fig. 13. Stellaria humifusa. A, A branch with both dead and fresh leaves; a, a fresh-green portion. B, An old leaf, the veins are distinctly seen. (Upernivik). С, D, a leaf magnified and а seedling. (Egedesminde; 28. 7. 1884.) (E.W.) F. The Stellaria crassifolia type. To this Г refer Stellaria crassifolia and — with doubt — S. humifusa. This type, which is represented by S. crassifolia, differs from the Cerastium arvense-type by developing small aérial (and? subterranean) hibernacula, organs which live through the winter and consist of short, thick-leaved shoots, rich in starch (Fig. 12). They are developed during the autumn (late summer), are liberated the next year, and then grow out into new plants by developing adventitious roots. Otherwise it resembles Cerastium trigynum in its mode of growth; the leaf-shoots with long internodes proceed from the same Caryophyllaceæ. | 257 root and lie prostrate along the ground, and are of a length of about 25 cm or perhaps even more; some аге subterranean. These hibernacula are described by Norman (1893) in his “Flora Arctica”, р. 26 (Christiania Videnskabsselskab), and are figured by О. Norpstepr in 1909 (Bot. Notiser, р. 51). The germination is described by SyYLvEn (р. 298). A seedling develops in the first year a plagiotropic, very richly branched aërial-shoot-system, of which some of the shoots often become subterranean by being covered with sand. Stellaria humifusa should perhaps be most properly placed by the side of Cerastium trigynum. From the base of the primary stem proceed, as in Cerastium trigynum, numerous prostrate, richly branching and rooting, entangled branches. On these are found small, thick-leaved shoots, with short internodes, which are no doubt winter-shoots, and when the buds open in spring, they do not seem to elongate. In specimens collected on July 27th, far north- wards (Upernivik), some of these shoots had unfolded and flowered, while others had not (see Fig. 13). This species has a distinct, fairly long, but slender pri- mary root. In addition to the creeping aërial shoots, sub- terranean shoots may occur. With the exception of the above-mentioned short shoots, it is hardly to be called evergreen. The shoots flower after a vegetative stage varying from two to several years. G. The Stellaria media type. The last growth-form is represented by the present species, which is widely distributed, but cannot be included among the truly Arctic species. It has, however, been found in several places in Greenland, north of the Arctic Circle, 258 Euc. WARMING. and, according to PorsiLp, far from present and former settlements, and in other places for instance together with Ranunculus hyperboreus. It 15 a hapaxanthic, summer-or winter-annual species, and does not appear to deviate from this rule in Arctic countries. The fullest account of it has been given by Wırrrock (1908). According to him, in Central Sweden, the winter-annuals may begin to flower even during the autumn; they continue flowering in the milder period of the winter and bring it to a termination in the spring, with an abundant seed-production. There, the winter-annuals are the most vigorous, owing to the fact that numerous adventitious roots are developed during the autumn, on account of the increased dampness of the soil. Wirrrock’s figure shows that the adventitious roots proceed from the leaf-axils, and that they may branch freely. A very distinet account of these con- ditions is given by Korsmo (Fig. in his valuable work “Ugræsplancher”). If the branches have been detached from the parent-stems, in some way or other, vegetative propaga- tion may thereby take place. In such a case it may perhaps also become pollacanthic, as an observation made by Box- NIER also seems to indicate. Other species also, in addition to being pollacanthic, are sometimes described as hapaxanthic, for instance Sagina procumbens and Sagina Linnaei. It is not improbable that this may be the case. I. Leaf Anatomy. With regard to the anatomical structure of the Arctic leaves the reader is more particularly referred to BONNIER and BORGESEN, and in “Meddelelser om Grönland’, vols. 36 and 37, to a series of treatises by HENNING E. PETERSEN, Caryophyllaceæ. | 259 О. GALLÔE, A. MENTZ, KNup Jessen and MATHIESEN: See also Tu. Hozm; LAZNIEWSKI may also be consulted. The foliage-leaves are dorsiventral or almost isolateral (central). The difference in the leaves is evidently connected with their breadth and the direction of their growth. In the broad- leaved species (the Melandrium-type) dorsiventrality prevails with palisade-tissue on the ventral side; see for instance + ® Or р © SE LS Pr PIE LET UT Fig. 14. Viscaria alpina. (West Greenland; 24. 7. 1887; Lytzen.) A, In the transverse section of the leaf is seen on the dorsal side a large lacuna, and on the ventral side a distinct palisade-tissue (sa): the crosses (xx) indicate cells with crystals of calcium oxalate; round the vascular bundles are sheaths. B, Transverse section of a leaf. C and D, Details from ventral and dorsal surfaces, more highly magni- fied. The epidermis of the ventral surface is distinctly higher than that of the dorsal surface; there are stomata on both surfaces. Pores are seen in the lateral walls of the epidermal cells. E and F, Epidermis of ventral and dorsal surfaces; the walls are highly porose. G and H, Section through the palisade-tissue and the spongy-tissue. (E. W.) Viscaria alpina (Fig. 14 and the explanation of the figure); here the dorsiventrality is expressed by the fact that on the ventral side there are two layers of palisade-cells, while on the dorsal side there is a spongy-tissue of rounded cells, with 260 Eve. WARMING. a large lacuna (la Fig. 14, А, В and D); then again in the height of the epidermis of both surfaces (B, C and D); in the form of the epidermal cells (E and №); in the form of the cells of the mesophyll (G and H); and lastly in the height of the cells which form the sheaths of the vascular bundles, since these cells are higher on the dorsal than on the ventral side (B). ae Similar, but not quite so decided dorsiventrality is i; Fig. 15. Melandrium involucratum. (From West Greenland; 30. 7. 1887; C. Ryder.) A, Transverse section of leaf; the ventral side (v) has higher epi- dermis than has the dorsal side; compare the somewhat more highly magnified В and С. D, Hair. Е, Epidermis of ventral surface and PF, of dorsal surface. (Е. W.) found in Melandrium involucratum (Fig. 15, A, В and С, as also EF and F). Also in the much more narrow-leaved species Silene acaulis decided dorsiventrality is found; this is probably connected with the fact that the leaves are more or less horizontally placed (Fig. 3). Here also the dorsiventrality is expressed in the structure of the mesophyll (Fig. 16, А); in Caryophyllaceæ. 261 the epidermis (it is devoid of stomata and is higher on the dorsal than on the ventral side: A, С, H and J; a few stomata, however, may occur towards the edge of the leaf), and in the sheath round the vascular bundles (А and D). For the rest, these features may vary somewhat, probably according to the prevailing conditions of the surroundings, but, on the whole, I found the structure of the leaves to be similar in plants fromW. and Е. Greenland, Iceland and Norway. Fig. 16. Silene acaulıs. (From Iceland: A, В, С, F, С and H. From W. Greenland: D and Е.) A, Transverse section of leaf; lib, bast; 4 is the dorsal surface, which has no stomata (see G) and is slightly higher than the ventral surface (v), which has numerous stomata (see H). The forms of the epidermal cells are seen in the same figures. В and С, Transverse sections of leaf. D, Transverse section of vascular bundle; lib, bast. E, Transverse section of leaf, four vascular bundles are seen and three cells with crystals of calcium oxalate. F, Stoma. J, Epidermis of ventral surface; the cell-walls are less undulating than in H. (E. W.) That it depends on the direction of the leaves (their position with reference to the light) whether they become isolateral or dorsiventral, was also pointed out by me in 1891, in the case of the Honckenya; they are isolateral when directed upwards, and dorsiventral when horizontal. bo 62 Buc. WARMING. The leaves of Silene acaulis are often somewhat oblique in transverse section (Fig. 16, B). | Cerastium alpinum also may have distinctly dorsiventral structure (Fig. 17); this is probably connected with the posi- tion of the fairly broad leaves which are found also in this species. The difference between the epidermis of the dorsal and the ventral surface is not so great as in the species already mentioned: there are stomata on both surfaces; but Fig. 17. Cerastium alpinum. (From Greenland and Iceland.) A, Transverse section of leaf, the two kinds of hair are indicated. There are two lacunæ (la). B, A similar section of a leaf from a bud. C, The transverse section shows that there are stomata on both sur- faces, that palisade-cells are indicated in the thin-walled, lacunose mesophyll, and that there is a sheath around the vascular bundle. The epidermal cells are nearly of equal height on both surfaces, per- haps if anything, a little higher on the ventral surface. D, From ventral side. The base of a hair is seen. E, Cell with crystal of calcium oxalate. F, Three portions of one and the same hair. G, Transverse section of a leaf, recently cleared of snow; la, lacuna on the lower side. (Uper- nivik; 7.5.1887; С. Ryder.) H, A glandular hair (dorsal surface). J, Both leaf-surfaces have highly undulating epidermal cells and stomata. А, Epidermis and stoma of ventral surface. Г, Same of dorsal surface. (Е. W.) the epidermis of the ventral surface is somewhat thicker than that of the dorsal (Fig. 17, К and Г). Whether the Caryophyllaceæ. 263 high degree of isolaterality shown in Fig. 17, G is due to the fact, that this leaf had recently been cleared of snow, or to some other cause, I cannot tell. The specimens from Ice- land showed essentially the same structure as those from Greenland. Fig. 18. Minuartia verna. (From N.E. Greenland; 27. 6. 1908; A. Lundager.) А, А small shoot with a branch. В and F, Transverse sections of leaf. There is strong bast tissue on the dorsal side of the vascular bundles. C, A glandular hair. D, Transverse section of leaf. The sheath of the vascular bundle is of the same thickness all round; see E, in which the bast tissue, sieve tissue and the tracheal tissue are shown. G, Epi- dermis with stoma. H, Epidermis of ventral surface. (E. W.) In the Minuartia verna figured in Fig. 18, D, decided dorsiventrality also prevails, but here the palisade-tissue is on the dorsal side — probably in harmony with the erect position of the leaves (A). 264 Euc. WARMING. In Minuartia groenlandica I found distinct palisade- tissue on the ventral side, but in the specimen investigated the leaves were, in contradistinction to those in M. verna, directed decidedly outwards. The epidermis of the dorsal surface was distinctly thicker than that of the ventral surface. Stellaria humifusa (Fig. 19). According to the transverse sections shown in Fig. 19, B and C, this leaf is but slightly 5 | Fig. 19. Stellaria humifusa. (From East Greenland; 15. 7. 1885; P. Eberlin.) A, Transverse section of leaf. В, The same, more highly magnified; v, ventral surface; d, dorsal surface. C, A transverse section, the epi- dermal cells are nearly equal in height, which is seen more distinctly in the more highly magnified H (from the ventral surface) and J (from the dorsal surface). D, The epidermis of the ventral surface, and E, the underlying mesophyll-cells. #, The epidermis of the dorsal surface and G the underlying mesophyll. (E. W.) J = dorsiventral, but а closer observation ргоуез the dorsiven- trality to be stronger. The epidermis of the dorsal surface is but little thicker than that of the ventral surface (Fig. 19, J and Н), but the ventral surface has cells with decidedly undulating walls (D), and is rich in stomata, while the dorsal surface has elongated cells with straight sides (Р). Caryophyllaceæ. 265 Moreover, the mesophyll which lies under the epidermis differed on the two surfaces (E and С). In a specimen from S.E. Greenland there was almost homogeneous, but indis- tinct, palisade-tissue on both sides. Similar weak dorsiventrality I found also in Cerastium trigynum (Fig. 20), and Minuartia biflora, which latter may however have decided isolaterality (Fig. 21, К and 5), but Fig. 20. Cerastium trigynum. (From Iceland; C. H. Ostenfeld.) A, Transverse section of leaf; pa, palisade-tissue. B, The transverse section shows the very lacunose structure of the thin-walled mesophyll; the epidermis of the dorsal surface (d) is higher than that of the ventral surface; there is a distinct sheath around the vascular bundle; there are about two layers of palisade-cells; g, a cell with crystal of calcium oxalate. C, Transverse section of stem. The black ring indicates the place of the mechanical tissue. There are five lacunæ in the cortex. D, From ventral side; epidermis and palisade-cells. Æ, From dorsal side. Magnification of D and Е is the same, so that it is distinctly seen that the epidermis of the dorsal surface is thicker than that of the ventral surface. F, The epidermis of the ventral surface and trans- verse section of palisade-cells. С, The epidermis of the dorsal surface. (Е. W.) it differs in the epidermis of the two surfaces, which, as it however appears, cannot be connected with the fact that the leaves may be directed upwards, with the ventral surface 266 Еос. WARMING. turned inwards, as in Fig. 21, А; they may also be directed straight out and curved backwards. More decided dorsiven- trality with two layers of palisade-cells on the ventral side, was found in a specimen of Cerastium trigynum from Iceland. Palisade-tissue on both sides Г have moreover found H Fig. 21. Minuartia biflora. A, From М.Е. Greenland; 27. 6. 1908; А. Lundager. В, Transverse section of leaf; the cross-hatched part is the mechanical tissue; the two leaves, С and D, which belong to В, show the venation; the fine dots indicate cells with crystals of calcium oxalate. E, Transverse section of leaf. F, Margin of transverse section of a leaf; to this belongs the more highly magnified G. H, Epidermis of ventral surface. J, Epi- dermis of dorsal surface. К, Transverse section of leaf, centric struc- ture. L, Epidermis of dorsal surface, seen from within, the outlines of the air-cavities under the stomata are distinctly seen. From N.E. Greenland: À, Е, F, С, H, J and Г; from У. Greenland (Holsteinsborg; 4. 8. 1885; Rosenvinge): В, С, D and К. (Е. М.) in Honckenya; it was however highest on the ventral side. Stomata occur on both surfaces. According to HEINRICHER, Caryophyllacee. 267 isolaterality is, for the rest, common in the Caryophyllacez, which must be connected with the fact that the leaves in many of the species are narrow, and can therefore be illu- minated equally on all sides. With regard to the mesophyll in general, the following may moreover be remarked. VESQUE had already recorded in 1883 that the mesophyll in the Caryophyllaceæ consists of large cells, which are loosely connected so that it is rich is intercellular spaces, and “spongy” (“spongieux’’). This is no doubt espe- cially applicable to the Arctic species. This feature is not only found in the species figured above, but in as high a degree in Cerastium nigrescens and Атепата ciliata (Figs. 22 and 23). In the latter figure it is even highly iso- lateral, with almost uniformly roun- ded cells. That it may be more com- pact is evident from what Tu. Нотм Fig. 22. > _ Cerastium nigrescens. “It has already been emphasized (From Dovre; 15th says regarding it. by others (BONNIER and BÖRGESEN) June; E. W.) A, The transverse sec- tion of leaf shows a the Arctic species, for the palisade- very lacunose tissue, with palisade-cells, and ‘ ; В а lacuna (la) in the differentiated, the layers few in num- spongy-tissue. B, Sto- ber, the height of the cells inconsider- ma of ventral surface. (Е. W.) that it is a feature common to all tissue to be most frequently imperfectly able, and the intercellular spaces large — at any rate, not so strong and compact as in species from more southern — for instance Mediterranean — countries. BÖRGESEN even found no palisade-tissue at all in specimens of Silene acaulis (1. с. Fig. 12) and Minuartia verna f. hirta; see also Arenaria ciliata in Tu. Нотм (Pl. XII, 2). 268 Euc. WARMING. This is evidently a structure which is due to the faint light and the damp surroundings, and which contrasts strongly with the leaf-structure of the Alpine plants (Wagner and Вогоезеп). In contradistinction to the Arctic Caryophyllaceæ we may refer е. ©. to Silene Otites from dunes near Montpellier, which has palisade-tissue on both sides (Warming, Halofyt- studier, 1897, р. 203) and Silene nicæensis from Cadix (ibid., — we (Og os @ D Fig. 23. Arenaria ciliata. (From N. Е. Greenland; 12. 7, 1908; A. Lundager.) A, Transverse section of leaf; two cells with erystals of calcium oxalate. B, A leaf; the dots indicate cells with crystals. C, D, Transverse section of leaf, showing the crystals. (E. W.) р. 198), together with Honckenya from Denmark which, when growing on sand strongly exposed to the sunlight, can have several layers of palisades on both sides (Warming, 1891, Fig. 12). Many records as regards anatomical differences in plants from different countries, and variations in the same species from different habitats, and in different countries, are found in HEINRICHER, BÖRGESEN, WARMING and others. It is evident that the local conditions prevailing in the habitats, are in epharmony with the differences in the individuals. As an example of this, the reader is referred е. g. to Jon. SCHMIDT’s Caryophyllaceæ. 269 experimental investigations of Lathyrus maritimus (Bot. Tidsskrift, Köbenhavn, 1899, р. 22). The different points of information concerning the leaf-anatomy of Alpine plants, may profitably be compared with those concerning that of the Arctic leaves. SCHRÖDER, in his excellent work “Das Pflanzenleben der Alpen”, mentions the first-named and gives a resume of BoNNIER’s and WaGNER’s results, which, in certain directions, show deviations from the Arctic species, especially as regards the palisade-tissue. When the Arctic plants have, on the whole, less differentiated palisade-tissue than the Alpine species, this must be especially attributed to the fainter light and the fogs in Arctic countries. The fact that the mesophyll in Arctic plants is always thin-walled, and has abundant intercellular spaces, has been emphazied by previous authors, for instance BORGESEN and Tu. Horm; and is evident from numerous figures which a number of authors have published in the “Morphology and Biology of Arctic Plants’, “Meddelelser om Grönland’, Vols. 36 and 37 (for instance О. GALLöE, 1910; KNUD JESSEN, 1911 and 1913; Carsten OLSEN, 1914; MATHIESEN, 1916; H. Е. Petersen, 1908, and others). In this, according to WAGNER, there is a conformity with the Alpine species, a con- formity which must be attributed to the greater moisture’. In Cerastium arvense (from willow copses in Kangard- luarsuk in West Greenland) the spongy tissue was of an unusual character, consisting more or less of stellately branch- ed cells, for which reason the lacune were particularly large. There was a distinct palisade-tissue with at least 3 cells in a vertical row. The epidermal cells had undulating walls and stomata on both surfaces. 1 For purposes of comparison see the anatomy of woody plants in WARMING (1887), Н. Е. PETFRSEN (1908), and Menrz (1909). XXXVII. 18 270 Euc. WARMING. A peculiar lacuna is found in the leaves of many of the species. It is caused by the sub-epidermal cell-layer of the dorsal surface separating, for a shorter or longer dis- tance, from the mesophyll lying internal to these cell-layers. The cells bordering on the lacuna are rounded off inwardly towards it. I have already in 1884 pointed them out in Silene acaulis, Arabis petræa, Dryas octopetala, Draba incana and Saxifraga oppositifolia, and they are found figured by Tu. Нотм in Stellaria humifusa (PI. X, Figs. 2 and 8), Salix reticulata (Pl. XII, Fig. 12) and species of Saxifraga (PI. Х, Figs. 3 and 5). See also BORGESEN l.c. In Heıcı Jonsson (1895, р. 291) we find the accompanying figure (Fig. 24), which shows a section through a bud of Silene acaulis. In the figures given above these lacunæ are shown in Fig. 14, A and В (Viscaria alpina); Fig. 24. Silene acaulis. Е 2 ‘ . (From Iceland; Helgi Fig. 17, A, B and G (Cerastium alpi- Tönen SEE EDT пит); Fig. 22 (Cerastium nigrescens). through a bud. In each leaf three vascular i р bundles are seen; А, cells thus superbus. I cannot imagine that with crystals. On the they are of any other use than as a dorsal side of the leaves air-containing lacunæ (l) are seen. I have also observed them in Dian- protection against cold, since they must be bad heat-conductors. Crystals. With regard to the mesophyll, it may further be remarked, that in many species cells are found with crystals of calcium oxalate. They are situated in that layer of the leaves, in which the vascular strands are found, or at any rate, in close proximity to them. They are no doubt found in all the species. They are figured in Fig. 14, А and В; Fig. 15, 4; Fig. 16, Е; 0 Fig ПЕ ЕО: Fig. 23, А, С and D, and also in В where they are indicated by dots; their position is similarly indicated in Fig. 21, С Caryophyllaceæ. 271 and D. I have seen them also in Silene venosa, Honckenya and Minuartia hirta (accompanying the three vascular bund- les), and several authors mention them for instance VESQUE (1883, pp. 130 and 133), Russen and HEINRICHER (see PI. 30, Silene inflata). There may be a difference in indivi- duals of the same species; in specimens of Silene acaulis from Finse in Norway there were many crystals of calcium oxalate, more than in the Icelandic individuals. Whether this difference be due to the season of the year, or to the habitat, I am not prepared to say. Mechanical tissue I have not found except in certain cases, and then within the sheath around the vascular bundle mentioned below, for instance, on the ventral side in Silene acaulis (Fig. 16, A and D), on the dorsal side of them in Minuartia verna (Fig. 18, B, D, E and F), M. groen- landica and M. biflora (Fig. 21), and on both the ventral and the dorsal side, or only on the dorsal side (of the erect leaves) of Melandrium involucratum (from Upernivik). Re- cords of this are also to be found in BÖRGESEN, VESQUE, and others. The epidermis always consists of one layer only, and the cell-walls are rather thin on all sides, but are naturally thickest on the outer side. The cuticle is thin. This must be an adaptation to the nature of the habitat, and can vary with it. W. Russet has demonstrated, that the leaves of Dianthus prolifer have much thicker epidermal walls in the Mediterranean countries, than near Paris, and therefore become much stiffer. Only in Viscaria alpina did I find the walls of the epi- dermal cells to be somewhat thicker, and the lateral walls porose (Fig. 14, C, D, E and F). The outer walls of the epi- dermis were found to be thick in Sagina nodosa from dunes in Denmark. 182 2712 Euc. WARMING. That the dorsiventrality can be displayed in the various degrees of thickness of the epidermal cells of the ventral and dorsal surfaces, has been stated above. It appears, as already mentioned, to be correlated, with the direction of the leaves (the epidermis of the surface which is turned upwards or outwards being the highest). The epidermal cells are arranged more or less regularly in longitudinal rows, at any rate in the narrow or linear leaves (see е. g. Fig. 14, E and F; Fig. 15, Е and F; Fig. 16, G, H and J; Fig. 18, H; Fig. 19, D and Е; Ше 20260204 Fig. 21, H, J and L). The lateral walls of the cells, when seen from above, are sometimes highly undulating on both leaf-surfaces, although more highly on that surface which limits the spongy- tissue than on the other surface (Fig. 15, Ё and F); more frequently the epidermal cells of the one surface have un- dulating lateral walls, while those of the other have only slightly undulating or quite straight walls; as also the cells of this surface are more elongated (Fig. 14, E and Е; Fig. 16, С, H and J; Fig. 19, D and F; Fig. 20, F ава с. ieee H, J and L). The cells with very highly undulating walls are moreover very short. The undulations indicate more damp or more shady conditions. The Stomata are lower than the other epidermal cells, and are placed on a level with the outer surface of the epi- dermis, or even project slightly above it (this is shown in many figures). Stomata may occur on both surfaces, some- times almost equally in number. In other cases, the one surface has no stomata or only a few. In some cases they are most numerous on the dorsal surface, in others on the ventral surface, for instance in Stellaria humifusa. All these circumstances have been fully treated of by VESQUE, BORGESEN, WAGNER and BONNIER (see also TH. Hom). As Caryophyllaceæ. 273 regards the number of stomata per unit of surface the reader may be referred, for instance, to BORGESEN. The apertures of the stomata, at all events in the narrow- leaved species, lie parallel with the longitudinal axis of the leaf. That the guard-cells are surrounded by the other epi- dermal cells in a peculiar manner, was, as far as | know, first pointed out and figured by VESQUE (Anm. sc. nat., 1883, 6. Sér., 15, p. 130): “Les stomates sont embrassés par deux cellules dont la cloison separatrice et perpendi- culaire à l’ostiole”. See the preceding Fig. 14, Е and F; 15, meander: 16, № and Л: 17, J; 19, D: 20, Р: and Fig: 21, J and JL, all of which, in addition, show, here and there, a small deviation of minor importance, in that an aperture may be surrounded by 3—4 cells, but each end of the stoma is always limited by a more or less crescent-shaped cell. Hairs are found in many of the species, in some even in such great numbers, that they become quite pilose, this is especially the case in Cerastium alpinum, which is otherwise exceedingly variable as regards the extent to which it is hairy. The hairs are of two kinds, both consisting of a single row of cells arising from a single epidermal cell, as already pointed out by Vesque (1883). The one kind of hair is thick-walled and dead (filled with air), and terminate in a point (Fig. 15, D and Fig. 17, Е); they are specially numerous in the form lanata of Cerastium alpinum. Their function must be to protect the plant against exces- sive transpiration. The other kind of hair is secretory, and its terminal cell is globular (see for instance Minuartia verna, Fig. 18, С, and Cerastium alpinum, Fig. 17, H and A). The Vascular bundles. It is, at all events in the narrow-leaved species, only a single bundle, which from the stem enters into the leaf, and there quickly develops a 274 Euc. WARMING. branch on each side, which may itself branch ogain (Fig. 21, C, D and Fig. 23, B). In the most narrow-leaved species this does not often happen, so that the entire number of bundles in а transverse-section will be three only (Figs. 18 and 21) — these are especially prominent in older leaves which have been emptied of their cell-contents, — but more frequently there are several, or even many, in the broadest- leaved specimens (Figs. 14, 17 and 20). As regards the vascu- lar tissue, I have no particulars to communicate. In some of the species mechanical tissue occurs on the ventral or dorsal side of the strongest vascular bundles, as mentioned above. It appears to be an exceedingly constant characteristic in the Caryophyllaceæ to have around the vascular bundles a thin-walled, sharply differentiated sheath consisting of one layer of cells, which are almost square in transverse section; this is illustrated in most of the figures given above. In the majority of cases, the cells are of equal height, but cases may occur in which those of the dorsal surface are higher than those of the ventral (e. <. Silene acaulis, Fig. OA D); For further particulars regarding the anatomy of the Caryophyllaceæ, the reader is referred for instance to REGN- AULT (1860; Ann. sc. nat., 4. Sér., 14); SEIGNETTE (Revue gen. I, р. 564); O.G. PETERSEN (Botan. Tidsskr., Koben- havn, 1888); HernricHer (1884); Лон. Erikson (1896); WARMING (1890, 1891, 1897) and SOLEREDER. As regards the structure of the leaves in the Alps, see WAGNER, Laz- NIEWSKI and C. SCHRÖETER. Ш. Adaptions to the Environment. The morphological and anatomical structure of plants is, in the first place, dependent upon their genetic relation- Caryophyllaceæ. 215 ships, but it will be difficult or rather impossible to demon- strate the origin or utility of their genetic peculiarities (for instance, why the leaves in the Caryophyllaceæ are opposite; the origin of the peculiar manner of branching with only one shoot, or at all events only one vigorous shoot, from each pair of leaves, and of the position of these shoots in a peculiar spiral; or the origin of the one-layered sheath, around the vascular bundle, just mentioned). As regards the Caryophyllaceæ VESQUE (1883, p. 133) has rightly stated that “La famille des Caryophyllées est des plus naturelles et des mieux limitées”. This is true both as regards the morphological and the anatomical structure and, as far as Г can see, cannot be explained ecologically. Also as regards the structure of the flowers, there are very great similiarities in the different genera. In the second place, the peculiarities of plants are dependent upon the ecological factors of the habitat, 1. е. the climatic and edaphic conditions which prevail there. It is consequently of importance that we should be accu- rately informed with regard to these points, but the truth is, we have information only about the general features of the climate and — possibly — of the soil. As regards the conditions which prevail in the special, local habitats in which the different individuals have been gathered we are, so to speak, never told anything. Unfortunately, there are many travelling botanists who find it sufficient to state, that such and such species have been found here and there, and to record the geographical situation; but as regards the conditions found to be prevailing in the habitats in ques- tion, they tell us nothing whatever. This is however а great drawback, for without doubt the ecological factors, or the dif- ferences which prevail in the habitats, are exactly that which sets a different stamp upon individuals of the same species. 276 Euc. WARMING. Of the botamists who have given the fullest information with respect to the natural conditions prevailing in the habitats in Greenland, [ must, as regards the Caryophyllaceæ, point out most particularly JENS VAHL (in the Arctic Her- barium belonging to the University of Copenhagen) and Cur. KRUUSE (in his papers on the plant-growth on the east coast of Greenland in “Meddelelser om Grönland” Vols. 30 and 49). To mention the numerous data to hand, would take us too far; here it must suffice to refer to the Floras and especially to Jou. LanGe’s “Conspectus”; I there- fore give the following short resumé. With regard to the climate of the habitat, we know, in а general way, for instance, that it becomes drier the farther we proceed from the south up towards the north in Greenland, and from the coast towards the interior. See for instance the small map (after W. Kress) reproduced in GUNNAR ANDERSON (1900). But the rainfall is hardly of as much importance as is the water accessible to the plants ecpecially from the ground and the fogs. The soil is usually rather damp. One hardly goes too far if one asserts that a great moisture of the soil is an essen- tial characteristic of Arctic nature. This is evident from the many descriptions, for instance in VAHL (“in locis humi- dis”, “in locis humidiusculis”, “in locis turfosis”, “in locis uliginosis”, “in pratis”, etc.), or the frequent occurrence, in KRUUSE, PorsiLp, LUNDAGER, and others, of expressions such аз “оп moist, barren ground”, “on wet sand”, “by preference in humid places”, “near lakes or running water”, “on sheltered, humid slope”, “on moist rocky flats’, “low- lying humid soil”, “on damp, manured places”, etc. Jon. LANGE has, as a rule, given tolerably clear information in his “Conspectus” about the nature of the soil, and his state- ments are presumably based upon Jens VaxL’s labels. Caryophyllacee. 277 His statements also verify the opinion that the majority of the species grow upon tolerably damp ground, many of them even upon wet or boggy ground. The cause of the general dampness of the ground, is apparent. It is, in the first place, the snow-covering, during the long winter, which lies over the whole country, so to speak, and which only disappears for a few summer-months, and then from a part of the coast-land only. The melting snow soaks into the soil and trickles down over the sloping ground, for a long time especially in those parts where the rays of the sun can only reach it with difficulty, for instance, on slopes facing the north. Different travellers have drawn attention to the fact that the northern sides of mountains or those sides of valleys facing the north, remain fresh and green much longer than those facing the south. From the Alps we hear much about “die Schneethalchen”’ and their peculiar flora (for instance SCHRÖTER, BROCKMANN- JEROSCH and other Swiss botanists); also from Scandinavia (for instance TH. REsvozL). These isolated snow-patches which lie long, many during the whole summer, retain a considerable amount of moisture for a long time, and may be prolonged sources of water which will trickle down over the land. Even if plants grow upon a sandy substratum, or even upon a rocky substratum, this will long be able to remain wet or damp. A stony or gravelly substratum also retains moisture for a long time, because the ground is pro- tected from evaporation by the covering or imbedded stones. Even if travellers record, with regard to an individual plant, that it has been found on “dry ground,” one must not take for granted that the ground has not been damp — even very damp. With regard to those plants, which have possibly been recorded to have been gathered “on manured 278 Euc. WARMING. 99 spots,” we may also suppose with tolerable certainty, that the ground on which they grew contained much water. Some of the species, however, are not very particular in their demands, since they are recorded as growing in “all loca- lities,” or “everywhere,” as for instance, Cerastium alpinum and Silene acaulis. The many varieties of the former perhaps stand in direct causal connection with small differences in the habitats. Transpiration. A factor which is connected with what has been said above, and is of no less importance, is the extent of the transpiration and of the atmospheric humidity. TH. Wüzrr stated (1912, p. 5) that “the Arctic Flora is of a decidedly xerophilous type.” There are, it is true, decidedly xerophilous types in the Arctic vegetation, which I have probably been the first to record (1887) since I, in chapter 5 (pp. 105—127), treated of “The Adaptations of Heath-plants to Drought”. It must, however, be remarked that the woody plants of the heath and the rocky-flat are what I especially treated, and amongst these one finds in reality a number of distinctly xerophilous types. I pointed out five different types of leaves, and even declared that similar types are met with in the steppes and in deserts — even in the Egyptian-Arabian desert, but I laid a great stress upon the fact that this essentially applied only to the woody plants of the heath. By far the greater majority of the species in question are evergreens, this especially applies to the many woody plants proper, the foliage-leaves of which remain for more than one year. With regard to these one can speak of phy- siological dryness, but hardly with regard to low-growing herbs. It is easy to understand that in woody plants the shoots must become xerophytic in structure, — which is, as usual, Caryophyllaceæ. 279 exhibited in the anatomical structure of the leaves, — be- cause they are higher in growth than are herbs, and often may be without а snow-covering during the winter; their leaves will frequently be exposed to the desiccating power of cold and dry wind and this in a higher degree than those of herbs, and the cold soil will for a long time prevent the roots from absorbing water. The winter, in particular, is a very dry season of the year, and the higher the latitude, the drier it becomes; but the danger of desiccation is much less as regards the vegetative organs of herbs, because they are generally lower in growth, and are, in addition, often protected by the dwarf shrubs. If we now regard the leaves of the Caryophyllaceæ, we find no such adaptation for withstanding desiccation. On the contrary, in the structure of their leaves they remind one more of those of aquatic plants, or of herbs which grow on damp, shady ground in woods. The epidermis has no specially thickened and cuti- cularised outer walls. The stomata are situated very super- ficially, occur often on both surfaces, and in great abun- dance; the epidermal cells are most frequently highly un- dulating; the mesophyll is very lacunose; aqueous tissue does not occur; the epidermis is on the whole glabrous, and in Cerastium alpinum alone a denser covering of hairs is found; and hard bast occurs rarely and only in a few thicker vascular bundles. These results agree, as already mentioned, with those previously arrived at by several botanists, as regards Arctic plants in general. The explanation is obvious: The rays of the sun are very oblique; the leaves are but slightly warmed, al- though the insolation may be very strong locally; the atmos- pheric humidity is great (80—85 °/,; see table in Börgesen); fogs and cloud-covered sky are very frequent, also during 280 Euc. WARMING. the summer; it is advantageous that transpirationis not lessen- ed by special structural features, so that the water-current can flow rapidly through the leaves and the assimilation increase in vigour and compensate for the briefness of the summer, with its fogs and weak light. The transpiration cannot be assumed to be specially strong during summer. It is also easily compensated for by the water which is accessible to the herbs from the fogs and from the water-contents of the soil. The precipitation undoubtedly varies greatly in the different regions. LuNDAGER records from Danmarks Havn in N.E. Greenland 28.4mm for the summer (April—Sept.) and 117.2 mm for the winter. Only !/,, falls as rain. The inconsiderable amount of rainfall must undoubtedly be compensated for by the water-contents of the soil and the fogs. Тновию WULFF has, in his paper of 1902, communi- cated some results of transpiration-experiments with Arctic plants made with the cobalt-test in the open in Spitzbergen, detached leaves being placed between cobalt-paper. WULFF particularly emphasizes the fact that the method is defec- tive, but no other could be employed. The water from the snow of the habitat situated on a high level “provided for a homogeneous distribution of the moisture of the soil.” Unfortunately, he did not investigate the temperature of the water and the soil, which of course plays an essential part as regards the amount of the transpiration. Among the ten species which were investigated, there was only one of Caryophyllaceæ, viz. Cerastium alpinum, the leaves of which were woolly with dense hairs. The transpiration was exceed- ingly sight, which Wurrr thinks was essentially due to the hairy covering, “as the anatomy of the leaf rather favours the belief in an excessive transpiration than tells against it.” Among the plants used for the experiment, Cerastium alpi- Caryophyllaceæ. 281 num belonged to those which had the least transpiration, and the slightest growth. The result does not appear to have been very conclusive, but he says that, on the whole, the Arctic plants which were investigated had only slight trans- piration, compared with plants from more southern regions. This must, beyond doubt, be due to the atmospherie humi- dity, which is certainly greater during the vegetative period than outside it. The leaf-structure of herbaceous plants agrees well with the humid air, the fogs and the abundant water-contents of the soil. As regards the physiological dryness in Arctic countries certainly too much has been said. Тновию WULFF generalized in too high a degree in saying, that the Arctic flora “is of a decidedly xerophytic type,” although this is true only of the woody plants. The above-mentioned conditions pertaining to the habitat also explain, at any rate some of the morphological conditions, e. g., that in so many cases adventitious roots are developed; that runners are formed may be due to the fact, that the low temperature has altered the growth-direc- tion of the shoots (Psychrokliny: Vöchting, 1898; Lidforss, 1903), the moisture, perhaps in connection with the pres- sure of the snow, having favorised a development of adven- titious roots. On the other hand, it must be the cold and desiccating winds which cause the frequently-occurring pulvinate growth; the shoots are arrested in their growth, and the branches become short and thick-set. That the buds are open and that bud-scales are absent, I regard as a generic character. This results in conditions favourable to the development of the pulvinate habit, espe- cially when the plants are covered with a layer of snow, which lasts throughout the winter; many leaves must be able to keep green for a long time, at all events until the snow has melted, and the youngest leaves are ready for 282 Euc. WARMING. quick expansive. In all Arctic Caryophyllaceæ it will certainly be found that the youngest leaves are, during the winter, protected by older leaves, which are more or less withered. That the peculiar condition of the protoplasm, and the different substances contained in the cells, can protect against the cold of the various seasons, must be taken for granted. With regard to this, however, I cannot give any new information. ТУ. Flower-biology and Notes on Seed-produetion. In by far the greater majority of the Arctic-flowering plants it appears that the flowers are formed during the year previous to that in which they open, and that they pass the winter in a more or less well-developed condition, — а circumstance which is а natural adaptation to the ex- treme brevity of the favourable season of the year (the extreme brevity of the summer-time). In the foregoing papers in “Meddelelser om Grönland” belonging to this series, the fact that the flowers are formed during the year previous to that in which they open, has been specially pointed out in a number of species by KNUD JESSEN (Ranunculaceæ, Rosaceæ) and others, and for in- stance, as regards the Caryophyllaceæ, this is found to be the case in Cerastium alpinum in Iceland and the Færües (Helgi Jonsson), Silene acaulis (idem), Cerastium trigynum in Norway (Th. Resvoll), Melandrium apetalum (eadem), and Minuartia biflora. On Greenland material, gathered on Aug.6th, I found buds which enclosed flowers for the next year. As an example I give here (Figs. 25) some figs. of Silene acaulis. The material was gathered by С. RYDER near Uper- nivik on ground which had recently been bared of snow. The buds which were present had evidently passed through the winter. The buds were 1/, mm in transverse section; LL 2; Caryophyllaceæ. 283 they were doubtless male flowers, all the leaf-whorls of which were formed, also that of the pistil (Fig. 25, the small figure to the left). For further particulars, see explanation of the figure. On material from South Greenland, gathered on Aug. 29th, I also found buds which enclosed flowers for the next year. That flowers buried in the snow, on the melting of the latter in the ensuing year, sometimes emerge “as fresh to look at as newly expanded”, is evident from what e. 8. LUNDAGER writes (1912) about Cerastium alpinum found in N. Е. Green- land, May 23rd, 1908 (loc. cit. p. 409). In “Botanisk For- enings Festskrift’’ (Co- penhagen, 1890), I published a series of Fig. 25. Silene acaulis. Young bud, highly magnified and made transparent, so that through the sepals biology of the Arctic (s—s), drawn with continuous lines, the calyx-stamens (sst), the petals (p) and the stamens (pst) placed opposite to them notes on the flower- and Danish Caryo- phyllacez, accompa- are seen. The pistil is in the middle; it nied by 29 figures. is at the stage of development shown in 3 the small figure. From Upernivik, 10. 5. notes, which 1887; С. Ryder. (Е. W.) were written in Da- nish, without any Resumé in a universally current Euro- pean language, appear to me to have remained almost unknown. I have, therefore, no hesitation in reprinting here such parts of them as concern Arctic species. My notes are based, for the most part, on observations made in the field in northern countries, and afterwards supplemented by studies made on material preserved in alcohol. I had been hoping to find some contribution towards the elucidation of the nature of gynodioecism, but this 284 Euc. WARMING. hope was not realised. It is to be hoped that others may be more fortunate in this respect, especially through expe- riments they may make. I begin with the Alsinex because I regard them as the most original type of Caryophyllacee: since their flowers are open and have polysepalous calyx and polypetalous corolla which are for the most part complete in number in all the whorls; can be visited by many kinds of insects; have many ovules in the ovary; and are hypogynous. From this proto- type there issue on the one side the ‘Silene, the flowers of which have gamosepalous calyx and are in a higher degree adapted to pollination by special insects; and on the other side, the Sclerantheæ, with more or less reduced flowers. 1. Cerastium arvense. L. Denmark. Gynodioecism; protandry. As the plant is propagated by underground runners, large patches are found covered with the one form only. The stigmatic papillæ extend almost to the base of the styles. $ are distinctly >! $, 13—15 mm in diam., petals 10— 11 mm long, while ? are 10—12 mm in diam., petals 7—8 mm long. Highly protandrous with the ordinary process of de- velopment; the cal.-anth. are the first to open, the cal.-st. standing erect, then the cor.-anth., before the cal.-anth. are emptied, and the cal.-st. bend backwards; ultimately, the styles bend outwards, and the stigmatic papillae grow out, as a rule before the pollen has fallen out entirely. Self-polli- nation may be able to take place. The anthers are usually 1 For brevety’s sake the mathematical signs indicating relations of quantities are used. For the same reason the following abbre- viations have been used: cal.-anth. for anthers opposite the sepals, cal.-st. for stamens opposite the sepals, cor.-st. for stamens opposite the petals, etc. =" es Caryophyllacee. 285 placed transversely to the filaments, with the pollen lying exposed on the side turned upwards (Fig. 26, a!) or even entirely outwards. In Germany, according to А. ScHULTZ, $ are not rare and are often the only form to occur over entire tracts. I do not know how frequent it is in Denmark. The anthers are generally distinct and almost normal in form, but they are white, transparent and empty. Sometimes a few (I have seen 1—3) anthers are met with among the others which are yellowish, dehiscent, and contain pollen-grains; the latter appear, however, to be useless, only a few being round and smooth, while the majority vary in size, are angular and thick-walled, and are therefore distinctly abnormal. It is also usual for such flowers to be somewhat larger (petals 9—10 mm) than those which are more decidedly female. In the neighbourhood of Copenhagen there is a growth of Cera- stium arvense in which the flowers vary between 8 and 2 and there fruit is sometimes set, but, as far as I have ob- served, rarely. Greenland (about 67° N. lat.). 3. Protandrous. Petals 8 mm, cal.-st. 5 and cor.-st. 41/, mm. Por 2 mm. Conse- quently, somewhat small-flowered, but otherwise normal. 2. Cerastium alpinum L. (Fig. 26). Almost as large-flowered as C. arvense, but has neverthe- less a decidedly homogamous and self-pollinating flower, which fact must undoubtedly be correlated with its northern (Arctic) and its Alpine homes. In Greenland I observed that the flowers after a two-days rain were somewhat more drooping, and more closed than usual. West Greenland (as far as about 73° N. lat.). Gyno- dioecism; Gynomonoecism. — $. Slight protandry with XXXVI. 19 286 Euc. WARMING. quickly ensuing homogamy and self-pollination (Fig. 26a, 61). The relative length and position of the stamens and pistil is shown in the figures. Before the cor.-anth. are opened the styles may be covered with germinating pollen. I have even seen homogamy set in immediately, even before the flowers had expanded fully, and while the styles were still Fig. 26a. Cerastium alpinum. From West Greenland (about 64°—70° N. lat.). 9. — A, al, a?: 8 in which four cal.-st. and one cor.-st. are abortive (at). B: 8, scarcely expanded, but the stigmas (5?) are covered with germinating pollen; three cal.-anth. have opened, but the cor.-anth. (61) are still closed. С, Style of 3 with pollen which has germinated; all the anthers have opened and are empty. ct shows the longest papillæ at this stage. — о. D, Style, ®”/, and not °*/, (from beyond 73° N. lat.; С. Ryder), and 41, 4?, 43 stamens in various stages of reduction; d! has fibrous cells in the wall, but no pollen. Е, Seed. (Е. W.) erect: in Fig. 1, B, 61, 6? the stigma is covered with ger- minating pollen although seven anthers are still closed. This Caryophyllaceæ. 287 is almost self-pollination in the bud. In specimens from Upernivik (C. Ryder) the flowers were closed and the styles were closely surrounded by the anthers. — © are very com- mon in Greenland, as far as 73° N. lat. They are < $, the petals being 8 (to 10) mm in length (sometimes only slightly longer than the calyx), while in $ they are 11—12 mm long; the pistils appear to be the same, but the stigmatic papillæ are larger in © (Fig. 26a, D) than in & (Fig. C, с"). The stamens may be abortive in varying degrees; the anthers may have a somewhat normal form, but be quite transparent, or the stamens might be, for instance, 4—4 mm long and have yellowish anthers, which is, however, devoid of pollen (Fig. 41—43). Near Godhavn I found one © in the same tuft as 8, consequently gynomonoecism. In several female flowers from Godhavn (69° N. lat.) I saw the stigmas densely covered with pollen-grains, so that insect-pollination must be able to take place. Intermediate forms also occur, for instance, flowers with two normal stamens and the others sterile; or four cal.-st. and one cor.-st. abortive, or even eight normal and two quite abortive. East Greenland. 8 (gathered by P. Eberlin), agree with those from the west coast, but the flowers are smaller (petals 9 mm). Fruit ripens in Greenland from the southern point up to at least Upernivik (seed, Fig. 26a, £). Spitzbergen. $ (gathered by Nathorst; 1.7. 1882), agree with those from Greenland. Fruit set, which ripens both here and I think in all other Arctic regions, е. ©. Grin- nel Land, Sabine Island, the north coast of Siberia, ete. Var. В cespitosum (Malmgr.), Fig. 26b, gathered by Nat- horst. The specimens are very small (foliage-leaves 5 mm long and 31/, mm broad), quite glabrous; appear to have 11} 288 Euc. WARMING. foliage-leaves which persist through the winter, and not to differ from the type in regard to the flowers. Iceland. (Reykjavik). Protandry. Norway. As in Greenland, sometimes even protogy- nous (Alten, about 70° N. lat.; Tromsö). Gynodioecism: Kongsvold (16. 7. 87). LinpMAN is of opinion that the anthers are placed so far above the stigmas that self-pollination cannot take place until in the last stage, when the fully grown and revolute styles may reach the anthers. — Sweden: Slight protandry (Axell). — The Alps: Gynodioecism ; protandrous (but Ф much rarer than 9: Ludwig). — Italy: Homo- gamy (Ricca). her [жеж à Fig. 26b. Cerastium alpinum 3 cæspitosum Malmgr. From Spitzbergen (Nathorst; 20. 8. 1882). A, A branch, natural size; lower part quite glabrous and with very small, close-set foliage-leaves; upper part covered with long, scattered hairs (not glandular hairs). The lowermost leaves have persisted through the winter. at, A pair of foliage leaves. В, A flower opened and its parts spread out artificially. C, Sepal. D, Petal. E, The basal part of a cal.-st. F, Style with germinating pollen-grains. There was still some pollen left in the anthers of the flower of which the style is shown in Fig. F. (E.W.) 3. Cerastium latifolium. L, Norway (Dovre; 15. 7. 1887). 8 have petals about 8 mm in length; the stigmatic papille extend downwards only 3/, or 4/; the length of the style. Slight protandry, but homogamy soon ensues, and in the specimens investigated Caryophyllaceæ. 289 the styles were bending backwards so decidedly that they almost touched the petals and were more-or-less dipping into the pollen, so that self-pollination took place. In one flower the styles were only slightly curved back- wards, but were already covered with germinating pollen; of the anthers, only the cal.-anth. were open and had shed a little pollen. Here, consequently, a slight degree of proto- супу, if anything, appears to have taken place. Switzerland: Protandry, with ultimate self-pollination (H.M.). 4. Cerastium trigynum Vill. (Fig. 27). (= Stellaria cerastoides L.) Greenland: Gynodioecism with decidedly homogamous or slightly protandrous or slightly protogynous flowers. 9 but rarely 2; gynomonoecism rare. Self-pollination. 8. The petals are only slightly outspread (Fig. 27, A) so that the flower becomes on the whole, bell-shaped; but it is possible that light and heat exercise some influence as regards this point. The flower is not infrequently somewhat irregular, some of the petals being smaller than the others. The petals grow decidedly during the flowering period; they are 6—10 mm long, but vary in length relatively to the sepals. The anthers and the stigmas attain almost the same height and are functional at the same time; as all the stamens simultaneously stand erect closely around the pistil, and the styles are more or less outwardly curved, self-pollination is inevitable (Fig. 27, В), and the numerous pollen-grains which are sometimes seen upon the styles (Fig. C) are also indicative of this; contact has also been directly observed (Fig. B). In fully developed buds the anthers are quite ripe, and they often dehisce immediately after the bud has been artificially opened; I have even found them dehisced in the bud itself, and as the stigmas also are fully developed, I 290 Euc. WARMING. do not doubt that as a rule self-pollination takes place im - mediately after the opening of the bud. The stigmatic papillae extend almost to the base of the styles; the latter 1 Fig. 27. Cerastium trigynum. A—C, from West Greenland (gathered partly by Rosenvinge). A: 3; length of petals 7—8 mm; styles erect. В, С: Parts of the same flower; self-pollination (6. 8. 1884). D—E, from East Greenland (P. Eberlin; 30. 7. and 5. 8. 1885). D, Two abortive stamens; one contains some pollen. E with ele: a remarkable, small-flowered form; all the cal.-st. are open and are partially empty; cor.-st. still full of pollen, but are in the act of opening. Е, а, from Norway (Tromsö). Е, ft: 2 with quite transparent anthers which are, however, almost normal in form. G, base of a stamen. se” Caryophyllaceæ. 291 are usually somewhat club-shaped, becoming broader at the apex, or even 2-lobed (Fig. 27 C; e?). According to LinpMAN the flower has a sweet perfume (Dovre), in Green- land I made the note “White, scentless.”’ A remarkably small-flowered form was gathered by P. EBERLIN in East Greenland (Kutek; 5. 8. 1884); it is figured in Fig. 27, E. The petals are in their last stage only 6 mm long and of the same length as the calyx. 2 occur, but rarely, as far as I can see. I have seen specimens from East Greenland, the stamens of which are shown in Fig. D, but as some of the stamens appeared to be normal, these flowers were consequently perhaps not purely $. Gynomonoecism: In the same tuft I found $ and © with large rudiments of stamens. Fruit is set in Greenland, for instance at 67°—70° N. lat. A © was found pollinated, so that insects must certainly be of assistance there. Norway. Gynodioecism (Tromsö, 21. 7. 1885); 8 ho- mogamous (West Finmark; Dovre; Tronfjæld); in all its parts as in Greenland. — $ has sometimes clear, transparent anthers, devoid of pollen, sometimes less decidedly reduced, and the same plant may show diversity of floral condition (Fig. 27, К). —Lınpman: Homogamy, regular self-pollination. I found transitional forms from 5 to 2 near Hammers- fest, for instance, three rudimentary cal.-st., the rest nor- mal; but I found cal.-st. < cor.-st. in a flower which occur- red on a plant which had otherwise entirely © flowers. In © flowers I have most frequently found the cal.-st. to be the shortest. % also occurred in which the stamens were so short that the backwardly-bent styles were not in contact with them, and in several 5 the cal.-st. were < cor.-st.; but the contrary is the rule. 292 Euc. WARMING. The Alps. Homogamy; in unfavourable weather self-pollinating to a high degree; specimens with partially shrunken stamens occur (H.M.). Italy. Homogamy (Ricca). — The flower appears to be exactly similar everywhere. Abortive stamens have hitherto been recorded only from Greenland and Norway. 5. Cerastium vulgatium L. (С. triviale Lin Nr.) Denmark. Gynodioecism. 3 slightly protandrous or almost homogamous; self-pollination takes place by the stigmas coming into contact with the anthers of the erect stamens. Sometimes all the cal.-anth. may be open, while the styles are still rolled up in the middle of the flower. The stamens usually stand more or less erect, especially the cal.-st. The anthers are white. The flower varies in diameter accord- ing to its age; at first it is about 5 mm in diameter, and afterwards as much as 8 mm. The styles bend to the right in the flower. In a few cases protogyny appears to occur. 2 with stamens aborted in varying degrees, some about 2mm long, with rather large and almost normal anthers, others only 1—11/, mm long with transparent anthers which are, however, somewhat normal in form. Styles large, twisted (to the right), papillose down to their base, and filling the throat of the corolla. The corolla somewhat smaller than in 8, the petals being 31/,—41/, mm, while those in $ are 5—7, most frequently 5—6 mm. The pistil is not, on the whole, more vigorous than in §, if anything smaller; the‘ovary is about 2mm long in both. © plants are rarer than $. Specimens with partially aborted stamens occur. Iceland (Reykjavik; 1884). Gynodioecism. $: petals 5mm, cal.-st. 3mm, ovary 2 mm and styles 2mm. Some flowers had normal anthers, that is, they opened and con- tained pollen, which, however, appeared to be almost useless. Sweden. Slight protandry; fertilized by self- pollination (Axell). Germany. Gynodioecism; Gynomonoecism. Slight protan- Caryophyllaceæ. 293 dry, rarely pure homogamy (Н. М.; А. S.); in Riesengebirge A. Scauzz observed in the same individual both decidedly homoga- mous and entirely protandrous flowers; entirely © are said to be rare. — This species is one of those in which © flowers seem to be rare; they are not much smaller than the 3 flowers. ON à ra ea; Fig. 28. Stellaria longipes. 9. — A, from Greenland (about 69° N. lat.); а?, pollen grain. В, from $ in its last stage. (differently magnified). С, Most fully devel- oped papille from 8 observed by me. Ф. — D, 41, d?, and Е, e!: 2 flowers from West Greenland (about 67° М. lat.). Е: Stigmatic papille of $ (longer than those of ¢ shown in С); from Greenland (about 67° М. lat.). С: Cor.-st. of © from about 69° N. lat. H, №1 (Pollen), №? and #3 from 8 from Spitzbergen. J, i1—i8, from © from Spitzbergen. (E. W.) 294 Euc. WARMING. 6. Stellaria longipes Goldie (St. Edwardsii В. Br.) Fig. 28. Greenland. Gynodioecism; $ > $. Slightly protandrous or homogamous flowers. 3 appear to be more or less decidedly protandrous or decidedly homogamous. The specimen figured in Fig. 28, A had all its anthers open, but the stigmas were still unripe. In the fully developed flower the stamens stand at a distance from the pistil, so that the anthers are not in contact with the stigmas which are never markedly revolute. The flower secretes honey abundantly. I am not prepared to say whether self-pollination takes place regularly, but judging from the position of the parts of the flower, there does not appear to be anything to prevent this, and the anthers are not empty when the stigmas ripen. The stigmatic papille extend half the distance down the styles (Fig. 28 В). The number of the styles varies, being sometimes as many as four. ® occur as far north as Upernivik (about 73° N. lat.), according to material gathered by C. Ryper (3. 9. 1886). They are much smaller than $ (compare Fig. 28, D and Е with A, or, for instance, the petals 41, el with at); while 8 is as much as 18—14 mm in diameter, $ is only 7—10 mm; the petals of $ are 7—71/, mm, of 2 5—6mm. But other- wise the © flowers vary as regards the relative size of the parts. Specimens occur in which the petals are longer than the calyx, others in which they are of equal length. The stamens (Fig. G) are smaller than, or at most of the same size as the ovary (Fig. D). The pistil is so large in proportion to the corolla that the apices of the styles protrude (Fig. D, Е). The stigmatic papillæ are longer in © than in $ (Fig. Е and 185 compared with В and с). I have no note of having seen ripe fruit from Greenland. Spitzbergen. Gynodioecism; § > (Fig. 28, H and J). It is remarkable the small size which the flowers (some- Caryophyllaceæ. 295 times?) attain here, which can be seen by comparing Fig. H—J (from Spitzbergen) with the others which are from Greenland. The anthers and pollen-grains however are almost equal in size (see a? and ht). The petals in some of the speci- mens investigated were less deeply cleft than usual (Fig. ne Fig. 29. Stellaria graminea. А, В, from 8 from Altenfjord, gathered late in the evening of 14. 7. 1885. Markedly protandrous, the stigmas still far from ripe. C—J, from Denmark. С, D, Е: 8. Е: Intermediate form. С, H, J: $ (Scale given applies to A, В, С, Е, F and Н). (Е. W.). Н, 12). $ are protandrous, the corolla is > the calyx (Fig. H). — 2: the length of the corolla is equal to that of the calyx (Fig. i! and 12), about 4mm. The stamens may have rudimentary pollen-grains of varying size and form (Fig. 7°, 296 Bue WARMING. some are perhaps sometimes normal); but in the highest degree of reduction the anthers are transparent, without the least trace of pollen-grains. The stigmatic papille are also here longer than in $ (see Fig. 16 and 13).. 7. Stellaria graminea L. p. p. (Fig. 29). Denmark. Gynodioecism; $ > 9, distinctly or less distinctly protandrous. $: Protandry, with the stamens performing movements towards the middle of the flower and back again. The anthers turn more or less outwards or even backwards. As a rule self-pollination may ultimately take place, but I have seen flowers which had lost their anthers before the styles spread out. The anthers are flesh-coloured or become ultimately somewhat brownish, the pollen is yellow in colour. The styles are covered with stigmatic papilla along the upper 3/, of their length (Fig. 29, G). At first they are bent inwards or almost rolled up, ultimately spread out, scarcely rolled backwards. The flower varies greatly in size and length relatively to the calyx (Fig. C and for instance £); this is also still more noticeable when compared with flowers from other countries, for instance Fig. A from Finmark. 2 appears to be as common as 3. Its size relatively to & is seen by comparing Fig. 29, 7 ($) with С and Е. The petals in © are 31/,—4!/,, rarely 51/,—6 mm; in 8 5—7. 9 is about 5—6 mm in diameter, while 8 is 9—13 mm, but near Copenhagen $ are found which measure 31/,—4 mm in diameter, and have petals 21/, mm in length, i.e. smaller than the calyx. It should be borne in mind that the petals grow in size during the flowering period. In some of the flowers the anthers (Fig. J) are quite transparent with crys- tals of calcium oxalate, in others the anthers are of a yel- lowish colour, or only their one-half is so coloured, and - Caryophyllaceæ. 297 pollen has begun to develop within them. Several variations may be found in the same tuft. The styles project beyond the corolla. Intermediate forms are often found in which a few or several stamens are aborted, and the others are normal (Fig. F). There does not seem to be much difference between the stigmatic papillæ of 5 and $ (Fig. 29, D and С). Norway (Altenfjord). Gynodioecism, 8 > 2. Protandry. 8: petals about 5 mm; corolla when approaching 3rd stage 9—10 mm in diameter. A particularly large form is shown in Fig. 29, A. 8. Stellaria media (L.) Vill. I have occupied myself very little with this species as I expect it to be fully treated by another writer. The fol- lowing is for the most part taken from the literature on the subject. The flowers open and close according to the weather; VAUCHER states that the flower lasts for one day only. Gynodioecism, Gynomonoecism. $: Protandrous in various degrees, self-pollination found especially in plants which flower during winter. As is well-known it flowers all the year round, even during the whole of autumn and winter, and sets fruit continually, probably in the closed flowers, thus forming a transition to cleistogamy (as regards 5. Bo- raeana Jord., which is said to be a form of S. media, see A. Schulz; it is cleistogamous, with petals either small or absent, like S. media var. brachypetala, and with 2—3 sta- mens). Here self-pollination appears to give excellent results. The flowers vary considerably in size and in the length of the petals relatively to the calyx. LANGE mentions a form apetala. Some of the flowers, viz. the largest, have all the Euc. WARMING. cO N Stellaria humifusa. Fig. 30. Caryophyllaceæ. 299 8: АР, Г. 5: G—K. A—C: $ from Greenland (about 68° N.lat.). D, Parts of $ from East Finmark. E, dimerous pistil with ripe stigmas,. e! (Greenland). Е, Different parts of ÿ from East Greenland (P. Eberlin); protogynous, the anthers not yet opened, but the stigmas well-developed. (The petals vary somewhat in form, which is an unimportant detail; in the same glass, specimens were found with petals of the usual forms), G: from West Greenland (about 67° N.lat.); ©. H, from the same place, larger © flower (compare the two flowers, both mag. 3'/s; also compare h! with gt). J, from Greenland. К, from West Green- land; 2 with stamens abortive in various degrees; pollen-grains formed in the three largest anthers, but thick-walled and irregular in form (kt). Г: 8 from Spitzbergen (Nathorst; 6. 8. 1882). (Е. W.) 10 stamens. But the number is reduced in the smaller flowers and especially in those that flower during autumn and winter; it is in the first instance the cor.-st. which are subject to this reduction, but afterwards the cal.-st. also often share the same fate, so that the number of functional stamens may be as few as two (H. M. in Weit. Beob.). But the nectaries of the suppressed cal.-st. usually remain. Ac- cording, however, to VAUCHER and A. SCHULZ the nectaries may also vanish, or, at any rate, they fail to secrete honey, even if present. BENNETT and others have shown that they are self- fertilizing during winter or in cold weather and so display a transition to cleistogamy. © are rare according to A. ScHurz. According to LoEw a few 2 are to be found on hermaphrodite plants. West Greenland (about 65° N. lat.). Appear to agree exactly with European forms. 9. Stellaria humifusa. Rottb. (Fig. 30). West and East Greenland. Gynodioecism; gynomo- noecism, $ > 2; protandry, as a rare exception protogyny . (East Greenland), then homogamy. 300 Euc. WARMING. $. Protandry with the ordinary movements of the stamens and styles; the cal.-anth. open immediately on the flower expanding. The anthers turn transversely (Fig. 30, C). But as a rule homogamy no doubt ensues very soon, and the backwardly bent styles are in contact with the anthers. Protogyny has been observed in specimens from East Greenland (Fig. 30, Е, f!—f?); the petals were, in addition, less deeply cleft than is ordinarily the case. The styles may number from 2 to 4 (Fig. E); in consequence of the long stigmatie papille at the apex they are rather club-shaped both in $ and 2. The papille extend about 1/, or ?/, way down the styles. Ф are considerably smaller than 8 (compare Fig. А (5) with 7 and С ($). The petals are 3—5mm in 9, but 41/,— 5—6 mm in $ (sepals 4 mm). The diameter of the 2 is 5—6 mm, but that of the $ 7—12 mm. The smallest of the © flowers (from about 67° N. lat.) is figured in Fig. G; the sepals and petals are about 21/, mm in length. The stamens have trans- parent, quite empty anthers; a few may be found with worthless pollen (angular, thick-walled and smaller than is normal; Fig. А, kt). The nectaries are not reduced in size (Fig. К). The stigmatic papillæ are longer in 2 than in 3 (see Fig. ı! and e+). The ovaries are equal in size (see Fig. h® and d!), or smaller in the © (see Fig. 2? and dt). I have noted down a single case of gynomonoecism (Godhavn). Intermediate forms occur with a few or several abortive stamens and the rest normal; as for instance one normal and nine barren. In Fig. К there are three anthers which are apparently normal (one cal.-st. and two cor.-st.), but the pollen-grains are thick-walled and irregularly angu- lar (kt). I have not observed any fruit which had set; it is possible that the species has compensation in its profuse vegetative shoot-formation. Caryophyllaceæ. 301 Norway. Protandry with ensuing homogamy; the styles become functional almost simultaneously with or a little after the cor.-st. (Altenfjord, 10. 7. 1885). Russia (Kola). 8: petals 4!/, mm, slightly longer than the sepals. Diameter about 7mm. Slight protandry with ensuing homogamy. Spitzbergen. Gynodioecism. The very small flowers which occur here are remarkable. In Fig. 30, L are shown parts of a $ with the same magnification as A and С. Sepals and petals are here somewhat more than 3 mm. Pistil small with rather erect styles, which have papille only at their apex. The anthers are normal, open. Can this be a form transitional to 4? An equally small $ has been found in Greenland. The © flowers are even somewhat smaller; the sepals and petals 21/, mm, but the stigmatic papille upon the slender styles are almost of the usual length. Tu. Fries states also that the flowers are smaller and the petals narrower in specimens from Spitzbergen than in those found on the coasts of Finmark. 10. Stellaria borealis Bigel. (Fig. 31). or Gynodioecism. Homogamy, self-pollination. 3, 4 styles (Fig. 31, С, Е). Perigynous (Fig. А, С). Greenland (about 65° М. lat.). 5: Corolla absent. The whole flower 3—4 mm in diameter. Homogamy. Peculiar to this species are the long styles which curve backwards and are thrust among the stamens, so that they touch the anthers (Fig. A, С, Е); the styles have stigmatic papille down to their base (Fig. D). Sometimes 1—3 stamens are abortive. In specimens from Greenland, collected by ROSENvVINGE the anthers were attacked by Ustilago antherarum and the pistil was checked in its development. XXXVII 20 302 Euc. WARMING. Norway (Dovre, Altenfjord). Homogamy. 3: Diameter as in Greenland, sometimes as much as 5 mm. Petals most frequently absent, but they may be present and are then much smaller than the sepals. Stamens 10, with palish- brown anthers. Self-pollination as in Greenland, the styles being directly in contact with the anthers. Fig. 31. Stellaria borealis. A—C: 8 from Greenland. D: 8 from Dovre. E—G, from Kola (Dr. Brotherus). А: $, the anthers have fallen ой. С, Nine anthers have opened and are partially in contact with the widely outspread styles. E—G, А few or several stamens abortive. (Е. W.). Russia (Kola, material from Dr. Е. У. BROTHERUS). Fig. Е, Е, С). Agree with those mentioned above, were also devoid of corolla. But there occurred in addition: ® and forms intermediate between $ and 2 with stamens partially abortive; in Fig. К, they were all abortive except- ing one (а cor.-st.); in Fig. С, the cor.-st. were abortive. The nectaries were not reduced, even when the filaments Caryophyllaceæ. 303 and anthers were reduced (Fig. Г, С). The styles have per- haps somewhat larger stigmatic papillæ than in $ (compare ft and D). According to Lindman, in Norway and Sweden, Stellaria Friesiana Ser. var. alpestris Fr. greatly resembles S. boreale on account of its long styles, papillose a long way down; but it is either protandrous or protogynous. Self-pollination takes place easily. Sagina. Linnei Presl. My own investigations are too incomplete to be recorded. Lup- wic records gynodimorphism (“undeutlich”). A. Scuuxz records that there are two forms which differ as regards the size of the flowers. a: decandra Fenzl (macrocarpa Rchb.), petals = or slightly > than sepals. Most frequently 10 stamens. Homogamy and self-pollination. © occur singly. — b: micantha Fenzl. which is, for instance in Riesen- gebirge, much commoner than decandra; the petals not even half as long as the sepals. Some, but rarely all, of the cor.-st. are always un- developed. 9 occur singly. Pollination often takes place in the closed flower. The flower is closed during dull, rainy weather, and even then self-pollination takes place with good result; consequently, this form approaches cleistogamy. See also A. Scauzz in “Sitzungsber. naturf. Freunde’, 1888, no. 4, p. 184, where he states that in 1886 he was staying several weeks in Hampelbande, but never saw an expanded flower, although several patches of ground were thickly covered with the plant; and the same was the case in 1887 “im westlichen Theile des Gebirges.” Forma decandra, on the other hand, had always open flowers. — In Greenland it sets entirely ripe fruit. 11. Sagina procumbens L. (Fig. 32). The flowers do not open in dull or in cold weather. Denmark; Greenland. I have figured some Greenland specimens in Fig. 32, A—E, for comparison with our Euro- pean specimens. 8: Petals, small, 1?/, mm in length, smaller than the sepals (Fig. 32, A) which are 21/, mm in length. The cal.-st. secrete honey as usual. Both the cal.-anth. and the cor.- anth. may be found in contact with the short, slightly out- spreading styles which have very long, stigmatic papille which extend to their base (Fig. Е, F). The Greenland spe- cimens were found full of germinating pollen, and there can 20* 304 Вос. WARMING. scarcely be any doubt that this is self-pollination, which I have also directly observed in Denmark; even in closed flowers I have seen (June 25th) the anthers shed a quantity of pollen on the stigmas; that self-pollination gives a good . result seems to be evident from the fact that every flower appears to set fruit with hardly any exception. $ I have seen from Denmark, and in Fig. 32, F I have figured a style which differs slightly from that of the 3, but beyond that I have nothing further to state as unfortu- nately I made no notes. Fig. 32. Sagina procumbens. A—E: $ from East Greenland (P. Eberlin). Ff, from Ф from Denmark (Е and 7° =7,). м 12. Sagina intermedia Fenzl. Dovre. Pollen was found abundantly on the stigmas of styles covered to their base with long stigmatic papillæ, and appeared to have got there by self-pollination. Fruit ripens on Spitzbergen and on the north coast of Siberia. 13. Sagina cæspitosa (J. Vahl) Lange. Greenland (The district of Upernivik; 27. 7. 1888; С. Ryder). 3: The petala are either equal to ог < the sepals; homogamy: pollen may be found both in the anthers and Caryophyllaceæ. 305 also germinating simultaneously on the stigmas. The styles are thick, short, erect and covered with long papille down to their base. Self-pollination principally takes place by the erect position of the stamens, whereby the anthers approach the styles. Sets fruit which ripens. Norway (Dovre: S. nivalis* cæspitosa?). Gynodioecism. 2 have small petals, and the stamens are abortive in varying degrees, as some of the anthers contain worthless pollen, and others are quite transparent. But there are also inter- mediate forms with some of the stamens normal and the others abortive, for instance with one normal and the rest abortive. 14. Arenaria eiliata Г. (Fig. 33). Styles 3, 4, 5. In Greenland it sets fruit which ripens on Disco. Norway (Alten, Skjærstad). Gynomonoecism, 8 about — 9, $: Protandry with ensuing homogamy (Fig. 33, С and cl); the cor.-anth. may be partially open before the stigmas are fully developed. Stigmas and anther may be found in contact. The cal.-st. are rather erect, and the cor.-st. are also bent somewhat inwards; they are almost equal in length (Fig. 33, С and ct). The styles are covered with stigmatic papillæ along their upper ?/; part. They ultimately bend so decidedly outwards that they reach across the flower, but at that time the anthers are probably always empty (Sakka- bani, Alten; 10. 7. 1888). © were found in a tuft together with $ flowers, although not with perfect anthers; 2 do not differ constantly from $ as regards size (see Fig. E and a?) and have three outspreading styles (Fig. С). $ in the last stage are 4—5 mm in diameter. The petals of $ are 4—41/, mm in length. The stamens of $ are but slightly abortive; they 306 Euc. WARMING. are about half the size of the normal ones (Fig. Е, G com- pared with C), and there is pollen in the anthers, but it does not appear to fall out and seems to be worthless. The styles appear to be stouter in $ than in 8; they have in particular somewhat longer stigmatic papillæ (Fig. F and D), Fig. 33. Arenaria ciliata. A, from Spitzbergen (Nathorst). 8. f, capsule from the previous year; the darkly shaded leaves also from the previous year; g, a bud; a, sepal; a?, petal (”/ı); the latter is almost 6 mm in length. B—G, from West Finmark (Sakkabani). В: 8; the darkly shaded leaves are from the previous year. Flower scarcely fully ex- panded. С and D from 8. Е, Е, G, from 2. (Е. W.) but this must be verified by further investigations, the dif- ference, as shown by the figures, being slight. Caryophyllaceæ. 307 Spitzbergen (17.7.1882; Nathorst). Fig. 33, А. 8: Appear to be protandrous. Do not differ essentially from the Norwegian specimens as regards size, and appear to agree exactly with them in other respects. MALMGREN records that the Spitzbergen specimens have larger flowers than those found in Russian Lapmark. (“Öfver. af K. Vet. Ak. Fôrh.”, 1862). 15. Moehringia lateriflora (L.) Fenzl. East Finmark. Appears protandrous or homogamous. Ultimately both the stamens and the styles spread out. The styles are hairy along their lower half. 16. Minuartia groenlandica (Retz.) Fenzl. (Fig. 34.) Styles 3, 4. West and East Greenland. 3: Slight protandry, begins no doubt often with homogamy. In large buds I found the anthers of the erect stamens partially open and closely surrounding the erect, spreading styles, whose stigmas were not yet quite ripe (Fig. 34, D). But I have also observed the same position in expanded flowers, and then self-polli- nation is almost inevitable (Fig. Ё). The stamens and styles may be somewhat more outspread as shown in Fig. F, which is evidently connected with conditions of weather. The thin, white petals are 7—9 mm long and 5 mm wide, about twice the length of the sepals, which are 4mm (Fig. А, В, С). Diameter 10—13 mm. I found the seeds to be sometimes finely warted (Fig. G, from Ilua) and sometimes finely wrinkled (Fig. 7, from Dronning Louise’s Island). I am not prepared to say the reason of this difference, which is however but slight. 308 Euc. WARMING. 17. Minuartia biflora (L) Shinz & Thell. (Fig. 35.) The flower is never spread out in a decidedly stellate manner because the sepals keep very erect (Fig. 35, A, B). Gynodoecism (or Gynomonoecism?); 8 = ©. West Greenland (about 69° N.lat.). 8: Slightly pro- tandrous, perhaps entirely homogamous. In a flower like Fig. 34. Minuartia groenlandica. All from Greenland. A (about 1'/.) with a leaf (fol) from Umanak (Rosenvinge). B, from Sukkertoppen (idem). C, Sepal and petal. D, from a bud, two cal.-anth. have opened. E, e: All the anthers have opened with the exception of three cor.-anth. (here, as is so often the case elsewhere, a shrinking of the anther-wall takes place, on dehiscence, which must aid in the shedding of the pollen-grains). F, from an old flower, East Greenland (P. Eberlin). G and H: Seeds from Ilua in South Greenland (Mrs. Lundholm) and from East Green- land (P. Eberlin). — (E. W.) the one figured in Fig. C, the cal.-anth. only were open and shed pollen directly on the stigma (Fig. ct). A similar flower with higher cal.-st. is shown in Fig. Е. When afterwards the cor.-st. also open they are of about the same height as Caryophyllaceæ. 309 the cal.-anth., and at a level with the apices of the styles. The anthers and stigmas are often in intimate contact with each other, the styles being thrust among the anthers; and self-pollination is inevitable. The filaments become ultimately more spreading. In some of the individuals the petals are narrow, almost spathulate, in others obovate (Fig. D, F); Fig. 35. Minuartia biflora. From Greenland. A—E, Parts af 8. From Norway. F (Altenfjord), f! and f? from 9. G ($ which has finished flowering) and Н (seed) from Dovre. they are rounded at the apex, and frequently emarginate or slightly toothed (Fig. D). The sepals are 5 mm in length, and the petals 7 mm; in a flower from Finmark the sepals were only 31/, and the petals 41/, —5 mm long. It seems that slight protogyny may occur, as the stigmatic papille may be fully developed even in the bud, 310 Euc. WARMING. and I have seen pollen-grains retained, but not germinating, on the stigmas of a Greenland specimen, the cal.-anth. of which were on the point of opening. Fruit is set in Greenland, as it seems regularly and abundantly, at least as far as 70°—71° М. lat. Norway (for instance Altenfjord, about 70° М. lat.). Gynodioecism. 3: Slight protandry, as the cal.-st. bend inwards towards the middle of the flower and open their anthers over the still unripe styles; or homogamy and self-pollination; the anthers may be found full of pollen at the same time as there is germinating pollen upon the stigmas, and they may be found in contact with the latter. It is possible that proto- gyny also occurs. 2: All the anthers abortive, but in various degree; they may be found to be shrunken; brownish, without any pollen at all, but with indications of the four pollen-sacs, as there is a transverse fold in each half (Fig. F, f+). 2 = 8 as regards size, but the stigmatic papillae in 2 appear to be constantly longer than in $ (Fig. f? and ct), and perhaps the pistil is, on the whole, somewhat larger, but this I have not measured precisely. A young fruit and a seed is shown in Fig. G and Н. — Linpman finds that they correspond with those of the Alps. Spitzbergen. Appear to agree with Greenland speci- mens, but the flowers appear to be smaller (sepals 31/, and petals 41/, mm). In the specimens from Spitzbergen the anthers were on the point of opening even in the bud, and scarcely has the flower expanded, when the anthers dehisce. — Sets ripe seeds. 18. Minuartia stricta (Sw.) Hiern. Greenland. The plant evidently sets ripe fruit regu- larly, at least as far as 70°—71° N. lat. APE Caryophyllaceæ. 311 Norway (Dovre). Homogamy prevails at least for а time, and as the anthers are at a level with the three highly papillose styles, and as fruit is so abundantly and regularly set, I must presume that self-fertilisation of the small flowers takes place with good result. Fig. 86. Minuartia verna (L.), var. hirta (Wormskjold). From Greenland (Disco, about 69° N. lat.). A with a! (foliage leaf 7/,), а?, a? and af. From Norway. В (Altenfjord) and С (Dovre; 14. 7. 1887), c! and с? (seed). From Spitzbergen (Nathorst; 1. 7. 1882). D (Е. W.). 19. Minuartia уегпа (L.) Hiern.var. hirta (Wormskjold). (Fig.36). Styles 3, 4. Greenland (about 69° N. lat.). A specimen, twice mag- nified, is shown in Fig. 36, A; the length of the petals is almost equal to that of the sepals. 8: Self-pollination and no doubt decided homogamy. Sets fruit which ripens. 312 Euc. WARMING. Norway (Dovre, Altenfjord, about 70° N. lat.). The petals may be still shorter than in the Greenland specimens (Fig. С, D); sepals 31/,, petals 21/,—3 mm; the entire flower may attain a diameter of 5mm. Specimens from Alten have petals only about half as long as the sepals (Fig. В). — 5: Homogamy. In younger flowers the stamens stand erect with the anthers gathered closely above the styles, and pol- len is found simultaneously in the anthers and germinating on the stigma. Finally, the styles curve backwards, and can even then touch the anthers (Fig. a3, cl, D). Self-pollination may take place easily. Fruit is set abundantly. A seed is shown in Fig. c?; it is finely wrinkled. A specimen from Norway (Foldal) appeared to be pro- togynous; the styles were spread out and were quite ripe, all the anthers were closed and stood at a lower level than the ovary, perhaps this was a 9, or a form transitional to it. — I have seen $ only. Spitzbergen (Fig. D). The specimens preserved in alcohol which I have investigated, had their flowers closed, and were homogamous; the open anthers stood at a level with, and were probably pressed closely against, the styles which were somewhat curved backwards (Fig. D); there was a rather large quantity of pollen on the stigma (Fig. D), so self-pollination had no doubt taken place there. Ripe fruit was developed; the seeds resemble exactly those from Dovre. The capsule ‚is only very slightly longer than the calyx. 20. Honckeruja peploides (L) Ehrh. (Fig. 37.)+ Styles 3, 4, 5 (6). 4-merous and 6-merous flowers occur (Fig. 37, С), for instance with 6 sepals, 6 petals, 6 -+ 6 sta- mens, 4 carpels. 1 I do not think there is sufficient cause to place this species in a separate genus (Ammodenia Gmel., a name older than Honkenya Caryophyllaceæ. 313 Greenland, Iceland, Spitzbergen, Nova Zem- bla, Norway and Denmark. I found the flower-biology to be the same everywhere, because the many variations as regards size which are found to occur in the flowers I must regard as local, I have observed these variations to occur in flowers from all countries without it having been pos- sible for me to formulate any rule concerning this. $ and © flowers are common everywhere, no doubt equally frequent, while $ flowers are rare, perhaps even very rare, and in this North America differs markedly from Europe (see Hook. et Benth. Genra pl.). Dioecism, polyoecism (5, 4, $), monoecism; 5 and 3 are. ©. & (Fig. 37, A—D, F—H). At first the cal.-st. are more erect, and in this position they open their anthers; after- wards they bend backwards and become outspread ; the cor.-st. perform the same movements, but I do not think they become so erect; the anthers do not turn. The cal.-st. have very large and yellow, glistening glands (Fig. A, d!, h!, ete.) which evidently secrete honey (by the river Alten I saw a Coccinella search for the honey). The variations in size are very considerable (compare Fig. A, B, C and D). The sepals are 5!/,—7 mm in length; the diameter of the flower is as much as 12—13 mm; petals 3—5 mm long. But in addition Ehrh.; Halianthus Fr., and others), because the fleshiness of its vegetative parts and of its calyx is an adaptation-character to a saline substratum; it has no fleshy disc in “the hollows of which the stamens stand fixed,” as several authors record, but simply the same glands on the cal.-st. as the other Alsinee have (Fig. 37, d!). True, there remain the large seeds, the partially many- chambered capsule, and the development of 2 and also 3, but as regards the kind of flower, in North America they are mostly ©, and the capsule character varies, many small seeds may occur, and the dissepiments may disappear even during the flowering. It can at most be referred to a sub-genus under Minuartia. EuG. WARMING. 314 Honckenya peploides (L.) Ehrh. Fig. 37. Caryophyllaceæ. 315 (The magnification the same, viz. about 21/, for -A, В, С, D, Е, К, L, and ’’, for the majority of the isolated stamens, etc.). 6: A, from Greenland (about 60° N. lat.). All the stamens are open and full of pollen. Detached petal (at), pistil (a?) and pollen- grain (а3). В, from Denmark. С: © from Reykjavik (about 7 mm in diam.). D: from Spitzbergen (Nathorst; 9. 8. 1882); about 6 mm in diam.; cal.-anth. open, cor.-anth. closed; see dt. 8: Е, from Spitzbergen (Nathorst; 9. 8. 1882), taken from a branch on which was a large fruit. Sepals 5 mm, petals 23/, mm. There was germinatiag pollen upon the stigmas; to this belong et and e?. 6: Е, from Tromsö, with style (f!). С, nat. size, with g! and g?; from Altenfjord. H, from Kola (Dr. Brotherus), nat. size; to this belong й1 (one cor.-st. abortive), h? and 13. ©: J, from Kola (Dr. Brotherus) with й and :?; these flowers from Kola are remarkable on account of their long and pointed sepals. К: 9, from Spitzbergen (Nathorst; 9. 8. 1882). Pollen rather abun- dant on the stigma and some of it germinating. The anthers (k!) were quite transparent. L: © from Tromsö; to this belong И, /?, 18. (Е. У.) the relative length of the calyx and the corolla varies, as Bown in Fig. 37 (cal. >cor. in Fig: А, С, D; cal.=cor. in Fig. В); the petals may be as long as the sepals: or much : shorter, for instance, half as long, and examples of both kinds of flowers may be found on the same shore. Flowers are also recorded in which the petals are entirely absent (f. apetala) or in which the petals have partially dis- appeared. The diameter of the flower varies greatly according to the degree to which the sepals have expanded, and in- creases during the flowering period; it varies between 7— 14 mm. The petals vary also in form, and the narrower they are, the more reduced is the venation (compare Fig. a!, d! and g?). I have not found the form in which the corolla is entirely wanting (apetala Rostr.). The length of the stamens varies also, relatively to the corolla and the calyx. The pistil is remarkably large, usually 3—4 mm in length and appears to be normal, but the styles are very short (1/,— 11/, mm), almost conical, erect and minutely papillose, and they always remain so (see Fig. a?; Е, р; #3). 316 Euc. WARMING. The figures show also that they vary in length (the styles of а? are for instance much longer than those of F and #3). The ovary characters also appear to be fairly normal; the ovules are large, perfectly formed (Fig. 37, F) and curved, the integuments are formed, even the long hairs which serve to guide the pollen-tubes are there, but the nucellus appears to be more or less shrunken, and on closer inspection the whole conveys the impression of being abnormal. The ovules are not everywhere abortive to the same extent, even in the same ovary they may be found to vary in size, and in some cases I must even assume that the ovules had been quite normal, but fruit-setting was nevertheless impossible, because the stigmatic papillæ were зо undeveloped (f!). All these flowers die also without setting fruit, and in such withered and closed flowers the size and position of the ovary and the styles is found to be unaltered. Near Altenfjord (Norway) I gathered unusually large 3 flowers (Fig. 37, С): in one 6-merous flower the diameter of the calyx was about 16 mm, and the length of the petals 5 mm. As regards other features see Fig. 21, 2. I have gathered unusually small 4 flowers, for instance оп Sjælland: diameter only 6 mm, petals 3—31/, mm, cal.-st. 5 and cor.-st. 4 mm. In the latter locality the small-flowered plant was growing about 15 degrees south of the large-flowered one. $: & flowers with a few rudimentary stamens occur, but they seem to be rare. Purely $ flowers are common. Here also the flowers vary considerably in size and in the relative length of their different parts, but on the whole the © are <4 (com- pare Fig. h? with J; as also К and L with $ uppermost in the plate: A—D). The diameter of the calyx is for instance 7—9 mm and that of the corolla about 5 mm, and the petals in particular are always very small, usually only 1 to 2 mm long (see Fig. 11, kt, [1). The same figures show their length Caryophyllaceæ. 317 relatively to the stamens. The stamens usually have very short filaments and transparent anthers, but the size of the glands is about as in $. The styles are 2—3 mm in length, and finally they are straight, laterally outspread and highly papillose (compare Fig. ı?, К? and ft). АП the © described above set fruit abundantly, at any rate in Denmark, where, consequently, insects must be present which pollinate them. Fruit is also set in West Greenland and in Iceland. I have seen very small 2 from Spitzbergen (Fig. К) and Tromsö (Fig. L). The sepals were only 31/, mm. There was germinating pollen on the stigma of the flowers from Spitzbergen, so there, also, pollinating insects must be present, and a female plant had fruit on it (9. 8. 1882, Nat- horst). 3 seem to be very rare; I have seen specimens only from Spitzbergen and northernmost Norway; according to my pocket-book notes they are not rare in Greenland; but as re- gards Greenland a confusion with 3 is possible, and my state- ment in ‘“‘Bestovningsmaaden af nogle grønlandske Blomster”, p. 130, needs to be revised (see however North America; note on p. 231). A specimen from Spitzbergen is shown in Fig. Е. They are of medium size and resemble more closely 4, but the styles are longer, outspread and more highly papillose; most likely the papille would have grown longer than those in Fig. e? as the flower is hardly yet in its last stage. As it is they stand midway between ¢ and 5. Monoecism. Near Molde on the coast of Norway I saw fruit on a plant on which there were also withered ¢ flowers. The case mentioned by me in 1878 in ‘‘Bot. Tidsskrift”, vol. 10, of vegetative shoots growing out after flowering and developing vigorously above the uppermost flowers, I have often observed since, for instance also in West Finmark; I have now even seen these shoots begin themselves to flower. XXXVII. 21 EN AX 318 Euc. WARMING. 21. Visearia alpina (L.) С. Don. (Fig. 38). Styles 4—5. West and East Greenland. Fragrant flowers (the perfume recalls slightly that of vanilla). Gynodioecism. 3 > $. 9: Protandry with the ordinary movements of the | stamens and styles, in three sets, finally homogamy. Self- pollination will be able to take place. Scarcely does the flower begin to open, then the cal.-anthers are seen at the throat. The stigmatic papille occur only on the upper half in a con- tinuous row, but farther down small ones occur at intervals. During the flowering period the stamens are thrust far out and are exposed, so that they are quite bared to the effects of rain and dew (Fig. 38, A). The pollen-grains are greenish and ordinary in structure (numerous pores; Fig. a!). Protogyny occurs; at first the styles spread out and then curve backwards; their papillæ are bigger than those in the protandrous flower. Then the cal.-st. spread out, and finally the cor.-st. But this is very probably a form transitional to 9. In the same inflorescence I found both quite sterile flowers and also protogynous ones with stamens partially abortive. 2: Seem to be much rarer than 8. Even before the flower is fully expanded the apices of the styles project beyond the throat (Fig. G). The following figures show the relation between the $ and: the’2: — 5 Petals 91/,—11 mm in length. © Re er en 3 Corolla about 10—12 mm in diameter. PME — 1—8 — —- 3 Corolla-tube (up to the scales at the throat) 4—5 mm. == = == == — 31/5 mm. | In 2 the anthers of the stamens may be quite transpa- ei Caryophyllaceæ. 319 rent, thin and devoid of fibrous cells, but may however attain a length of 2—3 mm (Fig. Е, Е); or they may be fairly normal, have fibrous cells, and well-developed, but useless pollen-grains, as in the intermediate forms (Fig. c!). Fig. 38. Viscaria alpina. A—H, from Greenland; J—K, from Norway (Altenfjord; about 70° N. lat.). — A, & in its second stage; at normal pollen-grains. В, 3 with protogyny; cal.-st. have become outspread, but only one anther has opened. С, 8, with only one normal anther; in all the others the pollen-grains are small and abnormal (ct). D, 2, with fairly normal anthers, but the pollen is worthless. The styles are withered; the ovary is evidently swollen. Е and F, 2 with transparent anthers. G, © in a bud; the styles are already protruding. H, Stigmatic papille of 8. J, it, i (ovules) and 7? from. К, k!, К? (ovules) and k from 4. (E.W.) 21* 320 Euc. WARMING. The pistil is perhaps somewhat larger than in the $, and the styles reach right across the petals, twisting to the left. The stigmatic papille are larger than in the 3 (see Fig. i? and 7). Transitional forms from $ to © occur, the anthers being only more or less abortive in some of the stamens, and normal in the others. Among the abortive stamens there often occur some with pollen and fibrous cells, but the pol- len-grains are smaller than ordinary, thicker-walled, and of irregular outline; the pores are also more distinct (Fig. c+). The barren stamens do not become elongated (Fig. C). In a flower (Greenland, about 67° N. lat.; 13. 7. 1884) one of the cal.-anth. was seen to be open while all the others were closed, and the styles were already entirely outspread; if this was not an imperfect 8, then there must have been protogyny (Fig. B). g. In “К. Danske Vid. Selsk. Overs.”, 1886, р. 129, I recorded the occurrence of 4 in Greenland, but I am now rather doubtful as regards this point. Ripe fruit is commonly set in Greenland. Norway (Dovre and Alten); Sweden (Fig. 38, J. К). Gynodioecism and Gynomonoecism, Andromonoecism; slightly fragrant in the evening. %, protandrous, corolla-tube 41}, —5 mm; diameter sel mm, exactly as in Greenland. A 4-merous flower was found together with 5-merous ones, viz. 4 seps., 4 pets., 4 + 4st. and 3 carp. The number of the styles are 3—4—5. 2: Diameter 7—9 mm, corolla-tube 5—51/,. In specimens from Alten the ovaries were smaller in © than in &, viz. 31/, mm as compared with 4—41/, mm; the styles, on the other hand, were slightly longer. The petals 10 mm in length in ¥ and 61/, in 2. Near Alten female plants were found (15. 7. 1885) which had set numerous fruit the year before. Caryophyllaceæ. 321 Andromonoecism occurs on Dovre. In a dense in- florescence the lowermost flowers were purely 5, while the uppermost ones were more or less distinctly 8. $ > 2 (see Fig. К and /). In $ the petals were 91/, mm and in ? 8 mm in length. In & the pistil is smaller, the ovules are smaller and compressed, but appear otherwise to be fairly normal, only the stigmas are undeveloped (Fig. kt, k?, КЗ, compared Bath 21. 12,12). [Note. Linpman and AxELL: Gynodoecism, 2 common on Dovre; slight protandry. | 22. Melandrium! apetalum. (L.) Fenzl. (Figs. 39 and 40). Greenland. Sets ripe fruit, at least as far as 70°— mic ON lat. Norway (Kongsvold, Alten). $ and 5. According to LinpmMan dimorphous 3, protogynous- homogamous. He has described more closely the variations which occur in the structure of the flowers: (a) some flowers are larger than the others, but are more inconspicuous and have shorter stamens, the petals do not reach to the throat of the calyx; they are dull dark-red (Fig. 39, A, A, at). The flowers are protogynous, and the styles twisted; (b) others are smaller, have protruding petals and longer stamens (see Fig. D, С); he regards them as approaching to 3. The calyx is thinner and lighter coloured. The anthers stand at a higher level than the stigmas. 1 The characters on which the genus Wahlbergella is founded are very insignificant, and the genus ought not to be retained; also the characters are not correct; as for instance the record in Hart- mann’s Flora that the claws of the petals of Wahlbergella are wingless, but those af Melandrium are winged, or that the seeds have a membranous edging; for M. triflorum, which evidently stands very near to the two Wahlbergella’s, has wingless seeds (see Fig. 41, H). 322 EuG. WARMING. 2 flowers are also mentioned by Linnman, they belong to form a. . Г have found the same forms, not only the different ÿ flowers, but also the 9. As regards the former, reference should be made to Lınpman’s figures; I shall only remark that according to specimens which I have seen cultivated Fig. 39. Melandrium apetalum. A, A!, a1: 6 with short petals (f. brevipetala) ; Dovre, 14. 7. 1887 (calyx 18 mm long, petals 13 mm; cal.-st. 8—9 mm. The stigmas are covered with pollen). В, b!, b?, 63: © from Altenfjord (about 70° N. lat.), 18. 7. 1885. The petals and the styles are just at the throat of the calyx (61); 53: anther with crystals of calcium-oxalate. С, D: Flowers from a cultivated plant (seeds from Petrograd); the petals are protruding; in D the calyx is greatly swollen. Æ, Seed (Dovre). F, Dorsal view of petal. (E. W.) (seeds from Petrograd), there occurs a form with much more inflated calyx, becoming almost globular, even during the Caryophyllaceæ. 323 flowering period; it has protruding petals (Fig. С and D). In this form the flowers do not become erect after flowering, but remain nodding. Self-pollination appears inevitable, as the anthers no doubt lie for a long time quite close to the stigma, or else the pollen falls from them down upon it; an- thers may also be found which are in direct contact with the stigmas. I give a few figures of © which show that the anthers may be fairly normal in form, but they are transparent, devoid of pollen and with crystals of calcium oxalate (Fig. bl, b?, 63). It is smaller than $ (Fig. В and A); the total length about 12 mm against 17mm; petals about 9 mm, cal.-st. scarcely 5 mm and cor.-st. 3 mm, styles 2 mm, thicker than usual (Fig. 51). The ovary also appears larger, about 6 mm long. Tu. Fries has established a variety arctica (“Tillägg. til Spitzbergens Fanerogamflora” in Öfvers. of K. Vet. Ak. Förhandl., 1869, p. 133); it has protruding petals, conse- quently, it seems to come near to the above-mentioned form which LinpMAN describes as being more staminate, and is perhaps identical with it. In the flowers of this form from Spitzbergen, which I have investigated, the limbs of the petals were protruding and the anthers were standing just at the throat, at a level with the uppermost tips of the styles. They appeared to be protogynous. As an indication of self-pollination may be mentioned that pollen was found only on the upper end of the style, just where the cal.-anth. could touch it, as long as the cor.-anth. which were standing at a lower level, had not dehisced. On the other hand, in a flower in which the cor.-anth. also had dehisced, without the cor.-st. having as yet become outspread, there was also pollen on the middle of the style, separated off from that of the upper part, and its situation corresponding to the level of the lower anthers; it had not yet germinated, which, 324 Euc. WARMING. on the other hand, the pollen situated higher up had. Thus, there was both as regards the position of this pollen and the degree to which it had germinated, every reason to believe that it came from the anthers of the same flower. For the rest, an immense quantity of pollen may be found deposited on the styles. Petals total length 15 mm, cal.-st. 10 mm long, cor.-st. 8 mm, pistil 8—81/, mm, of which the styles about 3 mm. Fig. 40. Melandrium apetalum. From Nova Zembla, gathered by Th. Holm. F, Anthers open; many pollen-grains in them, some of which are germinating. (E. W.) I have seen flowers from Nova Zembla (Fig. 40) which in a way stand intermediate between the two forms described by Linpman. The petals are just visible at the throat of the calyx (Fig. F), consequently, they are longer than in the one of his forms, but shorter than in the other; the anthers are at the level of the stigmas; much as in one of his forms, open, and quite evidently able to carry out self- pollination. In addition, the pollen was germinating in abundance not only in the anthers, but also on the stigmas. In Nova Zembla fruit is set which ripens. Caryophyllacesæ. 325 23. Melandrium triflorum (В. Br.) J. Vahl (Fig. 41). Greenland. I have described it in “Overs. over d. K. danske Vid. Selsk. Forhandl.”, 1886. Slight perfume. $: Slightly protogynous; self-pollinating. Both the eal.-anth. and the cor.-anth. lie finally close to the styles and are in contact with them. The styles, surrounded by i # RTS Yi RSA: Yes N; K AN м er: EHG Fig. 41. Melandrium triflorum (В. Br.) J. Vahl. A, The limbs of the petals are more or less outspread. The appen- dages of the throat (see Е) are hollow. В, On the slightly club-shaped styles the stigmatic papillæ are fully developed (see Г), but the anthers are closed and stand at a lower level than the styles. In C the cal.-st. have straightened, and the anthers have opened at the level of the middle of the styles, almost at the throat of the flower. In D the cor.-st. have also straightened and opened their anthers (the parts of the flower are spread out artificially). As the stigmas are still ripe, self- pollination will certainly be inevitable. In flowers at this stage of development I always found a great quantity of pollen on the stigmas, partially germinating (K). F and G almost ripe fruit. H, a seed. Some of the hairs have been omitted. From Kristianshaab, W. Green- land. (E. W.) 326 Euc. WARMING. the anthers, are situated just at the narrow throat. A large quantity of pollen-grains may be found on the stigmas, some of them germinating. Six styles may occur. Petals 12—121/, mm. Throat-scales hollow. Dehisced cal.-st. 81/, mm, pistil 71/, mm to the tip of the styles. Diameter of corolla 12 mm finally. Fruit appears to be set very regularly. (See expanation of Fig. 41). Fig. 42. Melandrium triflorum. From Greenland (about 73° N. lat.; С. Ryder, 5. 7. 1887). Parts of 2; al and a?, anthers showing different degrees of reduction. ай, Dorsal view of petal. I have received © flowers of this species also, gathered by Lieut. С. RypER at about 73° N. lat. in West Green- land; one of them is illustrated in Fig. 42. Four of its cor.-st. contain a little pollen, but it is not normal; in another flower only one or two of the cor.-st. contained a little pollen. All the cal.-st. are smaller than the cor.-st. and more reduced. Petals 13 mm, cal.-st. 6 mm, cor.-st. 81/, mm, styles 5mm, diameter of corolla 8—10 mm. Transitional forms from 8 to $ occur with a few or several anthers developed, for instance, all the cal.-st. sterile, and all the cor.-st. normal and longer than the cal.-st. Caryophyllaceæ. 327 In Greenland fruit is set which ripens even as far north as 76° N. lat., and fruit-setting appears to begin very early. 24. Melandrium affine J. Vahl. (See Warming in “Oversigt over D. Kgl. Danske Vidensk. Selsk. Forhandl.’’ 1886, р. 129). Fig. 43. Melandrium affine J. Vahl. A, B, Hermaphrodite flowers. C, a seed. D and E from male flowers; the filaments are developed, but not the anthers, and the petals differ in form from those of 5. From Altenfjord, northern Norway. (Е. W.) Greenland. ÿ: Slight protogyny. Self-pollination. Calyx 12 mm, corolla-tube about 10—11 mm, cal.-st. 91/,mm, cor.-st. 81/,—9 mm. The calyx-anthers reach to the tip of the styles. The anthers shed masses of pollen upon the stigmas, with which they are in direct contact. Petals white. Sets ripe fruit in West Greenland, regularly it appears, for instance at 70°—71° N. lat., and even as far as Grinnell Land (Discovery Bay, according to a specimen from the 328 Вов. WARMING. Herb. Upsala, which was forwarded as L. apetala L. var., but according to Tu. FRIES is a true Mel. affine). Norway (Altenfjord, about 70° N. lat.). 8, 9. Protogyny. 3: The anthers of the cal.-st. in my Fig. 43, В are open and pressed against the styles to which they adhere by means of the pollen tubes. The styles (4—5—6) are slenderer than in 2; the cor.-anth. have not yet dehisced. "In another $ all the anthers were of equal height and at a level with the upper one-third of the styles, consequently they were all like the cal.-anth. in my Fig. 43, B. The stig- matic papillæ extend down to the base. | $ (Fig. 43, D, Е): The petals are narrow, smaller and less notched at the apex, and the scales at the throat are absent, the stamens sterile, either with anthers (pale, trans- parent) or even entirely without anthers, but terminating in long hairs. Styles 4, thick, twisted. The stigmatic papillæ are short, and fairly equal in length from base to apex. The ovary is considerably larger (both longer and thicker) ш $ than in $ (Fig. Е and В). Spitzbergen (15. 7. 1882; Nathorst). 3: Protogyny. In a well-developed bud the styles were spread out and the stigmas almost ripe, and the anthers had so far developed that the pollen-grains were lying isolated. In older flowers the stigmas were found covered with germinating pollen before the cor.-st. had elongated and opened their anthers; this pollen comes evidently from the cal.-anthers. Fruit ripens in September; also in Nova Zembla. 25. Silene inflata (Salisb.) Sm. Flower scentless; moth-pollinated. The petals are not involute?). 1) Улоснев has a great many good biological observations on the Sileneæ which have not been noticed in the newer literature, for Caryophyllacee. 329 Denmark. Trioecism and gynomonoecism; protandry. The scales at the throat consist of two low protuberances with corresponding shallow depressions on the dorsal side. Zygomorphy is often distinctly seen, the stamens and styles being bent downwards in the lower part of the flower; the flower is what Delpino calls zygomorphic of the Ist degree. $: Highly protandrous, the stamens appear gradually not exactly in two sets. The stamens project far out. 4 are almost equal to 5 in size; they are hardly common. $ < 8. Diameter of corolla 13—14 mm, in § 15—18 mm “(the corolla grows during the flowering period), but other- wise the difference is but slight. In both flowers the calyx is 12—13 mm in length, the corolla-tube in 2 9—10!/, mm, in $ 11 mm; petals in 2 13—14—16 mm, in $ 14—16!/, mm, the limb in 2 4—5 mm, in 8 51/, mm. The ovary is 31/,— 5mm in length in both; and the thickness and the length of the style are the same in both. Nor can I see any differ- ence as regards the ovules; there are about 12 in each row, both in 5 and in 9. The calyx-stamens in © are often remark- ably long, viz. 41/,—5—61/, mm in length, so that they project far above the ovary, and they have distinctly marked- off, dorsified anthers; the length may diminish to 2—21/, mm. The cor.-st. are as usual somewhat smaller. The anthers are instance with regard to the involution of the petals (those that become involute in the evening he calls “meteoric corollas”); and with regard to protandry, direct or indirect pollination (e. g. self-pollination and crossing); and about movements before, under and after flowering, etc. In one place he writes as follows: — “Du reste, on peut croire qu’en observant les divers Silene sous ce point de vue, on y découvrirait d’autres arrangements qu’on ne soupçonne point encore, et qui donneraient une idée bien plus grande de la richesse et de la variété que le créateur à mises dans ses ouvrages, que ne peuvent le faire les differences de feuilles ou d’inflorescences”. 330 Еоб. WARMING. transparent, thin and without the least indication of pollen. The styles protrude 8—10 mm and are on the whole 15— 16 mm long. LANGE indicates the female plant as {. micropetala. Gynomonoecism: А few © together with $8. Norway, Sweden: Gynodioecism. I found 2 in Oster- dalen and оп Dovre, as also in the neighbourhood of Stock- holm; it may have very rudimentary stamens, and the dia- meter of the corolla may be 17—18 mm, consequently much larger than recorded above. % protandrous. Both zygo- morphic. Trioecism with $ and 2 > 6 according to AXELL. 26. Silene maritima With. Norway (Altenfjord). $: Rather markedly protandrous with the usual development in three sets, but the cor.-st. may spread out singly and almost simultaneously with the cal.-st. Calyx 15—16 mm in length; diameter of corolla 13— 15 mm, corolla-tube about 13 mm, stamens 14 mm in length, protruding beyond the throat. As the styles ultimately come into contact with the anthers, and the latter may still con- tain pollen, self-pollination may take place. The scales at the throat consist of such low protuberances as to be almost wanting. At night the flower stood fully expanded, white and fragrant. After rain there was much water in the base of the flowers, even among the basal portions of the stamens. Iceland (Reykjavik; 4. 8. 1886, A. Feddersen). Rela- tive size of parts about as those given above, but the dia- meter of the corolla 18mm. Decided protandry. Ultimately the styles protrude 6—7 mm. Various flies and plant-lice were found between the calyx and the corolla. 27. Silene rupestris L. Norway (Romsdalen, Dovre). %: Protandrous in three sets, rather markedly. Calyx about 4—5 mm, petals 41/, mm, Pr Caryophyllaceæ. ; aa corolla 6 mm in diameter, corolla-tube 3—4 mm, the throat is about 11/, mm wide, so that entrance is easy for insects with short proboscises. Switzerland. The corollas are not “meteoric”. 3: Protandry, but styles and stigma develop while the cor.-anth. are shedding their pollen. Possibility of self-pollination (Vaucher; H. M.). 28. Silene acaulis L. (Fig. 44). Fig. 44. Silene acaulis L. A—F, from Greenland; G, from Norway. — A, & (about 69° N. lat.); all the anthers open; the parts shown in D belong to this. B, 2; Е belongs to the same form. С, 2 (about 67° N. lat.); Е belongs to the same form. D, from 3 from Greenland; petal with stamen and pistil 31/,; styl, apex of style. Е, from © (Upernivik; С. Ryder); the petals and pistils 31/,, the staminodes (std.) 37/,. Е, from © from East Green- land (11. 8. 1888; P. Eberlin); the pistil with the corolla 31/,; styl, style; sid. sterile stamen. С, from © from Tromsö. H, Dorsal view of petal, showing the openings into the cavities in the scales at the throat and that petals with lateral projections may occur (compare G). Wherever I saw it, I found the colour to vary; white specimens occurred here and there. Slight perfume in Green- 332 Euc. WARMING. land. In the Greenland specimens I found flowers formed the year previous to that in which they were to open. In specimens gathered on May 10th, 1887, by Lieut. RYDER, near Upernivik, young flowers, about !/, mm in diameter, were found far down among the leaves. (See p. 282, Fig. 24). Greenland. Trioecism: %, $ and $; as 2 sets fruit, we have here evidence of insect-pollination. I have not seen $ flowers, but LANGE mentions them in “Conspectus Flore Groenlandiæ” on the authority of VAHL. $ are > 2 (Fig. 44, A, В, С), which is especially seen in the length of the petals and in the diameter of the corolla. In $ the petals are 8—1N mm, in 2 51/,—61/,, of which the claws are about 4mm long and the limb scarcely 2 mm (see Fig. D, Е, F). The calyx tube is generally larger in & (5— 7mm) than in 2 (4—5 mm). The corolla-tube in & is 6— 61/, mm, in © 4—5!/, mm, the diameter of the corolla in g is 7—11 mm, in 2 6—8 mm (see Fig. A, В, С). & has always a small pistil with styles 11/,—2 mm in length, which I have not seen grow and separate (Fig. D), as they do in the Alps according to H. MÜLLER. There are shrunken ovules, the inte- guments of which are however formed. The cal.-st. are the first to develop, then the cor.-st. The styles have at the top some very short stigmatic papillae, which never grow longer than those shown in Fig. D. 2 flowers have always rudiments of stamens, but in vari- ous degrees of development; in some cases they are as much as 2 mm long, and have distinctly marked-off anthers (Fig. E, std), which are however shrunken, transparent and devoid of pollen; in other cases they have still shorter stamens, and the anthers are hardly indicated (Fig. F). Correspondingly with this there appears to be a different degree of develop- ment of the size of the flowers, the size of the petals, and the degree of development of the stigmatic papillæ, so that the Caryophyllaceæ. 333 flowers which have the larger stamen-rudiments are larger and have shorter stigmatic papillæ, while the flowers which have smaller rudiments are smaller and have longer stigmatic papille (compare Fig. В and E with С and F); the former are nearer to the hermaphrodite condition, while the latter are more decidedly female. About 3/, of the style is covered with papille. The pistil is 4—6 mm long, the styles about 3 mm, and the ovary 2—31/, mm; consequently, there is no great difference between the size of the ovary in the 5 flower and $, as Fig. D, Е, Е also shows. Ripe fruit is set in Greenland; in the capsules investi- gated I have however found many aborted ovules besides the ripe seeds, of which there were often only 2—4 in each capsule. This is indicative of defective pollination. Spitzbergen. Dioecism. The $ flowers agree with those from Greenland. Petals 10 mm and cal.-st. 8 mm. AURIVILLUS is of opinion that the passage to the honey is shorter in Spitzbergen flowers than in those from southern countries; this I regard however as doubtful. Iceland (North coast; St. STEFANSSON). Monoecism, dioecism, andromonoecism. $ protandrous; the corolla-tube short, 3—4 mm; 2: the tips of the styles protrude before the flower opens. Petals as much as 7 mm, corolla-tube 5 mm. Sterile stamens as much as 41/, mm, pistil 8 mm, of which the styles occupy 3—4 mm; consequently, unusually large female flowers. STEFANSSON has gathered flowers from the same place, which are 2, but approach closely to 9: the anthers are transparent and without fibrous cells; the cal.-stamens are altogether 8 mm long; the pistil is 5 mm, and the styles are somewhat shorter than in ©. Norway, Sweden. (I have examined specimens from Finmark, Nordland, Dovre, Tronfjeld, Areskutan and Jemteland). Trioecism, but $ and 2 have been found on the XXXVII 22 334 Euc. WARMING. same individual (trioecism and monoecism: Axell). 8: the stigmatic papille were larger than in 4. According to LinpMAN the high degree of protandry makes self-pollination impossible. 9: There are two sizes of © flowers (corolla 5—11 mm in diameter) according to LinpmAn. From Tromsö I have seen specimens with corolla-tube 5 mm, staminodes 1 mm, ovary 3 mm in length and corolla 7—9 mm in diameter. From Dovre with diameter 51/,—7, corolla-tube 5—6 mm, petals 9 mm, calyx-tube 5 mm; the pistil altogether 81/,—9 mm; the rudi- ments of the stamens 21/, mm. Even before the flower has opened the styles protrude. $: like those from Greenland. Ovules rudimentary. Corolla 8—9 mm in diameter, on Dovre even 12 mm, indeed a few as much as 15 mm, calyx- tube 61/,—8 mm. The pistil-rudiment may attain a length of 31/, mm, petals 14 mm, and the corolla-tube 7 mm. Near Hammersfest and Tromsö I saw many ripe fruits in the female plants; on the Skaadavara mountain in Altenfjord, on the other hand, they had remained barren. The Alps (Vaucher, Koch and H.M.). Trioecism, monoecism. The 5 flower is as a rule larger than the 3, which is larger than the 9. Marked protandry, self-pollination is possible in the case of the 3 flower (H. M.). Butterfly pollination. 8 is found only singly, here and there, but appears nevertheless to be of far more common occur- rence than in the North. 29. Dianthus superbus L. This large-flowered species with fragrant, finely divided petals seems to occur with $, 2 and $ flowers. According to VAUCHER the flowers open in the evening and close irregularly during the warm hours of the day. Denmark. 8: Protandry. $ (?): the form mieropetalus Lge. is probably a female plant. Kola (specimens from Dr. Brotherus). At the transition- point between the limb and the claw there are long unicellu- Caryophyllaceæ. 330 lar hairs with undulating walls and filled with sap. The corolla-tube is about 2 cm in length. $ and 2 occur, or at any rate forms which approach much to $; their stamens reach only to the upper part of the ovary; the anthers are about 11/,—2 mm long and well-developed, but the pollen contained in them is evidently useless, shrunken, etc. A fly was found in one flower. Germany. SPRENGEL figures it; gynodioecism; self-pollination impossible. § > 9. — The Alps. Gynodioecism, androdioecism (?; Vaucher); butterfly-pollination with large flowers in plains and small ones on mountains. 3: corolla 50—-60 mm in diameter, tube 20—25 mm. Marked protandry with movements of stamens. Self-pollination im- possible (Vaucher; H.M.). 9: It does not appear.to be a constant feature for the 2 flowers to be smaller than the 5; H. MürLer, however, records the diameter as 36—45 mm; they are many times rarer than 2. Rudiments of stamens occur; useless pollen may be present. 3: VAUCHER writes that “les stigmates souvent avortent”. General remarks concerning the Biology of the Flower. The Formand Venationofthe Petals. A single vein enters the petal at its base, and immediately divides into three branches (cf. the foliage-leaves), which may themselves branch, becoming reticulate in large petals (Silenex); (Figs. is, 21, 23). The Appendages of the throat (ligules). In the Suleneæ they are in some cases solid, and in others hollow as in the Boroginee. The Growth of the Petals. The petals often grow during the expansion of the flowers, — this is at all events frequently the case in the Alsineæ, — therefore, conclusions with regard to the occurrence of large-flowered or small- flowered types, should be formed with great caution. Secretion of Honey. Honey is secreted in all the Caryophyllaceae; in the Alsinex it is from the gland-like, 22% 336 Euc. WARMING. swollen, yellowish base of the calyx-stamens; it accumulates in the sometimes hollow base of the sepals (Spergularia and the Paronychieæ deviate from this rule). In the Sileneæ honey is secreted by the inside of the ring which unites the bases of the stamens and the petals, perhaps also at the base of the ovary. In the female flowers, also, the size of the nec- taries is essentially unaltered, even when the anthers are considerably reduced. The Anthers in the Alsinex are often twisted in such a way that the pollen-covered surface is turned upwards or even outwards, in correlation with the fact that the honey occurs between the petals and the stamens (see H. MULLER and A. SCHULTZ). In self-pollinating individuals this feature will hardly be found. The Styles are, for a longer or shorter distance, covered with stigmatic papillae, varying in length, being longest т the large-flowered Sileneæ. I have, however, never found germinating pollen far down, near the base of the style; but whether this is because the papilla occurring there, are not functional, I do not know. In Cerastium trigynum the styles are very broad towards the apex, even sometimes bi-lobed. In older flowers they are twisted spirally. either to the right or the left, most markedly in the Sileneæ. VAUCHER has al- ready explained this as being suitable to the purpose: “afin de recevoir plus facilement le pollen des fleurs mâles.” The Development of the Flowers after Expan- sion. Protandry is very common in the Arctic Caryo- phyllaceæ, and the movements of the stamens are the same as elsewhere: first the calyx-stamens stretch out, and bend inwards towards the middle of the flower, then the corolla- stamens perform the same movements, and lastly the styles ripen and spread out; sooner or later however homogamy ensues in almost all the species, and the flower may be entirely Caryophyllaceæ. 397 homogamous from the beginning. Protogyny is very rare (Melandrium apetalum, M. triflorum and M. affine, Cera- stium trigynum, and according to Шмомлм Stellaria Frieseana var. alpestris). Abortion. When the stamens are aborted in a pro- tandrous flower, the corolla-stamens are the first to be re- duced, and they are also always even smaller and younger than the calyx-stamens. In protogynous flowers the order of development is reversed. Everywhere did I find confirmation of the rule established by H. MÜLLER, viz., that protandry or, on the whole, the degree of dichogamy, is correlated with the size of the flower. The larger the flower is, the more pronounced is the protandry, provided the conditions are otherwise similar. In the foregoing | have arranged the species belonging to each genus in the order of the decreasing size of the flower, begin- ning with the largest-flowered. An examination of some of the genera will prove the correctness of Mürrer’s rule — of course with exceptions, for there should be laid stress on the fact that the external conditions must be similar. Some species are homogamous almost from the begin- ning; these are either Arctic or Alpine, or else autumn and winter flowers, or they are very small-flowered. In the Paro- nychieæ pure homogamy seems to be frequent. These facts appear to me to be connected with the fact that the Arctic- Alpine and the small flowers are probably not so quickly developed as in southern countries, or as are the large flowers. It is above all in the physical conditions that we must look for the explanation of these differences in the development. Self-pollination is by no means rare; it appears to be least common in some of the large-flowered Sileneæ, for instance Dianthus superbus; the vigorous development of the corolla is there followed by a quicker development of the 338 Euc. WARMING. stamens. But in many homogamous or slightly dichogamous and small-flowered species, self-pollination is a process which regularly takes place, and which produces an excellent result. It is seen here, as in the Greenland Crucifere, Saxifra- ge, etc., that fruit-setting follows immediately after flowering, and that there is a gradual succession of the sizes of the fruits, according to the age of the flowers. Polygamy (Pleogamy). Many of the High Arctic, and even small-flowered and homogamous species have poly- gamy, especially gynodicecism (for instance Cerastium alpi- num, С. trigynum, Stellaria longipes, 5. borealıs, Minuartia biflora, and others, even Arctic species of Melandrium). Poly- gamy must be regarded as an unfavourable feature where the country is as poor in insects as are the Arctie countries. From the preceding detailed descriptions it is evident that few species have hermaphrodite flowers only, and con- tinued investigations will probably result in the reduction of even their number; also probably many more combina- tions of $, 2 and 3, than are now known, will be found. A peculiarity common to the gynodicecious species is the varying degree of reduction of the stamens in the female flowers; many examples of this have been mention- ed and figured above. There is every possible degree of reduc- tion of the anthers, according as the development which has once begun, stops earlier or later; I have seen no female flower which was entirely without rudiments of stamens. To this must be added those hermaphrodite flowers in which a few or several stamens are abortive. Several successive stages of more or less decided abortion may thus be demon- strated, and that often in one and the same species. Another peculiarity which accompanies the one mention- ed above, is that the corolla is diminished in size more or less in all the female flowers. There are extremely few Caryophyllaceæ. 339 Caryophyllaceæ of which the female flowers are similar to the hermaphrodite flowers in size, as they are in Minuartia biflora. The alterations in the structure of the flowers in indivi- duals of the same species, of which so many examples have been given in the foregoing, will be understood, to a certain extent, if the ontogeny of the flower is known. Those organs which have developed first, or at an early stage, grow most quickly and vigorously in the flower (especially the calyx and pistil and the calyx-stamens), and will have the best chance of becoming developed, should any check occur; while the last-developed or slow-growing organs may become dwarfed, or are even completely suppressed — all according to the point of time when the check occurred. In the pro- tandrous flowers it is the corolla-stamens which are sup- pressed whenever the androecium is reduced, which is in perfect harmony with the ontogeny. In the protogynous flowers the pistils hasten as usual, and attain full maturity before the stamens, and so, if the check occurs very early, a purely female flower may be produced. The size of the flowers evidently has also a certain con- nection with the longevity of the individual: annual species generally have smaller flowers than perennials; this is also confirmed with regard to the Caryophyllaceæ as a whole. Further, the numbers in the flowers are dependent on the size of the flowers. But the causal conditions are still problematie in many respects, viz., with regard to the dependency of the form of the flower on the prevailing conditions, especially those relating to nutrition. Dichogamy or pleogamy in the Caryophyllaceæ must, as a whole, be regarded rather as a sign of weakened constitution, or of unfavourable conditions of life, than as an advantage. 340 Euc. WARMING. Principal Literature. ANDERSSON, GUNNAR, 1902. Zur Pflanzengeographie der Arkties. Geogr. Zeitschr. BONNIER, G., 1888. Étude expérimentale sur l’influence du climat alpin sur la végétation et les fonctions des plantes. Bulletin de la Société botan. de France 1888. — 1890. Cultures experimentales dans les hautes altitudes. Compt. rend. de l’Académ. — 1890. Influence des hautes altitudes sur les fonctions des vé- gétaux. Ibid. 111. — 1894. Sur les plantes arctiques comparées aux mêmes espèces des Alpes et des Pyrenees. Revue génér. de Botanique, 6: 505. — 1895. Recherches expérimentales sur l’adaptation des plantes au climat alpin. Ann. d. sc. nat. (7), 20. BorGEsEn, F., 1895. Bidrag til Kundskaben om arktiske Planters Bladbygning. Botan. Tidsskr. 19. Kjobenhavn. 3 pl. — 1895. Sur l’anatomie des feuilles des plantes arctiques. Journal de Botanique. Eriksson, JoH., 1896. Studier öfver Sandfloran i östra Skane. Bihang t. k. Vet. Akad. 22. GALLOE, OLAF, 1910. The Biological Leaf-Anatomy of the Arctic spe- cies of Saxifraga. 29 figures. Meddelelser om Gronland, 36. Hartz, N., and Cur. KRUUSE, 1911 (see KRUUSE). HEINRICHER, E., 1884. Über isolateralen Blattbau mit besonderer Berücksichtigung der europäischen, speciell der deutschen Flora. Pringsh. Jahrb. 15. m. 5 Taf. Hoım, Тн., 1887. Novaia Zemlias Vegetation. Dijmphna-Togtets Zool.-Bot. Udbytte, med 12 Tavler. (Résumé francais). Kje- benhavn. i JESSEN, Knup, 1911. Ranunculaceæ in „The Structure and Biology of Arctic flowering Plants”. 58 figures. Meddelelser om Gren- land, 36. — 1913. Rosaceæ; 41 figg.; ibid., 37. Jonsson, Heuer, 1895. Optegnelser fra Vaar- og Vinterexkursioner i Ost-Island. Bot. Tidsskr. 19: 273—294. Krutman, A. O., 1890. Pflanzenbiologische Studien aus Russisch- Lappland. Acta Soc. Faun. Flora Fenn. 6. KJsELLMAN, R., 1901. Se Botan. Notiser. KRUUSE, Снв., 1905. List of Phanerogams and vascular cryptogams found on the coast 75°—66°20’ lat. N. of East Greenland. Med- delelser om Gronland, 30. Caryophyllaceæ. 341 KRUUSE Снв., 1906. List of Phanerogams and vascular cryptogams found in the Angmagsalik District (65°30’—66°20’ lat. N.). Ibid. 30. — 1911 (and N. Hartz). The vegetation of Northeast Greenland. Meddelelser om Grenland. 30. Kjabenhavn. — 1912. Rejser og botaniske Undersøgelser i Østgrønland о. $. У. Meddel. om Grønland, Bd. 49. 4 Tavler, 44 Textfig. LANGE, Jo#., 1880. Conspectus Flore Groenlandice. Meddel. om Grønland, 8. LAZNIEwskI, W. v., 1896. Beiträge zur Biologie 4. Alpenpflanzen. Flora, 82. Leist, K., 1889. Über den Einfluss des alpinen Standortes auf die Ausbildung der Laubblätter. Mittheilungen der Naturforsch. Gesellsch. von Bern. LUNDAGER, ANDR., 1912. Some notes concerning the Vegetation of Germania Land, North-East Greenland. Danmark-Ekspeditio- nen til Grønlands Nordostkyst. 1906—1908. Meddelelser om Gronland, 43: 347—414. MATHIESEN, Ев. J., 1916. Primulaceæ; 25 figg. Meddel. om Gron- land, 37. MENTZ, A., 1909. Empetraceæ. (Structure and Biology of Arctic Flowering Plants. I.). Meddelelser om Gronland, 36. Kjo- benhavn. OLSEN, CARSTEN, 1914. Cornaceæ; 13 figg. (The structure and biology af Arct. Flowering Plants. П.). Meddel. om Grønland, 37. Kobenhavn. OSTENFELD, С. H. and Амов. Lunpacer, 1910. List of Vascular Plants from North-East Greenland (N. of 76° N. lat.). Dan- mark-Ekspeditionen til Gronlands Nordostkyst. 1906—1908. Meddelelser om Gronland, 48: 1—32. PETERSEN, HENNING E., 1908. The biological Anatomy of the Leaves and of the Stem (of Arctic) Ericineæ: 39 figures. In The Struc- ture and Biology of Arctic Flowering Plants. Meddelelser om Grønland, 36. — 1908. Diapensiaceæ; 9 figures. Ibid. PETERSEN, Q. G., 1888. Momenter til Caryophyllaceernes Anatomi. Botan. Tidsskr. 16: 216. Kobenhavn. Porsitp, M., 1902. Bidrag til en Skildring af Vegetationen paa Øen Disko. Meddel. om Gronland, 25. Tab. 1—6. Rés. franc. — 1910. The plant-life of Hare-Island off the coast of West- Greenland. Meddel. om Gronland, 47. — 1912. Vascular Plants of West Greenland between 71° and 73° N. Lat. Meddel. om Gronland, 50: 349. ReEsvoLL, THEKLA, 1917. Om Planter som passer til kort og kold Som- mer. Archiv for Mathem. og Naturvid., 35. 64 Textfigurer. Russet, W., 1895. Influence du climat méditerranéen sur la structure des plantes communes en France. Ann. d. sc. nat. (8), 1. 349 Euc. WARMING: Caryophyllaceæ. SCHROETER, C., 1908. Das Pflanzenleben der Alpen. Zürich. Simmons, H. G., 1909. A Revised List of the Flowering Plants and Ferns of North Western Greenland. Report of the second Norwegian Arctic Expedition in the “Fram” 1898—1902, no. 16. Videnskabs-Selskabet i Kristiania. — 1913. A Survey of the Phytogeography of the Arctic American Archipelago. Roy. Physiographic. Society. Lund. N.F. (2), 9. SOLEREDER, H., 1899. Systemat. Anatomie der Dicotyledonen. Stutt- gart. SYLVÉN, N., 1906. Om de svenska Dikotyledonernas första Förstärk- ningsstadium. K. Sv. Vetenskapsakad. Handlingar. 40. WAGNER, A., 1892. Zur Kenntniss des Blattbaues der Alpenpflanzen und dessen biologischer Bedeutung. Sitzungsber. d. kais. Akad. d. Wissenschaften in Wien. Bd. 101, Abt. 1. WARMING, Euc., 1877. Smaa biologiske og morfologiske Bidrag. Botan. Tidsskr. 10 (3. R., 2). — 1884. Om Skudbygning, Overvintring og Foryngelse. Natur- hist. Forenings Festskrift. Kjobenhavn. — 1886. Om Bygningen og den formodede Bestovningsmaade af nogle gronlandske Blomster. K. Danske Vidensk. Selskabs Oversigt, S. 101—159; 15 Figg. — 1887. Om Gronlands Vegetation. Meddelelser om Gronland, 12. — 1888. Uber Grünlands Vegetation. Englers Jahrb. 10. — 1890. Om Caryophy!laceernes Blomster. Botanisk Forenings Festskrift. Kjøbenhavn. , — 1890. Botan. Exkursioner. 1. Fra Vesterhavskystens Marsk- egne. Vidensk. Meddel. fra d. naturhist. Foren. Kobenhavr. — 1891. Botan. Exkursioner. 2. De psammophile Formationer i Danmark. Vidensk. Meddel. fra Naturh. Foren. Kjøbenhavn. — 1897. Halofytstudier. Danske Videnskab. Selsk. Skrifter, GE. S: — 1906. Dansk Plantevækst. 1. Strandvegetation. Kjøbenhavn. — 1918. Om Jordudlobere. (Underground runners). K. Danske Vidensk. Selsk. Skrifter, 43 figg. 8.R., II, 6. VESQUE, Jul., 1883. Contributions à l’Histologie systématique de la feuille des Caryophyllinées. Ann. Sc. nat. (6.8.), 15. Wittrock, У. B., 1908. Om... Stellaria media. Vetenskapsakade- miens Ärsbok. Stockholm. Wurrr, Тновию, 1902. Botanische Beobachtungen aus Spitzbergen. Lund. 1. De arktiska växters transpiration. 14. сеет а ее. | = | Morten Р. Porsild. 1920; f the order Liliales several species, belonging to differ- ent families, advance to the arctic limit of forest occasionally even crossing it, not attaining, however, a wider distribution in the arctic territory where they do not belong. We find also that they in the alpine regions do not, or only occasionally, cross the limit of forest, cf. remarks in SCHROE- TER: Pflanzenleben р. 354—5. The principal of those species are the following: 1) Veratrum album L. and the closely related У. viride Ait. extend respectively in the river-valleys of Northern Eurasia and Northern America to the coasts of the Arctic sea. 2) Streptopus amplexifolius (L.) D.C. is an Е. Asiatic —N. American forest plant; the fruit is a berry, and it has probably, through the agency of birds, been brought to the subarctic copses of the southernmost part of Greenland. 3) Lloydia serotina (L.) Sweet — closely related to Gagea — advances on both sides of Bering’s strait farther into arctic territory than any of the others. Thus it is found at Cape Lisburn and extends even as far as the purely arctic New Siberian Islands. 4) Allium schoenoprasum L. (including A. sibirieum Г.) is found in the island of Kolgujew, in the river-valleys of N. Asia up to the Arctic sea, and on the north coast of Alaska. 346 Morten P. Porsitp. 5) Iris sibirica L. (and related species?) extend to the coast on both sides of Bering’s strait. As to the structure and biology of the majority of the above-mentioned species the reader is referred to KIRCHNER, Low & SCHROETER’S: “Lebensgeschichte” where — on Alpine material — these questions have recently been treated. Widely distributed in the arctic territory are only the two following species of the genus Tofieldia: 1) T. palustris Hups., circumpolar, common from the subarctic territory far into the arctic, probably without reaching the purely high-arctic regions, however. 2) T. coccinea Ricu., N. Asiatic, N. American, found besides in the northernmost two-thirds of Greenland, every- ` where rare or perhaps overlooked; numerous of its Asiatic occurrences are only known from previous erroneous deter- minations, corrected by OsTENFELD (Fl. Arctica р. 32). A third species: Т. calyculata WAHLENB. is distributed throughout nearly all the mountain regions of Central Europe, whence it often descends into the lowland, whilst it only in a few places crosses the Alpine limit of forest: the variety glacialis REICHENB. The structure and biology of 7. calyculata is extremely well-known, and the same applies partly to 7. palustris. An account of this, based upon Alpine material, is found in Loew & KırcHner’s: “Lebensgeschichte” р. 229 ff., a copious literature being quoted here, too. Cp. also the description of the closely related Narthecium ossifragum Hups. ibid. p. 244 Й., and RAUNKIÆR р. 138 ff. The following observations, Liliales. 347 having all been carried out upon arctic, living and alcohol material, chiefly from Greenland, will therefore serve mostly as a complement, especially with regard to the least known of the species: T. coccinea. Occurrence in nature. Tofieldia palustris and T. coc- cinea grow on moist spots among heath- and bog vegetation, more rarely on fresh plantless moraines. They occur most frequently here as firm cushions or cakes which are held together by the vigorous secondary roots (Fig. 1, А). They attain their highest development when growing among bog- mosses. Here the leaves become longest, and the floral shoot tallest, but on the other hand, the stalks are able to creep better and more widely, and consequently we do not get here as large and firm cakes as otherwise. The habitats are covered by snow during winter. The shoots are short-jointed, flat, frequently erect, more rarely, through want of space, obliquely ascending and, in vigorous specimens, each shoot bears from 6—8 leaves. During several years the shoot remains on a purely vegetative stage until finally ending in an inflorescence. Some of them remain on the vegetative stage. Structure of shoot ср; Fig, A—D. The leaves are all uniform, apart from a slight difference in length, the lowermost being the smallest, evergreen, two- rowed ensiform, equitant with a long sheath. Below the green leaves some withered ones (often black-spotted by fungi) are found, and at the base setaceous, vascular strands are seated originating from older marcescent leaves. Scale-leaves are not found. The lateral shoots are formed in the uppermost axils of the leaves, normally only one or two are developed. 348 Мовтем P. Porsitp. Fig. 1. А, В, С. Structure of shoot in Т. palustris (Finse, Norway). A. Habit drawing of two flowering shoots which only cohere by the aid of the roots. b, the scape with its sheath-formed bract. В. The main-shoot, I, has two withered leaves (1—2) and six fresh (3—8). In the axil of 2 a shoot: II, is to be seen, and in the axil of 3 another one II, of which 4 leaves are visible. Magnified, detailed drawing of it in the following (С). С. Shoot with its subtending leaf: f, and with 3 leaves marked 1, 2, 3, 4. 1 and 2 turn their back to the mother-axis. D. Т. coccinea. Foliage shoot with withered remains of leaves, from unfavourable habitat. (Dove Bay, E. Greenland; leg. A. Lundager, 27. 6. 1908). Е. Germination in T. calyculata. (Drawn by E.W.) 349 Liliales. In T. coccinea there is as a rule 9(—11) vascular bundles (Fig. 2), in T. palustris 11(—13). In the first-mentioned the leaves are often more flaccid (Fig. 1, D), frequently partly decumbent, especially in exposed places, whilst they, in the latter, always are rigidly erect. The structure of the leaf is in all essentials identical in the two species. The sheath is constructed like а normal leaf, above this the leaf is isolateral, has an equal right and left side, an upper and lower margin (Fig. 3). The epider- mis Consists, except on the upper (inner-) side of the sheath, of ra- ther small cells, the walls of which are slightly undulating, highly thickened, porous and covered by a well-developed cuticle which also spreads to the guard-cells of the stomata (Fig. 3 B, D). The Nordre Strömfjord, Greenland. T. coccinea. Fig. 2. stomata are arranged longitu- dinally, their slit is quite diminu- the beneath them is quite small. The tive, intercellular cavity base of the sheath consists of hyaline parenchyma; alittle higher up, on the morphological lower side, we find the chlorophyll a—e show sections in various height through a _ foliage leaf. The xylem of the vas- cular bundles is dark. The dotted line shows the bound- ary between the chlorophyll tissue and the aqueous tis- sue; a, showing the basal part of the sheath, has no chlorophyll. å Bec (Drawn by M. P. P.) tissue consisting of several lay- ers of cells, whilst the upper side is constantly non-chlorophyl- lose. Above the sheath the chlorophyll tissueis homogeneously developed on both sides, the non-chlorophyllose part occupy- ing the centre of the leaf, extending right to the tip of the leaf (Fig. 2, 3, A, В). In contradistinction to Narthecium (vide RAUNKIÆR) no air-chambers are found in the chloro- " phyll layer on transverse sections. XXXVII. 23 350 Morten P. PorsiLp. The cells in the chlorophyll tissue are parallelepipedic and frequently no intercellular spaces are seen in transverse sections; in longitudinal sections, however, quite small ones are to be seen. The cells of the chlorophyll tissue contain chlorophyll grains filled with starch. In the isolateral part the vascular bundles approach by twos to each other, and the uppermost one is entirely connated of two (Fig. 2e, ЗА). In transverse sections the large, crescent-shaped coverings of bast are to be seen. DAC AE Tee N ААА» TF Fig. 3. Anatomy of leaf. А, В, С. T. palustris; from Jakobshavn, W. Greenland. A. Transverse section of a foliage leaf above the equitant part. The hatched am is the chlorophyll tissue; the epidermis and the central aqueous tissue are light. 3 B. A part of the same section, more highly magnified; ep. Epidermis with thick cuticle and a stoma with very minute air cavity. Chloro- phyll tissue almost without any intercellular spaces, the cells filled . with starch; m aqueous tissue. C. Epidermis with two stomata, longitudinally placed. The walls of the cells are highly thickened, slightly undulating. D. Transverse section of a leaf of T.coccinea (Danmarks ©, Е. Green- land leg. N. Hartz). (Drawn by E. W.) The scape of inflorescence is as a rule leaf-less, yet in vigorous specimens 1—2 leaves may be found, the upper- most one at least having the character of a bract (Fig. 1). In Г. coccinea 2—3 stem-leaves are nearly always present (Fig. 4 A), the uppermost one having, in this case, less of the character of a bract. as Liliales. 331 The inflorescence in T. palustris is nearly always a globular head (Fig. 1 A), more rarely a subcylindric spike. In T. coccinea the elongated spike is the normal form, the globular the rarer one (Fig. 4 A, B), but it varies according to the quality of the habitat. In 7. coccinea, moreover, Fig. 4. T.coccinea. Nordre Strömfjord, Greenland. Habit-drawing of a weakly flowering specimen growingin Sphagnum. Inflorescence of a more richly flowering specimen. Single flower, shortly before the hermaphrodite stage. E, F. Almost ripe fruits in various positions; D and E the same specimen; note the asymmetry. Transverse section of the ovary. Uppermost part of the inflorescence, showing the downwards- directed carpotropie curvature, characteristic of the species. (Drawn by Thorbjern Porsild). Sop. mm single isolated flowers are frequently to be found below the spike, just asin 7. calyculata and in other species of the genus. 23% 352 Morten P. Porsitp. The flowering season is the middle of the summer, consequently the species do not belong to the earliest flowering. In an average year they will, on the latitude of Disko, be out from the end of July, and then flowering specimens are to be met with all through the rest of the sum- mer, because the floral shoots from the same tuft are not all of them developed simultaneously. Under each flower a small 3-dentated bractlet is present which often envelops the peduncle. In 7. palustris Fig. 5. T. palustris (Norway.) Part of inflorescence; note the small, 3-dentated bractlet. Flower at the beginning of the hermaphrodite stage: the stamens arise from the spoon-formed perigonial leaves. Stamen in a perigonal leaf when the flower bursts. Anther with open valves. Carpels in the stage of pollination. A single carpel, more highly magnified, showing pollen-grains on the stigma. (Drawn by E.W.) SI aan it is always shorter than the peduncle (Fig. 5 A), in T. cocci- nea just as long as this or longer (ABROMEIT). In the latter a fairly well-developed “calyculus” is as a rule present, a bi-symmetrical whorl of 3 small leaf-formations of which one is turned outwards, the other two transversal (Fig. 6). The morphological interpretation is uncertain. In 7. palu- stris it is normally absent, but may also, according to ABROMEIT, occur occasionally. Liliales. 858 Diagram. Етснгев’$ diagram of T. calyculata shows, besides the zygomorphy in the calyculus, a slight irregularity in the mutual size of the carpels. This also applies to the arctic species, especially to Т. coccinea (Fig. 6), where it appears to be even more strongly developed than in EIcHLER’S diagram. This difference is undoubtedly connected with the carpotropic movements (see below). In 7. palustris the perianth is purely white, or with a greenish rather than a yellowish tinge, in 7. coccinea it is purely white on the inner side, deep purplish on the exterior, especially along the median line of the leaves. Also the upper part of the scape and the carpels are of а beautiful purple-red colour. This bi-colouring is in live specimens very conspicuous, and when the species grow together they are easily Bo distinguished from each other by the colour Floral diagram of T. coccinca. Note the red pigment often disappears entirely, the asymmetry in the carpets. c (Drawn hence the numerous confoundings and erro- py М.Р.Р.) of the flowers alone. But in herbariums the flowers in both species becoming yellowish, neous determinations. The biology of the flowers (Fig. 5 B—E, 4 С) has been studied by H. Mürter in the Alps on T. calyculata which he found proterogynous, while 7. palusiris was almost homogamous. In spite of a greater abundance of honey- secretion in the latter he found a greater number of insects visiting the former. Greenlandic flowers of 7. palustris and Т. coccinea were almost homogamous with a slight indication of proterogyny. Visiting insects I have never seen there. In Т. palustris MÜLLER draws the stamens as freely projecting in the flower, in Greenlandic living material the filaments were curved downwards into the cavity of the perigonial 354 Morten P. Porsitp. leaf (Fig. 5 B, b, 4 C), this was especially prominent in the case of T. coccinea. During calm weather the cavity of the peri- _ gonial leaf is filled with pollen. When the flower has been open for some time the filaments, as well as the perigonial leaves, start curving upwards, and then the introrse anthera cannot avoid touching the stigmas. Thus cross-pollination is possible, but self-pollination the rule. Fig. 7. T. palustris (Disko, Greenland). A. Hibernated inflorescence with upwards-directed capsules, cha- racteristic of the species. B. Single, complete fruit of the same. С. Almost ripe inflorescence gathered late in autumn. D. Fruit, almost ripe, but not yet open. (The perigonial leaves removed). Note the almost perfect symmetry, as against Т. coccinea. Е. Hibernated capsule with walls partly fallen off. typical for the species. Some seeds are still seated on the placentas. F and С. A “normally” opened fruit (a rare case in the species). #. in moist condition. G. dry, seen from above. H. Seeds. (Drawn by T. P.) After the p ollination the inflorescence, which hitherto has been slightly nodding, straightens itself up, and the Liliales. 355 зсаре becomes rigid. The head or spike is stretched а little (Fig. ТА). The perigone withers, but persists for a long time round the carpels (Fig. 7 С). In Т. palustris the capsule is ovate, often a little brownish in colour and bent upwards (as in T. calyculata) (Fig. 7 À, С), whilst it in Т. coccinea is shorter, subglobose and bent downwards (Fig. 4 H). The lowest fruits in the inflorescence bend right downwards parallel with the scape, the uppermost obliquely downwards or the top one only horizontally. This carpotropie movement begins immediately and is continued during the ripening of the fruit. The ripening of the fruit is very late. Even on Disko, that lies far to the south of the Polar limit of the species, it is very difficult, even after favourable summers to find ripe fruits in the autumn, and if the unripe fruits are brought home the ripening is not continued. Exceptionally I have, however, found single, quite ripe fruits of T. palustris late in the autumn; the dehiscing is septicidal, forming a slight open- ing so that the seeds might be shaken out (Fig. 7G). I have had no chance of seeing 7. coccinea late in autumn. The majority of fruits in 7. palustris, and probably also in the other, do not thus attain to ripen their fruits before the snow comes. Consequently the ripening must take place under the snow. I have looked through a large material of hibern- ated fruits of both species and only succeeded in finding a single “normally” opened fruit. Instead, the very thin cap- sule-walls had gone to pieces, and the seeds had got out through the holes which had come into existence in that way. Often the capsules may be quite “skeletonized” so that only the backs of the carpels are left (Fig. 7 Е). SERNANDER (,,Spridningsbiologi’”’ pag. 354) has gathered T. palustris with seeds in the capsules in spring at the time when the snow was beginning to melt, and he therefore includes 356 Morten P. Ровзи. them among the winter-standers. То а certain extent this is right, as it does not attain to ripen its fruits till the snow comes, and consequently may happen to spread its seeds on the snow. But the white, paper-thin capsules of Tofielda are to a less extent adapted to this than for instance, the Luzula species, the brown capsules of which are saturated with a sub- stance which makes them proof against the various attacks of the first winter and preserves their hygroscopicity. What brings about the skeletonizing of the Tofieldia capsules, I do not know. If it had been effected by mechanical wear of drifting snow in the spring-time there would always be found some which had been lying in sheltered places and remained intact. The apertures are irregular, as if the thin parchment-like capsular membranes had, through iterated freezing or exsiccation, become so fragile that even the slight- est touch was enough to cause the apertures. The production of seeds is no doubt abundant in both species, judging by the number of well-developed ovules and by the occurrence of the species in nature. Vegetative propagation is practically excluded, because even the strong- est stream issuing from melting snow would hardly be able to break asunder the cakes which are made up of the strong roots felted together. Only in loose growing mosses a libera- tion of the lateral shoots takes place when the rhizome dies away at the back. The seeds (Fig. 7 H) in T. palustris are small, yellow, a little triangular and slightly curved, with few wrinkles faintly marked on the testa. They weigh 0.03 mg. In 7. coccinea, of which I have gathered but a few seeds in hibern- ated capsules, they were of about the same size and appear- ance. Germination. The first stages have been seen by KLEBs (quoted in “Lebensgeschichte’”’) and later on by WAR- Liliales. 357 MING in T. calyculata (Fig. 1 Е). After having emerged from the testa the cotyledon is geniculate bent, and at a very early stage a whorl of roothairs is produced at the base of the young root. The later stages are not known, but they might agree with those in Narthecium described by BucHEnAu. Fig. 8. A. Seedling of T. palustris at the end of the second period of vegetation. B. 3—4 years old seedling of the same species. Both from Disko, Greenland. С. Retarded inflorescence, extricated from an almost withered spec- imen of T. coccinea; g. small vegetative lateral shoot. Nordre Strömfjord, Greenland. (Drawn by Th. P.) Seedlings (Fig. 8) are easily found in nature, but it is very difficult to find the very first stages because they are so small. The earliest formed foliage leaves wither rapidly, and the young plants pass through a period of strengthening, which lasts for many years, before they attain to flowering. BIBLIOGRAPHY. ABROMEIT, J.: Botanische Ergebnisse der von der Gesellschaft f. Erd- kunde .... ausgesandten Grönlandsexpedition. Phanerogamen Bibliotheca Botanica 42. 1899. Low, Е. und KircHNer, O.: Tofieldia in KiRCHNER, LOEW u. ScHROETER: Lebensgeschichte der Blütenpflanzen Mitteleuropas Lief. 14. Stuttgart, 1911. In this work an extensive bib- liography is found. MÜLLER, H.: Alpenblumen, ihre Befruchtung durch Insekten und ihre Anpassungen an dieselben. Leipzig 1882. RAUNKIÆR, C.: De danske Blomsterplanters Naturhistorie I. Kjoben- havn 1895—99. SCHROETER, C.: Das Pflanzenleben der Alpen. Zürich 1908. SERNANDER, R.: Den skandinaviska florans spridningsbiologi. Upsala 1901. Besides, I have made freely use of notes and drawings by Pro- fessor В. WARMING. ag By Fr. J. Mathiesen. 1921. Preface. a the present paper, partly on the basis of notes on the subject found in previous publications, and partly on the basis of my own investigations, I have given a descrip- tion of the ramification, shoot-structure, flower-biology and anatomy of some Arctic and Subarctic Scrophulariaceae; in addition to which, under each species, the geographical distribution has been briefly mentioned, and, as far as the literature on the subject rendered it possible, the nature of the habitat. Besides the herbarium and alcohol material belonging to the Botanical Museum of the University of Copenhagen, I have, through the courtesy of Mrs. THEKLA REsvoLL, Dr. phil., with respect to several of the species, had an opportunity of investigating some excellently preserved material, collected on the mountains of Norway. The subject-matter of the present paper was worked out in the Plant-anatomical Laboratory of the University of Copenhagen, and I wish to express my grateful and heart- felt thanks to the Director, Professor С. RAUNKIER, for his advice and help during the work. | The majority of the figures illustrating the structure of the flowers, are drawn by Professor WARMING, to whom also my most sincere thanks are due, partly on account of the interest he has taken in my work, and partly because he so 362 Ев. J. MATHIESEN. generously placed his notes on the Arctic Scrophulariaceae at my disposal. Fig. 20, B, C and D, Fig. 40 and Fig. 44, which are published here for the first time, have been re- drawn after Professor WARMING's originals, in order to be reproduced according to the method employed here. The remainder of the figures are drawn by myself. Mr. M. PorsıLp, mag. sc., Director of the Danish Arctic Station, Greenland, has very kindly read through great parts of my manuscript, and made several additions of a biological nature, besides having also kindly permitted me to use his notes on the geographical distribution of Arctic plants. For the help thus rendered to me Г herewith beg him to accept my best thanks. The following species have been investigated: — Veronica fruticans Crantz. — alpına L. — _ officinalis L. f. glabrata Fristedt. Castilleia pallida (L.) Kunth. Euphrasia arctica Lange. Bartschia alpına L. Pedicularis lapponica L. — sudetica Willd. — euphrasioides Steph. = Sceptrum carolinum L. — capitata Adams. — hirsuta L. — lanata (Willd.) Cham. & Schlecht. — flammea L. — Oederi Vahl. Scrophulariaceae. 363 Principal literature. ABROMEIT, J., 1899: Bot. Ergebnisse der... unter Leitung von Dr. von Drygalski ausgesandten Grünlandsexpedition. B. Samen- pflanzen (Phanerogamen). (Bibliotheca botanica. 42. Stuttgart). ÅNDERSSON, С. and Н. Hessetman, 1900: Bidrag till kännedomen om Spetsbergens och Beeren Eilands kärlväxtflora, grundade ра iakttagelser under 1898 ars svenska polarexpedition. (Bihang till K. Svenska Vet.-Akad. handlingar. Bd. 26. Afd. III. No. 1). AURIVILLIUS, C., 1883: Insektlifvet. 1 arktiska länder. (Nordenskiöld: Studier och forskningar föranledda af mina resor i höga norden. VI. Stockholm). AXELL, S., 1869: Om anordningarna för de fanerogama växternas befruktning. Stockholm. Втутт, A., 1906: Håndbog i Norges Йога. Kristiania. Ввомогм, J: А. Z., 1898: Bidrag till kännedomen om de svenska fanerogama örternas skottutveckling och öfvervintring. Dis- sertation. Upsala. Borcesen, F., 1895: Bidrag til Kundskaben om arktiske Planters Bladbygning. (Botanisk Tidsskrift. Kjøbenhavn. Bd. 19. — Resumé français: Journal de botanique. Vol. IX. 1895). CLEVE, Astrip, 1901: Zum Pflanzenleben in Nordschwedischen Hoch- gebirgen — Einige ökologische und phänologische Beiträge. (Bihang till К. Svenska Vet.-Akad. handlingar. Bd. 26. Afd. ME No..15). Dusén, P., 1901: Zur Kenntnis der Gefässpflanzen Ostgrönlands. (Bihang till К. Svenska Vet.-Akad. handlingar. Bd. 27. Afd. ТЫ No>3). Easrwoop, ALICE, 1902: A Descriptive List of the Plants collected by Dr. Е. Е. Blaisdell at Nome City, Alaska. (Botanical Gazette. Vol. 33). Exstam, О., 1897: Einige blütenbiologische Beobachtungen auf Novaja Semlja. (Tromso Museums Ärshefter. 18. 1895). — , 1899: Einige blütenbiologische Beobachtungen auf Spitz- bergen. (Tromso Museums Ärshefter. 20. 1897). FREIDENFELT, 1904: Der anatomische Bau der Wurzel in seinem Zu- sammenhange mit dem Wassergehalte des Bodens. (Bibliotheca botanica. 61. Stuttgart). GoEBEL, K., 1897: Morphologische und biologische Bemerkungen. 7. Ueber die biologische Bedeutung der Blatthöhlen bei Tozzia und Lathraea. (Flora. Bd. 83). 364 Ев. J. МАТНТЕЗЕМ. Groom, Percy, 1897: On the Leaves of Lathraea squamaria and of some allied Scrophulariaceae. (Annals of Botany. Vol. XI). HABERLANDT, G., 1897: Zur Kenntniss der Hydatoden. (Pringsheims Jahrbücher. Bd. 30). Hartz, N., 1894: Botanisk Rejseberetning fra Vest-Gronland 1889— 1890. (Medd. om Gronland. 15. Hefte. Kjbh. 1898). — , 1895 (I): Ostgronlands Vegetationsforhold. (Ibidem. 18. Hefte. Kjbh. 1896). — , 1895 (II): Fanerogamer og Karkryptogamer fra Nordost- Grønland, с. 75°—70° N. Br., og Angmagsalik, с. 65°40’ N. Br. (Ibidem. 18. Hefte. Kjbh. 1896). — and Cur. KRUUSE, 1911: The Vegetation of Northeast Green- land, 69° 25’ lat. n.—75° lat. п. (Ibidem. 30. Hefte. Kjbh. 1911). HEINRICHER, H., 1898—1910: Die grünen Halbschmarotzer. I—VI. (Pringsheims Jahrbücher. Bd. 31, 32, 36, 37, 46, 47). HozLsrTein, О., 1907: Beiträge zur vergleichenden Anatomie der Stengel und Rhizome von dicotylen Alpenpflanzen. Dissertation. Gött- ingen. Ногм, Тн., 1885: Novaia-Zemlia’s Vegetation, særligt dens Phane- rogamer. (Dijmphna-Togtets zoologisk-botaniske Udbytte. Kjbh. 1885). HovELAcQuE, M., 1888: Recherches sur l’appareil végétatif des Big- noniacées, Rhinanthacées, Orobanchées et Utriculariées. Paris. HucHEDÉ, 1907: Veronique et Gratiole. These. Paris. JUNGNER, J. R., 1894: Klima und Blatt in der Regio alpina. (Flora. Bd. 79). JORGENSEN, E., 1919: Die Euphrasia-Arten Norwegens. (Bergens Museums Arbok. 1916—17. Naturvidenskab. række. No. 5). Kerner v. MarıLaun, А. 1898—1900: Pflanzenleben. Bd. I—II. Leipzig u. Wien. KJELLMAN, Е. R., 1882 (I): Sibiriska nordkustens fanerogamflora. (Vega-expeditionens vetenskapliga iakttagelser. Bd. I. Stock- holm). — , 1882 (П): Fanerogamfloran pa Novaja-Semlja och Wajgatsch. (Ibidem). — , 1882 (III): Om tschuktschernas hushällsväxter. (Ibidem). — , 1882 (IV): Asiatiska Beringssunds-kustens fanerogamflora. (Ibidem). — , 1883: Ur polarväxternas Ш. (Nordenskiöld: Studier och forskningar föranledda af mina resor i höga norden. VII. Stock- holm). — andA. N. Lunpstrém, 1882: Fanerogamer fran Novaja-Semlja, Wajgatsch och Chabarova. (Vega-expeditionens vetenskapliga iakttagelser. Bd. I. Stockholm). Кмотн, P., 1899: Handbuch der Blütenbiologie. Bd. II. Teil 2: Lobe- liaceae bis Gnetaceae. Leipzig. Scrophulariaceae. 365 Косн, E., 1895: Über die systematische Bedeutung der anatomischen Charaktere der Scrophulariaceen. Dissertation. Erlangen. Kruuse, Снв., 1898: Vegetationen i Egedesminde Skjergard. Medd. (om Gronland. 14. Hefte. Kjbh. 1898). — , 1905: List of the Phanerogams and Vascular Cryptogams found on the coast, 75°—66°20’ lat. N. of East Greenland. (Ibidem. 30. Hefte. Kjbh. 1911). — , 1906: List of Phanerogams and Vascular Cryptogams found in the Angmagsalik District on the East coast of Greenland between 65°30’ and 66°20’ lat. N. (Ibidem). — ,1911: Rejser og botaniske Undersøgelser i Ost-Gronland mellem 65°30’ og 67°20’ i Årene 1898—1902 samt Angmagsalik-Egnens Vegetation. (Medd. om Gronland. Bd. XLIX. Kjbh. 1912). LANGE, J., 1871: Bemærkninger om froenes form og skulptur hos beslægtede arter i forskellige slægter. (Botanisk Tidsskrift. Kjbh. Bd. 4). — , 1880: Conspectus Florae Groenlandicae. (Medd. om Gron- land. 3. Hefte. Kjbh. 1880). — , 1887: Tillæg til Fanerogamerne og Karsporeplanterne i Consp. Flor. Groenl. (Ibidem. 3. Hefte. Fortsættelse. Kjbh. 1887). Leıst, 1889: Ueber den Einfluss des alpinen Standortes auf die Aus- bildung der Laubblätter. (Mitth. der naturforsch. Gesellschaft von Bern). Linpman, С. A. M., 1887: Bidrag till kannedomen om skandinaviska fjellväxternas blomning och befruktning. (Bihang till K. Sven- ska Vet.-Akad. handlingar. Bd. 12. Afd. III. No. 6). Lunpacer, A., 1912: Some Notes concerning the Vegetation of Ger- mania Land, North-East Greenland. (Medd. om Grønland. Bd. XLIII. Kjbh. 1917). MÜLLER, H., 1881: Alpenblumen, ihre Befruchtung durch Insekten und ihre Anpassungen an dieselben. Leipzig. NATHORST, A. G., 1883: Nya bidrag till kännedomen om Spetsbergens kärlväxter och dess växtgeografiska förhällanden. (K. Svenska Vet.-Akad. handlingar. Bd. 20. No. 6). Norman, J. M., 1895: Norges arktiske Flora. II. Kristiania. ÖSTENFELD, С. H., 1901: ‘‘Phanerogamae and Pteridophyta” in “Botany of the Ferées’. Vol. I, Kjbh. — , 1908: The Land-Vegetation of the Færåes. (Ibidem. Vol. III). — , 1915: Plants collected during the First Thule Expedition to Northernmost Greenland. (Medd. om Gronland. Bd. LI. Kjbh. 1915). — and A. Lunpacer, 1910: List of Vascular Plants from North- East Greenland (N. of 76°N. lat.), collected by the Danmark-Ex- pedition 1906—1908. (Medd. om Grønland. Bd. XLIII. Kjbh. 1917). SKK VIL. 24 366 Ев. J. MATHIESEN. Porpius, В. R., 1903: Blombiologiska iakttagelser. (Acta soc. pro fauna et Йога fennica. 25). Porsitp, M. P., 1902: Bidrag til en Skildring af Vegetationen pa Øen Disco etc. (Medd. om Grenland. 25. Hefte. Kjbh. 1902). — , 1910: The Plant-Life of Hare-Island off the coast of West- Greenland. (Ibidem. Bd. XLVII. Kjbh. 1911). — , 1912: Vascular Plants of West-Greenland between 71° and 73° N. lat. (Ibidem. Bd. Г. Kjbh. 1912). — ,„assisted by A. Erring Porsitp, 1920: The Flora of Disco Island and the adjacent coast of West-Greenland from 66°— 71° N. lat. etc. (Ibidem. Bd. LVIII. Kjbh.). RAUNKIÆR, C., 1907: Planterigets Livsformer og deres Betydning for Geografien. Kjbh. Resvorr, THEKLA R., 1917: Om planter som passer til kort og kold sommer. (Archiv for mathematik og naturvidenskab. Bd. XX XV. Kristiania). RosEenvincE, L. KoLpErup, 1892: Andet Tillæg til Grønlands Fanero- gamer og Karsporeplanter. (Medd. om Gronland. 3. Hefte. Fortsettelse. Kjbh. 1887). — , 1896 (I): Nye Bidrag til Vest-Gronlands Flora. (Ibidem. 15. Hefte. Kjbh. 1898). — , 1896 (II): Det sydligste Grønlands Vegetation. (Ibidem 15. Hefte. Kjbh. 1898). ScHERFFEL, A., 1888: Die Drüsen in den Höhlen der Rhizomschuppen von Lathraea squamaria L. (Mitth. a. 4. Bot. Inst. zu Graz. У. 1888. Jena). SCHRÔTER, C., 1908: Das Pflanzenleben der Alpen. Zürich. Siren, F., 1905: Blombiologiska iakttagelser i Kittila Lappmark. (Medd. af soc. pro fauna et flora fennica. 31). Simmons, H. G., 1906: The Vascular Plants in the Flora of Elles- mereland. (Report of the second Norwegian Arctic Expedition in the “Fram’’, 1898—1902. No. 2. Kristiania). — , 1913: A Survey of the Phytogeography of the Arctic American Archipelago, etc. (Lunds Universitets Arsskrift. N.F. Afd. 2. Ва. 9. №. 19). SKOTTSBERG, C., 1901: Einige blütenbiologische Beobachtungen im ark. Teil von schwedisch Lappland 1900. (Bihang till K. Svenska Vet.-Akad. handlingar. Bd. 27. Afd. III. №. 2). SYLVÉN, N., 1906: Om de svenska dicotyledonernas första förstärk- ningsstadium etc. I—II. (К. Svenska Vet.-Akad. handlingar. Bd. 40. No. 2). WAGNER, A., 1892: Zur Kenntniss des Blattbaues der Alpenpflanzen und dessen biologischer Bedeutung. (Sitzungsber. der kaiserl. Akad. der Wissenschaften in Wien. Matem.-naturw. Classe. Bd. CI. АБТ. Scrophulariaceae. 367 WARMING, Euc., 1884: Om Skudbygning, Overvintring og Foryngelse. (Naturhistorisk Forenings Festskrift. Kjbh.). — ,:1886: Om Bygningen og den formodede Bestovningsmäde af nogle grønlandske Blomster. (Oversigt о. d. Kgl. Danske Vid. Selsk. Forhandl. Kjbh.). — , 1888: Om Grønlands Vegetation. (Medd. om Grønland. 12. Hefte. Kjbh. 1888). — , 1890: Biologiske Optegnelser om grønlandske Planter. 3. Scro- phulariaceae. (Botanisk Tidsskrift. Kjbh. Bd. 17). R. v. WETTSTEIN, R., 1896: Monographie der Gattung Euphrasia. Leipzig. VOLKART, 1899: Untersuchungen über den Parasitismus der Pedicu- laris-Arten. Dissertation. Zürich. 24* 368 Ев. J. MATHIESEN. Veronica fruticans Crantz (V.saxatilis Scop.). Alcohol-material from Greenland (Julianehaab, leg. Lür- ZEN, 3. 7. 1887; Præstefjældet, leg. Е. WARMING, 2.8. and 6. 8. 1884) and Northern Norway (Käfjord, leg. Е. WARMING, 15. 7. 1885). — Herbarium-material from Greenland, Iceland, the Faeröes and Fennoscandia. Lit.: Axeır, 1869, р. 102; Lance, 1880, р. 73; 1887, р. 201; Мбттев, 1881, р. 267; WARMING, 1888, pp. 35, 75 and 87; 1890, р. 203; RosENVINGE, 1892, р. 685; 1896 (II), pp. 128, 144, 161; WAGNER, 1892, рр. 8 and 21; Hartz, 1894, pp. 9 and 57; 1895 (1), рр. 273, 289; 1895 (II), р. 335; Hartz and Квоозе, 1911, pp. 352, 357, 359 and 409; Косн, 1895, р. 117 et seq. and р. 127; Norman, 1895, р. 452; "KNUTH, 1899, р. 167; Porsırp, 1902, р. 230; 1920, р. 141; Hucnené; 1907, р. 72, fig. ГУ, В* fig. У, CE and Ge ae Syıvfn, 1906, р. 80; Kruuse, 1905, р. 175; 1906, р. 248; 1911, in part IV pp. 202, 207, 243, 261 and 262, besides many notes in the preceding parts; SCHRÖTER, 1908, рр. 221 and 656. An evergreen, nanophyllous!, sympodial under-shrub- chamæphyte which has a main root that dies away early (according to SYLVÉN), and develops adventitious roots fairly abundantly; the winter-buds are raised at most a few cm above the surface of the ground; older, vigorous specimens form small tufts (Kruuse, 1906, р. 248). The shoot-development extends over two years; during the first year a horizontal or obliquely-ascending part is de- veloped, only a few cm long and furnished with a few (3—7) pairs of foliage-leaves; the next year the shoot 1 C. RAUNKIÆR: Om Bladstorrelsens Anvendelse i den biologiske Plantegeografi (Bot. Tidsskrift, Kjbh., Bd. 34, 1916). Scrophulariaceae. 369 usually continues its growth in a vertical direction, larger leaves are developed, and the growth can be terminated by a few-flowered raceme without a terminal flower. Special bud-scales do not occur; during the winter-rest the point of the axis is protected by the uppermost, not yet expanded pair of foliage-leaves, which are in con- tact with each other by their hairy margins. After the fruit has ripened, the stem dies as far down as to some- what above the boundary line be- tween the Ist and 2nd year’s growth; the perennial basal portions bear the innovation shoots. Principal buds proper do not occur, but the uppermost buds appear generally to be the most advanced, and it is evidently especially these buds which produce the flower-bearing axes, while the lower ones often produce only small, weak, few- leaved shoots which — as recorded by Wanne (1890, p. 205) and de as also shown in Fig.l — may 2.8.1884). (About nat.’size.) be somewhat runner-like and fur- | nished with only a few small leaves. The specimen illustrated in Fig. 1 is rather scantily branched; each of the floral shoots has only two real “innovation-buds”, and of their parent- shoots the one to the left has also had two, of which one has developed into a vegetative shoot; the parent-shoot to the right is somewhat more richly equipped: in addition to the lowermost quite small shoot it bears two opposite floral shoots, and in the axils of the next pair of leaves two more shoots, of which the one (cut-off) was floral. Such a difference 370 Fr. J. MATHIESEN. between shoots from two opposite leaves, that the one becomes vigorous and floral, and the other vegetative only, is very common; it appears to be the rule that the shoot which turns outwards towards the periphery of the tuft, becomes more vigorous. It happens not rarely, however, that the short, vegetative shoots are later on instrumental in the formation of floral shoots, in that the latter arise as lateral shoots upon them; in such a case the part of the parent- shoot above these lateral shoots dies away, as is the case in the floral shoot. The development of the lateral shoots usually takes place in the second year of the parent-shoot, and consequently simultaneously with the flowering. The Leaves are rather thick, entire or slightly serrate; on the median rib and along the margin there is a sparse covering of non-glandular hairs; such are also found on the stem and in the floral region, viz., on the calyx and ovary. In the Botanical Garden in Copenhagen the leaves on the new innovation-shoots, developed during the last summer, remain green and fresh during the winter; on the flowering shoots, in my herbarium-material, the leaves from the previ- ous year were withered in some cases; most often, how- ever, they were green. As a rule, the basal portion of the shoots takes root abundantly, and is drawn down by the roots to the surface of the ground; in the individual illustrated in Fig. 1, which has a somewhat more erect growth than is generally the case, adventitious roots are only scantily produced; the roots are developed at the earliest in the second year of the shoots. The basal portions of the shoots have growth in thick- ness and may live several years; an individual from Greenland (Ikalik) had a prostrate stock, 4mm thick, which showed 16 annual rings. The Flower-biology has been investigated by H.Mür- Scrophulariaceae. >71 LER (the Alps) and Е. WARMING (Greenland and Northern Norway). The flowers are large and conspicuous; also in the Arctic regions, as a rule, they are of a pure and bright colour; but according to M. PorsiLp, in northern Greenland a white- flowering variety may commonly occur. The ultimate dia- meter of the corolla is very constant, being 10—12 mm in material from the different localities; I found it to be parti- cularly large in some material from Iceland, viz. 14mm (in the Alps, according to Müller, only 6—7 mm). “The corolla- tube and the throat is white, and furnished with a wreath of hairs; then follows a dark, brown or reddish ring and then the deep-blue limb. The corolla-tube is 11/,—2!/, mm in length (according to Müller 3 mm). As in the Alps the filaments taper very much at the base, are about 5 mm in length and, as the pistil is of the same length, the stigma and anthers stand at the same level. The flowers are homo- gamous. In fully expanded flowers self-pollination appears to be able to take place only with difficulty, as the stamens diverge so much laterally and the anthers are thereby re- moved from the stigma on the straight outstretched style; but self-pollination might have taken place at an earlier point of time, for I have seen a flower (from Greenland at 67° М. lat., Aug. 6) which, perhaps on account of gloomy and rainy weather, was only slightly expanded and in which the one anther was open, so that the pollen fell out of it down upon the stigma which stood close to it and which appeared to be fully ripe. The ovary is covered with small upwardly directed, adpressed hairs, and the style is 2—3 times longer than the ovary. Ripe fruit is produced in West Greenland at least up to 70° N. lat.” (Е. Warmine, 1890, p. 203). V. fruticans is homogamous also according to AXELL and H. Murer. The style, however, appears to me to be always slightly longer than the anthers; the outspread posi- 312 Ев. J. MATHIESEN. tion of the latter is distinctly seen in Fig. 5, С; the figure is drawn from living material from the Botanical Garden in Copenhagen. According to M. Ровзи, near its northern limit in Green- land it flowers late and even sets fruit; but it is only in favourable, dry autumns, without too much frost, that its seeds ripen. Unripe fruits live through the winter, but perish without developing further. The structure of the shoots ex- plains, as in V. alpina (TH. ResvorL), the late flowering. There was, however, ripe fruit to hand from all localities. KRUUSE (1906, р. 248) in the Angmagsalık-distriet notes “abundant ripe fruit”. Geographical Distribution according to LANGE: Greenland, the Ural Mountains, Lapland, Finmark, Norway, Iceland, Great Britain, the Alps and the Pyrenees. Besides this it is recorded from the Færûes. Its certainly known north limit in West Greenland is 70°17’ on the continent (Vaj- gattet), from 70°—67° it occurs in isolated specimens only, and not until further southwards is it common; consequently, in Greenland it belongs to the southern types (M. PorsiLp). In East Greenland it occurs as far north as the Scoresby- Sound-district (Hartz and Kruuse, 1911). According to WARMING. (1888) and RosENviNGE (1896 (II)) in West Greenland the species grows in willow-copses, birch-copses and on “herb-slopes”, and in places whence snow melts early it can ascend as high up in the moun- tains as 750 metres (LANGE, 1880: in grass-covered, open places, in rocky clefts). М. PorsıLp informs me in con- firmation of this that its natural habitat is the edge of the willow-copse, there it grows on warm, sunny slopes which are not too dry; the species thrives but badly when over- shadowed; it needs snow-covering throughout the winter but must be early freed from snow; at its north limit it Scrophulariaceae. Эт does not at any rate belong to the real snow-flora. In the Scoresby-Sound-district it is found ‘Чт particularly well sheltered luxuriant, humid, herby slopes with high snow- cover in winter,’ but here it is very rare (Kruuse, 1905, р. 175). In the Angmagsalik district, where it is somewhat more common (KRUUSE, 1906, p. 248), besides growing on the herby-slopes, it is also noted from the “steps of steep rocks above the slopes.’’ Ac- cording to Norman, in Northern Norway the species is found on all kinds of stony substrata rather than SS) ON on grassy ground, and в. ces DO) it very much prefers ЕЕ 7 OS SIE EERE IOS the sunny side, especi- RUE DO 0 ally the side facing СО СЕ SOMO! D SO directly south, and _ occurs only very rarely on the indifferent Fig. 2. Veronica fruticans. De, Transverse section of an adventitious root sides”. In the Alps (about ?3°/,). (Greenland) it is recorded from “Schneeblössen’’ (SCHRÖTER, р. 656), spots where the wind (eastern and western) sweeps the snow away in winter time. Anatomy. The Root. The epidermis of the adventi- tious roots dies away early; the outermost layer of the cortex is developed as an exodermis with cuticularised walls. Even before any secondary growth has taken place in the stele, a cork-cambium is developed in the layer under the exo- dermis, which is instrumental in the formation of a few- 374 Ев. J. MATHIESEN.| layered cork (cf. Fig. 2). The walls of the primary cortex attain a rather considerable thickness; this is especially the case as regards the endodermis; both in this and in the os q ne Ion Me dr D EU otre О. TORI onto iio (©) SD; ® 17 oo О 6 on G oe 025. OSSE 2 nee Fig. 3. Veronica fruticans. Transverse section of the stem (about 299%/,). (Greenland.) other layers, radial walls are formed during the secondary growth of the root. In the wood of the 2-year-old root illustrated in Fig. 2, an annual ring is formed outside the central part which was developed during the first year; the limit between them is ee ee ee Me u > Scrophulariaceae. 375 easily discernable, since they are separated by a zone in which the vessels and wood-fibres are intermixed with non- lignified, axially elongated parenchymatous elements. The Stem. In the lower persistent parts of the stems, a cork-cambium commences activity during their second year; it appears in the outermost layer of the cortex and forms a few-layered cork (Fig. 3). The cells of the primary cortex are rather thick-walled and show division by radial walls; they contain chlorophyll-grains. The endodermis has very distinct Casparian dots (as in the root). In the pericycie small groups of hard-bast cells are found. The figure shows a portion of the transverse section of a stem with apparently two growth-rings; between the two rings consisting of vessels and wood-fibres there is a zone inter- mixed with thin-walled and non-lignified cells; however, in reality the stem is 3 years old; during the first year only the innermost zone of the xylem with the small scattered vessels being developed. According to whether the shoots during the first year succeed in becoming more or less vigorous, so also does the thickness of the xylem of the first year vary; shoots may sometimes be found in which a continuous wood-ring has been developed as in the follow- ing years. In the upper part of the shoot which dies away, no cork-formation takes place, the cortical cells are less thick- walled than in the lower part and richer in chlorophyll and Casparian dots are less distinctly developed. Hard bast is ‘wanting or is scantily present in the pericycle, and the wood- ring is quite narrow. From all the axial organs medullary rays are quite absent. The Anatomy of the Leaf has been investigated by Косн and HucHEDE; my investigations entirely bear out the conclusions arrived at by them. The epidermal cells of 376 Ев. J. MATHIESEN. both the upper and lower surface have undulating lateral walls, those of the epidermal cells of the lower surface are however more strongly undulating; under a higher magni- XY OY så, ne SOSSODAS HBOS tn Fig. 4. Veronica fruticans. A, Epidermis of the upper, and B of the lower surface of the leaf. C and D, The uppermost and lowermost layer of the mesophyll respec- tively, shown in surface view. Е, Epidermal cells from the upper sur- face of the leaf, more highly magnified. F, A non-glandular hair from > the mid-rib. С, Transverse section of the leaf (A and В about °°/,; С, D, Е and С about 115/,; Е about 375/,). (Greenland.) fication the lateral walls are seen to be as shown in Fig. 4, Е: pores and nodose thickenings alternate. The outer walls have fine cuticular striations. Stomata are almost equally Scrophulariaceae. 377 distributed on both surfaces of the leaf, I found them how- ever (contrary to Wagner’s statement) to be slightly in the majority on the lower surface; the absolute number per square unit proved, however, to vary somewhat, but I did not succeed in finding any fixed rule as regards this point. The stomata are on a level with the surface of the leaf. One of the non-glandular hairs occurring on the leaf- margin and the under-side of the midrib is shown in Fig. 4, F; they are 2—4 celled, thick-walled and have cuticular warts ‘on the surface. Glandular hairs with a one-celled stalk and two-celled head occur in great numbers and are equally abundant both on the upper and the lower surface of the leaf (Fig. 4, A and В). A transverse section discloses 2—8 layers of short and broad palisade-cells with rather large intercellular spaces; the spongy parenchyma consists of ovate to slightly branched cells; all the cells of the mesophyll are abundantly filled with chlorophyll-grains. Palisade tissue and spongy parenchyma are shown in surface view in Fig. 4, C and D. The mid-rib has on its under-side a thin layer of stereom. Veronica alpina L. Alcohol-material from Norway (Tromso, leg. E.WARMING, 21. 7. 1885; Muggrubskampen, Röräs, leg. Tu. ResvoLı, 29. 7. 1918) and Greenland (Nunatsuk, 11. 8. 1885; Ivigtut and Dronning Louises ©, leg. P. EBERLIN, 21. 8. 1883 and 8. 8. 1885). — Herbarium-material from Greenland, Iceland, the Færôes and Fennoscandia. Lit.: AxELL, 1869, р. 102; Lange, 1880, р. 72; 1887, р. 261; MüLLer, 1881, р. 270; Lınpman, 1887, р. 81; WARMING, 1888, pp. 31, 35, 39, 75, 87, 93 and 142; 1890, р. 204; RoseEnvinGeE, 1892, р. 685; 1896 (II), pp. 128, 161, 168; Wacner, 1892, pp. 9 and 20; Hartz, 1894, pp. 9, 49, 50 and 57; 1895 (I), pp. 137, 170, 179, 266, 271, 288 and 304; 1895 (II), р. 385; Hartz and KRUUSE, 1911, рр. 346, 359, 364, 409, 417 and 423; Junener, 1894, р. 275; Косн, 1895, pp. 117 et seq. and 378 Fr. J. MATHIESEN. 128; Norman, 1895, р. 452; Kerner, 1898, р. 350; Кмотн, 1899, р. 171; Ствув, 1901, рр. 12, 16, 25, 40, 57 and 89; Easrwoon, 1902, р. 292; Porsitp, 1902, рр. 119, 181 and 209; 1910, р. 26773220: р. 140; Kruuse, 1905, р. 175; 1906, р. 248; 1911, in part IV pp. 196, 202, 229, 230, 242, 247, 255, 261, 262, besides many notes in the preceding parts; SYLVÉN, 1906, р. 80; HucHep£, 1907, р. 73, fig. ГУ, Аз, fig. VI, C1; ScHRÖTER, 1908, pp. 221, 226, 468 and 493; Tu. RESVOLL, 1917, р. 208. Nanophyllous sympodial chamæphyte with primary root which dies away early (SYLVEN), and abundant development of adventitious roots from the stem-bases, this in conjunction with the fact that the older shoots gradually die away, de- termines the vegetative reproduction, enabling the plants to form lax, but large tufts (Kruuse, 1906, р. 248). The winter-buds either rest upon, or are slightly raised above, the surface of the ground. The Shoot-development has been described by Tu. RESVOLL; it is a process of two years duration, as in the foregoing species; the first-year’s part of the shoots is either erect or obliquely ascending or quite horizontal, as in TH. Resvorr’s Figs. 60 and 61; most often only 1—2 cm long and bearing a few small leaves. The next year the growth of the shoot is continued in a vertical direction, while essentially larger leaves are being formed; the shoot is frequently ter- minated by an inflorescence; the flowers are already formed during the autumn of the first year (1. с.). Special bud- scales do not occur, the end of the axis is only pro- tected by the uppermost pair of leaves. After the fruit has ripened, the axis dies away to slightly above the “innova- tion-buds’’; the latter occur, however, often rather far down on the shoots, even in the axils of the very first pair of leaves, which causes the branches of the sympodia to be- come very short, and the stems crowded. Of the two shoots in the axils of two opposite leaves, the one may be far more vigorously developed than the Scrophulariaceae. 379 other; this is no doubt usually the case, and then it appears to be the shoot which is turned towards the periphery of ‚the tuft which becomes the more vigorous and floriferous. When the plant grows in damp moss, the internodes of the horizontal, first-year portions of the shoots — as mentioned by Tu. ResvoLı, and as I myself had an opportunity of verifying in the mountains of Norway — may become elon- . gated, so that the shoots become almost runner-like; in this case the plant is capable of spreading considerably. The longevity of the branches of the sympodium is greatly restricted; a growth in thickness of the axial organs, continuous for years as in V. fruticans, does not take place. Adventitious Roots are developed in the second growth-period of the shoots, they arise in the neighbourhood of the nodes. The foliage-leaves are either entire or slightly ser- rate; I am not prepared to say whether the leaves occurring at the base of the shoots remain green throughout the winter, in the following summer they are at any rate always found in a withered condition. The Flower. H. Малев, LinpmMan, Е. WARMING and KERNER have in the works cited above described the structure and biology of the flower, which according to these authors agree in the Alps, Scandinavia and Greenland, nor have 1 been able to find any differing features. “The small, dark blue flowers are at first only 2.5—3 mm in diameter, but may ultimately become 5—5.5 mm. They are protogynous-homo- gamous, and appear to be well-adapted to self-pollination. While the corolla is still almost tubular or funnel-shaped and consequently only slightly open, the anthers may be ` open and lie close to the stigma, which may be seen to be covered with pollen-grains, many of which are germinating. Afterwards the anthers are slightly removed from the stigma by the filaments bending backwards, but not so decidedly 380 Ев. J. MATHIESEN. as in Г. saxatilis; the anthers are, indeed, not far from remaining parallel with the style, and are therefore con- stantly near the stigma’? (WarminG, 1890). TH. REsvoLL. records self-pollination (at Röräs, in sched.); Fig. 5, A and B show that this may easily take place; moreover, in scarcely expanded flow- ers I too found the anthers open and the stigma covered with pollen. The ovary varies between the glabrous and finely-hairy con- dition. LiNDMAN remarks that the flowers open two at a time. They are, however, on the whole, not very con- Fig.5. A, and B, Flowers of Veronica alpina (Röräs, Norway, */,). A, In front- nes view; В, in‘side-view, half of the calyx ViSItS are scarce. and of the corolla is removed.: (About By reason of the 5/,). С, A flower of Veronica fruticans. (Hort. bot. Hauniens.) (About 4/,.) spicuous, and insect- fact that there is no division of labour be- tween the vegetative and purely floral shoots, the flowering period occurs late (TH. ResvorLL). In Greenland У. alpina sets fruit abundantly, and ripens its seeds normally; the latter are scattered in autumn; wherever in Greenland V. fruticans and V. alpina occur in the same locality, the latter is always the first to flower (M. PorsiLp). Geographical Distribution according to LANGE: East and west coasts of Greenland, Labrador, the Rocky Mountains, western North America, Siberia, Arctic Russia, Scrophulariaceae. 381 Scandinavia to Lapland and Finmark, Iceland, Great Britain and the Alpine regions of Southern Europe. To these should be added the Ferées (OstENFELD, 1901, р. 56). In West Greenland the northernmost limit is found on the Nugsuak- peninsula (70°42’); in East Greenland it passes over the Scores- by Sound. Habitat. In West Greenland V. alpina grows in copses, on ; ЕН ae “herb-slopes’’, in snow- 7 = > troughs (Snelejer) where the snow does not lie till very late in the summer, and on sandy flats near the shore (WARMING, 1888; ROSENVINGE, 1896 (II). In East Greenland it is noted too from “herb- Fig. 6. Veronica alpina. slopes” grassy-slopes Transverse section of an adventitious root ‘ (about 239/.). (Norway.) and copses; besides, KRUUSE, in the Angmagsalik district, has found it growing on the steps of steep rocks above the slopes. It is always well covered with snow during its winter-rest (cf. also KRUUSE, 1905, р. 175 and Ровзи, 1920, р. 140). Tu. Res- VOLL mentions the species as а common plant of the snow- troughs in the whole of Norway; according to NorMAN it is also found by mountain-streams, in birch-wood glades, on inundated river-banks, on the coast and on mountain-sum- mits. A. CLEvE records: “Häufig und üppig ausgebildet in XXXVII. 25 382 Ев. J. MATHIESEN. Sumpfwiesen, . . . Ferner häufig in der Moosmatte, dort noch mit Blüten und Knospen am &/, und in rein vegetat. Individ. näher an dem ewigen Schnee als irgend eine andere Blütenpflanze beobachtet. Meidet dagegen ent- schieden trockene Hei- den und Wiesen”. Tu. RESVOLL, also, states that, in unfavourable localities, where the snow lies till late in the summer, V. alpina occurs as sterile speci- mens. In the Alps, as recorded by ScHROTER, the species is “einer 255% | der verbreitetsten Ra- LEAKS RES senbestandteile von AN? SS = SH der subalpinen bis zur nivalen Region (1500— 3185 m in der Schweiz, 1560—2870 m in Bay- ern) alle Bodenarten Fig. 7. Veronica alpina. bewohnend, indifferent Transverse section of a stem (about ?%0/,). gegen Düngung und (Norway.) Humusgehalt”. Be- sides this, it occurs in snow-troughs (Schneethälchen). Anatomy. The Root. The epidermis becomes cuti- cularised and persists a long time. The outermost layer of the cortex consists of thin-walled cells with cuticularised outer and lateral walls; the two succeeding layers within the cortex (Fig. 6), have, on the other hand, fairly thick- walled cells which unite mutually and with the exodermis pt: sæden Scrophulariaceae. 383 without intercellular spaces; the walls show the reaction of cellulose. During the growth in thickness of the stele, the endodermal cells are stretched tangentially, and become sub- divided by 2—4 thin radial walls; the walls of the original endodermal cells show Casparian dots, but only faintly. The cortical cells contain starch; root-hairs occur in great num- bers. In the xylem portion I have not found indication of annual rings (the root figured is several years old) as in the foregoing species; consequently, the roots complete their development in the course of one year. The xylem is composed of vessels and wood-fibres. The Stem. As in V. sazatilis the xylem, in the first year of growth, when the shoots are but weak, frequently consists only of scattered vessels with intervening non- hgnified parenchyma, and not until the second year is a continuous ring of vessels and wood-fibres formed, as shown in Fig. 7, which shows a portion of a transverse section of the basal, persistent part of the shoot. I never found more than 3 years’ growth in thickness, not even in shoot-bases apparently older. The vessels are comparatively numerous, and the wood-fibres are more sparingly present and also more thin-walled than in the foregoing species. In the cells of the long-lived epidermis tannin is found; cork-formation does not take place, neither in the stem nor in the root. The cells of the cortex contain chlorophyll-grains. The young shoots are beset with glandular hairs of the same type as those on the leaf. In the upper part of the stem, which dies away after flowering, the wood-ring is of course formed entirely during one growth-period, here it is thinner than in the lower part; the cortical cells are thinner-walled and richer in chlorophyll, and the epidermis contains no tannin. The Leaf. The epidermal cells of the upper surface have slightly, those of the lower surface more highly, un- 25* 384 Fr. J. MATHIESEN. dulating lateral walls; they are thin and without pores (Fig 8, D). Stomata occur almost equally on both sides of the leaf, frequently they are slightly in the majority on the Fig. 8. Veronica alpina. A, Epidermis of the lower surface of the leaf. B, Epidermis of the upper surface of the leaf. C, A non-glandular hair from the leaf-margin. D, An epidermal cell more highly magnified. Е and К, The upper and lower layers respectively of the mesophyll, in surface view. G, Transverse section of the leaf (A and В about 9°/,; С, Е, F and С about 115/,; D about 375/,). (Greenland, Dronning Louises ©, 8. 8. 1885.) upper surface; the stomata are on a level with the surface of the leaf. Both leaf-surfaces are furnished with glandular hairs of the usual type — a two-celled head on a one- Scrophulariaceae. 385 celled stalk; non-glandular hairs (Fig. 8, G) usually occur only along the leaf-margin. The transverse section shows 2—3 layers of short and broad palisade cells with rather large intercellular spaces; the cells of the spongy parenchyma are somewhat more branched than in the foregoing species (cf. surface view, Fig. 8, F). Chlorophyll-grains are abundantly present in the whole of the mesophyll; they are also found in the epidermis of the lower surface. Veronica offieinalis L. (f. glabrata Fristedt). In the collections preserved in the Botanical Museum in Copenhagen there is some material of Veronica officinalis collected in the Ferées, part of which belongs to the main form with hairy leaves, and part to f. glabrata Fristedt (Væxtgeografiska skildr. af Södra Ängermanland, Upsala, 1857). Of both forms there were individuals of normal size, as well as dwarf individuals. The Morphology of the species has been described several times (e. g. WARMING, 1884, р. 58; Brunpin, 1898, р. 83), most exhaustively by the first-named author, from whose description the following is quoted: ‘The creeping and rooting shoots bear only foliage-leaves (evergreen); at the apex they are frequently bent slightly upwards in a curve, especially when the plant grows among moss; but gradually as the adventitious roots are developed in ascending succession, the stems are drawn down to the ground. From the basal por- tion of the year’s shoots there proceeds shoots which re- semble the parent shoot; in this species no lateral shoots are found which can be indicated as special “assimilatory shoots”, which are such only, without taking at the same 386 Ев. J. MATHIESEN. time part in the vegetative propagation; from the leaf- axils which follow next, racemose inflorescences proceed. But after a short pause during the flowering period, the parent shoot continues its growth at the apex and, under favour- able conditions, may also produce new lateral shoots, as also it may, it is true, continue its growth throughout the winter, as soon as thetemperature has reached a certain degree of warmth. I have, however, observed several cases in which the main shoot had been biologically arrested through flowering, and had died away as far down as below the inflores- - cences.” Fig. 9. Veronica officinalis Г. А dwarf speci- f. glabrata Fristedt. men of f. glabrata Kirkebökamp, Strömö (The Færûes). es > 2 (About 2/,). (Strömö, Kirkebö- i kamp, leg. C. H. OSTENFELD, 8. 6. 1895) is illustrated in Fig. 9; it is drawn almost twice the natural size. The individual did not flower; the few leaves from the previous year which are still remain- ing, are recognizable by their being shaded. The branching is seen to be abundant — these dwarf individuals therefore often form rather dense tufts. Uppermost in the figure the last-formed leaves are seen to bend over the apex of the Scrophulariaceae. 387 shoot; the reason of this is unknown to me. Upon the oldest portion of the shoot-system which has been figured, adventitious roots are seen to be rather abundantly devel- oped. According to RAUNKIÆR (1907, р. 46) Veronica offici- nalis is an active chamæphyte; the species is nanophyllous, and evergreen (see above). In “Field-notes on the Biology of some of the Flowers of the Færües” (Botany of the Feerdes. Vol. III, р. 1065) Е. WARMING writes concerning Veronica officinalis: “The diameter of the flower is 7—8 mm. The corolla is pale-lilac with stripes of a darker colour. Homogamous. The anthers and the stigma occur at the same level, but as the stamens are spreading, insect-pollination appears to be necessary for the setting of fruit.” Geographical Distribution: Europe, Western-Asia, North-America. According to OSTENFELD (1901, р. 57) in the Fieröes Veronica officinalis is “Rather common in low-lying regions on hill-slopes and on rocky-ledges; also occurs at high levels and there mostly as f. glabrata Fristedt . . . Fl. beginning of July. Fr. August, but bears fruit sparingly; often fails altogether in the hills.” The specimen illustrated in Fig. 9 was investigated with respect to its leaf-anatomy. The epidermis of the upper surface (Fig. 10, B) had lateral walls, from straight to slightly undulating, and that of the lower surface (Fig. 10, A), rather strongly undulating lateral walls. Glandular hairs with one-celled stalk and two-celled head occurred abundant- ly on both leaf-surfaces. In Fig. 10, С some epidermal cells are shown under higher magnifying power; the two smaller and thin-walled ones have borne glandular hairs (the small 388 Ев. J. MATHIESEN. circles inside their outlines indicate the connection of the stalk of the gland with the outer wall of the cell), the lateral walls of the others are seen to be porose with nodose thick- enings between the pores; in the main form I found these thickenings to be somewhat more strongly developed; they were very pronounced in individuals collected in dry habitats in Denmark. A transverse section of the leaf is shown in i II Dal) a aes <> LE АХЛ а ЕЯ == at Fig. 10. Veronica officinalis L. Î. glabrata Fristedt. A, Epidermis of the lower, and B of the upper surface of the leaf.\ C, Epidermal cells from the upper surface of the leaf, more highly magni- fied. D, Transverse section of the leaf. (A and В about %/,; С about 2); D about 15/,). (Kirkebökamp, Strömö (The Færôes)). Fig. 10, D; in the upper part of the figure there are two layers of short and broad palisade-cells; the spongy paren- chyma consists of rounded or slightly branched cells. Any special difference in the structure of the mesophyll in the Fzeröese and the Danish plants, could hardly be demonstrated; the description given by Kocx (1895, p. 134) also agrees with that given above. Scrophulariaceae. 389 Fig. 10, À and B, which are drawn from epidermis-pre- parations of one of the leaves with the under surface turned upward, seen uppermost in Fig. 9, show the proportion between the number of the stomata on the upper and lower surface of the leaf to be about 2 to 3; in the normally orien- tated leaves of the same individual, there were, however, generally about 3 times as many stomata on the lower as on the upper surface. Оп the whole, I always found in all the individuals of Veronica officinalis investigated by me, by far the greater number of stomata on the lower surface of the leaf, viz., 3 to 5 times as many as on the upper sur- face — this feature, as well as the absolute number of the stomata per unit of area, may however vary in the different leaves on the same individual. The non-glandular hairs, which are found in greater or fewer numbers on the leaves and stems of the main form, are multicellular, thick-walled, and have cuticular warts. Castilleia pallida (L.) Kunth. Alcohol material from Kola (the Voronej River, leg. BROTHERUS, 2. 7. 1887) and Arctic America (King Point and Herschell Island, leg. A. H. Linpstrém, 1905—1906). Her- barium material from Arctic America (Port Clarence, King Point and Herschell Island (var. unalaschkensis Cham.), Labrador, the coast of Hudson Bay (Ranken Inlet, Churchill (var. septentrionalis (Lind.) Gray)), Lapponia imandrae. Lit.: LANGE, 1880, р. 79; Е. Warmine, 1890, pp. 220—228, fig. 34; RoSENVINGE, 1892, р. 687; P. Кмотн, 1899, р. 193; Simmons, 1913, pp. 122 and 138. - Spot-bound, nano-microphyllous, sympodial proto-hemi- eryptophyte with a slightly branched primary root of long duration. Only one shoot-generation reaches maturity in each growth-period. The perennial basal portions of the shoots 390 Ев. J. MATHIESEN. live for several years, and can have growth in thickness; adventitious roots are developed but sparingly. The shoots are erect and hairy especially above in the floral part; if they are capable of flowering — which ordinarily appears to be the rule — they terminate in a raceme, the subtending leaves of which are large, the lower ones similar to the foliage- leaves, but often having long lobes. In the axils of the 4—7 radical scale-leaves (bud-scales) of the shoots “innovation-buds” occur, but of these only 1—2 are further developed; the uppermost of them appear on the whole to be the most vigorous, but none of them is a decidedly principal bud. The foliage-leaves differ greatly in shape and size in the different forms in which the plant occurs; in var. unalaschkensis they are large, ovate-lanceolate (50—60 mm long, as much as 17 mm broad); in var. septen- trionalis they are smaller, linear-lanceolate 30—40 mm long, 3—4 mm broad); the principal form is intermediate. Fruits with ripe seeds occurred in my material from the locali- ties in Labrador and Hudson Bay; the fruit ripens in July— August. The seeds are small and light with a reticulated, pitted testa; this pattern is produced by the very thin outer walls of the outermost, large-celled layer of the seed-coat, sinking down into the cavity of each cell, while to the inner and lateral walls rigidity is given by a network of anasto- mosing flange-like thickenings. The structure and biology of the flower has been ex- haustively described by Е. WARMING, (1. c.), from whom I quote the following (compare also the accompanying figure with explanation which has been taken from the paper in question): “The hairy calyx is deeply cleft into two lateral lobes, which again are cleft into a larger anterior and a smaller posterior lobe, both oblong. Here the corolla is only of about the length of the calyx, tube-shaped and *two-lip- Scrophulariaceae. Bl ped. The lower lip is shorter, with three ovate obtuse lobes which are erect or slightly reflexed; at the base of each lobe the lip is inflated into a bipartite arch (A, B and G). The upper lip is erect, boat-shaped, and terminates in an entire apex; a little distance below the apex it is provided on Fig. 11. Castilleia pallida (From Kola at the White Sea). A and B, A flower seen from the right side and from the front; the anthers (anth.) have grown out of the flower and are open; the stigma protrudes above them; behind it the upper lip (s) is seen; the middle part of the lower lip is marked m, one of its lateral lobes (the right one) /. The hairs are indicated only along the periphery. С and D, Parts of а young flower, seen from the left side and from the front, after having been artificially opened. The anthers have not dehisced; m is the middle lobe of the lip, J one of its left lateral lobes. E, A flower from the same inflorescence as A, seen from the left side, but much younger and smaller; even here the stigma is large. F, Ovary of A. G, Each stigma in the whole inflorescence to which this flower belongs is as much enclosed as it is here; the anthers (anth.) of the long stamens are open, those of the short ones are closed. H, The relative positions of the stamens and stigma in the lowermost flower (14 mm long) in an inflorescence; the two uppermost anthers are open and are situated just above the stigma; the next flower appears to be similar; the third one was almost a bud, but had its stigma protruding to the usual distance. (Drawn by Е. WARMING). 392 Ев. J. MATHIESEN. either side with a rather thin wing (D and G), which in its upper part may be erect or curved slightly inwards around the anthers and stigma, while in its lower part the edge is revolute (G) as in Pedicularis, but the revolute portion (“die Rolle”) is quite smooth here. As both lips are erect and close tightly, there is, between them, a very narrow entrance to the flower, through which, on the one hand, only very small flies and other small insects can creep in, and on the other hand, only butterflies and bees with long proboscides can gain entrance into the flower. The anthers of the long stamens open before those of the short ones (7), and the anthers of the latter always remain enclosed, while those of the long stamens protrude more or less. Here also the style appears to vary in length; in some inflorescenses I found it protruding far, even in quite young flowers — sometimes so young that they must rather be called buds (Е); and as the stigma here also appears to be developed early, these flowers must be called longistylous—protogynous. In other flowers I found the stigma to protrude less, but nevertheless to be higher than the anthers, while in others again the long stamens were so long that their anthers were in con- tact with the stigma, or even protruded above the latter (Н).” (The author then goes on to say, that among flowers from the same inflorescence there may be some difference as regards the relative length of the style, the stamens and the corolla which apparently cannot be referred to differ- ence in age); “in flowers such as A, B and E (from the same inflorescence) self-pollination will be able to take place only with diffieulty, the stigma being throughout higher than the anthers and the flower standing erect, hidden behind the large bracts provided with marginal lobes. Nor will self- pollination be easy in a flower like G, but in flowers like Scrophulariaceae. 393 H, self-pollination will be able to take place easily, as the pollen will almost inevitably fall upon the stigma.”’ To the above description I can add, that in all cases, I found in fully developed flowers, with the exception only of the flowers in two inflorescences from Kola, — which hap- pen to belong to the material upon which WARMING has based his description — the stigma projecting more or less beyond the apex of the posterior lip, usually as in Fig. 11, A and B, sometimes projecting even more, and in quite young flowers somewhat less; in several cases I verified the presence of germinating pollen upon the stigma. It has not been possible to demonstrate conditions indicative of hetero- styly. It does not appear to be uncommon for the pollen, at any rate in the anthers of the two longest stamens, to germinate even before the flower has entirely expanded, 1. e. while it is still hidden by the large subtending leaf, and the corolla is surrounded by the lobes of the calyx. As the stigma, even at that time, protrudes beyond the lobes of the corolla, the distance between it and the anthers will, however, in all probability be so great that self-pollination will not be able to take place; in material from Kola there was a flower, like that figured in 7, which had germinating pollen in the anthers. Both calyx and corolla have a dense covering of shorter and longer non-glandular and glandular hairs, similar in form to those described below under the anatomy of the leaf. According to Е. WARMING (1. с.) the flowers are visited by small wasps (the genus Pteromalus); he found such in several flowers. The Geographical Distribution of the plant is, ac- cording to LANGE: Labrador, Canada, the Rocky Mountains, western Arctic North America, Kamchatka, East and Arctic 394 Ев. J. MATHIESEN. Siberia, and Arctic Russia. This author records it also from Greenland, but as it has not been found since, it ought to be omitted from the flora of Greenland (ROSENVINGE, M. Por- SILD). According to Simmons the following localities from the Arctic North American archipelago can be added: Banks Land, Baring Land, Victoria Land. Anatomy. The Root: The epidermis dies away quick- ly. The primary cortex follows for a long time the growth in thickness of the stele by tangential elongation of the cells, accompanied by divisions in these by radial walls. During the secondary growth the cambium forms outwards a large amount of parenchyma, arranged in very regularly radiating rows, while the formation of sieve-tissue is extremely spar- ing. The cells of the parenchyma are entirely filled with starch. In the woody parts the single growth-zones are easy to distinguish, in an inner narrow portion of the annual ring the vessels being surrounded by thin-walled cells, while in the outer larger part the vessels are accompanied by wood- fibres. Also in the thin-walled cells in question starch occurs. Parenchyma-rays are absent. The Stem: As shown in Fig. 12, A and B the ring of wood in the lower, persistent part of the year’s shoots is considerably thicker than in the upper, perishable part. In that part of the stem, from which the section shown in Fig. 12 A is taken, the epidermis, together with the outer- most thick-walled cortical layer connected with it without intercellular spaces, has partly separated from the rest of the cortex; this happens especially at the nodes, thus a leaf occurred on the axis just above the plane of the section on the side turned upwards in the figure. At the base of the shoots slits are formed at some depth in the cortex (Fig. 12, B and C); this explains the fact that the cortex can so easily be rubbed off in flakes. The persistent basal portions of the Scrophulariaceae. 395 shoots have growth in thickness, but probably only to a very limited degree. In Fig. 12 C two annual-rings are seen, the last-formed is very excentric, the fact often being that the most vigorous growth in thickness takes place on that side of the stem which is opposite to a developing bud. а и) nz == I LIION =»), © SL Jess Fig. 12. Castilleia pallida. A and В, Transverse sections of the stem of the year’s shoot; A is taken from half-way up the stem, B from near the base. C, Trans- verse section of a 2-year-old, persistent stem-base. D, А portion of A, more highly magnified. (King Point.) (А, В and С about 1°/,; D about 73°/,.) Fig. 12 D represents a part of A more highly magnified, the epidermis and the outermost layer of the cortex is thick- walled; the cortex is few-layered, its cells contain chloro- phyll-grains; stomata occur in the epidermis. The endo- dermis is only demonstrable by faint Casparian dots. 396 Ев. J. MATHIESEN. < ©. D - A OH ae, Cote =o EX sd ХХХ 0 Zu | Wr CY } CR I OSES Oe ae III DI : Eu) = Le у ic A À As N Ja A N > @ rs I _ > \ | À LS À <=> SAS ÈS = = | Ca 5 SEG 5 STEDEN Fig. 13. Castilleia pallida. A and В, Transverse section of leaves (5/1). A, var. septentrionale (Churchill); B, the principal form (Kola). C, Transverse section of the midrib of the leaf (Kola). D and E, The uppermost and lowermost layers of the mesophyll, seen in surface view. F, Transverse section of leaf. G, A glandular hair, more highly magnified. H and I, Epi- dermis from the upper and lower surface of the leaf respectively. (D, Е, Е, H and I King Point; G Kola.) (С, D, Е, Е, H and I about 90/,; G about 220/,.) Scrophulariaceae. 397 In the wood, medullary rays are entirely wanting (the radiating lines in Fig. 12, A, B and C indicate only the radial arrangement of the elements); exteriorly, the ring of wood consists chiefly of wood-fibres; interiorly, vessels become more frequent, and are accompanied by partly lignified paren- chyma. In the basal portion of the stem the vessels are relatively more numerous than shown in Fig. 12, D, and the cells of the stereom are thinner-walled. The pith consists exteriorly of lignified, rather thick-walled, distinctly porose, axially elongated parenchymatous cells, towards the centre the cells of the pith are thinner-walled, non-lignified and die away, so that the stem becomes hollow. The Leaf: Any difference as regards anatomy could scarcely be demonstrated in the leaf-types of the different forms. Transverse sections, slightly magnified, of the leaf of var. septentrionale and of the principal form are shown in Fig. 13, A and B, the three main veins and the numerous fine anastomoses are seen. The vascular bundle of the main veins has on its under- side a covering of somewhat collenchymatously thickened elements; on the upper side the surface of the lamina dips down towards the vascular bundle. The structure of the mesophyll is fairly homogeneous, thus in Fig. 13, F it is hardly possible to demonstrate a palisade-layer; in specimens from Kola the uppermost layer of the mesophyll was how- ever elongated in a somewhat palisade-like manner. Chloro- phyll-grains occurred abundantly in the entire mesophyll, and besides this they were found in the epidermal cells of both surfaces. The epidermal cells of the upper surface have from straight to slightly undulating lateral walls, those of the lower surface can be more strongly undulating. Both the outer and lateral walls of the epidermal cells are thin; the cuticle XXXVII. 26 398 Ев. J. MATHIESEN. of the outer walls is but slightly developed. Stomata occur in almost equal number on both sides of the leaf; the guard- cells are surrounded by 4—5 cells, and are on a level with, or slightly raised above, the surface of the lamina — on the large subtending leaves of the flowers I found them sometimes raised high above the surface. The leaves are densely covered with hairs; the following types of hair occur: (1) Non-glandular, thin-walled, pointed and with fine cuticular striations; they may either be short, (1—2 celled) or long (multicellular) and even elongated like a whip, the last-mentioned are numerous especially in var. unalaschkensis, and particularly in the top of the shoots. (2) Glandular hairs with multicellular stalk, and with 1—2 celled head; the basal cell has sometimes cuticular striations. (3) Glandular hairs with unicellular, quite short stalk and 2-celled head (in Fig. 13, G such a hair is seen in lateral view, highly magnified; in H and J they are seen from above), in the cells of the head and the stalk comparatively large nuclei occur, and a highly granular protoplasm. These various forms of hair occur intermixed (cf. Fig. 13, F, H and J); they are also found on the surface of the stem and, as already mentioned, in the floral region. Euphrasia aretiea Lange. (Euphrasia latifolia (Pursh) Wettst. ) Alcohol material from West and East Greenland (Fre- drikshaab, leg. ROSENVINGE, 15. 8. 1886; Sydpröven; Неа Havn, leg. N. Hartz, 13.8; Unartok and Tasiusak). Herbarium material from West and East Greenland, the shore of Hudson Bay, Northern Norway, Lapponia murman. Lit.: LANGE, 1880, р. 79; 1887, р. 264; WARMING, 1886, pp. УП and 43; 1888, рр. 34 and 59; 1890, р. 226; RosEnvinGe, 1892, р. 687; 1895 (1), р. 68; 1895 (II), pp. 162, 168, 219 and 245; Harız, 109% Scrophulariaceae. 399 рр. 9, 15, 20, 46 and 57; 1895 (I), рр. 146, 170, 179 and 289; 1895 (II), рр. 335, 359, 372 and 377; Hartz and Kruuse, 1911, рр. 359, 364, 409, 416, 423 and 428; WETTSTEIN, 1896, р. 136; ABRoMEIT, 1899, р. 46; Dus£n, 1901, р. 40; Porsitp, 1902, р. 197; 1912, рр. 382 and 387; 1920, р. 141; Kruuse, 1905, р. 176; 1906, р. 250; 1911: in part ТУ, рр. 196, 229, 242, 247 and 261, and besides these many notes in the preceding parts; JØRGENSEN, 1919, р. 99. As regards the Morphology, Biology and Anatomy of the Euphrasia spp. see: MÜLLER, 1881, р. 279; HOVELACQUE, 1888, рр. 400, 454 and 477; Linpman, 1887, р. 81; Косн, 1895, рр. 140—144; WETTSTEIN, 1896; HEINRICHER, 1898 and 1902; Кмотн, 1899, рр. 202—206; and Kerner, 1900. Nanophyllous therophyte; the seed germinates during early summer. The main root is rather scantily branched; above the cotyledons there are 1—3 pairs of leaves separated by rather long internodes; “the first flowers generally occur in the axils of the 3rd or 4th pair of leaves, frequently in that of the 2nd pair, more rarely not until in that of the 5th, and as an exception, even in that of the Ist pair” (JORGENSEN). To- wards the apex of the stem, the distance between the leaf- pairs becomes very short; the leaves become broader, and their teeth longer and more pointed than are those on the lower leaves of the stem; they all subtend flowers. The branching is, as a rule, scanty; sometimes a few weakly- developed branches are seen to proceed from the uppermost pair of leaves below the inflorescence; occasionally, however, branches, almost as vigorous as the main axis, may be devel- oped from the axils of the two lowermost pairs of leaves — in this case it is evidently a matter of rather late-flowering individuals (August—September). Stem, leaves and calices are more or less densely covered with hairs. “Die Pflanze variiert sehr viel, von sehr klein und ein- fach, mit wenigen Blüthen — so häufig an der Witterung 26* 400 Ев. J. MATHIESEN. ausgesetzten alpinen Standorten — bis sehr gross, grob und verästelt mit sehr grossen und breiten, grobgezähnten Blät- tern, die bis über 2 em lang und ebenso breit werden” (Jör- GENSEN). Similar variations were found in the Greenland material. From the above-said it follows that the individuals generally agree in habit with the growth-type indicated by WETTSTEIN as “frühblütige Form” (l.c. Fig. 1, р. 44); the comparatively short period of vegetation with which the individual may be compelled to be satisfied, in consequence of the geographical distribution of the species, makes, as JORGENSEN (I. с. р. 104) remarks, this growth-type necessary in Arctic regions. The flower-morphology and -biology of the Euphrasia spp. have been described so often that I can here confine myself to the following remarks: E. arctica belongs decidedly to the small-flowered forms; in no case did I find the corolla to be more than 7mm in length, more frequently it was shorter. The flowers are protogynous, and in the recently expanded flower, the stigma generally stands somewhat in front of the anthers, afterwards the style curves downwards and backwards, so that the stigma is brought into close contact with the anthers which are now quite open (WAR- MING, 1890). Sometimes, however, the stigma, even at the time when the flower is expanding, is so near to the anthers, that self-pollination must be able to take place easily; I even found pollen on the stigma of a not yet expanded flower. WARMING (1890, р. 227) mentions that a length-incre- ment of the corolla-tube can take place — as in larger- flowered species —, by which the anthers are carried forward and approach the stigma; in some few of the flowers I did find the stigma protruding so far that a movement of this kind may very probably be of importance. Geographical Distribution according to JORGENSEN: Scrophulariaceae. 401 Greenland, Cumberland, Labrador, Iceland, the Ferées, Scot- land, Sweden southwards to Herjedalen and Jämtland, Nor- way (common towards the north, rare south of the Trond- hjem Fjord), Northern Finland and Russia, in Siberia at least as far towards the east as Jenisei. Ровзию (1912, р. 382) records the northern limit of the species in West Greenland to be Tartusag (71°25’ N. lat.). In East Green- land E. arctica extends almost to 74° N. lat. (Lille-Pendulum Island: Dus£n); from the district of Angmagsalik north- wards it occurs most frequently, and grows more luxuriantly at the head of the fjords, whilst it is rarer and stunted in growth along the coast outside the fjords and on the islands off that coast. The Habitat is given by the various authors in unison as being willow copses, heaths, grassy slopes and “herb- slopes”; RosENVINGE (1896 (II), pp. 219 and 245) also found the species growing on gravelly flats (Igaliko) and on knolls in mossy bogs; Hartz found it in the district of Scoresby Sound flowering as late as the middle of September. The fruit-setting appeared everywhere to be abundant and good. The Anatomy of the root and the stem has been very exhaustively described by HOVELACQUE as regards “E. offi- cinalis”; in Е. arctica I found nothing which differed from his description. The Root has only a few root-hairs; the cortex is thin and few-layered (3—4 cell-layers in thickness), its elements are greatly elongated in a tangential direction; in the endo- dermis the Casparian dots are distinct; in proportion to its diameter, the xylem part in the full-grown root forms a very thick bundle, composed of rather thin-walled vessels and wood-fibres. The Stem has a rather thick-walled epidermis, upon which occur 2—3 celled, non-glandular hairs with fine cuti- 402 Ев. J. MATHIESEN. cular striations, and small glandular hairs with one-celled stalk and two-celled head. The cortex is few-layered (3—4 cell-layers in thickness) as in the root; in the endodermis the Casparian dots are faint, but nevertheless demonstrable Fig. 14. ÆEuphrasia arctica. A, Epidermis of the upper, and B of the lower surface of the leaf. С, Transverse section of the leaf. D and ЕЁ, Glandular hairs from the lower surface of the leaf. F, A glandular hair from the margin of the leaf. G, A 2-celled, non-glandular hair. (А, В, С, D, Е, Е and С about 410/,). (Greenland.) with Sudan III. The pericycle is 1—2 layered. The outer part of the rather broad ring of xylem is composed exclusively of wood-fibres. The peripheral cells of the pith have some- what thickened walls. Medullary rays are absent. The Leaf. A transverse section of one of the lower foliage-leaves is shown in Fig. 14, С; there occur 1—2 layers Scrophulariaceae. 403 of short palisade-cells, with rather large intercellular spaces; the spongy parenchyma consists of only slightly branched cells. The mesophyll of the upper leaves of the stem (i. е. the leaves subtending the flowers) appears, on the whole, to be more lacunose in structure than that illustrated in Fig. 14, С. The margins of the teeth of the leaves are somewhat bent over, whereby a concavity is produced on their under side. In this concavity the leaf-surface is densely covered with glandular hairs, and the cells of the epidermis differ in character from the other epidermal cells of the leaf. For whilst these, both on the upper and lower surface — and especially on the upper side of the teeth of the leaves — have strongly undulating lateral walls (Fig. 14, A and B), which only here and there show a small flanged thickening, the epidermal cells in the concavities, and especially those which bear the glandular hairs, have straight or only slightly wavy lateral walls (Fig. 14, D and Е), which are often (not, however, in the places illustrated in Fig. 14, D) furnished with flanged thickenings, the one by the side of the other. Stomata are almost equally abundant on both leaf-surfaces, in the concavities of the lower side of the teeth, however, they are less frequent; the guard-cells are on a level with the surface. In the teeth the veins break up into a fine net- work of tracheids. The glandular hairs which occur on the lower side of the teeth are of two types: (1) a smaller one with one-celled stalk and two-celled, globular head, and (2) a larger one with two or more frequently four-celled cupola-shaped head, seated upon a short and broad stalk-cell which partly sinks below the leaf surface, the inner-wall of the stalk-cell, during the development of the hair, becoming rounded inwards, thus pressing the layer of epidermal cells, situated under the stalk- 404 Ев. J. MATHIESEN. cell, in the same direction (Fig. 14, C). The stalk-cell of the latter type of hair rests upon the connecting parts of 4—8 epidermal cells, which are produced by the division of a single cell, viz., the epidermal cell which has been the mother- cell of the glandular hair. Below the stalk-cell small, circular intercellular-spaces are formed between the lateral walls of the cells of the epidermal layer (cf. Fig. 14, С and О; in D that part of the lateral walls of the epidermal cells which lie under the stalk-cells, is indicated by dotted lines — the dotted circle inside the head of the small glandular hairs indicates the outline of the stalk-cell). These large, peltate glands appear, therefore, to belong to the same type as the “Schilddriisen” of Lathraea (and Bartschia), the development of which has been so thoroughly described by SCHERFFEL; the intercellular spaces in the layer under the stalk-cell have not been perceived by HovELACQUE or WETTSTEIN. The small glandular hairs belonging to type 1 are also found in other places on the leaf, but nowhere in such abun- dance, as in the concavities on the under surface of the teeth. Glandular hairs with 1—2 celled head and multicellular stalk (Fig. 14, F), such as those required for the diagnosis of species, are found both along the margin of the leaf and also — but scantily — upon the lower surface. Non-glandular hairs occur upon both leaf-surfaces, and are especially numerous towards the margin and on the veins; they are most frequently one-celled, but can also be 2—3 celled, pointed, and have fine cuticular striations (Fig. 14, A, C and G; C and G are seen in optical section; the thickness of the walls is evident from the figures). The epi- dermal cell or cells (2—4) which bear them are distinguished from the surrounding cells by having less strongly undulating lateral walls, and outer-walls which are more or less strongly arched outwards. Scrophulariaceae. 405 In the concavities on the under surface of the teeth I found the epidermis covered with crystal aggregates, which proved to consist of calcium-carbonate. — The same observation has been made by WETTSTEIN (1. с. р. 19). I have sometimes also found crystals of the said substance in the cells of the large glandular hairs, as shown in Fig. 14, Е. The palisade-tissue and spongy parenchyma were rich in chlorophyll-grains ; and such were also found in the epi- dermis of the lower surface. Bartschia alpina L. Alcohol material from Norway (Aursundsöen, leg. TH. RESVOLL, 31. 8. 1918; Bosekop, Kaafjord, Alten, leg. Е. WARMING, 11, 17. 8. 1885), Greenland (Sukkertoppen, 5. 7. 1884; Lyngmarken, leg. L. К. RosENVINGE, 25. 7. 1886; Кап- gerdluarsuk, 5. 8. 1884 and 30. 7. 1885; Holsteinsborg, 15. 7. 1884; Sarfanguak, 15. 7), and Iceland (Gullfos and Tungafos, leg. A. FEDDERSEN, 25. 6. and 30. 6. 1886). Herbarium material from Fennoscandia, Island of Kol- gujew, the Færües, Iceland, Greenland, the coast of Hud- son Bay. Lit.» Axevr, 1869, р. 102; Lance, 1880, р. 78: 1887, р. 263; MÜLLER, 1881, р. 283; WARMING, 1886, pp. II, VI, 7 and 43; 1888, рр. 34, 39, 67, 75, 93 and 188; 1890, р. 226; Linpman, 1887, р. 82; HOVELACQUE, 1888, pp. 403, 451 and 478; RosENviNGE, 1892, р. 687; 1896 (II), рр. 128, 144, 161, 165 and 168; Hartz, 1894, pp. 9, 15, 27, 43, 57; 1895 (I), р. 302; 1895 (II), pp. 359, 372 and 392; JUNGNER, 1895, pp. 224 and 238; Norman, 1895, p. 455; ABROMEIT, 1899, p. 45; Кмотн, 1899, р. 197; Kerner, 1900, Bd. I, pp. 107, 129, 131 and 32, 175 and. 606; Bd. II, рр. 116, 156, 251 and 301; Greve, 1901, pp. 8 9. 23, 40,57, 11, 76 and 89; Heinricher, III, 1901: PoRsıLno, 1902, рр. 181 and 194; 1912, р. 382, 387; 1920, р. 141; Kruuse, 1905, р. 177; 1906, р. 249; 1911: in part IV, pp. 202, 229, 242, 244 and 254, and besides these numerous places in the preceding sections; SYLVEN, 1906, р. 86; Horıstein, 1907, р. 126; Эснвбтев, 1908, pp. 465—68 and 772. A nanophyllous to microphyllous proto-hemicryptophyte with subterranean runners; the shoot-development usually extends over 2 years. 406 Ев. J. MATHIESEN. According to HEINRICHER and SYLVEN the seed germin- ates in early summer. The first-named author, who has studied Bartschia very thoroughly, has pointed out that it belongs to the comparatively few plants which produce buds in the axils of the cotyledons. Whilst the uppermost part of the primary shoot dies away, the following year one of these buds develops into a foliage-leaf-bearing shoot, in the basal portion of which, small innovation-shoots arise, which, lke their parent-shoot, will continue their growth in their second year, and become foliage-leaf-bearing shoots. As was the case with the primary shoot, the uppermost part of the axes of the succeeding shoot-generations dies down to above the innovation-shoots. Even after the 4th vegetative period, HEINRICHER’S cultures did not yet bear flowers; the young plants had formed small sympodial, subterranean rhizomes, composed of the persistent, basal portions of the innovation- shoots. HEINRICHER is of opinion that, in Nature, the first vege- tative-stage is of 4—5 years duration. With regard to the morphology of the full-grown plant, in the following description the statements given in the cited work of HEINRICHER will be found verified; for criticism of preceding investigations (HOVELACQUE, KERNER), reference may be made to the same author. Fig. 15, A and D show the structure of the shoot, and the mode of branching. In Fig. 15, A, each of the two shoots, IT! and II”, had terminated in an inflorescence. The shoots are in their 2nd year of development; during the first year each of them had developed a subterranean portion, bearing decussated scale-leaves which have now withered; the upper- most pairs of these scale-leaves has served as bud-scales for the winter-buds. The first year’s growth-increment especially of II? is seen to have the character of a runner, the growth- oe ae Scrophulariaceae. 407 direction has been horizontal, and the distance between the nodes is comparatively considerable. The apex of such a runner is shown in Fig. 15, С. In Fig. 15, D, ИГ? has not borne flowers; the sister-shoot ///! has been arrested in its Fig. 15. Bartschia alpina. A, The lower portion of a flowering plant; both ГЛ and I 7? have borne inflorescences (Holsteinsborg, 5. 7.1884). В, A foliage-leaf. С, The apex of a runner. D, Portion of a sympodium in which the runner- like portion of shoot II is particularly long; г is an adventitious root. (С and D, Aursundsôen, 31. 8. 1918.) (About 4/3.) development, the two small basal pairs of leaves are from the previous year, and withered, but the bud is in a fresh condition; like shoot // shoot / has had a long runner-like portion. The innovation-shoots are always developed from the portion of the shoot of the first year; in case the latter 408 Ев. J. MATHIESEN. shoot is elongated in a runner-like manner, the new shoots generally appear near the spot where it begins to bend up- wards, but И it 13 short and vertical, the new shoots most frequently appear at the base — sometimes, however, also higher up; as an exception, shoots (“enriching-shoots’’) may appear in the axils of the lowermost foliage-leaves: such shoots, however, will probably be always purely vegetative and terminate their development in the same year as the parent-shoot. After flowering and fruit-setting, the latter shoot dies down to immediately above the innovation-shoots. Innovation-shoots are generally developed to the number of two on the parent-shoots; //! and //? in A and 7711 and I/II? in D are seen to have arisen in the axils of two opposite scale-leaves; this is most frequently the case, and explains the fact why two flower-bearing axes are frequently found to stand together. In Fig. 15, D а non-expanded bud is seen in the axil of the first scale-leaf to the right of the subtending leaves of the innovation-shoots. The innovation-shoots are developed rather late in the Arctic summer; in the abundant herbarium-material from Greenland which I have at my disposal I never found the young shoots to be visible in individuals collected in June— July, whereas in those collected in August they were about one centimetre long. This does not quite agree with the statement of HEIN- RICHER, according to which they already expand simultane- ously with the foliage-leaves of the mother-shoot. The dif- ference may perhaps be explained by the more unfavourable conditions of growth, presented by the Arctic summer, in which the building-up of new organs can be compelled to extend over a longer space of time. The same author states, that small buds for the inno- vation-shoots are already to be found the year previous to a Scrophulariaceae. 409 that, in which they expand; the same is the case with the buds for the flowers. As already mentioned, the portion of the shoot devel- oped during the first year of growth, may either have elong- ated internodes and be runner-like, or — as e. ©. figured by KERNER оп р. 129 (I. c.) — it may have quite short inter- nodes with close-set scale-leaves; the first 2—3 internodes of the stem are, however, generally somewhat longer than those following; and here the scale-leaves are smaller. If by means of the latter form of shoots the ramification should be continued through several shoot-generations, the indivi- dual becomes gradually cæspitose in habit. A tuft of this kind collected by M. Ровзию in West Greenland (“heathy slopes”) had about 50 older and withered shoots, and some 20 fresh ones — both vegetative and flowering. The indivi- dual had sometimes formed quite short runners, adventitious roots were scantily developed, and the entire system was borne by a single root (whether main root or adventitious root was not clear) about 5mm thick. The adventitious roots spring — often abundantly — from the nodes. The rhizome, probably as a rule, dies away rather quickly from behind; however, in a specimen from East Greenland with cæspitose habit, a part of it had attained a thickness of fully 5mm, and the whole system was evidently old. The plants are densely covered with hairs especially at the top; hairs (non-glandular) also occur on the subterranean portions of the shoots, KERNER ascribes importance to these hairs as organs of absorption; according to HEINRICHER, however, there is no reason to assume that they have such a function. The Flower-biology has been dealt with by H. MULLER, Е. WARMING, LinpMAN, KERNER and others. The accompa- 410 Ев. J. MATHIESEN. nying figures are borrowed from WARMING, 1886, and the following is cited from his description: “The Scandinavian and Greenland individuals of this species appear to me to be exactly alike in every way, but differ from those of the Alps in at least one respect, that is, provided that H. MULLER has informed us of all the forms occurring there. Fig. 16 shows Fig. 16. Bartschia alpina. A and В from East Greenland (Eseruın, 30. 7. 1885); С from Kaa- fjord near Alten in West Finmark; D from Godthaab in West Green- land (28. 6. 1884); all the anthers were open, and pollen had already germinated upon the stigma. E and F are from Holsteinsborg in West Greenland (15. 7. 1884). A—F are twice the natural size; the others, which are all drawn from Greenland material, are in various magni- fications. For further particulars see text (WARMING, 1886). the structure of this flower, partly from Norwegian and partly from Greenland material (in the majority of cases the hair- covering has been omitted or only indicated along the outlines of the figures). The conspicuousness of the dull purple-violet —— ENE Scrophulariaceae. 411 corolla, which like the calyx, is densely covered with glandular hairs, is increased by the fact that the subtending bracts are also of a dull violet. I have not noticed any scent, but honey is secreted by the greenish nectary on the front of the ovary (Fig. 16, С). MÜLLER has described how humble-bees and other insects, by thrusting their proboscis and also their heads into the flower (see Fig. /, the corolla seen from the front) strike against the pointed lower ends of the anthers (Fig. 16, К and Г), whereby the anthers which adhere by their hairs, are violently torn from one another, by which means the dry and light pollen-grains are scattered and fall upon the insect, which then easily conveys them to another flower and deposits them upon the stigma (Fig. 16, A—E, М) which protrudes even in the bud (Fig. 16, B). Specimens with flowers, the structure of which is exactly like that of the flowers of the Alps, have also been found to occur both in Norway and in Greenland!, these specimens have evidently marked insect-visited flowers which either cannot pollinate ‘themselves, or can only do so with great difficulty. My figures (which have all been drawn from full-grown flowers) show that both the length of the style and the size of the whole flower vary considerably, apart from the difference of age. Thus specimens are found which have so short a style through- out their whole life, that the stigma does not become visible outside the corolla; in some, however, the stigma is situated immediately inside the throat of the flower (Fig. 16, К, К), but in others it lies even as far inside as above the hind- most anthers (Fig. 16, L). In these cases the stigma lies against the anthers, and self-pollination, as far as I can see, is inevitable; and in such flowers I have also distinctly seen 1 MULLER does not mention that the style is beset with stiff hairs directed forwards, except at its uppermost end (Fig. 16, H); but it is probable that it has this feature also in the Alps. 412 Tee) Me pollen-grains deposited especially upon the lower part of the stigma — the part which is in contact with the anthers (Fig. 16, M). For the rest, the pollen-grains fall out very easily, and it is very common for masses of pollen-grains to be found scattered everywhere in the flower: on the corolla, the filaments, the style, etc. Even if the stigma ripens shortly before the anthers — which I found to be the case in the distriet of Holsteinsborg — the two organs nevertheless very soon become simultaneously functional and this throughout the greater part of the life of the plant, so that self-polli- nation is possible. That Bartschia, in many places in Green- land, produces ripe fruit, I saw in 1884, from the fragments of such fruit which had remained over from the preceding year.” According to KERNER (II, р. 301) wind-pollination takes place regularly in Bartschia, in a similar manner аз for in- stance in Lathrea, at the end of the flowering period: the style in the older flower withers, the filaments are elongated and carry the anthers outwards, the latter separate and pollen, which may have been left in the anthers is scattered by the agency of the wind, and may pollinate the younger (upper) flowers of the inflorescence. This probably explains the fact mentioned by Linpman that the anthers may pro- ject beyond the edge of the corolla. This author also men- tions the same variations in the length of the style as those of which WARMING speaks. Also in the material at my disposal there were flowers in which the anthers protruded, and it was generally the lowermost (oldest) flowers of the inflorescence in which this was the case. Аос. ScHuzz differs from WARMING in finding that in the Riesengebirge (Bibl. bot. No. 10, 1888, p. 74 — here cited from Knuth) the length of the style, in relation to the corolla, Scrophulariaceae. 413 differs in older and younger flowers, so that the style of the younger flower protrudes farther than that of the older, the corolla-tube of the latter having increased in length. The elongation may be as much as 5 mm and bring the anthers — the filaments follow the corolla in growth — into contact with the stigma, so that self-pollination takes place. War- MING (1890, p. 226) admits that such a growth can take place, and disturb the relations of the organs in question, but main- tains that individual differences also occur in the relative length of the style and the corolla. I found the same forms that have been illustrated by WARMING (1886) in his Fig. 3, А Е, К and L (reproduced in this paper as Fig. 16), and even if there might possibly occur some difference in the length of the corolla of older and of younger flowers of the same inflorescence, yet I never found the style pro- truded far in the young flowers and enclosed in the older ones; that, at any rate under more northerly latitudes, it is really а matter of individual differences, 15 supported by а statement of Mr. NyHuus of Tromsö, who in a letter to Е. WARMING in 1885 writes: “On Aug. 24th I found Bartsia there (1. e. Dalfjæld in Marknæsdalen) in abundance (towards the south) from a height of 2500 feet and down over the whole mountain; but from about 2500 to 1500 feet I did not find a single flower with protruding style although I can truthfully say that I examined several hundred plants. On the other hand, lower than that, long-styled flowers became constantly more frequent” (Е. Warming, 1886, р. 10). I found the fruit-setting to be good and abundant every- where. “Die geflügelten Samen werden durch den Wind ver- breitet.’’” (SCHRÖTER). Geographical distribution in the Arcties according to LANGE and Porsizp: Labrador, Arctic Russia and Siberia, (according to Е. В. KJELLMAN's lists of the flora, in Siberia XXXVII. 27 414 Ев. J. MATHIESEN. it does not seem, however, to occur at or near the coast), Scandinavia, Iceland, the Ferées and Great Britain, in addition, the species extends over “ein mittel-europäisches Gebirgsareal von den Pyrenäen bis zum Balkan mit Vor- stössen zu den deutschen Mittelgebirgen und zahlreichen Standorten im Vorland’ (Зснвотев). In West Greenland its northern limit is the Ignerit Fjord (Umanak), in East Green- land it does not reach so far as to the region of Scoresby Sound. Kruuse (1905, p.177) states that the northern limit there is at 68°8’ М. lat. Habitat. According to WARMING in West Greenland it grows in willow-copses and on “herb-flats’’; in the most southerly West Greenland ROsENvVINGE found it in willow and birch copses, on “herb-slopes” and grassy slopes. As regards the conditions pertaining to the growth of Bartschia in East Greenland Kruuse (1905, р. 177) writes: “it is very rare north of 66°20’ and only 10 cm. high, but yet it sets flower everywhere and as far as can be judged from the collected material also fruit. It shuns here the coast, and all finding places are well sheltered, exposed to the south and have the character of herby slopes, while the species more to the south prefers heath”, and in the Ang- magsalik distriet it is (1906, p. 249) “commonly distributed on herby slopes and in fertile heath”. Norman records it from Northern Norway as growing “in bogs, on damp cliffs, in birch-woods, on flats at the river- banks and on the beach, as an exception below the upper- most belt of sea-weeds; it occurs chiefly on the sunny sides, somewhat more seldom on the indifferent (eastern and west- ern) sides and as an exception on the shady sides”. A. CLEVE records that on the mountains in the north of Sweden it is “Ein sparsamer, exclusiver Bewohner der trockenen Blüten- wiese mit vollen S-Exposition . . . Ein typischer hydrophyt Scrophulariaceae. 415 der Weiden- und Waldregionen Lapplands, kann aber diese Art im Hochgebirge sich nur an den viel Wärme darbieten- den Stellen behaupten, mögen letztere auch viel trockenen sein, als es gewöhnlich in dem Tieflande der Fall ist.” In the Alps it grows on “alle Formen alp. Wiesen und den Quellfluren von den Vorbergen bis gegen die Nivalregion (SCHRÖTER)’. In the Arcties Bartschia belongs to the group of plants of later summer, which is easily understood when one con- siders the considerable vegetative work the shoots must carry out, before flowering can take place. Anatomy. The Root. An adventitious root, about 1 mm thick, showed the following structure: The epidermis had died away, only here and there a few of its cells remained in a collapsed condition. The outermost layer of the primary cortex had cuticularized cell-walls; in many of its cells a division had taken place during growth, and in these cells one very thin, non-cuticularised radıal wall was found. The cells in the inner layers of the cortex were distinctly tan- gentially elongated and were likewise divided secondarily by thin radial walls — in the cells of the endodermis as many as four such walls were found. The wall of each of the ori- ginal endodermal cells is in its whole circumference furnished with a cuticularized layer, but as in the exodermis, the thin radial walls which had developed later, were not cuticularised here either. The root was 3-rayed; in transverse section the xylem-part now formed a circular bundle, consisting of ves- sels and wood-fibres, since the cambium had commenced its activity all the way round. The epidermis evidently dies away very early; even in a root hardly 0.5 mm in thickness, the epidermal cells were found in a collapsed condition. The root of 5 millimetres thickness, with which the indi- vidual from West Greenland, mentioned above, was fur- 27% 416 Ев. J. MATHIESEN. nished, had a xylem-cylinder, 3 mm thick, which consisted for the most part of vessels, thin-walled and rather wide; in addition thin-walled wood-fibres occurred. Parenchyma rays did not occur, neither were there any indication of growth-zones. In the secondary cortex sieve- tubes were rather scantily present. Cork was not devel- oped. In the roots both the endoder- mal and the exor- dermal cells were full of a brown sub- stance. The Stem. In HovELAQUE’s work it is particularly the PT à anatomy of thestem Fig. 17. Bartschia alpina. \ Diagrammatic transverse sections of the stem: which has been ex- A and В, from runners; С, from the aerial haustively treated, stem taken between the 3rd and 4th pair of foliage-leaves; D, taken at the point of tran- sition between rhizome and the aerial stem on this point, since (the black ring in C and D is the hard-bast, whilst the wood-ring in all the figures is radi- ally shaded). (About 18/,.) ated by HOLLSTEIN. To this I have noth- ing new to add, and can therefore confine myself to the fol- and Bartschia has, been again investig- lowing brief description. A section taken from a little above the middle of the foliage-leaf-bearing part of the stem (a diagrammatic repre- sentation is given in Fig. 17, C) shows a thin-walled epider- mis, beneath which there is a cortex, the outermost layer of which has somewhat thickened cell-walls; several of the cor- tical cells have undergone division by means of thin radial L Scrophulariaceae. 417 walls. In the endodermis — contrary to HOVELAQUE, who mentions its cells as being “sans cadres” — I found distinct, although rather faint, Casparian dots. The outermost layers of the stele consist here of rather thin-walled lignified prosen- chymatous cells; the sieve-tissue forms a continuous ring outside the xylem-part, which in the whole extent of its periphery consists almost exclusively of wood-fibres. The pith consists of thin-walled cells which collapse early in the centre of the stem. A section taken at about the middle of the subterranean part of the axis, is distinguished from the above by having a more distinct endodermis, the cells of which have, as in the root, a suberised lamella in the whole circumference of the wall — the contents of the endodermal cells have a brown colour —, by entirely wanting hard bast, or at any rate, by having it only in a far more slightly developed degree (in the two diagrammatic figures, Fig. 17, A and B, the line between the outer edge of the xylem-ring and the periphery is the endodermis, the black dots within this, in Fig. В, are small groups of hard-bast cells — in A these are totally wanting), and lastly, by a comparatively nar- rower pith and broader xylem-ring, in which wood-fibres are less numerous and thin-walled. At the point of transition between rhizome and the aérial stem, the hard-bast appears successively (the cells of the hard-bast are here somewhat more thick-walled than at the top of the shoot), the pith widens, the xylem-ring becomes thinner and richer in wood-fibres throughout its periphery. The sieve-tissue is divided into four larger groups, since the hard bast and the wood-fibres of the xylem-ring at four points which decussate, tend to approach one another, and may practically coalesce at these points (this phenomenon is connected with the passing-out of the leaf-trace-bundles 418 Ев. J. MATHIESEN. from the stele, compare Fig. 17, D, which is a diagrammatic representation of a part like the one just described — the section was taken between the uppermost scale-leaf and the first pair of foliage-leaves), small groups of sieve-tissue A B AVE Fig. 18. Bartschia alpina. A and В, Transverse sections of foliage-leaves. С and D, Portions of the epidermis of the upper and of the lower surface respectively. (West Greenland.) (A and В, about 14°/,;; С and D, about *?9]..) elements, however, are always found enclosed in the stereom (well-illustrated in HoveLaquE’s Fig. 334, р. 405). Medullary rays do not occur. The epidermis of the rhi- zome persists for a long time; cork is not developed (HorL- STEIN); in the older rhizomes a successive dying-away of the outer layers of the cortex takes place. The Leaf. The epidermal cells of both the upper and the lower surface have undulating lateral walls, this being somewhat more strongly the case with those of the lower surface than with those of the upper (Fig. 18, С and D). Scrophulariaceae. 419 Stomata occur on both leaf-surfaces, but they are more numerous on the lower surface; the proportion between the number of stomata on both surfaces proved to be about 1 to 2. On the upper surface the lateral walls of the epidermal cells are slightly porose, especially near stomata; the guard- cells are surrounded by 3—6 cells, and are on a level with the surface, or project slightly (Fig. 18, A and B). The transverse section (Fig. 18, A and В) shows 1—3 palisade-layers. Both A and B were drawn from sections of leaves taken from specimens from West Greenland, and in both cases the leaves were chosen from the middle of the stem. A is seen to be somewhat thicker than B, and to have an assimilatory tissue consisting of as many as 3 cell-layers, the elements of this tissue are somewhat irregular and some of them are slightly branched; B probably represents the more common type. The transverse section of the leaves of the Norwegian plants resembles, on the whole, the trans- verse section shown in B, with the exception that the pali- sade-cells are slightly higher. The spongy parenchyma con- sists of rather copiously-branched cells. In the palisade- tissue and spongy parenchyma chlorophyll-grains are pre- sent in abundance. The leaves are rather densely covered with hairs; the following types of hair occur: — (1) Non-glandular hairs, with one or many cells, thin- walled, glabrous and pointed (Fig. 19, C and E); (2) Glan- dular hairs with longer, two- or many-celled stalk and a 2—4 celled head (Fig. 19, A, В and D); (3) Glandular hairs with short, one-celled stalk and 2 or commonly 4-celled head (Fig. 19, F); (4) Glandular hairs of the kind illustrated in Fig. 19, G, Н, I and К with a low somewhat sunken stalk- cell, and a cupola-lıke extremely thin-walled 4-celled head. By comparision of the figures in question with Fig. 14: 420 Ев. J. MATHIESEN. С, D and Е, it becomes evident, that this form of glands in Bartschia — as already pointed out by HEINRICHER (cf. his addition to the cited paper of SCHERFFEL) — belongs to the Lathraea-type like the similar ones in Euphrasia. Fig. _19, I and К show these glandular hairs seen from above; Fig. 19. Bartschia alpina. Different types of non-glandular and glandular hairs from the leaf. (A—E, about %/,; F—K, about 79°/,.) below the stalk-cell the lateral walls of the epidermal cells are indicated by dots, in К two intercellular spaces are formed between them, 2—4 of these are found rather commonly under the hairs. A glandular hair, similar to that which is seen from above in К, is shown in H in transverse section, here also 2 intercellular spaces are distinctly seen. G and / have no intercellular spaces of this kind. Glandular hairs of type 4 occur only on the lower surface of the leaf, under the branches of the veins of higher order; in the same place there is found in addition to this a great number of hair Scrophulariaceae. 421 of type 3, but the latter kind of hair is also found on the upper surface of the leaf, located at the margin of the leaf in the innermost part of its indentations, i.e. above the extreme points of the veins (cf. Fig. 15, B). The non-glan- dular hairs and the long-stalked glandular hairs occur on both leaf-surfaces, and they grow most densely near or upon the veins. Upon the scale-leaves at the base of the shoots glandular hairs of the type shown in Fig. 19, F predominate, they are especially numerous on the lower surface, where the majo- rity of them are found arranged in two broad stripes, one along each of the margins of the leaf; mingled with them are found glandular hairs of type 4. The stems are covered with multicellular non-glandular hairs, and glandular hairs which frequently have multicellular stalks, Pedieularis lapponica L. Alcohol-material from Norway (Aursunden, leg. Tu. В. RESVOLL, 8. 8. 1918; Skadavara in West Finmark, leg. Е. WARMING, 7. 7. 1885). Herbarium-material from West and East Greenland, Lapponia murmanica, Nova Zembla, Labrador, Arctic Sibe- ria and the coast of Hudson Bay. Lit.: AXELL, 1869, р. 102; Lange, 1871, рр. 256 and 265, tab. III, fig. 16 (the seed); 1880, р. 74; 1887, р. 262; WARMING, 1886, pp. 44 and 47; 1888, pp. 35, 39, 59 and 87; 1890, pp. 207, 208, 210, 211 and 219; Linpman, 1887, р. 84; Rosenviner, 1892, р. 686; 1896 (1), p. 68; 1896 (II), р. 78; Hartz, 1894, pp. 37 and 39; 1895 (I), pp. 137, 175, 219, 271 and 288; 1895 (II), рр. 335, 359 and 372; Hartz and KRUUSE, 1911, pp. 347, 359, 365, 388, 409, 419, 423 and 428; Exstam, 1897, pp. 119, 167 and 168; Kruuse, 1898, pp. 373, 380 and 394; 1905, р. 175; Asromeıt, 1899, р. 42; Кмотн, 1899, р. 190; Dusen, 1901, р. 39; Norman, 1901, р. 457; Cıeve, 1901, pp. 9, 39, 57, 70, 11, 76 and 89; Зкоттзвевс, 1901, р. 8; Porsitp, 1902, pp. 114, 167, 175 and 187; 1910, рр. 267 and 271; 1912, рр. 382 and 387; 1920, 422 Fr. J. MATHIESEN. р. 142; Poprpıus, 1903, р. 48; SyLVÉN, 1905, р. 88; SILÉN, 1906, р. 92; Simmons, 1906, р. 34. Hemicrytophyte, with a primary root which dies away rather quickly; it spreads by means of slender, subterranean runners. As a rule, these are monopodial: their short ver- tical portion bears leaf-rosettes, usually through a limited number of growth-periods (1—3), the floral shoots arising as lateral axes on the rosette-axis. _ As shown in Fig. 20, A and В the horizontal portion of the runners has elongated internodes, with small scale- leaves; in the axils of these, buds are formed which grow out into new runners. New runners arise only in the hori- zontal portion of the parent-shoot. The runners are provided with adventitious roots, which must be described as “re- stricted to one position only”, inasmuch as they arise in connection with the axils of the scale-leaves, as a rule one at each. Haustoria were found on the roots. When the runner bends upwards, its internodes become short, and the scale-leaves pass without intermediate forms, into long-stalked foliage-leaves, of which it bears 3—7 in the first year (at x in Fig. 20, A and B the dead stalks of the rosette-leaves of the previous year are seen); at the end of the season of growth a number of scale-leaves are again formed, which function as bud-scales for the winter-bud. Some runners pass over directly into their winter-rest, with- out having succeeded in forming leaf-rosettes in the first year. Both the scale-leaves under the rosettes, and the rosette- leaves, as well as the scales of the winter-bud, can subtend buds (Fig. 20, A, B and D); the lowermost buds must prob- ably be regarded as reserve-buds; any true principal bud does not occur. At the beginning of the next season of growth, one or several — as many as 3 — of the buds can grow out into Scrophulariaceae. 423 Fig. 20. Pedicularıs lapponica. A, A part of a flowering plant. On the runner two new runners (a and 6) have developed as lateral shoots, and an erect, floral shoot, of which only the lower half has been figured; 7, adventitious roots. The uppermost of the two buds seen just below the floral shoot is the terminal bud of the runner. The runner a is in its second year of growth; the withered remains of the rosette-leaves (x) of the previous year are seen under the scales of the winter-bud. B, The apex of a runner; under the leaf-rosette one sees the scales of the winter-bud, — which scales subtend small buds, — and the withered remains of the foliage- leaves (x) of the previous year; a floral lateral shoot has been devel- oped. C and D, Apices of two runners; C with dead, terminal bud 494 Ев. J. MATHIESEN. and two lateral shoots (floral), D with a floral lateral shoot and four reserve-buds. E, One of the leaves of а floral shoot. (A and E, Green- land.) A and В about nat. size; С and D about 4/,; Е about 3/1. (В, С and D are drawn from sketches by Е. WARMING.) floral shoots; as shown in the figures, they bear scale-leaves (ex parte bud-scales) at their base; these pass gradually into foliage-leaves in the middle of the shoot. The foliage-leaves of the floral axes differ somewhat in form from the rosette- leaves; the midrib is broader, and the incisions of the margin are less deep (Fig. 20, Е). The inflorescence is a crowded, almost capitulate raceme, without a terminal flower; accord- ing to WARMING, the flowers in the middle of the inflores- cence, are the first to expand. The terminal bud of the run- ner, as a rule, continues its growth by forming a second leaf- rosette; exceptionally, it may produce a floral shoot, which consequently becomes terminal; the latter has only scale- leaves at its base. In addition I must remark, that in a single case, I found a runner which, without a previous vegetative-stage (leaf-rosette), terminated directly in an in- florescence. A runner hardly ever forms leaf-rosettes for more than 3 successive growth-periods. Also in the axils of the scale- leaves of the floral shoots buds occur; in cases in which these develop further, they frequently form, in the following year, shoots of similar structure to the parent-shoot. WARMING mentions that at the base of the floral shoots also, rosettes of foliage-leaves may be produced; in some cases I suc- ceeded in demonstrating such leaf-rosettes, but they are evidently of rare occurrence. The structure and biology of the flower are explained by Lınpman and Е. Warming, of whose descriptions the following is an extract: The lower lip of the almost horizont- ally projecting flower incline very decidedly to the right Scrophulariaceae. 425 (seen from the front); besides this, the whole flower is subject to torsion around its axis to the right; this causes the upper lip to slope even more to the right, while its helmet-shaped apex, and the part of the style turned downwards, point to the left; the lower lip stands almost in the vertical plane (Linpman, р. 84, Pl. IV, fig. 47, A and В). _ “The calyx is dark-red, the corolla pale yellow or brim- stone-coloured. On the lower lip there are two conspicuous convexities covered with hairs, the corolla-tube is 6—8 mm long” ... “The slit of the upper lip is about 1—11/, mm wide; the front stamens are slightly hairy or glabrous. Fig. 21. Pedicularis lapponica. The style may be enclosed A flower with a far-protruding style. (and this is the case at the (Greenland, about 67° N. lat.) (Е. WARMING, 1890.) same time as the anthers are open), but usually it protrudes more or less, from 1.5 to 2.5 mm or even more (without being moved away from its natural position up to the dorsal side of the upper lip). Even in the bud it protrudes somewhat, so that the stigma comes into contact with the lower Пр; the former. is even at that time papillose, and appears ripe, consequently, slight protogyny appears to prevail. Afterwards the anthers are seen to have dehisced, without the length of the style having altered; consequently, self-pollination appears to be able to take place, but not easily, because although it is true that the anthers come to stand vertically above the stigma, yet the distance between them is comparatively © great”. (WARMING, р. 219, fig. 33, reproduced here as Fig. 21). All authors join in attributing perfume to the flowers of P. lapponica; LiNDMAN mentions its “strong perfume’’, 426 Ев. J. MATHIESEN. and Hartz (1895 (I), р. 271) writes, that P. lapponica and Viscaria filled the air with perfume (East Greenland, 7.7.92). Pollination takes place by the agency of humble-bees, which, according to LiINDMAN, alight on the upwardly turned, left side of the lower lip; the stigma is consequently, by the torsion described above, turned towards the visitor. Ac- cording to the same author, the humble-bee thrusts its pro- boscis into the flower, at the widest part of the slit of the upper lip, at about the middle of it, just above the revolute portion of the edge of the upper lip, but not into the narrow groove between the two convexities of the lower lip. Several of the specimens from West Greenland had one- year-old fruits, from East Greenland I saw fruit-bearing specimens from Danmarks © (N. Hartz leg.). According to Ровзию (in litt.) visits of humble-bees are rare, and fruit-setting (in contradistinction to the other Pedi- cularis spp.) 18 generally not good; many of the capsules are empty, and in others only a few seeds are found. EKSTAM's observations from Nova Zembla agree with those of Ровзи. A description and beautiful illustration (PI. III, Fig. 16) of the seed is found in LANGE, 1870. P. Lapponica is a decidedly middle-summer-flowering species. Geographical Distribution: Greenland, Arctic Ame- rica, Labrador, Kamschatka, Arctic Siberia, Nova Zembla, Arctic Russia, Scandinavia and Lapland. RosEnvinGe (1892) records, that in West Greenland it is rather common north of 64° N. lat., but very rare south of this. “The northern limit in West Greenland is still unknown, in the fjords at 72°23’ it was so commonly distributed that the northern limit hardly can be here” (Ровзи, 1920). In East Green- land the species is only found between 69°25’ and 73°10’ М. lat. (Квоозе, 1905, р. 175). Scrophulariaceae. 427 Habitat: In Greenland it grows in mossy bogs, and in damp places on heaths; the runners creep in the moss, or in slightly decomposed vegetable matter (M. PorsiLp, in litt.); it is “covered by thick layers of snow during winter” (PorsiLp, 1920). WARMING, 1888, mentions it as inhabiting willow copses, herb-slopes and heaths. On Nova Zembla Ex- STAM found it growing in dry localities, which were exposed to strong insolation. A. CLEVE finds it on mountain heights in Lule Lappmark, avoiding wet ground and the poorest heath; she writes: “Sie gedeiht am besten in den Waldregionen und sucht im Hochgebirge relativ trockene Stand- orte auf, was schon Wahlenberg aufgefallen war.” An о RU AS The SZ Anatomy. Root: The structure is shown in the transverse section figured in Fig. 22. The secondary growth Fig. 22. Pedicularis lapponica. Transverse section of an adventitious is not considerable, and por on), does not greatly exceed what is shown in the figure in question. The endodermis is а very beautifully developed Casparian sheath. The epi- dermis dies away quickly; the outer walls of the exodermis were found to be cuticularised. On the surface of the roots, dark-coloured, fungal hyphæ were frequently found. The Stem: (1) Runners: The epidermal cells have fairly thick outer walls; they are filled with а homogeneous brown mass, which gave the reaction of tannin. The outer- most cortical layer adjoins the epidermis without inter- 428 Ев. J. MATHIESEN. cellular spaces; the cells of the cortex show radial divisions by thin walls of secondary formation. The endodermis is exactly as in the root. The pericycle is 4—7 layered, its cells, like those of the cortex, have somewhat thickened walls. The wood-cylinder is continuous, without parenchyma- rays, but is often somewhat excentrically developed, so that the wood on that side of the shoots which is turned down- ) 1 GA eG Seok BD DE DES FAZIT DATE Bas es (KB SDD0S20000 2% 0x OK, Bae Eee SØGE LOC) ыы @ CORDES OO 000. SO) RS Ç ER, LS ARCS EN: (2 SI + Fig. 23. Pedicularis lapponica. A, Portion of transverse section of runner. B, Portion of transverse section of flower-bearing stem. (Greenland.) (About 14°/,.) wards, is thinner than on the upturned side. As shown in Fig. 23, A, annual rings can be formed in the runners. The two annual rings are separated by thin-walled parenchyma, mixed with small vessels, and the vessels are much larger in the second than in the first annual ring. Some stereom (wood-fibres) is seen, especially in the inner annual ring. (2) The Flower-bearing stem (Fig. 23, В) differs from the runners in that the cells of its cortex and of its only 1—2 layered pericycle, are much thinner-walled than in the runners, moreover, by the Casparian spots in the endodermis being much fainter, by the narrower vessels of the xylem- Scrophulariaceae. 429 ring, and the almost continuous ring of wood-fibres on the outer side of the latter, and by the outer layers of cells of the pith being thickened and lignified. The central part of the pith has died away, and the stem is consequently hol- low. Stomata occur in the thick-walled epidermis, and chloro- A oa Ste Е | я м bd à Fig. 24. Pedicularis lapponica. A, Transverse section of the leaf. B, Epidermis of the upper, and C of the lower surface of the leaf. (Greenland.) (А about 119]; В and С about ?220/..) phyll in the cortical cells. In the cases in which the basal portions of the flowering-shoots persist and bear lateral shoots, a second growth-ring —often excentric— may be formed here. The Leaf: As regards the anatomy, there appears to be very little difference between the rosette-leaves and the leaves of the floral shoots. The epidermis of the upper sur- face (Fig. 24, B) has straight to slightly undulating lateral XXXVII. 28 430 Ев. J. MATHIESEN. walls; those of the epidermis of the lower surface (Fig. 24, С) are more strongly undulating. The lateral walls of the epi- dermal cells of both the upper and lower surface, although thin, are yet very distinctly porose, those of the upper sur- face, more strongly so. Stomata occur only upon the lower surface, the guard-cells are surrounded by 4—6 cells. In trans- verse section (Fig. 24, A) 1—2 layers of palisade-cells are seen; the spongy parenchyma is composed of highly-branched ie i © Sg, cells (the small circles inside the cells of the spongy paren- chyma indicate the branches ; Cu hdl (2 | FS P Le” ve, of Е Е < which have been cut through N RUN а N S P Fo Ge on preparing the section); N Re à и с | CEE 7 chlorophyll-grains are present abundantly in the whole of Fig. 25. Pedicularis lapponica. A portion of a leaf-section seen from the lower surface; the dermis of the lower surface; glandular hairs are seen to be the stomata occur on a level crowded under the veins (about 35). with, or are raised slightly the mesophyll, also in the epi- above, the leaf-surface. The final branchings of the veins end in an epithema-like tissue of somewhat inflated, thin-walled tracheids. Very characteristic of the leaves of the Pedicularis spp. are the numerous glandular hairs of the type shown in Fig. 24, A and C. Fig. 25 shows a portion of a leaf-section seen from the lower surface, — the glandular hairs are found here only —, the veins are drawn as if visible through the mesophyll, and the glandular hairs are seen to be crowded under the veins, the same feature is recorded by HovELAcQUE (1. с.) in the case of other species of Pedicularis. In Fig. 24, С three glandular hairs are seen from above; the epidermal Scrophulariaceae. 431 cells from which they have developed, have far thicker and more highly porose lateral walls than the surrounding cells; the finely-drawn circle which surrounds the thickly-drawn glandular-head, indicates the place of contact between the basal cell of the hair, and the outer wall of the epidermal cell; the smaller circle at the top of the head indicates that the cuticle at this place is lacking, thus forming a pore only spanned by the cellulose wall, as described by Percy Своом for P. palustris. Pedicularis sudetica Willd. Herbarium-material from the shore of Hudson Bay (Churchill), King Point, King William’s Land, Southhampton Island, Port Clarence, Taimyr Peninsula, Chabarowa, Nova Zembla and Lapland. Lit.: LANGE, 1880, р. 75; Кзегтмам, 1882 (I), р. 256; 1882 (II), р. 325; 1882 (III), pp. 361, 363 and 366; 1882 (IV), р. 509; K3ELLMAN and Lunpström, 1882, р. 303; Hozm, 1885, р. 43, tab. VI, fig. 10, tab. УП, figs. 1—7; WARMING, 1890, pp. 207 and 215; Exstam, 1897, pp. 118, 166, 168, 176 and 180; Кмотн, 1899, р. 191; Simmons, 1913, pp. 123 and 138. Sympodial hemicryptophyte of semi or entirely rosette type, the above-ground part of the shoot bearing the in- florescence being either furnished with a few foliage-leaves, or being totally devoid of them. The species may form tufts, (a specimen from Herschell Island belonging to forma lanata Walpers, had numerous rosettes of foliage-leaves, even seven flowering shoots and 10 withered peduncles from previous years) and in this case it has a short mesocorme (“rhizoma multiceps’’), a fairly thick primary root, sometimes probably of long duration, adventitious roots being only few in number, but rather vigorous; more frequently it appears, however, that the basal portion of the main shoot dies away, perhaps even 28* 432 Ев. J. MATHIESEN. after the first flowering period; the main root is then less vigorously developed, and slender adventitious roots, also from the rhizome-portions of the innovation shoots, are abundantly developed, a circumstance which permits those shoots to continue their existence as independent individuals, after iso- lation from the main shoot — consequently, the species can be propagated vegetatively (cf. ЕкзтАм). This difference in the growth is probably conditioned by the nature of the habitat; the tufted individuals must be assumed to have grown on drier soil, while the other form must have lived in damper localities; the fact of my having found Sphagnum in connection with the rhizomes of many specimens of the latter type, indicates moist moss- tufts as their habitats. As regards the duration of the first vegetative stage, it is not possible for me to state anything with certainty, owing to want of material; in a comparatively weakly-developed individual, I estimated its extent as three years. Fig. 26, 7 shows an individual produced from an iso- lated innovation shoot; (the figure has been drawn from soaked herbarium-material, and is somewhat diagrammatic). At the base of the relatively-main axis, we see the separation surface (scar) along which it has loosened itself from the parent-shoot; during its first year, the shoot has formed a few-leaved rosette of foliage-leaves, and completed the year’s growth by the formation of scale-leaves, for the protection of the winter-bud; (at / some fragments of the first year’s leaves are seen; the bud A is dead). Next year a piece of stem, about 1 ст long and only bearing 2 leaves, was first produced, next a rosette of foliage-leaves, followed in turn by scale-leaves, of which three still remain (two only can be seen). The process is repeated in the following (3rd) year, and simultaneously the buds C and D in the axils of the Scrophulariaceae. 433 scale-leaves of the 2nd year, expand and produce a few foliage- leaves, some of which still remain in a withered condition (с, and d, respectively). In the fourth year the individual at last flowered; /V indicates the lower portion of the pedun- cle. С and D which, like the terminal bud of the shoot, Fig. 26. Pedicularis sudetica. 1, Ваза! portion of а plant, somewhat diagrammatic; as regards the explanation of letters and numbers cf. text оп pp. 482—434. 2, A flower seen in side view. 3, The upper lip seen in side view; the reciprocal position of the anthers and pistil is indicated. 4, Stigma, more highly magnified. (Nova Zembla.) (1 about nat. size; 2 and 3 about °/ı; 4 about 50/,.) have been protected during the winter-rest by scale-leaves, are forming their second leaf-rosette, the leaves of which are indicated by c, and d, respectively; and B, which rested during the previous summer, has now developed а foliage-leaf, 65; 434 Ев. J. MATHIESEN. the now dead leaves x and y are from the leaf-rosette formed the year before the main shoot flowered; from the ах! of one of the uppermost bud-scales, an innovation shoot E has also been developed, and is expanding its first two foliage-leaves. Here, consequently, the vegetative-stage has lasted 3 years, and this appears to be the rule in individuals of this type; in the tuft-forming type in which, in dried material, this question is far more difficult to decide, it appears to me that the vegetative stage can be restricted to two years. It still remains to be noted that, although the majority of the individuals which result from isolated innovation shoots, had elongated internodes inserted between the ro- settes, yet only rarely were they as long as is shown in Fig. 26, 1; the phenomenon is probably due to the endeavour of the plant to keep its terminal bud in а certain position rela- tive to the surface of the substratum. In Fig. 26, 1 adventi- tious roots are developed only at the base of the shoot, but these may also be developed below the other rosettes, often in great numbers. They are 1—2 mm thick. Г never found haustoria on such adventitious roots, but they occurred on the root-branches in the specimen from Herschell Island, mentioned above. From Nova Zembla plants were to hand of which the peduncle was only 5 em high during flowering; in specimens from Chabarowa it was 17cm, and in one from Churchill, even 25cm high; after flowering an elongation of the axis of the inflorescence takes place. Fig. 26, 2 shows a flower seen from the side. The posi- tion of the stigma should be noted, and the two pointed lobes in which the front part of the helmet terminates, and the ridge, with the veins which proceed from it into the helmet, also the slightly warty revolute portion of the edge of the upper lip, and the two convexities on the lower lip; Scrophulariaceae. 435 as stated by Е. WARMING, the latter is somewhat oblique. The calyx is drawn as densely hairy, this is the case in forma lanata Walpers. In Fig. 26, 3 the position of the stamens inside the helmet is shown. It must be borne in mind that the natural position of the flower is almost horizontal, whence it follows that self-pollination must be able to take place with comparative ease, for the pollen will be sprinkled out of the front opening of the helmet by accidental movements of the plant, caused for instance by the wind, and will fall upon the hinder side of the stigma, which, as shown in Fig. 26, 4 is almost globular. Exstam has studied the species in Nova Zembla, and he writes concerning this: “Ziemlich häufig — häufig an beson- ders nassen Orten. Besitzt einen ziemlich starken jasmin- ähnlichen Wohlgeruch und hell- bis dunkelrote Blüten. Schon in sehr jungen Knospen ragt die Narbe aus der Ober- lippe hervor — in einer ganz aufgeblühten Blume etwa 21/,—3 mm — und wird bald papillös, gewöhnlich bevor die Blume ganz aufgeblüht ist, und die Antheren sich geöffnet haben. Selbstbestäubung dürfte dadurch ermöglicht werden, dass die kugeliche Narbe rund umher papillös ist, und in der Falllinie der Pollenkörner gelegen ist. Nachdem die An- theren sich entstaubt haben, bleibt die Narbe noch eine Zeit lang glänzend, papillés. Blüth im Hochsommer. Mit reifen Früchten beobachtet. Anemophile Samenverbreitung. .... Wurde am %1/, von einer kleinen Fliege, am 6/; von einem grossen Bombus hyperboreus Schönh., am 18/, ebenfalls von einem grossen Bombus hyperboreus Schönh. besucht. Im Sommer 1891 trotz fleissiger Beobachtung kein einziger In- sektenbesuch bemerkt.” In somewhat vigorous specimens, the exceedingly dense inflorescence, with the deep-red corollas, cannot fail to be very conspicuous. 436 Ев. J. MATHIESEN. The Geographical Distribution, according to Sim- mons and LANGE: Western Arctic and Subarctic North Ame- rica, (both the continent and the archipelago), Arctic Siberia, Nova Zembla, Arctic Russia and besides on mountains TY 2 Fig. 27. Pedicularis sudetica. A and B, Portions of transverse sections of the rhizome of the individ- ual illustrated in fig 26, 7 (cf. text). С, Epidermis and the outer layers of the cortex of the same rhizome. D, Transverse section of the leaf. (Nova Zembla.) A and B about 14°/,; C about 7/,; D about au) (Sudeten and Riesengebirge) of Central Europe. Does not occur in Greenland. According to KJELLMAN (1882 (III)) the flowering year’s- shoots of P. sudetica in a pressed and fermented condition, and the roots, rhizomes and buds in fresh condition, are used as food by the Chuckes in Eastern Arctie Siberia. Anatomy. An adventitious root about 2 mm thick Scrophulariaceae. 437 showed the following structure: Of the exceedingly thin- walled epidermis only scattered fragments were left, the cells were dead and had collapsed. The outermost layer of the cortex had cuticularised outer and lateral walls (it must certainly be this layer which Tu. Нотм in his Fig. 7, Tab. ПТ, designates epidermis); the layer just within the outermost layer adjoins the latter without or with small intercellular spaces between them; the endodermis has distinct Casparian dots. The intervening part of the cortex consists of a very lacunose tissue, which corresponds exactly with that in the rhizome (shown in Fig. 27, C). Such slender adventitious roots complete their development during one season of growth; their structure is 2-rayed, and by the secondary growth there is formed a bundle of xylem-tissue, about 0.5 mm thick consisting of vessels with intervening non-lignified tissue. Of the individual figured in Fig. 26, 1 a transverse section of the axis was taken from the part between / and //, and also one between JJ and III; the former part must be assumed to be 3 years old, and the latter 2 years. In the part between J and //, there was found in the stele a circle of vascular bundles, separated by broad, non-lignified, parenchyma-rays (cf. Fig. 27, A, which shows the xylem part of a vascular bundle). In the second year of growth an increase of the vascular elements has evidently taken place, the lowermost (1. e. in the stem inner- most) group of wide vessels probably indicates the limit of the first year’s growth; what is found outside that belongs then to a later (probably the second) year. Wood-fibres are quite absent; here the pith does not die away. Fig. 27, B is drawn from a section taken from the part between // and III, at about the middle of it. Here the axis is hollow, because the central part of the pith, as in the peduncle, has died away; the walls of the outermost layers of the pith have 438 Ев. J. MATHIESEN. become thickened, lignified and porose, the same is seen to be the case as regards the inner part of the parenchyma- rays, which separate the vascular bundles. Around the ves- sels non-lignified elements occur, but the limit of the first- year’s growth is shown in the vascular bundles by a layer of wood-fibres, 2—3 cells thick, which adjoins the lignified parenchyma rays, so that a continuous stereom results. The few scattered vessels outside this I regard as the growth of the second year; such small groups of vessels may also occur in the space between any two of the original vascular bundles. A second individual exhibited a similar structure, but in a third I found no trace of a second-year’s growth; unfortu- nately the scantiness of the material prevented my studying this point more thoroughly. A portion of the cortex is shown in Fig. 27, C; with the exception of the two outermost layers it is very lacunose in structure. The epidermis is small- celled, and furnished with a rather thick euticle. The Leaf. The epidermis is of the usual Pedicularis- type, the lateral walls of the cells are thin and only slightly porose, very similar to those in P. hirsuta and lanata. Sto- mata occur only on the lower surface of the leaf; the guard- cells are on a level with the other epidermal cells. A trans- verse section is shown in Fig. 27, D; the part which has been illustrated is typical and also agrees well with Tu. Hozw’s Fig. 4, Tab. VII. The upper half of the mesophyll consists of 2—8 layers of short, thick palisade-cells, and the lower half of spongy parenchyma composed of rather deci- dedly branched cells. Chlorophyll occurs abundantly in the whole of the mesophyll. Glandular hairs of the same form and size as in the other species of Pedicularis, occur also in P. sudetica on the lower surface of the leaf — not as TH. Horm (l.c. p. 44) writes “here and there,” but in abun- dance below the branches of the veins of tertiary and higher Scrophulariaceae. 439 order. On the lower surface of the leaf, and especially in the middle of the sections, thin-walled 1- to 3-celled non- glandular hairs occur. Pedicularis euphrasioides Steph. Alcohol-material from Greenland (Itivnek near Hol- steinsborg, leg. Euc. WARMING, 13. 7. 1884). Herbarium-material from Greenland, the shore of Hud- son Bay (Churchill), Port Clarence near Behrings Strait, and Siberia (by the Boganida River). Lit.: LANGE, 1871, рр. 254 et seq, 264, tab. II, fig. 10 (the seed); 1880, р. 74; 1887, р. 262; WARMING, 1886, pp. 19, 44 and 58; 1890, рр. 206, 210, 211 and 218; RosEenvinGe, 1892, р. 686; 1896 (I), р. 68; 1896 (II), р. 78; Harrz, 1894, р. 39; Кмотн, 1899, р. 190; PorsiLn, 1920 p. 142. Spot-bound hemicryptophyte of the semi-rosette type, biennial to pluriennial hapaxanth, or — what is certainly the most common — а pollacanth, with scantily-branched main root, the longevity of which corresponds with that of the individual. Should the species appear as a pollacanth, innovation shoots are developed at the base of the flowering shoot. The length of the first vegetative-stage varies from one to (probably at most) two years. During the vegetative stage a leaf-rosette is developed every summer and — towards the end of the growth-period — bud-scales for the protection of the young organs during the winter-rest; in the summer in which the plant flowers, the shoot forms no rosette of foliage-leaves. The shorter vegetative-stage (of one-year’s duration) appears especially to result in weaker individuals, which probably often end their lives by fruit-setting in the second year. This species differs very much in habit from the other species deseribed here, owing to the formation of numerous 440 Ев. J. MATHIESEN. branches which, during the flowering year, arise in the axils of the foliage-leaves on that part of the stem which has elong- ated internodes. These lateral shoots increase in length down- wards to some way below the middle of the parent-shoot, and then again become shorter towards its base; here they are often purely vegetative, while the upper ones bear, in addition to 2—3 decussated pairs of foliage-leaves, a small inflorescence corresponding with that of the parent-shoot. In the pollacanthous individuals the innovation shoots are produced in the axils of the uppermost bud-scales; they usually develop only a few pairs of opposite foliage-leaves, then bud-scales, and then the following year, they most often terminate their growth by the formation of an in- florescence. I cannot deny that there is a possibility of the innovation shoots also flowering without any preceding rosette-formation, and therefore, in their first year, develop- ing only bud-scales; in this case the species would be a transitional form between the semi-rosette type and the protohemicryptophyte-type. In individuals in which а “rhizoma multiceps” is devel- oped, the uppermost part of the main root may become as much as 2 cm thick. The flower-bearing stem may become as much as 25 ет high; after flowering, the axis of the inflorescence is some- what elongated. On the basis of material from Greenland (about 67° N. lat.) Е. WARMING gives the following description of the structure of the flower (ef. Fig. 28): “This species, which is in no wise common in Greenland, is very conspicuous owing to its numerous inflorescences with yellow, fragant flowers; the tip and the helmet of the yellow corolla is slightly brown- ish. The flowers are very oblique, the lower lip being highly oblique (Fig. 28, A, B, D and H); but the upper lip also is Scrophulariaceae. 441 twisted in such a manner that its two-toothed tip turns somewhat to the right (Æ). The ridge reaches the edge slightly above the beginning of its revolute portion (A and G). The slit between the two edges of the upper lip is usually rather narrow. The front filaments are sparingly hairy. The stigma protrudes more or less above the two teeth (В, D, Е and Н). Fig. 28. Pedicularis euphrasioides. (From Greenland; about 67° N. lat.) А, В, А flower seen from the left and the right side; in À the position of the style and the stamen is indicated. С, Stigmatic papillæ. D, A flower in front view. Е, The upper lip seen in front view. F, The ovary. G, The upper lip seen in side view; the venation is indicated. H shows the same аз A. Г, Transverse section of ovary. (Drawn by Е. WARMING, 1890.) Judging from the structure of this almost horizontally- placed flower (A and B), I must assume that it is only with difficulty that it pollinates itself. It produces ripe fruit in Greenland at 67° N. lat.” Geographical Distribution. West Greenland be- tween 62° (but only in a few localities south of 64° N. lat.) 442 Ев. J. MATHIESEN. and 69°30’ N. lat. (RosENvINGE, 1892), оп the whole rather rare there; Labrador, the Rocky Mountains, the continent of Arctic America, (but not mentioned by Simmons from the Archipelago), Siberia and Dahuria. In Greenland it grows on luxuriant, damp heaths; the species requires a snow-covering during its winter-rest (M. Ровзил, 1920). Anatomy. The root is of the type described later on in Р. hirsuta and lanata and, as in these, the secondary cortex is traversed by numerous radial clefts. There is, how- ever, a difference between P. euphrasioides and the two spe- cies in question as regards the nature of the xylem-part, because, in the two former it is composed of a large quantity of non-lignified tissue and of relatively few vessels only, whilst in P. euphrasioides one always finds in the root several concentric rings, each consisting of numerous vessels with intervening Попе stereom (thin-walled wood-fibres). These continuous rings of lignified elements, the outlines of which are however somewhat undulating, are separated by portions in which the vessels are not so wide or numerous, and are surrounded by non-lignified tissue; here clefts occur between the radiating lines of vessels. That these rings of vessels and wood-fibres really in- dicate zones of annual growths, is evident from the investiga- tion of some specimens chosen from among the individuals in my material, which had flowered for the first time: such as — judging from the leaf-scars — had had only a vegeta- tive-stage of one-year’s duration (that is to say, flowered in the second year) showed two rings; others which, as far as I could see, had had a vegetative-stage of two-year’s duration, showed three rings. The greatest number I suc- ceeded in demonstrating amounted to three. The cavity of the stem is comparatively large. The Scrophulariaceae. 443 cortex is thin; in the xylem-ring the wood-fibres dominate in the 3—4 outermost layers; the cell-layers in the periphery of the pith are lignified, thickened and porose. On the epi- dermis there are found non-glandular hairs of the same type as on the leaves. The Leaf. A transverse section of one of the leaves Fig. 29. Pedicularis euphrasioides. A, Epidermis of the upper, and B of the lower surface of the leaf. C, Transverse section of the leaf. D and E, Non-glandular hairs from the surface of the leaf. (Greenland.) (A and В about ?*/,; С about 10}: D and.E about, 19.) from the middle of the stem, is shown in Fig. 29, C. There is generally found only one layer of rather elongated pali- sade-cells, below these there is usually a layer of cells which, although elongated, are nevertheless branched (this is di- stinctly seen to the left in Fig. 29, C), and which must there- fore most naturally be included in the spongy parenchyma; the lower layers of the latter consists of richly-branched cells. The epidermal cells of the upper surface has thin, slightly 444 Ев. J. MATHIESEN. undulating lateral walls, those of the lower surface are still more thinwalled, their lateral walls more highly undulating and, as in the epidermis of the upper surface, furnished with flange-like thickenings, which I found to be particularly well- developed in the bracts of the flowers, and, on the whole, in the upper leaves on the axis. Stomata occur only on the lower surface; the guard-cells are on a level with the other cells of the epidermis; they are surrounded by 3—6 cells. Chlorophyll is abundantly present in the whole of the meso- phyll and is also found in the epidermis of the lower surface. On the lower surface there are glandular hairs of the usual kind, and with the typical distribution; and besides there are non-glandular hairs both on the upper and the lower surface (Fig. 29, D and E); they are multicellular and thick- walled, often however with one or two cells unthickened and therefore collapsed. Pedieularis Sceptrum carolinum L. Alcohol material from Aursundsöen, Norway (leg. Tu. Resvort, 11. 8. 1918). — Herbarium material from the district near the river Boganida, Alten, Mortensnæs in East Finmark, and the Murman Coast. Lit.: Lance, 1871, pp. 254 et seq., 267, tab. III, fig. 25 (the seed); WARMING, 1890, р. 215; Norman, 1895, р. 456; Kerner, 1898, Bd. П, р. 655; Кмотя, 1899, р. 191; Sire," 1900 р. 32 А sympodial hemicryptophyte of rosette- or semi- rosette type, with vertical or obliquely-ascending mesocorme, which dies away behind, and from which adventitious roots, 1—2 mm thick, are abundantly developed. 2-year-old specimens, from Aursundsöen, showed that the plant in its first year forms a few-leaved rosette; the winter-bud is protected by scale-leaves, and the following year the growth is continued by the formation of a more Scrophulariaceae. 445 luxuriant rosette. [ can state nothing as regards the duration of the first vegetative-stage, the 2-year-old plants in question were still rather frail. The hypocotyl was somewhat elon- gated, the adventitious roots had already begun to develop, and in one of the individuals, a lateral shoot had been pro- duced in the axil of a rosette-leaf of the first year. On therhizomes, innovation-shoots, from 1to 3 in number, occur; their vegetative stage, judging from the material at my disposal, which is however but scanty, extends over 2 years: During the first year 2—4 foliage-leaves are developed, and then scale-leaves, which serve as bud-scales for the winter- bud; during the next year a somewhat more dense rosette is developed, again succeeded by scale-leaves; and finally the shoot, during the 3rd year, after having formed a lux- uriant rosette, terminates in an aerial, flower-bearing portion, which is often quite destitute of foliage-leaves, and may attain a height of even 60 cm. The fact that the rhizome dies away behind, in con- junction with the abundant development of adventitious roots, must allow of vegetative propagation, since the lateral shoots gradually become free and independent by the dying away of the parent rhizome. Е. WARMING (1. с. р. 215 et seq, figs. 30 and 31) gives the following description of the structure and biology of the flower, on the basis of material from Kola, and from Oster- dalen in Norway; his two groups of figures are reproduced here as Fig. 30 and Fig. 31. “The flowers are . . . . as much as 32 mm long; they stand erect, the subtending bracts are closely adpressed, and in addition each flower appears to be completely closed. The lower lip (Fig. Æ) is erect, 14—15 mm long, pressed closely to the upper lip, which is a little longer (16 mm), and closes the throat completely, but is easily pressed down. The visit- XXXVI. 29 446 Ев. J. MATHIESEN. ing insects must evidently be large and vigorous, such as humble bees and moths, and they must, in а way similar to that in Antirrkinum and Linaria, press their body or head down between the two lips, by which means they primarily come into contact with the stigma; Fig. 30, B shows (the lower lip having been removed) that the stigma can be so protruding that contact with it must be regarded as inevit- Fig. 30. Pedicularis Sceptrum carolinum. From Kola near the White Sea (collected by Dr. BroTHERUS). A, Flower, natural size and position. B, The same after the lower lip has been removed (D belongs. to this). С, Part of a 25 mm long flower, seen in front view; the stigma protrudes less than it does for instance in B; the anthers have not yet opened (see F); to this belongs the tip of the style shown in H. Е, The lower lip of В. С, Ovary. v indi- cates the revolute portion of the edge of the upper lip. (В. WARMING, 1890.) able; that cross-pollination must take place by the agency of large insects, is also evident from the fact that the distance from the stigma to the nectary may be as much as 2.5 cm. I have also several times seen the stamens torn off the upper lip, evidently by the violence of a large insect. In one of the flowers (C) which has been illustrated, the anthers had Scrophulariaceae. 447 not yet opened (F), and as the stigma appears to be ripe (H), protogyny must consequently prevail. When the anthers have afterwards opened (D) it must evidently be the insect knocking against the filaments, which shakes the pollen down over it; strange to say, both the fila- ments, anthers, style as well as the revolute portions of the edge of the upper lip, are quite glabrous (see B, C and D) so that there is no special contri- vance which by its resistance increases the shock. On the other hand, hairs occur — which are probably of im- portance in preventing the pollen-grains from falling out at the sides — viz., along the edges of the upper part of the upper lip (B, C, D), consequently, in the same place as ordinarily, but on another organ. For the rest, the width of the slit of the upper lip, and the degree to which the stigma protrudes, differ also. While the stigma protrudes far out in the quite young flower B, it does not pro- trude so far in the young C, and still less in Fig. 31 (also in this paper Fig. 31) of which the anthers are already open .... Self-pollination appears to be able to take place only with diffi- culty. As the species sets fruit abun- Fig. 31. Pedicularis Sceptrum carolinum. From Kola. The upper lip in front and side view. The anthers are open. In the front there is a flat or incurved part, and the slit is tightly closed above. In this case the stigma does not protrude, but stands just within the uppermost end of the slit. (E.WARMING, 1890.) dantly in so northerly a region as near Alten (about 70° N. lat.) it must undoubtedly be humble-bees that effect pollina- tion.” Conditions pertaining to the pollination of the flower 29+ 448 Ев. J. MATHIESEN. have afterwards been studied by SILÉN, who writes: “Were much-visited by humble-bees, which forced themselves into the closed flowers, which were thereby torn open on the one side, seen from the front, the right side, whereby the lobes of the lower lip are bent outwards. After the visit of а humble-bee, the flower cannot close again. The smaller humble-bees penetrate entirely into the flowers, whilst the stomach of the larger ones remains visible outside the flower.” In flowers from Aursundsöen I found the lower lip to be relatively somewhat longer than is figured by WARMING; the length of the style was, on the whole, the same as that shown in Fig. 30, B. Many of the older flowers in the in- florescences had been opened by the visits of insects. GeographicalDistribution: Fennoscandia, Lapland, Western and Northern Russia, Siberia as far eastward as to the river Kolyma, Mantchooria and Japan. Besides this, the species has an area of distribution in Central Europe (Germany, with a southern limit in the Bavarian Alps); here it is possibly a glacial relict (KERNER). According to Norman, in Northern Norway the species grows in bogs, on boggy plains, on the coast and in willow copses. Anatomy. The Root. The epidermis remains long intact. The cells in the outermost layer of the cortex are provided with a cuticularised lamella along the whole of their circumference; between this exodermis and epidermis there are no intercellular spaces, and such are wanting also between the exodermis and the outermost layer but one of the cortex; further inwards, the cortex becomes rather lacun- ose. Starch occurs in the cortical cells. The endodermis is a typical Casparian sheath. When the secondary growth in the stele is completed, a small central group of vessels, cir- cular in outline, is found to be developed. Scrophulariaceae. 449 The Peduncle. A portion of a transverse section is shown in Fig. 32, A. The layer of complete cells seen upper- most in the figure, is the endodermis. The pericycle is several-. ау ered; sieve-tissue occurs rather sparsely. The vigorous development of the stereom of the peduncle, is characteristic Fig. 32. Pedicularis Sceptrum carolinum. А, Portion of a transverse section of peduncle. В, Transverse section, and C longitudinal section of the upper part of a rhizome; B and С diagrammatic, cf. text. (Aursundsöen, Norway.) (A about 22/,; В and С about $6/;.) of the species. In the mass of wood-fibres only here and there a radial row of vessels is inserted — in the figure the 4th row from the right hand is such a radial row. In between the groups of the primarily formed vessels, of which one is seen in the middle of the figure, the wood-fibres internally adjoins the peripheral portion of the pith, which also consists of strongly thickened, lignified and porose cells, so that in the fully developed stem the original groups of vessels are 450 Ев. J. MATHIESEN. found completely surrounded by a compact mass of stereom; only the tissue immediately surrounding the protohadrome remains unthickened and unlignified. The anatomical structure in that portion of the stem which bears the rosette-leaves deserves а more special de- scription. A transverse section of an axis of an innovation-shoot in its first growth-period, about 3mm thick, exhibited a distinct Casparian sheath, and in the stele a circle of small vascular bundles, separated by comparatively broad rays of intervening parenchyma; at a depth of about 3—5 layers of cells within the cortex there had been formed, even at that point of time, a cork-cambium which had produced a 3—6 layered cork (Sudan III). The transverse section of a 2-year-old rhizome, with its richer 2nd year’s leaf-rosette developed, exhibited in its uppermost part a structure similar to that described in the case of the one-year-old shoot, only, the diameter of the organ was larger (about 5mm), the vascular bundles were more numerous and also individually broader, and they had on the outer side of the xylem-portions a thick cap of wood- fibres. In the basal portion of this shoot, which had been developed the year before, and which was consequently 2 years old, there occurred, however, a distinct growth-zone in the vascular bundles, since, either as a direct continuation of the xylem-portion developed during the first year of growth, and furnished on the outer side with a cap of wood- fibres, or else separated from that xylem-portion by 1—3 layers of thin, unlignified elements, one more xylem-portion was found, which also had a cap of wood-fibres along its peripheral portion, and evidently represented the growth of the 2nd year. Frequently, such xylem-groups of the 2nd year are also developed without being in connection with the vey ge ре Scrophulariaceae. 451 xylem-groups of the 1st year, viz., opposite the parenchyma- rays which separate the original vascular bundles. The cork in the primary cortex is constantly few-layered only. Lastly, a portion of a rhizome of a 3-year-old shoot with a terminal peduncle, was investigated. The basal portion was found to be unaltered, as described above in the case of the 2-year-old shoot; in the part developed during the 2nd year of growth (illustrated diagrammatically in Fig. 32, B), fundamentally, the same development has taken place as that which, in the second year, took place in the basal portion, the only differences being that here the vascular groups are larger and more numerous, and the stereom-caps much thicker, and also that the two zones of growth are separated by a ring of non-lignified tissue (parenchyma) as much as 15 cell-layers thick, in which a layer of cork, con- sisting of 3—5 layers of cells, is developed; this layer of cork surrounds, in a tube-like manner, the xylem developed during the first year of growth. In the figure the cambium is consequently situated immediately outside the outermost circle of xylem-groups; the endodermis is in part distinct, and in the figure, is indicated by a fine line; the two thickly- drawn circles indicate the outer and inner cork-layer; within the pith there is another one like it, also indicated by a thickly-drawn line; in the outer circle of xylem-groups, the inner, dotted part indicates that this part consists especially of vessels, the outer radiately-shaded part, that there wood- fibres predominate, in the inner circle where the situation of the vascular- and stereom-elements is less sharply distin- guishable, the whole of the xylem is shaded. Where the rhizome bears the rosette of the 3rd year, it shows a transition to the structure of the peduncle, and gradually merges evenly into the latter — the wood-fibres form an almost continuous ring, and cork-development ceases. 452 Ев. J. MATHIESEN. Fig. 32, C shows a diagrammatic, longitudinal section of that portion of the 3-year-old rhizome described above, in which the boundary line between 2nd and 3rd year’s growth occurs — here, also, the thickly-drawn lines indicate the cork-layers, and the xylem is shaded. As is evident from IN Fig. 33. Pedicularis Sceptrum carolinum. A, Transverse section of leaf. B, Palisade cells, and C, spongy paren- chyma in surface view. D, Epidermal cells from the upper, and E from the lower surface of the leaf. (A, В and С about !1%/,; D and Е about ??2/,.) (Aursundsöen,- Norway.) the figure the inner layer of cork is continued horizontally in an inward direction, and forms a plate right across the pith; the outer layer of cork shows various irregularities. A closer investigation of the process of development was not possible, owing to the very scanty material; as regards the development of a continuous layer of cork between the growth-zones of the xylem, it must be noted, that a similar Scrophulariaceae. 453 structure, according to the investigations of L. Kocu? and Е. STRASBURGER, is found in the aereal stems of Sedum populifolium L. The Leaf. The epidermis of the upper surface has lateral walls, varying from straight to slightly undulating; only along the edges of the sections are the undulations more marked, very much as on the lower surface (Fig. 33, D and E). The lateral walls are thin, and faintly and distantly porose; only those cells on the lower surface which bear glandular hairs, have, as is also the case in the other species, thicker and more highly porose walls. The transverse section shown in Fig. 33, A, exhibits the lacunose structure of the mesophyll: the 1—2 layers of short palisade-cells, which tend to be almost isodiametric, and the usually abundantly- branched cells of the spongy parenchyma; a surface-view of the latter cells is shown in Fig. 33, С. Stomata occur only on the lower surface; the guard-cells are on a level with, or project only slightly above, the epidermal cells, which surround them to the number of 4—6. Chlorophyll occurs abundantly in the whole of the mesophyll, and also in the epidermis of the lower surface. Non-glandular hairs are totally absent; glandular hairs of the usual type oceur in great numbers under the secondary veins, and the veins of higher order of the leaf-sections. Pedicularis eapitata Adams. Herbarium-material from St. Lawrence Bay, King William’s Land (Gjöa Harbour), Ellesmere Land, Kotzebue Bay, Island of Iglorlik and from near the Taimyr River. 1 I. Koch: Untersuchungen über die Entwicklung der Crassula- ceen. Heidelberg, 1879, and Е. Srraspurcer: Ueber den Bau und die Verrichtungen der Leitungsbanen in den Pflanzen. Jena, 1891. pp. 324 —326. 454 Ев. J. MATHIESEN. Lit.: LANGE, 1880, р. 78; 1887, 262; KJELLMAN, 1882 (I), р. 257; 1882 (IV), р. 511; RosEenvinGe, 1892, р. 687; Easrwoon, 1902, р. 288; SIMMONS, 1906, p. 26; 1913, pp. 125 and 139. А sympodial hemieryptophyte of the rosette-type, which spreads by means of slender runners. The runners bear small scale-leaves and, rather scantily, adventitious roots, which are given off at the nodes; on some of the roots haustoria were found. The length of the vegetative period of the innovation- shoot varies from 1 to 4 years; during the first year of growth it develops at its apex either а winter-bud only, or it first forms a small rosette of 2—4 leaves. In one single case I found the shoot already flowering in the following year, but as a rule a rosette of foliage-leaves is formed in one, two, (which appears to be the most common), or three more years. The flower-bearing axis is terminal, as much as 9cm high, most frequently quite leafless, rarely with 1—2 foliage-leaves; at its apex it bears 2—6 flowers in a capitate raceme; some fresh leaves are always found at its base, since the shoot, also in the year in which it flowers, begins the year’s growth by the development of a small rosette of foliage-leaves. The winter-buds are protected Бу scale-leaves. In that part of the shoots where the leaf-rosettes are developed, branching seldom takes place; the new shoots arise most frequently from the portion of the runner imme- diately below the rosettes, or from the part of it, where it turns upwards. The runners are as much as 2 mm thick; how long they may become, I am not prepared to state, owing to the scantiness of the material, which would illu- strate this point; several runners measured 3—4 cm. Some- times they may be quite short, or a runner-like portion is entirely lacking at the base of the innovation-shoot, so that several flowering shoots may occur close together. Scrophulariaceae, 455 The structure of the flower is shown in Fig. 34, A—D; аз I had only dried plants at my disposal, these drawings are somewhat diagrammatic. As in P. Sceptrum carolinum the lower lip is directed upwards, on its inner side (morphological upper side) are two convexities. The maximum distance from the base of the corolla-tube to the tip of the helmet, was 3 cm; the ia Fig. 34 Pedicularis capitata. A, A bud and B a fully expanded flower; in B the position of the anthers is indicated. C, Corolla-tube and upper lip cut through the median plane of the flower, the lower lip has been removed. D, Uppermost half of the upper lip, more highly magnified in order to show the vena- tion of the helmet; lowermost in this figure one sees the uppermost parts of the ridge and of the revolute portion of the edge of the upper lip respectively. (Ellesmere Land.) (A, В and С about ?/,; D about 3/1.) colour of the corolla is yellow. The position of the ridge is shown in Fig. 34, C, and the venation of the helmet in Fig. 34, D; in the latter figure the uppermost part of the ridge is seen, and a little of the narrow revolute portion of the edge of the upper lip (entirely black in the figure). The revolute portion appears as if it were smooth, but, under the microscope, some of its epidermal cells are seen to be furnished with quite small, warty protuberances with cuti- cular striations; such protuberances are also found along the 456 Fr. J. MATHIESEN. edges of the helmet, above the revolute portion. The corolla- tube is somewhat Вашу within, in its lower part. The length of the style and the position of the stigma varied greatly, even in flowers from the same plant. The relatively longest-styled flower in my material was the one shown in Fig. 34, B; the plant to which it belonged came from Ellesmere Land; Fig. 34, A shows a not yet open flower from the same plant. From Ellesmere Land I have, how- ever, also had flowers in which the style was so short that it hardly reached outside the tip of the helmet, whilst in others it protruded 1—2 mm outside, but was bent more or less decidedly upwards. In several specimens from King Point there were flowers in which the end of the style was bent so strongly inwards, that it came to lie within the edges of the helmet, close up to the anthers. Some flowers from the Island of Iglorlik showed that, in the older ones, the point of the style may be bent more inwards than in the younger ones of the inflorescence; anything corresponding to this I have, however, not succeeded in confirming in flowers from the other localities. With the exception of the cases in which the stigma protrudes slightly, self-pollination by falling pollen must easily be able to take place, naturally, more especially where the stigma, is bent close under the anthers. According to Simmons P. capitata flowers on Ellesmere Land at the beginning of July, and in most years does not, in all probability, succeed in setting fruit there, for which reason it will often be under the necessity of reproducing itself vegetatively by means of its runners. As regards its habitat, Simmons remarks, that the spe- cies grows chiefly in marshy soil, but may also be found in drier localities where, however, the inflorescence becomes poorer (1—2 flowered). Scrophulariaceae. 457 According to the same author, the Geographical Di- stribution of P. capitata is as follows: North-western Green- land, Arctic American Archipelago, Arctic America, Una- laschka, East-Arctic Siberia to Taimyr Peninsula, and Kam- schatka. Anatomy. The Root: The structure of evidently full- Fig. 35. Pedicularis capitata. A, Portion of a transverse section of root. B, Transverse section of a runner, diagrammatic, cf. text. С, The peripheral layers of a runner in transverse section, showing the cork-formation. D, Two non-glan- dular hairs from the lower surface of the leaf. Æ, Transverse section of leaf. (A and С about 22°/,; В about 16/,; D about 14°/,; E about 17°/,.) grown adventitious roots, about 0.5mm thick, was as fol- lows: The epidermis appears to die away rather quickly; the cells in the outermosi layer of the cortex have cuticular- ised outer and lateral walls; the cells under the exodermis were frequently divided by a thin tangential wall. The 458 Ев. J. MATHIESEN. cortex is few-layered (5—7 layers), the cells in the 3 outer- most layers unite closely without any intercellular spaces; starch is found in the cortical cells. During the secondary growth of the root, the endodermal cells are elongated tangentially, as are the other elements of the cortex and, are divided as the exodermal cells, by thin radial walls, into 2—4 cells; moreover, in the inner walls, and in the primary radial walls, there is formed a cuticularised lamella (the thickly drawn line in Fig. 35, A), so that the stele becomes, so to speak, surrounded by a tube of cuticularised material, only interrupted by non-cuticula- rised passage-cells opposite to the protohadrome-rays; (the roots investigated were 2—3 rayed). The secondary growth of the root stops when a thin bundle of xylem-tissue, cir- cular in outline, and surrounded by a narrow zone of sieve- elements, is developed in the stele. The Peduncle presents nothing particular in its ana- tomy. Chlorophyll-grains are found in the cortical cells, the cells of the outer layer of the cortex have somewhat thickened outer and inner walls; the endodermis has recognizable Cas- parian dots. In the periphery of the xylem-ring the wood- fibres dominate; the outer layer of the pith has thickened, woody, porose cell-walls; the central cells of the pith are thin-walled and die away. On the surface multicellular non- glandular hairs occur. On the other hand, the anatomical structure of the run- ner-like part of the shoots deserves a fuller treatment. A diagrammatic transverse section is shown in Fig. 35, B; the thickly drawn line slightly within the periphery indicates a cork-layer, a portion of it is shown in Fig. 35, C under higher magnification. The outermost (uppermost) layer of cells in the latter figure, is the epidermis; this, together with the 2—3 outermost layers of the cortex, naturally dies away as Scrophulariaceae. 459 soon as cork-formation commences, yet how early in the development of the runner I am not prepared to state, but it probably happens rather early. The amount of cork developed is small; more than five layers of corky cells (the thin-walled elements placed in radial rows seen in the figure) never occur, in fact, often only 2—3 layers; sometimes the cortical layer situated immediately under the epidermis, forms the cork-cambium. This cork-formation is continued from the runner-like, into the rosette-bearing part of the shoot. In the stele, which is surrounded by an endodermis of a similar kind as that of the root (in Fig. 35, В the inner- most circle indicates the endodermis), two groups of xylem are found; the vessels are intermingled with non-lignified paren- chyma, and the groups are separated by two broad medul- lary rays, diametrically opposite to each other, in the outer part of which a small group of narrow vessels occurs (leaf- traces). The pericycle is several-layered; like the cortex, it consists of cells with somewhat thickened walls. In the peri- fery of the xylem-groups a few wood-fibres sometimes occur. The Leaf. The epidermis of the upper surface has lateral walls varying from straight to slightly undulating, comparatively thick, but only slightly porose; on the lower surface of the leaf the lateral walls of the epidermal cells are found to be slightly undulating; the cells are essentially more thin-walled than those of the upper surface. Stomata occur only on the lower surface of the leaf; they are sur- rounded by 4—9 cells. A transverse section of the leaf is shown in Fig. 35, Е; as far as I could judge from the material at my disposal, the portion which has been selected for illustration is typical: 2 layers of palisade-cells, and a few-layered spongy paren- chyma with cells, which do not, on the whole, branch very copiously, are seen. 460 Ев. J. MATHIESEN. On the lower side of the leaf-lobes, glandular hairs of the usual Pedicularis-type are found, below the branches of the veins of higher order; the dense covering of non-glandular hairs on the lower surface of the leaf as well as on the leaf- stalk, is very characteristic of the species; the hairs (Fig. 35, D) are 2—5-celled, rather thick-walled, pointed and smooth, they are especially numerous below the branches of the veins of lower order, and along the leaf-margin, but are also found scattered here and there among the glandular hairs, and on the spaces between the veins. The epidermal cells from which they originate (Fig. 35, D) may be very thick-walled; they are often arched cupola-like. Pedieularis hirsuta L. Alcohol-material from numerous localities both in East and West Greenland, collected by Е. Warmine (11. 7. 1884), C. RYDER (11. 7. 1886 and 5. 7. 1887), S. Hansen (2. 7. 1888), N. Hartz (1. 9. 1891) and А. Гомрлсев (29. 6. 1907, 17. 6. and 20. 6. 1908); also from Spitzbergen (Belsund, leg. NAT- HORST, 1. 7. 1888) and Norway (Talvik near Alten, leg. E. WARMING, 17. 7. 1885). Herbarium-material from Greenland, Ellesmere Land, Grant’s Land, Spitzbergen, Nova Zembla and Northern Nor- way (Alten). Lit.: Lange, 1871, pp. 253 et seq, 266, tab. III, fig. 21 (the seed); 1880, р. 76; 1887, р. 262; NATHorst, 1883, р. 10; WARMING, 1886, pp. УП, VIII, 21, 44 and 54; 1888, pp. 35, 39, 59, 71, 14, 8%, 97 and 103; 1890, pp. 206, 208 and 213; RosEnvinGe, 1892, р. 686; 1896 (II), р. 78; Hartz, 1894, pp. 32 and 34; 1895 (I), pp. 116, 128, 167, 170, 171, 175, 225, 247 and 288; 1895 (II), pp. 335, 359 and 372; Hartz and Kruuse, 1911, pp. 339, 342, 343, 345, 347, 348, 353, 359, 364, 370, 375, 376, 378, 385, 402, 405, 409, 416, 418, 419, 423 and 428; Norman, 1895, р. 458; Exstam, 1897, pp. 118, 168 and 180; 1899, рр. 7, 33, 37, 40, 48 and 50; Квоозе, 1898, pp. 350, 373, 380, 386, 394 and 398; 1905, p. 176; 1906, p. 249; 1911, in part IV pp. 196, 202, 244, 254, 266, 269 and 272, besides notes in the preceding parts; Scrophulariaceae. 461 ABROMEIT, 1899, р. 43; Кмотн, 1899, р. 190; Andersson and HEs- SELMAN, 1901, р. 16; CLeve, 1901, pp. 10, 39, 57, 69, 70, 72, 76, 82, 83 and 88; Dusén, 1901, р. 39; EAstwoon, 1902, р. 290; PorsıLD, 1902, pp. 110, 114, 124 and 216; 1910, pp. 259, 267 and 271; 1912, pp. 382 and 385; 1920, р. 143; БугуЕм, 1905, р. 88; Simmons, 1906, р. 27, tab. II, figs. 7—8, tab. III, fig. 1; 1913, pp. 123, 137 and 164; OSTENFELD and LuNDAGER, 1910, р. 31; OsTENFELD, 1915, р. 381; Lunpacer, 1917, рр. 360, 394, 399 and 402; Resvozz, 1917, р. 214. Spot-bound, sympodial hemicryptophyte of semi-rosette type, with primary root of long duration, attaining as much as 2 em in thickness and often furnished with some vigorous branches. As recorded by Tu. RESVOLL, in older specimens adventitious roots can be developed. SYLVÉN describes only mere seedlings. The seed ger- minates in spring; during the summer a rosette consisting of a few small foliage-leaves, is formed, and the young plant enters upon its winter-rest with its shoot-apex protected by 4—6 bud-scales (TH. ResvoLL): the following summer а rosette of more abundant leaves is formed, then again bud- scales and so on, until the time of flowering is reached. A flower-bearing shoot-portion with elongated internodes ter- minates the growth of the main axis. Also as TH. RESVOLL remarks, during the flowering year, no rosette of radical leaves is formed. As regards the length of the first vegeta- tive stage, the said author has found, that two-year-old plants had not yet any floral organs developed in the ter- minal bud. At the rhizome-portion of the floral shoots, 1—3 innova- tion-shoots are developed — most frequently two; the suc- cession of shoots with short internodes which are thereby formed, gradually brings about the formation of a “rhizoma multiceps” (a mesocorme). The innovation-shoots pass through a vegetative-stage of two years duration before flowering; every summer, a rosette consisting of a few foliage-leaves, is formed and, as XXXVII. 30 462 Ев. J. MATHIESEN. in the main shoot, bud-scales, for the protection of the winter-buds. In the adventitious shoots, also, no rosette of fresh foliage-leaves occurs in the flowering year. I have been so fortunate as to have Arctic material at my disposal, collected as late as on the Ist of September (Danmarks @, leg. Hartz); it contained both buds which would flower next year, and also younger buds. The former were of a considerable size (1.5 cm long, 1 cm thick); ex- > © Fig. 36. A—E, Pedicularis hirsuta. A, À bud-scale. B, Transitional form between bud-scale and foliage- leaf. C and D, Leaves from a part of the stem with elongated inter- nodes. Æ, Rosette-leaf. (Greenland). F—H, Pedicularis lanata. F, À lower and G an upper leaf from the rosette of the same year. H, А leaf from the part of the stem with elongated internodes. (Green- land). All about nat. size. ternally they were protected by the somewhat broadened bases of the dead leaves of the rosette, then followed 10—12 arched, pointed bud-scales, 10—13 mm long, and 5 mm broad, and hairy along their edges; the outermost of these bud- scales were withered. The young foliar organs were already large, and in the axils of the subtending-leaves of the in- florescence were floral buds about 0.5 mm in size, all enveloped . Scrophulariaceae. 463 in non-glandular hairs. The purely vegetative buds were in all cases smaller, and the bud-scales fewer in number; also in.them the young foliar organs were well-developed. In Fig. 36 are shown 2 bud-scales (of which the one, B, was uppermost in a bud, and shows a rudimentary leaf- blade), a rosette-leaf and 2 leaves from the part of the shoot with elongated internodes (both kinds of foliage-leaves vary greatly in form). The leaves are hairy, especially at the base, and the stems also are abundantly covered with hairs, especially in the upper part. The height of the flowering plant may vary greatly. From Spitzbergen (leg. Tu. Wurrr) I had specimens scarcely 2 cm high, which nevertheless bore as many аз 4 flowers of normal size. It appears that during the end of the flowering period, a rather considerable elongation of the stem, especially of the upper part, takes place; two fruit-bearing specimens from Greenland (Amitasiarsak Fjord) were as much as 36 cm high. The development of а “rhizoma multiceps” causes a somewhat tufted growth of the species. One specimen from Hvalrosodde (North-east Greenland, leg. ANDR. LUNDAGER) had at the same time as many as eleven flowering shoots, numerous leaf-rosettes and ten dead inflorescence-axes from previous years. This tuft-forming growth must afford ad- ditional protection to the winter-buds, the latter being covered by the numerous withered leaves. Haustoria are found on the roots; according to RosEn- VINGE this species is parasitic on Vaccinium uliginosum and Salix herbacea, for example. In the fresh condition the root has a slightly yellowish colour, and a somewhat sweetish taste (Мдтновзт). The radiating clefts, which are formed in the tissue of the root, 30* : 464 Ев. J. MATHIESEN. especially in the cortex (see below), are, during the winter, found to be full of ice (M. PorsiLp, in lutt.). The structure of the flower will be evident from Fig. 37 (= WARMING, 1886, fig. 12). A shows a flower seen from Fig. 37. Pedicularis hirsuta. A and В (the same flower) as well аз C—F and K—N are drawn from Greenland material; Z from a flower from Northern Norway (leg. Му- Huus), О and P from flowers from Spitzbergen (leg. Naruorst). (Cf. text.) (Е. WARMING, 1886.) the side; В the same flower, and К another flower, both in front view: in A the densely hairy calyx should be noted, the short corolla-tube, the short, warty revolute portion of the edge of the upper lip and the ridge which passes down- wards from its uppermost point, into the back of the corolla- tube; also the two convexities on the lower lip, and the Scrophulariaceae. 465 furrow between them; in B the slightly oblique lower lip should be noted and the wide slit (as much as 2 mm) between the two revolute portions of the edge of the upper lip: here the stigma protrudes just beyond the tip of the helmet. С shows the curving of the upper part of the style, in E is seen its natural and usual position in relation to the helmet and anthers. A flower of somewhat different structure 15 shown in /: there the style is shorter, less curved, and does not reach down underneath the anthers. D and / are opened anthers, С stigmatic papillæ and Н germinating pollen-grains. Honey is secreted by the three-lobed swelling on the front of the oblique ovary (P). Peloria may occur: such a flower is shown in N; there the corolla is irregularly 6-lobed, the stamens are of equal length, and the style protrudes far, and is erect. In very small-flowered specimens brought by NATHORST from Spitzbergen, Е. WARMING found as close contact of stigma and anthers, as shown in ©. “The corolla is of a pale rose-colour, with darker throat and tip of helmet, more seldom the colour is pure white (for instance in Upernivik). The flower is set almost hori- zontally. The corolla-tube is... . 7—8 mm long; the total length of the flower is about 13—14 mm. Slight protogyny appears to occur, and the stigma probably protrudes some- what more in the younger flowers than later on, so that cross-pollination may perhaps take place in these” (Е. War- MING, 1890, p. 213). According to information given by M. PorsiLp it is common in Northern Greenland for the corolla to be of a whitish shade or, at any rate, of a pale pink colour. EKSTAM records this species from Nova Zembla as scentless, but from Spitzbergen, as having a strong perfume. Like all the species of Pedicularis mentioned here, it must be assumed originally to have been adapted to pollination by humble- 466 Ев. J. MATHIESEN. bees, but insect-visitors are rare in Greenland (M. PorsıLp, in litt.), nor are they common in Northern Norway (EKSTAM, 1897, р. 175); from Spitzbergen humble-bees are absent (AURI- VILLIUS). In spite of this, the species regularly sets fruit everywhere; it must also be said to be the one of the Arc- tic Pedicularis spp., which is best fitted for self-pollination. In flowers like E and О (less, it is true, in J) self-polli- nation is inevitable. The seeds are described and figured in LANGE, 1871 (tab. III, fig. 21). Geographical Distribution: West Greenland from 64° and northwards, rare between 64° and 68°; East Green- land from 65° and northwards (rare in the distriet of Ang- magsalik); Arctic North America (the continent and the islands), Arctic Siberia and Russia, Nova Zembla and Spitz- bergen, the mountain summits of northern Scandinavia. In Greenland, according to the numerous records in the literature, it appears that the species makes no great demands as regards the nature of its habitat. Thus it can be found on heaths, mossy flats, herb-slopes, in coppices and in fissures in fields where the surface is cracked into poly- gonal cakes (rudemark) (LANGE, WARMING, HARTZ, KRUUSE and М. PorsiLp), according to the last author also in gravelly barrens (1920, р. 143). In Greenland this species can dis- pense with a snow-covering during winter, and it is there a decided sun-plant, (M. PorsıLp), and has not its proper home in the snow-troughs. According to Norman, in Northern Norway it prefers gra- velly, not grass-covered localities, and there it has not hitherto been found where the aspect is sunny, but only on the north- ern and indifferent (eastern and western) sides of the mountains ; it belongs there to the flora of the snow-troughs (TH. REs- VOLL). A. CLEVE also remarks: “Bedarf sehr wenig Insolation -Scrophulariaceae. 467 und meidet sogar die sonnig trockene Wiese’’; the species is otherwise characterized as “häufig in der Heide, insbesondere die frischere Andromeda-H., gern in schattigen, schneereichen Mulden, vom Plateau das Gebirge emporsteigend, Ferner in der Sumpfwiese und auf den Moorhiimpelen, aber einerseits ins Moor nicht hinabsteigend, anderseits nicht auf trockener, vegetationsarmer Felsenflur. Ein verfilzter Humusboden ist unbedingt notvendig.” In Greenland this species is a decided- ly spring-flowering plant, but, as an exception, flowering specimens may be found throughout the summer according to the point of time at which the snow melts at the place in question. Harrz found it on Danmarks © with the flowers still in bud on the ist of September, and expressly remarks that it was “on quite recently-bared ground, at the foot of the snow-drift.” Anatomy. The Root. Fig.38, A and B, shows portions of the transverse section of a root, about 4mm thick. The epidermis decays very quickly — as described by Hove- LACQUE for the root of Pedicularis palustris — the epidermal cells have already collapsed in roots not more than one mm in thickness, and in the root figured they have disappeared completely. The uppermost layer in Fig. 38, A, is the outer- most layer of the primary cortex, which, on the whole, only consists of 4—5 cell-layers. The cells have been very greatly elongated tangentially during the secondary growth in thick- ness, and are divided by a great number of thin radial walls; the primary cortex persists for a long time, I found it even on roots about one cm thick. The outer walls of the outer- most layer are cuticularised. At the stage of development figured in Fig. 38, an endodermis is not recognizable; in quite young roots faint Casparian dots can be demonstrated. The structure of the secondary cortex is very characteristic, since there is found in it a circle of radiating clefts, of which two 468 Ев. J. MATHIESEN. are seen in В (26 such clefts — most of them larger — occur- red in the root figured) by which all the phloëm-tissue of secondary formation is split into a corresponding number of radi, which in their outermost parts ap- pear to be some- what crumpled to- gether, evidently by pressure of the tis- sue later on formed by the cambium. In these radii the groups of sieve- tubes are flanked on both sides by parenchyma. With- LE LE in the cambium ve Rt some groups of ves- sels are seen; cor- ray of tissue in the cortex 1 or 2 rays of groups of vessels occur in the xylem, Fig. 38. Pedicularis hirsuta. A, Transverse section of the outermost layers M of the cortex of the root. B, Transverse sec- by broad radii of tion of a portion of the stele of the root. parenchyma, in divided reciprocally Greenland.) (A and В about 129/,. : (Greenland.) (A an cape ey which the clefts from the phloém may be continued, and may reach almost to the centre of the root. | The portionof the stem with elongated internodes shows a thin-walled epidermis with stomata; the cortex is few- layered, and its cells contain chlorophyll-grains. When the responding to each Scrophulariaceae. 469 stem has reached its full development (1.е. at the end of the flowering period), there is found in the stele a continuous, but narrow ring of vessels and a few layers of wood-fibres, strengthened internally by the outermost layers of the pith, III > = > ae { ] PT 6 Sees Le) ge: | |) hd X OA KAS Fig. 39. Pedicularis hirsuta. A, Transverse section of the leaf. B, A non-glandular hair from the basal portion of a leaf. C, Epidermis of the upper, and D of the lower surface of the leaf. (Greenland.) (A about 12°/;; В about ®/,; С and D about ??°/,.) the cell-walls of which become somewhat thickened and ligni- fied; the central part of the pith dies away. In the basal part of the stem which bears the leaf-rosettes (the mesocorme), a transverse section shows a circle of vascular bundles with 470 Ев. J. MATHIESEN. intervening parenchyma-rays. Secondary growth in thick- ness takes place, and in connection with this a similar cleft- formation in the phloëm as in that of the root. The pith is wide, and does not die away. The Leaf. The epidermal cells of both the upper and lower surface have thin, finely porose and almost straight to slightly wavy lateral walls (Fig. 39, C and D). Stomata occur only on the lower surface; they are surrounded by 4—6 cells. The transverse section shows 1—2 layers of well- developed palisades and a spongy tissue consisting in part of rather copiously branching cells (Fig. 39, A). Non-glan- dular hairs of the type shown in Fig. 39, B occur both on the stem and leaf; they are found in particularly great num- bers on the basal portions of the foliage-leaves of the portion of the stem with elongated internodes, and on the leaves subtending the flowers, but more sparingly on the basal portion of the rosette-leaves proper, and only singly on the leaf-lobes, and then always on the upper surface. On the epidermis of the lower surface, under the branches of the veins of higher order, glandular hairs of the usual type are found; the epidermal cells upon which the glandular hairs are seated, have straighter, thicker and more highly porose lateral walls, than have those which surround them. Chlorophyll was present in abundance in the whole of the mesophyll and was also found in the epidermis of the lower surface. No difference could be observed as regards the anatomical Structure of the leaves of the rosette and of those on the portion of the stem with elongated internodes. ‚Serophulariaceae. 471 Pedicularis lanata (Willd.) Cham. & Schlecht. Alcohol-material from Greenland (Amerdlok, 15. 7. 1884, leg. Tu. Horm; Kingartak, 25. 7. 86, leg. RypER; Proven, 24. 6. 1888, leg. MYHRE). _ Herbarium-material from Greenland, King William's Land, Ellesmere Land, Arctic North America, Nova Zembla and Arctic Siberia. Lit.: . LANGE, 1880, р. 76; 1887, р. 262; NATHORST, 1883, р. 10; WARMING, 1886, pp. VII, VIII, 21, 44, 47 and 54; 1888, pp. 59, 74, 87 and 104; 1890, pp. 206, 208, 210 and 213; RosEnvinGE, 1892, р. 687; 1896 (I), р. 68; Hartz, 1894, р. 45; 1895 (I) р. 306; 1895 (II), р. 372; Kruuse, 1898, рр. 350, 373, 380, 394 and 398; 1906, р. 249; 1911, pp. 103, 132, 196 and 207; ABROMEIT, 1899, р. 44; Exsram, 1899, pp. 7, 32, 37, 40, 48, 49 and 50; Кмотн, 1899, р. 193; ANDERS- son and HEsseLman, 1901, р. 16; Eastwoop, 1902, р. 291; Ровзию, 1902, pp. 114, 178; 1910, pp. 259, 267 and 270; 1912, pp. 382 and 385; 1920, р. 143; Simmons, 1902, р. 29, tab. II, figs. 1—8; 1906, рр. 124, 137 and 164. Spot-bound, sympodial hemieryptophyte of the semi- rosette type; like the foregoing closely related species, it has a thick, scantily branched main root of long duration; ad- ventitious roots are probably, as a rule, only slightly deve- loped. As P. hirsuta, this species can form large and dense tufts. It appears that the first vegetative-stage may be con- fined to two years, so that the main axis of the young plant, after the formation of two consecutive leaf-rosettes, can, in the third summer, form a shoot-portion with elongated inter- nodes, although a comparatively weak one, terminating in an inflorescence. Generally, the necessary vegetative-stage preceding the flowering is, no doubt, of longer duration. The vegetative stage of the innovation-shoots extends as a rule over 2 years. P. hirsuta and P.lanata differ essentially from each other by the fact that the latter has no specially-developed _ En RU AS IR QGILA J PC" Le 5 a ®* Ps 472 Fr. J. MATHIESEN. bud-scales, but protects its young buds during the winter with the broad, arched, very persistent bases of the rosette- leaves developed during the growth-period of the preceding summer, which bases are densely hairy along their edges; the rosette-leaves formed towards the end of the growth- period, are especially distinguished by their large basal-por- tions. Fig. 36, G shows such a leaf. On the other hand, the stalk and blade of the leaf illustrated in Fig. 36, G are es- sentially smaller than those in Fig. 36, F, which is a pre- viously developed leaf from the same rosette; the leaf-blade may be even smaller than in G, but I do not think it is ever wholly wanting. The leaves upon the uppermost portion of the shoots where the internodes become elongated, have not such broad bases (Fig. 36, H). The rosettes are rich in leaves; at the base of a fairly vigorous flowering shoot I counted 60 withered leaves or remnants of such; also the above-ground portion of the stem is abundantly covered with leaves, especially below, although it cannot be said that the shoots have any true radical rosette of fresh leaves in the flowering year. The rosette-leaves, when withered, are very persistent, and may remain for several years, black and crumpled, at the base of the shoots (Fig. 36, F and G are two such withered leaves); naturally, this contributes towards making the tufts denser, and the protection of the young buds more effective. The root is intensely yellow in colour, and has, as already stated by Млтновзт, a sweetish taste, almost like that of carrots. It is used as an article of food by the Smith-Sound Eskimos? and by the Chuckes (Kjellman, 1882 (III), р. 366). The structure of the flower is shown in Fig. 40, A—D 1 P. Freucuen: Om Plantekost hos Smith-Sund Eskimoerne. Geografisk Tidsskrift. Kjöbenhavn. Bd. 24. 1917—18. p. 310. Scrophulariaceae. 473 (the drawing is executed after а sketch by Euc. WARMING made from material from Greenland). In A and В the flower is seen from the side, the position of the stamens is indicated in А; in В the ridge reaches highest above the upper- most edge of the revolute portion of the margin of the upper lip; of the three lobes of the lower lip, the middle one is Fig. 40. Pedicularis lanata. the smallest. The lower 4 and В, Flowers seen in side view; a : in A the position of the anthers and lip is not oblique; the the style is indicated; the hairs of the two convexities are quite calyx are drawn only along the out- lines of the figures. C, The uppermost } Г part of a flower in front view. D, A flower in front view, the pair of anthers. (Greenland.) (А, В and narrow furrow between C about IE D about I) (After E. WARMING.) glabrous. C shows a the convexities should be noted. D, a pair of stamens; here the filaments of both stamens are hairy; this feature varies, sometimes only the two longest are hairy. In this species also there occur nectary- protuberances at the base of the ovary. “This species is still more conspicuous (than P. flammea and hirsuta) on account of its richly and densely flowering inflorescence, and the bright pink colour of the flowers; the tip of the upper lip and the throat are darker in colour than the rest of the flower. It has also a slight scent, and the flowers are considerably larger than those of the two foregoing species (P. flammea and hirsuta), having, namely, a total length of 2 cm, and a corolla-tube-length of 12—13mm .... The flower stretches out considerably, it may even be almost horizontal. The stigma projects out of the flower, 474 Fr. J. MATHIESEN. and will evidently be easily touched by an insect visitor. the anthers are placed immediately behind the stigma and, on account of the stretched-out position of the flower, partly above it, so that self-pollination will probably take place” (Е. WARMING, 1890, р. 213). The same author describes a flower from Spitzbergen with pelorial development. On Spitzbergen (forma dasyantha) Exstam found the end of the style to be rolled up (as in P. hirsuta, cf. Fig. 37), so that the stigma is in contact with the anthers. In my material I always found it to be protruding, as also shown in the figure. Varieties with entirely white flowers may occur, but they are very rare (М. PorsiLp, 1920). The species is fragrant (Exstam and WARMING). It sets fruit everywhere abundantly and regularly, and also in Spitz- bergen, where humble-bees, its natural pollinators, are want- ing (AURIVILLIUS); according to. information given by M. PORSILD, insect-visitors are rare in Greenland. Among the Eskimos the plant is known for the abundance of honey contained in its flowers; at Cape York its popular name is: “The food of the humble-bees;” and children pick the flowers and suck the honey out of them or even eat them entire (FREUCHEN, l.c.). During the end of the flowering-period, the axis of the inflorescence elongates. Geographical Distribution. The species is common in West Greenland, from Disco Bay and northwards. “In №. Strömfjord it is restricted to alpine stations and northern slopes and rather scarce (P & E). South of Holsteinsborg observed several times down to Itivdlinguaq, 66°30’ (P & Е)” (Porsizp, 1920); from East Greenland it is totally absent (PorsiLp, 1. с.). It is also found in Arctic North America (the continent and the islands), the western Subarctic North . Scrophulariaceae. _ 475 America, (the Rocky mountains), Arctic Asia, Nova Zembla and Spitzbergen. The species prefers fairly dry habitats. M. Porsizp records that it is found “on rocky flats and on heaths”” (1912, р. 382). “In poor and open heath, often in gravelly barrens far away from other plants. When growing A 8 ICh > LE Fig. 41. Pedicularis lanata. A, Transverse section of the leaf. B, Epidermis of the lower, and С of the upper surface of the leaf. D and Е, Non-glandular hairs from the lower surface of the leaf. F, A glandular hair from the stem. G, А non-glandular hair from the basal portion of leaf. (Greenland.) (A SOON ib, СЕ and F about 220]. ; G about *°/,.) amongst other plants, however, the roots also of this species are provided with haustoria’ (1920, р. 143). During the winter, although the plants in themselves are often bare of snow (1. c.), yet the dense tufts always catch some snow between the numerous decaying leaves, and this naturally 476 Ев. J. MATHIESEN. affords some protection to the winter-buds (M. Ровзио т litt.). According to the same author, in the same locality, the species flowers 1—2 weeks before Р. hirsuta (1920, р. 143). On Spitzbergen, according to Екзтам, it is found “auf trockenen, starker Insolation ausgesetzten Abhängen.” In anatomical respects, the root and stem agree closely with the corresponding organs in P. hirsuta. The anatomy of the leaf is shown in Fig. 41, A—G. In the transverse section 1—2 layers of short and broad palisade-cells are seen, and a spongy parenchyma, which consists of distinctly branched cells. B shows a portion of the epidermis of the lower surface, and С a similar portion from the upper sur- face of the leaf. The lateral walls are thin and finely porose; stomata are found only on the lower surface of the leaf and the guard-cells occur on a level with the leaf-surface. On the basal portion of the leaf, the cuticle is thicker than on the stalk and the lobes, and it is furnished with fine stria- tions. Chlorophyll occur in abundance in the mesophyll, and also in the epidermis of the lower surface. The glandular hairs in this species are almost exclu- sively confined to the edges of the leaf-sections; in A is seen exteriorly to the left, a group of the tracheidal tissue in which the branches of the veins end, and in connection with which the glandular hairs occur. The structure of these is similar to that in the other species; the epidermal cell from which the glandular hair is developed, has not, however, in P. lanata, much thicker or more highly porose walls, than have the epidermal cells, which surround it. Non-glandular hairs of the type shown in G occur in abundance on the edges of the leaf-bases, on the stem and, in the floral region, on the scale-leaves and calyx, more scantily on the leaf- stalk, and on the upper surface of the leaf-blade. On the lower side of the leaf-lobes there are small non-glandular KE 22 Scrophulariaceae. 477 hairs (Fig. 41, В, D and Е) intermingled with the glandular hairs; on the petiole and stem are glandular hairs of the form shown in Fig. 41, F. Pedicularis flammea L. Alcohol-material from Greenland (Godthaab, leg. E. WARMING, 29. 6. 1884; Christianshaab, leg. 5. Hansen, 2. 7. 1888; Proven, leg. Мунве, 2. 7. 1888; Danmarks ©, leg. М. Hartz, 18. 2. 1892), Iceland (Hvitarvattn, leg. А. FEDDER- SEN, 3. 6. 1886), and Northern Norway (leg. NyHuus, 1885). Herbarium-material from numerous places in East and West Greenland, from Norway, Iceland and Arctic America. Lit.: LANGE, 1870, pp. 254 et seq., 266, tab. III, fig. 22 (the seed); 1880, р. 75; 1887, р. 262; WARMING, 1886, pp. VII, VIII, 44, 47 and 54; 1888, pp. 27, 34, 39, 59, 87, 90, 104, 130 and 133; 1890, рр. 207, 208 and 211; Воземутмсе, 1892, р. 686, 1896 (II), р. 128; Harız, 1894, pp. 11, 32 and 43; 1895 (I), pp. 165, 170, 171, ‚175, 188, 247, 256, 288 and 290; 1895 (II), рр. 335, 359 and 372; Hartz and Kruuse, 1911, рр. 338, 342, 347, 357, 359, 378, 379, 380, 385, 409, 411, 423 and 428; Borcrsen, 1895, pp. 223 and 225; Norman, 1895, р. 458; KRUUSE, 1898, pp. 350, 373, 379, 394 and 398; 1905, p. 176; 1906, p. 248; 1911, in part IV pp. 196, 243, 254, 255 and 274 besides notes in the preceding parts; Аввомеит, 1899, р. 43; KNUTH, 1899, р. 190; CLeve, 1901, pp. 18, 39, 57, 72, 76 and 88; Duse£n, 1901, р. 39; Porsıno, 1902, pp. 114, 175, 178, 187, 216; 1910, р. 267; 1912, pp. 382 and 387; 1920, р. 142; OsTENFELD and LUNDAGER, 1910, p. 31; Simmons, 1913, pp. 124 and 140. Spot-bound, sympodial semi-rosette hemicryptophyte, with short, vertical mesocorme which dies away behind, and a rather abundant development of adventitious roots. The first vegetative-stage fairly commonly appears to extend over 4 years — in the case of one individual (Disco) I found it to extend over 3 years, and in the case of another (Iceland) 6 years. A short, vertical mesocorme is developed during the first vegetative stage; the main root is still present during the first flowering-period, but dies away afterwards. XXXVII. 31 478 Ев. J. MATHIESEN. The winter-buds are provided with a number of bud-scales; these as well as the foliage-leaves leave real leaf-scars on the mesocorme on decaying. Adventitious roots are devel- oped rather abundantly, especially from the basal part of the mesocorme, they generally occur in connection with innovation buds; the adventitious roots are somewhat swollen. The innovation-buds arise from the axils of the scale- leaves; the lower buds on the mesocorme do not generally develop any further, but remain as reserve-buds; as in 2. Oederi it is the buds which occur in the axils of the inner- most scale-leaves formed during the summer previous to that in which the parent-shoot flowered, which are the most vigorously developed. Usually only two of these buds ex- pand, but the number may be more, in a single case I counted as many as 6. During their first summer the adventitious-shoots always bear two fully-expanded foliage-leaves; as in the main- shoot the terminal bud is protected by scale-leaves. In case flowering does not take place in the following year, — it may occur, but no doubt only exceptionally, — a rosette of foliage-leaves is developed, followed in turn by scale- leaves, and then, after passing through still another winter the shoot, in the majority of cases, succeeds in flowering. In a specimen from Ignerit (Greenland) in which 2 of the basal buds on the mesocorme had developed further, one of them had formed rosettes of foliage-leaves for 4 suc- cessive summers, and the other for even 5. In a winter-bud (the specimen in question was collected by N. Harrz, on the 18th of February 1892, on Danmarks 0, East Greenland) there were found young foliar organs and flowers, all of which were highly developed; in the young flowers the separate parts were quite distinguishable Scrophulariaceae. 479 under the microscope. The shoot which was terminated by this winter-bud, was one year old — consequently, the length of the purely vegetative period in this case will only have been one year — during the previous year it had, as is usually the case, unfolded 2 foliage-leaves; the bud-scales numbered 5. The maximum height of the flower-bearing shoot was found to be 28cm; from Iceland and Greenland I found small individuals, flowering for the first time, which were only 3 cm high. A specimen from Disco differed entirely by a total absence of foliage-leaves on the above-ground stem, which was 20 cm high, and by having flowers, the lowermost of which were long-stalked almost down to the base; the axis was comparatively very thick (about 8 mm in diameter). Е. Warmine found the flowers from Norway (W. Fin- mark), Iceland and Greenland to agree entirely in structure; from his descriptions I quote the following: “The flowers are not very conspicuous; their lower part stands erect, almost parallel with the axis of the inflorescence, their upper part is bent slightly forward (Fig. 42, A); the colour is yel- low, the upper part of the helmet is, however, more or less reddish-brown. The corolla-tube is only 7—8 mm long; the lower lip is very small, with three rounded lobes, and often appears smaller still by the fact of its edges being bent backwards (Fig. 42, С). The cleft of the upper lip is some- times very narrow (1/, mm, see Fig. 42, D), but sometimes wider, even so wide that there is а large entrance to the interior of the flower, as for example in flowers with an upper lip curved strongly forward. This species appears to be as highly adapted to self- pollination, as it is poorly adapted to insect-pollination, which also agrees with the fact that it is so little conspicuous. I never saw the stigma protrude from the blunt, untoothed tip; so it is placed immediately in front of, or slightly under 31* 480 Ев. J. MATHIESEN. the anthers (Fig. 42, F). The fact that in many flowers I have found the stigma covered with pollen-grains, some of which were germinating, and have found abundance of pol- len around the anthers, even in a flower like A—B, in which the stigma is situated so unusually high above the anthers, is a justification for regarding this as brought about by self- pollination. I must, however, point out the fact that the Fig. 42. Pedicularis flammea. (Greenland). A, А flower in natural position in the inflorescence; В (5/,), the upper part of the same in longitudinal section. C, Lower lip and the part around the throat seen in front view. D, The upper part of a flower in front view. Е, Lower Пр, slightly magnified. F аз В, but with the stigma in another position. G, А flower in natural position, larger than A, showing the venation of the upper lip. (Е. WARMING, 1890.) stigma is not placed in the same position relative to the anthers in all the flowers of the same inflorescence, and it appeared to me to be the rule for it to be situated slightly more forward in younger than in older flowers; . . . . the filaments are glabrous. In many places in West Greenland, as far as Upernivik (RYDER), it was found with abundant fruit. Even in July (1884) new fruit was set at 64° and 67° М. lat.’ ers Scrophulariaceae. 481 Г can add to the above that in older flowers the helmet often appears to bend forward, as described by KERNER in the case of Р. Oederi. In the material at my disposal, both from West and East Greenland, there were many fruiting specimens. That it really is the rule for the stigma to be enclosed in the helmet, is seen among other things, by the fact that Bryrr in his “Haandbog i Norges Flora” gives this as a specific character. Geographical Distribution: West and East Green- land (rare in South-west Greenland according to RosEn- VINGE, in East Greenland even as far north as Germania Land according to OSTENFELD and LUNDAGER), Iceland, Arc- tic and Subarctic North America, Arctic Russia and the mountainous regions of Northern Scandinavia. As regards the habitat of P. flammea, many notes are to be found scattered in the literature on the subject. A. CLEVE states that in the district of Northern Sweden in- vestigated by her, it was found “nur auf den Moorhügeln und in Sumpwiesen zusammen mit der vorigen (P. hirsuta), welche diese Art nicht bis in die trockene Heide zu begleiten vermag, ebensowenig wird sie auf dem N. Abhang oder in der Nähe von spät schmelzendem Schnee gefunden.” In Northern Norway it is found, according to Norman, only on the northern and on the indifferent (eastern and western) sides of the mountains. In West Greenland it grows, accord- ing to Е. WARMING, in willow-coppices where it may be said to have its home, and also on herb-flats, on heaths (parasiti- cally on Vaccinum uliginosum and Salix herbacea) and in grassy bogs. Ровзил has also found it on slightly damp moss-heaths. N. Hartz and KRUUSE record it from East Greenland as growing on herb-slopes and rocky flats, in carpets of prostrate Betula nana and Vaccinium uliginosum, on knolls in bogs, and on damp heaths. According to Harrz 482 Ев. J. MATHIESEN. (1895 (II), р. 335), its habitat is snow-covered during winter; the plant is mentioned as common in the interior of the fjords, but absent further eastwards, out towards the open sea. DuseEn records practically the same, and he also knows it only from a single habitat in the coast-land outside the fjords (Cape Borlase Warren). А. CLEVE considers it to belong to the “Gruppe der späte- A CoD ren Frühlingspflanzen.” In Greenland it is a middle- summer-flowering plant (M. PorsILD). The Anatomy of the root and stem agrees ex- actly with that of the cor- responding organs in P. Oederi. The parenchyma of the root contained a great abundance of amylodextrin- starch. The surface-view of Fig. 43. Pedicularis flammea. A, Epidermis of the upper surface of the leaf; В, of the lower surface of the leaf is shown in Fig. of the leaf. С, D and Е, Glandular { hairs; С and D from the upper 23, 2 and Sa RENE surface and E from the lower sur- the lateral walls were al- face of the leaf. (Greenland.) (A ways found to be rather and !B about *°/,; С, D and Е about 7°/,.) upper and lower epidermis faint. The transverse sec- tion of the leaf agrees ex- actly with that of P. Oederi. Non-glandular hairs are ab- sent : in addition to the glandular hairs of the usual Pedicu- laris-type, such forms are also found as are shown in Fig. 43, С, D and Е; С and D occur scattered, especially on the upper surface of the leaf, and Æ on the lower surface of the segments, especially along their main vein and lateral veins i Scrophulariaceae. 483 of 1st order. Chlorophyll-grains occur abundantly in the whole of the mesophyll, and also in the epidermis of the lower surface. Pedicularis Oederi Vahl. Alcohol-material from Norway (Kongsvold (Dovre), leg. Е. WARMING, 13. 7. 1887; Muggrubfjeldet near Roraas, leg. THEKLA RESVOLL, 29. 7. 1918). Herbarium-material from Nova Zembla, Arctic Siberia (Khabarowa and from near the Taimyr river) and St. Law- rence Island. Lit.: AxELL, 1869, р. 102; KıeLıman, 1882 (I), р. 257; 1882 (II), р. 325; 1882 (ТУ), р. 510; KJELLMAN and Гомозтвбм, 1882, р. 304; AURIVILLIUS, 1883, р. 451; WARMING, 1886, р. 47; 1890, рр. 207, 208, 210 and 214; Linpman, 1887, pp. 82 and 99, tab. IV, fig. 46; Kerner, Bd. II, 1898, р. 337; Кмотн, 1899, р. 186; SyLVÉN, 1906, р. 89; ScHrö- TER, 1908, рр. 454, 455 and 458; ResvoLz, 1917, р. 210. Spot-bound, sympodial semi-rosette hemicryptophyte with a short, vertical mesocorme which dies away behind. Adventitious roots are rather abundantly developed. According to SYLVEN the seed germinates in spring. “The elongated hypocotyl unites with the main root, scantily branched throughout, into a peg-shaped food-storing organ which becomes rather thick at an early period. In more advanced young-plants adventitious roots, thickened by the storage of reserve food-material, like the main root, some- times appear to be developed from the hypocotyl, or from the base of the epicotyl.” Tu. ResvoLL has studied the whole development from seed to flowering plant: During the first summer two small foliage-leaves expanded, the terminal bud was protected by a few scale-leaves, and contained the rudiments of the foliage-leaves of the following summer; the plant passed the winter in this stage. The foliage-leaves fell off late in autumn, and left leaf-scars. “After having passed 484 Ев. J. MATHIESEN. through the winter, the young leaf-organs developed into а rosette of foliage-leaves, few in number, . . . even early in the summer the development of foliage-leaves stopped, in- stead of which the apex of the stem developed a terminal bud covered with scales. In this manner the plant continues through several years. The leaves, which are exceedingly small in the first summer, having blades only about 1 cm long, become larger every summer, and the number of the leaves also increases with the age of the leaf-rosette. The last summer before the plant flowers I have seen 4—8 foliage- leaves.” The shortest first-vegetative-stage observed by the author in question, extended over 3 summers, in other individuals it proved to have extended over as many as 6 summers; in one individual (from Nova Zembla) I found 3 vegetative years, another showed 7. As the leaves fall off, leaving a regular leaf-scar (as already described by Е. War- MING (1890, pp. 207 and 209) both for this species and for Pedicularis palustris) and the broader and more distantly- placed scars of the foliage-leaves can be easily distinguished from the narrower and more closely-placed scars of the scale-leaves, the individual year-growths of the mesocorme become very marked and easily observable (TH. Resvorr). “It is characteristic of this plant, that the development takes place, so to speak, suddenly, as long as the plant is in the rosette-stage. Thus all the leaves come out almost at once, and not one by one during the summer; and as soon as the rosette has unfolded, the bud for the next summer is already seen in the middle of it. The same is also the case as regards the rosette-stage of the lateral shoots.” (Tx. Resvorr). Even in the bud, the rudiments of the organs which are to expand during the next summer, attain a high degree of development. In an individual which TH. RESVOLL in- DT ne Scrophulariaceae. 485 vestigated on Knutsho (Dovre) on July 23rd, the terminal bud proved to contain “the young stem for the coming summer, covered by the rudiments of the stem-leaves; within these were the subtending leaves of the flowers; the separate parts of these were already distinctly discernable.” The flower-bearing stem-portion has, as a rule, no foliage- leaves at its base, but such may occur as an exception. In the interval between germination and flowering the plant develops a vertical mesocorme, as much as 1 cm long and 0.5 cm thick, from the lower part of which there proceed a varying number of adventitious roots (I counted as many as 8); as in P. flammea they are somewhat swollen, as a rule especially in their proximal half, but sometimes also along their entire length, but narrowing abruptly towards the base; they become as much as 10 cm long, and as much as 6mm in diameter; most frequently they are furnished with slender, filiform branches, or, at any rate, scars left by such are present; more rarely with 1—2 vigorous root-branches. The main root dies away; in the majority of cases it probably does not survive the first flowering period of the individual. As mentioned by WARMING (1890, р. 207), the majority of the adventitious roots arise at the base of a bud on the mesocorme. According to the obser- vations of WARMING and Tu. ResvoLL — which observations I am able to confirm fully — the lateral buds are developed especially in the axils of the scale-leaves, and are of the same age as these. The buds in the lower part of the stem develop two small foliage-leaves during their first summer, but usually do not develop any further; as a rule they are found as small formations covered with a few scale-leaves; they must be regarded as reserve-buds, for, as seen in a specimen from Nova Zembla, under certain circumstances they may develop further and even produce flowers. 486 Ев. J. MATHIESEN. The most fully developed buds are always found (Тн. RESVOLL) in the axils of the uppermost set of scale-leaves, which are found to the number of 5—10 at the base of the flower-bearing stem-portion. It is commonly the innermost scale-leaves which subtend the innovation buds, and usually only two such are found, more rarely several (as many as 7, TH. ResvoLL). During their first summer they develop 2—3 rather large foliage-leaves, often followed by 1—2 quite small ones, and lastly bud-scales for the protection of the winter- bud. The buds open almost at the same time as the flowers of the main shoot. According to TH. ResvoLı, the duration of the vegeta- tive-stage of the lateral shoots may be restricted to one year only; it is, however, longer, as a rule; for instance, on Knutshö (Dovre) the said author found 4-year-old lateral shoots, which had not as yet young floral organs in the terminal bud. In several cases also I found a vegetative stage of only one year’s duration (among others in a specimen from Nova Zembla), more frequently, however, 2 or 3 years. As the parent-rhizome dies away behind, and the lateral shoots develop adventitious roots, there is a possibility of vegetative propagation, although, of course, the plant cannot spread much by this means. SCHRÖTER records this species from the Alps as occur- ring parasitically especially on Carex firma. The structure and biology of the flower was first des- cribed by Lınpman; since that Е. WARMING and KERNER have dealt with this species. Some unpublished drawings of the flower have been kindly placed at my disposal by E. WARMING; the illustrations in Fig. 44, with the exception of G and H, have been made after them. A is characterized by its long corolla-tube, its densely-hairy calyx, (the hairs have been indicated only along the edge of the calyx), Be . Scrophulariaceae. 487 and by its but slightly protruding stigma; in В the main features of the venation of the helmet are seen, to which a system of raised ridges on the outer side of the helmet cor- responds. The dots on the side of the helmet in Fig. 44, B indicate small glandular hairs, which are especially numer- ous in the front part of the helmet. Fig. 44, H shows such glandular hair in lateral view, highly magnified; the head ‚is 4-celled. D is the lower lip in surface view; the two con- Fig. 44. Pedicularis Oederi. А, В, С, D and Е are drawn from flowers from Sweden; F is a flower from Kongsvold (Dovre). G, Transverse section of one of the warts from the revolute portion of the edge of the upper lip. 7, A glandular hair from the helmet. (A, B, C, D, E and F slightly under ?/,; G and H about “°/,.) (Drawn by Е. WARMING.) vexities are distinctly visible here, as also in Fig. 44, C and F. Fig. 44, F shows a flower with a very far-protruding style; in Æ the position of the stamens is seen, the filaments of the foremost pair of stamens are hairy along their upper part. One of the warts from the revolute portion of the upper lip is seen under high magnifying power in Fig. 44, G; on its outer surface it has strong cuticular striations, as have also the cells which surround it. The calyx, subtending leaf and axis of the inflorescence vary from glabrous to 488 Ев. J. MATHIESEN. hairy, even in individuals from the same locality (Muggrub- fjeldet near Röraas); the hairs are multicellular, pointed and glabrous. I quote the following from Linnman's exhaustive des- cription:— “The flower is 20 mm long, yellow or whitish- yellow in colour; the tip of the upper lip is dark-red on its inner side, and this colour more or less penetrates through and gives the corresponding spot on the outer side a greyish- red tint; the upper lip has often in addition on the outer side a dark-red spot on each side. The flower is quite erect and adpressed to the редипее. .... The stamens agree with those in the other species; their pollen is consequently dry and loose, is fully developed even in the flower-bud, and is contained in a receptacle formed by the four convergently dehiscing anthers, which are held together by the laterally highly compressed upper lip. . . . . As in the other species the pollen is shed when an insect penetrates into the flower, and thereby widens the slit of the upper lip, which is rather narrow in this species. . . .. The nectary is in the same place as in the other species; the corolla-tube is 10—13 mm high.” The position of the stigma in the flowers figured by Linp- MAN is very much like that in Fig. 44, B and F, given here; in addition LinpMAN mentions flowers in which the style was so short, that the stigma remained quite inside the helmet, and did not even reach the anthers; the stigma was normally developed, however, in these flowers. According to KERNER (p. 337) the whole of the upper lip curves forward so strongly at the end of the flowering period, that it looks as if it were broken; even if it originally stood as a continuation of the lower part of the corolla-tube enclosed in the calyx, it stands then at an angle of 70°, or even 90° with this, and by bending forward it drags down with it the style and the stamens, so that the stigma no ur ee Scrophulariaceae. 489 longer stands in front of, but below the anthers; these are no longer united, and shed their pollen spontaneously, which is scattered down upon the stigma, and effects self-pollination in case cross-pollination should fail. LinpmMan and WARMING mention nothing regarding this point; I observed it in many inflorescences but must, however, add that the curvature may also take place in the uppermost part of the corolla-tube. Kerner (Die Schutzmittel der Blüthen gegen unberu- fene Gäste, 1876) and after him H. MüLrer, describes the pollination of P. recutita, a species very similar to the pre- sent one, and, in so doing, states that the insect (humble- bee) on its visit thrusts its proboscis into the narrow furrow between the two convexities on the lower lip, and forces these and the edges of the slit of the upper lip apart from each other; while LınpmAan maintains that the furrow in P. recutita is too narrow an opening for the introduction of the proboscis of the humble-bee, as this is not even pos- sible in the larger-flowered P. Oederi, where the proboscis can only be thrust in higher up, at about the middle of the revolute portions of the edges of the upper lip (“die Rolle’’). Visitors: LınpmAan observed Bombus nivalis and Bombus alpinus; AURIVILLIUS mentions frequent visits of humble- bees. According to WARMING, fruit is set abundantly in Scandinavia. Geographical Distribution. In the Carpathians and the Alps (“ganze nördliche Kette von St. Gallen bis Waadt — fehlt den Centralalpen,’’ SCHRÖTER), the mountainous districts of Scandinavia, Arctic Russia and Siberia (recorded by KJELLMAN from several localities along the north coast of Siberia), Nova Zembla, Central Asia and the mountainous distriets of East Asia. Habitat: Втутт (Håndbog i Norges Flora) records 490 Ев. J. MATHIESEN. “boggy places in mountainous districts, . . . . from the birch- zone up to the lichen-zone; more rare below the pine- limit”; Tu. Везуотл, has now and then observed it in snow-troughs; it is, however, no typical snow-trough-plant. It does not ascend very high in the Alps; SCHRÔTER (р. 255) mentions it as occurring in the transition-zone between his Agrostidetum-trifolietum and Curvuletum (Carex-curvula- assoc.). In Arctic Eastern Asia KsELLMAN found it growing on damp coastal plains. Anatomy. In anatomical respects the root very much the radiating cleft-formation in the stele — at any rate in the adventitious roots, which Fig. 45. Pedicularıs Oedert. A, Transverse section of the Are the only ones that have peripheral layers of the root. heen investigated — is not, B, Section through a leaf-scar. À ; (A about %/,; В about 2*/,.) however, so decided as in P. hirsuta. The epidermis decays early; during growth the cells of the cortex are greatly elongated in a tangential direction, and undergo divisions by thin radial walls. The portion figured in Fig. 45, A is taken from a transverse section of an adventitious root, 6 mm thick; the endodermal region lies at a depth of only 3—5 cell-layers; the Casparian dots are still recognizable here and there in the endodermis, the cells of which, as those of the primary cortex, are seen to be very greatly elongated in a tangential direction, and to have undergone secondary division. In the secondarily formed cortex the parenchyma predominates over the sieve-tissue. In all the parenchymatous cells of the root and meso- resembles that of P. hirsuta, © ye se Scrophulariaceae. 491 corme (except the outermost layer) reserve food-material was found in the form of grains, as much аз 10 in size, round, or slightly angular, and with a small central cavity. With a solution of iodine-iodide of potassum they assumed a red-violet to brownish-red colour, which disappeared on il Pian RER, ARE | ERE i: A Fig. 46. Pedicularıs Oederi. A, Transverse section of. the leaf. В and С, Epidermal cells, В, from the upper, and C from the lower surface of the leaf. D, A non-glandular hair. (A about 119],; В and С about 22%/,; D about °°/,.) heating, but reappeared after cooling; they stand out after treatment with potash-lye, and at last disintegrate altogether. Judging from the above-mentioned reactions, they must be assumed to consist of amylodextrin-starch. The structure of the above-ground stem is fundamentally as in the other species. The thickening and lignification of the wood-fibres between and outside the vessels, and of the 492 Ев. J. MATHIESEN. peripheral cells of the pith, does not appear to take place until towards the time of ripening of the fruit. The Leaf: Some of the cells of the upper and lower epidermis are shown in surface view in Fig. 46, B and C. Here, the great thickenings on the lateral walls of some of the epidermal cells of the upper surface and, as regards the lower surface, especially the walls of the cells which bear the glandular hairs, is very striking. This feature, however, is not constant in the species; as regards this point the individuals from Dovre were, on the whole, as shown in the figures in question, but the thickenings were far fainter, or even entirely wanting, in the specimens from Roras. As usual in the Pedicularis spp. the stomata occur only on the lower surface of the leaf, and the epidermal cells which surround them have highly undulating lateral walls. A transverse section of the leaf is shown in Fig. 46, A. The spongy parenchyma consists of copiously-branched cells, especially in its lower layer. As seen in the figure, the epi- dermis of the lower surface, in the spaces between the branches of the veins, is highly convex and in places without direct connection with the spongy parenchyma, so that, on the lower surface of the leaf, there are hollow spaces — air- chambers — in the roof of which the stomata occur. In the figure, a glandular hair is seen under each of the veins; two such hairs are shown in surface view in Fig. 46, C. D is a non-glandular hair. Chlorophyll-grains occur abundantly in the whole of the mesophyll, and in the epidermis of the lower surface. A section through one of the scars which the leaves at their fall leave on the rhizome, showed that here a cork of at least six layers was formed; outside the leaf-scars no cork- formation takes place. A small portion of such a section is shown in Fig. 45, B. Scrophulariaceae. 493 Summary and General Remarks. A. Account of the Growth-form and remarks on the Structure of the Shoot, the Vegetative Reproduction and the W inter-stage. I. Undershrub-Chamephytes: Veronica fruticans and alpina. The shoot-development extends over 2 years; especially in V. alpina the first-year portions of the shoots often assume the character of runners; this, in conjunction with the abundant development of adventitious roots, and the dying away of the older portions of the stem, makes abundant vegetative reproduction possible in this species. Special bud- scales do not occur. Both species may form lax tufts. Nanophyllous winter- greens. In the Arctic regions they require a snow-covering during their winter-rest; V. fruticans requires also a habitat which is freed from snow rather early in spring. II. Hemicryptophytes. a. Proto-hemicryptophytes: Castilleia pallida and Bartschia alpina. The former species has a main root of long duration; in Bartschia it dies away early, and adventitious roots are abundantly developed; this in conjunction with the forma- tion of runners allows rather abundant vegetative reproduc- tion. When Bartschia grows in comparatively dry and firm soil the formation of runners is checked and it forms tufts; in that case it evidently forms a transition to the under- shrub-chamephyte type. Nano-microphyllous. During the winter-rest the growing-point is protected by the scale-leaves on the portion of the shoot formed the foregoing summer. In the Arctic regions Bartschia (and probably also Castilleia) requires a snow-covering during its winter-rest. XXXVII. 32% 494 Ев. J. MATHIESEN. b. Rosette- or semi-rosette plants: Pedicularis spp. 1. With Runners: Р. lapponica and P. capitata. In Р. lapponica the runners are most frequently mono- podial, and during а limited number of growth-periods they form few-leaved rosettes at their apex, while the floral shoots, without any preceding vegetative stage, arise later- ally on the rosette-axis. Sometimes the runners may ter- minate their growth by the formation of а floral portion. Floral shoots may also arise laterally on the basal portion of older shoots of the same kind, so that a small rhizoma multiceps is developed. In P. capitata the runner terminates its growth after a vegetative period of 1—3 years, by the formation of a floral portion, often without foliage-leaves (a peduncle). The winter-buds of both species are protected by scale- leaves. Adventitious roots are developed and vegetative reproduction takes place. In the Arctic regions P. lapponica (and certainly also P. capitata) requires a snow-covering during its winter-rest. 2. Without Runners: P. sudetica, euphrasioides, Scep- trum carolinum, hirsuta, lanata, Паттеа and Oedert. All these species have a short vertical rhizome (“meso- corme”, Е. WARMING"); with reference to the duration of the main root, they may be divided into two groups, as stated by Е. Warmine (1890, р. 206 et seq.), viz., those in which it is of long duration, a fact connected with its vigour, and often rather copious branching (P. euphrasioides, hir- suta and lanata), and those in which it is of short duration (P. flammea and Oederi, to which may be added P. sudetica and Sceptrum carolinum). 1 Е. WARMING, 1918: Om Jordudlöbere (Underground Runners). With a résumé in English. Kgl. Danske Vid. Selsk. Skrifter. 8. Række. ILANo%G: Scrophulariaceae. 495 In those species in which the main root is vigorous and branched, the development of adventitious roots is repressed, and does not, at any rate, occur except in old individuals (cf. Tu. Resvorr, 1917, р. 215); in the species of the other group adventitious roots are more or less abun- dantly developed, certainly according to the nature of the substratum. The length of the first vegetative-stage varies, as a rule, it may be of several-years duration. The variations must be assumed to be due to the varyingly favourable conditions afforded by the habitat in question, as for in- stance, early disappearance of snow in spring. In P. lanata the winter-buds are protected by the large persistent bases of the foliage-leaves of the year’s rosette; in the other species special bud-scales occur. P. euphrasioides appears, perhaps exceptionally, to be able to occur as a biennial hapaxanth. The growth of the main shoot — in all the species which have been investigated — terminates by the development of a floral portion, fre- quently furnished with a few or several foliage-leaves, but sometimes quite devoid of these (individuals of P. sudetica and Sceptrum carolinum). The adventitious shoots arise most frequently in the axils of the scale-leaves; previous to their flowering they usually pass through a vegetative stage of some years duration. Only in P. euphrasioides the shoots usually appear to be able to achieve flowering in their second year. After fruit-setting the upper part of the shoot dies away; the complex of the persistent basal-portions forms a rhizoma multiceps (a mesocorme), which is especially large and vigorous when the main root is of long duration, as in P. hirsuta and lanata; these species also form the largest tufts. In those species in which the main root and rhizome die away quickly behind, and adventitious roots are abun- dantly developed (especially in P. sudetica and Sceptrum caro- 32* 496 Ев. J. MATHIESEN. linum) some vegetative reproduction takes place; but natur- ally by this means the plant can spread only to a very limited extent. According to M. PorsıLp’s observations P. hirsuta and lanata can dispense with a snow-covering during winter; all the other species of Pedicularis here mentioned no doubt require а snow-covering, but at the same time certainly also require that the snow in their neighbourhood shall melt rather early in spring; therefore, in the Arctic regions these species hardly belong to the real snow-trough Йога, as is for instance the case with Р. hirsuta in Norway. In southern Greenland, however, several of the species ascend rather far up the mountains (Е. WARMING, 1888, р. 87). Ill. Therophyte: Zuphrasia arctica. Germinates in spring. B. The Structure and Biology of the Flower. In the large winter-buds of Р. sudetica, hirsuta, flammea and Oederi the floral organs are found to be highly developed in the year previous to that in which they expand, and the same is probably the case in the rest of the species of Pedi- cularis. Also in Veronica alpina and Bartschia alpina floral organs have been demonstrated in the winter-buds. The corollas are small and only slightly showy in Vero- nica alpina and Euphrasia arctica; in the former species the colour is deep blue, and in Euphrasia pale lilac to white, with yellow and violet markings. Veronica fruticans has bright blue corollas, which are even as much as 14mm in diameter. Bartschia alpina, Castilleia pallida and the Pedi- cularis spp. have rather large flowers with brightly coloured corollas (deep bluish violet, red, yellow, and yellow with red or brown markings), collected into crowded and often large 3 SET ra er Cte ER | Scrophulariaceae. 497 inflorescences; Bartschia and Р. lapponica have in addition coloured bracts. Forms with white corollas have been found in Veronica alpina, fruticans, P. hirsuta and lanata. Honey-secretion, either by a secretory ring around the ovary (Veronica) or by protuberances at its base, has been demonstrated in the majority of the species. Perfume is present in Р. lapponica, sudetica, euphrasioides and lanata. The species which have been investigated varied from homo- gamous to slightly protogynous. In Veronica alpina and Euphrasia arctica self-pollination is no doubt customary. In the Arctic regions self-pollination no doubt plays an important part as regards the Pedicularis spp. since, accord- ing to several investigators (EKSTAM, PORSILD, etc.), insect- . visitors are. very scarce or even totally wanting there, and in Spitzbergen, as mentioned by AurivirLıus, and as has since often been pointed out by others, none of the natural pollinators of these flowers (humble-bees) are found; in spite of this absence, P. hirsuta and lanata regularly set fruit there. When, as in P. flammea, the stigma is always included in the helmet on account of the shortness of the style, cross- pollination by insects seems to be impossible. In P. hirsuta (no doubt frequently) and lanata (sometimes) the upper part of the style can be found to be so strongly curved that the stigma is pressed against the under-side of the anthers; per- haps this is due to movement which has taken place during a later stage in the flowering, so that originally cross-pollina- tion has been possible; in the majority of the flowers of P. lanata which have been investigated, the stigma was found to protrude slightly beyond the tip of the helmet, as also in P. sudetica. In these three species the flowers are set almost horizontally, and the stigma is therefore placed under 498 Ев. J. MATHIESEN. the anthers, so that at any rate the dry and light pollen, when it falls out of the anthers by accidental shaking of the plant, will easily be able to fall upon the almost globular stigma. Self-pollination must take place with somewhat more difficulty in Р. euphrasioides and lapponica owing to the great obliquity of the flower. The latter species, which has in addition the most protrusive stigma, is no doubt also the worst self-pollinator, and according to PorsiLp it is also the poorest fruit-setting species in Greenland. In P. Oederi, in which the flower, as in Р. flammea, stands almost in a vertical position, the upper lip curves forward and downward at the end of the flowering period, by which means the stigma is brought in a position that enables it to receive the pollen when it falls. Pedicularis Sceptrum carolinum, as Castilleia pallida, no doubt normally requires insect-visits for pollination. Bart- schia alpina is characterized by great variation in the length of its style; in the short-styled forms, especially in those in which the stigma is included in the helmet, self-pollination must easily be able to take place; any appreciable growth of the corolla tube has not been demonstrated in the Arctic regions, so the variations must be regarded as individual. On the other hand, in several instances of older flowers, the anthers were found to project outside the corolla so that in these cases we may perhaps assume with KERNER that some amount of cross-pollination takes place between older and younger flowers by the agency of the wind. With the exception of P. capitata I found, in my material, ripe fruit in connection with all the species. The seeds were small and light in all the species. Scrophulariaceae. 499 C. Parasitism. As is well-known the Zuphrasia spp., Bartschia alpina and the Pedicularis spp. are green semiparasites. As regards their capability of assimilating the carbon dioxide of the air, HEINRICHER (1910, VI, p. 574) states that in the Euphrasia spp. this is almost normal, and the same author also attributes active assimilation to Bartschia. Vot- KART (1899, p. 30) sums up his results as regards the Pedi- cularis spp. (P. recutita, foliosa and verticillata) investigated by him as follows: “Die Versuche zeigen, dass die Starke- speicherung der Blatter der untersuchten Pedicularisarten normal und ohne grosse Abweichung von denjenigen anderer Pflanzen vor sich geht. Eine weitgehende Herabsetzung der Assimilationstätigkeit durch den Parasitismus, wie sie Bon- NIER aus seinen Versuchen ableitet, findet nicht statt.” In this connection I wish to draw attention to the fact that the mesophyll, in all the Rhinantheæ spp. investigated by me, was found to be normal in structure, and its cells contained a normal supply of chlorophyll grains. Bartschia alpina according to HEINRICHER’S investiga- tions, is an obligate parasite. According to VOLKART, the Pedicularis spp. may show gradations in their dependency on other plants; from P. palustris, which he describes as “Starkster Parasit mit stark entwickelter Neigung zur sapro- phytischer Ausnützung der Nährpflanze” (by the destruction of the roots attached), to P. comosa, which is said to be a “Relativ selbeständig entwichlungsfähige Art”; as regards P. Oederi he writes “Saprophytische und hydroparasitische Ausnutzung der Nährwurzel gleichgestellt.” From the Arctic regions we lack accounts of thorough investigations regarding this matter; but as regards P. lanata there is a note to hand which might suggest that under 500 Ев. J. MATHIESEN. certain circumstances this species may at times lead an independent life, since М. Ровзил remarks (1920, р. 142) that he has found P. lanata growing “in gravelly barren, far away from other plants”; where it occurs among other plants, this species, also, has haustoria on its roots (1. с.). According to VOLKART, in the Alps the Pedicularis spp. do not appear to be very particular as regards host-plants. From the Arctic regions ROSENVINGE and WARMING records Р. hirsuta and flammea as parasitic on Vaccinium uliginosum and Salix herbacea; Hartz (1905, р. 219) writes that in East Greenland Р. lapponica “is closely connected with Betula, and probably occurs parasitically on its roots.” A rather interesting circumstance connected with the present question is recorded by Cur. KRUUSE (1911, р. 61) from the district of Angmagsalik. On some slopes facing north and cut into terraces, where a rich vegetation of “espalier-shrubs” was growing on the vertical sides of the terraces, he found the platforms to be “covered with coarse gravel which bore at great intervals (3—5 metres) wind-affected tufts of Salix (Salix glauca)’, and in these tufts a Pedicularis (hirsuta or flammea) always occurred. “Where the platforms consist of sand, Salix herbacea is always predominant, and the latter also, with its crooked branches 5—10cm long, protects a Pedicularis.” | D. Anatomy. The Root. Both the species of Veronica have a thin- walled, cuticularised exodermis; V. fruticans has a peripheral cork-formation in the root (as in the stem), whilst this 1s lacking in V. alpina. In both species the primary cortex persists a long time and, especially in V. fruticans, consists of rather thick-walled cells; the original endodermal cells are elongated during growth and undergo divisions by radial mer moto. ones à do ee hit Scrophulariaceae. 501 walls which are quite thin in V. alpina, but of the same thickness as in the other cells of the cortex in V. fruticans. In the latter species there are distinct growth-zones in the xylem-portions of the stele, whilst such growth-zones are lacking in V. арта. It is common to the roots of Euphrasia, Bartschia and the Pedicularis spp. to have the primary cortex few-layered and its cells frequently very thin-walled; root-hairs are ~ usually wanting (in some specimens of Euphrasia arctica, alone, I found a few scattered root-hairs). In the species of Euphrasia the epidermis persists a long time; according to HovELACQUE the same is the case in Bartschia, but in this I found that it may sometimes die away early. With the exception of P. Sceptrum carolinum it dies away very quickly in the Pedicularis spp. which have _ been investigated; even in roots a millimetre thick, I found the cells collapsed and partly thrust off; the outer walls of the layer below the epidermis become cuticularised, and where the growth in thickness of the root is considerable (P. euphrasioides, hirsuta, lanata, flammea and Oederi) the cells of the exodermis as well as of the other 3—6 layers of the primary cortex become greatly elongated tangentially and divided by thin radial walls. The primary cortex persists a long time. In Bartschia alpina and P. capitata a cuticularised lamella arises in the wall of the endodermal cells, in the former species all the way round, in the latter only in the inner and radial walls. The Casparian dots were very distinctly discernable in Euphrasia arctica, and in P. lapponica, whilst they were quite faint in P. hirsuta and lanata. Lacune, such as are found, for instance, in the primary cortex in P. palustris (HOVELACQUE), were, in the species 502 Ев. J. MATHIESEN. investigated, found well-developed in the adventitious roots of P. sudetica only. In the roots of P. Euphrasioides, hirsuta and lanata, which had a vigorous growth in thickness, the rather thick secondary cortex was found to be traversed by radiating clefts (Fig. 38, B); this was less well-marked in the adventiti- ous roots of Р. flammea and Oederi. In P. hirsuta and lanata in which the thin-walled parenchyma — аз is also the case in P. flammea and Oederi — ıs also dominant in the xylem, the clefts are often continued far into the latter. In the xylem of P. euphrasioides distinct growth-zones were seen, each probably corresponding to one year’s in- crement. The thick roots and rhizomes in several of the Pedi- cularis spp. certainly contain a considerable amount of reserve food material, no doubt especially during the winter- rest; the sweet taste of these organs in P. hirsuta and lanata is well-known. In Р. flammea, Oederi and in the specimen of P. hirsuta mentioned on page 462, which was collected in late summer, I found the roots full of amylodextrin-starch. The condition of the material at my disposal did not allow of a closer study of the haustoria. The Stem. As has frequently been pointed out (Hove- LACQUE and SOLEREDER) the absence of parenchyma-rays from the part of the xylem formed by the cambium, is a character widely distributed among the Scrophulariaceæ. Parenchyma-rays were also absent from all parts of the shoot in the Veronica spp., Castilleia pallida, Euphrasia and Bartschia alpina, and from the above-ground parts of the axes in the Pedicularis spp. (in P. lapponica also from the runners). In these cases the cambium very quickly develops inwardly a continuous ring of vessels and wood-fibres, which bridges over the leaf-gaps so that these become demonstrable Scrophulariaceae. 503 оп the inner side of the xylem-ring only аз thinner portions of the latter among “les faisceaux reparateurs”. A transverse section of the rhizome in P. sudetica, Sceptrum carolinum, hirsuta, lanata, flammea and Oederi showed, on the other hand, a circle of vascular bundles mutually separated by parenchyma-rays, as is commonly the case in fleshy dicotyledonous rhizomes, especially such as are abundantly covered with leaves. With respect to P.Sceptrum carolinum this feature has been already described and figured by Cuatin.t In P.capitata the runners had two rather broad diametrically opposite parenchyma-rays. In the above-ground parts of the axes in the Pedicularis spp. the peripheral portions of the pith consisted usually of thickened, lignified and porose cells; this was most decided- ly the case in P. Sceptrum carolinum, in the up to one- metre-high stem of which the stereom is, on the whole, very vigorously developed. When an elongation of the axis takes place during flowering and fruitsetting, the lignification of the stereom naturally does not begin until this is accom- plished. The central portions of the pith are very thin-walled and die away early. Hard bast was present in the periphery of the stele in Veronica fruticans, and very abundantly in the above-ground parts of the axis in Bartschia alpina; in the rhizome of the latter species the hard bast was present more sparingly, and sometimes totally absent. Continued growth in thickness through several growth- periods could be observed in the persistent basal portions of the shoots of the two Arctic species of Veronica (in V. fruticans the growth in thickness is continued throughout many years — as many аз 16 were counted), in Castilleva 1 CHATIN: Anatomie comparée des vegetaux. Plantes parasites. Dicotylédones. Р. 192. Pl. XLII. 504 Ев. J. MATHIESEN. pallida (the increment was small and irregular) and some- times in Bartschia; the runner-like portions of the shoots in Р. lapponica occasionally show two annual rings; in the rhizomes in Р. Sceptrum carolinum the xylem of the vascular bundles always showed two growth-layers, since in the second year of the rhizome-portion a group of vessels (and stereom) is developed outside that produced in the first year; the two groups are distinctly separated, some layers of thin-walled unlignified parenchyma being found between them. Thin-walled and unlignified cells are also present in the Veronica spp. at the boundary between the annual rings. Cork-development occurs sub-epidermally in Veronica fruticans in the basal, persistent portions of the shoots, and in Р. Sceptrum and capitata further within the primary cortex; the cork is always only few-layered, but P. Sceptrum carolinum develops, in addition, cork-cambiums in the paren- chyma between the annual increments in the vascular bundles and also within the pith. In the majority of cases the endodermis was demonstrable by the presence of Cas- parian dots, or else a cuticularised lamella was found in its walls, either all the way round (as in the runners in Bart- schia) or only on the inner and radial walls (runners of P. capitata). The Leaf. АП the species investigated have leaves mesomorphie in structure. The cuticle is thin; in Veronica fruticans a moderate thickening of the walls of the epidermal cells takes place, the other species have thin walls. The lateral walls of the epidermal cells of Castilleia pallida are almost straight, of the other species more or less decidedly undulating; in Euphrasia arctica and Bartschia alpina the difference between the epidermis of the upper and lower surfaces is in this respect inconsiderable; in the species of Veronica and Pedicularis I found that the epidermal cells of pe Scrophulariaceae. 505 the lower surface had the most decidedly undulating lateral walls. The lateral walls of the epidermis in Veronica alpına, Castilleia pallida and Euphrasia arctica are without sculp- ture; in the other species they are porose; in Veronica fru- ticans and Pedicularis euphrasioides, flammea and Oederi the pores are distinet and alternate with flanged thickenings, whereas they are fine in Bartschia alpina and the other Pedicularis spp. It should be noted, however, that the epi- dermal cells of the lower surface which bear the large glan- dular hairs have (except in Р. lanata) very vigorously devel- oped flanged thickenings in the Pedicularis spp., and usually also in Euphrasia and Bartschia; also their walls are, on the whole, distinctly thicker than those of the surrounding epi- dermal cells. Stomata occur exclusively on the lower surface of the leaf in the Pedicularis spp., nearly equally on both surfaces in Veronica fruticans, V. alpina, Castilleia pallida and Eu- phrasia arctica, and more abundantly on the lower surface in Veronica officinalis and Bartschia alpina. The guard-cells are on a level with, or raised only slightly above the leaf- surface; they are surrounded by 3—8 cells. The mesophyll in all cases is rather lacunose and consists of thin-walled cells. In Castilleia pallida it may be almost homogeneous; 2—3 layers of short palisade-cells occur in the Veronica spp., Euphrasia arctica and Bartschia alpina. Of the Pedicularis spp., P. euphrasioides, capitata, hirsuta, flammea, Oederi, and frequently lapponica, had 1—2 layers of typically developed palisade-cells; they were shorter and broader in P. Sceptrum carolinum, sudetica and lanata. In P. flammea and Oederi there were large air-chambers between the spongy parenchyma and the epidermis of the lower surface (cf. Fig. 46, A). In all the species investigated, 506 Ев. J. MATHIESEN. the cells of the mesophyll contained а normal number of chlorophyll-grains. In the majority of the cases chlorophyll- grains were also present in the epidermis of the lower surface. Only Р. lapponica, Sceptrum carolinum and flammea appear to be totally devoid of non-glandular hairs. The leaves were densely covered with hairs in Castilleia pallida, Euphrasia arctica, Bartschia alpina, P. hirsuta, lanata (in the last-mentioned two species especially along the edges of the petiole on the part near the base of the leaf) and capitata (on the lower leaf-surface and the petiole). The non-glandular hairs are from one to several-celled; in Veronica fruticans and Pedicularis capitata they are rather thick-walled, in the others thin-walled, smooth or with fine cuticular striations; only Veronica fruticans had hairs with coarser cuticular warts. Glandular hairs were present in all the species: small glandular hairs with a one-celled stalk and two-celled head in the Veronica spp., Castilleia pallida and in especially great numbers in Euphrasia and Bartschia alpina; the last-men- tioned two species had in addition large, peltate glands with a low sunk stalk-cell and a 2—4 celled flat head, like those in Lathrea (SCHERFFEL, HABERLANDT and GOEBEL; com- pare my Figs. 14 and 19). As previously mentioned by Hove- LACQUE and Percy Groom, large glandular hairs of peculiar structure occur also in the Pedicularis spp.; here the low stalk-cell is not sunk, and the head is 2-celled (see for in- stance Fig. 23, A and C). All the species of Pedicularis which have been investigated had such glands. The above-mentioned special forms of large glands are found only on the lower surface of the leaf and, as pointed out by the above-mentioned authors and as may be seen from my figure 24 and from most of the transverse sections of leaves figured, they are localised under the vein-branches Peer Scrophulariaceae. 507 of higher order which at their tips consist only of a few thin- walled tracheids. According to information to hand it seems unquestionable that these glands are organs for the secretion of water. As to the Pedicularis-glands it should be noted that special water-paths from the tracheids to the basal cell of the glandular hair, limited by the cuticularisation of cell- walls, as have been demonstrated in P. palustris by Percy Groom, have not been able to be demonstrated in any of the species investigated here. A bursting and crumbling- away of the cuticle on a small circular spot in the middle of the glandular head, such as the author in question men- tions in connection with P. palustris, was very common in P. lapponica and Р. Sceptrum carolinum, and was occasion- ally also found in other species. Glandular hairs with а 2 to several-celled stalk and a 2—4 celled head were present in Castilleia pallida, Euphrasia arctica, Bartschia alpina, and also in P. flammea. Ann ес та й \ i in Vp een à | By | YT wh Uae (eh É fe i i th | р 1 { a), ) u im ur Ai р А RD ‘à AN TET en Vee | / ” у FÅ - и Ean ANT ONE et wie TS Hal ors ari i | uw Г AA Wa ОА ER м if мы ad Mt вое ‘a Nat SLA FAIR 2 I YA Ta у MÅ RN MAN ога pee næ . ! | \ Г | , р 414 N 4 АА ee LE hh ы | DE REQUIS LE a she} hf | NCA МУ у : in р 3 À Г | i ål 5 } Ud I ; | 4 Re LETTER NUT. есь 1 WENN ike } re ity i MIRE TANT vw 1 3 d р $ (ng Ted м | u Е € IE | м д J : К A) | mir i ” wil; ter fl} ip, RAL at rå + | ss i å i Ny a2 q a ri ae! A м м {pa i | HP J т 1 Ca . : h Es TO TON x Г ti CP i и) j i ’ } | Au, Ge uf? Meddelelser om Gronland, udgivne af Commissionen for Ledelsen af de geologiske og geographiske Undersøgelser i Grønland. #7 Syv og tredivte Hefte. Kjøbenhavn. C. A. Reitzel, Boghandel. Bianco Lunos Bogtrykkeri. 1921. Pris: 12 Kr. , + «®.; så FORTEGNELSE over de af KOMMISSIONEN FOR LEDELSEN AF DE GEOLOGISKE OG GEOGRAFISKE UNDERSØGELSER I GRØNLAND udgivne MEDDELELSER OM GRØNLAND udkommer som Regel 1 Gang aarlig og faas portofrit tilsendt ved Henvendelse til Hovedkommissionæren С. А. Reitzel, Boghandel, Løvstræde 7, København K. Ws mer] De ei rial DT 02964 CEE LC < à к ae i =», « ы = See & CC EE НЕС << CE <<< < DES FA A AA АА AR AA AA) AAA AMIN ? AA A Ar \ AINA AS x DAR Nn AAR Ve Fole an < 4 | E N | Se i | 2 ' + ы re ee < = d 2 a > ae re - ” x > С a EN eed QU € са ot ER < СК <=: TETE Cl CAN LE = > d « AGT KE LES EE 5 et. "ar pce ARE NE 4 xan < 5, SEC" i = CC KER, a 6 ö Жк < rams i ER | \ CE QO Ох QUES, ECC re CC Oe EE LCR LIA A р AA AR AAR a A Le EC : ANY, Aw RP I \ < с S a. : gts SRE ER «Са 4 zur, < a : А À Ye