iili k 1 i'lliinl iji'ii liHiiiiHiiHiii m ill, M ®tj^ E 1. Hill IGtbrary Nnrtb (Earoltna ^latp (ToUpg? QK45 K4 1902 V.2 NORTH CAROLINA STATE UNIVERSITY LIBRARIES "'||l|||||l||l||f llllll" S01131251 E 5136 This book is due on the date indicated below and IS subject to a fine of^^WE CENTS a day thereafter. * •. iAY 3 aS4 MAY 2 0 1964 DEC 3 0 1974 THE NATURAL HISTORY OF PLANTS The Natural History of Plants Their Forms, Growth, Reproduction, and Distribution From the German of the late ANTON KERNER von MARILAUN By F. W. OLIVER, MA, D.Sc, With the Assistance of LADY BUSK, B.Sc. and Mrs. M. F. MACDONALD, B.Sc. With about Two Thousand Original Woodcut Illustrations VOLUME II The History of Plants LONDON BLACKIE & SON, Limited, 50 OLD BAILEY, E.G. GLASGOW AND DUBLIN 1902 .f/3^ CONTENTS OF VOLUME SECOND. INTRODUCTION. Sources of a History of Plants, - 1 I The Language of Botanists, Page 3 THE GENESIS OF 1. Asexual Reproduction. Spores and Thallidia, 8 Buds on Eoots, - 25 Buds on Stems, 28 Buds on Leaves, ------ 37 2. Reproduction by means of Fruits. Definition and Classification of Fruits, - 46 Fertilization and Fruit-formation in Crypto- gams, 49 The Commencement of the Phanerogamic Fruit, - - - ... 70 Ovaries on a Conical Receptacle, - - 75 Ovaries on a Flat or Excavated Receptacle, 76 Stamens, ------- 85 Pollen, 95 Protection of Pollen, 104 Dispersion of Pollen by the Wind, - - 129 Dispersion of Pollen by Animals, - - 152 Allurements of Animals with a View to the Dispersion of Pollen, - - - - 167 PLANT-OFFSPRING. ! The Colours of Flowers as a means of At- tracting Animals, - - - - 182 The Scent of Flowers as a means of Attract- ing Animals, 198 Opening of the Passage to the Interior of the Flower, 2(i9 Reception of Flower-seeking Anima 13 at the Entrance to the Flower, - - - 221 Taking up of Pollen by Insects, - • - 243 Deposition of Pollen, 27G , The Crossing of Flowers, .... - 287v Autogamy, - - - - - - - 331 Fertilization and Formation of Fruit in Phanerogams, 401 3. Change in Reproductive Methods. Fruits Replaced by Oflfshoots, ... 452 Parthenogenesis, 463 Heteromorphism and Alternation of Genera- tions, * 469 THE HISTORY OF SPECIES. 1. The Nature of Species. Definition of Species, 486 The Specific Constitution of Protoplasm, - 487 2. Alteration in the Form of Species. Dependence of Plant Form on Soil and Climate, - 495 The Influence of Mutilation on the Form of Plants, 514 Alteration of Form by Parasitic Fungi, - 518 Alteration of Form by Gall-producing Insects, 527 Genesis of New Forms as a Result of Crossing, 554 3. The Origin of Species. The Genesis of New Species, - - - 576 Derivation of Existing Species, - - - 595 Subdivisions of the Vegetable Kingdom, • 600 Classification of Plants proposed by the Editor, - 616 /3^ CONTENTS. 4. The Distribution of Species. Page 790 The Distribution of Species by Oflfshoots, - The Dispersion of Species by means of Fruits and Seeds, 833 Limits of Distribution, Plant Communities and Floras, - 5. The Extinction of Species, - Page - 878 - 885 - 899 CLASSIFICATION OF PLANTS. Phylum 1. — Mtxothallophtta. Alliance 1. — Myxomycetes, - - - 618 Phylum 2. — Thallophyta. Class I.- Schizophyta, .... 820 Alliance 2. — Cyanophvceje, the Blue-green Algse, - - -' - - - - 621 Alliance 3. — Schizoraycetes, the Bac- teria, ------- 622 Class II. — Dinoflagellata, Peridinese, - - 625 Alliance 4, 625 Class III. — Bacillariales, - - - - 625 Alliance 5, 625 Class IV. — Gamophyceae, - - - - 627 Sub-class I. — Chlorophyceae, the Green AlgEe, ------- 627 Alliance 6. — Protococcoideae, - . - 628 Do. 7.— Siphonese, - - - - 641 Do. 8. — Confervoidese, - - - 648 Do. 9.— Conjugatae, - - - - 654 Do. 10.— Charales, - - - - 659 Do. 11. — Phaeophycese, ... 661 Do. 12.— Dictyotales, - - - 666 Do. 13.— Florideae, Eed Sea-weeds, - 666 Class V.— Fungi, 668 Sub-class I. — Phycomycetes, - - - 668 Alliance 14. — Oomycetes, ... 668 Do. 15. — Zygomycetes, - - - 673 Sub-class II. — Mesomycetes, - - - 674 Alliance 16, 674 Do. 17, 674 Sub-class III. — Mycomycetes, - - - 676 Alliance 18, 676 Do. 19, 684 Additional Group of Fungi, Lichenes, - 691 Phylum 3. — Archegontat^. Class I.— Bryophyta, 696 Alliance 20. — Hepaticae, Liverworts, - 696 Do. 21. — Musci, Mosses, - - - 699 Class II. — Pteridophyta, Vascular Crypto- gams, 704 Alliance 22.— Filices, Ferns, - - - 705 Do. 23. — Hydropterides, Rhizocarps, 709 Do. 24. — Equisetales, Horsetails, - 711 Do. 25. — Lycopodiales, Club-mosses, 713 Phylum 4. — Phanerooamia, Flowering Plants. Sub-phylum A.— Gymnospermae, - - 718 Class I. — Cycadales, Cycads, • - - 718 Alliance 26, 718 Class II.— Coniferse, 720 Alliance 27, 720 Class III. — Gnetales, . . . . 726 Alliance 28, 726 Sub-phylum B. — Angiospermae, - - 728 Class I. — Monocotyledones, - - - 728 Alliance 29.— Liliiflorae, - - - - 728 Do. 30. — Scitamineae, - - - 734 Do. 31.— Gynandrge, - - - 736 Do. 32. —Fluviales, - - - - 738 Do. 33.— Spadiciflorae, - - - 740 Do. 34.— Glumiflorae, - - - 745 Class II. — Dicotyledones, - - - - 748 Sub-class I. — Monochlamydeae, - - •• 748 Alliance 35. — Centrospermae, - - - 748 Do. 36.— Proteales, - - - - 750 Do. 37.— Daphnales, - - - 752 Do. 38.— Santalales, - - - 754 Do. 39.— Eafflesiales, - - - 755 Do. 40. — Asarales, ... - 755 Do. 41.— Euphorbiales, - - - 756 Do. 42. — Podostemales, - - - 757 Do. 43.— Viridiflorse, - - - 758 Do. 44. — Amentales, ... 762 Do. 45.— Balanophorales, - - - 762 Sub-class II. — Monopetalse, ... 763 AUiance 46. — Caprifoliales, ... 763 Do. 47. — Asterales, . - - - 765 Do. 48. — Campanales, - - 767 Alliance 49 — Ericales, - Do. 50. — Vacciniales, Do. 51.— Primulales, Do. 52. — Tubiflorae, Class III.— Polypetalae, Alliance 53. — Eanales, - Do. 54. — Parietales, Do. 55. — Malvales, - Do, 56.— Disciflorse, CONTENTS. vii Page 768 Alliance 57. — Crateranthse, Page - 779 770 770 Do. Do. 58.— Myrtales, - 59. — Melastomalcs, - - 781 - 783 771 Do. 60.— Lythrales, - - 784 774 Do. Do. 61.— Hygrobiae, 62.— Passiflorales, - - 784 - 785 774 774 Do. Do. 63.— Pepones, - 64.— Cactales, - - 785 - 786 776 Do. 65.— Ficoidales, - 787 777 Do. 66.— Umbellales, - 788 ILLUSTRATIONS. FROM ORIGINAL DRAWINGS BY E. HEYN, E. v. RANSONNET, J. SEELOS, F. TEUCHMANN, 0. WINKLER, AND OTHERS. Page 11 14 15 Ferns and their spore-cases, &c., Horse-tails (Equiseta), . - - - Mosses and their fructiiication, - Swarm-spores of Saprolegniacese and Chytri diacese, Moulds and their fructification, - Discomycetes (Morel and other Fungi), Basidiomycetes (various Fungi), Thallidia of Muscinese or Mosses, Formation of Thallidia in the cells of Hydro dictyon and in those of Pediastrum, - Helwingia rusciflora, with flowers seated upon the foliage-leaves, . - - - - Formation of Buds on the apices of the Fronds of Ferns: Asplenium Edgeworthii, • Formation of Buds on Fronds and Foliage-leaves, Fertilization and fruit-formation in Ulothrix zonata (partly after Dodel-Port), Fucus vesiculosus (Bladder-wrack), various sec tions, ...--- Fucus vesiculosus, do. do., Fertilization and fruit- formation in Mucorini, Siphonacece, and Floridece, - Fertilization, fruit-formation, and spore-formation in the Peronosporese, . - - - Fruit- formation in Stoneworts (Characeis), Structure of Phanerogamic Ovaries ; various ex amples in section and otherwise, - Structure of Phanerogamic Ovaries : various ex amples in section and otherwise, - Structure of Phanerogamic Ovaries: various ex amples in section and otherwise, - Structure of Phanerogamic Ovaries: various amples in section and otherwise, - Antholysis and Structure of the Ovary, Ovules and Foliaceous Carpels, - Stamens of double and monstrous flowers, - Stamens, numerous examples, - Curved anthers in the flower of Phyllanthus Cy clanthera (after Baillon), - - - idtamens, numerous examples, - Pollen-grains, fourteen varieties, Pollen-grains, nine varieties, Pollen-grains and pollen-tetrads united by threads of viscin, ------- 101 41 50 51 73 Protection of Pollen from Wet in Sea-buckthorn, &c., Protection of Pollen from Wet in Globe-flower Digitalis, &c., ----- Protection of Pollen from Eain,- Protection of Pollen in Crocus multifidus, - Protection of Pollen in Carlina acavZis, Protection of Pollen in Herb-Robert, Bell-flowe Scabious, ------ Protection of Pollen in various plants. Flowers of Vallisneria spiralis floating on the surface of water, . - . - The common Alder {Alnus glutinosa) and its flowers, ------ The Paper Mulberry-tree (Broussonctia papyri fera), ------- The Ash {Fraxinus excelsior) and its flowers. Flowers of Avena elatior, - - - - The Elm (Ulvius campestris) with its flowers anc seed vessels, -.-.-• Flowers, &c., of Mountain Pine (Pinus Pumilio) Male Flowers of Yew (Taxus baccata), Hazel (Oorylus Avellana) with flowers and fruits. Curled Pondweed {Potaviogeton crispus) in the act of pollination, . - - - - Flowers of Arrow-grass (Triglochin palustre), The Nottingham Catchfly (Silene nutans) in tl: daytime, .-.--- The Nottingham Catchfly (Silene nutans) by night a flower being visited by a moth. Transport of Pollen by Egg-laying Insects, Arum conocephaloides, with the front wall of th spathe removed, - - - - - Flower of Birthwort (Aristolochia ringens), Honeyless Flower of Argemone. Mexicana with abundant pollen, . . - - Flowers of the Snowflake (Leucojum vernwu). Honey-secreting tissue in flowers, Nectaries in several flowers, ... Flower of the Snowdrop (Galanthus nivalis), Flower of Narcissus (Narcissus Pseudonarcis sus), ------- Flower of the Wild Valerian ( Valeriana officinalis), cut through longitudinally, - Concealment of Honey in Flowers, 110 112 113 117 121 125 132 135 137 133 139 143 144 145 147 148 149 155 157 164 166 168 170 173 174 176 177 178 180 ILLUSTRATIONS. Paga Concealment of Honey in Flowers, - - - 181 Examples of Colour-contrasts in Flowers, - - 184 Two New Zealand Haastias or" Vegetable Sheep" [Haastia pulvinaris and Sinclairii), - - 188 Colour-contrast in the flowers of the Bean ( Vicia Faha), 189 Narcissus {Narcissus poeticus), showing colour- contrast, - - - - - - - 190 Preparation of Flowers for Insect-visits in the Laburnum {Cytisus Laburnum), • - - 223 Arrangements for the reception of Insects at the entrance to the Flower, .... 226 Arrangements for the reception of Insects at the entrance to the Flower, .... 227 Wood Anemone (Anemone nemorosa), - - 229 American Dogwood (Cornus florida), - - - 231 Sticky glands as a protection to Flowers, - - 236 Sticky Bristles at the edge of the Calyx as a pro- tection to Flowers, 237 Tufts of Hair as a Protection to Flowers, - - 240 Capitula of Serratula lycoptfolia protected by Ants from the attacks of a Beetle, - - - 242 Contrivances for loading insects with pollen, - 246 Longitudinal section through a flower of the Evening Primrose {(Enothera biennis), - 247 Contrivances for ensuring the deposition of pollen on insect-visitors, ..... 249 Withdrawal and deposition of pollinia in the flowers of an Orchid, ... - - 255 Clip-mechanism for fastening the pollinia of As- clepias Cornuti to the feet of insects, - - 258 Apparatus for pumping pollen on to the bodies of insects, ....... 261 Transference of pollen to the bodies of insects by means of mechanism of the percussive type, 262 Explosive apparatus for the transfer of pollen to the bodies of insects, ..... 265 Explosive apparatus in a papilionaceous flower, - 266 Transference of pollen to the bodies of insects by means of explosive apparatus, - - - 267 Expulsive apparatus in Orchid-flowers : flower of Catasetum tridentatum, - , - - - 269 Flower of Pedicxdaris recutita, .... 272 Sprinkling apparatus of various plants, - - 273 Sprinkling apparatus, ..... 275 Arrangements for the Retention of the deposited Pollen, 279 Deposition of the Pollen in Mimulus luteus, • 280 Evening Primrose (CEnothera biennis), - - 282 Calabar or Ordeal Bean (Physostigma venenosum), 285 Types of the 1st to 10th classes of the Linnean System, 289 Types of the 13th, 14th, 15th, 16th, 18th, and 20th classes of the Linnean System, - - ■ 292 Types of the 11th, 12th, 17th, and 21st classes of the Linnean System, .... - 293 Type of a monoecious plant: Common Oak (Quercus pedunculata), .... 298 Tyi>e of a dioecious plant: Crack Willow (Salix fragilis), Heterostyled flowers : Primrose and others, Change of Position of Anthers and Stigmas. Flower of the Rue {Ruta graveolens), - Completely dichogamous Flowers [Geranium. Parietaria), .... Dichogamy in Saxifraga rotundifolia, Incompletely dichogamous Flowers {Epilobium, Eremurus), .... Geitonogamy with adherent pollen, - Geitonogamy with dust-like pollen, . Autogamy effected by the inclination of curved stamens, ....... Autogamy effected by inclination of curved stamens {Circaa, Agrimonia), Autogamy brought about by elongation of the the pistil (Epimedium alpinum), - Autogamy effected by means of an inflection o the style, ...... Autogamy in the flowers of the Willow-herb (Epilobium angustifolium), - Autogamy by means of spiral twistings of stamens and style, ..... Autogamy by means of a crossing or a bending back of the style-branches, - 1 Autogamy effected by means of an inflection o: the style-branches (Arnica, Senecio), - Autogamy effected by tlie petals, Autogamy effected by means of the corolla (Gen tiana asclepiadca), ..... Autogamy effected by means of the corolla (Pedl cularis incarnata), ..... Autogamy caused by inflection of the flower-stalk and the adjustment of the under-lip to form an inclined plane down which pollen slides to the stigma: Calceolaria Pavonii, Autogamy caused by the combined inflections of pedicel and stamen-filaments: Pyrola unijora, Autogamy through inflection of the pedicel and disarticulation of the corolla : Phygelius capensis, Autogamy resulting from an inflection of the pedicel accompanied by spiral torsion of the filaments: Cobcea scandens, .... Autogamy resulting from inflection of the pedicel combined with inclination of the style to the place where the pollen has been deposited: Alliuvi Chamcemoly, ..... Autogamy resulting from inflection of the pedicel combined with the folding up of the corolla : Gentiana Clusii, ...... Development of Pollen. tubes in Lilium Martagon and Averia elatior, ..... The course of the pollen-tubes in a Rock-rose (ffelianthcinuin marifoliuin), Chalazogamic fertilization in the Hornbeam (Car- pinus Betulus), ...... Page 299 302 305 306 307 308 309 320 342 343 349 851 354 357 360 363 369 372 375 388 409 411 412 ILLUSTRATIONS. Page Chalazogamic fertilization in the Alder {Alnus glutinosa). Diagrammatic, - - - 413 Fertilization in Ephedra and Ornithogalum, - 415 Embryo-sac of Monotropa, • - - - 417 Seeds with a Reserve-tissue, .... 422 Seeds with winged and hairy appendages, - - 423 Polar Willow {Salix polaris) with opened fruits showing masses of hairy seeds escaping, - 424 Seeds with caruncles and hilar scars, - - - 425 Branch of Mezereon (Daphne Mezcreum) with berries. Fruiting branch of the Lime [Tilia) with downy hairs investing the nut-like fruits, 426 427 428 429 431 432 433 434 435 436 437 438 439 Indehiscent fruits and schizocarps, Winged fruits or samaras of the Ash, &c., - Flowering branch of Banksia serrata with thick walled dehiscent capsules, - - - . Various capsular fruits, Achenes provided with a plume or pappus, The Hornbeam (Carpinus Betvlus) in fruit. Fruits with persistent receptacles, Fruits with Cupules: Acorns, .... Fruits in whose structure the receptacle and pedi eel take a share: Cashew-nut, Collective and aggregate fruits: Betel, Sweet-sop, and Sour-sop, ..---- Branch of the Bread-fruit Tree (Artocarpus in- cisa), with flowers and fruit, The Lotus Lily [Nelumhium speciosum), Flower of the Lotus Lily {Nelumhium speciosum), 440 Fruit and Seed of Coniferae: Silver Fir, Scotch Pine, Larch, ...... 441 Fruits and Seeds of Conifers: Yew, Arbor Vita, Juniper, ....... 442 Coniferous Fruits and Seeds, .... 443 Protection of ripening seeds against animals: Anatto, ....... 444 Protection of ripening seeds against animals : Mimosa, Chestnut, &c., - - - - 445 Protection of seeds against wet, .... 448 Branch of Mangrove Tree {Rhizophora Mangle), with flowers and fruits, - - • - 451 Bulbils replacing flowers and fruits, - - - 455 Flowers and fruits replaced by tubers and bud-like offshoots, 460 Flowers and fruits replaced by bulbils. The Coral-root (Dentaria bulhifera), - 461 The Annual Dog's Mercury [Mercurialis annua), 466 Alternation of Generations in Ferns, - - - 472 Tree-ferns (Alsophila) in Ceylon, - - - 473 Rhipidopteris peltata, with sterile and fertile fronds, ....... 474 Stag's-hom Fern (Platycerium alcicorne), - - 475 Alternation of Generations in Mosses, - - 477 Alternation of Generations in Mosses, - - 479 Asexual and sexual reproduction in Saprolegni- aceae, ........ 480 Asexual and sexual reproduction in the Mucorini, 481 Larch trees (Larix europcea), - . - - 483 Myxomycetous Fungi and their fructification Magnified Examples of Desmidieae, - Fungus-galls on Juniperus coramunis and Aronia rotundifolia, .... Various Galls and their effects on plants, A Witches' Broom on the Silver Fir, produced by ^Ecidium elatinum, Galls in section and otherwise, - Various Galls on Leaves, .... Galls on Spruce, Poplar, Oak, &c.. Solid Galls on Lime and Celastrus, Various Oak-galls, ..... Spirophyton from the Upper Devonian Rocks, Riella helicophylla growing under water, Myxomycetes or Slime-fungi, Various Examples of Bacteria, - Highly-magnified Examples of Diatoms, Hydrodictyaceae, with Zoospores, Ulothrix zonata, with Gametes and .Zoospores, Various Examples of Desmids. - Spirogyra, • Structure and reproduction of Chara fragilis, Laminariaceae, with perforated fronds, Fucus vesiculosus (Bladder-wrack) : female con ceptacle, &c., ...... Fucus vesiculosus: male conceptacle, &c., . A branch of the Gulf -weed, Sargassum baceiferum with leaves and air-sacs, ... Chytridiaceae and AncylistaceiE, Swarm-spores in Saprolegniaceae and Chytridiaceae The False Vine-mildew, Peronospora viticola, Achlya lignicola and its Reproduction, Entomophthoreae: Entomopjhthora and Empusa, Mucor and its Reproduction, Ascomycetes, ..... The Ergot of Rye, Claviceps purpurea, Various Ascomycetes, The Morel and other Discomycetes, - Portion of a lamella of an Agaricus with a dial layer, . . . . - Basidiomycetes: various species of Fungi, Gasteromycetes: Puff-ball, Earth-star, &c. Lichens, ...... Gelatinous Lichens, .... Crustaceous Lichen: Lecanora esculenta, Vertical section through an air-chamber of the Liverwort Marchantia polymorpha, Jungermanniaceae: Scale-mosses, Mosses and their Spore-capsules, &c. Mosses, with details of structure, Spore-capsules of Mosses, - Various Ferns, .... Life-history of a Fern, Hydropterides or Rhizocarps, - Equisetacese or Horsetails, Club-moss: Lycopodium annotinum, Lycopodiales, with details of structure, A group of Cycas revoluta, Pase- 491 492 ILLUSTRATIONS. 737 739 741 Page Temale Cone and Scales in Abietineas, - - 721 Flowers, &c., of Mountain Pine (Pinus Pumilio), 722 The Scotch Pine (Pinus sylvestris), - - - 723 The Arolla Pine (Pinus Cembra), • - - 724 Wehcitschia mirabilis in its natural surroundings, 727 Liliiflorae: Examples of the Lily Alliance, - - 729 Asjjhodelus ramosus at Paestum (Southern Italy), 730 The New Zealand flax (Phormium tenax), - - 731 Adam's Needle ( Yucca gloriosa), - - - 732 ^chmea paniculata (family Bromeliacese), - - 733 The Traveller's Tree (RavenalaMadagascariensis), 735 Angrcecum eburneum, an Orchid epiphytic on a tree-trunk, --.--.. ■Curled Pondweed (Potamogeton crispus), Primeval forest in Ceylon with Climbing Palms (Calamus) and Areca disticka, Aroids (Arum Tnaculatum, Colocasia antiquorum, &c.), 742 Eaphidophora decursiva climbing in a primeval forest of the tropical Himalayas, - - 743 Climbing Aroids (Philodendron pertusum and P. Imbe) with cord-like aerial roots, - - 744 Papyrus Plant (Papyrus antiquorum) in the Upper Nile, 747 Marvel of Peru (Mirabilis Jalapa), - - - 749 Proteales: Flower and Fruit of Banksia, - - 751 Daphnales: Camphor, Flowers of Cinnamon and Mezereum, .--.--- 753 "Living bridge" formed of the aerial roots of the India-rubber and other kinds of Figs in Sikkim-Himalaya, ..... Amentales: Catkins of Birch and Hornbeam, Flowers and Fruit of Oak (Quercus sessilijlora), - The Common Beech (Fagus sylvatica), ■Caprifoliales : Ipecacuanha (Cephaelis Ipecacu- anha), ....... Ericales : Flowers and Fruit of Strawberry-tree (Arbutus Unedo), ..... Acanthacese: Acanthus mollis, • Eanunculaceae : Black Hellebore and Mousetail The Indian Lotus (Nelumbium speciosum), grow- ing in a marsh, near Pekin, Parietales: sections of flower-bud and ovary, Disciflorae: Euonymus, Quassia, &c., - Crateranthas: sections of flowers, Myrtales: Flower of Melaleuca; Flower-bud and Fruit of Eucalyptus, - Melastomaceae: Melastoma Malabatliricum, Cactacese: several species of Cactus, - Umbellales: Umbellifers with Fruit and Flowers, Fairy Rings in a meadow near Trins in the Tyrol, formed by the ascomycetous fungus Spathu- laria flavida, ...... Plants with tubers and bulbs whose mode of growth leads to the formation of colonies arranged in lines and clusters, 758 759 760 761 764 769 772 773 775 776 778 779 782 783 787 789 r91 796 Page A section through soil permeated by the proto- nemal threads of the Moss Pottia intermedia, 799 Formation of a clustered colony by means of aerial runners in Saxifraga flagellaris, - 801 Frogbit (Hydrocharis Morsus-rance): winter buds in process of detachment, .... 804 Frogbit : winter buds rising to the surface and young plants developed from them, - - 805 Trichia clavata, showing capillitium and spores, - 813 Dispersal of spores by wind : spore-capsules in damp and in dry weather, - - - - 814 Spores of the Horse-tail (Equisetum Telmateja), - 815 Polygonum viviparum and its aerial tubers, - 819 Houseleek (Sempervivum soboliferum) with ball- shaped offshoots, ..... 821 The formation of offshoots in Sedum dasyphyllum, 822 The formation of offshoots in Kleinia articulata, 823 Distribution of spores by expulsive mechanisms, - 825 Distribution of detached sprout-like offshoots by means of animals, 829 Sling-fruits: Ecballium elaterium, Oxalis Accto- sella, 834 Sling-fruits: seven examples, .... 835 Catapult fruits: five examples, - . - - 841 Fruits which creep or hop along the ground, - 843 Creeping and hopping fruits, .... 844 Fruits which open upon being wetted with water, ------- 845 Examples of fruits and seeds dispersed by the wind, 843 Fruits and seeds dispersed by the wind: Plantago Cretica, 849 Fruits and seeds dispersed by the wind, - - 852 Fruits and seeds dispersed by the wind : thirteen examples, --..--- 853 Dispersion of fruits and seeds by the wind : ten examples, ------- 854 Fruits and seeds dispersed by the wind, - - 855 Fruits and seeds dispersed by the wind (Cotton, &c.), 856 Fruits and seeds dispersed by the wind (Groundsel, Dandelion, &c.), 857 Fruits and seeds dispersed by the wind, - - 858 Seeds of the Orchid Vanda teres, - - - 859 Dispersion of fruits and seeds by the wind (Thistle), 860 Dispersion of fruits and seeds by the wind ( Vanda, TUlandsia), §62 Examples of sticky fruits, 870 Fruits which hook on to or stick into passing objects, - 8^1 Fruits furnished with hooks, .... 873 Fruits with hooks, 874 Fruits with needle-like spines, - - - - 875 Bamboo Forest in Ceylon, §90 Mangrove Forest in India, - - • -891 THE NATURAL HISTORY OF PLANTS. INTRODUCTION. Sources of a History of Plants. — The Language of Botanists. SOURCES OF A HISTORY OF PLANTS. From the sixteenth to the latter part of the eighteenth century, " Historia plant- arum" was the customary title for botanical works. Most of the scholars of that time took as their authorities and models the writings of Theophrastus, the cele- brated pupil of Aristotle, together with the thirty-seven books constituting Plin3^'s " Historia naturalis". Thus it came about that the titles of the new books were similar to those of Theophrastus and Pliny. However, all these books are anything but histories of plants, if in the idea of a history we include an account of the changes which occur within the limits of space and time. In reality the bulky folios of Clusius, Bauhin, and Haller, the title-pages of which bear the inscription "Historia plantarum", contain descriptions merely of the external characters of plants, accompanied by only sparing details of the situations in which these plants had been found growing wild. Works of this kind, dealing with limited areas of country, were later on distinguished by the name of Floras. By this name they are still known. Although the authors of the Flora had no such purpose in view, their Avorks furnished the starting-point for a real history of the vegetable world. A com- parison of the Floras of neighbouring regions shows that certain plants inhabit a greater, others a lesser area; that the boundaries of the species confined to a distinct district coincide with territories inhabited by various races of mankind; that the boundaries of this and that species coincide and stand in relation to various climatic and other conditions. All plants have the power of pi'opagating themselves. They send their oftspring forth as colonists towards all points of the compass, and endeavour in this way to enlarge their areas of distribution. Suppose that a species hitherto subsisting in localities where there are seven months of snow and five months of vegetation in the year multiplies, and that its descendants are scattered in all directions, what would happen if any of these emissaries reached places where frost and snow Vol. II. 2 SOURCES OF A HISTORY OF PLANTS. prevail for eight months instead of seven, and where the season for vegetation is confined to four months? They would succumb to the inhospitality of the climate; and it follows that a limit to the distribution of the species in question would be attained at a line connecting all places which possess a climate of equal rigour. This does not preclude the possibility of other causes constituting a barrier to the distribution of the same species in other directions. Peculiarities of soil, for instance, may prevent the naturalization of a plant; or, its spread may be baffled by the opposition of plants already long settled in the place invaded; or any other like impediment may operate as a check. Facts of this kind, being brought to light by the comparison of different Floras, led to detailed research into the means of reproduction and distribution in plants, to a study of the many contrivances for their propagation, and of the nature of the equipments which enable the descendants of a stock to enlarge the area where it grows. Side by side with these investigations into the history of individual kinds of plants, there was developed a special department of research with the view of determining the actually-existing boundary-lines of distribution — the so-called lines of vegetation — of particular species, and of ascertaining all the conditions of soil and climate afiecting plant-life which prevail along these lines, so as to take into consideration all the possible causes of limits to distribution. The range of observations was likewise extended to displacements of the lines of vegetation, to the advance of particular species in one direction or another, and the suppression and annihilation of others within historic times; thus a chronicle of plant migration was started. The unlooked-for discovery of the multitude of plants which flourished upon the earth ages ago, and have been preserved as fossils, led to a further comparison of forms — viz. of those now living with those that have perished. There was no evading the idea that existing species are derived from others now extinct; on the contrary it proved so attractive that it was followed up with the greatest interest and zeal. Then these inquiries into the parentage of species naturally led further to the whole problem of their origin — in short, to a study of the history of species. The range of vision continued to become yet wider. It is impossible that the dwarf willows and birches found living in Greenland at the present day should be descendants of the maples and beeches which grew there in the Tertiary Period, or that the alders or pines now flourishing on the soil above the beds of bituminous coal at Haring in Tyrol should have sprung from the Proteaceae and Myrtaceae which formerly covered the same ground, as we learn from the fossil remains found there. Local changes must have taken place, and the various floras must have undergone a process of expatriation on a large scale not unlike that of men at the time of the migration of tribes. New realms were then occupied by those floras in a manner corresponding to the formation of states by the struggling and ming- ling races and nations of mankind. The knowledge of the fact that a plant's form depends at the present day upon soil and climate entitles us, moreover, to infer that a similar connection existed in past times between the forms of plants and their THE LANGUAGE OF BOTANISTS. 3 conditions of life, and enables us to discover what gave rise to migrations and caused the redistribution of floras. These phenomena are the subject-matter of the History of the Plant World in the fullest meaning of the phrase; and their explanation is eagerly sought by modern botanists. In 1853 linger, to whom all branches of Botany were equally familiar, made the first attempt at such a history of plants. Since then a great number of new discoveries have been made both in the Old World and the New. Men with minds intent upon this object are everywhere searching for the fossil remains which throw such valuable light upon it; but, so far, this — the most recent branch of Botany — has not led to a comprehensive result. We find ourselves, as it were, in the midst of a stream in full flood owing to the number and magnitude of its tributaries, and it is no easy matter to steer clear of shoals and run safely into harbour. Some decades hence it may perhaps be possible to write an accurate and complete history of the plant world founded upon the mass of authentic evidence which will by that time have been winnowed from the records of past ages. At present I must content myself with sketching in general, and often ill-defined, outline the changes which take place in the world of plants. The foregoing introductory remarks concerning the sources from which materials for a history of plants are derived serve also to explain the arrangement of the subject-matter to be dealt with in the Second Volume of this work. The order of presentation of the different parts of the subject will follow the stages of develop- ment of the science. A history of the entire plant-world considered as a single great community must be preceded by a history of species. But each species is the sum of numberless individuals, which are alike in constitution and have the same external characteristics, and a history of species therefore presupposes a knowledge of the history of the individual. Accordingly, our first business is to describe the rejuvenescence, multiplication, and distribution of individuals, and to show by what means a plant, considered as a separate organism, maintains itself, takes possession of its habitat, and is enabled to keep its hold on that habitat up to the moment when it is replaced by descendants endowed with a vitality of their own. THE LANGUAGE OF BOTANISTS. Before entering upon a description of any of the above phenomena, I feel it necessary to say a few words respecting the technical botanical terms of which ] shall make use. The need of short and compendious names to denote particular forms, particular organs, and particular processes, has been always universally recognized, and more or less appropriate additions to our vocabulary have been made by men of science from time to time. As might be expected, these designa- tions are not only an expression of the particular standpoint to which, at the time of their invention, the actual knowledge of plant-life had most recently attained; but they are also liable to bear the stamp of theories advanced by eminent 4 THE LANGUAGE OF BOTANISTS. naturalists of the day, or of the hypotheses which happened to be then in vogue. The progress of true knowledge is too often hindered by the fact that men exalt their speculative theories to the position of " laws of nature ", and when they first encounter contradictory evidence twist and turn it until it appears actuall}^ to verify those theories. We need not inquire in these instances how much is due to self-deception and how much to prejudice and dogmatism on the part of the investigators. Certain it is that such a perverse method of research, especially when supported by the authorized beliefs of the thoughtless multitude, acts as a drag on true science. Fortunately, however, it is nothing worse than a drag. For, sooner or later, the conviction again asserts itself that our notions respecting the history of plants must be derived from the facts observed in their entirety and purity, instead of facts being made to fit a preconceived opinion — some being explained away as exceptions, whilst others are altogether ignored and sup- pressed. In all sciences for which it is requisite to invent technical terms — and in Botany no less than in others — we find that the terminology bears traces of ideas formed at earlier periods, and now rejected as being based on insufiicient experiment or imperfect observation, on self-deception or prejudice, as the case may be. The question has, therefore, repeatedly been raised whether it is better to retain such names and modes of expression, although they are likely to suggest wrong ideas to students, or to abolish them and substitute new ones in their stead. There ai-e strong arguments for both courses. The chief advantage of retaining the old terms is that readers of modern works are thereby enabled to understand more easily the writings of older botanists. We have also to consider the probability that in rejecting old terms and inventing new ones we may fall into the same errors as our predecessors. Any one who has worked in the field of Botany for more than forty years, as I have done, must have found that on an average every ten years prevailing ideas have undergone a change. He has seen how theories, which for a time influenced every branch of the science, and were actually standard conceptions in many departments of research, have sooner or later had to give place to new ones. He knows how often a naturalist is compelled, in consequence of fresh and unexpected discoveries, to let go a position whicli he has considered impregnable, and which has become endeared to him by long familiarity. Thus, experience teaches diffidence, and one learns to attribute only a temporary value, so to speak, even to one's own original theories, and to rest assured that, in a few decades, what now appears to be nearest the truth will be superseded by something else which comes still nearer to it. But if, whenever a fresh stage of knowledge were reached, all terms and phrases which had become antiquated and no longer quite applicable were abandoned and replaced by others, and if in addition new names were introduced corresponding to every modification in the results obtained by observing all the different processes and appearances with which we have to deal, our science would inevitably be rendered far less accessible — and this consequence would be much to THE LANGUAGE OF BOTANISTS. 5 be regx'etted. However strong the desire to understand the secrets of plant-life, it could only be satisfied at the cost of learning a special scientific language; and Botany would become, in an even greater degree than is the case at present, a close study for specialists, instead of being the common property of the many inquiring minds to whom the results of our researches by right belong. Accordingly, we shall retain so far as is practicable the recognized scientific terms. Where they are no longer quite suitable they will be briefly elucidated; and, when the conceptions to which they refer have been expanded or limited, the old established names will also be taken in a wider or a narrower sense as the case may be. New expressions will only be introduced where their use is productive of greater clearness and distinctness in the ideas involved; and even these additions must be in harmony with the terms already in existence. It is worthy of note too that many foreign words, which have been longest established and also subject to frequent use by botanists, originally meant some- thing altogether dififerent from what they are intended to denote at the present day. In the very first section of this volume a whole series of such words will be employed. The history of the plant-individual is there dealt with. What is an "individual"? The word comes from dividers, to divide, and denoted originally a thing which is not divisible. But there is no such thing as an indivisible plant. The survival of plants, their reproduction and multiplication, are all dependent on processes of division; and any species whose individuals were not divisible, would be doomed to inevitable destruction. The characteristic property of an individual cannot therefore lie in absolute indivisibility. A qualification has in consequence been inserted in the definition, and an individual is explained to be a thing which cannot be divided without ceasing to be, as heretofore, an organized being subsisting independently, in which each single part belong indispensably to the whole. Even this definition is not appropriate to a plant. The living protoplast of a unicellular plant — an organism which must without question be conceived as an individual — divides into two halves, which separate from one another and constitute two independent individuals. This instance affords, however, an indication of the true definition. A plant-individual is an organism which can and does live indepen- dently and without the aid of other organisms of the same form. There are plant- individuals each of which consists of a single protoplast, whilst others are com- posed of many protoplasts living together. In the latter there is for the most part a division of labour accompanied by a corresponding variety in the forms of the different parts of the individual; but even in these cases individuality is not necessarily destroyed by division. Where division of labour has been carried so far as it is in a plant provided with stem and leaves (c/. vol. i. p. 584), it used to be thought necessary to look upon the structure as an association of individuals. Each single shoot was conceived to be an individual because it possessed the power of continuing to live after it had been separated from the axis, and on that assumption each one of the higher plants was built up of such and such a number of separate individuals. Later on, however, inasmuch as every branch of a shoot 6 THE LANGUAGE OF BOTANISTS. is capable of living when separated from the rest and of producing a new inde- pendent plant, the parts of a shoot came to be considered as being individuals, and the term " Anaphytes " was applied to them. We shall see hereafter the extent to which this conception is of importance in relation to the subject of alternation of generations. It would be out of place to treat it more fully at present. Another conception of the plant-individual must also be mentioned here. When the impossibility of defining it as indivisible was recognized, the strange expedient was invented of assuming the existence of divisible individuals and of representing all parts which have been produced asexually and have become inde- pendent as belonging to a single individual. A potato puts forth thirty or forty fresh tubers in the course of a few years, and all these were considered as collectively constituting one single individual, as also were the countless young carnation- plants which are to be derived by means of cuttings from one old plant. The general rule was that only an organism produced by sexual process was to pass as an individual. Cuttings, tubers, and the like, detached from such an organism would be, according to this conception, merely parts of one individual, even though they were capable of living quite independently and at a distance from one another. This definition, the invention of philosophers, has never been taken seriously by naturalists, and I only cite it because it introduces another important problem which I purpose to treat in an exhaustive manner in the first three sections of this volume, namely, the question of the propagation or generation of plants. The modes of reproduction in plants have been subjected in recent times to most patient investigation on the part of botanists gifted with the keenest powers of observation,, and their researches have led to the conclusion that in most — probably in all — divisions of the vegetable kingdom two kinds of propagation occur. In each case a single protoplast forms the starting-point for the new^ individual; but, in the one, this protoplast does not require the special stimulus aflforded by union with another protoplast, whereas, in the other, in order that a new individual organism may be produced, a pairing — i.e. a union of the substances — of two protoplasts, which have come into being at different spots, must take place. The former is called asexual reproduction, the latter sexual reproduction. All reproductive bodies arising asexually are included under the name of brood-bodies, whilst those which are associated with the sexual process may be termed broadly fruits. There are all grades of brood-bodies, from the single cell to the completely formed plant. Brood-bodies, if unicellular, are termed sijores, if multicellular, thallidia; and those which constitute rudimentary shoots are called buds. The bud form of brood-body either detaches itself from the living parent-plant, or else, as more frequently happens, it becomes independent through the death of the plant from which it sprang. In the latter case the ofi-shoots remain in the immediate vicinity of the parent-plant. In the case of trees and shrubs the buds do not sever themselves from the axis on which they were developed, but continue their connec- tion with it as they gi'ow into shoots, and in this manner are formed those compound individuals to which reference has been already made. It is much less THE LANGUAGE OF BOTANISTS, 7 common for full-grown shoots to detach themselves from the parent-plant and act as brood-bodies. Fruits of all degrees of complexity are also found. They are sometimes single cells, sometimes groups of cells, and sometimes complete plants in miniature. Usually the fruit — or at least the most important part of it which contains the fertilized ovum or the embryo produced thereby — becomes detached, when ripe, from the parent-plant; but, in many groups of the vegetable kingdom, in Ferns, Mosses, Lichens, and Florideffi, for example, the fruit remains at its place of origin and preserves its connection with the mother-plant whilst itself developing into a new generation, which, however, does not produce fruits but spores. When asexual and sexual reproduction take place alternately in a definite manner, we speak of an Alternation of Generations. Hitherto the subjects of fruit-formation and of the alternation of generations in their relation to the History of Plants have remained unrecognized and unelucidated. In one of the following sections of this volume an attempt will be made to solve this great mystery. THE GENESIS OF PLANT-OFFSPRING, 1. ASEXUAL REPRODUCTION. Spores and Thallidia. — Buds on Roots. — Buds on Stems. — Buds on Leaves. SPORES AND THALLIDIA. In the chapters on ferns in the old herbals, attention is invariably directed to the extraordinary phenomenon that the plants in question do not produce flowers or fruit, and yet propagate their kind and multiply abundantly, and the remark ip made that these plants will often spring up quite unexpectedly in caves, or in the cracks of old walls, without any seeds having been previously perceptible there. Hence in Germany a fabulous story was invented that the seeds of ferns were formed in a mysterious manner at the time of the summer solstice only, and that these seeds could only be collected on Midsummer Eve by persons initiated in the mystery who made use of certain magic words on the occasion. Hieronymus Bock or Tragus, as he called himself in accordance with the then prevailing fashion of translating names into Greek, preacher and physician at Hornbach in 1532, was the first to contradict this childish superstition, and to convince himself of the possibility of obtaining "fern-seeds" without the use of incantations. In his Herbal, published in 1539, he gives the following account, which is in many respects interesting, of his endeavours to collect the seeds of ferns : " All our teachers write that the fern bears neither flower nor seed; nevertheless, I have four times looked for the seed in the night of Midsummer Eve, and have found early in the morning before daybreak small black seeds like poppy-seeds on cloths and on the broad leaves of mullein beneath the stems in varying quantities. ... I have used no charm or spell in this matter, but have looked for the seeds without any super- stition and have found them. One year, however, I found more than another, and I have sometimes been out without success. I have not gone alone to fetch the seeds, but have taken two others with me, and have made a great fire in an unfre- quented spot and let it burn all through the night. How the thing came to pass, and what secret nature intends to reveal by it, I cannot tell. I have stated all this because all our teachers describe the fern as being without seeds." There can be no doubt that by the brown seeds Hieronymus Bock meant those structures which, about two centuries later, were named " spores" by Linnaeus. But even in the time of Linnaeus the whole subject of spores, especially in their relation to fruit, was shrouded incomplete obscurity. The word "spore" is derived from SPORES AND THALLIDIA. y the Greek, and signifies etymologically precisely the same as "seed", and spores were considered to be peculiar seeds, formed by means of some mysterious processes of fructification. As late as fifty years ago the spore was defined as " that part of a cryptogamic plant which corresponds to the seed, and from which, although it contains no germ, a new plant can be developed ". The mode of fructification in the Fern, and, in general, the entire history of its development, were discovered for the first time in 1848. It was then shown that these plants pass through two kinds of regularly alternating generations. One of these is itself inconspicuous, but bears reproductive organs and produces fruits; the other, springing from the fruit, which continues its connection with the parent- plant, is distinguished by fronds and produces spores. Thus the fronds of Ferns bear no sexual reproductive organs, and the spores formed upon them cannot there- fore be looked upon as fruits or even as seeds, a seed being part of a fruit. Some people, it is true, treat the entire frond-bearing Fern-plant as a fruit and the spores on the fronds as part of this fruit, although such a theory involves the admission that fruits may strike root, multiply by means of runners and continue to grow for many years, putting forth annually new spore-bearing fronds. Accord- ing to this view, which I cannot endorse, a gigantic tree-fern, aged a hundred years, would be a fruit, and to be consistent it would be necessary to regard a whole grove of Horse-tails as belonging to one single fruit. Other botanists, whilst deny- ing that the Fern-plant with its roots and fronds is the fruit itself, are yet of opinion that the formation of spores in the Fern is a consequence of the process of fruiting, inasmuch as the Fern-plant would never make its appearance at all but for the formation of fruit by the previous generation; and they hold that the spores of Ferns, and of their allies the Horse-tails and Club-mosses, should on that account be distinguished from those of other Cryptogams. To this view there are two objections. In the first place, we know many cases wherein a Fern-plant with spore-bearing fronds is developed from the first generation without any formation of fruit having taken place, and the plant in these instances is in no way different from those which have sprung from fruits of the first generation. Secondly, it is difficult to see why the sporogenous generation should be more dependent on the fruit of the first generation in the case of Ferns than in many other Cryptogams, which similarly exhibit an alternation of generations. As the spores of Ferns, and of Cryptogams in general, are not the direct result of a process of fertilization, they are not parts of fruit, but are brood-bodies. They should be placed by the side of the bud forms of brood-body presently to be described, though differing from these in that they always produce a single layer (i.e. a thallus) only, and never a leafy, axial structure. They are just as characteristic of Cryptogams as buds are of Phanerogams or Flowering Plants, and as the name of Cryptogam is no longer quite appropriate, it is often replaced by the term "sporogenous plant". Before the discovery of the alternation of generations in Cryptogams, the name spore was applied to many fruits and rudiments of fruits, particularly where these happened to be unicellular, an error which we should be 10 SPORES AND THALLIDIA. careful to avoid at the present day. When we come to the description of fruits and their origin, we shall have occasion to return again to this subject. The places where spores originate are remarkably varied. In some plants nests of cells make their appearance in the interior of an extensive tissue; in others single cells are exposed on the surface. The task of spore-development devolves sometimes upon a part of a green stem or leaflet. Sometimes — in plants devoid of chlorophyll — upon the protoplasmic contents of a tubular structure, and some- times upon the abstricted ends of hyphal filaments. The best way to arrive at an idea of the extreme diversity in this respect is to classify spores in groups accord- ing to their mode of origin. (Cf. p. 20.) One group comprises all such spores as are formed in the cells of a tissue. Amongst these are the spores of Ferns, Rhizocarps, Horse-tails, Club-mosses, and the numerous kinds of Mosses and Liverwoi'ts. In one sub-group of Ferns papillae spring singly from the epidermis clothing the ribs of the fronds, each papilla being divided by a transverse wall into a free extremity and a stalk-cell. Both cells of the papilla become partitioned so as to form bodies of tissue, and the one that developes from the free terminal cell assumes an oval or spherical shape. In this latter ball of tissue a tetrahedral central cell and an envelope composed of several layers of cells may be distinguished. By internal partition of the central cell a little cluster of cells is formed, whilst, in the meantime, the inner layer of cells composing the envelope is dissolved, so that the whole now assumes the aspect of a receptacle inclosing a ball of cells embedded in a fluid matrix. Each cell of the cluster next divides into four compartments, and the protoplasts which constitute the contents of these chambers provide themselves with membranes and become disconnected upon the solution of the framework of their home. These separated cells are the spores. To the naked eye they have the appearance of a powdery mass. As has been said, of the cell-layers which formed the envelope of the sporogenous tissue, only the inner one was dissolved; the outer layer persists and constitutes a kind of capsule, to which the name of spore-case or " sporangium " is applied (see figs. 189 ^^ 189^*, 189^^). A collection of sporangia of this sort is called a "sorus". In the Polypodiacese — a family of Ferns to which the majority of European species belong — the sori may be seen on the backs of the fronds (see 189 ^). Upon the veins running through the green tissue are seated little cushion-like groups of cells. Each cell in one of these cushions is capable of developing into a stalked sporangium, and sometimes a single sorus consists of no less than fifty such stalked sporangia. In the CyatheaceaB also, which include most of the Tree-ferns, the sori are developed on the under side of the fronds, but in their case each is borne on a kind of peg projecting at right angles to the surface of the frond. The sporangia derived from the epidermal cells of this peg are very shortly stalked. An annular wall is produced from the green tissue of the frond and surrounds the sporangiferous peg, which consequently stands up from the middle of a cup (see figs. 189 ^°, 189 ^\ 189 '^). In the delicate and graceful Hymenophyllaceae — Ferns with a resemblance to SPORES AND THALLIDIA. 11 Fig. 189.— Ferns. Nephi-oLcpm Uiqli. '^ Tnchomanes Lyelli. 3 Sorus of the same Fern with cup-sliaped iuvestiiieiit seen in longitudinal sectioB. * Rhipidopteris peltata. ^ Polypodium serpens. ^ FinnA of Gleichenia alpina. t Schizceafistulosa. » Botrychium lancto- latum. 9 Under side of a pinna of Gleichenia alpina ; in the two upper cavities the sporangia are covered by leaflets, in tlie under ones they are exposed, lo, n Pinna of Cyathea elegans. 12 Longitudinal section through a Sorus and Cup of Cyathea. is Sporangium of Cyathea. " Sporangium of Polypodium. is Sporangium of Schizcea. »« Under side of the Prothallium of Spleenwort. », 2, 4, s, 6_ r^ 8 natural size ; », », 10, », 1-. ". ", ", '« niaguilled from 5 to 20 times. 12 SPORES AND THALLIDIA. Mosses, and belonging for the most part to tropical regions — the veins of the pinnae project beyond the margin of the green tissue and form styloid processes whose epidermal cells become the points of origin of sporangia. Each styloid process thus constitutes an axis bearing the sporangia, and the entire sorus has the form of a little spike. But the sorus itself stands in a cup formed by an upgrowth of the green tissue at the margin of the pinna (see figs. 189^ and 189^). In the three groups of Ferns above dealt with the sporangia arise from epidermal cells. In the Gleicheniaceae and Schizaeacese (two specimens of which are shown in figs. 189^ and 189'') the sporangia are modified leaflets. We must here remark that the fronds of Ferns in spite of their similarity to foliage-leaves are not to be regarded as such, but as phylloclades, whilst the scales upon the fronds must be considered to be leaves. We shall refer to this again later on. Now, in Gleicheniaceae and Schizseaceee some of these small scaly leaves are metamorphosed into sporangia which here take the form of rounded bodies set in rows of pit-like cavities hollowed out of the pinnae, whilst other scales constitute protective coverings to these sporangia. The relation existing between the various parts in the case of a pinna of Gleichenia alpina is shown very clearly on an enlarged scale in fig. 189^. In respect of origin and development the spores and sporangia are again quite diflferent in the group of Ferns comprised under the name of Ophioglosseae, one species of which — viz. the spear-shaped Moonwort (Botrychiu7)i lanceolatum) — is represented in fig. 189 ^ In these Ferns, the sporogenous portions take the form of nests of cells embedded in the tissue of the frond. The cells in these niduses become partitioned each into four chambers, and the latter contain protoplasts, which surround themselves with membranes and become spores. The spores are set at liberty as a consequence of the solution of the walls of the chambers, and they occupy, in the form of a fine powder, little vesicular cavities in the tissue of the pinnules. The epidermis of these pinnules now serves as the wall of the cavities, i.e. of the sporangia. Each plant in the group of the Ophioglosseae exhibits two kinds of frond: the one kind develops no spores and has the appearance of a green foliage-leaf; the other produces sporangia, which are arranged either like bunches of grapes or in spikes consisting almost entirely of the sporangia (see fig. 189^). A similar arrangement may be observed also in many Ferns belonging to other divisions, as, for instance, in the genera Allosorus, Struthiopteris, and Blechnum, representatives of which occur in the European Flora as well as in others. In other cases, such as the Flowering Fern (Osmunda regalis), for example, sporangia are only formed on the upper portion of a frond, whilst the lower segments are foliaceous. A very peculiar form is that of RJiipidopteris peltata, a fern indigenous in the mountainous regions of Mexico (see fig. 189*). Besides the flat, fan-shaped fronds which produce no sporangia, other fronds shaped like funnels or shallow bowls are developed, and the spore-cases are produced from the epidermal cells in the hollows of these fronds. In the last case it is worthy of note that the sporangia are formed on the upper SPORES AND THALLIDIA. 13 surface of the frond, for this is of very uncommon occurrence. They are usually situated on the under surface of the frond, the reason being that they are thus best sheltered from both rain and sun. Most instances exhibit in addition a further safeguard against excessive moisture or desiccation in the form of a special awning covering the sporangia. This awning is either an outgrowth from the cells forming the apex of the sporangiferous cushion or peg, and takes the form of a delicate membrane stretched over the whole sorus and known as an indusium, as in our male Shield Fern (Aspidium Filix-mas), or else small, scale-like leaflets spread themselves over the sporangia, as in the Gleichenias (fig. 189^), to which reference has already been made, and in the no less remarkable Lygodiums and Davallias. Sometimes five or six squamous leaflets stand in a circle round the sporangia and envelope them so that the whole looks deceptively like a closed flower, as in the genera Schizoccena, Hymenocystis, and Diacalpe; or, these leaflets form a sort of box, which opens with a lid, as in Cibotiuni. In other cases, again, membranous bands or borders grow up from the surface of the frond, and the sporangia, which are arranged in a long line, are covered over by them, as occurs in Lindsay a and Blechnum; or, the margin of the frond is split and the sporangia are hidden in the narrow cleft thus made, as in Vittaria and Schizoloma. Often the margin of the frond is folded over, thus covering the sporangia, which are here developed on raised cushions; AUosorus, Ceratopteris, Ceratodactylis, Parkeria, and many other genera exhibit this formation. The extreme variety prevailing in this class of adaptation is connected with differences in the climatic conditions of the habitats where the plants in question live. Any attempt to deal with individual contrivances here would lead us too far. The Rhizocarpese are a group nearly allied to Ferns, and they naturally follow the same lines as Ferns in the formation of their spores and sporangia. Salvinia reminds one to some extent of the Hymenophyllacese, at any rate as regards the outgrowth of an annular wall below the sporangia (the latter being in this case also borne on a fusiform axis), and also as regards the development of this wall, which becomes closed at the top into a complete box enshrouding the sporangia. Marsilea, on the other hand, exhibits stalked, bean-shaped capsules with cavities in which che sporangia are formed on raised cushions. The Club-mosses (Lycopodiace?e) also bear a striking resemblance in their mode of spore-formation to Ferns, especially to the various species of Lygodiuvi and Lygodictyon, genera of which mention has already been made. The first rudiments of the sporangia are swellings at the base of the little squamiform leaves, or on the axis just at their insertion. The internal tissue of this protuberance is marked oflf in the form of a roundish ball. The cells constituting the ball separate and then become segmented each into four chambers, the walls of which are subsequently dissolved. The protoplasts within the chambers inclose themselves in membrane's and become free spores. The epidermis originally clothing the swelling persists, and now forms the wall of a bean-shaped sporangium containing loose spores. The sporangium subsequently opens by means of a lid like a box. 14 SPORES AND THALLIDIA. Horse-tails exhibit a process of spore-formation quite peculiar to themselves. Two species of this group — namely, Equisetum arvense and E. sylvaticuvi are shown in figs. 190 ^ and 190^. At the top of the hollow stem there is a spike of peltate scales borne on short stalks and arranged in whorls, each of which must, in consideration of its origin, be looked upon as a metamorphosed leaf (cf. fig. 190^). Fig. 190.— Horse-tails. Summer Shoot of Equisetum arvense. 2 Vernal fertile Shoot of Equisetum arvense. * Spike of whorled sporangiophorea from the same Equisetum. * A single sporangiophore. ', « Spores. ' Equisetum sylvaticiim. « Prothalliuni of ;i Horse- tail. ', *, ' natural size ; » x 3; * x 6 ; *, « x 25 ; 8 x 30. On the inner surfaces of the scales — i.e. those turned towards the axis of the spike — little warts arise, which develop into sporangia (cf. fig. 190^). The outer cell-layers of these multicellular warts become the walls of the sporangia, whilst the inner tissue breaks up into cells. These cells then divide into four cells, each of which becomes a spore. SPORES AND THALLIDIA. 15 The last division of plants wherein the spores are formed deep down in a tissue is that of the Muscinese, which include Mosses and Liverworts. In these plants the spore-producing generation consists of a cellular body, which has arisen from the fruit, is usually seated on a stalk, and in shape is cylindrical, pyriform, or more or less spherical (c/. figs. 191 3,4,7,8,15-) ^g must here remark, by the way, that botanists used formerly to look upon this sporogenous generation of the Moss erroneously Fig 191.— Mosses. i Polytrichum commune, the sporogonium to the left concealed by the cap, the sporogonium to the right exposed, s The same Moss in an earlier stage of development. 8 Sporogonium of Polytrichum commune with its lid. ■• The same after the lid has fallen off. * Bryum ccespiticium. « Sporogonium of the same Moss with its cap. ' The same without the cap, but with the lid still on. s The same after removal of the lid, showing the teeth (peristome). » A piece of the peristome. 10 Antheridia, Archegonia, and Paraphyses of Bryum ccespiticiicm. n Hylocomium splendent. '2 Sporogonium of Hylocommm splendens. is Andrcea rupestris with burst sporogonium. " Sphagnum cymbifolium, its spherical sporogonia still covered by their lids in the left-hand specimen, is A single sporogonium of the same Moss, i, «, «, ", " natural size ; s, 4, 6^ 7, 8^ 12, 13, 15 X 5; 9, 10 X 150. as the fruit itself. The only structure rightly to be considered as the Moss-fruit is that in which the embryo is produced as a result of fertilization. If afterwards a new generation springs up from the embryo which has been formed in the interior of the fruit, this generation cannot any longer be described as a fruit even in cases 16 SPORES AND THALLIDIA. where it remains permanently connected with the mother-plant, as happens in Mosses. The cells composing the tissues of the cylindrical, pyriform, or spherical body above referred to develop in a variety of ways. Those situated near the outer surface form the wall of a receptacle, and those in the interior, which serve as a filling to the receptacle, form the spores. The process of spore-formation is here much the same as in Ferns. The cells of the central mass, at first united into a tissue, in time become isolated; each divides into four, and the spores are ultimately developed from these protoplasts. The spores are then left free in the form of a fine powder within the receptacle, which is called a sporogonium. In most Liverworts, a group nearly allied to the Mosses, certain other cells having a curious structure are formed from the internal tissue besides the spores. These are the so-called " elaters ", and they serve to scatter the spores. In a few Mosses a sort of central column remains in the middle of the sporogonium in addition to the spores when the whole is mature. Externally the sporogonia of Mosses differ very little from the cellular bodies out of which they were developed; like them, they are spherical, pear-shaped, or cylindrical as the case may be. But the part which subsequently opens and liberates the spores at the proper time exhibits in its more minute anatomy considerable differentiation. This subject and that of the elaters mentioned above will be again referred to in the section devoted to the distribution of spores. As with the sporangia in Ferns, so also in Mosses the sporogonia are protected; during development from injurious external influences, especially desiccation, and are wrapped in coverings which vary considerably according to their origin. In Mosses a kind of cap is usually to be seen covering the young and tender sporo- gonium (see fig. 191 ^), and this structure has its origin in the fruit from which the sporogenous generation (or sporophyte) has sprung, the coat of the fruit being torn away and its upper part carried up in the form of a cap by the sporophyte during its growth from the embryo. Later on, when the sporogonium is no longer in need of protection, and the presence of a cover would be detrimental in that it might prevent the spores from being scattered, the cap is cast off. All the spores hitherto discussed originate within a tissue, and their history involves the conversion of the protoplasmic contents of each compartment of the reproductive part of the tissue into a spore. A second group of spores is composed or those which arise from the breaking up of the protoplasmic contents of tubular, club-shaped, or spherical cells not united in tissues, and are set free from their birthplaces as soon as they are formed. The cells thus constituting the mother- cells of spores may, by analogy, be conveniently termed sporangia. The process of formation of spores within them appears to be much simpler than in Ferns, Club-mosses, Horse-tails, Mosses, and Liverworts. Speaking generally, the only striking differences occurring in these cases are such as afiect the number and shape of the spores which escape from a sporangium. As described in the first volume of this work (c/. vol. i. p. 23, and fig. 25a, a-d\ SPORES AND THALLIDIA. 17 the filamentous organism Vaucheria produces a single comparatively large green spore in each of the club-shaped outgrowths developed by the tubular branches of the plant, and each spore thus formed is able, when free, to swim about by means of its numerous short cilia. On the other hand, the mould-like Saprolegniaceee, which live under water upon decaying animals, develop a large number of colourless spores in their clavate filaments, and these after escaping from the tubes whirl aliout in the water by means of two long revolving cilia (c/. fig. 192), In both l\ Fig.. 192. —Swarm-spores of Saprolegniacese and Chytridiaceee. ' Achlya prolifera. ^, ", * Development and escape of swarm-spores of Achlya proli/era. * Chytridium Ola parasitic upon the oogonium of (Edogonium; development and escape of swarm-spores. ^ Saprolegnia lactea. i Development and escape of the swann-spores of Saprolegnia lactea (partly after De Bary and Pringsheim). ix20; a, », «x400; 5x300; 6x100; 'x300. these instances the spores themselves possess the power of movement and of swarming about in water, whence they are called "swarm-spores". The name "zoo-spores" (^'wo;^= animal) has also been applied to them on account of their decided resemblance in form and behaviour to certain Infusoria. The delicate, profusely-branched mycelia of the Moulds, included under the name Mucorini, give rise to special filaments which grow straight upwards. These erect hyphse divide into two cells. The upper cell becomes a spherical bladder, and the under a long slender stalk, the upper end of which protrudes in the form of a hollow stopper into the bladder supported by it (c/. fig. 193 ^). The protoplasm in the upper vesicular cell breaks up into a large number of spores and thus 18 SPORES AND THALLIDIA. becomes a sporangium. The increase in weight of the sporangium causes the filiform stalk to bend; the sporangium bursts, and the spores, together with the clear fluid in which they are suspended, issue through the rent in the sporangium (c/. fig. 193^). In the Moulds of the family of the Mucorini the sporangia are for the most part Fig. 1S3.— Moulds. Mucor Mueedo- x40. « Longitudinal section of a sporangium of ifucor ifwcerfo; x260. « Fruit-formation in Mucor J/uc«do; XlSO. * Aspergillus niger; x30. ' Longitudinal section of a sporophore of A«per(7i7ZMS Ju^er. « Fructification of /"enicit- lium crustaceum (after Brefeld). ? Fruit-formation in Aspergillus (after Eidani). » Penicillium crustaceum ; x 40. 9 Sporophore of Penicillium crustaceum ; x 200. closely crowded together, but they are never walled in by a tissue or surrounded by any particular envelope. They are, moreover, always separate, and have the appearance of a miniature plantation. A different state of aflfairs is found in that group of Fungi known as the Ascomycetes, a group which includes, amongst well- SPORES AND THALLIDIA. 19 known plants, the genera Morchella and Helvetia {of. fig. 194), Lichens, and also several mould-like forms, notably the Erysiphege, which produce Mildew, and Claviceps, which is the cause of Ergot of Rye. In these plants the ends of the hyphse stand up from restricted areas of the mycelium, some in the form of long clavate tubes, some as delicate filiform paraphyses, the group of tubes and paraphyses being surrounded by other cellular structures in such a manner that the whole has the appearance of a dish or cup or capsule. The protoplasm in the tubes breaks up and forms either ellipsoidal bodies arranged usually in linear series (cf. fig. 194^) or long fascicled threads, which, whilst still inclosed in the Fig. 194. — Discomycetes. The >rorel {Morchella escnlenta). 2 Longitudinal section from the hymenium of Morchella esculenta showing five fllnmenta e.ich containing eight spores and filiform paraphyses in between them, s Helotium Tuba. * Anthopeziza Winteri. 5 reziza vesiculosa. « Helvella In/ula. 7 Helvella fistulosa. 1, *, 5, 6^ 7 natural size ; 3x4; 2x120. tubes, put on a stout cell- wall. The name of asci (do-K6s = a leather bag) has been given to these sporangia, and asco.^pores to the spores which they contain. They are destitute of cilia, the distinguishing mark of zoospores, and have no power of independent motion after their extrusion from the tubes, which takes place through a rent at the top. There is great variety in the mode of grouping, as also in the envelopment of the sporogenous tubes in different genera and species. When the tubes grow from the bottom of flask-shaped excavations or pits, the whole structure is spoken of as a, perithecium ; if they stand in a shallow patellif orm cavity or on the surface the 20 SPORES AND THALLIDIA. term used is apotheciuvi. Perithecia and apothecia have been erroneously called fruits also. The same principles must here be applied as governed our consideration of Ferns and Mosses. Even if the genesis of perithecia and apothecia is really preceded by a process of fertilization, still the only part properly to be called a fruit is the tissue in which one or more protoplasts have become embryos in consequence of the act of fertilization. The outgrowth from this fruit is precisely the new generation; and it does not matter at all whether this new sporogenous generation preserves its connection with the previous fruit-forming generation or not. Perithecia and apothecia, and, in general, all so-called fruits in the Ascomycetes are therefore equivalent to the sporogonia in Muscineae and to the sporangiferous plants in Horse-tails, Club-mosses, and Fei-ns. We shall place together in a third group all spores which arise neither singly in the cell-compartments of a tissue nor through the breaking up of the protoplasm within a tube, but by abstriction and abjunction. The process of spore-formation in these cases is as follows: — The protoplasm, which is inclosed in a cell- wall, produces an internal partition whereby it is itself divided into two halves, and the cell-cavity into two chambers. As soon as this has taken place the partition-wall splits and the two cells fall asunder. If the cell which undergoes the process of bipartition is in the form of a blind tube or sac, and if the partition is intei*- calated near the tip of the sac, the effect produced is as though the end of the sac had been tied off or abstricted and had then dropped. The part remaining behind now constitutes another blind sac, and in some genera the process of abjunction from the extremity may be repeated over and over again. Basidium is the name given to a closed sac of this kind from which spores are abstricted, it forming in a manner a base for the spores. This term has hitherto only been employed by botanists in relation to the so-called Basidiomycetes (which includes the Fungi known as Mushrooms and Toadstools), but it is justifiable to extend its application to all other structures which play the same part. Abstriction of spores is exhibited at its simplest in the plant known as the Rust of Wheat, which at a certain stage of its development lives as a parasite in the green tissue of the leaves of our species of Wheat. For the purpose of spore- formation tufts of hyphae project beyond the surface of the infested leaves. At the extremity of each hypha, which is in the form of a closed sac, a single spore of comparatively large size is developed; and after the fall of this one spore the hypha or basidium has lost the power of abstricting others. A similar phenomenon is observed in the Fungi belonging to the genera Hydnum, Polyporus, Agaricus, and Clavaria, of which several examples are represented in fig. 195. Their basidia are club-shaped, and terminate in four slender filaments, the so-called sterigmata, and from the end of each sterigma one spore is abjointed (fig. 195^). These basidia, together with a number of slender sac-like tubes, to which reference will again be made when the Basidiomycetes are described in detail, beset certain structures projecting from the under surface of the cap-shaped sporophore — these structures being lamellse or spikes or tubes SPORES AND THALLIDIA. 21 as the case may be. Aspergillus niger (see fig. 193* and 193 ^X a Mould living chiefiy on the juices of fresh or preserved fruits, develops slender upright hyphse with swollen ends, which bear numbers of short peg-like processes — the sterig- mata — from which moniliform series of from five to eight spores are abjointed in » Clavaria aurea. ' Divdalea qvcrc. phalloides. ' Clavate basidia w hynieiiiiim of Amanita phallo, ' X 250. Fig. 195.— Basidiomycetes. na. i Marasmiiistenerrimus. * Marasminsperfnrans '■Craterclbis dnvntus. » Aniayiita th filamentous sterigmata, from the ends of whicli spherical spores are ahjointed (from the des). 8 Ijydmnn imbricatnm. 9 Polyporus perenitis. i, =, ». ■•, ', «. «, » natural size; rapid succession. These spores at first hang loosely together, and are arranged hke strings of pearls, but collectively these rows of spores form a spherical head. A shock of any kind, especially the disturbance occasioned by currents of air, will cause a severance of the spores, and the entire sphere consequently falls to pieces. 22 SPORES AND THALLIDIA, Nothing then remains but the hyphal filament with its swollen end beset with pegs and looking like a club armed with spikes (c/. fig. IQS'*). Also in Penicillium, the commonest of all Moulds, the spores are abjointed from the sterigmata in moniliform rows; but in this ease the erect hypha which bears the spores is septate and not clavate at the extremity, and terminates in forked branches, so that the chains of spores are grouped like the hairs in a camel's- hair pencil. A species of Penicillium — viz. P. crustaceum — is represented in fig. 193*^ and 193''). In the Peronosporeae, to which class belongs the parasite Cystopws candidus, celebrated for its fatal effects on cruciferous plants, moniliform rows of spores are abjointed from the basidia without the intervention of sterig- mata. The mode of arrangement of the chains of spores in this parasite is, how- ever, not quite like that in either Penicillium or Aspergillus. A further diversity in this kind of spore-formation by process of abj unction is introduced by the presence in several families of plants of special envelopes surrounding the abjointed spores. Particular cases of this are afforded by Gasteromycetes (Puff'-ball family) and Floride^ (Red Seaweeds) and by that stage in the development of the Rust-Fungus which is known by the name of JEcidium. The secidia make their appearance in the form of structures growing out from a mycelium infesting the green tissues of leaves. The basidia are formed by the ends of hyphae which stand up in dense crowds. Moniliform chains of spores are abjointed from the basidia and are enveloped by a sporangium-like wall developed from the cells surrounding the basidia. It is not till this enveloping capsule bursts that the spores are set free and can be distributed. In the large Puff'-Ball family (Gasteromycetes) the same process takes place, but the basidia and spores are not arranged so regularly, and amongst the spores are to be found other hair-like, cellular structures which constitute what is termed a capillitium and are of especial importance in relation to the distribution of the spores. Florideae develop their spores within receptacles peculiar to themselves, which frequently resemble urns or capsules, and might be designated sporangia for the sake of terminological uniformity. The spore-filled "sporangia" of Florideae, like those of Muscineas — and in particular of Liverworts — are to be conceived as a separate generation, and, moreover, as a generation springing from cells which have undergone fertilization and have thereby been converted into fruit. The description of the process of fertilization must be postponed to a later section of this book; we have only to notice here that short cells are put forth as branches from the fertilized cells, and that some of these branches abjoint clusters of spores whilst the others develop into a sheath enveloping the assemblage of spores thus produced. Under the name of Thallophytes are included all such plants as are destitute of vascular bundles and therefore are never developed into real plant-bodies (cf. vol. i. pp. 590-592). It often happens that Thallophytes form, in addition to the uni- cellular brood-bodies to which the name of spore must be limited, cell-aggregates which sever themselves from the thallus and become independent, the genesis of SPORES AND THALLIDIA. 23 which has not been in any way a result of fertilization. These aggregates of cells are, in a manner, structures intermediate between the unicellular spores and the buds, differentiated into axis and leaves, which occur in vascular plants. They are portions of the thallus which produced them, and are either very like it or assume the same form as soon as their further development is complete. Hence the most appropriate name for these bodies is that of thallidia (ea\\6s = young shoot; el5os = a likeness). They are also known as gemmce. Thallidia are some- Fig. 106.— Thallidia of Muscine®. Slarcnantia polymorpha with cups containing thallidia or gemmae. » Longitudinal section of thallidial or gemmiferous cup. 3 A single thallidium. •» Tetraphis pellucida. « A stem of Tetraphis bearing a cup containing thallidia. « Lon.Lritudinal section of a thallidial cup. ', s Isolated thallidia of Tetraphis. » A stem of Leucodon sciuroides with brood-bodies, lo A brood-body set free from the stem, n Development of a brood-body from the rhizoids of a leaflet torn from Campyloptit fragilis. 12, is, h Development of thallidia at the apex of a leaf of Syrrhopodon scaber. « Aulacomnion androgynuw. 16 A stem of the same bearing thallidia. ", is Single isolated thallidia. 1 natural size ; «, »« x 2 ; s, ", is x from 8 to 15 ; 6_ 6^ 9, 10, 14 X from 20 to 40; s, r, 8^ 17^ I8xl20. times in the form of rows of cells, as, for example, those developed on the leaflets of the Moss Syrrhopodon scaber (see figs. 196 ^2, is, uy^ sometimes they are nets, as in the Water-Net (Hydrodictyon, see figs. 197 ^- *• ^). In the Moss TetrajMs pellucida (see figs. 196 ^■^•^> 7- 8) they occur as plates of cells, and in other cases they assume the form of globular or ellipsoidal lumps of tissue, as, for instance, in the Moss Aulacom- nion androgynwm (see fig. 1 96 ^^' ^^' ^^' ^^). Sometimes the number of cells associated 24 SPORES AND THALLIDIA. in a brood-body of the kind is limited to two, as is the case in the so-called " teleutospores " of the Rust-Fungus; whilst those of Florideae sometimes have four cells and are known as " tetraspores ". Again, in other cases hundreds of cells are associated together to form a thallidium, an instance of which is afforded by the brood-body or gemma of Marckantia (see fig. 196 ^'2. 3). The "soredia" of Lichens must also be brought under this head — by the term soredia being under- stood certain bodies which arise upon the thalli of Lichens and consist of one or more green cells wi-apped in a net-work of colourless hyph93 (see vol. i. p. 248). I Water-Xet (Ilydrodictyon utricnlatum), natural size. 2 a piece of the Water-Net; x 50. ». <, ' Development and escape of a reticulate thallidium; x 300. « Pediastrum granulatum; development and escape of thallidia; the liglitlj-dotted cell chambeis already vacated. ', « Thallidia of Pediastrum after their escape ; x240. Thallidia may originate in the interior of a cell-cavity of the parent-plant and escape in the form of complete, though extremely minute, cell-aggregates. Instances of this are afforded by the Water-Net (Hydrodictyon utriculafum), which is shown in fig. 197 ^, and by Pediastrum granulatum (fig. 197 ^), an organism of frequent occurrence in pools. The alternative method of formation of thallidia is by the severance of groups of superficial cells, which, after an interval of peregrination of variable duration, fasten on to some spot or other and found a new colony. In many Liverworts and Mosses special pockets and nopmr imitAXY N. C State Collegt WWDPERTVOF KIM. COLLEGE LIBRARY. BUDS ON ROOTS. 25 cups are produced, within which thallidia are continuously developed in the manner shown in figs. 196 !■ 2.3- 4.5, 6,7,8. The formation of these brood-bodies by Lichens and Mosses may be induced by wounds or mutilations affecting the plants in question; but the stimulus is not here susceptible of being so clearly and surely inferred from its effects — and perhaps has hardly yet been so carefully investigated — as in the case of trees, shrubs, and herbs, which, being planted on a large scale, have afforded experience for centuries with the result that the practice of inducing the formation of buds by mutilation and of using them for the purpose of artificial propagation is extremely common in cultivation. Parasitic thallophytes receive an evident stimulus to the formation of brood-bodies upon the death of their hosts. As long as the host-plant is healthy and vigorous the parasites keep their hyphse and suckers buried within the nutrient tissue. They there consume all there is to consume, increase in size, and thread their way through wood and green tissue in ever-widening circles — but without ever forming brood-bodies. Not until the host is quite exhausted and languishing at death's-door does the parasite, to avoid the danger of perishing with its foster-parent, provide for its departure from the ruin, and it is then in the form of brood-bodies that it escapes from the tissue it has ravaged. Here and there some of the tubular cells grow quickly from the interior of the dying tissue of the host-plant and emerge to the surface through stomata or rotten cell-walls. All the substance contained in the cells of the parasite becomes concentrated at these new foci of formative activity, and here masses of spores and thallidia are developed and abstricted at the very points where most extensive distribution is rendered possible by currents of air and water. Thus, the parasite is resolved into a number of brood-bodies and abandons the mansion which it has brought to destruction. BUDS ON ROOTS. Just in front of the house in which I am writing there used to stand years ago a great Aspen. The tree was felled, the axe being laid so close to the earth that only a stump projecting a few centimetres above ground was left. In the following spring the stump became the centre of quite a grove of Aspens, slender shoots having pushed through the grass over a large circular area round the stump. At first the shoots appeared one by one, then by dozens, and at last by hundreds at a time. They grew up into trees, and now, instead of the single Aspen, there is a little wood composed of trees which have not sprung from seed, but from the subterranean roots of the felled Aspen. Before the old tree had been deprived of its trunk and foliage its underground roots produced lateral roots only, which grew in a plane beneath and parallel to the surface, and continued to spread so long as they did not encounter any insuperable obstacle. Suddenly there was a change 'n the processes going on in this root; its formative energy was no longer devoted to the development of lateral roots, but was directed 26 BUDS ON ROOTS, to the construction of buds from which green leafy shoots sprang up above the surface of the ground. A forester of the old school, whose attention I drew to the above phenomenon with a view to ascertaining how he would explain it, told me that when the tree was cut down the flow of sap destined for the nourishment of the trunk and its crown of foliage was arrested in the roots underground, and thereupon sought an outlet elsewhere. Lateral roots having become useless, the diverted juices did not form them, but instead sent a great number of delicate shoots above the ground, because this was the only manner of preserving the life of the Aspen. At first sight this may seem to some people a foolish answer, and I have even heard it called absurd. Nevertheless we are obliged, after impartial consideration, to admit that we are not in a position to give any explanation which is not essentially the same. If we conceive the living protoplasts in the formative tissue of the roots as being the "juices" referred to by the forester, there is no longer any difference between his explanation and that given by Science. At the very spots where formerly lateral roots would have been developed, leafy stems are produced. The same protoplasts which now work at the construction of a bud would, if the tree had not been cut down, have fashioned a lateral root. That this alteration in active function was caused by the felling of the tree is certain, although no mechanical explanation of this stimulus can be given. The only possible source of excitation seems to be the checking of the egress of formative material stored in the roots in the direction in which it was formerly accustomed to flow. Another special point of interest connected with the history of this Aspen is that for the most part the roots, after giving rise to a series of shoots, died and decayed, whilst the shoots developed into separate and independent trees, each furnished with roots of its own, so that they look as if they had been deliberately planted in the earth in rows. As a matter of fact, however, the Aspen itself pro- duced these saplings from its subterranean portions, and planted them out, thus not only renewing its own youth but multiplying. For such multiplication it is evidently necessary that some cell in that part of the root which possesses the power of growth should form the starting-point or rudiment of a new shoot. The cell chosen for the purpose divides into daughter-cells, and these again become sub- divided; but several adjacent cells also participate in the new fabrication, and we can picture to ourselves the process as the action of a group of protoplasts located within the Kmits of the living and formative tissue of the root, which separate themselves off" from the rest and form a confederation of mutually helpful associates with the common function of constructing the new shoot. Neither the protoplast in the mother-cell of the young shoot nor the adjacent protoplasts undergo any stimulation by neighbouring cells before beginning their work. No process of pairing takes place. The phenomenon of renewal and multiplication of the Aspen which goes on before our eyes must therefore be classed as a case of asexual repro- duction. The fact that a single root of the Aspen, instead of producing one sapling BUDS ON ROOTS. 27 only, gave rise to ten, obliges us to suppose that these protoplasts of the growing tissue of the root, which separated themselves off under the influence of the new conditions created by the felling of the tree, arranged themselves in ten groups and each group from that time forth devoted itself to the new task of furthering the growth of the shoot developing at its centre. On investigation we find that these aggregations of cells are invariably situated in the deeper layers of the rind. In the first place a delicate tissue is developed from a particular cell which dominates the entire group and governs the process of construction. This tissue pushes outwards, on the one hand, towards the superficial layers of the rind, whilst, on the other hand, it sends a shaft inwards into the cambium layer of the root. Immediately afterwards vascular bundles are developed, and the shaft-like rudiment of the young bud is through them placed in connection with the woody tissue of the root, and when all this is finished the rind is finally broken through, and a bud clothed with leaves behind its growing point bursts out through the opening. These buds, and the shoots arising from them, are termed radical buds and shoots. They are anything but rare, and it would be an error to suppose that they only occur on the Aspen because that tree has been chosen to illustrate the subject. Not only a great number of trees, but also many shrubs, and a host of herbaceous plants, great and small, exhibit this kind of revival and multiplication, and for many species it is the safest and most fruitful mode of reproduction. It would also be wrong to suppose that radical buds only arise when the aerial parts of the plant concerned have been injured or destroyed in consequence of some unusual occurrence. A shock of the kind is certainly the most frequent cause; but it is equally certain that of trees and shrubs not a few develop rudimentary buds on their roots when their time comes — i.e. when they have become decrepit, and one branch after another is dying — without their having suffered any injury from worm or weather, or from the woodman's axe. A profuse after-growth of young plants always springs from the roots and surrounds old and dry trees of the following kinds: — the Aspen {Populus tremula), the Tree of Heaven {Ailanthus glandulosa), the Tulip-tree {Liriodendron tulipifera), and the Osage Orange (Madura aurantiaca), and the same statement applies to the following shrubs when they begin to wither — the Raspberry (Rubus Idceus), the Sea-Buckthorn (Hippophae), the Hawthorn (Cra- tcegus), the Barberry (Berberis), the Lilac (Syringa), and the Rose (Rosa), and to many other woody plants; whereas, no such "breaking" from the root is seen on young specimens of the above unless there has been some previous injury to the parts above ground. The budding power of roots is made use of by gardeners for the purpose of arti- ficial propagation. They cut pieces from the roots of the plants they wish to multiply and insert them in soil which is kept moist, and they may then count almost with certainty upon the development of several buds on each separate piece of root. This mode of propagation by root-cuttings or slips, as they are called, is attended by particularly successful results when applied to the flowering trees or shrubs of Cydonia Japonica, Paulotunia imperialis, Tecoma radicans, Dais cotoni- 28 BUDS ON STEMS. folia, and to various species of Acacia, Halesia, Hermannia, and Plumbago. More- over, the development of buds on roots is observed to take place not only in trees and shrubs, but also in herbaceous plants; and, indeed, in some it is of regular, annual recurrence. As instances of this may be mentioned the Dwarf Elder {Sam- bucus Ebulus), Asclepias Cornuti, Sophora alojjecuroides, Lepidium latifolium, the Dock (Rumex Acetosella), various species of the Toad -flax and Spurge (e.g. Linaria pallida, L. genistcefolia, L. vulgaris, Euphorbia Cyparissias), and several Composites and Pelargoniums. In another series of herbaceous plants the pheno- menon occurs exceptionally as a result of special external conditions, and chiefly in consequence of injuries, as, for example, in case of damage to the roots of certain Orchids (Epipactis microphylla, Neottia Nidus -avis), or of the Adder's Tongue amongst Ferns (Ophioglossurti vulgatum). Nor must we omit to mention the buds which are formed on aerial roots. There is so regular a production of buds from the columnar aerial roots of tropical Fig-trees, and of leafy shoots from the buds thus developed, that at first sight one is inclined to take the root-columns for trunks. BUDS ON STEMS. Buds and shoots growing directly from a part of the stem are termed cauline buds and shoots. Any part of a stem may become the point of inception of a bud. The commonest positions occupied by buds are the regions of the stem which bear respectively scale-leaves and foliage-leaves, and this is especially the case with those buds which subsequently become brood-bodies. But also low^er down and higher up buds are observed to develop, and do so, indeed, without the occurrence of any apparent injury or other assignable external cause. Thus, for instance, it frequently happens that buds are developed on the hypocotyl of the Scarlet Pimpernel {Ana- gallis arvensis), which abounds in our fields and kitchen-gardens, and the same is true of the species of Spurge {Eupyhorbia Peplus and E. vulgaris) which grow as weeds in company with the Pimpernel, and likewise of young Toad-flax plants {Linaria vulgaris), and of a few Umbellifers. These buds grow out immediately into green leafy shoots. In all probability the phenomenon occurs in many other plants besides, but hitherto the subject has received only cursory attention. These buds on the hypocotyl are all the more worthy of notice because they emerge below the cotyledons and in no case from a leaf-axil, i.e. the angle formed by a leaf with the stem. In the region of the foliage-leaves it is comparatively rare for a bud to originate at any other spot than in the axil of a leaf. As instances may be mentioned the extra-axillary buds of the Nightshades (Solanaceae), the buds in Serjania, Medeola asparagoides, Szc, which spring laterally from the stem close to the foliage-leaves, and those in the Vine and Virginian Creeper (Ampelideae), which are set opposite to the foliage-leaves. But even in these cases the positions of the buds, relative to the foliage-leaves of the stem, are always such as to be most naturally explained by the need of the former to obtain the formative materials produced in the green tissue of the leaves, in order to complete their own develop- BUDS ON STEMS. 29 ment; and these materials are most directly conveyed to them if they are situated as near as possible to the spot where the vascular bundles of a green leaf lead into the stem. When a large number of foliage-leaves are packed closely together upon a stem, it is scarcely possible for a bud to be developed in every axil. On such occasions the buds appear always to possess the power of selecting the most convenient points of origin. The majority of leaf -axils are altogether destitute of buds, and it is only at spots where their inception would be most favourable to the plant's development that a few hardy buds are put forth. This is what happens, for example, in most species of Spurge, in the Toad-flax, in Pines and Firs, in Araucarias, and the rest of the numerous family of Conifers. Where buds are formed in the axils of leaves, either there is one to each leaf, or several are crowded together in an axil, and of these one is conspicuous owing to its central position, and also usually for its size, whilst the rest are subordinate. The occurrence on the leafy region of the stem of buds crowded together in this fashion — the meaning of which will be examined in detail in the next few pages — is confined to certain species belonging to the Flora of the Mediterranean, of Australia, and of various Steppe-lands. They are much more commonly found on such regions of the stem as bear scale-leaves, especially in bulbous plants, which sometimes exhibit as many as a dozen little buds springing from the short, thick stem in the axil of one of the expanded scaly leaves of the bulb. The buds produced in the floral region of the stem (or inflorescence) usuall}- develop into flowers, and their function being the production of fruit, they cannot be considered until a later section of this work is reached. Meanwhile the bud-form of brood-body is not entirely absent from this region of the stem. Grasses, Saxi- frages, and Polygonums afibrd a great number of examples of their occurrence in that position. A wound may cause the formation of a bud at any altitude upon the stem. The bud invariably springs from the injured spot and often no relation can be detected between its point of insertion and the position of the leaves. An instance is known where the herbaceous stem of a Sea-Kale {Crambe maritima) was cut through transversely, and, after the pith had decayed, buds were formed on the inner surface of the vascular-bundle ring from the tissue of the so-called vascular- bundle sheath, and from the buds shoots eventually developed. If the main trunk or a branch of an Angiospermous tree, such as an Oak or Ash, is cut off' smooth, a mass of tissue is formed from the cambium, thus exposed, at the boundary between wood and bast; this tissue gradually creeps out from the margins of the wound and swelling up takes on the form of a circular rampart. The wood- cells which have been cut through and left bare within the circumference of the rampart have not the power of dividing and multiplying so as to initiate a new structure, but are dried up by exposure to the air and perish. The tissue foi'ming the rampart continues, however, to increase in breadth, and encroaches upon the dead interior of the section of the stump so completely that the cut surface of wood 30 BUDS ON STEMS. is quite covered over by the new growth. The latter is termed "callus", and may be compared to the tissue which is developed when an arm or a foot is amputated, and which grows from the ligaments beneath the skin until it gi-adually covers the whole stump. The callus in plants derives a special interest from the fact that within it are formed the rudiments of fresh buds, from which subsequently spring the shoots which " break " so plentifully. A longitudinal section through an Oak stump thus overgrown shows the callus wedged, as it were, between the old bast and the old wood; and we find that it consists of cork and parenchymatous cells, whilst vascular tissues, springing from the wedged portion of the callus, have also been developed, and, descending in bent and tortuous lines, establish an organic connection with the old trunk. The buds arising in the callus do not stand in any relation of any sort to the leaves, as has already been mentioned; nor do the intervals between them follow a geometric law, as is the case with the buds which take their rise from the axils of leaves. They are for the most part in aggregations and are produced anything but simultaneously. A callus of the kind may con- tinue to produce buds at appropriate spots year after year, and shoots of many different ages may be seen springing from it. One cannot contemplate such a callus growth, covering a stump and sending out shoots as direct ofF-shoots of the decapitated trunk, without being involuntarily reminded of trees that have been "ennobled" by grafting in the manner described in vol. i. pp. 213, 214. There is also an analogy to certain parasitic plants, such as Loranthus, in which the connection with the host is established in exactly the same way as that between callus-buds and tree-stump by means of a tissue interposed between wood and bark (cf. vol. i. p. 211). A formation of callus ensues upon the excision of the cortex from the side of a stem in the same manner as when the entire trunk is sawn through; and the process of covering up the exposed wood with callus, derived from the tissue lying between the bark and the wood, goes on similarly in the case of lateral injuries to the trunk. Some trees in addition exhibit a formation of callus without external damage having been received, as, for instance, the Ash, which has a bark liable to split and break open here and there spontaneously, whereupon a tissue of the nature of callus is formed in the open places. Oldish trunks of the North-American Ash {Fraxinus nana) are invariably covered with swellings and callosities of the kind, and most of them furnish starting-points for a score or more of buds. The buds which spring from growths of callus on trunks must not be confounded with those called by foresters " dormant eyes" and "dormant buds". Nor must we fail to distinguish them from the structures which have been termed superposed and collateral buds, which whilst exhibiting extreme diversity in their various modes of development, yet all constitute contrivances for the preservation of the plants from destruction in that their function is to replace dead shoots. With reference to the part played by these structures, it is most convenient to classify them under the name of "reserve-buds". They either originate simultaneously with those which they are destined in certain circumstances to replace, or they BUDS ON STEMS. 31 are only subsequently formed in the cortex in the immediate neighbourhood of the points of origin of shoots which have already withered. The latter is of com- paratively rare occurrence. In Spartium scoparium, which is represented in vol. i. p. 331, one bud only is produced in each axil. The following year, this bud grows out into a long switch, and at the same time a new bud is initiated in the cortical tissue just beneath the base of this shoot. If the first shoot dies next year, as often happens, especially in the case of plants growing near the northern limit of the Mediterranean region, the second bud produces a shoot, and close under its base is formed once more the rudiment of a bud for future substitution. This may go on for several years until at last a whole row of withered stumps are to be seen above the last substituted shoot. This mode of growth, which has been observed not only in Spartium, but also in several allied Papilionacese belonging to the Mediter- ranean Flora, is very prejudicial to the freshness and vigour of the plant's appear- ance. The presence of a number of withered remnants crowded together produces an impression of disease and starvation; else, as an alternative, one is tempted to suppose that the bushes have been cropped by cattle, or annually truncated by man, whereas they themselves accomplish all these changes without any damage of the kind being inflicted. In Robinia Pseudacacia, the plant known by the name of Acacia, a single bud is formed at first in the axil of each foliage-leaf. But later on the stem close to the thickened base of the petiole becomes hollowed out, and in the cavity thus formed little knobs arise underneath the first bud. Sometimes there is one only, sometimes there are two or even three. These knobs are nothing more or less than first rudiments of reserve-buds which develop in this position where they are sheltered and protected by the remaining portion of the petiole. If, as is often the case, in the following year the shoot put forth by the first bud dies, it falls to the uppermost reserve-bud to develop into a substitution-shoot, which may perish in its turn and be replaced by the next reserve-bud. The different species of the genus Gleditschia behave in precisely the same way as Robinia Pseudacacia, but in them the reserve-buds are only partially hidden beneath the remnant of petiole, and the power of forming new buds at the ends of the branches is here almost unlimited. In some species of Gleditschia, e.g. G. Caspica, a substitution of shoots, one for another, as they successively dry up, takes place for a period of ten or more years. The consequence is that the long branches of these trees are nodulated at the seats of origin of the buds, and the dried stumps of upwards of twenty short branches dating from previous years are seen crowded close together on these nodes. In Pterocarya Caucasica, a Caucasian tree allied to the Walnut, a single bud is formed every year in the axil of each foliage-leaf, and this bud has the peculiarity of being elevated from 1-5 cm. to 2 cm. above the leaf-insertion. Whilst it is j growing next year into a shoot, the rudiment of a reserve-bud is formed just above ' the original leaf-insertion, but it only develops in some subsequent year in the event of injury to the first shoot. Far more common than the above are the cases where the buds which sprout in 32 BUDS ON STEMS. the first year and those which remain dormant until called upon to replace the « earlier ones originate all together simultaneously. In the Common Elder (Samhucus nigra) two buds are formed one above the other in each leaf -axil; in the blue- berried Honeysuckle (Lonicera coerulea) and in several- of its allied species, three buds of almost equal size are superimposed one above another in a straight line in each axil. In the year following their formation, usually only one of them grows out into a shoot; the others stop as they are, and maintain their vitality for a couple of years in reserve and only then develop if the first shoot has met with destruction. The North- American False Indigo, species of which (e.g. Amorjyha fruticosa, A. glauca, and A. nana) are cultivated as ornamental shrubs in European gardens, produces two buds of different sizes above each foliage-leaf, the larger of the two being placed just above the smaller. The former sends forth a shoot in the following year, the latter remains in reserve. If the shoot first developed withers, as very often happens, the reserve-bud sprouts, and the withered stump of the first shoot is then visible just above the fresh one. The North-American tree Gymnocladus Canadensis also exhibits on the upper ramifications of its powerful branches two superimposed buds above the insertion of each leaf; the larger is situated above the smaller, and the latter only develops into a shoot in the event of its being required as a substitute. Several other woody plants which, though their stems become very thick, possess neither the growth of a tree nor a symmetrical crown of foliage — such as the Judas-tree (Cercis Siliquastrum) and the Forsythia viridissima of Japan — put forth long switch-like shoots, the upper halves of which often die off during the winter. The buds on the lower surviving half of each shoot are very close together, and generally they are in pairs, the upper one in each pair being close upon the lower. Only the upper one of a pair is at first developed in the next year; the lower bud does not develop unless the other fails. It is sometimes the case that the axil of every leaf produces three buds set side by side instead of one above another. The middle bud sends out a shoot in the following year whilst the lateral ones are left as a reserve. The year after, if the shoot has died, what happens is either that one of the two accessory buds develops — as, for example, in Lonicera fragrantissima and in the case of the long shoots of the Nettle-trees (Celtis Tournefortii, G. orientalis, G. occidentalis), or both accessory buds develop simultaneously — as in the Southern Reed (Arundo Doiiax) and in several species of the genus Bambusa. The species belonging to the genus Xan- thoxylon form in each leaf-axil the rudiments of from nine to eighteen buds, of which the middle one is the biggest and grows out during the following year into a short or long shoot. The other smaller buds are kept in reserve in the cortex at the base of the shoot. In the Tree of Chastity (Vitex Agnus-castus) four buds are set in the axil of each foliage-leaf. The central bud is the largest and a smaller one is situated under- neath it, whilst the other two — also smaller — are posted to the right and left respec- tively of the first. Next year a shoot is put forth from the large central bud whilst i BUDS ON STEMS. 33 the other three remain dormant. By the second year this shoot has probably perished, and in that case the little reserve-buds sprout. Their development is not infrequently simultaneous, so that here and there upon the tree we have tufts, each consisting of four slender shoots— one withered and three green — which all radiate from one point. If the three later shoots dry off at the ends, the buds on their basal parts produce fresh shoots, and the bushes present a bristly and not very ornamental appearance like besoms, especially when they are destitute of foliage. A curious development of reserve-buds may also be observed in A traphaxis, a ragged shrub indigenous to the Steppes of Southern Russia. Four buds are formed simultaneously and in close proximity to one another in the axil of every foliage- leaf. Of these a very small one is immediately above the insertion of the leaf; it has a large one above, and two of medium size on either side of it. The large bud becomes a leafy shoot and the small one a blossom. The two latei-al buds are kept in reserve unchanged during the second year, and in some circumstances during the third also. If the shoot dies, the development of the lateral reserve-buds is pro- ceeded with ; but as soon as they begin to sprout, the rudiments of fresh reserve- buds are formed in the cortex to the right and left of those that are thus developing. Here again, the ragged habit of growth of the shrub is connected with its peculiar mode of bud-formation. The following case is also very common. Of a crowd of axillary buds, placed either side by side or one upon another, one or more produce flowering shoots. When the fruits generated in the flowers have dropped — an e\ent in this connection equivalent to the fall of the shoots which bear them — and the spots of detachment are scarred over, the reserve-buds come into play for the first time. In Spiraea crenata there is only one such reserve-bud; in the Dwarf Almond {Amygdalus nana) and the Mahaleb (Prunus Mahaleb) there are two or three. The diversity amongst plants in this respect is almost endless, but the compass of this work does not admit of the subject being treated in greater detail. Seeing, however, that the facts involved have not received due consideration on the part of botanists, I should like to draw attention to the peculiar phenomena of development in Buddleia, Rhodotypus, Fontanesia, Philadelphus, Rubus, Berheris, Caragana, Alhagi, Lycium, and Ephedra, and also to point out that amongst woody, shrubby and sufiruticose Steppe-plants, which are especially liable to frost- bite and desiccation, many exhibit highly interesting characteristics in their development of reserve-buds. In Willows we find a form of reserve-bud which difiers from all the rest. It is obvious at a glance that every bud on an annual shoot of a Willow is entirely shrouded by a single scale shaped like a hood. This bud-scale originates in the outer layers of the cortical tissue, jind is, so to speak, a raised piece of the cortex covering the rudimentary bud. The large bud wrapped in this scale possesses an axis which has arisen laterally from the axis of the branch which bears the bud, and the vessels and cells of the wood may be followed uninterruptedly from the branch to the base of the bud. But, close to the latter, we also notice some very small bud- rudiments with no bundles running into them from the branch. They take their Vol. II. 63 34 BUDS ON STEMS. rise in a special cellular tissue intercalated in the cortex, and on a branch in its first year are not externally visible, because they are covered by the large hood-shaped scale. The tissue of cells from which these small buds spring might be compared to a callus if it were not produced on wholly uninjured branches and long before the formation of cracks and fissures in the bark. In the second year, when the large central bud begins to produce a lateral branch, throwing off the hood-scale and elongating its axis, the small buds also become visible in the form of spherical or oval knobs at the base of the new side-branch springing from the large bud. They do not, however, get larger or smaller, but remain completely dormant and unaltered. There is even a possibility of their never developing further, but in the event of the branch at the base of which they were produced receiving an injury and dying, they are aroused from their lethargy and grow out into leafy ramifications. It is obviously their function to replace such of their predecessors as fall victims to unfavourable external conditions. The Crack- Willows derive their name from the extraordinary fragility of their branches. The hard bast and wood at the base of their one-year-old and two-year- old branches exhibit a peculiar structure, the result of which is that a slight shock is sufficient to sever the tissue, so that the branch breaks across at its base and drops off. It seems to be an advantage for these Crack- Willows to get rid of certain leaf- less and useless twigs which bear nothing but the scars of shed catkins, and are merely an encumbrance. Thus much is certain, that several kinds of Crack- Willow cast off spontaneously a number of these branches, and that the buds above described as lying dormant in the cortex put forth leafy shoots as substitutes. Similar phenomena may be observed in Poplars. But in them the twigs break off at a little distance from the base, and the substitution of green, leafy branches for those covered with dead excrescences is effected by means of reserve-buds pre- formed in the axils of former bud-scales. There can be no question of mutilation in these cases any more than in the autumnal shedding of leaves which takes place spontaneously for the benefit of the plants concerned, and is not susceptible to the influence of external conditions except inasmuch as the latter may accelerate or retard it. In all the cases hitherto described, the substitution-buds are developed in the cortical tissue. At first, there is no direct connection between them and the woody tissue of the stem; it is only when these buds are roused from their lethargy, and called upon to put forth shoots, to replace anterior or collateral shoots which have fallen, that communication with the wood, and to that extent also with the current of crude sap, is set up by means of special conductive strands. There is, however, another form of accessory bud, which is connected from the very beginning with the wood of the stem appertaining to it, and maintains this during its whole life. In forestry the name of "dormant eye" already referred to is employed in particular for this form of bud. If a year-old branch is examined, it is found that the buds in the leaf -axils of its upper half are strikingly larger and more vigorous than those near the base; indeed, above the point of insertion of the BUDS ON STEMS. 35 lowest scale-leaves of the branch, it is not even possible in most cases to detect so much as a swelling that might be construed into the rudiment of a new bud. It is not till a longitudinal section is made through the lowest part of the branch that one perceives the existence of buds, here, too, in a very rudimentary condition and buried in the cortical tissue. The large buds to be seen at the close of the first year about the middle and at the extremity of the branch develop next year into fresh branches, the lower parts of which are again clothed with bud-scales, and the upper parts with foliage-leaves; but the small, inconspicuous or invisible buds at the base of the first year's shoot are left undeveloped and completely dormant. They are preserved practically unaltered in size or shape at the spots where they originated within the cortex, in some cases showing above the surface, in others concealed by the outer coats of the bark; and the only change which takes place is that the bundles leading from the wood of the branch to the dormant buds elongate yearly to the extent of the thickness of the new woody I'ing. These bundles exhibit the same disposition as those within the shoots which are visible on the surface, and so far, we might look upon them as latent lateral axes or side branches imbedded in the wood of the main branch and terminating in dormant buds. The analogy is confirmed by the fact that the lateral axes buried in the wood are capable of rami- fying in the same manner as those which project beyond the periphery of the stem and send their branches out into the air. The rudiments of fresh buds may also be formed on the concealed branchlets within the wood of the continually thickening main axis; and in many trees densely-branched structures terminating in dormant buds are formed in the wood of the stem, and exercise a disturbing influence on the course of the surrounding tubes and fibres of the wood of the main stem, causing them to bend and twist about to a very great extent. In this manner knobs of various sizes are formed, composed of the branched latent shoots which terminate in dormant buds and of winding wood-fibres. These nodules are found interspei'sed amongst the elements of the wood, which pursue a normal course, and they are known as "bird's eyes". Sections of such bird's-eye timber were much in demand some decades ago for use as veneering in cabinet-making, owing to the curious traceries exhibited by them, which usually take the form of eyes surrounded by rings and of serpentine lines — the former corresponding to latent branches, the latter to sinuous wood-fibres. As already mentioned, in many trees and shrubs it is particularly the buds pertaining to the axils of the lowest leaf -structures that are kept back in a dormant condition. A striking deviation from this habit is exhibited by the Tamarisks (Tamarix). The young branches, covered with innumerable little leaves and an assemblage of buds — usually three in number — are formed in the axil of each leaf. Want of space would of itself be sufficient to make it impossible that all these buds should produce shoots in the following year and develop simultaneously; about a thousand lateral branches would in that case be produced simultaneously from an axis little over a metre in length. As a matter of fact only comparatively few of the buds produce shoots, and these are so aptly distributed that no one of them 36 BUDS ON STEMS. restricts the freedom of another by pushing it aside or competing for its supply of air and light. Hundreds of rudimentary buds, not only at the base but scattered over the entire length of the branch, remain dormant in the Tamarisk branch, as it grows thicker and thicker, and thus is explained the fact that shoots springing from such branches have an almost inexhaustible store of lateral shoots, and are capable of producing every year afterwards hundreds of fresh shoots. Those reserve-buds which are formed in the cortical tissue and have no connec- tion with the wood of the stem which bears them, for the most part maintain their vitality only for a few years. The dormant buds at the extremities of latent branches may, on the other hand, preserve their capacity for development for many years, although they undergo no change either in shape or in size. No doubt many of them die in the course of a year or two without being replaced by others; whilst many others which perish have their places filled by new ones developed at the ends of embedded branches. But these are rare occurrences in comparison with the large number of cases where dormant buds retain their vitality for many years. Suppose the case of a tree one hundred years old, which has been shattered by a violent storm. With its crown of foliage torn down and its great branches broken off and strewn upon the ground, it reminds one of the ruins of a building of which roof, gables, battlements, and walls have been partially demolished. Where previously thousands of leafy boughs formed a spreading crown, now a few riven stumps are seen standing in dreary solitude. The tree has the appearance of being hopelessly destroyed, and one would anticipate that its trunk would dry up completely in the following year. Yet, marvellous to relate, fresh life quickens in the old and shattered trunk. Buds which have lain dormant in the cortex during scores of years stretch out, push their way through the fissures in the bark and develop into vigorous branches, and within a twelvemonth the thick stumps of the old trunk and branches are covered over with a drapery of fresh shoots which have buds set in the axils of their leaves. After another year has passed lateral branches develop from some of these buds, and this process continues until, in about ten years, the maimed tree becomes furnished with a new, densely-ramifying crown of foliage. Who, after witnessing such a phenomenon as this, can doubt that the arrested development of a portion of the cauline buds is an adaptation to ensure trees and shrubs against destruction in case of their being fractured by the wind or otherwise mutilated, or that dormant buds are to be looked upon as a reserve to meet possible accidents in the future! The fact that twigs which have shed themselves or succumbed to adverse external influences are replaced out of the store of dormant buds or by the buds of the callus, has led to various interferences on the part of man with the natural growth of cultivated plants, and has given rise to a whole series of methods of propagation, which have been employed by farmers and foresters ever since ancient times. To this class of operations belongs, for example, the method employed to promote the growth of underwood, which mainly depends on the development BUDS ON LEAVES. 37 upon the stumps left when the wood is cut, of new shoots from the callus or from the dormant eyes, shoots which in the course of thirty or forty years replace the old plantation, that is to say, the mass of wood which has been taken away. Pollarding is another instance. Pollard-trees are kept cut down to a particular height, and in consequence become thickened at the top, as may be seen in the case of Poplars, Ashes, and more particularly Willows. The pruning of Vines and Fruit- trees is likewise of this category, and the same process is applied also to the woody plants trained to form espaliers or hedges when a park is being laid down or an estate inclosed. All these manipulations have in view, on the one hand, a develop- ment of more vigorous shoots from the stumps that are left behind and the acquisition of as abundant a yield of timber, forage, or fruit as possible; on the other hand, a denser growth of the tree-top, or a stunting of the tree, such as is required for gardens in the old French style, with their formal green walls, obelisks, and marvellous ornamentation. Seeing, however, that each of the various trees and shrubs has peculiarities of its own in relation to the formation of callus and dormant eyes, many different modes of pruning are applied to them. We cannot generalize from one case to all the rest, and it would be a great mistake, for example, to try to pollard Apple-trees like Willows, or to convert Pines into under- wood. Climatic conditions must also be taken into account in connection with these intentional mutilations of cultivated plants. To give one instance of their effect, it may be mentioned that vine-pruning in Hungarian vineyards is quite different from the corresponding process employed on the Rhine, whilst the latter again differs from the method practised in Northern Italy, which, in its turn, is not the same as that of Southern Italy. In each locality the kind of treatment most adapted to prevailing climatic conditions has been found out in course of time. BUDS ON LEAVES. Hitherto only such buds as are developed on roots or on the various regions of the stem have been dealt with; but an enumeration of these does not nearly exhaust the multiplicity of bud-forms which exist. Buds and shoots may also spring from the tissues of leaves — particularly foliage-leaves. These are termed epiphyllous buds or shoots, and they are classified in several groups according to their places of origin. Befoi-o discussing this classification it is necessary to note carefully that epiphyllous buds must be strictly distinguished from those which occur en the foliage-leaves of Helwingia and on the leaf-like cladodes (or phylloclades) of Butcher's-broom, &c. As regards Helwingia (see fig. 198) careful investigations prove that certain strands proceed from the leaf-bearing axis to the buds seated upon the leaves. Each of these strands represents a lateral axis, but instead of being free it is bound up (or fused) with the midrib of the leaf from the axil of which it springs. The lateral axis thus adnate to the midrib first abandons its connection with the latter at a spot on the lamina, about a third of its entire 38 BUDS ON LEAVES. length from the base. It there terminates in a bud, or, if it divides, in several buds, and inasmuch as these are flower-buds, it may be looked upon as a flower- stalk. These buds cannot therefore be said to be epiphyllous, i.e. to spring direct from the tissue of a foliage-leaf. In reality each is borne upon a structure of the nature of a stem, only the peduncle, stalk, or axis has partially coalesced with the midrib of a leaf. Willdenow, who was the first to describe it, named the plant, represented in fig. 198, the Butcher's-broom Helwingia {Helwingia Tuscifiora), ^^*-K.: Fig. 198.— Helwingia rusciflora, with flowers seated upou the foliage-leaves. because the floral buds here as in the Butcher's-broom (Ruscus) were borne by foliaceous structures (c/. vol. i. p. 333). The two cases are, however, essentially different. The green leaf -like structures in the Butcher's-broom, which carry floral buds upon their upper surfaces, are not leaves at all, but leaf-like shoots, that is to say axes, and the buds upon them are, therefore, not epiphyllous but cauline. The same statement applies, of course, to other plants with flat, expanded shoots, a few representatives of which are shown in the illustration of p. 335 of the first volume, and in this category must be included Ferns also, if we look upon their fronds as phylloclades, and not as foliage-leaves. It would be quite out of place here to enter into the question of the nature of fern-fronds, or to set forth the reasons why they must be considered as phylloclades. The proof cannot be BUDS ON LEAVES. 39 conveniently introduced until we come to the description of Ferns themselves. It is sufficient to mention here that buds very frequently occur on the fronds of Ferns; indeed, certain species, e.g. Asjdenium bulbiferum (see fig. 200) develop buds on almost all their fronds. In most cases they spring from the surface of the green pinnre, but in Ceratopteris thalictroides, a common denizen of swamps in the East Indies, it is from the little stalks of the ultimate green lobes, in Fig 199 — Foimation of Buds on the apices of tlie Fionds of Feins Aipleniuiii EdjeuoiUu Gleichenia from the angles of the forkings of the fronds (cf. fig. 189^), and in AspleniuTYi Edgeworthii (see fig. 199), from the apices of the fronds, that is to say from the extremities of the cladodes. The last-mentioned Fern grows upon the bark of trees, and the tips of its fronds are endowed with the property of avoiding the light, in other words, they bend towards the darkest parts of their substratum, creeping into the fissures in the bark, where they become firmly adnate, and each develops a bud above the point of contact. This bud gives rise once more to fronds, of which, however, one only, as a rule, develops vigorously. After it has unrolled itself, this new frond in turn searches with its apex for a dark rift. The 40 BUDS ON LEAVES. process is repeated over and over again, and results in the trees, upon the bark of which the Aspleniuvi has established itself, being regularly encircled and woven over by fronds, as is shown in fig. 199. The separate fronds of the fern in such circumstances have a strong resemblance to the runners of certain species of Veronica, Ajuga, and Periwinkle, which have their leaves arranged in two rows. Unlike the above cases — viz. the buds of Helwingia borne on special stalks adherent to the leaves, those growing on the cladodes of the Butcher's-broom, and those on the fronds of Ferns, all of which must, in spite of their extreme similarity to epiphyllous buds, be looked upon as cauline — true epiphyllous buds always arise from cells of a true leaf and have no connection with adjacent axes beyond that involved in the fact of the bud-producing leaf being derived like all other leaves from a stem. Epiphyllous buds are even produced by leaves severed from the axis; indeed, in many instances, the severance of the leaves is itself the apparent cause of the development of the buds. This phenomenon is exhibited, for example, by Bryophyllum calicinuTn, a plant of the House-leek family which belongs to the tropical parts of the Old World, but has long been cultivated in our greenhouses and has attained a certain celebrity even in non- scientific circles, owing to the fact that Goethe interested himself in it and mentions it repeatedly in his writings. The foliage-leaves of this Bryoiihyllura (see fig. 200 ^) are deeply divided, the separate lobes being oblong-obovate and conspicuously notched. Every full-sized leaf exhibits in each notch of the margin a group of cells, which is perceptible as a dot to the naked eye. So long as the leaf remains upon the stem there is usually no further development of these cell-aggregates, but if the leaf is plucked oflf and laid on the earth an active process of division is set up in them, the result of which is the formation of a little plant with stem, leaves, and roots, as is represented in the figure opposite. The leaves of Bryophyllvjm calicinum are thick and fleshy, and contain when mature such an abundance of reserve material and water as to render it super- fluous that any absorption of nutriment from the environment should take place. It is not till later that the little plants which spring from the notches of the leaf, having used up the materials stored in the latter, are driven to seek food from the environment by means of their roots. If the leaf has been laid on moderately damp earth, the rootlets of the young plants, developed in its notches, penetrate the ground and, in the event of the tissue of the leaf being in the meantime exhausted and withered, all the little plants become independent and develop into full-sized individuals. Phenomena similar to those exhibited by Bryophyllum calicinum are also observed in other plants with thick, fleshy leaves, particularly in Echeverias. Young plants also make their appearance sometimes on the fleshy leaves of Rochea falcata after they have been picked. There is, it is true, the noteworthy diflference that the phenomenon is not foreshadowed, as in Bryophyllum, by the existence of special groups of cells at the points of origin; but Bryophyllum, Echeveria, and Rochea have this in common, that in all cases the need of materials for the construction of the young plants is met BUDS ON LEAVES. 41 by the succulent leaf for some time after its severance from the stem, so that it is not necessary to place the leaf in communication with damp earth with a view to its deriving the requisite water therefrom. They are thus exempted from conditions to which the greater number of plants propagated by gardeners by means of so-called leaf -cuttings are subject. This method of propagation by leaf-cuttings has long been recognized, and has been particularly applied to Citron and Orange trees, as also to the Wax Fig. 200.— Formation of Buds on Fronds and Foliage-leaves: 4, 3 on the pinnules of Aspleiiium bulbiferum; * on the margins of the lobes of the leaves of BryophyUwn calicinum ; « on the foliage-leaves of Cardamine pratensis; 5 on the margin of foliage-leaves of Malaxis paludosa. « Two buds on the margin of a leaf of Malaxis paludosa. >, *, *, * natural size ; 2x2; « x 20. Flower {Hoy a carnosa), to Theophrasta Jussieui, a plant belonging to the Myrsinese, to the Aucuba Laurel (Aucuba Japonica), to the beautiful Clianthus puniceus, and to various other plants besides. But it is only quite recently that it has been practised on a vast scale, since the discovery that the Begonias, introduced from the tropical parts of America and now so fashionable as orna- mental foliage-plants, and the Gesneracese from Brazil with their splendid flowers, are capable of being propagated with extreme facility and in immense numbers by means of their leaves. The cultivator has only to pick one of the foliage- leaves and place it in contact with moist sand or sandy soil, and in a short time 42 BUDS ON LEAVES. young plants sprout from the leaf and may be transplanted as independent growths. We will briefly describe what takes place. The first change observed in a leaf which has been cut off" for the purpose of forming cuttings is the desiccation of the cells lying next the cut surface. Beneath the layer of dried-up cells a cork-tissue is formed, whilst the dead, outer layer is converted into a bark. A parenchymatous tissue is next formed from the part beneath the cut which is still living; indeed, it is the epidermal cells nearest to the dead layer of cells that initiate this formation of tissue. They grow in a radial direction, elongating and dividing by means of the insertion of transverse walls, the result being a uniform thickening coextensive with the surface of the wound. A little later some of the living cells in the middle of the cut, which are still covered over by the dead layer, begin to divide; and as the tissue there grows in size, it tears the overlying dried layer into shreds and pushes it off" in parts. This exuberant tissue has received the name of callus. Whilst the formation of callus is proceeding, suckers are developed at the points of contact of the leaf-cutting with the sand, their numbers being particularly abundant along the projecting ribs of the leaf. In form and function these suckers are entirely similar to the absorbent cells lying close to the growing extremities of roots, and called root-hairs. They are of the greatest importance to the leaf-cuttings in their subsequent processes of development. So long as the leaf adhered to the axis it was supplied with a sufficient quantity of watei from that which was ascending through the stem; the aqueous vapour lost through evaporation was replaced by moisture absorbed by the roots from the damp soil and afterwards conducted through the stem to the leaf in question. But when the leaf has been cut off" it is no longer able to derive any material from the earth through the intervention of the stem, and as its ordinary epidermal cells have not the power of taking up from the damp soil, which serves as sub- stratum to the leaf-cutting, as much water as is lost by evaporation, the cutting is exposed to the risk of desiccation in spite of its being in contact with a wet substratum. In order to escape this danger and save itself from destruction the leaf treated as a cutting furnishes itself with absorbent cells. By their instrumentality the water, which is particularly needful for the formation of callus, is put by. Even if the materials necessary for the construction of the cells of the callus may be present in abundance in the cells of the leaf, it is of little avail unless these materials are diluted and conducted to the places where they are used up, and for this a much greater quantity of water is requisite than could be retained by the severed leaf. When the callus has reached a certain size numerous roots make their appearance. They usually take their rise from cells of the parenchyma adjacent to a vascular bundle of the leaf, break through the callus, and grow rapidly in length. Only after the development of these roots, which absorb liquid copiously from the substratum by means of their suction-cells, are buds produced on the upper — less frequently also on the under — surface of the leaf-cutting. In Begonias it is chiefly cells of the epidermis BUDS ON LEAVES. 43 that give rise to buds; in other plants, particularly in the Gesneracese, in the species of Peperomia, a genus belonging to the Pepper order, in Tournefortia. Citrus, &c., it is cells of the callus that divide and become the rudiments of buds, and indirectly of shoots. In the case of Begonias isolated buds occasionally spring from the callus in addition to the others, but this is not inconsistent with the fact that in these plants the epidermal ceils are the favourite places of inception. Especially are those epidermal cells preferred which are situated above the bifurcation of a vascular bundle in the lamina. If an entirely uninjured leaf is laid upon moist sand, the buds develop just above the base of the lamina where the strands radiate out from one another. It is a common custom of gardeners, however, when making use of Begonia-leaves to propagate the plant, to set the petiole in wet sand and to make a number of transverse cuts across the larger veins of the lamina, which is laid flat upon the sand. After this operation quite a host of buds — i.e. new plants — take their rise all along the course of the intersected vein, some immediately in front of the cut, which is covered by a callus, but frequently others again at a distance from that spot. From this we may conclude that the new formation depends principally upon the conduction of material by the veins. No doubt its relative position with regard to the roots developed from the callus to the stock of reserve materials and so forth, also play an important part. The upshot is, however, that numberless cells of the epidermis of the leaf become the seats of inception of new plants, and that buds are able likewise to develop from deeper-lying cells of the callus. Whether the development of an epiphyllous bud has begun in one place or another, there is always in the inceptive area a concomitant pro- duction of vascular bundles, which establish a connection between the axis of the bud in process of formation and the previously-developed roots; and it is not long before the axis produces green foliage-leaves capable of assimilating in the presence of light. The leaf -cutting, upon which a miniature plant is now seated, in most cases retains its vitality for a considerable time longer, but at length it begins to turn yellow and gradually it dies. Only that part which produced the buds and roots persists in the form of a pad, forming in some species, foi example, in Begonias, a thick, fleshy, cellular body, looking almost like a little tuber. The phenomenon above described as ensuing in consequence of artificial manipulations takes place sometimes spontaneously in nature in a few plants, and that without the leaf concerned in the process being separated from the axis. Ex- amples of plants which have been observed to bear occasional epiphyllous buds when growing wild in their natural habitats are Cruciferae (Cardamine pratensis. Nasturtium officinale, Roripa palustris, Brassica oleracea, Arabis pumila), Papaveracese (Ghelidonium majus), Water-lilies (Nymphcea guianensis), Gesneraceae {Episcia hicolor, Ghirita sinensis), Lentibularieae (Pinguicula Backeri), Aroideae {Atherurus ternatus), Orchidacese {Malaxis monophyllos and 31. paludosa), Liliacese (Fritillaria, Ornithogalum, Allium, Gagea, Hyacinthus) and Amaryllideae 44 BUDS ON LEAVES. {Cnrculigo). In many cases the buds which arise in the form of little papillae grow straightway into miniature plants, as in the case of the Cuckoo-flower (Gardamine pratensis, see fig. 200*), or else little bulbs are formed in the first instance, as in the various species of Garlic and in the Crown-imperial (Allium and Fritillaria), or small tubers, as in the above specified instances of the genus Malaxis. In the one case cells situated in the middle of the lamina — usually above the point of bifurcation of a vascular bundle — are the seat of origin of buds, as, for example, in the Cuckoo-flower, already so often referred to; in other cases, such as Curculigo, the buds spring from the extremity of the midrib. The little orchid Malaxis paludosa (see fig. 200 ^), which is a native of mooi'lands in North- western Europe, develops its diminutive buds principally on the surface and margins of the upper portions of the green foliage-leaves, and these buds appear in such large numbers that several botanists state in their descriptions that the leaves of Malaxis paludosa are for the most part "shortly ciliated". Of all the manifold kinds of epiphyllous leaves these little structures produced on the green leaves of the Orchid in question possess a surpassing interest on account of their form. Each bud (two of which are shown in fig. 200 ^) consists of a yellowish-green cellular body, shaped like a kernel, and of a layer of cells hanging loosely together and enveloping the kernel like a sac. At the free extremity the cells of the envelope form a kind of ring, which constitutes the rim of a round depression. The resemblance of these buds to the seeds of Orchids, especially to those of Malaxis paludosa, is obvious on the most cursory examination, and it will again be referred to in a subsequent section. Buds are found much less frequently on scale-leaves and floral-leaves than on the green foliage-leaves. Sometimes they may be observed to spring from bulb-scales if the latter are stripped off" the axis and put into moist sand. In these cases they are invariably developed at the spots where the scales have been cut and injured. Dutch cultivators of bulbs make use of this property to propagate hyacinths direct from the bulb-scales. They cut out the axis of the bulb, remove also any rudiments of floral axis which may be present, and cut transversely through the lower part of the bulb-scales. Not infrequently the bulb-scales are also partially divided longitudinally. One would think that after such treatment the bulb must sooner or later perish; but, on the contrary, a crowd of small bulb-like buds are produced on the scales at the edges of the cuts, and cases are known of over a hundred young bulbs being obtained in the manner described from the scales of a single hyacinth bulb. Of all epiphyllous buds those originating in the tissue of floral-leaves are, as stated, the least common. Minute buds have, however, been repeatedly observed to be developed, instead of seeds, on the carpels in the interior of the fruits of several species of Crinum and Amaryllis. They were seated on round bodies of tissue, which were not distinguishable from little tubers. When laid on damp soil, each produced a new plant. We need only allude here to the cases of parthenogenesis, which will be discussed later on, wherein seeds capable of BUDS ON LEAVES. 45 germinating are developed without fertilization from the ovules concealed in the ovary. The instances of bud-formation above enumerated, when considered with respect to their origin, show that not only cells of roots, but also those of all refj-ions of the stem, and of scale-, foliage- and floral-leaves may become initial cells of buds, or, in other words, of rudimentary shoots. Hence we may draw the conclusion that all the living protoplasts which are capable of division in whatever part of the plant their cells are situated, from the root-tip to the highest apex of the stem, and from the scale-leaves to the ultimate floral-leaves, have the power of undertaking the function of renovation without previously undergoing fertilization. Under ordinary circumstances, no doubt, it is only protoplasts in the cells of the axis, close to the spots where the foliage-leaves emerge, which turn into rudiments of shoots, and the most natural explanation of the selection of these places is that the constructive materials prepared or temporarily deposited in the foliage-leaves may there be turned to account at first hand; but in extra- ordinary circumstances — i.e. as a consequence of unfavourable climatic conditions, or of dangerous injuries, and particularly under the influence of approaching peril of death — the important task of initiating new plants devolves also upon cells situated at most widely different parts of the parent stock, cells which otherwise would certainly not have assumed this function. In these cases it is astonishing to see how stress of external circumstances results in an entirely new division of labour in the cells of the tissue affected thereby; how in one place a protoplast, originally destined to play an altogether diflferent part, divides and becomes the starting-point of a fresh plant, whilst the protoplasts of neighbouring cells convey constructive materials to that particular member of their fraternity and are regularly consumed by it. Very different would have been the order of things and the kind of co-operation of adjoining protoplasts under ordinary conditions! 46 DEFINITION AND CLASSIFICATION OF FRUITS. 2. REPRODUCTION BY MEANS OF FRUITS. Definition and Classiticatiou of Fruits. — Fertilization and Fruit-formation in Cryptogams. — The Commencement of the Phanerogamic Fruit. — Stamens. — Pollen. — Arrangements for the Protec- tion of the Pollen. — Dispersion of Pollen by the Wind. — Dispersion of Pollen by Animals. — Allurements of Animals with a view to the Dispersion of Pollen. — The Colours of Flowers considered as a means of attracting Animals. — The Scent of Flowers considered as a means of attracting Animals. — Opening of the Passage to the Interior of the Flower. — Reception of flower-seeking Animals at the entrance to the Flower. — Taking up the Pollen. — Dispersion of the Pollen. — Cross -pollination. — Autogamy. — fertilization and Fruit-formation in Phanerogams. DEFINITION AND CLASSIFICATION OF FRUITS. To all appearance there is no difference between the protoplasts which develop into brood-bodies and those which are the points of origin of fruits. Nevertheless, it has been ascertained by experience that the protoplast, which is the starting- point of a brood-body, evolves its constructive energy without receiving any special stimulus from the protoplasm of a second cell of distinct origin, whereas for the development of a fruit the necessity of such a stimulus is a characteristic and distinctive feature of the phenomenon. Brood-bodies may spring from any part of a plant. If the parent-stock as an individual is in danger of perishing, brood- bodies are developed from protoplasts which otherwise would never have been called upon to play such a part. Brood-bodies may develop on roots, stems, and leaves, on foliaceous prothallia, and on hyphal filaments. They may be formed above or below the ground, and upon or beneath the surface of water. Their origin may be from superficial cells or from cells deeply seated in a tissue. It is scarcely going too far to say that in cell-aggregates of large dimensions the protoplasm of every young cell is potentially the starting-point of a brood-body. If a fruit is to arise, the ooplasm, i.e. the protoplasm destined to initiate a new generation, must unite with the fertilizing protoplasm, which is called spermato- plasm. The two protoplasts concerned in this phenomenon originate at separate spots, and if they are to coalesce the space between them must be surmounted. One at least of the two protoplasts must accomplish a change of place, and this locomotion must take place in a definite direction. The union of two protoplasts which have been formed at places separated in space from one another constitutes the essence of the process of fertilization, and it results in a change in the ooplasm which, in accordance with our idea of the minute structure of the substances in question, may be looked upon as a displacement of molecules and an alteration in their grouping. Sometimes this internal rearrangement is plainly manifested externally by a change of form and colour, or by an increase in size; and where this occurs it ensues immediately upon fertilization. But for the most part no alteration in the fertilized ooplasm is perceptible at first, and it would be difficult to specify any certain signs whereby the fertilized ooplasm may be distinguished from the unfertilized. It is, however, known by experience that in most cases DEFINITION AND CLASSIFICATION OF FRUITS. 47 the unfertilized ooplasm dies without developing further, whereas the fertilized ooplasm, after a longer or shorter period of rest, exhibits a characteristic growth aud becomes the point of origin of a young organism, the new generation. The ooplasm rendered capable, by fertilization, of this particular kind of growth is to be considered as an embryo, even in cases where no outwardly-visible change in form, size, or colour has taken place. Both ooplasm and spermatoplasm are formed in special cells at definite spots on a plant which is preparing to reproduce itself by means of fertilization. The cell-chamber wherein the ooplasm is developed, and which is itself adapted to the reception of foreign matter, and constitutes the point of origin of the embryo, will in future be called an oogonium {4ov = egg; 76i'os = parentage); the cell wherein the spermatoplasm is brought to the proper form and composition for the purpose of fertilization is called an antheridium in the case of a Cryptogam, and a pollen- grain in the case of a Phanerogam. In a few instances the ooplasm is set free from the oogonium and fertilized outside it; the oogonium has then, of course, nothing more to do with the subsequent processes of development. In other cases fertilization takes place within the oogonium; the oogonium persists in a more or less altered form as the immediate envelope of the embryo, and is then designated by the name of "carpium" (/cap7r6s = fruit), or briefly "carp". In yet other instances it is possible, at the very earliest stages of development, to distinguish a special multicellular envelope surrounding the oogonium. To this envelope we may apply the term " amphigonium " in order to simplify the terminology. If the amphigonium is later on converted into the coat of the carpium, it may be called an "amphicarpium". In many plants this envelope to the oogonium is succeeded externally by a second called a " pericarpium ", which will be the subject of more detailed study later on. Now what ought we to take to be the fruit? To try to conform to ordinary usage, or to adopt the terms employed in other sciences, would cause fatal confusion. The most expedient course, therefore, seems to be to put aside the names and definitions adopted in other departments and to lay down an independent and unambiguous definition of the plant-fruit, and apply it to all plants. Thus, from the botanical point of view, we consider every structure to be a fruit which is the product of fertilization, and at the same time constitutes the first step towards the renewal of the fertilized plant. This definition includes the ooplasm, which is fertilized outside the oogonial envelope, and forms the starting-point of a new individual; there may, therefore, be fruits each consisting of nothing more than an embryo. But usually the ooplasm is enveloped by a coat, which may be single or double, or even threefold. Fertilization then takes place within these coverings, and the influence of the spermatoplasm extends more or less beyond the ooplasm to its investments. In such cases the coats also are involved in the process of fruit-formation. They are stimulated to grow in a particular manner and take the form of a mantle clothing the embryo, of a protective cover, or of some contrivance which promotes the further development of the embryo and its full 48 DEFINITION AND CLASSIFICATION OF FRUITS. expansion into a new generation. Fruits of this kind have sometimes a very- complicated structure. In them we are able to distinguish a complex outer coat, and within, the embryo with its tightly adherent covering, the latter portion of the fruit being that which has from ancient times borne the name of seed. Fruits thus come before us as a series of forms, of which the members at opposite extremities of the series differ greatly, but are linked together by a large number of intermediate forms. At one end of the chain we have the unicellular fruits of the microscopic Desmids, at the other the fruits of the Cocoa- nut, which is differentiated into seeds on the one hand and several envelopes on the other, and is as large as a man's head. As already stated, the spermatoplasm acquires the composition and form whereon its fertilizing power depends within the confines of certain special cells. Extreme variety is, however, found to prevail in this connection. In some plants, especially those which conduct the process of fertilization under water, the spermatoplasm takes the form of minute particles usually furnished with special motile cilia to enable them to swim about. These have received the name of spermatozoids. They escape from the cell-chambers in which they were formed into the water, rush about for a short time or are carried by currents in the water, and finally reach the ooplasm, whereupon they place themselves in contact with it, and enter into combi- nation with it in a manner which may best be likened to the merging together of two drops of oil floating upon the surface of water. In another category of plants the spermatoplasm does not escape from the cell in which it has been developed, but this cell itself enters into combination with the oogonium as a whole, and a possibility is afforded in a variety of ways for the two kinds of protoplasm to coalesce within a single enveloping cell-membrane. A third category of plants is remarkable for the fact that the spermatoplasm does not coalesce as a whole with the ooplasm, only a portion of it passing to the ooplasm. The above prefatory remarks give some idea of the extreme variety which exists in the processes of fertilization, and it is no easy matter to give a short and concise, and at the same time accurate, presentation of the facts involved, especially if one tries not to use more than is absolutely necessary the innumerable technical terms invented in recent times. Even taking into account only the most important of the plienomena above referred to, we find twelve different processes or types of fertiliza- tion and fruit-formation, and it will be the object of the next chapter to present these in order, beginning with the simplest cases and concluding with the most complicated. It will materially conduce to clearness of exposition if, in considering these phenomena, we adhere to the old classification into Cryptogams and Phaverogams, which was introduced by Linnaeus. According to the etymology of the words, Cryptogams are plants which are fertilized secretly, whilst in Phanerogams the process of fertilization is apparent. Since the microscope has been perfected and brought into common use this distinction has no doubt lost its significance. If, however, we adopt a somewhat different interpretation of these terms, we may I FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 49 continue to use them with advantage. Thus, under the name of Cryptogam we shall include all plants destitute of flowers in the ordinary sense and possessing organs of fructification which are not clearly visible excepting under the microscope* whilst the term Phanerogam will comprise such plants as bear flowers, and have organs of fructification which are visible without aid from the microscope and are of the nature of metamorphosed leaves. The retention of these old and familiar terms is rendered all the more desirable by the fact that another important distinction, which is inherent in the process of fertilization itself, and has not as yet received sufficient attention, is involved in the separation of Cryptogams and Phanerogams, namely, that in Cryptogams fertilization takes place in water or in a watery medium, whereas the process in Phanerogams is accomplished almost exclusively in the air. FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. In the mountain districts of Central Europe, after the winter snow has melted and the turbid water derived from it has gradually cleared itself up, a beautiful sight is afforded, especially when a ray of sunshine strikes the water, by the dense crowds of short delicate filaments of a bright emerald-green colour, whicli every- where form a coating to the stones at the bottom of streams and to the sides of the troughs used to convey spring-water from the heights. These green threads belong to a plant named Ulothrix. Each separate filament consists of numerous cells joined together so as to form a chain, as is shown in fig. 201 \ When these filaments are 1 mature, and the time has come for the production of fruit, the protoplasmic contents of the separate cells break up into a number of spherical green masses, which, how- ever, continue to be held together in a rounded cluster by means of a colourless substance. An aperture is now formed in the wall of each of the cells in question, and through this opening the conglomerate mass escapes into the surrounding water (see figs. 201 ^ and 201^). The individual masses of protoplasm which compose the con- glomerate are set free shortly afterwards, and each exhibits at its anterior extremity a pair of revolving cilia, by means of which it swims about in the water (fig. 201*). When in the course of their peregrinations two protoplasts which originated in jone and the same cell-cavity encounter one another they get out of each other's way; if, on the other hand, the protoplasts from cells belonging to different filaments meet, far from avoiding one another, they come into full collision with their anterior ciliated extremities, turn over, and lay themselves side by side and coalesce, forming a single body with four cilia (see fig. 201 °). A little later the cilia vanish, and the product of the coalescence comes to rest. This fusion is the simplest conceivable case of fertilization in the whole realm of plants. The product of fertilization is ^^he fruit. It consists in Ulothrix of the little lump of protoplasm formed by the iprocess of coalescence just described, which now surrounds itself with a thick cell- JTiembrane, and fastens on to some stationary body under water (see fig. 201 ^). Wl* lave nothing to do at present with the subsequent development of this fi-uit ; it is Vol. II. 50 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS, sufficient to remark in order to explain the illustration that the attached unicellular fruit does not produce again immediately a string of cells, but that first of all swarmspores are developed from its protoplasm (see figs. 201'' and 201^''), and these fasten on to appropriate spots, inclose themselves in cell membranes, divide and ulti- mately initiate new filaments composed of cells arranged in linear series as before. In Ulothrix and allied genera the protoplasts which pair as a first step to the formation of fruit do not diflTer from one another in form, size, colour, or mode of locomotion, and it would be impossible to determine from outward appearances which of them acts as fertilizer and which is fertilized. The terms ooplast and Fig 201.— Fertilization and fiuit-fonnation in Ulothrix zonata (partly after Doilel-rort). • Two filaments composed of cells joined together in chains. 2 Escape of conglomerated gametes. 3 Spherical conglomerate of gametes after it has escaped. * Separation of the gametes. ' Gametes swimming about and pairing. « Fruits (products of the pairing of gametes) attached to a substratum. '-9 Subsequent development of fruit. i<> Two swarmspores produced by fruit. 1 x 250; 2-io x 400. spermatoplast are therefore not applied to them, but they are called gametes, and the entire process described in connection with them may be spoken of as fruit- formation by pairing of gametes. This process of pairing is, so far as it can be apprehended by our senses, a mutual permeation of the two protoplasmic bodies, and we may suppose that a rearrangement of molecules is caused thereby, which endows the product of pairing with the power of developing independently. This assumption is supported in particular by the fact that if any gametes, after being set free from the conglomerate, fail to pair they undergo no subsequent develop- : ment but deliquesce in the surrounding water and perish. The Wi'acks or Fucaceas, which grow profusely in the sea, resemble Ulothrix\ inasmuch as the protoplasts, destined to act as fertilizers, escape from their cell- cavities, fertilization consisting of a fusion of freo protoplasts disconnected from the I mother-plant. But these Wracks dififer very strikingly from Ulothrix and allied forms in that the protoplasts are of two kinds, there being an obvious diversity in FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 51 T a*^ '■m- ^ ■^,r /^^^ size and form between ooplasts and spermatoplasts. The thallus in all species of Fiicus is tough and leathery, brown in colour, foliaceous, and dichotomously branched or lobed, and has interspersed here and there air-containing swellings which serve as floats. The apices of the lobes are punctate, and each spot corresponds to an internal cavity which has the form of a globular pit (see fig. 202 1). SectioiLs through these cavi- ties show that a large number of segmented filaments known as " para- physes " spring from the lining- layer of the cavity. In Fucus vesicu- losus (figs. 202 and 203) these filaments remain concealed in the cavity; in some other species of Fucus they pro- trude through the narrow orifice (osti- ole) of the cavity in the form of a pencil of hairs. Amongst the fila- ments other struc- tures are also formed within the cavity. A few of the cells lining the cavity swell into papillfe, and each becomes divided by the intercalation of a transverse septum into two cells, one of which is spherical, whilst the other assumes the form of a stalk bearing the upper one (see fig. 202 ^). The protoplasm in the spherical cell is dark brown, and breaks up into eight parts, which round themselves off" and con- stitute the ooplasts. The thick wall of the spherical cell resolves itself into two layers, of which the inner one incloses the eight rounded protoplasmic bodies like a bladder. This bladder stuffed full of ooplasts next detaches itself entirely, and glides upward between the paraphyses until it reaches the orifice of the cavity. Fig. 202.— Fucus vesiculosus. Longitudinal section through one of tlie cavities in the thallus. - A vesicle surrounded by paraphyses from the bottom of the cavity, s a detached vesicle containing eight ooplasts ; the inner lamella swollen up. * Liberation of the ooplasts from a rent vesicle. (After Thuret.) 52 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. Here the bladder splits into two lamella, and finally the inner lamella becomes inflated, bursts and shrivels up, leaving the eight ooplasts free (see figs. 2023and 202 ). Whilst a certain proportion of the individual plants of Fucns ves^culosus develop ooplasts in the cavities in their lobes, other individuals give rise to spernmto.oxds ^ in similar cavities (see fig. 203^). The cells lining the hollows de- velop papillose protuber- ances which grow longi- tudinally, divide and form a ramifying mass, of cells as is shown in fig. 2032. Here and there the extremities of branches in this mass of cells have a dark brown colour, and their proto- plasmic contents are broken up into a number of minute portions (the spermatozoids). These vesicles become detached and collect at the orifice of the cavity. This hap- pens especially at the time when that zone of the sea-shore where the wrack grows is left dry, and the Fucus plants are lying flat upon the stones, and look like brown and faded leaves. At the recurrence of high-tide, when the wracks are again submerged, the cells full of spermatozoids burst, and the tiny spermatozoids formed from their protoplasmic contents swarm out into the surrounding water. Each spermatozoid has a sharp and a blunt end, exhibits a so-called eye-spot, and is furnished with two long cilia by means of which it swims about in the water (see fig. 2033). Analogy to similar processes which take place in Mosses makes it seem probable that the ooplasts above described as lying near the orifices of cavities in the thallus secrete some compounds or other— presumably organic acids— which attract the spermatozoids swarming in Fig. 203.— FitcMS vesiculosus. I Longitudinal section through a portion of a thallus including a cavity full of antheridia. "- Antheridia extracted from a cavity of the kind. » Spermatozoids escaping from the antheridia. " spherical ooplast covered with spermatozoids. 1x50; «xl60; 3,4x350. (After Thuret.) FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 53 the water. The actual fact is that spermatozoids which come into the vicinity of the spherical ooplasts adhere to them in such large numbers that a sphere is some- times entirely coated with spermatozoids (see fig. 203 *). It has also been observed that the spherical ooplasts are set rolling by the adherent spermatozoids, and are thus removed from the places where they pre- viously lay stranded. The fertilizing effect exercised by the spermatozoids, one of Fig. 204. — Fertilization and Fruit-formation in Mucorini, Siphunacecv, and Fiufidfcv. «-* Conjugation and fruit-formation in Sporodinia grandis. *, « Vaucheria sessilis. ' Fruit-rudiment witli trichogyne of Dudresiiaya coccinea. » Antlieridia of tlie same plant with sperniatozoids in the act of abiunction. » Fruit of the same. i-^XlSO; 6,6x250; ?, 8x400; 9x250. (V-9 after Bornet.) which, as it appears, coalesces with the ooplasts, consists doubtless in a rearrange- ment of molecules, and the first outwardly visible result of this rearrangement is the envelopment of the ooplast in a tough cell-membrane. The body must now be considered to be a fruit — a unicellular fruit, which remains unaltered in a state of rest for some time, but at length bestiru itself, and stretching out attaches itself firmly to the ground by means of root-like outgrowths. It then divides and gra- dually develops into a fresh Fucus plant. In the two cases just described, the ooplasts are not fertilized till after they have 54 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS escaped from the cells of the mother-plants into the surrounding water, and at the time of fertilization they are destitute of any special coverings of their own. In the plants to be dealt with next, on the other hand, the ooplasts at the time of fertilization are still in connection with the mother-plant. The cell-membrane, which maintains this union, persists as an envelope to the protoplasm which is to undergo fertilization. There are two ways in which a fertilizing protoplast may exercise its influence upon a protoplasmic body thus inclosed in a cell-membrane. Either a piece of the envelope is broken through and a free passage made for the spermatoplasm to the ooplasm, or else, if a true fertilization takes place, it must be by osmosis through the envelope. The solution and removal of part of the cell-membrane enveloping the ooplast, and the opening up of a passage in which the spermatoplast can unite with the ooplast, is observed to occur in the Mould-fungi known as Mucorini, and also in the innumerable little green and brown water-plants which, on account of their characteristic mode of fertilization, have received the name of Conjugatce. In these plants the coalescence of the two kinds of protoplasts is always preceded by a process of "conjugation", that is to say, the envelopes surrounding those protoplasts come in contact and grow together, and a special cavity is thereby created in which the fusion of the protoplasts can take place. This method of fertilization is shown in the clearest manner in fig. 204 12, 3, 4^ ^j^g instance being that of Sporodinia grandis, a Fungus belonging to the Mucorini. Two more or less parallel tubular hyphse put forth lateral protuberances (fig. 204 ^) which stretch out towards one another until their free ends come into contact and cohere. As soon as this union is effected, a transverse wall is formed on either side of the plane of contact, and it is now possible to distinguish in the limb connecting the two hyphse a median pair of cells supported by the two basal portions of the outgrowths (see fig. 204 2). The con- necting limb is usually likened to a yoke (^vydv). The wall arising from the junction of the outgrowths, and now separating the two cells in the middle of the yoke, dissolves, thus producing a single cell-cavity (instead of the two), which is called a " zygogonium ". The two protoplasts inhabiting the pair of cells were hitherto separated, one being derived from the hypha to the right, and the other from the hypha to the left; they are two different individuals, but, upon the dissolution of the wall between them, they coalesce within the zygogonium. This coalescence is to be looked upon as the act of fertilization. The membrane of the median cell, which surrounds the blended mass of protoplasm, thickens, and, in the selected instance of Sporodinia grandis, becomes warted, whilst in Mucor Mucedo (fig. 193^) it becomes rough and wrinkled, and in other Mucorini even spinose. It also acquires a decided dark coloration. Lastly, the dark median cell detaches itself from the basal portions of the original outgrowths, which have held it up to that time, and thus becomes free and independent (see fig. 204 *). It then drops just as a cherry does from the twig of a tree, and, like the cherry, it must be designated as a fruit, although it consists of a single cell onl3^ Fruits of this kind have received the name of " zygotes ". FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 55 It is no more possible to say which of the two protoplasts uniting in the zygo- gonium of Sporodinia grandis is fertilized and which acts as fertilizer, than it is to predicate of the pairing protoplasts of Ulothrix, that the one is the ooplast, and the other the spermatoplast. Theoretically we must assume there is a difference, and it probably consists in peculiarities of molecular constitution, but no perceptible difference can be detected in size, configuration, or colour, nor is there any apparent distinction in respect of origin. In the Besmidiacece also, of which two examples (Closterium and Peniivm) are given in vol. i. fig. 25a, i and k, and in the Diatomacece, whose species are reckoned by hundreds, no perceptible external difference exists between the protoplasts which unite for the purpose of fertilization. Only in the Zygnccraacece is it pos- sible to look upon a particular one of the combining protoplasts as an ooplast, and the other as a spermatoplast, and the distinction is in this case founded on their relative positions. An instance of the mode of fertilization prevailing in these plants is well shown in fig. 25a, I, in the first volume, the case chosen for illustration being that of Spirogyra arcta, which consists of green filaments of a slimy con- sistence, and occurs very commonly in our ponds. The cells are arranged in linear series, and from some of them are formed lateral outgrowths like those produced by the tubular cells of Sporodinia grandis. As in Sporodinia, the out- growths from opposite cells come into contact, coalesce, and form a kind of yoke. Usually a number of the opposite cells of two filaments floating close together in the water establish connecting links of the kind, which resemble the rungs of a ladder (see vol. i. fig. 25a, I, to the right). The wall formed by the coalescence of the two apices of the outgrowths is removed by solution, and a channel con- necting the opposite cell-chambers of the Spirogyra-^\sLm&nts is thus opened up. In the meantime the protoplasm in each of these cells undergoes a change. Hitherto it has been occupied by a chlorophyll-body in the form of a spiral band, but now it assumes the form of a dark-green spheroidal mass, which is destined to unite with the one opposite to it. In Spirogyra this coalescence does not take place in the middle of the connecting canal as in Mucor and Sporodinia, but the green ball of protoplasm from one cell glides through the transverse passage into the opposite cell-chamber, and there coalesces with the second protoplasmic mass which has remained at rest and not changed its position. It is permissible to call the resting protoplast an ooplast, and the one which moves across to it a spermato- plast; but it must again be expressly stated that in Spirogyra no difference in size, shape, or colour can be detected between the two uniting protoplasts. It is worth j- of note that the zygote produced by the coalescence, and now assuming an ellip- soidal shape, is not equivalent in bulk to the two protoplasts, as one might expect but that its volume is obviously smaller. We may infer from this that at the moment of coalescence a fundamental change in the molecular structure of the entire mass takes place. The characteristic property of fertilization in the Con- jugatce — of which Sporodinia grandis and Spirogyra arcta have here been chosen as examples — consists in the union of two separate individuals by means of the- 56 FERTILIZATION AND FRUIT- FORMATION IN CRYPTOGAMS. formation of a yoke between opposite cells which put forth lateral outgrowths towards one another for the purpose; this is the reason why this kind of fertilization is called conjugation, and the plants concerned are named Conjugatce. Similar to conjugation, but differing from it in several essential particulars, is the mode of fertilization by means of a protruding outgrowth from the antheridium, which pierces through the wall of the oogonium. This method is observed to occur in particular in the destructive parasites comprised under the name of Peronosporeae. The species named Peronospora viticola, which is repre- sented in fig. 205, has attained a melancholy notoriety as a parasite on the Vine, and to the same group belong Peronospora infestans, which causes the Potato- disease, Cystopus candidus, known as a deadly parasite on Cruciferous plants, the various species of Pythium, &c. Tubular hyphse develop directly from the spores of these Peronosporese, which attack the fresh foliage, green shoots, or young fruits of the particular flowering plants that they select to serve as hosts. The hyphse bore into the green tissue, piercing through the cell-walls and growing in the intercellular spaces, where they ramify extensively. Segmentation of the hyphse by the introduction of partition-walls is comparatively rare, but very frequently little suckers, called "haustoria", are sunk into the interior of the living cells of the host (see vol. i. p. 165, fig. 32 ^). These hyphae infesting the green tissues of the host-plant swell up at their blind extremities into globular heads, and a septum is introduced in each case to partition ofi" the terminal sphere from the rest of the tube, which preserves its cylindrical form. The splierical cell is an oogonium, and the protoplasm forming its contents is the ooplasm. The latter differentiates itself into two portions, namely, a central darker ball and a clearer transparent enveloping mass. The antheridia containing the sper- matoplasm develop in the form of lateral clavate outgrowths from another tube, or more rarely from the same tube. These protuberances grow towards the oogonium and apply themselves to its surface. As soon as the antheridium touches the oogonium it sends out from the point of contact a conical or cylindrical hollow process which pierces the wall of the oogonium and penetrates to the dark ball in the middle of the ooplasm (see fig. 205 ^). Meanwhile the protoplasm in the antheridium has differentiated itself into a parietal lining on the one hand and the true spermatoplasm on the other. The antheridial process, which has received the name of " fertilizing-tube ", opens at the extremity buried in the interior of the oogonium; within an hour or two the spermatoplasm has flowed through this channel to the ooplasm and become so completely merged with it that it is no longer possible to recognize any boundary between the two. A short time afterwards the fertilized ooplasm incloses itself in a thick cell- membrane composed of several layers. The outermost laj'^er is usually rough and warty, and in some species is even beset with spikes. The fruit thus formed is unicellular and remains so. It frees itself from the now decaying oogonium — thus effecting its separation from the mother-plant — and then enters upon a long period of rest. The new generation developed from the fruit begins as a tube FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 57 which subsequently, in some cases, puts out sac-like processes and branches and fashions itself into the likeness of the mother-plant without passing through any intermediate stage; or in others, the tube, which represents the embryo, produces first of all from its protoplasm a number of swarmspores. These roam about for a period and then seek out a convenient spot where they come to rest and develop into new individual plants. The additional production by Perono- sporeae of spores on dendritically-branched hyphae growing out through the Fig. 205.— Fertilization, fruit-formation, and spore-formation in the Peronosporese. ' A bunch of grapes attaclced by the Vine-Mildew. 2 Spores on branched stalks projecting through a stoma of a Vine-leaf. 3 Fertilization in Peronospora vilicola. * A single spore. « A single spore the contents of which are dividing into swarm- spores. 6 A single swarraspore. ' natural size; 2x80; *-6x350; 6x380. (*-« after De Bary.) stomata of the green host-plants is shown in fig. 205 ^ but an opportunity will occur later on of discussing the details of that process. The Siphonaceae exhibit a different mode of fertilization from those processes which involve the preliminary construction of a fertilization-tube and a conjugation- canal respectively. All the Siphonaceae live in water or on damp, periodically submerged earth; they contain chlorophyll and are neither parasites nor sapro- phytes. We may take as a type of this group of plants, which includes forms of great diversity, a species of the genus Vaucheria (see vol. i. figure 25a, a, and text p. 23) and use it also to illustrate the processes about to be considered. 58 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. If a green filament of Vaucheria is examined under the microscope it is found to consist of a single tube without septa, but with numerous saccate branches. The sac-like outgrowths serve a variety of purposes; those at the base fasten the tube to the substratum, those at the free extremity develop swarmspores, whilst those springing laterally from the filament have the functions of fertilization and fruit-formation. The lateral outgrowths are of two kinds (see figs. 204 ^ and 204^). One form is short, thick, and oval, and usually projects obliquely; the other is a slender cylinder curved like a chamois horn or wound round in a spiral, and sometimes it is subdivided into several little horns. The protoplasm in these sacs severs itself from the protoplasm of the main tube and a partition of cellulose is inserted in the plane of disjunction in each case. We have thus corresponding to each protuberant sac a cell-cavity or receptacle which incloses the protoplasm destined to take part in the formation of fruit. The obliquely- oval receptacles contain ooplasm and are oogonia, the curved, cylindrical receptacles inclose spermatoplasm and are antheridia. Their development is accomplished rather rapidly. It usually commences in the evening, and by the following morning the oogonia and antheridia are already completed. During the course of the fore- noon an aperture appears at the apex of the oogonium, whilst simultaneously the ooplasm within it contracts into a sphere. The spermatoplasm in the antheridia has meanwhile broken up into a large number of oblong spermatozoids, with a cilium at each end. After this has happened the free extremity of the antheridium bursts open, and the minute spermatozoids are expelled in a swarm into the sur- rounding water. Some of them reach a neighbouring oogonium, pass through the opened summit into the interior of the receptacle, and there coalesce with the ooplasm which has contracted into a green sphere. In connection with this phenomenon there is the following very striking circumstance to be noted. Where, as is usually the case, an oogonium and an antheridium are developed in close proximity to one another on the same tube, they seldom open simul- taneously, and this circumstance most effectively prevents the fertilization of the ooplast by spermatoplasm of the adjacent antheridium; but on the other hand it usually happens that the spermatoplasm from the antheridium of one tube reaches the oogonium of another tube, and in this manner a crossing of the two takes place (figs. 204 ^ and 204 ^). As soon as an ooplast is fertilized it surrounds itself with a tough cell- membrane; the green colour of the protoplasm changes to a dirty red or brown, and the fruit is to be seen imbedded in the oogonium in the shape of a reddish- brown, unicellular sphere. The oogonium dissolves or else breaks off* with tlie fruit inclosed in it. In either case the product of fertilization is removed from the tube whereon it developed and sinks to the bottom, where it undergoes a comparatively long period of rest often lasting through an entire winter. When the unicellular fruit germinates, the outer layer of the cell-membrane splits, and out of the rent emerges a tube of like form to that which produced the fruit. In every case of cryptogamic fertilization hitherto discussed a union of the FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 5& spermatoplasm with the ooplasm occurs. The protoplasts set aside for the purpose of coalescence forsake the cell-interiors when they have attained to maturity, or at least one of the sexual cells liberates its protoplasm so that it reaches the other unfettered and is enabled to effect a union of their two masses. For this result it is necessary for a part of the cell-membrane enveloping the protoplasm in question to be previously removed, for otherwise it would not be possible to effect the kind of union to which the phrase coalescence of protoplasm is properly applicable. On the other hand, many cases exist in which there is no obvious perforation of the wall, although the changes usually fol- lowing true fertilization take place. Under these circumstances it is difficult to resist the view that if fertilization (i.e. a fusion of protoplasts) really happens (as to which difference of opinion still prevails) it is accomplished by means of osmosis. With this qualification we may say that fertilization by means of osmosis is observed in its simplest form in the Erysiphese, popularly known as Mildews, in the Moulds allied to Aspergillus and Penicilliivm, a description of which in relation to their methods of spore-formation is given on pp. 21, 22, and in several Discomycetes, including the curious Fungus named Ascobolus, which will be dealt with more thoroughly when we come to the subject of the mechanisms for dispersing spores. The Mildew occurring on the surfaces of green foliage-leaves appears under the microscope as a peculiar kind of mycelium. The hyphae, which are filiform, colourless, and densely interwoven, do not penetrate into the intercellular spaces of the tissue of the host-plant, but satisfy themselves with sinking little suckers into the superficial cells of the leaves and stem (see vol. i. p. 165, fig. 32 -). Here and there these hyphal tubes rise erect from the substratum and abstrict monili- form rows of spores; others put forth short, lateral outgrowths which become partitioned off by the insertion of a transverse wall in each, so that the protoplasm in the outgrowth is shut off from the rest of the protoplasm in the tube. Some of these latter structures are oval or club-shaped, and they contain ooplasm and are to be considered as oogonia; the others are cylindrical and sometimes bent into the form of hooks, and they contain the spermatoplasm and constitute antheridia. In a few species the upper, somewhat swollen end of the outgrowth filled with spermatoplasm — i.e. the antheridium — bends over the top of the oogonium and attaches itself closely thereto, without, however, sending any special fertilization-tube into the interior of the oogonium; in other Fungi of the Mildew family both cells — the oogonium as well as the antheridium— are spiral and are coiled round one another, and at the same time pressed tightly together. On the assumption that a true fertilization now occurs, this must, as already indicated, be by a diffusion of the spermatoplasm through the cell-mem- branes to the ooplasm, causing a change in its ultimate structure which corresponds to fertilization. The ooplasm is thereupon converted into an embryo. The cell inclosing the embryo neither dissolves nor severs Itself from the parent-hypha, but divides and becomes differentiated into an upper swollen cell and a lower short. 60 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. stalk-like cell, and below the stalk fresh tubular outgrowths develop from the hyphal filament in question which become septate and ultimately form a voluminous multicellular envelope round the embryo. The now mature fruit preserves its connection with the parent-hypha, and is to be seen seated upon it in the form of a minute sphere. When a large number of fruits are developed simultaneously on the hyphal reticulum — as is the case in Sphcarotheca Castagnei, which is parasitic on the leaves of Hops — the grey mildew spread over the foliage has the appearance of being studded with the tiny globular heads. From the embryo a new generation is produced. In the species of the genus Podosphcera it develops, within the cellular mass just referred to as investing the fruit, into a single tube (ascius). The protoplasm within the ascus breaks up and fashions itself into true spores, which abandon the tube and are distributed by the wind. In Erysiphe, on the other hand, the embryo becomes septate, and takes the form of a simple or branched chain of cells, and it is not till after this stage that tubes are produced whose protoplasm is transformed into a group of spores. The tubes in question are long, erect, and club-shaped, and they spring from the cells of the aforesaid chain. The manner of fertilization and fruit-formation in Penicilliu7)i, and generally in all the forms of Mould which are comprised under the name Aspergilleae, is the same as that described in the case of Mildews (Erysipheae). In them also the extremities of tubular hyphse which contain the ooplasm and spermatoplasm, respectively, come into close contact. They are either spirally twisted and wound round one another, or else the extremity corresponding to an antheridium is hooked and grasps the other, as is shown in fig. 193^ (p. 18). Fertilization takes place by osmosis. The embryo produced by the spiral oogonium is septate and multi- <;ellular, and develops club-shaped or egg-shaped outgrowths, whose protoplasm breaks up into round or ellipsoidal balls (fig. 193^). This structure becomes surrounded by a continuous multicellular tissue, which owes its origin to the upgrowth of a number of hyphae from the cells at the base of the oogonium. These hyphse elongate rapidly, ramify, become intertwined, and develop septa until they constitute a spherical envelope round the embryo. The fruit thus constructed is in Penicillium about half a millimetre in diameter. The Floridese, or Red Seaweeds, are likewise fertilized by means of osmosis. The details of the process are, however, intrinsically different from those observed in Mildews and in the Moulds classed as Aspergillese. The organs developed for the purpose of fertilization have also quite a different form in Florideas. Their most striking feature is the so-called " trichogyne ", a long filamentous cell which projects far above the fruit-rudiment. From this structure the characteristic mode of fertilization in Florideae is called fertilization by aid of a trichogyne. In some Floridese the cell containing the ooplasm leads directly into the trichogyne; in others the fruit-rudiment which incloses the ooplasm is septate, that is to say, it consists of a row of broad cells which together form a short branch of the ramifying thallus, and adnate to one side of this row of cells is the long, delicate, FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 61 filamentous cell called the trichogyne (fig. 204 ''). Rudimentary fruits of this kind are produced on one individual, whilst antheridia are formed upon another. It is much less common for fruit-rudiments and antheridia to be developed on the same individual, and in the few species which do exhibit this combination, self-fertilization is rendered practically impossible by a retardation of the development either of the fruit-rudiments or of the antheridia. The antheridium always takes the form of a limited portion of the thallus, from which separate round cells filled with spermato- plasm are thrown off. Fig. 204 ^ represents an antheridial branch of Dudreanaya coccinea. A slender branch of the thallus terminates in a group of cells arranged dichotomously, and the outermost of these cells, which become rounded off and detached, contain the spermatoplasm, and must be looked upon as spermatozoids. Unlike the spermatozoids of Vaucheria and those of the Characese (Stoneworts), MuscineaB, and Ferns, to be described presently, these have no cilia, and do not move by virtue of any power of their own in the surrounding water, but are carried about by currents which are always more or less prevalent at the places where the Florideae live. Through the action of these currents in the sea, the spermatozoids (or spermatia as they are called) reach one of the trichogynes and adhere to it, as is shown in fig. 204'^. The question as to how far attractive forces emanating from the ooplasm come into play in order to eflfect this conjunc- tion must remain undecided. It is not impossible that substances may be secreted by the ooplasm and be given off into the environing water, and that they may take part in the phenomenon. Nothing more definite is known beyond the evident fact that the spermatozoids adhere much more commonly to trichogynes than to other objects floating in the neighbourhood. Part of the protoplasm of the adherent cells passes apparently by osmosis into the protoplasm of the trichogyne. The change ensuing upon this process is transmitted to the protoplasm occupying the ventral enlargement at the base of the trichogyne, and in many cases even further to the protoplasm of adjacent cells. Although this propagation of the change in the molecular structure of the protoplasm cannot be directly observed, it may be assumed on various grounds, and we may fairly suppose that the action of tlie absorbed constituents of the spermatoplasm upon the ooplasm is comparable to that of certain enzymes, which have a convulsive effect upon any protoplasm in their vicinity, and even when they are separated from it by cell-walls cause a displace- ment and rearrangement of the molecules (see vol. i. p. 464), That the change affecting the protoplasm at a particular spot in the fruit-rudiment is capable of being propagated so long as any protoplasm susceptible of the same change is present, is proved by the fact that it is not the trichogyne itself but the ventral enlargement at its base and the cells adjacent to this enlarged portion which undergo subsequent development. They increase in bulk, whereas the trichogyne shrivels and dies. The cells which contain the protoplasm fertilized through the intervention of the trichogyne must, in my opinion, be looked upon as the fruit. Any subsequent structure arising from them is no longer fruit but a new generation. In Florideae, as in so many other cases, this new generation preserves its connection with the ■62 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. mother-plaut, but differs conspicuously in form from the generation from which it sprang. This stage having already been dealt with on p. 22, it is here only necessary to mention briefly that the cells of the fruit begin to shoot out after a period of rest of variable duration and abstrict a mass of spores, and that in addition, in most Floridese, linear series of cells grow from the cells at the base of che fruit, and form a capsular envelope around the young spores. The Cryptogams that we have still to deal with, viz.: the Characeoe, Muscinese, und Vascular Cryptogams, differ from those already described in that the oogonium is wrapped up in a special sheath before fertilization takes place, and that the entrance-passage provided for the spermatoplasm is consequently modified in a characteristic manner. This sheath, to which we shall apply the term amphi- gonium (also known as archegonium), is in the main of the same construction in all the plants exhibiting it; but as regards the penetration of the spermatozoids into the amphigonium, and the behaviour of the fruit produced from the fruit- rudiment, there are very considerable differences amongst the groups in question. To follow out these diversities in minute detail is not possible within the narrow limits of this book, and I must content myself in the following pages with giving a brief sketch of the most important phenomena. To begin with the Stoneworts (Charace^), we find that in them the fruit- rudiment is ellipsoidal in shape, and is borne on a very short unicellular stalk. This stalk is seated upon the so-called "nodal cell", a short discoid cell which forms the pedestal of the large ellipsoidal oogonium, and also gives rise to five tubular cells arranged in a whorl, and twisted spirally round the oogonium, thus enveloping it in a sheath of great beauty (see fig. 206 ^). From the ends of these investing tubes, which project above the oogonium, small cells are separated off, and together constitute a little roof for crown to the amphigonium. Beneath the crown the enveloping tubes are drawn together so as to form a neck which incloses a narrow cavity, and this is the part where at the time of fertilization fissures are formed between the otherwise connate tubes of the envelope, thus enabling the spermatozoids to penetrate into the interior of the amphigonium, and to reach the ellipsoidal oogonium there matured. The mode of genesis of these spermatozoids is extremely remarkable. They are produced in certain red, globular structures, which are slightly smaller than' the fruit-rudiments and have a like origin — that is to say, they take their rise amongst the whorls of lateral offshoots. In some species they are formed on the same individuals as the fruit-rudiments (cf. figs. 206 ^ and 206 ^), in others the two kinds of structure develop on different individuals, and are thus separated from one another; hence we distinguish Characeae into monoecious and dioecious species. Each red sphere is composed of eight plates, outwardly slightly convex. Each of these is in the shape of a spherical triangle with indented edges and folds running radially from the centre to the notched margins (see fig. 206 "*). The plates are joined together into a sphere, the notches of the margins fitting into one another so as to form a regular dovetailed suture. From the centre of the gently FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 63 concave inner face of each plate a cylindrical or conical cell projects, carrying upon its summit another, capitate, cell. Each of these head-cells is surmounted by long strings of cells, of which the lowest segments are spherical or cylindrical, whilst the rest are short discs (see fig. 206 ^). The whole structure may be likened to a whip with many thongs, and the stalk-cell projecting from the plate has hence been called the " manubrium " or handle. So long as the eight plates of the sphere are Fig. 206.— Fruit-formation in Stoneworts (Characese). Chara fragilis. 2 Piece of the same with amphigonia and antheridia on the branches, s A single branch with amphigonia and antheridia. ■* An antheridium. » A plate of the antheridium with manubrium and cells grouped in the form of tliongs and containing spermatozoids. « Several cells from one of the whip-like filaments ; the cells in tlie middle contain each a sperniatozoid, the spermatozoid is escaping from the uppermost cell, the lowest cell is already vacated. ' A single sperma- tozoid. 8 Amphigonium inclosing the oogonium. 1 natural size ; 2 x 10 ; » x 15 ; * x 35 ; ' x 100 ; « x 300 ; ? x 600 ; » x 50 closed, these manubria project towards the centre of the hollow sphere, and the rows of cells proceeding from the manubria are conglomerated into a ball. But as soon as the plates separate and the sphere falls to pieces, the ball is untwisted and its parts assume the appearance shown in fig. 206 ^ By this time a spiral sper- matozoid has developed from the protoplasm in each of the discoid segments of the filaments, and may be seen lying within its cell (see fig. 206^). But almost immediately afterwards these cells open, and the spermatozoids, which are provided at one end with a pair of long cilia, escape and whirl about in the surrounding 64 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. water (soe fig. 206^). The spermatozoids then pass through the fissures already described as existing beneath the crown of an amphigonium, and so reach the interior of the latter. Here, in the middle of the cavity is the oogonium (i.e. the great cell containing the ooplasm), and over it there is a slimy gelatinous mass, which occupies more particularly the neck of the amphigonium. The cell-membrane of the oogonium is attenuated and almost liquefied, and these soft and swollen masses of mucilage do not interfere in any way with the progressive motion of tlie spermatozoids. The latter reach the ooplasm, and, so far as we can see, a coalescence of the two kinds of protoplasm takes place. The changes set up in the fruit-rudiment by fertilization first manifest them- selves externally in an alteration in colour. The chlorophyll-bodies, hitherto green, assume a reddish-yellow tint; the spiral cells of the amphigonium become thickened and nearly black, and the amphigonium constitutes a hard shell which acts as an outer envelope inclosing the inner envelope of the fertilized ooplasm, now converted into an embryo. The entire structure next detaches itself from the stalk-cell, sinks under water, and remains for a considerable time — usually through the whole winter — lying unchanged at the bottom of the pond. The embryo does not germinate till the following spring, when it begins by developing a linear series of cells, the so-called pro-embryo, and from one of the cells of this pro-embryo is pro- duced a Stonewort plant with branches in whorls as before (see fig. 206 ^). The fruit-rudiment in Muscinese (Mosses and Liverworts) exhibits in many respects a resemblance to that of a Stonewort, although its origin is quite different. It takes its rise from a superficial cell of the Moss-plant, and the cell belongs, according to the species, either to the foliaceous or to the cauline portion of the thallus. This cell projects in the form of a papilla above the adjoining cells, and becomes partitioned by a transverse wall into an under and an upper cell, the former of which serves as a pedestal to the body of tissue developed from the upper cell. The cellular body referred to is differentiated, by repeated insertion of longitudinal and transverse walls, into a central row of cells and an envelope. Amongst the central cells one situated somewhat low down in the series is conspicuous for its size; it contains the ooplasm, and must be looked upon as an oogonium. The central cells, which are placed in succession above it, are called the canal-cells of the neck. The name is derived from the fact that they occupy the constricted portion or neck of the envelope. The cellular envelope, which incloses the central row of cells and constitutes the amphigonium, is shaped like a flask (see fig. 191 ^*'); the lower, enlarged, ventral portion conceals the oogonium, the upper constricted portion is filled up by the neck-cells, and the whole structure, which received from the earlier botanists the name of " archegonium ", is closed at the top by a lid composed of several cells. When the time for fertilization arrives the canal-cells of the neck swell up and are converted into mucilage. The lid-cells open and part of the mucilage is forced out; what remains offers no impediment to the admission of the spermatozoids to the ooplasm in the centre of the fruit- rudiment. FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 65 The antheridia arise in the same manner as the fruit-rudiments. A superficial cell of the thallus is enlarged into a papilla, and, by the repeated partition in all directions of its first segments, a body of tissue is produced, which includes a delicate stalk and a thickened upper portion, either clavate or spherical in shape. The latter part consists of a multicellular sac-like envelope and a parenchymatous filling-tissue inclosed within the envelope. In each cell of the internal tissue the protoplasm fashions itself into a spirally-bent spermatozoid, and shortly afterwards the entire filling-in tissue is resolved into its separate cells. The antheridium now opens at the top, and the loose cells with the mucilage in which they are embedded are ejected into the surrounding aqueous medium composed of rain or dew-drops. The spermatozoids then escape from their delicate cell-membranes, and swim about the water by the help of the two long cilia wherewith each is furnished (see vol. i. p. 29, figs. 7® and 7 ^°). Passing down the open neck of the amphigonium, now filled with mucilage only, they succeed in reaching the oogonium in the enlarged base of the fruit-rudiment and apply themselves closely to its surface; a constituent portion of the spermatoplasm is absorbed into the ooplasm with the result that the latter becomes fertilized. Usually several antheridia are situated close together. In Mosses they are mingled with paraphyses, structures resembling hairs, the significance of which has not yet been explained. In many species one individual develops only anther- idia, another only amphigonia; but in other species antheridia and amphigonia are developed side by side on the same Moss-plant. Where the latter is the case either the oogonium exhibits an earlier development than the antheridium, or the reverse is the case. Either the passage leading to the oogonium through the neck of the amphigonium is opened whilst the adjacent antheridia are still closed, or else the spermatozoids are set free from the antheridia at a time when access to the oogonium is still barred by the lid-cells of the amphigonium. As in so man}^ cases of a similar kind this contrivance prevents a union between the ooplasm and the spermatoplasm produced by the same individual, and favours cross-fertilization Ijetween different individuals. In some Liverworts the antheridia and amphigonia are surrounded by annular walls, and these organs then appear to be sunk in depressions of the thallus. In other Liverworts separate lobes or branchlets of the thallus are transformed into stalked shields or discs, and the antheridia and amphigonia are formed in special niches and compartments on the surface of the shields. Those Muscinese which have their thalli differentiated each into a cauline axis and cellular laminaj resembling leaflets, develop antheridia in the axils of the leaflets, or else in pitcher- shaped cavities at the tops of the stems. In Mosses the principal or secondary axes terminate in groups of antheridia or amphigonia, and specialized leaflets act as envelopes or roofs and constitute the " perichsetium ". Sometimes these leaflets have the appearance of floral leaves, as, for instance, in the Hair-Mosses (Poly- trichum), several species of which may be included amongst the commonest of our Mosses. The antheridia and amphigonia are here distributed on diflferent individuals. Vol. II. 65 66 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS, The investing leaflets at the summit of those stems which terminate in antheridia are ci'owded close together; they are short, broad, and of a brownish-red colour, and look like small floral-leaves seated upon a disc-shaped receptacle. Polytrichum is a typical instance of the Mosses which exhibit a conspicuous contrast between the investing scales of antheridia and those of amphigonia. The perichaetium in individuals which produce only amphigonia possesses an altogether different form and arrangement of parts from the corresponding structure in antheridia-bearing individuals. They form a flask-shaped enclosure rather than a coloured flower-like structure, and later, from their midst arise stalked capsules in which the spores are boi'ne. These capsules are the direct product of the amphigonia after they have been fertilized. The colouring of the antheridial involucres gave rise to the sugges- tion that insects might be concerned in fertilization. But there does not appear to be any evidence in support of such a view. As before said, there is a close resemblance between Muscineae and Characeoe as regards the position of the ooplasm to be fertilized in the middle of the amphi- gonium, the genesis and form of the spermatozoids, and, lastly, the process of fer- tilization. But from the moment of fertilization the course of development is altogether different. The fruits of Characeae become detached from the mother- plant, whereas those of Muscineae remain in connection with it, and this connection is not merely mechanical but organic. The generation developed from the Moss- fruit continues to derive the nutritive substances requisite for its growth and completion from the mother-plant, and without the support of the latter it would inevitably perish. The word support may here be used in a wide sense; for the mother-p^nt is actually the bearer or stay of the new generation, which is produced from the ooplasm converted by fertilization into an embryo, and it may be com- pared to a tree with Mistletoe growing upon its boughs. In Characeae the separate stages of development are always quite distinct; the stage of maturity in par- ticular being characterized by the falling away of the fruit from the mother-plant. This is not the case in Muscineae. Since no separation in space takes place, it is also difficult to establish time-limits and to say when the fruit has attained maturity, and the difficulty is increased by the fact that no sufficient indications are afforded by alterations of shape or colour. It is best to look upon the forma- tion of fruit as being complete as soon as fertilization has taken place; from this moment the ooplasm must be considered to be an embryo, and its envelopes to be fruit-coats. Evidence in favour of this conception of the phenomenon is afforded by the circumstance that after the union of ooplasm and spermatoplasm development is arrested, and a period of repose ensues, whereas both before and afterwards the outward manifestations of change follow one another in rapid succession. A description of the subsequent changes has been already given (see pp. 15, 16), and we need only repeat here that the generation which springs from the Moss-fruit develops spores and, after having scattered them abroad, dies away. The strongest likeness exists between the fruit-rudiments and antheridia of Muscineae and those exhibited by Ferns, Horse-tails, Rhizocarps, and Club-Mosses, FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. 67 all of which are classed together under the name of Vascular Cryptogams, on account of the presence of vascular bundles in their stem -structures and phylloclades. The first generation of these Vascular Cryptogams, whereon are developed the antheridia and fruit-rudiments, also resembles in an unmistakable manner the first generation in certain Liverworts. In Ferns, which constitute the most extensive section of the Vascular Crypto- gams, and may be taken as their type, the first generation makes its appearance in the form of a flat, green, foliaceous structure, usually reniform or heart-shaped, lying in close contact with the nutrient soil (see fig. 189 ^^). Inasmuch as the tissue of this first generation nowhere contains vascular bundles, it must be looked upon as a thallus, and has received the name of prothallium. The Fern-prothallium bears the fruit-rudiments as well as the antheridia upon its under surface, which is in contact with the nutrient soil, and which adheres to it by means of a number of delicate hair-like suction-cells. Some Ferns develop the fruit-rudiments and antheridia on separate prothallia; others produce them both on the same prothallium. In the latter case the fruit-rudiments are situated near the sinus of the prothallium, and the antheridia on the part remote from the sinus. Each fruit-rudiment may be compared to a flask in shape, and arises from a superficial cell of the prothallium which is only slightly arched outwards. This cell is divided by the insertion of two partition-walls into three cells, each of which is again segmented in definite directions. From the uppermost cell is produced a tissue which forms the neck of the flask-shaped fruit-rudiment; the middle cell gives rise to three cells, of which the two upper, the canal-cells, occupy the neck, whilst the undermost one becomes the relatively large and subsequently rounded ooplast. The daughter-cells de- veloped from the lowest primary cell take the form of an investing wall round the ooplast, or, to return to the analogy of a flask, constitute the wall of the ventrally enlarged portion of the flask. The protoplasm of the ooplast is the ooplasm, and is now to be seen surrounded by a pluricellular tissue, which, as in the case of Characese and Muscinese, may be called an amphigonium. Only the neck of the amphigonium projects above the other adjacent tissues of the prothallium; the enlarged ventral portion is, as it were, sunk in the substance of the prothallium. The antheridia are also developed from cells upon the surface of the pro- thallium. These cells project in the form of papillae above the surrounding tissue and undergo division by the introduction of partition-walls. The outermost daughter-cell becomes enlarged and assumes a globular shape, and from the proto- plasm in its interior are formed spiral spermatozoids. Another mode of origin consists in the formation of a papilliform or hemispherical protuberance of tissue which shows unmistakably a diflferentiation into central cells destitute of chloro- phyll and enveloping cells containing chlorophyll. The former divide up and a filling-in tissue is formed, the small constituent cells of which contain spermato- plasm. After the development of a spermatozoid in each of these small cells, the whole of the filling-in tissue falls to pieces, that is to say, the individual cells separate from one another and remain for a short time disconnected but still in 68 FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAMS. contact. At length the top of the antheridium opens; the loose cells are discharged into the surrounding water derived from rain or dew, and from each of them is set free a spirally-coiled spermatozoid furnished as regards its anterior half with bristling cilia (see vol. i. p. 29, fig. 7 ^^). The spermatozoids manifestly direct their course to an amphigonium as they whirl about in the water. Meanwhile the neck canal-cells of the amphigonium have been partially converted into mucilage; some mucilage is discharged into the environing water, and it seems that concomitantly with this organic acids have been evolved in the region of the amphigonium, which exercise an attractive influence on the spermatozoids. What is known as a fact is that the spermatozoids accumulate in this mucilaginous mass and also penetrate through the slimy substance left behind in the canal of the amphigonial neck. Thus they reach the ooplasm which is hidden in the oogonium at the bottom of the fruit-rudiment. As it has repeatedly been observed that spermatozoids make their way into the ooplasm and there disappear, we may assume that the delicate envelope of the ooplast is pierced by the spermatozoid, and that thereupon a coalescence between the two kinds of protoplasm takes place (c/. also figs. 346 1' 2. 3, 4). The fertilized ooplasm now subdivides into several cells with partition-walls inserted between them, and thus is produced a multicellular embryo which remains embedded in the unaltered amphigonium. This structure, though scarcely differing at all from the fruit-rudiment, must be considered as a fruit. After a short period of rest the embryo germinates, and the new generation, which gradually makes its appearance as stem, roots, and fronds emerging from the embryo, continues for a short time to receive its food-stuffs through the mediation of the parental pro- thallium. At length, when the new generation has grown suflBciently strong, and is capable of taking up food-stuffs directly from the surrounding air and soil, and of transforming them into constructive materials, the assistance of the prothallium becomes superfluous. The prothallium then withers, and by the time the sporo- genous fronds have developed it has vanished, and no trace of it remains. The Horse-tails (Equisetacese) have, in the main, the same features as the Ferns just described as typical of the Vascular Cryptogams in all that relates to the forms of prothallium, antheridia, and fruit-rudiments. The prothallium produced from the spore is at first delicate and ribbon-shaped, but later becomes multifariously lobed, and in form recalls the thallus of certain Liverworts, or sometimes even resembles a little curled foliage-leaf. In most species antheridia and fruit- rudiments grow on different prothallia. Where this is not the case, fertilization of the ooplasm by spermatoplasm arising from the same individual is rendered impossible by means of a disparity between the organs concerned in respect of the time at which they mature. The prothallia which give rise to antheridia are always much smaller than those which produce the fruit-rudiments. The antheridia develop from superficial cells at the end or on the margin of the lobate prothallium, whilst the fruit-rudiments, on the other hand, are derived from super- ficial cells in the recesses between the lobes (see fig. 190^). The spermatozoids FERTILIZATION AND FRUIT-FORMATION IN CRYPTOGAJIS. 69 i have a spatulate enlargement at one extremity, and carry on the other, attenuated I end a regular mane of extremely fine cilia. Far more important are the characteristics which distinguish from Ferns the Rhizocarpese and Lycojjodiales, especially the genera Salvinia, Marsilia, and Selaginella, in all of which the development has been studied with great care. The antheridia-bearing prothallia are, in the last-mentioned genera, extremely ditFerent in point of size from those which bear fruit-rudiments. Both prothallia, it is true, have spores for their starting-points, but these spores themselves have ; different dimensions, and are distinguished as rnicrospores and Triacros'poves {i.e. i small spores and large spores). The microspores are the parts of the plant where antheridia are formed, and the macrospores those where fruit-rudiments are formed. In a microspore the protoplasm divides into several parts, and partition-walls are inserted between them, thus forming a tissue composed of a very few cells, the greater part of which remains concealed in the interior of the spore. Only one or two superficial cells of this tissue push out through rents made here and there in the coat of the spore, and these protruded cells constitute the antheridia. The apical cell of the antheridium becomes filled with a tissue, and in each cell of this tissue is formed a spirally-coiled spermatozoid. The opening of the antheridium and the escape of the spermatozoids then ensues in the same manner as in Ferns. The prothallium which originates from a macrospore and is the seat of formation of fruit-rudiments, although it is larger and composed of more cells than that just described, does not forsake the interior of the cavity of the macrospore to any greater extent, but only protrudes a little at one place where the tough outer coat of the macrospore is ruptured. Two kinds of tissue are in reality developed within the limits of each macrospore, viz.: the one above referred to as emerging between the torn edges of the outer spore-coat, and a tissue of reserve material deposited at the bottom of the macrospore. The latter is very rich in starch and oil, and serves as a storehouse of nutriment for the prothallium at least until it ia in a position to get food for itself out of the environment. The fruit-rudiments (amphigonia) appear on the protruding portion of the prothallium, and are entirely buried in its tissue. The development of the fruit-rudiment, the formation of canal-cells which subsequently turn into mucilage, the penetration of the spermato- zoids, and the act of fertilization, are in all essential respects the same as the corresponding processes in Ferns, and therefore a description of them in detail may here be dispensed with. The tissue produced from a macrospore in the Rhizocarpese and Selaginelleae has been compared to the ovule as it occurs in the Phanerogams which will be the subject of the next chapter, and certain actual analogies have been brought out which are exhibited by the ooplasm when converted into an embryo, the store- chamber for food-stuffs, and the protective envelope in each case. Having regard to the identity of object aimed at through the instrumentality of these structures in the most widely different sections of the Vegetable Kingdom, such analogies are really a matter of course, and if naturalists limit themselves to proving that organs 70 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. which have the same functions, however greatly they may differ in form, yet always possess certain similarity, and that this similarity increases in a conspicuous degree when the external conditions of life are the same, no objection can be made to the generalization. But if it is made the basis of far-reaching speculations and of hypotheses concerning the evolution of one group of plants from another, the descent of Phanerogams from Cryptogams, for example, I must enter an emphatic protest against any such proceeding. THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. Long experience has shown us that the propagation of plants is accomplished with much greater certainty by means of Brood-bodies than by Fertilization and production of Fruit. For a fruit to be formed, two portions of protoplasm which have arisen separately must be brought together. Such a union denotes that at least one of the two protoplasts in question is endowed with a capacity for translation, that the male cell is not obstructed on its way to the female, and that facilities are present to promote its union with that cell. But there's many a slip 'twixt the cup and the lip! Adverse winds, unfavourable currents, long- continued drought, uninterrupted rain, these and many another unexpected cir- cumstance may bar the way to fertilization. Often enough fertilization is hindered from causes such as these, and in consequence the young fruit-rudiment atrophies, the embryo is not formed, and the plant, in order to propagate, must rely on its brood-bodies. That fruits do not miscarry oftener than they actually do is due to the fact that the difficulties of the situation from external cause, are to some extent met by the position of the egg-cell and the form of the young fruit. In other words, the fashioning of the organs concerned in the production of fruit is adapted to the circumstances of the environment. Perhaps the obstacles are at a minimum in the case of plants in which fer- tilization is accomplished under water. The cells in question here require no especial protection. The surrounding water maintains them in the proper position, brings food to them, and protects them from drying up. In it they both live and move. Thus it is intelligible why so many plants which live under water, or which use water for the accomplishing of fertilization, are destitute of any but the simplest envelopes for their spermatoplasm and ooplasm. Com- plicated investments are valueless under such circumstances, possibly even dis- advantageous; in any case they are superfluous. Nor is it usual in plants to produce superfluous structures. As we know, aquatic plants do not possess woody stems and branches. And for this reason. Tissues of this kind are not required, since the surrounding water buoys them up in the proper position so that hard wood and bast are not needed. So also with the ooplasm and spermatoplasm. Cryptogamic plants which fruit under water do not possess complex ovaries like Phanerogams, as they are unnecessary. Just before the time of fertilization the THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 71 spermatoplasm is segmented up into many fragments; these escape from the antheridium and reach the simple fruit-rudiment by swimming. Since the sper- matozoids are attracted to the young fruits by certain excretions which the latter pass out into the water, the multifarious devices associated with aerial fertiliz- ation are unnecessary. Protective coats around the sexual organs, sheaths to limit evaporation, brightly-coloured or sweet-smelling floral-leaves to attract insects that they may transfer the pollen from flower to flower — all these are wanting in plants which are fertilized under water. Now it is just these accessory protecting structures which constitute what are called blossoms. Thus we can say that these water-plants have no blossoms. To avoid misconception it must be stated that although they have no blossoms they have flowers. For although, popularly, blossoms and flowers are used as synonymous terms, under flowers are compre- hended the organs which are concerned in fertilization, under blossom merely the leaves which inclose the essential organs and which guard and protect the young fruits and stamens. It is these latter which produce the sexual protoplasts. Their union is promoted by the leaves of the blossom. Sometimes they catch the pollen-grains as they are blown by the wind, or by the production of honey and scents attract insects which remove the pollen in their visits. In other cases, by projecting ridges and corners, they are instrumental in detaching the pollen from these same insects, and in a thousand ways protect and assist the difficult process of aerial fertilization. In the above lines we have been speaking not of aquatic plants generally, but of such as are fertilized under water. And these should be carefully distin- guished. Many aquatics, which pass their lives under water, send up their flowers to the surface so that their fertilization is aerial. On the other hand, strange though it may seem, the fertilization of most aerial Lichens, Mosses, and Ferns which grow on the sand of desolate moors, on the sunny rocks of mountain sides, or on the dry bark of old tree stems, is accomplished under water. Plants of this sort may be exposed to drought for many months, and the movement of sap within them may be suspended; but when they are moistened with rain or dew they are quickened and rejuvenated, and form their young fruits and antheridia. Things are so arranged that the liberation of the spermatozoids coincides with the moment at which these plants have access to sufficient moisture. Thus we see that it is literally true of these plants — whether growing on the bough of a tree or in a ravine on a mountain side — that their fertilization is accomplished under water. The only really important distinction between plants permanently submerged and such as are thus situated from time to time, is that in the latter the young sexual organs are protected against desiccation during the periods of exposure by means of sheathing structures and leaf-like scales, as is particularly well shown by the Mosses. Blossoms in the usual sense, however, are not found amongst Ferns and Mosses, and we can make the following three general statements: — (1) That Cryptogams are fertilized under water and most Phanerogams in the air; (2) that 72 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. Cryptogams lack blossoms, since these are not necessary for aquatic fertilization; (3) that almost all Phanerogams, on the other hand, possess blossoms, since they are required to protect and promote aerial fertilization. The very complicated structure of the parts immediately adjacent to the region where the sexual protoplasts are developed depends upon the fact that fertilization is aerial. The portions of protoplasm destined for fertilization can only be adequately elaborated if their enveloping membranes are thin and delicate, and suited for the osmotic transfer of materials. Such a membrane, however, is incapable of protecting the protoplasm from the drying influence of the air; it is absolutely essential that both the spermatoplasm and the ooplasm shall be protected during the critical period by a suitable envelope. Thus one finds in all Phanerogams — quite apart from the perianth — a protective mantle developed around the sexual cells. This mantle has its cell-walls suitably thickened; its outer layers afford the necessary resistance to desiccation, whilst deeper down an ample supply of water is maintained. These characters are well shown in that constituent of the ovary from which the seed will be ultimately produced. This portion is known as the ovule. Every ovule consists of a mass of tissue, the nucellus of the ovule in which the ooplasm or egg-cell is concealed, and an enveloping sheath, the integument, which may be either single or double. Such ovules are borne in the genus Gycas (figs. 208'' and 208 ^) without further covering than a fretwork of hairs which protects them against too great drying up. In other Cycads and in the majority of Gymno- sperms, of which the Cypress and Juniper, the Pine and the Fir, may be quoted as examples, the leaf-like scales of the young fruit are so arranged that the ovules produced on their surfaces are hidden from view and secure against outside danger. In the other Phanerogams (the Angiosperms) the ovules are concealed in a closed chamber — the pistil — the lower enlarged portion of which is known as the ovary. In the construction of this chamber the chief part is taken by the floral axis and by the floral-leaves known as carpels. So unequal, however, is the share taken by these parts in the structure of the ovary that in some cases it is formed almost entirely from the floral axis, and in others almost entirely from the carpels. In consequence the apex of the floral axis, which is known as the floral receptacle, shows an extraordinary variety of form. Thus in one series of plants the receptacle is not excavated, but solid, assuming the form of a knob, hemisphere, or cone (figs. 207^ and 207^); whilst in others it is concave and excavated (figs. 208^ and 208^). The forms met with in nature can be produced artificially by taking a conical mass of soft wax and flattening its summit, then gradually pressing it down into a saucer-like shape, and so on until one has produced a hollow bowl. So in nature we have at one extreme the solid cone, at the other the hollow vessel. Between these two extremes, between the conical and excavated receptacles, we have the flat or disc-like receptacle. It is hardly necessary to point out that in the growth and differ- THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 73 entiation of the living plant the excavated receptacle is not the result of any actual hollow ing-out process as in the lump of wax, but is due to unequal growth of the different parts of the receptacle — the peripheral parts growing up as a circular wall around the central parts, so that the form of a cup or urn is gradually assumed. When one speaks of the excavation of the receptacle one is speaking figuratively — there is no excavation in a literal sense. The configuration of the receptacle is further complicated by the fact that Fig. 207. — Structure of Phanerogamic Ovaries. * Dehisced fruit of Miltonia stellata. 2 Ovary of Miltonia cut across transversely, s Ovary of Alignonette {Reseda) cut across transversely. ■• The same ovary intact. * Longitudinal section of the ovary of the Jerusalem Artichoke (Uelianthus tuberosus). « Ovary of the Violet ( Viola odorata). ? The same, cut across. » Receptacle and carpels of Myostinis mi7ii- mits. 9 The same in longitudinal section, lo Young fruit of Potato (Solanum tuberosum), n The same cut transversely. All the figures considerably magnified. the centre of the receptacle does not always cease growing, but grows up as a cushion or peg; thus we have a receptacle having the form of a conical peg with a peripheral, urn-like wall around it. In describing the relations of the floral-leaves to the receptacle it will be simplest to commence with the conical receptacle. Here the floral-leaves are found arranged in whorls above one another or in a continuous spiral. At the top are the carpels, below these the stamens, and below these again the leaves 74 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. of the perianth. Of these various kinds of leaves there may be developed one, two, or even more whorls. When several whorled carpels are united together so as to inclose a single chamber, the tip of the receptacle may be produced above the point of insertion of the carpels and project into the ovarian cavity, or it may penetrate the ovary as a central column. On the other hand, each carpel may give rise to a separate chamber, in which case one finds a whorl of distinct Fig. 208.— Structure of Phanerogamic Ovaries. I Excavated receptacle and carpels of a Rose (Rosa Schottiana). 2 The sarae in longitudinal section. « A single carpel of the same in longitudinal section. ••Ovary of the Apple (Pyrns Maltis) in longitudinal section. 6 The same in transverse section. « Transverse section of a ripe Apple. ' Carpel of Cycas revolnta with ovules. 8 Longitudinal section of an ovule ot Cycas. 1, «,', 8 natural size ; 2,4,5x3; *x8. ovaries at the tip of the receptacle (fig. 210''); or there may be numerous small ovaries spirally arranged around the receptacle (figs. 207 ^ and 207 ^). In order that the position and mutual relations of the various floral-leaves on disc-like and excavated receptacles may be intelligible it is necessary that we should return to the lump of wax. Let the cone of wax be pressed down so that it assumes the form of a disc or cup. Assuming the floral-leaves to be present upon it during this process — covering the cone from base to apex — when the disc stage is reached the leaves formerly present at the apex will occupy the centre, those at the base the periphery of the disc. If the wax be further moulded into a cup the leaves previously at or near the apex of the cone will THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 75 occupy positions within the cup — those immediately at the apex being at the centre — whilst those near the base will be found on the edge of the cup. According as the leaves are inserted spirally or in whorls upon the receptacle, whether they are present in single or double cycles, whether they are fused with one another or with the receptacle — all these offer almost infinite possibility of variation in form, so manifold, indeed, that their complete description is quite beyond the limits of the present work. Here the forms described must be limited to a series of more or less typical cases; they are for the most part selected from well-known and widely -distributed plants readily accessible to any one. To avoid repetition the seventeen selected cases are arranged in two groups, of which the first group includes forms with a conical receptacle, the second such as have a disc-like or excavated receptacle. Each of these groups is further sub- divided, according as the carpels are all of one sort or of two sorts. OVARIES ON A CONICAL RECEPTACLE. Carpels all of One Sort. (1) The carpels are inserted spirally on the receptacle. Each carpel contains one or several ovules. The receptacle is either much elongated, as in the Mousetail {Myosurus, figs. 207^ and 207^), or conical, as in the Tulip-tree (Liriodendron), or button-like, as in the Crowfoot (Ranunculus). (2) The carpels are inserted in whorls upon the receptacle, their margins are infolded and fused with the prolonged apex of the receptacle. Since they are also fused with one another below, they collectively form a multilocular ovary. Each carpel bears ovules over its inner surface. As examples may be quoted the Yellow Water-Lily {Nujphar), and the Flowering Rush (Butomus, figs. 210^ and 210^). (3) The carpels are inserted in a whorl at the summit of the receptacle and are fused with one another. The receptacle does not project into the ovarian cavitJ^ Each carpel bears ovules either along its margins, as in Mignonette (Reseda, figs. 207 ^ and 207 *), or on its internal surface, as in the Sundew (Drosera), or basally, as in Dioncea, Drosophyllum, and in Caylusea (Resedacese). In Reseda the ovary is open above. Carpels of Two Kinds. (4) The carpels arise at the tip of the receptacle in two alternating whorls of two each. The two upper carpels are reduced to midribs on which the ovules are borne in two rows. A delicate membrane is stretched like a tympanum between these two midribs which form the frame. The two lower carpels are destitute of ovules and are fixed like valves to the upper pair. This form is met with in numerous modifi- cations in the Cruciferse. (5) The carpels arise in two whorls at the tip of the receptacle. Those of the lower whorl are destitute of ovules and form the ovary, those of the upper whorl are modified into strings or cushions, and are fused with the inner surface of the- 76 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. lower carpels. They bear the ovules. Examples:— the Violet {Viola, figs. 207^ and 207 ^), the Celandine (Chelidonium), and the Poppy (Papaver). (6) The lower whorl of carpels are united edge to edge, inclosing the ovarian cavity. They are destitute of ovules. The tip of the receptacle projects a very short distance into the ovary, and bears a single ovule-bearing carpel which is apparently terminal upon it. Examples: — the Rhubarb (Rheum), and Dock (Eumea;, fig. 212 23). (7) The lower whorl of carpels are united edge to edge like staves, forming the ovary into which the apex of the receptacle projects as a central column. The upper ovuliferous carpels are metamorphosed into cushion-like structures consoli- dated with the i-eceptacular column. These cushions are either arranged spirally, as in Glaux (figs. 211 ^ and 211 ^), or in a whorl, as in Primula Japonica. (8) The lower carpels are inserted in a whorl, and have their margins infolded, and are fused together so as to form a multilocular ovary. The upper, ovuliferous carpels arise from the tip of the receptacle, which is continued through the centre of the ovary. The ovules project into the cavities of the ovary. Examples: — The Spurge {Euphorbia), Azalea, Foxglove {Digitalis), Potato {Solarium, figs. 207 ^*^ and 207^1). OVAEIES ON A FLAT OR EXCAVATED RECEPTACLE. Carpels of One Sort. (9) The carpels are arranged spirally upon a raised central cushion of the flat receptacle. Each carpel forms a distinct ovary containing one or more ovules. Examples: — Drijas, Potentilla, the Raspberry {Rubus Idceus, figs. 210 ^^ and 210^^). (10) The carpels are arranged spirally within an excavated receptacle. Each carpel forms a distinct ovary containing one or more ovules. There is no fusion between the walls of the carpels and that of the receptacle. Example: — The Rose {Rosa, figs. 208 1- 2. 3). (11) A single ovuliferous carpel is inserted in the centre of an excavated receptacle. It is apparently terminal upon the axis, and is not fused with the excavated receptacle. This condition prevails in the Cherry, Plum, Apricot, and Almond {Amygdalus, figs. 209^ and 209 7). (12) The carpels arise in a whorl from the end of the axis at the base of an excavated receptacle. Their margins are infolded, and they are fused together into a multilocular ovary. The ovary fills the whole cavity of the receptacle, with the inner wall of which it is fused. Ovules are borne by the infolded margins of each carpel. Examples: — The Medlar {Mespilus), Pears and Apples {Pyrus, figs. 208 ^- 5- 6). (13) The carpels arise from the tip of the axis at the base of the excavated receptacle. The receptacle has a remarkable structure; it is Kke a bottle in shape with three portions of the wall removed, so that it is reduced to three ribs which join above and bear the other parts of the flower. The apertures in the receptacle are occupied by the three carpels. Thus the ovary consists of three carpels and THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 77 three receptacular ridges. The ovules are borne on longitudinally-running cushions on the carpels. This class of ovary is found in great variety amongst the Orchidacese (figs. 207 ^ and 207 2, and figs. 212 1-^. 3.4)^ Carpels of Two Kinds. (14) One series of carpels, destitute of ovules, arise from the margin of the deeply-excavated receptacle, roofing it in. Another series, metamorphosed into Fig. 209.— Structure of Phanerogamic Ovaries. Longitudinal section of the ovary of Cereus grandifiorus. 2 Ovules on a branched placenta from the base of the ovary of Cereus. » Longitudinal section of the ovary of Uedychium angustifolium. * Dehisced fruit of the same plant. « Trans- verse section of the ovary of the same. « Longitudinal section of an Almond ^ovter (Amygdalus communis). ? Lonj:i- tudinal section of the ovary of the same. ^, » Transverse and longitudinal sections of the ovary of the Willow-herb (Epilobium angustifolum). 1 natural size ; », *, 6, 6 slightly magnified ; 2, 7, 8^ 9 x 10. ovule-bearing strings, arise spirally from the inner wall of the receptacle and project into the ovarian cavity. Examples are afibrded by the Cactacere, e.g. Opuntia and Cereus (figs. 209 ^ and 209 ^). (15) One series of carpels closes the mouth of the excavated receptacle, as in (14). The other series, bearing the ovules, are filamentous, and arise as a whorl from the base of the receptacle; they are consolidated with a thread-like prolonga- tion of the tip of the axis which runs up as a central column. Example: — The Willow-herb (Epilobium, figs. 209 ^ and 209 9). 78 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. [These two figures are slightly inaccurate in that the partitions of the ovarian cavity are not indicated. In the cross-section, fig. 209 ^, they would run diagonally from the comers to the central column. In allied forms they are sometimes incomplete. — Ed.] (16) One series of carpels as in (14) and (15). The other series are metamor- phosed into ovuliferous cushions spirally inserted on a continuation of the axis Fig. 210. — Structure of Phanerogamic ovaries. », « Aiitholysis or Chloranthy of a Larkspur (Delphinium cashmirianum). » Ripe dehiscing fruit of same. * Longitudinal section of a single carpel of same. * Longitudinal section of an ovule of the same. <* A single foliaceous carpel of same. ? Pistil of Butomus umbellatus. 8 pistil of same dissected. » Young ovule of same, i" Full-grown ovule of same in longitudinal section, n Vertical section of flower of Raspberry (Riibns Idceus). 12 Longitudinal section of a single carpel of the same. 1, 2, s natural size ; *, «, f, " magnilied 2-6 times; ', «, », 10, 12 magnified 6-8 times. which rises up from the base of the receptacle. Example: — Hedychium (figs, 209 3. *. 5). (17) As before, one series closes the mouth of the receptacle. From the tip of the axis at the base of the receptacle a single apparently terminal carpel arises which bears a single ovule. This condition obtains with many variations in the Compositae, e.g. the Sunflower (Helianthus, fig. 207 ^). The account of the structure of the ovary just given differs in several fundamental points from the current views of the best authorities in plant morphology. Especially is this so in two points. Firstly, in that the wall of THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 79 so-called "inferior ovaries" consists, for the most part, according to my own investigations, of a deeply excavated receptacle and not of carpels invested by the tube of the calyx or perianth. That the latter condition occurs (as in many Fig. 211.— Aiitholysis and Structure of the Ovary. *-• Longitudinal sections of the ovaries of "monstrous " flowers of Prinmla jajionica ; the outer carpels form the ovarian cavity and are destitute of ovules; the inner carpels show all transitions between ovuliferous cushions, concrescent with the extremity of the axis, and isolated leaf-structnres, the marginal teeth of which correspond to ovules. ' A single "mon- strous" flower of Primula japonica. 8 Longitudinal section through the ovary of Glaux maritima. » View into the ovary of same after removal of the front wall. ' natural size ; the others magnified 6-8 times. Saxifrages) is not here denied, but more frequently is it the receptacle which is raised as a circular wall to form a closed ovary. On the ripening of the fruit the capsule in many cases opens by means of valves which strikingly resemble the valves formed from true carpels. It is, however, but a resemblance comparable to that existing between the phylloclades of Butcher's-broom and true leaves {cf. vol. i. p. 333). 80 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. A second divergence from recognized views is the assumption that two kinds of carpels take part in the formation of many ovaries, i.e. an outer series, destitute of ovules, forming the ovarian cavity, and an inner, ovuliferous series variously metamorphosed into cushions, strings, ridges, &c. This view is supported not only by extensive investigations into the development of ovaries, but also by a number of cases of antholysis which throw considerable light on obscure points of ovarian morphology. As we shall refer frequently to this state of Antholysis it will be well to state at once, briefly, exactly what is meant by the term. Everyone is acquainted with the "double flowers" of Roses, Snowdrops, Carnations, Primroses, Tulips, &c., so common in cultivation. Into the cause of their origin we shall inquire later on; here it is suflacient to note that in double flowers we find (1) that the stamens are entirely or in part transformed into petals, occasionally into carpels; (2) that a multiplication of the perianth-leaves, stamens, and carpels is apparent, and (3) that with this change is often combined a greening of the parts not usually green, and (4) a general loosening and separation of parts which in ordinary, single flowers are fused with one another. Especially do we find those leaf-structures which normally are united to form the ovary loosened and increased; they are produced as flattened structures, having much the appearance of green leaves. One finds frequently all possible transitions in one and the same flower, so that the various stages in the conversion of carpels into green leaves can be readily followed. In cases of antholysis where the parts of the ovary show a transformation into green leaves, one feels justified in regarding the structures in question as foliar in nature. Especially is this so when none of the ascertained facts of development militate against this view. In the same way such parts as never assume the forms of leaves in these " loosened " or segregated flowers may be interpreted as stem- structures — always provided that developmental history harmonizes with this view. In the cursory review of types of ovarian structure given in the last few pages it was stated that in some cases carpels of one kind only are present, whilst in other cases carpels of two kinds contribute to the formation of the ovary. This statement is based in part on facts gleaned from an examination of these loosened, antholytic, or so-called "monstrous" flowers. The antholytic flowers of a Larkspur (Delphinium cashmirianum) reproduced in figs. 210 ^"^ show unmistakably that only a single whorl of carpels is present and that each of them bears ovules on its margins. Similarly those of the Japanese Primrose (Primula japonica) represented in figs. 211 ^' 2- ^> *• ^' ^' '^' demonstrate that here two sorts of carpels are concerned, i.e. outer foliaceous ones destitute of ovules, and inner ovuliferous ones modified into a cushion. Having described the chief forms assumed by the ovarian cavity, we may pass on to speak of its most important contents, the ovules. All ovules agree in this: that at the time of fertilization they consist of masses of tissue, exhibiting a difier- entiation into central and peripheral cells, and also in the fact that one of the cells of the central portion is destined to become an embryo. In the majority of THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 81 flowering plants we find a definite central mass of cells, the nucellus, surrounded by a well-marked sheath, the coat or integument. Generally the integument is double, as in Delphinium and Butomus (cf. figs. 210 ^' ^' ^°), in other cases it is single, as in Compositse, Umbelliferae, Hippuris and Cycas revoluta (cf. fig. 208^). In most Orchids the nucellus is inclosed in a large-celled, inflated and transparent integument, through which it is distinctly visible (cf. fig. 212^). In not a few epiphytic Orchids, however, this contrast of parts is only imperfectly shown, whilst in the BalanophoresB and various other parasites no trace of the distinction into nucellus and integument is found. In all cases where an integument is present it is discontinuous at one point, where the nucellus is uncovered. This is the micropyle. Sometimes the micropyle is at the apex of the ovule, but in a very large number of cases the whole ovule is bent over so that the micropyle is situated close to the point of attachment of the ovule. The ovule may be attached to its support (placenta) by means of a filamentous cord, or it may be directly seated upon it. The common condition of an inverted ovule fused with its filamentous stalk is shown in figs. 208^ and 210^*'. The filamentous stalk is technically known as the funicle, and the ridge where it is fused with the ovule as the rajihe (cf. vol i. p. 644). The cells of the nucellus of the ovule show a very unequal growth. One of them enlarges in a conspicuous manner, and is known as the Embryo-sac. In Conifers it attains relatively to the other cells of the nucellus enormous dimensions, whilst in most other flowering plants as it grows it encroaches upon the other cells of the nucellus till only a single layer remains surrounding it. And even this layer may be in part absorbed, so that the embryo-sac actually penetrates to the micropyle. The protoplasmic contents of the embryo-sac is richly vacuolated, but at the end directed towards the micropyle vacuoles are absent, and the protoplasm breaks up into several distinct protoplasts, each of which is provided with a con- spicuous nucleus but in the first instance with no cell-membrane. As a rule three such protoplasts are found at the micropylar end of the embryo-sac; of these one only gives rise, after fertilization, to an embryo. This cell is the ooplast or "ger- minal vesicle", the other two are named synergidce (cf. also, figs. 315 and 316). In the ovaries of Orchids, as shown in figs. 212^'2'3>*, the ovules arise in great numbers upon peculiar furrowed ridges of the carpels. They arise from the super- ficial cells of these ridges, and are not provided with any vascular-bundle connec- tions; in fact, they are comparable to those epidermal structures known as hairs or trichomes. This analogy is emphasized by the fact that in the ovaries of many Orchids real hairs are present, as, for instance, in Loelia Perrinii and Coslogyne plantaginea, transverse sections of which are represented in figs. 212^-2.3,4. j^ these remarkable species six ridges project from the wall into the ovarian cavity, and from all of these hair-like structures are developed. The three ridges belong- ing to the curious excavated receptacle, already described, alone bear ordinary anicellular hairs, the others bear ovules, one of which is shown in fig. 212^. The ovules of Cycads are very differently developed, as may be seen from a 82 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. reference to fig. 208 ^, Here no ovarian cavity is formed, the carpels are distinct from one another, and are spirally inserted upon the termination of the caudex; they are deeply lobed, certain of the segments being transformed into ovules. Thus, while the ovules of Orchids seem to be equivalent to hairs, those of Cycads represent leaf-segments. In both cases the relations of the parts seem obvious. But in a great many cases the significance of the ovules is by no means so obvious, especially when the developmental history admits of various interpretations. In such doubtful cases antholysis offers a welcome assistance — that is, where this " loosening" and "greening" involves not only the ovary but also the ovules. Especially valuable in this respect are certain cases of antholysis of the flowers of the Sundew (Drosera). Whilst in the normal flowers of this plant the ovules arise on the inner surface of the united carpels, in the foliaceous or antholytic ones they are borne upon the open and isolated carpels as glandular tentacles, like those usually occurring upon the leaves of this plant (cf. fig. 212^). On many of the carpels these glandular structures are fused together in little clusters (212^), and these fused structures show various transitional stages leading up to inverted ovules (figs. 212 ^'®' ^°' ^^' ^2). From a study of these cases one may infer that the integument of the ovule here is equivalent to a group of tentacles. Very different is the case of the Larkspur (Delphinium). In normal flowers the ovules arise from the infolded margins of the carpels, each of which forms an ovary (cf. fig. 210*). But in the foliaceous flower the carpels are open and their margins lobed (cf. fig. 210 ^ and fig. 212 ^^). They recall the carpels of Cycas (fig. 208 7) and agree with it in that some of the segments are converted into ovules. And it must be especially noted that the leaf-segments are so folded that a pit-like excavation is formed (cf. figs. 212 ^* and 212 ^^). Thus it appears that in the Larkspur the ovular integument is formed by the folding of the leaflet-like segments. Different again is the case of the Clover (Trifoliwni), of which an antholysis is shown in fig. 212 ^^ The ellipsoidal ovules, which are borne along the fused margins of the infolded carpel in the normal flower, are here replaced by little, leafy structures resembling leaflets on the margin of the open carpel (cf. figs. 212 ^^ and 212^''). These leafy structures are neither rolled up nor folded, and from each projects the nucellus of an ovule, or rather a mass of tissue corresponding to a nucellus, surrounded by an enveloping wall (cf. figs. 212 ^^•^^'2°'^^). This wall may be regarded as representing the inner integument of the ovule, whilst the outer one is replaced by a leaflet. The monstrous ovules in the ovary of the Common Sallow (Salix Caprea, fig. 212^^) show similar relations, except that the green, leafy structure upon which the nucellus of the ovule is inserted is folded along its midrib and has a fimbriatecJ margin (fig. 212 ^°), Of especial interest are the monstrous flowers of Rumex scutatus (cf fig. 212 2*- ^^- 2^- 2^' ^s), a plant common on the debris slopes of limestone mountains. In the normal flower of this plant the ovary is egg-shaped, and consists of three carpels united edge to edge (figs. 212 ^^ and 212 ^). But in these monstrous cases it is enlarged from six to tenfold, and modified into a funnel-shaped tube open above (212 2*. 25, 26, 27 y From this the ovule, also modified into a tube, sometimes THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. 83 Fig. 212.— Ovules and Foliaceous Carpels. Transverse section of the ovary of Loelia Perrinii; natural size. 2 A portion of this section ; x6. » Transverse section of the ovary of Ceelogyne plantaginea. 4 A portion of this section ; x6. » A seed of Ccelogyne plantaginea. « Aiitholysis of the flower of Sundew (Drosera intermedia). (After Planchon.) f-12 Isolated portions of this liower. "-is Isolated por- tions of a similar flower of Delphinium elatum. (After Cramer.) is Antholysis of Trifolium repens. "-2> Isolated por- tions of the same. (After Caspary.) 22 Flower of liumex scutatus. 23 The same flower in longitudmal section : magnified. 24-28 Isolated portions from an antholysis of Rumex scutatus. (Partly after Peyritsch.) 29 Longitudinal section through the pistil of a " monstrous " flower of Salix caprea. so Foliaceous ovule from this pistil, e-so slightly magnified. )rojects (212 ^*), or it may remain concealed within (212 2^). Inside the ovular tube rises a little protuberance which may be regarded as equivalent to the nucellus of 84 THE COMMENCEMENT OF THE PHANEROGAMIC FRUIT. the ovule. It is sometimes attached to the wider end of the tube (212^^), but more frequently it arises from the narrowed base as a tiny, conical projection inclosed in a circular envelope of its own (212-^). This envelope corresponds to the inner, and the tube to the outer integument of the ovule. From a study of these monstrous flowers it would appear that when the ovule possesses two integuments, the outer one corresponds sometimes to the whole apical portion of a carpel, sometimes to but a segment of a carpel; the former being the case when carpels of two kinds are present, and when, at the centre of the floral receptacle, above the outer non-ovule-bearing carpels, only a single fertile carpel is produced. The inner integument, on the other hand, arises like a corona from the leaf -like outer one. The nucellus of the ovule arises in many instances {e.g. in Orchids) from a mass of tissue produced by the division of a single epidermal cell, but in by far the majority of cases at the margin or upon the surface of a leaf or leaf-segment, resembling in all respects a foliar bud. That the ovule can be produced directly from the floral receptacle is not yet ascertained with certainty, though such an origin would appear to be not improbable in the Pepper family. That is no good reason why ovules should behave differently from bud-like brood-bodies, which arise sometimes from leaf- and sometimes from stem-structures. So great is the analogy between ovules and detachable buds, that ovules formerly received from Botanists the name of "seed-buds". In this con- nection it is very instructive to contrast the ovules in the ovary of certain Orchids with the foliar buds produced on the leaves of some of these plants. In Malaxis paludosa (cf. fig. 200^, p. 41) the foliar buds are found partly on the upper surface of the leaf, partly on the margins, forming in the latter case a fringe. They con- sist of a compact, central portion inclosed in a large-celled envelope which is so fashioned that the whole structure resembles an ovule (cf. fig. 200^). So striking is this resemblance, that anyone unacquainted with the fact that these buds arise from foliage-leaves would unhesitatingly regard them as ovules. Later on, of course, differences appear, in that in the ovule an independent embryo is produced, whilst the bud gives rise to a shoot, which must be regarded as a branch of the parent plant. This is, of course, an important distinction, and applicable to the majority of cases, though not quite to all. The parthenogenetically produced brood- bodies, to be treated fully by and by, have both the form of true embryos and occupy the same position in the ovule beneath the micropyle. Were it not known that the hard, indehiscent fruit (achene) of Gnaphalium alpinum ( = Antennaria alpina), with the rudiment of another generation which it contains, is produced without the intervention of pollen, without fertilization, it would certainly not be apparent from its structure. From this we may conclude that the distinction between bud and ovule, between brood-body and fruit, cannot be based on purely structural characters, and that fruits and brood-bodies are sometimes interchange- able— facts of great importance in solving the question of the importance of fertilization in the origin of new species. STAMENS. 85 STAMENS. As the last patches of snow disappear from the fields, the Snowdrop raises its white bells, and the catkins of the Willow break through the bondage of their bud- scales; in the copses likewise, where the warm March sunbeams penetrate, the Hazel begins to blossom and sheds its powder. These are the signs that spring is coming, and that the long winter is over. For some time the flowers both of the Snowdrop and Hazel have been ready — in the Snowdrop under ground, wrapped up in sheathing leaves; in the Hazel on the twigs as short, cylindrical, dusky catkins. With the advent of spring the catkins stretch and their crowded flowers are separated, they becoming flexible and hang like golden tassels from the branches, swaying in the wind and giving ofl* their clouds of dust. To this powder, long known to be connected with the fruiting of plants, the name of flower-dust has been given. This term, suitable in so many cases, has been used in others for a substance which, although corresponding in function to the flower-dust of the Hazel, differs from it in appearance. The cells which take the form of dust in the Hazel assume in other plants the form of sticky, viscous lumps, of spindle-shaped masses or granulated bodies, to which the designation dust is quite inappropriate. Were the species of plants whose flowers do not produce dust but few the term could stand, but when we find belonging to this category many of the principal families of plants — ten thousand Composites, eight thousand Orchids, five thousand Labiates, four thousand Rubiaceae, three thousand Papilionacese, and thousands of Umbellifers, Rosacese, Crucifers, &c.; that, roughly speaking, two- thirds of Flowering Plants do not produce dust, it is evident that the term cannot have a general application. Consequently, Botanists speak of Pollen and not flower-dust. It is true this word simply means flour, and that its selection has not been a very happy one. Still the term has entered into botanical terminology, where it will remain. It is given to all those cells produced in the flowers of Phanerogams, which contain the spermatoplasm. Pollen, then, consists of cells which contain spermatoplasm, and may be compared to the antheridia of Cryptogams. A definite portion of the substance of certain leaves of the floral axis is appropriated to the production of Pollen. These leaves, known as Stamens, resemble the other leaves of the floral axis in that they are inserted in whorls, or one above the other in a much-flattened spiral. Very few species of plants possess only a single stamen in each flower. The majority of flowers contain stamens arranged spirally or in whorls. As a rule stamens are inserted according to the i or f system (c/. vol. i. pp. 399, 400). In many cases their number and insertion resembles that of the petals and carpels of the same flower, though more frequently there is a difierence. Thus, in the flowers of the Tulip-tree (Liriodendron), whilst the perianth-leaves have a divergence of ^, the stamens are arranged according to the ^ system. In Ranunculus the leaves of the perianth are arranged on the f plan, the stamens on the ■^; in Polygonum the former on the f, the latter on the f system. 86 STAMENS. Since in every species of plant the number of stamens remains constant, thus in the Mare'stail {Hipimris) there is one, in Lilac two, in Iris three, in the Woodruff four, in the Violet five, and in the Tulip six stamens, their number has been made the basis of a classification of flowering plants at once convenient and popular, though not strictly scientific. In the well-known System of Linnfeus plants are arranged into groups called Classes, in which the first class {Monandria) includes all forms with a single stamen, the second (Diandria) those with two stamens, and so on. The aggregate of stamens in a flower is termed the Androeciuvi. As a rule the Fig. 213. — Stamens of double and monstrous flowers. Vertical section of a green flower of Primula japonica. « Vertical section of a double flower of Primula spectabilis, »-8 Isolated stamens from the same flower, s stamen from a green flower of the Tiger Lily(Lt7t!(m tigrinum). lo, " Folia- ceous stamens from a flower of Campanxda Trachelium. 12 Green flower of Saxifraga stellaris. " A single stamen from the same flower. (All the figures enlarged.) androecium is inserted between the leaves of the perianth and the carpels, so that from without inwards the sequence is perianth, stamens, carpels. Sometimes the carpels are wanting, so that the stamens constitute the inmost members of the flower ; similarly also carpels may be present but no stamens. We distinguish in a stamen that portion which is concerned in the production of Pollen — the Anther — and its stalk, the Filament. The stamens in many flowers are partly metamorphosed into petals; indeed, there are grounds for believing that all petals have been originally modified from stamens. What are known as "double flowers" are often flowers in which the stamens have given place to petals. All intermediate stages between stamens and petals can be seen in double-flowered Roses, Carnations, and Primulas (c/, figs. 213 2- *■ ^' ^' ^). Not infrequently, at the STAMENS. 87 place where a petal narrows into its stalk or " claw ", a little yellow swelling or callosity may be seen; this may be regarded as a reduced anther, and now and then it possesses the character of an anther, and contains actual pollen. It is frequently observed in double flowers that a multiplication of the leaf accompanies the con- Fig. 214.— stamens. Empleurmn serrulaturn. ^ Hypericum olympicum. ^ Juglans regia. * Soldanella alpina. i Viola odorata. ^.^ Artemisia Absynthium. s Haminia (after Baillon). » Ahies excelsa. lo Euphorbia canariensis. n, i» Platanus orientalis. 13, 14 Juniperus Sabina. is Ualiimiocnemis gibhosa. is Halantium Kulpianum. i? Sanguinaria canadensis. " Allium sphoerocephalum. is Actcea spicata. 20 Aconitum Napellus. 21 Salvia officinalis. «2 Viscum album. 2S Mirabilis Jalapa. 2* Tilia ulmifolia. 25 Thymus serpyllum. 26 Acalypha (after Baillon). 27 Bryonia dioica. 28 Jliciiin.i com- munis. ^^ Corydalis capnoides. ^0 Polygala amara. ^^ Doryphora (after Baillon). i^ Paris quadri/olia. (All fljiures somewhat enlarged.) version of stamens into petals. In the place of a single stamen we may find two stamens partially converted into petals, or there may be a greater number of petal- like leaves, standing one behind another, or, finally, we may have the appearance shown in figs. 213 ^ and 213 ^ of a double Primula. By the action of parasitic Aphides and Insects stamens often assume a leaf- like appearance, they become green like the carpels described on p. 80. Such instances are of value in comparing the various parts of a stamen with those of the 88 STAMENS. hypothetic fundamental type of leaf -structure. At the first glance it might be sup- posed that the filament is a metamorphosed petiole, and the anther a metamorphosed lamina. But these monstrous flowers seem to indicate that such is comparatively rarely the case. Thus in the green stamens of Campanula Trachelium (figs. 213 ^° and 218 ^') there are scattered everywhere over the green substance of the lamina yellow excrescences and warts containing reduced pollen-cells, and occasionally these occur fused together into actual portions of anthers ; hence it may be inferred that in this case the anther may be regarded as equivalent to a green lamina. But far more frequently in such cases the pollen-producing tissue is found at the base of the lamina only, at the upper extremity of the leaf-stalk, where these two parts articulate. From this we may conclude that in the majority of cases anthers corre- spond to that portion of a leaf at which the stalk runs into the lamina. In such stamens the lamina is entirely suppressed, or is represented by a continuation above the pollen-producing region. A few forms of this continuation above the anther, which we regard as repre- senting a leaf lamina, are illustrated in fig. 214. Figs. 214^ and 214 ^ show it as a small shot-like grain, 214^ as a truncate cone, 214* as a two-pronged fork, 214^'^'^ as a flat, triangular scale, 214^ as a toothed, sword-shaped process, 214^' ^^' ^2. is. 1* as a curved membraneous scale, 214^^ and 215^^ as a coloured bladder for attracting insects; and, finally, figs. 214^^ and 214^"^ as a long, whip-like bristle. That the filament of the stamen, or at any rate its lower portion, corresponds to a leaf -stalk seems so obvious, that it is hardly necessary to prove it by comparison with monstrous cases. Its name of filament indicates its character in a great number of flowers. Examples of these are Hemp, Hop, Wheat, Rye, Rice, Maize, Flax, and many others. For many cases no doubt the term filament is unsuited, as, for instance, in the thick, abbreviated stalks in the Violet and Bryony (figs. 214^ and 214 ^"). Similarly the filament may be strap-, spindle-, or club-shaped. The last is the case in Thalictrum aquilegifolium., Bocconia, Sangidnaria, and Actcaa spicata (cf. figs. 214 ^"^ and 214 ^^), and it has been observed that the stamens very readily oscillate at the moment of liberation of pollen with the slightest breath of air. Like the foliage-leaves of the Orange, the stalks of which are provided with a peculiar joint, many Spurges and Labiates have hinged filaments {cf. figs. 214^° and 214^^). These hinges are wonderfully fashioned in many species of Salvia, reminding one of the articulation of the feet of insects; their importance in fertili- zation will be described in a later chapter. In the Linden the filament forks immediately below the anther (fig. 214 2*), whilst in Corydalis it is band-like, and divides into three (fig. 214 ^^). In the Castor Oil Plant (Ricinus), and many other Euphorbiaca?, it is much divided and branched (fig. 214 ^^). These divided filaments are not to be confused with fused ones, for occasionally we find that the filaments of adjacent stamens unite with one another into a ribbon or tube, as for instance in Mallows, Papilionacese, and Polygalaceoe (cf 214 ^*'). Attached to the sheath of foliage-leaves curious appendages, the stipules, are often found (cf vol. i. p. 595). In the case of stamens these are but rarely met STAMENS. 89 with. They occur, however, in certain species of Orni-'iiogalum (e.g. Ornithogalum nutans and chloranthum), in Allium rotundum an'- sphoerocephalum, and in the Monkshood (Aconitum). Occasionally such staijain .1 stipules are modified as honey- secreting glands at the base of the stamen, e.g. DcW^^'^^ i^f- %«• 214^^ and 214 ^o). It sometimes happens in monstrous flower^ that the stamens are transformed into carpels, or we may find here and there '-n isolated stamen, which is partly so modified and partly still polliniferous. Ir such monstrosities it usually happens that it is the upper part which forms p'^Hen, and the lower part which produces ovules (cf. figs. 213 ^ and 213 ^). From inis and other facts it has been inferred that the ovary corresponds really to the s^^eaths, the style to the petioles, and the stigma to the laminae of the floral-leaves ojncerned. The monstrous flower of a Saxifrage (figs. 213^- and 213^^) shows th-it anthers and ovules can be produced from the same part of the leaf-stalk, 'rhis flower (213 ^^) produces at the periphery five sepals and five narrow, green petals; in the centre two carpels (shaded dark in fig. 213 ^-) as in normal Saxifrage flowerg. Between the petals and carpels, i.e. where the stamens are usually founr^, there are ten structures which, whilst resembling both carpels and stamens to some extent, remind one forcibly of the excavated leaf- rachis of so many of the Pitcher Plants (cf. vol. i. pp. 125-133.) One of these is represented in fig. flS^^. Its free extremity consists of an irregularly serrated scale, which may be compared either to a stigma or to the continuation of an anther, and may be } egarded as the metamorphosed lamina. The excavated portion below may be regarded as the j)etiole. In its cavity are four rows of yellow protuberances, which might at first sight be taken for ovules. Closer investigation shows, however, that they contain pollen-mother-cells, each inclosing four pollen- grains. Here, then, we find the petiole consisting partly of carpel and partly of anthers, from which it may be concluded that that portion of the carpel which produces ovules corresponds entirely in position to the pollen-producing tissue. The parts of the anther which produce Pollen in special chambers are known as Pollen-sacs, the tissue which binds these together as the connective. The connective is a direct continuation of the filament, and, like this, is penetrated by a vascular bundle. The pollen-sacs may be arranged like niches around the columnar connf^ctive, which itself terminates in a sort of little shield, as in the Yew Tree (cf. fig. 234 -), or they may be situated symmetrically right and left of it. In the latter c^ise the pollen-sacs may lie at the edge of the connective in one place, as in the Juniper (figs. 214^^ and 214^*), or they may be in pairs, i.e. two pollen-sacs to the right and two to the left of the connective (fig. 214 ^). This latter form is by far the most frequent, and occurs in certainly 90 per cent of ail Phanerogams. It must be pointed out that the two pollen-sacs of each pair are separated from one another by a partition-wall only in the young anther. This disappears hiter on, and in the mature anther one finds, instead of four, only two sacs filled with pollen. Sometimes all four pollen-sacs run together in this "^^y> by the breaking down of the parti- walls, as in Sundew (Drosera), Moschatel (Adoxa), Monotropa, and especially in Globularia (cf. figs. 216 ^^ and 216 ^^). Id 90 STAMENS. Orchids, on the other hand, -.he number of pollen-sacs is reduced to two, a number which remains unaltered at maturity. The pollen-sacs in the anthe-s of the Mimoseae are very curiousl}^ formed. In the anthers of Acacia, Albizzia, (Jalliandra, and Inga, there are eight spherical chambers in which pollen is producc^d, whilst in Parkia we find longitudinal rows of lenticular cavities in which balls of pollen lie embedded. The anthers also of the Rhizophoreffi show several long>,udinal rows of such chambers, amounting in all to as many as thirty. The anthers of the Mistletoe {Viscum, fig. 214-') contain as many as forty to fifty pollen-cha.nbers. In the majority of the Laurels (Lauraceae) each anther is divided into four cavities, which stand in pairs, one above the other. As a rule, all four open to^vards that side by which insects visiting the flower for honey have to pass. Many marked variations in the form of the aither are due to the relative dimensions of connective and pollen-sacs. Thus in the majority of Ranunculacese, Magnoliaceae, Nymphfeaceae, and Papaveraceae, the con- nective is broad, the pollen-sacs forming only a narrow rim to the anther (c/. fig. 21 ti"). In the Skull-cap {Scutellaria), Calamint (Calaminiha), Thyme (Thymus), and many other Labiates, as als^ in several Rosacese (Rosa, Agrimonia, &c.), the connecvive has the form of a three- to six-sided mass of tissue in w^hich are embedded Fig. 215. -Curved anthers In the the sphcrical or egg- shaped pollen-5:acs. Such anthers SL^aVS B«. "'■ frequently resemble an insect's head with two lateral eyes. It is not always possible to distinguish the limits of con- nective and filament, the whole stamen resembling a truncate column or anvil (figs. 216 -'^ and 216 =^-). Sometimes the connective assumes the form of a bar or lever running transversely to the filament, to which it is attached by a movable joint. This is notably the case in certain species of Salvia, to be described hereafter. Such a connective moves very readily upon its fulcrum. In many Liliaceae (e.g. Tulips, Lilies, and Crown Imperials) and several Gentians (Gentiana ciliata, nana, Szc), the anther is united with the filament by an extremely delicate joint, so that the slightest touch sets it in vibration (versatile anthers). As examples of bulky pollen-sacs and much reduced connective, Mirahilis Jalapa (fig. 214 23) and Solarium Lycopersicum (fig. 216^) may be quoted as examples. It stands to reason that the character of the anther, indeed of tlie whole stamen, is correlated with the form of the pollen-sacs. All possible stag(3s occur between globular and egg-shaped, and between egg-shaped and linear pollen-sacs. The drawings of sixty-four different stamens in figs. 214 and 216 give a good idea of the variety in this respect. Very curious are the curved anthers of Phyllanthus Cyclantkera (fig. 215), and those of Acalypha, which resemble a ram's horns (fig. 21426); ^jjQ same remark applies to the undulating anthers of many Cucurbitaceae, of which those of Bryonia dioica may serve as an example STAMENS. 91 (fig. 214 -7). There are forms allied to this last-named plant in which the anthers show very complicated convolutions — like those of the human brain. When the time draws near for the pollen to leave its place of origin, its cells — whether in a loose powder or sticking one to another — become free from the inclosing wall of the anther, and lie embedded in the cavity of the pollen-sac, as it were in a purse or pocket, awaiting their release. The pollen-sac, hitherto stamens. Calandrinia compressa. ^ Solanum Lycopersicum. ^ Galanthus nivalis. * Cyclamen europceum. ^ Ramondia pyreiMica. ^,T Cassia lenitiva. » Pyrola rotundifolia. ^ Arctostaphylos Uva-ursi. ^^ Arctostaphylos alpina. » Vaccinitim uliginosum. n Pyrola xmiflora. is Medinilla (after Baillon). i* Vaccinium oxy coccus. i* Calceolaria Pavonii. 1' Tuzzia alpina. ^t, is Sihbaldia procumbens. is Galeopsis angustifolia. -o, 21 Erythrcsa Centaurium. ''-, 23 Jleli^sa officinalis. 2*, 25 Calla palustris. 26 Nyctandra (after Baillon). 27, 28 Globularia cordifolia. 29, so Theobroma Cacao. SI Pinguicula vulgaris. S2 Garcinia. (All somewhat enlarged.) closed, now opens, and the pollen is liberated. This opening or dehiscence of the pollen-sacs is accomplished in various ways. It has been already explained that most young anthers contain four sacs which rarely all remain distinct, but, by the breaking down of the parti-walls between each pair, become merged into two cavities. These two cavities may be spoken of as anther-halves. In cases where the 92 STAMENS. four cavities remain distinct, a curious aperture is formed above each of them, as, for instance, in Theohroma Cacao (figs. 216'-^ and 210^*^). When, however, the aforesaid fusion takes place (e.g. Calla palustris, figs. 216 ^^ and 216-^), two openings only are formed. The anthers of Glohularia have a very small punctiform connective and four pollen-sacs joined into an ellipsoidal body. After the disappearance of the parti-walls, leaving a single cavity occupied by pollen, a gaping, transverse slit arises, so that the anther is transformed into an open vessel (c/. figs. 216 ^^ and 216 ^^). After the removal of the pollen the remains of the original parti-walls can be distinguished as two intersecting ridges. Similarly in the Butterwort {Piiiguicula, fig. 216^^) and in the majority of so-called one-celled anthers. In many Labiates, in which the anthers of adjacent stamens are in contact, and to some extent united together (syngenesious), the openings of the pollen-sacs in each anther unite, with the result that a pair of somewhat sinuous niches are presented, borne on the two curved filaments (c/. figs. 216 22 and 216 ^s). Dehiscence is accomplished sometimes by the formation of holes or pores, •sometimes by slits. Of anthers with porous dehiscence, the greatest variety is exhibited by the Heath tribe and Pyrolaceae. In the anthers of the Bilberry, Bog Vaccinium, Cowberry, and Cranberry {Vaccinium Myrtillus, uliginosum, Vitis-idcea, Oxycoccos), as also in Winter-green (Pyrola), the pouch-like pollen-sacs ure drawn out into shorter or longer tubes, each of these tubes opening at its extremity by small circular pores (c/. figs. 216 ^'^^'^2. i4^^ g^^ much more frequent is a dehiscence by means of slits. These are either longitudinal or transverse, or they may be sinuous or semicircular. When they are semicircular a valve or trap-door is cut out of the anther- wall. At its first formation the slit resembles one cut by a sharp knife (cf. fig. 216 ^). In a number of cases the margins of the slit remain together, so that the aperture retains the form of a narrow crack; more frequently, however, the slit gapes, its margins roil up outwards or are folded back like a lid or folding-door. The longitudinal slits reach from end to end of the pollen-sacs (fig. 216 ^), or they may take the form of short gaping clefts near the free extremity of the anther. In the latter case (several examples of which are represented in figs. 2iQ 2, 3, 6, 7, 9, 10, 13, 15, 16^^ ^j^g g]j^g very much resemble pores, from which they can only be distinguished in some cases by their mode of development. Occasionally the short, gaping clefts of adjacent anther-halves unite into a single opening, with a heart-shaped or rhomboidal outline, by which the whole of the pollen of both anther-halves escapes (examples are Cyclamen smd Ramondia, figs. 216^ ^nd 216^). Transverse slits are met with most frequently in the stamens of Euphorbiaceae, Cyclanthaceae; also in a few Rosaceae (Alchemilla and Sibhaldia, figs. 216 ^'' and 216 ^^), in the Golden Saxifrage and Moschatel {Chrysosplenium and Adoxa), in Glohularia, Malva, and others. On the whole, however, this method of dehiscence is rare. Of still rarer occurrence is that form of dehiscence in which semicircular slits arise in the anther-wall, producing valves or trap- STAMENS. 93 doors. This is known as valvate dehiscence. It is met with generally in Berberidaceae (e.g. Berheris and Epimedium) and Lauracece. In the Bay Laurel, Camphor, and Cinnamon Trees {Laurus nohilis, Camphora officinalis, and Cinna- momum) and Nyctandra (fig. 216 2^) are found little apertures on one side of the stamen, each with its trap-door or valve, which is raised up in dry, but shut down in wet weather. The anthers of Mimulus, Galeopsis, and Garcinia (figs. 216^^ and 216^^) resemble little tubs or boxes, which on opening raise their lid-like valves. The dehiscence of the anthers in many plants is accompanied by yet other changes. The two anther-halves may become partly separated from their attach- ments and become twisted or diverge at right angles. If the anther-halves separate at the base only, as in Convolvulus, Gentiana, and Menyanthes, the anther assumes the form of an arrow-head; if they separate both above and below, and at the same time become somewhat bent, we have an X-shaped anther, found in many Grasses. In many Crucifers {Diplotaxis, Sinapis, &c.) the anthers become spirally twisted after dehiscence, a feature very pronounced in the Centaury {Erythrcea, figs. 216 ^*^ and 216 2^). A very striking phenomenon is the shortening which not a few anthers with longitudinal slits undergo after dehiscence. The anthers of most Liliaceous plants are long and linear; they dehisce by means of slits from above downwards. In the course of a few hours they are transformed into globular bodies, covered with pollen. In Gagea lutea these balls have a diameter only one-third of the previous length of the anther, whilst the anthers of the Crown Imperial (Fritillaria imperialis) shorten from 20 to 10, those of Narcissus poeticus from 11 to 4, and those of Scilla hifolia from 2 to 1 millimetres. Each one of the various occurrences which accompany or succeed dehiscence depends upon some structural character of the anther-wall. The relations are simplest in those anthers which open by means of pores. The pores arise from the absorption of limited portions of the wall. Further changes, such as the shrivelling or shortening of the anther, or the expanding of the apertures, do not occur. There is a corresponding simplicity of the tissues of the anther-wall. Similarly, in anthers (e.g. Orchids) in which a splitting arises along a previously- indicated line, or in consequence of the absorption of a row of cells, no peculiarities are noticeable on the wall. But where slits with movable lips and valves are developed, cells of characteristic structure are present, which may be termed the contractile cells. One series consists of more or less cubical cells, and exhibit, on portions of their walls, fibrous or rod-shaped thickenings. The wall of one of these cells directed towards the cavity of the anther is equally thickened, that towards the outside is thin and delicate, easily folded, and destitute of thickenings. The side-walls, however, are characteristically strengthened by rod-like thicken- ings, which in their distribution may be compared to a hand, in the position usually employed in grasping an apple; the palm corresponds to the strongly- thickened inner wall, and the fingers to the tapering, rod-like thickenings of the y4 STAMENS. side-walls. As the cells dry a contraction of the rod-like thickenings supervenes, leading to a movement like that of the afore-mentioned hand when the tips of the fingers approach one another. Simultaneously the thin outer walls are thrown into folds, so that where a number of these cells are present, side by side, the whole outer surface will contract. These cells, being appropriately distributed over the wall of the anther, will cause the slit-margins to fold back or the valves to be raised. Besides these, other forms of contractile cells are present, difiering from those described chiefly in form rather than in their mode of action. It must suffice here to mention only a very few instances. The anther-wall in Conifers consists of a single layer of contractile cells, whilst that of Agave reaches the other extreme, there being six to eight layers of such cells present. As a rule the contractile layer is covered externally by a layer of delicate, thin- walled cells, known as the Exotheciuini; the contractile layer constitutes the Endothecium. The lining of the pollen-chambers consists of yet a third layer, the tajMal cells. In anthers which have dehisced this last-mentioned layer is rarely demonstrable, it having been already absorbed. Of the various layers it is the middle one, the endothecium (contractile cells), which is active in the various movements under discussion. In the discharge of the pollen from the opened anthers a great variety of methods prevails. In the Nettle and Mulberry the filament of the stamen uncoils like a spring at the moment of dehiscence of the anther, and the pollen is forcibly scattered (fig. 229). The whole event is instantaneous, and to the observer resembles an explosion. In other plants dehiscence is accomplished quietly, and the pollen, which escapes slowly, may be first of all stored up temporarily at definite spots within the limits of the flower. This storage occurs a good deal more frequently than is generally supposed, and stands in relation to various events which will be fully discussed later on. In Papilionacese the liberated pollen is deposited in the hollowed apex of the Keel; in the Violet it is stored in the grooves of the lowest, spurred petal; in the Poppies, Roses, and Buttercups, it falls, at any rate in part, on to saucer-like depressions of the petals. The dust- like pollen as it falls from the anthers of the catkins of the Walnut, Hazel, Birch, and Alder, is received temporarily on the upwardly-directed under-surfaces of the flowers standing below (c/. fig. on p. 742, vol. i.). In Composites, Cam- panulas, and several Stellatse, the pollen is stored on the style or stigma, but not, as was previously supposed, upon the receptive portions of this organ. On the contrary, it is retained here by various hairs and papillae, specially designed for the purpose. Then, in the Proteaceae again, the pollen is deposited, whilst the flower is still in bud, upon the summit of the stigma, without, however, coming into contact with the receptive spot ; the stigma in this case serves, at the commencement of flowering, as a temporary depot for the pollen. In Sarracenia the pollen falls upon the stigma, which has the form of an expanded umbrella, and here for a while it remains, but not in contact with the receptive points. We Bhall hardly overstep the mark in saying that in some 20,000 species of plants POLLEN. 95 the pollen is temporarily stored in some portion of the flower and preserved for future use. More frequently, however, the pollen remains within the opened anther. Usually these flowers are visited by insects which disturb the anthers and release the pollen, or they dust themselves over with it and carry it off" to another flower. The fact that the anthers are directed sometimes inwards, sometimes outwards, is correlated with these insect- visits. Where the slits or pores of the anthers are directed towards the periphery of the flower, one speaks of outwardly-directed anthers (extrorse), where toward the centre of the flower, of inwardly-directed (introrse). These relations are of importance in respect of insect-visits. If the honey is situated outside the whorl of stamens, the insects must pass between the stamens and petals to secure the nectar, as in Colchicuvi, Iris, Convolvulus, Epimedium, and Laurus. Here it will be advantageous for the anthers to be extrorse. On the contrary, when the honey is between the ovary and the bases of the stamens, and the insects have to penetrate to this region, as in Gentians and Opuntias, the stamens will be introrse. It is of great importance that the pollen exposed in the anthers should be rubbed oflT by the insects and carried to other flowers, a result only obtainable when the dehiscent side of the anther is placed in the way of the insect as it enters or leaves the flower. Numerous other peculiarities aflfecting the structure, position, and movements of stamens will be dealt with later on, when treating of the removal of pollen from and to flowers by insects and other animals. POLLEN. Like all other leaf-structures, stamens arise in the first instance as convex projections from their points of insertion on the stem. These projections consist of a homogeneous, small-celled tissue. They soon, however, assume a club-shaped form, and the outlines of anther and filament become recognizable. A vascular bundle is found traversing the entire length of each stamen, and the anther, which increases in size more rapidly than the filament, shows symmetrically-arranged, longitudinal grooves, with projecting portions between, arranged in pairs. The cells situated immediately below the surface of the young anthers become now marked out into tissues of two kinds. Towards the outside three layers of cells become distinguishable, and these, with the outermost, enveloping layer give rise to the wall of the anther; within, large cells become conspicuous, and form what is known as the archesporiurri. These archesporial cells are arranged either in nests or in longitudinal rows embedded in the surrounding tissue. In the latter, the more usual, case, there are four, rarely two or eight, such rows arranged in pairs right and left of the central vascular bundle. Although at this stage of development all the cells of the anther hang together into a continuous mass, the existence of the future pollen-sacs — now POLLEN. filled with the archesporial cells — is easily recognized. As time goes on the contrast between the wall of the anther and the contents of the chambers becomes more pronounced. The archesporial cells divide, giving rise to the pollen-mother- cells which entirely fill the pollen-sacs. Of the layers of the anther-wall, the inmost is usually dissolved, so that the mother-cells are bathed in a fluid mucilage; thus the wall comes to consist solely of the outmost, enveloping layer and of the contractile cells ("fibrous layer") within. Changes continue in the chambers or pollen-sacs, and in the partition-walls between them. The walls of the pollen-mother-cells become thickened, and often show a stratification. The protoplasm within divides into four parts, arranged frequently, though not invariably, in the corners of a 4-sided pyramid (i.e. in tetrads). Each of these cells becomes invested with a wall of its own, at first thin and delicate, but afterwards thickened and stratified. These are the pollen-grains. Their protoplasm possesses the property of a fertilizing agent, and is termed the Spermatoplasm. In most plants a further division of the protoplasm in the pollen-cells takes place. This is conspicuous in the Conifers and Cycads, but relatively obscure in the majority of flowering plants. Of the two or more cells thus arising within the pollen-grain one only takes an actual part in fertilization. How long the spermatoplasm retains its fertilizing properties unimpaired has not been sufficiently investigated. It has been stated of the plants enumerated below that this property is lost as follows: — In Hibiscus Trionum after 3 days. In the Larger Periwinkle ( Fi'nca „ The Wallflower (C/^eM-a?i BtiplUhalmiim grandiflorum. '' Uibiscus tsrnatus. » Malva rotundifolia. '^ Campanula persieijolia ; x200. poles are distinguishable. The number of the grooves is constant for a given species, and even for whole families of plants. A single furrow is characteristic of the grains of the Tulip-tree, Magnolias, and Water Lilies (fig. 218^), of the Meadow Saflfron, Tulip, Lily, Iris, Narcissus, and Snowdrop, of Palms, Grasses, and, indeed. Monocotyledons generally. Two furrows are found on the pollen-grains of Calycanthus, several climbing Smilacinese (Tamus, Dioscorea), and several species of Amaryllis. A very great number of plants have three grooves, e.g. Rock-roses, Violets, Poppies, Ranunculacese, Roses, Almonds, many Papilionacese, Beeches, Oaks, and Willows, Solanaceae, Gentians, Scrophulariaceae, and many Composites (cf. figs. 217^^ and 218 2). Four grooves have been noticed in several Boragineae (Anchusa, Nonnea), some Labiates {Teucrium montanum, Sideritis scordioides), in Houstonia, Platonia, Blackwellia and Cedrela odorata; six in most Labiates (fig. 217^*), nine or ten in Sherardia, Borago, and Sym,phytu7n; twelve in Crucianella latifolia ; sixteen in Polygala chamcebuxus ; twenty-one to twenty- three in Polygala myrtifolia. On crystal-like pollen-grains the grooves are ex'remely delicate, and their number depends on the number of angular ridges. 100 POLLEN. A very conspicuous feature of many pollen-grains is the infinitely varied sculpturing, &c., of their walls. Sometimes this takes the form of a delicate dotting of the wall, as in Asarum, Meadow Saffron, Rue, Salvia, Gentians, and Euphorbias, many Aroids and Musacese (cf. figs. 217 ^^ and 217 ^■*); or the projecting ridges may be transversely striated as in Saxifraga aizoides; or, again, delicate striations may run in meridian-like circles (e.g. Brugmansia arborea). Sometimes dotted lines are found arranged in various ornamental reticulating patterns. On the smooth surface of the grains of Thesium alijinum and rostratuTn reticulations occur, and in the centre of each mesh a distinct dot. Similiarly in Thrift and Sea Lavender {Armeria and Statice), and in the Corn Cockle {Agrostemma Githago). Often the surface presents considerable unevenness. In Cuphea platycentra the outer coat is prettily ridged, whilst in many other cases it is finely granulated. The little projecting granules may be either scattered equally over the whole surface, or they may be arranged in networks — which is specially the case in Cruciferse {Capaella, Raplianus, Sinapis). In the Passion Flowers (e.g. Fassifiora Kermesina, fig. 217 *) these networks are inclosed in shallow, ring-like depressions, whilst in Cobcea scandens (fig. 217 ^) the surface has a honey -combed appearance. Sometimes the whole surface is dotted over with little wart-like projections, as in Centaurea Jacea, Mistletoe {Viscum album), White Water Lily {Nympltoba alba), and the tropical Bauhinias (Bauhinia armata, furcata, cf. figs. 218 ^ and 218 -); or it may be covered with sharp, needle-like prickles, as in the pollen-grains of Composites, Scabiouses, Campanulas, Cucurbitacese, Malvaceae; also in the genera Armeria, Amaryllis, Canna, Lonicera, Ipomcea, and Convolvulus (cf. figs. 218^ and 218^). It is only the superficial layer of the pollen-grain which shows these sculpturings and projections, the inner layer, which abuts immediately upon the protoplasm, is homogeneous. The wall of pollen-grains is, as a rule, three-layered. These three layers are: — the internal one or inline, the middle one or extine, and the external one or perine. The extine and intine arise from the protoplasm of the pollen-cell itself; the perine, on the other hand, is deposited from the matrix in which the young pollen-grains lie embedded. It comes about in this way. The young grains first clothe themselves with delicate walls, which are in due time thickened. This is the extine. Within this they form a second layer, the intine. Lastly, the perine is deposited upon them from without. The intine and extine can generally be readily distinguished as separate layers, but between the extine and perine the boundary is by no means so well marked. The various sculpturings, prickles, and other unevennesses of the outer coat really appertain to the perine. It sometimes happens at definite spots on the wall of a pollen-grain, from a separation of the molecules there, that little spaces or actual canals arise which open externally by tiny pores. This may be well observed in Thesium, Prunella, I])omcea, and Gentiana. In these canals a yellow (rarely colourless) oil is con- tained, which oozes out in the form of minute drops when the grains are moistened and absorb water. Such at any rate is the behaviour in Prunella grandifiora and Gentiana ciliata. In many other plants the whole .surface of the grain is saturated POLLEN. 101 with this oil, I ascertained that in about 400 out of 520 species investigated by ine the outer surface was overlaid with oil. The layer is so thin that with dry pollen-grains it is not visible, but when they are placed in water, the coating is resolved into a number of minute, strongly -refringent droplets, which adhere to the swollen pollen-grains like tiny beads. There is no doubt that this coat consists of a fatty oil, since it is soluble in alcohol and olive-oil, and with osmic acid it turns dark-coloured and becomes congealed. More rarely are pollen-grains found with masses of a sticky, structureless substance adhering to them. This substance does not form droplets with water, nor does it dissolve in alcohol and olive -oil. It may be termed Viscin, from the similarity which it presents tc the bird-lime obtained from the berries of the 1 (k ^^-# Fig. 219.— Pollen-grains and pollen-tetrads united by threads of viscin. K ^ Rhododendron hirsutum. ^ CEnothera biennis. * Epilobium angustifoliiim. ixS; --*x50. Mistletoe ( YiscuTn). Such a viscin is met with on the surface of the pollen-grains of Fuclisia, ClarJda, Circcea, Gaura, Godetia, (Enothera, Epilobium — indeed, throughout Onagracege and in Azaleas, Rhododendrons, Orchids, and Asclepiads. It is very sticky, and on the slightest touch can be drawn out into delicate threads. The contents of the anthers, as they escape, in the Evening Primrose (GEnothera) and Willow-herb {Epilubium angustifoliura) resemble fringes and tattered ribbons, or a broken net hanging from the adjacent anthers. Under the microscope this sub- stance is seen to consist of pollen-grains, joined together by the sticky strings of viscin (fig. 219 ^ and 219 *). The phenomenon is even more striking in the numerous species of Rhododendron. In Rhododendron hirsutum all the pollen-tetrads of an anthei-- cavity are held together by a mass of sticky viscin. The anther dehisces by two terminal pores, and from these the pollen-tetrads ooze out to some extent. If the sticky mass be touched with a bristle it adheres, and the whole contents of the anther can be readily withdrawn (fig. 219^). Its appearance under the microscope is shown in fig. 219 ^. In many species, as for instance in the elegant Rhododendron Chamcecistus of the Northern Limestone Alps, and in the large-flowered Himalayan 102 POLLEN. species, strings and filaments are woven from the anthers a centimetre long, and insects visiting the flowers touch the strings, stick to them, and carry off with them to another flower generally the whole of the contents of the anther in question. The sticky substance is probably a mucilage formed from the outer wall of the pollen-tetrad, or from the broken-down walls of the mother-cells. Not to be confused with the little pores communicating with the canals containing the oil are the thin portions of the outer layer, into which the intine projects, caecum- like, as it swells up in water. It often looks as though the extine were actually perforated at these places; this, howe\'er, is not the case, and it is not till later, when the intine pushes through and the pollen-tube is formed, that these places are burst and true apertures arise. The variety exhibited by these spots is as remarkable as that of the sculpturings. The outmost layer often thins out at those spots where the wall is grooved. As the grain swells up in water, the extine often bursts at the thin region, and may actually peel oW {cf. fig. 217^°). In Mimulus and Thunbergia the thin region has the form of a spiral, or it may run into loops and convolutions, as shown in 217^. When the intine swells up and bursts the extine, the pollen-grain looks as though it had been pared. In the Passion-Flower the thin places are ring-like, so that with the swelling of the intine, the extine comes away in the form of little lids. The same thing happens in the Gourd, where the lids are very small, and are provided each with a little spine (fig. 217 ^). A curious condition obtains in Morina Persica (allied to the Teasel, cf. fig. 217 ^). Each of the pollen-grains has at its equator three projections, resembling closed bottle-necks with swollen, circular mouths. Very frequently the thin places are disc-like, and may be compared to the glazed port-holes of a ship. It is especially this form whicn suggests that the outmost layer of the wall is perforated from the beginning. In Umbelliferse, Rosacese, Papilionacese, Violets, Rutacese, Hypericinese, Scrophulariacese, and other groups of plants, the little circular windows lie hidden in the grooves; in Cohcca scandens (217^) they are in the "cells" of the honey-comb, and in Onagracea?, e.g. Enchanter's Nightshade (Circcea), the outer coat is continued as a thin invest- ment over the tops of the projecting warts (fig. 217 ®). The number of windows varies from plant to plant. Cyperaceae have one; Bromeliacese and the Meadow Saffron, Figs and Brugmansias two; Nettles, Oaks, and Beeches, Evening Primroses and Willow-herbs, and many other plants three; Alders and Birches four to six; Currants eight to twelve; Convolvuluses fifteen to eighteen; Carnations, Oraches, and Mezereons twenty to thirty; and Nyctaginese over thirty. Having concluded the description of the walls of pollen-grains, the question arises, for what purposes are all these remarkable structures, these grooves and striae, these chinks and furrows, thorns and spines developed? What is the meaning of the coats of oil and viscin? What of the thin places, and windows, and tiny lids? Of these the last question is the easiest to answer. As observation shows, pollen-grains swell up with lightning rapidity when they are placed in water. The POLLEN. 103 protoplasm within, destined for fertilization, takes up water from the environment very quickly and energetically. In consequence it swells rapidly, and must have an inclosing wall which will not impede its rapid stretching. For this purpose the thin places and folds are admirably suited. Through them fluids readily pass to the interior, and simultaneously the grooves (previously folded inwards) become inflated, and the pollen-grains come to occupy two to four times the space they previously did. The thicker portions, saturated with oil, play a purely passive role in these events. Water cannot enter by these parts, nor do they stretch with the swelling up inside. Later, when the intine has grown out and assumed the form of a tube, the outer wall is not essentially altered. The thin spots have been raptured, and where lids are present, they are raised; the protoplast, enveloped in the tube-like intine, vacates the extine by one of the thin spots, much as a germinating embryo does its seed -coat. Just as it is of advantage in germination for the seed-coat to be fixed on the substratum, whilst the young plant gets a good hold of the ground, so here it is of value to the young pollen-tube as it quits the extine of the pollen-grain that the coat should be fixed firmly; for this purpose the various ridges, teeth, and spines possess a high significance, serving as a means of anchoring the pollen-grain whilst the pollen-tube is being formed. But the most important service rendered by the sculpturings and inequalities of the walls consists in the fact that thereby considerable quantities of pollen-grains are enabled to cohere in crumbling masses to the slits of the opened anthers, and to become attached to insects and other animals visiting the flowers for food. Con- trasting with this clinging pollen is the already-mentioned dusty pollen, with smooth and non-adhesive surface. Dusty pollen does not cohere in clusters, nor does it readily attach itself to foreign bodies. On the other hand, the least dis- turbance or breath of air carries it away in clouds It is sufficiently obvious that globular or ellipsoidal pollen-grains with smooth surfaces will be distributed in the form of dust more readily than grains possessing an angular or crystalline form. The former have a smaller surface of contact than the latter. When the surface is, in addition, variously sculptured and raised into folds and inequalities, the points of contact are of course enormously increased- The little projections of the surfaces of adjacent grains interlock like the wheels of a watch; longer ones become entwined like fingers; thus it comes to pass that hundreds of neighbouring pollen-grains hang together like burs. That such masses will readily attach themselves to the hairs, bristles, probosces, and legs of insects hardly needs further demonstration. This capacity for clinging is much increased when the surfaces of the grains are [saturated with oil. The sticky property of the viscin has been already enlarged upon. We may thus summarize the whole matter in the statement that the crystalline forms, the various sculpturings, spines, and other projections, as well as the presence of oil and viscin on the surface are arrangements in virtue of which the adhesiveness of the pollen-grains is increased. According as one or other of these arrangements is present or absent we find lOi PROTECTION OF POLLEN. every degree of cohesiveness in pollen — dusty, floury, crumbly, clotted, glutinous, waxy. A marked contrast is noticeable between flowers the anthers of which produce dusty, and those which produce coherent pollen. So pronounced is this, that we shall treat of the pollination of these flowers, and in particular of the transmission of the pollen from flower to flower, under separate headings. Here it need only be added that this distinction between dusty and coherent pollen is found not only with isolated pollen-grains but with tetrads. When the stamens of Heaths (Erica) are disturbed the pollen escapes in clouds of dust, just as it does from the catkins of the Hazel. This dust, however, consists, not of isolated pollen- cells, but of tetrads. In Azaleas and Rhododendrons, on the other hand, the pollen-tetrads cling together into sticky filaments, just as do the isolated grains of the Evening Primrose and Willow-herb. Why it is that the pollen is in some cases in tetrads and in others in isolated grains, why its adhesiveness is promoted by such various means as those enumerated, is diflttcult to say. These differences are perhaps connected with the varying form of the insect-visitors which carry the pollen away, and of the stigmas upon whicli it is deposited. That the sculpturings protect the pollen against untimely wettino- will be shown in the following chapter. PROTECTION OF POLLEN. The approach to Venice from the mainland is by a long embankment, on either side of which the traveller commands an endless vista of marshes full of reeds and rushes broken here and there by expanses of brackish water — the famous lagoons — which themselves exhibit a luxuriant vegetation consisting principally of Pond- weeds and Naiadacese. One plant in particular, the Grass-wrack (Zosfera), is conspicuous for its abundance in the lagoons, covering, as it does, extensive tracts of the sandy mud at the bottom of the shallow water. The leaves are submerged, ribbon-shaped, and of a brownish-green colour somewhat resembling sea-weed, and, when collected and dried, they are known in commerce by the name of " Sea-grass", and are used in the packing of glass, and of late years also for stuffing mattresses and cushions. These Grass-wracks, of which there are two known species, differ so greatly from other Phanerogams, not only in appearance, but also in development and in the mode of pollination, that one might almost be induced to assign to them and their immediate allies a special class, were it not that the fact of the existence of numerous intermediate forms and connecting links tells against their isolation. In tlie first place, the pollen in Zostera does not possess the outer coat which is so characteristic of the cell-membranes of most pollen-cells. Further, from the moment the pollen-cells are set free by the opening of the anthers — an event which occurs under water — they exhibit the form of elongated cylindrical tubes. In the plants most nearly related to the Grass-wi acks, namely, the genera Posidonia and C ijnwdocea, some species of which grow in brackish and some in salt water, the long hypha-like pollen-cells lie in complicated coils and curves within the anther, and PROTECTION OF POLLEN. XQo when they escape from it, and are carried by the water against the long fiUform stigmas they adhere to those structures as do the spermatozoids (spermatia) to the trichogyne in the Red Sea- weeds (c/. pp. 60, 61). The filamentous pollen of Halo- phila is even divided by transverse septa into several chambers, or, more accurately, the pollen-cells are aggregated into long strings. The pollen-cells are intercepted under water by the filiform stigmas and grow down them into the ovaries. In the difierent species of Naias as also in those of Zannichellia the pollen-cells are spherical or ellipsoidal in shape so long as they are inclosed in the anther, but when the anther opens they assume the form of tubes, and are wafted about by the water until they reach the stigmas. The stigma in Zannichellia is triangular and com- paratively large, and owing to the fact that three or four such stigmas have their edges in contact, a sort of funnel is formed, which serves to collect the pollen-cells as they float about. The plants above referred to, about fifty species in all, were classed together by the older botanists under the name of Kaiadece, but are now grouped into the families of the Potamogetonacege, Naiadacese, and Hydrocharidacese. They are all aquatic plants, but it would be erroneous to suppose that all the members of these groups possess the same kind of pollen as is exhibited by the Grass-wracks, and the various species of Halophila, Posidonia, Cyviodocea, Naias, and Zanni- chellia, that is to say, a filamentous pollen destitute of external coat which is con- veyed to its destination by currents of water. On the contrary, thousands of aquatic plants discharge their pollen above the surface of the water and not beneath it. The pollen-cells are spherical or ellipsoidal, have a distinct external coat, and are transported to the stigmas not by flowing water but by the wind or by insects. This is the case even in plants whose leafy parts remain under water throughout their lives. Aldrovandia, Hottonia, and Utricularia, many Pond- weeds {Pota- mogeton) and Water-crowfoots (Ranunculus), not to mention many others, always raise their flowers above the surface of the water, so that the pollen may escape into the air and be blown or otherwise conveyed from one flower to another. I have observed that even in the case of the various species of Water-starwort {Calli- triche), which were formerly said to accomplish their fertilization under water, the ant/hers open only in the air, and that the staminal filaments grow in length accord- ing to circumstances until the anthers project above the surface. If they fail to do so, then the anthers of the flowers in question do not open at all; the spherical pollen remains inclosed and decays, together with the anther and its filament, beneath the water. The far-famed Vallisneria (see vol. i. p. 667), too, to which we shall return again later on, only emits the pollen from its anthers into the air. jThe staminiferous buds, it is true, develop under water; but they detach themselves from the axis of the inflorescence in the form of closed bladders, and do not open until they reach the surface. The stamens then project out of the floating flowers into the air, the anthers burst, and the pollen is set free (c/. fig. 227). If the buds are kept submerged artificially, neither they nor the anthers open, but they decay, and the pollen perishes under the water. And, as in the case of these aquatic 106 PROTECTION OF POLLEN. plants, so also in that of the multitude of plants which germinate and flower on dry land, if the pollen happens to fall into the water or is purposely kept immersed, it is destroyed. It is thus the fact that the pollen of Phanerogams, with the exception of about fifty species, of which the Grass-wracks may be taken to be the type, is injured by prolonged immersion or subaqueous transport. This obviously suggests an inquiry as to the reason of the hurtful action of water upon cells which require an especial abundance of liquid materials for the development of the pollen-tubes. There is, however, a great difference between the absorption of pure water and the absorption of the liquid substances yielded by stigmas. A pollen-cell deposited upon a stigma gradually takes up the liquids there available, and the pollen-tube pushes out comparatively slowly. If, on the other hand, the pollen-cell is put into water, or is in nature so wetted by rain or dew as to be practically immersed in a water-bath, absorption of water takes place almost instantaneously; the intine is pushed out wherever no resistance is offered by the extine, and in a moment the pollen-cell swells up. Such a process cannot properly be called a development of the pollen-tube. No real growth can take place in so short a time, and what has occurred is simply an expansion of the intine and a smoothing out of the folds which have hitherto lain tucked in. Frequently, indeed, the limits of elasticity art exceeded; the projecting part of the intine bursts, and the spermatoplasm flows out into the water in the form of a finely granulated, slimy mass. In that event the pollen-cell is destroyed, and comes to nothing. But even if the intine does not burst, the pollen undergoes such complete alteration through the rapid absorption of water that its protoplasm loses the power of fertilization. It seems as if the protoplasts inclosed in pollen-cells, subjected to prolonged immersion, were literally drowned. Thus much is certain, that the immense majority of pollen-cells perish under water, and that even if wetted they incur great risk of destruction. This danger, which may be of daily occurrence in case of rain or heavy dew, has to be avoided. In order to preserve the pollen fit for use it must be secured by protective apparatus against the injurious effects of moisture, especially against atmospheric deposits; it must be able to develop under conditions from which this factor — in so far as it is harmful — is, generally speaking, excluded. In regions where there is a regular alternation of rainy and rainless seasons— in the llanos of Venezuela, the Brazilian campos, the dry districts of India and the Soudan, above all, in the parts of Australia to the south of the tropic where the rainfall is limited to the winter and afterwards ceases for months — the climate itself indirectly affords security to the pollen against risk from water; or, in other words, any apparatus to protect from rain the pollen of plants which flower in rainless seasons would be superfluous. The trees which wave above the grass of the wonderful savannahs of Australia, as also the numerous dry and rigid shrubs which belong to the adjacent "scrub", do not flower until the rainy season is over, when the flowers do not run any risk of being drenched with rain. In the absence of the danger the necessity for any direct means of defence against it also PROTECTION OF POLLEN. 107 disappears, and in Australia the numerous Mimosese and Myrtacese and the Proteaceae, which constitute the principal part of the dense copses just referred to, are accordingly destitute of any contrivance capable of acting as a protection to the pollen. These plants preserve their rigid character even during the flowering season; the filaments bearing the anthers project in large numbers far beyond the small floral envelopes in the Acacias and in the innumerable species of Callistemon, Melaleuca, Eucalyptus, Calothamnus, and Metrosideros, and the styliform prolongations of the ovaries in Proteaceae, on the top of which the pollen is deposited when set free from the anthers, spring up and stretch out unprotected far beyond the restricted perianth. Flowers which inhabit a region where moisture is deposited from the atmo- sphere in greatest quantity in the flowering season exhibit an entirely dififerent form. In the mountains of Central and Southern Europe, where this coincidence occurs, the plants whilst in flower must be prepared for daily showers. In addition every plant drips with dew in the early morning, and drops of water are deposited on leaves and flowers in the course of the day by the mists as they roll by. It must often happen that the pollen remains for days together hanging to the opened anthers before it is carried away by bees or butterflies to the stigmas of other flowers. Here if anywhere is an instance of the necessity of ample shelter for the pollen. Examine the plants composing the smaller brushwood of such a region, and you will find how great a contrast they aflTord to the plants of the thickets of Australia. The flowers of the Heather (Calluna vulgaris), and of the Bilberry, Bog Whortleberry, and Cowberry {Vaccinium Myrtillus, V. uliginosum, V. Vitis-Idcea) have bell- or cup-shaped corollas which hang down from curved stalks with the mouths of the flowers towards the earth, and so cover the pollen-laden anthers. Similarly, we find the Alpine Rhododendrons ("Alpine Roses "), which clothe the mountain sides, with flowers inclined at a right angle to the erect stalks so that the anthers are perfectly sheltered beneath their tilted bells. All the many contrivances whereby pollen is directly protected from wet are of the same nature as the above, the method of protection being by some such roofing in or envelopment of the anthers. That these adaptations should exhibit so much variety in detail in spite of the uniformity of their object is due to the condition that the envelopment must itself not be carried too far. On no account must the dissemination of the pollen or its transport by wind or insects to the stigmas of other flowers be hindered; nay, the very same parts of a flower which shelter the pollen from rain frequently have the additional function of assisting I the dispersion of the pollen when the rain is over. ' In the enumeration of arrangements for warding oflf injury to pollen from i wetting, the various coverings and protections are described as equally effective for rain as for dew. But this, of course, is not for the same reason. A roof protects structures from rain by intercepting the drops, and from being bedewed since it diminishes radiation from the bodies beneath and thus keeps them at a 108 PROTECTION OF POLLEN. higher temperature than would otherwise have been the ease. This explanation must be borne in mind. We find, therefore, an amount of variety in the forms of safeguard against wet corresponding to the multiplicity of the adaptations which subserve the purpose of pollen-transport by the wind or by butterflies, bees, beetles, or flies, as the case may be. The means of protection are diversified also according to whether the cover is placed immediately over the pollen or over an entire group of flowers, whether it shelters the newly-opened, pollen-laden anthers or that part of the flower whereon pollen liberated from the anthers is temporarily deposited, and again they vary according as it is the anther-walls, stigmas, petals, involucre, or foliage-leaves which have to serve as roof to the pollen. The Lime-tree affords an instance of the last-mentioned arrangement, its flowers being invariably so placed that at the time when pollen is yielded by the anthers they are covered by the broad, flat foliage-leaves. However sharp the showers to which a Lime-tree is subjected the rain-drops roll off" the blades of the leaves, and it is only by exception that any one of the many flowers stationed beneath them is wetted. The same provision is met with in a few species of Daphne (e.g. D. Laureola and D. Philippi), in several Malvaceae (e.g. Althcea pallida and A. rosea), and in the Iinnpatiens Nolitangere, a plant which possesses other remarkable features and will be the subject of further discussion by and by (c/. fig. 220 ^). In Impatiens the flower-buds are held by their delicate stalks above the surfaces of the leaves from whose axils they spring, and the leaves are at first folded upwards like erect troughs. Subsequently, when the buds get bigger and their stalks longer, the latter slip down to one side of the leaves and hide beneath them, whilst the leaf-margins still continue to be curved upward. The leaf then flattens itself out and fixes the drooping flower-stalk by means of one of the lobes of its heart-shaped base, and thus indirectly keeps the suspended bud in position, so that when later on the bud and its anthers open, which they do simultaneously, they are roofed over by a smooth lamina, off" which the rain-drops roll without ever wetting the flowers or their pollen (fig. 220 ^). In many Aroideae the spadix is completely covered by the large sheathin^^ leaf or spathe at the time when the anthers burst, as, for instance, in the curious Japanese Arisema ringens, where the spathe curves over the inflorescence like a Phrygian cap, and in Ariopsis peltata, where the spadix is protected fi-oni rain and dew by a sheathing leaf resembling a boat with the keel uppermost (cf. fig. 221^). Genetyllis tulipifera, a shrub belonging to the Myrtacese, bears at the ends of slender, woody twigs inflorescences which at first sight might be taken to be pendent tulips. On closer inspection it appears that the large white leaves with red veins which recall the leaves of the tulip perianth are involucral bracts which cover the closely-crowded flowers and shield them from the rain. Similarly in the case of the Banana and its allies {Musa, Ravenala) the flowers are covered over when the pollen is mature by large involucral sheaths which ■subsequently, after the pollen has been used up and there is no longer any neefl PROTECTION OF POLLEN. 109 for protection, detach themselves and drop to the ground. Fig. 220 ^ shows the male flowers of the dioecious Sea-Buckthorn (Hi^ypojohae rhamnoides), which are arranged in spikes and are seated in the axils of scaly bracts at the bases of the I young lateral shoots. In each flower are four anthers which discharge their 1 abundant powdery pollen whilst the flower is still closed like a bud and has the appearance of a little bladder (fig. 220 ^). This pollen is of an orange colour, and drops to the bottom of the flower, where it remains (figs. 220 * and 220 ^) awaiting ia dry wind to transport it to the stigmas of the female flowers growing on other plants often at a considerable distance. Several days may go by before Impatiens Nolitangerf. 2-5 Ilippophae rhamnoides , 2, 8, T, 8 natural size; 3. *, = slightly magnified. Fig. 220.— Protection of Pollen from Wet. Coiivaltaria ynajalis. ' Euphrasia offic ' Iris sibirica. his kind of wind sets in, and meanwhile there is the danger of the store of ollen being soaked by rain or dew and rendered unfit for dispersion. To obviate his risk the pair of curved perianth leaves, which have their concave surfaces limed towards one another, and form, as has been already mentioned, a kind f bladder inclosing the anthers and pollen, dehisce at the sides only. Thus two pposite gaps (figs. 220 ^ and 220 ^) are produced, whilst at the top the two valves 3main joined together and form an arch completely sheltering the mass of fallen ollen from atmospheric deposits. When the needful wind arises it blows the ollen out through the chinks in the bladder and conveys it to the stigmas of fcher plants of the same species. Plants of the Globe-flower (TroUius) genus, whose species grow in the Arctic igions in damp situations and also further south in mountainous districts of tl.e 110 PROTECTION OF POLLEN. Old World, are daily exposed to rain or heavy dew. Nevertheless their pollen is never wetted, the anthers being completely shut in by the perianth-leaves, which are spirally inserted on the receptacle and closely furled one upon another. These flowers have a ring of stalked nectaries round the stamens, and insects which visit them for the sake of the honey are obliged to break through the roof formed by the overlapping perianth-leaves in order to reach the inside of the flower. The pliability of these leaves enables bees by their weight to effect an entrance, whilst falling drops of rain cannot penetrate, but roll off" the flower. Fig. 221.— Protection of Pollen from Wet. « Ariopsis peltata. 2 Flower of Trolling europmus. « The same with some of the floral-leaves removed. ♦ Digitalis luUscens. 6 A single flower of Digitalis hUescem in longitudinal section. « Aretia glacialis. ' Single flower of Aretia glacialis in longitudinal section (magnified). Also in Corydalis, Calceolarias, Toad-Flax and Snap-dragon (Corydalis, Calceolaria, Linaria, Antirrhinum) the corolla forms a closed envelope round the anthers; and again in papilionaceous flowers the pollen is, up to the moment of an insect's visit, hidden in the cavity formed by the two petals of the keel. \y The majority of lipped flowers — Butterwort, Yellow-rattle, Cow- wheat, and Eye-bright (Pinguicula, Rhinanthus, Melampyrum, Euphrasia, cf. fig. 220^)— as also the Violet {Viola), Monkshood (Aconitum), and innumerable other plants whose flowers open laterally, do not regularly inclose the pollen, but protect it against rain or dew by means of an arched portion of the flower which forms a roof over it. In Acanthus the flowers are inclined laterally, and, though PROTECTION OF POLLEN. HI resembling bi- labiate flowers in general appearance, possess no prominent upper lip, the protection of the pollen being effected by a sepal which stretches out at the place where the upper lip would be. A curious arrangement for the protection of pollen by sepals may be observed in the inflorescence of Hydrangea querci- folia (fig. 222^), a native of Florida allied to the Hortensias. The flowers of this plant grow in handsome bunches, and are of two kinds: the one kind includes stamens and pistil, but only a very small, greenish perianth incapable of shielding the pollen of the adjoining stamens from rain or dew; the other has neither stamens nor pistil, but has very large, white, expanded sepals which are arranged so as to constitute with their erect stalks a sort of umbrella. The flowers of the latter type are developed on the outermost and uppermost branches of the inflorescence, and are always in a position to stop the rain from falling upon the umbels of small pollen-bearing flowers which are situated underneath them. In rare cases the stigmas act as pollen-protectors. The most striking instance is that of the genus Iris. The stigmas in the Iris are petaloid, and consist of three foliaceous structures gently curved outwards, and each terminating in a pair of dentate apices (c/. fig. 220^). The upper surfaces of these foliaceous stigmas are convex and usually somewhat keeled along the middle line, the under surfaces are concave. Beneath each stigma one finds a pollen-laden anther nestling close against the concave surface, and so perfectly concealed that it is impossible that it should ever be reached by a drop of water however heavy the rain. Flowers of the form called " hypocraterif orm " by botanists are adapted to the protection of their pollen on an essentially different principle. The species of Phlox and Daphne included in this category, the delicate species of Primulaceae belonging to the genera Androsace and Aretia, which dwell amid mountain-mists, and the pretty, erect-flowered Primulas (e.g. Primula farinosa, P. denticulata, P. Cash- miriana), all bear flowers which are not roofed in, but have the mouths of their corollas open to the sky, the tubular part of the corolla passing abruptly into an expanded limb (c/. figs. 221 ^ and 221 ''X so that drops of rain or dew collect on the Hmb surrounding the mouth of the tube. Here it seems inevitable that some drops of water should reach the anthers inserted in the tube. Yet, as a matter of fact, the pollen is kept dry. For, at the place where the tube passes into the limb of the corolla it is abruptly contracted, besides being often also studded with callosities, in consequence of which the opening is so narrowed that, although insects with fine probosces gain access to suck the honey in the flower, any rain-drops that may happen to be lying upon the limb do not gain admission because the air cannot escape from the tube. If flowers of Aretia glacialis (fig. 221^), a plant growing on the moraines of glaciers, are examined after a shower, it is found that every one has a drop resting upon it whicli slightly compresses the air in the narrow tube of the corolla, but cannot reach the pollen upon the anthers lower down the tube. A subsecjuent shake or puff" 112 PROTECTION OF POLLEN. of wind causes the drops to roll off the limb of the corolla, or else thej^ art- got rid of by evaporation; in either case, the flower becomes once more accessible to insects. In none of the instances hitherto described does any change take place in the relative positions of the foliage-leaves, petals, or petaloid stigmas, whereby the pollen shall be the better protected. On the other hand, there is a long list of plants wherein the protection of the pollen is effected exclusively by means of changes in the position of some one or other of the leaves in question. This occurs especially in all those species which, like the forms last mentioned. fig. 22 -Protection of Pollen from Rain. ' Flower of Eschscholtzia Califomiea opened in the sunshine. 2 xhe same closed in wet weather. 8 Floral capitulnm of Uieracium Pilosella, closed. * Single flower of the same plant. * Capitulum of tlie same, open. ^ Longitudinal section tlnough a closed capitulum of Catananche ccerulea. ? Single flower taken from the capitulum in the last stage of flowering. 8 Poition of inflorescence of Uydrangea querci/olia. » Young closed flower of Eranthis hiemalis. 10 Old closed flower of the same. have the mouths of their flowers exposed to the incidence of rain, or unshielded, so that radiation is not diminished and dew is formed, but, unlike them, exhibit no sufficient constriction of the tubular part of the corolla to prevent drops of Avater from falling into the flowers. Such unconstricted, cup-shaped, urccolate, infundibular, or tubular flowers would, if upright, constitute regular rain-collectors, and the water would at once saturate the pollen within the flowers. If flowers of the kind close up temporarily and keep their petals or involucral leaves arched over the interior so long as there is any risk of water collecting there, the requisite security from inundation is attained by very simple means. As a matter of fact, protection of pollen is effected in numerous cases by the closing of flowers. Examples of this are afforded by the flowers of Meadow Saffron, Sternbergias, and Crocuses (Colchicuvi, Sternhergia, Crocus, cf. fig. 228), PROTECTION OF POLLEN. 113 which lift the cup-shaped limbs of their corollas above the ground in the spring or late autumn, the Gentians of Alpine meadows and their allies of the Centaury genus (Erythrcea), a host of Bell-flow^ers wnth erect blossoms (Campamila glomerata, C. spicata, C. Trachelium, Specularia Speculityin, &c.), the Peonies, Roses, Flaxes, Opuntias, Mammillarias and Mesembryanthemums, numerous species of the Star of Bethlehem, and Thorn-apple genera (e.g. Ornitho- galum umbellatum, Mandragora veymalis, Datura Stramonium). The floating flowers of the Water-lily (Nymphcea), and the large flowers which are borne I % Fig. 223.— Protection of Pollen. Flowers of Crocus multifidus. On the right, flowers open in the sunshine ; on the left, flowers closed at night or In wet weather. One of the three closed flowers has some of its perianth-leaves removed. an the branches of Magnolias also belong to this group of forms. Throughout the day when the sun is shining the floral cups or funnels of these plants are kvide open and often even expanded into stars, whilst swarms of insects hover bound them; but at dusk when the dew "falls" the petals close up again and pverlap one another so as to form a case (cf. fig. 223) upon which any amount f dew may be deposited without affecting the interior of the cup. In damp )r rainy weather these flowers do not as a rule open. Thus the period of their )eing closed coincides with a time when most honey-seeking insects are absent, laving either gone to rest for the night, or retired to their hiding-places for helter from the wet. It is a very interesting phenomenon that petals which close over the anthers Vol. II. 58 114 PROTECTION OF POLLEN. in the evening grow much larger in the course of the flowering period. In many species they become double as long as they were at the moment the flower first opened. The enlargement of the petals takes place pari passu with certain pro- cesses in the development of the anthers to be protected. Some Ranunculacea^ with erect flowers — e.g. the Hepatica (An,emone Hepatica) and Winter Aconite {Eranthis, cf. figs. 222'' and 222 *<')— have their pistils surrounded by a crowd of stamens, and these again encircled by concave perianth leaves (petaloid sepals) which are wide open by day but closed at sunset, forming a dome over the stamen.s. The anthers of these plants do not open simultaneously, but only by degrees. The pollen on the outermost anthers nearest to the sepals is set free first of all, and this happens at a time when the filaments bearing those anthers are still short. It is obvious that comparatively short sepals suffice to shelter these stamens. Gradually, however, the anthers nearer the middle of the flower open; their filaments elongate, and the sepals would now be no longer of sufficient size to form a dome over all the pollen-laden anthers at night time. They accordingly grow in length day by day, until the anthers next to the carpels yield up their pollen. In the case of Eranthis the sepals lengthen in this way from 11 to 22 millimetres {cf. figs. 222 '^ and 222^°), and in that of Anemoiw Hepatica from 6 to 13 millimetres; that is to say, they actually double their original length. A curious instance of the closing of petals is that of Eschscholtzia Californica (cf. figs. 222^ and 222^). By day the four golden-yellow petals are expanded, the pollen falls from the stamens, which grow in a bunch in the middle of the flower, on to the concave petals, and rests on them in a floury layer as much as 1 millimetre in depth. When evening comes the anthers in the centre, having already lost their pollen, are left unprotected, but each petal furls itself up longitudinally in the prettiest manner conceivable, and thus the fallen pollen is sheltered under four little tents. The flowers composing the capitula of the Dandelion {Taraxacum), Lettuce {Lactuca), Chicory {Cichorium), Nipple-wort {Lapsana), and many other Com- posites, of which we may here select the Mouse-ear Hawkweed {Hieraciam^ Pilosella (fig. 222) as type, have tubular bases, but above are produced unilateral!}' into a strap-shaped structure to which the term ligule is applied. From thej bottom of each ligulate flower spring five stamens whose anthers are connate into a tube. This tube is early filled with pollen discharged introrsely, i.e.\ towards the centre of the flower through longitudinal slits in the anthers. Thel style is embedded in the tube, and as soon as the pollen is liberated it elongates, and, acting like a chimney-sweep's brush, pushes up the pollen which fills thel anther-tube until it rests above the opening at the top. The pollen resting onl the top of the style is brushed off" by insects when they settle upon the capitula.] But it is not certain that insects will make their appearance within a few hours! of the extrusion of the pollen, and even if they do they only brush lightly overl the flowers, and are sure to leave some of the pollen behind, and this pollen isl PROTECTION OF POLLEN. 115 then 1-eserved for another destiny which we shall have to consider more carefully later on. In any case the pollen adherent to the projecting end of the style, near the mouth of the tube composed of the connate anthers, must be protected before nightfall, when there will be condensation of dew, or in case of rain being imminent. This protection is, in fact, afforded to each floret by the ligule of the adjoining corolla, which stretches out laterally and constitutes an umbrella. In the Hawkweeds (Hieracium) the ligule bends so as to form a covering over the pollen to be protected (cf. figs. 222^ a.nd 222^). In Catananche, another Composite, each ligule is spread out flat whilst the sun shines, but in the evening becomes concave and at the same time arches over the pollen belonging to its own flower (fig. 222 ^). We cannot here go into all the diflerences in detail which occur in connection with this form of adaptation. We must not, however, overlook the fact that in these Composites the ligules of the peripheral florets of a capitulum are always much longer than those of the central florets, and that the pollen of the latter shares therefore the protection from wet afforded by the bending over of the outer ligules. We do not mean to say that the short ligules in the middle of the capitulum are not required to take any part at all in sheltering the pollen. In most instances they, too, stand up and curve over inwards, and act in conjunction with the longer outer ones in preventing the entrance of water. The adaptation of the flowers of Catananche is carried so far that the long ligules of the peripheral florets cease to bend inwards when there is no longer any pollen to protect in those florets — that is to say, when the pollen has been brushed off and the florets have entered into their last stage of development {cf. fig. 222"). The short ligulate florets in the central part of the capitulum must then of course see to the protection of their pollen themselves. This is the reason why one sees only the central ligules of old heads of Catananche arched inwards, whilst those near the margin remain motionless and stand out in rays during the dewy night just as they do under the noontide sun. The mechanism for the protection of the pollen is well worthy of notice in those Composites also in which the central florets of the capitula are all tubular and the peripheral florets all ligulate, and in those where the tubular florets ai'e crowded together on a round disc and encompassed by an involucre of stiff leaves which resemble petals. The Marigold (Calendula) may be taken as type of the first group, and the Carline Thistle (Carlina acaulis) as type of the second (fig. 224). In these plants the style grows and pushes the pollen out at the top l^f the tubular florets, just as in the case of the ligulate flowers above described ft was pushed up through the hollow cylinder formed by the connate anthers, jind above each floret a little lump of pollen is seen resting upon the free end of the style. These tubular florets are, however, incapable of securing their pollen jigainst bad weather, and a division of labour is therefore in some degree instituted kithin the limits of each capitulum, the ligulate florets or radiating marginal ^"•iicts, as the case may be, which produce no pollen, being turned to account or the purpose of sheltering the pollen-bearing florets of the centre. In fine 116 PROTECTION OF POLLEN. weather the ligulate florets and bracts stand out in rays from the periphery of the capitulum, but in bad weather and at night they are raised and actually bent over the central tubular florets. They are either disposed so as to form together a hollow cone over these florets, or else they overlap one another like the tiles on a roof ; often, too, they are twisted together in apparent disorder into a tuft, but they are always so arranged as to afford complete shelter to the central florets and to the pollen exposed by them. It is a remarkable fact that the length of these incurving rays stands in a definite relation to the diameter of the capitulum. Heads with large discs and great numbers of tubular florets have relatively long marginal rays, those with small discs and few tubular florets have relatively short rays. Moreover, at first when the florets in the middle of the disc are still closed, and only the tubular florets set near the margin have extruded their pollen, the ligulate florets of the ray and the radiating bracts are still short because they only have to shelter their nearest neighbours ; but as soon as the flowers in the middle of the disc open, the peripheral florets lengthen so as to be able to cover them also. Thus the roof here actually grows in proportion to the dimensions of the surface requiring shelter. The changes affecting the position of petals, ligulate florets, and bracts, which have been briefly described and which are classed together under the name of closing movements, take place in most plants in from thirty to fifty minutes, but in a few cases they are much more rapid. Sometimes the process of closing is com- pleted in the course of a few minutes. With Alpine plants it may happen that tht. flowers shut and open several times within an hour. The warmth imparted by a casual ray of sunshine is sufficient to cause the flowers of Gentiana nivalis to spread out their deep-blue petals, but no sooner does the sun disappear behind a cloud than the petals wind themselves round one another in a spiral and close up, forming a hollow cone. If the sun comes out again the corolla is once more open in the course of a few minutes. In plants with funnel-shaped, tubular, or bowl-shaped corollas, as, for example, the Thorn-apple, Gentians, and the Venus' Looking-Glass (Datura, Gentiana, Specu- laria), the phenomenon of closing is attended by a complex folding, bending, and twisting of the petals; but as a rule the position assumed by the petals on such occasions is the same as that which they previously exhibited in the bud. Generally speaking, most flowers and heads of flowers when closed at night have the same appearance as they had in the bud state. For the proximate cause of the movements of closing we must undoubtedly look to alterations in the tension of the layers of tissue involved in the operation. These alterations are due chiefly to variations of heat and light. Fluctuations in the degree of moisture of the air may also partly contribute to the result. In the Carline Thistle (Carlina acaulis), indeed, the opening and closing of the heads depends solely on this condition, and temperature is only a factor inasmuch as the relative moisture of the air is generally diminished as the heat increases in the parts of the world where the plant grows. Owing to this property of PROTECTION OF POLLEN. 117 Carlina acaulis, its large heads of flowers are used as hygrometers and weather- glasses. When the dry bracts surrounding the tubular florets of the capitulum stand out in rays dry weather and a clear sky are indicated, but when the hygroscopic bracts become erect and subsequently converge, so as to form a hollow cone, wet and cloudy weather is anticipated (cf. fig. 224). The significance of these movements of the radiating bracts or involucral leaves to the plant itself is as follows. By day when the air is warm and dry the rays have an outward curve and are spread out widely so as to turn their inner surfaces, which are silvery white, to the sky, and they glisten so brightly in the sunlight that they are visible from a great distance. They thus act as a means of alluring insects Fig. 224 - I'lotection of rolkn. Capitula of the Carline Thistle (Carlina acaulis), the one on the right open as in the sunshine, that on the left closed as at night or in bad weather. to the inconspicuous tubular florets of the disc, and these visitors whilst sucking the honey also load themselves with the exposed pollen and subsequently convey I it to other flowers. A large number of humble-bees alight on the open capitula of the Carline Thistle, suck the honey from the florets, and at the same time remove the pollen. If at that moment there were to be a sudden shower of rain the florets of the disc would inevitably be wetted and the pollen ruined. But owing to their hygroscopic sensitiveness the rays rear themselves up on occasion of even a slight increase of moisture in the air such as precedes rain, and, bending inwards, unite into a compact tent, off" which the drops of rain roll without being able to do any mischief. j Alterations in the form and position of certain tissues of the stamens due to the taking-in and giving-out of water also aflford a means of protection for pollen against wet in the case of Plane Trees, and of many Conifers, Yews, and Junipers 118 PROTECTION OF POLLEN. in particular. The pollen-cases are in these plants borne on squamous or peltate stalks, which are attached to an axis in a manner similar to the scales of a fir-cone. They also possess in common with the scales of a cone the property of closing and bringing their margins into contact when they are moistened, whereas when quite dry they stand away from one another, leaving wide intervening gaps {cf. figs. 226'^ and 226^*5 with figs. 226 ^' and 226 ^S). The pollen-dust which is developed in little spherical pollen-cases on the inner faces of the scales, is very liable to be shaken out of these gaping interspaces, but such an occurrence, as will be presently more fully explained, is only advantageous to the plant if dry weather prevails. In damp weather, and especially during rain, such escape would be equivalent to de- struction of the pollen. To avoid this risk the gaps close up, an operation which is effected by the scales absorbing moisture and swelling until their edges are in con- tact, so that the little pollen-cases attached to their inner surfaces are covered up. In the flowers hitherto described the parts adapted to the protection of the pollen from wind and wet are all leaf-structures or scaly or peltate outgrowths from the connectives of the stamens, and the adapted structure is bent or hollowed out, expanded or folded, as the case may be. Another group of floral forms, scarcely less considerable than the foregoing in point of numbers, secures this protection in a still simpler manner by bendings of the stalks and stem which convert bowl and cup-shaped flowers into pendulous bells. Usually the inflection occurs shortly before the blossoming of the flower, and then the flower retains the drooping position so long as its pollen is in need of protection. Man}- Campanulas (e.g. Campanula harbata, C. iJevsicifoUa, G. jmsilla), Solanaceaa and Scrophularinese (e.g. Atropa, Brugmansia, Cestrum, Physalis, Scopolia, Digitalis), Primulacese and Boragineas (e.g. Cortusa, Lysimachia ciliata, Solda- nella, Mertensia, Pulmonaria), Alpine-roses, Winter-greens and Whortleberries {Rhododendron, Moneses, Vaccinium), Ranunculacene and Dryadese (e.g. Aquilegia, Clematis integrifolia, Geum rivals), and many Liliaceous plants (e.g. Fr'diUaria, Galantlius, Leucojum, Convallaria) may be seen with their flower-buds supported on erect stalks and turned to the sky so long as they are closed. But before the flower is quite open the stalk curves downward, and the mouth of the flower is thus directed more or less towards the earth. No sooner has the flowering period expired, and with it the necessity for shielding the anthers concealed in the interior of the flower, than the stalks, in most instances (e.g. Digitalis, Soldanella, Moneses, Fritillaria, Clematis integrifolia, Geum rivale), straighten out again, and the fruit developed from the flower — especially if a dry fruit- is once more borne at the end of an erect stalk. The phenomenon is illustrated in figs. 221 •* and 221 ^ It is common to hundreds of plants belonging to most widely different families, and exhibits a great variety of modifications. The limits of this work forbid our discussing all these secondary forms of adaptation, which vary partly according to the structure of the stem and flower-stalks, partly according to the form and disposition of the leaves, petals, and stamens We can only give a brief account of some of the most striking cases. PROTECTION OF POLLEN. 119 If the filaments supporting the anthers charged with pollen are sma.l and short, the perianth, which in the inverted flower constitutes their protective cover, is also of small size, as may be seen, for instance, in the case of the Lily of the Valley (Convallaria majalis, cf. fig. 220^). A much longer envelope is assigned, on the other hand, to stamens with long filiform filaments. Flowers of the kind possessing large petals but seldom need to be completely pendulous in order to shelter their pollen, it is usually sufiicient for them to nod, i.e. to droop a little to one side. Thus, for example, the stalks of Lilium candidum bend in the flowering season only just enough to incline the mouths of the flowers in a lateral direction. Usually the form of the protective cover is such that the rain can trickle oflf it in drops. A contrivance far less common is for the petals covering the anthers to form a receptacle out of which the water is periodically emptied. An instance of this is afforded by the South African Sparmannia (Sparmannia Africana). The flower-buds are grouped together in umbels, and are borne on stalks, which are curved in a semicircle outwards and downwards away from the main axis, so that the flowers are inverted and their anthers are turned towards the ground and covered over by the petals. When the flower is open, however, the petals are not simply spread out like an umbrella, but are slightly tilted back, i.e. upwards. The margins of the petals overlap one another, and their outer surfaces, which, in consequence of the inverted position of the flower are uppermost, thus form a basin open to the sky. When it rains this basin placed above the anthers fills with water, thus adding to the weight borne by the stalk, and as drop after drop increases the strain upon the latter a point is at length reached when the basin tips over, letting the water flow over its edge without wetting the cluster of stamens suspended beneath it. This mechanism preserves the pollen clinging to the dehiscent anthers of Sparmannia from rain and dew in spite of their apparent exposure, which to a hasty observer seems to render it inevitable that the stamens should be wetted. In some plants whose flowers are arranged in racemes a process of inflection takes place before the floM^ers open, which does not affect the pedicels themselves but the axis from which they spring, the result being that the entire racemes or spikes become pendent. All the flowers are then inverted, and the petals act as a roof in sheltering the pollen adhering to the anthers. This is the case in the Cherry Laurel {Prunus Laurocerasus), the Bird Cherry (Prunus Padus), the Barberr}- (Berheris), and Mahonia. In the Walnut, the Birch, the Hazel, the Alder, and the Poplar {Juglans, Betula, Corylus, Alnus, Populus) also, the rachis of the spike changes its position shortly before the dehiscence of the anthers thus providing a shelter for the pollen as it becomes free. The male flowers of these plants whilst in the bud condition are crowded closely together, and form a stiff' erect cylindrical spike. But before the flowers open the rachis of the spike grows in length slightly and becomes pendent, whilst the flowers it bears are consequent!}' separated a little from one another and become inverted, so that the floral envelopes, which are composed of little scales and perianth-leaves, are uppermost 120 PROTECTION OF POLLEN. and the anthers below them (see tig. vol. i. p. 742). Whilst thus suspended beneath the scales the anthers open and the pollen rolls out. It is not, however, imme- diately blown aw^ay, but falls vertically and collects first of all in trough-like depressions which occur on the external surfaces of the separate flowers. Here it remains until there is dry weather and a puff of wind blows it away to the stigmatic flowers, this being accomplished in a manner that will receive closer consideration later on. Up to this moment its resting-place is sheltered from rain and dew by the flowers situated above it on the same spike, and the appendages of each flower thus constitute, on the one hand, a receptacle for the pollen of the higher flowers, and on the other, a roof over the pollen which has fallen upon the grooved backs of the lower flowers, as is shown in the illustration representing the flowers of the Walnut already referred to. A special interest attaches to those flowers and inflorescences which assume periodically an inverted position and whose stalks possess the faculty of bending, stretching, or turning concomitantly with the alternations of day and night, and of fine and wet weather. Such plants might quite properly be described as weather-cocks. They include forms belonging to most wddely different families, but possessing the common attributes — first, that their flowers or inflorescences are borne on comparatively long stalks, and secondly, that they offer their honey and pollen to the flying insects which visit them in shallow cups or flat saucers, or even on plane discs. In the daytime in fine weather when flowers and inflorescences of this kind straighten out and turn their open surfaces towards the sun, they are plentifully visited by such insects as refuse to enter pendent bells and tubes from underneath, and only alight from above on wide, open, and easily accessible flowers, and thus is effected the important function of pollen- dispersion. On the other hand, by becoming pendent at night and in rainy weather — i.e. at a time when insects are not commonly on the wing — they ensure security for their pollen and honey against wet. Hence the periodic movement of the axis appears to achieve a double advantage. In many Campanulacese and Geraniacege it is the stalks of individual flowers that bend. The widely-distributed species, Campanula patula and Geraniwni Robertianum have been selected from the list of those orders for illustration (c/. figs. 225^ and 225^ with figs. 225^ and 225*). The same phenomenon occurs in many species of Wood-sorrel, Poppy, Pheasant's Eye, Isopyrum, Crow-foot, Wood Anemone, Cinquefoil, Starwort, Chickweed, Saxifrage, Rock-rose, Anoda, Potato, Pimpernel, Jacob's Ladder, and Tulip (e.g. Oxalis lasiandra, Papaver alpinum, Adonis vernalis, Isopyrum thalictroides, Ranunculus acris, Anemone nemorosa, PotentiUa atrosanguinea, Stellaria graminea, Cerastium cJdoroifolium, Saxifraga Huetiana, HeliantheTnum, alpestre, Anoda hastata, Solanum tuberosum, Anagallis phoenicea, Polemonium coeruleum, Tulipa sylvestris). In the Scabious given in the illustration opposite (Scahiosa lucida, figs. 225"^ and 225^), and in several Composites (Bellis, Doronicum, Sonchus, Tussilago, &c.) it is the peduncles bearing the capitula which bend; in many Umbelliferous plants (e.g. Astrantia PROTECTION OF POLLEN. 121 ulpina, A. carniolica, &c.), it is the stalks of the umbels, and in some Cruciferous plants (e.g. Draba aizoides, Arabis Turrita, Sisymbrium Thalianuvi), the axes of the racemes. The above-mentioned Scabious and Composites exhibit a periodic inversion of the entire inflorescence in consequence of the inflection of the axis, and the radiating ligulate florets set round the margin of the capitulum serve to shelter the pollen of the central florets. Similarly in the Umbellifers named, the involucres of the separate umbels, being comparatively large, act in the same way. The fact is also worth notice that in some Willow-herbs (e.g. Ejoilobium \ ^^/--f' ^W '* Fig. 225.— Pi-otectiuii uf Tollen. • Flowers of the Herb-Kobert (Geranium Robertianum) in the daytime ; the pedicels erect. « The same plant with its flowers pendent on curved pedicels, the position assumed during the night and in wet weather, s Bell-flower (Campanula patula) by day ; the flower on erect pedicel. ■* Flower of the same plant inverted for the night or for wet weatlier, the pedicel being curved. « Capitulum of a Scabious (Scabiosa lucida) in the daytime ; the peduncle erect. « Capitulum of the same plant at night or during rain, the peduncle curved and the capitulum inverted. hirsutum,, E. m.ontanuQn, E. roseum.), the flower-stalks themselves do not bend, but the long stalk-like inferior ovaries curve downward and straighten out again, periodically causing the flowers, which are of a flat salver shape, to alternate between a pendent and an erect position. The inflection of flower- stalks, or, of their substitutes, the ovaries, ceases as soon as the pollen of the flowers concerned has been removed by one means or another, and a shelter for it is no longer needful. The flower-stalks of Saxifraga Huetiana only con- tinue to bend so long as the anthers in the flowers they support are covered with pollen, and the long ovaries of the Willow-herbs mentioned above only curve towards the earth on two successive evenings; the third evening, 122 PROTECTION OF POLLEN. when there is no longer any pollen to protect from rain and dew, they remain erect. All these phenomena of inflection and straightening on the part of flowering axes and inferior ovaries are brought about in the same way as the periodic movements of petals and bracts by alterations in the tension of the tissues. These variations of tension are again due partly to vicissitudes in respect of heat and light, and of the degree of moisture of the air. But mechanical stimuli also play an important part, especially such shocks to the flower-bearing axis as are occasioned by the incidence of drops of rain and by gusts of wind. The fact that drops of water are found resting on the nodding or drooping flowers, if the latter are examined before sunrise when there is a heavy dew, or after a shower, tempts one to look upon the inflection merely as a consequence of the strain imposed upon the stalks by the increased weight of the water- laden flowers. No doubt this strain has something to do with the inflection, but it is equally certain that the drooping state does not disappear at once when the water has evaporated and the strain due to its weight has terminated. This persistence of the inflection at all events must be attributed to an alteration in the tension of the tissues of the stem, and no more than the first impulse can be derived from the weight of dew or the impact of drops of rain. Additional evidence of this is afl'orded by the facts that the process of bending is set up by rain falling on flowers and stem, even when it rolls ofl" immediately, and that pedicels and peduncles also bend over whenever the entire plant i.'^ caused to sway about by the wind which precedes a downpour, the stems on these occasions always curving away from the' direction of the wind, or, to use a nautical expression, to the lee side. This phenomenon of the bending of stalks and drooping of flowers before the rain has actually begun looks almost as if the plant had the power of foreboding the approach of bad weather and of adapting itself beforehand in such a manner as to prevent any injury being subsequently inflicted upon it by that destructive agency. Such is the opinion of the peasantry in parts of Europe, and they look upon the inflections above described, as well as the closing of the heads of the Carline Thistle, which was mentioned further back. ■ as a sign of imminent rain. There is, however, as already said, a mechanical explanation of the phenomenon dependent on a change in the tension of the tissues of the stem induced by the oscillations of the plant when subjected to the gusts of wind which usually precede rain, the change of tension beini: manifested externally by the persistence of the stem's inflection. Moreover, this ; lasting curvature of the stem may also be produced artificially by inducing | the same kind of strain as is caused by the weight of the rain-drops or the | vibration caused by rain and wind. If, for instance, you bend the pedicels of ! various species of Oxalis from the erect position they occupy in the middle of , the day and hold them down for a time, or if you shake or knock them, the ' tissues forthwith undergo a change of tension which results in those stalks PROTECTION OF POLLEN. 123 becoming curved and the flowers drooping towards the ground instead of facing the sky as before. The same is true of the stalk of a Tulip (Tulipa), of the long peduncles of BoronicuTn, of the flower -bearing stems of Asperula arvensis, Astrantia major, Gardamine pratensis, Lychnis flus-jovis, and Primula cortusoides. If you try to straighten the stalks again afterwards you run a risk of breaking them. An interval of some hours elapses before this inflexibility disappears and the tensions existing before the act of mechanical stimulation are re-established and the stems become straight again. These different changes in the direction and position of petals, bracts, flower- stalks and stems, which take place concomitantly with the alternations of night and day, of storm and calm, cloud and sunshine, often imply a complete trans- formation in the aspect of the vegetation within a very brief space of time. On warm summer days, when the sky is clear and the air still, the green of the meadows is sprinkled with the colours of innumerable open flowers. The stellate, salver-shaped, and cup-shaped flowers and inflorescences of Anemones, Ranunculuses, Potentillas, Gentians, and Composites are all wide open, so that the upper brightly-coloured surfaces of their flowers are visible from a great distance. Most of them are turned towards the sun, which enhances their brilliancy; several of the flowers and inflorescences — as, for instance, the Rock- rose (Helianthemum) — follow the sun, and face the south-east early in the morning, the south at noon, and the south-west in the afternoon. Countless flies, bees, and butterflies swarm and buzz round the flowers in the sunshine. When the sun sets a cool breeze springs up, and there is a copious deposit of dew on leaves and flowers. The insects withdraw to their homes to rest for the night, and the flowers seem to fall asleep too. Petals fold up, heads of flowers close, flowers and inflorescences bend towards the ground and exhibit the inconspicuous outer surfaces of their floral envelopes to the onlooker. Whilst the night lasts the meadow, drenched in dew, continues in a state of torpor, from which it is awakened once more by the warmth imparted by the sun when it rises next morning. A similar change of aspect occurs when a storm is brewing, when the meadow is swept by wind and rain falls upon the flowering plants. In this event also most flowers cover over or wrap up the parts liable to destruction in time to prevent material damage being done to their pollen. Comparatively few among ordinary meadow plants appear to be in no way affected by these alterations in external conditions. Some seem to be able to dispense altogether with contrivances for protecting their pollen, for when once the flowers have opened the pollen-cases are left free and uncovered even on occasion of heavy showers. Thus, for example, in Plantago and Glohidaria the anthers are borne on long filaments and project in both good and bad weather out of the small flowers, which grow close together in spikes and capitula, and it would seem as though their pollen were exposed to inevitable destruction in case of wet. But closer inspection reveals that even these plants are not destitute of apparatus for the protection of the pollen. To the anthers themselves 124 PROTECTION OF POLLEN. is due the security enjoyed by the pollen developed from their tissues. For if dewy nights or wet weather occur after dehiscence has taken place and whilst the pollen is exposed at the apertures in the anther-cavities, the latter <;lose up again and encase the pollen once more. The mature pollen is then protected from wet just as effectually as it was during the period of its maturation, for no injurious effect can be exercised by rain or dew through the walls of the anther upon the pollen-cells concealed within. When thei'e is a return of warm, dry weather the anthers open afresh in the same manner as on the occasion of their first dehiscence. Precisely the same processes as were described on pp. 91-93 are repeated. If the anthers are unilocular with transverse dehiscence, like those of Olohularia and the Lady's Mantle (Alche- milla; see figs. 226 ^•'^•'^'^■^■^°), the sutures open and shut like lips. If the dehiscence is opercular, as in the Bay Laurel (Laurus nohilis; see figs. 226 ^^' ^^' ^^' ^*), the valves shut down again and force the pollen adherent to them back into the open recesses of the anthers. Lastly, if the dehiscence is longitudinal and the anther- walls open outwards like folding dooi-s and at the same time become revolute, as in Thesium and Bulhocodiwm (cf. figs. 226 ^' ^' ^' *), the movement is reversed in wet weather, and the two valves close completely tocjether ajjain. In the Arctic regions and amongst the mountains of Central Europe where copious deposits of moisture occur during the flowering season common to most plants, the number of species possessing anthers which open and shut periodically is not great. Besides those already named, i.e. Bulhocodiwm, Thesiwni, and the Alchemilla, only the Plantains (Plantago) and Ranunculaceae, especially those with pendulous anthers (Thalictrum), remain to be mentioned as exhibiting this phenomenon particularly clearly. It appears to be much commoner in warmer parts, especially in sub-tropical and tropical regions; at all events, this periodic opening and closing of the anthers is exhibited to perfection in the following plants : — Cinnamon-trees, the Camphor-tree, the Laurel and Lauraceous plants generally, Araliacese and Cycadese, the various species of Ricinus and Euphorbia, Cistus, the Vine (Vitis), and indeed the majority of Ampelidere, the Tulip-tree and Magnolias (Liriodendron, Magnolia), and lastly, amongst Conifers the genus Cephalotaxus. The phenomenon in question is the result of changes in the condition of the air in respect of moisture, and depends upon the contraction and expansion of the hygroscopic cells which we noticed in the last chapter as being developed underneath the epidermis of the anther-walls. As in the case of the movements of the involucral bracts on the capitula of the Carline Thistle, the process is only affected by heat inasmuch as the relative degree of moisture in the air alters with a rise or fall of temperature. Seeing that under ordinary conditions variations of temperature and increase or decrease of humidity are connected with the alternation of day and niglit, it is clear that a periodicity will also be manifest in the opening and closing of anthers, and that in the evening when the degree PROTECTION OF POLLEN. 125 of moisture is increased the anthers will close, remain shut throughout the night, and not begin to open again till after sunrise, when the degree of moisture is diminishing. In cases where both the anthers and the petals of a flower open and close periodically, the corresponding movements are for the most part accomplished simultaneously; but if the cause of the movement is different for petals and anthers it may happen that there is no such unison. For instance, after prolonged rain, the petals of Bulbocodium may open under the influence of a Mi ;> -j; « 1 ^ Fig. 226.— Protection of Pollen. 'Flower of the Bulbocodium with the perianth and the anthers open as they are when the sun is shining and tlie air dry. 2 An anther from the same. 3 Flower of Bulbocodium in moist air ; the perianth half open, the anthers closed. * An anther from the same. 6 Flower of the Lady's Mantle (Alchemilla) with its anthers open in a dry atmosphere. 6, 7 Antliers from the same, s Flower of the Lady s Mantle with its anthers closed in rainy weather. ^, i" Anthers from the same, n Flower of the Bay Laurel (Laurus) with its anthers open in a dry atmosphere. 12 An anther from the same, is Flower of the Bay Laurel with its anthers closed in wet weather. 1* An anther from the same. 15 Staminiferous flowers of Junipanis Virginiana in a dry atmosphere. 1^ The same magnified, i? Staminiferous flowers of Juniperus Virginiana in wet weather, is The same magnified. 1, 3, is, 17 natural size. The rest x 2 to 8 times. warm spell of sunshine, whilst the anthers still remain closed owing to the excessive moisture of the atmosphere. Anthers close up much more quickly than petals on the approach of danger. They usually take only a few minutes, and in many cases not more than half a minute. The anthers of the Bastard Toad-flax {Thesiwm alpinum) shut up within thirty seconds of their being moistened. In this plant the process of closing is rendered additionally interesting by the fact that the moistening of the anther-walls is effected by peculiar tufts of hairs projecting from the perianth. The briefest possible description of this phenomenon will be given here. The open flower of Thesium has the limb of its perianth turned to the sky. This position is maintained unchanged day and night, and even the occurrence of bad weather does not cause any alteration in the direction of the flower-stalks or the position of the flowers. Hence rain-drops falling from above and the dew formed on 126 PROTECTION OF POLLEN. clear nights must inevitably rest on the open flowers. The immediate wetting •of the entire flower is, however, prevented by peculiarities in the form of the limb. The anthers close with great celerity upon the deposition of the drops, the expla- nation being that the perianth-lobes are connected with the anthers standing in front of them by a bunch of twisted hairs which not only are themselves peculiarly susceptible of being wetted, but conduct the water to the anthers and so cause the anther-walls to close. A characteristic manner of protecting the pollen by means of the anther- walls after the pollen has been set free, and when it is ready to be carried away by insects, may be observed in several Composites (e.g. Onopordon, Centaurea). There is no material difference between these plants and the other Composites discussed on p. 114 in respect of the structure of the tube of syngenesious anthers, the discharge of the pollen into that tube, or the structure of the stj'le and its situation inside the anther- tube; but an essential distinction exists in the fact that the pollen is conveyed to the mouth of the tube not through the elongation of the style but the contraction of the filiform supports of the anther-cylinder. These filaments in Onopordon and Centaurea contract in response to mechanical stimuli, and in shortening they pull down the anther-tube with them. The top of the style thereupon becomes visible, for the style is sheathed in the tube, and does not shorten when the filaments do so nor change its position. The pollen resting on the st34e is consequently exposed, and appears in the form of a pulverulent mass on the top of the stjde surmounting the anthers. If the mechanical stimu- lation of the filaments is due to the hovering of an insect about the capitulum, the pollen is no sooner exposed than it is brushed off" by the insect, and the entire contrivance is obviously so devised that the same insects as cause (by the touch of their legs or probosces) the contraction of the filaments, the retraction of tlie anther-tube, and the exposure of the pollen may be themselves loaded with the pollen. Up to the moment of the insect's visit, however, the pollen is hidden in the sheath formed by the anthers, and this position is of advantage to it inasmuch AS it is there sheltered from rain and dew. The Composites in question have their capitula erect. The capitula of Onopordon include neither movable ligulate •ray-florets nor radiating bracts capable of closing. Centaurea has trumpet-shaped marginal florets, but they do not possess the power of arching over and protecting the tubular florets of the centre. The stalks of the capitula become neither pendent nor nodding in wet weather. In short, the pollen of these particular Composites is destitute of any of the various means of protection which are present in other genera of the same family and which have just been discussed. But instead, the anther-tube itself undertakes the task of sheltering the pollen after the latter is liberated until the moment when the insects which are to carry it away alight upon the flowers. We need only notice incidentally that extrorse anthers, which turn their recesses filled with coherent masses of pollen towards the earth and their backs to the sky are also to a certain extent protected against wet. A more impor- PROTECTION OF POLLEN. 127 tant provision at all events consists in the fact that the injurious effect of rain or dew on the pollen-cells may be obviated by certain special sculpturings on the surfaces of these cells. Reference has already been made to such cases at the conclusion of the last chapter. They are on the whole rare, and are limited apparently to plants of the tropical and sub-tropical regions. The pollen of the beautiful climbing Cobcea scandens (cf. fig. 217^), one of the PolemoniacejB, will serve as an example. On the surface of this pollen may be observed a number of little pits with angular rims which make it look at first sight almost like a honey-comb. The pits are not, it is true, so deep as those of a honey-comb, but they are deep enough to prevent the air with which they are filled from being displaced by water dropping upon the pollen. Thus air remains in the pits and thereby affords protection from wet, for it forms an intermediate layer separating the thin parts of the cell-membrane from the water. The thick layers of the cell-membrane which project in ridges are still liable to be wetted, but water cannot penetrate at once through them into the interior of the cell, and such an entrance it is that constitutes the greatest danger to the pollen. A gradual absorption of watery liquid — especially that which is derived from the cells of the stigma — is not only not avoided, but is even necessary for the ■subsequent development of the pollen-cells. The instances chosen hitherto for the exemplification of the numerous contriv- ances whereby the pollen in flowers is protected against wet belong, for tlie most part, to the category of those which have developed one form of protective apparatus ■only. Frequently, however, two or even three methods of defence co-exist, so that in case one contrivance should fail there is another in reserve. This occurs in ca.ses whore the plant has only a meagre stock of pollen, where the number of flowers on one individual and the quantity of pollen-cells produced from each flower are small, and therefore there is not much pollen to waste, where the time allotted to a plant in which to unfold all its flowers is extremely limited, and where the transport of. the pollen from flower to flower is accomplished exclusively by flying insects, whose visits are sometimes delayed for several days when the weather is unfavourable. To mention a few instances with more than one means of protection, in many Anemones and Crow-foots, the Hepatica, the Rock-rose, and the Wood-sorrel (Anemoiie, Ranunculus, Hepatica, Helianthemum, Oxalis), not only do the petals close over the pollen-laden anthers, but the flower-stalks also bend, causing the flowers to nod. In the Daisy (Bellis), the Corn Sow-thistle (Sonchus arvensis) and many other Com- posites not only do the ligulate florets of the ray incline towards one another and form a roof over the pollen of the central florets in cloudy weather and in the evening, but in addition the peduncles become bent or pendent. In Podophyllum ■peltatum the pollen is sheltered by the bell-shaped flower, but in addition to this the peltate foliage-leaves are also spread out over the flowers and act as umbrellas. Tiie ■synchronous closing of both anthers and petals over the pollen when rain threatens is a phenomenon that may be easily observed in a number of plants, as, for Instance, in BuXhocodium {cf. figs. 226 ^- ^•^•*). 128 PROTECTION OF POLLEN. The fact is also worthy of note that identical means of protection have not always been evolved by members of the same family of plants. One has one method of defence, another another. This diversity is exhibited particularly by the various genera of Solanacese, and by the multifarious species of the genus Campanula. In the Solanacese we find the following variety of contrivances according to the genus. The flowers of the Potato {Solanum tuberosum) fold up in the afternoon and assume an inverted position owing to the curvature of their stalks for the night, but only maintain it whilst the night lasts. The next morning the flower-stalks straighten, and the flowers unfold again. The Deadly Night-shade {Atropa Belladonna) has its flowers inverted during the whole of the flowering season, and it is therefore not necessary for the corollas to open and shut. The flowers of the Mandrake (Mandragora vernalis) remain erect, but in the night and in rainy weather the tips of the upright corolla-lobes close over the pollen-covered anthers inside. As regards the different Bell-flowers {Campanula), those which have very long peduncles — e.g. Cam^panula carpathica and Campanula patula (cf. figs. 225^ and 225*) — are only pendent in the night and in bad weather; by day and in fine weather they are erect. They exhibit pronounced periodic movements resulting in the curvature of their axes. In other Bell-flowers with shorter stalks — e.g. Campanula persicifolia, C. pusilla, C. rotundifolia — the buds nod before they open and continue in this position throughout the time of flowering, whilst in those species wherein the flowers are crowded together in heads and have very short stalks — e.g. Cam.panula Cervicaria, C. glom,erata, C. spicata — there is in general no curvature of the axes, but the flowers remain upright and guard themselves against rain by means of an inflection of the points of the corolla towards one another which closes the mouth of the bell. Lastly, in the Venus' Looking-Glass, a plant nearly related to the Bell- flowers, the flower closes by means of deep folds formed in the corolla. When contrivances have to be described which subserve several purposes at the same time, it would lead to confusion to attempt to say everything that there is to be said about them in one place. In such cases it is much more to the purpose to keep one object alone in view even at the risk of appearing one-sided to a hasty reader. This remark is particularly applicable to the means of protection just described as being adopted by plants to preserve their pollen from wet; for there is no question but that most of these contrivances are capable of rendering other services to the plants in question besides the one specified. In many cases the closing of petals effects not only the protection of the pollen, but also its transference to neighbouring stigmas in the event of a dearth of insect-visitors, as will be explained in a subsequent chapter. If a flower-cup filled at the bottom with honey remained open to the rain the honey would be immediately spoilt and would no longer act as an allurement to insects. Hence we may infer that the shutting of the entrance to the interior of the flower, the construction of the corolla-tube, and the change to a nodding position in the case of melliferous flowers preserve not only the DISPERSION OF POLLEN BY THE WIND. 129 pollen, but also the honey from being spoilt by the wet. The narrowing of the corolla-tube and the barricading or complete closing of the entrance to the flower also serve, on the other hand, to keep out certain honey-seeking creatures whose visits would not be advantageous to the plant. Finally, these same con- trivances may ward off also such insects as would remove the pollen without conveying the least particle of it to other flowers. In connection with this last function there exist, no doubt, special adaptations besides, one of the most striking of which occurs in the Monkey Flower (Mimulus) and in the Hemp^ Nettle (Galeopsis), and is shown in the illustration of a stamen of Galeoi^sis angustifolia (fig. 216 ^^ p. 91). In this instance the anthers are furnished with two lids which can only be opened by a certain proportion of the insects visiting the flowers. Insects with bodies of such a size that when they enter the flower they rub the pollen from the anthers on to their backs are able to lift the lids of the anthers by brushing against them, and they thus expose the pollen. On the other hand, smaller animals which would not load their backs with pollen on visiting the flowers in question or would not convey it to the stigmas of other flowers are not strong enough to open the anthers. Thus the pollen is effectively protected by means of these lids against the detrimental action of small-sized plunderers. DISPERSION OF POLLEN BY THE WIND. At the beginning of the last chapter it was stated that the medium wherein the transport of the pollen to the stigmas takes place is, in the great majority of Phanerogams, the air. For the conveyance of pollen between flowers situated at a distance from one another there exist two main agents, viz. the wind and insects. Hence Phanerogams have been distinguished by botanists into "anemo- philous " or wind-fertilized, and " entomophilous " or insect-fertilized plants. But these terms, which are adopted in most works on Botany, can only be used in a strictly limited sense. It is no doubt true that there are plants in which the I transference of the pollen to the stigmas is eflected exclusively by the wind, and I others in which the equivalent process takes place solely through the intervention of animals; but, on the other hand, it has been ascertained in the case of a large number of plants that whereas shortly after the flowers open small creatures carry off" the pollen and convey it to other flowers, later on, when the flowering period is drawing to a close, the pollen is committed to the wind and by it transferred to the stigmas of neighbouring blossoms. The best instances of this are aflbrded by several of the Rhinanthacese, as, for example, Bartsia and the Tooth wort (Lathrcea), and by many Ericaceas, such as Calluna vulgaris and Erica carnea, but many more could be mentioned. The conformation of the various parts of these flowers when they first open renders a dispersal of the pollen by the wind impossible; but in fine weather insects visit them in large numbers, and in the act of sucking the honey load themselves with pollen 130 DISPERSION OF POLLEN BY THE WIND. which they afterwards convey to the stigmas of other flowers. Subsequently, however, the conditions are reversed, the supply of honey is exhausted and insects stay away; but, on the other hand, the filaments bearing the anthers have elongated, the pollen-sacs are consequently exserted above the mouth of the corolla, the pollen contained in them is laid bare, and, at the proper time, is blown away by the wind to the stigmas of younger blossoms. Plants of the kind thus appear to have a second contrivance in readiness in case the first fails, so that in any circumstances the object of flowering may be attained. This is indeed a matter of urgent necessity. How easily may it happen that insects ax-e kept away for a long time by unfavourable weather or that they pay but a few visits. Most plants, therefore, take the precaution to provide that under such circumstances the expenditure of energy involved in the production of flowers shall not have been in vain. It would be inconsistent with the plan of this book to discuss here all the remarkable adaptations which have been evolved for the purpose of providing a supplementary means of dusting the stigmas with pollen in the event of an absence of insects, but it is necessary to make preliminary mention of this one ai-rangement whereby many flowers, originally entomophilous, subsequently become anemophilous, because it enables us to determine the proper degree of significance to be attached to the division of plants into anemophilous and entomophilous species. As would naturally be expected, it is, speaking generally, only pollen whicli is of dusty or floury consistency that is transported by the wind. If it is true, as gardeners assert, that the pollen of Azaleas, which oozes from the anthers in the form of sticky fringes, has on occasion been torn away and conveyed to the stigmas of neighbouring flowers by the wind, the occurrence can only be looked upon as accidental. In ninety-nine cases out of a hundred the viscid strings, if detached by the wind, would not be conveyed to the stigma of another flower, but would adhere to the outside of the calyx and petals, or to the leaves and stem, and would there perish. The same remark applies also to pollen-cells which are bound together into little lumps by oil and viscid substances, or by acicular processes on the outer layer of the cell-membranes. Only in the rarest instances are they carried by the wind to the stigmas of flowers in the vicinity. These are primarily adapted to becoming attached to the bodies of winged insects. All the more remarkable, therefore, is the fact that in certain water-plants the pollen, though cohering in sticky masses, is blown by the wind on a kind of little boat to the stigmas which are raised above the surface of the water. The phenomenon was first observed in the case of Vallisneria spiralis, an aquatic plant which grows in still water, and is widely distributed in Southern Europe. It is, for example, very luxuriant in the ponds, canals, and shallow inlets along the shores of the Lake of Garda, and we will select it as an illustration in the account which follows. The reader is requested first of all to look at the figure on p. 667 of vol. i. It represents a plant living under water with strap- DISPERSION OF POLLEN BY THE WIND, 131 shaped leaves arranged in fascicles at the ends of the creeping stems which are attached to the mud by root-fibres. In the axils of these leaves a variety of buds are produced — in some cases one only which constitutes the starting-point of a new creeping shoot; in othei's three close together, one of which grows in length parallel to the bottom and develops a foliage-bud at its extremity, whilst the two otliers grow straight upward, or there may be two, of which one elongates in a horizontal direction, whilst the axis of the other rises towards the surface of the water. Each of the upward-growing shoots terminates in a kind of bladder composed of two concave and somewhat transparent bracts, one of the pair overlapping the other so as to close the bladder securely. Within these bladders are the flowers. Of the individual plants some develop female flowers only, others male flowers only. The former occur singly in the bladders. Each possesses a long cylindrical inferior ovary crowned by three relatively large stigmas with bi-lobed apices and fringed margins. The stigmas are surrounded by an envelope consisting of an upper whorl of three small abortive petals and a lower whorl i of three large ovate-lanceolate sepals. These floral segments are invariably so disposed as to allow the finely-fringed margins of the stigmas to project somewhat beyond the perianth-lobes so that pollen may be caught by the fringes from the side. This is also the reason why the three inner perianth-lobes are stunted, for if they were as large as the outer three the stigma would be covered in at the side and no adhesion of pollen could take place. When the stigmas have reached the stage of being adapted to the reception of pollen, the top of the bladder investing the flower splits; the ovary elongates, flower and stigma are pushed up above the envelope, and appear on the surface of the water, where they are spread out in the medium of the air (see fig. 227). The phenomenon described is only rendered possible by the fact that the stalk of the pistilliferous flower lengthens to an extraordinary extent, and does not cease growing until the flower it bears has reached the surface of the water (c/. vol. i. p. 667). The case of the staminal flowers is utterly different. They are not solitary, but grow in large numbers in a bunch on an axis which stands up in the middle of the investing bladder. The two leaves composing the bladder become disjoined under water, and expose the raceme of spherical buds. The buds are still in situ jon the rachis, which remains quite short, the inflorescence being held at a height lof about 5 centimetres above the mud, as is shown in fig. 155, p. 667, vol. i. Shortly afterwards one of the most wonderful processes exhibited b}^ the vegetable world is gradually accomplished. The flower-buds hitherto connected with the axis of the raceme by diminutive stalks become detached, ascend in the water, and float about on the surface. At first they are still closed and globular, jbut soon afterwards they open. The three concave leaflets (sepals) forming the outer whorl of the perianth, which have up to that time been arched like cowls 3ver the stamens, are thrown back and assume the appearance of three boats connected together at one spot, and the stamens, which were originally three in lumber, but of which only a pair are now furnished with anthers whilst the thii-d 132 DISPERSIOX OF POLLEN BY THE WIND. has remained rudimentary, project obliquely up into the air (see fig. 227). The opening of the petals is immediately followed by the dehiscence of the authors. The coat of the anther shrivels up rapidly, leaving nothing but a little flap upon which the pollen-cells rest. There are generally only 36 pollen-cells contained in each anther. These are comparatively large and very sticky, they cohere together and form a mass of pollen which is borne upon the thick stamen. Notwithstanding the fact that they are very near the surface of the water, the masses of pollen-cells are not easily wetted. The three sepals underneath them form, as has been said, three boats which respond to the slightest movements of the water without upsetting, and therefore protect their freight from wet to Fig. 227. Flowers of Vallisneria fpiralis floating on the surface of water. In the niiiMle a female flower with several male flowers on eitlier side of it in various stages of development; some still closed, some in process of opening, some open with their boat-shaped perianth-lobes thrown back. Projecting from the open flowers are the stamens. An open anther is attaching its pollen to the fringed stigmatic margin of the female flower, x 10. perfection. These little floats are blown hither and thither by the wind and accumulate in the neighbourhood of fixed bodies, especially in their recesses, where they rest like ships in harbour. When the little craft happen to get stranded in the recesses of a female Vallisneria flower they adhere to the tri-lobed stigma, and some of the pollen-cells are sure to be left sticking to the fringes on the margins of the stigmatic surfaces. Directly after the adhesion of the pollen, which takes place in the manner shown in fig. 227, the female flower is draw^n down under the water. The long flower-stalk assumes a spiral form, and its coils close up so tightly together that the ovary, or young fruit as it now is, is brought to rest at quite a small distance above the muddy bottom of the water. Up to the present time the conveyance by the wind of adhesive pollen on floats composed of the perianth of the flower is known to exist in the widely- DISPERSION OF POLLEN BY THE WIND. 133 distributed Vallisneria spiralis, in Vallisneria alternifolia, which is indigenous in tropical Asia, in JEnalus acoroides, which grows in the Pacific and Indian Oceans, in Hydrilla verticillata, Elodea Canadensis, and a few species of the genus Lagarosiphon, native at the Cape and in tropical Africa — only 13 species all together, comprised in the little family of Hydrocharidacese. This number is almost inappreciable compared with that of the species which produce pollen in the form of fine dust or loose flour, and wherein the pollen is dispersed exclusively and throughout the period of flowering by the wind which blows it away in clouds. It would not be far out to put the number of wind-pollinated plants at 10,000, i.e. at about a tenth of the total number of Phanerogams. To this category belong the Conifers, Oaks, Beeches, Hazels, Birches, Alders, Poplars, Walnut-trees, Mul- berry-trees, Planes, and the majority of Palms. All these are of the nature of lofty trees, and usually grow in numbers together, each being associated with others of its own kind so as to form extensive woods or plantations, characterized by a close association of individuals. To these must be added the Grasses produced in meadows, prairies, and savannahs; the Sedges, Reeds, and Rushes characteristic of marshes; the Cereals of our fields; Hemp, Hops, Nettles, and Plantains; the common Pondweeds growing in still or running water, and many other plants belonging to families of the most widely different kinds. One striking characteristic of these exclusively wind-fertilized plants is the absence of fragrant and bright-coloured flowers. The floral-leaves are compara- tively small, of a greenish or yellowish colour, and stand out very little, if at all, from the foliage. The interior of the flower is destitute of honey and perfume. It is of no advantage to these flowers to be visited by insects, and accordingly they have no need of any of the means of alluring bees, butterflies, or flies. Hence the absence of odorous substances, of sw^eet juices, and of brightly-coloured corollas contrasting with the green foliage and visible from afar. We do not mean to say, however, that the flowers of the plants in question are altogether shunned by insects. Many insects covet the pollen itself, and not infrequently they may be seen hovering about the catkins of Hazels and Birches, on the spikes of Plantains, the panicles of Grasses, Rushes, and Reeds, collecting or devouring the pollen. But these visitors play but a very subordinate part in the dispersion of the pollen. By knocking against parts of a flower that are covered with pollen-dust they may, of course, cause some to fall out, but in so doing they only render a service to the plant if the right wind happens to be blowing at the moment and conveys the pollen to the stigmas. If there is no wind, or it blows in a wrong direction, the plants are more likely to be injured than aided by the insects' visits; for, as the stigmas are not brushed by the pollen-seeking insects and therefore receive no deposit of pollen from them, and, on the other hand, the pollen that they shake out is not likely to be carried to the corresponding stigmas if the air is still, it usually happens that anemophilous plants of the kind thus suffer a loss of pollen without ■obtaining any compensating advantage. 134 DISPERSION OF POLLEN BY THE WIND. As has been above implied, however, it is not eveiy aerial current that is adapted to serve as an agent for transferring pollen to stigmas. The least favourable winds are those which are combined with atmospheric deposits. Besides the fact that the pollen-dust would be washed away from its resting- places by the rain and carried to the ground, it must perish in consequence of the soaking. Storms of wind without rain are also anything but beneficial, for they forcibly whirl away any pollen that they encounter and carry it in one direction only, and, as but a small proportion, if any, of the stigmas requiring to be fertilized lie in the path of the wind, the greater part of the pollen is wasted. The result aimed at is best achieved when the pollen-dust, after being removed from the spot where it has been produced or deposited, is distributed uniformly over an ever-extending area, becoming, in a manner of speaking, diluted and forming a cloud of gradually increasing dimensions but diminishing density, so that the thousands of loose pollen-cells which have up to that time been crowded together within the province of the flower and contained in a space about the size of a pin's head are scattered over an area many million times as great. A gradual dispersion of the kind is only occasioned by a gentle wind. The light breezes which sweep through valleys shortly after sunrise, ascending air- currents such as one perceives quivering over heated plains at noon, the alter- nating land and sea breezes of the coast-winds which, in passing over cornfields, set the com in gentle waving motion, and in woods cause a scarcely audible rustle — such are the most propitious agents of pollination. It is easy to observe how, at the proper season, under the influence of a gentle wind of the kind one little cloud of dust after another detaches itself from the flowers of the plants in question and slowly soars away. Owing to the fact that the motion of aerial currents is undulatory and undergoes at short intervals alternate augmentation and diminution, the first motion of the pollen as it dissipates itself is also in waves; but the little cloud is soon withdrawn from observation as it proceeds on its way, and the only thing we can clearly discern is that pollen, like dust raised on a road, ascends in an oblique direction. The form and distribution of the stigmas to be covered with dust-pollen are also in harmony with these conditions. Most plants, whose pollen is in the form of dust, and transported entirely by currents of air, have dioecious or monoecious flowers, and those which develop hermaphrodite flowers exhibit complete dichogamy, that is to say, the androecium and gyncecium ripen at different times, so that when mature pollen is being discharged from the anthers of a flower the stigmas of the same flower are already withered, and therefore no longer in a condition to receive the pollen-cells, or else they are still so immature that they cannot be covered with pollen. Any possibility of the transference of pollen from the anthers to the stigmas situated close to them in the same flower being attended with success is as effectually excluded in dicho- gamous plants as it is in monoecious and dioecious species, and the pollen has to be blown to other flowers in the neighbourhood whose stigmas happen to be in DISPERSION OF POLLEN BY THE WIND. 135 the receptive stage of development. In all these dichogamous plants the flowers with stigmas in the receptive condition are situated higher than the anthers from which mature pollen is being committed to the wind. If you look at any of the species of Plantain (Plantago) a few days after they have begun to flower, you find that only the styles with their stigmas ready to receive the pollen project from the uppermost flowers in each spike, whilst the flowers from which pollen is being shaken by the wind occupy the lower parts of the spike. ^Mi^^M' // i/" %' Fig. 228.— The common Alder (Alnus glutinosa). Branch with flowers that open before the leaves are unfolded; the male flowers grouped in the form of pendent catkins, and above them the female flowers grouped in tlie form of little spikes. 2 Leafy branch at the top of which are the rudimentary inflorescences for the following spring. In these lower flowers the stigmas are already withered, in the upper ones the anthers are still closed. Therefore, in order to reach the receptive stigmas, the pollen must travel upwards. The same conditions are found in most species of Sorrel (Rumex), in the Wall-Pellitory (Farietaria), in Saltwort (Salsola), in Arrow-grass (Triglochin), and in Pondweeds (Potamogeton), and many other plants with hermaphrodite but perfectly dichogamous flowers (cf. figs. 236 and 237). This phenomenon is still more strikingly exhibited by monoecious plants, i.e. where male and female flowers occur on the same individual. In the Oak, the Beech, the Alder, &c., the catkins of mature polliniferous flowers hang down 13b DISPERSION OF POLLEN BY THE WIND. from the branches in the form of swinging tassels whilst the flowers containing mature stigmas are always above them, whether situated on the same or on adjoining branches (c/. fig. 228). In Fir-trees, only the pendent lateral branches of the boughs bear the male inflorescences, which at a distance look almost like red mountain-straw^berries, wdiilst the female inflorescences stand up in the form of little cones on the top of the same boughs like tapers on a Christmas-tree; indeed, many Fir-trees bear the female flowers only on the highest branches close to the summit, and on the lower boughs none but male flowers, and under such circumstances pollen could not possibly reach the stigmas if it were only carried by the wind in a horizontal direction. Even in dioecious plants {i.e. where the male and female flowers are on distinct individuals) this relatively inferior situation of the staminal flowers is often to be observed, the end being attained by the fact that the individuals bearing male flow^ers grow less high than those bearing female flowers. Thus, for example, in Hemp-fields one may see that the plants discharging pollen never reach the same height as those w^hose flowers are to receive the pollen. Exceptions to the rule do, it is true, appear to exist in the Bulrush (Tyjjha), the Bur-reed (Spargaiiium), and many species of Sedge (Carex), which possess monoecious flowers, inasmuch as in them the male flowers are situated above the female; but in consequence of the non-simultaneous elon- gation of the axis, it usually comes about that the mature female flow^ers of a plant whose stem is amongst the older and taller ones rests at a higher level than the male flowers of the individual next to it whose stem is younger and shorter, and it is easy to convince one's self by observation that here also the pollen is not conveyed by the wind in a horizontal direction but obliquely upwards, and is wafted to the stigmas of neighbouring plants. This must not, of course, be looked upon as implying that when pollen is dispersed by the wand none descends; but it is unquestionably true in the majority of cases that the clouds of pollen which are carried off" by moderate winds at first soar upwards and either reach the stigmas awaiting them at a higher level direct in their way, or else, later on, when the air is still and the pollen-cells are scattered over a wider space, they sink slowly down, leaving a deposit on the stigmas, just as when dust is raised in a room it ends by slowly falling again and covering the furniture with a uniform layer. In some species at the very moment when the anthers burst open the pollen is ejected violently into the air and ascends obliquely in the form of a little cloud of dust. In this country a good example of this phenomenon is aflforded by the Nettles. Anyone standing in front of a bed of Stinging Nettles on a bright summer morning, and waiting until the first rays of sunshine fall on the flowers, will be surprised to see small pale-coloured clouds of dust ascending here and there from amidst the dark foliage. At first the clouds are solitary, and are given off" at measurable intervals; by degrees they become more frequent, and at times one may see five or six or more arising at the same moment and at no great distance from one another. But gradually the little explosions become less frequent again, and DISPERSION OF POLLEN BY THE WIND. 137 in another half-hour there is an entire cessation of the phenomenon. On inspection one easily discovers that it depends on the fact that the filaments bearing the anthers are coiled in the bud, and suddenly spring up at the same moment that the dehiscence of the anthers takes place. The species of the genus Parietaria and many tropical Urticaceae behave in the same manner in this respect as our Nettles. As an instance may be taken Pilea viicTophylla (also known under the name of Pilea muscosa), which grows Fi„'. 229— The Papei Mulbeiry tree (Brotissoiietia papynjei a) » Leafy branch with capitulum of female fioweis. - Piece of a branch stripped of its foliage with spike of male flowers. « An unopened male flower in longitudinal section. * An open male flower in longitudinal section; two of the Jilanieuts are still tucked in, one has sprung up and is e.\pelling the pollen from the opened anthers. ' An open male flower with all its stameus already uncoiled and the pollen discharged from the anthers. « xwo female flowers witli long hairy stigmas. 1, 2 natural size; s-s x4-5. ; native in Central America, and is often raised m botanic gardens with a view to j demonstrating the phenomenon here alluded to. One only has to sprinkle the plant with water at a time when it is covered with flower-buds and then take it out of the shade into the sunshine, and the phenomenon is immediately exhibited. All I over the plant the flower-buds explode, and a whitish kind of pollen is discharged I into the air in the form of a little cloud. Many Moreae also display this phenomenon, as, for example, the Paper Mulberry-tree (Broussonetia iDapyrifera), an illustration of whose flowers is given in fig. 229. The male flowers are arranged in spikes (229^), and each flower consists of a sepaloid perianth with four stamens 138 DISPERSION OF POLLEN BY THE WIND. upon it. The filaments are very thick and, in the closed bud, are tucked in (229^): they are in a state of tension like a spring, but as soon as the cup-shaped perianth opens the filaments spring up one after another, whilst at the same instant the anther-cavities burst open and the pollen is ejected with force into the air (229*). When all the anthers are empty the filaments curve backwards (229^), and soon afterwards the entire spike of flowers drops off the axis, it being no longer of any value to the plant. In all these plants ejection of the pollen only ensues when a light, dry wind blows at sunrise and causes an altera- tion in the tension of the tissues con- cerned. If there is no wind at all, or the air is close and damp, or if it rains, the opening of the flowers and ejection of the pollen do not take place, or rather they are postponed until the atmo- sphere has become dry again and a breeze arises which causes the flowering branches to sway about. The results of actual observation are of importance to a proper understanding of the dis- persal of pollen-dust. For it thus ap- pears that the air in motion has to start two processes which supplement ' one another, and must operate in rapid succession if the pollen-dust is to reach the right place and not be lost. The same current of air which causes a liberation and expulsion of the pollen by shaking the flowering axes and by | altering the tension of the tissues of the flowers, also carries the pollen away from the spot where it has been pro- duced and conveys it to its destined goal; and this statement applies to the full extent not only to the case of resilient stamens, but also to all other instances of anemophilous pollination where the pollen is in the form of dust. A similar phenomenon is observed in the case of plants with short, thick filaments and comparatively large anthers filled with pollen of a floury consist- ency. The Phillyrea, the Pistachio-nut (Pistacia), the Box-tree (Buxus), and most Ashes, especially the common Ash (Fraxinus excelsior, see fig. 230), may serve to illustrate this group of plants. The development of the carpels in each flower precedes that of the pollen. At a time when the relatively large , fleshy stigmas stretch out far beyond the limits of the inconspicuous floral I envelope, and are already capable of taking up the pollen, the anthers may be; i'l Fig. 230.— The Ash {Fraxinus excelsior). Small bifurcating branch, the left-hand limb of the fork bear- ing male flowers, the right-hand limb bearing hermaphro- dite flowers. 2 Hermaphrodite flower. 3 Two anthers ; the upper one open, the lower one still closed. i natural size ; 2 8 X 6. DISPERSION OF POLLEN BY THE WIND. 139 seen to be still tightly closed (230^ and 230 -). The latter do not open till two or three — often even as much as four — days later, and only then in the event of the air being dry. Dehiscence is accomplished by longitudinal Assuring of the anther-lobes. The edges of these fissures contract very speedily, so that each of the pair of anther-lobes is converted into an open recess wherein the pollen lies in the form of a floury or powdery mass (230^). Just before dehiscence the Fig. 231. — Avena elatior. I ' A closed anther, a An open anther. ^ Spikelets on a calm day with glumes distended and anthers pendulous. * Spikelets ! in a wind. The pollen escaping from the pendulous anthers in the spikelets to the riglit ; in that to the left (and below> I the anthers (two only remaining) have shed their pollen ; in a third flower (in the same spikelet as the last-mentioned) I the anthers are still closed and in process of being exserted. i, 2 x 12 ; s, * x 5. I janthers place themselves in such a position as to ensure the fissure being turned ■upwards, so that the recesses full of pollen are not emptied so long as tlie air lis still. It is only when the flowering branches begin to sway to and fro that ithe pollen falls out of the loculi and is blown away in the form of a cloud of dust by the same breeze as set the boughs in motion. In another group of plants the anthers are borne on long filaments, and are set oscillating and vibrating by the least breath of wind, the pollen being in 140 DISPERSION OF POLLEN BY THE WIND. ■consequence discharged in little pinches as though from a sugar-sifter. If the flowers of this kind of plant contain pistils as w^ell as stamens, the relative development of the two sets of organs is always so regulated that the stigmas are already perfect and adapted to the reception of pollen at a time when the : anthers of the same flowers are still hidden beneath the floral or involucral ■ envelopes and the pollen is consequently immature. By the time the pollen is ' completely developed and is in a state to be discharged from the opened anthers, the stigmas of the flower in question are withered and are no longer capable of taking up the pollen. Hence it follows that in these plants the pollen-dust must be transported to other flowers which happen to be at a younger stage of development if fertilization is to be brought about. This is what occurs in nature through the instrumentality of gentle breezes wdiich impose a tremulous motion upon the anthers. | In the first rank of plants belonging to the above category stand the Grasses. Their mode of pollination is so remarkable that it is worth while to look into it a little more closely. One group of Grasses — of which Avena elatior, represented in fig. 231, is an excellent example — commences the process under i discussion by a sudden distension of the bracts (knowm by the name of glumes) through the instrumentality of a special turgid tissue situated at their base. The result is that the anthers, till then concealed, are exposed, and it becomes possible for them to be exserted beyond the glumes into the air. This exsertion I is effected by an extraordinarily rapid longitudinal growth on the part of the j filaments. It has been calculated that in some grasses the filiform filaments elongate to the extent of 1-1 "5 mm. in the course of a minute, and that usually in ten minutes they are three or four times as long as they were originally. In one subsection of these plants the filaments grow downwards, in another horizontally, and in a third straight upwards tow^ards the sky. The turgidity of the cells in these delicate filaments is so great as to enable even those which , grow vertically upwards to support the weight of the anthers without bending. In the case of those Grasses whose stamens grow downwards from the beginning it does no doubt look as though this direction were assumed in ' consequence of the weight of the anthers. This is not, however, the fact. A ' high degree of turgidity exists here also, and if one inverts the inflorescences of this kind of Grass, the stamens which have just completed their longitudinal grow^th remain quite stiff", in spite of their extreme slenderness, and project ' straight up. Soon after, it is true, this condition ceases. The filaments become slack; those that were erect nod and droop, those that w^ere horizontal fall : down, and the anthers are then all suspended at the ends of oscillating threads. The dehiscence of the anthers is accomplished synchronously with these changes in the filaments. As long as the anthers lay hidden beneath and pro- tected by the glumes they w^ere straight and linear in form (see fig. 231 ^). Each ; anther consists of two contiguous parallel lobes, and each lobe has a line running longitudinally down it, along which dehiscence takes place. This operation DISPERSION OF POLLEN BY THE WIND. 141 invariably commences after the anther has assumed a pendent position. The filaments and anthers are joined together by a slender connective, and the tissue of this connective is, as it were, articulated so that the anther is capable of turning freely without becoming detached (a condition termed versatile). Hence under any circumstances the requisite position can be assumed; that is to say, the at first uppermost ends of the anthers can be made to hang down whether they are on pendent, or on horizontal, or even on erect filaments. When this inversion has been accomplished the anther-lobes open along the sutural lines already referred to. The slits only gape open for a short distance from that extremity of the anther which is now lowest. This partial opening is in some measure dependent on the further circumstance that at the dehiscent portion the two anther-lobes separate from one another and curve round in opposite directions, as is shown in fig. 231 ^ The significance of this inflection lies in the fact that the powdery pollen is prevented from falling out of the loculi the moment the slits are formed. For the curved ends of the anther-lobes assume the shape of little hollow boats in which the pollen may rest for quite a long time if the air is still (fig. 231^). It is not till a gust of wind sets the anthers swinging that the pollen -dust is blown away in the form of a small cloud (fig. 231*, to the right). On the first occasion only the tiny heap pertaining to the dehiscent extremity of the anther is removed, but this is immediately replaced by fresh pollen pouring down from the upper indehiscent portion of the anther. This new supply naturally has no long time to wait, but is blown away by the very next gust. The process may be repeated several times, and generally does not cease until there is no longer any pollen left. When the anthers are quite emptied they drop oflT the filaments in the form of dry husks. Usually, however, this detachment of the anthers does not take place till several hours after pollination, and in the majority of Grasses, plants which have flowered in the early morning or during the day still have their empty anthers hanging to the spikes or panicles, as the case may be, at sunset. The changes preceding pollination are much more markedly dependent on the weather in Grasses than in other plants. The temperature and hygro- scopic condition of the air in particular play an important part. Rain and low temperatures may delay the splitting asunder of the glumes and the extrusion and dehiscence of the anthers not merely for hours, but for days. A very dry atmosphere accompanied by a high temperature also has the eflect of retarding the processes above described. The most favourable conditions for pollination in the case of most Grasses prevail in the early morning at an hour when there is still some dew lying on the meadows, when the first rays of sunshine fall obliquely upon the flowers, and the temperature is rising gently and a light breeze sets the spikes and panicles in motion. Under such external conditions as these the phenomena of flowering and pollination are accomplislied with astonishing rapidity. In some Grasses an observer may see the glumes relax and spring open, the stamens grow out, the anthers open and the pollen scat- 142 DISPERSION OF POLLEN BY THE WIND. tered, all in the space of a few minutes. The earliest discharge of pollen begins between 4 and 5 a.m. in the height of summer, and the plants which take part in it thus early are the Meadow-grass (Poa), Kaeleria, and Avena elatior. A little later, between 5 and 6 o'clock, comes the turn of the Quaking-grass (Briza media) and Aira cwspitosa, and of Wheat and Barley (Triticum, Hordeum). Between 6 and 7 pollination occurs in Rye and in a great number of different Grasses which grow in meadows, such as Cock's-foot-grass (Dactylis), Andro- 2?ogon, the Brome-grasses {Brachypodiumi), and many species of Fescue (Festuca). Between 7 and 8 o'clock the pollen is liberated from Oats of the Trisetum group, from the Fox-tail-grass {Alopecurus), Timothy Grass {Phleum), and the Sweet Vernal Grass (Anthoxanthum). An interval now intervenes, at least amongst the indigenous Grasses. Of exotic species which are cultivated in gardens the following discharge their pollen in the course of the forenoon, viz. the Millets {Panicurti milliacewm and SorghuTn) between 8 and 9 o'clock; Setaria Italica and the Brazilian Pampas-grass {Gynerium, argenteum) between 9 and 10 o'clock. Towards noon indigenous Grasses come again into play. About 11 o'clock pollination takes place in most species of the Bent-grass genus (Agrostis), and between 12 and 1 in Melic-grass (Melica), Molinia, Mat-grass (Nardus), Elymus, Sclerochloa, and several species of Calamagrostis. In the course of the afternoon the process takes place in a few isolated species, as, for instance, in some Brome-grasses at 2 o'clock, in a few species of Oat (Avena) at 3, in Agropyrum at 4, and in Aira flexuosa between 5 and 6. It is worthy of note that the Soft -grass (Holcus), under favourable atmospheric conditions, opens its glumes, pushes forth its anthers, and liberates pollen twice a day, once in the morning at about 6 o'clock, and a second time in the evening at about 7 — provided always that the temperature of the air is not less than 14° C. The entire process lasts in most cases from 15 to 20 minutes for each flower. With the opening back of the glumes and extrusion of the anthers are often connected alterations also in the position and inclination of the stalks which bear the spikelets. For example, the pedicels of the spikelets of Agrostis, Apera, Calamagrostis, Koeleria, and Trisetum divaricate from the axis, so as to form with it angles of from 45° to 80° for the period of pollination. But as soon as the pollen is discharged all these stalks move back tow^ards the main axis of the inflor- escence, and the panicle, as it were, contracts. These movements are obviously designed to give sufficient room to the anthers when they are exserted, in order that they may oscillate freely and so disperse their pollen. In those Grasses where the flowers are crowded together in close spikes, and also in the large Carex section of the Cyperacese, the bracts do not spring open but only relax, and sometimes merely to such a slight extent that it is scarcely noticeable on cursory inspection. The thread-like filaments are also only partially visible in cases of the kind, the anthers arc pushed forwai'd and raised above the glumes through the rapid growth of their filaments. As soon as a filament reaches the proper length its upper DISPEllSION OF POLLEN BY THE WIND, 143 extremity becomes pendulous, and the anther hangs from it and encounters no obstacle to movements such as are required to shake out the pollen. As in the case of Grasses and Sedges, so also in Hemp and Hops (Cannabis, Humulus), and in numerous species of Sorrel and Meadow-rue (e.g. Rumex alpinus and R. scutatus, Thalictrum alpinum, T. foetidum, T. minus) the pollen-dust is shaken out of anthers which are pendulous at the ends of delicate filaments; only, .^ilS. Fig. 232. -The Elm (Ulrmm campestris). With ilowere. « With fruits. I in these plants not glumes but small perianth-leaves form the protective envelope j round the anthers before they open. Moreover, in Hemp and Hops, and the above- I mentioned species of Meadow-rue, the anther-lobes do not burst wide open when i they dehisce, but exhibit parallel slits which are at first so narrow that the pollen jean only shake out little by little. Plantains (Plantago) also have their pollen 1 shaken out of the anthers, which are borne on long filaments, by the wind. The j filaments are tucked in so long as the flower is in bud, but when the petals unfold the filaments straighten out and project beyond the floral spike. The versatile 144 DISPERSION OF POLLEN BY THE WIND. anthei-s borne by these filaments are broad and for the most part heart-shaped; the two lobes of which each anther is composed only open on the side turned to the sky, Fig. 233.— Mountain Pine (Pintts Pwnilio). *■ A single poUiniferous schle (stamen) seen from above. 2 Three polliniferous scales, one above the other, seen from the side. The pollen falling from each anther alights on the upper surface of the stamen next below, s Xwo spil- ' and cylinders. A short time before the. anthers burst the axis of the spike elongates and becomes pendent, causing all the flowers seated upon it to assume an inverted position with their originally upper faces turned to the ground and their backs upwards. The back of each flower is so contrived as to catch the pollen falling from the anthers of the flowers above it, and retain it until the tassels are set swinging by a gust of wind, and the pollen is in consequence dissipated {cf. vol. i. p. 741). Sometimes the hollow upper surfaces of sepals, petals, or bracts serve as landing-stages for the pollen when it is discharged. This is the case, for example, in various species of the Pondweed genus (Potamogeton), in the Arrow-grass (Triglochin), and the Sea -Buckthorn {Hippopltae). In the Curled Pondweed {Potamogeton crispus), a plant which lives submerged in ponds and slow running .. . PROPERTY OF DISPERSION OF POLLEN BY THE WIND. 149 brooks, and in the height of summer raises its flower-spikes above the surface of the water (see fig. 236), the large, fleshy, reddish-brown stigmas are ah-eady ripe to receive the pollen at a time when the anthers close beside them are still closed. The perianth-leaves of the flowers concerned are indeed still folded together, and may be seen underneath the four projecting stigmatic lobes which are arranged in a cross, whilst the anthers are hidden beneath the perianth. The shortly -stalked, concave perianth-leaves do not open back until the stigmas have begun to wither. Almost at the same instant longitudinal slits are formed down the large, white anthers, and they are speedily converted into gaping fissures, out of which flows a copious supply of yellow pollen of mealy consistency. If a fresh, dry wind is blowing at the moment of the dehiscence of the anthers part of the pollen is at once carried ofi" from the spikes of the Pondweed as they project above the water; but if a calm prevails a certain amount i i- v\.V.I,l/ll/ of the pollen drops into the cavity of the particular perianth -leaf immediately below the anthers. , '* S' Here the pollen may remain for hours together if there is no wind. It is only blown away by a strong puff" of wind, and is then conveyed directly to other spikes projecting out of the water whose flowers happen to be in a much earlier |^ tj' stage of development, the four Fig. 237. -Arrow-grass (rnpZocAui paZ«stre). radiating stigmatic lobes being in 1 a flower with brush-like stigma already mature ; all the anthers ,. f,- r, i. i.r still closed. 2 A flower with the stigma withered whilst the three a receptive condition, but the an- ^^^^^^^^ ^^^-^^^^ have opened and are depositing their poUeu in thers yet indehiscent and the peri- ^^e concave perianth-leav-es at their bases. lu both flowers the •^ ^ lower front penanth-leaf has been cut off. x 8. anth-leaves still closed (see fig. 236). A still more striking instance of the temporary storage of pollen in concave perianth-leaves is found in the Arrow-grass (Triglochin). Here, too, the develop- ment of the stigmas precedes that of the anthers by two or three days. During the whole period that the brush-like stigma at the top of the ovary is sound and in a receptive condition the anthers are closed, and they only open when the stigmas have faded and turned brown (c/. figs. 237^ and 237 2). The stamens, six in number, are in two whorls of three each, situated one above the other (cf. vol. i. p. 646), and underneath each stamen there is a deeply-concave perianth- leaf. As soon as the anther opens the pollen rolls into the receptacle thus prepared beneath it, whilst in the meantime the perianth-leaf has moved a little away from the axis and somewhat loosened its connection with it. The pollen rests in its hollow until a puff" of wind sets the slender floral spikes swaying to and fro and blows away the pollen. It is a noteworthy circumstance that all six anthers of a flower do not open at once, but that first the lower whorl of stamens comes into play, and that after their pollen has been carried away 150 DISPERSIOX OF POLLEN BY THE WIND. by the wind as above described both the empty stamens and the perianth-leaves at their bases drop off. Only after this has happened does the upper whorl of perianth-leaves relax; the anthers of the three upper stamens burst open, their pollen glides into the bowl-shaped perianth-leaves below, and exactly the same process is repeated as took place in the case of the superior whorls. The case of the Sea-Buckthorn {Hippophae; cf. figs. 220-'2'*'^ p. 109), is worth mentioning as a third example of the same nature. The flowers of this shrub are conglomerated in little tufts on the sides of woody branches. Each male flower is composed of four stamens and two opposite concave scales; the latter have their edges in contact, so that they form a little bladder within which the four stamens are concealed. The pollen is of an orange-yellow colour and mealy consistency, and is set free from the anthers at a time when the bladder is still closed. It falls into the cavity, and is there completely sheltered from rain and dew by the overarching scales. When a warm, dry wind sweeps over the shrubs the bladders open by two opposite chinks, and the pollen is blown out from its resting-place in small quantities at a time. In damp weather the two scales close up quickly and protect what remains of the pollen from wet; on the return of dry weather they move apart again, leaving a free passage for the wind, which then carries off" the rest of the pollen. This simple mechanism ensures the safety of the pollen in the event of rain, whilst enabling it to reach the stigmas of neighbouring shrubs whenever the external conditions are propitious. A close connection exists between these various contrivances to ensure that pollination shall only take place at the best possible moments, and the mainten- ance of a free passage in the direction in which the pollen is to be transported by the wind, and further between these adaptations and the shape of the stigmas devised for the reception of the pollen. It is obvious that no barrier must be interposed in the path of the little clouds of pollen-dust on their journey to the stigmas. If the flowers of the Arrow-grass, of Poudweeds, or Grasses were w-rapped in large foliage-leaves a great part of the pollen would adhere to these leaves and would be as irretrievably wasted as if it had fallen to the ground or into the water. On this account also all flowers which have ■ their pollen blown out of them by the wind are arranged in spikes and panicles at the upper extremities of the shoots and project freely into the air, but are never clothed with a mass of foliage. Particular attention may be drawn to \ the fact that a large number of plants wherein the pollen is in the form of \ dust flower before coming into leaf; that is to say, yield up their pollen to '< the wind at a time when the green foliage is still folded up in the buds or is ' just emerging from them. The Sea-Buckthorn, the Alder, the Ash, the Elm, the ' Hazel, the Birch, and the Aspen all flower and discharge their pollen at a season ' when the branches are bare of leaves {cf. the illustrations on pp. 109, 135, 138, 143, and 147). Were these plants to begin to blossom after the complete develop- ment of their extensive foliage the wind-transport of the pollen would be rendered DISPERSION OF POLLEN BY THE WIND. 151 almost impossible. The way to the stigmas would be stopped by innumerable barriers, and the pollen would inevitably be deposited upon these obstacles and stranded. As regards the stigmas, we find that in plants with dusty pollen they are invariably fashioned so as to catch the dust. In one case they are fleshy and swollen and have the surfaces which are exposed to the wind covered with a velvety coating (see fig. 236), in another they are in the form of tufts of long papillose or capillary filaments, as, for instance, in the Paper Mulberry -tree {cf. figs. 229^ and 229 ^ p. 137); sometimes they assume the shape of delicate feathers {cf. fig. 231, p. 139), sometimes of camel's-hair pencils and brushes (fig. 237). At the time when pollination takes place they are always fully exposed to the wind and so placed that when the pollen-cells are blown against them they are caught like midges in a spider's web. Yet, in spite of all these contrivances, it would remain very doubtful whether the stigmas would be dusted with pollen through the action of wind were it not for the concurrence of another circumstance. The wind is but an uncertain means of transport, especially in the case of a passive object incapable of exercising any influence on the selection of a route. It is, therefore, important that the pollen should be disseminated broadcast in as thorough a manner as possible, and this is only possible if the number of pollen-cells is excessively large. Supposing that only two thousand pollen-cells were produced in a Nettle-inflorescence and these were surrendered to be the sport of the wind, it would be only by a lucky chance that a single one of these cells would be caught by the stigmas of a plant at a distance of 5 metres; but, inasmuch as the number of the cells constituting the pollen-dust of a Nettle amounts to millions, the probability of successful pollination is increased to a proportionate extent. If the stami- niferous flowers of Conifers, Hazels, Birches, Hemp, or Nettles be picked before the dehiscence of their anthers and placed on a suitable substratum until the anthers open, the mass of pollen-dust which is liberated is quite astonishing. It seems scarcely credible that so large a quantity of pollen could have been developed in anthers which are themselves so small, and the apparent anomaly only becomes intelligible when one remembers that the cells were packed closely together in the anthers, but afterwards lie simply in a loose heap. In years peculiarly favourable to the flowering of Conifers vast clouds of pollen are borne on gentle winds through the Pine-forests, and are often swept right beyond them, so that not only the female flowers, needles, and branches of the trees in question are powdered over with the yellow pollen, but also the leaves of adjoining trees and even the grasses and herbs of the meadows around. In the event of a thunder-shower at such a period the pollen may be washed off" the plants and run together by the water as it flows over the ground, and then, after the water has run off", streaks and patches of a yellow powder are left behind on the earth, a phenomenon which has given rise on various occasions to the statement that a fall of sulphurous rain has taken place. 152 DISPERSION OF POLLEN BY ANIMALS, DISPERSION OF POLLEN BY ANIMALS. If this book were ornamented with pictorial initial letters illustrative of the contents of each section, we should have at the head of this chapter a group of flowers with bees and butterflies swarming round them, whilst into the scrolls of the capital would be woven a representation of the quiet life of field anil forest as manifested on bright summer days — a subject which plays a prominent part in the poetic descriptions and pictorial art of all unsophisticated nations. Even in these days, pictures of butterflies fluttering about bright-coloured flowers, or of bees engaged in collecting the materials for their honey-combs, still find an appreciative public. Young people especially take pleasure in subjects of the kind, and, since youth never entirely dies out, there will always be people who prefer to see the beautiful lines and tints of flowering meadow and shady wood depicted in miniature than the bold outlines of a landscape. If, however, mere casual observa- tion of the relations between flowers and their insect visitors is sufficient to cause aesthetic pleasure, and has stimulated people of every age and nationality to the production of works of art, it may be imagined how great must be the incentive to scientific study supplied by a deeper insight into these phenomena, and what extreme pleasure is derived from the successful discovery of the reasons for these wonderful relations, and from tracing their connection with other facts of science. It may be confidently asserted that the careful investigation of the processes connected with the visits paid by insects and other animals to flowers has brought the solution of the main problems of modern science considerably nearer, and we have good ground for hoping that the prosecution of these researches will succeed before long in raising the veil which still conceals the truth in the case of a number of unexplained phenomena Zoologists are quite justified in their assertion that many of the developments of insects' bodies are correlated with the forms of particular flowers. But equally true is the conclusion to which botanists have arrived that many of the properties of flowers are likewise in correlation with the shape and habits of flower-seeking insects. Now, these flower-loving animals which would perish if for a single year the earth were destitute of blossoms, vary to an extreme degree in size and shape, in the nature of their external coatings, in what they require for nutrition, and in respect of their time of flight, and of a large number of other habits dictated by soil and climate. From the tiny midges to humming-birds, from the thrips, which are scarcely 1 mm. long, and live and die with the flowers, to the gigantic butterflies of Ceylon, Brazil, and New Guinea, whose expanded wings measure 16 cm. across, and which flutter cumbrously from flower to flower, a long and graduated series extends which corresponds with a perfectly similar series in the floral world. The diversities of colour in the creatures which visit flowers, the various kinds of mechanism of flight exhibited by beetles, flies, bees, butterflies and birds, the multiplicity of organs by means of which they extract their food from DISPERSION OF POLLEN BY ANIMALS. 153 the flowers, their means of attachment to the blossoms, their fur and bristles for brushing off the pollen, have all their corresponding variations in form and colouring amongst flowers, and consequently there is an equally long and apparently parallel series in the realm of plants. Contemporaneously with the opening of the earliest spring flowers occurs the escape of the first pioneer butterflies from their cocoons; the same sunny day which rouses hive-bees and humble-bees from their winter sleep, sees the Willow- catkins protrude from their brown bud-scales and offer their honey and pollen to the world at large. Many flowers which open early in the morning are only visited by particular butterflies which forsake their nocturnal haunts at the same hour; as soon as the flowers close at sunset the insects in question also seek their quarters, fold their wings, and remain the whole night fast asleep. Other flowers do not open till sunset, when day-flying butterflies are already gone to rest, and they are visited by Hawk-moths, Silk-moths, Owlet-moths, and other Noctuse which have remained throughout the day concealed in shady nooks and commence their ramblings when dusk sets in. These instances of the mutual relations existing amongst vital phenomena obtrude themselves annually on the notice of the most superficial observer, and have been described time after time. We need not occupy ourselves any longer at the present day with an account of the facts themselves, but rather with the inquiry into the causes both proximate and remote of all phenomena which are presented to our wondering senses. First of all, the question arises: what is it that induces insects and small birds to visit flowers, and what advantage accrues to a plant from the visits with which its flowers are favoured? The answer is, that the inducement is in some cases care of young, in others the desirability of securing themselves against dangers from storms, and, most commonly of all, it is the craving for food. Flowers, however, do not provide animals with breeding-places, with temporary shelter, or suitable nutriment without claiming a reciprocal service, but have their parts so adjusted that their visitors become laden with pollen, which is then transported to other flowers and deposited on their stigmas where it initiates a series of changes result- ing in the setting of the seeds. The next few pages will be devoted to the eluci- dation and proof of this general answer by aid of individual instances. As regards the choice of nests for their young it has long been known that the nocturnal Lepidoptera of the genus Dianthcecia, and also some species of the genus Mamestra lay their eggs in the flowers of Caryophyllaceous plants, e.g., in those of the Nottingham Catchfly, the Bladder-campion, Ragged Robin, and Common Soap- wort {Silene nutans, Silene injiata. Lychnis Flos-cuculi, Saponaria officinalis). The eggs, which are brought forth through a comparatively long ovipositor, produce tiny caterpillars which move about freely in the undivided cavity of the ovary, and there enjoy not only complete shelter but suitable nutriment, for they live on the ovules and young seeds which are seated upon the central placenta situated in I the middle of the ovary. When they grow up they bite a hole in the side wall oi j the ovary, creep through it and descend to the ground, where they pass into the 154 DISPERSION OF POLLEN BY ANIMALS. chrysalis condition. One may see, frequently, on examining the ripe fruit-capsules of the Catchflies, the perforations by which the moth-larvse have gained their freedom. If the caterpillars of Dianthoecia devoured all the seeds in the ovaries, the species of plants frequented by them would derive no benefit, but, on the con- trary, an injury from their visits. Owing to the large number of ovules, howevei-, they are very seldom completely destroyed, and even if all the seeds in one of the capsules were to be consumed there would always be other capsules in the same plant which would develop plenty of seeds capable of germination. The majorit}^ of the Caryophyllaceous species here in question, the Nottingham Catchfly {Silene nutans, see figs. 238 and 239) amongst the rest, flower at night, their blossoms opening at dusk, remaining expanded all night, and closing at sunrise. This is repeated in the case of each flower at least three times. On the first evening the petals which have hitherto been rolled up and folded in the bud, spread themselves out in rays and bend somewhat back (fig. 239); five anthers are rapidly exserted from the middle of the flower, and these soon afterwards de- hisce, become covered with adhesive pollen, and remain the whole night in that condi- tion. In the course of the following morning the filiform filaments bearing the anthers belonging to the outer circle of stamens bend back, and the anthers fall oft' or, less commonly, are left hanging to the ends of the reflexed filaments in the form of empty shrivelled sacs. The next evening the second whorl of stamens included in these flowers comes into play, and just in the same manner as before, five anthers, which dehisce at nightfall, are exserted from the mouth of the flower and expose their pollen. The third day these stamens likewise bend back and usually let their anthers drop, and when dusk sets in the long velvety S-shaped stigmas, which have till then been concealed inside the flower, are pushed out. Certain changes of position aflecting the petals proceed simultaneously with these mutations. It has already been mentioned that the petals rolled up in the bud unfold on the first night, and assume a stellate and re- flexed attitude. At this time also the flowers emit a delicate perfume like that of hyacinths, which attracts a large number of nocturnal insects, but only lasts from 8 o'clock in the evening till about 3 a.m. At daybreak the petals begin to roll up again, the operation taking place faster when the temperature is moderately high and the sky clear than when the weather is cold and the sky overcast. In the pro- cess of involution the petals fall into longitudinal folds and become wrinkled and grooved, so that they hang like five crumpled bags round the mouth of the flower, and Fig. 238.— The Nottingliam Catchfly {Silene nutans) in the daytime. DISPERSION OF POLLEN BY ANIMALS. 155 by their appearance might lead one to think that the flower had faded (see fig. 238). But as evening approaches the wrinkles vanish, the petals unfold, spread tliemselves out into a star, and become slightly reflexed once more. One peculiarity of these flowers is that the inner surface of the petals is white, whilst the outer surface is always of some inconspicuous colour, such as dirty -yellow, greenish, brown, dull red, or ashen-grey. Hence the radiating petals with their white inner surfaces exposed are very striking in the evening darkness, whereas in the daytime the crumpled petals with only their backs visible are anything but conspicuous, and give the impression of being already brown and withered, as may be seen in fig. 238. They are consequently not noticed by insects in the daytime and receive no / *** visits from them. This appears to be exactly what is aimed at. Such insects as visit flowers by day in order to suck their honey would be the reverse of welcome to the Catchfly. The filaments are reflexed, the anthers shrivelled and empty or dropped, and there is no pollen in the flower to be brushed oflf. A honey - sucking insect could not either take up or deposit pollen in the daytime, and the honey would therefore be sacrificed in vain. Indeed, the flowers would be worse off" inasmuch as, being despoiled of their honey, they would possess one less means of attrac- tion in the ensuing night. On the approach of night the pollen-laden an- thers and velvety stigmas appear in front of the entrance to the interior of the flower where the honey is concealed, the scent and white colour act as allurements, and the visits of insects are welcome, provided the size of their bodies is such that they rub against the pollen or stigmas and fly quickly from one flower to another. Those which are too small, or are destitute of wings, are still kept at a distance, this being eflfected by means of contrivances which will be the subject of discussion later on. Of all the welcome species the best adapted in respect of size and shape of body, length of proboscis, and various other structural characteristics are the Owlet Moths (Noctuas), and of these in par- ticular those of the genus Dianthoecia, one of which is represented as visiting the flower of the Nottingham Catchfly in fig. 239. These little moths pay frequent visits to suck the honey whilst the females also lay their eggs in the flowers. It sometimes happens, too, that the females become loaded with pollen from a flower upon which they have rested and taken a meal of honey, and that afterwards they fly with the pollen to other flowers where, instead of sucking any mere honey, they li_ _ > 111, \ t'li _h nil Cif titl\ (S!/f)! J I /(/( )1 \ iiuht a flower being \isited b> the moth Dianthacia albimanda 156 DISPERSION OF POLLEN BY ANIMALS. lay their eggs, and in so doing dust the stigmas with their freight. To sum up, the flowers of the Nottingham Catchfly and of other species of Caryophyllaceae above referred to are adapted to the small Noctuse of the genera Dianthoscia and Mmnes- tra, and are visited exclusively, or, at any rate, principally, by those insects. The NoctujB obtain honey from them, and the females find in them homes suitable for their eggs. The return made by the moths to the plants consists in the conveyance of pollen from flower to flower and the consequent conversion of ovules into seeds which would not be effected spontaneously. The relations just described occur also among several other groups of plants and Lepidoptera. A number of species of the small blue butterflies belonging to the genus Polyommatus stand in the same relation to Leguminosae and Rosaceae. The beautiful Folyommatus Hylas visits the flowers of Lady's-fiugers {Anthyllis V al- neraria) and in doing so transfers the pollen from one plant to another. The female lays her eggs in the ovaries of the flowers she visits, and from the eggs issue cater- pillars which feed on the young seeds. When mature the caterpillars forsake the ovaries and retire underground to pass through the chrysalis stage. The same relation exists between Polyommatus Bceticus of Southern Europe and the Bladder- Senna {Colutea arborescens), between Polyom^matus Areas and the Great Burnet (Sanguisorba officinalis) and in many other cases; only, besides the butterflies named, others alight with a freight of pollen on the flowers of these plants, but do not lay eggs in the ovaries, and only receive honey in return for their conveyance of the pollen, so that these cases are really only partially of the same category. On the other hand, the life-history of one of the moths living on the capsule- bearing species of the genus Yucca, and named Pronuba yuccasella, has been made out, and must here be dealt with in some detail, as it aflfords one of the most wonderful examples of the dispersal of pollen by means of egg-laying insects. The flowers in all species of Yucca are arranged in large panicles (vol. i. fig. 154, p. 659), and each is bell-shaped and suspended at the end of a smooth, green stalk. The perianth-leaves, six in number, are yellowish-white and are consequently visible from a considerable distance in the dusk and on moonlight and starry nights. After the flower-buds open, which happens regularly in the evening, the perianth forms a widely-open bell (c/, fig. 240^). The dehiscence of the small anthers, which are supported on thick and velvety filaments, takes place simultaneously with the divergence of the petals, and a golden-yellow adhesive pollen is to be seen in the spiral slits of the anthers. Each flower is wide open for one night only; by the next day the free extremities of the six perianth-leaves bend towards one another causing the flower to assume the form of a balloon or bladder with six narrow lateral apertures (fig. 240^). In the twilight and by night, numerous small yellowish- white moths (Pronuba yuccasella; see fig. 240*) which have a metallic glitter in the moonlight flutter about the flowers of the Yucca plants. The females penetrate into the interior of the wide-open bells and there endeavour to possess themselves of the pollen, not with a view to devouring it, but that they may carry it away. For this purpose they are furnished with a special implement. The first DISPERSION OF POLLEN BY ANIMALS. 157 joint of the maxillary palp is lengthened to an extraordinary extent, and its inner surface is beset with stiff bristles and can be rolled up like a trunk (see fig. 240 ^ ). It is used to seize the pollen, to conglomerate it into a ball and afterwards to hold Fig. 240.— Transport of Pollen by Egg-laying Insects. ' Branch from the inflorescence of Yucca Whipplei; the middle flower open, that beneath it was open the previous night and i» now closed again, the rest of the flowers in bud. 2 Single flower of the same plant visited by a moth of the species Pronuba yuecasella ; the three front perianth-leaves removed. 3 stigma of Yucca Whipplei. * Pronuha yuccasella flying to a flower of Yucca Whipplei. s Head of Pronuba yuccasella with a ball of pollen held by the coiled maxillary palp. « Twig with inflores- cence of Ficuspumila ; the urn-shaped inflorescence (or syncouium) cut through longitudinally. ? Single female flower from the bottom of the syuconium of Ficus pumila. », 9 Stamens of the same plant from the upper part of the synconiuni. 1" Synconium of Ficus Carica full of gall-flowers produced by Blastophaga, cut through longitudinally; near the mouth of the cavity is a Fig-wasp (Blastophaga grossorum) which has escaped from one of the galls, n Synconium of Ficus Carica full of female flowers, cut through longitudinally ; near the mouth of the cavity are two Fig-wasps, one of which has already crept into the cavity wliilst tlie second is about to do so. 12 Male flower, "a Long-styled female flowers of Ficus Carica. h Gall produced from a short-styled gall-flower, is Blastophaga grossorum escaping from a gall, is .A. liberated Blastophaga. n The same magnified. 1, ^ 4, 6^ 10, 11, le, natural size; s x 2; 3x20; 7, «, 9, 12, is x 5; i*, is, w x 8. it. In a very short time a moth collects by its means a ball of pollen, which is held by the rolled-up palpi close underneath the head and resembles a great crop. Laden with this lump of pollen, which is sometimes three times as large as its head, the 158 DISPERSION OF POLLEN BY ANIMALS. moth abandons the despoiled flower and seeks another forthwith. Having found one, it circles nimbly round it, making a sudden spring off and on, and ends by settling on two of the thick reflexed filaments, sprawling its legs out upon them. It then seeks to reach a favourable spot on the surface of the pistil with its ovipositor and there deposits its eggs. The ovipositor is composed of four horny bristles, and is adapted to pierce through the tissue of the pistil. After the eggs are laid and the ovipositor is withdrawn, the moth darts to the top of the infundi- buliform stigma (fig. 240^), unrolls its trunk-like palpi, and stufis the pollen into the stigmatic funnel, moving its head to and fro repeatedly during the operation {fig. 240 ^). It is alleged that the same moth repeats the processes of alternately laying eggs and stuffing the stigma with pollen several times in the case of the same flower. Most of the eggs introduced into the pistil are deposited in the vicinity of the ovules. They are of oblong shape, narrow and transparent and increase rapidly in size, soon revealing in each a coiled-up embryo. On the fourth or fifth day the larva is hatched and at once begins to devour the ovules in the cavity of the ovary. Each grub requires from 18 to 20 ovules to nourish it during the period of its development. When it is grown up, it bites a hole in the still succulent wall of the ovary, crawls out through the aperture, lets itself down to the ground by a thread, burrows into the earth and spins an oval cocoon underground in which it remains till the following summer. Fourteen days before the time of flowering of the Yucca, it begins to show signs of life, and the moment the flowers of that | ■. plant open the silvery moths escape from their pupal envelopes. I An important element in the interpretation of the relations subsisting between the Yucca and the Fucc