e 1° 1 1 THE WATERLILIES CARNEGIE INSTITUTION OF WASHINGTON. WATERLILIES, PLATE 1. NYMPH/EA GIGANTEA VIOLACEA. «* THE WATERLILIES £ A MONOGRAPH OF THE GENUS NYMPHAEA BY HENRY S. CONARD Senior Harrison Fellow in Botany University of Pennsylvania Published by The Carnegie Institution of Washington 1905 mici CARNEGIE INSTITUTION OF WASHINGTON Publication No. 4 * ZU Barb Qgafttmor* (Jptiet THE FRIEDENWALD COMPANY KAI.T1MORI. MD.. U.S. A. ?^ PREFACE. This work is the outcome of four years' study on the Harrison Foundation in the University of Pennsylvania. It would have been difficult or impossible without the generous support given by the Provost and the Board of Trustees. In the University Botanic Garden over thirty species and varieties of waterlily were cultivated, and several hybrids originated. These and other collections of living plants furnished excellent material for the present review of the group. Nothing like a complete synopsis of the Waterlilies has hitherto been put before the English-speaking world. DeCandolle made a special study of the genus Nymphaea for his Prodromus, and both Lehmann and Planchon gave much attention to the genus from 1850 to 1853. Beginning in 1855 and continuing until 1888, a number of exhaustive treatises by Robert Caspary went far toward perfecting our knowledge of these plants. Indeed, had the learned doctor of Konigsberg assembled his vast knowledge into one connected whole, the present work would be needed chiefly as a translation. Since his day, however, great strides have been made in the culture and hybridization of water- lilies, with the result that they now appear in almost every collection of plants. Already by crossing and recrossing and the use of horti- cultural names, serious confusion exists concerning the identities of the parent species. It has therefore seemed important to bring together the knowledge of the genus in all of its botanical relations and in its bearings on human life and history. In the present attempt there may be many omissions, in spite of the utmost efforts in collating the material, but the paper may serve to show how incomplete is our knowledge of any one group of plants, and it is hoped that it may make the way of the future investigator easier. In the adoption of the generic name Nymphaea, the writer regrets that he is at variance with a number of American botanists for whom he has the utmost respect ; but a glance at the synonymy of Nymphaea, in iv Preface. page 125, shows the leading facts. It seemed patent to the writer that all of the general, and several special, arguments against a change of name are particularly cogent in the present example. The new code, however, proposed by a committee of the Botanical Club of the American Association for the Advancement of Science, seems to sanction the nomen- clature here chosen.1 In the use of specific names, authorities have been omitted wherever possible, since the names adopted in Chapter V are followed. Citations of literature are based on the bibliography on pages 243 to 263. Throughout the time occupied by the preparation of this work, I have been deeply indebted to Professor John M. Macfarlane, to whom I desire to express my gratitude. To Mr. William Tricker, formerly of Riverton, N. J., and through him to the Henry A. Dreer Co., of Philadelphia, I am under obligation for the privilege of studying their magnificent collection of waterlilies. Acknowledgment is also due to Dr. J. W. Harshberger, of the Uni- versity of Pennsylvania, and Dr. J. N. Rose, of the National Herbarium, for suggestions and herbarium material ; to Mr. A. J. Mela, of Helsingfors, Finland ; J. J. Soar, of Little River, Fla.; Wm. Gollan, Saharanpur, India ; Wm. Fawcett, of Kingston, Jamaica ; R. M. Gray, gardener to O. Ames, North Easton, Mass. ; Gopal Chandradatta, Calcutta, India ; Sr. Alberto Lofgren, director of the Botanical Garden of Sao Paulo, Brazil ; the director of the Royal Botanic Garden of Tokyo, Japan ; to the Gray Herbarium, Boston Natural History Society, Massachusetts Horticultural Society, Columbia University ; to the Academy of Natural Sciences of Philadelphia for library facilities; and to John B. Lieberg, Joseph H. Painter, Chas. C. Abbott, Prof. Conway MacMillan, Albert J. Edmunds, E. D. Sturtevant, and C. G. Pringle for various courtesies. Mr. Frederick V. Coville and Mr. J. C. Bay, of Washington, D. C, have also given helpful aid and criticism while the matter was going through the press. In order to give the greatest possible accuracy to the taxonomic portion of the monograph, I was enabled by the trustees of the Carnegie Institution of Washington to examine important type material in several European herbaria, viz, in Kew, British Museum, Linnaean Society, Museum d' Histoire Naturelle, the herbaria of De Candolle, Delessert, and Barbey-Boissier, and the herbaria of Munich, Berlin, and Copenhagen. The results of this trip were very valuable, thanks to the courteous assistance 'Bulletin Torr. Bot. Club, 31: 249-290. May, 1904. Preface. v of those in charge of the various collections. The principal changes resulting in the manuscript were in the addition of a few species. That the book appears in its present form is also due to the Carnegie Institution, on account of their acceptance of the manuscript for publication. Were it not for this privilege, much of the illustrative material must have been omitted. These many obligations I have felt keenly, and have tried in a measure to meet by conscientious care in the execution of every detail. University of Pennsylvania, HENRY o. CONARD, Philadelphia, December i, 1904. Senior Harrison Fellow. CONTENTS. VI PAGE Preface iii-v Description of Plates ix-xi List of Text Figures xii-xiii I. History 3 Oriental Literature 3 Pre-Linnaean Literature 10 II. Structure 27 The Root 28 The Stem 37 The Leaf 54 <--The Flower, Fruit, and Seed 77 III. Development 95 Organogeny and Embryology 95 Germination and Early Growth. ... 106 IV. Physiology 113 V. Taxonomy 125 Nymphaeae Apocarpiae 127 Anecphya 127 N. gigantea 128 Brachyceras 131 N. elegans 131 N. ampla 134 N. flavo-virens 137 N. stellata 140 N. caerulea 141 N. micrantha 146 N. heudelotii 147 N. ovalifolia 150 N. calliantha 151 N. capensis 153 N. sulfurea 161 N. stuhlmanni 161 page V. Taxonomy — Continued. Nymphaeae Syncarpiae 163 Castalia 163 Xanthantha 163 N. mexicana 163 Chamaenymphaea 167 N. tetragona 167 N. fennica 170 Eucastalia 171 N. Candida 172 N. alba 175 N. odorata 179 N. tuberosa 189 Lotos 192 N. lotus 194 N. zenkeri 197 N. pubescens 198 N. rubra 199 Hydrocallis 200 N. amazonum 200 N. rudgeana 204 N. blanda 206 N. lasiophylla 207 N. gardneriana 208 N. jamesoniana 209 N. stenaspidota 209 N. tenerinervia 210 N. oxypetala 210 N. gibertii 210 VI. Distribution 213 Geographical 213 Geological 215 VII. Hybrids and Garden Varieties 219 General 219 Taxonomic 225 VIII. Culture and Uses 235 Bibliography 243 Index 265 /x DESCRIPTION OF PLATES. Plate i. Plate 2. Plate 3. Plate 4. Plate 5. Frontispiece. Nytnphaea gigantea violacea. in the Cape York Peninsula, Australia. From a painting made by Mrs. E. Rowan, Plate 6. Nytnphaea odorata. 1-7, development of leaf and stipules; 1, 2, front and back views of the same rudiment; 3, older; 4, 5, front and side views of one rudiment; 6, 7, frcnt and back views; 8, rhizome with roots and leaves trimmed off, natural size; 9, trans- verse section of rhizome, natural size. A. Koronski del. Nymphaea gigantea. Under side of leaf and details of flower. 1, anterior sepal ; 2, outer- most stamen ; 3, inner stamens ; 4, median petal ; 5, vertical section of ovary ; 6, outer- most petal. Photograph from plant grown at University of Pennsylvania. Nymphaea elegans. Flower and under side of leaf. From plant grown at University of Pennsylvania ; stock of the Dreer Co. A. Koronski del. Nymphaea ampla, from Caspary, 1878, pi. 28. 1, flower from hb. St. Petersburg, coll. B. Jaeger, No. 233, Hayti, with 149 stamens, 24 carpels ; 2, leaf from hb. Geo. Engelmann, coll. Wright, No. 1857, Cuba; 3, flower, seen from the side; 4, rays of stigma and carpellary styles; 5, stamens (2-5 from the same specimen) ; 6, two carpels of a flower in hb. St. Petersburg, coll. Karwinski, 1841, in Colipa River, Mexico. Nymphaea Havo-virens (N. gracilis Hort, not Zucc, 1832). Flower and under side of leaf, Photo, from plant grown at University of Pennsylvania, from stock obtained from the Dreer Co. Plate 7. Nymphaea stellata versicolor. Under surface of leaf and details of flower, a, sepal ; b-e, petals of successive whorls; f, stamens; g, closed flower. From plant grown by the Dreer Co. (Mr. W. Tricker), from seed sent by W. Gollan, Saharanpur, India. Plate 8. Nymphaea caerulea. Pennsylvania. Flower and under side of leaf. From plant grown at University of Plate 9. Nymphaea micrantho. 1, under side of leaf; 2, upper side with young plant developing; 3 caudex, with a branch ; 4, scars of leaves and of a flower, from surface of a caudex, enlarged ; 5, cross section of caudex ; 6, longitudinal section of a caudex ; 7, superficial tangential section through leaf cushions of Fig. 4; /, base of a petiole; f, base of a peduncle; 8, section a little deeper than 7; 9, same as 8, a little deeper. From a drawing in color by R. Caspary, now in hb. Berlin, by kind permission of the Directors. Plate 10. Nymphaea capensis. Under side of leaf. From plant grown at University of Pennsyl- vania; stock from Dreer Co. Plate ii. Nymphaea capensis zanzibariensis. Flower and leaf. Plant grown at University of Pennsylvania; stock of Dreer Co. A Koronski del. x Description of Plates. Plate ia. Nymphaea tulfurea. I, s, leaves viewed from the upper surface; 2, under side of leaf; 3, a closed flower; 4, fully open flower; 6, rhirome. From a co-type in hb. British Museum, by courtesy of the Curators. P. Highly del. Plate 13. Nymphaea mexicana. Flower and leaf, natural size. From plants grown at University of Pennsylvania, from roots collected in Little River, Fla., by Prof. J. M. Macfarlane. A. Koronski del. Plate 14. Nymphaea tetrogona. Flower and under side of leaf, natural size. From plants grown by the Dreer Co. Plate 15. Nymphaea alba rubra, 1-21 ; N. Candida, 22. Nymphaea alba rubra. I, partly opened flower; 2, fully open flower on its last day; 3-8, petals; 9-12, stamens of successive whorls from outside inward, viewed from front, side and back; 13, ovary stripped of sepals, petals and stamens, from the side; 14 stigma and styles from above; 15, section of ovary ; 16, pollen, from pole and from side ; 17-19, fruits ; 20, petal ; 31, stigmatic lobes. Nymphaea Candida. 22, flower cut in half. From drawings in color by R. Cas- pary, now in hb. Berlin, by kind permission of the Directors. Plate 16. Nymphaea lotus. Leaf, flower, transverse and vertical section of fruit. Photo- graphed by P. Highley from a specimen coll. Schweinfurth, Nov., 1887, at Damietta, Egypt. >n hb. British Museum, by courtesy of the Curators. Plate 17. Nymphaea pubescens. Flower and under side of leaf. From a plant grown by the Dreer Co. (Mr. Wm. Tricker), from seed sent by Mr. W. Gollan, Saharanpur, India. Plate 18. Nymphaea rubra. 1, flower of second day, between 9 and 9.30 a. m. ; 2, flower at same time of third day; 3, flower at same time of first day; 4, flower of fourth day, be- tween 8.45 and 9 a.m.; 5, sepal from outer and inner sides; 6, petals of first (at left) and second whorls; 7, petal of innermost whorl; 8-10, stamens from outer and inner sides; 11, 12, dry pollen; 13-15, pollen germinating in sugar solution; 16, 17, pollen, examined in citron-oil, 18, 19, in water, 20, in dilute sulphuric acid; 21, germinating pollen; 22, section of carpellary style; 23-26, hairs of ovary, 24 treated with potash; 27, pistil seen from above ; 28, ovary of Fig. 3, with stamens ; 29, ovary on second day. From color drawings by R. Caspary, in hb. Berlin, by kind permission of the Directors. Plate 19. Nymphaea amazonum. 1, closed flower after its first night; 2, fully open flower, second night between 4.30 and 4.45 a. m.; 3, the same flower on the first night be- tween 4.30 and s a. m. ; 4, sepal, inner side ; 5, petal of outermost whorl, from with- out; 6, 7, 8,9 petals of the next four whorls; 10-13, stamens of successive whorls; 14, is, side and section views of ovary after the second night; 16, ovary after the first night; 17, transverse section of ovary; 18, 19, pollen, dry and in water; 20, epidermis of ovary. From color drawings by R. Caspary, in hb. Berlin, by kind permission of the Directors. Plate 20. Nymphaea rudgeana. 1, flower on its first night, 10.15 P- rn. '• 2, the same between 9.20 and 9.40 p. m. ; 3, 4, ripe fruits ; 5, closed flower after first night ; 6, flower with 3 outer whorls of floral leaves removed; 7, sepal from within; 8, petal of outermost whorl; 9, petal of second whorl; 10, petal of fourth whorl; 11, stamen of outermost whorl of four; 12, stamen of second whorl of four; 13, stamen of third whorl of seven or eight; 14, 15, inner and innermost stamens; 16-19, views of ovary; 20, 21, 22, pollen in dilute sulphuric acid, dry, and in water; 23, ripe fruit with floral leaves removed; 24, transverse section of ovary; 25-27, sections of carpellary style at apex, just below the club-shaped apex, and at base ; 28, papilla; of stigma. From color drawings by R. Caspary, in hb. Berlin, by kind permission of the Directors. Description of Plates. XI Plate 21. Nymphaea blanda, N. blanda fenzliana. N. tenerinervia, N. stenaspidota. From Caspary, 1878, pi. 36. 1-7, N. tenerinervia, from a specimen coll. Martius at Joazeiro, Province of Bahia, Brazil. 1, flower; 2, lower surface of leaf; 3, apex of outer stamen; 4, of innermost stamen ; 5, carpellary style ; 6, upper surface of leaf, near the point of inser- tion of the petiole, showing series of raised dot1; and short raised lines; the arrow points to the insertion of the petiole ; 7, outer surface of sepal showing short crowded raised lines ; 8, outer surface of sepal of N. stenaspidota with raised dots between the rather long raised lines; 9-12, N. blanda fenzliana, specimen from San Juan tie Nica- ragua, in hb. Vienna. 9, young fruit; 10, 11, outer and median stamens, from hb. Ben- tham; 12, leaf; 13-16, N. blanda. 13, flower viewed from side, coll. Parker in British Guiana; 14, leaf of the same; 15, part of upper surface of same leaf, showing raphid-like lines and raised dots; 16, leaf of Meyer's type, coll. Dr. Rodscheid in the Essequibo River (from hb. Grisebach). Plate 22. Nymphaea lasiophylla and N. rudgeana, from Caspary, 1878, pi. 34. 1-7, N. lasiophylla. I, 2, flower from face and back; 3, leaf, showing lower side, with distant lobes; 4, part of a larger leaf seen from above, lobes touching ; 5, 6, outermost and innermost sta- mens; 7, part of lower surface of leaf marked with raised dots and branching lines; 8-19, N. rudgeana. 8, leaf of a plant from Jamaica, cult, in Berlin Botanical Garden; 9, 10, II, stipules of three leaves of a Jamaica plant, cult, in Konigsberg Botanical Garden; 12, fruit, raised at Konigsberg ; 13, another fruit, raised at Konigsberg, with persistent sepals and petals; 14, Konigsberg fruit in median section; 15, small flower, coll. Poep- pig, No. 3033, 1832, in the Island of Colares of the Amazon, in hb. Berlin; 16, sepal of a flower, coll. Sello at Victoria, Brazil; 17, 18, stamens of this flower; 19, apices of median and innermost stamens showing a narrow rim of the connective, but no append- age. Plate 23. Nymphaea jamesoniana, from Caspary, 1878, pi. 32. 1, 2, flower from front and back, in hb. Boissier; s, sepal; 3, 4, 5, innermost, outer and outermost stamens; 6, 7, leaves, from hb. St. Petersburg and hb. Lehm; 8, part of leaf from beneath, showing forked branching lines and raphid-like idioblasts. Plate 24. Nymphaea stenaspidota and N. gardncriana, from Caspary, 1878, pi. 33. 1-3, AT stenaspi- dota. 1, whole plant from type of hb. Lehm ; s, sepal ; 2, flower ; 3, part of leaf, upper surface, showing raised dots, the arrow points toward point of insertion of petiole; 4-n, N. gardneriana. 4, flower, from hb. Vienna; s, sepal; 5, 6, 7, upper part of outermost, median and innermost stamens; 8, 9, 10, leaves; 11, upper surface of leaf showing series of raised dots, with raphid-like lines interspersed, the arrow as in Fig. 3. Plate 25. Nymphaea oxypetala, from Caspary, 1878, pi. 31. I, 2, flower, from hb. Boissier, from face and back ; 3, 4, outermost and innermost stamens ; 5, three carpellary styles, with surface of stigma and axile process; 6, 7, leaves from hb. Lehm. and St. Petersburg; 8, globose idioblasts of inner face of anther at the point of Fig. 3. Plate 26. Nymphaea Pennsylvania = N. caerulea X zanzibariensis. From the original plant grown at University of Pennsylvania. A. Koronski del. Plate 27. Nymphaea capensis X zanzibariensis. Flower and under side of leaf. From plants raised at University of Pennsylvania. Plate 28. Nymphaea elegans X zanzibariensis. Flower and under side of leaf. From plants raised at University of Pennsylvania. Plate 29. Nymphaea Havo-vircns X zanzibariensis. Flower and under side of leaf. From plants raised and bred at University of Pennsylvania. Plate 30. Nymphaea omarana, hybrid. Stock from Dreer Co. A. Koronski del. LIST OF TEXT FIGURES. PAGE Tail-piece. — Guest at a feast, with lotus (after Wilkinson) i Figure I. — Nefer Turn (after Wilkinson) 6 2. — Egyptian lotus (designs after Wilkinson) 7 3. — Pleasure boat in a pond of lotuses (after Wilkinson) 7 4. — " Faba Aegyptia." Facsimile from Matthiolus 15 5. — " Nymphaea alba." Facsimile from Lobelius 15 6. — Longitudinal section of root tip of N. caerulea 29 7. — True and false epidermis of roots 30 8. — Root of N. odorata, transverse section 32 9. — Contractile root of N. Aavo-virens, transverse section 33 10. — Vascular bundles of the roots of N. odorata 36 1 1. — Development erf air-canals 37 12. — Diaphragms in the air-canals of roots 38 13. — Rhizome of N. tuberosa 39 14. — Tubers 40 15. — Perennating body of N. mexicana 42 i6j — Rhizome of N. odorata; transverse section of internode 43 17. — Microscopic section of rhizome of N. odorata 44 18. — Caudex of N. caerulea; transverse section 46 19. — Cork formation at base of root scar 47 20. — Details of leaf scar 48 21. — Transverse section of caudex of N. flava SO 22. — Course of vascular bundles at node of Nymphaea flava 51 23. — Vascular system of stolon of N. rubra? at a node 52 24. — Stipules 55 25. — Collenchyma in petioles 57 26. — Sections of petioles and peduncles 59 27. — Types of vascular bundle 63 28.— Leaf of N. ftava 65 29. — Epidermis of leaf, N. tetragona 67 30. — Leaf of N. rubra, vertical section 68 31. — Arrangement of vascular bundles in veins of leaf 71 32. — Development of mucilage hair of N. alba 72 33. — Hairs and hair-bases 73 34. — First leaves from germinating tubers 75 35. — Transverse sections of petioles of water-leaves 76 36. — Floral diagram for Nymphaea 80 37. — Details of anther and stigma 84 38.— Pollen 85 39. — Diagram of ovary (Brachyceras) in longitudinal section 86 40. — Diagram of ovary in transverse section 86 41. — The aril 91 42. — Sections of outer cell-layer of testa (after Weberbauer) 92 43. — Distribution of vascular bundles from the chalaza (after Chifflot) 93 xii List of Text Figures. xiii PAOB Figure 44. — Longitudinal section of seed 94 45. — Development of flower 99 46. — Ovule of N. odorata from a half-grown bud 102 47. — Embryo sac of N. odorata 102 48. — Embryology 105 49. — Seedlings of a hybrid Nymphaea of the Lotos group 108 50. — Seedling leaves no 51. — First broad (submerged) leaf of N. lotus seedling 112 52. — A strong-growing waterlily in shallow water 1 19 53. — Nymphaea elegans, details of flower 132 54. — Nymphaea Aavo-virens, details of flower 138 55. — Nymphaea caerulea, details of flower and fruit 142 56. — Nymphaea heudelotii, flower and leaves 148 57. — Nymphaea heudelotii nana, leaf and flowers 149 58. — Nymphaea ovalifolia, leaf and flower 150 59. — Nymphaea calliantha, leaf and flower 151 60. — Nymphaea copensis, details of flower 154 61. — Nymphaea capensis zanzibariensis, details of flower 158 62. — Nymphaea stuhlmannii, leaf and flower 162 63. — Nymphaea mexicana, details of flower and fruit 164 64. — Nymphaea mexicana, seedling leaves 165 65. — Nymphaea tetragona, details of flower 168 66. — Nymphaea Candida, leaf 173 67. — Nymphaea odorata, details of flower and leaf 181 68. — Nymphaea odorata minor, details of flower 184 69. — Nymphaea odorata gigantea, details of flower 187 70. — Nymphaea tuberosa, details of flower 189 71. — Nymphaea tuberosa, tubers and water-leaves 190 72. — Nymphaea tuberosa, details of leaf 191 73. — Floral organs of the Lotos group 193 74. — Early leaves of the Lotos group, from germinating tuber 195 75. — Nymphaea zenkeri, flower and leaf 197 76. — Nymphaea dumasii, leaf (after Schenck) 216 77. — Nymphaea pulcherrima, details of flower 224 78. — Nymphaea Pennsylvania, details of flower 225 79. — Nymphaea capensis X zanzibariensis, details of flower 226 80. — Nymphaea elegans X zanzibariensis, details of flower 226 81. — Nymphaea " Wm. Stone," details of flower 227 82.— Nymphaea " Mrs. C. W. Ward',' details of flower 228 ." THE WATERLILIES BY HENRY S. CONARD 3 CHAPTER I. HISTORICAL. ORIENTAL LITERATURE. Man was probably impressed from the first with the grace and beauty of the waterlilies ; for their distribution is worldwide. There is evidence enough that the seeds and roots (tubers) of various species have been used for food by many savage races, as in Australia, Madagascar, West Africa and Central America ; but naturally it remained for more civilized peoples to rise to anything like an aesthetic interest. A Chinese writer of the eleventh century a. d., Chou Tun-I, is thus quoted in Giles's " Gems of Chinese Literature," — Since the opening: days of the T'angr dynasty [600 a. d.], it has been fashion- able to admire the paeony ; but my favorite is the waterlily. How stainless it rises from its slimy bed ! How modestly it reposes on the clear pool — an emblem of purity and truth ! Symmetrically perfect, its subtle perfume is wafted far and wide ; while there it rests in spotless state, something to be regarded reverently from a distance, and not to be profaned by familiar approach. In my opinion, the chrysanthemum is the flower of retirement and culture ; the paeony, the flower of rank and wealth ; the waterlily, the Lady Virtue sans pareille. Alas ! few have loved the chrysanthemum since T'ao Yuan-ming ; and none now love the waterlily like myself ; whereas the paeony is a general favorite with all mankind. The waterlily of China is the little Nymphaea tetragona, which, though pure white and of pleasant scent, never attains to much more than a pigmy size ; and its odor is not nearly so rich as that of our own N. odorata. It is interesting also to note that our author is ex- pressing only a personal opinion, one which is not at all general, and which indeed may be regarded as heretical. India is much richer in gorgeous varieties of waterlily than any other portion of the globe. By day the blue N. stellata exhibits several shades of color, even running off by degrees into pink in parts of Bengal. At night and early in the morning the great white N. lotus and the deep crimson N. rubra expand their flowers 6 to 10 inches in 3 4 The Waterulies. diameter, and the display is enriched by every shade of color connecting the one form with the other. That these have found place in the rich literature of the Hindus is quite certain. But some confusion has arisen on account of the presence and importance of Nelumbo nucifera, the pink-tipped sacred lotus of India. Nelumbo is unquestionably the sacred plant of the Buddhists, the flower upon which Buddha sits, the padma of the Hindu prayer. The only question is whether the few statements assigning a religious value to Nymphaea are to be entirely disregarded, or whether plants of this genus did really receive a minor degree of worship. " Lotus-eyed," " lotus-hearted " and the figure of water on the lotus leaf, all so frequent in Sanskrit literature, certainly refer to Nelumbo.1 But when we read in the Mahabharata (Vana Parva, ch. 307, 19) of a hero "possessed of a complexion bright as the copper- colored lotus leaves," we think at once of N. rubra or its varieties. An examination of the Sanskrit, however,2 reveals that the root padma occurs here (padmatdmradalojjvalam), showing that the plant in the poet's mind was Nelumbo, and we may translate " lotus-petals " in place of " lotus- leaves," improving the simile. Again in Adi Parva (ch. I, 86) we read "As the Full-Moon with its mild light opens the buds of the waterlily, so does this Purana with the light of Sruti expand the human intellect." This waterlily is probably N. lotus or N. rubra, as they are night bloomers. Kalidasa makes a similar reference in Sakoontala, " And now While the round Moon withdraws his looming disc Beneath the western sky, the full-blown flower Of the night-loving: lotus sheds her leaves In sorrow for his loss, bequeathing naught But the sweet memory of her loveliness To my bereaved sight:" (Williams, 1856, p. 92.) There is some poetic license here ; for Nelumbo does not open at night, but does shed its petals when full-blown, while Nymphaea lotus opens at night, but retreats under water after three nights, taking its petals with it ! It seems, however, that N. lotus is the plant referred to; for Williams adds in a note (p. 14) that "the moon is often called the ' lover or lord of the lotuses '." lCf. Dhammapada, ch. iv, 58, 9; ch. xxvi, 401; Miiller 1881. Mahabharata, Adi Parva, ch. 122, 29 ; ch. 199, 11 ; Vana parva, ch. 5, 16 ; ch. 231, 13 ; ch. 307, 23 ; Dutt 1895. 1 Kindly made for me by my learned friend Albert J. Edmunds of Philadelphia. Historical. j Nymphaea stellata is alluded to in the same poem (p. 15) thus — " thoughtlessly attempts To cleave in twain the hard acacia's stem With the soft edge of a blue lotus leaf." Graceful and appreciative as are these literary references, and many- such might be added, they cannot be regarded as having any religious significance. The Sanskrit words utpala and kantala for the blue lotus (N stellata), kumuda the white night lotus (N. lotus) and pundarika white lotus flower, so far as we know, have no religious associations (Macdowell, Elwell, Lanman). Williams (p. 5, notes) states very pertinently that the lotus " is as favorite a subject of allusion and comparison with Hindu poets as the rose is with Persian." On the other hand, it is stated in an early number of the " Botanical Magazine" (1805) that when a Hindu once entered the study of Sir William Jones where a flower of Nymphaea lotus chanced to lie for ex- amination, he " made prostrations before it " as though this were an act of piety. This reverence is further said to " pervade Hindostan, Tibet and Nepal." Tratinnick (1822) considers the same species to be the sacred lotus of India, without giving evidence. Pickering (1879) records that N. stellata is said to be " distinctly figured in the cave-temples of Adjunta," and that he himself saw it figured in Braminical cave-temples elsewhere in India. Nymphaea mbra appears as the " autumn lotus " in Dhamma- pada, 285, according to Pickering, and he thinks that this species may be the padma of the Hitopadesa and Vetala panchavinsati I ; Graham is quoted as having observed it around Bombay, " in tanks, particularly in the neighborhood of caves or temples." But to the writer, none of these records present any certain evidence of the religious importance of Nymphaeas in India. The account in the " Botanical Magazine " is the most suggestive, but is by no means conclusive, and the use of the flowers as decorative designs for temples indicates no more than an appreciation of their beauty. The presence of the plants themselves " especially near temples," if the observation is trustworthy, may be explained both by the beauty of the flowers and the food-value of seeds and tubers. In the absence, therefore, of any direct evidence that waterlilies are regarded as sacred in India, I believe we have no sufficient reason for attributing this regard to them. The Waterlilies. EGYPTIAN ARCHEOLOGY. In Egypt the relations of the waterlilies to the social and religious life were more extended. But after an impartial examination of the facts available, I am strongly of opinion that they received no worship or even reverence here. The teachings of archaeological research were brought together long ago by Duppa (1813-1816), and to some extent by Tratinnick (1822), and more recently by Dr. Pleyte (1875), whereas the critical observations of Schweinfurth from 1882 to 1886 have added many interesting facts, and placed many points beyond the reach of doubt. In spite of a complete unanimity among scholars, considerable confusion exists in the popular mind as to the identity of the so-called Sacred Lotus of Egypt. In America, at least, Nelumbo nucifera is commonly styled Sacred or Egyptian Lotus. But Pickering, Pleyte, Joret and Schweinfurth from the bo- tanical side, and Wilkinson, at least, among archaeologists, unite in the opinion that Ne- lumbo is never found on the ancient monu- ments, and that it was not known in Egypt before the advent of the Persians. Not until the Roman period did it find a place in Egyptian art ; it does become more or less prominent at this time. But this comes within the limit of recorded history and is discussed in another place. We must understand, therefore, that the " lotus " of the palmy days when Egypt was truly Egyptian was a waterlily native to the Nile Valley and Delta. Two species are figured on the monuments and tombs, N. lotus, the white night-lotus, and N. caerulea, the blue diurnal one ; of these the latter occurs very much more frequently. Pleyte found the white lotus only on a single tomb, belonging to the XII dynasty, about 2500 B.C.;1 this example, however, was plainly distinguishable from the allied species. Schweinfurth (1884 c) found petals of the white lotus along with those of the blue in the funeral wreaths of Ramses II and Amenhotep I, but he never observed any carvings or pictures of this plant, though he says Unger found such at Beni-Hassan. The lotus borne on the head of Nefer Turn in one of his manifestations (Wilkinson, 1883, 2: 180; Wiedemann, 1897, p. 138) is long and obtuse and suggests Fio. 1.— Nefer Turn (after Wilkinson). ' Petrie's dates are followed throughout. Historical. 7 N. lottts rather than N. caerulea, but it is too much conventionalized to offer any conclusive argument (Fig. i, 6). Since the leaves of the blue species are entire and its petals acute, whereas the white one has sharply dentate leaves and broad petals rounded at the apex, the difference is usually evident even in very crude representations (Fig. 2). We recognize the blue one easily in manifold applications. It occurs, according to Schweinfurth (1883 b), on all the ancient monuments of Egypt. The use of its tuberous rhizome for food, is said to have been given to the people by Menes, or perhaps even by Isis (Pickering ; Diodorus). At convivial meetings a flower of the blue lotus was presented to each of the guests, and in feasts for the dead, the feasters delighted themselves with the color and odor of this lotus (Buckley). In a carving figured by Wilkinson, the guests are distin- w Fig. 2.— Egyptian lotus (designs after Wilkinson). Fig. 3.— Pleasure boat in a pond of lotuses (after Wilkinson). guished from the hosts and servants partly by their lotuses (see page 1, underneath title). In most of these picturings the flowers are considerably conventionalized, but Schweinfurth saw them in the temple of Ramses II at Abydos, and on coffins of the Ptolemaic period, distinctly painted blue. One of Champollion's figures shows a lotus with green sepals, blue outer petals and red inner petals, all of the parts being tipped with black ; the red and black must have been inserted for love of variety ! An interesting Theban picture shows a pleasure boat, being towed round a pond by three slaves ; the water is represented by the characteristic wavy lines, and is made more vivid by a liberal sprinkling of lotuses, both leaf and flower, over the water-surface1 (Fig. 3). Just as the blue lotus appears in figures of social life and in recreation, it occurs as a favorite flower in religious observances. It is figured among offerings to the gods in the IV dynasty (Pickering ; 1 This lotus has also been found depicted on the pavement of the palace of Sardanapalus, but has evidently been copied by the Assyrian artists from Egyptian sources (Bonavia, 1894). 8 The Waterlilies. Pleyte), and stands in front of Osiris at the judgment of the dead (Buckley). "When the Egyptians approached the place of divine worship," Diodorus tells us, " they held the flower of the ' agrostis ' in their hand, indicating that man had proceeded from a well-watered or marshy land, and that he required a moist rather than a dry aliment " (Wilkinson, /.• 428). This " agrostis " Wilkinson considers to be another name for the lotus (1. c. J: 350). The god Nefer Turn "was figured as a man crowned with an upspringing lotus flower (Fig. 1), a symbol of the resurrection, and of his power to grant continuous life in the world to come" (Wiedemann, p. 138). In one manifestation this is distinctly the blue lotus (Fig. 1, a). Ahi, an infant deity is "sometimes represented as sitting on the lotus flower," the vase from which the plant grows repre- senting a lake of water (Wilkinson j: 132-3). "Osiris, too, swam on a lotus leaf, and Harpocrates was cradled in one " (Whewell). Hapi, or the Nile, a god widely known and honored, " was sometimes resolved into the Nile of Upper and the Nile of Lower Egypt, the lotus being considered emblematic of the former, and the papyrus of the latter ; a design con- sisting of both plants tied together formed a favorite subject for the decoration of the royal throne, as typifying the king's rule over Upper and Lower Egypt" (Wiedemann, p. 145-6). The use of flowers in funeral decorations seems to have been very prominent in the XIX to XXI dynasties. The custom was to lay wreaths and semicircles of flowers on the breast of the enwrapped corpse until the sarcophagus was quite packed with these floral tributes. Flowers of Nymphaea caerulea on petioles 18 to 20 inches long were fastened between the bands encircling the mummies of Ramses II and the priest Nisboni, scattered singly all over them (Schweinfurth 1883 b). Breast- wreaths consisting mostly of petals and sepals of the same plant, sometimes also with petals of N. lotus, were found in the coffins of Ramses II, Amenhotep I, Ahmes I (1580 b. c), the priest Nisboni, the princess Nzi-Khonsu, and the mummy of the case marked kent, from the XX and XXI dynasties, found at Deir-el-Bahari (Thebes) in July 1881. These are probably the " Egyptian wreaths " of Pliny and Plutarch, the "lotus garlands" of Athenseus. Most of these plant-remains date from nearly 2000 b. c, but those of Ramses II were renewed about 1100 b. c. ; for in moving these sarcophagi at that time into their secret resting-place in the Valley of the Tombs of the Kings, to avoid the marauding hordes from the desert, that of Ramses was accidentally broken ; a new coffin and new floral decorations were therefore supplied. The wreaths of Historical. 9 Amenhotep are perhaps older than the Trojan war ! (Schweinfurth, 1883 b). In every case the preservation of the vegetable relics was as perfect as in newly-dried herbarium specimens ; they were easily soaked out in water, and Schweinfurth was able not only to identify them beyond doubt and to find out the manner in which the wreaths were fastened together, but also to prepare and mount examples for the Boolak Museum, where his success is attested by eleven cases of plants. Specimens were also sent to the British and several continental museums. A wreath of lotus petals from the coffin of Ramses II was figured in Nature, in 1883 (Schweinfurth, 1883 b). Such a wreath consists of foliage leaves of Mimusops schimperi folded in two or in four and stitched together with fibers from the date-palm in such a manner as to clamp and hold the lotus petals without piercing them ; the whole was then strung on strips of leaves of the date-palm. The mummy marked rent had small whole flowers of N. caerulea woven in the wreaths by means of papyrus threads. Some- times, as in the case of the princess Nzi-Khonsu, the whole body of the mummy was covered several layers deep with these garlands laid in concentric semicircles from the chin downward. It is interesting to note in passing, that Schweinfurth was not able to detect the slightest differ- ence between these ancient plant remains and the species now growing in Egypt. Either these species are not suffering transmutations, or the changes are infinitely slow. Of real importance, however, is the con- clusion that since the species are unchanged, the climate also of Egypt must be very nearly the same as it was four thousand years ago. If this be true, the absence of Nelumbo and Papyrus from modern Egypt cannot be attributed to meteorological changes, but we are strengthened in the conclusion that these plants only existed there in ancient times through artificial, i. e. human, agencies, namely by cultivation. Several additional references to the Egyptian lotus are unquestion- ably meant for a species of Nymphaea, but whether the blue or the white, the writer cannot now say. Bes, probably a god adopted from the negro tribes, was, at the end of the New Kingdom (500 b. c ?) " fused with . . . and endowed with all the attributes of the young sun-god, and represented like him as sitting upon a lotus flower" (Wiedemann, p. 167). Plutarch says the Egyptians " characterize the sun as though it sprang every day fresh out of the lotus plant" (Wilkinson, 3 : 132-3), and Proclus pretends that the lotus was particularly typical of the sun, which it appeared to honor by the expansion and contraction of its petals (1. c, p. 350). Wilkinson indicates that this lotus is always N. lotus L. (p. 132-3), but io The Waterlilies. in this he is certainly mistaken ; for the one referred to by Proclus is evidently diurnal, and therefore must be N. caerulea. The reader may also be warned against the statement of Prichard that " Nymphaea lotus and N. Nelumbo [= Nelutnbo nucifera\ appear plentifully on the sculp- tures of Thebes." The unanimous voice of later observers sufficiently negatives this. So it appears that the lotus was enjoyed in many ways, but especially as an ornament, a thing of beauty. It was the emblem of the Nile God, for it was the evident product of the river. It was offered to Osiris as any treasured object is consigned to the gods. It was laid upon the bodies of the dead, but other plants could take its place at seasons of the year when the lotus was out of bloom. Thus we see no decided evidence that it was a sacred flower, or an object of worship. It was the " favorite flower in the hands of the Egyptians" (Wilkinson, j: 132-3) as the rose or others might be in the hands of any modern people. It was the symbol of Nefer Turn and the resurrection just as Lilium longi- florum has a similar significance with us and is used for decoration at Easter. pre-linn,ean literature. The facts of the foregoing account are drawn from the researches of scholars during the past hundred years, and have had no influence upon the growth of botanical science. The beginnings of our present knowledge of the waterlilies are to be found in early Greek and Roman literature, from which a growth can be continuously traced. It is true, however, that classic writers became acquainted with one or two African species of Nymphaea by traveling in Egypt and noting the customs of the people and the plants which surrounded them. For example, Herodotus (Bk. 2, § 92) tells us that when the Nile " is full and has made the plains like a sea, great numbers of lilies {xpivov), which the Egyptians call lotus {Xmro^j, spring up in the water : these they gather and dry in the sun, then having pounded the middle of the lotus, which resembles a poppy, they make bread of it and bake it. The root also of this lotus is fit for food, and is tolerably sweet, and is round, and of the size of an apple." This evidently refers to a Nymphaea, probably to N. lotus L. ; for the white-flowered form would be more likely to appeal to the writer as a " lily " than the blue N. caerulea. The description applies equally well to either species ; perhaps, like most other early writers, Herodotus did not distinguish between the two, or did not notice Historical. i i that there were two varieties. He goes on to mention other " lilies, like roses," with fruit " like a wasp's nest," and edible seeds " of the size of an olive stone " ; this is evidently Nelumbo nucifera. He also uses the word "lotus" in book iv, par. 177, to designate a tree whose fruit is the food of the Lotophagi (Lotus-eaters) and their neighbors the Machyles, tribes of Northern Africa. This, the Forgetful Lotus, is a species of Zizyphus. Theophrastus (Bk. 9, ch. 13), in a chapter dealing with various plants and roots, speaks of some which are " sour, others bitter, others sweet." " The sweet [one] called nymphaea (wfiyaia) is accustomed to grow in pools and marshes, as in the Orchomenian country, and Marathon, and the island of Crete. The Boeotians call it madonia, and eat the fruit. It bears large leaves upon {supra) the water ; [these] ground and placed on wounds are said to stop blood. It is useful also as a drink for intestinal disorders." This has been generally accepted, I think cor- rectly, to refer to Nymphaea alba, the common white waterlily of Europe. Theophrastus also describes Nelumbo, under the name of xua/xo^. A brief mention of the "lotus " of Egypt, evidently Nymphaea, occurs in the Historical Library of Diodorus Siculus, written about the beginning of our era. He says (Bk. I, ch. 3) it grows in great plenty in the delta of the Nile, and is used for bread. He too speaks of Nelumbo, but under the name of xtftmpwv. Celsus gives " lotus " in a list of therapeutic agents for " cooling the body." The word may not belong to the original manuscript, but may have been added in later copies. It is considered to stand for the Egyptian lotus as above. Dioscorides, in his Materia Medica, devotes a chapter (112) in book iv to " the Egyptian lotus (kwTbt; 6 'at AquTcry j-eww^svoc). which grows in plains flooded by the water " of the Nile. " The stem resembles that of xua/ioi; [Nelumbo], having a white flower like a lily (xpivov) which, they say, opens at sunrise, and closes at sunset, and the head is altogether hidden under water, and again emerges at sunrise. The head is like a very large poppy ; in it is seed (xo/otoc ; grand) like millet, which the Egyptians dry and beat into bread. It has a root like a quince, which is also eaten both raw and cooked ; when cooked it resembles the yolk of an egg." The identity of this with our Nymphaea lotus L. can scarcely be doubted. We find a difficulty in that N lotus is a night- bloomer, and therefore could not possibly be thought by those familiar with it to sink under water at night. But N caerulea grows with N lotus in Egypt, and is a day-bloomer, closing into a very insignificant bud by j 2 The Waterlilies. evening ; its flowers, however, are not white, but pale blue. We think, however, that N. lotus is the one meant by Dioscorides, as its tubers are larger than those of N. caerulea, it fruits more freely, and the fruits when ripening are devoid of all the floral envelopes except the bases of the outer stamens, exposing a large poppy-like ovary. The fruit of N. caerulea is surrounded by its four sepals and four alternating sepaloid petals, so that its ovary, though equally poppy-like, remains hidden. The story of its behavior at night is a traveler's tale, enlarged upon, as we shall see, by Pliny, and copied many times afterward. Three other kinds of " lotus " are given in Dioscorides : one is the tree of Herodotus' Lotophagi ; the other two are leguminous plants near to or part of our present genus Lotus. Nymphaea {wptpaia) appears in book m, chapter 138. He says it grows in marshes and standing water ; has leaves like xifiwpwv [Nelumbo], but smaller and more oblong, several from the same root, some floating and some submerged ; flower white, like a lily (xpivov), with a bulb (xpoxwdei;) in the middle. " But when done blooming, there comes a round black body of the shape of an apple or a poppy-head, which contains copious, black, crowded seed, of a glutinous taste. Stem smooth, slender, black, like the stem of xtfiibptov. Root (/!>i£a) [Rhizome] black, rough, like a club {p'apaloidrjz) ; gathered in autumn." Dried and taken with wine, it is recommended for dysentery and bowel complaints ; it is also powerfully anaphrodisiac.1 For all these uses the seeds are also good. It is called Nymphaea because it " loves a watery place " ; it is found plentifully {ttoMjj) in Elis in the river Anigros, and in Aliartia of Bceotia. This description all together is quite recognizable for Nymphaea alba; the black ovary is the only discrepancy. There is, continues Dioscorides, another nymphaea [which is called nymphona {yufupCova), and its flower nouphar (voixp «/>)] with leaves like the former, but the root [rhizome] white and rough, flowers yellow, brilliant, resembling a rose. This can be no other than a species of Nuphar. He describes further, in book n, two kinds of xiaptoz, the Greek and the Egyptian ; the former appears to be a kind of bean, the latter Nelumbo ; the fruit of this (which is used for food) he calls xiftwptov or xtfiioztov, and the root, which is like that of a cane, is also eaten both raw and cooked, and is called xoXoxaala. Pliny in several places in his Natural History, uses the word "lotos" for four different plants, the herb lotos, the Egyptian lotos, a shrub lotos (perhaps the pomegranate) and the tree lotos. The fruit of the second of 1 " Eadem contra veneris insomnia bibitur, quae adimit : quin et aliquot diebus continuo pota, genitale flaccidum et enerve reddit." Edit. Sprengel, /.' 479. Historical. 13 these (Bk. 22, chap. 21) he calls "lotometra" and states that the Egyptian bakers knead the flour of -its seed with milk or water to make bread. He adds, to give the quaint English of Philemon Holland's free translation (p. 125). "There is not any bread in the world (by report) more wholesome and lighter than this, so long as it is hot ; but being once cold, it is harder of digestion and becommeth weightie and ponderous." In book xiii (ch. 17) the " Egyptian lotus " (a Nymphaea) is described in almost the same words as those of Dioscorides : its occurrence and habitat, its poppy-like head, and seed resembling millet. He tells us that the Egyptians gather the heads, lay them in heaps until they putrefy, then wash the seeds clean, dry and grind them. Pliny repeats the story of the closing of the flower at sunset, stating only that it is covered with leaves until sunrise ; also that the flower is white. But though he is more reserved in statement here than is Dioscorides, he goes farther in telling of the behavior of the " Egyptian lotus" in the Euphrates (chap. 18). There, he says, the flower is plunged beneath the water at evening, so deep that a man cannot reach any part of it with his hand ; but after midnight it rises slowly, appearing above water again at sunrise. He tells us in addition that its root is covered with a black rind, and is good to eat, especially when boiled or roasted, and that it is excellent for fattening hogs. Nelumbo is mentioned by Pliny in two places under the names " colocasia," "cyamos" and Egyptian bean {/aba). The description of Nymphaea in Bk. 25, ch. 7, reads much like that of Dioscorides. A few new features are introduced. The name and the plant, it is said, originated from a nymph who died of jealousy through love of Hercules ; hence the plant is called by some Heraclion ; but by others it is called rhipsalos, for its club-like rhizome {radix). It grows in water and has broad, floating leaves. The other features of the plant, its names, locali- ties, and uses, and the yellow Thessalian variety are exactly as given by Dioscorides, and the identities of all these are subject to the same doubts and assurances as those given in discussing that writer. At that remarkable banquet imagined by Athenaeus, the conversation turned on Alexandria, whereupon one of the feasters tells about the " lotus " which grows in the marshes near that city in summer ; " it bears flowers of two colors, one like that of the rose, and it is the garlands woven of the flowers of this color which are properly called garlands of Antinous ; but the other kind is called the lotus garland, being of a bluish {xuaUav) color" (Bk. 15, ch. 21). The first plant is probably Nelumbo, the second Nymphaea caerulea Sav. ; we have here the first 14 The Waterlilies. mention of Nelumbo under the name of lotus, and the earliest Latin or Greek record of a blue Egyptian lotus. Nelumbo is frequently men- tioned elsewhere by Athenaeus, once again under the name of lotus, but more often as xtfttbpiov, xuapoz atyunuoz, or xolozaoiu (Bk. 3, ch. 2). He says, at variance with all other writers, that the Egyptians call this plant lotus, but quotes the other names from Nicander, Theophrastus, Diphilus the Siphnian, and Phylarchus. Galenus adds nothing to the facts and theories of his predecessors concerning these plants. He recognizes three kinds of lotus (Bk. 7, ch. 11), the herb, the tree and the Egyptian lotus (Nymphaea), and two kinds of Nymphaea (Bk. 7, ch. 13), one with black rhizomes and one with white ; these correspond with Dioscorides' descriptions of Nymphaea alba and Nuphar. He simply mentions Nelumbo, under the name of /aba Aegyptia (Bk. 1, ch. 20). From this time onward we find nothing of note for our science until it revived with the Revival of Learning. William Turner in 1548 gives this quaint note in his catalogue of *' The Names of Herbes : " " Nymphea is also named in greeke nymphaia & madonais & is called in english water Roses, & some with the Potecaries cal it nenufar. The duch me cal thys herbe Seeblumen. Boeth the kyndes of water Roses growe in standyng waters." Mr. Britten (cf. Turner, 1 548, p. 1 30) has identified " boeth the kyndes " as Nymphaea alba L. and Nuphar lutea Sm. Three years later Fuchsius and Hieronymus Bock (Tragus) in Germany published their botanical classics. The former has a rude woodcut of the white waterlily with the Latin name Nymphaea Candida and the Greek Nu/ifaia huxy. On the next page occurs "Nymphaea lutea" with a figure which identifies it with Nuphar. Tragus goes into the subject much more extensively and shows evidences of personal observation. There are two kinds of " Seeblumen," the white and the yellow, of which the former is commoner. He compares the flower of this to a beautiful double {gefult) lily or rose, with each petal shaped like one's thumb ; and in the midst of the flower is a golden yellow sun (ein schone goltgale Sonneri), and it is odorless. He describes the shape of the bud, and its green "outer leaves." The broad, round, disc-like foliage leaves, he says, are stiff and leathery, and grow on smooth hollow {lucken) stems. His figures are quite crude but recognizable, the " Weiss Seebluomen " as Nymphaea alba, the other Historical. *5 as Nuphar. A long list of common names is given : Wasser Gilgen, Haarwurz, Haarstrang, Nenuphar in apothecary books, Clavus Veneris, Digitus Veneris, Alga palustris, Papaver palustre. The account of the medicinal virtues of these plants does not materially differ from those previously given. Matthiolus, in his Commentary on Dioscorides (1554), transcribes the chapter on Nymphaea from that writer with numerous slight verbal Fig. 4.— "Faba Aegyptla.' facsimile from Mftttliiotas. Fig. 5.—" Nymphaea alba.' Facsimile from Lobelias. changes. He understands " Nymphaea " to signify our N. alba, and gives " Nymphaea altera," also called " blephara," for Nuphar. The two are figured side by side in rude but quite recognizable cuts ; but these may have been inserted by C. Bauhin, whose notes were added to the original in the edition which I have seen. On the identity of the Egyptian Nymphaea and Nelumbo, Matthiolus was hopelessly at sea ; for Nelumbo had been so long extinct in Egypt that nothing could be found there to correspond with the descriptions given by the classic writers. So, after much weighing of evidence, our author invented 1 6 The Waterlilies. a unique figure (Fig. 4) to show what this confused plant must be — a figure which formed the basis of a heated discussion for the next three quarters of a century. We can only refer to this in passing, since Nelumbo, and Arum, and Colocasia are the genera most nearly concerned. Lobelius' " Observations of Plants " (1576) figures two " Nymphaeas," alba and lutea, and gives their medicinal virtues at length. The cuts are rude, but original and easily recognized (Fig. 5) as the species still bearing those specific names. In the " Adversaria Nova," Nymphaea lutea is given before alba, with most of the synonyms used by Tragus and in addition Nenuphar officinarum, plompen (Dutch), Escudetes del Rio and Higos del Rio (Spanish), Water lilie, blanc d'eau and Lis d'estang. From this time forward, waterlily is a term consistently used as a synonym for Nymphaea to the present day, except that in Devonshire, Eng., Iris pseud-acorus has shared the same common name (Friend, 1882). Nymphaea alba is not directly given any common name by Lobelius, but it is evident that some of those given above apply to it ; he says it is sometimes called "major" to distinguish it from "N. alba minor," by which some small white flower, probably Limnanthemum, is designated. A " Nymphaea lutea minor Septentrionalium " is also figured ; it may be our Heteranthera. " Nymphaea," therefore, would seem to be definable as an aquatic plant with evident flowers and floating leaves ; the idea represented by the word is acquiring a more distributed content. " Lotus Aegyptia " {Nymphaea lotus L.) is discussed on pages 385 and 386, but Lobelius is unable to interpret the descriptions given by the older writers, and regards the whole thing as a fabrication. He seems to have no idea of such a plant as Nelumbo. Lyte's Herball (1578) {fide Britten and Holland, 1879) gives a single name for the white and yellow waterlilies, viz., nenuphar. This name was also corrupted into " ninnyvers." In 1589, Petrus Bellonus published observations from Greece, Asia, Egypt, &c. ; he found nothing in Egypt to correspond with the old descriptions of Nelumbo, but found the word " colocasia " used for the root of an aroid which was used for food, as was already known by Matthiolus. So he concluded that Colocasia, Lotus, and Faba Aegyptia were synonyms, and made vain endeavors to interpret Herodotus, Theo- phrastus, Dioscorides and Pliny in harmony with this view. A great beam of light was shed on the subject, however, by Alpinus (1592), who made a special study of the Egyptian lotus. Finding no sign of Nelumbo, but only nymphaeas and aroids in Egypt, he considered the Historical. 17 ancient writers wrong in their synonymy ; colocasia, he said, was an aroid. " Lotus Aegyptia " he believed to be the white flower which, from its resemblance to the European waterlily, he had no hesitation in calling a " larger nymphaea " ; in fact he regards it as practically identical with the species so familiar about Venice. For both have the poppy-like head, and seed like millet, and both, he said, retreat under water at sundown to emerge again in the early morning, " as is observed by everyone in the common nymphaea " {quae in communi nymphaea ab omnibus animadver- titur). Indeed he was surprised that Pliny and others thought this behavior strange in the Egyptian lotus. Both flowers and leaves are submerged at night, according to Alpinus, the flowers emerging first, at sunrise, and a little later all the leaves appear. Of course this does not really occur; but the flower of Nymphaea alba closes into a tight bud during the afternoon and is covered by the dull green sepals, to open again with the light of the next day ; the disappearance of the leaves, however, must have been purely owing to darkness, and their reappear- ance after the flower must have been due to their darker color. Alpinus stated that the white Egyptian lotus was called by the Arabs " nuphar," and by the Hebrews " Arais el Nil." After the publication of the work just discussed, Alpinus received numerous tubers of the Egyptian lotus from a Venetian doctor in Cairo ; they were of the size of a walnut, ovate, with fibrous roots, blackish outside, yellowish within ; they were said to be eaten freely by the Egyptians, both raw and cooked ; the sender also said that the plant resembled the lesser nymphaea (JV. alba L.) except that tHfe flowers were blue. " The which, if it be true, this is the lotus niloticus of Athenaeus." Alpinus figures it with entire leaves, insuring its identity with N. caertilea Sav. At a still later date (1735 b), the same writer summed up the knowledge of the Egyptian lotus (" Lotus Aegyptia ") in a " dissertation " on the subject. As synonyms are given Lotus niliacum, Nymphaea, and Nuphar Aegyptium. The flower and its stalk are called Arais el Nil, the leaf and petiole Bis el Nil, the root (tuber) Biarum. Four plates are devoted to it ; one shows the whole plant, with its dentate leaves, ovoid tuber and the fruit devoid of floral leaves ; another shows the calyx and fruit on a larger scale ; the third shows two fully open flowers, and the fourth gives an enlarged view of a leaf and a tuber. The flower is described as having the four outer leaves green outside and white within, each with five to seven or more longitudinal nerves, spread- ing wide open when in bloom. Within these are about 12 smaller leaves [petals], pure white ; then 25 to 40 long acute leaves [outer stamens], 3 1 8 The Waterlilies. white below and yellow above ; then a fourth series of slender yellow leaves [stamens] and a fifth series of almost capillary members [innermost stamens], followed by a row of narrow incurved bodies [carpellary styles]. At the middle is a small, hard globule, and a head containing the seed. The fruit is round like that of Nymphaea alba, with a green cortex, and seed like that of cabbage. The peduncles are hollow, and of the size of one's finger. The leaves are crenate (the crucial point of the description), of a dark green color, and so numerous as to cover the whole surface of the water ; they die away in November, so that Alpinus considered the plant annual, in spite of the fact that it comes up again at the next flooding of the Nile. The root, he says, is ovate, of the size of a hen's egg, blackish outside, yellowish inside, and tastes at once sweetish and astringent. This excellent description of Nymphaea lotus deserves perma- nent recognition. It represents a masterly attempt to bring order out of the chaos which then existed, and that by the true scientific method of observation. Had subsequent descriptions been as well written, many of our difficulties in nomenclature would not have arisen. We may mention here, following the chronological order, that Phile- mon Holland, in 1601, translates the word nymphaea used by Pliny as "Nemphar"; this name for the waterlily must have been more or less common, therefore, in England at that time. Clusius in his earlier works (1576 and 1601) continues the fruitless discussion of the relations of Lotus, Faba Aegyptia and Colocasia. Concerning Lotus Aegyptia, however, he quotes at length from the valuable text of Alpinus ; but, ignoring the latter's excellent illustrations, he copies the crude figure of the white waterlily from Matthiolus (in fact, the same block was evidently used in the printing), and introduces confusion by labeling it " Nymphaea alba. Lotus Aegyptia Alpini." Beside it appears Matthiolus' cut of Nuphar, marked " Nymphaea lutea major." In 1605, however, he raised the cloud which obscured the Egyptian bean of classic writers by figuring and describing a fruit of Nelumbo, brought from Java by an intelligent sea-captain. Dodonaeus (1583) recognizes only two species of Nymphaea, namely alba and lutea. His figures of these are almost if not quite identical with those given by Lobelius, and his description and synonymy bear a striking resemblance to the text of Hieronymus Bock. The white species (candidd), he says, has great, broad and nearly round, smooth, floating leaves on long, terete, smooth, porous petioles ; flowers solitary, on similar stalks, composed of many oblong and acuminate leaves, in Historical. 19 the midst of which are many yellow stamens. After the flower, a round head is produced, in which are very many black, shining seeds, like millet. The rhizomes are thick, nodular, black outside, white and spongy within, fastened in the mud by many fibrous roots. In addition to the synonyms given by Lobelius, Dodonaeus tells us that Nymphaea is Nenuphar of the shops, and that Apuleius gives the Latin names " mater Herculania," " alga palustris," " clavum Veneris," " digitum Veneris," and the Gallic name Baditin. To these Dodonaeus adds Wassermahnen and Haar-strang of the Germans. The medicinal powers of the herb are discussed and the subject is then dismissed without any reference to the Egyptian species. Besler's magnificent folio (161 3) illustrates in color three kinds of Nymphaea : N. alba major, N. alba minor, and N. lutea, known now as Nymphaea alba, N. Candida, and Nuphar lutea respectively. The descrip- tion of " N. alba major " is given with singular accuracy in many details : leaves broad, smooth, thick, green, cleft to the petiole, etc. ; radix spongy within, and marked with blackish scars (foliorum vestigiis), etc. In the figure, the characteristic round root-scars are very plain. The material is now in hand for a satisfactory summing up of the knowledge of botany as determined by the progressive increase of accuracy and the widening horizon of science up to this time ; and the summing up was duly made by Caspar Bauhin. In the Pinax (1623) he divides Nymphaea into two "genera," one with white flowers and, black rhizome, the other with yellow flowers and white rhizome ; the * latter, called " Nymphaea lutea " includes Nuphar and some other types. The former, " Nymphaea alba," is divided into four species. The first of these, with its synonymy, is : " Nymphaea alba major Nenuphar album, Brunfelsius. Nymphaea alba, Brunf. Matth. Ang. Dodon. Cord. hist. Lon. Ad. Lobel. Taber. Thai. Ger. Clus. hist. Nymphaea Candida, Trag. Fuch. Tur. Caes. Nymphaea major alba, Lugd. Cast. Eyst. Lotus Aegyptia, Alpini." The last synonym is unquestionably wrongly placed. The second species of Bauhin is " Nymphaea alba minor," being " Nymphaea parva, Matth." Whatever the plant may be, it is not a Nymphaea in the present sense ; and the same may be said of the third and fourth species of Bauhin. The consideration reduces for our present purposes to a single species, in which, however, two, and possibly three, quite distinct plants are 20 The Waterlilies. included, namely, the European species N. alba, and perhaps N. Candida, and the Egyptian white lotus, N. lotus L. The blue Egyptian lotus, IV. caerulea Sav., mentioned by Athenaeus and Alpinus, is quite left out. A unique volume is Gerard's " Herball or Generall Historie of Plantes " published in 1636. Of his five "Water Lillies" only the first would now be placed in the genus Nymphaea ; this is " Nymphaea alba, White Water Lillie," given in the index of the work, however, as " N. alba major." The second species is "Nymphaea lutea," our Nuphar; the Others have no relationship with the Nymphaeaceae. The figures of the first, second, fourth and fifth species were apparently printed from the same blocks as were those of Matthiolus, and the text is but an abbreviated translation of the same writer. The Egyptian lotus is not mentioned at all by Gerard. Vesling, editing Alpinus in 1640, recognized the difference between the European and Egyptian white waterlilies and discussed their relationships. He tells us also that the white lotus flower is woven into triumphal crowns. In the same year Parkinson's fascinating " Theatrum " appeared ; his archaic English is most entertaining. He gives a burning paragraph on the " Egyptian bean " question, having full knowledge of that plant (which we know as Nelumbo), and arraigning Matthiolus for his " false " and imaginary figure of it (Fig. 4). At no point does he refer to this plant as " lotus." In chapter 29 he gives " Nymphaea, the Water Lilly," with seven species : 1. N. alba major vulgaris, The great common white Water Lilly. 2. N. alba major Aegyptiaca sive Lotus Aegyptia, The great white Water Lilly of Egypt. 3. N. alba minor [Not a Nymphaea as now understood]. 4. N. alba minima quae et Morsus ranae vocatur. 5. N. lutea major [Nuphar] . 6. N. minima lutea \ rxT . XT , , ,-\ ,_ , ,, . >• LMot Nymphaea as now understood J. 7. N. lutea llore minore ) The first is of course our N. alba L. ; it is illustrated by a copy of Lobel's figure (Fig. 5). The second is our N. lotus L., and is accom- panied by Alpinus' figure of the whole plant. Referring to the resting period of this plant in winter, Parkinson suggests — perhaps in criticism of Alpinus — that the plant has been therefore " called annual, but I would call it restible." A compilation of former observations with valuable comments, was made by Bodaeus a Stapel in the 1644 (Amsterdam) edition of Theophras- Historical. 21 tus. He considered the word Nenuphar a corruption of the older Greek nouphar. But his description of the European waterlily is the valuable feature. The leaves, he says, are on long, terete, glabrous, porous peduncles, are approximately round, firm, almost coriaceous, for the most part floating on the water ; upper surface smooth, distinctly veined beneath. Those which lie concealed below the water are thinner and softer. Flowers solitary, on stems like the petioles, made of many floral leaves, so that sometimes a single flower is composed of twenty-five to thirty or even forty leaves, each of the shape of one's thumb or a leaf of the greater Sedum. From the middle of the flower many yellow stamens project. Bud oblong, its outer leaves [sepals] purplish green. Flowers scentless, shining like the sun. Here, curiously enough, reappears the myth of its retreat below the water at evening and emergence at sunrise. The account continues : After the flower, there is produced a head like a poppy or round apple, with black shining seed, larger than millet. The rhizome {radix) is about as thick as one's arm, knotty, black outside, white and spongy within, odorless, tasteless, beset with and fastened in the mud by fibrous roots. Bodaeus calls this plant " Nymphaea alba major." Two other varieties of " N. alba" and four of " N. lutea " are next de- scribed, but they are all outside our present genus Nymphaea and need not be considered here. At another place Nymphaea lotus is fully treated under the name of " Lotus Aegyptia," giving all of Alpinus' figures. Bodaeus thinks this plant should not be classed with Nymphaea on account of its bulbous root. He thus differs in opinion from Alpinus, and especially from C. Bauhin. Piso (1648) gives us the earliest account of an American Nymphaea, in a Brazilian species. " Among those plants," he says, " which are common to Europe and the western world is Nymphaea, called by the Brazilians Aguape, by the Portuguese Galvaon, which is noticed everywhere floating on the surface of pools and still waters. The leaves are similar to our Nymphaea, with a great network of veins beneath." The flower has a pleasant odor, and consists of four green sepals and about twelve narrowly oblong, acute, white petals. Its medicinal virtues are also detailed. Just what species is here referred to cannot be positively identified, nor does the crude figure accompanying assist. But from the character and number of the petals, and the fact that we know of but one white day-blooming waterlily in Brazil, we have but little hesitation in referring it, as did Caspary (1878), to N. ampla DC. Bontius, whose accounts of the East Indies were bound in with Piso's 22 The Waterlilies. of the West Indies in 1658, devotes considerable space to " Nymphaea Indica flore purpureo"; the figure and description relate very clearly to Nelumbo and may be omitted here. Another new feature is introduced by Flacourt (1661) in his " History of the Great Island of Madagascar." He found there a blue waterlily, which was considered a delicious food. It is called by the natives " Tan- tamou." Instead of being anaphrodisiac like the European species, this one provokes to excesses. The white waterlily, says Flacourt, is called in this country " Laze-Laze," and the yellow flowered one " Talifouc." The roots of all three are used for food. The identities of all of these are in the highest degree uncertain. The blue is probably N capensis or N. madagascariensis. The white may be N. lotus, or a mere variety of the blue. The yellow one may be a Nuphar ; at any rate, no yellow Nym- phaea is known at present from that island. Returning to the European flora, we may gather a few points from J. Bauhin's " Histoire des Plantes " (1670). Curiously enough he quotes the name Nymphaea alba minor, with the synonym N parva Matth. from C. Bauhin, as signifying the European white waterlily, and adds a number of -the common names given by Lobelius. The real identity of the plant is rendered certain by a crude picture. The peculiar mixing of the names remains unexplained. Chabraeus (1677) records five species of Nymphaea, viz.: N alba vulgaris, N lutea, N minor lutea Lobelii, N. minor lutea flore fimbriato, N. alba minor. The last four would not now be classed in the genus Nymphaea at all. The figure of N. alba vulgaris, is an enlarged and very slightly modified copy from Fuchsius. John Ray's "Catalogue of Plants of England" (1670) mentions five Nymphaeas, of which only the first, "Nymphaea alba',' belongs in the genus as now constituted. The only information given in addition to the name is the words " In fluviis. Florem habet plenum!' The " Methodus Plantarum" of the same author (1682) is especially interest- ing for its freedom from slavery to the classic writers. The twenty-fifth genus contains "Anomalous herbs and those of uncertain position"; the first section of these is headed "Anomalae Aquaticae," and the first of these is Nymphaea, plants with round to oblong, floating leaves, differing in their flowers. Three varieties are mentioned, of which " N. alba vulgaris " is the only true waterlily, the common white one of Europe. The " Historia Plantarum" (1688) follows the classification of the " Methodus " in general, but describes the European white water- Historical. 23 lily as N alba simply. In many points the description is only a repetition of earlier writers ; the flowers are said to be odorless, with the outer leaves greenish as in Ornithogalum (ut Ornithogali). A " Nymphaea Brasiliensis," known to the natives as Aguape, and described in nearly the same words as used by Piso, is given here, on the authority of Marggrav's observations, and classed as belonging to N alba. A fifth species is added, N alba major Aegyptiaca. The specific description of this refers to the " crenate leaves " which identify it with N lotus L., and the statement is added, " flores . . . semper supra superficiem aquae." The object of this addition comes out when in the succeeding paragraph the stories of Theophrastus, Pliny, and Alpinus are given regarding the retirement of the flower far under water at night, and the writer boldly concludes, " Nobis certe incredibilia videntur." Jacobus Breynius (1680) is one of the few among the early botanists who considered new and foreign plants. His second Prodromus gives "Nymphaea flore coeruleo odoratissimo, Cap. bonae spei ; nobis," and " Nymphaea flore suave purpurascente Japonica ; nobis." The former is doubtless N. capensis Thunb., the latter Nelumbo. The magnificent "Flora Malabarica" of Van Rheede (1692) contains figures of a number of water plants. In vol. 11, plates 26, 27, 28 and 29 represent plants with floating leaves, all under the general Malabar name of Ambel. The first, called simply Ambel, is Nymphaea pnibescens; the sec- ond, Cit-Ambel is the source of Willdenow's Nymphaea stellata, a blue- flowered waterlily. The other two, Nedel-Ambel and Tsjeroa-Ambel are Limnanthemums. All are quite fairly recognizable from the engravings, at least as to their generic position. It is interesting to note that the natives of the Malabar coast classed Nymphaea and Limnanthemum under the same type or genus, just as did their European contempor- aries. The text accompanying the plates refers chiefly to the medicinal properties of the plants. In 1696 Commelin published Latin names and synonymy for the species illustrated by Van Rheede. To Ambel he gave the title " Nym- phaea indica, flore candido, folio in ambitu serrato," although he regarded it as identical with the Egyptian white lotus according to Parkinson, Alpinus, and Vesling. Cit-Ambel was dubbed "Nymphaea malabarica minor, folio serrato" and was a new species to European botany. Perhaps the most magnificent and exhaustive botanical work of the pre-Linnaean period is the " Phytographia " of Plukenet (1691-96), with its exquisite engravings and its several companion volumes of text. We 24 The Waterulies. cannot tell exactly what this writer understood by the term Nymphaea ; for of the four varieties figured in the Phytographia, one appears to be Trientalis, another is plainly Marsilia, a third is Nelumbo, and the fourth Limnanthemum. The genus is still more comprehensive in the " Amal- theum " (1705), including in addition to those above mentioned our Nym- phaea alba and N lotus as well as Nuphar, Hydrocharis, Podophyllum, and some others whose brief descriptions are unrecognizable. Several of these had been much more scientifically disposed of by earlier authors with whose writings Plukenet was fully conversant. In the " Mantissa " ( 1 700) Brasenia and Pistia are added as species of Nymphaea, additions which one can easily understand. Nymphaea capensis is also given under the name and on the authority of Breynius as quoted above, but it is confused with Nelumbo. Finally an American terrestrial plant classed by previous writers as a Brassica, is here classed as a species of Nymphaea. The white Nymphaea or Laze-Laze of Flacourt is referred by Plukenet in the "Amaltheum" to his "Nymphaea Indiae" (Nelumbo). Plukenet brings in no new information for our subject. In 1 707 Sloane described a white flowered Nymphaea with the edges of the leaves "deeply cut," from the "Fresh River" in Jamaica; he says its leaves agree with the description of the white lotus of India, and hence he speaks of it as " N Indira flore candido folio in ambitu serraio" of Commelin. The identity must have been mistaken, since none of the Old World species are native in the new. Caspary has rightly identified this with N ampla DC. In discussing the names, Sloan suggests that the Indian and Egyptian white lotus seem to him to differ very little ; English botanists now rank them as one species. He also states, quoting authority, that his Jamaica waterlily was carried to the Indies by way of merchandise. Perhaps this is an attempt to explain its supposed identity with the Egyptian plant. By all odds the best pre-Linnaean classification of plants, as is well known, was that of Tournefort (1700). The waterlilies are placed in his Sixth Class, " herbs and suffruticose plants with rosaceous flowers." Nymphaea is the last (eleventh) genus in the fourth section of this class, while Nelumbo is separated from it as the first genus in section five. In this respect Tournefort came nearer our present ideas than Linnaeus, who made Nelumbo a species of Nymphaea. The genus Nymphaea is defined in the " Institutiones " as having a circle of many petals, and a pistil rising from the midst which forms a globose or conical multilocular fruit filled with many oblong seeds. The floral leaves and the fruit are putrescent, and Historical. 25 the genus is " easily recognized " by the broad sub-rotund floating leaves. " Nenufar " is given as a generic synonym. Four species are named : 1. " Nymphaea alba, major C. B. Pin.," our N. alba. 2. "Nymphaea Americana, foliis circinatis, maximis, acute crenatis, subtus purjmreis, flore albo Plum." 3. "Nymphaea Americana, foliis circinatis, minoribus, obtuse crenatis, subtus viridibus, flore albo Plum." 4. " Nymphaea lutea, major? our Nuphar luteum. In vol. 2, p. 137-8, the flowers, petals, ovary, fruit and seed of Nym- phaea alba and Nuphar are figured, with also a cross-section of the fruit of each. We are not prepared to settle the identity of Plumier's American species ; it seems likely that No. 2 may be referred to N. ampla DC, and No. 3 to N rudgeana Mey. ; Caspary so placed them (1878). We cannot explain the absence of N. lotus from the list, unless it was meant to be included under the species N alba major, as classed by C. Bauhin, who is quoted as authority for this name. Finally Ludwig's (1737) arrangement of the genera may be mentioned. He separated Nymphaea from Nelumbo, as did Tournefort, and described the former genus as having five sepals and many small petals ; this is doubtless our genus Nuphar. He made an additional genus " Leuconym- phaea " which has many petaloid stamens, and covers our present genus Nymphaea, as was recognized by Kuntze (1891) and MacMillan (1892). The early plant anatomists seem to have neglected Nymphaea entirely. Grew and Malpighi make no mention of the genus. Vaillant in his essay on the structure of flowers (17 17), refers to the "Grand Nenufar-blanc " as illustrating the condition where the stamens arise from the sides (reins) of the ovary. Older writers, as we have noted, referred occasionally to the "spongy" rhizomes and the "porous" petioles and peduncles. The Linnaean and post-Linnsean literature will be cited throughout the course of the present paper. CHAPTER II. STRUCTURE. The habits of waterlilies are too well known to need more than a brief statement here. These plants are found in the shallows of slow streams or still water all around the world. The long rhizomes of the Castalia group creep along in or upon the mud bottom of ponds or in a tangle of vegeta- tion overlying this, in a depth of water from a few centimeters to as much as 5 or 6 meters, branching here and there, and dying off behind as they advance by apical growth. By far the greater number, however, have an erect, tuberous stem or caudex, which never acquires any considerable length, but stands with its growing apex about on a level with the mud in which its lower portion is buried. From the stem in every case a great many stout, fibrous roots pass downward and anchor the plant in place, while the leaves and flowers rise upward. The former lie flat upon the water surface, the upper side being dry and exposed to the air. Under some conditions the petioles may be strong enough to hold up the leaves and make them totally aerial. All of the species, at certain periods of growth (germination and beginning of each annual growing season), produce small, thin leaves which lead a submerged existence. The flowers of about half of the species open at the surface of the water and seem to float upon it, those of the other half are raised on strong peduncles some distance above the water level. Only in one or two species (N. rudgeana, N. amazonuni) is cleistogamy known to occur. The flowers in such cases never reach the air, or only do so for a few hours, after self-pollination has been effected. After anthesis the fertile flowers are drawn down into the water by movements of the peduncle, and here the seeds ripen. By means of a buoyant aril, the ripe seeds from the irregularly bursting fruits rise to the water-surface and are floated far and wide. These peculiar conditions of life have impressed themselves strongly upon the forms and functions of the waterlilies. Manifold adaptations have arisen in every part of the plant. We shall discuss these in order in the succeeding pages, taking up first the gross morphology, then the finer anatomy of root, stem, leaf, and flower, and finally the physiology of the genus. 27 28 The Waterulies. THE ROOT. The primary root in Nymphaeaceae is of short duration ; in Nelumbo, indeed, it never becomes more than an embryonic rudiment. The entire root-system, therefore, of mature plants of this family is adventitious. The following description relates to conditions in mature Nymphaeas, the course of affairs at germination and immediately afterward being reserved for a later chapter. On digging out a well-grown rhizome of a waterlily, we find it beset by a great number of long, terete, white, spongy roots. In large tropical species (Al. caerulea, lotus, flavo-virens, zanzibariensis, capensis, and hybrids) these may be a centimeter or more in diameter, and they are nearly as large in strong specimens of N. alba; in N. odorata they measure usually 0.3 to 0.6 cm. in diameter. Each root is decidedly contracted at its point of junction with the stem, but reaches its greatest diameter about 5 cm. out, or less ; thence it tapers almost imperceptibly to its apex 30 to 60 or 100 cm. away. In species with a short caudex (Apocarpiae ; Lotos, Hydrocallis) these roots may be so numerous and close together as to be deformed at the base into prismatic shapes ; in Eu-castalia the internodes are so elongated as to cause the roots to appear much more scattered. For the roots all arise from the bases of the petioles, or (in Eu-castalia) from an upraised cushion-like portion of the rhizome upon which also the petiole is inserted (Fig. 20). In N. flavo-virens and its hybrids a thick irregular excrescence remains after the decay of the petiole, projecting from the tuber a centimeter or more, on the lower side of which the roots are borne, to the number of 5 to 13. In N. odorata 3 to 5 or 6 roots occupy a triangular area on the cushion behind each petiole ; the apex of the triangle is farthest from the petiole and is occupied by the first formed root ; this one is of medium size. Next to it is a second and much larger root, and the base of the triangular area is covered with a few much smaller ones. Some of the roots of young plants, and of adult specimens of N. flavo-virens, are transversely wrinkled in their upper parts, showing a contractile power. No co-ordinate branches are given off by the roots as a rule, but they are all clothed with a great number of slender rootlets, and the larger of these may again be beset with capillary branchlets ; this arrangement may compensate for the total absence of root-hairs through- out the adult plants. A root-cap is present on the tip of every root and rootlet, often showing the long, thimble-like shape so frequent in water plants ; it may be black or brown in color, but is frequently colorless or whitish on the smallest rootlets. Structure. 29 The root-cap is said by Henslow (1892, p. 507-8) and Van Tieghem (1898) to resemble in structure that of monocotyls rather than dicotyls. In N. flavo-virens and N. caerulea, however, the tip of the root agrees exactly with the fourth type of structure given by De Bary (1884, p. 12), Fio. 6.~ Longitudinal section of root-tip of IV. caerulea. a, air-canul ; s, space between root and upper portion of cap. From a photomicrograph. i. e., there is a transverse layer of initial cells from which all of the tissues are developed (Fig. 6). Caspary and Nicolai assert that the cap continues to grow only for a very short period, after which no farther addition to it is made, though it is persistent, and not deciduous as in Nuphar (De Bary 1884, p. 413). In a root of N. flavooirsns from a plant just breaking from a strong tuber, the cap is short-conical and rounded at apex, with 3Q The Waterlilies. long thimble-like prolongation around the root. Its outer surface is bounded by a distinct, smooth layer of cells, which are much larger than the underlying ones (cf. Fig. 6). At the apex these outer cells are columnar in shape and arrangement and are removed from the region of initial cells by 12 or 13 elongated elements (columella) ; they become nearly cubical opposite the initial region, where the cap is five to six cells thick, and elongated in the upper free portion of the cap, where for some distance this layer alone is present. The relative size of the root-cap in this species is shown by the following measurements. The cap extends 0.035 mm. beyond the initial cells, and 1.33 mm. above this around the root; its diameter opposite the initials is 0.51 mm., the same as the diameter of the root above the cap. A somewhat larger root-tip of N. caerulea, (Fig. 6), taken from a young root of a mature plant, shows a structure exactly similar to that of N. flavo-virens ; in one specimen, however, the apical portion of the cap had been worn away in irregular fashion so that the primitive outer layer of large clear cells is visible only on the sides. All of the cells of the cap are arranged in longitudinal rows. The peripheral por- tion of the cap is continuous with the upper, sheathing part, and is composed of narrower cells than those below the body of the root. The root-meristem and root-cap are formed by a transverse row of cells as described above. A regular epidermis of columnar cells is found near the apex of the roots (Fig. 7 (a) e) ; the outer walls of these cells are cutinized, but their partition walls are quite thin. They increase somewhat in size as the root enlarges, but their outer ends become rounded and somewhat drawn apart (Fig. 7 (b) e). In most of the species examined (N. flava, tuberosa, odorata minor, marliacea-chromatella, lotus) they become almost cubical in shape (i. e., in cross section ; they are elongated in the direction of the axis of the root), then are ruptured in their outer walls, the radial walls remaining as a black, irregular fringe on the surface of the root (Fig. 7 (d) e), giving it a powdery appearance. In N. elegans X zanzibariensis, the epidermal cells become greatly flattened and epithelioid before rupturing. Flo. 7.— True and false epidermis of roots. N. marliacca chromatella. (a) from very young, (b) older, (d) mature root In transverse section; (c) surface view of (b), t , epidermis, h, hypodermal layer, c, cortex. Structure. 31 In all cases, however, the epidermis persists on the narrower portion of the root within about 2.5 cm. of the rhizome. Its loss is the direct result of the great expansion of the root, caused by the enlarging and maturing of the air-canals. Meanwhile the subjacent outermost layer of cortex has developed in an especiai manner (see Fig. 7 h). At a very early stage this layer is composed of columnar cells closely resembling the epidermis at the same period. But as its cells become broader and more nearly cubical, the walls become collenchymatously thickened, as also do those of the next one or two layers below (N. lotus, tuberosa, marliacea-chromatella). The outer walls of the hypodermal cells unite to form a fairly even surface, but the inner walls meet at various angles, owing to the varying depths of the cells. Thus an even hypodermal covering is formed in place of the lost epidermis. The cortex may be divided into five layers : (1) the single hypodermal layer ; (2) exo-cortex, a layer of from one to three or four tiers of closely packed cells without intercellular spaces ; (3) medio-cortex, a tissue composed chiefly of large air-canals and constituting nine-tenths of the bulk of the root ; (4) endo-cortex, consisting of one or two layers of cells closely set together, lying adjacent to (5) the endodermis (Figs. 8, 9). The first of these layers has already been described. The exo-cortex is continuous, between the air-canals, with the medio-cortex. In N. odorata (Fig. 8) there are between the outermost air-canals and the hypodermis two series of cortical cells with thin walls and in size of lumen equal to or larger than the hypodermal cells. N. tuberosa and N. marliacea-chromatella have usually three series of cells in the exo-cortex ; they are rounded, with collenchymatous walls, and larger by half than the hypodermal cells. The exo-cortex in N. lotus consists of a single row of cells (four or five near the base of a mature root), in N. /lava of one or two rows. In N. elegans x zanzibariensis there are two or three series of cells which are thin-walled and three or four times as large in cross- section as the hypodermal cells ; in longitudinal section, however, the hypodermal cells are much the longer, being about 0.025 mm. long; the exo-cortical cells are only 0.0175 mm. in length, and the cells of the medio- cortex about 0.0076 mm. N. flavo-virens, at least near the proximal end of the root, has five or six cell-layers in the exo-cortex. The outstanding characteristic of the medio-cortex is the presence of large air-canals. These are elliptic or rounded or more or less hexagonal in cross-section, and run parallel throughout the length of the root, 32 The Waterlilies. beginning in the meristem immediately back of the initial layer of cells Fig. 6 a). They appear to be without communication of any kind with one another, i. e., the cells bounding them fit together without transverse inter- cellular spaces. A single perforation 0.0025 mm. in diameter has been observed in the side wall of a canal in a root of N. tuber osa, but this must for the present be considered exceptional. At more or less frequent intervals (0.127 mm. to 3.0 mm. for N. tuberosa, 1.2 mm. to 5 mm. for N. elegans X zanzibariensis) diaphragms lie directly or obliquely across the canals. A diaphragm is a very thin plate of epithelioid cells (0.0076 mm. to 0.013 mm. thick for N. tuberosd) with sinuous margins ; the pro- tuberances of the margins of the cells meet one another, leaving spaces be- tween the sinuses (Fig. 12 /). In jV. lotus the cells are more even in out- line, with large, round, in- tercellular spaces. Thus the passage of gases is left practically unimpeded. The walls of the canals are a single cell-layer thick, and bear stellate cells (idio- blasts) here and there. In A^. tuberosa these occur about 0.32 cm. apart on the average along each canal ; in N. elegans X zanzibariensis they are Fia. 8.— Root of N. odorata, transverse section. From photomicro- graph ; e, remnants of epidermis ,* h, false epidermis. Structure. 33 two or three times as frequent, but somewhat more slender, and, though of nearly the same actual size, are much smaller in proportion to the size of the canals ; they are few in the root of N. flavo-virens, numerous in N.flava and N. odorata, very plentiful and of massive build in N. marliacea- chromatella. Transverse sections show that they occur only in angles where three partition walls meet, and their arms project into the three adjoin- ing canals (Fig. 8). The canals vary in actual and rela- tive size in the different species, but are always largest at the middle of the medio-cortex and smaller toward exo- and endo-cortex. N. odorata (Fig. 8) has six or seven concentric rings of canals, i. e., that number can be counted along any radius of the root ; they are all quite large and distinct. N. tuberosa has seven to ten rows, very irregularly placed ; the middle and outer ones are approximately circular in outline, those near the endo-cortex are greatly elon- gated radially, and the innermost and outermost ones are scarcely more than large intercellular spaces. N.flava and N. lotus have six or seven rows of rounded and quite regular canals in the main part of the root, becoming irregular in N. lotus near the stem. N. marliacea- /m/ Fio. 9.— Contractile root of N . flavo-virens, trano- CkrOtnatella has nine rOWS, arranged verse section ; en, endodermis ; ft, false epidermis ; i • \T i L .i j • i t, latex or tannin cell. much as in N. tuberosa; the radial elongation of the inner canals causes the partition walls in this region to look like radial bead-like strings of cells. N. elegans X zanzibariensis has eight or nine rows of large canals, each of a rounded-hexagonal shape ; N . flavo-virens shows a similar structure in its nutritive roots. The cells of the partition walls are nearly square in side view (longitudinal section). The angles of the canals are occupied mostly by rounded-hexagonal cells ; three alternate sides of these face the canals, the other three are joined to the neighboring partition cells. Between two angle-cells is usually a single square or rectangular cell ; but often 4 A 34 The Waterlilies. this divides into two smaller ones. Occasionally the angle is formed not by a single cell but by three partition cells which meet at a point and are connected by beveled ends. In N. flavo-virens, N. tuberosa, and to some extent in N. marliacea-chromatella, all of these cells are greatly rounded out and turgescent, but their areas of union with one another (common wall) are narrow and bounded by slightly curved lines ; in transverse sections these lines give the appearance of small tubular and branched cells lying between larger rounded cells ; longitudinal sections reveal the true state of affairs. The contents of the medio-cortex cells may be very small indeed, as in N. flavo-virens and N. elegans X zanzibariensis, or there may be starch in the inner parts, as in N. tuberosa and N. marliacea- chromatella, or the whole medio-cortex may be loaded with starch, as in N.flava. N. marliacea-chromatella is intermediate in this respect between tuberosa and flava. The grains of starch are rounded and simple or elongate and double. The endo-cortex is an ill-defined layer, most distinct in N. flava and N. elegans X zanzibariensis ; in the former it consists of one or two cell-layers, in the latter of but one ; it is composed of small cubical cells, and has rather large intercellular spaces, both between its own cells and between these and the endodermis on the one side and the medio-cortex on the other ; it should not, perhaps, be separated in the description from the medio-cortex. Bounding the cortical tissues on the inner side, an endodermis with thickened radial walls may always be recognized. Its circular outline is interrupted by occasional slight sinuosities ; it is most evident in N- lotus, odorata (Fig. 8) and elegans X zanzibariensis, least so in N. flavo- virens and tuberosa. Its cells contain starch in N. flava and marliacea-chromatella. A typical radial bundle system passes along the middle of the root. Its outer- most cell-layer constitutes the pericambium. Against this the xylem and phloem patches abut ; their number in different species is shown in the accompanying table. In the center of the bundle an area of pith remains ; this may extend over one-third of the diameter of the bundle in N. odorata and lotus, one-fourth in N.flava, tuberosa, and marliacea-chroma- tella, or one-sixth in N. elegans X zanzibariensis. In a general way this Name. No. of rays of xylem. N. odorata minor 7 N. ? 8,8 N. flava 6,7 N. flavo-virens X zanz. 5 N. elegans X zanz. 7, 7. 7. 7 N. caerulea X zanz. 6,6 N. tuberosa 8 N. caerulea 6, 7- 7 N. flavo-virens 5, 7-8 N. lotus 6, 6, 9, 4 N. amazonum 7. 7 N. rudgeana 7 N. marl.-chrom. 8,8,8 Structure. 35 shows a greater amount of xylem tissue in species with a greater extent of leaf surface. The first-formed and smallest xylem tracheae are in contact with the pericambium ; one such trachea occurs in each xylem plate, and from it a single row of tracheae continues radially inward. These are in direct contact with one another in N. tuberosa, elegans-zanzibariensis, flavo- virens X wnzibariensis , and sometimes in N. odorata ; parenchyma cells may be interposed between them in N. marliacea x chromatella, lotus, and odorata ; in N. flava the radial lines are quite irregular and difficult to follow, the tracheae having much the appearance of being scattered through- out the parenchyma. Only spiral and spiro-reticulate tracheae have been observed (A7", tuberosa, elegans X zanzibariensis). Large phloem patches with sieve-tubes and companion cells, but no fibers, lie between the xylem plates ; they are rounded on the inner side and do not extend so far toward the center of the bundle as the xylem plates ; that is, the inner- most tracheae project more deeply into the pith than do the inner bound- aries of the phloem masses. All of the cells of the bundle appear polygonal and nearly isodiametric in transverse section, but even the parenchyma is much elongated in the direction of the axis of the root. Contractile roots differ from the more numerous nutritive roots in that the xylem is less in quantity in the former, and the air-canals of the cortex are much smaller in proportion to the cells bounding them. In a root of an adult plant of N. flavo-virens (Fig. 9) taken up in October, the false epidermis and exo-cortex are apparently normal. The outer part of the medio-cortex is collapsed wherever a transverse wrinkle occurs, but it consists for the most part of comparatively dense tissue, in which the intercellular spaces are not larger than the sur- rounding cells. The vascular system is cut off from the cortex by a distinct endodermis. Each of the 5 to 9 xylem rays consists of a single line of tracheae occupying about one-third of the diameter of the vascular cylinder. The pith in the center of the cylinder is sharply marked off from the surrounding tissues by an endodermis with very distinct radial thickenings. This occurrence of an intrafascicular endo- dermis seems very anomalous in a flowering plant, but was constantly found in contractile roots of adults of N. flavo-virens. In non-contractile roots of this species the xylem rays are often two cells wide, they extend farther toward the center of the vascular cylinder, and there is no inner endodermis. In none of the roots of other species was an inner endodermis found. As stated before, a radial vascular bundle system traverses the entire 36 The Waterlilies. length of each adventitious root. Branches are given off from it at right angles and pass straight out through the cortex to the side rootlets. Where the root joins the stem, the root-bundle passes in and bends sharply backward (away from the stem apex) and runs more or less obliquely into the plexus of bundles of the stem. In N. odorata the bundles of all of the roots from a single leaf-base unite in the peripheral region of the stem into one strong trunk (Fig. 10) ; the appearance of a pith in the midst of this trunk in N. alba candidissima gives rise to the " root-bearing stele" of Gwynne-Vaughan (1896 b). In N. lotus each root-bundle passes in separately and breaks up into a number of branches before joining the plexus of the stem. It is notable that there is no direct connection between the bundles of any leaf and the roots which spring from its base, close as their proximity is. The innumerable small rootlets which cover the main roots seem to repeat entirely the struc- ture of the latter. The vascular system (in N. elegans x zanzibariensis) is hexarch, with uniseriate pericambium and a tiny central pith ; endodermis and endo-cortex are uniseriate and distinct ; the meso-cortex has only one or two series of very small and irregular air-canals ; hypodermis and epidermis occur as above described. The rudiments of roots are laid down below the bases of the petioles at a very early period. In a strong rhizome of N. odorata a leaf bud less than 0.3 cm. long, including the petiole, had four subtending roots already laid down in the tissues of the cushion just below it ; the posterior two of these had the root-cap well defined and had destroyed some of the overlying cortical cells of the stem. After this time, however, their development is very slow ; at the base of a petiole over 8 cm. long the root apices had just penetrated the epidermis, but were still concealed in the short pubescence of the leaf-cushion, No pericambial tissue has yet been distinguished in Nymphaeaceous stems for the source of adventitious roots. Reference has been made to cell-multiplication in the apex of growing roots, and to the early origin of the air-canals. These canals are at first surrounded by six cells. Between every two of these another cell is cut off, giving one at each angle of the canal, and one on each side (Fig. 1 1 b). Fio. 10.— Vascular bundles of the roots of N. odorata. Longitu- dinal section of leaf-base. Structure. 37 The angle-cells are, in this stage, triangular and nearly equilateral (Fig. 1 1 a) ; the lateral cells are narrow and brick-shaped, with their ends abutting on the canals. It is largely by the great extension of the lateral cells that the air-canals and the roots themselves increase in diameter ; for the shortest side of the young cell becomes much the longest in maturity ; meanwhile the angle-cells have expanded also, their three points becoming first rounded, then flattened, to give the cell its hexagonal shape (cf. Fig. 8 and n). A diaphragm of an air-canal may be traced back (Fig. 12) to a single lateral cell near the tip of the root ; this extends out into the lumen of the canal, by distension of its wall, as a rounded papilla. The papilla is cut off, and soon divides longitudinally ; each half then divides transversely, and again longitudinally. Thus by growth a plate of cells is formed, ex- tending across the canal. As these cells enlarge, inter- cellular spaces are formed between them until the ma- ture diaphragm is produced. Stellate cells develop from the angle-cells in /&fc/ (b) the USUal way, viz., Fl°- n.— Development of air-canals. Transverse section of root^tip of N. jlavo- , fire/i*. From photomicrograph, a, earliest stage ; b, division of one angle-cell ; by extension OI the e, cell divisions nearly completed; e, epidermis; h, hypodermal layer; (a) is 11 1 farther from apex than (b), but from the same root. wall at several points, with subsequent thickening and deposition of crystals. This is exactly as described for Nuphar lutea by Trecul (1845). Lateral rootlets take their origin while the root tissues are still very young, even within the root cap. They spring from the pericambial layer opposite each xylem ray, and make their exit in the usual way, but not until a considerable time after their inception ; they do not occur in a mature state near to the root apex. THE STEM. Externally the stems of Nymphaeas offer some striking differences, which are at times characteristic of subgenera or even of species. Members of the Eu-castalia group have always a stout, spongy, horizontal rhizome which decays in its older parts as it elongates by apical growth. fa,J 3» The Waterlilies. In JV. tuberosa (var. parva Abbot) and JV. odorata mhtor this may be only 5 to 10 cm. long and 1.2 to 2 cm. in diameter, but in the typical forms of these species it becomes 30, 60 or even 100 cm. long and 4 cm. in diameter, while in the larger forms of JV. alba and Candida it may reach a diameter of 10 cm. In general shape these rhizomes are terete, though the largest one I have seen (JV. alba candidissima) was somewhat flattened dorso-ventrally. In the American species they are of a dull whitish color, though this is nearly obscured by fine black hairs ; in the European Fid. 13.— Diaphragms in the air-canals of roots ; a, d, e, early stages from transverse ; b, c, from longitudinal microtome sections of the root of N. flavo-virens ; /, mature diaphragm with beadlike intercellular spaces from the root of ZV. elegana X zanzibarlemi*. species the color is brownish or nearly black. All of the species of this group except JV. tuberosa branch rather infrequently in their stems, and the branches are stout and attached to the main shoot by a broad base. In JV. tuberosa, however, the rhizome (Fig. 13) is often crowded with short, tuber-like shoots from 1.2 to 8 cm. long and 1.2 to 2 cm. in diameter, attached to the parent axis by a short and very slender isthmus, sometimes only 0.3 cm. in diameter. In external appearance these do not differ from the main rhizome, except in size and a darker color. One such tuber often has others, one, two, three or four, of nearly equal size, growing out from its sides and base, so that bunches of tubers are formed. Being light enough Structure. 39 to float on water and easily detached from the parent plant and from one another by reason of the brittleness of the tissues of the isthmus, these tubers serve to distribute the plant widely. Waterlily leaves (" lily-pads ") are eaten by large herbivorous animals (deer, &c), which wade about in the shallow water ; these must detach numberless tubers and leave them floating about, as a person does when wading among them. Doubtless also wading birds, turtles, bull-frogs and many kinds of fish accomplish the same end. Special mention of the structure of these tubers will be made later. Each of the Nymphaeas except the group just mentioned possesses a short, thick, erect caudex. Where this remains throughout the year Fio. 13.— Rhizome of N. tuberoaa. Natural size. in a wet and more or less vegetative condition, it dies off in the lower part as it elongates above, about keeping pace in its elongation with the amount of sedimentation taking place around it. Such conditions habitually surround the Xanthanthae and Chamaenymphaeae, and these cannot otherwise survive. Species of the Hydrocallis group often have a like perennial growth, as also sometimes have the Lotos and apocarpous groups in cultivation. But the habit of all of the truly tropical members (Lotos, Hydrocallis, Apocarpiae) is to be dried off completely at one season of the year. In this case the large mature blooming plants die and rot away. N. flavo-virens alone, with some of its hybrids, is able to withdraw from vigorous flowering growth into a dry, resting tuber. Young plants, however, readily store away all of their nourish- ment in the caudex, which then forms a tuber from the size of a pea to that of a hen's egg (or much larger in N. flavo-virens). Such resting 40 The Waterlilies. tubers are usually of an ellipsoid shape, or spheroidal in N. gigantea and elegans (Fig. 14). They are of a brownish black color, smooth and rounded in the lower half, but more or less acute above, because the apical region is protected by a dense upward growth of long fine hairs intermingled with dried and shriveled bases of petioles. These tubers live through the dry season without injury, and begin growth again in a characteristic manner on the return of sufficient moisture and heat. Mature stems of the types under discussion are, like the tubers, of an ellipsoid shape, and have grown in our Botanic Garden to a length of Fio. H.— Tubers : a, N. amazonum ; b, b', N. gigantea, from two sides 90 degrees apart ; c, e', N. eapensis, two specimens; d, a", N. elegans, two specimens; e, N. tanzibariensis rosea; f, N. lotus, seedling; g. If. lotus, lateral tuber from mature plant. All natural size. 20 cm., and a diameter of 10 cm. in N. rubra var., 12.7 by 7.5 cm. in N. dentata, caerulea and zanzibariensis ; I have seen a resting tuber of N. Mrs. C. W. Ward (N. flavo-virens X zanzibariensis rosea) in the Dreer Nurseries which measured 21 cm. by 10 cm. All of these are of a brownish black color. With us N. flava has not exceeded 7.6 cm. long by 1.6 cm. broad, and has a yellowish color, but Mrs. Treat (1877) records the length of an old caudex in the St. Johns River, Florida as 60 cm., and diameter 10 cm. N. tetragona stem is rather slender, 7 to 12 cm. long by 2.5 cm. in diameter, and is of a blackish color ; N. fennica is similar in appearance but stouter, 10 cm. long by 3.75 cm. in diameter. No branching of the stem, that is, no development of vegetative buds, occurs in N. tetragona, and none in any of the apocarpous species Structure. 41 ( neglecting some hybrids), except in breaking out from the resting tuber. N.fennica may rarely fork, as one of our specimens shows. The Lotos and Hydrocallis species, however, give rise more or less freely to buds from the lower part of the stem, and these develop into young plants entangled among the roots and leaves of the parent. On the approach of winter or the dry season, these lateral shoots harden off into resting tubers for the perennation of the plant. Tubers thus produced (in N, lotus) are very irregular in shape (Fig. 14 g), appearing knotty and gnarly, but they still show in their lower half a smooth surface, and on the upper parts a dense growth of long protective hairs intermingled with remains of dead petioles. A remarkable development of lateral shoots occurs in N. mexicana. Long stolons originate from the lower part of the stem, even in very young plants, and run horizontally just beneath the surface of the mud to form a new plant at a distance of 15 to 60 cm. from the parent. No sooner has the young member established a few roots and leaves than it in turn sends out runners, and so on. Three or more such runners may be found attached to a single plant. Their minuter details will be described in another place. In autumn the ends of these runners develop a peculiar form of perennating stem (Fig. 15); it consists of an axis 1 to 1.6 cm. long and 0.48 cm. in diameter lying horizontally or obliquely in the soil ; on its upper surface is a continuous row of 3 to 6 small leaf buds, while from the under surface depends a bunch of thick, swollen, starch-laden roots 1 to 3.5 cm. long and 0.3 to 0.6 cm. in diameter ; these lie crowded over each other somewhat in rows like bananas on a bunch ; at the tip of each a flattened and dis- tended root-cap remains. A last type of stem exhibited by Nymphaea is that produced from resting tubers when they resume growth. From the center of the upper end of the tuber, in apocarpous species, a short, smooth, white stem (stolon), 0.16 to 0.3 cm. in diameter, rises a centimeter or so, and at its summit a tuft of leaves and roots is formed ; this may now be broken off and will grow into a perfect plant, while the tuber will repeat the process, if strong enough, as many as 8 or 10 times (N. flavo-virens and hybrids). In Lotos species a much greater area of the tuber, in fact nearly all of its upper surface, develops a very large number of tiny shoots, crowded all over it. One, two, three or four of these start forward at nearly the same time and each develops a long internode 1.2 to 8 or 10 cm. in length — a smooth, slender, white or reddish, often much curved stolon 0.16 to 0.3 cm. in diameter — bearing a simple lanceolate or subulate bract 0.6 to 2.5 cm. 42 The Waterlilies. long at its summit. This is followed by a second internode, much shorter, 0.3 to 1.3 cm. long, and this bears the tuft of leaves and roots which are the beginning of a new plant. If these shoots are detached, other shoots develop, and so on almost indefinitely, even from a tuber no bigger than the end of one's thumb. Phyllotaxy in the Nymphaeas presents some peculiarities. As might be expected, forms with elongated rhizomes (Eu-castalia) have the simpler orders of leaf arrangement, and where dorso-ventrality appears leaf scars are only half as numerous on the lower as on the upper side. In N. tuberosa and odorata, the phyllotaxy seems to be as low as f or f, and the internodes are 1.3 cm. or more long. N. alba shows a much higher order, probably /r or \\ (Raciborski 1894). But the erect tuberous stems are so densely crowded with leaf attachments as to leave only narrow, sinuous lines of stem-sur- face visible ; the order of the spirals I have not determined. One can speak of a definite anthotaxy also in the genus Nym- phaea, since a plant once in flower produces its bloom continuously until it dies. In N. gigantea the peduncular attachments form sepa- rate and distinct secondary spirals alternating with two parallel spirals of leaves (Caspary, 1865). In other species the flowers occur in the place of leaves, as members of the leaf-spirals. In every case the peduncles arise entirely independently of any leaf axil and there is no visible trace of a bract. The relation of leaves and flowers in the primary leaf-spirals is shown in the following series taken from Raciborski (1894). The numbers represent leaves, the letter f stands for a flower. A rhizome of A^. alba taken up in autumn gave 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, f, 12, 13, f, 14, 15, f, 16, 17, f, 18, 19, f, 20, 21, f, 22, 23, 24, 25, 26, 27, 32> 33- f. 34. 35. 36, 37. 38, 39, 40, 41. 42, 43. f. 48, 49, f, 50, 51, f, 52, 53, f, 54, 55, f, 56, 57, f, 62, 6$, f, 64, &c. ; the flowers up to 45 have already bloomed, the others would come out the next season. In N. zanzibariensis, a very floriferous species, leaf and flower alternate regularly after the plant once begins to bloom. " N. stellata" gave 1, 2, f, 3, 4, 5, f, 6, 7, 8, f, 9, 10, 11, f, Fig. 15.— Perennating body of N. mexieana. Seen from nearly opposite sides. Natural size. 28, 29, f, 30, 31, f, 44, 45, f, 46, 47, f, 58, 59, f, 60, 61, f, Structure. 43 12, 13, 14, f, 15, 16, 17, f, 18, 19, 20, f, 2i, 22, 23, f, 24, 25, 26, f, &c; the primary spiral seemed to be sometimes xr» sometimes higher. N. rubra in the Munich garden (which Raciborski considered was more probably a hybrid) showed 1, 2, f, 3, 4, 5, 6, f, 7, 8, f, 9, 10, 11, 12, f, 13, 14, 15, 16, f, 17, 18, f, 19, 20, 2i, 22, f, 23, 24, 25, 26, f, 27, 28, f, 29> 30, 31, 32, f, 33, 34, 35, 36, f, 37, 38, f, 39, 40, 41, 42, f, 43, 44, 45, 46, f, 47, 48, f, 49, 50, 51, 52, f, 53, 54, 55, 56, f, 57, 58, f, 59, 60, 61, 62, f, &c.' The structure of the stem in Nymphaeaceae has long been a puzzle, and for the most part still remains so. A transverse section of any vegetative stem of Nymphaea (Fig. 16) shows to the unaided eye three well-marked regions, viz., (1) an outer wall of dense cellular tissue, about 0.16 cm. thick, consisting of epidermis and exo-cortex, (2) a medio-cortex full of large lacunae or air-spaces, (3) a dense central core traversed in every direction by vascular bundles. The second layer is traversed by numerous trabecular of denser tissue connecting the exo-cortex with the central core. The following description applies only to normal vegetative stems, not to stolons or resting tubers. Microscopically examined, a well-marked epi- dermis is always present. In N. lotus and odorata /T^. ..... ,. 111 Fig. 16.— Rhizome of iV. odor- (rig. 17) this is an irregular layer as to the level ata; transverse section of in- of the cells on the outer and inner sides, and as 5^£££253£ to their size, though they are approximately cubical afeasvyNaaturaair8eurSCUlar bUn in shape. The walls are equally thin on all sides. In N. flava the cells are much shallower than long or broad (about 2:1) and form a more even surface. With the exception of a very minute area at the growing point of the stem, its whole surface is covered with hairs. Where these first appear outside the naked area, every epidermal cell bears one. At a short distance back of this point, however, many of the epidermal cells undergo an oblique division, cutting off a wedge-shaped segment with its apex at the outer surface on one side of the cell, and a base nearly equal to half the base of the mother cell ; the apex of the wedge widens out, and soon the new cell is cubical. A single cell thus comes to intervene between two formerly contiguous 1 Raciborski's TV. flava, having no runners, but an ordinary branching rhizome as in other Nymphaeas, was almost certainly N. odorata sulfurea (N. odorata X flava) and not the true N . flava. His A'', slellala was probably N. caerulea Sav. 44 The Waterlilies. Specie*. Diameter of bain at apex of stem. N. caerulea N. tuberosa N. lotus N. flava N. odorata O.024-O.030 mm. 0.035 mm. 0.021-0.028 mm. 0.021-0.033 tarn. 0.028-0.033 mm. and hairs. Such hairs are slender lines of terete cells, of almost even diameter throughout, but tapering at the apex, and attaining a length of 0.48 cm. in N. caerulea, 0.6 cm. in N. tuberosa, or 1 cm. to 1.3 cm. in N. lotus. The basal epidermal cell is cubical in shape, and if we represent its depth by unity, the lengths of the succeeding cells will be approximately 0.8 to 1.3, 0.5 to 2.2, 2.2 to 10, so on up to 15 or 20 times as long as broad. The relative and actual diameters of these hairs in differ- ent species is shown in the accompany- ing table. In addition to these, hairs with thick multicellular base and slender shaft occur in N. tuberosa and lotus. The lower part of such a hair consists of 1 to 3 tiers of polygonal cells about one-third to one-half the diameter of the basal column of the hair ; then follow 2 to 5 short cells in linear series, and finally a row of long cells as in the hairs first described. These complex hairs may arise from the simpler type by upgrowth of the tissues supporting them. Those of N. lotus are much larger than of N. tuberosa ; indeed in the former species the surface of the stem is completely occupied by these hair-protuberances . Much more sharply than in the root, we recognize in the stem three cortical regions (Figs. 16, 17, 18), an outer, dense, sub-epidermal layer 0.16 to 0.3 cm. thick, a median spongy layer 0.6 to 2 cm. thick, and an inner dense layer continuous with the parenchyma which makes up the central core of the stem. The spongy layer is not continuous, but is cut into irregular disconnected blocks by plate-like trabeculae of dense tissue passing radially from outer to inner cortex. The cells of the outer cortex in N. odorata (Fig. 17) are nearly isodiametric, thin-walled, and contain 6 to 8 starch grains each; there are 12 to 14 layers of Fig. 17.— Microscopic section of rhizome of JV. odorata; showing exo-oortex and 3 chains of the spongy medio-cortex, e, epidermis; t, tannin (latex) cell. Camera lucida sketch. Structure. 45 such cells, with occasionally a small intercellular space, and here and there a typical stellate cell many times as large as the cortical cells. In N. lotus the exo-cortex is 23 or 24 cells thick ; each cell is ellipsoidal, the radial axis (with regard to the stem) being shortest, and the horizontal tangential axis longest. The walls appear thickened, and probably are colloid. There are no starch grains, no intercellular spaces and no stellate cells. The relative amount of lacunar cortex is very different in the different types of stem. In the Castalia group it is quite prominent (Fig. 16), in the others rather insignificant ; the following table shows the relation in five types examined. " Central core " includes endo-cortex and all within it. Species. Width of exo-cortex Medio- cortex Central core. N. flava N. tuberosa N. odorata N. alba candidissima N. lotus 0.1 cm. . 0.16 0.13 0.16 0.2 0.4 cm. 0.8 0.3 1 0.8 0.48 cm. 0.8 45X3.8 6.4 In N. lotus (cf. Fig. 18 m) the lacunar cortex might be said to occur in isolated patches of irregular shape around the periphery of the stem. The greater part of these regions is air space, and this is traversed in every direction by anastomosing chains of cells, after the manner of a loose, spongy parenchyma of a leaf. The cells are thin-walled, small, and nearly isodiametric in N. odorata and alba candidissima, and the chains may consist of one or more rows of these (Fig. 17), being sometimes rather plate-like ; the nodes of the network are always multicellular. In N. lotus, on the contrary, the cells are three to four times longer than broad, or they may have a rounded body with slender arms reaching out to connect with neighboring cells. The inner cortex and the tissue forming the groundwork of the central core of the stem alike consist of medium-sized spherical cells, containing numerous starch grains, and enclosing large intercellular spaces. There are about 25 cells radially outside the vascular region in N. alba candidissima, constituting the endo-cortex. In N. lotus the endo- cortex may be 0.4 cm. deep in a large stem, and consists of very many cell layers. The cells are wide apart, reaching out toward each other by short, narrow processes, and are surrounded by a large amount of air space ; the tissue is of a whitish color. The cells and spaces are smaller, but otherwise it does not differ from parts of the brown medio-cortex. 46 The Waterlilies. The endo-cortex contains some starch. The ground tissue of the central core of the stem consists of spherical cells with large intercellular spaces, as in N. alba candidissima, and is densely filled with starch, even in a large vegetative stem taken up in autumn and otherwise destined only to decay. In N. flava (Fig. 21) the endo-cortex and ground-tissue are constituted as in N. alba-candidissima, but the cortex cells seem more crowded together than the central tissue of the core, and more densely laden with large starch grains. Stellate cells are occasional in the central core, and very plentiful in the cortex of this species. The trabecular which cross the medio cortex (Fig. 16) are composed for the most part of tissue similar to the endo-cortex, but at the outer ends they gradually take on the character- istics of the exo-cortex. The number of trabecular in a cross-section is about 5 in N. flava, 7 or 8 in N. odorata, 8 in N. tuberosa, 14 to 15 in a very large rhizome of N. alba candidissima, 1 5 or 1 6 in a large caudex of N. lotus. On stripping off the exo-cortex of a stem o{ N. odorata, I found eight bands of trabecular tissue running longitudinally throughout the length of the portion studied, connected by trans- verse or oblique cross-trabeculae at intervals of a half-inch or more. The petioles and peduncles are always situated over a longitudinal trabecula, in which the vascular bundles travel to join the central vascular plexus ; but these leaf and flower insertions stood in no discoverable relation to the transverse and oblique trabecular ; in two cases carefully dissected there were no cross-trabeculae near the insertion of the leaf. In N. flava slender trabeculae divide the medio-cortex into fusiform areas, each lying with its longer axis horizontal and tangential to the stem ; a leaf or flower or stolon stands over each such area, and a stout vertical trabecula crossing the short diameter of the area bears the central leaf trace to the vascular cylinder of the stem, while the lateral traces travel inward near the points of the fusiform area, surrounded by a thin layer of dense cortex. Caspary (1857 b) states that such a division of the spongy cortex into areas is common to Victoria, Euryale and Nymphaea. Fig. 18.— Caudex of N. caerulea; transverse sec- tion. Heavy lines represent vascular tissue; e, exo-cortex ; /, peduncle ; m, medio-cortex ; n, endo- cortex ; p, petiole ; r, root ; 8, stipule. Natural size. Structure. 47 In waterlilies which are adapted to enduring a dry season, the resting tubers differ markedly from the vegetative caudex. In the case of young seedling plants, which alone form resting tubers in most apocarpous species, the tuber acquires a smooth, corky bark on its lower parts. The leaf and root scars are obliterated by intrusion of cortical cells, which absorb the spiral thickenings of the xylem (Caspary, 1859) and give rise to several layers of typical cork. Although I have never found these in section to have the regularity of arrangement shown in some trees with a well-developed cork cambium (Ginkgo), I have often with my hands peeled off the bark from a living tuber, and found it to separate as smoothly as the bark of a willow twig in April. The interior tissues are dense and hard with their load of starch, and the vascular system remains intact except in its peripheral parts. The irregular lateral tubers produced upon the sides of large plants of the Lotos group seem to possess much less vascular tissue than the vegetative caudex and to have much more cortex, heavily laden with starch. The lacunar cortex is wholly absent from resting tubers, probably because, as in N. flavo-virens, the cork layer is formed inside the lacunar tissue, and all of this region dries or decays away. T t- .. , r^, , ... Fio. 19.— Cork forma- ln h.u-castaha and Chamaenymphaea, subenzation tion at base of root soar, takes place only over the scars of petioles, peduncles *— ■ and roots, and on the boundary of the living and dead tissues at the old decaying end of the stem. The root scars are healed over by a curved surface, concave on the outer side, in the middle of which the vascular bundle projects slightly. The cortical cells along this concave surface (which lies four or five cells farther in than the proximal ends of the air-canals of the root) undergo one to three parallel divisions (Fig. 19) to form the cork layer, and not until some time afterward do the adjoining root tissues decay away. I do not know how the vascular bundle of the root is closed. In N. alba candidissima, where the above facts were observed, the old rhizome, decaying at its posterior end, seems to have its epidermis shriveled and more or less disintegrated and its exo-cortex suberized in its cell- walls, but usually without cell-division. When this layer peels off, carrying with it the leaf scars, a rounded corky invagination or pouch is seen (Fig. 20 b) at the base of each main air-canal of the petioles and peduncles. The largest pouches extend inward a quarter of an inch, and point slightly backward ; they are hollow, and wide open to the exterior, and are usually full of mud. The whole surface of the 48 The Waterlilies. £LX *f* JL (c) a>) TVS— scar is suberized in a manner identical with that of the root. In one case (Fig. 20 c) the spiro-reticulate elements of a bundle in a leaf scar were closed by large thyloses from the neighboring parenchyma. Since but 3 to 5, usually 3, layers of cork are formed, the protection is only temporary ; and when the water breaks through into the underlying tissues, a thin corky layer is thrown across the cortex in front of it, sometimes simply by modification of the cell-walls without division of their contents. Soon the water breaks through this, and again a partition is thrown across farther forward. So in the older parts of the stem, the plant keeps up a continually losing conflict with death, resulting in a gradually advancing decay of the tissues. In the outer, corky cells of the scars, and to a less extent in the inner ones, crystals are found. Some of these are irregular and rounded in outline, but most are beautifully rect- angular, prismatic or tabu- Fio. 20.— (a) Old leaf cushion on rhizome of N. alba candldUsima, Jar. The Tegular CrVStals showing: base of petiole with air-canals, insertion of stipule, », and i root scars, r; (b) transverse section of (a) through the two largest may be transparent, homO- air-canals ; (<•) thylosis closing spiral vessel in old leaf scar. , , . . geneous and sharp-angled, or they may be surrounded by a bunch of needles projecting from the middle line toward either end, or they may present much the appearance of bundles of raphides. Similar crystals and crystal masses are found in the suberized exocortex of the older parts of the stem. They are very resistant to reagents, and consist probably of calcium oxalate. They are often imbedded in masses of mucilage which becomes pinkish in chlor-zinc-iodide. The irregular, crystal-like bodies blacken with ferric chloride solution, and doubtless contain tannin. Throughout this research cork has been tested by its yellow color when treated directly with chlor-zinc-iodide, and by a blue color with the same reagent after boiling in potash solution and washing in hot alcohol (Meyer 1898). Before taking up the vascular system of the stem it will be of advantage to describe some special types of stem-structure. The remarkably developed stolons of N. Jiava and the lesser ones of N. lotus have the same general structure. Inside of a thin but cuticularized epidermis of small cubical cells, and about eight layers (JV. flavd) of closely fitting cortical cells (two to three layers in N. rubra var.), the Structure. 49 space is filled with longitudinal air-canals separated by partitions of a single cell-layer in thickness. Throughout the length of the stolon three to five vascular strands pass along, each surrounded by endodermis with the characteristic thickenings on the radial walls. Gwynne-Vaughan (1896, a, b) called these strands "steles," and considered the stolons polystelic. He states that the narrow isthmus connecting the tubers of N. tuberosa with the parent axis are similarly polystelic. I am not prepared to express an opinion upon the stelar nature of these strands, especially since the present idea of a " stele " seems to be of doubtful value, and, to say the least, needs more accurate definition. It will suffice to describe the appearance of the vascular strands in transverse section, and to mention their constituents. In N. flava a strand has at center a large, regular and thickly cuticularized air-canal, with two or three (usually three) projecting vascular lobes. Each lobe has one to three (sometimes none) spiral xylem cells next to the cells bounding the air- canal ; the rest of the area consists of phloem, with large sieve-tubes and small companion cells. The vascular strands of N. rubra stolons have the large air-canal with only one or two vascular lobes. In N. tuberosa the very short stolon resembles the rhizome in its epidermis and superficial hairs, and the large, starch-laden cells of the interior ; these latter cells include very large intercellular spaces. There are about three vascular strands, bounded by endodermis. At the middle of each, or near one side, is a large group of spiral elements, corresponding in size and position to the air-canal in N. flava. A mod- erate amount of phloem lies about this, either concentrated in four phloem lobes, or half surrounding the xylem and bearing two or three lesser lobes, or completely surrounding the xylem and scarcely lobed. Two or three of these conditions are seen in a single section. The vascular system of nymphaeaceous stems has always been described as consisting of scattered bundles (astelic), resembling mono- cotyledonous structure. In most cases this is about all that can be said (cf. Figs. 16, 18). Trecul (1845) studied the stem of Nuphar for a year with considerable success, but was far from being satisfied with his results. Nor do his researches shed much light upon the vascular structure of Nymphaea. Although we can not yet unravel the complicated network of tracheae and phloem strands, which seem to run in every imaginable direction and position throughout the inner tissues of the stem, some points are brought out by an examination of the stolons and young plants of N. rubra, and the general and nodal structure of N. flava. This latter 5 5o The Wateri.ilies. species has an additional interest as representing probably the most primitive member of the syncarpous series. In a transverse section of the caudex of N. flava (Fig. 21), one finds a pretty well marked cortex, a broad vascular ring with one to three leaf- gaps, and a well defined pith which is continuous with the cortex through the leaf-gaps. The vascular ring, however, presents a most peculiar Fig. 21.— Transverse section of caudex of JV. flava. b, base of leaf with bundles and air-canals ; e, exocortex ; g, leaf gap ; I, point of branching of leaf trace ; m, medio- cortex ; n, endocortex ; p, pith (medulla); r, root trace ; t, central leaf trace. structure. Projecting slightly into the pith from the inside of the ring are seen seven or eight vascular bundles, t, of varying size, cut transversely, each with its phloem against the pith, and its xylem (of equal or greater amount) next to the vascular ring ; these are the central leaf-traces. The greater part of the "ring " consists of horizontal or oblique elongated elements, mostly of phloem, but with some spiral tracheae ; these are derived from root-traces and lateral leaf-traces. In this ring and sometimes on the outer side of it, next the cortex, are additional bundles Structure. 5i cut transversely, with xylem inside and phloem next the cortex ; their origin or end I was unable to learn. On the surface of the stem at each node or point of union of a petiole with the caudex, there is in N. flava a prominent conical projection, on the upper face of which the petiole is attached, and on the lower face the roots. Within the tissues of this projection the bundles from the attached roots unite into a single trace, and this passes horizontally through the cortex to the lower edge of the leaf-gap. A few cells join the vascular cylinder of the stem directly, but most of the trace is contained in two large lateral forks, one right and one left, which start almost horizontally into the vascular cylinder and could not be followed. The petiolar Fio. 22.-Course of vascular bundles at node of Nymphaeaflam. «i) base of petiole ; (b) base of peduncle • g leaf gap; rr, roots; ». vascular cylinder of stem, semi-diagrammatic; (c) diagrammatic projection of (a); p, petiole. bundles are nine in number, three in a median vertical plane, and six smaller lateral ones. Of the first three, one is anterior, one median, and one posterior. The others are right and left antero- and postero- lateral, with one directly right and one left of the centre of the petiole. On entering the stem, the posterior and median bundles fuse, and are shortly joined by the anterior, the whole forming a large central leaf-trace. The right and left antero-lateral bundles are joined at about the same level as the above, by the directly right and left bundles, and the two fused bundles then join the central leaf-trace. This trace passes horizontally in through the cortex and comes in contact with the root- trace, then passes over the brink of the leaf-gap and turns sharply down- ward next to the pith, giving the appearance noted in the cross-section. About midway on its course through the cortex, the central leaf-trace 52 The Waterlilies. gives off a strong branch to either side. Each branch passes about 45 degrees around the stem and then runs obliquely into the vascular cylin- der. Meanwhile each has been joined near the middle of its course by a postero-lateral bundle from the petiole, which, on first entering the stem, bent sharply away from the other bundles and ran along horizontally in the exo-cortex to the corner of the leaf-area before turning inward. The whole is made reasonably clear by a glance at Fig. 22, (a), (c). A very similar arrangement of bundles was found at the node of a stolon of N. rubra. It is usual in the Lotos group to find a slender leaf a short distance below the young plant near the apex of the stolons. This is the only place in any Nymphaea, I believe, where a leaf can be found at sufficient distance from others to admit of simplicity in the arrangement of the bundles. At some distance below the leaf in question there are four bundles (steles of Gwynne-Vaughan) passing longitudinally in the stolon. Nearer the leaf one of these forks, making five bundles, nearly equidistant from one another (Fig. 23, upper end). Opposite the leaf- base, the bundle on that side of the stolon divides into two parts which run almost in contact through A single root is present here, and its The lower end is toward apical trace, r, passes obliquely inward toward the space region of stolon. , . „ -_ ,. . . between the recently split bundle, and part joins each half. Three bundles enter the stolon from the leaf, a central, c, and two lateral, b, e. The central bundle passes in parallel to the root- trace, and a part joins each half of the split stem-bundle. Midway of its course inward, however, it gives off a large branch right and left and these join the next nearest stem-bundles on either side. Each lateral leaf-trace is joined by another small bundle (probably stipular) just within the stem tissues, and the two unite with the right and the left branches respectively of the central bundle. Above this leaf (below, in Fig. 23) the two parts of the split bundle move farther apart, and finally they fuse successively, each with its nearest neighbor, giving rise to five and then the original four bundles of the stolon. Shortly, however, these fused bundles again divide in making ready for the insertion of the next two leaves. The bundles from these leaves are in number and mutual relations identical with those of the first leaf; but the leaves here become crowded and the details of the course of the bundles Fig. 28.— Vascular system of stolon of N. rubra ? at a node ; b, e, lateral leaf traces ; ("0- while in the center of a large petiole (or peduncle) there may be three phloem strands about a single air-canal. The four or six large peripheral bundles of the petiole usually have two phloem strands and one or two canals. These air-canals are quite small, scarcely visible to the naked eye, and are bounded by a very regular row of clear cells, whose walls next to the lumen of the canal are convex and strongly cuticularized. In very young petioles remnants of spiral elements may be seen in these canals, showing that they bear the same relation to the bundle as the well-known canals in the xylem of monocotyls, Zea mays, etc. Where the petiole joins the leaf several spirals are found at the sides of the canals, and these rapidly increase in number as we approach the collar until the canal is wholly obliterated, and there is, instead, a mass of spiral elements nearly or quite as large in cross-section as the phloem strands. The neat wall of the canal dis- appears pari passu with the encroachment of the spirals (cf. Caspary, 1858, p. 382, 385 ; 1888, p. 2 ; Solereder, 1898). The lamina of the waterlily leaf (Fig. 28) varies in size from 5 to 7 cm. in diameter in N. tetragona and odorata minor to 40 cm. or 60 cm. in the large tropical forms. It is cleft nearly to the center where the petiole is attached, and has been termed fissi-cordate. Nearly orbicular in most species, it is somewhat oval in N. caerulea, and decidedly so in N. elegans, pubescens, and tetragona. The margin is entire in Eu-castalia and Cha- maenymphaea, slightly wavy at base in Xanthantha ; in Hydrocallis and Brachyceras it varies from entire in N. amazonum and elegans to deeply sinuate in N. rudgeana, ampla, flavo-virens, capensis, and zanzibariensis ; in N. gigantea and the Lotos group the leaves are sinuate-dentate, with Structure. 65 wide rounded sinuses and sharp, almost acuminate teeth. The apex is usually rounded, but somewhat emarginate in N. caerulea and sometimes in N. odorata, especially the variety gigantea. In the sinuate-margined Apocarpiae, the tendency is for the apex to lie in a sinus, while in Lotos the midrib is continued out into one of the serrations. At base the two lobes may overlap or just touch or divaricate in the same plant or species, so that this feature is of very little systematic value. Near the petiole the lobes may be fused by their edges for a distance of 2.5 to 5 cm. in Lotos, 1.3 cm. in some Apocarpiae, or almost none in Castalia ; this causes the leaf to be more or less peltate, and the fused portion I have called the pelta (Fig. 28, p). At the apex of the lobes they tend to develop distinct points marking off the borders of the sinus (the space between the lobes) from the rest of the margin of the leaf; this point is acuminate and sharp in N. gigantea, capensis, and the Lotos group, obtuse but still produced and evident in Castalia, altogether absent and the lobes broadly rounded in N. amazonum. The thickness of the leaves of most species gives them a leathery feeling, but they are rather brittle and easily torn. Mostly lying flat upon the water, we have exceptions in the crisped margins of JV. jlavo-virens, capensis, zanzibariensis and the Lotos group. N. odorata gigantea often has the margin of the distal half of the leaf turned up at right angles to the water-surface to the height of 2 or 3 cm., somewhat after the manner of Victoria. A rich chlorophyll-green is the usual color of the upper surface of the leaves. In N. flava, mexicana, tetragona, rubra-rosea, amazonum and zanzibariensis, however, there is more or less of brownish-red blotching, and in N. rubra the whole leaf is of a bronzy-red hue. The under surface of the leaf is quite different. Dull green in N. tuberosa, it is Fio. 28.— Leaf of JV. flava, showing upper surface on right, lower surface on left half ; a, angle of lobe ; I, lobe; p, pelta; r, principal area; 8, sinus; t, astomatic area ; 1, 2, 3, 4, 5, primary veins. 66 The Waterlilies. reddish to dark crimson-brown in N. odorata, alba, tetragona and flava, dull brownish in the Lotos group, and blue-purple in N. zanzibariensis and gigantea. Dark brownish or blackish blotches and specks mark the under surface in N. flava, caerulea, ampla, elegans and amazonum. The leaves are rolled inward from the sides (involute) in venation. They come up through the water in a more or less erect position, and only on reaching the surface does the lamina bend over at right angles i to the petiole (Frank, 1872). The venation, as inferred above, is in general palmate, but there is a distinct midrib with pinnate branching. The lowest pair of veins turn backward into the lobes, and others radiate out at various angles ; it is sometimes difficult to decide whether the most anterior of these are distinct or branches of the midrib. Five {N. flava, tetragona) to twelve {N. zanzibariensis) veins on either side of the midrib originate thus directly from the base of the leaf, and were called by Caspary (1865) primary veins (Fig. 28, 1-5). They give off small branches along their course, but sooner or later fork into two nearly equal parts. A strong cross-vein soon connects the forks of adjacent veins, enclosing a distinct area between each pair of veins. The area whose long axis stands most nearly at right angles to the midrib was called by Caspary (1. c.) the principal area (Fig. 28, r), and its length in relation to the radius of the leaf was used as a standard measurement in all of his taxonomic work. The vein below it is the principal vein (Fig. 28, 3). Beyond these primary areas there is a series of similar meshes, becoming rapidly smaller out to the margin of the leaf. Here the outer boundaries of the ultimate meshes make an irregular, submarginal vein in all but the Lotos group ; in these, three or four veinlets run together to a point in the apex of each tooth of the leaf-margin. Great variety exists in regard to the prominence of the veins on the under side of the leaves. In Castalia the veins are easily seen, but they are marked by very slight and gradual elevations of the leaf surface. In N. amazonum (Hydrocallis) the veins are prominent beneath as narrow, low, but quite abruptly raised lines. Among the Apocarpiae, N. gigantea has the veins quite abruptly raised above the leaf surface, and standing out as a prominent reticulum ; N. zanzibariensis is similar save that the ultimate veinlets at the leaf margin do not project, and so we pass through N. flavo-virens to A^. caerulea, in which the primary veins alone are markedly prominent, and the tertiary branches are raised very little if any above the leaf surface. The Lotos group, however, has an extremely prominent network of veins. The Structure. 67 midrib near the base forms a ridge a centimeter high and 0.3 to 0.5 cm. wide, with vertical, parallel sides and rounded summit. The primary veins are but little lower ; the ultimate branches also, especially where they run out into the serrations, stand out like little cords. Thus by a system of underlying rafters the lamina is rendered firm and stiff. In the large leaves of Euryale and the gigantic foliage of Victoria this method of strengthening is carried to a remarkable degree of development. In describing leaf-structure we may speak of (1) the upper epider- mis, (2) the palisade layer, (3) spongy parenchyma, (4) idioblasts, (5) vein- structure, (6) lower epidermis, and (7) external hairs. Fig. 29.— Epidermis of leaf, N. tetragona: (a), lower epidermis with bases of hairs ; (b), upper epidermis with average stomata ; (c), upper epidermis with very large stoma from margin of astomatic area. The upper epidermis is perfectly smooth and even and somewhat cuticularized on its outer surface. The cells are square or polygonal and from 0.0127 to 0.0178 mm. across (in eight species and varieties). Their outlines are quite wavy in N. flava, and mexicana, less so in N. tetragona and alba candidissima, and nearly straight in N. amazonum, odorata, tube- rosa and zanzibariensis x . The cells vary in depth, but are about 0.01 mm. deep in N. rubra and odorata, 0.01 to 0.013 mm. in N. tuberosa, 0.013 in N. amazonum, 0.01 to 0.0178 mm. in N. tetragona, and 0.0178 to 0.02 mm. in N. zanzibariensis X . Stomata are present only on the upper surfaces of the floating leaves. They are usually small and round, Fig. 29, (b), and separated from one another by two to three or four epidermal cells ; their openings are comparatively large. The guard-cells stand on a level 68 The Waterlilies. with the surrounding epidermis. The size is quite constant in the eight species studied, namely about 0.0176 mm. in diameter ; those of N. tetra- gona, however, are slightly oval and 0.02 to 0.025 mm. wide by 0.025 to 0.028 mm. in length. The average number of stomata per square centimeter is, in round numbers, 340 for N. tetragona and odorata, 390 for N. tuberosa, 500 for N. zanzibariensis X , 530 for N. mexicana, 720 for N. ama- zonum, 960 for N. alba candidissima, 1,000 for N. flava. Immediately over the insertion of the petiole there is an area totally devoid of stomata, Fig. 28, (t) ; it is of a rounded shape with a slender prolongation over Fio. 80.— Leaf of N. rubra, vertical section ; o, air chamber ; Id, idioblast ; p, palisade layer ; 8, stoma ; up, spongy parenchyma ; v, a small rein. each primary vein, and varies in size from 1.2 cm. to 3.8 cm. across. Here the epidermal cells are slightly larger than elsewhere. At the edge of this area the stomata are at first far apart and extremely large, Fig. 29, [c) ; in N. tetragona I found them 0.05 to 0.052 mm. long by 0.038 to 0.045 mm- wide, in iV. alba candidissima 0.045 mm- l°nl» by 0.038 mm. wide, in N. odorata 0.043 to 0.045 mm- l°ng by 0.038 to 0.043 mm- wide, in N. zanzibariensis X 0.038 to 0.05 mm. long by 0.03 1 mm. wide, in N. ama- zonum 0.031 to 0.038 mm. long by 0.028 to 0.033 mm. wide. From these exceptionally large dimensions the stomata rapidly become smaller at greater distances from the petiole, reaching the average size within the length of 0.5 to 0.7 cm. A border of about 0.16 cm. in width around the margin of the leaf is also devoid of stomata, but there is no increase in size of those nearest the margin. Structure. 69 The palisade layer of waterlily leaves is very deep, and composed of parallel strings of cylindrical cells. Between these are long, narrow air- chambers, opening through the stomata above, and connecting with the lacunae of the spongy mesophyll below. In N. tetragona the palisade layer is about 0.09 mm. deep, in N. rubra (Fig. 30) o. 1 to o. 14 mm. in depth, being in each case one-half the thickness of the whole mesophyll, and consisting of two to three series of long, narrow cells. In N. odorata the palisade occupies a little more than one-third of the thickness of the mesophyll and consists of five to seven series of cells with an aggregate depth of 0.14 to 0.15 mm. In N. tuberosa there are four or five series of cells with a depth of about 0.16 mm., or about one-half of the mesophyll. N. zanzibariensisY. has five series of cells, making about 0.14 mm., or a little less than half the mesophyll ; finally, N. amazonum has four series of cells, altogether about o. 14 mm. deep, or half the depth of the mesophyll. Where the palisade layer is more than two cells thick, the uppermost ones are shorter than the rest, often but little longer than wide. Ameling (1893, p. 189) gives 0.0166 mm. as the average diameter of a cell of the palisade parenchyma of N. alba, and points out the interesting fact that this is exactly the same in a leaf 230 mm. in diameter and in one only 95 mm. across. In another table (p. 198) a leaf 190 mm. across had the palisade cells 0.019 mm. in diameter, and this is compared with a like measurement of 0.0197 mm. in a leaf of Victoria regia 900 mm. across. The close agreement in size of cells is striking. The lower ends of the strings of palisade cells rest against a very loose, irregular layer of mesophyll cells, the upper members of the spongy parenchyma (Fig. 30). They are large, thin-walled and turgid, and of the most varied forms ; they connect with trabeculae or rods and plates of soft cells which pass vertically from this layer to the lower epidermis. Great air-chambers lying between the trabeculae occupy five-sixths of the space of the spongy layer. Bounding and supporting the lower epidermis in N. odorata, tuberosa, fiava, marliacea chromatella, and zanzibariensis X. is a 1 to 3 seriate layer of smaller, fairly regular, and close-fitting parenchyma cells. This layer may be absent in N. rubra, amazonum, and tetragcma, the air-spaces of the mesophyll being bounded below only by the epidermis. Immediately over the insertion of the petiole the mesophyll cells are nearly cubical, with walls evenly colloid-thickened, and without intercellular spaces. A very striking feature of the mesophyll is the presence of innumera- ble idioblasts. In the spongy layer these are usually of the multipolar or stellate form. They project from every side into the air-spaces, and •jo The Waterulies. assume the most grotesque shapes ; one might truly say that of the thousands one sees, no two are alike. Since most of the arms end freely in open cavities, as do those in the air-canals of the roots, petioles, and peduncles, it is impossible to assign to them any function of strengthening the tissues as Arcangeli (1890, a) suggested, although bipolar and H- shaped idioblasts in leaves may have that use in a measure. Arcangeli's additional opinion that they serve as depositories for waste calcium oxalate seems to be a good one, although Kohl (1889) reverses the case and considers that in all such cases the salt is a by-product of the deposition of thickening carbohydrates on the cell-wall ; he imagines that the glucose or dextrose is transported as a calcium compound. There are, however, great numbers of idioblasts in the leaf, of still another type, which seem to have a definite supporting office. These are cells of great size, with one end spread out in three or four lobes against the inner side of the upper epidermis ; from this broad base a stout shaft runs down between the strings of palisade cells (Fig. 30, id). Reaching the spongy parenchyma, it may break up at once into several radiating arms, or may continue straight on down to the lower epidermis and attach itself there by a narrow base ; all imaginable intermediate conditions may be found. The walls of such idioblasts are extremely heavily thickened. In dried specimens these make prominences on the upper surface of the leaves, and since their number differs according to the species, they were used by Caspary (pachycysts, 1865) for systematic distinctions. The veins which ramify copiously throughout the leaf present several interesting features. The midrib is marked by a line of modified paren- chyma. In the palisade region this line is three to five cells wide. The cells are nearly cubical and very poor in chlorophyll ; the upper ones have thickened walls. These grade off insensibly into a plate of larger, close- fitting parenchyma cells which runs through the leaf to the lower epidermis. There is a broad arc of corner-collenchyma just within the epidermis on the lower side of the rib. Embedded in the plate of parenchyma are the vascular bundles, and along its sides are two {N. odoratd) to several {N. rubra, zanzibariensisX.) air-canals. Each vascular bundle is sur- rounded by a sheath of parenchyma cells much smaller than those of the neighboring tissue. Although a well-developed xylem of spiral tracheae and companion cells is present, it is less in amount than the phloem. Near the base of the midrib several types and positions of bundle are met with (Fig. 31). In N. odorata there is shortly below the epidermis a tiny bundle, and just below this a much larger one, both with xylem above and Structure. 7i phloem below. Under this is a double bundle with xylem in the middle and phloem above and below, then another double bundle, placed trans- versely, and finally a simple bundle with xylem uppermost. The lowest bundle is underlaid by a transverse band of thin yellowish collenchyma. Near the apex of the same midrib the lowest bundle with its band of collen- hyma still persists ; above it is an inverted simple bundle, and above this again, near the base of the palisade layer, is a simple bundle with normal orientation. Lateral bundles seem to arise along the midrib from the median double bundles, and as they go farther from the rib they swing round so as to present a normal orientation of xylem and phloem. At a point nearer the petiole than the section first described, the place of the transverse double bundle is taken by an irregular vascular mass, 1 » ft (c) i f*J$ Fio. 31.— Arrangement of vascular bundles in veins of leaf: (a) section near the base; (b) near apex of midrib, N. odorata; (c), (d) from iV. eapemi* X; e, upper epidermis; p, parenchyma; pal, palisade layer ; pe, lower epidermis and parenchyma ; ph, phloem ; *j>, spongy parenchyma ; x, xylem. consisting of one or two xylem areas, each surrounded by two to four phloem lobes ; each phloem lobe represents a bundle which emanates from the vascular collar, several of which fuse to form the transverse double bundle of the rib. The primary veins are traversed by a lower large simple bundle with phloem below and xylem above, and a similar smaller bundle above this. In all of the lesser veins the orientation of bundle tissues is strictly normal. The lower curve of the vein also is lined within with corner-collenchyma, but there is no interruption in the palisade layer above ; the bundle has a close sheath of small cells, and a wider investment of parenchyma. In all but the finest ramifications the veins lie nearer to the lower surface of the leaf and are made fast there by the continuity of their surrounding parenchyma with that which lines the lower epidermis ; for there is such a lining layer near the veins, even in N. rubra. The smallest veins lie just beneath the palisade layer (Fig. 30, v), where the exchange of crude and elaborated materials takes place. I have not observed any free endings of veins ; anastomosis seems to be universal. 72 The Waterlilies. Via. 32.— Development of mucilage hair of N. alba; 1-5, successive stages. (After Schilling.) The lower epidermis (Fig. 29, a) has been frequently referred to in previous pages. It is a perfectly continuous layer, 0.01 to 0.018 mm. thick, quite smooth on the outer surface, and consisting for the most part of polygonal cells, many of which are about twice as long as broad (0.018 X 0.036 mm.). In these is located the pigment which colors the under side of the leaf. More or less plentifully interspersed with them are thickened circular cells (suberized, according to Schilling, 1894) representing the bases of hairs. Frequently the other epidermal cells are arranged around these in a radial manner. This caused Barthelemy (1874) to speak of them as pseudostomata, and he thought he could see in the center of some of them a small perforation. His opinion lacks confirmation. The circular cells are about 0.02 mm. across in JV. flava, mexicana, amazonum, odorata, 0. rosea, and tuberosa ; 0.02 to 0.028 mm. in N. zanzibariensis X and alba j( || ~~^/f ~~~*f^yf candidissima, and 0.025 to 0.029 mm. in ^^-^^ A^ tetragona. Their relative number is shown by the following figures, represent- ing the number visible in a single field of my microscope (Leitz, oc. 2, obj. 7) : TV. flava 32, mexicana 19-20, amazonum 52, tetragona 21, odorata 36, o. rosea 34, tuberosa 28, rubra 46. In the last-named species permanent hairs stand on many of these basal cells. Hofmeister in 1868 mentioned N. alba in company with Fagus, Salix, Quercus, &c, as having hairy leaves only in the bud. From what has already been said concerning the occurrence of basal cells of hairs in the epidermis of the petioles and lower leaf-surface, it will be seen that hair structures always occur on these parts at some time in their history ; the main ribs of the leaves and their upper surface are alone totally devoid of them. On a leaf rudiment of N. flava a trifle over o. 1 6 cm. long, some of these hairs are already complete, but others continue to develop, so that on a leaf three times as long all stages may still be found. A young epidermal cell swells outward beyond its neighbors and is cut in half by a transverse wall ; the outer cell continues to swell, and is soon divided by a wall which is strongly concave outward, making a termi- nal spherical cell (Fig. 32). A second concave wall may follow the first in the apical cell. Now, according to Schilling (1894), who investigated N. alba and other aquatics, the cellulose walls of this cell become mucilagin- ized, the cuticle is ruptured, and the outer cell breaks down into a mass of Structure. 73 mucilage. Cylindrical and clavate mucilage hairs are common in N. alba (Schilling, 1. c), and appressed finger-like ones occur in N. rubra. They are shed as soon as the tissues of the leaf begin to elongate rapidly and become mature, being of use only to keep the developing organs em- bedded in a coat of slime. This phenomenon was noted by Irmisch in 1859. Subsequent to the development of mucilage hairs, or along with and among them, longer hairs are formed in most Nymphaeas. The early stages of these are exactly as just described ; the apical cell, how- ever, instead of remaining spherical, elongates and may undergo division Fio. 33.— Hairs and hair bases : (a) on petiole of N. odorata ; (o) on petiole of iV. rubra, the shaft being: shed ; (c) on petiole of iV. tetragona, showing- also the collenchyma Immediately below the epidermis ; (d), (e), from rhizome of N. tuberosa. into a number of cylindrical elements as in Eu-castalia. The terminal cell is bluntly rounded in Eu-castalia or acute in Lotos. These longer hairs are persistent all over in members of the Lotos group, but only at the bases of the petioles in Eu-castalia, and as a specially developed ring of long hairs at the summit of the petiole in N. amazonum. In Lotos the shaft of the hair is unicellular. Its base may be very much wider than the basal epidermal cells in the petiole of N. rubra and sturtevantii, or but little wider, as on the under side of the leaves of N. rubra. The hairs of the ring on N. amazonum consist of 1 2 to 15 cells, and reach a length of 0.48 cm. Stahl (1888) considered the mucilage and other hairs as intended for protection against animals (snails, &c), and Goebel (1893) suggested that they served to prevent contact of the young organs with one another in the bud ; but Schilling (1894) regards them rather as useful in preventing 74 The Waterlilies. the access of water to the embryonic tissues. The first idea could apply to the hairs on mature leaves as well as to the younger. It would be use- less to assert that either purpose is unimportant. In a few cases the leaves of Nymphaea have been viviparous. Plan- chon (1852, c) records this for "N. stellata" (probably N. micranthd) as grown by Van Houtte. The plant came from English cultivators, with whom it had never shown this tendency. The new plant springs from the upper surface of the leaf just over the insertion of the petiole. "Already visible," says Planchon, "though very small on the young leaf, the axis but a tubercle with very rudimentary leaves, [the bud] remains almost station- ary on the adult and functional leaf so long as this retains its connection with the living plant. If the leaf is detached artificially or by decay of the petiole, it puts out roots, develops its leaves rapidly, and seems to hasten to live an independent life." ' Further details of the process have not been recorded. It was observed in JV. micrantha Guill. & Perr. and given thus in the description of the species : " Folia . . . inter lobos saepius bulbifera, bulbillis sphaericis pubescentibus." Hooker (1850) also records the fact in a few words in describing JV. micrantha in the Botanical Magazine, where the leaf figured is viviparous, and Paxton (1853, a) mentions it in his " Flower Garden." Lehmann (1853, a) went so far as to establish a " Tribe III Bulbophyllon" for his JV. vivipara, JV. guineensis Thonn., and JV. mi- crantha Guill. & Perr. We have not considered these as specifically distinct. Water-leaves are produced in all seedling Nymphaeas to the number of 2 (JV. flava) to 15 or 16 before any floating leaves. Of these more will be said in another chapter. In species which have dry resting tubers the first leaves produced on resuming growth are also entirely submerged. JV. flavo-virens alone, with its huge tubers, is at times an exception to this rule ; small tubers follow the rule. The perennating bodies of JV. flava (doubtless also of JV. mexicand) begin the growth of each bud with a few water-leaves, and the other Castalia species produce them either in late fall or early spring, or both. These members also lose their floating leaves and produce only water-leaves in unfavorable conditions2 of culture, ' Deja visibles, mais tres-petits sur la jeune feuille, pourvus d'un axe en forme de tubercule, mais avec des feuilles encore tres-rudimentaires, sur la feuille adulte et pleine de vie, ils restent presque stationnaires tant que la feuille tient a son pied nourricier. Qu'on detache artificiellement cet organe ou qu'il s'isole spontanement par la decomposition de son petiole, aussi tot le bourgeon pousse des racines, developpe rapidement ses feuilles et semble se hater de vivre d'une vie indepen- dante. (Flore des Serres, 8: 285.) 'Bonnier (1800) found in several lakes of the Alps shoots of A^. alba without floating leaves, but with an unusually large number of submerged leaves, "as is often seen in streams with rapid current." Structure. 75 while the tuberous species in like circumstances lose all of their foliage and form tubers, from which they come out again first with water-leaves. A^r^ Fia. 84.— First leaves from germinating tubers: 1, first, 2, second submerged leaf; 3, first, 4, third floating leaf of If. gigmitea ; 5, second floating, 6, first, 7, second submerged leaf of N. flavo-virens ; 8, first, 8, second, 10, third submerged leaf of If. eaerulea ; 11, 12, first submerged leaf from brood body of N. mexicana and If. Jlava, respectively ; 13, second floating, 14, second, 15, first submerged leaf of If. elegans ; 16, first, 17, second, IS. third submerged. 19, first floating leaf of If. zamibariemiis rosea ; 20, 21, first and second submerged leaves of N. "Mrs. C. W. Word" (N. fiavo-virens X zanzibarienaU). All natural size. Wachter's (1897) observations on N. stellata, dentata, rubra, thermalis, and alba showed, by cutting off the roots or leaves from growing plants, that production of water-leaves could be induced by enfeebled nutrition, 76 The Waterlilies. especially by lack of elaborated materials in the absence of mature foliage ; cutting off the roots had a much less marked effect. I have noted also in N. rubra that young shoots acquire floating leaves much sooner if left attached to the parent tuber than if early broken off and planted out. It seems quite certain, therefore, that the occurrence of water-leaves in fall and spring in Castalia is directly due to the cutting off of the leaves by frost, and the suspension of vitality in very cold water ; and, except in the large and powerful tubers of N. flavo-virens, the young shoots of other species must gather strength in themselves before they are able to pro- duce floating leaves. In other words, a leaf-rudiment will, up to a certain ^3 Fig. 85.— Transverse sections of petioles of water-leaves; a. N. rubra, from tuber; b, N. ortorata, from rhizome in spring ; e, N. tuberosa, from tuber. age, develop into a submerged or a floating leaf according to nutritive conditions. Such submerged leaves as are now under discussion assume a more or less deltoid shape, varying to hastate (Lotos), cordate, or nearly orbicu- lar (Fig. 34). Specific shapes, so far as known, are given in the taxo- nomic section. The petioles are usually 2.5 to 7 or 8 cm. long, though sometimes longer, and rather stout in proportion. Structurally they show but slight differences from the petioles of floating leaves. The thickly cutinized epidermal cells contain many bright-green chloroplasts in the upper part of the petiole, where it is not covered by mud. The funda- mental tissue, too, is rich in chlorophyll throughout. Its outer two to four layers are more or less colloid in the angles, but only slightly so at most ; here also chloroplasts are most plentiful. The arrangement of air-canals differs from that of petioles of floating leaves, but in the larger water- leaves this character approaches nearer and nearer to the typical arrange- ment for the species. The tendency in petioles of water-leaves is toward the condition with two main canals (Fig. 35) and a great many smaller ones around and between these, with partitions but one cell thick. This approaches the adult structure of N. flava and tetragona, and points also to a probable primitive state in which, like Nuphar and the roots of Structure. 77 Nymphaea and Nuphar, the fundamental tissue of the petioles was traversed by innumerable small and equal canals. Specialization has enlarged some of these at the expense of others in Nymphaea, Euryale, and Victoria. The distribution of vascular bundles in petioles of water- leaves follows the plan laid down for floating leaves. Stellate cells occur in the air-canals, but no thyll-like bodies have been found. The lamina is of a reddish color in Lotos and Eu-castalia, but is plain green in N. tetragona, fennica, and amazonum, and Apocarpiae. Its texture is extremely thin and fragile in Lotos, much firmer, though still very thin, in Castalia, Hydrocallis, and Apocarpiae. The upper epidermis is composed of rather larger, somewhat sinuous cells, and possesses a few stomata. Arcangeli noted stomata on water-leaves of N. alba in 1890 (b), but Brand (1894) failed to find them, and considered Arcangeli's specimens to be arrested floating leaves. Wachter (1897) corroborated Arcangeli's observation by examining leaves in chloral hydrate, and extended the work to N. rubra, dentata, stellata, and thermalis, in all of which a few stomata were found on the upper surface of the leaves. The lower epidermis is like that of floating leaves. Great reduction has taken place, however, in the mesophyll. Arcangeli (1. c.) found the palisade layer in N. alba reduced to two strata of globular or ovoid cells, and Wachter found a single layer of crescent-shaped cells representing the palisade in N. rubra and thermalis. The spongy parenchyma is greatly reduced in quantity, and consists of loose cells with large intercellular spaces. Idioblasts are present, but few in number, and the vascular network is simple. The ordinary involute vernation is found in the water-leaves. They are, therefore, to be considered as reduced floating leaves, and no indication is offered of the wholly aerial habits which must have belonged to the distant ancestors of the Nymphaeaceae. Even the large aerial leaves of Nelumbo, from having stomata still restricted to the upper surface, must be considered as derived from floating leaves, and in no close relation to those of the probable terrestrial ancestors of the order. THE FLOWER, FRUIT, AND SEED. Few genera of plants offer such variety in size and color of flowers as is found among the waterlilies. The primitive yellow, but of a rich hue, is preserved in Xanthantha and the newly discovered N. sulfurea and stuhlmanni; white prevails in Castalia, varying to pink in N. fennica, odorata rosea, and alba rubra. All shades from pure white (N. dentata) to deep crimson (N. rubra) occur in the Lotos group, and from white 78 The Waterlilies. (A7", ampla) to pink (N. versicolor) and royal blue (N. zanzibariensis) in Apocarpiae. In Hydrocallis a creamy white prevails. Orange tints have been obtained in hybrids of N. flava and alba rubra (JV. aurora, seignoreti). In size, N. tetragona, fennica, and odorata minor stand at one end of the line, with flowers 3.5 to 5 cm. across, and at the other end are A^. rubra, dentata, zanzibariensis, and giganlea, whose flowers measure 30 cm. or more in diameter. There is, of course, great variation within a single species, according to the food supply and other conditions. Other general characters will be brought out in the following description of the flower and flower-stalk. So close is the resemblance between petiole and peduncle in Nym- phaea that we shall greatly abbreviate the present section by referring back to the previous one. The peduncles are always terete, and in color and presence or absence of pubescence are like the petioles of the same species (no brown stripes occur on the peduncles of N. tuberosa). In Eu-castalia the likeness extends to the length, diameter, and number of air-canals of the two organs. In Chamaenymphaea the likeness includes length and diameter. The same is true of Xanthantha when in deep water ; in shallow water, the peduncles are stouter. These species are also peculiar in that the upper part of the flower-stalk is somewhat four- furrowed with shallow rounded grooves. In all other species the peduncles are 1 to 2.5 cm. in diameter, and stiff enough to raise the massive flowers 5 to 30 or 40 cm. above the water. The more slender peduncles of N. amazonum (Hydrocallis) are yet quite stiff and reach to or very little above the water surface. In exceptionally shallow water the same may be true for all of the Castalia group. The epidermis presents nothing characteristic. It bears a large num- ber of hair bases. Beneath it there is a single layer of thin-walled cells in N. dentata, rubra rosea, devoniensis, amoldiana, and zanzibariensis. The inner wall of this layer is somewhat collenchymatous in N. zanzibariensis X , marliacea chromatella, alba candidissima, and gladstoniana. Following this layer in the above-named forms, or including it in N. rubra, lotus, slurle- vantii, caerulea, odorata, and tuberosa, there are 3 to 1 1 layers of collen- chyma as in the petiole. The character of the remaining tissue is also like that of the petioles. In Apocarpiae the cells contain a good deal of starch, but not so in Lotos and Eu-castalia. Four main air-canals traverse the peduncles of Castalia, but six is the standard number for the remaining groups ; these numbers are subject to an occasional variation of one more or less, but much oftener of one more. The canals are equal Structure. 79 and arranged symmetrically about the centre of the peduncle (Fig. 26). Outside of each of these are two rather prominent but much smaller canals, and in N. zanzibariensis each of these is followed by two more quite evident canals. Many smaller scattered ones are always present. Idioblasts are extremely scarce in the Lotos group, frequent in Apo- carpiae, plentiful in Castalia ; fibrous (bipolar) idioblasts occur throughout the peduncle in Eu-castalia and around the periphery of the upper part in Brachyceras. Thyll-like masses of soft cells are frequent on the sides of the larger air-canals in Apocarpiae, occasional in Lotos, and not found in Castalia. Vascular bundles are distributed on the same principle as in the petiole (Fig. 26). In the centre is a large double or triple bundle. Oppo- site each of the 4 or 6 main partitions there is a large double bundle, often with two well-formed air-canals in Apocarpiae and Castalia, but with the second canal a mere fissure without definite wall in Lotos. Midway between these bundles are smaller ones, and so on in several series, according to the size of the peduncle. Examples are shown in detail in the diagrams. The constituents of the bundles are exactly as described for petioles. Just beneath the flower the peduncle widens out to form a receptacle for the insertion of the floral organs. This widening extends over about 1.3 cm. in Lotos, making an inverted frustum of a cone with a slope of about 45°. In Apocarpiae, Hydrocallis, and Eu-castalia the transition is more sudden, almost like the limb of a salver-form corolla. The resulting receptacle is usually round, but in Chamaenymphaea its upper margin is square, with distinct and prominent angles. In Xanthantha there are four rounded prominences on the sides of the receptacle, with distinct grooves between ; the grooves are continued down a short distance on the petiole, but they swell again quickly above, so that the receptacle is nearly round at the line of insertion of the sepals. The color of the receptacle is usu- ally like the peduncle, but paler. Within the upper part of this region the vascular bundles of the peduncle unite into an extremely complex horizontal network, from which bundles are separated out for the floral organs. The complexity of this anastomosis is too great to admit of any definition or unraveling as yet. The flower proper has four sepals, of which one is anterior and outer. The two lateral sepals are covered on one edge by the anterior one, and on their posterior sides they overlap the posterior sepal. At times three or five sepals may occur. I have seen the smaller number in N. rubra 80 The Waterlilies. and the larger in N. elegans, and Eichler (1878, 2 : 183-186) records both numbers for N. alba. A trimerous flower of N. atnpla was the type of Gaudichaud's N. trisepala. Hexamerous and heptamerous flowers are known. When five sepals occur, two stand in place of one in the posterior position ; when only three are present, the posterior one is lacking (Payer, 1857 ; Eichler, 1. c). The number of petals varies from 7 (A7", tetragona) to 33 (Al. capensis). The 4 outermost are largest and alternate with the sepals ; over each interval between these, i. e., opposite each sepal, is a pair of smaller petals, making a whorl of 8. This seems to be followed in Brachyceras and Lotos by another alternating whorl of 8. Of this row usually only 4 or 5 are present in N. caerulea, while an additional row is found more or less complete in N. capensis and zanzibariensis. In Hydrocallis the petals are all regularly arranged in alternating whorls of 4. Here, also, the outermost stamens follow in one or two whorls of 4 ; then come one or two whorls of 8, the first of which has two members over each interspace between the preceding 4; within this it is impossible to make out any order. fio. «.-Fiorai diagram The stamens of the Lotos group follow on in more or forNymphaea. jegs definite whorls of 8, and the same may frequently be seen in Brachyceras. In old flowers of N. caerulea, long after in- florescence, when the organs are spread apart by the growth of the ripening ovary, I have several times found the stamens arranged in 16 vertical lines, with mostly 4 stamens in each line. The whorled arrange- ment is not distinct in Castalia, and Eichler (1. c.) states for A^. alba that there are 4 petals alternate with the sepals, and 4 alternate with these, i. e., over the sepals. These 8 are the beginning points for oblique lines of successively smaller petals, rising right and left with equal inclination. Each series has, on the average, four members. The inclination of the spirals is such that each successive petal stands over the space between two neighbors, so that in all 16 vertical rows of 2 petals each are formed. The arrangement of the stamens was not made out. The inner petals and stamens are inserted at various levels up the sides of the ovary (receptacle). There is a narrow, naked zone between the petals and stamens in Lotos and Anecphya, but in other genera the two sets of organs are in continuous series, except in some forms of N. Candida and tetragona. The carpels are always in a circle in the middle of the flower, abutting against a central prolongation of the floral axis. Structure. 8i We will now take up some details relating to the individual floral organs. The line of insertion of the sepals is usually smooth and rounded ; but in N. Candida the bases of the sepals are slightly produced downward, and in N. fennica the sepals make a sharp, clean-cut angle with the receptacle. The general shape of the sepals is ovate. They are always greenish and chlorophylloid outside, and soft and petaloid inside ; the margins which are covered in the bud are wholly petaloid. Specific details of shape, size, color, and texture are given in the taxo- nomic portion of this paper. The epidermis of the outside of the sepals has a very smooth surface, and consists of irregularly polygonal cells in sur- face view, nearly square in section. Bases of mucilage hairs are frequent in N. tuberosa and plentiful in N. odorata. In the latter species, also, stomata are present. They begin at or just below the middle of the sepal and are more plentiful toward the apex and are a little larger than those of the leaf (0.015 by 0.015 mm.). The cells of the inner epidermis are rounded and prominent on the exposed side, of approximately rectangular shape, and shallower than those of the back of the sepal. Here also rounded cells like the bases of mucilage hairs are frequent, but no stomata. The parenchyma of the sepal is traversed longitudinally by vascular bundles and in the lower parts by air-canals. Inside the outer epidermis there are in N. zanzibariensis x about 8 tiers of close-fitting, nearly spherical cells containing chlorophyll ; they become gradually larger from without inward. Then follow 1 or 2 flattened cells and a vascular bundle, or, between the bundles, 2 to 4 flattened cells and an air-canal. About 5 roundish cells intervene between the air-canals and the inner epidermis. At the apex of the sepals the structure differs from this only in the absence of air-canals. Just within the epidermis on each side two or three layers of cells are feebly collenchymatous, but most noticeably so on the inner side of the sepal. The air-canals in this hybrid are rather large, round or oval, and arranged in a single row between the middle of the parenchyma and the inner epidermis. The partitions between them are two or three to several cells thick. The vascular bundles have a well-developed protoxylem, and at some distance from this, in the larger ones, a small air-canal. There is a row of larger bundles about the middle of the parenchyma and a row of smaller ones between the air-canals and the inner epidermis, with still others between the canals. Many stellate idioblasts project into the air-canals, but none was found elsewhere in these sepals. Thyll-like ingrowths also occur. One of these consisted of but two dense clavate cells, another had a multicellular base surmounted 7 82 The Waterlilies. by a crown of ovate cells with dense granular contents. Intermediate forms were seen. In N. odorata the sepal is quite thick at the base, but very thin above. In the basal region there are about 1 2 tiers of cells between the outer epidermis and the air-canals, of which the first 4 or 5 are slightly collenchymatous. They are all rounded in outline and become gradually larger from without inward. Among them are very many bipolar idioblasts lying lengthwise of the sepal. The air-canals are very numerous and irregular in shape and size. They are separated by partitions only one cell thick, and occupy a zone about as thick as the denser parenchyma just described. Great numbers of idioblasts of irregularly stellate forms occur here. Between this zone and the inner epidermis there is a zone of about 6 layers of parenchyma. A single irregular row of vascular bundles traverses the partitions between the air-canals. They have well-developed protoxylem and, in all the larger ones, an air-canal. The same structure is continued in the apex of the sepal, but with fewer tiers of cells and much smaller air-passages. Nearly all of the parenchyma cells contain chloroplasts, especially the two or three outer layers, where the cells are crowded together and polygonal in outline. A section near the middle of the sepal of N. tuberosa shows a still more leaf-like arrangement. Within the outer epidermis are two to four layers of round cells with large intercellular spaces and containing many chloroplasts. Then comes a zone 3 to 5 times wider than the first, made up of branching lines of cells and large intercellular spaces. Through this zone the vascular bundles travel, accompanied by one or two layers of close-fitting parenchyma. Phloem and xylem are well represented, but no air-canal was seen. The orientation of the bundles is normal in all parts of the flower. Stellate idioblasts are plentiful in the air-chambers. The inner epidermis is backed by a single continuous layer of rounded cells. Near the apex of the sepal the air-chambers are quite small, the intervening partitions are multicellular, and idioblasts much fewer. The total thickness of the sepal, however, may be greater than at the middle. Most of the general features of the petals have been already men- tioned, and specific details are sufficiently covered in the chapter on taxon- omy. The outer four petals are usually of about the size and shape of the sepals. They are thicker at base than elsewhere, and are traversed by a number of longitudinal veins which divide up into a fine network in the upper half or two-thirds of the petal. There may be some longitudinal air-canals in the basal parts. No distinction of blade and claw occurs, but Structure. 83 the oval to lanceolate lamina tapers gently to a comparatively broad inser- tion. Where the petals are brightly colored, the lower parts are usually paler, and often yellowish, even in blue flowers. The margins of the petals, as of the sepals, are always entire and are of softer texture than the sepals. The inner ones are shorter and narrower than the outer. Th^ structure of the basal and sepaloid portions of the outer petals of N. odorata is exactly like that of the sepals, even to the presence of large stomata on the back. The softer parts of these petals, and all of the inner ones, are much thinner. The epidermis of both surfaces is like the inner epidermis of the sepals, and has numerous hair-bases. Parallel and anastomosing veins traverse the parenchyma, accompanied in the middle line by idioblasts of varied shapes, but without air-canals. The stamens are quite characteristic in the different groups of waterlilies. In Castalia, especially N. alba, all gradations are found, from inner petals with a callous yellow tip or with one or two tiny anther cells, to stamens with petaloid base and a small but perfect anther, and so on to the innermost stamens, where the anther is large and thick and broader than the filament. N. amazonum shows similar but more rapid gradations ; the innermost filaments are filiform and the anthers narrowly linear. In N. gigantea, however, all of the filaments are slender, and anthers narrow. Lotos also shows no gradation from petal to stamen, but the outer filaments are as much as half an inch wide at base ; their insertion is wide and somewhat decurrent at the sides. The anther is broad and ligulate, with blunt, round apex. The inner stamens are shorter and narrower, ligulate in shape, and not decurrent. In all of the foregoing groups the connective is not continued above the anther cells more than as a limiting border to the anther, with the exception of the outermost stamens of N' gigantea, which bear a tiny subulate tip. In Brachyceras, however, there is a prominent prolongation of the connective at the apex of a large number of the outer stamens. This is least in N. elegans (0.08 cm.), small in N. flavo-virens, but may reach nearly 1.3 cm. in length in N. zanzi- bariensis. In this group, also, transitional forms between petal and stamen are sometimes seen. The outer filaments are ovate, often shorter than the lanceolate or linear anthers ; these grade by slow degrees into the innermost stamens, which are short and slender and without appendage. In all species the outer stamens are colored like the petals, excepting the yellow anther cells in Castalia and Apocarpiae and the reddish-brown anthers of some Lotos species. The innermost stamens are wholly yellow in all but N. rubra, where the reddish-brown color persists. 84 The Waterlilies. The structure of the filament resembles that of the inner petals. The epidermis bears hair-bases, the parenchyma is made of soft, rounded cells, and air-spaces are present. There are one to three or more longitudinal vascular bundles. The mature anther is covered by a single layer of square cells, a little deeper than wide, with thickened inner and lateral walls ; the outer wall is unthickened. The anther cells are parallel and introrse, with longitudinal dehiscence (Fig. 37, a, b). In Lotos and Brachyceras the division between the two loculi of each cell is quite evident after dehiscence, as the pollen dust remaining after most has been shed lies in two pairs of grooves on the inner face of the anther. ChifRot (1902) has recently made a minute study of sporo- phyll structure in Nymphaea- ceae, including nine species of Nymphaea. He finds slight differences in the venation of the stamens. He sums up his work (which I have not had time to verify) in a classifica- tion based on the distribution of idioblasts (" sclerites ") and stomata in the stamen, and of idioblasts in the carpels (1. c, p. 280). We quote this in translation. The names are correct and mostly well chosen, unless in the case of "N. blanda." The material of this came from the Kew Herbarium, where N. rudgeana occurs under the name of N. blanda, and is much more plentiful. Syncarpiae : Idioblasts rare in the filament, always present in the anther a. Idioblasts absent from filament, anther, and ovary (Xanthantha) N. flava. Idioblasts present in the filament and outer part of ovary (Castalia) c . a. Stomata present on upper epidermis of filament (Lotos). Idioblasts absent from mesophyll of ovary, present in carpellary styles A^. Ihcrmalis . a. Stomata wholly absent from the stamen (Hydrocallis) b. b. Idioblasts of but one kind, present in filament and anther "N. blanda." b. Idioblasts of two kinds, localized in the anther N. oxypetala. c. Idioblasts absent from the anther, but numerous in the carpellary rays d. c. Idioblasts present in anther, rare in carpellary ray N. tuberosa. d. Stomata absent from whole stamen N. alba. d. Stomata present at base of loculi on both faces of filament. N. pygmaea [ = N. tetragonal . Fia. 8?.— a, b, trausverse sections of anther of N. flava near base and near apex, respectively ; c, vertical section of stigma, showing papillae, N. tuberosa. All after Chifflot. Structure. 85 Apocarpiae : Idioblasts rare in the filament, always present in the anther. Stomata absent from all the surface of the filament (Brachyceras). Idioblasts absent from all parts of the mesophyll of the ovary and carpellary styles TV. caerulea. Idioblasts absent from filament and anther (Anecphya) N. gigantea. The dry microspores (pollen grains) of N. alba are oval and deeply folded i»»on one side (Kerner, 1895, fig. 218); in Brachyceras (Fig. 38, b-e) they are spheroid, with greatly flattened poles, or, in N. capensis, lentic- ular, and the edges may be folded under from three sides. Fischer (1890) speaks of them as folded in (eingefaltet) in N. dentata, lotus, rubra, and alba, the infolded half being separated from the other by an equatorial line. Along this line the exine separates in germination. The exine on the infolded half is thinner and smoother than on the other half, and has parallel to the equatorial line a ring of similar thickness and sculpture to the outer half (Fischer, 1. c). In water the grains become spherical in a A Fio. 88.— Pollen: a, young grain of S. flavo-virtns ; b— e, N. caerulea; b, just matured ; c, nearly dry ; d, e, as shed from the anther, seen from opposite sides. few minutes, and along the equatorial line in N. capensis, caerulea, zanzi- bariensis, and hybrids the two halves of the exine separate slightly ; they seem to be too small to cover the whole swollen grain. In pollen of N. odorata mounted in glycerine jelly, some grains have the exine caps still touching at one edge, but wide apart at the other, while some are equidis- tant all round, and some are unbroken. Fine rods, points, and granules cover the exine in Eu-castalia, and one-half of it in N. tetragona and flava (Caspary, 1865 ; 1888). It is smooth in Brachyceras, and, according to Caspary (11. a), in Anecphya, Lotos, and Hydrocallis. Fischer (1890), however, found it granulate (kornig-walzig) in N. rubra and dentata, and smooth in N. lotus, and I have found it smooth in N. mexicana. In germination Caspary frequently refers to an operculum, and in one place (1888) to the "smaller half" of the exine ; but the two parts which I have noticed in Brachyceras and N. odorata are practically equal, and Fischer speaks simply of two pieces (zwei Stilcke) in N. alba, and elsewhere of the two " halves " of the exine. The term " operculum " is therefore misleading. The pistil is composed of 5 to 35 carpels, the number being extremely variable for any one species. They are placed radially round a central 86 The Waterlilies. prolongation of the floral axis and are immersed in the excavated receptacle with about half of the ventral suture exposed above and the other half fused with the central axis. The back of each carpel is more or less prolonged above the receptacle (except in N. gigantea) into a stylar process. We shall speak of this as a carpellary style (Fig. 39, c). The exposed portions of the ventral sutures are connivent or united into a broad, round disc, cup, or funnel, the whole surface of which is stigmatic. At the center of this the floral axis projects as a process (Fig. 39, a) which is ovoid or nearly spherical in Eu-castalia, narrowly ovate in N. tetragona, linear and rounded in Xanthantha, linear and acute in Hydrocallis, broadly conical and obtuse in Lotos and Apocarpiae. It is of a dull white color. Fig. 39. Fig. 40. Fig. 38.— Diagram of ovary (Brachyceras) in longitudinal section : a, axile process ; c, carpellary style ; n, air-oanals ; r, receptacle : a, stigma tic papllhe ; t, torus, with bases ot floral leaves. Fig. 40.— Diagram of ovary in transverse section : a, sy ncarpous ; b, apocarpous. The stigmatic surface (Fig. 39, s), is usually yellow and densely papillose. The papillae completely cover the upper surface of the small rounded dorsal projections of the carpels of N. gigantea, and in other species they extend for a short distance up on the inner face of the carpellary styles. The styles themselves are, as indicated, wanting in N. gigantea. In Brachy- ceras they are short, stiff, orange-colored bodies, grading from mere points in N. elegans to erect processes in N . flavo-virens or stout, incurved bodies 0.6 cm. long in N. zanzibariensis. In Xanthantha and Chamaenymphaea they are more or less ligulate, and stand out horizontally in the basal part. ■ In Eu-castalia and Lotos they are 0.6 to 1.3 cm. long, ligulate and flexible, and in Hydrocallis of about the same length, slender and terete below, with a large clavate end. In the structure of the ovary Caspary (1865) was first to recognize two fundamentally different types. In most of the old group Cyanea of De Candolle and Planchon he found the walls between the cells of the ovary Structure. 87 are composed of two distinct lamellae, which easily fall apart along their smooth surfaces of contact. In all other species there is no such division in the walls, but the tissues in each partition are perfectly continuous. The first group was called on this account Lytopleura [Xuzo<; free, nhupd, side), with the definition "carpella toto latere libera tantum externe ad dorsum et interne ad processum axilem coniunctum enata." The remain- ing group was called Symphytopleura (oufupuToz, connate), and defined "car- pella toto latere coniunctum enata." It seems desirable, however, to get rid of names as far as possible, so we have adopted the simple descriptive terms "apocarpous" and "syncarpous" for the two groups. It is valuable to keep in mind that apocarpy occurs in a large group of Nymphaeas, as the Cabombeae and Nelumboneae are all apocarpous. The dehiscence of the fruit of Nuphar, N. flava, and the fossil Ancectomeria into separate carpels suggests a recent apocarpous ancestry.' The absence of any definite placenta in Nymphaea has long been a subject of remark. The ovules are attached in large numbers all over the walls of the ovary cells, both lateral and dorsal. They are anatropous, and, at the time of flowering, are suspended on the funiculi with the micro- pyle upward. The cells of the ovary are oval in side view, wedge-shaped in transverse section, and are filled with a stiff, colorless mucilage. Externally the ovary (or receptacle) is densely covered with fibrous and mucilaginous hairs. Those of the first type are few and early decidu- ous in A^. zanzibariensis X , but plentiful in the other species examined ; they are plainly visible in the Lotos group. The longest I noted in N. odorata was 0.34 mm., and in N. 0. minor 0.66 mm., composed of two shaft- cells ; the longest in N. tuberosa was 0.3 mm., with three shaft-cells ; in N. zanzibariensis X 0.25 mm., five-sixths of its apical part being a single cell. The epidermis is about 0.15 mm. thick, composed mostly of cubical cells. Within this is a layer of dense tissue about 0.34 mm. thick, of small polyhedral cells with very minute intercellular spaces, followed by a loose, spongy tissue with very large air-spaces. In N. zanzibariensis X the air- space in the latter zone barely exceeds the cellular tissue in volume and there are no idioblasts. In N. tuberosa and odorata the air-space greatly preponderates, and in the latter stellate cells abound. The spongy layer is much narrower and idioblasts very scarce in N. odorata minor. The epidermis bounding the ovary cells is backed by a single continuous layer of soft parenchyma cells. In Syncarpiae this layer and the lacunar per- enchyma are alone present between the epidermal layers in the partitions of the ovary. In Apocarpiae the condition is very different. The line of 88 The Waterlilies. division between the carpels lies in the middle of each partition and extends from a point in the wall of the ovary a little farther in than the back of the ovary cells to a point in the central axis distinctly farther in than the inner edge of the cells. On each side of this line is a smooth face of square or columnar epidermal cells ; the two sides are in contact. This epidermis is backed by two layers of closely packed rectangular cells ; then comes a line of open spongy tissue, then a single layer of parenchyma supporting the epidermis which lines the cells of the ovary. This epidermis is composed of short columnar cells of highly glandular nature ; each one bulges out into the cavity with a round, turgid end. They doubtless secrete the mucilage which fills the ovary. The axis of the ovary is composed of spherical cells with large intercellular spaces in N. zanzibariensis X ; in Eu-castalia (A^ alba candidissima, odorata, (uberosa) the spaces are so large that the cells are drawn out into several broad arms. Many stellate idioblasts were present in N. odorata. In the upper parts the tissues of the axis become solid. The process in which it termi- nates is firm and composed of thin-walled parenchyma ; the cells are often slightly elongated and arranged in rows bending from the center out to the periphery. A regular epidermis of small cubical cells covers the process. The stigma is marked with a number of narrow radiating furrows, one between every two carpels and one along the ventral suture of each. Indeed, the ventral sutures are scarcely closed and are easily torn apart. The stigmatic papillae (Fig. 37, c) are crowded as close as they can stand and consist of rows of 2 to 8 short cylindrical cells. The terminal one is rounded and contains large, oily-looking bodies. In Hydrocallis the cells separate from one another and the stigma is covered with a moist, powdery mass of oily-looking cells, some of which remain attached to each other in strings. The carpellary styles have numerous mucilage hairs on the epidermis and are made up internally of spherical cells loosely laid together. On the clavate apex of the styles of N. amazonum, the epidermis has a slightly corrugated cuticle, and the cells contain a pink-purple pigment. The vascular system of the ovaiy presents certain regular features. From the complicated plexus in the base of the receptacle the immediate connections have not been worked out, but sooner or later there is separated a ring of bundles which ascend with numerous branchings and anastomoses in and just outside of the lacunar tissue of the ovary wall, and a ring which ascend in the central axis near the edges of the cells of the ovary. In N. odorata another bundle passes up the centre of the axis Structure. 89 to terminate in the axile process. From the bundles of the outer ring branches are given off to the petals and stamens on the one side and to the walls and partitions of the ovary and ultimately to the ovules on the other. Finally, a bundle extends into each carpellary style. In Syn- carpiae the bundles of the inner ring, like those of the outer, are irregularly placed with regard to the carpels, and they seem to unite among them- selves by oblique branches ; but in Apocarpiae there are one or two bundles in the inner edge of each carpel, and these are separated from neighboring ones by the divisions between the carpels. The inner bundles branch copiously into the partitions between the cells of the ovary in order to connect with the ovules. The bundles are so oriented as to bring xylem next to the ovary cells and phloem away from these, that is, normally with regard to the carpel itself. Consequently the bundles of the outer ring have xylem on the axial side, but the inner ring has phloem toward the axis. In the partitions the bundles nearest one cell have their xylem next to that cell ; since those on the other side of the partition are similarly arranged, the two sets have their phloem sides turned toward one another. The central axile bundle of N. odorata has one or two spiral tracheae with phloem all round, but especially thick on two opposite sides. Some peculiar vascular bundles occur in the ovary wall. The largest bundles of this type have an outer mass of phloem, then a band of xylem ; this is followed by a band of small-celled parenchyma, and this by another xylem band equal to the former. These are comparable with the smaller veins of the leaves where two bundles stand one above the other ; in the ovary the phloem of the upper bundle is suppressed and the intervening parenchyma reduced. The fruit of a waterlily is technically a berry. The outer wall of the ovary remains firm, but its inner tissues become soft and mealy, i. e., the cells easily separate from one another. The ovary-cells are filled with a greater or lesser number of hard spheroid or ovoid seeds embedded in mucilage, and each surrounded with an aril. Dehiscence is irregular ; the fruit simply bursts open by the growth of the arils and swelling of the mucilage. In N. flava, however, there is a decided tendency for the car- pels to separate, still containing the seeds. The ripening of the fruit takes place under water, during which time various changes occur in the floral leaves. In Castalia these completely rot away, leaving only small scars. In Lotos about half an inch of the bases of the outer stamens remains attached to the fruit. In N. amazovum the sepals persist and the other parts of the flower may be only partly decayed. In Brachyceras 90 The Waterlilies. the sepals and outer four petals remain and become stiff and turgid and develop a good deal of additional chlorophyll. The fruits of N. tetragona And/lava are small (2 to 2.5 cm. long), and the persistent carpellary styles form only an insignificant crown around the top. In Eu-castalia the stigma grows very little if any after anthesis, but is about half as wide as the diameter of the spheroid fruit. But in Brachyceras and Lotos, where the fruits become 8 to 12 cm. across, the stigma keeps pace with the growth of the rest of the pistil, so that the fruit has much the same shape as the ovary of the flower. The seeds of waterlilies offer marked characteristics which are often of great systematic value. They were made an object of special study by Weberbauer (1894), from whose valuable paper many of the following details, especially regarding N. alba, are taken. Some important species, however, e. g., N. tetragona and flava, did not come under his notice. The largest seeds are those of Xanthantha and Anecphya, then N. tetra- gona, followed by N. tuberosa, Candida, alba, and odorata ; those of Lotos and Brachyceras are considerably smaller than the preceding; the smallest are found in Hydrocallis. They vary in number inversely as the size. N. flava may have only seven perfect seeds to a fruit, or one to three per carpel ; N. tuberosa has sixty to seventy, while N. lotus, rubra, caerulea, capensis, zanzibariensis, and amazonum have thousands. They are nearly globular in N. flava, flavo-virens, elegans, and amazo?ium, and ellipsoid in the rest. The micropylar end is always slightly pointed, and in Eu-castalia and Chamaenymphaea the raphe forms a gentle swelling along one side. In Castalia the color of the seed is usually greenish black, but brown is the regular color in N. tuberosa. Dark brownish-olive tints are found in all others that I have seen, except N. amazonum, which is reddish brown, covered with silvery hairs. The micropyle and hilum lie near together at one end of the seed. The character of the surface, whether smooth or ridged or hairy, can best be discussed in describing the microscopic details of structure. The aril which is attached to the ripe seed is a bell-shaped outgrowth of the funiculus. It is whitish and translucent, mucilaginous, and holds between its two coats tiny air bubbles. It is much shorter than the seed in N. tuberosa, a little longer than the seed in N. odorata (Fig. 41, a), lotus, caerulea, capensis, zanzibariensis, and amazonum, but in these the seed is visible at the open end of the aril. In N. alba (Weberbauer) and glad- stoniana the seed is completely inclosed and hidden by the aril, while the aril of N. flava is a large, wrinkled bag, big enough, if stretched out, to Structure. 9i cover two seeds with ease. Its cells (Fig. 41, c-e) are tubular and stretched out in the direction of its growth, thin-walled, and somewhat pointed. In N.flava they are about three times longer and wider than in N. odorata and caerulea. The inner layer of cells adheres closely to the seed, and its cells are much smaller than those of the outer layer. The walls are supplied with fine pores (Weberbauer). The aril serves, by its mucilage and inclosed air, to float the seed for a short time ; it drops the seed in the course of a few hours (JV. tuberosd) or a day or so {N. caerulea), partly by being bursted and torn through absorption of water, partly by decay (Fig. 41, b). The seed proper is protected by a firm shell consisting of two parts, the outer a single layer of thick-walled, close-fitting cells, the inner a Fig. 41.— The aril. N. odorata.- a, diagrammatic section of aril and seed ; b, portion of aril after the seed is dropped ; c, cell of the aril where the outer wall bends In, In continuity with the inner wall of the aril ; d, cell of inner wall ; e, cell of outer wall. few layers of thick-walled, but loosely placed cells. The latter layer is of the same thickness as the former ; its cells are much flattened tangen- tially and contain large tangential air-spaces ; the walls are lignified, with the exception of a thin layer next to the lumen. The cells of the outer layer are also strongly compressed tangentially (except in N. flava). On account of the great induration, which is thickest on the outer wall, they have but a small lumen (Fig. 42). The walls are penetrated by pore- canals. They are composed of lignin, excepting a thin layer of cellulose next to the lumen and an outer thin cuticula which projects inward along the boundaries of the cells. In the cells there are (in N. alba) bluish-green granules which give color to the seed. In surface view the outer cells are arranged in longitudinal rows, and, except at the micropylar end, they have very sinuous walls. Over the raphe the cells are elongated in the direc- tion of the rows, but elsewhere at right angles to the rows. At the micro- pylar end of the seed the cells of the hard (outer) layer are nearly square and in rows, but interrupted over the hilum, which lies (in N. alba) about 92 The Waterlilies. three cells from the micropyle. At the hilum the thick-walled cells of the inner seed coat are rounder and more closely packed than elsewhere. Based on the examination of the seeds of about twenty species of Nymphaea, Weberbauer distinguished four types of testa, as follows: (i) Surface bare, smooth, and shining: N. alba, Candida, tuberosa, odorata. (2) Bare, with longitudinal ridges ; adjacent edges of rows of cells are alternately raised up or level ; the ridges may be interrupted at middle of seed {N. rubra (Fig. 42, a), pubescens). Fio. 42.— Sections of outer cell layer of testa: a, N. rubra; b, N.fiavo- virens (" N. gracilte " Weberb.) ; c, N. stellata. After Weberbauer. (3) The seed bears longitudinal rows of hairs in the position of the ridges of the second type ; the rows may be similarly interrupted (A7, blanda, amazonum, gigantea, caerulea, stellata (Fig. 42, c), capensis, ampla, madagascariensis); N. lotus, thermalis, and dentata are only slightly hairy and show transitions to the second type. (4) Both longitudinal and transverse cell walls have short papillar out- growths (N. Jlavo-irirens, Fig. 42, b). Differences occur in the degree of flattening of the outer cells and thickening of the walls, as also in the number of hairs and character of the surface under them. In N. tfiermalis the amount of thickening is less than in other Lotos species, and in N. gigantea the outer walls of the cells are scarcely thicker than the lateral and inner walls. In N. rubra. the cells are but little compressed. N. amazonum has a dull, roughened surface beneath the hairs, but in "A^. blanda " (probably N. rudgeana) this is very smooth and shiny and the hairs are bent at the tip. N. ampla Structure. 93 and madagascariensis show an inclination to the formation of ridges as in the second type, with a reduction of the hairs. The outer coat of N. flava is peculiar in several points. It is cov- ered with long, slender, appressed hairs just visible to the naked eye, springing from all sides of the cells. It thus combines the characters of the third and fourth types of Weberbauer. The cells are extremely wavy superficially, and so irregular in shape as to render the longitudinal rows obscure. They differ from those of all other species investigated in being nearly twice as deep as wide. A great amount of induration of the outer walls leaves the lumen as a small pear-shaped space at the inner end of the cell. This is a striking approach toward the narrow columnar structure of the hard layer of Victoria and other genera of Nymphaeaceae. I find that N. tetragona belongs plainly to the first type of j seed coat according to Weberbauer, and N. elegans to the fourth type. The inner seed coat is pressed to a thin layer, except at the tip of the nucellus. The walls are FlB-« ru'8'rib"tion °f ~ r vascular bundles from the thickened on the inner side, but as it approaches the ebaiaza, in seed coat of n. . . il • « ,, alba. After Chlfflot. micropyle the thickening extends up the side walls, until at the micropyle only the outer walls are unthickened. With this thickening there is an increase in the height and sinuosity of the side walls. The micropyle is closed by slightly thickened rounded cells and a layer of thinner cells overlying the inner coat (cf. Fig. 44). A vascular bundle passes along the raphe, and spreads in the seed coat irregularly (Fig. 43), as figured by Chifflot (1902). Nine-tenths of the space of the seed is occupied by perisperm (nucellar tissue). This consists of large, thin-walled polyhedral cells (Fig. 44, p), densely packed with starch. The outermost cells of the nucellus are shriveled and compressed together into a thin membrane, which is thickest between the embryo and the seed coat. The outer food-storing cells are small and somewhat flattened, but in most of the nucellar tissue they are elongated and disposed somewhat in lines {Leitungsbahnen of Weberbauer) running obliquely upward toward an air-space beneath the embryo. The central axis of the seed is filled with looser, thin, elongated starch-laden cells. The nuclei of the cells, where distinct, are pressed over against one wall. The starch is in simple or compound granules, and these, in N. gladstoniana and odorata, are aggregated into spherical masses. The grains are very small, and the spherical masses are not 94 The Waterlilies. larger than single grains of some plants. Each spherule is surrounded by a pellicle of protoplasm, and a network of the same substance ramifies among the constituent granules. This type of aggregation exists in a less perfect condition in N. jiava, but was not seen elsewhere. In N. odorata stained sections of the perisperm show a network of thicker plasma strands between the starch spherules near the middle of each cell. In the micropylar end of the seed a small lenticular or globose space (Fig. 44) is occupied by the endosperm and embryo. The former is a single layer of cubical cells (e) inclosing the latter. The endosperm cells are thin-walled, but Weberbauer was able to find the walls covered with pits so that the thicker portions formed a fine network. They have moderate nuclei, and contain a rich supply of proteids and oil, but no starch. Over the radicle the endosperm is very thin and poor in contents. The embryo has two equal, fleshy, hemispherical cotyledons, ex- cavated on the inner side to receive the plumule. Between their symmetrical bases is the tiny caulicle, and a slight prominence at the end of this is the radicle. Food- storing tissue of isodiametric cells and elongated conducting cells are distinguishable ; an axial strand of the latter is found in the caulicle, with a branch into each cotyledon. No starch is found in the embryo, but it is richly supplied in all its parts with oil and protein. In a crushed embryo of N. flava the oil globules run together in considerable drops which stain easily with alcannin. The plumule consists of two rounded papillae, the leaf rudiments, placed transversely to the cotyledons. One of these is about four times as large in volume as the other. Fig. 44.— Longitudinal section of seed : c, cotyledon ; e, endosperm ; g, tegmcn ; m, micropyle ; p, perisperm ; I, testa. CHAPTER III. DEVELOPMENT. ORGANOGENY AND EMBRYOLOGY. If it were not for the long hairs which clothe the apex of the stem of Nymphaea, there could hardly be a more favorable object for the study of the development of the organs of a mature plant. We cut off the broad tip of a rhizome of N. alba, odorata, or tuberosa i cm. back and pin it down to the wax bottom of a dissecting pan. Then with the aid of a dissecting microscope the larger leaves and their stipule's are cut away. The hairs may be pulled out with fine forceps. This is the most difficult part of the task. Finally one discovers that the stem apex is slightly basin- shaped. At the center is a very minute elevation, the growing point, and around it the rudiments of leaves and flowers are arranged in spiral order. The figures of leaf development given by Payer (1857) for N. alba are equally applicable to N. odorata, and would need but little alteration for N. rubra. I have examined the last two by dissection, and have studied N. caerulea, odorata, and flava briefly in serial microtome section. The youngest rudiments at the stem apex are mere rounded papillae, without indication of their future destiny. They gradually elongate and on reach- ing a length of 0.08 cm. the foliar or floral character becomes evident. The leaf rudiment (cf. PI. II, 1-7) takes on a more and more conical form, while the floral rudiment remains broad and rounded and develops the anterior sepal. A leaf of N. rubra var., 0.16 cm. long, is about three-quar- ters as wide at base, and tapers to a rounded apex ; on the inner face a broad triangular depression is sharply marked off by two lateral ridges which are continuous over the apex and curve inward a little below, ending about one-fifth of the length of the rudiment above its base. Below this is the petiole ; the upper part is the lamina. The rudiment becomes longer and more acute, and the lateral ridges become slightly rolled inward until in a leaf 0.3 cm. long the former triangular depression has become a narrow groove, rounded above. At the base of the petiole there is now on either side a narrow rounded auricle, three times as long as 96 The Wateri.iues. broad, placed so that one end touches the rhizome and the other is a very short distance above. These are the stipules. In a leaf 0.64 cm. long the lamina is still more than twice the length of the petiole ; the inrolled sides now meet in the median line and the veins are prominent. The lobes of the leaf extend down along the front of the petiole and are slightly divari- cate. The petiole is about as broad as long and somewhat contracted at each end. The stipules are slightly larger and more nearly vertical. The coating of hairs on petiole, stipules, and lamina is well developed, and the swollen collar at the base of the leaf is becoming evident. Subsequent development is, to outward appearances, chiefly a matter of size, except with regard to the stipules. On a leaf of N. rubra whose petiole is already more than 5 cm. long these form lunate auricles, with nearly verti- cal line of attachment to the petiole, and a lobe extending upward a short distance above the line of insertion. Such a stipule is about one-third as wide as the diameter of the petiole and twice as long as wide. After this it decreases in proportionate width, while greatly elongating to reach the mature form. In N. odorata (PI. II) and N. alba the young leaf rudiment, less than 0.16 cm. in length, is more rounded and less acute than in N. rubra and the stipules are quite different. About the time the flattening of the inner side of the rudiment to form the lamina occurs, two elongated ridges appear, one on either side of the base of the leaf, extending along toward each other just in front of the rudimentary petiole. The adjacent ends soon meet in the median line and the beginning of the stipular plate of the Castalia group is evident. Being shorter in the median line than at the sides, its outline is obcordate. As the whole ridge grows higher, it overtops its leaf at the length of about 0.3 cm., and on either side of the middle line an S-shaped fold occurs, which throws the median part of the stipules away from the leaf, or, to state it otherwise, forms a groove into which the leaf fits. A great thickening along these folds makes two strong keels on the inner side of the stipule. Later this broad plate elongates to the typical adult form. A leaf rudiment oi N. odorata 0.3 cm. long had downward pointing hairs covering the back of the lamina and the petiole, but on the next younger leaf these were present only on the lamina. The formation of the principal air-canals takes place at an extremely early period. The first beginning of separation of the cells was seen in N. caerulea in a rudiment 0.5 mm. long. The canals were seen with the naked eye in a leaf rudiment of N. flava about 1 mm. long ; the bounding walls were composed of columnar cells. At 1.6 mm. long all of the canals are indicated. In the smaller leaf the fusion of the three main bundles CARNEGIE INSTITUTION OF WASHINGTON. WATERLILIES, PLATE II. A. KORONSKI DEL. NYMPHAEA ODORATA. 1-7, development of leaf and stipules; 1, 2, front and back views of the same rudiment; 3, older; 4, 5, front and side views of one rudiment; 6, 7, front and back views; 8, rhizome with roots and leaves trimmed off, natural size; 9, transverse section of rhizome, natural size. Development. 97 from the petiole to a single trace and the separation of this again into the central and two lateral traces was seen. The bundles were chiefly composed of procambial cells, though some spirals were present in the three branches within the inner fork. The lamina of the older leaf mentioned had elongated cells along the lines of the veins, but the parenchyma was almost totally undifferentiated and was without intercellular spaces. A leaf of-iV. tuberosa richardsoni whose lamina and petiole are each 1.3 cm. long has the air-canals well developed along the midrib. Between the veins the mesophyll consists of five layers of cubical cells arranged in ver- tical lines from the upper to the lower epidermis. The layer next to each epidermis is continuous, but between the vertical rows in the three middle layers long intercellular spaces are present. Idioblasts are already pres- ent near the midrib. A leaf of N. odorata with lamina 2.5 cm. long and petiole 3.2 cm. has the air-chambers of the mesophyll well marked off by plates of cells ; the palisade consists of a single layer of dense cubical cells, underlaid by two or three layers of flat cells elongated parallel to the surface of the leaf ; the mucilage hairs are large and functional. When the lamina is 8 cm. long the palisade still consists of a single cell-layer ; the big idioblasts have not yet appeared. Stoma-mother-cells are present in full number and nearly mature size ; some, indeed, have just divided to form the guard cells. They are round, granular, and highly refractive. The maturation of the stomata occurs only a day or two before the leaf finally expands. Caspary (1854; 1858), in announcing first the apical growth of leaves, showed that this apical growth continues longer in Nymphaea than in other plants, and that the margin of the leaf in all Nymphaeaceae is the latest part to complete its development. The tissues of the stem begin to differentiate within 0.25 mm. of the point of apical growth. No plerome is at any time distinguishable. The lacunar cortex acquires intercellular spaces while all of the tissues are still meristematic, and the early establishment of desmogen strands, adven- titious roots, and external hairs has already been referred to. Concerning the development of the flower, we are able to corrob- orate the work of Payer (1857) and Goebel (1886) on N. alba by obser- vations on N. caerulea and lotus, and to add some details. The discussion of important morphological problems will be deferred whenever possible until the observed facts are set forth. The earliest rudiment of a flower that I have seen was a smooth, thimble-shaped upgrowth at the stem apex in N. caerulea. On such a papilla, according to Payer, there appears in N. alba the rudiment of the anterior sepal ; then the two lateral sepals 98 The Waterlilies. arise on the sides, and later the posterior sepal. Hence comes their peculiar aestivation. Four papillae next appear alternate with the sepals. They arise simultaneously, grow rapidly, and are always larger than the inner floral organs, which originate later. They are at the base of a hemi- spherical receptacle whose surface, naked hitherto, becomes covered little by little from base to summit with a number of new papillae. These are arranged in ten vertical rows, one abutting on each petal, one on each lateral sepal, and two on the posterior and anterior sepals. The lowest ones soon flatten out and become petals ; the uppermost become typical stamens ; between these two sets one finds all gradations from stamen to petal. The stamen rudiments, however, do not cover the whole recep- tacle. A bare space remains at the summit, around the edges of which a row of papillae appears. These (the carpels) become horse-shoe shaped, and their inner edges join, making a sinuous line. The sinuosities extend somewhat toward the center, and a thick process of the receptacle grows up to fill the central space, closing the ovary-cells. Now the sepals and the four outer petals remain stationary, as it were, while the rest of the receptacle elongates, carrying with it the carpels, stamens, and inner petals. Thus the ovary comes to be inferior with regard to the stamens and inner petals while the outer petals and the sepals have an inferior insertion. There is no fusion of the floral leaves outside the carpels with the ovary or receptacle, as Koehne {fide Goebel) believed. According to this account from Payer, the ovary would bear the carpels at its summit, and its lower parts would be of axial origin. Goebel (1. c.) noted this difficulty and showed that before the carpels are mapped out there is a hollowing of the receptacle as is common in inferior ovaries. From the edge of the axial excavation, but reaching to the bottom of it, arise the carpel-rudiments. They are never free from the receptacle on the outer side. The whole wall of the ovary-cells, however, is of carpellary origin, and no part is formed from the vegetative apex of the flower stalk. Our own observations on N. caerulea and lotus support Goebel's view. In N. lotus the time relations of the origin of the outer floral organs differ slightly from that given by Payer. In the apex of a large flowering plant we found three rudimentary flowers of successive ages and sizes, in all of which only the anterior and lateral sepal-rudiments were present (Fig. 45, a, b). This shows that these organs originate much earlier than those farther in. The anterior sepal was much larger than the others, and was inserted considerably farther down. In such a rudiment one could hardly speak of a peduncle, for this sepal seemed almost sessile on the Development. 99 parent stem. The next older flower noted (Fig. 45, c) had a broad papilla representing the posterior sepal, a rounded one for the right antero-lateral petal and a partly formed papilla for the left antero-lateral petal. The right petal was about half as large as the posterior sepal, and this was about half as large as a lateral sepal, and this again half as large as the anterior sepal. The next flower had the two anterior petals about equally developed. The remaining organs (except the carpels) originate one by one ; two postero-lateral petals come next, then four petals opposite the sepals ; then four opposite the first petals. The exact order of these was not determined (Fig. 45, d). In early stages the bud exceeds the peduncle in length, but after the total length of the two reaches 2.5 cm. the peduncle begins to exceed the bud. Previous to this stage the anterior and lateral Fio. 45.— Development of flower, N. lotut var.; a, end view; b, side view of very early stage ; c, with rudi ments o f sepals and two anterior petals ; d, sepals, two whorls of petals, and one member of the next whorl laid down. sepals alone cover the bud ; later they are pushed apart and the posterior sepal is exposed. The carpels are rather late to develop. In a bud 1 cm. long they resemble little folded leaves, fused half way up the back with the hollowed receptacle, but open along the ventral suture, and, even in N. lotus, wholly free from each other. The above facts throw light on several questions, and first upon the kind of inflorescence found in Nymphaea. It will be remembered that the flowers stand in place of leaves on the stem, but without any subtending bract. Payer says the case is very simple : Each flower arises in the axil of a leaf, but on account of crowding in the bud, the flower comes to be at some distance from its leaf, and it requires some experience (" il faut quelque habitude ") to find the leaf to which a given flower belongs. Schumann (1894) maintains the same view and cites the case of Victoria, in which there are parallel spirals of leaves and flowers on the stem, and at the apex each leaf has a flower to its left as they originate. Raciborski (1894) opposes this with the theory that the bract is entirely suppressed in Nymphaea and Victoria. This better accords with the exact location of the flowers in the leaf spirals ; and in Nuphar there is at times, if not always, a rudimentary bract subtending the young flowers. Eichler (1875), however, ioo The Waterlilies. on the authority of Caspary and Braun (in personal letters), reluctantly accepts the belief that the flowers are axillary, but the bract and bracte- oles are carried up on the summit of the peduncle as anterior and lateral sepals. The argument for this has been based largely upon the uncertain ground of teratology. Planchon (1850, e) figured a flower of N. caerulea " whose peduncle showed accidentally [at the lower end] three spatulate bracts, with petioles wavy here and there on the margins as the sepals are at base." * Caspary {fide Eichler) found in the same species a flower with the anterior sepal at the base of the peduncle as a ribbon-shaped bract, the lateral sepals also pushed down, and the posterior sepal in its proper place. This was followed by four sepals in a whorl, placed diago- nally like the second whorl of a normal flower. Caspary found similar cases in JV. alba, gigantea, rubra, and hybrids of M. caerulea and capensis. My own observations on the development of the flower of N. lotus fully substantiate Caspary's view, and, I may say, completely turned me from an opponent to an adherent of his theory. The low insertion of the rudiment of the outer sepal and its very early appearance, followed at once by the two lateral sepals, are convincing. This also explains the peculiar relations of the sepals in the bud ; on any other theory, except a blank displacement, the anterior sepal should be, like the posterior, covered by the lateral sepals. It remains, however, to explain the presence of a single posterior sepal and the oblique position of the next four floral leaves. Caspary and Braun looked upon the single sepal as an " Erganzungsblatt. " But if the floral diagram of Nuphar as given by Baillon (j : 82) can be depended on, in which the outermost of the five sepals is posterior and the petals " are inserted along a spiral," we have the most perfect starting point for the development of the flower of Nymphaea. The outermost sepal of Nuphar homologizes with the posterior sepal of Nymphaea, and a very slight shifting of the remaining four sepals away from the median anterior line gives them the oblique position of the four outer petals of Nymphaea. The strongly sepaloid character of these petals, especially in Brachyceras, indicates such an origin. As to the inner petals, from the acycly of Nuphar, Nymphaea shows a continually increasing tendency to cyclic symmetry. In N. lotus the rudiments of all of the floral leaves (excepting the three outer sepals) arise in series, one after another, but Payer's account ol N. alba places the four outer petals in a whorl, and the inner * dont le pedoncule offre accidentellement trois bractees spathulees, a petiole ondule c_a-et-la sur les bords, comme les sepales le sont a leur base. Development. ioi petals and the stamens in alternating whorls of five ; probably this plant should be worked over anew. In the mature flower the cyclic arrange- ment is least evident in Castalia and but little more so in Lotos ; but, as stated, in old flowers of Brachyceras, and open flowers of Hydrocallis it seems to prevail in all of the parts (cf. Fig. 36). Probably cyclic arrange- ment is always to be considered as derived from acyclic, and Nymphaea seems to be in the transition stage. For such flowers the term hemicyclic or spirocyclic (Engler, 1898) is best. It goes beyond the truth to place Nymphaeaceae among acyclic orders as Goebel did (1887, p. 412). While still maintaining that the terms sepal and petal as commonly used for waterlilies are physiologically correct, the morphological value of the petals must be considered. We have shown that the anterior sepal represents the bract, the lateral sepals the bracteoles, and the inner sepal with the four outer petals the primitive sepals. For the last homology I am indebted to a suggestion from my instructor, Dr. John W. Harsh- berger. In development no distinction can be made between the rudi- ments of the remaining petals and the stamens, while in N. alba the four outer petals do have a distinct and simultaneous origin. Payer, therefore, pointed out, and rightly so, that the inner petals are really modified stamens, and the waterlilies are habitually double flowers in the same sense as roses, almonds, etc., are doubled in cultivation. This at once explains and is supported by the extremely variable number of petals and the occurrence of transitional forms between petal and stamen. It seems probable, however, that in last analysis all petals are modified stamens, and in Lotos and Anecphya, where a naked space occurs between the insertion of stamens and petals, and intermediate forms rarely if ever occur, the distinction between the two classes of organs is being rapidly drawn, and the homology is more distant than in Castalia. The fixation of definitive petal and stamen characters upon the floral rudiments is taking place before our eyes in the genus Nymphaea. It is striking, but on second thought only natural, to find this kind of specialization going on alongside of the change from spiral to cyclic symmetry. The two phe- nomena are but manifestations of the same law of adaptation and differ- entiation in form and function. After the stamen rudiment is established it grows considerably before the archesporium appears. It has been claimed that the anther cells are at first extrorse and are pushed around to the inner side by uneven growth of the connective. In N. lotus, at least, this is not the case. Two pairs of archesporial lines are differentiated in the substance of the anther, 102 The Waterlilies. each consisting of a single straight row of cells. They are removed four or five cells from the inner epidermis and seven to twelve cells from the outer epidermis, and between each pair of lines eight cells intervene. The cells immediately surrounding the archesporia take on a more or less concentric arrangement, and the development goes on in the usual way for flowering plants. In N. caerulea the outer stamens develop slightly in advance of the inner ones ; where the former are 1.3 cm. long and the spore-mother-cells are already separate from one another and becoming rounded off, the mother-cells of the inner stamens form a continuous tissue. The ovules originate after the ovary has attained a considerable size, say 0.5 cm. in diameter in N. caerulea. At first a tiny rounded column Fig. 4«.— Ovule of N. odo- Fio. 47.— Embryo sac of N. odorata; longitudinal section of ovule at time of rata, from a half-grown bud, flowering. From a photomicrograph, showing archesporium. From a photomicrograph. projects into the enlarging ovary cell. This swells at the end into a knob, an integument of two cell-layers grows up around the knob, and then a second integument outside the first. When the inner integument has reached the height of the nucellus, the outer one is about half as high (cf. Fig. 46). Meanwhile the knob has turned over at right angles to the supporting column in its change toward anatropy. A large archesporial cell is now visible at the apex of the nucellus, covered by a layer of cells. The archesporial cell presents five times as large a surface in longitudinal sec- tion as the surrounding cells ; its nucleus is as large as an entire cell of the nucellus, is rather poor in contents, and has a single dense nucleolus. Its further development has been described for N. odorata by Cook (1902). According to his account, a tapetal cell is cut off at the outer end of the archesporial cell (cf. Fig. 46), and from the former several tapetal cells arise by a variable series of divisions. Three potential megaspores are next cut off from the mother-cell. These rapidly degenerate, leaving a Development. 103 single large megaspore. This cell elongates to form the embryo-sac. Returning to the stage in which only the archesporial cell is found, we note an elongation of cells in the center of the funiculus which foreshadows the vascular bundle. This was seen in an ovary of N. odorata 0.6 cm. in diameter ; the ovules were just visible to the naked eye. In an anther from the same flower the microspores are formed, but have not received the chj -acteristic markings on the exine ; the tapetum appears as a layer of disintegrated matter around the walls of the anther cell ; the thicken- ings are not yet laid down in the anther wall. After this the ovule rapidly grows to its mature form. The outer integument becomes three cells thick and loses its protoplasm, while the inner integument consists of two layers, of which the inner alone remains richly protoplasmic. The nucel- lus consists of many cells arranged in radial rows ; the outer ones are tangentially flattened, the central ones rounded. The raphe is prominent and is traversed by a slender vascular bundle with a spiral trachaea on the side next to the nucellus. The funiculus is covered at base with a glandu- lar epidermis like that of the ovary, but the glandular cells end at a sharp line near the seed, beyond which the epidermis is plain and smooth. Between the epidermis and the vascular bundle the round parenchyma cells enclose large intercellular spaces. The embryo-sac (Fig. 47) is prominent as a clear area below the micropyle, and is shaped like a flask with the neck extending down into the axis of the seed. The ratio of the diameter of the body and neck of the flask is as 5 to 1 in N. odorata and zanzibariensis, but in N. lotus the whole ovule is longer and narrower, and the embryo-sac is lanceolate in longitudinal section, with rounded ends. A single layer of columnar nucellar cells covers its outer end. A day or two before the flower opens the egg apparatus is complete. The embryo-sac is slightly narrowed at the micropylar end, where it is covered, as before, with one layer of columnar cells. In this upper part of the sac three similar cells are crowded, two above and one below. The rest of the sac is lined already by a few cells around the sides, visible only in my preparations as compressed nuclei. Hofmeister (1858) doubt- less referred to the same thing in Nuphar, " whose embryo-sac," he says, "even before fecundation shows clearly a cellular wall" (Zellstoffhaut). A pair of nuclei indistinctly seen in the extreme end of the narrow inner part of the sac in N. odorata doubtless represent the antipodal cells; a single cell with very large nucleolus occupies the same position in N. zanzibariensis. The outer integument now consists of two (or near the raphe three) layers 104 The Waterlilies. of slightly flattened cells, poor in contents. The inner integument has both of its layers very thin, and the outer one nearly empty. Dahmen (1892) critically examined the funiculus of the ovules of Nymphaea. Since the so-called aril originates as a swelling on the funiculus before fecundation occurs, he considers that it is not an aril in the usual sense, restricting the use of the term to outgrowths which occur when the seed is ripening. The distinction, however, is extremely unimportant, especially as the aril of Nymphaea does not reach maturity until the seed is ripe, and most of its growth occurs after the seed has attained its full size. In N. lotus Dahmen noted two swollen zones on the funiculus near the ovule, but only one in N. rosea [quid?) and zanzibariensis. The epidermis of the funiculus has slightly thickened walls ; within it is a thin-walled parenchyma and a vascular bundle with phloem and a little xylem. 'The bundle has no connection with the aril. Small round starch grains are found temporarily in the parenchyma, epidermis, and phloem. Intercellular spaces occur only in the arillar swellings according to Dahmen, but I find them plentifully elsewhere in N. odorata. The tissues are all filled with mucilage, which is only precip- itated by strong alcohol. The protoplasm is dull yellowish. Asparagin occurs in the funiculi, but no sugar was found. Unfortunately we cannot accurately trace all of these tissues to their mature state, but the changes are not important in any part except the embryo-sac. The early stages were worked out by Hofmeister (1858) in Nuphar luteum and advena, which differ from Nymphaea in that the narrow inner part of the embryo-sac is much longer, and extends after fecun- dation to the chalaza ; an axial tube remains open even in the mature seed (Weberbauer, 1894). In Nuphar there are three cells at the top of the embryo-sac, two upper synergidae and an ovum below. About the time of fecundation a transverse wall divides the narrow part of the embryo sac from the upper wide part, and the latter is soon filled by two or three endosperm cells. The fertilized ovum now divides transversely, and a suspensor of one to five cells is formed above the embryo. The obser- vations of Mr. Cook (1902) on N. odorata seem to be corroborated up to the first cell-wall laid down in the embryo-sac. His views on the develop- ment of the embryo, however, which he considers to be monocotyledonous and without suspensor, are totally at variance with my own preparations. I cannot imagine that his interpretations are correct. This portion of Mr. Cook's paper is based chiefly on Nuphar advena, in which Caspary's drawings (in Hb. Berlin) seem also to indicate the absence of a suspensor. Development. 105 The earliest fertilized ovule of Nymphaea that I have seen is of N. lotus, and has already reached nearly its mature size. The integuments are considerably hardened, and the perisperm is loaded with starch. A broad space at the micropylar end of the seed is occupied by the endo- sperm, the cells of which are very thin and watery (Fig. 48, a). A distinct layer of perisperm still bounds the endosperm on the side next to the seed c-^at. From a point near the micropyle a row of three cells projects into the endosperm, representing the embryo. An older seed of the same species shows the endosperm occupying an approximately spherical space, somewhat pointed above and flattened below ; its cells are large and Fio. 48.— Embryology : (a), (b) successive stages, N. lotus; (c) spherical embryo with suspensor, N. odorata ; (d) early differentiation of cotyledons and plumule, N. caerulea. distinctly nucleate, but very thin walled and poor in contents. The embryo (Fig. 48, b) consists of a line of three suspensor cells and a terminal knob of probably eight cells (four in longitudinal section). The uppermost suspensor cell is hexagonal, the others square (in section). After this the endosperm enlarges very little. As the embryo encroaches upon it, the outer cells become denser and more regular until they reach the mature condition, and the inner ones are crushed to a thin layer of remnants between the permanent outer layer and the cotyledons. The embryo proper (cf. Conard, 1902) becomes accurately spherical as it enlarges, until it consists of some hundreds of cells (Fig. 48, c). Then the lower end becomes flattened by reason of the outgrowth of cells on each side of the lower pole. In section the embryo now appears triangular and almost equilateral. The two lobes increase symmetrically and form the two equal cotyledons. At first these spread out nearly horizontally (Fig. 48, d), while in the middle between them a slight cone-like papilla io6 The Waterlilies. represents the plumule, and a similar prominence opposite (next to the suspensor) is the radicle. Subsequently the cotyledons spread round the walls of the endosperm and fill up all of the limited space available, while between them the two leaf-rudiments of the plumule are differentiated. GERMINATION AND EARLY GROWTH. Corresponding with the habits of life of the various species of Nym- phaea are certain requirements for the proper growth of seedlings. Seeds of the Eu-castalia group quickly lose the power of germination if dried ; in nature only those which are, by chance freshets, stranded hopelessly above the normal water level ever become dry. N. tetragona can withstand drought fairly well. Seeds from Kashmir communicated by Mr. Wm. Gollan, of Saharanpur, India, germinated quite well, but of several sendings received in the early spring of successive years from Japan, only one seed has ever sprouted. The former were sent to us directly after they were gathered, while the latter have doubtless been kept in a seed-house since the previous summer. Seeds of the Lotos and Brachyceras group can with- stand a great deal of drought ; they are accustomed to it ; but they are kept much more successfully, as Caspary (1877) recommended, packed in clay and air dried. Under these conditions a considerable amount of moisture is retained, even when the mass appears to be perfectly dry. N. zanzibariensis was introduced into Europe from Zanzibar by seeds packed in chalk ; and of a box of seeds of N. dentata, mingled with soil, re- ceived last summer (1901) from Jamaica, nearly every one sown germinated. In all of the Castalia group the seeds should be kept in bottles of water. It was thus, also, that Victoria regia was successfully carried to England after a long series of costly failures. For germination the seeds of waterlilies need to be submerged in 5 to 30 cm. of water. Those of the Castalia group prefer a temperature of 150 to 180 R., but the tropical species do best at 230 R. (Siber, 1883). A very irregular length of time is required. Of a quantity of N. elegans X zanzibariensis seeds sown as soon as collected, a large number came up in two weeks ; after more than a year some earth from this pot chanced to be placed in a warmer tank, and numerous new and vigorous seedlings appeared. In the Dreer Gardens at Riverton, N. J., I have seen flowering plants of N, caerulea which came up spontaneously out of doors where two years previously that species had been grown ; the seeds had lain dormant over all one summer. In this case Mr. Tricker informed me that the bottom of the pond had been dug over each spring. Probably the Development. 107 seeds were buried in the first digging and brought to the surface in the second. In midsummer N. zansibariensis seeds will germinate quickly after maturing, and the bags of seed must be gathered as soon as the fruit bursts ; but in the summer of 1900 numerous seedlings of this plant sprang up in the pond of the Botanic Garden (University of Pennsylvania) and blossomed before frost ; they must have come from seed ripened in the previous autumn. Mr. Waters (1886) gives an interesting account of the germination of N. odorata. Seeds gathered in the fall of 1 883 were sown at once in an aquarium kept in a living room with north light. In March, 1884, many were germinating. In June, 1885, the vessel was placed where it received direct skylight with sun for an hour at noon ; a half dozen new plants appeared in August. In the fall they were returned to the living room, and about Christmas another seed had germinated. From February to April, 1886, four others sprouted ; these had therefore lain dormant for about two and a half years, although they were constantly in conditions which would permit germination. The general features of a seedling Nymphaea were noted by Titt- mann in 1821. Treviranus, Caspary, Klebs, Goebel, and others have since remarked upon or figured such seedlings. The first visible evidence of life in the seed is the loosening of an operculum at the apex (Fig. 49, 8, 6). This is a conical body bearing the micropyle and hilum ; its edge is ragged by reason of the sinuous margins of the cells of the testa. The operculum is pushed aside by the protruding embryo and its wrappings, and lies on these at one side near the edge of the opening from which it was torn. The seed coat is also rent with short longitudinal fissures, dividing the lip of the aperture first formed into numerous acute teeth (Fig. 49, 8). As the embryo emerges it is covered by a white sheath produced by the swelling and straightening out of the overlying " rumpled up " portion of the inner seed coat ; this protects it from the sharp and ragged edges of the testa. The cotyledons remain permanently within the seed and show no changes in the inner parts except in the emptying of the cells. It is by the elongation of the bases of the cotyledons into a kind of petioles that the radicle and plumule are passively carried out of the seed. These coty- ledonary petioles also curve geotropically downward until the caulicle is in a vertical position, and then they cease to grow (Fig. 49, 3, 4). Through all this time the plumule and radicle have remained dormant, but now both start into activity, the former very rapidly. The hypocotyl does not elongate at all. The epicotyl and first leaf shoot upward as a slender subulate body (Fig. 49, 3, 4). The elongation of the epicotyl io8 The Waterlilies. Fig. 49.— Seedlings of a hybrid Nymphaca of the J,otos group. 1, tuber formation at summit of epicotyl well advanced. 2, the first broad leaf expanded, the primary root, with root hairs, mature. 3, the first (filiform) leaf nearly mature, and rhizoids springing from cotyledon. 4, a slightly older plant, from a seed which was buried more deeply in earth ; the epicotyl is much elongated. 6, first secondary root. 6, small ripening tuber ; epicotyl rotted off. 7, beginning of second secondary root and third broad leaf (4th leaf). 8, soon after germi- nation ; the tegmen (dotted) extends beyond the hard seed coat, bearing the detached operculum. 9, plant let of same aye as 3, with cotyledons broken off from their petioles, a, first (filiform) leaf ; o, operculum. Development. 109 depends upon the depth to which the seeds are buried ; it reaches up just to the surface of the soil. If they are uncovered at the bottom of the water, the epicotyl is scarcely visible (Fig. 49, 3) ; it is about as long as the width of the cotyledonary petioles. In other circumstances this inter- node may be 2.5 cm. long (2, 4, 5). The stimulus which effects this difference is light. Seeds of N. gladstoniana and flava grown in glass bottles of water in the dark sent out very long epicotyls reaching far above the layer of seeds in the bottom of the bottle. The first leaf produced by the seedling is a slender awl-shaped object, like a spear of grass (Tricker, 1897) ; it looks like a prolongation of the epicotyl (Fig. 49, (a); the boundary between the two is marked by a tiny rudiment of the next leaf. While these changes are going on in the plumule, a large tuft of root-hairs is formed on a broad prominence at the base of each cotyle- donary petiole. These hairs serve to anchor the plantlet to the substratum and doubtless to absorb some food during the inactivity of the radicle (Goebel, 1893). After the subulate leaf has nearly reached maturity, and the second leaf-rudiment is swelling, the radicle begins to elongate (Fig. 49, 4). Hitherto it has been a tiny papilla between the bases of the cotyle- dons, but it soon grows out into a primary root 5 or 8 centimeters long (5). The second leaf of the plantlet arises opposite the first and develops a petiole and an ovate to linear lamina (2, 5). At its base a single adven- titious root comes out and descends into the earth ; it only becomes evident after the leaf is nearly mature (5). Subsequent leaves are broader and broader in the lamina, becoming deltoid in Lotos and cordate to nearly orbicular in Brachyceras (cf. Fig. 50). Several of them remain on short petioles and are typically aquatic, but finally small floating leaves are sent up. Each one bears one or more adventitious roots at the base. The petioles of all the earlier leaves are winged and clasping at the base ; no stipules are developed until later. The plantlet being thus pretty well established, the primary root ceases to function and soon dies away. As the nutriment is baled out from the seed, the central portions of the cotyledons are first emptied, then their outer parts, and then the endosperm. Finally the solution of the perisperm begins at the end next to the embryo and works steadily back until the supply is exhausted. The nutriment furnished to the young plant, therefore, is at first highly proteinaceous, and at the last wholly carbohydrate. The cotyledons remain of the same size as before germina- tion, or very nearly so. When the seed has been emptied of its food, the epicotyl shrivels and dies as the primary root did before it (Fig. 49, 1, 6). no The Waterulies. The internodes above the epicotyl are not developed, but the stem gradually thickens and enlarges into a miniature tuber. In nearly all of the genus the plant is now essentially mature ; it needs only to enlarge. But in Eu-castalia a marked change takes place. The internodes after a time Fio. 60.— Seedling leaves : (a), N. amazonum ; a, first leaf ; b, second (first broad) leaf ; c, third leaf ; e, epicotyl ; p, primary root ; si, first secondary root ; (b). If. catrulea ; 1. 2, 8, 4, first and following broad leaves ; (e). first floating leaf, X. odorata giyantea ; (d), If. lotus var., 1, 2, », 4, first and following leaves, the last very young ; (e), N. zanzibariensis ; 1, 2, 8, first and following broad leaves ; If), If. tetroyoiia, 1, 2, 8, first and following broad submerged leaves ; 1 fi, first floating leaf. Twice natural size. regularly, but slightly, elongate, and the stem suddenly becomes plagio- tropic ; thus a tiny horizontal rhizome is produced. From this point onward maturity is simply a matter of size. In structure the characteristic features of the order are evident at a very tender age. While the primary root is still a mere papilla, the root- cap consists of a single layer of cells, and behind it are three to four close Development. i i i layers of meristem cells ; immediately back of these, large intercellular spaces abound. The epidermis of the primary and early adventitious roots differs from all mature rootage in having a copious supply of root- hairs (Fig. 49, 2, 5). It will be recalled that root-hairs are generally said to be wholly lacking in Nymphaea, Nuphar, Euryale, and Victoria. As we have observed them in N. dentata, lotus, caerulea, zanzibariensis, amazonum, elegans, odorata, capensis, and hybrids, they are doubtless universal in the genus. On one occasion also they were found on early roots of a plant springing from a tuber of N. caerulea. Victoria lacks them even on the primary root. The epidermal cells of the early roots (in N. dentata) are long and rectangular in surface view, and about as deep as wide ; where root-hairs occur, however, they spring from cubical cells. No shedding of epidermis occurs in these roots. The root-cap is long and thimble-shaped. Inside the epidermis there are in the primary root about six layers of thin-walled cortex cells and a small (diarch ?) vascular bundle. The adventitious roots are like this, but successively larger. In the earliest of these the diarch bundle is quite plain and is surrounded by a distinct endodermis ; the two xylem patches are continuous across the middle of the bundle. The appearance of the central vascular strand of the hypo- cotyl and lower part of the epicotyl is exactly the same as this, but it is described here as " two opposite bundles with their xylems confluent " (cf. Gwynne-Vaughan, 1897). A short process of procambial cells extends toward the base of each cotyledon, but no vascular tissue extends into them. They are composed of elongated thin-walled cells, and have a rather narrow insertion. The bundles of the stem are turned with their sides toward the cotyledons, i. e., a line connecting the bases of the coty- ledons would pass through the confluent xylem and between the phloem masses. Higher up in the epicotyl the two bundles become more distinct, and each, instead of being semicircular in section, is nearly round ; the xylems are still in contact, though torn asunder to leave an irregular air- canal in the middle. The first two leaves are placed opposite the phloems of the two bundles and at their point of insertion the vascular system spreads out toward them on each side. In close succession the bundles of the first adventitious root, the acicular leaf, the first broad leaf, and the second root and broad leaf join the widened portion of the system, with anastomoses across from bundle to bundle. Each of the above-named organs has a single bundle. No central strands exist above this, but the bundles spread out through the now fleshy stem on their way to the various organs. The communication from vessel to vessel which easily ii2 The Waterlilies. occurs in ordinary plants by reason of their proximity is made possible in these thick stems with widely separated bundles by the profusion of cross- branches. Gwynne-Vaughan (1897) noticed in the center of the young stem of N. zanzibariensis a cylinder of meristem tissue suggesting a stele, in which later desmogen strands were differentiated. He evidently con- sidered this as furnishing a kind of transition from the "monostelic" epicotyl to the " astelic " stem. The epidermis of all parts of the plant above the cotyledons, except- ing the upper surface of the leaves, is supplied with mucilage hairs. In N. Jlava these are so numerous on the acicular leaf as to make a plainly visible coating of jelly all over it, as is seen on petioles of Brasenia. The hair-bases remain on the full-grown parts as in mature plants. The inner tissues are very loose throughout. In the epicotyl the intercellular spaces are often as large as the rounded cortex cells. The acicular leaf is similarly constructed. The first broad leaf of N. dentata has Fio. 61.— First broad (submerged) leaf of N. lotus seed- r 11 j 1 1 1 . , lin?: a, stoma In upper epidermis; b, portion of tip of tOUr Well-aevelOpeQ air-CanalS in tUe leaf, showing distribution of stomata by dots. petiole ; the second has six canals, which are at first of almost equal size. The lamina of these leaves is extremely thin and fragile. Stomata are found scattered over the upper surface in N. lotus (Fig. 51), caerulea, elegans, amazonum, zanzibariensis, and capensis, but I failed to discover any in N. odorata or rubra ; probably the observation is at fault, for they are at best rather difficult to see. The best success was had with material fixed in chrom-acetic acid and mounted in balsam ; probably Wachter's (1897) chloral hydrate method would prove equally effective. The palisade-layer of these leaves is reduced to a single layer of hemispherical cells against the upper epidermis. Between these and the lower epidermis there are branching partitions two cells high, enclosing large air-spaces (Constantin, 1886). The venation is very simple. Between the bases of the second and third leaves a few long protective hairs appear as on the adult stem-apex. The young stem is at first devoid of lacunar tissue ; its fundamental parenchyma consists of rounded cells in which starch is rapidly deposited. From this time on the plant is in a position to go into a state of rest on any emergency. CHAPTER IV. PHYSIOLOGY. Special observations on the physiology of waterlilies are not numer- ous. That such plants have the usual "senses of direction " — geotropism and heliotropism — goes without saying. The primary and adventitious roots are all alike highly sensitive to gravitation at first, but the latter soon become more or less passive and run obliquely downward, while the secondary and tertiary roots ramify throughout the soil from its surface to a depth of 30 cm. or more. The peculiar type of root spoken of on a previous page as " contractile " exhibits a striking physiological adapta- tion. Such roots have been found only on young plants of the Lotos and apocarpous groups, excepting N. flavo-virens, in which adults also have contractile roots. In all of the plants mentioned, however, resting tubers are produced at the close of the season. The object evidently is to draw these tubers well down into the earth for protection against drought and herbivorous animals. For the adult plants, with the one exception named, perish when the flowering season is over, and consequently are not sup- plied with any special contractile organs. N. flavo-virens y^zanzibariensis bears a few contractile roots on the lower part of the tuber of adult plants. This hybrid usually has persistent tubers as in N. flavo-virens ; its contrac- tile roots, however, show none of the peculiar structure which character- izes those of N. flavo-virens. As a rule it is the lowest (oldest) root from each of the older leaf-bases which is contractile ; in N. lotus one or two upper roots occasionally contract. This lowest root is strongly geotropic, while the others are rather feebly so. It also persists in a living condition after all of its companion roots have completely rotted away. After con- traction it becomes fusiform in shape. The greater density of the cortex in contractile roots indicates that this is the layer where contraction takes place, probably, as Rimbach (1902) has shown for liliaceous plants, by shortening and widening of the cells. Thus we see a strong differentiation, apparently greatest in N. flavo-virens, into nutritive and mechanical roots. In stems the great density of tissue in all but the Castalia group, and especially in resting tubers, has already been shown. This is distinctly related to the storage of starch which takes place so largely in such plants. 9 IX3 ii4 The Waterlilies. I know of no experiments upon the geotropism of stems outside of the Eu-castalia group. Among these plagiotropism is very pronounced. A stem of N. odorata 3 cm. in diameter and 10 cm. long planted upright in a pot in early spring had turned its apex to a horizontal position by mid-summer. Young petioles and peduncles of different species turn through 900 in the course of twenty-four hours if placed in a horizontal position. This tendency is, of course, well developed in all species. The leaf-stalks and flower-stalks, however, are often so slender that their positions are mostly passive, except at the ends. The buds in Castalia, and the leaves in all species, by their buoyancy make their way to the surface of the water unless opposed by other forces. We have already noted that the leaf in coming up through the water has its midrib in a straight line with the petiole, and on reaching the air the petiole bends over at the top nearly at right angles to the lamina (Frank, 1872). That this is not a purely passive condition is shown by the stiffness and turgidity of the organs. Also, when plants are moved close together or placed in small tanks, where the upper parts of the petioles are thrown out of their former directions, they bend the lamina down until it is more or less submerged. In the course of twenty-four hours, however, the leaf is again lying flat upon the water surface. A corresponding bending has taken place in the upper part of the petiole. This bending may be due simply to readjust- ments of a mechanical kind in the elastic collenchyma of the petiole, or it may be the result of a distinctly vital action. In the latter case it is related to the phenomena next to be described. It has been noted that the " length of the petioles and peduncles depends upon the depth of the water in which the plant is growing." Not only is this true for plants in different localities, but a given plant, if moved from shallow into deeper water, where all of its leaves are sub- merged a few centimeters, will be found on the following day with all its leaves floating in the normal manner. Plants moved from deep to shallow water need no adjustment except in the angle between the petiole and lamina ; but leaves produced afterward have petioles only long enough for the new conditions. Now, it is certain that throughout the life of a leaf its petiole is continually growing longer, even though the water level remains constant. This is necessary in order to make room for the new leaves which are rising in succession from the center of the plant. One need only look at a large plant of any tropical Nymphaea as grown in gardens to be convinced of this. We see an area ten or twelve feet in Physiology. i i 5 diameter covered with the foliage from a single well-grown specimen. If the gardeners have not been too careful we see that the outer border of this ring is occupied by yellow and decaying leaves on very long petioles. Within these comes a zone of large green and shining healthy leaves, then smaller tender ones, and, at the center, one or two are only partly unrolled from the bud. The rate of elongation of a petiole of N. rubra detached from th parent plant was 2.5 cm. per day. This, with perhaps a certain amount of stretching of the collenchyma and softer tissues, may account for the rise of leaves on plants moved from shallow to deeper water, but it does not reach the case of adjustment from deep to shallow water. Frank (1872) investigated the causes of this phenomenon, giving especial attention to Hydrocharis morsus-ranae, and concluded that gravitation (as expressed by water pressure) and light, though not the causes, are used by the plant as signs by which to measure how much it has yet to accomplish before a growing organ has reached its most advantageous position (" Gravitation und Licht sind nicht die Erreger jener Wachsthums- formen, sondern die Pflanze bedient sich ihrer nur als Merkmale, an denen sie abmisst, wieviel sie noch zu leisten hat, bis das durch Wachsthum zu richtende Glied seine vortheilhafteste Lage erreicht hat." p. 85). The expression is an excellent one, even if the " signs " are insufficient. For is it not so with all stimuli and all organisms ? No external stimulus is in itself a final cause of any physiological action ; that is, the irritability of protoplasm is the indispensable condition of all vital functions. Karsten (1888), experimenting on Hydrocharis and Ranunculus sceleratus, showed the insufficiency of Frank's explanation, and proved that it is the oxygen of the air which inhibits the growth of the petioles as soon as the leaf-blade reaches the surface of the water ( " — dass es der Sauerstoff der Atmos- phere ist, welche bei den Schwimmblattern jene constatirte Hemmung im Wachsthum ihrer Stiele bewirkt, sobald sie die Wasseroberfiache errei- chen." p. 577). In Nymphaea, however, the inhibiting influence of the air still permits the slow elongation of the petioles by which the older leaves are moved farther and farther away from the center of growth. Once spread out upon the water surface, the leaves are subjected to the stress of currents in air and water. Jahn (1896) pointed out four requirements for such foliage: (1) lightness and firmness of lamina; (2) the greatest possible amount of surface ; (3) insertion of petiole at center of leaf ; (4) the petiole must make a large angle with the lamina. The first is attained by the great amount of air space, and in Lotos and Apo- carpiae by the girder-like veins. But in strong winds the leaves are often n6 The Waterlilies. raised up at one edge or blown completely wrong side up. The toughness of the lamina prevents its tearing. Breadth of leaf surface is carried to its ultimate extent in the circular leaves. The larger the leaves the less easily they will be overturned, and a further need will be appreciated when it is noted that waterlilies must have all of their foliage in one plane ; a distribution in space of three dimensions is not available to them. That the rounded outline has been acquired by a backward extension of the basal lobes of the leaf is due to the third condition. A round lamina with marginal attachment of the petiole is quite a possible thing. But if such a leaf were to exert any tension upon its anchorage, the effect would be to submerge the basal portion. The central attachment obviates this difficulty ; and peltation is a further advance toward the same end. Various other adaptations of the plant in irritable response to its peculiar environment have been referred to in previous pages. A number of interesting activities of the floral parts will be described shortly. The nutritive relations of waterlilies have not been specially worked out. They require a large amount of nitrogenous food, as every cultivator can testify, but prefer it in the form of nitrates rather than ammonia. Excess of calcium, as in limestone waters, is not beneficial. Doubtless these substances are absorbed with water by the roots directly, since no root-hairs are present. And the importance of the roots is shown by the great reduction in vigor of plants when moved from one pot to another, though they soon regain their strength. Wachter (1897) has shown, how- ever, that complete removal of the roots is less injurious than removal of the leaves. Water is also absorbed elsewhere than at the roots. A leaf of N. rubra 35 cm. across, severed from the parent plant and left floating in the water, with the end of the petiole tied to a stick above the water level, not only remained apparently healthy for three days, but grew 1.3 cm. in diameter, and its petiole elongated 6.3 cm. Whether or not this leaf was transpiring moisture through the stomata I cannot say ; but that a large amount of transpiration occurs from waterlily leaves was shown by a small tank in my father's garden. This was a water-tight iron vessel four feet square, and the surface was almost completely covered with leaves so that very little free evaporation could occur. After every clear summer day it was necessary to add one to three buckets of water to keep the level constant. So reduced is the xylem in the petioles that one questions how and where any sap can ascend. Some have thought the small air-canals in the vascular bundles are filled with liquid, but we find no evidence of this. But, in a leaf placed with the end of the petiole in Physiology. 117 watery eosin, the phloem portions of the bundles were stained as though the coloring matter had traveled up in them. The whole subject of the course of water in aquatic plants needs investigation. Respiration is to be looked upon as causing some of the most charac- teristic structures of water plants, namely the air-canals. In August, 1867, Lechartier experimented on N. alba as it grew naturally in the rivers. He found that on cutting off a leaf just below the water surface a flow of gas takes place from the air-canals of the petiole. This flow continues throughout the day, ceases at dark (7 p. m.), and begins again the next morning (8.30 a. m.). It may continue even against a pressure of 26 cm. of water, and after all of the leaves have been cut off the plant. In one instance 220 cc. was collected from a single petiole in an afternoon, and in another 262 cc. The total amount of gas from a single plant was 1,028 cc. in fifty-three hours. In these plants all of the leaves were submerged during the experiments. A plant with several floating leaves was next used. A petiole was cut off near the leaf, and when held 1 cm. below the water surface no gas came out ; when raised in a closed inverted tube of water 10 cm. above the surface of the river, the flow amounted to 600 cc. in fifteen minutes. This gas was tested in portions of 60 cc. collected successively, and at three times of day, as follows : The first, fifth, and tenth portions of air contained — AT 6.30 A. M. AT 11.30 A. M. AT 2.30 P. M. I. V. x. I. V. X. I. V. CO, 0. N. I 7-7 91-3 3 8.1 88.9 2.5 8.2 89.3 100 o-5 9 90-5 2-5 9-7 87.8 100 2.4 9-7 87.9 100 o-S 16.8 82.7 2 10-7 87-3 100 100 100 100 100 From these figures it appears that the air drawn out last at each period (V, X), and therefore probably from the stem and roots of the plant, contains more carbon dioxide than that in the petiole, and that the oxygen content of the inclosed air increases greatly while the plant is exposed to sunlight. At best, though, the proportion of oxygen present is much less than in atmospheric air. It appears also from the first experiments that the gases are not necessarily, if at all, derived directly from the air. Barthelemy (1874) attempted to repeat Lechartier's work, but found that the flow of gas diminished rapidly after the first gush ; he u8 The Waterlilies. therefore believed the flow to be due to a higher pressure in the plant than outside. He agreed with Lechartier in believing that there is a move- ment of gases in the air-canals of Nymphaeaceae toward the leaf, to make exit through the stomata, and concluded that stomata in general are for the exit rather than for the entrance of gases. Goebel's (1893) experi- ment seems to differ from these. On a dull autumn day petioles of N. rubra and stellata (?) were cut off under mercury near the water surface and held as nearly as possible erect. The smaller intercellular spaces became in every case injected with mercury up to the lamina, even in pieces 60 cm. long. This would indicate a considerable negative pressure. Recognizing the positive pressure as noticed by Sachs (though without mention of Lechartier or Barthelemy), Goebel concludes that when the amount of assimilation is slight there is a rarefaction of air in the canals caused by the continuance of respiration ; but when assimilation is active, part of the oxygen is given off into the intercellular spaces, and reaches the air-canals and finally extends to the rhizome and root. This view at once harmonizes the observed phenomena and explains the presence of the air-canals. It is the more plausible when we consider the peaty soil in which our waterlilies grow and the general absence of free oxygen in submerged muddy material. Since the behavior of the flowers has for its object the successful reproduction of the species, and since there is in this connection a very distinct series of movements and activities, the whole relation of these parts has been reserved to this point, although there would be good reason to take them up under the head of irritability. The rise of the flower bud from the rhizome to the surface of the water takes place in the same man- ner as that of the leaf, and the peduncle must be regulated in length by the same stimuli that regulate the petiole ; it is even capable, in Castalia, of a slight elongation if the water level rises after the flower first comes into bloom. A peculiar adaptation occurs, on the other hand, in plants of any species, when grown in very shallow water, say 10 to 15 cm. This often occurs in our gardens, where we have the tropical kinds as near to the surface as possible in order to keep them warm. Now, the flower bud is first pushed out from the apex of the plant in a very immature state, and it is to finish its development as it is rising through the water. In very shallow water, therefore, it reaches the surface when it is by no means ready to be exposed to the air. The situation is recognized at once by the plant, and the peduncle bends to one side and becomes S-shaped Physiology. 119 (Fig. 52) ; thus the bud maintains its erect position and is kept submerged. Then in due time, the bud having matured and the peduncle having reached a suitable length, the latter straightens up and lifts the former into the proper position for opening. In N. flavo-virens, where I have noticed this most plainly, the flower stands about a foot above the water. The opening of the flower occurs at a particular time of day for each species of Nymphaea. Indeed, we have attempted in the accompanying Fig. 62.— A strong-growing waterlily in shallow water. list * to imitate Linnaeus's floral clock in this genus alone. After remain- ing open a few minutes (N. amazonum) or hours, the flower closes at a certain time, and this is repeated on two (N. flavd) or three {N. odorata, caerulea, etc.) to six or seven (N. gigantea) successive days. This is the foundation on which was based the statement of Pliny and others that the flowers retreat under water at night. The process of opening and closing usually occupies nearly an hour. On its first day the flower always opens * Floral clock for the latitude 8- 4 a. m. N. amazonum (new flower) opens. t- 8 N. amazonum (old flower) opens wide. 6-6 N. amazonum closes; N. marliacea car- nea opens. 6-7 N. alba candidissima and odorata minor open. N. gracilis and elegansXzanzibariensis open. 7- 8 N. elegans, capensis, and caeruleaXzan- zlbariensis open. N. caerulea, marliacea chromatella, and tuberosa open. 8-9 N. marliacea chromatella (new flower) opens. N. capensisXzanzlbarlenBis opens. 8-10 N. zanzibariensisX opens. 10-11 N. mezicana opens. N. rubra and rubra rosea close. 11-12 m. N. tetragona and zanzibariensis open. N. omarana closes. of Philadelphia {40° N.). 13- 1 p. m. N. odorata, tuberosa, caerulea, and ele- gans close. N. alba candidissima closes. N. elegansXzanzibariensis and marliacea chromatella close. N. mexicana closes. N. capensisx zanzibariensis closes. N. tetragona, gracilis, and Pennsylvania close. N. capensis, zanzibariensis, and zanzi- bariensisX close. 1-2 2-3 8-4, 4-6) 6-6 6-7 7- 8 8- 9 9-10) 10-lH 11-12 N. dentata and omarana open. N. amazonum (old flower) opens sepals. N. rubra, rubra rosea, and devonlensis open. N. rudgeana opens and closes (Casp.). 120 The Waterlilies. about an hour later than on subsequent days, closes earlier, and spreads but half as wide. An exception to this must be made for N. amazonum, whose very peculiar habits are detailed in the description of the species. N. rudgeana has also shown a very short nocturnal period of opening. Observations on other Hydrocallis species would undoubtedly yield interesting results. The characteristics of each species, so far as known, are given in the taxonomic chapter. The stimulus to these actions is to be attributed, in diurnal species at least, wholly to light. If plucked and taken into a room they rapidly close, even though the normal time has not come. On the following morning, however, they reopen even in a very dimly lighted room. Subsequent movements of the cut flower soon become irregular, and it finally ceases to move. This is" probably due to the death of the protoplasm in the sensitive or motor centers, though the flower may remain open without wilting for some days. That light is the normal stimulus is also shown by the earlier opening of the flowers in more northern latitudes. Thus Kerner (1895) gives 7 a. m. as the time for opening of N. alba at Upsala, and 8 to 9 a. m. at Innsbruck, 130 farther south. Heat seems to have no effect upon the Eu-castalia group in ordinary atmospheric quantities ; for the time of opening and closing remains the same on days and nights when the mer- cury is continually above 8o° F. and on cool, cloudy days when the temper- ature does not rise above 650 F. In tropical species as grown out of doors in our climate heat does have a marked effect. One evening in 1900, when the mercury registered about 950 F., N. zanzibariensis remained open nearly an hour later than usual ; and in autumn, when the night temperature reaches 400 F. and the plants are looking the worse for wear, the diurnal tropical species become very irregular and open but a little way. The conclusion to be drawn is that the plant itself must be in a certain condition of health or tonicity to enable its parts to perform their functions. The effect of heat is essentially systemic rather than local. In Lotos it seems likely that heat may have a more distinct influence upon the flowers, but it is still obscure. They do not close until the heat of the sun is making itself felt, but on exceedingly hot nights they open to perfection, and on much cooler days they close at the usual time. But when the cold of autumn injures the plants, they sometimes remain open two or three days and nights continuously. On one occasion at Riverton, N. J., I saw about sixty large flowers of this type cut in the morning and placed in jars of water in a shaded glass house ; they closed very slowly and incompletely. We must conclude, therefore, that light is again the Physiology. 121 principal stimulus, causing in these closure, in other species opening, of the flower. The difference in effect is paralleled in positive and negative heliotropism and positive and negative geotropism. The influence of different colors of light has never been tried. Chemical stimuli were used by Tassi (1884). He found that flowers of N. alba and other plants were unable to perform their movements under the influence of anaesthetics, and in -some plants the color changed. The large and showy flowers of waterlilies are in many ways designed for the visits of insects. The colors are mostly conspicuous and attractive. N. caerulea is comparatively insignificant ; and N. rubra makes very little showing at night, but in the early morning it is magnificent. N. rudgeana, according to Caspary (1878), is often cleistogamous. The possibility of such a thing is suggested by its regular occurrence in Euryale ; and N. amazonum, with its odd habits, may well be considered as approaching the same condition. Certain it is that close pollination occurs regularly in this species. In cold weather, too, the flowers of N. amazonum seem to be fertilized without ever rising to the water surface. The occurrence of similar irregular cleistogamy in N. rudgeana may have been due to unnatural conditions. Observations are totally lacking on Hydrocallis species in their native haunts. In Castalia and Apocarpiae the flowers have a sweet ethereal odor. This is richest and most delicious in N. odorata, faint in N. tetragona, scarcely perceptible and only present on the first day of opening in N. alba. The Lotos and Hydrocallis groups have very little odor ; in the latter it is pleasant and ethereal, but in the former it is dull and rather offensive. The flowers are always pistillate on the first day of opening, and the stigmatic cup is filled nearly to overflowing with a large quantity of slightly sweetened water. This is excreted by the stigma itself and by the axile process. The stamens at this time stand erect around the stigma, leaving a narrow vertical pit in the center of the flower, bounded all over the sides by the still closed anthers of stamens of different lengths. All the stamens are stiff in Lotos, but in Castalia the innermost filaments are very slender and flexible, scarcely able to hold up the heavy anthers. The carpellary styles also stand erect. In flowers of the second and later days the stigma is quite dry and the papillae shriveled (in N. odorata and tuberosa ; Robertson, 1889), though in N. alba, according to Schulz (1890) still recep- tive. The carpellary styles, when long enough, have bent inward over the stigma until they lie horizontally or point downward, and the inner stamens meet in the middle so as more or less to hide the pistil. The 122 The Waterlilies. stamens now dehisce. Where the flower is open only for two days, as in Xanthantha, all of the stamens dehisce on the second day, but the outer ones precede the inner by a few hours. In the rest of Castalia the inner stamens dehisce on the second day of opening ; they do not completely cover the entrance to the stigma from above, but they hang over far enough to drop their pollen upon it, and sometimes close fertilization occurs. In Brachyceras, as a rule, the entrance to the pistil is totally obstructed on the second day by the inner stamens, and the outer ones are dehiscing ; subsequently the inner stamens bend outward and dehisce until, on the last day, the passage to the stigma is wide open again. In spite of this, N. caerulea and stellata are said by Caspary (1877) to be capable of self-pollination, the latter while still in the bud. The stiff stamens of the- Lotos group dehisce simultaneously on the second day, and undergo no considerable change of position. The exact manner of effecting cross-pollination and the insects concerned are unknown for the Lotos group, but the white and light-pink varieties are extremely fertile when grown out of doors in the vicinity of Philadelphia. The agency of insects in Hydrocallis is likewise undetermined. N. alba is, in the opinion of Delpino, fertilized by scarabaeid beetles (Cetoniae and Glaphyridae); he regards Nymphaea and Victoria as especially adapted for beetles. Kirchner, in his "Flora von Stuttgart," page 275, mentions "flower-beetles" as visitors {fide Miiller, 1883, p. 93; Schulz, 1890), and Picciola found Donacia (a chrysomelid beetle) abundantly. But Schulz found N. alba perfectly fertile with its own pollen, and considered that the flies and beetles which visit it always cause close fertilization. N. tuberosa is visited by small mining bees (Andrenidae), flower-flies, and beetles (Robertson, 1889), and N. odorata by Halictus and others. In our own experience, mining bees and flower-flies are the only insects found. They are drowned in large numbers in the stigmatic fluid, and are found in the closed flowers after anthesis. In a single bloom of N. capensis X zanzi- bariensis I once counted thirty-two little bees ! Bacon (1874) noticed dead insects in N. odorata and supposed that they were caught by the closing of the flower ; Delpino attributed their death to the heavy odor of the flower, and Planchon to the accumulation of carbon dioxide in the floral cup {fide Robertson). But Robertson is undoubtedly correct in considering it simply a case of drowning. The course of events is about thus : The insects which visit the flowers are in search of the copious pollen ; no true nectar is secreted. The earliest flowers to open each day are the older ones, in which pollen is plentiful and the stigma is dry. The insects fly Physiology. 123 from flower to flower in safety gathering their loads. When an hour later the new flowers open with their receptive stigmas, the insects are well dusted over. Coming to the narrow opening of the new flower, they rush in with confident haste; but reaching the slender inner stamens, these bend down under the weight, and the visitor is dropped into the pool of liquid in the stigmatic basin. As he struggles to get out by crawling up the styles, he must again lay hold of the treacherous inner stamens, and be again let down into the water. Thus pollen is washed off the bee into the fluid, and the need of the plant is met ; it is a matter of strength and endurance on the part of the visitor whether or not he gets out alive. When the stigmatic fluid is absorbed, the pollen grains settle down among the papillae and are in a position to germinate. In some species a consid- erable amount of heat is evolved in the newly opened flower, but never in anything like the quantity found in Victoria regia. On the last day of opening the flower begins to sink back into the water again. In Hydrocallis it takes but 2 to 5 hours for the closed flower to become totally submerged ; in N. odorata it requires 6 to 1 2 hours, and the flower remains nearly wide open. In Lotos and Brachyceras also the flower often sinks while still open, but on its last day it has already moved down near to the water surface, and is not totally submerged until 1 2 to 20 hours later. The submergence of the flower is due to active, even power- ful, movements of the peduncle, and occurs most pronouncedly in fertilized flowers. Infertile flowers scarcely get below the surface of the water, and then quickly disintegrate ; sometimes the peduncle decays before the floral parts, showing how great is the transmitted effect of fertilization. Under the most favorable circumstances only about half of the ovules proceed with development to form mature seeds, but the presence of 10 or 12 fer- tile ovules in each cell of such fecund species as N. lotus or caerulea is suffi- cient to stimulate the movements of the peduncle. In Lotos the peduncle simply bends over and carries the flower down, and the ripening fruit lies on its side. In N. amazonum there is a double curve, so that the ripening fruit is upright. But in other species more or less of spiral coiling accom- panies the bending. The stout peduncles of Brachyceras descend in the same manner as N. amazonum, but there is in addition a strong curving of the lower half of the peduncle to the right or left. The result is that we find the fruits near the bottom of the water, 15 to 35 cm. away from the crown of the plant and on the opposite side from the point of insertion of the fruit-stalk. N. flava in shallow water (15 cm.) makes the same kind of a bend ; as the stalk is more slender, the bends are sharper and the 124 The Waterlilies. fruit is pressed close against the mud. N. tetragona resembles N. flava in these respects. In deep water (i to 5 meters) N. flava, odorata, alba, and tuberosa carry the spiral turning to its highest development. From a few centimeters above the rhizome to within a few centimeters of the flower, the peduncle is coiled into a close helix from 2 to 8 centimeters in diameter, and with 2 to 8 or 10 turns ; thus the fruit is drawn down within about half a meter of the bottom of the water, and often quite into the mass of algae and other vegetation overlying the mud. Just such an adaptation has given the specific name to Vallisneria spiralis. Besides removing the fruit from the drying influence of the air, the purpose of these movements is doubtless twofold : First, to hide the fruit from the many aerial, aquatic, and amphibious animals which might feed upon it, for the majority of such creatures get their food at the surface rather than in the depths of the water ; secondly, to reduce the danger of the fruits being broken off, as would easily occur with the long, slender peduncles of Castalia. •* CHAPTER Y. TAXONOMY. ' Genus NYMPHAEA (Linn.) J. E. Smith. Generic diagnosis — Sepals 4 (rarely 3 or 5), inferior. Petals many (12 to 40), multiseriate, the lowest alternate with the sepals. Stamens many (20 to 700), multiseriate, inserted above the petals on the upraised torus which surrounds the carpels ; outer filaments broader, often petaloid ; innermost slender or filiform ; anthers introrse, bilocular. Carpels many (8 to 35), sunk in the cup-shaped, fleshy receptacle, fused dorsally with it, and produced above it into shorter or longer processes, the carpellary styles (except in Sec. Anecphya); margins of carpels fused centrally with the floral axis, which extends above them as a short axile process. Stigma broad, concave, radiate. Ovules very many, anatropous, pendulous from the sides of the ovary cells. ' Fruit a spongy berry, ripening under water and burst- ing irregularly. Seeds with a floating sac-like aril, open at the apex; endosperm very small ; perisperm copious. Embryo small, straight. Aquatic herbs with perennial elongated or tuberous rhizome, rooting in mud at the bottom of ponds, streams, etc. ; leaves floating, ovate to orbicular, fissi-cordate, 5 to 60 cm. in diameter ; flowers solitary, showy, floating or raised above the water on stiff scapes, 2.5 to 30 cm. in diameter, white, blue, red, or yellow, in all shades. Founded by Smith on N. alba L. Nymphaca, Tournefort 1700, in part. Linnaeus 1753, in part. Nymphaea, J. E. Smith 1809. DeCandolle 1821 b, 1824. Planchon 1853 b. Lehmann 1853 a. Caspary 1855, 1865, etc. Bentham & Hooker 1862. "Gray, etc. Not Salisbury 1806 a. Castalia, Salisbury 18060. Greene 1887 a and b, 1888. Britten 1888 a. Lawson 1889. Britton & Brown 1897. Leuconymphaca, Ludwig 1737. Kuntze 1891. MacMillan 1892. Nymphea, Rafinesque 1830. 1 This chapter, together with most of the bibliography at the end of Chapter VIII, was sub- mitted to the University of Pennsylvania in partial fulfillment of the requirements for the degree of Doctor of Philosophy, May 1, 1901. At that time I had placed the members of the Lotos group in a single species, N. lotus. The separation of these into three distinct species, the fusion of N. Hava and N. mexicana, the addition of some African species, and the substitution of " N. fiavo- virens " for " N. gracilis " are the only material changes made since that time. A few citations and measurements have been added in several species. 125 126 The Waterlilies. In order to present a tabular view of species and subgenera the following synopsis and key are inserted. The genus Nymphaea divides readily into two main divisions, which are subdivided into five subgenera. Group I. Nymphaeae apocarpiae. Subgenus (i) Anecphya. (2) Brachyceras. Group II. Nymphaeae syncarpiae. Subgenus (3) Castalia. (4) Lotos. (5) Hydrocallis. Analytic Key. Carpels free at sides, i. e., walls between ovary cells double (Fig. 40) {Nymphaeae apocarpiae) 1 Carpels fused at sides, i. e., walls between ovary cells single (Fig. 40) {Nymphaeae syncarpiae 13 1 . Carpellary styles wanting ; stamens unappendaged (Anecphya) N. gigantea 1. Carpellary styles present, thick and fleshy ; stamens appendaged (Brachyceras) 2 2. Flowers yellow 3 2. Flowers never yellow 4 3. Leaves 4 to 6 cm. across ; veins inconspicuous N. sulfurea 3. Leaves 20 to 25 cm. across ; veins very prominent N. siuhlmannii 4. Leaves entire or nearly so 5 4. Leaves dentate or crenate 10 5. Leaves narrowly elliptic (twice as long as broad) , blotched N, ovalifolia 5. Leaves ovate or orbicular 6 6. American species ; flowers pale violet N. elegans 6. African species 7 7. Mature leaves bulbilliferous N. micrantha 7. Not bulbilliferous 8 8. Leaves deep purple beneath A^. calliantha 8. Leaves green or purplish-green beneath , spotted with black 9 9. Petals s to 10 ; stamens 11 to 16 N heudelotii 9. Petals 14 to 20 ; stamens 50 to 75 N. caerulea 10. Flowers white 11 10. Flowers blue or pink 12 11. Petals narrow, acute ; leaves pure green beneath N. flavo-virens 11. Petals elliptic, obtuse ; leaves black-spotted beneath N. ampla 12. Petals 20 to 30 ; stamens 150 to 275 N. capensis 12. Petals 1 1 to 14 ; stamens 33 to 55 A'', stellata 13. Sepals prominently veined ; styles linear ; leaves spinose-dentate (Lotos) 14 13. Sepals obscurely veined ; outer filaments petaloid 17 14. Leaves 10 to 15 cm. long, thin and membranous N. zenkeri 14. Leaves 20 to 25 cm. in diameter, stiff ; venation very prominent 15 15. Flowers red ; leaves bronzy N. rubra 15. Flowers white 16 16. Leaf orbicular, smooth or finely puberulent beneath N. lotus 16. Leaf ovate, densely pubescent beneath N. pubescens 17. Styles clavate ; flowers nocturnal (Hydrocallis) 18 17. Styles ligulate ; flowers diurnal (Castalia) 26 18. Sepals and petals long acuminate A^. oxypetala 18. Sepals and petals acute or obtuse 19 19. Leaves sinuate N. rudgeana 19. Leaves entire 20 20. Petiole bearing a ring of long hairs at top A'', amazonum 20. No ring of hairs at top of petiole 21 Taxonomy — Nymphaeae apocarpiae. 127 21. Styles 12 to 16 mm. long A^. gardneriana 21. Styles 4 to 10 mm. long 22 22. Sepals with black markings 23 22. Sepals marked with dark crimson lines 24 22. Sepals marked with fuscous striae 25 23. Spots few, large ; leaves large, broadly peltate N. lasiophylla 23. Spots few, linear ; leaves small, narrowly peltate N. tenerinervia 24. Leaves green above and below N. blanda 24. Leaves marked beneath with dark purple forked and branching lines. . . . A^. jamesoniana 25. Leaves very narrowly peltate, small N. stenaspidota 25. Leaves larger and more peltate N. gibertii 26. Rhizome erect 27 26. Rhizome horizontal ; flowers white or pink 29 27. Flowers yellow (Xanthantha) N. tnexicana 27. Flowers white, small (Chamaenymphaea) 28 28. Leaves plain green above A'', fennica 28. Leaves more or less blotched with brown A^. tetragona 29. Leaves crowded on the rhizome ; European or Asiatic 30 29. Leaves scattered on the rhizome; American 31 30. Line of attachment of sepals to receptacle sharply angular A^. Candida 30. Line of attachment of sepals rounded A^. alba 31 . Petals spatulate ; rhizome tuberiferous N. tuberosa 31. Petals elliptic ; rhizome not tuberiferous N. odorata Group I. NYMPHAEAE APOCARPIAE (= Lytopleura Casp. 1865, 1878, 1888). Carpels free from one another at the sides, fused along part of the suture with the axis of the flower, and dorsally with the perigynous torus. Tropical species with diurnal flowers ranging in color from blue through pink to white, raised 7 to 30 cm. above the water, on stiff peduncles. Venation of sepals inconspicuous. Outermost stamens first to dehisce, innermost last. Principal air-canals in the peduncle 6 (5 to 7), in the petiole 2; idioblasts very few; multicellular ingrowths rather frequent in the air-canals. Rhizomes erect, tuberous, drying off in the resting season. Plants gla- brous throughout, except on the apical portion of the tuber and the base of the petioles and peduncles. Subgenus 1. ANECPHYA Casp. 1865, 1888. Carpellary styles absent. Stamens, very many, with narrow filaments and short, curved anthers, without appendage, or simply mucronate, inserted very densely at the summit of the torus around the stigmatic disc, the outermost being at some distance from the insertion of the innermost petals. Seeds larger than in the next section. Flowers appearing subtended each by two leaves, i. e., there are on the rhizome oblique series of flower bases alternating with two series of leaves. One species, in Australia and New Guinea. Sec. Cyanea, Planchon 18526, 18536 (in part). Sec. Appendiculatae, trib. Cyananthos, Lehmann 1853a (in part). Sec. Inappendiculatae, trib. Castalia, Lehm. 1853a (in part). 128 The Waterlilies. Nymphaea gigantea Hook. (Plate III.) Diagnosis. — See under subgenus Anecphya. Nymphaea gigantea, W. J. Hooker 18520, Ad. original specimen in hb. Kew. Planchon 18516, 1852 d, 1853 6. Rev. Hortic. 18526. Lehmann 1853 a. Paxton 1853 d. F. Muller 1861, 1862. Bentham & Muller 1863. Caspary 1865. Garden 1883 b. G. & F. 1893 a, with good half-tone. Tricker 1897. Moenkemyer 1897. G. C. 1900. Conard 1901a. Not seen: Lemaire 1853; Duchartre 1858; A. G. 1894, 1897. Castalia stellaris, Salisbury 18066 (in note to tab. 14, referring to specimen collected by Banks at Endeavor River). Castalia gigantea, Britten 1888 a. Lawson 1889. Victoria Fitsroyana, Hort. (fide Hooker, 1. c). Description. — Flower 6 to 30 cm. in diameter (12.5 to 19 cm. as cultivated about Philadelphia; 30 cm. in the type specimen, dried), opening for 7 days from 9 a. m. to 7 p. m. (Casp. 1865), inodorous. — Bud ovoid, rounded at apex. — Peduncle terete, rather slender (1 to 2.5 cm. in diameter), tapering upward, smooth, light green, rising 10 to 35 cm. (18 inches, Robinson 1865) above the water; air-canals as in group Apocarpiae (figured by Hooker 18520, as petiole). — Receptacle short, small, dark green. — Sepals 4, oval or elliptic, rounded at the apex, breadth : length = 1 : 2.5 to 3 (average) ; persistent in fruit. Anterior (outermost) sepal pure green without, with a few blackish lines and dots; margins blue-purple above, shading to carmine-purple at point of attachment. Posterior (innermost) sepal with a broad green area in mid- dle of back; a wide border (3 to 5 mm.) around apex and sides, narrowing to the base of the sepal, covered by the lateral sepals in the bud, of a dark royal-purple color in apical part, shading to sky blue at sides and carmine-purple at base. Lateral sepals with one side covered, and similar in coloring to posterior sepal. Five strong veins and several smaller ones visible at base of each sepal. Inside of sepals sky blue, shading to royal purple at apex, dark blue on lateral margins, and carmine purple be- low.— Petals 18 to 51, longer than the sepals and not entirely covered by these after the first day of opening; outermost petals obovate, obtuse, narrowed at base, deeply concave, not at all sepaloid, outside sky blue, shading to royal purple at apex and base, similar in color, but paler, within. Intermediate petal obovate to spatulate, rounded at apex, tapering at base, colored like outer petals, but paler throughout; innermost petal oblanceolate, rounded at apex, tapering to the base, pale whitish blue, tinged pure blue at apex and base. All of the petals thin and fragile, finely 1 to 7 nerved, satiny and crumpled like chiffon, fading in color on later days of opening. — Stamens 367 to 745, incurved at summit through an arc of 450 to 900, half as long as the petals, separated from these by a wide interval and inserted densely on the upper part of the ovary. Anthers bright yellow, about 75 outer ones bearing a short, thin, acute apical appendage (or all unappendaged, Casp. 1865). Filaments paler yellow; 10 or 12 outermost ones broader than their anthers; all the rest are nearly or quite thread-like and longer than their anthers. " Outermost anthers first to open, inner- most last. Pollen smooth" (Casp. 1865). — Carpels 12 to 20, quite distinct from one another and rather easily separated from the receptacle, stigmatic over all their upper free surfaces ; carpellary styles wanting. Ovules large, few in number. — Seed " large, CARNEGIE INSTITUTION OF WASHINGTON WATERLIUES, PLATE 3 tit NYMPH/EA GIGANTEA. i. Anterior sepal. 2, Outermost stamen. 3, Inner stamens, 1. Median petal. 5. Vortical -eriion of ovary. 6, Outermost petal HFIIOfYPE CO., B03TON. Taxonomy — Nymphaea gigantae. 129 dark colored, ellipsoid, apiculate, usually covered with interrupted longitudinal lines of hairs " (Casp. 1865). (See table of measurements of flowers, below.) Submerged leaves from sprouting tuber (Fig. 34) 2 or 3, entire, green above, red- dish beneath. First leaf triangular-hastate, with rounded apex and angles, and nearly straight sides; sinus, very broad and shallow. Second leaf ovate-sagittate, broadly rounded at apex ; sinus broadly triangular, lobes narrowly rounded. First floating leaf ovate, with deep open sinus and rounded apex and lobes; margin slightly irregular- wavy. Second and third leaves similar, with narrower sinus and margins distinctly sinuate-dentate on outer side of lobes; first two dull green beneath, third suffused with purplish blue, and with a few blackish spots. Leaf of mature plant very narrowly peltate, orbicular-ovate or elliptic, 60 cm. or less in diameter, breadth : length = 1 : 1.4 ; margin flat or slightly wavy, sharply den- tate; teeth short, acute (or obtuse, Casp. 1865 and Mrs. Rowan's paintings), 2.5 cm. apart and 0.6 cm. long ; upper surface green and veiny, minutely elevate-punctate ; un- der surface glabrous, brownish pink when young, becoming Tyrian purple when full grown. Veins prominently reticulated on under side of leaf; primary nerves 9 or 10 on each side; length of principal area: length of radius of leaf = 1: 1.33. Depth of sinus =0.37 of length of leaf; margins entire, touching or slightly overlapping within, then curving out, becoming 7.5 to 15 cm. apart at periphery of leaf; angles produced into a fine subulate tooth. — Petiole smooth, terete, 60 to 150 cm. long, with 2 large air-canals, 2 smaller ones before and behind these, and a surrounding ring of 10 still smaller ones ; 2 double bundles anteriorly and posteriorly, with a peripheral ring of 1 1 bundles. — Stipules absent, but petiole winged at base, the wings gradually vanishing above. Rhizome (tuber) ovoid; small ones (Fig. 14) nearly spherical, 1.6 to 1.75 cm. in diameter, contracted above, the vegetative bud springing abruptly from the rounded summit; larger tubers with prominent leaf-bases. Each flower subtended by two leaves, the peduncular scars on the rhizome forming an oblique series alternating with two series of leaf-scars (Casp. 1865). Measurements of Flower from University of Pennsylvania Botanic Garden, 1900 (in cm). Sepals. Petals. Stamens. Length. Breadth. Length. Breadth. Length of Filament. Length of Anther. Width of Filament. Width of Anther. Length of Appendage. 7 (outer.) 7 (inner.) 2.2 2.4 8 (outer.) 7-3 (median.) 6.7 (inner.) 2.6 2.2 1.6 2.55 1.9 1-45 0.33 I-I5 0.95 0.65 0.3 0.25 0.075 0.04 0.025 O.16 O.08 O.08 0.04 0.16 0.075 0 0 10 '30 The Waterlilies. Measurements of Leaves (in cm.). First submerged*.. Second submerged* First floating* Second floating*... Third floating* Mature Mature Length of Midrib. 24 2.23 2.55 2.87 3-2 Length from Apex to Angle. 2.87 3-2 4 5-1 4.8 38.2 46 Depth of SfntiB. 0-5 0.8 1.6 1.9 2.23 Width of Leaf. % 2.5 3-2 30.6 40 Width of Sinus. i-9 1.9 1. 12 1.42 1-3 * From a small tuber. Geographic Distribution. — Australia and New Guinea. "Lakes and marshes throughout tropical Australia" (Benth. & Mull. 1863). "At river mouths in S. Australia," Schomburgk 1875. " Moreton Bay, Queensland; Clarence River, N. S. W.," Benth & Mull., 1. c. Papua, Mull. 1875. " New Guinea, coll. Zippel ; Endeavor River, coll. Banks," Casp. 1865. Rockhampton, E. Australia, coll. A. Dietrich, in hb. Berlin. King's Creek, Queensland, coll. Miiller, in hb. Munich. Type collected by Bidwill, No. 39, at Wide Bay, Queensland, in hbb. Kew, Berlin. Forma alba — with white flowers. Forma rosea — with pink flowers. Benth. & Mull., 1. c. Benth. & Mull., 1. c. Notes. — Introduced into England in 1852, and fruited by Van Houtte (Ghent) in 1855. Flowered in open air, in heated water, at Glasnevin (Dublin) in 1865. Fre- quent in cultivation in America, flowering freely out of doors all summer. Tubers are said to germinate better in deep water, i. e., about 60 cm. ; but this is not necessary if good bottom heat is furnished. Var. violacea (Lehm.) new comb. (Plate I, Frontispiece.) Smaller than the type. Flower 12 to 15 cm. in diameter. Bud conical-ovoid, acute. Sepals coriaceous, ovate to oblong, tapering above with nearly straight sides to the obtuse apex ; outer surface dark green more or less densely marked with inter- rupted heavier and lighter black lines; inner surface pale. Petals many, rich violet, slightly shorter than the sepals ; outer ones larger, obovate oblong, tapering above. Stamens very many, much shorter than the corolla; appendage obsolete. Carpels about 10. Leaves subovate, repand, light green above, deep purple beneath ; petiole to apex 9 to 11 cm.; width 11.5 to 1 2.7 cm.; lobes 6.4 cm. long, obtuse; primary nerves 10 to 12 ; veins numerous, reticulate, prominent beneath, green. Sinus margins curved, open at periphery. Rhizome cylindrical, the size of one's little finger. (Details mostly from Leh- mann.) Nymphaea violacea, Lehmann 1853 a, Ad. original specimens coll. John Macgillivray, Oct. io, 1848, Cape York, Voyage of the Rattlesnake, Botany, 410, in hb. Kew. N. Banksii, Cunn., fid. original specimen, coll. A. Cunningham, June, 1819, Endeavor River, N. E. Australia, in hb. Kew. Taxonomy — Nymphaea elegans. 131 N. serrata, Muller, fid. original specimen, coll. F.. Miiller, July 15, 1855, Arnheim's Land, Aus- tralia, in hb. Kew. N. repanda, Muller, fid. original specimen, coll. F. Muller, July 16, 1856, Arnheim's Land; also Head of Sturt's Creek, Australia; both in hb. Kew. N. versicolor, Muller, fid. original specimen, coll. F. Muller, July 15, 1856, Arnheim's Land, Aus- tralia, in hb. Kew. Not Roxb. 1809. N. stellata, Muller 1861, fid. original specimens, sent by F. Muller from Rockingham Bay, Queens- land, in hb. Munich; from Escape Cliffs, N. Australia, coll. W. Halls, in hb. Kew; from Bar ron River, Queensland, coll. W. Sayer, in hb. Paris ; also in hb. British Museum. Not Willd. 1797. N. Brownii, F. M. Bailey, 1899, fid. original specimen in hb. British Museum. Geographic Distribution. — Queensland, Arnheim's Land, Australia. Notes. — Caspary considered this merely a small form of N. gigantea; but the paintings of both types made by Mrs. Rowan in the native habitat, which I saw at the Philadelphia Museum, April 16, 1901, show the marked peculiarities cited above, which certainly entitle the plant to varietal rank. Mrs. Rowan considers it a very distinct type, and says it is restricted to the Cape York Peninsula, a region practically unin- habited by white people. We are especially fortunate in having opportunity to repro- duce Mrs. Rowan's painting from the original, which is now the property of the University of Pennsylvania. Examination of the European collections gave fully conclusive evidence on the rank of this form. (Cf. also its synonymy.) Subgenus 2. BRACHYCERAS Casp. 1865, 1878, 1888. Carpellary styles present, short, stiff, and fleshy. Stamens numerous, following the petals in order on the torus directly or without much interval ; exterior anthers long, more or less strongly appendaged with a process of the connective, the corre- sponding filaments being flat and more or less expanded. Seed smaller. Twelve spe- cies of world-wide distribution in the tropics. Cyanea and part of Lotos DC. 1821 b. Cyanea (mostly) Planchon 1852 b, 1853 b. Section I. Appendiculatae Lehmann 1853a (in part). Nymphaea elegans Hook. (Plate IV; Fig. 53.) Leaves entire or slightly wavy at base, narrowly peltate, broadly ovate to ovate- orbicular; under surface dark purple; about 18 cm. long. Flower 7 to 13 cm. in diam- eter, pale violet, open from 8 a. m. to 1 p. m. Buds ovate. Sepals marked with black lines and dots. Petals 12 to 20, ovate, obtuse. Stamens about 100, stout, yellow; appendage minute; filaments narrow. Nymphaea elegans, Hooker 1851, fid. original specimen, cultivated at Kew, from Texas, in hb. Kew. Revue Horticole 1851 b. Paxton 1853 b. Sterns 1888. Tricker 1897. Conard 1901 a. Castalia elegans, Greene 1888. Lawson 1889. Rose 1895. Nymphaea Mexicana, Gray 1850. Not Zucc. 1832. Description. — Flower 7 to 13 cm. in diameter, open widely on three successive days from 8 a. m. to 1 p. m. ; odor sweet, moderately strong. — Bud ovate, obtuse or rounded at apex. — Peduncle terete, slender, stiff, rising 12 to 18 cm. above the water, 132 The Waterlilies. reddish brown, with 5 or 6 main air-canals, each with two very small canals outside. — Receptacle very narrow, sloping. — Sepals 4 (or sometimes 5) lanceolate, breadth : length = 1 : 34 to 3.6, obtuse or sub-acute ; outer surface dark green, marked with fine brownish-black lines and dots, especially near the margins ; 9 longitudinal veins visible by looking through the sepal at a light; inner surface dull greenish-white. — Petals 13 to 24, lanceolate, tapering to base and apex, breadth : length = 1 : 3.7 in outermost petals, crowded on lower part of torus. Outer surface of outermost petal sepaloid c a Fig. 58.— Nymphaea elegant; a, sepal ; b, petal of outermost whorl ; e, petal of second whorl ; d, petal of Innermost whorl ; e, outer stamen ; /, successively smaller stamens ; g, vertical section of ovary. Natural size. Floral Organs. No. No. No. No.* Sepals . . . Petals . . . Stamens . Carpels . . Total .... 4 21 145 22 5 20 118 22 5 24 139 25 4 13 15 192 165 193 * Small flower, early in season. in midline, green, with blackish lines and dots, with 5 strong veins visible on looking through; edges and inner surface pale violet, nearly white, fading to dull whitish on lower third of petal. Second row of petals a little shorter than first, pale violet above, nearly white in lower half, 5-veined. Innermost petal narrowly lanceolate, colored like second series, but very pale. — Stamens 80 to 145, stout, yellow, about one-half the length of the petals, inserted around the top of the ovary, leaving a bare space above the petal insertions; appendages very small, only present on the outer stamens. Outermost filament narrowly elliptic, slightly broader than anther ; innermost tapering CARNEGIE INSTITUTION OF WASHINGTON. WATERLILIES, PLATE 4. A NYMPH/EA ELEGANS. Taxonomy — Nymphaea elegans. 133 from the width of the anther at top of filament to about one-half as wide at base. All of the anthers longer than the filaments. (See measurements at end of descrip- tion.)— Ovary nearly hemispherical, with a naked ring on the sides between petals and stamens.— Carpels 15 to 25; styles yellow, erect, minute, 0.8 mm. long, a little broader than high, acute, the tip projecting as a very small point beyond and above the broadly rounded and centrally grooved stigmatic ray. Stigma more or less curved funnel-shape, with large central depression, in which the slender, conical, acute axile process stands free. Diameter of stigma 0.8 cm. ; depth of depression 0.3 cm. ; height of axile process 0.2 cm. — Fruit rounded, much depressed, cheese-shaped, 3.2 cm. in diameter by 2.2 cm. high, to 3.5 cm. in diameter by 1.9 cm. high, of a pale greenish color, subtended by the persistent sepals, contracted above to the low crown of hard, inwardly- directed styles; stigma basin-shaped, about 1.6 cm. in diameter, black, radially striate- sulcate; axile process broadly conical and obtuse. — Seeds nearly globose, about 0.13 cm. in diameter, mucronate at hilum by persistence of a portion of the funiculus, dull olive green, smooth, with rather strong raphe. Aril about as long as the seed, and closely investing it. Germination may occur immediately on ripening of seed, or after a period of drought. First leaf of seedling filiform, 1.6 to 2.8 cm. long. Second leaf ovate, narrowed above, obtuse, 1.12 cm. long by 0.48 cm. wide, rounded at base; petiole 0.8 cm. long. Third leaf like second, but with slightly cordate base. Fourth leaf broader and more cordate at base. Fifth leaf broadly ovate, cordate, with rounded apex and basal lobes. First submerged leaf from small tuber (Fig. 34) narrowly triangular, acute at apex ; base truncate with rounded angles. Second leaf typically cordate, acute. First three floating leaves entire, ovate, broadest about midway of length, with rather wide sinus; apex broadly rounded, angles rounded; first two leaves green on both sides, third green above, purplish beneath. Fourth floating leaf speckled above with small brownish blotches, under surface brownish red with a number of fine darker specks ; margin wavy on outside of lobes. Measurements of Leaves from Small Tuber (in cm.). First submerged . . Second submerged Second floating . . . Fourth floating . . . Length of Midrib. 1.6 2-5 2.6 Length from Apex to Angle. 1-55 2.9 4-5 5-7 Depth of Sinus. O 0-45 1.9 2-55 Width of Leaf. 0.5 1.9 3-0 4.15 Width of Sinus. 1.0 1.2 Leaf of mature plant broadly oval, or at times nearly orbicular ; breadth : length = 1: 1.3; usual length 15 to 18 cm., peltate by 0.6 to 2.0 cm., of rather thin texture, entire in upper half, very gently sinuate-dentate in basal half; apex rounded. Upper surface dark green, with small blackish-purple markings when young; under surface bright purplish-red with frequent blue-purple spots, greenish along the midline. Veins prominent out to the sixth grade; principal nerves 6 to 8 on each side; length of prin- cipal area : radius of leaf = 1 : 1.76.— Sinus open, depth : length of leaf = 1 : 3 134 The Waterlilies. (approx.), margins straight or slightly concave, converging toward the periphery; angles about 3.2 cm. apart, tapering, acute. — Petiole slender, light brown, attached to the lamina by a slender " collar " ; principal air-canals 4, two larger and two smaller ; stellate cells numerous on walls of larger canals. Rhizome (tuber) erect, ovoid, black, thickly covered with leaf scars, about 7.5 cm. long by 3.5 cm. thick, ripening early in the fall. No stolons formed in vegetative period. Small tubers (Fig. 14), 1.2 to 1.35 cm. in diameter, spherical-ovoid, smooth in lower half, tapering above to the scaly-woolly vegetative end. Measurements in centimeters. Stamens OF Flower A.* Sepals. Petals. Total Length. Width of Filament. Length of Anther and Appendage. Width of Anther. Length of Appen- dage. Length. Width. Length. Width. Remarks. 1-75 I.42 1. 13 0-95 O.64 0.2 0.l6 0.13 O.I O.06 0.47 0.64 0.48 O.16 0.13 0.13 O.13 0.13 0.08 0.04 0 0 0 3-8 <*3-8 3-65 3-8 2.9 I.I ao.96 1. 1 O.96 0.8 3-5 3-3 3-2 0.9 0.9 0.8 Outer, a Median, a Inner, a * Measurements in this table and those marked a in adjoining tables are from the same flower. Geographic Distribution. — Texas, Mexico to Guatemala, chiefly in warm low- lands. " Near head of Leona River " (Gray 1850) ; Waco, Texas (Sterns 1888) ; Rio Grande, El Paso, Texas (Coulter 1890) ; Guatemala (Smith 1893) ; Tamaulipas, Vera Cruz, and Mazatlan, Mexico; Brownsville, Texas (coll. Pringle, Flora of Texas, 1888, No. 1955). Also coll. Berlandier, No. 2402, June, 1834, at Palo Alto, Rio de las Mines, Mexico, in hbb. Kew, Paris, and No. 972 in hb. Berlin ; coll. Hartweg, Mexico, No. 1972 in hb. Paris and 1592 in hb. Delessert. Notes. — Types collected by Dr. C. Wright on Leona River in June, 1849, and sent to A. Gray, and thence with seeds to Hooker, who raised plants and described the species from them. It was rediscovered in 1887 by Misses Trimble and Wright at the Waco locality in eastern-central Texas. The flower is beautiful and delicately colored ; the plant grows easily and blooms and fruits profusely. Nymphaea ampla (Salisb.) DC. (Plate V.) Leaves 15 to 40 cm. in diameter, sub-orbicular, narrowly peltate, sinuate or nearly entire, with small black spots above and below, at least when young; under surface red-purple; lobes acute. Flowers white, 7 to 13 cm. in diameter; sepals ovate-lanceo- late, obtuse or somewhat acute, marked with blackish lines; petals 7 to 21, lance- ovate, obtuse; stamens 30 to 190, yellow, outermost ones much longer than innermost. Castalia ampla, Salisbury 1806 a, 1806 b. Nymphaea ampla, DC. 18216, 1824 (fid. leaf of original specimen, in hb. British Museum). Spren- gel 1825. Don 1831. Macfadyen 1837. Schomburgk 1848. Gray 1850. Planchon 1853 b. Grisebach 1864. Caspary 1878. Garden 18836. Watson 1883. Hemsley 1888. Pringle 1890. Gray, etc., 1895. Conard, 1901 0. )3«- CARNEGIE INSTITUTION OF WASHINGTON WATERLILIES, PLATE 5 NYMPH/EA AMPLA. i. Flower. 2, Leaf. 3, Flout r si-en from side, 4, Rays ofsti^iiKi and carpellary styles. ;, Stamens. 6, Rays ofstiynifl from another (lower. MELIOTYPE CO., BOSTON. Taxonomy — Nymphaea ampla. 135 N. Lotus, Linn. 1762 (the American plant). Aublet 1775. Willdenow 1797. Loudon 1855. N. Candolleana, Lehmann 1853 a, as to the leaf (fid. original specimen from hb. Lehmann now in hb. Berlin). Flower is of N. rudgeana. N. foliis circinnatis maximis, Burmann 1759. N. ampla var. Plumieri, Planchon 1853 b. Grisebach 1857. Caspary 1878. N. trisepala Gaudich. ? fid. Eichler 1878. Description. — Flowers white, 9 to 13 cm. across. — Sepals 4, coriaceous, narrowly oblong, breadth: length =1:2.3 to 5-8 (average of 51 sepals from 43 flowers = 1 : 3.6, C-.sp. 1878), obtuse, acute or acuminate, outer surface green, marked with short blackish lines. — Petals 7 to 21, the outermost tinged yellowish-green. — Stamens 30 to 190, outermost much longer than innermost ; anthers appendiculate. — Carpels 14 to 23 ; styles short-conical, gradually narrowed or abruptly apiculate-acuminate, acute, sub- acute, or obtuse, breadth : length =1:1.5 on the average, =1:1 to 2.5 in extremes; axile process rounded, breadth : length = 1 : 0.09 on the average, = 1 : 0.3 to 2.0 in ex- tremes; stigma extending out on styles in short, rounded or rarely acute rays. — Seed sub-globose-elliptic, about 1 mm. long, with longitudinal rows of hairs. — Leaves large, narrowly peltate (3 to 31 mm.; average 10 mm.), sub-orbicular, 15 to 40 cm. in diameter, sinuate or nearly entire, green above, with small black spots ; under surface red-purple; veins prominent, principal nerves 5 to 14 (average 10) on each side of leaf ; length of principal area : radius of leaf =1:2. Geographic Distribution. — Tropical and sub-trop- ical America, from 260 N. in Texas, through the An- tilles, to 8° S. in Brazil; Fort Clark and Spofford, Texas (Plank 1896). Lampasas, Mexico. Neuvo Leon; Lakes between Ocuiltzapotlan and Famulte de las Sabanas; La- gunas de Macultepec (very large specimen, coll. J. N. Rovirosa, 1889, in hb. Acad. Nat. Sci. Phila., with petals 9.5 cm. long and leaf 56 cm. long by 48 cm. wide). Mexico, coll. Houston, No. 173 1, in hb. British Museum. Springs and streams near Monterey, Mexico (Pringle, Plant. Mex., 1889, No. 2581). La Paila, New Granada. St. Domingo island. Vera Cruz, Hazienda de Mamulique, Hazienda de Tamatoc, Laguna Verde, Mexico (Casp. 1878). Aguada del Labach, Yu- catan, coll. Schottmuller, No. 528, in hb. British Museum ; also Gaumer, No. 428, in hb. Boissier. Cuba; St. Miragoane; Alligator Pond, Jamaica; Rio Ulna, Honduras; Belize, Honduras, coll. J. Robertson, Dec. 27, 1889, No. 20, in hb. British Museum. Hog Island, Bahamas, coll. Eggers, No. 4081. Martinique, coll. Plee, in hb. Paris. Guatemala (Smith 1 891, 1893) fid. original specimen in hb. Kew. Encruzilhada, near Pernambuco, Brazil, coll. H. Schenck, Herb. Brazil, No. 4152, in hb. Berlin, and coll. Gardner, No. 915, near Alinda, in hb. Kew. Demarara, British Guiana, coll. Parker, in hb. Kew. Trinidad, coll. Fendler, No. 207, in hb. Kew. St. Vincent, south end, coll. Smith, No. 1708, in hb. Kew. Prov. Quito, Peru, coll. Hartweg, No. 1592, in hb. Boissier. Notes.— In the arrangement of this species we have followed Caspary as closely as possible; he examined the type material in nearly all of the European collections. DeCandolle's type specimen consists of a leaf of this species with a flower of N. rudgeana Mey. Sepal Dimensions. Breadth. Length. cm. 0.9 2.9 2.4 cm. 2.6 9.2 9.4 136 The Waterlilies. The true species (var. Plumieri Planch., etc.) is of powerful growth, but has not yet been introduced into cultivation. We have discarded the varietal name in this and in other species, allowing the specific name alone to stand for the form originally described. Var. speciosa (Mart. & Zucc.) Casp. Of medium size; stamens 42 to 129; leaves less coriaceous, margins more or less repand-sinuate or dentate, with obtuse teeth ; nerves slightly prominent beneath ; lobes divergent, overlapping or approximate. Tuber sub-globose. N. speciosa Martius & Zuccarini 1832, fid. original specimen, coll. Martius, in aquis S. Christoph., Prov. Rio de Janeiro, Brazil, in hb. Munich, in sched. No. 28. JV. ampla, Hooker 1849. N. ampla P Hookeri, Planchon 1853 b. N. ampla Y Salzmanni, Planchon 18536 (flower). N. tropaeolifolia, Lehmann 1853 a (flower), fid. original specimen in hb. Berlin. Leaf is of N. rudgeana. N. nervosa, Lehmann 18530, from hb. Steudel, fid. Casp. and original specimen in hb. Berlin. N. ampla var. speciosa, Caspary 1878. N. Leiboldiana, Lehmann 1853 a, as to the American plant in hb. Vienna, fid. Caspary 1878. Speci- mens " ex herb. Lehm." in hbb. Kew and Berlin are from Africa, and appear to be N. copensis. N. gracilis, Zuccarini 1832 a, b, fid. original specimens (2 sheets) coll. Karwinski, Aug. 1827, "ex lacu Mexicano," in hb. Munich. Coll. Pringle, Plant. Mex. 1891, State of Jalisco, No. 3891. N. undulata, Lehmann 18530, fid. original specimens, coll. Galeotti, No. 4846 (4840 in Kew), Nov., 1840, " lacs pres Oaxaca, Mexico," ex hb. Lehmann in hbb. Berlin, Kew. Hemsley 1888. Geographic Distribution. — Tropical America from 190 N. to 230 S. Valley of Mexico (Karwinski) ; Oaxaca, Mexico (Galeotti) ; Guadalajara (Pringle). Caguas, Porto Rico, coll. P. Sintenis, No. 2577, Sept. 5, 1885. St. Domingo, coll. Poiteau, 1802, in hb. Delessert. Jamaica; Nevis; St. Croix; Lebanon in Antigua; Maracaibo; La Paila in New Granada; Merida; Demerara in British Guiana; Kaw River in French Guiana; Pernambuco, Santa Cruz, St. Christophe, Para, Rio de Janeiro, Cabo Frio, Capocabana, and Bahia in Brazil (Caspary 1878). Notes. — This is the form cultivated in Europe as N. ampla, sent to Kew from Jamaica by Macfadyen in 1847 or 1848, and to Caspary from Caracas by Dr. Ernst in 1869. Specimens from St. Christophe and Maracaibo have the lobes of the leaf widely divergent; those cultivated in Europe from Jamaica stock have the lobes touching; nearly all the others in collections have an open but very narrow sinus. The type specimens of N. ampla y Salzmanni Planch, and N. tropaeolifolia Lehm. consist each of a flower of this variety with a leaf of N. rudgeana Mey. (Casp. 1878). Var. pulchella (DC.) Casp. Flowers of medium to small size ; stamens 30 to 50 ; margin of leaves sub-entire or slightly repand-sinuate ; veins slender, scarcely prominent. N. pulchella, DC. 1821 b, fid. original specimens, without data, in hb. DeCandolle, but marked coll. Pavon, at Guayaquil (DC. ms.) in hbb. Paris and British Museum. Planchon 1853 b. Lehmann 18530. N. lineata, St. Hilaire 1833, fid. Caspary 1878, and original specimen in hb. Paris. N. ampla var. pulchella, Caspary 1878. Taxonomy — Nymphaea flavo-virens. 137 Geographic Distribution. — South America, from 2° to 230 S. Guaniva, Porto Rico, coll. Sintenis, No. 3662, det. Caspary, in hbb. Berlin, Kew, Munich. Guayaquil, Peru; Province of Rio de Janeiro (between Sitio de Paulista and Sitio do Porier), Brazil (Caspary 1878). Nymphaea flavo-virens Lehmann. (Plate VI; Fig. 54.) Leaves sub-orbicular, 30 to 45 cm. in diameter, narrowly peltate, deeply and irregularly sinuate or nearly entire, plain green on both sides; angle of lobes rounded. Flowers white, 10 to 15 cm. in diameter; sepals pure green, or with 2 or 3 scattered short black lines near the margins ; petals 16 to 20, narrow, acuminate; stamens about 60, deep yellow ; outermost filaments short, broad, petaloid ; anthers with a yellowish- white appendage. Nymphaea flavo-virens, Lehmann 1852 b, 1854, fid. specimens from the Hamburg Botanic Garden, 1864, in hb. Kew, and from hb. Lehmann in hb. Berlin. Hooker 1901, fid. original specimen in hb. Kew, cult, in Kew Garden. N. gracilis Hort. Watson 1887? G. & F. 18946. Tricker 1897. Conard 1901 a. Not Zuccarini 1832 a. N. mexicana Hort, not Zuccarini 1832, fid. Hooker, 1. c. Description. — Flowers 10 to 15 cm. in diameter, open from early morning to 5 or 6 p. m., sweet scented, resembling Convallaria majalis. — Peduncle terete, slender, glabrous, tapering upward, rising 20 to 30 cm. above water, with 6 small main air- canals and 12 secondary ones, all with thick partitions between. — Sepals 4, lanceolate, tapering to an acute apex, 8.3 cm. long by 2 to 2.4 cm. wide, green and obscurely nerved outside, sometimes with a few scattered short black lines ; inner surface white. — Petals 18 or 19, of firm texture, outermost opening nearly horizontally, about as long as the sepals, narrowly lanceolate, acute, breadth : length = 1 : 5.6 to 6.1 (=1.27 cm. : 7.8 cm. ; and 1.4 cm. : 8.3 cm.). Outermost petal greenish-white outside, with 2 or 3 red- dish lineolse. Inner petals smaller, pure white. — Stamens about 60 (56 to 61), yellow; outer ones longer, with short, broad, petaloid filaments, and appendages of medium length, stout, light yellow in color. — Carpels 12 to 15 ; styles short, stout, fleshy, orange- yellow, becoming elongated and incurved in fruit. — Fruit nearly globose, 2.2 cm. in diameter by 2.87 cm. high, crowned by the small circle of styles (1.9 cm. in diameter at base, outside measurement), and slightly depressed, basin-like stigmatic disc; axile process broadly conical with rounded apex, about twice as wide as high. Cells of berry oval, about 1.9 cm. high by 0.95 cm. wide radially. — Seed nearly spherical, with insignificant raphe, about 0.2 cm. in diameter. Floral Organs. M. ATI: UK Leaves. Organs. No. No. Length. Breadth. Sepals 4 18 60 IS 4 19 56 15 cm. 43-3 46.0 46.3 cm. 38.9 40.0 42.7 Petals Stamens Carpels '38 The Waterlilies. Submerged leaves from small tuber 2 or 3; first one ovate, acute, or oftener triangular-hastate with acute apex and angles ; second leaf much larger, angles slightly rounded, lobes long and narrow, widely diverging, thin and crisped or even twisted. First floating leaf ovate, abruptly acuminate, broadest opposite the insertion of the petiole, slightly crenulate; sinus deep, margins straight or slightly curved, slightly divergent or touching or overlapping, angles narrowly rounded; upper surface green with a few brown blotches, especially near the margin ; under surface green, blotched with brown. Leaf of mature plant narrowly peltate (1.2 to 1.9 cm.), orbicular-ovate, 30 to 45 cm. in diameter, deeply and irregularly sinuate-dentate, with obtuse or rounded teeth a- i Via. 54.— Symphaea flavo-viretu : a, sepal ; b, petal of outermost whorl ; c, petal of second whorl ; d, petal of third whorl ; e, outer stamen ; /, inner stamens ; g, vertical section of ovary. Natural size. 1.2 cm. long and 2 cm. apart, rarely nearly entire; pea-green above, paler below; veins prominent beneath, primary nerves 7 to 9 on each side; length of principal area: radius of leaf =s i: 1.6 to 1.75. — Sinus, depth: length of leaf = 1 : 2.5 to 2.8; margins more or less convex, divergent or overlapping; angles rounded. — Petiole terete, 1.2 cm. in diameter, 60 to 150 cm. long. Rhizome (tuber) erect, stout, 3.8 cm. long by 2.9 cm. thick to 21 cm. long by 10 cm. thick ( ! ), brownish-black, very rough with projecting leaf cushions, each capped by the dried base of the petiole and marked on the under side with root scars; upper CARNEGIE INSTITUTION OF WASHINGTON WATERLILIES, PLATE 6 NYMPH/EA FLAVOVIRENS. HEUOTYPE CO., BOSTON. Taxonomy — Nymphaea flavo-virens. 139 quarter of tuber clothed with long light brown hairs, above which the stout soft leaf- rudiments may project visibly even in the dried state. Phyllotaxy of a high order, but the leaf scars are 1.2 to 1.9 cm. apart, and by their dry projecting habit give the tuber somewhat the appearance of a rough pine-cone. Measurements of Leaves from Small Tuber, in Centimeters. Leaf. Length of Midrib. Length from Apex to Angle of Lobes. Diameter of Leaf. Width of Sinus. Depth of Sinus. First submerged 1.6 2-5 30 5.6 7.3 5-4 3-0 3-0 2.2 4.0 5-6 1.6 4.0 3-35 3-0 4.8 1.6 4.0 0.6 0-5 1-25 2.ZC 1.0 2.2 0.8 3-35 0.6 2.2 0.8 2.2 1.0 2.2 3-5 2.9 1.K w «/ Geographic Distribution. — Uncertain. Notes. — This species has been known for ten years in American gardens as N. gracilis. An examination of Zuccarini's types shows this name to be mistaken, while the plants agree well with the types of N. flavo-virens both of Lehmann and Hooker. Concerning its original home there is some doubt. Lehmann's type in hb. Kew is marked " Hort. Bot. Hamburg 1864, ex hb. Reichenbach fil." The plant figured in the Botanical Magazine was grown at Kew from stock received in 1892 from W. N. Pike, Floral Park, N. Y., under the name of AT. mexicana. The present stock in Amer- ican gardens was collected by Mr. C. G. Pringle and introduced by Mr. E. D. Sturte- vant, then of Bordentown, N. J., previous to 1894 (G. & F. 1894 b). Of Mr. Pringle's Plantae Mexicanae, No. 3891 is labeled "Nymphaea gracilis Zucc," and agrees well with the types. Concerning the herbarium specimens and the living plants, how- ever, Mr. Pringle writes as follows under date of September 18, 1904 (received while the present chapter was in press) : " No. 3891 of my Plantae Mexicanae was collected in an artificial pond near the station of El Castillo some fifteen miles east of Guadala- jara ; and from the same station I gathered, at the end of the season, when they were lying dry on the surface of the soil and resembling pine cones (the water of the pond had then been drawn off) the tubers which I suppose were supplied to Dr. Sturtevant through my associate, Mr. F. H. Horsford, of Charlotte, Vermont. I am sure of these facts." In G. & F. 3:415 (1890) Mr. Pringle writes: "Nymphaea gracilis I only know in the region drained by the Lerma, yet this is a wide region, extending from Aguas Calientes and San Luis Potosi on the north to near the City of Mexico on the east and south and to beyond Guadalajara on the west." Is N. flavo-virens Lehm. merely a cultivated form of iV. gracilis Zucc. (=N. ampla DC.) ? This is a robust species, and very resistant in the resting state. It hybridizes very readily with pollen of N. zanzibariensis, from which cross beautiful garden varieties have arisen and been named in several places, e. g. N. Astraea, N. Greyae, N. mauvii, N. Wm. Stone, N. Mrs. C. W. Ward, N. gracilis azurea, N. g. purpurea, N. g. rosea perfecta, N. g. rubra. 140 The Waterlilies. Nymphaea stellata Willdenow. Leaves elliptic-orbicular, rather broadly peltate ; margin irregularly sinuate; lobes obtuse; upper surface green; pink or blue-violet beneath. Flower 5 to 12 cm. across, pale blue (varying to pink or white), open from 8 a. m. to 2 p. m. Buds ovate; sepals with minute blackish dots. Petals II to 14, dull white at base ; stamens 33 to 54 ; appendage blue, anthers and filaments pale yellowish. Nymphaea stellata, Willdenow 1797. Andrews tab. 330. Aiton 181 1. DeCandolle 1821b, 1824. Wight & Arnott 1834. Planchon 1853 6. Rev. Hortic, 1854 a, 1855 b, col. plate. Caspary 1865 in part ; 1877. Conard 1901 a. Castalia stellaris, Salisbury 1806 a. N. malabarica, Potret, in Lamarck 1802, fid. original specimen coll. Sonnerat, Isle de France, from hb. Lamarck, in hb. Paris. N. stellata var. parvifiora, Hooker & Thomson, 1855. Description. — Flowers small, 5 to 13 cm. across, open on three successive days from 8 a. m. to 2. p. m., nearly odorless. — Bud oblong, ovate, rounded at apex. — Peduncle terete, slender, rising well above the water-surface. — Sepals 4, oblong, sub- ovate-triangular, breadth : length = 1 : 2.8 to 3.5, margins nearly straight toward the apex; outer surface light green, uniform or marked with minute dark purplish lines; inner surface bluish-white. — Petals 10 to 16, lanceolate, acute, shorter than the sepals, pale blue above, yellowish-white at base. — Stamens 33 to 54, with small blue appen- dage; anthers and filaments pale yellow. — Carpels 10 to 17; styles short-triangular, usually apiculate, obtuse; breadth : length = 1 : 1 to 1.5; stigmatic ray terminating on the style in a short obtuse triangle, breadth : length = 1 : 0.5 to 0.7. Stigma slightly con- cave.— Seed elliptic globose, diameter : length = 0.8 : 0.9 mm. ; = 0.8 : 1.0 mm., gray. Leaf of adult plant rather broadly peltate ; breadth of pelta : depth of sinus = 1 : 6 to 10; outline sub-orbicular-elliptic, 12 to 15 cm. in diameter, irregularly sinuate- repand, or subentire, rounded or slightly retuse at apex; upper surface bright green, sometimes faintly brownish blotched; under surface purplish to deep blue-violet, with green veins. Primary veins 9 to 11 on each side of leaf; length of principal area: radius of leaf =1:1.4 to 1.6. — Sinus deep, usually open; margins nearly straight; angles slightly produced, acute or obtuse. — Rhizome ovoid. Geographic Distribution. — Southern and southeastern Asia and neighboring is- lands, Java, Borneo, and the Philippines. Malabar, Tanjore in India (Hainan, coll. C. M. & A. Henry, No. 8382). Var. cyanea (Roxburgh) Hooker & Thomson. Flowers of medium size, blue, with faint odor or none. Nymphaea cyanea, Roxburgh 1824. Wight & Arnott 1834. N. stellata /?, Sims 1819. Planchon 18536. Var. versicolor (Roxburgh) Hooker & Thomson. (Plate VII.) Flowers of medium size, pink ; stamens numerous. Leaves sinuate, pink beneath, marked above and below with pustules 1 to 2 mm. in diameter. Nymphaea versicolor, Roxburgh, in Sims 1809. Roxburgh 1814, 1824. Planchon 18536. N. punctata, Edgeworth 1845 = N. Edgeworthii Lehmann 18526, fid. original specimen, coll. M. P. Edgeworth, 1844, Sadhoura, India, in hb. Kew. Not N. punctata Kar. & Kir. 1841 ? CARNEGIE INSTITUTION OF WASHINGTON WATERLILIES, PLATE 7 iTo NYMPH/tA STELLATA VERSICOLOR. . Sepal. >,-■ )' ; whorls, f, Stamens of dosed flower. >,- Closed Hover, HEUOTYPE CO., BOSTON. Taxonomy — Nymphaea caerulea. 141 N. rhodantha, Lehmann 18530, fid. specimen coll. dimming, No. 702, "in insulis Philippin.," from hb. Lehm. marked in the handwriting of Lehmann, in hb. Berlin. N. Hookeriana, Lehmann 1853 a, fid. original specimen marked in handwriting of Lehmann, coll. Hooker & Thomson, Dec. 19, 1850, at Noapolly, Bengal, from hb. Lehm. in hbb. Berlin, Kew. N. bella, Lehmann 1853 a, as to the flower ; leaf is of N. pubescens, fid. original specimen from hb. Lehm. in hb. Berlin. Notes. — Hooker & Thomson record, and the herbaria bear them out, that no sharp line could be drawn between the varieties of this species as it occurs in India, but they were certainly wrong in including the African plants. The three extreme Indian forms are themselves fairly recognizable. The margin of the leaf may be al- most entire. The pustules so evident on the leaf of our N. versicolor cultivated from seeds sent by Mr. Gollan of Saharanpur, India, may be present at times in the other varieties. Our plant also had a great number of bulbils on the caudex, as shown in several herbarium specimens both of blue and pink forms. Willdenow's description was based solely on the figure and text of Van Rheede's Flora Malabarica 1 1 : 53, tab. 27, and was very brief : " N. foliis integerrimis, lobis divaricatis, acutis, calyce tetraphyllo petalis acutis longiore." Andrews figured the plant in color and later Sims gave a rather larger form than that of Andrews. Planchon figures a plant evidently identical with that of Andrews and Van Rheede, and Caspary (1877) places its description clearly in parallel columns with the allied forms N. caerulea, capensis, and zanzibariensis. Of these it comes much nearest to N. caerulea, but is plainly distinct. I have been unable to find it in American gardens. The flowers of this species, according to Caspary (1877), are regularly self-polli- nated, often while still in the bud ; it refused to cross with N. caerulea under Caspary's management. Rose-colored and white varieties of N. stellata sent from Canara to Caspary by the missionary Zeigler in 1869 were long cultivated in the botanic garden at Konigsberg. Nymphaea caerulea Savigny. (Plate VIII; Fig. 55.) Leaves entire or slightly wavy at base, orbicular or ovate-orbicular, narrowly peltate, becoming 30 to 40 cm. in diameter; green above; under surface green with small dark purple spots, purplish all round near margin. Flowers 7 to 15 cm. across, open from 7.30 a. m. to 12 m. Buds conical. Sepals thickly marked with black lines and dots. Petals 14 to 20, lanceolate, light blue above, lower half dull white. Stamens 50 to 73; appendage long (0.5 cm. on outermost stamens), slender, pale blue. Nymphaea caerulea, Savigny 1802, fid. original specimens coll. Delile, 1798, near Cairo. Ventenat 1803. Delile 1813. DeCandolle 1821 b. Planchon 1851 a, 1853 b. Caspary 1877. Wilde- man & Dur. 1900. Conard 1901 a. N. stellata, Caspary 1865 in part, 1873. Boissier 1867, 1888. Oliver 1868; and of many authors and collectors. Not Willdenow 1797. N. maculate, Schum. & Thonn. 1829, fid. original specimens from Copenhagen, kindly loaned by the curator of the herbarium. Welwitsch ms. in hb. Paris. ■/V. poecila, Lehmann 1853 a, fid. original specimens from hb. Lehmann seen in hb. Berlin. N. scutifolia, Tricker 1897. Not Castalia scutifolia, Salisbury 1806 a. N. discolor, Lehmann 1853 a = JV. caerulea, Kotschy MS. (see note). Castalia stellaris, Salisbury 18066 (in part). C. caerulea, Tratinnick 1822. 142 The Waterlilies. Description. — Flowers 7 to 17 cm. across, light blue, with a faint and character- istic sweet odor, open on 3 days from 7.30 a. m. to 12 m. (4 days, 8 a. m. to 2 p. m. Casp. 1877). — Bud distinctly conical, with straight sides, obtuse, abruptly contracted below to the receptacle, 7 cm. long by 2.5 cm. in diameter at base. — Receptacle nearly three times as wide as peduncle, spreading out at about 450 from the vertical, pale purplish green. — Peduncle terete, 0.6 to 0.8 cm. in diameter, 18 to 38 cm. long, most slender about 1.25 cm. below flower, smooth, dull brownish-green; main air-canals 6, surrounded by a circle of 12 smaller ones. — Sepals 4, broadly lanceolate, breadth : length = 1 : 2.8 to 4.7, the sides nearly straight in upper two-thirds, somewhat converging below to the broad base; apex rounded. Outer surface dark green, thickly flecked with purplish-black dots and lines most dense near margins of sepal and midway of length ; the green ground-color assumes a yellowish tint near base of sepal. Seven u e Fig. SS.—Nymphaea eaerulta: a, dosed flower; b, sepal; e, d, e, outer, median, and inner petals; /, h, stamens ; (/, carpellary style viewed en face ; (, vertical section of ovary ; k, vertical section of fruit. Reduced about one-half. longitudinal veins can be seen by looking through the sepal at the light. Inner surface dull white, tinged blue on midline near apex, greenish and veiny and semi-translucent in lowest three-quarters of length. — Petals 12 to 20, of moderately firm texture, open- ing about 300 above horizontal, about as long as the sepals. They stand in 3 series: first, a whorl of 4, alternate with the sepals ; second, a whorl of 8, one on each side of each outer petal ; third, a whorl of 8 alternate with the second, but usually imperfect, consisting of only 2 to 5 petals on one side of the flower, rendering it so far unsym- metrical. Outermost petal narrowly elliptic-lanceolate, obtuse, with rather broad in- sertion (0.8 cm. in Flower No. 3, see table of measurements), not at all concave; outer surface colored on lower half and on the midline dark green with black spots exactly like the sepals ; one-third of width at margin on each side, above, colored pale blue ; inner surface white on lower half, becoming translucent and 7-veined at base, shading above to a pale sky-blue at apex and on upper margins. Petals of second whorl shorter than those of the first, rhombic-lanceolate, narrowed at apex and base, 5-veined CARNEGIE INSTITUTION OF WASHINGTON. WATERLILIES, PLATE 8. If^ 'fi ~" j t * • # .• > .« NYMPH/EA CCERULEA. Taxonomy — Nymphaea caerulea. 143 below, white on the lower third, pale blue above, pure sky-blue at apex, the blue and white shading together insensibly, the latter predominating in two-thirds of length on inner surface of petal. Petal of third (innermost) whorl linear-lanceolate, obtuse, blue at apex; outer surface with white margins in lower three-fourths of length, the blue continuing half way down the midline ; on inner surface, white prevails on lower five- sevenths of length. — Stamens 50 to "J2» sometimes appearing inserted without order, sometimes indistinctly spiral, often in about 16 vertical ranks of 3 to 5 stamens each. Outerm.~,t stamens with short and rather broad filaments and long acuminate anther with two long, nearly parallel anther cells and proportionately long, acute, 3-angled ap- pendage, one angle pointing inward ; blue at tip and half-way down, yellowish-white below. Median stamen has short, broadly elliptic filament, very long parallel anther cells, and short appendage; anther and filament bright yellow, only the appendage blue. Innermost stamen bright yellow all over, short, stout ; anther broader and longer than filament; no appendage. — Ovary nearly hemispherical (e. g., 1.9 cm. in diameter by 0.95 cm. high ; 0.4 cm. by 0.24 cm.). — Carpels 14 to 21 ; styles short (0.3 cm. long by 0.08 cm. in diameter), fleshy, nearly erect, subacute, concave inward, all alike, pure yellow, papillose up to about 0.08 cm. from apex, the papillose area ending roundly. Stigma nearly flat, bending upward on the styles, yellow. Axile process rising ab- ruptly from the stigma, broader than high (height: breadth = 1 : 1.5; = 0.24 cm. : 0.38 cm.) and somewhat conical in shape, with obtuse tip, of a whitish color. — Fruit large, round (4.5 to 6.4 cm. in diameter by 2.5 to 3.8 cm. high), truncate above, with deep radiating fissures between the carpels, flattened or even excavated beneath around the peduncle, of a pale green color, becoming translucent and brownish, crowned with the hard, slightly enlarged styles, and surrounded by the sepals and outer 4 petals, all of which are dark green and spotted as in the flower. The peduncle makes a large, rude spiral turn, holding the fruit still erect but with its base nearly or quite resting on the earth. — Seed ellipsoidal, acuminate at the hilum, 0.17 cm. long by 0.12 cm. in diameter, dull olive brown ; surface marked with about 14 interrupted longitudinal lines of minute hairs ; raphe evident, not prominent. Aril longer than the seed. Germination may take place immediately on ripening, or the seeds will withstand drying over winter. First leaf of seedling filiform, 1.6 to 1.9 cm. long. Second leaf broadly elliptic on a petiole about 2.5 cm. long; blade about 1.27 cm. long by 0.6 cm. wide. Third leaf broader and shorter. Fourth leaf deltoid-ovate with rounded angles and apex, and truncate base; lamina 1.6 cm. long by 1.27 cm. wide at base; later leaves finally deltoid-cordate, with rounded apex and angles, and rather deep sinus. First floating leaves nearly orbicular, 2.5 to 3.8 cm. in diameter, entire ; sinus deep with straight and touching margins ; under surface marked with irregular blackish spots. First leaves from tuber submerged, 2 or 3 on each shoot; the first one may have a minutely denticulate margin, 1 cm. long by 0.3 cm. wide ; or, more commonly, the lam- ina is entire, broadly deltoid, 3.2 cm. long by 3.2 cm. wide at base, with rounded angles, and petiole 2.5 to 3.8 cm. long. A third one may be produced, of oval shape, fissi cordate, with deep open sinus and rounded apex, 3.8 cm. long by 2.9 cm. wide. First floating leaves from tuber like those of seedling, but larger ; perfectly entire. Leaf of mature plant very narrowly peltate (1:14 to 16, Casp. 1877), ovate- orbicular to orbicular, 30 to 40 cm. in diameter (33 cm. long by 30.6 cm. wide), soft, 144 The Waterlilies. thin and quickly withering, slightly wavy or sinuate in the basal half, apex slightly emarginate, dark green and obscurely veiny above, color darker over insertion of peti- ole. Under surface paler green, with numerous small dark purple blotches, which are larger near the midrib than at the periphery ; margins purplish, shading from dark red- purple at the extreme edge to green, the transition taking place in the course of about i cm. near sinus angles, in about 2.5 cm. at apex of leaf. Veins prominent out to the fifth grade ; primary nerves 6 to 10 on each side of leaf. Petiole attached by quite a strong " collar." Length of principal area: radius of leaf = 1:1.3 to 1.7. — Sinus usu- ally closed; depth: length of leaf = 1 : 2.7 (= 13.5 cm. :3s. 7 cm.); margins doubly curved, convex and overlapping above, separating about 2.5 cm. from periphery of leaf and becoming parallel ; angles subacute, slightly produced, about 0.6 cm. apart. — Petiole terete, or flattened near the leaf, about 0.6 cm. in diameter, dull brownish-green ; with 2 larger upper air-canals, and 2 smaller ones below, and a ring of 12 still smaller ones outside. — Stipules smooth, fused at base, short, distant above, acuminate, very acute; 1.9 to 2.2 cm. long, 1.3 cm. wide at base; free portion 1.1 to 1.6 cm. long, 0.5 cm. wide ; tips 0.9 to 1.2 cm. apart. Rhizome (tuber) erect, thick, ovoid, 2.5 to 7.5 cm. long by 2 to 6.4 cm. in diam- eter, blackish, apex protected by leaf-bases and long fine hairs. Phyllotaxy of a very high order, the leaf-bases completely covering the larger rhizomes, leaving only about 0.3 cm. around each one. Roots springing from the bases of the petioles. Offsets formed only when the tuber breaks out from the dried state; 2 or 3 shoots may come from one tuber; the stolon is terminal, very short, the point of origin of new leaves remaining within the coating of long hairs. Number of Floral Organs. Ideal. Average. Sepals Petals Totals 4 17 6l 17 4 14 SO 15 4 16 16 4 20 57 20 4 16 69 17 4 16 73 21 4 14 52 IS 4 20 4 16+ 60 17+ 99 83 94 101 104 114 85 104 97+ Measurements of Sepals. Flower. Length. Breadth. cm. cm. Flower 1 6.8 1.9 6.2 1 -75 (8.6 2.6 ^8.8 2.9 1.8.8 2.7 Taxonomy — Nymphaea caerulea. 145 Measurements of Petals (in centimeters) . Flower I. Flower 1. Flower 8. Flower 5. Outermost Whorl. Second Whorl. Innermost (Im- perfect) Whorl. Length. Breadth. Length. Breadth. Length. Breadth. Length. Breadth. Length. Breadth. Length. Breadth. 6.7 i-4 6.S OI.4 8.45 2.0 7-8 2.9 7-3 1-75 5-9 01.4 6.4 ••35 6.3 MI.3 8.45 1.8 8.0 2.8 7-3 1-5 5-6 in 1.4 6.2 1.2 8.3 Z2.0 8.0 2.8 7.2 1-75 5-4 MI. I 5-7 0.8 8.45 1.9 7-8 2.8 7.0 1-3 5-26 z'0.8 5-6 0.7 7-65 7-65 7-65 7-65 2.0 2.0 2.0 1.9 4-95 10.7 1= Inner. m = median. 0 = outer. Measurements of Stamens (in centimeters). Flower. Total Length. Breadth of Filament. Length of Anther Cells. Breadth of Anther. Length of Appendage. Flower 5. ... • 3-75 2.7 i-4 0.88 3-7 2.7 0.88 3-5 2.1 1. 12 O.56 0.4 0.16 0.13 O.56 O.48 0.13 0.48 O.32 O.16 1.6 1.6 0.8 O.48 1.6 1.6 O.48 1.6 1.42 0.72 O.32 O.24 O.16 O.16 O.32 0.24 O.16 O.32 O.24 0.21 0-5 0.3 O.08 0 O.56 O.24 0 0.48 0.16 0.04 Geographic Distribution. — Northern and Central Africa. Cairo, Savigny 1802. Damietta and Rosetta, coll. Schweinfurth. Guinea, Schum. & Thonn. Nubia, Kotschy (see note). Sona Gungu (Distr. of Cataracts), Congo, Wild. & Dur., 1900. Angola, Welwitsch. Omkenen, Kordofan, coll. Dr. Hendur, No. 890, in hb. Berlin. Cyre- naica, coll. M. Pacho, 1826, in hb. Paris. Var. albiflora, Caspary. Flowers white ; sepals without black spots. Nymphaea caerulea var. albiflora, Caspary 1877. N. abbreviata, Guill. & Perr. 1830. N. ampla, Kotschy MS. =N. nubica, Lehmann (see Note). Not N. ampla, DC. 1821 b. Geographic Distribution. — With the type. Rare. Damietta, Egypt, coll. Sieber, fid. Caspary 1877. Egypt, coll. Schweinfurth, 1885, in hb. British Museum. Central Africa, coll. Schweinfurth, No. 2329, in hb. Berlin. Kounoun, Cape Verde, coll. Per- rottet, 1 83 1, in hb. Delessert. Note. — The name of this species has had a varied application, as is shown by the synonymy of this and the preceding and following plants. The cause lies partly in 11 146 The Waterlilies. the close resemblance of N. caerulea and N. stellata, and partly in the fact that these and N. capensis were introduced into cultivation near the same time, and the descriptive character of the name caerulea made it equally applicable to all. The geographical prox- imity of this and N. capensis may have added to the confusion. Indeed, examination of herbarium material shows that N. caerulea is connected by intermediate forms with N. micrantha on the west and N. capensis on the south. We have not been able to secure living material of N. micrantha, but the capensis type is very easily distin- guished from the present species. Their hybrid (N. pulcherrima Tricker) is entirely sterile. Other related forms are discussed in their proper places. Many botanists have included all the blue waterlilies of Africa and Asia under the name of N. stellata, as did Hooker & Thorns. (1855) ; Caspary, ten years later, separated off N. capensis, and in 1877 gave exact descriptions of four species in parallel columns; he stated that his old arrangement (1865) needed complete revision, but unfortunately this work was left undone. The Kotschy collection from Nubia has long been a stumbling block to students of Nymphaea. They are marked No. 166 N. caerulea Sav. and No. 167 N. ampla DC. — " Kotschyi Iter Nubicum. In stagnibus pluvialibus ad radices eurontas montis Cordofani, Arasch Cool, Jan. 5, 1841, d. Oct. 12, 1839." The first is represented in hbb. Kew, Paris, Boiss., Deless., and Berlin. It was marked by Caspary "N. capensis Thunb." The Boissier specimen has narrower petals and fewer stamens than the type of capensis; the Berlin specimens differ between themselves — one flower has spotted sepals, and the leaf gently sinuate; another has leaves typical of N. caerulea but no spots on the sepal. No. 167 was marked "N. caerulea Sav." by Caspary in hb. Boiss., with the addition of " var. albida " in hb. Paris. The leaf is gently sinuate in the Kew, Deiessert, and one Paris specimen. It is also represented in hbb. Berlin and Boissier. The leaf may or may not have black spots and some purple coloration beneath. The flower resembles that of N. caerulea, but was evidently white, and in the Kew specimen, at least, the outer stamen is provided with an unusually long, lanceolate ap- pendage. The two plants seem strikingly intermediate between N. capensis and N. caerulea. No. 167 is the type of Lehmann's N. nubica. Self-fertilization occurs regularly in this species. Reference to the importance of N. caerulea as the Blue Lotus, the lotus par excel- lence of Egypt, has been made in Chapter I. Nymphaea micrantha Guillemin & Perrottet. (Plate IX.) Leaves small, round-cordate, peltate, entire, petiolate ; petioles very long, slender ; lobes divaricate, acuminate ; under surface reddish, marked with violet-black dots ; upper surface glabrous, pale green, in mature plants bulbiferous at the summit of the petiole ; stigmatic rays 15. N. micrantha, Guillemin & Perrottet 1830. Hooker 1850. Paxton 18530. N. caerulea, Guillemin & Perrottet 1830. Not Savigny 1802. N. caerulea var. albida, Guillemin & Perrottet 1830 = N. micrantha, Hook. N. rufescens, Guillemin & Perrottet 1830. N. stellata var. bulbillifera, Planchon 1852 c. Not N. stellata, Willdenow 1797. N. vivipara, Lehmann 1852 a, b, 1853 a, 1854. N. guineensis, Schumacher & Thonning 1829, fid. Chifflot 1902. CARNEGIE INSTITUTION OF WASHINGTON WATERLILIES, PLATE 9 4V NYMPH/EA MICRANTHA. i , Under aide of leaf ; 2. Upper side with young plant developing, t, Caudex, with a branch. |, Scar* nf leave* and of a flower, from aarface of a caudex, enlarged : 5, Cross section of caudex ; 6, Longitudinal section of a caudex. 7. Superficial ftagential section through m* of fig. 4; /,. base <>t ;i petiole; '. base «»f .1 peduncle; B, Section a tittle deeper than 7. o, tame ;t^ S, a little deeper. HELIOTYPE CO., BOSTON Taxonomy — Nymphaea heudelotii. 147 Description. — Sepals 5, narrow, lanceolate, acuminate, with violet dots. — Petals 10, 1.3 to 5 cm. long, a little shorter than the sepals, lanceolate, narrow, blue to white. — Stamens numerous; filaments whitish, narrow at base; anthers of various lengths, twice as long as the filaments; appendage short, conic-terete. — Ovary round. Seeds very numerous, very small, blackish. — Leaves 7.6 cm. long, 5.1 cm. wide (in the type). — Rhizomes clustered, with soft tawny wool around the bases of the petioles ; roots very long. Geographic Distribution. — West Coast of Africa. In swamps of the penin- sula of Cape Verde, near n'Batal ; in the Galam region ; in the kingdom of Cayor, etc. Flowers almost all the year. Types. — Coll. Perrottet, 1830, in hbb. Delessert, Kew, DeCandolle. Also coll. E. Holub, Eastern Bamangwato territory, Bechuanaland. Notes. — The most striking difference between this and N. caerulea Sav. lies in the viviparous habit. Chifnot (1902) states, and Planchon's description indicates, that this character does not appear until the second year of the plant's growth. The first year it has nearly white flowers, and corresponds to N. guineensis Sch. & Th. ; the second year bulbils appear, but the flowers are small and pale, and the plant corre- sponds to N. micrantha Guill. & Perr. Only in the third year, according to ChifBot, does the plant acquire its definite morphological characters. We do not find either in the specimens or descriptions of Guillemin & Perrottet any important differences between the several viviparous forms they enumerate. N. rufescens is a larger plant; sepal 7.6 cm. long, petals 18 to 20, stamens very numerous with petaloid filament; carpels 18 to 20; leaves 12 to 25 cm. in diameter, more or less spotted beneath, primary veins 12 to 20, prominent; coll. Perrottet, 1830, Senegal, seen in hbb. Delessert, Boissier, DeCandolle ; also coll. Roger, Dec, 1823, Senegambia, in hb. Kew (with the note " Les graines servent de nourriture aux negres ") ; also coll. Buchholz, Eliva Sonanga, Cape Lopez, West Africa, in hb. Berlin. N. caerulea Guill. & Perr. has " leaves strongly peltate, impunctate, red beneath, lobes long, free, acute " ; types coll. Perrottet, 183 1, Senegal, in hbb. Delessert, Kew; also coll. Leprieur, Sept., 1821, " in paludosis prope Ngalete, regio Cago, Senegambia," in hb. Delessert; also coll. C. Barter, Baikie's Niger Expedition, No. 7064, 1858, at Nupe, in hb. Paris. N. caerulea var. albida Guill. & Perr. has larger flowers than the type, with narrow lanceo- late petals, nearly white ; types coll. Perrottet, Walo, Senegambia, in hb. Delessert ; also coll. Leprieur in same district in hb. Delessert. This is apparently identical with N. micrantha Hook. Caspary's " N. micrantha Hook." (notes and herbarium, now in hb. Berlin) has an orbicular leaf with wide sinus and obtuse lobes, with or without black spots beneath; petals ovate, white, 3.8 to 4.5 cm. long. This plant was much used by Caspary in hybridization, giving well-marked intermediates with the closely allied iV. caerulea Sav. (seen in hbb. Kew, British Museum, Berlin). Nymphaea heudelotii Planchon. (Fig. 56.) Flowers small, 2.8 to 5.1 cm. in diameter. Sepals lance-ovate, with black spots outside. Petals 5 to 10, lanceolate, acute, bluish-white. Stamens 11 to 16; anthers much longer than the filaments. Carpels 4 to 10. Leaves almost or quite entire, ovate i4» The Waterlilies. or orbicular, cleft nearly to the petiole, 2.5 to 8.2 cm. long, reddish beneath, with black specks. Seed smooth. N. Heudelotii, Planchon 1853 b, according to the original specimen. N. pseudo-pygmaca, Lehmann 1853 a, according to the original specimen in hb. Kew, which con- sists of a submerged leaf, with flower and fruit, much dilapidated, coll. Heudelot and communicated by hb. Paris. Description. — Planchon gives the following : Dwarf, glabrous ; floating leaves cordate-suborbiculate, very narrowly peltate, with the sinus narrow and nearly closed Via. S6.—Nymphaea heudelotii ■ (a), (b), leaves ; (c), flower ; 1, 2, 8, 4, sepals. From a drawing of the types in the notes of B. Caspary, now In hb. Berlin, by kind permission of the Directors. Natural size. (in the membranous submerged leaves wide open) ; margin slightly repand ; under sur- face purplish, violet spotted. Flowers small ; sepals ovate at base, becoming rather long attenuate, with violet lines outside ; petals 5 to 8, bluish, appendiculate, anthers five to six times longer than the rather broader filaments; rays of the stigma 8 to 10, short, triangular, acute. Fruit globose, shorter than the calyx; seed elliptic-globose, smooth (not costate). Geographic Distribution. — Senegambia, in shallow flowing streams ; Fouta Dhiallon. Types. — Coll. Heudelot, No. 844, in 1837, seen in hbb. Delessert, Berlin, and Kew. In hb. Kew I saw another specimen consisting of a half-closed flower and two frag- Taxonomy — Nymphaea heudelotii. 149 mentary leaves, coll. Sierra Leone Boundary Commission, Dec. 22, 1891, in "small streams, Luseniga, Samu." In hb. Berlin is a specimen, coll. Alex. V. Mechow's Exped., No. 480, Jan., 1880, at Malange, West Africa, marked N. stellata; also coll. Pogge, No. 559, Jan. 2, " auf der Ueberschwemmung ausgesetzten Stellen am Lulua," and No. 558, Nov. 31, 1881, " Wassertumpel in der Campine am Lulua"; also. coll. Schweinfurth, No. 2329, Oct. 30, 1869, " Land der Djue : Gr. Scriba Ghattes in der Sumpfsteppe im Wasser; Bl. reinweiss." Var. na^a new var. (Fig. 57.) Sepals 4, lanceolate, obtuse, 1.2 to 2 cm. long by about 0.32 cm. wide, green out- side, with small and distinct or large and diffuse black spots; nearly white within. Petals 7 to 10, lanceolate, acute, shorter than the sepals, 1.5 to 1.7 cm. long, 4 mm. wide, white. Stamens 11 to 14, yellow, with very short appendage, all nearly alike and about half as long as the petals, apparently in two series. Stigma of 5 to 6 broad carpels, with short, broad styles. Fruit nearly globose, surrounded by all of the floral parts. Seed gray-green, ellipsoid, 1. 3 mm. long, 0.88 mm. in diameter. Fig. 57.— Nymphaea heudelotii nana: a, open flower, 1, 2, 8, 4, sepals ; b, under side of leaf; c, flower pressed on one side showing outside of sepals; 4o 11.48 12.75 4.6 3.8 2.55 3-5 1.2 9.9 10.2 10.2 11.6 7-65 Depth of Sinus. cm. 5-6 5-7 5-75 5-1 4-5 3-5 4-15 4-5 2.07 1.9 1-3 1.6 o Length Midrib. Length of Prin- cipal Area. cm 5-1 6 Length Radius. cm. 7.0 7-8 7-65 6.4 5-6 Pelta. cm. 0-3 0.32 0.32 0-3 0.3 0.48 0 0.64 0.8 Number of Veins. 10 8 Taxonomy — Nymphaea tetragona. 167 Geographic Distribution. — Florida, Texas, and Mexico. "Lagoons of the lower Rio Grande southward to the Valley of Mexico, and westward to the valley of the Lerma, State of Jalisco" (Pringle 1890). Lake Xochimilcho, Valley of Mexico, coll. J. W. Harshberger. Brownsville, Texas, coll. Pringle, Flora of Texas, 1888, No. 1956. In Florida : St. Johns River, from Jacksonville southward for 40 miles (Treat, 1877, fid. original specimens) ; Curtis, North American Plants, No. 101 a and Second Distribution No. 4815. Miami River; Prairie Creek (both in hb. Acad. Nat. Sci. Phila.). Little River, coll. Dr. J. M. Macfarlane, cult, in Botanic Garden, Univ. of Penns) 'r-vania. Reservoir Pond, Palm Beach, coll. Mr. Thomas. Also coll. Rugel, " ad fl. Alachua, Florida or." July, 1848, in hb. British Museum. Notes. — Zuccarini's description of N. mexicana ascribes white flowers to it, whereas the leaf agrees with that given above. Examination of the types leaves no doubt as to the identity of the plants now so familiar in cultivation ; the white color must have been purely a guess, as Professor Radlkofer suggested to me. The dried specimens are brown like the average herbarium specimens of the European waterlily. Our conclusion to unite the Florida and the Mexican plants comes only after pro- longed consideration and comparison, and it is to be regretted that the very suggestive name N. fiava has to be abandoned. It has seemed to us that the leaves of the Mexi- can plant are thicker and rounder, with less of purple coloration beneath. The floral leaves also are a little broader than those of the Florida plants. A fruit on Pringle's No. 1956 in hb. Acad. Nat. Sci. Phila. seems to have been nearly spherical, while the fruits of the Florida plants in our garden were ovoid. We have not been able to secure living plants direct from any Mexican locality. The Prairie Creek specimens, collected by Garber in 1878, have leaves more nearly orbicular than those from other Florida localities ; this is probably because they grew in shallow water. We see in cul- tivation like variations from this cause. The name N. fiava Leitner rested on the floral decorations and name on Audu- bon's plate of the great white swan; the leaves accompanying the yellow flowers are evidently those of a Nuphar. Excepting Rugel's unrecognized specimen, the species was known only by that plate until Dr. Edw. Palmer in 1874 sent a dried specimen to Har- vard, with a statement of its yellow color. Mrs. Treat, however, in 1876, again dis- covered the plant, and made it known by sending ample material both fresh and dried to Harvard ; from there it was sent to Kew, where it flowered first in July, 1882. B. CHAMAENYMPHAEA Planchon 1853 ft. Flowers white or rosy, opening near noon and closing in the late afternoon. Innermost stamens first to dehisce. Pollen aculeate or tuberculate. Seed of medium size. Leaves entire, obovate to oval. Principal air-canals in peduncle 4, in petiole 2; idioblasts stellate only. Rhizome erect, short, entirely without offsets (at least in N. tetragona) . Native of the North Temperate Zone, especially in China and Japan. Nymphaea tetragona Georgi. (Plate XIV; Fig. 65.) Flower small, 2.5 to 5 cm. across, white, with a small number of floral parts ; re- ceptacle distinctly tetragonal. Leaves ovate, small (2.5 to 12 cm. long), green above with brown blotches when young, under surface dull red; sinus open, lobes acute. Petiole with 2 main air-canals. Rhizome erect, covered with projecting leaf-scars. 1 68 The Waterlilies. Nymphaea Utragono, Georgi 1775, Ad. original specimen from hb. Pallas, in hb. British Museum; :8co. Caspary 1865. Robinson 1896. Conard 1901 a. N. tetrogonanthos, Pallas MS., in Sims 1813, Ad. specimen cited above. N. pygmaea, Alton 181 1. Sims 1813. Smith 1819. DeCandolle 182 1 b, 1824. Sprengel 1825. Don 1831. Ledebour 1842. Turczaninow 1842. Planchon 1852 b; 18530 and b. Lehmann 1853 a. Hooker & Thomson 1855. Loudon 1855. Hooker 1872. Garden 1883 b. Tricker 1897. Moenkemyer 1897. N. foliis cordatis integerrimis and N. alba minor, Gmelin 1769. N. pygmaea alba, Hort, Gerard 1890 a? Castalia pygmaea, Salisbury 1806 b. C. Leibergi, Morong 1888, Ad. original specimens. G. & F. 1888. C. odorata var. minor, Macoun 1891 (in part) Ad. specimen coll. J. M. Macoun, July 17, 1886, No. 94 from hb. Geol. and Nat Hist. Survey, Canada, in hb. British Museum. Not N. odorata var. minor Sims 1814. C. Utragona, Lawson 1889. Britton & Brown 1897. Fia. 66.— Nymphaea tetraaona : a, section of ovary ; b, section of receptacle ; c, outer, d, inner sepal ; e, outer, /, median, 0, innermost petal ; h, stamens. Natural size. Description. — Flower floating, 3.3 to 7.6 cm. across, open on 3 or 4 days from noon until 5 p. m., sweet-scented (or odorless, Britton & Brown 1897). — Peduncle slender, 0.3 to 0.4 cm. in diameter, terete, 10 to 30 cm. long, with 4 main air-canals. — Receptacle enlarging abruptly from the scape, distinctly tetragonal, with rounded angles, green. — Sepals 4, ovate to oblong-ovate or broadly lanceolate, breadth : length = 1:1.16 to 3.61, obtuse or slightly acute; opening horizontally or less, putrescent in fruit; green and veinless outside, white within; when dry, marked outside with short elevated lines. — Petals 8 to 17, rather thin, pure white (or faintly striped with purple lines, Morong 1888), about as long as the sepals, oblong, obovate or broadly lanceolate, or oblong-elliptic, apex obtuse or somewhat acute, narrowed toward the base, grading into the stamens. — Stamens about 40, with golden yellow anthers, shorter than the petals (stamens 12 to 16 in the American plant, in 3 or 4 rows, running up the ovary about half way, Morong 1888) ; outer stamens shorter than median ones; outermost filaments elliptic or obovate to lanceolate-oblong, very broad, narrowed toward the base (inserted up to the summit of the ovary, Casp. 1865 ; Turcz. 1842) ; outermost anthers shorter than median and inner, cells diverging below, sometimes only one cell developed. Median stamen with filament as in outer, but nar- rower; cells of anther parallel. Innermost stamens very short, filaments slender, cells CARNEGIE INSTITUTION OF WASHINGTON WATERLILIES. PLATE 14 ' £ NYMPH/CA TETRAGONA. HEUOTVPE CO., BOSTON. Taxonomy — Nymphaea tetragona. 169 of anthers as in median stamens. — Pollen yellow, finely tuberculate, operculum smooth ; diameter 0.03 to 0.04 mm., average 0.034 mm. — Carpels 6 to 8 (4 to 12, as hitherto described) ; styles thin, subcoriaceous, ovate or oblong-elliptic, short and blunt, breadth: length = 1 : 1.25 to 3, concave; stigmatic area extending on their bases as an upraised cushion, ending obtusely. Stigma little or not at all depressed. — Fruit spherical, naked, green, crowned with the small stigma. — Seed large, 0.3 to 0.2 cm. long by 0.16 to 0.17 cm. in diameter, dark brown, smooth. Seeds which germinate in late summer form small plants, and at approach of winter, I*se their leaves and retire into a small tuber. These germinate, producing first a series of submerged leaves, then small floating leaves as follows : Submerged leaves broadly "obovate, rounded at apex, thin, pale green; sinus deep, margins nearly straight near the petiole and parallel or diverging or slightly overlapping, curved out- ward below to the broadly rounded lobes. First floating leaf obovate-oblong, rounded at apex; sinus overlapping, lobes narrow, rounded at angles, suggesting rabbit-ears; green above with a brown line around margin; under surface greenish-red, densely marked with fine bright red marblings. Leaf of mature plant cleft to the petiole, more or less coriaceous, broadly oblong, ovate or elliptic, 5 to 10 cm. long by 3.3 to 7.6 cm. wide, entire, rounded at apex (rarely emarginate, Casp. 1865), dark green above, more or less mottled (at least when young) with reddish-brown ; under surface green to dull red ; primary veins 5 or 6 on each side of leaf, sunk in the tissues; length of principal area: radius of leaf = 1 : 1.5 to 1.9 (average 1: 1.3). — Sinus depth: length of leaf = 1 : 1.8 to 2.2 (average 1:2), lobes unequal, margins nearly straight or slightly convex, touching, or diverging and 1.9 to 3.8 cm. apart at periphery, angles more or less produced, rounded to sub-acute or rarely acuminate. Astomatic area very small; stomata larger and not so crowded as in Eu-castalia, scattered toward the point of insertion of the petiole ; cells of upper epidermis large. — Petiole smooth, terete, very slender, 0.24 to 0.4 cm. in diameter, 10 to 30 cm. long, with 4 main air-canals, 2 larger and 2 smaller, with 4 still smaller ones anteriorly and posteriorly, and one on each side; 2 double bundles are present (1 anterior and 1 posterior) with 3 or 4 smaller ones on either side; stellate cells projecting into the 6 larger air-canals, rays very slender, long, densely covered with small crystals, quite brittle. — Stipules oblong-ovate or elongate-triangular, acute, not at all keeled. Rhizome erect, 7 to 15 cm. long, blackish, covered with long hairs (1.3 cm. long) ; apex covered with leaf-bases and stipules. Entirely without branches or offsets. Leaves from Small Seedling-Tuber (in centimeters). Measurements of Sepals (after Caspary, 1865). Breadth. Length. cm. 0.6 1-3 0.85 0.8 cm. 1.8 3-3 1.9 3-1 Leaf. Length. Breadth. Submerged . . . Submerged . . . Submerged . . . Submerged . . . First floating 4.8 1.4 3-2 3-5 3-7 3-35 1.1 2.1 2.4 2.4 Depth of Sf epth ol Sinus. 2-55 0.8 1.6 1 -75 1.9 Width of Sinus. Half- At Angles. way up. 0.16 O.08 81 1.1 0.3 0.6 0.3 0.6 a Overlapping. !7o The Waterlilies. Geographic Distribution. — Eastern Europe, Asia and North America, south to Australia. Russia (Baikal region), Georgi 1775. Khasia Hills, and marshes at Non- kreem, alt. 5,600 ft., India, Hook. & Thorn. 1855, fid. specimens coll. Hooker, Duthie, Stewart, etc. N. Queensland, Australia, Midler 1887? Siberia, Gmelin 1769. Com- mon in China and Japan, fid. numerous specimens. Granite Station, Kootenai County, Idaho, U. S. A., coll. Leiberg, Heller and the writer. Misinaibi River and along Sev- ern River, Canada (Ontario), Macoun 1891. Notes. — Caspary (1865), from whom the measurements in the text were mostly taken, divides the species thus: I. Lata: Sepals and petals shorter and broader, outermost petals slightly exceed- ing the sepals ; depth of sinus less than half the length of the leaf. Leaves not spotted. Pollen smaller. Found in eastern and northern Asia. II. Angusta: Sepals and petals longer and narrower; depth of sinus equal to or exceeding half the length of the leaf; pollen larger. This varies into (1) orientalis, outermost petals a little shorter than the sepals ; lobes of leaves long-produced, rarely acuminate ; leaves not spotted. China and Japan. (2) indica, outermost petals slightly exceeding the sepals; lobes of leaves long-produced, tapering and acuminate; leaves usually brown-spotted above. Northern India (Nonkreem). Georgi's specimens are of the broader sepaled type. Those with narrower petals and sepals may deserve varietal rank as var. angusta Casp. The American plant, from its small number of stamens and their insertion below the summit of the torus, may deserve the rank of a variety. N. tetragona was introduced into England from China in 1805 by the East India Company, in the ship " Winchelsea," Capt. Campbell. It is hardy in our climate. Being of small size it is suitable for aquaria. The seeds are used for food in Japan, and the leaf buds are also eaten, dressed with vinegar (Rein 1886). Nymphaea acutiloba, DC. 1821, is of uncertain identity, but as it came from China it is probably N. tetragona Georgi. Nymphaea fennica Mela. Flower cup-shaped, small, 3.8 to 6.9 cm. across, white or rosy ; receptacle large, square, sloping upward, with a keel running from the peduncle to each corner; sepals forming a sharp angle at their line of insertion. Petiole with 2 main air-canals. Nymphaea fennica, Mela 1897, fid. specimens coll. A. J. Mela, in hb. Univ. of Penna. G. C. 1899. Description (taken from Mela's paper). — Flower cup-shaped, small. — Bud short ovate, square at base, constricted just above the base, then rounded and rapidly drawn in to the acute apex. — Receptacle characteristic, very broad ; breadth of receptacle : length of sepals = 6 mm.: 14 mm. (small flower) ; = 9 mm. : 19 mm. (medium) ; = 12 mm. :25 mm. (large flower). Looking at the receptacle with the apex of the bud turned away, the surface slopes gently, roof-like, from the top of the peduncle, with a low, narrow keel running to each angle. Edges of receptacle not projecting downward as in N. Candida, but plane beneath, making a sharp angle on the side. In fruit the receptacle is a little wider than the berry and calyx, these being completely hidden Taxonomy — Nymphaea fennica. 171 when the fruit is viewed from beneath. — Sepals elliptic-ovate, a little contracted below the apex, often very broad; length : breadth = 3:2 (18 to 19 mm.: 12 to 13 mm.) ; calyx somewhat contracted just above its origin, so that the edges of the receptacle, which are sharp-angled outwardly, appear to project. — Petals 10 to 15, very concave, and crowded, with apex directed upward so that the corolla is cup-shaped; lanceolate or narrowly-lanceolate. — Stamens about 50. Anthers very short and broad (3 mm. long by 1 mm. wide) ; filaments petaloid, with broad, elliptic, deep yellow blade and strongly contracted, whitish claw ; blade often with 1 to 3 longitudinal folds, with 1 or 2 trans- verse folds between blade and claw. — Stigma 5 to 9 rayed, with yellow, often violet- tipped, styles. Stigmatic surface usually (not always) deep violet, with generally 3- pointed rays. — Fruit broad conical-ovate, broadest at base, but not so broad as the receptacle which forms a square disc, whose edges (and especially the corners) ex- tend noticeably beyond the circumference of the fruit; the calyx persists as a 4-sided pyramid surrounding the fruit. Leaf small, 7 to 12 cm. long by 5 to 9 cm. wide, elliptic-cordate ; lobes acute or obtuse, margins overlapping or approaching or more or less wide apart (3 cm.). Primary veins of the lobes usually divergent, even when the lobes overlap, though they are sometimes parallel or even converging. — Petiole slightly flattened, with 2 large air-canals. Rhizome erect, 5 to 10 cm. long, about 2.5 cm. in diameter, conical at apex, cov- ered with long black hairs. One of our specimens has two equal branches. Geographic Distribution. — Middle and East Finland, mostly in small forest ponds, sometimes in running water, from the shallows to a depth of 1 or 2 meters, often accompanied by large and small forms of N. Candida. Found also in Southern Finland. (Cf. specimens coll. Ldnnbohm, Aug., 1898, No. 3601, " locus classicus," in several herbaria.) Notes. — The receptacle and air-canals of petiole place this species in near relation to N. tetragona; in several other respects it approaches N. Candida semiaperta. One is inclined to suggest that N. tetragona of China shades off through the Siberian forms into N. fennica, this then runs into N. Candida, and it again into N. alba of Germany and England. A careful review of these forms, in the living state by cultivation and hybridization, as also in histological details, is highly desirable. The seeds of N. fennica have not as yet been described. Numerous stations for this species throughout Finland are given in Mela's paper. Red-flowered specimens occur "im See Keskimmainen im Kirchspiel Lapinlahti," and in Pieni Musto and Jalajarvi. Through the kindness of Mr. Mela, we have a fine series of specimens of N. fen- nica; of several rhizomes sent to us alive, only one survives, and it has never bloomed. C. EUCASTALIA Planchon 1853 b. Flowers white or rosy to deep carmine, opening in early morning, closing from noon to mid-afternoon. Innermost stamens first to dehisce. Pollen aculeate or tuber- culate. Seed of medium size. Leaves entire, nearly orbicular, of solid color above. Principal air-canals in peduncle 4, in petiole 4; thickened bipolar idioblasts more or less numerous in the tissues of leaf and flower stalks, this type of cell being absent iy2 The Waterulies. from other syncarpous members of the genus. Rhizome horizontal, long and creeping, more or less branched. North temperate zone, south to Algeria, Cashmir, and British Guiana. Nymphaea Candida Presl. (Plate XV; Fig. 66.) Leaves floating, crowded on the rhizome, phyllotaxy about 13 on 34; lowest pair of veins on under side of leaf (veins of the lobes) curved toward one another. Flower odorless, white, 6.4 to 7.6 cm. across. Line of insertion of the sepals angular, projecting downward. All of the filaments broader than the anthers. Ovary much contracted and destitute of stamens immediately below the stigmas. Fruit ovoid, seed large. Nympkaeo Candida, Presl 1822, Ad. specimens coll. Topitz, Flora exsiccata austro-hungarica, No. 1282, "in stagno Egersee ad pagum Sonnberg prope urbem Gratzen (locus classicus)." Reichenbach 1845. Planchon 1853 b. Schuchardt 1853. Celakovsky 1867. Caspary 1879. Grisebach 1882. Gerard 1892. Tricker 1897. Moenkemyer 1897. Conard 1901a. N. alba, Linnaeus 1753; 1762 (in part), and other European authors. N. alba oocarpa, Caspary 1855 = N. alba oligostigma Casp. 1863 ; 1865. N. biradiata, Sommerauer 1833, Ad. specimens coll. Sommerauer, in Tricbnersee in Styria in hbb. Paris, Boissier, and coll. Wettstein, Flora exsiccata austro-hungarica, No. 1281. Coll. Statzer /in I. Dorfler's Herb. Norm., No. 3502; coll. Keller in Magnier, Fl. select, exsicc No. 3453! coll. F. G. Strobl, " in aquis stagnantibus : Triebner — vel Gaishornsee vallis Paltinae (locus classicus)." Reichenbach 1845. Fries 1846. Planchon 1853ft. Schuchardt 1853. N. semiaperta, Klinggraeff 1848, Ad. F. Schulz, Herb. Norm. nov. ser. Cent. 3, No. 207, and specimen coll. Caspary, Aug. 9, 1859, near Konigsberg, in hb. Boissier. Schuchardt 1853. N. neglecta, Hausleutner 1850, Ad. specimen coll. R. Fritz, July, 1870, in a pond near Rybnick, Ober-Schlesien, in hb. Berlin. N. Kosteletzkyi, Palliardi MS., Ad. Lehmann 1852 b, and specimens sent by Palliardi to Lehmann, from hb. Lehmann in hbb. Kew, Berlin. Schuchardt 1853. N. pauciradiato, Bunge, in Meyer & Bunge 1829 = N. alba Patrin, Ad. Meyer & Bunge, 1. c ; Ad. specimens coll. Ledebour, 1845, in the Altai region, in hbb. Paris, Munich. N. intermedia, Weiker, in Reichenbach 1842. N. Basniniana, Turczaninow 1842, Ad. specimens sent by Turczaninow from hb. St. Petersburg, in hbb. Kew, Boissier, DeCandolle = N. bashiniana, Steudel 1841. N. Cachemiriana, Cambessedes in Jacquemont 1844, Ad. original specimens in hb. Paris = N. alba var. Kashmiriona, Hook. & Thorns. 1855. Hooker 1872. N. punctata, Kar. & Kir. 1842, Ad. Casp. 1879, from Semenow 1854, and specimens coll. Kar. & Kir. " in Songoriae lacubus prope fluvium Lapsa," in hbb. Berlin, Paris, DeCandolle. AT. alba, L. var. minor DC, Ad. specimen coll. Fischer, 1819, in hb. DeCandolle. Besler 1613, cf. Flora 1833. Schuchardt 1853. =AT. minor Reichenbach 1842; 1845. =N. permixta, Boreau 1859. N. Milletii, Boreau 1859, Ad. specimen coll. Ed. Longeron, June 12, 1861, Maine et Loire (locus classicus?) in hb. British Museum. N. Wenzelii, Maack in litt, Ad. plate from Ruprecht 1859, in hb. British Museum. Description. — Flower floating, 6 to 13 cm. in diameter. — Sepals 4 (or 5) oblong or ovate-oblong (3.8 cm. long by 1.3 cm. wide), narrowed at apex, acute or obtuse,about 7- veined, green outside, white within. — Petals 12 to 20, white, oblong to oblong-lanceo- late, obtuse or somewhat acute, 7-nerved, outer ones nearly as long as the sepals. — Stamens 32 to 70, shorter than in N. alba, grading inward from the petals, orange- yellow; filaments lanceolate (outer) to linear-lanceolate (inner), long acuminate, never narrower than the anthers, usually inserted only on the sides, not on the summit, of Taxonomy — Nymphaea Candida. 173 the ovary. Pollen granulate with smooth operculum, larger than in N. alba. — Ovary ovate or roundish, usually contracted and destitute of stamen insertions below the stigma. — Carpels 5 to 14; styles linear, flattened, breadth : length = 1 : 1.5 to 2.5 ; apex rounded or somewhat acute. Stigma concave, yellow, red or violet, the rays terminat- ing usually in 3 teeth (a central and two lateral) on the base of each style. Axile pro- cess slender. — Fruit ovoid or spherical ; in the former case, narrowed below the stigma, which stands as it were on a short thick column surmounting the berry; greenish or reddish; subtended by the persistent sepals; diameter 1.9 to 4.4 cm. — Seed larger than in N. albit, ellipsoid, brownish, 0.3 cm. long ; aril longer than the seed. Leaf cleft at base to the petiole, sub-orbicular to oval, 10 or 12 to 30 cm. long and 25 cm. wide, entire. — Sinus, depth : length of leaf = 1 : 2.5 ; lobes nearly equal, inner margins curved, overlapping a little above, and spreading out toward periphery of leaf (or straight, and parallel or touching) ; angles more or less acute. Veins on under sur- face of leaf prominent, the lowest pair (i. e., those running into the lobes) curved, and, if produced, would cross, inclosing an oval area. — Petiole terete, smooth. Rhizome stout, 5 cm. or more in diameter, horizontal, with strong lateral roots, but often without side- branches ; apex surrounded by peti- oles, peduncles and stipules. Phyl- lotaxy about 13 on 34 with very short internodes ; the surface of the rhizome may almost be said to con- sist of large leaf-cushions. Geographic Distribution. — Northern and arctic Europe and Asia, south to the mountains of Ger- many and the Himalayas. Krumme- See, near Grubno, in West Prus- sia (Caspary 1884). Canton St. Gallen, in Switzerland (Gremli 1882). Lojo in Poland (Lindberg 1881). Herjedalen in Sweden (Behm 1887). Far north in Finland (Saelan 1879). Egersee, near Gratzen, in Austria (Kerner 1886). Koniggratz, in Bohemia (Hansgirg 1880). Veszto, in Hungary (Borbas 1881). Ascends 500 meters in southern Bavaria (Schenck 1885). East slopes of Urals (Cler 1873). Trieb- nersee, Palatinathal, Austria (Kerner 1886). Bohemia, in ponds and ditches (Presl 1822). Austro-Hungary (Leopoldskron, coll. Stohl), Flora exsiccata austro-hungarica, Fig. 66.— Nymphaea Candida : Under side of leaf of specimen in ho. British Museum, from the classic locality (Flora exsiccata austro-hungarica. No. 1882), by courtesy of the Directors. P. Highley, del. Reduced one-quarter. 174 The Waterlilies. No. 1283. Franzensbrunn : 3 Nomenteiche, coll. Caspary, Aug. 25, 1855, in hb. Kew. Lakes of the Vosges near Sembach, Bavaria, coll. Koch, July 20, 1859. Hammerteich near Falkenberg, Ober-Schlesien, fid. specimen in hb. Kew. Teplitz in Bohemia, coll. Winkler, 1853, in hbb. Boissier, DeCandolle. Near Stockholm, Sweden, coll. Ljo- gruen, 1845, in hbb. Munich, DeCandolle. Lithuania, coll. E. Rudominowna, in hb. Kew. Neusattel near Ellnbogen, coll. Caspary & Ortman, 1855, in hb. Kew. Schwarz- see near Kitzbuhel, Tyrol, coll. Traunsteiner, No. 1608, marked " N. alba," in hbb. Boissier, Paris. Lakes of the Baikal region, Siberia, coll. Fischer, 1836, in hb. Paris. Cashmere Lake, coll. J. E. Winterbottom, May 13, 1847, and cou- Falconer, both marked " N. alba" in hb. Kew. Notes. — The species varies widely as indicated above in the description, and as may be inferred from the lengthy synonymy. Presl's distinctions were not widely known, or were lost sight of by most botanists. He separated N. Candida from N. alba largely on account of the absence of stamens from the summit of the ovary. The same name had been used for a European white waterlily by Tragus, Fuchs, Caesal- pinus and others. N. pauciradiata was characterized by its small flowers and few car- pels (8 to 9). N. biradiata was noticed on account of the blood-red star on the stigma. N. basniniana was separated from N. alba on account of having oval leaves instead of round, and its small number (9) of carpels. N. cachcmiriana has also only 7 to 8 carpels ; the original description is quite complete. N. semiaperta has as its out- standing trait the fact that even at noonday the flowers remain only half open. N. neglecta was marked off from alba by its small oval fruit, few rayed stigma, and the converging course of the lowest pair of veins of the leaf; from Candida in having petals longer than the calyx, and ovary only one-third bare, instead of two-thirds as in Presl's Candida; from semiaperta by smallness of ovary and color of stigma. It is interesting to notice how various observers have been impressed with different feat- ures of the plant as they appear more or less exaggerated in certain localities. Probably the best disposition of the case is that of Caspary in 1879, when he first granted specific recognition to this group of forms. His paper of 1855 placed these plants in variety oocarpa of N. alba, with the following divisions: N. alba var. oocarpa. E. angustata — styles longer and narrower; Konigswerth in Bohemia. F. semiaperta — styles shorter; flower opening one half. a. erythrocarpa — fruit reddish (=N. semiaperta and biradiata) b. chlorocarpa — fruit greenish (= N. neglecta) G. aperta: a. erythrocarpa (=N. kosteletzkyi and cubogermen, from Franzensbrunn) (=N, Candida Presl from Pardubitz in Bohemia) b. chlorocarpa (=N. Candida Ortm. from Ellnbogen, Bohemia) (=N. Candida Presl from Pardubitz, Bohemia) The name oocarpa was changed by the author to oligostigma in 1863, and finally, on discovering that crosses between N. alba oocarpa and N. alba sphaerocarpa were true hybrids, with imperfect sexual apparatus, the two were made distinct species, and the following disposition of the one in hand was given, with a few, mostly Swedish, localities. Taxonomy — Nymphaea alba. 175 N. Candida Presl. Var. I. oocarpa; fruit ovoid, higher than broad. 1. aperta; sepals and petals opening to form an obtuse angle. A. xantho stigma; stigma yellow. Nerike; Fagertarn; E. Prussia; Konigsberg; a few lakes in West Prussia. B. erythrostigma; stigma more or less dark crimson, especially near apices of stigmatic rays. Sodermanland and Norbotten. a. erythrocarpa. Norbotten. b. chlorocarpa. Norbotten. 2. .iimiaperta; sepals and petals forming an acute angle when open; fl. campanulate. Prus- sia, east of the Weichsel, and Bohemia. Erythrocarpa and chlorocarpa, N. biradiata and N. semiapcrta. Var. II. sphaeroides; fruit spheroidal, broader than tall. 1. aperta; sepals and petals opening to form an obtuse angle. A. erythrocarpa. Norbotten. B. chlorocarpa. Norbotten. 2. semiapcrta; sepals and petals forming an acute angle when open; fl. campanulate. A. erythrocarpa. Norbotten; Prussia. B. chlorocarpa. Prussia. N. alba X Candida occurs in East Prussia at Neuhausen near Konigsberg, near Goldap, and elsewhere (Caspary 1879). Sernander (1894) claims to have found a red variety of N. Candida in Fagertarn, Sweden. Nymphaea nitida, Sims 181 1 a, with entire, suborbicular leaves, with obtuse lobes; flowers white, cup-shaped, may belong here. Its habitat is unknown, and it has not recently been positively identified anywhere. The figure of the Botanical Magazine was " drawn at Mr. Vere's garden " in Kensington-Gore, Mr. Anderson, gardener. A specimen exists in hb. Caspary (now Berlin), which was presented by Prof. M. T. Masters in 1854 from the botanic garden of Oxford. Here it was received in 1840 from Kent, whither it came directly from Anderson. This specimen appears to belong to N. Candida as here defined. Nymphaea alba (L.) Presl. Leaves nearly orbicular, floating, red when very young, 10 to 30 cm. in diameter, crowded on the rhizome, phyllotaxy about 13 on 34 ; lowest pair of veins straight and diverging. Flower odorless, pure white (in the type), 10 to 12 cm. across, opening fully. Line of insertion of sepals rounded, not prominent. Stamens covering the ovary to the summit. Filaments of inmost stamens not wider than the anthers. Axile process spherical. Fruit more or less spherical, seed small. Nymphaea alba, Linnaeus 1753, fid. original specimen in hb. Linnaeus, without data; 1762 (in part). Willdenow 1797 (in part). Aiton 1789; 1811 (in part). Smith 1800; 1809 (in part), and all writers previous to 1822, and many since. N. alba, Presl 1822. Reichenbach 1845. Hentze 1848 a; 1852 a, b. Schuchardt 1853. Caspary 1879. De Halacsy 1900. Conard 1901 a. N. alba melocarpa, Caspary 1855 = AT. alba polystigma Casp. 1863. N. Aesopii, Orphanides MS. in Boissier 1888; fid. specimens coll. T. G. Orphanides, Flora Graeca exsiccata, No. 1051 (hb. Orphanid., No. 264) Feb., 1862, "ad lacu Baldovo supra Vodena Macedoniae," in hbb. Paris, Boissier. N. Dioscoridis, Heldr. MS., in Boissier 1888; fid. specimens coll. Heldreich, Herbar. Flor. Hellenic, May 23, 1878, at " Aetolia : in lacu Lysimachion," in hb. Boissier. 176 The Waterlilies. N. venusta, Hentze 18486; 1853 a, fid. specimens coll. Hentze from hb. Lehmann, in hbb. Kew, Berlin. N. rotundifolia, Hentze 1848 b; 1852 a, fid. specimens coll. Hentze, from hbb. Lehmann and Caspary, in hbb. Kew, Berlia N. erythrocarpa, Hentze 1852 6 = AT. alba Hentze 18480; 18520, fid. specimen coll. Hentze, from hb. Lehmann, in hb. Berlin. N. parvifiora, Hentze 1848 b; 18520, fid. specimens coll. Hentze, from hb. Lehmann, in hbb. Kew, Berlin. N. spUndens, Hentze 18480; 18520, fid. specimens coll. Hentze, from hb. Lehmann, in hbb. Kew, Berlin. N. urceolata, Hentze 18486; 1852 a, fid. specimen from hb. Lehmann, in hb. Berlin. N. alba var. intermedia Gave 1892; fid. original specimen coll. P. Gave, June 30, 1891, St. Andre, Haute Savoie. Castalta speciosa, Salisbury, 1806 a. Britten 18880. C. alba, Woodville & Wood 1806. Link 183 1. Greene 1888. Lawson 1889. Description. — Flower floating, 7 to 15 cm. in diameter, open on 4 days from 7 a. m. to 4 p. m. (7 a. m. Upsala; 8 a. m. Innsbruck, Kerner 1895), (5 p. m. Schulz 1890), faintly sweet-scented on first day of opening, odorless afterward. — Peduncle terete. — Receptacle rounded along the line of insertion of the sepals. — Sepals 4 (rarely 3 or 5, Eichler 1875) oblong or ovate, obtuse or somewhat acute, smooth, opening out horizontally or more, putrescent in fruit, green mostly shaded with reddish-brown out- side, whitish or pale reddish within, stiff, obscurely 7-nerved. — Petals 12 to 24, of rather firm texture, outermost opening horizontally, grading inward in size, shape and position to the stamens. Outermost petals about as long as the calyx, lanceolate to oblong, or ovate, obtuse, white, often striped with green, gray, or reddish on the back — Stamens 64 to 100 or more, covering the torus to its summit, innermost first to ripen. Outer ones 15 to 17 mm. long, innermost 4 to 5 mm. long. Anthers sulphur-yellow to orange-yellow, sickel shaped with the convexity inward, inserted at an angle with the filament, the angle being slight in the outer stamens, reaching 900 in the innermost; lines of de- hiscence facing inward and downward ; outermost anthers 10 to 12 mm. long, inner- most about 3 mm. Filaments curved, convex side outward, bringing the anthers o»er the middle of the flower ; outer filaments white, petaloid : innermost filament fifitnnn, yellow, narrower than the anther. Pollen aculeate, operculum smooth; 0,038 mm. in diameter. — Ovary slightly contracted below the stigma, covered with stamens to the summit. — Carpels 8 to 24 (mean = 12 to 20) ; styles fleshy, triangular-ovate, rounded at apex, more or less concave and 1 to 5 sukate on inner side, subphur yeDow, often tinted orange. Stigmatic area scarcely produced above to the bases of die styles, end- ing acutely. Stigma more or less deeply depressed, basin or funnel shape, sulphur- yellow. Axile process conical, acute, yellowish-white. — Fruit globose or depressed- globose, crowned with the styles, which are bent sharply inward. Inner wall of cells whitish, spongy ; tissues green or blood-red. Rq»e fnrit yellowish-green devoid of floral leaves.-^Staf 3 mm, long by 2 mm. in diameter, «*lKp«ni^»i o dark olive green, smooth ; raphe evident. Aril inclosing the seed, open at end < hirum. and with irregular margin. Smbmtrgti leaves of adult plant 2 or 3, broadly defend; lamina about n can. across, with stomata on upper epidermis; petiole 10 cm. long: leaves ptodtcd in fal, or in unfavorable conditions, (Cf. Arcangeh 1890 k) Taxonomy — Nymphaea alba. 177 Size of Leaves. Length. Width. cm. cm. 27 24.9 27" 30.6 35-1 37 34-4 36.4 31.2 34-4 29-3 95-23 9-5-23 Mature floating leaf circular or round-ovate, entire or slightly emarginate at apex, somewhat leathers-, usually dark green above, paler in astomatic area and over primary veins ; yellowish-green or reddish beneath ; reddish or even dark red when very young ; marginal line brownish-red.— Sinus reaching to the petiole, open or closed, margins curved, or rarely nearly straight, separated, touching or overlapping; angles obtuse. Veins of the lobes, if produced, diverging. — Stipules large, lanceolate, narrowly emar- ginate, bicarinate, entirely free, far overreaching the apex of the rhizome. Rhizome horizontal, often very thick (5 to 8 cm.), becoming 90 to 120 an. long, black, with prominent leaf cushions. Apex protected by the long stipules and the usual long fine hairs. Phyllotaxy of a high order, about 13 on 34. Branches few, stout, strongly attached. Geographic Distribution. — Europe and Northern Africa. Great Britain (Smith 1800, etc.). Denmark, France, Germany, Italy, Spain, Greece. Algeria (Planchon 1853 b) . Sweden. E. g. Heldreich, Herbar. Graecum Norm., No. 1 106, coll. in " Euboea Septentrionalis : in stagnis Suvala ad cumen Tympanon in Telethrii." Steinitz, Flora exsiccata Austro-Hungar., No. 2056, " Hun- garia : in aquis stagnantibus prope Rakos in ditione Pestinensi." C. Scholz, No. 775 of A. Callier : Flora Silesiaca exsiccata, Editio 1895, " Guhrau : Wald bei Gabel." Lignier, No. 3452 in Magnier, Fl. select, exsicc, 1893-4, " Loir-et-cher : Villeherviers, etang au Glandier." Notes. — This is the Nymphaea par excellence of the classic botanists. It has a wide distribution and varies much in size of leaf and flower, as well as in the color of all the parts, and the shape of the fruit. Hentze ( 1848-52) distinguished several varie- ties which so fully retained their characteristics in cultivation and in propagation from seed that he regarded them as practically of specific rank. N. splendens was so named for its orange-yellow instead of sulphur-yellow stamens, and was collected in ponds in " Reinhardswalde in Niederhessen " and between Beberbeck and Hombrcssen. N. urceolata was characterized by the deeply urceolate stigma; collected between Rodcl- heim and Bockenheim, and at Bergen near Frankfurt a. M.; also at Darmstadt ("grossen Wog"), Buchrainweihern near Heissenstamm, and Bulau near Hanau. N. venusta, with many (24) petals and (90) stamens, and a cup-shaped flower, seemed especially showy ; collected in the so-called " Grosse Judenteiche " at Gosslar in the Harz. N. rotundifolia was named for the shape of its leaves ; collected near Heringen, " Kleinensee im Werra-Thal." Hentze's typical N. alba came from Sweden (Upsala). N. erythrocarpa was characterized by the blood-red tissues of the interior of the fruit, and came from " Bremerhafen, aus dem Lande Hadeln." N. parviilora was distin- guished by its small flowers (7 to 10 cm. across, the others being 10 to 15 cm.), accom- Size of Fruit. Height. Diameter. cm. 3-8 3-2 cm. 4.0 3-5 14 178 The Wateri.ilies. panicd by leaves of the usual size; it was collected at two stations near Lich. All of these were supported by numbers of specimens in all stages of flower and fruit. In Greece, N. aesopii of Macedonia was separated off by Orphanides on account of its huge size, and N. dioscoridis from Lake Lysimachia in Aetolia for its smallness. Caspary began his work on waterlilies by reviewing this species. He rightly maintained that the smallness or largeness of the plant is not a safe guide to classifica- tion, but that there were other characteristics of a more profound and constant nature on which divisions could be founded. Accordingly in 1855 he gave the following arrangement : N. alba. I. melocarpa. A. dcpressa. a. chlorocarpa = N. venusta Htze. = JV. rolundifolia Htze. b. erythrocarpo = N. erythrocarpa Htze. B. circumvallata. C. sphaerocarpa. a. chlorocarpa. a. ftava ==N. parviHora Htze. = N. alba Htze. P. splendens = N. splendens Htze. b. erythrocarpa. D. urceolata = N. urceolata Htze. II. Oocarpa (see under N. Candida Presl). In 1863 Caspary changed the main headings, substituting I. Polystigma and II. Oligostigma in the above table. Finally in 1879, as referred to under N. Candida, he raised the two varieties to specific rank under the names of N. alba (L.) Presl and N. Candida Presl, dividing the former species and giving a few localities, thus : N. alba. I. sphaerocarpa. Fruit subglobose. 1. platystigma. Stigma more than half the diameter of the fruit (commonest form). A. chlorocarpa. Fruit green (widely distributed). a. Aava. Stamens sulphur color. b. splendens. Stamens orange. B. erythrocarpa. Fruit more or less suffused crimson. a. vulgaris. Petals white, or outer ones pale rosy. b. rosea. Petals and stamens deep crimson == N. alba and rosea Hartm. Handb. Skand. Fl., 1864, p. 76. 2. engystigma. Stigma, measured from the backs of the processes, equal to or less than half the diameter of the fruit (rare) ; West Prussia, near Danzig. II. dcpressa. Fruit depressed-globose ; alt. : lat. = 2:3 to 5:6. 1. chlorocarpa — here and there in Germany. 2. erythrocarpa — here and there in Germany. III. urceolata. Stigma deeply infundibuliforin in fruit, sunk in to middle of fruit— in Ger- many, very rare. Chlorocarpa and erythrocarpa. IV. oviformis. Fruit ovoid ; upper one-fourth to one-third of height destitute of filaments— in some lakes of West Prussia. :--••£; e r~ "." :•. :r •• -:- •■:--:■. "E-_ . E5 "_i_E ': NYWHvEA ALBA RUBRA, • ?