5 ;■><*: Botany EDITED BY DUKINFIELD HENRY SCOTT, M.A., LL.D., D.Sc., For. Sec. R.S. LATELY HONORARY KEEPER OF THE JODREI.L LABORATORY, ROYAL BOTANIC GARDENS, KEW JOHN BRETLAND FARMER, D.Sc., M.A., F.R.S. PROFESSOR OF BOTANY, ROYAL COLLEGE OF SCIENCE, LONDON FRANCIS WAFL OLIVER, M.A., D.Sc., F.R.S. QUAIN PROFESSOR OF BOTANY, UNIVERSITY COLLEGE, LONDON AND ROLAND THAXTER, M.A., Ph.D. PROFESSOR OF CRYPTOGAMIC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S A. ASSISTED BY OTHER BOTANISTS (/' 1 \fo^ ,p A C> v^\. ifVv-' X y Annals of VOLUME XXX With Fifteen Plates and Two Hundred and Eighty-one Figures in the Text IS LONDON Humphrey milford, oxford university press AMEN CORNER, E.C, Edinburgh, Glasgow, New York, Toronto, Melbourne, and Bombay iqi6 PRINTED IN ENGLAND AT THE OXFORD UNIVERSITY PRESS CONTENTS ro.SMO. FAGE David Thomas Gwynne-Vaughan. With Portrait i-xxiv Alfred Stanley Marsh xxv-xxvii No. CXVII, January, 1916. WlLLIS, J. C. — The Evolution of Species in Ceylon, with reference to the Dying Out of Species. With two Figures in the Text ........... 1 Leitch, I. — Some Experiments on the Influence of Temperature on the Rate of Growth in Pisum sativum. With Plate I and ten Figures in the Text . . . . .25 Laidlaw, C. G. P., and Knight, R. C. — A Description of a Recording Porometer and a Note on Stomatal Behaviour during Wilting. With three Figures in the Text . 47 Knight, R. C. — On the Use of the Porometer in Stomatal Investigation. With seven Figures in the Text .... ......... 57 Brenchley, Winifred E. — The Effect of the Concentration of the Nutrient Solution on the Growth of Barley and Wheat in W'ater Cultures. With Plate II and four Diagrams in the Text .............. 77 Barratt, Kate. — The Origin of the Endodermis in the Stem of Hippuris. With six Figures in the Text ........ ......91 Davie, R. C. — The Development of the Sorus and Sporangium and the Prothallus of Pera- nema cyatheoides, D. Don. With Plate III and two Figures in the Text . . .101 Stopes, Marie C. — An Early Type of the Abietineae (?) from the Cretaceous of New Zealand. With Plate IV and seven Figures in the Text . . . . . .111 Holden, H. S. — Further Observations on the Wound Reactions of the Petioles of Pteris aquilina. With four Figures in the Text . . . . . . . . .127 Fritsch, F. E. — The Morphology and Ecology of an Extreme Terrestrial Form of Zygnema (Zygogonium) ericetorum (Kuetz.), Hansg. With three Figures in the Text . . 135 Takeda, IT. — Dysmorphococcus variabilis, gen. et sp. nov. With fifteen Figures in the Text . . . . . . . . . . . . . . . 1 5 1 Takeda, H. — Scourfieldia cordiformis, a New Chlamydomonad. With five Figures in the Text ........ ..... 157 Stapledon, R. G. — On the Plant Communities of Farm Land ...... 161 Fraser, Mary T. — Parallel Tests of Seeds by Germination and by Electrical Response. (Preliminary Experiments.) . . . . . . . . . . .181 NOTES. Small, James. — Anomalies in the Ovary of Senecio vulgaris, L. W7ith three Figures in the Text . . . . . . . . . . . . . . .191 Doyle, Joseph. — Note on the Structure of the Ovule of Larix leptolepis. With one Figure in the Text 193 No. CXVIII, April, 1916. Takeda, H. — Some Points in the Morphology of the Stipules in the Stellatae, with special reference to Galium. With twenty-seven Figures in the Text . . . . .197 Hill, Arthur W. — Studies in Seed Germination. The Genus Marah (Megarrhiza), Cucurbi- taceae. With Plate V and two Figures in the Text . . . . . . *215 Hind, Mildred. — Studies in Permeability. III. The Absorption of Acids by Plant Tissue. With eleven Figures in the Text . . . . . . . . . .223 A 2 IV Contents. PAGE DE Fraine, E. — The Morphology and Anatomy of the Genus Statice as represented at Blakeney Point. Part I. Statice binervosa, G. E. Smith, and S. bellidifolia, D.C. ( = S. reticu- lata). With systematic and ecological notes by E. J. Salisbury. With Plate VI, twenty-eight Text-figures, and four Tables ........ 239 Delf, E. Marion. — Studies of Protoplasmic Permeability by Measurement of Rate of Shrink- age of Turgid Tissues. I. The Influence of Temperature on the Permeability of Protoplasm to Water. With seventeen Figures and five Tables in the Text . . 283 Groom, Percy. — A Note on the Vegetative Anatomy of Pherosphaera Fitzgeraldi, F. v. M. With one Figure in the Text . . . . . . . . . . . 31 1 Sampson, K. — The Morphology of Phylloglossum Drummondii, Ivunze. With five Figures in the Text . . . . . . . . . . . . . . . 315 Sutherland, Geo. H., and Eastwood, A. — The Physiological Anatomy of Spartina Town- sendii. With seven Figures in the Text ......... 333 NOTES. Doyle, Joseph. — On the ‘Proliferous ’ Form of the Scape of Plantago lanceolata. With two Figures in the Text ............. 353 Salisbury, E. J. — On the Relation between Trigonocarpus and Ginkgo .... 356 West, Cyril. — Stigeosporium marattiacearum, gen. et sp. nov. ...... 357 No. CXIX, July, 1916. Jeffrey, Edward C., and Cole, Ruth D. — Experimental Investigations on the Genus Drimys. With Plate VII ........... 359 Takeda, H. — On Carteria Fritschii, sp. nov. With ten Figures in the Text .... 369 Fritsch, F. E., and Takeda, H. — On a Species of Chlamydomonas (C. sphagnicola, F. E. Fritsch and Takeda — -Isococcus sphagnicolus, F. E. Fritsch). With fourteen Figures in the Text .............. 373 Acton, Elizabeth. — Studies on Nuclear Division in Desmids. I. Hyalotheca dissiliens (Sm.), Breb. With Plate VIII and four Figures in the Text 379 Paine, Sydney G. — On the Supposed Origin of Life in Solutions of Colloidal Silica. With Plate IX .............. 383 Blackman, V. H., and Welsford, E J. — Studies in the Physiology of Parasitism. II. Infection by Botrytis cinerea. With Plate X and two Figures in the Text . . . 389 Brown, William. — -Studies in the Physiology of Parasitism. III. On the Relation between the ‘ Infection Drop ’ and the underlying Host Tissue ...... 399 Bayliss-Elliott, Jessie S., and Grove, W. B. — Roesleria pallida, Sacc. With eleven Figures in the Text ............. 407 Welsford, E. J. — Conjugate Nuclei in the Ascomycetes. With four Figures in the Text . 415 Rusiiton, W. — The Development of ‘ Sanio’s Bars 5 in Pinus Inops. With four Figures in the Text ............ ... 419 Stiles, Walter. — On the Interpretation of the Results of Water Culture Experiments . .427 Willis, J. C. — The Distribution of Species in New Zealand. With a Diagram in the Text . 437 Smith, Gilbert Morgan. — Cytological Studies in the Protococcales. I. Zoospore Forma- tion in Characium Sieboldii, A. Br. With Plate XI and two Figures in the Text . 459 Smith, Gilbert Morgan. — Cytological Studies in the Protococcales. II. Cell Structure and Zoospore Formation in Pediastrum Boryanum (Turp.), Menegh. With Plate XII and four Figures in the Text ........... 467 NOTE. Barratt, Kate. — A Note on an Abnormality in the Stem of Helianthus annuus. With three Figures in the Text 4^ 1 Contents. v No. CXX, October, 1916. Prefatory Note to two Unpublished Papers by the late Professor D. T. Gwynne-Vaughan Gwynne-Vaughan, D. T. — Observations on the Anatomy of the Leaf in the Osmundaceae. With Plate XIII Gwynne-Vaughan, D. T. — On some Climbing Davallias and the Petiole of Lygodium. With Plate XIV and eight Figures in the Text ........ Worsdell, W. C. — The Morphology of the Monocotyledonous Embryo and of that of the Grass in particular. With ten Figures in the Text ....... Salisbury, E. J. — Variations in Anemone nemorosa. With three Figures in the Text . Dutt, C. P. — Pityostrobus macrocephalus, L. and H. A Tertiary Cone showing Ovular Structures. With Plate XV and two Figures in the Text . . Ridley, H. N. — On Endemism and the Mutation Theory ....... Davey, A. J. — Seedling Anatomy of certain Amentiferae. With eighteen Figures in the Text Takeda, H. — Some Points in the Morphology of the Stipules in the Stellatae, with special reference to Galium. (Additional Note.) With seven Figures in the Text NOTE. Sampson, K. — Note on a Sporeling of Phylloglossum attached to a Prothallus. With two Figures in the Text PAGE 485 487 495 5°9 525 529 55i 575 601 605 INDEX A. ORIGINAL PAPERS AND NOTES. Acton, Elizabeth. — Studies on Nuclear Division in Desmids. I. Hyalotheca dissiliens (Sm.), Breb. With Plate VIII and four Figures in the Text . ... . Barratt, Kate. — The Origin of the Endodermis in the Stem of Hippuris. With six Figures in the Text A Note on an Abnormality in the Stem of Helianthus annuus. With thiee Figures in the Text .... ........ Bayliss-Elliott, Jessie S., and Grove, W. B. — Roesleria pallida, Sacc. With eleven Figures in the Text ............. Blackman, V. H., and Welsford, E. J. — Studies in the Physiology of Parasitism. II. In- fection by Botrytis cinerea. With Plate X and two Figures in the Text Brenchley, Winifred E. — The Effect of the Concentration of the Nutrient Solution on the Growth ot Barley and Wheat in Water Cultures. With Plate II and four Diagrams in the Text ............... Brown, William. — Studies in the Physiology of Parasitism. III. On the Relation between the ‘ Infection Drop’ and the underlying Host Tissue ...... Cole, Ruth D., see Jeffrey, E. C. Davey, A. J. — Seedling Anatomy of certain Amentiferae. With eighteen Figures in the Text Davie, R. C. — The Development of the Sorus and Sporangium and the Prothallus of Peranema cyatheoides, D. Don. With Plate III and two Figures in the Text .... DE b raine, E. — The Morphology and Anatomy of the Genus Statice as represented at Blake- ney Point. Part I. Statice binervosa, G. E. Smith, and S. bellidifolia, D.C. ( = S. reticulata). With systematic and ecological notes by E. J. Salisbury. With Plate VI, twenty-eight Text-figures, and four Tables ........ Delf, E. Marion. — Studies of Protoplasmic Permeability by Measurement of Rate of Shrink- age of Turgid Tissues. I. The Influence of Temperature on the Permeability of Protoplasm to Water. With seventeen Figures and five Tables in the Text Doyle, Joseph. — Note on the Structure of the Ovule of Larix leptolepis. With one Figure in the Text -On the ‘Proliferous’ Form of the Scape of Plantago lanceolata. With two Figures in the Text ............ Dutt, C. P. — Pityostrobus macrocephalus, L. and H. A Tertiary Cone showing Ovular Structures. With Plate XV and two Figures in the Text ...... Eastwood, A., see Sutherland, G. H. Fraser, Mary T. — Parallel Tests of Seeds by Germination and by Electrical Response. (Pre- liminary Experiments.) . ........... Fritsch, F. E. — The Morphology and Ecology of an Extreme Terrestrial Form of Zygnema (Zygogonium) ericetorum (Kuetz.), Hansg. With three Figures in the Text and Takeda, H. — On a Species of Chlamydomonas (C. sphagnicola, F. E. Fritsch and Takeda — Isococcus sphagnicolus, F. E. Fritsch). With fourteen Figures in the Text .............. Groom, Percy. — A Note on the Vegetative Anatomy of Pherosphaera Fitzgeraldi, F. v. M. With one Figure in the Text ........... Grove, W. B., see Bayliss-Elliott, J. S. Gwynn e-Vaughan, D. T. — Observations on the Anatomy of the Leaf in the Osmundaceae. With Plate XIII On some Climbing D avail ias and the Petiole of Lygodium. With Plate XIV and eight Figures in the Text ........ PAGE 379 91 481 407 389 77 399 575 101 239 283 193 353 529 181 135 373 311 4S7 495 > Index. Hill, Arthur W. — Studies in Seed Germination. The Genus Marah (Megarrhiza), Cucurbi- taceae. With Plate V and two Figures in the Text ....... Hind, Mildred. — Studies in Permeability. III. The Absorption of Acids by Plant Tissue. With eleven Figures in the Text .......... Holden, PI. S. — Further Observations on the Wound Reactions of the Petioles of Pteris aquilina. With four Figures in the Text ......... Jeffrey, Edward C., and Cole, Ruth D. — Experimental Investigations on the Genus Drimys. With Plate VII . . Knight, R. C. — On the Use of the Poromeler in Stomatal Investigation. With seven Figures in the Text .............. see Laid law, C. G. P. Laidlaw, C. G. P., and Knight, R. C. — A Description of a Recording Porometer and a Note on Stomatal Behaviour during Wilting. With three Figures in the Text Leitch, I. — Some Experiments on the Influence of Temperature on the Rate of Growth in Pisum sativum. With Plate I and ten Figures in the Text . . . . . Obituary : David TH'omas Gwynne-Vaughan. With Portrait i Alfred Stanley Marsh xxv- Paine, Sydney G. — On the Supposed Origin of Life in Solutions of Colloidal Silica. With Plate IX ....... ........ Prelatory Note to two Unpublished Papers by the late Professor D. T. Gwynne-Vaughan Ridley, H. N. — On Endemism and the Mutation Theory ....... Rushton, W. — The Development of ‘ Sanio’s Bars ’ in Pinus Inops. With four Figures in the Text ............... Salisbury, E. J. — On the Relation between Trigonocarpus and Ginkgo . . . . Variations in Anemone nemorosa. With three Figures in the Text . see de Fraine, E. Sampson, K. — The Morphology of Phylloglossum Drummondii, Kunze. With five Figures in the Text - ■ Note on a Sporeling of Phylloglossum attached to a Prothallus. WTith two Figures in the Text Small, James. — Anomalies in the Ovary of Senecio vulgaris, L. With three Figures in the Text Smith, Gilbert Morgan. — Cytological Studies in the Protococcales. I. Zoospore Formation in Characium Sieboldii, A. Br. With Plate XI and two Figures in the Text . — Cytological Studies in the Protococcales. II. Cell Structure and Zoospore Formation in Pediastrum Boryanum (Turp.), Menegh. With Plate XII and lour Figures in the Text ........... Stapledon, R. G. — On the Plant Communities of Farm Land ...... tiles, Walter. — On the Interpretation of the Results of Water Culture Experiments . Stopes, Marie C. — An Early Type of the Abietineae (?) from the Cretaceous of New Zealand. With Plate IV and seven Figures in the Text ........ Sutherland, Geo. H., and Eastwood, A. — The Physiological Anatomy of Spartina Town- sendii. With seven Figures in the Text ......... Takeda, H. — Dysmorphococcus variabilis, gen. et sp. nov. With fifteen Figures in the Text Scourfieldia cordiformis, a New Chlamydomonad. With five Figures in the Text ............... Some Points in the Morphology of the Stipules in the Stellatae, with special reference to Galium. With twenty-seven Figures in the Text ..... On Carteria P'ritschii, sp. nov. With ten Figures in the Text • Some Points in the Morphology of the Stipules in the Stellatae, with special reference to Galium. (Additional Note.) With seven Figures in the Text see Fritsch, F. E. Welsford, E. J. — Conjugate Nuclei in the Ascomycetes. With four Figures in the Text see Blackman, V. H. West, Cyril. — Stigeosporium marattiacearum, gen. et sp. nov. ...... vii PAGE 215 223 127 359 57 47 25 -xxiv -xxvii 383 485 55r 419 356 525 3L5 605 191 459 467 161 427 hi 333 I5I 157 197 369 601 4i5 357 VI 11 Index. PAGE Willis, J. C. — The Evolution of Species in Ceylon, with reference to the Dying Out of Species. With two Figures in the Text ........... i The Distribution of Species in New Zealand. With a Diagram in the Text . 437 Worsdell, W. C. — The Morphology of the Monocotyledonous Embryo and of that of the Grass in particular. With ten Figures in the Text ........ 509 B. LIST OF ILLUSTRATIONS. a. Plates. Portrait of Professor D. T. G wynne- Vaughan. I. Pisum sativum (Leitch). II. Water cultures (Brenchley). III. Peranema cyatheoides, D. Don (Davie). IV. Planoxylon Hectori, nov. gen. et sp. (Slopes). V. Marah (A. W. Hill). VI. Statice (de Fraine). VII. Drimys (Jeffrey and Cole\ VIII. Hyalotheca dissiliens (Acton). IX. Colloidal Silica (Paine). X. Botrytis cinerea (Blackman and Welsford). XI. Characium Sieboldii, A. Br. (Smith). XII. Pediastrum Boryanum (Smith). XIII. Osmundaceae (Gwynne- Vaughan). XIV. Davallia and Ligodium (Gxvynne-Vaughan). XV. Pityostrobus macrocephalus (Dutt). b. Figures, i. Distribution in Ceylon of the earlier VR, R, and RR species from Trimen’s Flora (Willis) . . . . . . . .12 2. Distribution diagram for the genus Doona (Willis) .... 14 1, 2. Apparatus used in experiments on the influence of temperature on the rate of growth in Pisum sativum (Leitch) ..... 27 3. Diagram showing relation of growth to temperature in graphical form (Leitch) .......... 28 4, 5. Diagrams showing relation of growth to temperature in graphical form (Leitch) .......... 33 6. Diagram showing relation of growth to temperature in graphical form (Leitch) 35 7. Diagram showing relation of growth to temperature in graphical form (Leitch) 36 8, 9. Diagrams showing relation of growth to temperature in graphical form (Leitch) . .3 7 10. Growth curve, Krogh’s curve, and Kuijper’s curve compared (Leitch) 40 1. Recording porometer (Laidlaw and Wright) .... 48 2. Diagram showing result of experiment on Maranta coccinea (Laidlaw and Wright) 52 3. Result of experiment on Phaseolus vulgaris (Laidlaw and Wright) 53 1, 2. Effect upon the stomata of the continued passage of air. Begonia and Eucharis Mastersi (Knight) ... ..... 60 3. Effect upon the stomata of fixing a porometer chamber. Eucharis Mastersi (Knight) ......... 62 4. Diagram showing the form of double chamber used (Knight) . . 65 5. Effect of the intercellular spaces of the leaf upon porometer readings (Knight) ... 67 6. Behaviour of stomata of different portions of a leaf under similar condi- tions (Knight) 72 7. Behaviour of stomata on different leaves under similar conditions ; Knight) . 74 1. Curve 1 (Brenchley) Si 2. Curve 2 (Brenchley) 83 3. Curve 3 (Brenchley) 84 4. Curve 4 (Brenchley) ......... 86 1. Longitudinal section of tip of stem of Hippuris vulgaris (Barratt) . 93 Index. IX Figures. \ 2. Transverse section of same tip at region where longitudinal section ends (Barr att) 3. Transverse section showing innermost layer of periblem after division into inner and outer cells (Barratt) ...... 4. Transverse section showing further addition of periblem cells (Barratt) 5. Transverse section in region of a node, about same level as Fig. 4 (Barratt) 6. Transverse section through older stem (Barratt) . 1. Vertical section through a semi-mature sorus of Peranema cyatheoides (Davie) 2. Diagrams showing development of the sporangium (Davie) 1. Planoxylon Plectori, sp. nov. (Stopes) ...... 2. ,, ,, ,, j, ...... 3- 4* 5- u, t. 1. 2. 3* 4- 1. 2. 3- 1-15. 1-5* 1, 2. 3- 1-8. 9-i i* 12. 13. 14-18. 19-22. 23-27. 1. 2. 1. 2. 3- 4- 5- 6. 7- 8. 9- 10. 11. 1. 2. 3- 4- 5- 6. 7- Planoxylon Lindleii (Witham) (Stopes) . » # 5J >> • • • Wound reactions of petioles of Pteris aquilina (Holden) n Zygnema (Zygogonium) ericetorum (Fritsch) Dysmorphococcus variabilis, Tak., gen. et sp. nov. (Takeda) . Scourfieldia cordiformis, Tak., sp. nov. (Takeda) Anomalies in ovary of Senecio vulgaris (Small) ?> n r> 55 Longitudinal section of upper part of ovule of Larix leptolepis (Doyle Stipules of Galium gracile, Bunge (Takeda) .... Galium paradoxum, Maxim. (Takeda) ..... Galium saxatile, L. (Takeda) ..... . . Asperula asterocephala, Bornm. (Takeda) .... Interfoliar stipules of Didymaea mexicana, Hook. fil. (Takeda) Double and forked stipules of Asperula arvensis, A. sherardioides Galium leiophyllum, and A. aspera (Takeda) Asperula trifida, Makino (Takeda) ...... Marah horridus. Petiole of one cotyledon from split portion of petiol tube (Hill) ...... .... Marah horridus. Unit strand of split petiole with active pericycl (Hill) Potato in hydrochloric acid (Hind) . Potato in nitric acid (Hind) ....... N Potato in nitric acid (Hind) ...... 1000 v ' Potato in sulphuric acid (Hind) Potato in oxalic acid (Hind) ....... Potato in formic acid (Hind) . N Potato in formic acid (Hind) • 1000 Potato in acetic acid of various strengths (Hind) N Potato in acetic acid (Hind) ...... 1 000 N Bean in nitric acid (Hind) ...... 1000 v »» 5* »> ?> ...... Plant of Statice binervosa after shingling (de Fraine) Leaves of S. binervosa. a-c, narrow-leaved form ; d-e, broad-leaved form (de Fraine) Calyces of S. binervosa and S. reticulata ( = bellidifolia) (de Fraine) Bracts and bracteoles of S. binervosa and S. bellidifolia (de Fraine) Transverse section of part of apex of a Main bank plant of S. binervosa (de Fraine) .......... Mucilage gland from base of leaf sheath of S. binervosa (de Fraine) Mettenius gland from the leaf of S. binervosa (de Fraine) page 94 95 96 96 98 102 105 112 n4 115 116 “7 118 1 29 130 >3* 132 13*1 142 146 152 J57 1 9 * 192 J93 J99 200 200 200 204 206 209 220 220 225 226 227 228 229 230 230 231 232 234 2 35 242 246 247 248 252 253 255 X Index . Figures. page 8. Seedling plants of S. binervosa (de Fraine) ..... 258 9. Transverse section of a root of S. binervosa (de Fraine) . . . 259 10. Transverse section of part of the xyletn of the root of S. binervosa. Main bank plant (de Fraine) ....... 259 11. Diagram of part of a root of a Main bank plant of S. binervosa (de Fraine) 260 12. A = diagram of the transverse section of the root of a Main bank plant of S. binervosa; B = diagram of the transverse section of a root of a plant of S. binervosa cultivated in garden soil from seed (de Fraine) 261 13. Diagram of the longitudinal section through the apex of rosettes of S. binervosa (de Fraine) . . ..... 264 14. Diagram of part of stem of S. binervosa (de Fraine) . . . 265 15. Diagram of part of the leaf-blade of S. binervosa and S. bellidifolia in its middle region (de Fraine) ....... 267 16. Diagram of the transverse section of the middle region of the petiole of S. binervosa and S. bellidifolia (de Fraine) ..... 268 17. Sclereides from the petiole of S. binervosa (de Fraine) . . . 369 18. Sclereides from the leaf-sheath of S. binervosa, narrow-leaved plant (de Fraine) 269 19. Transverse sections of part of the leaf-blade of S. binervosa and of S. bellidifolia (de Fraine) 270 20. Transverse section of the leaf-margin of S. binervosa and S. bellidifolia (de Fraine) 271 21. Upper epidermal cells of S. binervosa and of S. bellidifolia (de Fraine) 272 22. Diagram of part of the inflorescence axis of S. binervosa (de Fraine) 274 23. Part of the inflorescence axis of S. bellidifolia, narrow-leaved binervosa, and broad-leaved binervosa (de Fraine) ..... 275 24. A, Gland in surface view from the inflorescence axis of S. bellidifolia ; B, Part of the epidermis ; C, Stoma of S. binervosa (de Fraine) . 275 25. Transverse section of part of the inflorescence axis of broad-leaved S. binervosa and of S. bellidifolia (de Fraine) . . . .276 26. Diagram of a transverse section of a root of S. bellidifolia (de Fraine) . 277 27. Detail of part of xylem in root of S. bellidifolia (de Fraine) . . 278 28. Transverse section of the stem of S. bellidifolia (de Fraine) . . 278 1. Apparatus for measuring rate of tissue-shrinkage (Delf) . . . 285 2. Diagrammatic representation of cross-section of middle region of leaf of onion (Delf) .......... 289 3. Transverse section of middle region of onion leaf (Delf) . . . 289 4. Longitudinal section of middle region of onion leaf, taken between the bundles (Delf) .......... 290 5. A single cell from a longitudinal section of a turgid onion leaf, show- ing stages in plasmolysis (Delf) . . . . . . .291 6. Diagrammatic representation of transverse section of middle region of dandelion scape at flowering period (Delf) ..... 292 7. Part of transverse section of middle region of dandelion scape taken be- tween the larger bundles (Delf) . . ..... 292 8. Curves comparing effect of 0*18 and 0*73! grm. M. concentration of sugar on the plasmolytic contraction of onion leaves (Delf) . . 294 9. Curves showing effect of temperature on onion in distilled water at 36° C. and plasmolytic contraction at same temperature (Delf) . . 296 10. Curves showing effect of temperature on onion in distilled water and effect of subtonic sugar solution (0*18 grm. M.) (Delf) . . . 297 1 1 . Chart of the course of the shrinkage- time curves of onion leaf at different temperatures (Delf) 298 12. Curve showing the rates of shrinkage of tissue of onion leaf under uni- form external osmotic compression but at different temperatures (Delf) 299 13. Curves showing percentage contraction of dandelions at I9°C. with subtonic solution (0-3 grm. M. sugar) (Delf) .... 301 14. Chart of shrinkage-time curves of dandelion scape at different tempera- tures (Delf) 302 J 5. Curve of relation to temperature of the rate of shrinkage of tissue of dandelion scape (Delf) ......... 304 16. Relation of temperature and permeability of protoplasm (Delf) . 306 17. Course of plasmolysis of elder pith in 0'73i grm. M. saccharose solu- tion at different temperatures (Delf) ...... 307 F IGURES. Index. xi 1. 2. 3- 4- 5- i. 3- 4- 5- 6. 7* 1. 2. I-IO. I-I4. 1-4. 2. 3* 4-6. 7-9* 10, 11. 1-4. 1. 2. 3* 4- Transverse section of leaf of Pherosphaera Fitzgeraldi (Groom) Phylloglcssnm Drummondii, Kunze (Sampson) 99 99 99 99 99 >9 99 9 9 99 9 9 99 9 9 99 99 19 99 8 Physiological anatomy of Spartina Townsendii (Sutherland and Eastwood) 99 99 99 99 99 99 9 9 99 99 99 9 9 99 99 99 99 99 99 99 99 9 9 99 First specimen of vegetative abnormality on the scape of Plantago lanceolata (Doyle) Second specimen of abnormality on Plantago lanceolata (Doyle) Carteria Fritschii, Tak., sp. nov. (Takeda) .... Chlamydomonas sphagnicola (Fritsch and Takeda) Illustrating successive changes in the division of the chromatophore and pyrenoid (Acton) ....... Infection by Botrytis cinerea (Blackman and Welsford) Ascophores of Roesleria pallida on the roots of Willow (Bayliss- Elliott and Grove) Vertical median (microtome) section of an ascophore, showing the hemispherical hymenial disc (Bayliss-Elliott and Grove) Section through ascophore of R. pallida after the older paraphyses forming the peridium have been brushed away (Bayliss-Elliott and Grove) Portion of hymenium showing ascospores ; anastomosing paraphyses forming the peridium ; asci containing ascospores (Bayliss-Elliott and Grove) ........... A young ascus, an ascus dehiscing, and a young paraphysis; spores; a very young ascophore (Bayliss-Elliott and Grove) Pilacre faginea ; P. Petersii (Bayliss-Elliott and Grove) Conjugate nuclei in Botrytis cinerea (Welsford) .... Transverse section of Pinus Jnops, showing ‘ Sanio’s bars’ crossing xylem, cambium, and phloem (Rushton) ..... Radial section of P. Inops, showing bar of Sanio crossing xylem, cambium, and phloem (Rushton) ...... Series of tangential sections passing from cambium through the tracheides (RUSHTON) Series of tangential sections, showing bar in section in cambium and through the two tracheides nearest to cambium (Rushton) . Diagram illustrating distribution of species in New Zealand (Willis) . Outline drawings of pyrenoids showing their irregular contour (Smith) Uninucleate cells showing that the relative position of nucleus and pyrenoid is not constant (Smith) ....... Young colonies showing that the nucleus and pyrenoid of a cell are not definitely located with respect to the margin of the colony (Smith) Portions of cells showing irregularly shaped pyrenoids (Smith) Outline drawings, at three-minute intervals, of the changes taking place in the first few minutes after the cessation of zoospore movement (Smith) Surface view of a colony showing 16-nucleate cells adjacent to 2- and 4-nucleate ones (Smith) .... . . Abnormality in the stem of Plelianthus annuus (Barr ATT) 9 9 9 9 9 9 9 9 99 Lygodium scandens (Gwynne-Vaughan) Lygodium japonicum (Gwynne-Vaughan) 99 # 99 99 Lygodium dichotomum (Gwynne-Vaughan) Lygodium volubile (Gwynne-Vaughan) Davallia fumarioides (Gwynne-Vaughan) •n i> >> )1 J5 PAGE - 3L3 ■ 3T7 319 320 321 323 337 339 34i 344 345 347 347 353 354 370 375 382 395 4° 7 407 408 410 412 413 416 422 422 423 423 442 461 465 469 470 474 476 481 482 483 497 498 4 99 5°° 501 502 504 5°5 XI 1 Index. PAGE Figures. i. Zea Mais (Worsdell) . 509 3. Grass-cotyledon, showing successive developmental stages (Worsuell) 5 1 1 4. Zizania aquatica (Worsdeli.) . . . . . . . . 5 1 2 5. A, Pan i cum miliaceum ; B, Oryza saliva (Worsdell) . . . 313 6. Hordenm vulgare (Worsdell) . . . . . . . .515 7. A, Stipa arenaria ; B, Elensine coracana (Worsdell) . . . 516 5. Agapanthus umbellatus (Worsdell) . . . . . .520 9. Cyrtanthus sanguineus (Worsdell) . . . . . .521 jo. A and B, Heterachta (Commelvnaceae) ; C, Tinnantia (Commelyna- ceae) ; D, Tamus communis (Dioscoreaceae) (Worsdell) . . 522 1. Perianth segment of abnormal flower (Salisbury) .... 526 2. Anemone nemorosa, var. robusta (Salisbury) ..... 526 3. Anemone nemorosa, var. apetala (Salisbury) . . . . .527 1. Pityostrobus ovatus (Dutt) ........ 534 2. ,, < . * * • • • • • • 5 36 1. Diagrams illustrating modifications of the diagonal type (Davey) . 578 2, 3. Myrica californica (Davey) . . . . . . . . 581 4, 5. Myrica Gale (Davey) . . . . . . . . *582 6. Juglans nigra (Davey) ......... 583 1 • 35 33 33 ......... 584 8. Juglans Sieboldiana (Davey) ........ 585 9-11. Alnus cordifolia (Davey) .... ..... 588 12, 13. Carpinus Betulus (Davey) ........ 589 14, 15. Castanea sativa (Davey) ......... 592 16, 17. 3 3 33 33 • • • • • • • • • 593 1 8. Seedlings of Amentiferae (Davey) ....... 597 28. Galium kamstchaticum, Stell., a hirsutum, Takeda (Takeda) . . 602 29, 30. Galium kamtschaticum, Stell., 0 oreganum, Piper (Takeda) . . 602 31, 32. Rubia grandis, Kom. (Takeda) ....... 602 33, 34. Asperula odorata, L. (Takeda) ....... 602 A. Diagram constructed from serial transverse sections, to show the sporeling and the prothallus in vertical longitudinal section (Sampson) 606 B. Transverse section through the sporeling and part of the pro- thallus (Sampson) 606 Annals of Botany, vol. xxx David Thomas Gwynne- Vaughan. With Portrait. DAVID THOMAS GWYNNE-VAUGHAN 1 was born on March is, 1871, at Royston House, Llandovery, his mother’s home. He was the eldest child of his parents, Henry Thomas Gwynne-Vaughan, of Cynghordy, and later of Erwood Hall, Breconshire, and Elizabeth, second daughter of David Thomas, of Royston House, Llandovery, who died in 1874. His father, who had two daughters and a son by a second wife, died in 1890. The Vaughans are an ancient Welsh family, descended from Sir Roger Vaughan, who was killed at Agincourt, and was one of the ‘three valiant Welshmen . . . who had rescued the King, and were knighted by him as they lay bleeding to death Going back to more mythical days the Vaughans trace their descent to Cradoc of the Strong Arm, one of the Knights of the Round Table, and further back still. In remote days the G Wynnes of Cynghordy and the Vaughans are said to have had a common ancestor in Aulach, great-grand- father of Cradoc, who again was the descendant of Gwarldeg, King of Garthmadryn (now Brecon) (a. d. 230 ?).2 That extraordinary person Thomas Vaughan, the Rosicrucian and alchemist (1621-65), belonged to another branch of the family. David Gwynne-Vaughan attended a preparatory school at Kington, Herefordshire, and later, the Monmouth Grammar School. In the October term of 1890 he entered Christ’s College, Cambridge, with an exhibition from his school, and in the following year obtained a scholarship in Science from the College. He was thus a member of Charles Darwin’s college, an association which, as it seems to me, had a special appropriateness. He took a First Class in Part I of the Natural Science Tripos in 1893. Gwynne-Vaughan did not go on to Part II of the Tripos, and thus had no opportunity at Cambridge of showing his real powers. During the year after he left Cambridge, he was engaged in teaching, as Science Master at a school. 1 I am indebted to Mrs. H. C. I. Gwynne-Vaughan, D.Sc., F.L.S., for particulars of her husband’s life and family, and for much kind help in other ways. 2 I understand that much more information about these old families is to be found in Jones’s History of Brecknockshire and Nicholas’s Annals and Antiquities of the Counties and County Families of Wales. [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] A* ii Obituary .—David Thomas Gwynne- Vaughan. I first became acquainted with Gwynne-Vaughan in the autumn of 1894. It was at the invitation of the Director, Sir W. T. Thiselton-Dyer, that he came to work at the Jodrell Laboratory, of which I was then Honorary Keeper. After a couple of preliminary visits he started work on Oct. 12, beginning, by way of practice, with an examination of the roots of Pandctnus . We used to go round the houses and pits together, and he soon became familiar with the treasures of Kew. He made a good many observations on Cycads, and took part in the hunt for centrosomes, on which in those days we were all bent, led by the flattering tale which had been told so well. Gwynne- Vaughan’s skill as a microscopist soon became evident, and from the outset he was very keen. For the first three months or so his work was rather miscellaneous — he was getting into training. In January, 1895, he began to settle down to definite research. His first subject was the anatomy and morphology of Nymphaeaceae, on which he made some curious and interesting observations, afterwards embodied in his Notes of 1896 and paper of 1897, referred to further on. Kew . provided him with some specially interesting material for his work ; in particular, a whole series of seedlings of Victoria regia , at suc- cessive stages, was placed at his disposal. In suggesting the group for investigation, I had specially in view the peculiarities of its anatomy, to which Gwynne-Vaughan was soon able to add some new and unexpected features ; at the same time his attention was further directed to points in the external morphology A little later, in May of the same year, he took up another anatomical subject, the structure of polystelic Primulas ; as the result of his observa- tions he was able to make a considerable advance on the views of Van Tieghem. .About this time there was some question of his going on an expedition to South Brazil, but it was not till a couple of years later that his hopes of tropical travel were realized. In September, 1895, Gwynne-Vaughan attended the meeting of the British Association at Ipswich, the first at which a separate Section for Botany was constituted. On October 8, 1895, W. H. Lang (now Professor of Cryptogamic Botany in the University of Manchester) first came to the Jodrell Labora- tory. This was the beginning of the close and life-long friendship between him and Gwynne-Vaughan. In November of that year Gwynne-Vaughan showed me the draft of his Nymphaeaceae paper, the first written product of his work ; it was considerably extended before publication, and some of the most interesting results were recorded meantime in two Notes in the ‘ Annals of Botany ’ (1896). Obituary —David Thomas Gwynne- Vaughan. iii Lang was working at apogamy in Ferns, and on August 20, 1896, made his astonishing discovery of the occurrence of sporangia on the prothallus. Gwynne- Vaughan was present, and when they told me what Lang had found, I did not believe a word of it, but thought they were ‘ chaffing \ However, it was quite serious, and within a few days another case of the same kind came to light in a different Fern. I mention this here, because Gwynne-Vaughan and Lang were working in the same room, and the former was just as keen on Lang’s results as the discoverer himself. In September, 1896, Gwynne-Vaughan was present at the Liverpool meeting of the British Association, and this visit had an important influence on his after career. He read a paper on ‘ The Arrangement of the Vascular Bundles in certain Nymphaeaceae ’. Professor Bower was impressed by this ‘ peculiarly lucid preliminary statement and at once offered Gwynne-Vaughan the Junior Assistantship in his laboratory at Glasgow. If Kew may claim to have first started Gwynne-Vaughan on original work, it was at Glasgow that he developed his most characteristic line of research. During the ensuing autumn Gwynne-Vaughan again worked at Kew, completing his paper on Nymphaeaceae and his investigation of the anatomy of Primula. The former was passed for publication by the Council of the Linnean Society on Dec. 17. Next day I said good-bye to him at Kew, and with the New Year (1897) his Glasgow career began. It was not long, however, before his energies were diverted for a time into other channels. He accepted a commission to undertake a journey in South America, up the rivers Amazon and Purus, to report on the Rubber production of the district for a commercial syndicate, starting on Oct. 9, 1897. He travelled 2,500 miles up the great river and its tributaries, and reached the Bolivian frontier. I am permitted to make some extracts from letters written on this and on the subsequent Malayan journey, to his half-sister in England. The first is from a letter, written between Obidos and Serpa, and post- marked 4/12/97 : £ But the journey up the Amazon is the most fascinating thing I have yet experienced. In parts we steam along a biscuit toss from the margin of the stream, which is bordered on either side by a solid wall, 60 ft. high, of virgin, impenetrable Forest of the most luxuriant tropical vegetation. Palms of various kinds stand out in delicate relief against the darker mass of giant trees of India-rubber, Mimosae , Bertholletia , &c., and the very water’s margin is crowded with gigantic plants of the Arum family, reaching even to 20 ft. high. It is vain to attempt to describe the posi- tively wanton prodigality of nature in these regions. Most of this Forest A* 2 IV Obituary . — David Thomas G wynne- Vaughan. is 2-6 ft. sunk in water for the extent of many miles on either side of the free stream.’ ‘ Every common weed in the streets is completely new to me, and the size and beauty of the flowers in the waste spaces are very impressive — every one is new to me, and I am feeling very confused with it all, when I think that the wealth of the Virgin Forest has yet to be en- countered/ The next extract is from a letter written just after his return to Kew : ‘ For a description of the journey along thousands of miles of river, flanked by ancient Forest over ioo ft. high, the strange beasts in the water and on the land, the occasional settlements we met with, and their in- habitants, the Indians almost white and quite pleasant fellows that we came across, the miles that I have walked in the half-flooded, damp, and heated “Gapo” or low-lying Forest, with the thermometer at 90° in the shade, alone with a single Indian guide and a flask of the local “ 40-rod exterminator ” — of all these anon. Of snakes and sun — but as Rudyard would remark, “ That ’s another story ”. These I must keep until the fortunate hour in which we next meet. ‘ Myself and my companion were travelling against time, so I’m sorry to say my botanical results were only meagre, and I am haunted by an appropriate quotation : “ Ah, fool was I and blind ; The worst I stored with titter toil , The best I left behind.” I’m going to fire this off at a good many ; an apt quotation is the best sort of excuse.’ On Sept. 1 6, 1898, he was back in England, and returned to Kew to work for a time at the laboratory again. During this visit he was doing the anatomy of L ox soma, and afterwards of other Ferns, for his solenostely papers. We had many discussions on questions of stelar morphology and the comparative anatomy of Ferns. He stayed at Kew over Christmas, leaving on Jan. 3, 1899. The following month he started on his second tropical journey, that to the Malay Peninsula. It is characteristic of him that the interval between these two adventurous expeditions was filled up with regular laboratory work. From The Times , Spring, 1899: ‘ Cambridge Expeditions in the Far East . ‘An expedition, under the leadership of Mr. W. W. Skeat, left Cambridge a few days ago for Bangkok. The members of the party include Mr. Gwynne- Vaughan, of Christ’s College, and Messrs. Evans & Annandale, of Oxford, and it will be reinforced at Singapore by Mr. Bedford, of King’s College. V Obituary . — David Thomas Gwynne- Vaughan . The object of the expedition is to make a scientific survey of the little- known country lying south of Siam and north of the protected States of the Malay Peninsula.’ As his experiences on the Malayan trip were of considerable interest, rather full extracts from his letters, above referred to, are given. Approaching Singapore , March 8, 1899. ‘ I dread Singapore ; we have an appalling number of engagements to make there, in the least possible time. If nothing else, you have a peace- able time at sea, a time that I love, though Malayan has done its best to spoil this trip as Portuguese did my former one ; still I live in hopes of returning. If I do I am afraid this will upset the last chance of settling down and quiet work in England. I greatly fear the East, for was I not born a tramp- Royal ? ’ As we shall see, this anticipation was very far from being fulfilled. Bangkok , Siam , March , 1899. ‘ I have dined at a Siamese Prince’s, Prince Naret, in dress clothes and a stick-up collar, eaten a splendid European dinner, conversed with his wife, who is a delightful little person, but she chews betel to such a degree that an average European would have a fit, and talked over the odds on the ’Varsity boat-race with his son, Prince Charoon, an old Cambridge man and a very nice fellow too, while looking on at a Siamese dramatic per- formance commanded at the Prince’s house for our benefit, and that of the officers of a British warship at present in port. ‘ Last night we dined on board the same, after having spent a long day at Ayuthia, the former capital of Siam, a wonderful place, a city which formerly contained about a million inhabitants, but now only a few majestic ruins scattered among a jungle only penetrable with difficulty. Huge palaces and temples, an image of Buddha 40-50 ft. high, and sitting at that, alone, huge, stately, and impassive in its vast unroofed and crumbling temple, while the irresistible jungle is steadily covering, destroying, and breaking down all things around it. I now know, as few do, what is meant by the “ The Letting in of the Jungle ”. I stood in the hand of this gigantic idol and declaimed “ The Karela, the bitter Karela”, &c. ‘ The Malay is grievously misjudged in England. I distinctly like the race. They are a folk who are proud and sensitive, the only race in the East who look you straight in the face when they speak to you. In their manners they are naturally gentlemen. They admit a possible superiority in Englishmen, but hold all the Eastern races in strong contempt. It is true that they have some excusable little eccentricities, such as running amok and using their kris too freely, but then no race is altogether perfect.’ vi Obituary . — David Thomas Gwynne- Vaughan. S in g or a , A pril , 1899. £ When we landed at Singora we arranged for some boats and a dozen Siamese boatmen to take us and our band of brigands up the Ta Lei Sap, a large inland sea here which only two Europeans have ever seen before. This we did, thanks to our credentials from the Siamese Government, pretty comfortably, and having reached a village half-way up we started our collections. Then the party divided. I and another one, Evans, started in the boat, and the other two went inland to investigate a tribe of very primitive savages with peculiarly repulsive and interesting methods of burying. I and Evans went in search of another lake to the north of Ta Lei Sap, called Ta Lei Nawi, found and entered it, being the first Euro- peans to see that fine sheet of water. That night I attained the Ultima Thule of a traveller’s desire, that is, I reached a spot where I was the very first white man to tread the soil. I should like to draw you a picture of our evening meal, after thirteen and a half hours’ fast too, as we sat and wolfed, with a ring of about seventy natives seated around at a respectful distance, and observing us with interest and sundry guttural remarks. My gun caused considerable comment next day (I do the shooting of the Expedition, and do it extremely well). N.B. I have had a lot of luck in shooting lately, and must “ blow ” a bit. I had some difficulty in doing much there, because the whole village insisted on accompanying me, and made enough noise to frighten all the game into the next continent.’ Biserat in Jalor (undated). ‘ Here I have made the acquaintance of a most villainous old brigand in a village near a mountain just here ; he has instituted himself as my companion and mentor in shooting excursions and ascents of the hill, and he has just shown me, as a great treat, the entrance of a cave which runs direct into the hill. We have been yearning for caves for some time, and I am very pleased, because the very next day I got torches and, accom- panied by my hired assassin, we have explored it. He, the betel-chewing old assassin in question, is the only man here who dares enter, because there are a large number of exceptionally ferocious and horrible “ Hantus ” (Demons, or Devils, mixed up a good deal with ghosts). A strong wind comes out from the mouth of the cave, which is their breathing. The brigand did his utmost to dissuade me for my own good, advancing numerous weighty reasons, such as that the cave ended in a pit which dropped through to the infernal regions, and that it was not the right time of year. Somewhat incongruous, isn’t it ? He only screwed himself up to accompany me by the offer of bullion, and by assurances that I had paid special attention to cave-demons, and was capable of coping with any likely to appear ; indeed, I openly confessed to intimate acquaintance with several cave-devils of the higher class. Further, he besought me to bring a little . Obituary .- — David Thomas Gwynne- Vaughan. yii stick (very ordinary stick) that I have had given me, and which is uni- versally known to be peculiarly efficacious for almost all kinds of Hantus. This reads funny, but the old boy genuinely believes it, or fancies he does. We said charms before we entered, both of us. Mine (the genders of Latin nouns ending in -is) pleased him vastly, and he would very much like to learn it, especially when I told him that it could be used in other cases too, as well as cave Hantus, especially when prefaced with “ Feminine -do -io -go -is, -as -ans and -x, will show -Es, if no increase be needed, -s by consonant preceded ‘ The cave turned out to be really fine, with magnificent Cathedral-like domes, far into the mountain. Since then I have found a simply majestic Cathedral cave, with a small opening at the top and others imitating clerestory windows, the floor strewn with huge stalagmites imitating white marble tombs. I have never seen anything like it. I also found a deep pot-hole, which I descended by a rope. I got about 150 ft. down, when it stopped slanting and dropped sheer. I laid myself out on a tooth of rock over- hanging the black mouth of the pit and dropped [a stone?]. I counted ten before it struck bottom, and simultaneously became nervous about the stability of the jutting rock, thought of Rider Haggard, and precipitately retired. I was very, very careful in climbing back up, and I told the brigand that the Hantu at the bottom of this cave was a very big and dangerous one indeed. He instantly agreed and said that he had seen him once, something like a goat and something like a man, and very big. I said he was quite right, and that we’d better leave him alone till I’d worked up a suitable charm to floor him with. Apart from this personal interest, the caves have a scientific one, for we have discovered a “ cave fauna” in them, about twenty species, including a snake which we caught in a butterfly net, yet he was 6 good feet long, and probably very poisonous. All these never saw the light, being sightless. They feel their way in con- tinual darkness, existing in a strange little world of their own. Highly interesting, isn’t it ? ‘ I have been deer-hunting, pigeon-decoying, bull and cock fighting, and all sorts of weird things here. I hope, when Skeat arrives, to get a wizard I am acquainted with, to give an exhibition of devil-raising. I do not say these things flippantly. There are some very strange things out here among the jungles, cliffs, and caves.’ Kota Bharu , in Rahman , c. June 15, 1899.. ‘ I like the Malays very much indeed ; we have long talks at night, and they tell me all the ghost stories, and about “ Hantus”, Spirits, and Jinns, and all sorts of Devils. They have an enormous mythology which they really believe, with the weirdest and most incongruous conceptions imaginable. viii Obituary. — David Thomas Gwynne- Vaughan. ‘ They admit, and really believe, that I am an advanced devil doctor (Pawang) myself. I barefacedly claim this honour. I have been away in a Malay hut, half a dozen waist-bare folk variously squatting on mats, in a closed space some 8 foot square, lighted by a couple of smoky dammars, all of us smoking or chewing betel, and having chanted and tambourined our nerves into a high state of tension, I have seen a brother Pawang drum himself into a positive epileptic fit, uttering short screams and mirthless laughs, throwing about his head, and twisting his limbs in a most unearthly manner, while we rhythmically drummed and incanted all the faster. The statements he made when in the fit were acted upon with the intention of amending the sick person whom we were professionally attending. It is known that devils had entered into the Pawang while in that state, and while he is coming to they speak through him. I must say he made a tolerably close approach to my preconception of demoniacal possession, but when the devils spoke in friendly and appreciative terms of myself, I was forced to confess it was not the real thing. They said that they had informed me of the best medicine. I stated that I had been informed, and that it was well. Then, since the person had dysentery, I administered chlorodyne, produced a rapid cure, and by the Pawang himself (who really seems half to believe it all ) I am believed to control devils of extraordinary capabilities. I have been told that there are folk who would very much like to see my devils enter me. But I refuse because I am aware that I could not anything like approach the performances of the native pro- fessionals. I explained that my “ Hantus ” are exceedingly shy (“ malu ”) of entering any one, and they only enter me under the most persuasive circum- stances. Further to keep up their reputations, that if they, by mistake, entered any one else, they would almost certainly “ bust ” him. ‘This may appear rather nonsense to you, but it is good sound, rational, and believable stuff out here, based on a knowledge of the popular opinions and surmises. It isn’t Alice in Wonderland, though it is rather like.’ Kelantan River , July 26, 1899. ‘Twice Hadji Sirath and I have had to work the raft on alone, half the time wading and half punting it along. The river is full of rapids ; several times I have passed over rapids or waterfalls of 4 foot fall in almost as many yards, the passage being between rocks hardly apart enough to allow the raft to pass ; twice indeed the bamboos of the raft touched on both sides, and once indeed even mounted the rock at one side. Between the wild swirl and rush of the waters, the swift descent of the raft twisting and turning with all its bamboos creaking and splitting, our cries of encourage- ment and warning, the waving of poles and the splashing of water, these rapids make the most exciting sort of time in their shooting. I would like to have a photograph of our descent immensely. If we made a really bad IX Obituary .- — David Thomas Gwynne- Vaughan. i shot, the rafts would come instantly to pieces and all my kit would probably be ruined. No other kind of craft would stand it. These rafts are only made of a few bamboos lashed together by rotan, and they bend like fishing rods before they break. Twice we have had to moor out our kits and descend in the raft alone, because it was too risky. ‘ Hadji and I sleep together at night on the raft, wherever night over- takes us, which it has up to now in the middle of the jungle. When it rains we put up my waterproof sheet, and go on getting wet as before. You will hardly believe it that, in spite of all this, I am in uproarious good health, and am enjoying it immensely, although I am frightfully savage at having the rest of the trip spoiled.’ ‘ Now Tomoh is a gold country, worked by a Chinese colony, and one day on the march through it we changed our path from the bed of a large stream to a path (hardly better) through the jungle, when after a while we were petrified by meeting in the midst of the forest a little locomotive engine, and one or two trucks with English gold-mining machinery in them, all broken and decayed, the plates of the engine all burst and broken in all directions, but with its little rusty smoke-stack still erect. There it stood, as though suddenly slain in the midst of action, and now the forest had rapidly overgrown, crowded in and suffocated it, a stout liane twining round the smoke-stack painfully suggesting strangulation. The trail of the white man stood before us, the little engine looked pitiful and melancholy, and it felt as though we were deserting a disabled comrade in the face of a relentless and cruel foe, when we went on, leaving the work of our hands alone again and utterly desolate, to suffer the revenge of the violated jungle. The tale came out later on. Some time ago six “ Orang putih ” — whites — came here to work for gold. They stayed three years, they found none worth working, and then three of them went back, two died of fever and one by the fall of a rock. They must have been pretty good men because they left a good reputation among the Malays, who speak well of them.’ ‘ This morning, after great difficulty in obtaining two men, we floated down on our raft very slowly, using paddles, the water now having become too deep to punt, until I got to Kluat a village. There I managed to buy a small prahu, a much swifter craft. Now there are no more rapids, the river being a fine one, about 1 50 yards wide, running through a hilly and densely afforested country, every now and then coming to a solitary house or Kampong. It is really excessively beautiful, and not in the least mono- tonous, as was the Amazon. I also managed to get three men to paddle, and actually got them to work nearly all night, chiefly because I could not sleep myself, since the boat leaked, and everything got soaking wet, to say nothing of the rain. We have made excellent progress, and I expect to get to Kelantan by night. One of the men I picked up is a certain Hadji Said, x Obituary. — David Thomas Givynne- Vaughan. a wicked old chap who has helped me a great deal, and undoubtedly stolen my pyjama coat. We have become great friends, although he prays fervently and by mention each night, that my head may burn violently in Jehanum. Islam is a frightfully uncompromising religion.’ The following extract from Gwynne-Vaughan’s Journal gives an account of an adventure in the Malayan forest : Near Tremangan. 1 The jungle consisted of fairly high trees, and below a swamp, thick-set with Pandanus ; we had to cut every yard of our way through. I had Akib and Uda with me, and we took turns at leading the way and using the parang. We were waist-deep in very malodorous swampy mud, swarming with leeches, trying to stand on Pandanus stumps and crawling along fallen and rotten tree-trunks, every now and then missing our footing and slipping into the horrid, repulsive mud over our middles. The Pandanus leaves cut our hands, lianes tripped us up, and the forest mosquitoes and leeches did their utmost to add to our discomfort, and a very melancholy spectacle we made when we got out into a clearing. ‘ One afternoon while here Akib and I got into a nasty adventure. While in a large padang, about its centre, we looked up and saw the leader of a herd of krebau, which had been grazing in the distance, had approached us uncomfortably near and looked distinctly uneasy. We tried to scare it off without result ; at last it stamped its foot impatiently, Akib yelled out •“diamon terkan” and lit out for the nearest jungle in fine style, and the krebau made for us ; the whole herd, which had quietly crawled up close, also charged at his heels. ‘ I did a lot of quick thinking, during which I recollected I had only one shot cartridge in my gun, and then I took after Akib, being just in time to see him take a flying leap into the jungle. I put in a record ioo yards, getting home a little to the right of Akib, who, as I shot past, I observed was shinning up a tree as hard as he was able. I went on through the jungle, once nearly knocked down by my collecting tin getting stuck between two trees, and fortunately got through into another padang through which I cut at half-mile pace only. At the edge of the next jungle I stopped, and Akib shortly joined me, saying that the krebau had followed me and had not gone through far. I would have much liked to climb a tree myself, but I had the gun to look after.’ Gwynne- Vaughan was back in London on September 26, 1899, having been absent a little over seven months. Caledonian Hotel , Strand , September 26, 1899. J I am glad to get back of course . . . but really I feel strangely that I am an outsider looking on at the show, and hardly having any definite Obituary. — David Thomas Gwynne- Vaughan. xi part in it to play, my interests being still right away back with the honest, clean, and simple savagery of the Malay/ H is friend Lang, who was himself in the Malay Peninsula a year or two afterwards, writes to me as follows on the subject of Gwynne-Vaughan’s travels : ‘ The expedition, as you know, had a pretty adventurous time in the unexplored Siamese States as well as in better-known Eastern States of the Peninsula. The experience made a deep impression on Gwynne-Vaughan’s life and thought, but not on his work, which he continued on the straight line on his return. Apart from collecting for the expedition and loyally handing over all the results, he drank deeply of the life with the Malays, who were after his heart in many ways. The Malay attitude to many things was always with him afterwards, and since I was there, a little later, we kept up an interest in the subject and for a time hoped to return. Malay proverbs and phrases were always coming from him as the appropriate expression of his thought in circumstances little dreamt of in the East. ‘ The Amazon expedition was a much rougher and less scientific experience, but he was younger, and got great joy of adventure in pulling through tight places in which the whites were often more savage than the natives. The experience appealed to his joy in adventure, which was always a current parallel to that of his almost too conscientious discharge of the duties of his various posts. { This side of him came out in his fishing expeditions to out-of-the-way spots in Ireland and the Hebrides, where he loved to meet with all sorts and conditions of men, and not to be suspected of having been “ academic There is nothing to add to these words, which seem to me to express perfectly the significance of Gwynne-Vaughan’s journeys in his life. After he returned from the Malay expedition he took up the thread of his work, where he had dropped it, and his most important publications date from the period which then opened. These will be considered after- wards ; his career may be first shortly sketched. He remained at Glasgow till 1907 — ten years in all. He assisted his Professor in preparing the second edition of the ‘ Practical Botany for Beginners ’, which appeared in 1902 under their joint names. I believe this was the only work in book form in which he took part. In 1904 he became lecturer in Botany at Queen Margaret’s College, Glasgow, in addition to his other University duties. . In 1907 his long and happy association with the University of Glasgow came to an end, and he accepted the post of Head of the Department of Botany at the Birkbeck College, becoming a Recognized Teacher and Internal Examiner in the University of London. He did not, however, remain for very long in the ranks of London botanists, for two years later, in 1909, he became Professor of Botany in the Queen’s University at xii Obituary, — David Thomas Gwynne- Vaughan. Belfast, a post which he held for five years. On his appointment he wrote to me : ‘ It will be a great change at first from the “ intensive’’ kind of work we do here in London. In spite of the much too hard work I have had to do here, I shall leave London with great regret. I am very fond of the South.’ He speaks also of the ‘ keen and appreciative students ’ he had at the Birkbeck. Although somewhat isolated from his botanical colleagues, he was happy in his relations with his fellow professors at the University of Belfast. He accomplished much good work while there, especially continuing the joint research with Kidston on the Fossil Osmundaceae, begun at the end of the Glasgow period. On December 7, 19 11, he married Dr. H. C. I. Fraser, the distinguished cytologist, who had succeeded him as Head of the Department of Botany at the Birkbeck, a post which she continued to hold during their married life and still occupies. In July, 1914, to the great joy of his friends in England, he accepted the chair of Botany in University College, Reading, which unhappily he was only destined to hold for about a year. At Belfast the teaching work had been much concentrated in the Summer term, and was then very hard ; at Reading it was more spread over the whole year, and would probably have proved less exacting when he had once got settled in the new position. He had long been seriously troubled with neuralgia, though few of his friends were aware of it. In spite of failing health, he was able to carry out to the end the duties of his first academic year, completing the work of the Summer term, though with some difficulty. During the vacation he became seriously ill, and, after some fluctuations, the end came on September 4, 1915, the immediate cause of death being a rapid onset of tuberculosis. His friend Lang, who was with him almost at the last, says : ‘ He was very weak and changed, but, when he talked, absolutely himself. It was the bravest end to a long fight.’ I saw him last myself, about the end of January, when he was in his usual health, and we had a long and lively talk, partly on botanical subjects, especially that strange fossil, Tempskya , and partly on the war. His position as a botanist, and the great services to science which he was able to render in the course of a too short life, will best be considered after a brief survey of his published work. Before going on to this, a ofew words may be added about his relations to the British Association and other Societies. When the Association met at Glasgow in 1901, Gwynne-Vaughan was one of the Secretaries of Section K. He was again Secretary at Winnipeg, Sheffield, and Portsmouth (1909-11), and at Dundee and Birmingham Obituary . — David Thomas G wynne- Vaughan. xiii < n (1912-13) he was Recorder of the Section. He contributed greatly to the success of the Botanical Section while an officer, and would have made an excellent President if he had lived. In 1909 he received the MacDougall Brisbane Medal for Research from the Royal Society of Edinburgh, and in 1912 became a Member of the Royal Irish Academy. In 1914 he was appointed External Examiner to the University of Glasgow. He was elected a Fellow of the Linnean Society in 1907, and was a member of Council at the time of his death. Published Work. Gwynne-Vaughan’s first publication was a Note in ‘ Science Gossip ’ (June 1894) on the rare Crucifer, Arabis stricta , Huds., the Bristol Rock Cress. He had found it on an ancient camp, near Llandrindod, Radnor- shire, and sent it to the Editors, who confirmed his identification, adding that ‘ any records of so rare a plant are particularly interesting ’. So far as I know, it had previously been recorded only from Clifton and Cheddar, and its speedy extinction was anticipated by Bentham and Hooker. The first published result of his research work at Kew was a Note in the c Annals of Botany ’ (1896. 1), ‘ On a New Case of Polystely in Dicotyle- dons \ At that time we were very keen on polystely and such phenomena, under the influence of Van Tieghem’s anatomical conceptions. Gwynne- Vaughan showed that the tuber-bearing stolons in certain species of Nymphaea have a clearly polystelic structure, which he compared with the structure of the main stem in the related genera Cabomba and Brasenia. What was still more important in those days, he was able to demonstrate, within the family, a complete series of transitions between polystely and astely. The paper read at the British Association that year, which so impressed Professor Bower, covered nearly the same ground, but added the new fact that ‘ root-bearing steles ’ occur in the rhizomes of Victoria and Nymphaea spp. (1896. 2). The full paper 4 On the Morphology and Anatomy of the Nymphaeaceae’ was published in the Transactions of the Linnean Society (1897. 1). I had to read it for him (on Feb. 18) as he was then busy at Glasgow. It is a rather miscellaneous paper, for he began with the morphology of the leaf, showing how the ontogeny of the individual adult leaf repeats the successive forms of the seedling leaves, constituting a special case of Recapitulation. He also investigated the origin of the peltate form of leaf. He pointed out the remarkable absence of any primitive stages, whether morphological or anatomical, in the seedling of Nelumbium , so different from the conditions in the Water-lilies. xiv Obituary,- — David Thomas Gwynne- Vaughan. The anatomical portion of the paper is an extension of the results communicated in his preliminary notes, illustrated by excellent drawings. From a purely anatomical point of view, Gwynne- Vaughan’s other Kew paper, ‘ On Polystely in the Genus Primula ’ (1897. 2), *s perhaps more important. He paid much attention to the variations of stelar structure in the same species and in different parts of the individual plant. He was able to show that gamostely is not an advanced condition due to the fusion of steles, but is more primitive than dialystely, and that a gamodesmic structure (with the bundles united, but not completed to form steles) probably preceded either. He followed the anatomical development of the seedling in gamostelic and dialystelic species with much care, and found different modes of transition from the original monostely to the perfect or imperfect polystely of the adult stem. His results marked a distinct advance on Van Tieghem’s interpretations, and had a direct bearing on his own later work on the Ferns. This work, as we have seen, was in full progress during the interval between his two tropical journeys. Publication did not begin till some time after his return from the Malayan expedition, when the first paper on the Anatomy of Solenostelic Ferns appeared, dealing with the remark- able New Zealand Fern, L ox soma Cunninghamii (1901. 1). He revived the term ‘solenostely ’ invented but afterwards discarded by Van Tieghem, in place of the then familiar c gamostely ’, on the ground that the latter term implies a fusion of steles supposed to be originally distinct ; in most cases this is the reverse of the truth, the solenostelic preceding the dialy- stelic condition, as shown by the development of the individual plant. The conception which lies at the root of Gwynne- Vaughan’s Fern anatomy thus follows directly from the results he first attained by his study of the Primulas. The solenostele is defined as a single hollow vascular cylinder, inter- rupted only by the departure of the leaf-traces. The conception is thus narrower than that of gamostely. In Jeffrey’s terminology, Gwynne- Vaughan’s solenostely forms ‘ a special type of amphiphloic phyllosiphony ’, i. e. a vascular tube with internal as well as external phloem, interrupted by leaf-gaps. The comparative rarity of solenostely is pointed out, the Ferns offering the chief examples. The detailed investigation of Loxsoma not only established an excellent type of solenostely, but proved interesting in other ways, especially in the demonstration of two kinds of protoxylem, endarch and spiral in the leaf-trace, exarch and scalariform in the stele. The affinity of the genus was regarded as closest with the Dennstaedtias, while some relation to Gleicheniaceae, Hymenophyllaceae, and Schizaeaceae was also recognized. XV Obituary.— David Thomas Gwynne- Vaughan. The second part of the Solenostelic work, which appeared two years later (1903), had a much wider scope, and extended to Ferns of other than solenostelic structure. It is, in fact, one of the chief existing contributions to Filicinean anatomy. Gwynne-Vaughan showed that the dictyostely (typical polystely or dialystely of Van Tieghem), so common in Ferns, is ‘ primarily due simply and solely to the overlapping of the leaf-gaps in a solenostele ’, though other gaps in the tube may occur, as in Dicksonia rubiginosa . He traced the origin from the solenostele of internal vascular strands, and of internal tubes, such as were found by him in Pteris elata , and were previously known- in Matonia . Incidentally he called attention to a curious historical point. c Having first come to the conclusion that Ferns do not possess true seeds, Cesalpino proceeded to deduce the fact that they cannot possess true stems either.’ This, then, was the strange origin of the belief, so obstinately main- tained, even down to our own times, that the ‘ caudex ’ of a Fern is merely a sympodium of leaf-traces. It also became the source of the f phyton ’ theories of Gaudichand and others. Gwynne-Vaughan showed that the apparent segmentation of certain Ferns is really to be regarded as a late development rather than as a primitive feature. In connexion with this question he investigated the course of the internal strands of Cyatheaceae and other Ferns, and found, in agreement with Trecul, that they are not decurrent leaf-traces, but strictly and essentially cauline. He thoroughly worked out the development of the dictyostele in the young plant of Alsophila excelsa , and showed that after the protostelic stage a core of phloem first appears, then the internal pericycle, next the endodermis, and finally the central ground tissue, the structure thus becoming a solenostele. The leaf-gaps as they begin to overlap convert this into a dictyostele, and ultimately the complexity of the mature structure is attained by the nipping off of the internal strands from the main vascular tube. When lateral shoots are developed they repeat, more or less imperfectly, the ontogeny of the plant as a whole. The paper, in which I have only indicated a few characteristic points, concluded with a theoretical discussion of the theory of the stele. Gwynne- Vaughan strongly inclined to the view that the origin of the central parenchyma in the higher . Ferns was to be regarded as stelar, though he admitted that Jeffrey’s opinion, that it arose from the cortex, was theoreti- cally possible. The diagrams of the stelar structure in the solid are an excellent feature of the plates, and so are the combined photographs and drawings, an under-exposed print being used as a earner a-lucida sketch. The paper is altogether an admirable type of modern anatomical work, xvi Obituary , — David Thomas Gwynne- Vaughan, under the influence of the Theory of Descent — the comparative anatomy of the Darwinian period at its best.1 Going back to the year 1901, there is a Note ‘On the Nature of the Stele in Equisetum ’ (1901. 2), a subject on which he never published in full, though he continued to give it much attention. He regarded the vascular bundle of Equisetum as a compound structure — ‘ of the three strands of xylem present in each bundle of the internode, the carinal strand alone passes out at the node as a leaf-trace \ ‘ So the xylem of the so-called vascular bundle of Equisetum consists of three strands, two of which are lateral and cauline, while the median, or carinal, strand is common to both stem and leaf. The fact that only a small portion passes out as a leaf-trace, and not the bundle as a whole, constitutes an essential point of difference between it and the bundle of a Phanerogam.’ He further pointed out that the lateral xylem-strands in E. giganteum gave a strong impression of centripetal development. He compared the stele of Equisetum with the protostele of Sphenophyllum , and with the various degrees of medullation found in the Lepidodendreae, and suggested that the lateral xylem-strands in the recent genus ‘ may perhaps be taken to represent the last remnants of a primitive central mass ’. His idea was that the lateral strands in Equisetum might correspond to the prominent points of the primary xylem in Sigillaria , in the hollows between which the protoxylems lie. This Note must not be taken as representing Gwynne-Vaughan’s final view of the morphology of the stem in Equisetum, Judging from letters of his (1912), and from manuscripts he left, it appears that while he always maintained his view of the triple nature of the bundle, and of the essentially protostelic character of the vascular system, he was inclined to give up the centripetal development of the lateral strands, and to admit the possible occurrence of leaf-gaps in certain cases. The whole question is an inter- esting one, and it is much to be regretted that Gwynne-Vaughan’s mature views were never fully recorded. In a Note (1902) ‘ On an Unexplained Point in the Anatomy of Helmin- thostcichys Zeylanica\ Gwynne-Vaughan described for the first time the curious canals which lead down through the cortex almost to the stele, one in front of each leaf. He thought it ‘ possible that they represent the last indications of vestigial axillary buds ’, a suggestion which has since been confirmed by the work of his friend Lang.2 We owe to Gwynne-Vaughan (1905. 1) the first description of the anatomy of the new genus of Marattiaceae, Archangiopteris , discovered by Dr. Henry in China. The structure was compared with that of Kaidfussia , 1 The British Association Note (1901. 3) is preliminary to the paper just considered. 2 W. H. Lang, Studies in the Morphology and Anatomy of the Ophioglossaceae. Ill, p. 14. Ann. of Bot., January, 1915, vol. xxix. Obituary. — David Thomas Givynne- Vaughan. xvii both being relatively simple forms anatomically. Interesting conclusions were drawn regarding the anatomical distinctions between the Marattiales and Filicales ; among the characters peculiar to the former group the internal protophloem is the most striking. Another paper of the same date (1905. %) has a curious title : ‘ On the Possible Existence of a Fern Stem having the Form of a Lattice-work Tube ’. The epidermal pockets found in the Ostrich Fern and other species are compared with the endodermal pockets which started the lattice-work structure of the vascular system in so many Ferns. If the epidermal pockets likewise became continuous, the whole stem might assume the form of a lattice-work tube. Both series of changes are due to the in- creasing dominance of the leaf and the leaf-traces. It is an ingenious little paper of an unusual kind. Leaving the joint series with Kidston for future consideration, we will next take Gwynne-Vaughan’s much-disputed work ‘ On the Real Nature of the Tracheae in the Ferns 5 (1908). In this paper he endeavoured to show: (a) That the pits in the Ferns investigated are open. (b) That the middle lamella of the tracheal wall is absorbed, leaving an open vertical passage between the bars of thickening. The investigation was suggested by the appearances observed in some of the fossil Osmundaceae. He summed up his conclusions as follows : ‘ The xylem elements of the Pteridophyta are, for the most part, vessels with true perforations in their longitudinal as well as in their terminal walls.’ ‘ In the Osmundaceae, N ephrodium Filix-mas , and probably others, a special type of vessel occurs which is characterized by the complete disappearance of the primary tracheal wall at certain points, so that the cavities of the pits are vertically continuous in the middle of the wall.’ Gwynne-Vaughan’s results were controverted by Halft in 1910, and by Miss N. Bancroft in the following year. Their work leads to the conclusion that the old view was right, i. e. that the pits in Pteridophyta are, as a rule, closed, and that the middle lamella is persistent. Gwynne-Vaughan was in touch with Miss Bancroft’s work, and helped her with criticism and advice. This is not the place to discuss the question, but I may be permitted to quote the following passage from a letter of Mrs. Gwynne-Vaughan’s, which defines her husband’s position in the matter : ‘With regard to the Fern-tracheae I think I can best summarize my husband’s view by saying that he thought the question required further investigation. I do not think he would ever have undertaken this himself ; it did not interest him as much as the other things he had worked on.’ We have now reached the period of Gwynne-Vaughan’s co-operation with Kidston, a most happy conjunction which gave rise to much good xviii Obituary .—David Thomas G wynne- Vaiighan. work, and especially to the fine series of memoirs on the Fossil Osmundaceae. I am told that Kidston and his future colleague first met at Cambridge, during the British Association Meeting of 1904, and at once became friends. Their joint work began soon after while Gwynne-Vaughan was still at Glasgow, but their first paper was published in 1907, the year he left, and from that time onwards they had to arrange special meetings during vaca- tions, usually at Stirling. A well-known photograph shows them hard at work together in the study at 12 Clarendon Place. The series on Fossil Osmundaceae is of exceptional interest, for the authors trace back an existing family from recent times to the Permian, on sound evidence, mainly anatomical. Nothing quite like this has been done for any other group. There are five Parts, ranging in date from 1907 to 1914, all published in admirable form in the Transactions of the Royal Society of Edinburgh. Part I (K. and G.-V., T907) deals with four species of Osmundites : O. Danlopi and Gibbiana , new species from the Jurassic of New Zealand, O. Dowkeri , Carr., from the Lower Eocene of Herne Bay, and O . skidega- tensis , Penh., from the Lower Cretaceous of the Queen Charlotte Islands. The first three are more or less typical members of the family ; O, Dunlopi approaches Todea , but was found to have a ‘ practically continuous xylem- ring ’ ; the other two are nearer Osnnmda. The Pacific species, O. skidcgate?isis, intermediate in age between the New Zealand and the Herne Bay fossils, showed a much more complex structure than either, or than any living species. The wonderful preserva- tion exhibited the structure to perfection, and the plant proved to have well-developed internal phloem, continuous through the leaf-gaps with the external phloem-zone. In fact O. skidegatensis comes very near the dictyo - stelic ancestor of the Osmundaceae, postulated by Jeffrey. Kidston and Gwynne-Vaughan, however, regarded the plant rather as marking the culminating point in the development of the family. In this first memoir the authors anticipated the discovery of ancestral types with a continuous zone, or even a solid mass of xylem, and suggested a common origin of the Osmundaceae with Botryopteris and Zygopteris. Their anticipation was soon realized ; in their second memoir (K. and G.-V., 1908. 1) they described two species from the Upper Permian of Russia, establishing the new genus Zalesskya for them. In Z . gracilis (Eichwald) they found a wide, continuous ring of wood, divided into two zones, the inner of which consisted of short and broad tracheides. In if. diploxylon , sp. nov., there was evidence that the xylem was solid, the inner zone extending to the centre of the stele. This may also have been the case in the former species, though there the internal tissue had entirely Obituary . — David Thomas Gwynne- Vaughan. xix perished. The rest of the structure, especially that of the petioles, appears to leave no doubt of the relationship to Osmundaceae. In the third memoir (K. and G.-V., 1909) other species from the Russian Permian are described, of which Thamnopteris Schlechtendalii (Eichwald) is much the most important. In this fine fossil conclusive evidence was found for protostelic structure, the inner region of the solid xylem consisting here also of large, thin-walled, reticulate tracheae. A full account was also given of the structure of the leaf-trace, which passes from mesarchy to endarchy in its outward course. A preliminary communication on this point had been published in the previous year (G.-V, and K., 1908. 2). The origin of the endarch from the mesarch leaf-trace was held to be general in the Filicales. In the ‘general conclusions’ of Memoir III the authors laid stress on the approach in the stem of Thamnopteris to the structure of Zygopteris corrugata , and predicted the discovery of a Zygopteris with a solid xylem and central short tracheae ; this was almost exactly realized immediately afterwards by Dr. Gordon’s discovery of the stem of Diplolabis , a Lower Carboniferous Zygopterid. The fourth memoir (K. and G.-V., 1910) is concerned with Osmundites Kolbei, Seward, from the Wealden of the Cape, and O. Schemn itzensis (Pettko) from the Miocene of Hungary. The latter is well preserved, but of ordinary structure. I n O. Kolbei , described externally by Seward in 1907, the pith - was much compressed, but the authors found that it contained undoubtedly tracheal elements ; they add : ‘ There is no doubt whatever that the tracheal elements are true and real constituents of the central tissue.’ They regarded this tissue, therefore, as a ‘ mixed pith ’, essentially similar to that in the stele of Zygopteris Grayi , if. corrugata , &c. Another point of interest in O. Kolbei was that the leaf-traces pass out without immediately causing an interruption of the xylem-ring, i. e. that the leaf-gap only begins some little distance above the outgoing trace. In this memoir the geological distribution of all the forms so far described is tabulated. The general discussion is of great value. The importance of Zenetti’s remarkable work on the structure of Osmund a, written long before there was any knowledge of the fossil history, is fully recognized. The Zygo- pterideae are discussed at length, in the light of Paul Bertrand’s elaborate investigations, and their relation to the Osmundaceae considered. The peculiar Zygopteridean frond, so unlike anything in recent Ferns, is regarded as derived from the type common to the Osmundaceae and the Filicales generally. The possibility is recognized that the Permian genera Zalesskya and Thamnopteris may eventually require a new Order, though on existing evidence they must be included in the Osmundaceae. xx Obituary. — David Thomas G ivy nil e- Vaughan. These four memoirs constitute a complete whole, but the observation of additional forms led to the publication, after an interval of four years, of a fifth memoir (K. and G.-V., 1914). The chief species described are Osmundites spetsbergensis (Nathorst) from the Upper Tertiary of Spitzbergen, and O . Carnieri , Schuster, a Para- guay species of doubtful age (Jurassic — Tertiary). The Spitzbergen species, though it shows no stem, is interesting from the fact that the structure of leaflets and sporangia is preserved, as well as that of the petioles ; it is a modern type, resembling Osmunda Claytoniana in petiolar structure, and O. regalis in the form of the frond. A solenostelic Fern stem was found burrowing among the roots of the Osmundites. O. Carnieri • is more remarkable. It is unusually large, the stem measuring 90 mm. and the stele 35 mm. in diameter. The vascular ring is in the form of several meristeles, each surrounded by its own endodermis, and including a varying number of bundles. The preservation is imperfect, but it is probable that the phloem extended all round each meristele, the structure thus being completely dictyostelic. This form is of great interest in conjunction with O. skidegatensis , and may lead to important conclusions when more is known about it. This completes the main series on Osmundaceae, but there are two small papers by Gwynne-Vaughan alone, dealing with recent members of the same group. The earlier of these, ‘ Some Remarks on the Anatomy of the Osmundaceae ’ (191 1), is concerned with the possible retention of the primitive features of fossil forms in the young plants of recent Osmundaceae. The chief definite character found to be retained is the mesarchy of the leaf-trace, occasionally occurring in early leaves of O. regalis. The ontogenetic evidence also favours the intrastelar origin of the pith and medullary rays. A full and acute discussion of the general question of Recapitulation is of remarkable interest. The final paper on the Order is ‘ On a Mixed Pith in an Anomalous Stem of Osmunda regalis ’ (1914). The abundant medullary tracheae met with in this specimen were evidently formed in response to an injury. The value of the observation depends on the credence attached to Jeffrey’s theory that traumatic characters tend to be ancestral. Accepting this, the evidence of the injured specimen tells in favour of the protostelic theory of Osmundaceous anatomy. The argument is a good one, as addressed to the Jeffrey school, but the morphological value of traumatic characters is still very much of an open question. The Osmundaceae work as a whole is perhaps unique as a study of the comparative anatomy of recent and fossil members of a definite group, ranging over so long a period of geological time. Certain points are still the subject of controversy, and some of the facts may admit of diverse Obituary . — David Thomas Gwynne- Vaughan. xxi interpretations, but the value of the work accomplished by Kidston and Gwynne-Vaughan is solid and lasting. Two more palaeobotanical memoirs by the same authors remain to be noticed. The first of these (K. and G.-V., 1911) gives the first intelligible and satisfactory account of the structure of Tempskya , a group of fossils which had puzzled palaeobotanists ever since the year 1824. Tempskya consists of a dense mass of Fern roots, traversed by a number of larger organs, variously described as petioles or stems. The authors investigated an unusually good specimen of a Tempskya from Russian Turkestan, which they named T. Rossica . They found that it consisted of a number of solenostelic Fern stems, running longitudinally through a dense felt of their own adventitious roots. The stems bear two rows of leaves on one side and roots on the other, thus showing themselves to be of dorsiventral structure. They keep parallel to each other and to the general course of the roots, but face indiscriminately in all directions. They branch dichoto- mously. The authors interpretation of the facts is that the numerous stems with their felt of roots together formed an erect ‘ false stem the individual true stems becoming free at the top and bearing the leaves. A restoration of the plant as it is conceived to have appeared in nature was prepared ; it was not published with the paper, but was shown at the British Asso- ciation Meeting at Sheffield in 1910, and has been reproduced in Dr. Marie C. Stopes’s British Museum Catalogue of the Cretaceous Flora, Part II, p. 15. This interpretation of the structure receives support from the analogy of a recent Tree-fern, Hemitelia cremdata , described by Schoute in 1906; in this strange plant a number of branches are actually felted together by a mass of their own roots, to form a false stem. The excellent paper on Tempskya , as it appeared in the Transactions of a Russian Society, is not very accessible to English readers, but a full abstract will be found in Dr. Stopes’s Catalogue, above referred to. The memoir on Stenomyelon tuedianuni , Kidston (K. and G.-V., 1912), was intended to form the first of a series on the Carboniferous Flora of Berwickshire. So many important forms remain to be described, that it is to be hoped that Dr. Kidston, in spite of the grievous loss of his collaborator, will see the scheme through. Stenomyelon , originally found by Matheson in 1859, and re-discovered by Kidston, who obtained much better specimens, in 1901, is of Lower Carboniferous age, and is a remarkable and isolated form, referred to the transitional group which we call Cycadofilices. The structure is well preserved, and the investigation does it justice. The memoir is clear, full, concise, and perfectly illustrated, and may well serve as a model of structural work in palaeobotany. We have now passed in rapid survey through all the published work in which Gwynne-Vaughan took part. xxii Obituary . — David Thomas G wynne- Vaughan. Two papers, I am informed, were left unpublished, but in a sufficiently advanced state to warrant the hope that they may soon see the light. Their subjects are : ‘ On some climbing Davallias and the Petiole of Lygodimn ’, and ‘ Observations on the Anatomy of the Leaf in the Osmundaceae 5. There is also a considerable amount of manuscript with many Figures, relating to the genus Equisetum , in which he had so strong an interest. The chief part of this material was prepared with a view to a joint work on Equisetales and other groups which Gwynne-Vaughan and I had in con- templation. Its fate is still undecided. Gwynne-Vaughan’s chief subject as an investigator may be described as the comparative anatomy of Plants from the point of view of Descent. His work belongs essentially to the Darwinian period ; if he had lived, it would have been interesting to see how' far he might have been influenced by the modern tendency in the direction of Experimental Morphology. But I think the phylogenetic interest would always have been the main one to him, as it is to the present writer. He became an anatomist at Kew ; it was at Glasgow that he was led to specialize, with such admirable success, on the anatomy of Pteridophyta and above all Ferns. In this subject, so characteristic of modern English Botany, no one surpassed him. The phylogenetic study of Fern anatomy demands a knowledge of fossil forms, and here his association with Kidston was of the happiest influence, and gave rise to the best of his later work. He made the utmost use of his opportunities, and I was often impressed by his own accurate knowledge of fossil types. He had accomplished so much, and his judgement was so thoroughly sound, that great things might have been looked for from him, if all had not ended so much too soon. Gwynne-Vaughan’s original work lay within a well-defined field, but he was actively interested in other branches of his science, and, in particular, was very keen on the Algae, at which he worked hard on his occasional visits to the coast, from Glasgow and elsewhere. With students, who are generally good judges of character, he was, I am told, remarkably successful. At Glasgow, for example, he took a great interest in the athletic life of the University, and was very popular. In all his various posts he established a tradition among his students of trying to obtain first-hand information and to demonstrate things for themselves. Both in the laboratory and on excursions he had the power of inspiring interest and of making difficulties clear. Those who knew him only from his published work, necessarily of a technical character, would have formed little idea of him as he really was. Obituary. — David Thomas Gzvynne- Vaughan. xxiii Like Darwin, he was a sportsman even before he was a naturalist ; we have seen how it was he who did the shooting on the Malay expedition ; he always kept his taste for open-air pleasures. Though so much of his life was spent at Universities, he was very far from having an academic mind ; his love for Natural Science was inborn. This is shown by the way he came to Kew, entirely of his own choice, as soon as he was free to follow his natural bent. Another side of his character is shown by his marked interest in Malayan demonology, and generally in the customs and sayings of a primi- tive people. Here, again, his interest was sympathetic, and not merely that of an anthropologist. Gwynne-Vaughan’s personality was altogether attractive and original, and those who knew him felt that he was somehow different from the usual type of scientific man of his generation. He was a delightful and amusing companion, while at the same time his manner had a pleasant touch of old-world courtesy. He made warm friends, as was natural. Now that he has gone, so prematurely, we feel that, whether personally or scientifically, there could scarcely have been a greater loss to English Botany. D. H. SCOTT. List of Published Researches. 1896. r. 1896. 2. 1897. 1. 1897. 2. 1901. 1. 1903. 1901. 2. 193I- 3* 1902. 1905. 1. 1905. 2. 1908. 1911. 1914. On a New Case of Polystely in Dicotyledons. Note. Annals of Botany, 1S96. On some Points in the Morphology and Anatomy of the Nymphaeaceae. Annals of Botany, 1896. On the Morphology and Anatomy of the Nymphaeaceae. Trans. Linn. Soc. Lond , 1S97. On Polystely in the genus Primula. Annals of Botany, 1897. Observations on the Anatomy of Solenostelic Ferns : Part I .—Loxsoma. Annals of Botany, 1901. Part II. Annals of Botany, 1903. Remarks on the Nature of the Stele of Equisetwn. Annals of Botany, 190!:. Some Observations upon the Vascular Anatomy of the Cvatheaceae. Annals of Botany, 190L. On an Unexplained Point in the Anatomy of Helminthostachys Zeylaiiica. Annals of Botany, 1902. On the Anatomy of Arcliangiopteris Henryi and of other Marattiaceae. Annals of Botany, 1905. On the Possible Existence of a Fern Stem having the Form of a Lattice-work Tube. New Phytologist, 1905. On the Real Nature of the Tracheae in the Ferns. Annals of Botany, 1908. Some Remarks on the Anatomy of the Osmundaceae. Annals of Botany, 1911. On a Mixed Pith in an anomalous Stem of Osmumia rcgalis. Annals of Botany, 1914. XXIV Obituary. — David Thomas G Wynne- Vaughan. Jn collaboration with R. Kidston, Esq., LL.D., F.R.S., F.G.S. On the Fossil Osmundaceae : Kidston and Gwynne-Vaughan. 1907. Part I* — Osmundiles , sp. novae. Trans. Roy. Soc. Edin., 1907. ,, ,, 1908. r. Part II. — Zalesskya. Ibid., 1908. „ ,, 1909. Part 111. — Thamnopteris , Bathypteris , Anomorrhoea. Ibid., 1909. 1910. PartIV. — Osmundiles Kolbei. Zygopterideae. Ibid., 1910. ,, ,, 1914. Part V. — Osmundites spp. Ibid., 1914. Gwynne-Vaughan and Kidston. 1908. 2 : On the Origin of the Adaxially Curved Eeaf-trace in the P'ilicales. Proc. Roy. Soc. Edin., 1908. Kidston and Gwynne-Vaughan. 1911. On a New Species of Tempskya from Russia. Ver- handlungen der Russ. Kais. Mineral. Gesellschaft, Bd. xlviii, 1911. „ ,, 1912. On the Carboniferous Flora of Berwickshire. Parti. Stenomyelon tuedianum. Trans. Roy. Soc. Edin., 1912. The portrait is from a photograph taken in 1911 by Frederick Hollyer, 9 Pembroke Square, Kensington, W. Alfred Stanley Marsh, ALFRED STANLEY MARSH was born at Crewkerne on February 1, l \ 1892. He died on January 5, 1916, being shot through the heart by a German sniper, who had found a weak spot in the parapet of the trench. Death was instantaneous. One who ran to his aid was killed in the same way. Marsh was educated from 1903 to 1909 at Sexey’s School, Bruton, and here his taste for Natural History received encouragement. He obtained a First Class in both the Junior and Senior Oxford Local Examinations, and in both he gained distinction in Botany (among other subjects). In 1908* when only sixteen years of age, he gained the County Scholarship and an Exhibition and Subsizarship at Trinity College, Cambridge ; but he did not enter Cambridge University until 1909, when he also won the Soley Scholarship. He surmounted the obstacle of the Latin and Greek tests in the Little-go, being placed in the third class, and wiring to his Head master, characteristically- — ‘ Sorry I have passed one class higher than I intended.’ As an undergraduate at Trinity his record was a distinguished one. In 1911 he obtained a First Class in the Natural Sciences Tripos, Part I. He then specialized in Botany, and in 1913 he was placed in the First Class in Part II of the same Tripos. He was thereupon offered the Frank Smart Studentship in the gift of Gonville and Caius College, accepted it, and migrated from Trinity to Caius. This Studentship afforded him the oppor- tunity of doing research work in Botany, of which he availed himself with energy and enthusiasm. He acted as a demonstrator in Professor Seward’s classes in Elementary Botany, and on Dr. Moss’s field excursions, his keenness and good humour endearing him to the students. Marsh’s published work consists of four papers which appeared in I9M-L5 : (1) ‘ Notes on the Anatomy of Stangeria paradoxal New Phyt., xiii, pp. 18-30. 1914. Marsh demonstrated an interesting mode of vascular supply of the leaves in this Cycad, and followed the behaviour of the centri- petal and centrifugal xylem strands of the leaf throughout its whole length. Plis conclusion was that ‘a close relationship can be argued between the modern Cycadales and the fossil Cycadofilices ’. (2) ‘The History of the Occurrence of Azolla in the British Isles and in Europe generally.’ Proc. Cambridge Phil. Soc., xvii, pp. 383-6. 1914. This paper gives an account of the occurrence of A. caroliniana [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] xxvi Obituary. — Alfred Stanley Marsh. and A. filiculoides in Britain, and is a valuable basis for the study of these two interesting and rapidly spreading aliens. (3) ‘ The Anatomy of some Xerophilous Species of Cheilanthes and Pellaeal Ann. Bot., xxviii, pp. 671-84. 1914. In this paper Marsh gave a comparative account of the stelar anatomy of several closely related Ferns ; and in the case of the peculiar vascular structure of the petioles he showed that they could all be derived from Sinnott’s primitive type with three endarch protoxylems. (4) ‘ The Maritime Ecology of Holme-next-the-Sea, Norfolk.’ Journ. of Ecology, iii, pp. 65-92. 1915. This is the report of a valuable piece of work undertaken by a group of young Cambridge men, altogether ten in number. A short account of the work was given in 1913 before the British Association by P. ‘H. Allen (‘ by whose untimely death ’, Marsh writes, ‘ we lost a dear friend and a splendid fellow worker ’). Marsh was the chosen spokesman of this group of (in more senses than one) synecologists, and his paper is an admirable example of vegetational survey work and a demonstration of the possibilities of co-operation in ecological research. Marsh took a commission in November, 1914, in the 8th (Service) Battalion, Somerset Light Infantry. He was made First-Lieutenant in April, 1915, and got his Captaincy in August, being then only twenty- three years of age. Throwing himself whole-heartedly into his new duties he quickly won popularity with his men and the affection of his fellow officers. Letters from the Front, from officers and men alike, testify to the deep sorrow which his death inspired even among those to whom bereave- ments are of the texture of daily life. While devoting himself entirely to his military duties he longed for the time when he should be able again to take up the botanical work that he loved. His plans for the future included a continuance of the synecological work at Holme, a series of experiments on the relations of closely allied species to their different habitats, and a share (with his friend R. C. McLean) in an investigation of flower structure in the Ranales. His botanical interests were broad, and to every discussion he brought fresh and suggestive ideas and a power of seeing the distant implications of hypotheses. His earliest botanical enthusiasm was floristic, and he aston- ished one by his uncanny ability to state accurately the ‘census-number’ of every British flowering plant. His Cambridge training opened his eyes to the manifold avenues of research which Botany presents ; and though he had not yet settled down to any one definite line of work, it would probably have been some branches of Phylogenetic Anatomy and Ecology which would have claimed him. By his death botanical science is the poorer for the loss of a keen vision, an agile imagination, and an enthu- siastic capacity for work. XXV11 Obituary .—Alfred Stanley Marsh. His personality was of the frankest, merriest and most good-humoured. He radiated gaiety and good fellowship. Always ready for a jest, his quaint wit enlivened many a conversation. His sense of the odd possi- bilities of words led him into the paths of humorous verse and prose. He was editor of, and responsible for a great part of the matter in the latest number of that celebrated periodical ‘The Tea Phytologist \ His cheerful goodwill and appreciative outlook upon life made his presence the best possible antidote to depression. All who knew him sincerely mourn the loss of his bright and stimulating personality. To his many devoted friends his death brought bitter grief for the passing of one whose comradeship is among the happiest memories of their lives. R. H. C. Cambridge, March , 19 1 6. Vol. XXX. No. CXVII. January, 1916. Price 14s of Bota EDITED BY DUKINFIELD HENRY SCOTT, M.A., LL.D., D.Sc., For. Sec. R.S. LATELY HONORARY KEEPER OF THE JODRELL LABORATORY, ROYAL BOTANIC GARDENS, KEW / JOHN BRETLAND FARMER, D.Sc., M.A., F.R.S. PROFESSOR OF BOTANY, ROYAL COLLEGE OF SCIENCE, LONDON FRANCIS WALL OLIVER, M.A., D.Sc., F.R.S. QUAIN PROFESSOR OF BOTANY, UNIVERSITY COLLEGE, LONDON AND ROLAND THAXTER, M.A., Ph.D. PROFESSOR OF CRYPTOGAMIC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S.A. ASSISTED BY OTHER BOTANISTS * LONDON HUMPHREY MILFORD, OXFORD UNIVERSITY PRESS AMEN CORNER, E.C. Edinburgh, Glasgow, New York, Toronto, Melbourne, and Bombay 1916 Printed in England at the Oxford University Press CONTENTS ' PAGE David Thomas Gwynne-Vaughan. With Portrait i-xxiv Willis, J. C. — The Evolution of Species in Ceylon, with reference to the Dying Out of Species. With two Figures in the Text i Leitch, I. — Some Experiments on the Influence of Temperature on the Rate of Growth in Pisum sativum. With Plate I and ten Figures in the Text 25 Laidlaw, C. G. P., and Knight, R. C. — A Description of a Recording Porometer and a Note on Stomatal Behaviour during Wilting. With three Figures in the Text . 47 Knight, R. C. — On the Use of the Porometer in Stomatal Investigation. With seven Figures in the Text ............. 57 Brenchley, Winifred E.— The Effect of the Concentration of the Nutrient Solution on the Growth of Barley and Wheat in Water Cultures. With Plate II and four Diagrams in the Text 77 Barratt, Kate. — The Origin of the Endodermis in the Stem of Hippuris. With six Figures in the Text 91 Davie, R. C. — The Development of the Sorus and Sporangium and the Prothallus of Peranema cyatheoides, D. Don. With Plate III and two Figures in the Text . . 101 Stopes, Marie C. — An Early Type of the Abietineae (?) from the Cretaceous of New Zealand. With Plate IV and seven Figures in the Text . . . . . .111 Holden, H. S. — Further Observations on the Wound Reactions of the Petioles of Pteris aquilina. With four Figures in the Text . . . . . . . . .127 FritsCH, F. E. — The Morphology and Ecology of an Extreme Terrestrial Form of Zygnema (Zygogonium) ericetorum (Kuetz.), Hansg. With three Figures in the Text 135 Takeda, H. — Dysmorphococcus variabilis, gen. et sp. nov. With fifteen Figures in the Text 1 5 1 Takeda, H.— Scourfieldia cordiformis, a New Chlamydomonad. With five Figures in the Text .... ........ ... 157 Stapledon, R. G. — On the Plant Communities of Farm Land . ..... 161 Fraser, Mary T. — Parallel Tests of Seeds by Germination and by Electrical Response. (Preliminary Experiments.) 1 8 1 NOTES. Small, James. — Anomalies in the Ovary of Senecio vulgaris, L. With three Figures in the Text . 19 1 Doyle, Joseph. — Note on the Structure of the Ovule of Larix leptolepis. With one Figure in tfie Text . . . . . . . . . . . . 193 NOTICE TO SUBSCRIBERS. The subscription-price of each volume is thirty shillings, payable in advance : the Parts, four in number, are supplied as they appear, post free to subscribers in the United Kingdom, and with a charge of is. 6d . per annum for postage to subscribers residing abroad. The price of individual Parts is fixed at a higher rate. Intending subscribers should send their names, with subscription, to Humphrey Milford, Oxford University Press, Amen Corner, London, E.C. As the earlier volumes of the Annals of Botany are becoming scarce, Vol. I will only be sold as part of a complete set ; and Parts will not as a rule be sold separately, after the publication of the volume to which they belong. A few extra copies of particular Parts at present remain on hand, for which special application must be made to the Editors, Clarendon Press, Oxford. NOTICE TO CONTRIBUTORS. In future, English contributors should address their papers to Professor F. W. Oliver, University College, Gower Street, London, W.C. ; and contributors in America to Professor R. Thaxter, Harvard University, Cambridge, Mass. Papers sent in with a view to publication must be type-written and must conclude with a summary of the contents. In view of the increasing numbers of manuscripts submitted and the inevitable limitations of space, the Editors desire to impress upon contributors the importance of conciseness in the treatment of their subject-matter. Whilst they are not prepared at present to draw a hard and fast line, the Editors suggest that the longer papers should not exceed 12,000 words. Illustrations. These where possible should be suitable for reproduction as line- blocks in the text. Where lithographic or collotype plates are required the figures should be planned so as to fill properly a qto or an 8vo plate. The maximum space available for figures in a 4to plate is 8Jx iiJ inches, in an 8vo plate 8UX5J inches. Attention to this will conduce to the rapid publication of papers if accepted. Each contributor to the Annals of Botany is entitled to receive gratis one hundred separate copies of his paper, and may purchase additional copies if he informs the Editors of his wishes in this respect when he returns corrected, proofs. The price of these additional "ill diiinand nnnn th° ^ nnmh^r.nf nlafps in the naner. M. The Evolution of Species in Ceylon, with reference to the Dying Out of Species. BY J. C. WILLIS, M.A., Sc.D. With two Figures in the Text. IN a paper recently published 1 I have brought forward conclusions which have such far-reaching bearings upon many branches of botany (and probably of zoology also) that it will be well to re-enunciate them in connexion with the further deduction here made from the figures, and which was briefly indicated in that paper, that there is little evidence to show that any species of Angiosperms are dying out. In a recent paper2 upon the Podostemaceae and Tristichaceae I have endeavoured to show that these families, living as they do (and must always have done) under perfectly uniform conditions, cannot owe their evolution to Natural Selection. At the same time they show wide and extraordinary distinctions between species and genera, of the ordinary ‘ Linnean ’ type. In a further paper 3 on the origin of these families I have endeavoured to show that they must have arisen from land plants growing at the sides of the streams, and in any case that the first change necessary to give rise to their ancestral forms must have been a ‘ large ’ change, which could not therefore have been due to Natural Selection. In a series of papers published in Ceylon4 from 1906 to 1911 I have devoted attention to the very interesting endemic species of that island, and have endeavoured to show that they cannot be regarded as local species owing their origin to the operation of Natural Selection in response to local needs or conditions. These endemic species are not a casual 1 The Endemic Flora of Ceylon, with reference to Geographical Distribution and Evolution in general. Phil. Trans., B, vol. ccvi, 1915, p. 307. 2 On the Lack of Adaptation in the Tristichaceae and Podostemaceae. Proc. Roy. Soc., B, vol. lxxxvii, 1914, p. 532. 3 The Origin of the Tristichaceae and Podostemaceae. Ann. of Bot., vol. xxix, 1915, p. 299. 4 Six papers in Ann. Perad., vols. iii-v, 1906-11. [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] 2 Willis.— The Evolution of Species in Ceylon , assortment of species occupying spots characterized by special local conditions, but are distributed in the island according to certain fairly definite rules, which obtain equally well in other parts of India and the East, and which I found with much interest to hold also in the case of the numerous endemic species of the state of Rio de Janeiro. These rules, as I have already indicated, and shall endeavour further to prove below, appear to me entirely out of harmony with the current idea that the great differences in the geographical distribution of species are largely due to the operations of Natural Selection. When one works with any number exceeding, say, 15 or 20 allied species of similar distributional origin, one finds that any one group behaves like any other group. In one of the papers referred to 1 I published the statement that the Ceylon endemic species were rarer than those of wider distribution that occurred amongst them, and made up my mind to enumerate the whole flora in this respect at the first opportunity. This arose with the publication of my £ Revised Catalogue of the Ceylon Flora and at the same time it struck me that the evidence I was collecting would be rendered much more conclusive were the non-endemic species divided into two groups, those found also in Peninsular India, and those with yet wider dispersal than this. Marking the species in the Catalogue thus, and entering for each the degree of rarity given by Trimen, who in his great Flora of Ceylon divides all species into six classes— Very Common, Common, Rather Common, Rather Rare, Rare, and Very Rare— -I had not done many pages before I realized that I had come upon a general law, which shows as clearly in the figures as does Mendel’s Law in any table of results of crossing. Pages II to 15 of the Catalogue, for instance, show the following: Table I. Ceylon spp. Total. Ceylon- Peninsular- Indian. Total. Wider. Total. 1. VC — — — — 1 1 — 1 — — 1 2 4 3 2 3 4 16 2. C 1 — 1 1 2 5 — — — — 1 1 5 8 7 3 3 26 3. RC — 1 — 2 5 8 - 1 - - 3 4 2 3 6 3 H 4. RR — — — — 0 O 3 — — — — — — 3 2 1 — — 6 5. R 2 — 1 2 3 8 — 1 1 1 1 4 2 1 2 — 5 6. VR 8 212 4 17 - - 3 ~ ~ 3 3 2 1 3 ***“ 9 If VC (Very Common) be marked 1 and the others up to VR 6, we may easily calculate the average rarity by multiplying the total under each head of VC, &c., by the mark for that head, and dividing the grand total by the total number of species. This shows that the mean rarity (for these 1 Some Evidence against . . , Natural Selection, Ann. Perad., vol. iv, 1907, p. 12, 3 with reference to the Dying Out of Species . five pages) of the Ceylon endemics is 4*5, of the Ceylon-Peninsular-Indian species 3-8, and of the widely distributed species 2-8, in figures running • from 1 to 6. The first dozen pages were amply sufficient to show that a general law was making its appearance, and I went on eagerly to the end, when the grand total showed the figures already published, and which may be quoted again, with the addition of the percentages calculated crosswise. Table II. Ceylon. Ceylon- Peninsular- India. Wider. VC 19 spp. 6-66% 45 15-79% 221 77-54 % C 90 13*43 118 17-60 462 68-95 RC 139 25-04 103 i8‘55 313 56-39 RR 136 31-70 84 19-58 209 48-71 R 192 46-26 64 I5’42 159 38*3r VR 233 51-20 78 17-20 I44 31*65 809 492 1508 Rarity 4*3 3'5 3*o In other words, the widely distributed species are by far the commonest, and much commoner than the mean of the whole flora, which is obviously 3*5, the mean between 1 and 6. The species found also in Peninsular India, i. e. roughly as far as a line drawn from Bombay to Calcutta, are next most common, with the mean rarity of the whole flora, and the Ceylon endemic forms are very much rarer than this (4-3). This is a very important result, and it will be well therefore to point out how the figures have been arrived at in Trimen’s Flora, and that they are quite unassailable. A great number of species in Ceylon are found only in one small locality, not exceeding a few miles in diameter. These Trimen classifies as Very Rare. A number of localities of such endemic species are shown in the first of the maps given with this paper. So far as my personal experience of such species goes, they are not only confined to a very small locality, but are most often rare within that actual locality. The next stage is Rare, and examination of Trimen’s localities (which are all supported by specimens in the herbaria of Peradeniya and Kew) shows that such species occur in roundish areas of from 10-15 to 3° m^es or so in diameter, averaging perhaps 24. It is almost needless to remark that cases occur in which it is a matter of individual choice whether a species shall be looked upon as VR or R. Fortunately for this work, therefore, all the decisions (except for; Gramineae) were made by Trimen himself. The next higher stage is Rather Rare, which implies a roundish area of about 50 miles in diameter. Not only is the area larger, but the plants in general seem to be commoner in it, and the same may be said with regard to Rather Common, where the area occupied B 2 4 Willis.— The Evolution of Species in Ceylon , is yet larger again. In the case of Common, a species is usually found to occupy the whole area suited to it (the island shows Wet and Dry, Warm and Cool zones), and is common therein, whilst Very Common is the same as to area, but the species is yet more abundant. These figures are the result of over seventy years’ work by many excellent botanists, and it is not to be expected that more than perhaps one per cent, of them will be found to be erroneous. In any case, as they are based on actual herbarium specimens, it is impossible to lower a species in the classification, and consequently the ‘ wides ’ must remain much more common than the average of the flora. In the second place, even if some could be shown to be erroneous, the figures are so numerous that a few alterations would make no difference whatever. Ninety ‘ wides’ might each be lowered a class and yet leave the rarity 3-0. To equalize the endemics and the ‘wides’ would need 687 alterations in the former, each raising a species one class in the list, and 699 in the latter, each lowering a species one class, which we have just shown to be impossible. The mere fact that the figures come out in such remarkable arithmetical progression along the scales shows that they must on the whole be accurate, for the chances against such an arrangement turning up accidentally are inconceivably great. Now not only do the grand totals show these figures of rarity for the three groups into which we have divided the Ceylon flora, but (as might be expected from the way in which the law shows, almost page by page of the Catalogue) the figures for each family show the same thing, down to families with 14 endemic species, and the figures for the groups of families containing 12, 11, 10, 8 or 7, 6 or 5, 4 or 3, and 2 or 1 species respectively. These figures are given in detail in Table VI of the Phil. Trans, paper, and it will suffice to quote here the actual mean rarity of the endemic species in all these families or groups, which gives the remarkable figures 4-4, 4-3, 4-9, 4-4, 4-3, 4-5, 47, 4-4, 4-0, 4-2, 4-1, 3-9, 4-5, 4-5, 4-1, 4-5, 4-4, 4-2, 3-9, 4-3, 4-0, 4-0, 4-1. If we arrange these in numerical order, we get 3-9, 3-9, 4-0, 4-0, 4-0, 4-1, 4-1, 4-1, 4-2, 4-2, 4-3, 4-3, 4-3, 4-4, 4-4, 4-4, 4-4, 4-5, 4\5. 4’5. 4'5> 47. 4'9 ■ Such numerical results as these call for immediate explanation, if such be possible. One cannot pass them by as of no importance, as has been the custom with the usual statistics of geographical distribution, which give so many per cent, of Leguminosae, and so many of Orchidaceae, &c., as occurring in the locality under consideration. But to explain them in harmony with the theory of Natural Selection appears to me quite impossible. The further explanation which I put upon them is open to dispute, but the facts themselves are incontrovertible, and, so far as I can see, are very seriously out of accord with Natural Selection. And if the Ceylon flora cannot be explained upon that theory, it at least raises with reference to the Dying Out of Species. 5 a grave doubt as to the applicability of the theory in general, for there is not the least reason to suppose the Ceylon flora or the Ceylon conditions to be unique or isolated in any respect. The interpretation which I put upon the facts is simple. As all the plants, and all the families, behave alike, it is evident that their grouping and distribution must be the result of a cause which acts upon all with practically even pressure. Now Natural Selection could not do this, for in its essentials it is of a differentiating nature. The only cause that I can see which thus acts evenly upon all is age , and I am inclined, therefore, to think that the area occupied by any given species at any given time, in any given country, is to a large extent an indication of the age of that species in the country (not its absolute age, which has nothing to do with the question, so far as I can at present see). The widely distributed species, which must on the whole be the oldest, are the commonest, the Ceylon-Indian next oldest and next commonest, and the endemics, the youngest, are the rarest. In the case of a group of, say, twenty species of similar distributional origin, and the same family, as we have pointed out above, this is certainly extremely probable, but of course in the case of any single species numerous disturbing influences come into play. And it will be well to point out specifically that these remarks only apply as yet to the angiospermous species, and are based only on the flora of Ceylon, though the close similarity that I have observed at Rio de Janeiro leads me to believe that they apply generally to most floras. Still more true, in general, is this statement as regards those genera which contain a number of species, than as regards the actual species themselves. In two previous papers 1 I have gone into this question from the point of view of general geographical distribution, and may refer to them here. In other words, on the view of things thus propounded (which appears to me to have great probabilities in its favour, besides possessing the advantage of explaining numerous as yet unexplained facts in a simple way), endemic species confined to small areas are really species in the earlier stages of spreading , and, given time enough, they might ultimately be found covering large areas. Endemic species (all species are on this view endemic when young) begin as VR in some given country, and gradually extend their area, passing upwards through the stages R, RR, RC, &c. Already these views are meeting with numerous objections, and it will be well to deal with some of these in this paper. The first objection, that the endemic species are on the whole the oldest, and not, as I maintain, the youngest, is easily disposed of by a little consideration. Great numbers of well separated endemic species occur in such genera as Ranunctdus , 1 The Geographical Distribution of the Dilleniaceae, as illustrating the treatment of this subject on the Theory of Mutation. Ann. Perad., vol. iv, 1907, p. 69; Phil. Trans., 1. c., p. 335. 6 Willis . — The Evolution of Species in Ceylon , Polygala , Garcinia , Stellaria, Impatiens , Rhamnus , FzVzk, Crotalaria , Desmodimn , Poterium , Begonia , Dipsacus , Vernonia , Senecio, Symplocos , Swertia , Ipomoea , Justicia, Coleus , Scutellaria , Piper , Loranthus , Ficus , Dendrobium , Habenaria , Smilax , Areca , Eriocaulon, Carexi Panicum , &c. It is absurd to suggest that all these genera commenced in Ceylon, and yet this must have been the case if the endemic species are the oldest, unless one imagine that these genera commenced polyphyletically at the numerous places where they exhibit endemic species. The chief objection to my views comes from the supporters of Natural Selection, and is simply a restatement of their position (which as yet lacks proof) that endemic species are local species developed in response to local needs or conditions. I have already dealt with this question in a series of previous papers, but it will be well to add further arguments here, and especially an arithmetical argument which appears to me of a very conclusive nature. We have seen above that all the families with fourteen or more endemic species, and all the groups of families with fewer, agree very closely in the degree of rarity of those species, which only varies between 3-9 and 4*9. The mean rarity of the endemic species, from which no family departs very far, is 4.3. Now examination of the areas occupied by the various classes in Trimen’s Flora shows that Rather Rare species occupy on an average an area of about fifty miles in diameter, so that a rarity of 4-3 would indicate that the average area occupied by an endemic was about forty miles in diameter. Now in such an area, especially in the hilly south-west of the island, where the bulk of the endemics occur, it is impossible to talk of local conditions, for it includes every kind of soil, great range in local composition of flora, great differences of climate, and many other variations. This simple consideration alone makes a very strong case against Natural Selection. As I have elsewhere pointed out, why should a pinnate leaf suit one valley, and a simple one the next valley to it, in the same genus ? As Table II shows, the endemics increase in number down the scale from 19 Very Common to 233 Very Rare, while the species of wide distribution go in the opposite direction, and those of Ceylon and Peninsular India are fairly evenly distributed. Not only so, but in general, as we have shown, all the families with fourteen species or more, and all the groups of families with fewer, show the same thing. We may analyse the figures of Table VI of the Phil. Trans, paper, and exhibit the distribution of the endemics within each family in point of rarity, when we obtain the very striking result here shown : with reference to the Dying Out of Species. 7 Table III. Family. Endemics. VC C RC RR R VR Rarity . Orchidaceae 78 — O T9 18 20 18 4-4 Rubiaceae 71 6 6 1 1 8 17 23 4‘3 Dipterocarpaceae 47 — 3 2 8 l5 19 4’9 Euphorbiaceae 45 — 3 8 12 9 13 4’4 Acanthaceae 39 — 2 1 1 6 11 9 4‘3 Melastomaceae 36 — 3 6 5 11 11 4’5 Gramineae 3i 1 1 3 6 10 10 47 Myrtaceae 30 1 4 5 4 3 13 4’4 Lauraceae 23 — 5 5 1 7 5 4'° Anonaceae 21 — 1 7 3 6 4 4’2 Compositae 19 1 4 2 3 3 6 4’1 Geraniaceae 18 1 2 5 3 3 4 3'9 Scitamineae i7 — 1 5 — 6 5 4’5 Styraceae 17 — — 4 4 4 5 4*5 Anacardiaceae 15 — 5 1 1 3 5 4’1 Araceae 14 1 1 2 1 4 5 4*5 Total 521 11 44 96 83 132 155 4*4 All the families have some endemic species under every head from C to VR, with only two exceptions in eighty cases (RR in Scitamineae, and C in Styraceae). And the numbers on the whole increase in each family from top to bottom of the scale. Thus, with five-eighths of the whole number (809) of endemics, we find a wonderful likeness among the different families in the proportions of endemics in the classes C, RC, &c. The families with many endemics show greater average rarity than those with few, as has already been pointed out (cf. Table XVI of previous paper). It is not possible for the Natural Selectionist even to derive consolation from the remaining families with small numbers which are lumped together in Table VI just quoted. If we place them according to rarity, we get a simple table of variation of the usual trial and error pattern : Table IV. Rarity. Families. Species. No. of spp. per family .l Marks. 1*0 1 1 1*0 1 1-0 — 2*0 7 9 1*2 17 2*1 —3*0 6 J9 3*i 53 3*i —4-0 20 74 3*7 270 4‘i -5*o 32 161 5*° 736 5*i -6*o 9 24 2*6 132 75 288 1209 Or if we analyse them according to the classification of their contained endemics, we get, Table V. Families. Endemics. VC C RC RR R VR Rarity. 3 with 12 spp. 36 — 6 4 6 5 15 4-4 4 11 44 I 5 7 9 1 1 1 1 4*2 2 10 20 2 5 — 5 2 6 3*9' 7 8 or 7 5 ^ I 7 9 7 12 16 4*3 6 6 or 5 32 I 4 6 8 9 4 4*0 L5 4 or 3 47 2 9 8 7 8 13 4-o 38 2 or 1 57 I 10 9 11 13 13 4*i Total 288 8 46 43 53 60 1 00 1 4*^ 1 This shows very clearly that the wider deviations are mainly in the smaller families. 8 Willis.— The Evolution of Species in Ceylon , — a table exactly similar to Table III above, but with still lower average rarity (cf. Table XVI, Phil. Trans, paper). The same fact shows if one analyse the tables of rarity for the species Doona1 Stemonoporus 1 Impatiens Semecarpus Eugenia Memecylon Hedyotis Symplocos Strobilanthes Amomum VC i 1 2 Table VI. C RC RR I I 2 — — 2 3 4 4 4 5 24 1 4 4 1 1 13 3 1 5 4 4 7 _3 28 R 4 4 3 3 2 5 5 4 4 _5 39 VR 3 9 4 5 13 7 8 5 8 _3 65 ves, Total. Rarity. 11 4-6 15 5-4 15 4.2 13 4*3 29 4*4 21 4.6 16 4.8 17 4-5 25 4*4 11 4-7 173 4*5 The rarity varies only between 4*2 and 5-4, i.e. is never so low as the mean of the whole flora (3*5), nor so high as strictly local (VR = 6). And taking the classes C to VR, there are only eight cases out of fifty where there are no representatives of a genus (3 C, 2 RC, 3 RR). We may even take the genera with from five to nine endemic species, and get, Table VII. RC VC Acrotrema Goniotkalamus Calophyllum Dipterocarpus Shorea Vitis Sonerila Psychotria Lasianthus Vernonia Palaquium Diospyros Gymnostachyum Actinodaphne Litsea Loranthus Phyllanthus Glochidion Dicotyledons Oberonia Liparis Bulbophyllum Cirrhopetalum Eria Saccolabium Habenaria Calamus Eriocaulon Carex Garnotia Monocotyledons Grand Total C 1 2 1 17 1 1 1 1 2 1 1 1 3 14 2 2 1 1 2 1 1 10 RR 1 1 1 2 2 1 2 3 22 . 3 1 1 2 1 2 2 R 2 2 2 3 2 3 5 1 1 2 2 2 4 2 2 4 1 40 3 1 13 1 22 24 35 1 Endemic genera. 4 3 1 4 17 57 VR 3 3 2 1 2 4 3 2 1 2 3 1 2 4 3 1 37 1 3 4 1 1 1 1 5 2 1 20 57 Rarity. 4-8 5*3 4-2 4-6 4-6 4*2 5‘i 5’1 3*7 3*3 4*7 4*7 4- 6 3*8 3*8 5- o 5-i 3* 4*4 4*7 5*4 5*6 4-o 3*8 3*4 4-6 4*i 5*4 4-0 4-8 4*5 4*5 with reference to the Dying Out of Species . 9 Only two genera in this list ( Vernonia 3-3 and Sctccolabium 3-4) are equal to or above the average in point of rarity. One may even go lower than this, and take the genera with fewer than five species. Naturally any one genus may show any figures from VC to VR, but if one add them together in groups one gets the same result as before. We may take all the genera beginning with one letter of the alphabet, and get, Table VIII. VC C RC RR R VR 3 Rarity . Beginning with A 3 9 11 11 12 11 3*9 B — 4 5 2 10 4 4-2 C — 5 10 11 13 23 4«6 D — 1 4 5 3 9 4-6 3 19 30 29 38 47 There is no need to follow the comparison through the whole alphabet. It is sufficiently obvious from all these tables that the distribution of the species of the various endemic families and genera obeys a simple law which determines that the numbers under the different heads from VC to VR shall be distributed in proportions increasing from the former to the latter. The numbers simply vary about the suppositional numbers in the ordinary way of trial and error. The maximum is always or nearly always at VR, If one analyse the same tables as to the figures for the Ceylon-Indian and widely distributed species, one finds exactly similar results, but with the maxima differently placed. For instance, the first five orders (i.e. in number of endemics) show, Table IX. Orchidaceae Rubiaceae Dipterocarpaceae Euphorbiaceae Acanthaceae Total Ceylon-P. - India . 1 6 9 10 5 8 653 641 484 334 1 7 7 4 7 4 12 26 24 23 19 17 (121) Wide. 1 6 12 8 14 4 7 13 9 5 6 2 4 26 8 7 8 6 7 9 5 2 ~ 1 19 54 34 22 28 13 (170) For convenience it will be well to arrange these together with the endemics, thus : Table X. VC C RC RR R VR Total. Endemics 6 14 49 44 57 63 233 Ceylon-P.-I. 12 26 24 23 19 i7 121 Wide 19 54 34 22 28 13 170 Whilst the maximum of the endemics lies at VR, that of the Ceylon- Peninsular-Indian species lies at C, with no very great tailing off towards VR, and that of the Wides lies very markedly at C, with an immediate and sudden drop to RC and RR. Willis.— The Evolution of Species in Ceylon , 10 The mere fact that the numbers representing rarities come out with such simple arithmetical relations seems enough to show that whatever cause is operative in causing such relations it cannot be Natural Selection. So simple an explanation of evolution is impossible, though a more ‘ mechanical ’ explanation of its more obvious features than is Natural Selection is called for. [All the tables of rarity given in my previous paper show the same thing, varying in almost purely ‘mechanical’ ways. Table VI shows that the rarity of the endemics is much the same for all, taken family by family. Table IX shows much the same thing for genera, and if analysed gives the following figures : Table XI.1 Rarity i -2 2-1-3 3*i-4 4*i-5 5*1-6 spp. Spp. spp. spp. Spp. Genera, per Genera, per Genera, per Genera, per Genera, per genus. genus. genus. gemis. genus. Endemic 5 I *2 6 2*0 GO OO •°^7 recorded to date. This has always been a difficult matter to explain, and the Natural Selectionists have had two rival hypotheses, which it may be pointed out are mutually contradictory. The first is that Ceylon has a less ‘ tropical ’ climate than Malaya, having greater extremes of wet and dryness and of heat and cold. The second is that Ceylon has but a poor soil, with no variety in it, it being all the product of decay of gneiss and granite. On the first of these hypotheses the less variety in species is put down to greater variety in conditions, on the second to less. The first objection which occurs to one is that South India, with the same geology and a more variable climate, appears to have more species than Ceylon, and this is most remarkably supported by the case of the state of Rio de Janeiro in southern Brazil, which has just about the same area as Ceylon, has the same geology and soil, and, except in the narrow coastal belt, has a greater variety of climates, at any rate as regards heat and cold. The flora of Rio is something enormous, and as much richer than that of Java as the latter is richer than that of Ceylon. Dr. Lofgren, my late colleague, than whom no one better knows the flora of Brazil, estimates the flora of the state of Rio at 7,000-8,000 species. He calculates the flora of the single mountain of Itatiaya (10,000 feet), most of which is within this state, at 7,500 species. I explain this variety in the size of the floras mainly by the fact that the state of Rio has always, so far as geological evidence goes, been attached to large continental areas and is itself of enormous age. It has thus been open to the invasion of great numbers of foreign species, whilst its own mountainous configuration has tended to the development of large numbers of endemics within the state, just as in Ceylon every hill (and here also every little island off the coast) seems to have its own forms. Only a few days before I left Rio, Dr. Lofgren found on one of the nearer islands a most remarkable new species of Rhipsalis , which one could only describe as a pendulous shrub or tree, its stem being about four inches in diameter. From these and other similar facts I draw the conclusion that the number of species in a country depends upon its age from the time of its last submergence, and upon whether it has been attached to large areas with many species during most of its history, or whether it has been cut off at an earlier or later date. Species, or at least the majority of them, do 22 Willis . — The Evolution of Species in Ceylon , not appear to die out except by accident. A very small accident may kill out a species while at or below the stage represented in the Ceylon classification by VR, whilst it will need a geological submergence or some such accident to kill out one represented by VC. Whether under exactly equal conditions of age, attachment to other areas, and favourable climates, &c., a tropical area would have more species than one in the temperate zone of equal size must remain an unsettled question. Most of the evidence on which we have relied for an affirmative answer must now be regarded as incapable of bearing such a load, and the rich floras of South Africa and West Australia contradict the assumption. As yet we know of nothing that can be adduced as a reason why almost limitless species should not survive on an area with reasonably good climatic conditions. There is no evidence whatever that any of the angio- spermous species of the Ceylon flora are dying out, and from analogy we may imagine this to be generally true. Summary. The paper is a continuation of previous papers in which, among other things, I have sought to show that Natural Selection has but little to do with the geographical distribution of species or the areas they occupy, and that the area occupied at any given time in any given country depends mainly upon the age of a species in that country (not its absolute age). The figures of rarity of the Ceylon flora, derived from the statistics extracted from Trimen’s Flora of Ceylon, are considered, and it is shown in the first place that they are incontrovertible, being much too numerous and too well worked out. The rarity of the endemic species (in figures going from I, Very Common, to 6, Very Rare) is 4-3 or very close to that figure (Table II), that of the species found also in Peninsular India is 3-5, and that of the species of wider distribution is 3-0. Not only do the grand totals show these figures, but they come out family by family. Natural Selection cannot produce a result like this, acting with equal pressure on every family, and I therefore attribute the distribution"of species (taking them in groups of twenty or more) to age. Many objections are being raised to these views. The first, that the endemic species are really the oldest and not the youngest, is easily disposed of by the consideration that they belong to the same genera. The second, a restatement of the contention of the supporters of Natural Selection, that they are local species developed to meet local needs or to suit local conditions, is met largely by an arithmetical argument. Not only is the rarity much the same in every family, but all families show some species under every head from Common to Very Rare, and in increasing numbers (Table III). This holds for all families with fourteen or more endemic species, and for all groups of families with fewer. The 23 with reference to the Dying Out of Species. same facts show in all the genera with more than ten species (Table VI), with from five to nine species (Table VII), and even in the genera with less than five, when taken together in groups (Table VIII). Always the maximum is at or near VR. In the same way the Ceylon-Indian and Wide species show parallel figures, family by family and genus by genus, but with maxima at Common (Table X). It is inconceivable that Natural Selection (a differentiating agent) should thus act with uniform pressure on every family and genus ; the only factor that to me seems satisfactory is age. The same argument is then put into biological dress, with slight alterations, and maps of Ceylon are given showing a number of Very Rare, Rare, and Rather Rare species. The VR areas are small dots scattered over the map, the R’s little rings, and the RR’s larger ones. But all are scattered, and the circles overlap like the rings in a shirt of chain mail. Now it is impossible to find conditions varying in such a way as to cause such distribution as this, and it is much simpler to look upon the area occupied as an indication of the age. A general discussion is then given, and it is shown that the advocates of Natural Selection do not satisfactorily explain such facts, but rather pass them over as incapable of explanation. Other objections to my views are also dealt with. The question of dying out of species is then considered, and it is shown that the figures of distribution of the Ceylon plants give no reason to suppose that any angiospermous species are dying out at the present time, a supposition which is borne out by a comparison of the floras of Ceylon, Java, and Rio de Janeiro. Some Experiments on the Influence of Temperature on the Rate of Growth in Pisum sativum. BY I. LEITCH, B.Sc. With Plate I and ten Figures in the Text. THE first work of importance on the subject of the relation of tempera- ture to growth-processes in higher plants is that of Sachs (’60), on the effect of temperature on germination. His method is to compare the amounts of growth in a given time at different temperatures. The time is measured from the end of the soaking, and in Pisum the time- interval chosen is forty-eight hours. F or Pisum he concludes that the temperature at which germination proceeds most quickly is below 22°. This method seems to be defective, since the amount of growth of the roots in the rather long period of forty-eight hours will be much affected by the ease or difficulty with which the root bursts the seed-coat, there being great differences in this respect between peas treated in exactly the same manner and giving the same growth-curve afterwards. Koppen (’70) finds that alterations of temperature exercise a retarding influence on growth, and makes a number of determinations of the rates of growth at temperatures, for Pisum , from io° to 40°. These values show a great inconstancy, and besides, in view of the experimentation-time of forty-eight hours and the method employed, their value is small. Petersen (’74) points out the obvious defects of Koppen’s work, demon- strates that variations in temperature, as such, have no effect upon growth, and that the curve of growth at non-injurious temperatures must be a curve convex to the temperature-axis. Sachs (’87) investigates the occurrence of the Grand Period in different seedlings, and finds it to occur in Pisum on the ninth or tenth day. It is to be noted that in his experiments on the fifth to sixth day, the tempera- ture varies between io° and 19-8°, a variation sufficient to disturb the result considerably. He gives figures for the relation of growth to temperature in Zea Mais, and quotes a judicious selection from Koppen’s figures showing a relation agreeing in type with his own determinations. [Annals of Botany, Vol. XXX. No. CXVII. January, 19x6.] 26 Leitch . — Some Experiments on the Influence of Askenasy (’90) finds for Zea Mats that the Grand Period curve is very flat ; the maximum rate of growth is reached when the roots are 30 to 40 mm. long, and the rate remains constant till the appearance of the side- roots when the main roots are about 130 mm. long. He finds also, that in roots grown at a high temperature and suddenly subjected to a temperature of 30 to 6° the growth is suddenly stopped (contraction usually occurs), and that, on the return of the roots to the initial temperature, the rate is depressed below the normal rate, for a time depending on the lowness of the intermediate temperature, and the length of their subjection to it. True (’95) confirms these latter experiments of Askenasy, using Vida Fab a, Lupinus albas, and Pisum sativum , and finds further that a sudden change from a low to a high temperature is, if the interval be great enough (from 30 to 180), followed by a sudden elongation of the root, with a subsequent period of depressed growth. His experiments with Vicia Faba on the effect of transitions between 180 and 30° are of very doubtful value, since 30° is, for Vicia , 4 above the optimum \ More recently Schmidt (T3) has demonstrated for Humulus Lupulus the very close dependence of growth, under ordinary open-air conditions, upon temperature. His growth and temperature curves vary, in all cases, in the same manner, and often (cf. No. 17) almost exactly proportionally. Finally, Vogt (cited in Jost, T3) gives a series of determinations, but here nothing is told of the method of experimentation, and the experimentation-time is again 24 hours, a time, as will appear later, much too long, at least for high temperatures. The experiments to be described are on Pisum sativum , the material belonging to one sample, carefully mixed at the beginning ; and they are in two parts. The first part consists of experiments with the long experimenta- tion-time of 22J hours. Their purpose was, first, to map out the field, and second, by the use of much larger numbers than is possible in more accurate microscopic work, to afford an idea of the amount of variation in the material. In the second part, the experimentation-time is short, and the determinations made by microscope measurements. First Series. The time occupied daily in preparing material and making measure- ments in the first series of experiments being about one and a half hours, the experimentation-time became conveniently 22^- hours. At first (the experiments were begun in December), the peas were soaked and germi- nated at room temperature, about 150 C. during the day. But soon it was found that the temperature fell too low during the night to give a suitable length of root on the third day, and from then onwards the soaking and germination took place in an electric thermostat whose temperature varied slowly between 150 and 170 C. The peas germinated in an apparatus first Temperature on the Rate of Growth in Pisum sativum. 27 constructed by Professor Johannsen and not as yet described. A photo- graph of it appears in Plate I, Fig. 1, and the apparatus is made in the following manner : A dish of suitable size is chosen and fitted in the foot with a plate of cork. The whole is lined with paraffin, and pieces of wire, or needles, are stuck in the cork in pairs, at such distances that two rubber tubes, which are placed round them, press closely together. The dish is filled with plaster of Paris, and when it has set the rubber tubes are pulled out, the form removed, and the apparatus is ready. A lid as shown in the photograph is quite simply made with a similar lined dish and a smaller plate of cork, covered with paraffin, to press into the soft plaster of Paris. Convenient sizes for the germination of peas are : whole apparatus, 10 x 15 cm., and rubber tubes 0-7 and 0-5 cm. in diameter, o-6 cm. apart. For use the apparatus is thoroughly moistened and is set in a dish con- taining a little water. In the experiments proper, as distinct from the first a / / Text-fig. 2. two days’ growth, instead of the porous plaster-of-Paris lid, a bell-jar lined with filter-paper was used to cover the peas, and a petri dish with NaOH was placed above the peas to prevent the accumulation of C02. The peas were soaked for 22 \ hours, and then placed seventy alto- gether, each day, on this apparatus and left for 4 6\ hours. At the be- ginning of the third day after soaking, therefore, the thirty-five most uniform in length were selected and measured. The method of measure- ment was as follows : the triangular piece of the seed-coat burst up in germination was removed, and a piece of millimetre-paper placed behind the root so that a centimetre line came exactly behind the point marked a in Text-fig. 1. The length of the root was read off in millimetres, so that Professor Johannsen’s method of measuring ‘ up to ’ was used ; that is, the next higher millimetre was always taken as the reading. In Text- fig. 2, which shows the paper in place, the length is 15 mm., but, if the 28 Leitch. — Some Experiments on the Influence of root-tip had been between the 14 and 15 mm. lines, it would also be called 15 mm. The experiments were carried out in a gas thermostat, where it was possible to vary the temperature from n° upwards. The temperature was recorded by thermograph tracings, and even at the highest temperatures used the fluctuations shown did not exceed 0-5°. Temperatures below n° were obtained according to the weather in a room of very constant tempera- ture, where a thermograph tracing was a straight line for days if the windows were undisturbed and the door kept closed. Temperatures near to and below zero were obtained in the open air. The lowest was in an experi- ment where the temperature did not rise above — 20 and where the minimum Text-fig. 3. Temperature in 0 C. was — 50. Here no growth took place ; instead, a slight general contraction. The roots were frozen, but, on thawing at room temperature, continued to grow. It was intended that the experiments should be in two series ; the first was completed, and in the second— which began about 150— for some cause which I have been quite unable to explain, the values of the growth-rate suddenly rose and maintained a level uniformly above those of the first series to 28°, after which the usual fall of the rate took place, that is to say, at a temperature which is about two degrees lower than in the first series. At this point, in view of these results, the number of peas soaked each day Temperature on the Rate of Growth in Pisum sativum. 29 was reduced to the minimum possible. About ninety were required, since about 20 per cent, are injured or deformed or of extremely large or small size. It was thus hoped to eliminate unconscious selection of the soaked peas. From then onwards, too, the germinating apparatus and all dishes used were periodically sterilized as a precautionary measure. Under these conditions a third series was carried out, and it agreed exactly with the first. Detailed figures are given in Table I, and the results are plotted in Text-fig. 3. Table [I shows the average initial length, average growth in 22| hours, together with the standard deviation and probable error of the mean for each experiment. The experiments are arranged in chronological order. Second Series. In the second part, the peas were grown in the following way. They were soaked and germinated as before in the electric thermostat, but remained in the plaster-of-Paris apparatus for only one day, at the end of which the roots were about 5 to 10 mm. long. Then they were placed in tubes measuring i*8 cm. in diameter and 10 cm. in height, with corks bored eccentrically to admit the thermometer during the experiment, and having a triangular piece cut out at one side so that the air in the tube was in con- tact with the outer air at all times. In the cork a capillary glass tube was fixed so that it lay along the side of the tube, and the pea was fixed to the cork by a pin so that the root was in contact with the side of the tube, between it andja slip of filter-paper and just beside the capillary tube. The filter-paper dipped in water. The root was thus supplied with sufficient moisture at all temperatures and air had free access to it. It was found that without the capillary tube the filter-paper adhered to the glass, and the resistance thus offered to the growth of the root was sufficient to cause curvatures. Plate I, Fig. 2, is a photograph of peas so grown. By this method a large proportion of the peas grew diagrammatically straight, and only a small proportion were incapable of accurate measurement. A slight curvature, such as that shown in the tube to the right in the photograph, is of no significance, since the experimentation-time is so short that the inclination does not appreciably change. In these cases, readings are taken with the eye-piece slightly rotated so that the micrometer scale is perpendi- cular to the axis of the root. With the magnification used, a reading 1 on the micrometer scale represented 0*056 mm. O'l \ A beaker served as a water-bath, and the temperature was regulated by a micro-burner and Roux’s regulator. Except at low temperatures, the temperature of the water remained constant to within 0-5°. For tempera- tures below 1 5° a slow current of water flowed through the beaker ; a conical 30 Leitch . — Some Experiments on the Influence of flask was interposed between the beaker and the tap, and for temperatures between that of the water and 150 the micro-burner was used to heat the flask. Between i° and 50 ice was used. By these means a complete range of temperatures from i° upwards was available. The experiments from 150 upwards were performed on a table built into the ground and therefore free from any danger of disturbance by vibration or shaking. At lower temperatures they were carried out in a colder room where such an arrangement was not available, but the room was a basement one, and repeated tests showed no disturbance by shaking. In view of the great amount of trouble to be saved by conducting the microscope experiments in daylight, a series of experiments was made to test for a possible influence of a sudden transition from dark to light, or as Vogt (T4) incidentally records for Lupinus , in his paper on the influence of light on growth in Arena, for a possible increase of the rate of growth in light. The temperature of about 250 was chosen for these experiments as giving a high rate of growth likely to show reactions clearly, and as being subject to no injurious effect. Table A gives the results of these experiments ; the readings are half-hour readings, and the numbers micrometer divisions. In Experiments 3, 7, and 9, a 100-C.P. electric lamp at about fourteen inches was used in addition to daylight. Table A. ^ Commencement of light, f Stoppage of light. 1. i-7, 1 1-7, 2-o, f 1-9, 2-1. 2. i-6, >|r i*6, i*6, f i-8. 3- i*5> i I*5? i-5» l'5> f l'lA l'7> I*5> 1 1*5- 4. i*7, 1 1*6. 5* I*5> ^ I*5* 6. i*6, \|r 1*7. 7. i*i» !*4> 8. i*5»l 1-65. 9. i-7, f i*5, 1-3. The means are, in dark, i*6 ; in light, 1-55. These readings show that, neglecting for the moment Experiments 3, 7, and 9, the fluctuations in the readings show no constant relation to the presence or absence of light. The fluctuation is, indeed, such as takes place in experiments under uniform con- ditions. (Compare Experiments 3 and 4 on change of temperature, quoted below.) Experiment 9 shows a fall in light, and Experiment 7 a rise in darkness after illumination — which results do not necessarily confirm each other — and Experiment 3 under the same conditions shows, as the experiments in daylight alone do, no influence of the light. 1 Temperature on the Rate of Growth in Pisum sativum. 31 To illustrate the absence of any ‘ stimulatory ’ effect, Experiment 2 may be cited in full. Time. Growth. 'J' ^ 2-49, 3-19, 3-25, 3*31, 3-37, 3*43* 3‘49> 4*I9> 4*49- i*6 0-3 0*3 o*3 0*3 0-4 i-6 i-8 The fifth reading in light is due to the inaccuracy in reading, readings being ‘ up to 5 one micrometer division (0-056 mm.), and the rise to i*8 in the last half-hour is a difference such as occurs as a normal fluctuation under uniform conditions. Two of the experiments to be described later, on the Grand Period of growth, were also performed in darkness for comparison, and they showed no difference from those in light. The shoot was etiolated, but the rate of growth of the root was unaffected. Therefore, at least within the con- ditions under which all the rest of the experiments took place, light exerts no influence on the rate of growth. In the next place there is to be considered the effect of a sudden rise or fall of temperature. In all experiments the same results were found, namely that the roots assumed immediately on reaching the new temperature the rate of growth characteristic for it, and that, for temperatures up to 290, such fluctuations as take place afterwards occur without showing either a typical time or mode of occurrence, so that, taking the mean of a number of peas, the rate of growth, for at least six hours, is represented by a straight line. To illustrate this the following experiments are quoted : 1. Fall to low temper attire from thermostat temperature. Date. Time. Growth. Temperature. March 7- 2.27 p.m. to 2.57 p.m. o-i 4’9° 2.57 p.m. to 3.27 p.m. o-i March 8. 1 r. 42 a.m. to 12.12 p.m. o-i 4*5° 12.12 p.m. to 12.42 p.m. o-i 4’5° 12.42 p.m. to 1. 12 p.m. o-i 4’5° March 7- 3.30 p.m. to 4 p.m. o-i 4’9° 4 p.m. to 4.30 p.m. 0-2 4*9° March 4.30 p.m. to 5 p.m. 0-2 4*9° 8. 5 p.m. to 10.27 a.m. o-i i (mean) 3’5° 10.40 a.m. to n.ioa.m. o-i 3*8° 1 1. 10 a.m. to 11.40 a.m. 0-2 3-9° The fall here is from thermostat temperature, 150 to 170, to about 40. In both, the characteristic low rate of growth is assumed at once and maintained. In No. 2, where the growth during the night was measured, giving a mean rate of o-n, a fall is seen as compared with the rate from 3.30 to 5, which is 0*17, due to the fall of temperature from 4-9° to 3-5° during the night. In No. 1 the rate on the 8th is the same as on the previous day, it being measured at nearly the same temperature. 32 Leitch. — Some Experiments on the Influence of 3. Fall from high temperature to room temperature. Rate of growth given first for three half-hours, then in six-minute readings : Temperature. 270, 270, 27°, 270, 25-5°, 21*2°, 20*3°, 20*2°, 20*0°, 19*8° = room temperature. Tate of growth. i-8 i-8 2*1 0*2 0-2 0*3 0-2 0*2 4. Rise of temperature. Rate of growth given during first half-hour in six-minute readings, then in half-hour readings : Temperattire. 20*0°, 25*0°, 25*8°, 26*0°, 26-0°, 26*0°, 26-1°, 26*0° Rate of growth. 0-3 0-4 0*3 0*3 0-3 i-6 i-8 5- Temperature. 23*8 °, 26*5°, 27-0°, 27*0°, 27-0°, 27*0° Rate of growth. 0-3 0*3 0*3 0-2 0*3 6. Temperature. 23*9°, 26*1°, 26*5°, 26-6°, 26*6°, 26*6° Rate of growth. 0*4 0-3 0*3 0*4 0-3 7. Finally the following experiment, where the results given are the means of measurements of eight peas : the rise of temperature is from room-temperature, about 180 to 25*2°. In the first, second, and third interval of ten minutes, the mean rate of growth was: 0-50, 0-475, °*475 1 that is, practically identical ; and the rates in the first and sixth half-hours were 1-45 and 1*45. From these experiments it will be clear that a change of temperature is, in itself, without effect ; but that the rate of growth follows immediately and accurately any considerable change of temperature. The next point to be considered is the possible effect on the experi- ments of the Grand Period of growth of the roots. In view of Sachs’s determinations, it appeared at first as if, in my experiments, taking place always at the beginning of the third day of growth, the slight rise to be expected during that day would be negligible, at least in short-period experiments. But in the long-period experiments at higher temperatures it had already become probable to me that Sachs’s results did not apply, and therefore, at the beginning of the microscope experiments, a series of determinations on the time of occurrence of the Grand Period was begun and continued intermittently. The roots were grown in long test-tubes with exactly the same other arrangements as in the microscope experiments, and measurements were made by applying a millimetre scale to the outside of the tube, a method which avoids all disturbances of growth by handling. Table II gives a full account of the results. Measurements were never continued after the appearance of the side-roots. Temperature on the Rate of Growth in Pisum sativum . 33 The conclusions from Table II are (1) that the roots germinated and grown at a constant temperature (Expts. 9, 10, 11) show the Grand Period on the third day at 140 C., and on the second day at 230 C. ; (2) that roots germinated and grown for one day at a temperature of 150 to 170 C. and then transferred to a different constant temperature show the Grand Period also on the third day, but if they be transferred to a different higher temperature at the beginning of the second day, the Grand Period occurs on the second day. Thereafter, in all cases, the rate falls slowly to the time of appearance of the side-roots ; (3) that the side-roots appear at a definite length of the main root — in all cases except one, when the main root is between 80 and 90 mm. long. They therefore appear earlier, in time, at high temperatures than at low. This agrees with Askenasy’s results so far as the appearance of the side-roots at a particular length of the main root is concerned, but he found that the rate of growth of the main root maintained a constant value from the beginning of the Grand Period till then. Text-figs. 4 and 5 show the results of Expts. 9 and 10 respectively. With regard to the bearing of these facts on the experiments in general, all measurements, being made at the beginning of the third day of growth, are made in that phase where the rate of growth is, as nearly as possible, at its highest value and most constant. Further, as has already been deter- D 34 Leitch. — Some Experiments on the Influence of mined, the rate of growth, under the conditions of experimentation, is actually constant during the whole of the third day at very low temperatures and at 250 for at least six hours. Since, therefore, all points on the curve of growth are means of readings taken over usually three half-hours (some- times one or two ; never more than six), it may be taken that the difference between readings at different temperatures is entirely due to the difference between the temperatures. On the other hand, in the long-period experi- ments there will probably be a slight disturbance due to the passing of the Grand Period during the time of the experiment. At the commencement of the microscope experiments, I decided that most time should be devoted to low temperatures and to high, because, on the one hand, I thought that there might be a possible inaccuracy in the long-period determinations at low temperatures due to the length of time necessary for the apparatus to sink from the temperature of the laboratory, at which measurements were made, to the temperature of the experimental tion room or the open air ; and, on the other hand, the long-period method is impossible as a means of accurate analysis at high temperatures. The number of experiments at low and at high temperatures is therefore much greater than at medium temperatures, but the agreement of the determinations at medium temperatures shows them to be sufficient. At low temperatures it was found that the new results agreed with the long- period experiments: but already at about io° the two curves diverge, and separate more and more rapidly to 290. An analysis of the reasons for this has not been possible, but two things are probable, (1) that the Grand Period affects the long-period experiments to some extent, and (2) that the difference in the conditions of experimentation accounts for a considerable part of the difference. At least at temperatures above 20°, preliminary experiments, before the method described for the microscope experiments was determined upon, showed that the question of sufficient moisture is of the greatest importance, and that devices such as lining the tube with moist filter-paper (perhaps comparable to, though not so good as, the con- ditions in the long-period experiments) were quite insufficient to supply the roots with the necessary moisture. With the more perfect water-supply a higher rate of growth is to be expected, and was found. The experiments from o° to 20° were in one series, those from 20° to 28° in two series, and for higher temperatures the experiments were in several series. This precaution was adopted to prevent any possibility of disturbance through the operation of a particular set of conditions, as occurred in the experiments of Series I. No indication of any such appeared ; control experiments, repeated during the experiments at high temperatures to make sure of the good condition of the roots, always gave results agreeing exactly. The experiment at 250 quoted in connexion with the question of ‘temporary stimulation’ by a rise of temperature (Expt. Temperature on the Rate of Growth in Pisum sativum . 35 No. 7) was the last experiment done, and it shows conclusively that no alteration or disturbance in the material had occurred, for the value it gives lies exactly on the growth-rate temperature curve. Table III gives the results from o° to 290, and Text-fig. 6 shows the relation of growth to temperature in graphical form. Up to and including 28°, the growth-rate is constant during the time of the experiment. At temperatures between 28° and 30° a new factor comes into opera- tion. At this point, experimentation becomes very difficult. On raising the temperature to about 290, in most cases rapid and sharp curvatures took place. These curvatures are not the gentle curvatures already referred to, extending over perhaps half the length of the root and altering in direction so little and so slowly as to have no effect on the accuracy of the measure- ments. These occur over only about 5 mm. at the tip of the root and result in a sharp inward bend of the tip, which makes measurement impossible. More than one in two of the roots used curved in this way so that they could not be measured at all, and of the others, most could be measured only for two half-hours. Any number of measurements, however great at 36 Leitch . — Some Experiments on the Influence of this temperature, would therefore be an analysis of less than half the material. Indeed, this fact, together with the fact that above 30° a few similar curvatures still take place, shows that this point is a critical point at which, even if measurements in numbers were easy, an accurate analysis would be made very difficult, if not impossible, by the necessity for referring each individual to one or other of two types, according to whether a time- factor is or is not operating, as will be clear later. Above 30° measurements are again easy ; a few curvatures still occur, but the proportion is small enough to be negligible— about 1 in 15. From 30° onwards, an entirely new set of phenomena appears. I had expected that here it would be possible to get by short-time readings values above the readings at 290, and that a time-factor would be found operating in accordance with Blackman’s theory. A time-factor does operate, but the relation between the rate of growth and the time is not a simple one. Here, owing to the difficulty of the analysis and the necessity for many experi- ments, I considered it best not to experiment at random temperature inter- vals as before, but to restrict the experiments to a few temperatures. They are, approximately, 30°, 350, 40°, 43-5°, and 450, and Table IV gives the results for these temperatures. The results for 30°, 35°) and 40° respectively are plotted in Text-figs. 7, 8, and 9. For the experiments at 30° the standard deviation and probable error have been calculated for the readings in the first ten minutes, and for those in the first half-hour. They are : Mean . Standa?'d Deviation. Probable Error. No. First ten minutes. o*68 0*128 0-033 15 First half-hour. i-74 0*279 0*043 43 In both cases the probable error is, considering the small number of readings, satisfactorily small, and compares very well with the results in the earlier experiments, Temper attire on the Rate of Growth in Pistim sativum. 37 The meaning of the results may be summarized thus : The rate of growth in the first ten minutes at 3*30° is the highest rate obtained ; in the first half-hour it falls rapidly to a mini- mum, and thereafter a recovery takes place, giving a second maximum in the fourth half-hour. After that the rate falls uniformly and rather slowly, being about two-thirds of its initial value after eight hours. At 350 the course of events is similar, but the rise and fall are steeper. Here readings were taken in the five minutes which the temperature takes to rise from about 30° to 350. These gave a mean value of 0*33 in five minutes, or o*66 in ten minutes, in contrast to 0-38 in the next ten minutes. It is to be noted that this is nearly equal to, but lower than, that in the first ten minutes at 30°. At 40° things are again entirely dif- ferent : here the rate falls uniformly and rapidly, showing no recovery. Here also readings were taken in the five-minute Text-fig. 9. Half-hours. 38 Leitch. — Some Experiments on the Influence of period while the temperature rises from about 330 to 40°, and they give a value of 0-50 in ten minutes, which is more than twice as high as the value in the first ten minutes at 40°. At 42*7° the fall is similar but steeper. At temperatures between 44'5° and 450, growth can no longer take place. A slight increase in length takes place during the raising of the temperature, but in no case is there any after the temperature has reached 44*5°, and usually contraction occurs immediately. To quote an illustrative experiment : Readings at one-minute intervals for ten minutes ; then two readings, one after five minutes and one after thirty more minutes : 'Temperature. 390, 41-8°, 42-8°, 43-8°, 44*1°, 44-3°, 44-5°, 44-8°, 44-8°, 44-8°, 44-8°, 44-9°, 44-9° Growth . O'l 000 —o*i 0000 —o*i — o*2 —0*5 In this time, forty-five minutes, the root had contracted 0-45 mm., had, as was typical, become discoloured and flaccid and, left at room temperature for two days, showed no sign of recovery. The death-point appears there- fore to lie below 450. It must be borne in mind however that, with regard to this, the death- point itself is dependent on time. All roots after a shorter or longer exposure to temperatures above 30° became flaccid and discoloured, the normal colour of the root changing to a dull white, and the greeny-yellow of the root-tip to a dull brown. At 350 this required several hours ; but, while all roots left overnight at 350 were killed, the shoots recommenced to grow at room temperature and continued quite vigorously during the next two days. At 40° growth had stopped in most roots after an hour, and in none did it continue after one and a half hours. In two hours the roots were flaccid and discoloured, and neither the roots nor the shoots showed recovery in the next two days. At 4270 growth occurred only in the first half-hour, and after one hour there was no recovery. At or below 450, death seemed to be instantaneous. Discussion of Results. Considering the results first from the point of view of Blackman’s theory, it appears at once that no extrapolation according to Blackman’s method is possible. Even if there were not the sudden drop and recovery at 30° and 350, the rate during the raising of the temperature from 30° to 350 shows that the rate at 350 never does rise above that in the first ten minutes at 30°. It is even more evident that the rate at 40° never does exceed that at 30° or 350. Again, the coefficients for a rise of temperature of ten degrees are : 10/0 — 8-25, 15/5 = 4-07, 20/10 = 2*90, 25/15 = 2*38, 29/19 — a value probably greater than 2. They show a very distinct fall as the temperature rises ; only between io° and 290 do they lie between Temper attire on the Rate of Growth in Pisum sativum . 39 1 and 3. The curve representing the relation of growth to temperature cannot, therefore, be regarded as a van ’t Hoff curve. Regarding this point, Putter, basing his theory on experiments on skin-respiration in the Frog, has attempted to show that the relation between temperature and any life-process may be expressed graphically as a van ’t Hoff curve, and that deviations from this curve are due to the superposition of exponentials. His experiments are however open to objection on the ground that he neglects to take into consideration the effect of muscle-tone, a point shown by Krogh to be of fundamental importance. Kuijper has worked out the relation between temperature and respiration in Lupinus , Pisum , and Vicia, and regards the curve which expresses the relation as a van ’t Hoff. In this case, however, as in others cited by Putter, only the coefficients of the middle region show an approximation to the proper magnitude — between 2 and 3 — and this region is arbitrarily chosen as that on which most reliance is to be placed. That there is no question of a middle region of ‘ correct 5 coefficients giving place, at the two ends, to regions in which deviations are due to extreme conditions, is indicated by the fact that, in all cases, the coefficients fall continuously from the lowest to the highest temperatures. On the other hand, Krogh insists that for ‘ standard metabolism ’ in animals, the curve relating temperature to intensity is not a van ’t Hoff curve. If we plot my curve with Kuijper’s for respiration in Pisum, and Krogh’s for ‘standard metabolism ’, on the same axes, we find that the three are strikingly similar (Text-fig. 10) ; and this fact strongly supports the view enunciated by Krogh ‘ that the temperature relations of physiological processes do follow certain typical curves which seem to be identical or nearly related for processes of the same fundamental nature in different organisms.’ It is interesting, also, to compare Kuijper’s results at high temperatures with mine. At 30° and 350 he finds a fluctuating rate of respiration, and at 40° and above, a uniform fall in succeeding time-intervals, also without any possibility of extrapolation by Blackman’s method. The question of terminology may be referred to here. These growth experiments at once suggest the old terminology, minimum, maximum, and optimum. But the use of the word ‘ optimum ’ has, since Blackman’s (’05) paper on * Optima and Limiting Factors ’, fallen, if not into disuse, at least into disrepute. Now the process of growth shows in its relation to tempera- ture three well-marked points. — 2° C. is the lowest temperature at which growth takes place. Between 440 and 450 lies a point above which growth ceases practically instantaneously. About 2g° lies a point such that any increase of temperature means the introduction of a time-factor and the con- sequent continuous decrease in the growth-rate. The same three points are distinguishable in any of the other processes the relation of which to temperature has been studied. The first of these points is the minimum ; 40 Leitch . — Some Experiments on the Influence of the second the maximum . It is suggested that the term optimum should properly be applied to the third. The reason for the confusion which o "Z /?. 6o 40-8 25-8 3 26-3° i*6o 40*8 26-0 3 26-5° 1*70 42*6 26-3 4 26*6° i-8o 45*4 26-4 6 26-7° i*6o 40-8 26-5 1 27*6° i*6o 40-8 2 7*4 1 27-7° 1-65 41-6 27*5 7 29-3° i*99 48-6 29.3 Table IV. Relation of Growth to Time at High Temperatures . The first column gives temperature ; the second, third, and fourth, the growth in the first three ten-minute intervals ; the fifth column to the tenth give the growth in the first six half-hour intervals ; in the eleventh is given the growth in the eleventh half-hour ; and in the twelfth, that in the seven- teenth half-hour. The numbers in brackets indicate the number of peas measured in each case. Temp, Ten-minute Intervals . Half-hour Intervals. 1 2 3 4 5 6 7 8 9 10 11 12 30*3° 0-68 o*55 o*54 1-74 i-68 i-77 1-90 i-88 1-78 i*5 1-2 (15) (43) (32) (20) (12) (11) (10) (3) (8) 35*o° 0-38 0-38 0-31 1-09 1-14 i*3i 1.25 1*23 1-07 0*58 (13) (15) (8) (8) (11) (10) (9) (9) 49*5° 0*24 0-19 O-II o*54 o*i6 0'04 o-o (8) (8) 42*7° 0-12 0-03 0'02 0*17 o-o 44*5° o-o 4 6 Leitch. — Experiments on Growth in Pisum sativum. Literature cited. Askenasy: Ueber einige Beziehungen zwischen Wachstum und Temperatur. Ber. d. d. Bot. Gesellsch., Bd. viii, 1890, p. 61. Blackman : Optima and Limiting Factors. Ann. of Bot., vol. xix, 1905, p. 281. Eckerson : Thermotropism of Roots. Bot. Gaz., vol. lviii, 1914, p. 254. Jost : Vorlesungen iiber Pflanzenphysiologie. 3. Aufl., 1913, p. 401. Kanitz : Der Einfluss der Temperatur auf die pulsierenden Vacuolen der Infusorien und die Abhangigkeit biologischer Vorgange von der Temperatur iiberhaupt. Biolog. Centralbl., Bd. xxvii, 1907, p. 11. Koppen : Warme und Pflanzenwachstum. Bull, de la Soc. Imp. des Natural istes de Moscou, t. xliii, 2, 1870, p. 41. Krogh : The Quantitative Relation between Temperature and Standard Metabolism in Animals. Internat. Zeitschrift fiir physik.-chem. Biologie, Bd. i, 1914, p. 491. Kuijper : Ueber den Einfluss der Temperatur auf die Atmung der hoheren Pflanzen. Rec. d. Trav. bot. neerlandais, t. vii, 1910, p. 130. Pedersen : Plaben Temperaturschwankungen als solche einen ungunstigen Einfluss auf das Wachs- tum? Arb. d. Bot. Inst, in Wurzburg, Bd. i, 1874, p. 563. Pfeffer : Pflanzenphysiologie. II, 1904, p. 87. Putter: Temperaturkoefficienten. Zeitschr. f. allgem. Physiol., Bd. xvi, 1914, p. 574. Sachs : Physiologische Untersuchungen iiber die Abhangigkeit der Keimung von der Temperatur. Ges. Abh. iiber Pflanzenphysiologie, Bd. i, i860, p. 56. Abhangigkeit der Streckungsgeschwindigkeit von inneren und ausseren Ursachen. Vorles. iiber Pflanzenphysiologie, 1887, p. 551. Schmidt : Humlestsenglens Laengdevaekst og dennes daglige Periode. Medd. fra Carlsberg Lab. Kobenhavn, vol. x, 1913, p. 211. True : On the Influence of Sudden Changes of Turgor and of Temperature on Roots. Ann. of Bot., vol. ix, 1895, p. 365. Vogt : Ueber den Einfluss des Lichts auf das Wachstum der Koleoptile von Avena sativa. Zeitschr. f. Bot., Bd. vii, 1915, p. 193. Since this paper has been in course of publication a report has appeared in Botanisches Central- blatt (1915, No. 51, p. 662) of a paper by P. A. Lehenbauer [Physiol. Res. I, 1914, pp. 247-288] on Growth of Maize Seedlings in relation to Temperature. The paper, however, I have not been able to procure. For Descriptions of the figures on Plate I, illustrating Miss Leitch’s paper, see pp. 27 and 29. A Description of a Recording Porometer and a Note on Stomatal Behaviour during Wilting. BY C. G. P. LAIDLAW, M.A., AND R. C. KNIGHT, B.Sc.1 With three Figures in the Text. IN experiments involving the estimation of stomatal aperture it is often necessary to obtain a continuous record of stomatal behaviour over a considerable period. The necessity for making continuous observations by any of the usual methods is inconvenient, especially when observations of other phenomena have to be made. Balls (1) has used with success an automatic recording porometer, the Stomatograph, the chief objection to which is its high cost. More recently, Neilson Jones (2) has devised a recording porometer which can be constructed in the laboratory from inexpensive material. The apparatus here described is also of simple con- struction, and it has been found very satisfactory for stomatal investigations. The apparatus is essentially a self-recording modification of the aspirator porometer which has been described by one of us in a previous paper (3). In that paper it should have been stated that Balls (1, p. 34) has experimented with a recording aspirator, which, however, was subsequently abandoned for the type of apparatus of his Stomatograph. In the present apparatus a head of water in a constant-pressure aspirator is employed to draw air through the leaf, and the speed of the air-stream (and therefore the relative size of the stomatal apertures) is measured by the rate at which water flows from the aspirator. The apparatus is shown diagrammatically in Fig. 1. A is a wide- mouthed bottle fitted with a rubber stopper pierced by three holes, through 1 This work was undertaken by both authors jointly, but owing to the death of Mr. Laidlaw, who fell at Richebourg l’Avouee in April, 1915, the second author is alone responsible for the statements in the paper. [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] 48 Laidlaw and Knight.— A Description of a Recording which pass three glass tubes. Tube B is connected with the leaf-chamber attached to the plant, and passes almost to the bottom of the bottle. The air drawn from the leaf enters the bottle through B. C is the water-exit tube through which water is siphoned from the bottle through tube E, whence it flows .drop by drop on to the recorder. The siphon tube C is pro- vided with a three-way stopcock, F, by means of which the aspirator bottle can be refilled from the funnel when necessary, air escaping through the tube and stopcock D during the operation. The head of water used to draw air through the leaf is regulated by the difference in level between the lower ends of tubes B and E. E is connected to the siphon tube, C, by a piece of thick-walled rubber tubing, so that any required pressure difference may easily be obtained by raising or lowering E. When the rate of dropping is slow, each drop remains at the end of tube E for some time before falling, and there is a danger of its being pre- maturely removed by air currents. To prevent this, E is surrounded by a wide tube, R, secured by a rubber stopper. The inside of R is kept damp by lining it with wet blotting-paper, in order to reduce evaporation from the drop, which might conceivably be considerable under some conditions. The speed at which the water flows from E is measured by the frequency of the drops as determined by the recorder. This consists Porometer and a Note on Stomatal Behaviour during Wilting. 49 of a strip of mica, H, secured by a binding-screw, K, at one end, carrying a thin platinum wire, L, which is bent downwards at the free end and poised over mercury in a glass tube, M, which connects with the binding-screw N. The binding-screws are secured in their respective tubes by means of sealing- wax, and IC is separated from the mercury by a sealing-wax plug, P. Leads from K and N are carried to a battery and the terminals of a magnetic pen, writing on an ordinary clockwork recording drum. We have used and found satisfactory a drum supplied by the Cambridge Scientific Instrument Company. It is about 13 cm. in diameter and revolves once in about twenty-seven minutes, being long enough to record continuously for forty hours. The mica strip is placed so that the drops from E impinge upon its free end. Thus, when a drop falls, the platinum point carried by the mica is plunged forcibly into the mercury in M, and by this means a circuit is completed through the battery and pen, and the pen is deflected? making a record on the drum. The resilience of the mica immediately withdraws the platinum point from the mercury and poises it in readiness for the next drop, so that the current passes only momentarily. In order that water may not accumulate on the mica and so prevent the strip from springing back and breaking the contact, it has been found advisable to slope the mica strip so that the water drains from the end. The mercury retains a film or drop of water upon its surface, which, however, is no disadvantage, since even if the platinum continuously dips into the water, the low voltage (about two volts) required to work the pen produces very little electrolysis, and at the same time the effects of sparking, which occurs when the contact is broken, are lessened. It has been found that the mica strip tends to oscillate after a drop has fallen, thus making more than one record on the drum. This is not a serious drawback because it is readily detected when it occurs ; it is, however, easily prevented by raising the recorder till the tube R bears upon the mica, thus pressing the platinum point very close to the mercury surface. This is found to damp the oscillations and one record per drop is obtained. The chief difficulty with the apparatus is that when air has entered the aspirator bottle, temperature changes cause changes in the volume and a consequent alteration in the rate of dropping. To overcome this, it has "been found necessary to place the aspirator in a water bath at a constant temperature. An electrically heated bath, constant to +0-02° C., has been used and found extremely satisfactory, but such accuracy as this is not essential. It is obvious that a greater volume of air in the bottle entails less accuracy, so that an experiment should be started with the bottle full of water, and refilling should be resorted to as frequently as possible. E 50 Laidlaw and Knight. — A Description of a Recording Recently boiled distilled water is used in the aspirator in order to avoid the inconvenient accumulation of air bubbles in the tubes of the apparatus, when the temperature of the water is raised to that of the bath. A record obtained with this apparatus consists of a series of marks upon the clockwork drum, the distance between two adjacent ones repre- senting the time taken for a drop to be formed and discharged. The distances between adjacent marks will therefore be a measure of the rate at which air is being drawn into, and water out of, the aspirator bottle if the drops are all of the same size. Tests have been carried out to determine the weight of water per drop. Successive single drops were found to vary considerably less than i per cent., and this variation includes errors of collecting and weighing. Variation in the rate of dropping produces variation in the size of the drops — a quicker rate giving larger drops, but this variation is most marked with quick rates. Over the range which has been found convenient in experiment, i. e. not quicker than two or three drops per minute, the variation in size of the drops was again less than i per cent. In addition to the change in size of the drops, error may result from a change of temperature of the air in the glass or rubber connexions which are not in the water bath, namely the connexion from the leaf-chamber to the air intake-tube of the aspirator bottle. Such a temperature change would result in a change of volume, and consequently of the rate of drop- ping. To reduce this error to a minimum it is advisable to reduce the volume of air in these connexions to a minimum, by using narrow-bore thick-walled rubber and glass tubing, and to shorten the distance between plant and aspirator as far as possible. If the whole apparatus is shaded from direct sunlight sudden temperature changes will be avoided. A water jacket for these connexions has been contemplated, but it has not been found necessary to resort to this. Before employing the apparatus for investigation of stomatal changes, many preliminary experiments were carried out, using fine capillary tubes and platinum discs with small punctures in them, in order to estimate the accuracy of the apparatus under constant conditions. No purpose would be served by a full account of these preliminary investigations, but the conclusions drawn from them will be briefly outlined. Several types of air intake-tube were tried, the air being made to bubble from apertures of various sizes and shapes, but a square-cut end was found to be as satisfactory as any form. Similarly a square-cut dropping-tube (E in Fig. i) was found satisfactory. Slow rates of dropping tend to decrease the accuracy, possibly due to the fact that when a drop remains for some time attached to the dropping- tube there is a greater possibility of its being prematurely shaken off, and also greater opportunity for temperature changes to become effective. The most convenient and at the same time accurate rates are from one drop Parameter and a Note on Stomatal Behaviour during Wilting. 5 1 per thirty seconds to about one drop per 120 seconds. In work with plants, the rate can be arranged by selection of the leaf- chamber and regula- tion of the pressure difference in the aspirator. Naturally, owing to stomatal changes, the rate changes considerably during a day, and very slow rates are unavoidable, but for short experiments in daylight the above-mentioned speeds have been found easy to arrange. The experiments with fine capillaries and punctured platinum discs gave indications of the existence of a rhythm in the flow of air, and although it was realized that some of the observed irregularities might be due to the capillaries themselves, attempts were made both to determine and to elimi- nate this rhythm. It was thought to be due to the small pressure changes resulting from the relative sizes of the drop at E and the bubble at B, and various combinations of sizes of these two tubes were tried, but failed to give satisfactory results. As the period of the rhythm probably changes with temperature and rate of flow, attempts to calculate it were given up, especially as the later experiments upon stomata gave sufficient accuracy without allowing for a rhythm. It was evident, however, that the disturbing influence of the rhythm was less noticeable with large than with small pressure differences, so that in so far as is compatible with other conditions of the experiment, it is advisable to employ large pressure differences; from 5 to 15 centimetres have been found convenient. By taking the precautions outlined above, it has been found possible to obtain very satisfactory results with the apparatus: thus, on July 2, 1914, with the apparatus attached to a fine capillary tube and a pressure difference of 10 cm., a record running for two hours gave an average distance travelled by the drum during the formation of one drop of 5*0 divisions of the paper used (= about 55 sec.), the maximum distance for any drop being 5*2 divisions and the minimum 4-8. Typical records obtained with stomata are given in the second part of this paper. Test experiments have also been carried out to compare the records obtained with the apparatus with that obtained with the ordinary porometer by stop-watch readings, of which the following is typical : Exp. no. Five leaf-chambers were attached to one leaf of Maranta coccinea , var . floribunda at 4.30 p.m. on March 4, 1915. At 11.15 a.m. the next day, two were connected to two recording porometers, and the other three by means of a 4-way tube to a single bubbling aspirator for stop-watch readings (3), which were taken every fifteen minutes. The results are shown in Fig. 2, where the reciprocals of the distances between the two succes- sive marks on the drum and the reciprocals of the stop-watch readings are plotted against time. The recorders of course gave continuous records, but as each curve would involve about 350 points, only one point for each five minutes period is given. It has been shown in some work which it is E 2 52 Laidlaw and Knight —A Description of a Recording hoped to publish shortly, that the stomata upon different parts of a leaf may be considered to behave similarly under similar conditions of illumination, &c.j so that the curves in Fig. 2 are an indication of the relative accuracy of the recording apparatus and the ordinary porometer. The maxima in the curve of the latter are repeated in both recorder curves, and altogether here is a striking parallelism. Stomatal Behaviour during Wilting. It has been stated by Darwin (4, 5) and Darwin and Pertz (6) that, in the case of plants investigated by them, on severing a leaf from the stem and allowing it to wilt, there occurs a temporary opening of the stomata prior to the closure following upon wilting. This preliminary opening has been demonstrated by the above authors by three different methods at different times — by the horn hygroscope, by temperature methods (in which, however, the assumption is made that the amount of transpiration from a leaf is an indication of the condition of its stomata), and also by the porometer method. Lloyd (7), using his alcohol method of stomatal Porometer and a Note on Stomatal Behaviour during Wilting. 53 measurement, failed to find any indication of the temporary preliminary opening described by Darwin and Pertz. The recording apparatus described above seemed to adapt itself well to a question of this kind, and consequently experiments were undertaken to investigate it. The usual method adopted was to attach two leaf-chambers to different leaves of the plant to be investigated, and some hours later — generally the next morning — each was connected to a recording apparatus, and records were commenced. After an interval of an hour or more, one of the leaves was severed from the plant, the other being left attached to the plant to act as a control for temperature, humidity, and illumination changes. The experiment was continued, after severing one leaf, for a period determined by the result, and in many cases the second leaf was also severed later. The plants used included Maranta bicolor , M. coccinea , var . floribunda. Primus Laurocerasus , Eucharis Master si , Eupatoriuni adenophorum , E. Raffilli , Pelargonium (ivy-leaved), and Phaseolus vulgaris . With the exception of Eucharis Mastersi all these plants showed the preliminary opening described by Darwin. Only one experiment was carried out with the Eucharis , and it is doubtful if the leaf was appreciably wilted at the end of the- experiment, since the lamina is fairly fleshy and has a thick midrib which would hold much water, so that although no preliminary opening occurred in our experiment, further investigation might show that Eucharis is no exception. Fig. 3 shows the result of one of these experiments upon Phaseolus vulgaris . The time elapsing between severing the leaf from the plant and the opening movement of the stomata varies considerably with different plants and different conditions, and is apparently dependent upon the rate of wilting. In the case of a thin leaf which wilts quickly, the opening move- ment occurs very shortly after the leaf is cut from the plant, whilst in a thick leaf the opening may be long delayed. In the curves of stomata! 54 Laidlaw and Knight.— A Description of a Recording aperture shown in Fig. 3, the opening reached a maximum in both cases about five minutes after the severance from the plant, whilst in the case of the thick leaf of Prunus Laurocerasus , even when the temperature was 200 C., tending to produce rapid wilting, nearly twenty minutes elapsed before the maximum opening was reached. Similarly, with any particular plant, a higher temperature, tending to produce more rapid wilting, causes the maximum to be reached more quickly, and a roughly graded series may be obtained, as in the following experiments upon Maranta bicolor : 9 Experiment No. Mean greenhouse temperature (° C.). Twie elapsing before max. opening is reached. 100 16* 3 35 minutes. 102 A 18*0 13 102 c 18-0 T4 ?> 103 18-8 20 „ 106 20*5 9 The extent to which the stomata open on severing the petiole is variable, depending probably upon the plant used, and possibly also upon the rate at which the leaf wilts. The greatest opening observed was with a leaf of Phaseolus vulgaris , in which the stomatal aperture, as deduced from the speed of the air-stream, being represented by the value 168 before cutting the leaf from the plant, increased to 475 within five minutes after cutting. The results of these experiments are thus in close agreement with those obtained by Darwin and Pertz. In his first reference to the phenomenon, Darwin suggested (4, p. 617) that the temporary stomatal opening is a direct result of wilting, and is due to the guard-cells retaining their water, and therefore remaining turgid longer than the other epidermal cells. Thus the pressure of the rest of the epidermis upon the guard-cells is reduced early in the process of wilting, with the result that the size of the stomatal pore increases. The explanation offered by Darwin is no doubt the most probable one, but no experimental evidence was adduced in support of it, and alternatively there seem to be two other possible explanations which would fit the observed facts. The stomatal opening might be due, not to any effect of wilting, but to the shock sustained by the leaf as a result of severing it from the plant. In another connexion it has been demonstrated that the mere handling of some leaves may produce stomatal closure, but opening has never been observed to occur as the result of shock. Another possibility was that the change in the porometer readings on detaching a leaf from the plant might be due to some cause other than stomatal opening. In the unpublished work above referred to, it was observed that in Porometer and a Note on Stomatal Behaviour during Wilting. 55 at least one plant, Eucharis amazonica , it was possible in a porometer experiment with a normal attached leaf, that some of the air drawn from the leaf entered it through the petiole and not through the stomata, since, when all the stomata outside the leaf-chamber were blocked, a current of air could still be drawn through the leaf unless the petiole was severed and blocked, e. g. by immersing in water. It was thought possible therefore that when a leaf was detached from the plant and the petiole exposed, the air-stream might find the path through the petiole of less resistance than formerly. This would result in an increase in the speed of the air-current through the leaf, followed by a decrease when the stomata closed. If this were the case, it is to be expected that the temporary ‘ opening ’ would reach a maximum almost immediately after detaching the leaf. The results show that some minutes elapsed in every case before the maximum was reached, which also confirms the observations of Darwin and Pertz. This explanation of the phenomenon applies only to the results obtained by the porometer method, but Darwin’s experiments with the horn hygro- scope and temperature methods (4 and 5) showed clearly that an increase in transpiration occurred when the leaf was detached ; this he adduced as further evidence for the view that the stomatal apertures had increased. Dixon (8) has, however, suggested that the result of detaching a leaf is to reduce the tension in the water columns in the tracheae and thereby permit of more active evaporation from the mesophyll cells. Experimental tests of these possible explanations were undertaken, using chiefly Phaseolus vulgaris , Eupatorium adenophorum , and Maranta coccinea , var. Jioribunda. To prevent any flow of air through the petiole after severing it, the cut end was blocked by various means. Vaseline was found to be unsatisfactory owing to the difficulty of attaching it to the wet surface. Immersion in mercury, and coating the end with a wax mixture, stiff gelatine and stiff glue, were among the methods used to close the cut end of the petiole. The results showed that blocking the petiole by these means did not prevent the usual temporary opening, even though, in the case of the experiments with gelatine, the petiole was cut beneath the surface of liquid gelatine, and the cut end was therefore never exposed to the air. The phenomenon cannot therefore be attributed to the leakage of air through the petiole. If, however, the petiole is cut below the surface of water, and kept supplied with water, the temporary opening does not occur. Instead there is either a slight tendency to closure, or else the stomata behave quite similarly to those of the control leaf still attached to the plant. Under these conditions the leaf does not visibly wilt in the course of several hours, or at most is only slightly less turgid than the leaves attached to the plant. This result disposes of the possibility of shock being responsible for the 56 Laidlaw and Knight. — Description of a Recording Porometer . temporary opening, since severing the petiole beneath gelatine cannot cause more or less shock than severing it beneath water, and yet in the one case the opening occurs, and not in the other. At the same time it is shown that when the leaf does not wilt the effect of severing it from the plant is practically nil. From our experiments, therefore, it appears that when a leaf is detached from the plant and allowed to wilt, the stomata open for a short time before finally closing, and these results are quite in accordance with the observations of Darwin and Pertz. In addition, some direct evidence has been obtained in support of the explanation of this phenomenon offered by Darwin, viz. that it is the direct result of wilting probably due to the guard-cells retaining their turgor longer than the other epidermal cells. Neither stomatal closure due to shock nor the entrance of air into the leaf through the petiole can account for the increased porometer readings after detaching a leaf. Summary. 1. A description is given of a modification, making it self-recording, of the aspirator porometer described earlier. 2. Experiments on various plants confirm the observations of Darwin and Pertz — that on detaching a leaf from the plant and allowing it to wilt, the stomata open temporarily before finally closing. 3. Evidence is adduced in support of Darwin’s suggestion that this phenomenon is due to wilting. Department of Plant Physiology and Pathology, Imperial College of Science and Technology. References cited. 1. Balls, W. L. : The Stomatograph. Proc. Roy. Soc., vol. B. 85, T9I3> P- 33- 2. Neilson Jones, W. : A Self-recording Porometer and Potometer. New Phytologist, vol. xiii, I9I4? P- 353- 3. Knight, R. C. : A Convenient Modification of the Porometer. New Phytologist, vol. xiv, 1915, p. 212. 4. Darwin, F. : Observations on Stomata. Phil. Trans., vol. B. 190, 1898, p. 548. 5. A Self-recording Method applied to Movements of Stomata. Bot. Gaz., vol. xxxvii, 1904, p. 89. 6. Darwin, F., and Pertz, D. F. M. : New Method of Estimating the Aperture of Stomata. Proc. Roy. Soc., vol. B. 84, 1911, p. 149. 7. Lloyd, F. E. : Physiology of Stomata. Carnegie Institute Publications (Washington), No. 89, 1908, p. 81. 8. Dixon, H. H. : Transpiration and the Ascent of Sap in Plants. Macmillan’s Science Mono- graphs, 1914, p. 124. On the Use of the Porometer in Stomatal Investigation. V • BY R. C. KNIGHT, B.Sc., D.I.C. With seven Figures in the Text. IN the course of some investigations upon transpiration it was found necessary to record the changes taking place, under varying conditions, in the size of the stomatal pores of the plants under observation. The methods which have hitherto been employed for this purpose are various and widely divergent in principle. Some of the earlier observers assumed that the amount of transpiration from a plant was an index of the size of the stomatal apertures, and accordingly they expressed stomatal changes in terms of transpiration, as measured by experiments with the horn hygroscope (Darwin) (1) and cobalt chloride (Stahl) (2). Such methods must be rejected as untrustworthy in the absence of evidence in support of the assumption underlying them. At the other extreme is the direct method used by Lloyd (for Gossy- pium ) (3), in which the stomatal apertures are measured in situ in the living leaf by means of a microscope, a micrometer scale, and special illumination. The two methods now most commonly used are : (i) Lloyd’s earlier alcohol method (4) and (2) the indirect porometer method of Darwin and Pertz (5). The relative values of these two methods have been discussed by the two latter authors (loc. cit.), the chief points considered being the advantage in Lloyd’s method of taking observations over many different parts of the plant, and the advantages in the porometer of automatically averaging many stomata and of dealing with the same stomata through- out the experiment. In addition, the manipulation of the porometer is very simple, while, as will be readily admitted by any one who has attempted it, the accurate measurement of stomatal aperture under a microscope is very difficult and would seem to be liable to serious error. In the light of these considerations, therefore, the porometer method was adopted for the work in question, but at the same time it was realized from an a priori analysis of the conditions of the experiment, that there was ample opportunity for errors to creep in unless suitable [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] 58 Knight . — On the Use of the precautions were taken to obviate them. Accordingly, some preliminary experiments were carried out with a view to determining these possible errors and the steps to be taken to avoid them. The results of these investigations are embodied in the present paper. The majority of the experiments here quoted were carried out with the form of porometer described in a previous paper (6). The conclu- sions arrived at fall naturally into three main groups, and it is with these in view that the experiments have been classified, although many of them might be included in more than one group. Changes in Stomata. In dealing with such admittedly sensitive organs as stomata, it seemed possible that rough treatment might easily cause changes in the size of the pore, and some experiments were therefore carried out to determine whether the conditions of the porometer experiment affected the stomata. The main source of possible error appeared to be changes in stomatal size resulting from : (1) the mechanical strain on the leaf due to the reduced pressure ; (2) the passage of the air-stream through the leaf ; and (3) the effect of shock from the manipulation necessary for fixing the leaf-chamber. (1) One aspect of the question of mechanical strain has already been discussed in a previous paper, where it was pointed out that even a small pressure difference will cause a curvature of the portion of the leaf under observation, and this effect cannot therefore be neglected. Although there is no direct evidence on the question, it is possible that the stomatal pores in general will be affected by a change of curvature of the leaf. The acceleration (or retardation) of the air-current through the leaf, caused by this stomatal change, will obviously be incorporated in the porometer readings, and will constitute a regularly occurring error. If, however, the pressure difference employed to produce the air-current is maintained constant throughout an experiment, the leaf-curvature, and consequently the error resulting, will also be constant. Since this error will have a different value for every plant, perhaps for every leaf, and for each different pressure, it seems hopeless to attempt to determine it. As Darwin pointed out, the porometer gives comparative measurements of the direction and relative amount of stomatal changes, but not the actual dimensions of the pore ; this being so, a constant error, if small, will not detract from the comparative value of the readings, so that if the pressure differences used are as small as is convenient the curvature of the leaf is a factor which may be neglected.1 1 By supporting the leaf by means of a grid placed between it and the chamber the curvature might, no doubt, be avoided. Porometer in Stomata l Investigation . 59 (2) Errors arising from the effects of the passage of the air-stream through the leaf are, from their nature, unavoidable, but a few experiments have been carried out in reference to them. One effect of the air-stream is to replace the moist air in the inter- cellular spaces by drier air from without, which will result in an increase of transpiration from the leaf. If this increase is sufficiently large, it might conceivably result in wilting and stomatal closure, but if the air-stream is passing for only a short period — the usual procedure — the effects must be practically negligible. An attempt was made to determine experimentally any after effect resulting from continued passage of the air-stream through a leaf. Two or more leaf-chambers were attached to different leaves of the same plant, and porometer readings taken at intervals. The leaves were considered as mutual controls, the justification for which will appear at a later stage in this paper. Except when readings were actually being taken the leaf- chambers were in direct connexion with the outer air, so that there was no undue pressure upon the leaves and no artificial air movement within them. Having by this means determined the direction and slope of the curves of stomatal movement of each leaf during a period of two hours or so, air was drawn through one leaf at constant pressure for an hour or more, the periodic readings being continued for all the leaves. In some experiments the air-stream was then transferred to other leaves, the first leaf now acting as a control. By this means it was expected that any effect the air-current might have upon the stomata would be re- flected in the readings during, and after, the period in which the stomata were subjected to a continuous stream. The plants used were Ficus elasticai Eucharis M as ter si, and Begonia (‘President Carnot’), and the results were not the same for the three. Ficus and Eucharis were quite unaffected by the air-stream, whereas in the case of Begonia the stomata showed a distinct and sometimes con- siderable closure as a result of drawing air through, and when the stream was stopped there was a tendency for the curve to return to its former slope. Figs. 1 and 2 show curves obtained in experiments with Begonia and Eucharis respectively.1 Thus, although the conditions of the experiment were somewhat severe as compared with the ordinary porometer experiment, it is obvious that stomata may be affected by the air-stream used in the latter. The resulting error will be reduced to a minimum by ensuring that between two readings the air-current is stopped, and that, in addition, the time occupied in taking a reading is reduced to a minimum. 1 These and all other curves in this paper were obtained by plotting the reciprocals of the time between successive bubbles, not the square roots of the reciprocals, which was the method mainly adopted by Darwin and Pertz. 6o Knight. — On the Use of the Fig. i. Effect upon the stomata of the continued passage of air. Begonia (‘ President Carnot’). Fig. 2. Effect upon the stomata of continued passage of air. Eucharis Mastersi. In addition to the above-mentioned stimulation, which partakes of the nature of a cumulative effect, there is the possibility that a change in the stomata may be produced immediately the air-stream begins to pass through a leaf, an effect which is unaltered by the continued appli- Porometer in Stomatal Investigation . 61 cation of the stimulus. Such an effect may be the result of forcing apart resilient guard-cell walls which return to their original conformation imme- diately the rush of air ceases. By the porometer method it is obviously impossible to detect such an error, and equally impossible to avoid it, but the existence of any large error of its kind seems unlikely. (3) In view of the fact that a porometer experiment involves the attachment of the leaf-chamber to the leaf by means of some adhesive, a somewhat violent treatment, some experiments were carried out to obtain some indication of the response, if any, of the stomata to shock. In some experiments, in another connexion, it was observed that vaselining the upper surface of a leaf of Acalypha Wilkesiana var. mar - ginata with a brush caused rapid closure of the stomata, which are con- fined to the lower surface. This was attributed to shock, but the experiments were not followed up. In later experiments it was frequently noticed that shortly after fixing the chamber to the leaf, the size of the stomatal apertures, from porometer readings, was much less than, in the light of later readings from the same portion of leaf, seemed consistent with the general con- ditions of atmosphere and illumination. It was thought that, possibly, this was due to the fact that the manipulation necessary for the attachment of the chamber to the leaf caused the stomata to close. Confirmation of this was sought in a series of experiments, all of which gave the same results ; one of which is given in detail below. Expt. 56. 15. xii. ’14, and 16. xii. ’14. The plant used was Eucharis Master si, and at 2.0 p.m. on 15. xii. T4 two similar leaf-chambers, B and D, were fixed to different parts of the same leaf. Readings were taken at intervals from 2.30 p.m. onwards, and showed that the stomata opened rapidly until about 3.30 p.m., when they began to close, doubtless owing to the failing light (see Fig. 3). The readings were continued the next morning and throughout the day. At 11.35 a*m* another chamber, c, was attached to the leaf and readings taken from it as from the others. The most important point revealed in the resulting curves is that the stomata at C were at 13.10 p.m. almost closed, but they rapidly opened to a maximum at 1.30 p.m. at the same time as B and D, finally closing in unison with them. It may be assumed that the stomata at C behaved similarly to those at B and D on the morning of 16. xii. T4, and that at 11.35 a>m* t^ey were fairly widely open, but that they closed in response to the shock sustained in fixing the chamber C, but recovered by 1.30 p.m. The same conclusion applies to the closed state of the stomata at B and D at 3.30 p.m. on 15. xii. T4. 1 N.B. — The reason for lack of absolute coincidence, even under practically identical conditions, between b and d is discussed in a later section of this paper. 62 Knight.— On the Use of the The results of this series of experiments show that the treatment to which a leaf is subjected in affixing a chamber produces a definite stimulation of the stomata, causing them to close to a large extent. Recovery from the shock is fairly rapid, taking place in two hours or so, and is apparently complete, since it is found that a chamber can be left attached to a leaf for days, and readings may be taken daily, whilst the stomata still respond to light and darkness, and the leaf is not obviously injured. A second method was also adopted in the attempt to discover the effect upon stomata of shock. Two or more chambers were attached to Fig. 3. Effect upon the stomata of fixing a porometer chamber. Eucharis Masteri. different leaves of a plant, and when the curves of stomatal aperture had attained a regular form, one of the leaves under observation was tapped lightly several times with a pencil, or was shaken, and the effect upon the stomata was noted in the succeeding readings. It is naturally somewhat difficult to standardize tapping or shaking, and it was there- fore impossible to treat each leaf similarly. It is possible that this accounts for the variety of results obtained. Begonia showed no effect at all, the curves of stomatal aperture having similar forms whether the leaf was tapped or not. Ficus elastica showed no change in two cases, but in a third there was a tendency for the stomata to close as a result of tapping. Eucharis Mastersi fairly con- sistently closed its stomata, temporarily, in response to tapping. With all three plants several experiments showed slight effects which might be due to merely experimental error. Porometer in Stomata l Investigation. 63 Therefore to avoid errors from this source it is advisable to touch or shake the leaf as little as possible after affixing the leaf-chamber. It was with this end in view that a three-way stopcock was inserted between the leaf-chamber and the aspirator bottle in the apparatus used for the porometer readings (6) ; for by this means the chamber can be connected with the outer air with the minimum risk of shaking. It is obvious that Lloyd’s alcohol method of measuring stomata is not applicable to plants with sensitive stomata, since the manipulation required by his technique would be sufficient to cause change of aperture. Effect of Intercellular Spaces. The air drawn through the leaf in a porometer experiment must enter the leaf through stomata situated outside the leaf-chamber, must pass through a length of intercellular spaces, and finally leave the leaf through the stomata covered by the chamber. The rate at which this air moves is determined by the pressure difference, and by the resistance encountered, part of which is due to the stomata and part to the intercellular spaces. Any change in the size of the stomatal pores will be reflected in the resistance they offer to the air-stream, and therefore in the speed of the air, if the size of the intercellular spaces remains constant. In inter- preting the rate of air-flow as an indication of stomatal aperture, this constancy is necessarily assumed, but it would seem probable that when a leaf wilts, for instance, the intercellular spaces will be appreciably smaller than when the tissue is turgid. But in a normal daily cycle of changes, the variation is likely to be small, and the influence the spaces have upon the readings will depend upon the relative resistance to air-flow of stomata and spaces respectively. If the resistance is mainly due to the stomata, changes in the intercellular spaces will have but little effect upon the readings, but if a considerable proportion of the resistance is due to the spaces, then their changes will be reflected in the readings, and, in addition, changes in the size of the stomata will be appreciably masked by the part played by the spaces, and the sensitiveness of the method will be impaired thereby. The following series of experiments constitutes an attempt to estimate the relative importance of the parts played by the stomata and intercellular spaces in resisting the air-flow. From the nature of the problem, an accurate estimate of the relationship is almost impossible, especially as it is probably different for every plant and is continually changing, so that no quantitative generalization has been made. Four methods have been employed : j. Two chambers were fixed to the same leaf some distance apart, and periodical readings taken from each, preferably till a fairly constant value was reached in both. Vaseline was then smeared on that portion of the 64 Knight . — On the Use of the leaf immediately adjacent to one of the chambers, so as to block the stomata over an area in the form of a ring around the chamber. This treatment increases the length of intercellular spaces to be traversed by the air-stream, and the difference between readings taken before and after vaselining will be due to the added resistance of the extra intercellular spaces, provided that the control chamber gives readings similar to those taken previously. By vaselining a portion of the leaf-surface a number of stomata are blocked, but as long as the unvaselined area is considerably greater than the area under the chamber the rate of the air-stream will not be limited by the unvaselined portion. Expt. 8. 2. xi. ’13. Two chambers, A and B, fixed one on each side of the midrib of leaf of Eucharis Mas ter si. Area of leaf under chambers : A, 1-4 sq. cm. ; B, 1-3 sq. cm. Area of leaf unvaselined outside the chambers : on A side of midrib, 106 sq. cm. y> B v » j) 1^3 >> Width of ring of vaseline, approx. o-8 cm. Thus the increase of length of intercellular spaces to be traversed was o-8 cm., and the unvaselined area was 75 to 95 times the area under the chamber. Table I. The readings are given in seconds and represent the time required to draw a given volume of air through the leaf. The vaseline was applied at 12.30 p.m. B was the control. A. B. Time . Reading . Time. Reading. 12.15 p.m. 12*8 12.17 p.m. n*7 12.25 12'9 12.27 n*7 12.30 p.m. [Vaselined] — — 12.37 23-1 12.40 13*1 12.47 20*1 12.50 n*6 12.57 17*1 1.0 10*7 1.25 15*7 1.27 9-8 2.8 i6-i 2.10 io*5 The immediate effect of vaselining was to reduce the rate of air-flow in both cases, which was to be expected in view of the fact already established that handling causes stomatal closure in Eucharis Mastersi . The decrease in rate was however much greater for A than for B, and by 2.10 p.m., when both had apparently recovered and the stomata were closing, presumably in response to some natural stimulus, the air-stream in A was still much slower than in B. By comparing in the case of A the two lowest readings (12-8 and 15*7) we find that the rate has decreased by about 20 per cent, as a result of increasing the length of intercellular space traversed by only o-8 cm., at Porometer in StomcUal Investigation. 65 the same time the control reading has changed from 11-7 to 9*8, representing an increase of 16 per cent, in the rate ; the effect of vaselining is clearly much greater than the readings of A indicate. A modification of this method has also been employed with similar results. In this case the whole leaf outside the chamber was vaselined, except a long rectangular area, the shorter side of which was immediately adjacent to the leaf-chamber. Readings having been taken, the length of intercellular spaces to be traversed was increased by vaselining a portion of the rectangle nearest the chamber. In this case an increase of 3*0 cm. produced a decrease in rate of 35 per cent., the control remaining practically constant throughout. The plant used was Eucharis amazonica , which, as well as the other species of Eucharis used in the experiment previously described, has stomata only on the under surface of the leaves, with occa- sional ones on the petiole. The case of hypostomatous plants of this type is somewhat different from that of plants with stomata upon both leaf- surfaces, such as species of Helianthus , Saxifraga , and Richardia which have been used in later experi- ments. In the latter types the air-stream is at liberty to enter the leaf from one side and pass straight through into the chamber on the other side, reducing the length of intercellular spaces traversed to the thickness of the leaf. A vase- lining experiment, as described above, upon an amphistomatous leaf would require the whole of the stomata upon one surface to be blocked before measurements could be made. The question of amphistomatous leaves will be further discussed later. 2. A more convenient and efficient method of carrying out what is practically the same experiment as that described above has since been used. It depends on the use of a special type of double chamber of the form shown in Fig. 4. This consists of one chamber within another, so arranged that both can be fixed to the same leaf, each being provided with an independent outlet. Those used in the following experiments have been generally constructed in two parts and joined by means of rubber tubing or sealing-wax (at A in Fig.). The double chamber is fixed to a leaf and readings are taken from the inner chamber in the usual fashion. The air enters the outer chamber at C, passes into the leaf at D, and is drawn into the inner chamber at H. This constitutes an ordinary porometer reading from a single chamber. The outlet C of the outer chamber is now closed, and air must now enter the leaf at E — beyond the outer chamber, and to reach the inner chamber must traverse a longer path within the leaf, Fig. 4. Diagram showing the form of double chamber used. 66 Knight.— On the Use of the Readings are again taken, and the differences between these and the previous ones are an indication of the added resistance due to the extra length of intercellular spaces through which the air has now to pass. This method is easier of manipulation than the vaselining one, in that the extra length of tissue can be introduced at will and without any shock to the plant, an important consideration in view of the effect of shock upon the stomata already described. In constructing the double chamber it is advisable to ensure that the area of leaf between the outer and inner chambers (i. e. DD in Fig.) is somewhat greater than the area covered by the inner chamber (h), so that, when the outer chamber is open, the conditions may closely approximate to the ordinary single-chamber experiment. If DD is greater than H, there will be less tendency for air to enter the leaf at EE when C is open. The usual method of procedure was to take three readings in succession with the outer chamber open, the mean of these representing the stomatal resistance plus the resistance of a short length of intercellular spaces. The outer chamber was then closed and readings taken continuously till they became constant. Finally, the outer chamber was again opened and three more readings taken, in order to allow for stomatal changes during the experiment, the mean of the first and last three being considered to be the correct c open ’ reading. It was found that readings taken immediately after closing the outer chamber were not always the same as those taken a few minutes later, and this was attributed to the fact that at the moment of closing the outer chamber the air in it was at atmospheric pressure, but before any equilibrium could be set up the pressure in the outer chamber must approach that in the inner one. The difference between the ‘ open ’ and ‘ closed * readings was expressed as a percentage of the ‘ open ’ reading, this latter being regarded as the closer approximation to the indication of stomatal resistance. For each leaf investigated, several of these composite readings were taken at different times of the day, corresponding to different dimensions of stomatal aperture, and the series of percentages thus obtained were plotted as ordinates with the reciprocals of the ‘open’ readings as abscissae, i. e., roughly, the effect of the added length of intercellular spaces was plotted against stomatal aperture. It must be noted here that this method of plotting is entirely arbitrary, and the figures obtained only refer to one particular set of dimensions of the chambers used, but by expressing the results in this manner the conclusions drawn from them are easily demonstrated. Experiments with double chambers have been carried out with a variety of plants, including Hedera helix , Eucharis amazonica , E. Mastersi f 67 Porometer in Stomatal Investigation. Ficus clastica , Phaseolus vulgaris , Helianthus annuus , Begonia (“ Gloire de Lorraine ”), Begonia (“ President Carnot ”), Aucuba japonic a, Richardia aethiopica , Saxifraga c or di folia, Moms nigra, Eupatorium Raffilli, Dracaena Godsejfiana , and Piper dilatatum. These are hypostomatous plants with the exception of Helianthus , Richardia , Saxifraga , and Dracaena , which are amphistomatous. The following experiment shows a typical result obtained with a hypostomatous leaf. Expt. 38. 26. xi. *14. A double chamber attached to leaf of Begonia (‘ President Carnot ’). The dimensions of the chamber were as follows : External diameter of inner chamber, i-o cm. External „ „ outer „ 3-5 „ so that the extra length of path traversed by the air when the outer chamber was closed was 1-25 cm. approx. Readings were taken during both the morning opening and evening closing of the stomata. The curve obtained is shown in Fig. 5. 50 n id o £ 40 x I- V 5 30 20 o 10 1 1 RECIPROCALS OF OPEN' READINCS [STOMATAL APERTURE] Fig. 5. Effect of the intercellular spaces of the leaf upon porometer readings. The points do not lie upon a smooth line, but this is not surprising since slight changes in the size of the intercellular spaces are probably of frequent occurrence. The chief point for consideration is that, with increasing size of the stomatal aperture, the relative part played by the 68 Knight. —On the Use of the intercellular spaces also increases. This has been the case in all the hypo- stomatous leaves investigated, and is in complete accord with expectation. As the stomata open, the resistance offered by them to the air-flow decreases, and that offered by the intercellular spaces remains approximately constant, so that, relative to the total resistance, that of the intercellular spaces increases as the stomata open. Another phase of the same phenomenon is evident in the differences in different plants. A plant whose stomata are large or are capable of wide opening, such as species of Eucharis, shows relatively a much greater intercellular space resistance than a plant with fewer or smaller stomata, as Enpatorium Raffilli. With the double chamber of dimensions stated above, 59 per cent, was the highest value obtained for the relative resistance of the spaces in Eucharis amazonica, whilst in Enpatorium Raffilli the value never rose above 18 per cent. The case of the amphistomatous leaf is somewhat different from that of the hypostomatous one. The double-chamber experiment with the former has not the same significance as with the latter, unless the stomata upon the surface opposite to that on which the chamber is fixed are blocked. Experiments of this type have been performed with the amphistomatous leaves mentioned in the above list, and all of them gave results comparable with those obtained with hypostomatous leaves. In an ordinary porometer experiment with an amphistomatous leaf, there is available for the air-stream a path directly through the leaf from one surface to the other, and it is probable that this direct path offers less resistance than the longer one which the air necessarily traverses in a hypo- stomatous leaf. It is therefore likely that the air will take the path of least resistance and pass directly through an amphistomatous leaf. A test of this can be carried out by a double-chamber experiment without previously blocking the stomata of one surface. If the resistance of the direct path is relatively small, then the greater part of the air will pass along it, and closing the outer chamber will have little or no effect upon the readings ; but if the resistance of the direct path is relatively large, some portion of the air-current will be diverted by closing the outer chamber, and a decrease in speed will result. In the case of Richardia, when neither surface was vaselined, the effect of closing the outer chamber was never more than 3 per cent.- — probably within the limits of experimental error — but when the stomata on one surface were blocked, the effect rose as high as 29 per cent. Dracaena, on the other hand, showed an effect up to 12 per cent., whether one surface was vaselined or not, the double chamber used having the same dimensions as that previously described. 3. Owing to its simplicity of manipulation, the method just described Porometer in Stomatal Investigation. 69 has been used for a variety of plants, whilst other less simple methods have been used in one or two experiments on suitable plants. Of these the following method provides clear demonstration of the part which inter- cellular spaces may play in retarding the air-stream. Several chambers (four and five have been used) are fixed in a straight line on one leaf, preferably of uniform nature. Small chambers or a large leaf must be employed — Encharis Master si and E. amazonica have been found to answer well with chambers of 5 mm. internal diameter. The leaf is now smeared with vaseline over the whole stoma-bearing surface, except for a small area, 1 sq. cm. or so, just beyond the end chamber of the series and in line with them. This area serves for entrance to the leaf of the air- stream, which is drawn out through the chamber at the other end of the line by a constant-pressure aspirator in the usual manner. Each of the inter- mediate chambers is connected to a glass tube dipping into water and serving as a manometer. A stream of air entering the leaf at F and leaving it at A (see Expt. 52, p. 70) will encounter the resistance of the stomatal pores at those points, and also that of the tissue between them. If this latter resistance is not appreci- able, the passage of the air from F to A within the leaf will be easy, and the air pressures in the intercellular spaces at F and A will not be widely different. If, on the contrary, the resistance offered by the tissues is considerable, there will be a pressure-gradient in the leaf along the line FA, which will be indicated by the manometers B, c, D, E. In practice it is found that the movements of the water-columns in the manometers are slow, owing to the pressure having to be transmitted through the stomata, so that air is drawn through the leaf continuously for some hours and the manometers read at intervals. A rough estimate of stomatal changes is deduced from the rate of the air-stream. In all experiments in which extensive blocking of the stomata was resorted to — generally with species of Eucharis — it has been found expe- dient to cut the leaf from the plant and immerse the cut end of the petiole in water, on account of the tendency of air to pass through an attached petiole from other parts of the plant, when air is drawn from the leaf into the leaf-chamber. This can be demonstrated by fixing a chamber to a leaf and vaselining the whole of the rest of the surface, when it will still be found possible to draw air into the chamber. If the petiole is now severed and the end immersed in water, the air-stream stops. On cutting the petiole again above the water, air can be again drawn through, showing that the previous cessation of the stream was not due to stomatal closure. There is little possibility of the experiment being affected by severing the leaf, as is shown by the fact that three days after a leaf was detached and its petiole immersed, the stomata were still responding to the daily changes of illumination. 7o Knight. — On the Use of the Expt. 52. 10. xii. ’14. The plant used was Eucharis Mastersi. Five chambers (a, b, C, d, e) were fixed to a leaf and the leaf vaselined as described above. The chambers were about 2*2 cm. apart. Pressure difference in the aspirator was 18-5 cm. Manometers were read every thirty minutes ; the maximum heights reached, excluding capillary effects, were : B. C. D. E. F. io*o cm. 8*7 cm. 7*3 cm. 6*3 cm. (unvaselined area) Rough approximations from these figures give the pressure inside the leaf at A as 11-3 cm. below an atmosphere, and at F 5*5 cm. below, i. e. the difference of pressure on the two sides of the barrier of the stomata was 7-2 cm. (18*5-11-3) at A, and 5*5 cm. at F, the lack of agreement probably being due to the greater area of leaf exposed at F. Thus, whilst in passing from the leaf through the stomata into the chamber the pressure change is 7 cm., in passing through a centimetre of leaf-tissue the pressure change is at least 0-5 cm., showing that the inter- cellular spaces may offer a relatively large resistance to the air-current. 4. An extension of the pressure-gradient experiment has also been used to supplement the methods already described. A series of chambers is attached to a leaf, and the remaining leaf-surface is smeared with vaseline. The end chamber of the series is connected with the aspirator, and the others are provided with stopcocks so that air may be allowed to enter the leaf through any particular chamber at will. Whichever path the air-stream is made to traverse, it encounters two sets of stomata, one at the entrance and one at the exit, but the length of intercellular spaces to be traversed may be varied at will by means of the stopcocks, and the difference in speed caused by different lengths can be measured in the usual manner. It must be noted that for the correct working of this experiment the areas exposed under the leaf-chambers must be equal, and to ensure this is far from easy in practice. If, however, the areas are approximately equal, and if the areas beneath chambers farther from the exit chamber are not less than those beneath chambers nearer to it, the experiment will have a quali- tative significance. Expt. 53. 11. xii. ’14. Five chambers, A, B, C, D , E, fixed to a leaf of Eucharis Mastersi , and the remainder of the leaf-surface vaselined. The areas of leaf exposed under the chambers were in the proportion : A. B. C. D. E. 10 25 25 20 IO and the distances between the chambers are indicated below : A. 2*3 cm. B. 2-2 cm. C. 2-3 cm. D. 1-9 cms. E. 7i Porometer in Stomatal Investigation. E was connected to the aspirator with a pressure difference of 15 cm.1 Readings were taken with each of the chambers open in turn, but those from A were discarded, as indicated above. Only one set of readings is given here, which are typical of those throughout the experiment. Table II. Path of air-current. D to E C to E B to E Length within leaf. 1*9 cm. 4-2 „ 6-4 „ Time between bubbles. 12-9 sec. 15*6 „ i8'7 » A variety of methods having been used to demonstrate the resistance of the intercellular spaces, some consideration of the effect of this resistance upon porometer readings is necessary. In fixing a porometer chamber to a leaf a large number of stomata are of necessity blocked by the adhesive, and thus in a hypostomatous leaf the stomata by which the air enters are some distance from those by which it passes from the leaf into the chamber, and an extra resistance is therefore encountered. In an amphistomatous leaf the blocking of the stomata is of less account, as the air may pass from one surface to another, although, as has been shown (p. 68, Dracaena\ this direct path is not always traversed. For accurate work with hypostomatous leaves it is thus important to reduce the length of path within the leaf by blocking as few stomata as possible. In their original paper on the porometer (loc. cit.), Darwin and Pertz described the leaf-chamber used by them as having a flange at the mouth to facilitate its attachment to the leaf. Balls (7), in his work on the Stomato- graph, advocates a form of chamber which also involves blocking the stomata over a very considerable area. These forms may be suitable for amphistomatous leaves, but if used for others must introduce much inter- cellular space resistance, and tend to mask stomatal changes, if not actually to introduce errors. The form of chamber which has been generally found efficient is a piece of glass tubing of the required bore, tapering at one end to take the rubber connexion, and with the other end cut off square and ground flat, leaving to be attached to the leaf a surface equal in width to the thickness of the walls of the tubing, i. e. 1 to 1 mm. Darwin and Pertz, after experimenting with many adhesives, finally decided in favour of glue, whilst Balls has recommended paraffin. In the present work glue has been found quite satisfactory, but the consistency needs to be carefully adjusted, depending upon the plant used and the 1 In practice, readings were also taken using A, B, c, and d respectively as the exit chamber, and only after the experiment, when measurements of areas had been made, was it possible to determine which set of results was significant. 7 2 Knight. — On the Use of the temperature. Only occasionally was a leaf found to be injured as a result of fixing the chamber ; after recovery from the first shock, leaves generally remained quite healthy with the chamber attached. Behaviour of Stomata of Different Regions. There is another question with reference to the porometer method which has been mentioned by Darwin and Pertz and also by Balls, viz. the relative condition of stomata on different parts of a plant. Porometer readings indicate stomatal changes only in the immediate vicinity of the leaf-chamber, and Darwin and Pertz quoted a set of readings from a plant of Primus laurocerasus , which showed marked differences in the size of the stomatal pores in leaves of different age. Balls proposed to get over the Fig. 6. Behaviour of stomata of different portions of a leaf under similar conditions. Ficus elastica. difficulty by fixing several chambers to different parts of the plant under observation, and by connecting them all to one reading apparatus, a mean is automatically obtained. In this case, however, each chamber should include approximately the same number of stomata, otherwise changes in the stomata beneath the chamber including the greatest number are liable to take an undue share in the total result. Experiments have been carried out on the behaviour of stomata on different parts of a leaf, and on different leaves of the same plant. i. A series of similar chambers were fixed to different parts of a leaf, and periodical readings taken from each. To determine the quantitative relations of stomatal aperture in the different regions, the reading per unit area for each chamber was calculated from the area of leaf exposed and the average reading over the whole period. 73 Porometer in Stomatal Investigation. Expt. 54. 11. xii. ’14 and 12. xii. ’14. 1 Four chambers, A, B, c, D, were attached to one leaf of Ficus elastica at 1 1.0 a.m., 1 1. xii. ’14, A, c and D being on one side of the midrib at a distance from the apex of about one-quarter, one-half, and three-quarters, respectively of the length of the leaf. B was on the other side of the midrib, about one- third of the leaf-length from the apex. Readings were taken at intervals from 12.40 p.m. till 4.30 p.m., December 11, and from 10.15 a.m. till 4.0 p.m. on December 12. The curves in Fig. 6 are the results of the latter series, the first series, which is not given, showing the closure due to shock and the subsequent recovery, which has already been discussed and illustrated. In the curves here shown there is not absolute coincidence, probably partly due to the different areas included under the respective chambers, but there is a very striking parallelism in all four curves, a change of direction in one being accompanied by similar changes in the others. The frequent irregularities were probably due to the varying illumination. The areas and mean readings were : A. Area i*o sq. cm. Mean reading . . . 0*142 Reading per unit area 0*14 B. i*2 sq. cm. 0*171 0*14 C. o*8 sq. cm. 0*117 0*15 D. j*i sq. cm. 0*167 o*!5 These show good agreement, and indicate that if the numerical distribu- tion of stomata over the leaf-surface is regular, the stomatal pores in different regions of the leaf are open to the same extent under similar conditions. The distribution of stomata will be discussed later. Experiments upon Eucharis Mastersi and Eupatorium Raffilli gave similar results, and although exceptional readings are often found, it may be assumed that as a general rule in these plants the readings obtained with a porometer from one chamber on a leaf are indicative of the condition of stomata over the whole leaf. 2. The type of experiment just described was repeated, with chambers on different leaves instead of on one leaf, with comparable though not exactly similar results. Expt. 70. 1. ii. ’15. Three similar chambers were fixed to different leaves of Eucharis Mastersi at 1.0 p.m. on 31. i. T5. A, to an old leaf, brown at the edges. B, to a mature healthy leaf. c, to a young leaf. Fig. 7 shows the curves obtained. These do not show the same parallelism as those in Expt. 54 previously described, but there is the same 74 Knight . — On the Use of the tendency for changes in one curve to be accompanied by similar changes in the others. A. B. c. Area ... i-o sq. cm. i*o sq. cm. i*o sq. cm. Mean reading 0*128 0*317 0*193 It is clear that the stomata of the mature healthy leaf were more widely open than those of the very young or very old leaves, whilst from the curves it may be seen that the stomata of the mature leaf were capable of closing to an aperture as small as those of the other two. Similar results were obtained with Ficus elastica and Etipatorium Rafjilli. Fig. 7* Behaviour of stomata on different leaves under similar conditions. Eucharis Mastersi. The generalizations as to the relative size of stomatal pores on different parts of a plant are made on the assumption that the stomata are evenly distributed over all leaves. To test this, several countings were made of the number of stomata per unit area of epidermis on leaves of Eucharis Mastersi and Ficus elastica. From the results it appears that the region nearest the edges of the leaf, particularly near the apex, is more thickly covered with stomata than the more central portions, in some cases the ratio being as large as 1*4 : i*o ; but at quite a short distance from the edge the numbers are almost the same as those near the midrib. In the case of leaves of different ages there is a distinct gradation — the old leaves having most and young leaves fewest stomata per unit area, the largest ratio observed being 1*3 : i*o. Thus in the porometer experiments upon one leaf only, the chambers were never near enough to the margin to be affected by the irregular 75 Porometer in Stomatal Investigation. distribution of stomata, and in the experiments upon leaves of different ages, when the distribution is allowed for, there is still the same relationship between the mean readings. Therefore, whilst readings from one chamber are adequate indication of stomatal behaviour in a single leaf, yet when the whole plant is considered it is advisable to fix chambers to more than one leaf, as suggested by Balls. Summary of Results. The conditions of the porometer experiment involve the possibility of a number of errors which by means of suitable precautions can to a large extent be eliminated. i. Temporary deformation of the leaf is liable to occur owing to the pressure difference employed to draw air through. Such deformation may cause undesirable changes in the stomatal pores. Any such effect is reduced to a minimum by using small pressure differences, and if a con- stant pressure is maintained the effect will be constant. 3. Some stomata show a tendency to close when air is drawn con- tinuously through them. To avoid this, the air-current should be stopped when readings are not being taken, by placing the leaf-chamber in direct connexion with the outer air. 3. The stomata of some leaves are sensitive to shock, the handling involved in fixing chambers to the leaf causing the stomata to close almost completely, but recovery is fairly rapid, and two hours has been found sufficient. The mere tapping or shaking of some leaves may induce a closure of the stomata. It is therefore advisable that after the leaf-chamber is fixed, readings be not taken for two hours, and that the leaf be disturbed as little as possible. 4. The resistance offered by the intercellular spaces to the passage of air through a leaf is considerable, and may have a marked effect upon the _ porometer readings. In leaves other than amphistomatous ones, it is there- fore advisable to reduce as far as possible the length of tissue to be traversed by the air-stream, by using chambers of suitable construction. 5. In the plants investigated, stomata on different parts of the same leaf behave similarly under approximately similar conditions, and are open to about the same extent at the same time. Thus readings from one chamber on a leaf are sufficient indication of the stomatal behaviour of that leaf. 6. Stomata on different leaves in general behave similarly, but the agreement is not so close as between stomata on the same leaf. Stomata of a mature healthy leaf may open more widely than those of either a very young or very old leaf. Thus, when using a plant with several leaves, in order to obtain a comprehensive measure of the behaviour of its stomata, chambers 76 Knight. — Use of the Porometer in Stomatal Investigation . should be attached to more than one leaf, but for reasons already stated the chambers should be about the same size. It is with great pleasure that I record my indebtedness to Professor V. H. Blackman, at whose suggestion this work was undertaken, and under whose guidance it has been carried out. Department of Plant Physiology and Pathology, Imperial College of Science and Technology. References cited. 1. Darwin, F. : Observations on Stomata. Phil. Trans., 190, B., 1898, p. 531. 2. Stahl: Einige Versuche iiber Transpiration und Assimilation. Bot. Zeitung, 1894, p. I17* 3. Lloyd, F. E. : Leaf-water and Stomatal Movement in Gossypiiun , and a Method of direct Visual Observation of Stomata in situ . Bull. Torr. Bot. Club, 40, 1913, p. 1. 4. Lloyd, F. E. : The Physiology of Stomata. Carnegie Institute Publications, No. 89, 1908. 5. Darwin, F., and Pertz, D. F. M. : New Method of estimating the Aperture of Stomata. Proc. Roy. Soc., 84, B., 1911, p. 149. 6. Knight, R. C. : A Convenient Modification of the Porometer. New Phytologist, vol. 14, 1915, p. 212. 7. Balls, W. L. : The Stomatograph. Proc. Roy. Soc., 85, B., 1912, p. 40. The Effect of the Concentration of the Nutrient Solution on the Growth of Barley and Wheat in Water Cultures. BY WINIFRED E. BRENCHLEY, D.Sc., Lawes Agricultural Trust , Rothamsted. With Plate II and four Diagrams in the Text. F'OR some years past much discussion has taken place as to whether the concentration of the nutrient solution has any appreciable effect upon plant growth, and at the present time the controversy is far from settled. Brezeale 1 carried out numerous water-culture experiments with wheat, using the transpiration as the criterion of growth. He states that ‘ it is evident that there is an optimum physical concentration of the nutritive solution at which water cultures of wheat thrive best, aside from variation in the amounts present of the different nutrient materials ’. Cameron 2 inter- prets these results otherwise, and claims that Brezeale has shown that, ‘ in water-culture experiments with wheat, if a given ratio of mineral nutrients be maintained, relatively small effect is produced on the growing plants by varying the concentration over a wide range, in one case 75 parts per million to 750 parts per million, and this effect seems to be largely independent of the nature of the particular mixture of solutes ’. Hall and Underwood,3 however, obtained indications that with barley the concentration of the nutrient solution in water culture has a definite effect upon growth, the total dry weight of the plants decreasing with the strength of the solution. Recently Stiles 4 has made further inquiry into the matter and states that ‘ the variation over a fairly wide range of the con- centration of the nutrient solution of rye and barley growing in water cultures produces relatively little effect on the amounts of dry matter produced. Below a certain concentration there appears to be a definite falling off in the rate of growth.’ When the figures given by Stiles for the dry weights of barley in 1 Brezeale, J. F. : Effect of the Concentration of the Nutrient Solution upon Wheat Cultures. Science, xxii, pp. 146-9 (1905). 2 Cameron, F. K. : The Soil Solution, pp. 40-1. 3 Hall, A. D., Brenchley, W. E., and Underwood, L. E. : The Soil Solution and the Mineral Constituents of ^he Soil. Phil. Trans. Roy. Soc. 204, B. 307 (1913). 4 Stiles, W. : On the Relation between the Concentration of the Nutrient Solution and the Rate of Growth of Plants in Water Cultures. Ann. Bot., vol. xxix (T915). [Annals of Botany, Vol. XXX- No. CXVII. January, 191C.I y8 Brenchley —Effect of Concentration of the Nutrient Solution water cultures are compared with those for thousands of plants grown at Rothamsted during the last nine years, it is seen that they are remarkably low, so low as to suggest that some factor was in action at Leeds that was quite ignored or overlooked in the estimation of results. Plants are very sensitive to external influences other than those of food and water-supply, and the amount of light, variations of temperature, and the atmospheric conditions prevailing during the growing period all have definite action on the rate and quality of growth. Crowther and Ruston 1 have shown that the smoke pollution of the air at Leeds is so great that plant life is most seriously affected, considerable depression in growth being caused at the University. This factor must have operated upon the water cultures, and may, to some extent at least, have vitiated the results obtained. Whereas at Leeds in 1914 the mean dry weight of barley-plants grown from April 28 to June 6 was only 0-628 grm., at Rothamsted one series grown simultaneously from April 27 to June 9 averaged 2-516 grm. dry weight, another series averaging 2-252 grm., and this was in spite of the advantage gained by the Leeds plants in the frequent renewal of food solution, while the Rothamsted plants remained in the initial solution all through the experiment. It has been stated that £ plants growing in water cultures under exactly the same conditions are very variable ’,2 and this is used as an argument for discounting the value of water cultures as a method of experiment.3 As a matter of fact, the individual variation of plants within a single series is far less than with similar plants growing under natural conditions in the open field. It is only necessary to examine carefully a small area of barley in the field, plant by plant, and to compare with a number of water cultures growing at the same time, in order to be convinced of the truth of this fact. Dactylis glomerata is on the whole a bad subject for water-culture experiments, but even in this case the range of individual variation under such conditions is most obviously less than between plants growing on the experimental plots. Mean variation from series to series is fairly great, because the period of the year has a very great influence upon the rate of growth, and plants grown in January and February may possibly not reach one-quarter the development (as shown by dry weight) of similar plants grown in April and May for the same length of time. Experiments have shown that the differ- ence of even a week in putting plants in water cultures has a distinct effect upon the total dry matter that can be produced within a given time. Every experimental method has its disadvantages and its weaknesses, and while water-culture methods are far from perfect, and indeed make no claim to be so, yet they do afford those conditions that are the most under the control 1 Crowther and Ruston : Town Smoke and Plant Growth. Journ. Ag. Sci., vol. vi, Pt. iv, pp. 387-94. 2 Stiles, W. : loc. cit., p. 89. 3 Stiles and Jorgensen : Studies in Permeability, I. Ann. Bot., vol. xxix, p. 349 (1915). on the Growth of Barley and Wheat in Water Cultures, 79 of the operator, and for that reason, if for no other, they have a special value of their own. During the season of 1915 a number of water-culture experiments have been made to see if further light could be obtained as to the effect of vary- ing concentrations of nutrient solutions upon growth, barley being used as the test plant in the three main series, wheat being grown in one case only. Four strengths of nutrient salts were used, N, N/5, N/10, N/20, the N-solution being that in general use in the laboratory, containing— Potassium nitrate Magnesium sulphate Potassium di-hydrogen phosphate Sodium chloride Calcium sulphate . Ferric chloride Distilled water 1 grm. °*5 > > °'5 » °*5 j > °*5 » 0-04 „ to make up one litre. The range of concentration was thus approximately 3,000, 600, 300, 150 parts of food-salts per million, containing potassium, phosphate, and nitrogen as in the following table : Parts per Concentration of Solution million of N N/5 N/10 N/20 k2o 640 128 64 32 p2o5 204 41 20.5 10.25 N 138 28 H 7 All the usual precautions were taken ; the bottles were thoroughly washed, new corks were used in every case, the water was obtained from a silver still which was kept scrupulously clean and polished, and the food- salts were uniform all through the experiments and were weighed up separately for each unit of ten plants.1 The barley was a pure strain of c Plumage ’ obtained from Mr. Beaven, and the seeds were all graded between 0-05 and 0-06 grm. to reduce individuality as much as possible. The wheat was a pure line of ‘ Persian ’ wheat obtained through. the kindness of Dr. N. Vavilov of Moscow; these seeds had to be sown without grading, as the supply was very limited. In each experiment with barley 120 plants were grown in units of ten. (1) All concentrations (N, N/5, N/10, N/20), the solutions being changed regularly every four days. (2) All concentrations, the solutions being changed once, exactly half- way through the experiment. (3) All concentrations, the solutions being never changed. Great care was taken of the roots when the solutions were being changed. While the bottles were refilled, one by one, the plants were removed, and the roots laid in a saucer containing a little solution corre- 1 The salts used were Kahlbaum’s ‘ for analysis \ and the stock was specially reserved for this experiment in view of the impossibility of replacing them at the present time. 80 Brenchley. — Effect of Concentration of the Nutrient Solution sponding in strength to that from which the plant was taken, so that no check was caused, either by slight desiccation or by shock from the roots being laid in pure water or in an alien solution. Each test ran for seven weeks, and was repeated three times at intervals of three weeks, so that information was obtained for plants grown early and late in the season. The developmental history was carefully noted, and it was found that the differ- ence in growth of plants in different concentrations was not only shown by the ultimate dry weights, but was apparent to the eye through the whole course of the experiments, both with regard to the size of plants and type of growth, especially with the roots. Each plant was harvested separately, and the dry weights of roots and shoots recorded. First Series. Seeds sown, March 5. Plants put into solutions, March 15. Plants harvested, May 3. Solutions changed ‘ frequently \ at regular intervals of four days. Solutions changed ‘ once on April 8. Solutions frequently changed. Most of the plants started off with fairly normal root growth, but the N/20 1 began to vary within the first few days, remaining short, with short thin laterals, which gave the roots a square bunchy appearance compared to the usual long type. This ‘ bunchiness * persisted for several weeks, but eventually the laterals elongated more normally. An unusual feature of the root growth was seen in all concentra- tions at the end of about a month. In addition to the usual thin roots supplied with long thin laterals, there appeared a number of very thick long rootlets springing from the base of the plant, either entirely free from laterals or else furnished with a very few tiny ones. These rootlets were thickest and most numerous in the N-plants and persisted to the end, so that at harvest-time the roots were inclined to be thick and much less fibrous than usual. In the lower concentrations these roots were very prominent at the time of formation, but were overshadowed later on by the further development of fibrous rootlets, and at harvest the roots had regained a more normal type. The development of the shoots in the plants growing in the different concentrations was very similar for some long time, but gradually a falling off was noticed with the two lowest (N/10, N/20), and by harvest-time some indications of this appeared even with N/5 shoots. In the N-plants the shoots were of an exceptionally dark green colour to the very end, the lowest leaves remained green and healthy, and there was no sign of red coloration at the base of the stem. The N/5-plants showed similar 1 For convenience of reference, the plants in the different concentrations will be called N, N/g, N/10, N/20 plants. on the Growth of Barley and Wheat in Water Cultures . 81 jram 5 development, and were nearly as dark in colour when harvested, but some of the lowest leaves had begun to turn yellow, ‘ and a trace of red was visible in the stem a few days before cutting. In N/10-plants these phenomena were more marked. The withering of the lower leaves and coloration of the stem had set in at an earlier date and were more pronounced, also the general development was less good. With the lowest concentration (N/20) the shoots were very much smaller than in any of the others, and were of a yellowish green colour, while the lowest leaves had died off a fortnight earlier, at the same time as the red colour appeared in the stems. The general trend of these observations is reflected in the dry weights of the plants, which will be discussed later. Sohitions never changed . The difference in concentration affected root growth immediately, each strength of solution having a definite effect of its own. The N/5-roots fell behind the normal within a week, being short and rather bushy with laterals standing out from the rootlets at an angle. These laterals elongated later, and gave the roots a more typical appearance till they looked stronger than the N, but this appearance was falsi- fied by the dry weights. In the lower concentrations the roots were very poor at first, bunchy, with rather thick late- rals standing out on every side, giving the roots a ‘ stark * appearance, but later on de- velopment became more normal in type, though still weak. The abnormal development of thick unbranched rootlets seen with ‘ frequently changed ’ plants was not noticed in any instance where the solution remained un- changed throughout the course of the experiment. Shoot growth showed a regular depreciation as the concentration of the nutrient solution diminished. The weaker the solution, the earlier etiolation set in, and the sooner did the lower leaves begin to die off and the red colour appear at the base of the stem. Towards the close of the experiment the difference in the amounts of water lost by transpiration was very marked, hardly any being given off by the N/20-plants. Sohitions once changed . The single change of solution kept the N-plants growing better, so that at harvest-time the plants were more strongly developed and of better colour than in the ‘ never changed ’ set, G Curve i. Mean dry weights of ten barley-plants growing in nutrient solutions of different concentra- tions. Dotted lines show the limits of probable error. F, frequently changed ; O, once changed ; N, never changed. (March 15-May 3.) 82 Brenchley. — Effect of Concentration of the Nutrient Solution With all other concentrations the march of events was delayed, but not arrested. Considerable improvement occurred immediately after the change (the weaker the solution the more obvious the improvement), but the falling off in growth soon reasserted itself in each case, though the ‘ fillip ’ caused by the renewal of the food supply was well reflected in the dry weights (see Table I and Curve I). Series i. March 15-May 3. Solutions changed Solution. 1 Frequently. 1 Once. 1 Never. Shoot. Root. Total. Shoot. Root. Total. Shoot. Root. Total. N N/o N/io N/20 / 2-919 2-608 1-984 1-303 1-082 1-041 0-816 0-528 4-001 ± 0-142 3-649 + 0-074 2-800 + 0-049 1 -83 1 ±0-038 2-639 1-483 0-791 0-447 0-713 o-543 0-447 2-289 3-352 ±0-096 2-026 ± 0*07 2 1 -238 ± 0-048 0-736 ± o-ooi I-897 0-958 0-526 0-244 o-735 0-446 0-301 0-208 2-632 ±0.071 i*404±o-o20 0-827 ±0-015 0-452 ±0-011 Series 2. April 5-May 24. Solutions changed Solution. 1 Frequently. 1 Once. 1 Never. Shoot. Root. Total. Shoot. Root. Total. Shoot. Root. Total. N 4-442 i-33o 5'772 ± 0-193 3-746 1-324 5'07°±o-209 3-048 1 -1 59 4-207 ±0-113 N/5 3-365 I*I37 4-502 ±0-076 1-621 0-709 2-33o±o-ii7 1-136 0-631 1 *767 ±0-037 N/io 1-899 0-843 2-742 ± 0-097 0-841 0-470 1-311 ± 0-056 0-525 0-409 0-934 ± 0-040 N/20 1-283 0-683 i-966±o-042 0-363 0-318 o-58i ±0-025 0-253 0-258 q-511 ±0-015 Series 3. April 26-June 14. Solutions changed Solution. 1 Once. 1 Never. Shoot. Root. Total. Shoot. Root. Total. Shoot. Root. Total. N 5-137 1-320 6*457 ±0‘J44 3-218 1-400 4-618 ± 0-117 2-401 1-293 3*694±o-ioi N/5 4-574 1 '467 6-041 ±0-124 1-841 0-804 2-645 ±0-067 0-922 0-690 i-6i 2 ±0-080 N/io 2-286 0-806 3-o92±o-i75 ■0-428 0-220 0-648 ±0-032 0-285 0*233 0-518 + o-o 1 1 N/20 i-254 0-527 1-781 ± 0-074 0-179 0-113 0*292 ± 0‘OIO 0-085 o-o8i o-i66±o-oi5 Table I. Mean dry weights in grams of ten barley-plants grown in nutrient solutions of various strengths. on the Growth of Barley and Wheat in Water Cultures . 83 Second Series. Seeds sown, March 27. Plants put in solutions, April 5. Plants harvested, May 24 (Plate II, Figs. 1, 2, 3), Solutions changed 4 frequently at intervals of four days. Solutions changed 4 once ’, on April 30. The course of events was the same as in the first series, though growth was more rapid owing to the more favourable season for growth, and differences due to the varying concentrations were more marked than in the earlier experiment (Curve 2). The thick rootlets in 4 frequently changed ’ plants were less strongly developed. It seems probable that the frequent renewal of the nutrient salts caused the plants to put out the abnormal rootlets for some unexplained reason, particularly at the time of year when growth was fairly slow. When growth was more rapid, the root development remained more normal in type, though thick rootlets did appear to some extent. Later on in the year they were only produced by plants in r*~s high concentrations, the others bearing quite normal fibrous roots. It may very tentatively be sug- gested that the thickened rootlets provide a means of protection at certain periods of growth against the constant change of balance due to the frequent renewal of the food solution. As they are so badly provided with laterals, it may be that they are able to prevent the ingress of too great and sudden an influx of food material at the time of the change of solution, so that they act as a kind of control. When growth is more rapid, the plant can deal with extra food Curve 2. Mean dry weights of ten barley-plants ... . . growing in nutrient solutions of different concen- more readily, SO that the con- tvations. Dotted lines show limits of probable error. trolling function is of less impor- (April 5-May 24.) tance, and the thick roots are cor- respondingly less developed. This idea is also borne out by the fact that in the first series the N-plants produced very thick rootlets in quantity, rendering the root thick and much less fibrous than usual, thus indicating possibly that at that time of year the plant was never able fully to cope with such a constant renewal of food solution of high concentration, owing to the relative slowness of growth which entailed the utilization of a lesser quantity of plant food (see Table I). 84 Brenchley . — Effect of Concentration of the Nutrient Solution Third Series. Seeds sown, April 16. Plants put into solutions, April 26. Plants harvested, June 14. Solutions changed ‘ frequently every four days. Solutions changed ‘ once May 20. This series was started rather late in the season, so that in some ways the plants exhibited more variability than in the earlier experiments. With frequently changed solutions the life-histories were much as usual ; in the N-plants the roots were inclined to be rather short and bushy from first to last, and were well supplied with the typical thick rootlets ; for about a month the N/5-plants looked stronger than the N, but this did not con- tinue ; in lower concentrations a rapid falling off of growth was exhibited. The red coloration did not appear in the stems, but the lower leaves were dead in every case. When the solutions were rarely or never changed, the plants in the two lower concen- trations made little or no head- way for a long time. Root de- velopment was checked almost entirely for three or four weeks, after which very long thin roots were produced. The shoots were also long and very thin, with little or no tillering. The sharp falling off in growth below the N/5 concentration, even when the solutions were changed, is well shown in the graph (Curve 3 and Table I). Curve 3. Mean dry weights of ten barley-plants grown in nutrient solutions of different concentra- tions. Dotted lines show limits of probable error. (April 26-June 14.) Discussion of Results. An examination of the figures and curves of dry weight in all three series shows that, however the solutions are treated, there is a steady decrease in the dry weights of the plants as the strength of the nutrient solution gets less. This decrease in weight is very considerable and is always outside the limits of experimental error. The results run in the same direction in all the experiments, the differences being accentuated in the sets grown later in the year, when growth is more rapid. It is noticeable that the drop in dry weight from N to N/5 is far less when the solutions are changed frequently, on the Growth of Barley and Wheat in Water Cultures. 85 and in some cases (Series 1 and 3) the approximation is fairly close. This suggests the possibility that more frequent renewal of the solutions, main- taining more evenly the balance of the nutrient salts, would be followed by as much growth in the N/5 as in the N concentration, although the tendency is for growth to fall behind in the lower strength with small provocation. In other words, if it were possible to arrange an experiment in which the balance of the nutrient solution was kept constant by the automatic replacement of the food-salts absorbed, it is conceivable that plants in these two concentrations might produce equal quantities of dry matter. But there are indications that toxic effects would set in under these circumstances in the N solution, as some of the constituents might be present in so great a quantity as actually to put a brake on plant growth. In the N/5 solution, on the other hand, the probability of such action is considerably less, and the plants would continue to make full use of the food-salts and would approximate in growth to those in the N solution. If this supposition be correct it is not true to say that the plant is indifferent to the variation in the strength of these two solutions, but that it responds to increased strength by increased growth. With the highest concentration, N, however, another factor, that of toxic action, comes into play, counterbalancing the increased growth and reducing it to the level of that attained with the lower (N/5) concentration. Further experiments are being made to obtain more definite information on this point, and also to find out whether there is an optimum concentration for growth or whether the plant will grow equally well within a certain range of the higher concentrations. The main point at issue at present is not that of equal growth in varying concentrations, but that of the great dilution at which it is claimed that such equal growth can be made. With concentrations below N/5 a very different result is obtained. The more frequently the solutions are changed, i. e. the closer the con- centrations approach to a state of constant balance, the more marked is the drop in the dry weights as the strength of the solution decreases. This implies that with the lower strengths the plant is living in a condition of semi-starvation. When the solutions are changed frequently, the improve- ment of growth is the more marked the higher the concentration (up to a certain limit, N/5 in this experiment), owing to the sum total of food supplied approximating more closely to the needs of the plant for optimum growth. It is difficult to imagine that, even if a constant balance of food- salts could be maintained, the plants in the solutions below N/5 would in any way approach those in the higher strengths, as, if this were indeed the truth, some indications of it would have been obtained by some incipient approximation of dry weights, such as occurred with plants in N and N/5 solutions, instead of a marked and decreasing divergence of these weights as the concentrations rose towards N/5, when solutions were fre- quently renewed, As a matter of fact, the actual dry weights of the N/10 86 Brenchley. — Effect of Concentration of the Nutrient Solution and N/20 plants in frequently renewed solutions were practically the same, within the limits of error, in all three experiments, showing that the same amount of growth had taken place in each case, whereas normally plants grown later in the season form much more dry matter, owing to the increase in the rate of growth, the N and N/5 plants in the three series being about 50 per cent, heavier than those in the first series grown seven weeks earlier. This indicates that with the lower strengths the amount of growth was strictly limited by the quantity of food supplied, and that it was impossible for the plants to reach full development with such a re- stricted amount. In the single solution with wheat, the solution was frequently changed, and the plants were grown on for eight weeks. In this case again, a steady fall in dry weight occurred with decrease of concentration, but owing to lack of seeds it was not possible to carry the comparison so far as with barley. The drop is less marked with wheat, but this may be because it grows more slowly than barley, so that differences are less accentuated within the same limits of time, though they are none the less definitely marked (Curve 4 and Plate II, Fig. 4). Before proceeding to discuss the significance of these results further, it may be useful to summarize the effect of varying concentration of food- salts obtained by different workers. Worker. Plant. Brezeale .• Wheat Stiles Barley Hall and Underwood . . . Barley Brenchley Barley Brenchley Wheat Parts per million of Food-salts. 75°- 75 growth similar. 1,800 360 180 90 v " shows growth similar. slight decrease. 3,000 600 300 150 steady decrease in growth. 3,000 600 300 150 growth marked possibly decrease much same in growth (solutions changed). 3,000 600 300 150 steady decrease in growth (solutions not changed). 3,000 600 300 150 steady decrease in growth (solutions changed). Curve 4. Mean dry weights of six wheat-plants grown in nutrient solutions of different concentra- tions. Dotted lines show limits of probable error. Solutions frequently renewed. (March 16-May n.) on the Growth of Barley and Wheat in Water Cultures . 87 The above table shows how great are the discrepancies between the results of several workers dealing with the same species by the same method of water-culture experiments. In the explicit statement by Cameron quoted earlier, it is claimed that the effect of the varying concentrations is largely independent of the particular mixture of solutes. If this statement be true, it matters little what nutrient solution is used, provided plants will make good growth in it, and the argument cannot be advanced that the difference in composition of the solution explains the discrepancies between the results, concentration being the only point at issue, provided that balance or a given ratio of nutrient salts be maintained as far as possible. Stiles 1 probably came very close to the truth in saying that ‘ possibly in the American experiments something other than concentration of salts was acting as a limiting factor in all cases5, but apparently he failed to see that the same remark may have had a very pertinent bearing upon his own results, owing possibly to the smoke factor in the Leeds district. Brezeale found with wheat that about 300 parts per million of food-salts gave maximum growth, and that growth fell off as the concentration increased to 750 parts per million or decreased to 75 parts per million, whereas at Rothamsted wheat shows a steady depreciation of growth as the concentra- tion decreases from 600 to 1 50, well within the former range. Stiles maintains that barley grows equally well within the range of 1,800 to 180 pts. per. mil., and only shows a slight depression with as little as 90 pts. per. mil., whereas at Rothamsted a great and marked decrease in growth occurs from 600 to 150 pts. per. mil. It seems more than probable that some depressing factor must have been at work in the Leeds experiments, which tended to equalize the growth of the plants by hindering them in some way which prevented normal development and reduced growth to a dead level. Such a factor might be provided by the presence of minute quantities of a toxic body in the distilled water or in the salts used for making up nutrient solu- tions. Experience has shown that the presence of the merest traces of copper salts in differential experiments will vitiate growth to such an extent as to make comparison useless, and unless the water is prepared with the utmost care such toxic substances find only too easy an entrance. Unfavourable atmospheric conditions, unsuitable temperature, lack of cleanliness in working, growth of algae in culture bottles, and the admit- tance of light to the roots are a few of the factors which may adversely affect growth, and which have to be taken into consideration in estimating results if they come into play. It is impossible to generalize from water cultures to sand or soil cultures, or from one species to another, but so far as the growth of barley and wheat in water cultures is concerned, this last experiment at Rothamsted upholds the earlier contention of Hall and Underwood that the concentration of the nutrient solution influences very 1 Stiles, W. : Review, Journ. Ecol., vol. ii, p. 54 (1914). 88 Brenchtey . — Effect of Concentration of the Nutrient Solution greatly the rate of growth of plants. Not only is the rate of growth affected, but the amount of growth is strictly limited by the quantity of available food when the nutrient solutions are dilute. Little work has yet been done with higher concentrations, but it is possible (see page 85) that toxic action due to over-nutrition from too great a supply of food-salts comes in to counter- balance or replace the increase of growth caused by increase of nutriment which occurs with lower strengths. Although the experiments fail to corroborate the idea that concentra- tion is unimportant within very wide limits, still they fully support other observations made by Stiles. In every case there is a drop in the dry weight of plants grown in any concentration according to the frequency with which the solutions are changed, the £ frequently changed ’ plants being heavier than the ‘ once changed ’, and the ‘ once changed ’ than the ‘ never changed \ With the normal strength it is probable that there is a sufficient supply of food material even when no renewal of solution takes place. In one case, barley was grown in such a normal solution for over eight weeks, and analyses made at the end of the experiment showed that 25 per cent, of the initial nitrogen still remained in the solution, and as the nitrogen compounds are absorbed in greater quantities than other salts it is evident that an ample sufficiency remained, if the quantity of the salts were the only factor concerned. Thus it is probable that with this concentration the question of starvation does not arise, and that the steady decrease in weight is really associated with the change of balance of the nutrient salts, the plants being the better the closer the initial balance is maintained. With the lower concentrations, the drop in weight from ‘ frequently changed ’ plants to the others was much heavier. Since with the normal solution the decrease in weight due to the balance of the food-salts was so much less marked, it seems permissible to assume that the very heavy drop with the lower concentrations is due largely to quite another cause, that of varying degrees of starvation through lack of sufficient nutriment. When the solutions are changed as often as once in four days, twelve times altogether, the plant has access during its lifetime to a far greater' store of food material than when solutions are seldom or never changed. Consequently such plants suffer less from the starving effects due to the low concentrations of the food-salts in solution, but still the response corresponds strictly to the amount of food available at any one time. There- fore it seems evident that with the normal solution the change in the balance of food-salts has a hindering action upon the growth of barley, and that this hindrance is coupled with varying degrees of starvation as the con- centration decreases, being specially accentuated in those cases in which the solution is never changed. It has frequently been noted that the variation in the strength of the food solution not only affects the total dry weight of the plant, but on the Growth of Barley and Wheat in Water Cultures. 89 also has a very marked action upon the relative rate of growth of roots and shoots, and this is well shown in the experiments under discussion. All the way through, the shoot responds more sharply than the root to the changes in food supply ; consequently, as the solution decreases in strength the ratio between the dry weights of shoots and roots also decreases ; in other words, the weights of the roots and shoots tend to approximate more closely as the supply of nutriment gets smaller, until in some cases with very dilute solu- tions the root is as heavy as, or even heavier than, the shoot (Table II). The change in ratio takes place always, whether the solutions are changed or not during growth, but it is most marked in those cases in which great starvation has ensued owing to the low concentrations not being renewed. It seems as though the plant makes every endeavour to supply itself with adequate nutriment, and as if, when the food supply is low, it strives to make as much root growth as possible so as to offer the greatest absorbing surface for whatever nutriment may be available. Shoot / Root Ratio. Series I. March 15-May 3. N N/5 N/10 N/20 Solutions changed frequently . . 2-691 2-506 2-431 2-468 ,, ,, once . 3-702 2-731 1-770 1-548 „ „ never . . • 2-580 2-149 1-750 1-174 Series 2. April 5- -May 24. N N/5 N/10 N/20 Solutions changed frequently . . 33-4T 2-956 2.252 1.876 „ „ once . . . 2-830 2-288 I*79I 1-144 ,, „ never . . • 2-630 i-8oo 1-285 0-979 Series 3* April 26-June 14. N N/5 N/10 N/20 Solutions changed frequently . • 3-892 3*ii7 2*835 2-379 ,, ,, once . . 2-297 2-289 1-946 1-582 „ ,, never . . . 1-857 1*336 1-226 1-039 Table II. Showing the ratio between the shoots and roots of the barley-plants whose mean dry weights are given in Table I. Summary. When plants, such as barley and wheat, are grown in water cultures under favourable conditions, the concentration of the nutrient solution, up to a comparatively high strength, has a great effect upon the rate and amount of growth, even when the balance of the solution approximates to a constant level. Starvation effects, due to insufficient nutriment, are evident in much stronger concentrations than has been admitted by some 90 Brenchley . — Concentration of Water Cultures. other observers. The action of different high concentrations of constant balance has not yet been determined, and it is uncertain whether there is an optimum strength, above and below which growth falls off, or whether there is a range of concentrations between which the plants will make equally good growth. It seems evident, however, that if water cultures with wheat and barley are carried out under advantageous growth- conditions, complete and maximum growth cannot be obtained in a solution containing the amount of potash and phosphoric acid (K2q 28 p.p.m., P205 7 p.p.m.) stated by Cameron to exist in the soil solution. Rothamsted, October , 1915. EXPLANATION OF PLATE II. Illustrating Dr. Brenchley’s paper on Concentration of Water Cultures. Fig. 1. Photograph showing the growth of barley-plants in water cultures when solutions were frequently renewed. Concentration of solutions, N, N/5, N/10, N/20. Series II. Fig. 2. As above, but solutions changed once only. Series II. Fig. 3. As above, but solutions never changed. Series II. Fig. 4. Photograph showing the growth of wheat-plants in water cultures when solutions were frequently renewed. Concentration of solutions, N, N/5, N/10, N/20. The Origin of the Endodermis in the Stem of Hippuris. BY KATE BARRATT, B.Sc., Demonstrator in Botany , Imperial College of Science and Technology , London. With six Figures in the Text. THE question of the three germinal layers exhibited by the stem and root of Dicotyledons is a very old one, and has been the subject of investigation by many botanists. The idea that these three developing layers were the initials of definite tracts of tissue of the adult plant was first put forward by Hanstein. He supposed that of these three layers or histogens, the dermatogen gave rise to the epidermis, the periblem to the cortex, and the plerome to the central cylinder. The fact that in the adult organs the innermost layer of the parenchyma- tous tissue surrounding the central cylinder is frequently differentiated by special characters has given to this layer, the endodermis, a special significance. The reason that it has acquired so much importance is that it has generally been regarded as the innermost layer of the cortex and developed from the innermost layer of the periblem, forming thus the boundary of the stele. The importance of the endodermis as a morphological unit thus obviously depends on the uniformity of its mode of origin. This was realized by Schoute (1), who in 1902 published a general review of the stelar theory, and incidentally made a critical examination of a number of stems and roots. He showed that the majority of species examined exhibited a well-marked endodermis. The origin of this layer from the apical tissue of the stem was investigated in the case of only six species, in some of which no distinction between periblem and plerome could be made out. Among the plants examined was Hippuris vulgaris , in the stem of which the apical structure is so clear that it has become a classical type for use in the laboratory. Schoute, using series of transverse and longitudinal sections, traced the history of the development of the tissues from the apex to the mature part of the plant, and was led to the conclusion that here not only the endodermis, but several other layers of the cortex were derived from the plerome. [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] 92 Barratt. — The Origin of the This observation was directly opposed to the generally accepted idea of the origin of the tissues of Hippuris put forward by Sanio (2), according to which the endodermis was derived from the innermost periblem layer. Since Schoute’s observation has so far received no independent con- firmation, a re-examination of the critical case of Hippuris seemed desirable, and material was therefore collected in the spring of 1914 and 1915. Schoute’s methods of investigation were closely followed. Stem apices were selected from shoots of various sizes, care being expended to choose only straight ones. These were embedded and cut in series of transverse sections up to within a distance of about 100 n from the apex. The block was then rotated at right angles, and the remainder of the apex cut in a longitudinal series. The advantage of this method is apparent. A median longitudinal section through the apex shows very clearly the distinction between the periblem and plerome, which is made evident by the small number and extreme regularity of the layers of periblem cells which form a series of caps covering the central column of plerome, in which the cells are less regularly arranged. In transverse 'sections the distinction is not so clear in this region, and the use therefore of longitudinal sections through the extreme apex of a stem of which the lower part is cut transversely, enables one to determine the exact number of periblem layers concerned in the origin of the mature tissues. It is thus possible to trace with certainty the critical layer, the innermost of the periblem, and to identify the structure derived from it. In order to be quite certain about the position of the delicate cell-walls of the developing tissues, it is necessary to clear away the cell-contents, and this can only be done satisfactorily by the use of eau de javelle on the sections. The clearing agents failed to penetrate the fixed material in bulk, even after the lapse of several days ; on the other hand, the fresh materia], though sufficiently cleared, was unsatisfactory because the cell-walls were also affected, losing their original firm contour, and in some cases actually breaking down. In order to be able to use eau de javelle on serial sections, it was necessary to employ a fixative other than albumen ; water alone, collodion, and collodion with clove oil were tried, and a mixture of the latter in the proportion of 1 : 3 was found most satisfactory. With this fixa- tive it is necessary, first, to float out the sections on water and then to transfer them to a slide smeared with the mixture. It is essential to have absolutely clean slides and to use fixative recently made up. Of the many different stains employed, polychrome methylene blue gave the best results. It is, however, difficult to keep sufficient stain in the sections owing to its tendency to wash out in alcohol, unless the sections are mordanted after staining by placing them for a few minutes in a 10 per cent, solution of ammonium molybdate. 93 Endo dermis in the Stem of Hippuris . The advantages of polychrome methylene blue are twofold ; it is a very quick and easily controlled stain, and it differentiates well. The walls of the cells lining the air-passages stain differently from the others, and are thereby easily distinguished even in longitudinal sections. The structure of the apex of the stem as seen in median longitudinal section (Fig. i) exhibits an arrangement of cell-layers which follows dia- grammatically the scheme set forth for stem apices in general. The outside is clothed by a single layer of regularly arranged cells, which constitutes a typical dermatogen. Within this, in the specimen figured, are five regular layers of cells forming the periblem. The cells of these differ little in form from those of the dermatogen. Within these, the cells of the plerome are more irregularly arranged, owing to the fact that they divide by both periclinal and anticlinal walls. Fig. i. Longitudinal section of tip of stem of Hippuris vulgaris, d, dermatogen ; p, periblem ; pP plerome. There is considerable variation in the size of the stem apices, and the number of periblem caps in the material examined varies from three to six. The general arrangement in other respects is quite consistent. On turning to the transverse series and examining the first section (Fig. d) it is a simple matter to interpret the layers of cells in terms of the longitudinal section. In some cases there is more difference between the size of cells of the periblem and those of the plerome than appears in the specimen figured. This difference becomes more pronounced as the sections are traced further down the stem. The cells of the plerome at an early stage divide rapidly by walls in all directions, thus producing a large number of small cells. The method of division is fairly regular. One cell divides into two, and these again divide independently by walls nearly cit right angles, thus forming groups of four cells. These groups frequently 94 Barr ait, — The Origin of the continue segmenting-, but it is generally possible, for quite a long time, to trace the outlines of these groups which originate from separate initials. The cells of the periblem do not divide so early, but enlarge to keep pace with the increasing bulk of plerome, and later divide by anticlinal walls, thus preserving the original number of layers. The air-channels begin to form in the periblem as intercellular spaces at the angles between cells of different layers. As a result, one finds rings of small air-spaces appearing between successive layers of periblem cells. Fig. 2. Transverse section’of same tip at region where the longitudinal section ends. Such spaces are never found between the dermatogen and the first layer of the periblem ; moreover, the ring of canals between the first and second periblem layers develops a little later than the ones within. As can be seen by reference to Fig. 3, there are four rings of canals developing. At this stage the cells of the periblem begin to divide, and the subse- quent arrangement of the cortical cells and air-canals is dependent on the direction in which these periblem cells segment. The innermost layer becomes divided by periclinal walls which are laid down in such a way as to 95 Endodermis in the Stem of Hippuris. abut upon an air-space at either end. The relative position of the original cells is preserved, as very little growth of the daughter-cells takes place for some time subsequent to the division (Fig. 3). In the outer layers the original cells may undergo subdivision into two or three, rarely more. The walls may be laid down in any direction, which is however determined by the position of neighbouring air-canals, since the ends of these walls invariably abut upon these passages. Thus at this stage every intercellular space is a centre from which a varying number of walls radiate (Figs. 3, 4). With the increase in size of the stem which now follows, the cells of the develop- Fig. 3. Transverse section showing the innermost layer of periblem after the division into inner and outer cells. /5, innermost periblem layer. ing cortex undergo further subdivision, whilst at the same time the inter- cellular spaces enlarge enormously, owing to the rapid growth of these cells which surround, and eventually form single chains of cells separating them. The development of the system of intercellular spaces just described serves to differentiate, very markedly in the young stem, the nodes and internodes, since the air-canals are only developed in the latter. In the young nodes, on the other hand, cell-division takes place to a greater extent than in the internodes, first in connexion with the formation of the leaf-rudiments, secondly with the laying down of the pro- 96 Barratt. — The Origin of the Fig. 5. Transverse section in the region of a node, about the same level as Fig. 4. 97 Endodermis in the Stem of Hippuris . cambial strands of the leaf-traces, and thirdly to keep pace with the rapid increase in size of the young cortex consequent upon the formation of the air-canals in the internodes. For this reason the identity of the original periblem caps and their constituent cells is more difficult to trace in the node, though the contour of the original cells can as a rule be recognized by the thickness of their walls (see Fig. 5). It has already been indicated that the number of circles of air-canals is related to the number of layers of periblem, being one fewer ; thus, in the specimen figured with five periblem caps there are four rings of canals (Fig. 3). In the older parts of the stem, however, additional rings can be observed, and it is in studying the modes of formation of these additional intercellular spaces that the part played by the plerome in contributing to the inner region of the cortex, including the endodermis, becomes clear. It is necessary therefore to return and to consider in some detail the fate of the most internal layer of the periblem. As described above, the first dividing-wall runs tangentially across the cell from one air-canal to another (Fig. 3). Further segmentation takes place in the inner and larger of the two cells so formed, and in this the next wall is placed in an obliquely radial direction, with its outer end usually based upon an air- canal. Where the inner end of this wall intersects the wall of the outer- 1 most plerome cell a new intercellular space arises, and thus is initiated a fifth ring of air-canals (Fig. 4, ac.5). It is obvious that cells bounding these canals on the inner side must be derived from the plerome, and this can be readily confirmed. Moreover, it not infrequently happens that in the development of the fourth ring of air-canals the separating cell-walls of the two cells bounding it on its inner side may split apart, thus permitting the enlargement of a plerome cell which thus becomes the inner boundary of the canal. The whole plerome is at this stage actively enlarging, but the outermost cells divide for the most part by radial walls, and also increase in size in a tangen- tial direction. They thus form a fairly well defined layer distinct from the rest of the plerome. Although no regular sequence can be traced in their divisions, they sooner or later divide by tangential walls into inner and outer cells (Fig. 6). This stage can be traced for some distance down the stem. Both layers may undergo subdivision by radial walls. The cells of the outer layer eventually enlarge and become rounded off as intercellular spaces develop between them. Eventually the cells of the inner layer undergo a tangential division, and the relative position of the two layers of cells so formed remains unaltered in the mature stem. The innermost layer is the endodermis, and later develops the cuticularized folded band on the radial walls. Considerable variation may be seen from the process just described. Some of the divisions may be omitted or, in other cases, may be increased in number, but on the whole the general sequence of development in H 98 Barratt . — The Origin of the the outer part of the plerome is that described above, and thus normally three layers of the inner cortex, including the endodermis, take their origin from the outer plerome. This conclusion agrees with that of Schoute, and is thus directly opposed to the earlier explanation given by Sanio, who considered that the E- ND Fig. 6. Transverse section through older stem, showing the formation of the endodermis and final appearance of air-canals, end., endodermis. three layers in question originated by the subdivision of the innermost layer of the periblem. Although the observations do not coincide in exact detail with those of Schoute, this is probably to be accounted for by the fact that, owing to this investigation having been confined to this one species, it was possible to examine a large number of apices and thus to obviate the disadvantages 99 Endodermis in the Stem of Hippuris . connected with a description of one particular specimen. There is a wide variation in dimensions of the stem at the apex which is associated with dif- ferences in the number of periblem caps and bulk of the plerome. The number of the former may even vary in one and the same apex, as, for instance, in one specimen examined there were five caps on the one side of the stem and four on the other. The results of this investigation therefore confirm the observations of Schoute, and emphasize his conclusion that the endodermis cannot be regarded as a layer of definite morphological value. That is to say, its ontogenetic history is shown to be variable as elucidated by a study of cell-lineage. The only criterion of morphological identity in such instances would have to reside in the structure of the completely differentiated tissues, irrespective of their cellular history. This was, speak- ing generally, the position adopted by Van Tieghem, and whilst it illustrates the plasticity of the plant cells in differentiating into this or that form of tissue, it emphasizes the abstract character of the so-called morphology of the tissues themselves. References. 1. Schoute, J. C. : Die Stelar-Theorie. P. Noordhoff, Groningen, 1902. 2. Sanio, C. : Ueber endogene Gefassbiindelbildung. Bot. Zeit., xxii Jahrgang, 1864. The Development of the Sorus and Sporangium and the Prothallus of Peranema cyatheoides, D. Don. BY R. C. DAVIE, M.A., B.Sc., Lecturer in Botany in the University of Edinburgh. With Plate III and two Figures in the Text. IN a paper in the Annals of Botany, vol. xxvi, 1912, it was suggested that Peranema cyatheoides , D. Don, occupied a position intermediate between the Cyatheaceae and the Aspidieae group of Polypodiaceae. Various features of the mature plant, including those of the vascular system and the sporangium, suggested a relationship with the species of N ephrodium and especially with N. Filix-mas , Rich. Developmental stages of the sorus and sporangium have since been studied in material grown in the Glasgow Botanic Garden and kindly forwarded to me by Professor F. O. Bower, F.R.S. Fresh spores were sent from India through the kindness of the Director of the Calcutta Botanic Garden, and young plants were reared from them by Mr. L. B. Stewart, Plant Propagator in the Royal Botanic Garden, Edinburgh. In the Edinburgh Garden there are now half a dozen strong and healthy plants of Peranema . I tender my thanks to the gentlemen to whose courtesy this result is due. Development of the Sorus. P'rom the earliest stages available it appears that the sorus is very soon after its appearance covered almost completely by the indusium (PL III, Fig. 1 ). This indusium is composed, in the main, of a single series of cells though near the highest point of its curve and near its tip there are two cell- layers. It forms a scale attached at one side along a semicircular line to the under surface of the leaf and bent over the top of the receptacle. At one point the receptacle is, for a width of three cells, uncovered by the indusium. Below this point are one or two cells continuous with the series forming the epidermis of the leaf and with walls thickened like those of the cells of the indusium (Fig. 1). This suggests the presence of a small second flap. Sec- tions through older sori confirm the suggestion, since in these a distinct small second flap or edge of a cup is present (Figs. 2 and 3). In Fig. 3 the further curving over of the main indusial flap is shown. This is still more accen- tuated in Fig. 4, where the commencement of an extension of the receptacle at right angles to the leaf-surface can be distinguished. The second flap [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] 102 Davie . — The Development of the Sorus and Sporangium still is present as the sorus lengthens (Fig. 5). When the receptacle grows out further as the narrow stalk of the sorus, the curving over of the indusial flap is very marked. It is then curved back on one side, until the portion by which it is attached to the stalk is almost parallel to the surface of the leaf (Text- fig. 1). The earliest indica- tions of this recurving at the side are seen in Fig. 4. Of the second flap there remains in the older sorus only a mere knob, seen on the left side of Text-fig. 1, at the top of the stalk. The extreme tip of the indusium is bent in (Text-fig. 1), and between the recurved tip and the stalk is a narrow slit, which appears on the surface of the sorus as an elongated pore. This slit marks the edge of the ‘ cup ’ forming the indusium. In a trans- verse or oblique section of the mature sorus the two edges of the ‘ cup ’ may be seen (cf. PI. Ill, Fig. 6, and Ann. of Bot., xxvi, PI. XXIX, Fig- !5)- These sections through the sorus show that the in- dusium arises as a cup round the receptacle, that one side of the cup is strongly de- veloped and curved over the top of the receptacle, while a small portion on the other side is suppressed in de- velopment. On this side at maturity the edges of the remainder of the indusium curve in to form a pore. The receptacle, at first only slightly raised above the surface of the leaf (Fig. i), Text-fig. i. Vertical section through a semi-mature sorus, showing the 1 stalk ’ fully developed. The main indusial flap now completely covers the wide receptacle and is incurved at its tip. The small knob opposite to this incurved tip represents the remains of the laggard portion of the indusium. The ‘kink’ shown in PL III, Fig. 4 has developed as the series of cells on the right of the sorus and now extends almost at right angles to the ‘ stalk ’. The youngest sporangia occur among the stalks of the oldest at the summit of the receptacle ; quite young sporangia appear at its margins. The sequence of sporangia is thus at first basipetal, and later mixed, x 65. and the Prothallns of Peranema cyatheoides , D. Don . 103 becomes pushed out on the side at first uncovered by the indusium (Fig. 2) and may remain exposed until sporangia appear (Fig. 2). In some sori the indusium covers the receptacle entirely, even before the sporangia are fully defined (Fig, 3). As the sorus grows older, an elongation of the receptacle goes on (Figs. 3 and 5), the central part of the resulting stalk being continuous with the receptacle, and the peripheral part continuous with the indusium. The side of the indusium in Fig. 4 up almost to the £ kink ’ represents the portion which later becomes one side of the soral stalk (the portion up to the sharp bend below the start of the indusial flap in Text-fig. 1). The number of cells from side to side across the base of the sorus in Fig. 4 is exactly the same as the number across the stalk of the mature sorus in Text-fig. 1, while the number of cells from a to b in Fig. 4 corresponds with the number from a to b in Text-fig. 1. (A counting of the number of cells in length and width of the stalk of the sorus figured in Fig. 13, Ann. of Bot., xxvi, PL XXIX, gives precisely the same results. As this sorus was cut from another plant and the figure was made long before the material which forms the subject of this present paper was obtained, it is interesting to observe how closely the sori adhere to one type of construction. Comparison of herbarium specimens collected at wide intervals of time shows how constantly the sori preserve the same size and form.) The elongated condition of the cells of the stalk in Text-fig. 1 suggests what a comparison of Fig. 4 and Text-fig. 1 makes obvious, that, to produce the mature condition, there has been simply a lengthening of the cells at the base of the receptacle, and of the cells of the tissues which in the young sorus are continuous with the indusium. Apparently cell-divisions go on in the superficial layers of the young sorus until sporangia are initiated on the receptacle. Then a lengthening of the individual cells of receptacle and superficial layers produces the stalk of the mature sorus. The receptacle in the youngest stages examined has in tangential section the outline of an hour-glass. In median section it is shown curved at its tip towards the edge of the leaf (Fig. 1). It widens at its apex as it grows older (Fig. 3) ; its lower portion becomes constricted, its apex broadly dome-shaped (Figs. 4 and 5). The broad dome-shape is found in the fully developed sorus (Text-fig. 1). In even the youngest leaves which bear sporangia, the vascular tissue in the veins is defined. The young sorus is always produced below a vein, and the vascular tissue of the vein runs into the base of the sorus. In the earliest stages no tracheidal tissue is present in the upper portion of the receptacle or in the stalk (Figs. 1, 2, and 5), but the cells of the central series, particularly in the lower part of the stalk, become very much elongated, and are narrow even in a fairly young sorus (Fig, 5). When the stalk reaches its full length, the cells of its central region are more or less in process of change into tracheides ; tracheides are at the same time defined in the receptacle proper. In the mature sorus the tracheidal system runs as a narrow strip up through the 104 Davie, — The Development of the Sorus and Sporangium stalk of the sorus and ends as a fan below the apex of the wide receptacle. Through the base of the stalk this tracheidal system is continuous with the series of tracheides in the vascular strand of the vein upon which the sorus is inserted (cf. Annals of Botany, xxvi, PL XXIX, Fig. 13). Sporangia appear upon the receptacle as soon as it begins to lengthen (Fig. 2). The first sporangium occurs in varying positions on the surface — sometimes near the junction of the large indusial flap with the receptacle (Fig. 3) ; sometimes near the edge of the reduced flap (Fig. 2) ; sometimes midway between these positions. The first sporangium is rapidly succeeded by others, which arise close to it. Two sporangia of approximately the same age sometimes occupy positions near the apex of the receptacle, while others very slightly younger appear nearer to its margins (Fig. 4). A comparison of the condition shown in Fig. 4 with that in Text-fig. 1, in which the oldest sporangia occupy the summit of the receptacle, while young sporangia appear at the margins, leads us to conclude that the sequence is in the main basipetal, but that the sporangia succeed one another rapidly. In Text-fig. 1 the youngest sporangia figured are at the apex of the receptacle, among the stalks of the oldest. This mixed condition of the mature sorus was figured in Ann. of Bot., xxvi, PI. XXIX, Figs. 13, 14, and 15. The suggestion made in the earlier paper (ibid., p. 254) that the sorus is of a mixed type upon a Gradate receptacle is confirmed by the details shown in Fig. 4 and Text-fig. 1. The sequence of sporangia shown in these figures proves that the succession of the earliest sporangia is essentially basipetal. Development of the Sporangium. The earliest stages of the development of the sporangium are figured in Text-fig. 2, a-i. The cell which becomes a sporangium is often wedge- shaped (a). The first wall may be transverse (a and b) or oblique (c and d). The oblique wall more frequently occurs, to judge by the condition of various later stages (g and h), f suggests a later condition of the type shown in b. Where the first wall is transverse the one which immediately succeeds it is oblique ( e and f), and this second oblique wall commonly meets the first at its junction with the lateral wall (e and f). Where the first wall is oblique, it may meet the lateral wall at its junction with one of the walls forming the wedge-shaped base of the cell (g and h) ; it may meet the lateral wall about half-way down its length (d), or it may meet one of the basal walls (c), A wall next cuts the oblique wall at right angles (g), and that which succeeds this is parallel to the first oblique wall (h). In z is shown the central cell of the capsule fully formed, while the cells of the stalk are definitely delimited. The later stages in the development of the sporangium and the mature sporangium itself have already been described (loc. cit., p. 254). The mature sporangium is long-stalked, and has an and the Prothallus of Peranema cyatheoides , D. Don. 105 oblique annulus, the cells of which are continuous past the stalk (loc. cit., PL XXIX, Figs. 16 and 19). We can now consider how these features of the sorus and sporangium affect the systematic position alreadyassigned (loc. cit., pp. 264-5) to Peranema. The character of the vascular system and the structure of the mature sorus and sporangium led to the conclusion that Peranema occupied a position intermediate between the Cyatheaceae and the Aspidieae group of Polypodiaceae. A basal indusium of cup- type is characteristic of the genus Cyathea and other advanced types of the Cyatheaceae (Bower, ’99, p. 52). In Peranema the cup is ‘ developed unequally on its two sides and contracted at its rim, which is turned inwards ’ (Davie, T2, p. 253). The early stages show that it is of the Cyathea type, with a part of one side of the cup lagging behind the rest of the indusium (cf. above, pp. 101, 102). The stalk is apparently a late growth, due to an elongation of certain cells of the receptacle and superficial layers. The sorus with its basal cup-indusium is certainly of Cyatheaceous type. The form of the receptacle — widely dome-shaped at maturity — recalls that found among the Gradatae (Bower, Land Flora, p. 635). The first sporangia arise commonly at the apex of the receptacle ; the subsequent sporangia show clear indications of follow- ing in basipetal sequence. But the mature sorus is a mixed one. This shows a step in advance of the Cyatheaceae. The sporangium develops in the manner of one of the Gradatae (loc. cit., p. 638). The wedge-shaped Text-fig. 2. a. A young sporangium, showing a wedge-shaped base. The first division-wall is transverse, x 450. b. Another form of young sporangium, with a transverse first division-wall, x 450. c. A young sporangium, with wedge-shaped base and oblique first division-wall, x 660. d. Two young sporangia, with oblique first division-walls meeting the lateral walls of the sporangia, x 660. e. A young sporangium, of the type shown in a, with the second division-wall oblique and meeting the first transverse division-wall at its junction with one of the lateral walls. x 450. f A young sporangium of the type shown in b , with an oblique division-wall succeeding the first transverse division- wall, x 450. g. A young sporangium, of the type shown in d, with an oblique second division-wall meeting the first division-wall, x 450. h. An older sporangium, of the type shown in g, with the covering cell of the capsule defined and an oblique third division-wall meeting the second wall and parallel to the first, x 450. i. A sporangium with the stalk and the central cell defined, x 450. io6 Davie. — -The Development of the Sorns and Sporangium base and the oblique first wall correspond to similar features in the sporangia of Thyrsopteris and Alsophila (loc. cit., pp. 590, 603). The early segmentations in the sporangia in Text-fig. 2, ) )) >) }> 33 3) 33 33 Treatment of the Subject. It will be convenient to deal with the subject-matter under two distinct heads. Part I. The Weeds of Arable Land\ that is to say, the communities found under root crops, cereals, and ‘ seeds \ By { seeds J is here understood leys which are sown down with a mixture of grass and clover seeds for one year only, or if sown for a longer period, till the end of the first year. The above communities are under the direct control of man, and owing to the exigencies of the rotation are short- lived and consequently do not afford materialf or the study of the progressive stages in the colonization of bare land. Part II. The Weeds of Grass-land. This includes the communities found on long duration leys and on permanent grass, and affords material for the study of the progressive changes which occur in the colonization of land under the constant influence of grazing animals. For the sake of comparison two cases will also be given of the colonization of bare land from which stock have been withheld. PART I. The Weeds of Arable Land. Under this heading it is proposed to make the following comparisons: I. The aggregate arable flora of the Cotswolds with that of Mid- Wales. II. The results from Mid- Wales and the Cotswolds with those obtained elsewhere. III. The arable flora of Mid-Wales above and below 800 feet. IV. The weed flora under the several crops in the rotation. In order to facilitate these comparisons the frequencies found for a large number of species in both localities and in various places in the rotation are set out in the subjoined table (Table I). Stapledon . — On the Plant Communities of Farm Land. 163 Table I. DISTRIBUTION OF THE CHIEF SPECIES IN MID-WALES AND THE COTSWOLDS ' V ■ I ' ! i • Chief Species. Cotswolds. 2. Roots. •So <3 0 3- s\- ^ 0 . « 0 ^ 00 S 1 4- § £ <0 5> Cereals. 0 <3 O Mid- Wales , ^ above 800'. 7- S' 1 $ 8. Seeds. 0 Is 0 £ °° 5 § Mid- Wales , above 800'. Ranunculus repens 2 I 0-6 1 3(6) 3-9 0 0-2 0-4 R . parvijlorus 0 r — s r — — — R. arvensis r — — S-0 — — — — — Papaver Rhoeas 0 — — 2 (5) — — — — — P. dubium — — — s-o r — — — — Fumaria officinalis s— 1 r - 3-6 (9) r — r — — F. pallidifiora — 0 (2) r — O-I — — — — Erophila verna r — — (4) r — — — — Brassica arvensis O-I s (2) 0 2-4 0 (5) 1 (6) — — s B. nigra 0 — — 1 — — — — — Capsella Bursa-pastor is 2 0 0 1 S — r — — Thlaspi arvense 0 — — 0 — — — — — Viola tricolor (agg.) O-I 1 1 1 2 0 3 1 0-1 (3) Silene Cucubalus r-o — — i-3 s r — — Lychnis alba 0 — — 1-2 s r — — — L. Githago r — — 0 r — — — — Stellaria media 2 (4) 3-5 3-6 (9) 2-5 0 D 2 r r o-(5) Arenaria serpyllifolia 0 s — 0 r r o-3 (6) V — Spergula aniensis — 1 1-9 — 1-4 4-6 — r — Hypericum pulchrum — — — — — s-o — — — Geranium molle 0 s — s 0 s-o 1 (4) 0 (3) S-2 G. dissectum s r — S-0 0 s 0 (3) 0 (2) S-2 Erodium cicutarium — — — — — — (3) — — Trifolium minus — — — — — — o-3 i-4 (7) Ornithopus perpusillus — — — — — s-o — — — Vicia C race a — — s — — s — — — Lotus corniculatus — — r — — s — — s-o Lathyrus pratensis — — r — — s — — r Potentilla ereda — — 0 — — (6) — — S Alchemilla arvensis r S-0 r S-2 0 r i-3 2 S-l(2) Scandix Peden- Veneris 1 — — O — — — Aethusa Cynapium r — — I r — — — — Daucus Car ota s r — — — — — 0 — Galium saxatile — — — — — 0-2 — — s G. Aparine 2 r I r — 0 r — Sherardia arvensis s 0 s — 0 — 3 1 s Scabiosa arvensis r — — 1-2 r — — — S. succisa — — — — s 0 (3) — — — Beilis pwennis r s 1 — 3 2-9 s 1 0-2 (3) Gnaphalium uliguiosum — — 3 — — — — — Achillea Ptarmica — — — — — 0 — — — Chrysanthemum segetum — 2 0-4(7) — 0-4 i-7 — — — Matricaria inodor a r 0 s — s r * — r — Senecio vulgaris 1 1 0 O 0 r r — — Cnicus lanceolatus — — — — — — 0 0 — C. arvense 3 1 s 3 2 0 — — — Centaurea nigra s 0 — s-3 — — — C. Cyanus — — °l -1 — — — — Cichorium Intybus f i — — r 0 s Lapsana communis — 0 s 0 0 s — s-o s M 2 164 Stapledon . — On the Plant Communities of Farm Land . Chief Species. 1. Jo § 0 2. Roots. *o . • 0 e 0 ^3 § Mid- Wales, above 8oo'. • 4- l>* (Mid- Wales). Cichorium Intybus * ) Cnicus lance olatus * (Mid- Wales and Cotswolds). (b) Those which under good ‘ takes ’ are normally found to have high frequencies — frequencies for many of the species higher than found elsewhere in the rotation. Ranunculus repens .* Viola tricolor (agg.). Arenaria serpyllifolia. Cerastium spp.* Geranium molle.* G. dis sec turn .* Erodium cicittarium . A Icliemilla arvensis . Trifolium minus* (c) On poor soils and under moderate ‘ takes ’ the following are frequent : Lotus uliginosus (Mid-Wales on wet soils). Chrysanthemum leucanthemum * (Mid- Wales and Cotswolds). Veronica agrestis (Mid-Wales). Rtimex Ace to sella * (Mid-Wales, especially on peat). Holcus lanatus * (Mid-Wales and Cotswolds). Bromus hordeaceus (Cotswolds). Agrostis spp. (Mid- Wales). Beilis perennis . Cnicus lance olatus.* Hypochoeris radicata .* Taraxacum officinale.* 2 Myosotis arvensis* Veronica serpyllifolia. Prunella vulgaris.* Plan tago lanceolata . * ( d ) In dry seasons and wherever the ‘ take } may have been for any reason very poor the following are prone to occur : Capsella Bursa-pastoris. Fumaria officinalis (Cotswolds). Stellaria media.* Galium Aparine.* Veronica hederaefolia (Cotswolds). Polygonum Persicaria.* P. avicidarc. Euphorbia Helioscopia. Agropyron repens . 1 This point will be dealt with at length under grass-land. See Part II. 2 This plant has been noted by Brenchley to be absent or rare under ‘ seeds’ j this is by no means the case in Mid- Wales. Stapledon.—On the Plant Communities of Farm Land . 177 (e) Species, although abundant in a district which, are quite exceptional under even the poorest £ take \ Spergula arvensis .* Legousia hybrid a, Sonchus arvensis. Chenopodiiim album.* Chrysanthemum segetum .* * The seeds of species thus marked in the above lists are all more or less plentiful in poor seed mixtures. It appears from a consideration of the above lists that except for Cnicus lanceolatus , which is perhaps a typical weed of both young leys and older grass-land, the only plants met with under * seeds ’ that do not occur elsewhere in the rotation are certain exotics1 introduced with the clover seeds and which can grow under grass-land conditions. If the plants mentioned under headings (b) and (c) are regarded as being characteristic weeds in ‘ seeds ’ on soils that suit them, it would seem that the following growth forms are well adapted to compete with the sown turf-forming and gregarious species. A. Plants which produce Seedlings capable of attaching themselves closely to the ground. ( a ) Annuals. The most successful annuals are those which either in the first or second generation form little cushions on the ground (they appear thus in the late autumn) ; these autumnal plants do not, however, flower till the following spring — that is to say, they are, under the conditions obtaining, hibernal annuals. Good examples are : Viola tricolor (agg.) (often from second generation), Alchemilla arvensis (usually first generation), Trifolium minus (first or second generation), and Myosotis arvensis (usually first generation). Even more successful are those definitely hibernal annuals which form strong rosettes on the ground during the first autumn, e.g. Geranium rnolle and G. dissectum, and Er odium cicutarium (this plant often becomes a biennial or even a short-lived perennial). (b) Biennials and Perennials. The most successful are plants which during the first autumn produce considerable cushions, mats, or rosettes close on the ground, and subse- quently develop a spreading or creeping manner of growth, or send up comparatively long flowering stems. Examples of the first type are Ranunculus repens, Beilis perennis, Chrysanthemum leucanthemum , Veronica serpyllifolia (shortly creeping), Prunella vulgaris, and Rumex Aceto sella ; and of the second, Cnicus lanceolatus, Hypochoer is radio ata , Taraxacum officinale , * Exotic to the particular district. 1 78 Stapledon.—On the Plant Coinmunities of Farm Land \ and Plantago lance olata. It will be shown (Part II) that the majority of these plants are capable of gaining considerably on grass-land as the years go on. B. Plants the Seedlings of which do not attach themselves unusually closely to the ground. (a) G r amine ae. On poor soils Holcus lanatus and Agrostis spp. in Mid-Wales and Bromus hordeaceus on the Cotswolds may be fairly plentiful even in the first year of a ley, although they only become abundant as the rye grasses die off ; thus these grasses can compete to some extent with the sown species. (b) Other Natural Orders. The commonest plants ar eArenaria serpyllifolia. This annual is capable of spreading considerably over the ground, and under Sainfoin leys especially has a very gregarious habit. Veronica agrestis usually occurs on leys at high elevations in Mid-Wales ; observations suggest that it may there produce a few flowers in the autumn, but none the less live over the winter and flower more freely during the following spring. Some plants, although endowed with favourable growth forms (when judged by the above standards), are none the less rare or exceptional on leys. A good example is Agropyron repens , a tall-growing grass with an extensive system of stolons. This plant does not seem able to compete with other gregarious plants ; possibly it requires more light and air than is available under grass-land conditions. The plants which may be met with on poor leys and in dry seasons are for the most part ephemeral annuals which can thrive on bare patches, but the seedling plants of which, not having a cushion form of growth, are rapidly suppressed when the conditions are again favourable to the spread of the grasses and clovers. Colonies of these plants are frequent where the corn has been leyed and where, consequently, the seed e take ’ has been bad — very successful species then being, Stellaria media in Mid-Wales, and Veronica hederaefolia on the Cotswolds. In conclusion, it must be pointed out that a number of weed impurities are introduced with grass and clover seeds (especially in districts where the farming is poor), but if the ‘ take * is good only such as are capable of growing under ‘seeds’ will appear to any extent in the ley; exceptionally large amounts of Geranium spp. and other species marked with an asterisk in groups ( b ) and (c) are frequently to be attributed to this cause.1 1 A number of fields in Mid-Wales have been examined in the light of the impurities found in the samples of seeds sown. Stapledon.—On the Plant Communities of Farm Land . 1 79 Conclusion. One object of the present paper has been to show that considerable advantages are to be gained by studying the weeds of arable land on a statistical basis, and in the light of the community as a whole. The application of this method immediately shows that, apart from anything else, species differ very much in their powers of colonization. Certain species, although they may be generally distributed over districts, are never numerically abundant ; others are, however, capable of forming considerable carpets on the ground. For instance, Ranunculus repens ^ Rmnex Acetosella , Spergida arvensis , and Veronica hederaefolia , and a number of other plants under congenial surroundings may have frequencies as high as 9. Euphorbia Helioscopia , Lapsana communis , and other plants, although they may grow under equally congenial surroundings, seldom attain to frequencies as high as 2. Thus when contrasting the behaviour of species under different conditions, it is necessary to have in mind their inherent capabilities as colonizers ; con- sequently the presence of such weeds as Spergida arvensis , Ranunculus repens , &c., in very small amount may, in certain cases, be just as or more significant than the complete absence of a species with a low habitual frequency.1 It can be shown, furthermore, that a knowledge of the habitual frequencies of species makes it possible to gauge with some degree of accuracy their behaviour under unusual seasonal or other change. For instance, on the Cotswolds in the spring of 1912 (i.e. following the drought of 1911), a few species doubled and in some cases trebled their habitual frequencies, e. g. Lamium amplexicaule and Arenaria serpyllifolia , whilst Veronica hederaefolia everywhere attained to something approaching its maximum figure. Important, however, as it is to take frequencies into account, it is far more important to contrast whole communities, or at all events the chief contributing species of communities, rather than to interpret the influences of soil or of cropping in terms of the behaviour of certain ‘ index ’ plants ; this has been emphasized in the body of the paper, and was well exemplified when considering the flora of ‘ sour ’ soils. The results given would seem to show that the weed communities of arable land are (1) decidedly responsive to change in soil ; (2) are different near the altitudinal limits of cultivation, to what they are on the same soils at lower elevations ; this is, however, in part due to negligent husbandry ; (3) that they are also influenced by the crop under which they grow, but that this is largely due to the husbandry associated with the various crops. It has only been sought to compare the communities under Roots (including Swedes, Mangolds, and Potatoes), Cereals (including Wheat, 1 The presence of species with high habitual frequencies in small amount only may be due to the activity of the hoe — a source of error always to be guarded against. N 2 180 Stapledon.—On the Plant Communities of Farm Land . Oats, and Barley), Vetches and ‘Seeds’. It has been pointed out that under good farming the communities with roots are meagre, but that certain species are usually more luxuriant under roots than when associated with other crops. Under poor farming the communities met with in roots and cereals do not differ much from each other. c Seeds * have been shown to favour characteristic communities, and something definite can be asserted as to the growth forms of the generally successful plants ; the nature of the community is, however, considerably influenced by success or otherwise of the sown seeds. (4) In districts where inferior and unclean seeds mixtures are used, the communities not only under the seeds, but in subse- quent crops in the rotation, may be influenced to a large extent by the added impurities. The Faugan, Llanbadarn, Aberystwyth. Literature referred to. 1. Armstrong, S. F. : The Botanical and Chemical Composition of the Herbage of Pastures and Meadows. Journ. Agr. Science, vol. ii, Part 3, Dec. 1907. 2. Bravender, R. J. : On the Indications which are Practical Guides in judging of the Fertility or Barrenness of Soil. Journ. Roy. Agr. Soc., vol. v, 1845. 3. Brenchley, W. E. : The Weeds of Arable Land in relation to the Soils on which they grow. (a) Ann. Bot., vol. xxv, Jan. 1911 ; ( \b ) vol. xxvi, Jan. 1912 ; (c) vol. xxvii, Jan. 1913. 4. Buckman, Prof. : On Agricultural Weeds. Journ. Roy. Agr. Soc., vol. xvi, 1855. 5. Kinch, E. : Manurial Experiments on Permanent Grass. (See pp. 2 and 3 for Mr. M. Kershaw’s Chemical and Mechanical Analyses.) Royal Agr. Coll. Cirencester, Bull. No. 1, 1909. 6. Moss, C. E. : Vegetation of the Peak District. Camb. Univ. Press, 1913. 7. Salter, J. H. : Flowering Plants and Ferns of Aberystwyth and Neighbourhood. Aberystwyth, undated. 8. Smith, W. G. : Raunkiaer’s ‘Life Forms and Statistical Methods’. Journ. Ecology, vol. i, No. i, March, 1913. 9. and Moss, C. E. : Geographical Distribution of Vegetation in Yorkshire. Part I. Leeds and Halifax Dist. Geogr. Journ., xxi, 1903. 10. Stapledon, R. G. : Notes on the Weed Flora of some Arable Land. Roy. Agr. Coll. Cirencester, Bull. No. 2, 1910. 11. : Pasture Problems: Drought Resistance. Journ. Agr. Science, vol. v, Part 2, 1913. 12. : The Condition of the Seed Trade in the Aberystwyth College Area. Aberystwyth, Feb. 1914. # Parallel Tests of Seeds by Germination and by Electrical Response. (Preliminary Experiments.) BY MARY T. FRASER, B.Sc. Introduction. IN previous work it has been shown by Professor A. D. Waller1 that there is a definite electrical response in the case of living seeds, which is no longer given when the seed is dead. The present experiments were under- taken with the view of further developing this electrical test, in regard to its possible commercial application. When the necessary experimental pro- cedure has been settled, and the conditions studied more fully, the electrical response should furnish a more rapid and definite indication of the vitality of a set of seeds than the ordinary germination method. It would supply the practical man’s demand for a trustworthy answer c while you wait \ Method of Experiments. (i) Germination . — Germination was allowed to proceed in the usual way, the grains being placed between filter-paper moistened with a known quantity of distilled water, or with various food solutions, in suitable dishes and kept for a certain time (24 to 48 hours) at a known temperature. The electrical response was then recorded. Those having germinated at the end of the time were counted, and the percentage of the whole calculated. The average of a number of counts was taken as the germination value . 100 grains were found to be a convenient number to use, and the total germination value was recorded at the end of 6 to 10 days. In these experiments the grains used were all of one species— Hordeum vulgare , the common barley — on account of the greater ease of manipulation of the larger grain. The observations would necessarily be extended to other seeds, where special devices might have to be adopted on account of small size, &c. 1 Proc. Roy. Soc., vol. lxviii, 1901, p. 79. [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.] I§2 Fraser . — Parallel Tests of Seeds by Different samples of barley were very kindly sent for the purpose by Mr. Stapleton of Aberystwyth. (ii) Electrical Method. — The embryos were dissected from the grain, and connected to the circuit which Professor Waller describes in his paper on the vitality of seeds, in such a way that a current passing in the direction root — > stem causes a deflexion of the galvanometer from left to right. This direction is designated by the sign + , indicating a ‘ positive ’ response from B to A. The circuit consists of : (1) A compensatory circuit. (2) Two' galvanometers in series of greater and lesser sensitiveness respectively, the deflexions of which are thrown on to a specially constructed scale, which is then easily read. (3) An induction coil capable of giving single induction shocks of known strengths and direction. By a suitable arrangement of plugs any of these branch circuits can be short circuited. When the embryo is adjusted the galvanometer plug is opened, and any current in circuit due to the plant material noted. This is immediately balanced by throwing in, from the compensator, an equal current in the opposite direction to the accidental current. A standard for the electrical response is then obtained by ascertaining the deflexion of galvanometer caused by sending into the circuit o-oi volt from the compensator. From this the magnitude of the response can afterwards be calculated. The apparatus is adjusted so that the galvanometer deflexion is at zero. Next a single-break induction shock is sent into the plant, the galvanometer circuit being closed at the time of shock, and opened directly after, so that only the response of the plant may be recorded. A series of four shocks is applied, and the corresponding responses noted, the order in which they are given being: (1) + 1,000, (2) —1,000, (3) +10,000, (4) —10,000, in notation of a Berne coil supplied by two Leclanche cells. The response, if large, is read on the scale of the less sensitive galvanometer ; if small, on that of the more sensitive one. Any residual current aroused by an induction shock is balanced from the compensatory circuit before another is delivered. At the end of the series the value of o*oi volt is again recorded. This furnishes the routine of the experiments carried out, an average of ten experiments being taken as a rule. In Table I an example of ten experiments is given. Object of Experiments. To determine how far there was a correspondence between the average germination value and the electrical response. Germination and by Electrical Response . 183 Series of experiments to record the two values were carried out at the same time — and for each experiment — under the same conditions. The problem was attacked in several different ways : (i) Using the same samples of grain giving the same average germina- tion value. Records were taken to ascertain if there was a fairly consistent response under approximately similar conditions of experiment. Here, of course, exactly consistent responses could only be looked for if great care was taken to ensure exactly similar conditions as regards temperature, moisture, electrodes, &c. Small variations in these conditions would probably produce corresponding variations in the electrical response, and these would affect the absolute magnitude of the average of responses. (ii) Using the same sample of grain. Experiments were carried out to determine whether variation in the conditions affecting the germination favourably or the reverse affect the electrical response in the same way. Two parallel determinations are carried out in this case. (iii) Using different samples of grain giving different germination values. These were compared under constant conditions as far as possible. Results of Experiments. General. — The current of injury or accidental current is in the ‘ posi- tive * direction — from root to stem — although the opposite was noted in the majority of cases. The response to shocks is regularly in the 4 positive 5 direction, but occasional alterations in the negative direction were observed. The signifi- cance of this alteration has yet to be determined. It may be an important detail. The greatest response is noticed in the case of the first strong shock, the second strong shock giving just as regularly the smallest response — due, no doubt, to fatigue. The actual magnitude of the response varies with other factors as already suggested, and would be affected by the temperature at which the experiment was carried out, and the atmospheric conditions at the time. Further investigation would probably indicate more definitely the extent of the variation, if care was taken to record the temperature, pressure, &c., exactly. (i) The general nature of the electrical response for the same sample of grain was quite consistent. Even the germination values vary slightly, and it was found that the number of seeds growing in too was not always identical in the different experiments. 1 84 Fraser. — Parallel Tests of Seeds by M ON M C>\ 4" X n3 HH w h-J pq < H > a o 4-J a a o *Q GO Hi Ctf is rd . IN <3 •SP 8 fa o o o « o o r^. o CM o 1m 00 o o o o NO o CM M NO CO o >Cj X CM CM Hi CM CM CM CM W CM CM o o o o H- o 00 On >o ON o NO «S^ CM . o H ° 1 Hi CM *Q cO CO Hi NO ^ 2 1 I + o o + 4- + o 1 + + o o o iQ o hh o lO o o PtS lO CO H- HH o *o ON 00 .<3 o HH o o HH HH HH o o o o o o o 9 o O o 9 9 9 . o o o o o o o o o o + + + + 4" + + + 4- + o > . o Si O A 00 CM 00 CM oo o CM HH 00 < 3 2 f | + CM + + HH + CM 4- CM 4" M 4- o HH + HH + + I Th CO o o CO >o IQ CM iQ HH CO CO CO 1^. CO CO o o O o o o o O o o o o o o o o o O o 9 o d o o o o o o o + + + + + + + + 4- O > $2 I i>. On CM NO NO NO ON < kl w v 1 + + + + 4* + + o o 4- • . Hi 00 iQ H- iQ cO o CO o o CM cO oo CO CO cO o o HH LO *Q k<5> O o o O o HH HH o o o o A. o o o O o O O o o o 9 • o o o o o o o o o o o • + + + + 4- 4- 4- + 4- + + o > 0 o A *Q iQ) x>. CM HH CM o < H°- t HH CM HH HH tq M I + + + + +■ + + + + + 4- o o o CM o ir. o o o CM CO 00 o o *Q CO o >o o CM CO cO o o X CM CM CM CM CM CM CO HH CM CM *< 6 o NO o NO o oo CM ON o o Hi »-* CM 1-1 CM *-« Hi CM CM CM Hh o o o o o o o o o o 0 HH CO CO CO cO CO i'- O o o o o Hi o o o o 0 O o o o 9 9 9 o o o Vo O d b d o o d b d d 1 1 1 1 1 1 1 4- 1 1 Germination and by Electrical Response. 185 In Table II the average results of five sets of experiments carried out at different times are tabulated : Table II. No. of Average electrical response in volts, strength Direction No. embryos and direction of shock being given at head Date. tested . of columns. response. + 1,000 — 1 ,000 + 10,000 — 10,000 units. units. units. units. 1. 10 0*0023 0*0019 0*0074 0*0018 April 9 all + 2. 10 0*0051 0*0032 0*0033 0*0005 „ 12 a few — 3- IO 0*0020 0*0017 0*0039 0*0002 „ H all + 4. IO 0*0021 0*0016 0*0059 0*0013 » *9 a few — 5* IO 0*0020 0*0013 0*0033 0*0004 „ 22 all + Table II. — Average responses of embryos of Hordeum vulgare dissected from the grain after germinating for 30 hours on filter-paper soaked in distilled water at a temperature of 220 C. to 250 C. Average germination value = 85 %. Table III. Electrical response in volts. Strength and direction of shock at head of columns. + 1,000 — 1 ,000 + 10,000 — 10,000 units. units. [units. units. + 0*0009 + 0*0004 + 0*0036 0 + 0*0021 + 0*0025 + 0*0035 0 + 0*0035 + 0*0040 + 0*0030 0 + 0*0040 + 0*0028 + 0*0050 + 0*0011 + 0*0000 0 + 0*0008 + 0*0008 + 0*0028 + 0*0017 + 0*0066 0 + 0*0007 + 0*0011 + 0*0030 0 0 0 + 0*0040 0 + 0*0070 + 0*0045 + 0*0085 + 0*0005 0 0 + 0*0017 0 experiments: — +0-0021 + 0-0017 + 0 0039 + 0-0002 Table III. — Electrical responses of io embryos of Hordeumvulgare after treatment for 30 hours with distilled water. Temperature = 220 to 2 50 C. Germination value = 85 %• Table IV. 'Response in volts. Strength and direction of shock at head of columns . + 1,000 — 1 ,000 + 10,000 — 10,000 units. units. units. units. 1. + 0*0020 + 0*0020 + 0*0033 + 0*0013 2. 0 0 0 0 3* + 0*0030 + 0*0020 + 0*0060 0 4. + 0*0027 + 0*0008 + 0*0080 0 5* + 0*0150 + 0*0125 + 0*0025 0 6. — 0*0060 — 0*0030 — 0*0020 — 0*0010 7* + 0*0083 + 0*0050 + 0*0025 — 0*0012 8. + 0*0100 + 0*0050 + 0*0050 — 0*0020 9* 0 0 0 0 10. Average of ten + 0*0044 + 0*0020 + 0*0037 0 + 0*0051 + 0*0032 ±0*0033 + 0*0005 experiments : — Table IV. — Electrical responses of embryos of Hordeiim vulgare after treatment for 30 with distilled water. Temperature =3 2 2° tO 2 5° C. Germination value * 85 %. Tables III and IV are given as examples of the detailed experiments. It is noticed that the responses in Table III are all in the same positive direction. The responses in Table IV are also generally in this direction, but occasionally the direction changes to negative. i86 Fraser . — Parallel Tests of Seeds by In Table V is shown the average results of three sets of experiments on another sample which gave ‘ a slightly higher ’ germination value. Table V. No. of No. of Average electrical response. Strength No. embryos hours and direction of shock given at head Date. Direction, tested. soaking. of columns. + 1,000 — 1,000 + 10,000 — 10,000 units. units. units. units. I. IO 30 0*0041 0*0029 0*0062 0*0007 April 26 + 2. 8 30 0*0025 0*0023 0*0059 0*0009 April 29 some — 3- 10 27 0*0052 0*0032 0*0079 0*0010 May 3 some — Table V. — Average electrical responses of embryos of Hordeum vulgare after germinating on filter-paper soaked in distilled water at a temperature of 2 2°C. to 25°C. Germination value = 9° %• These responses are noticeably higher on the whole than those in Table II, and the negative response only occurred in two cases in the total number of experiments. (ii) Two sets of the same sample of grains were germinated under different conditions at the same time, and the electrical response obtained. In Table VI an example of the result is given. It was found that the use of culture solution instead of distilled water increased the percentage of grains showing signs of germinating energy in a certain time. There was a distinct relationship between this and the average electrical response. Table VI. % germinating at end of 40 hours. Average electrical response. Strength and direction of shock at head of columns. + 1,000 — 1,000 + 10,000 —10,000 1. Culture solution 76 0*0016 0*0013 0*0046 0*0023 2. Distilled water 40 o*oor 1 0*0013 0*0024 0*0014 Table VI. — Average electrical responses of embryos of Hordeum vulgare germinated for hours on moist filter* ■paper. Temperature =20° C. Table VII. Electrical response in volts. Strength and direction of shock at head of columns. + 1,000 — 1,000 + 10,000 — 10,000 units. units. units. units. 1. + 0*0055 + 0*0050 + 0*0125 0 2. + 0*0041 + 0*0012 + 0*0024 + 0*0012 3* 0 — 0*0010 + 0*0055 + 0*0020 4* + 0*0014 O + 0*0023 — 0*0014 5- + 0*0070 + 0*0060 + 0*0090 0 6. 0 0 — 0*0012 — 0*0017 7* + 0*0080 + 0*0053 + 0*0112 0 8. + 0*0013 + 0*0035 + 0*0026 — 0*0013 9* + 0*0070 + 0*0060 + 0*0035 + 0*0012 10. 0 0 + 0*0008 O Average of 10 experiments + 0*0034 + 0*0028 + 0*0051 + o*oooS Table VII.- — Electrical responses of embryos of Hordeum vulgare treated with distilled water tor 28 hours. Temperature — 23°C. Germination value = 85 %. Germination and by Electrical Response. 187 Table VIII. Electrical response in volts . Strength and direction oj shock at head of columns. + 1,000 — 1,000 + 10,000 — 10,000 units. units. units. units. I. [-0-0025] + 0-0087 + 0-0062 — 0-0025 + 0-0019 2. 0 0 0 0 3- 0 0 0 0 4- 0 0 0 0 5- 0 0 0 0 6. + O-OOIO + 0-0005 — 0-0014 — o-ooio 7- 0 0 0 0 8. 0 0 — 0-0013 0 9- 0 0 0 0 10. Average of 10 + 0-0054 + 0-0041 + 0-0091 + 0-0004 + 0-0008 + 0-0013 + 0-0018 + 0-0003 experiments : — Table VIII. — Electrical responses of Hordeum vulgare treated with distilled water for 28 hours. Temperature = 23°C. Germination value = 50 %. (iii) Using different samples giving different germination values — 50 per cent, and 85 percent, respectively. The results are most striking. Tables VII and VIII are examples of ten experiments on grains which were treated with distilled water on filter-paper for twenty-eight hours under the same conditions of temperature, moisture, &c., the electrical response of both samples being recorded the same afternoon to minimize any difference in the magnitude of the response due to slight differences of temperature, atmospheric conditions, &c., at the time of experiment. It is to be observed that the much lower average electrical response in the case of grains of the lower germination value is due to the larger number of embryos giving no response to electrical stimulus. Thus the 50 per cent, germination grain gives no response in about half the experiments. In Tables IX and X the record of the electrical responses in the case of the same samples of grain as above (85 per cent, and 50 per cent.) is tabulated. The galvanometric experiments were carried out when the grains had been soaked for twenty- five hours in distilled water at a tempera- ture of about 250 C. Here this point is brought out — that there is probably a time of maximum electrical response for these embryos just about the period when growth becomes obvious. This time would vary with the sample, and it is to be expected that the better the sample the sooner this response would be given. This is obviously a point of the greatest practical importance, and having ascertained that there is a parallelism between germinative capacity and electrical response, further experiments along these lines will evidently attempt to answer the question of the practical expert : ‘ What is the shortest time at which the electrical response will indicate the comparative vitality of a sample of seeds ? 5 Table XI gives the average electrical responses of the four sets of experiments described above. 1 88 Fraser . — Parallel Tests of Seeds by Table IX. Electrical response in volts. Strength and direction of shock at head of columns. + 1,000 — 1,000 + 10,000 — 10,000 units. units. twits. twits. I. + 0*0065 + 0*0040 + 0*0150 + 0*0030 2. + 0*0058 + 0*0058 + 0*0092 + 0*0046 3- + 0*0450 + 0*0430 + 0*0220 — 0*0110 4* + 0*0100 + 0*0100 + 0*0100 0 5- + 0*0177 + 0*0085 + 0*0085 — 0*0030 6. + 0*0008 0 + 0*0023 + 0*0008 7* + 0*0263 + 0*0173 + 0*0127 — 0*0027 8. + 0*0025 + 0*0090 + 0*0115 0 9* 0 0 — 0*0020 0 10. + 0*0100 + 0*0057 + 0*0121 — 0*0021 Average of 10 experiments : — + 0*0124 + 0*0103 + 0*0105 + 0*0027 Table IX. — Electrical responses of embryos of Hordeum vulgare treated with distilled water for 25 hours. Temperature = 230 C. Germination value = 85 %. Table X. Electrical response in volts. Strength and direction of shock at head of columns. + 1,000 — 1,000 + 10,000 — 10,000 units. units. twits. twits. I. 0 0 0 0 2. + 0*0050 + 0*0040 + 0*0150 + 0*0025 3- 0 + 0*0004 0 0 4* + 0*0020 0 + 0*0025 + 0*0005 5. 0 0 0 0 6. 0 0 + 0*0004 + 0*0004 7- + 0*0100 + 0*0220 + 0*0240 + 0*0010 8. 0 O 0 0 9* 1 — 1 1 0 6 0 M to 1 1 + 0*0016 + 0*0024 + 0*0040 + 0*0012 10. + 0*0013 + 0*0040 + 0*0073 O Average of 10 experiments : — + 0*0021 + 0*0032 + 0*0053 + 0*0005 Table X. — Electrical responses of embryos of Hordeum vulgare treated with distilled water for 25 hours. Temperature = 23°C. Germination value = 50%. Table XI. Electrical response in volts. Strength and direction of shock at head of columns. Germination + 1 ,000 — 1,000 + 10*000 + 10,000 No. of hours germinated. value. units . units . units. twits. 85% 0*0034 0*0027 0*0051 0*0008 28 85% 0*0124 0*0103 0*0105 0*0027 25 50% 0*0008 0*0013 0*0018 0*0003 28 50% 0*0020 0*0032 0*0053 0*0005 25 Table XI. — Average electrical responses of embryos of Hordeum vulgare treated with distilled water. Temperature = 23°C. Summary of Results and Conclusions. 1. The same samples of grain germinated under approximately the same conditions give results in which the germination value and the electrical response are quite consistent. Germination and by Electrical Response . 189 %. The electrical response of grains germinated under different con- ditions, such as food supply, temperature, &c., varies in the same way as the germination value. 3. Samples of grain giving good and bad germination values give average electrical responses which vary strikingly in the same way. 4. There is a certain amount of evidence that there is a time of maximum electrical response, probably corresponding with the time when growth becomes established. Experiments are being carried out to deter- mine this point with more precision. 5. In the samples giving a low germination value there are always a certain number of embryos which give no response to the electrical stimulus, indicating they are incapable of germination. A high proportion of such zero results under conditions which would normally produce a response would indicate a sample of low germinative value. 6. The electrical response can be ascertained in a much shorter time than the total germinative value. 7. There is an indication that the electrical response would discriminate not only between a ‘ live ’ seed and a ‘ dead ’ seed, but between a ‘ live ’ seed of high vitality and one of low vitality. I take this opportunity of expressing thanks to the Board of Agriculture, with whose assistance this work has been carried out. - "I : . . ' v , • ' ■: ' ■ ■ ' ' ... S NOTES ANOMALIES IN THE OVARY OF SENECIO VULGARIS, L.~ In all the species of Compositae examined hitherto, with the exception of Taraxacum officinale 1 and Zinnia spp.,2 only one ovule has been reported in each ovary. Schwere 1 figures two ovules with embryo sacs in T. officinale , but in this case they were apparently basal. In the present investigation of Senecio vulgaris several abnormalities have been observed which are of phylogenetic importance. The occurrence of biovulate ovaries was noted in several instances, two of which are figured (Figs, i and 2). In the former case the archesporial cell is quite distinct, and a wall two cells thick extends across the ovary. The occurrence of a biovulate, bilocular ovary is interesting confirmation of the derivation of the unilocular ovary of the Compositae from the former type. Both ovules in this case were on Fig. 1. Fig. 2. one side of the ovary, but in the latter case (Fig. 2), where there is no wall, the ovules are placed on either side of the ovary. The four ancestral placentae are thus indicated. In both cases the ovules are lateral. Numerous cases of single lateral ovules were observed in the younger stages (Fig. 3), and this is an abnormality of some importance. Particular care was exercised in determining that these ovules were truly lateral ; the presence of conducting cells and the arrangement of the adjacent parenchyma leave no doubt as to the actual attachment of the ovule to the side wall of the ovary. The arrow in the figures indicates the position of the axis. Fig. 3 is a longitudinal section, tangential to the axis. 1 Schwere, S. : Zur Entwickelungsgesch. der Frucht von Taraxacum officinale. Flora, vol. Ixxxii, 1896. 2 Don, D. : On the Origin of the Ligulate Rays in Zinnia. Trans. Linn. Soc., vol. xvi, 1829. [Annals of Botany* Vol. XXX. No. CXVII. January, 1916.] 192 Notes, Warming1 removes the Calyceraceae from the vicinity of the Com- positae, and places them near the Dipsaceae on account of the ovule, which he distinguishes as ‘ apotrope i. e. placed on the anterior wall of the ovary with the raphe anterior. He argues that if the basal ovule of the Com- positae, which he designates 1 epitrope *, were displaced so that it was pendant, the raphe would become posterior, instead of anterior as it is in the pendant ovules of the Dipsaceae and Calyceraceae. This assumes that the displacement is in the median plane of the flower, but the present observa- tions of the orientation of lateral, i. e. displaced, ovules show that dis- placement takes place in the plane at right angles to the median plane. Therefore, if the basal ovule of the Compositae were displaced so that it became pendant, the raphe would remain anterior and the ovule would have the same position as in the Calyceraceae and Dipsaceae. This displacement in the lateral plane is just what might be expected, con- sidering the obvious derivation of the unilocular ovary of the Compositae from a bicarpellary ovary in which the carpels were on the antero-pos- terior plane. Therefore the Calycera- ceae, in accordance with the usually accepted opinion, may be allowed to remain near the Compositae. Van Tieghem 2 also removes the Calyceraceae from the vicinity of the Compositae, and places them near the Rubiaceae on account of the ovule, but the value of the ovule in the classification of this portion of the Sympetalae has been somewhat over-emphasized considering the com- parative frequency of these lateral ovules, which form a transition to the Calyceraceae and Dipsaceae. JAMES SMALL. Armstrong College, October , 1915. 1 Warming, E. : Observations snr la valenr systematique de P ovule. Mindeskrift f. Japet, Steenst., 1913. 2 Van Tieghem, P. : L’oeuf des plantes. Ann. sci. nat., Bot., ser. 8, t. xiv, 1901. Fig. 3. Notes. 193 NOTE ON THE STRUCTURE OF THE OVULE OF LARIX LEPTO- LEPIS. — The writer began the study of the structure and life-history of the various Larches in March 1915. Such a study seemed profitable, as, except for occasional observations, the group has been practically neglected. Even such references as do exist are mostly detailed cytology, like Allen’s admirable account of the formation of the spindle in the reduction division of the pollen mother-cells of L. enropaea (Ann. Bot., vol. xvii, 1902). But the broad outlines of the life-history are still far from ascertained. Unfortunately, the work has not made sufficient progress this season to warrant a complete account, and this for many reasons. The absence of information on the group meant that the dates of the appearance of the various stages were unknown, and it was thought advisable to determine these dates as a preliminary to further work. The abnormal sterility of the specimens at the writer’s disposal was also a hindrance. It was no uncommon thing in July, when removing the hardened Longitudinal section of the upper part of the ovule of Larix leptolcpis. A, nucellus; k, embryo-sac; c, integument; 1, 2, 3, its layers in process of differentiation; D, micropylar con- striction ; e, fold with stigmatic hairs; F, a pollen-grain entangled ; G, thickened plate from which hair-like processes arise. integument, to cut into 200 ovules before meeting a good one. Finally, difficulties in procuring fixing chemicals, &c., owing to the war, also retarded progress. But there were one or two points which appeared which it is considered well to put at once into this preliminary note. I. The structure of the ovule was interesting and peculiar. Most worthy of record, however, is the fact that its structure was almost identical with that described by Lawson for the ovule of Pseudotsuga Douglasii (Ann. Bot., vol. xxiii, 1909). To describe its appearance and to emphasize the fact of its similarity to that of Pseudotsuga nothing can be better than to follow Lawson’s description of his ovule on the accompanying figure, which is a longitudinal section of the upper part of the [Annals of Botany, Vol. XXX. No. CXVII. January, 1916.3 O* 194 Notes . ovule of Larix lepiolepis. Lawson says : ‘ The pollen-receiving device in Pseudo- tsuga is quite peculiar and unlike anything yet described for Gymnosperms. For some little time after pollination the nucellus presents the form of a small protuberance with a perfectly rounded apex (a). The integument (c) extends for a considerable distance beyond the nucellus. At a point immediately above the apex of the nucellus the integument bends inwards in such a fashion as to partly close or narrow the micropylar canal, and then sharply bends out again. This results in the formation of adistinct stricture midway between the apex of the nucellus and the mouth of the micropyle (d). As a result of this peculiar curvature of the integument, the micropylar canal is not a straight passage of uniform width, but consists of two chambers, one immediately above the apex of the nucellus and the other near the mouth of the micropyle. In addition to this narrowing in the middle region of the micropyle, the integument is still further modified. The extremity of the integument which forms the mouth of the micropyle is folded inward (e). On the inner surface of this enfolding extremity numerous fine hair-like processes are present. A close examina- tion of these processes makes it clear that they were not cellular in structure, but were merely outgrowths from the external walls of the epidermal cells. They serve very effectively, however, as a stigmatic surface.’ The hairs in Larix lepiolepis are firmer than those figured by Lawson, and arise from a basal plate (g). He goes on to say that pollen-grains were never found on the nucellus : ‘ They were invariably found in the upper chamber of the micropyle and frequently entangled in the hair-like processes of the mouth.’ A similar condition is found in lepiolepis , the figure showing a pollen-grain so entangled (f). The quotation shows how like the two ovules are. The explanation may be, and most probably is, biological, but may also have some phylogenetic significance. The megaspore membrane in lepiolepis does not cover the upper end of the endo- sperm, a condition similar to Pseudotsuga. There are very small archegonial chambers in both. Lawson states the frequent presence of only one tier of neck- cells in Pseudotsuga . In Z. lepiolepis the prevailing condition seems to be one tier of five cells. There are typically five archegonia in Z. lepiolepis , four to six in Pseudo- tsuga. All these points are, of course, of minor importance, but with the peculiar sameness of the ovules there is an obvious temptation to magnify them. The detailed investigation may, however, settle the point. II. The archegonial jackets usually touch, so that two archegonia are only separated by two cell-layers. Frequently these coalesce to one, and even the arche- gonia may be separated only by the shrivelled remains of degenerated jacket-cells. III. Double pollen-grains are very plentiful in Z. lepiolepis , as already described by Coker (Bot. Gaz , vol. xxxviii) for Z. europaea , by Hutchinson (Bot. Gaz., April, I9I5) for Picea , and others. Their origin will be investigated next spring, as, even though material was first collected on March 15, practically all the pollen was shed on March 16, in spite of the fact that the spring was a phenomenally late one. Z. europaea presents similar features. IV. There is one last point — a vegetative abnormality. As is well known, the male buds of the Larch appear terminally on dwarf shoots from the second year on. While collecting such buds — already well developed in August — a case was noticed of a dwarf shoot carrying such a bud. The shoot was six years old, judging by the Notes. i95 leaf-scar rings on it. Growing out from the dwarf shoot from the axil of the third- year ring of leaves there was a very small secondary dwarf shoot with six or seven small leaves on it. Such an appearance is quite understandable, because, if a dwarf shoot under suitable conditions can become a long shoot upon which dwarf shoots are subsequently formed, there is no reason a priori why secondary dwarf shoots should not appear on an ordinary dwarf shoot. But the branching of dwarf shoots in the Abietineae has not often been described. Special thanks are due to Sir Frederick Moore, of the Glasnevin Gardens, Dublin, who has placed several Larch trees entirely at the writer’s disposal. JOSEPH DOYLE. Biological Laboratory, University College, Dublin. THE ANNALS OF THE BOLUS HERBARIUM (South African College) Edited by H. H. W. PEARSON, Sc.D., F.L.S. HARRY BOLUS PROFESSOR OF BOTANY IN THE SOUTH AFRICAN COLLEGE, CAPETOWN HON. DIRECTOR, NATIONAL BOTANIC GARDENS, KIRSTENBOSCH Now Ready : Vol. I, Parts lit and IV. Royal 8vo Paper covers. 5s net each Contents of Part III. — Notes on Lessertia, with descriptions of six new species and a Key. By L. Bolus. — On the Flora of the Great Karasberg : List of Plants collected ( concluded ). By F. Bolus, L. Bolus, and R. Glover. — Three Fungi collected on the Percy Sladen Memorial Expeditions of 1910-11 and 1912-13. By I. B. Pole Evans. — Key to the Flora of the Cape Peninsula: I. Spermaphyta ( continued ). By F. Bolus and L. Bolus.— Novitates Africanae. — Notes on Imperfectly knoWn Species. Bjr L. Bolus. — National Botanic Gardens : An Old Cape Frontier. By E. A. Walker, — Review. (With 8 plates.) Contents of Part IV. — Extra-tropical South African Acaciae. By R. Glover. — A Note on the Inflorescence and Flower of Gnetum. By H. H. W. Pearson. — Key to the Flora of the Cape Peninsula: I. Spermaphyta ( continued ). By F. Bolus and L. Bolus. — Novitates Africanae. (With 9 plates.) A prospectus giving full particulars of the Annals will be sent on request. London Cambridge University Press Fetter Lane CURTIS’S BOTANICAL MAGAZINE 1787—1916 Contains Coloured Figures so arranged as to show details of the plants and structural dissections, with full descriptions of new and rare plants from the Royal Botanic Gardens, Kew, and other botanical establishments. The Magazine was commenced in 1787 and forms the most extensive and authentic repertory extant. It is published monthly, with four or five hand-coloured plates and letterpress. Price 3s. 6d. net. Fourth Series. Edited by Sir D. PRAIN, C.M.G., C.I.E., LL.D., F.R.S., Director of the Royal Gardens, Kew. Vols. I-XI (1915), 42s. each. The set of eleven volumes at 36s. each. Annual Subscription for the current issue as published monthly , 42s. THE POTAMOGETONS OF THE BRITISH ISLES With Descriptions of all the Species , Varieties , and Hybrids By ALFRED FRYER, A.L.S. Continued from Mr. Fryer’s Notes by A. H. EVANS, F.Z.S., and concluded by ARTHUR BENNETT, A.L.S. The illustrations consist of 59 plates, mostly drawn from fresh specimens by the late Mr. Robert Morgan, F.L.S. , depicting each species, in many cases their various stages of growth, and with enlarged drawings of flowers, fruit, and sections. The plates of the last section are mainly by Miss M. Smith, of Kew. Complete with 39 hand-coloured plates , £5 5s. With plates uncoloured , £3 15s. Many other Standard Botanical Works and Colonial Floras. FULL CATALOGUES AND PROSPECTUSES ON APPLICATION. LONDON: L. REEVE & CO., LTD., 6 HENRIETTA STREET, COVENT GARDEN, W.C. THE JOURNAL OF ECOLOGY l ( . EDITED FOR THE BRITISH ECOLOGICAL SOCIETY BY FRANK CAVERS THE objects of The Journal of Ecology are (i) to foster and promote in all ways the study of Ecology in the British Isles, serving as the organ of the British Ecological Society; (2) to present a record of and commentary upon the progress of Ecology throughout the world. Its contents include original articles and notes on ecological topics of current importance, including methods of ecological study and research ; reviews and notices of publications of general ecological bearing and of work upon plant and animal communities in all parts of the world ; reports and notices of meetings of the British Ecological Society ; and the progress of the Nature Protection movement and of ecological work in Nature Reserves. The Journal of Ecology , being largely a Review, is indispensable to those desiring to keep abreast of the extensive and widely scattered literature of Ecology, of which it will take a comprehensive view, including in its scope not only work dealing with general principles, methods, surveys, and other ecological studies, but also publications dealing with Plant and Animal Ecology and Geography, Plankton- ology, the science of soils (soil physics, chemistry, and the investigation of soil floras and faunas), and other branches of Natural Science (Geology, Climatology, Agriculture, Forestry, &c.), so far as these have a direct bearing upon Ecology — the relationships of plants to their environment in the widest sense, including the inter-relationships between plants and animals. In brief, it aims at furnishing a comprehensive review of progress in the entire field of Ecology, touching, as this field does, kindred branches of science on various sides. The Journal of Ecology is published quarterly — in March, June, September, and December. The annual subscription price, including postage to any part of the world, for a single copy of each of the four issues making up the annual volume, is Fifteen Shillings (155.) net ; single copies, Five Shillings (5 s.) net each. Subscriptions for the Journal are payable in advance and should be sent to Mr. C. F. Clay, Cambridge University Press, Fetter Lane, London, E.C., either direct or through any bookseller. Members of the British Ecological Society should send their annual subscription to the Society, One Guinea (,£1 is.), which includes the supply of one copy of each of the four issues of the Journal , to the Secretary, Dr. F. Cavers, Goldsmiths’ College, New Cross, London, S.E., to whom all Editorial communications should also be addressed. Contents of Vo l. Ilf No. j. December igij RECOLONIZATION OF CULTIVATED LAND ALLOWED TO REVERT TO NATURAL CONDITIONS. By Winifred E. Brenchley and Helen Adam. ON THE ECOLOGY OF THE VEGETATION OF BRECKLAND. By E. P. Farrow. NOTICES OF PUBLICATIONS OF GENERAL BEARING. NOTICES OF WORK ON BRITISH VEGETATION. NOTICES OF WORK ON FOREIGN VEGETATION. TITLE-PAGE AND INDEX TO VOL. III. 3 Plates, 6 Text-figures. Vol. XXX. No. CXVIII. April, Annals of Botany EDITED BY DUKINFIELD HENRY SCOTT, M.A., LL.D., D.Sc., For. Sec. R.S. LATELY HONORARY KEEPER OP THE JODRELL LABORATORY, ROYAL BOTANIC GARDENS, KEW JOHN BRETLAND FARMER, D.Sc., M.A., F.R.S. A PROFESSOR OF BOTANY, ROYAL COLLEGE OF SCIENCE, LONDON FRANCIS WALL OLIVER, M.A., D.Sc,, F.R.S, QUAIN PROFESSOR OF BOTANY, % UNIVERSITY COLLEGE, LONDON AND ROLAND THAXTER, M.A., Ph.D. PROFESSOR OF CRYPTOGAMIC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S.A ASSISTED BY OTHER BOTANISTS LONDON HUMPHREY MILFORD, OXFORD UNIVERSITY PRESS AMEN CORNER, E.C. Edinburgh, Glasgow, New York, Toronto, Melbourne, and Bombay 1916 . ■ TV Printed in England at the Oxford University Press CONTENTS. PAGE Alfred Stanley Marsh xxv-xxvii Takeda, H. — Some Points in the Morphology of the Stipules in the Stellatae, with special reference to Galium. With twenty-seven Figures in the Text 197 Hill, Arthur W. — Studies in Seed Germination. The Genus Marah (Megarrhiza), Cucurbitaceae. With Plate V and two Figures in the Text 215 Hind, Mildred.— Studies in Permeability. III. The Absorption of Acids by Plant Tissue. With eleven Figures in the Text ......... 223 DE Fraine, E. — The Morphology and Anatomy of the Genus Statice as represented at Blakeney Point. Part I. Statice binervosa, G. E. Smith, and S. bellidifolia, D.C. (= S. reticulata). With systematic and ecological notes by E. J. Salisbury. With Plate VI, twenty-eight Text Figures, and four Tables 239 Delf, E. Marion.- — Studies of Protoplasmic Permeability by Measurement of Rate of Shrinkage of Turgid Tissues. I. The Influence of Temperature on the Permeability of Protoplasm to Water. With seventeen Figures and five Tables in the Text . . 283 Groom, Percy. — A Note on the Vegetative Anatomy of Pherosphaera Fitzgeraldi, F. v. M With one Figure in the Text 311 Sampson, K. — The Morphology of Phylloglossum Drummondii, Kunze. With five Figures in the Text 315 Sutherland, Geo. H., and Eastwood, A.—1 The Physiological Anatomy of Spartina Townsendii. With seven Figures in the Text 333 NOTES. Doyle, Joseph. — On the ‘Proliferous’ Form of the Scape of Plantago lanceolata. With two Figures in the Text . . ........ 353 Salisbury, E. J. — On the Relation between Trigonocarpus and Ginkgo . . . 356 West, Cyril. — Stigeosporium marattiacearum, gen. et sp. nov. . . . -357 NOTICE TO SUBSCRIBERS. The subscription-price of each volume is thirty shillings, payable in advance : the Parts, four in number, are supplied as they appear, post free to subscribers in the United Kingdom, and with a charge of is. 6d. per annum for postage to subscribers residing abroad. The price of individual Parts is fixed at a higher rate. Intending subscribers should send their names, with subscription, to Humphrey Milford, Oxford University Press, Amen Corner, London, E.C. As the earlier volumes of the Annals of Botany are becoming scarce, Vol. I will only be sold as part of a complete set ; and Parts will not as a rule be sold separately, after the publication of the volume to which they belong. A few extra copies of particular Parts at present remain on. hand, for which special application must be made to the Editors, Clarendon Press, Oxford. NOTICE TO CONTRIBUTORS. In future, English contributors should address their papers to Professor F. W. Oliver, University College, Gower Street, London, W.C. ; and contributors in America to Professor R. Thaxter, Harvard University, Cambridge, Mass. Papers sent in with a view to publication must be type-written and must conclude with a summary of the contents . In view of the increasing numbers of manuscripts submitted and the inevitable limitations of space, the Editors desire to impress upon contributors the importance of conciseness in the treatment of their subject-matter. Whilst they are not prepared at present to draw a hard and fast line, the Editors suggest that the longer papers should not exceed 12,000 words. Illustrations. These where possible should be suitable for reproduction as line- blocks in the text. Where lithographic or collotype plates are required the figures should be planned so as to fill properly a 4to or an 8vo plate. The maximum space available for figures in a 4to plate is 8JxiiJ inches, in an 8vo plate 8jx5j- inches. Attention to this will conduce to the rapid publication of papers if accepted. Each contributor to the Annals of Botany is entitled to receive gratis one hundred separate copies of his paper, and may purchase additional copies if he informs the Editors of his wishes in this respect when he returns corrected proofs. The price of these additional copies will depend upon the amount of text and the number of plates in the paper. /^Oisonian Instit f jUN 17 1916 Some Points in the Morphology of the Stipules in the Stellatae, with special reference to Galium. BY H. TAICEDA, D.I.C. With twenty-seven Figures in the Text. IT is hardly necessary to recapitulate here the results of the previous investigators on the nature of the leaf-like organs or c leaves 5 in the apparent whorls of Galium , and of many other members of the Stellatae (or Galieae, a tribe of the Rubiaceae). There is little doubt that in any whorl the two opposite ‘leaves’, one at any rate of which subtends an axillary shoot, are the true leaves, while the other members at the same node are stipules. Thus, in the case of a six-membered whorl there are two leaves, each of which is provided with two stipules. Where only five or four ‘ leaves ’ occur in a whorl, it is usually understood that in the first case one, and in the second case both, pairs of stipules have undergone a concrescence. If, however, more than six ‘ leaves ’ are present in a whorl, it is explained that one or more of the original four stipules have undergone chorisis, resulting in the production of supernumerary members.1 Eichler found in Galium Mollugo and also Rubia tinctorum , that there are often two primordia which fuse, giving rise to a single interfoliar stipule on either side of the stem.2 Goebel,3 on the other hand, found in 1 Cf. de Candolle, Vegetable Organography (Engl, ed.), vol. i (1841), p. 286; Le Maout et Decaisne, Traite general de Botanique, p. 15 (1868); Leunis, Synopsis der Pflanzenkunde, ed. 3, i, p. 193 (1883); Pax, Allgemeine Morpbologie der Pflanzen, p. 102 (1890); Velenovsky, Vergleichende Morphologie der Pflanzen, pt. 2, p. 433 et seq. (1907) ; Worsdel, Principles of Plant-Teratology, vol. i (1915), p. 172. It may be of some interest to mention that Wydler, who recorded the occurrence of a complete fission of a stipule into two separate organs in Galium Cruciata , Linn, (in Flora, vol. xlii, 1859, p. 10), suggested that each of the apparently single stipules in this species may be the product of a fusion of two separate organs. * If so,’ he says, ‘ the forma- tion of a midrib (indistinguishable from the midrib of the true leaves) on the common border of two fused stipules is remarkable. The midrib in the (fused) stipules would then correspond to commissural ribs, like those for example in a gamosepalous calyx.’ 2 Eichler, Entwickelungsgeschichte des Blattes, &c., p. 32, Taf. i, Fig. 18 (1861). Pax (1. c.) follows this view, and gives Galium roiundifolium and G. palustre as examples. 3 Vergleichende Entwickelungsgeschichte der Pflanzenorgane. Schenk’s Handbuch der Botanik, iii (1884), pt. 1, Fig. 48 b, p. 231. See also Goebel, Organography of Plants (Engl, ed.), pt. 2 (1905), p. 369. [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] - P 198 Takeda . — Some Points in the Morphology of the Galium palustre that the concrescence of two stipular primordia, as described by Eichler, is of very rare occurrence, and that, as a rule, only a single primordium gives rise to a single stipule. Franke,1 who had fresh material ol a good many species of Galium and other genera of the Stellatae at his disposal, has confirmed Goebel’s observations, and states 2 that the four-membered whorls in the Stellatae arise without exception from four uniform protuberances. He reached the conclusion by investigation of Galium aetnicum , G . boreale , G. Cruciata , G , lucidum . G. Molhigo , G. parisiense, G. physocarpum, G.pusillum , G. re- curvum , G. rubioides , G. saccharatum , 6\ sylvestre, G. tenuissimum , 6\ verum , and several other members of the Stellatae, including Rubia tine tor urn . It seems worth while, therefore, to inquire how far the feature described by Franke is general, and also whether there is any further evidence for Eichler’s statement. Fenzig3 records that in Rubia peregrina one often finds whorls in which two forked stipules occur opposite one another, or, where the fission is complete, that six ‘leaves’ (i. e. two leaves and the two pairs of stipules belonging to them) are present at a node. He also mentions4 that in Galium Cruciata the stipules are occasionally partially or completely divided, so that whorls with one or two of the laminae forked, or with five to six separate members, arise. The writer can add Galium gracile, Bunge,5 as giving an example of a similar phenomenon. This species, which is widely distributed over Japan and China, constantly possesses four ‘leaves’ at each node, two being the true leaves and the other two stipular. Although no specimen has ever been found with more than four ‘ leaves ’ at a node, examples are however fairly frequent in which one or both of the stipules of a whorl have two midribs, indicating their double nature. The apex of the stipules with double midribs is usually more or less indented, but cases are occasionally found in which the apex is almost entire. There also occur, though less frequently, stipules with a forked midrib. Figs. 1-8 have been selected from two specimens of this plant gathered by Swinhoe in the interior of Amoy, China, to illustrate these features. Fig. 1 represents an ordinary stipule which possesses a single midrib and two lateral veins, and assumes exactly the same shape and size as the true leaf. Fig. 2 is 1 Beitiage zur Morphologie und Entwickelungsgeschichte der Stellaten. Botanische Zeitung, vol. liv (1896), pt. 1, p. 33 et seq. 2 Franke, 1. c., p. 50. 3 Pflanzen-Teratologie, systematise!! geordnet, vol. ii (1894), p. 37. 4 Penzig, 1. c., p. 38. 6 Enumeratio plantarum quas in China boreali collegit (1832), p. 35. For a more detailed description see Makino in Tok\6 Botanical Magazine, vol. xvii (1904), p. 74. The plant has also been described as G. miltorrhizum : Hance, in Seemann’s Journal of Botany, vol. vi (1868), p. 114. Stipules in the Stellatae, with special reference to Galium . 199 a similar stipule, but the midrib is forked at the tip. obviously indicating the fused nature of the stipule. Fig. 3 shows the midrib still more deeply forked, the apex of the stipule being at the same time slightly indented. Fig. 4 has two complete midribs as well as the two lateral veins which are normally present, while the apex of the stipule is very shallowly indented. This is the type of the double stipule which commonly occurs in this species. Fig. 5 shows a shallow notch at the apex, while one of the two midribs is provided with two lateral veins ; one of the latter being in the middle region of the stipule. Figs. 6 and 7 are stipules with a deeper notch at the apex and exemplify a type of much less common occurrence. Fig. 8 shows the deepest cleft the writer has seen. Figs. 1-8. Stipules of Galium gracile, Bunge. All x 1*5. Explanation in the text. It is obvious that the appearance of a double midrib is not due to an abnormally strong development of one of the lateral veins, since the lateral veins are always present in the stipules, whether they have one or two midribs. Moreover, the midrib may sometimes fork into two, as shown in Figs. 2, and 3, and this phenomenon is held to be a step towards the pro- duction of two complete midribs. The stipule with two midribs and three lateral veins (Fig. 5) represents the nearest approach observed to a separa- tion of the stipule into two complete organs. Since the development of the leaves and stipules could not be investigated in the material at the writer’s disposal, it is impossible to state definitely whether the ordinary stipules (i. e. those with a single midrib) are always produced as the result of a true concrescence of two primordia. It is quite possible that each of the ordinary stipules is usually produced from a single primordium, and that two types of development may be found in one and the same species. Another instance is afforded by Galium par ado xum, Maxim.,1 which also is a Far Eastern species.2 This plant possesses four ‘ leaves ’ at each 1 Bulletin de l’Acad. Imp. de St.-Petersb., vol. xix (1873), p. 28 r. 2 This species was first discovered in Manchuria in i860. Later, in 1879, Franchet and Savatier (Enum. PI. Japon., vol. ii, p. 392) recorded it from Japan. Its occurrence in China (Hupeh) was first made known by Diels (Engler’s Bot. Jahrb., vol. xix, 1901, p. 583), and in Korea by Kamarov (Flora Manshuriae, vol. ii, 1907, p. 497). It is not uncommon in the mountainous distiicts of Japan, but is apparently scarce in China. From the latter country the writer has seen the following specimens: Hupeh, Chienshih (Henry, No. 5851), Patung (Henry, No. 6026), Changyang (Wilson, No. 1153), Yunnan; ad collem Yen-tze-hay, alt. 3,200 m. (Delavay, No. 3102). 200 Takeda . — Some Points in the Morphology of the node in the middle and upper parts of the stem, i. e. two true leaves and two stipules, the latter resembling the former, but easily distinguishable from the true leaves at the same node by the fact that they are often smaller, and also that they are always bearded at the base of the petiole. In the lower part of the stem there are again two interfoliar stipules, which, however, are quite distinct in appearance from the true leaves, being small and scale-like. They are subulate or triangular, membranous, often thinly ciliate on the margin, and also more or less barbate at the base (Fig. 9). In one of the specimens collected on Mt. Fuji in 1887, and sent to Kew from the Tokyo Imperial University, the writer found a small, scale-like stipule with a forked midrib at a node in the lowermost region of the stem, the apex of the stipule being at the same time slightly indented (Fig. 10). The writer has not been so fortunate as to find in this species Figs. 9. 10. 11. 1 2. Figs. 9-1 i. Galium paradoxum, Maxim., showing scale-like interfoliar stipules from the lower region of stem. Figs. 9 and 11, x 7, fig. 10 x 6. Fig. 12. Galium saxatile , Linn. Stipule with two separate midribs, x 6. Fig. 13. Asperula asterocephala , Bornm. Stipule with two separate midribs, x 2. any stipules with two separate midribs. A case has, however, been found in which two separate, small, scale-like stipules are present on one side of a node near the very base of the stem of one of the specimens gathered by Wilson in Changyang, in the province of Hupeh, China (Fig. 11). A third instance of ‘ four-leaved ’ species occasionally producing stipules with two midribs is presented by Aspenda asterocephala , Bornm.1 This handsome perennial, which wras discovered by Bornmuller in Kurdistan, attains a height of about 40 cm., and always bears four ‘ leaves ’ at each node, just as the first example given above. The stipules can readily be distinguished from the true leaves by the fact that they are always dis- tinctly smaller than the latter at the same node, and one-nerved, while the true leaves are often provided with one or two lateral veins on either side of the midrib. Towards the apex of the stem the stipules become very much reduced in size, often to such an extent as to appear almost scale-like, 1 In Mittheil. Thiir. Bot. Ver., N. Folge, vol. vii (1895), p. 7. Stipules in the Stellatae , with special reference to Galium. 20 r although they are always green and not scarious, thus differing from the scale-like stipules at the lower nodes in Galium paradoxum , Maxim. In one of the co-type specimens of A. asterocephala , which was distributed under J. Bornmiiller, Iter Persico-turcicum, No. 1337, the writer found at the third node from the base a stipule having two midribs, while its general appearance is very similar, except for size, to the true leaves at the same node (Fig. 13). So far as has been seen, the occurrence of stipules of this nature seems to be rare in this species. As another instance, Galium leiophyllum , Boiss. et Hohenack.,1 in which the writer found a stipule of a double nature, may be briefly described below. This plant is of moderate size 2 and bears five, or more often six ‘ leaves ’ to a node. In one of the specimens distributed under J. Bornmiiller, Iter Persicum alterum, No. 7105, 3 the writer found below the middle region of a stem (about the fifth node from the base) a whorl consisting of two true leaves and two stipules, one of the latter being furnished with two midribs (Fig. 21). The node immediately above and that below this whorl bear five ‘ leaves ’ each, and all the nodes in the higher region of the same stem are six-membered. The writer also found in a specimen of Galium saxatile , Linn., which he collected himself in England in 1915, a stipule with two midribs. It occurred in a four-membered whorl which was preceded by a five-membered whorl, and succeeded by a six-membered one (Fig. 12). Further cases of the occurrence of stipules of a double nature in four- membered whorls have been found in Asperula arvensis , Linn. (Fig. 19), A. sherardioides (Boiss.), Jaub. et Spach. (Fig. 20), and A. aspera (M. Bieb.), Boiss., to each of which further allusion will be made later on. It may, therefore, be not unreasonable to conclude from the instances above referred to, that in certain species of Galium and some other members of the Stellatae, such as those which were investigated by Goebel and by Franke, each stipule in a four-membered whorl arises from a single pri- mordium, while in some species such as G. gracile , G. paradoxum , G. leiophyllum , and G. Cruciata , and the three species of Aspemda above referred to, stipules are occasionally produced as the result of a coalescence of two primordia. The first case is regarded as a congenital concrescence,4 while the second is a true concrescence, and at the same time points towards the production of more than two stipules at a node. In regard to the whorl with five members, mention has already been made above of the result of the investigation carried out by Eichler,5 who found in Galium Mollugo that two stipular primordia on one side of a node 1 Boissier, Diagnoses, vol. i. 3 (1843) p. 36 ; Ejusd. FI. Orient., vol. iii. (1875), p. 51. 2 3°-45 cm. in height. 3 This specimen is to be referred to var. subvelutinum : Boiss., 1. c. 4 Goebel, 11. cc. - 5 1. c\, p. 32, Taf. i, Fig. 15. 202 Takeda. — Some Points in the Morphology of the fuse, and thus give rise to a single stipule, while each of the two stipular primordia on the other side of the same node develops into a separate organ. Franke,1 on the other hand, states that in this case the solitary stipule on one side of the node is produced from a primordium which is homologous with a stipular primordium of the four-membered whorl, while the paired stipules on the other side of the node arise from a common arc -like pro- tuberance, which is from the beginning distinguishable by shape and insertion. The latter phenomenon is, however, not to be considered as a fission in the proper sense. In other words, each of the three stipules in the five-membered whorl arises from a single primordium, and there is no evidence of a true concrescence of two stipular primordia giving rise to a single organ. So far as the writer’s own observations go, stipules of a double nature are very rarely met with in five-membered whorls. The only plant in which such an occurrence has been found is Asperula aspera (M. Bieb.), Boiss. (Fig, 22). This case, however, gives sufficient evidence for the conclusion that stipules are occasionally produced also in five-membered whorls as the result of a true concrescence of two primordia. A further reference to this plant will be made later on with regard to the distribution of double stipules. As to the case in which more than six members occur in a whorl, it is regarded that in addition to the two true leaves there are more than two stipules on one or both sides of the stem. Eichler 2 found in Galium Mollngo that a new tissue arises between the original two stipular primordia and produces an independent organ resembling the other stipules. If this phenomenon takes place on one side only of a stem, there would be five stipules produced. If, however, a new tissue between two primordia occurs on both sides of the stem, there would be six stipules formed, thus making the whorl eight -membered at a node. Franke 3 has confirmed Eichler’s observations, and states that he found in Asperula odorata and A. azurea a new primordium arising between two stipular primordia on one or both sides of the node, the former being distinguishable from the latter by its smaller size. Neither of these investigators has examined any whorl with more than eight members, although there are certain species of Galium , such as G . verum , Linn , and of some other genera, e. g. Asperula odorata , Linn., which often bear more than eight, and, particularly in the former plant, up to twelve, foliar members (i. e. two true leaves and ten stipules) at a node. Unfortunately, the writer’s own observations on these pleiomerous whorls in herbarium material cannot throw any light upon the subject. It may, however, be presumed that there arise as many primordia as there 2 1. c., p. 32, Taf. i, Figs. 1 6, 17. 3 i. c. , p. 51, Taf. i, Figs. 7, 8. 1 1. c., p. 50. Stipules in the Stellatae , with special reference to Galium. 203 are ‘ leaves ’ at a node. Thus, as many as five stipular primordia may occur on either side of a node. In this connexion mention may be made of some interesting features exhibited by Didymaea mexicana , Hook, fil.,1 which also belongs to the Stellatae. As a unique character 2 of this (monotypic) genus the plant has been described as having each of its (opposite) leaves provided with a pair of typically differentiated stipules, the latter being subulate or lanceolate and quite distinct from the true leaves.3 In Didymaea there are as a rule four scale-like stipules at a node,4 but very often three stipules are present on one or on both sides of the stem, and the middle or additional one is then distinctly smaller than either of the other two (Figs. 14, 15).5 This feature reminds one of the primordial stage of the seven- or eight-membered whorls in Aspentla .6 Although nothing is known about the development of the foliar organs in Didymaea , it 1 In Benth. et Hook., Genera Plantarum, vol. ii (1873), p. 150 ; Hooker’s leones Plantarum, vol. xiii (1878), p. 55, tab. 1271. 2 That is, unique among the Stellatae. 3 K. Schumann gives in Engler and Prantl’s Pflanzenfamilien, vol. iv (1891), pt. 4, p. 3, Rubia diphylla , K. Schum., as another example of a member of the Stellatae having ordinary triangular stipules. Neither the figure he refers to nor a description of this plant has ever been given. There is, however, little room for doubt that this name (nom. nud.) is a synonym of Relbunium diphyllum, K. Schum. (apud Martium, Flora Brasil., vol. vi, pt. 6, 1888, p. 1 1 7 ; also see under this genus in the Pflanzenfamilien, vol. iv. p. 154). This plant possesses a minute triangular interfoliar stipule on either side of the node, just as in Rubia ephedroides , Cham, et Schltdl. (Linneae, vol. iii, 1828, p. 231 ; also see Martius, FI. Brasil., vol. vi, pt. 6, p. 120, Tab. xciii, Fig. 1), and R. equisetoides , Cham, et Schltdl. (1. c., p. 231; Martius, 1. c., p. 119). These three plants are, however, apparently xerophytes and have their foliar organs very much reduced ; the true leaves being scale-like, three or four mm. in length, and only a little larger than the stipules, which are similar in shape. It is possible that if the true leaves could be induced to develop to a respectable size, the stipules would also become much larger, assuming exactly the same shape as the former. It is therefore obvious that these three species are not suitable as examples of the exceptional phenomenon among the Stellatae of scale-like stipules alternating with the true leaves, which are decussate and of normal appearance. So far as the writer knows, Galium para- doxum , Maxim., G. geminifolium , F. von Muller, Asperula geminifolia, F. von Muller, and Didymaea mexicana , Hook, f., are the only members of the Stellatae showing this rare phenomenon, and of these the first-named plant produces scale-like stipules at the lower nodes only, those at the middle and upper nodes of the stem being leaf-like. Both Galium geminifolium and Asperula geminifolia are Australian plants (cf. Bentham, Flora Australiensis, vol. iii, 1866, pp. 445 and 443), and usually bear at each node two opposite, narrow leaves, and two small scale-like stipules alternating with the former (cf. K. Schumann, in Engler and Prantl, Pflanzenfamilien, vol. iv, pt. 4, p. 15 1, Fig. 48 d). In G. geminifolium , however, the stipules are occasionally well developed, either becoming leaf-like or showing a transition. This is probably due to a change in some physiological conditions (such as might l}e caused by rain, &c.) at the time when the foliar organs are developing. It may be remarked here that Velenovsky (1. c., p. 434) gives Putoria as an example of the genera whjch belong to the Stellatae and bear interpetiolar stipules alternating with two opposite leaves. However, all the leading systematists agree to refer this genus to the tribe Anthospermeae ! 4 Cf. K. Schumann, in Engler and Prantl, Pflanzenfamilien, vol. iv, pt. 4, p. 147, Fig. 47, N, o. 5 Only once has the writer met with four stipules on one side of a node ; this was in one of the specimens distributed under Pringle, Plantae mexicanae, No. 4716 (1894). In this case also, the two middle ones are smaller than the other two stipules. 6 Cf. Franke, 1. c., p. 51, Taf. i, Fig. 8. 204 Takeda. — Some Points in the Morphology of the may be presumed that each of the three stipules arises from an inde- pendent primordium. The middle stipule, as a rule, shows no definite connexion with either of the adjacent stipules. The writer has, however, found a case in which one of the two stipules on one side of a node is divided about halTway into two unequal parts (Fig. 16),1 thus suggest- ing how three stipules may have arisen from two by fission of one of them, that is, if two be taken as representing the fundamental number of stipules on each side of the node. A feature of further interest has also been found in several speci- mens examined of Didymaea , the paired stipules on either side of the stem being not uncommonly united into a single organ, with the lamina either deeply or shallowly divided (Figs. 17, 18). There is, of course, Figs. 14. 15. 16. 17. 18. Figs. 14-18. Interfoliar stipules of Didymaea mexicana, Hook. fil. All x 5. (Figs. 15 and 16 were taken from inflorescence, the others from stem.) Explanation in the text. no shadow of doubt that these single stipules with a forked lamina are produced as the result of a concrescence of two separate primordia. The question then arises whether the presence of a single stipule on either side of the stem should be considered as the more primitive type, and if this were so, whether the occurrence of two to three separate stipules on each side of the node might be regarded as the result of a fission of the original single interfoliar stipules. A comparison of the different features found in Galium and other genera with those just described above for Didymaea will bring forward the more general question : whether the six-membered whorl in the Stellatae should be regarded as having been derived from a four- membered whorl, owing to chorisis of an original pair of opposite interfoliar stipules, or whether, on the other hand, the production of only two stipules at a node is due to reduction. An examination of the seedlings may throw some light on the problem, since early stages in ontogeny often show some ancestral characters.2 1 Unfortunately, the material was unsuitable for determining whether the two midribs of this stipule are completely separate or united near the base of the organ in question. 2 It is not maintained that characters observed in early stages of ontogeny are necessarily to be Stipules in the Stellatae , with special reference to Galium. 205 According to Lubbock 1 the seedlings of Galium saccharatum and G. tenuis - simutn, both of which usually bear more than six £ leaves ’ at each node, produce only four ‘ leaves ’ (i. e. two true leaves and two stipules) at the first node, five ‘ leaves ’ at the second, and often six ‘ leaves 5 (i. e. the two true leaves and four stipules) at the third node. The same author states 2 that the seedling of Sherardia arvensis possesses only a single stipule on each side of the stem at the first three nodes, and that four stipules (two on each side) occur at the succeeding nodes. Velenovsky 3 states that in the seedling of Asperula odorata the cotyledons are suc- ceeded by a four-membered whorl, in which two stipules (smaller) are distinctly enveloped by two opposite leaves. In the next whorl there are six ‘ leaves ’, and the two opposite leaves again surround the four stipules, which are produced by a fission of those organs which corre- spond to the original two opposite stipules at the first node. He further states 4 that similar features may be seen also in Galium sylvaticum , the seedlings of which however produce four-membered whorls throughout the first year’s growth. The writer’s own observations on seedlings have revealed more instances of the same phenomenon in other members of the Stellatae. Galium murale , All.,5 always produces four £ leaves ’ at each of the lower nodes, and five or six ‘ leaves ’ at the upper nodes. Both G. setaceum , Lam.,6 and Crucianella disticha , Boiss.,7 bear four leaves at the first, and often up to the third node, and six ‘ leaves ’ at each of the higher nodes. G. Vai/lantii, DC.,8 which possesses eight £ leaves ’ at each node in the upper region of the stem, produces four 6 leaves ’ at the first, and sometimes also at the second node, but five or more £ leaves ’ at each of the nodes at a slightly higher level. G. Aparine , Linn., usually produces four £ leaves ’ at the first, and sometimes the second node, and less commonly five or six ‘leaves’ at the first node. In Crucianella angustifolia , Linn., and C. patida, Linn., the first and often the second and third nodes have a whorl of four members, while each of the higher nodes is furnished with a six-membered whorl. C. latifolia , Linn., which is another ‘six-leaved ’ species, bears four accepted as ancestral (cf. Lang, Address to the Botanical Section, Brit. Ass., at Manchester, 1915, p. 6). Any given case, relating for instance to the leaves of a seedling, must be studied in the light of as much collateral evidence as is available, and a slight presumption in favour of ancestral character may be granted on the strength of a number of examples, such as the seedlings of the gorse, certain phyllode-bearing acacias, many Conifers (cf. Veitch’s Manual of the Coniferae, ed. 2, p. 22 et seq., 1900), &c. 1 Seedlings, vol. ii (1892), p. 77. 2 Lubbock, 1. c., p. 79. 3 1. c., pp. 434-5- 4 Velenovsky, 1. c., p. 435. 5 FI. Pedemont, i, p. 8, Tab. 77, Fig. 1 (1785) ; a Mediterranean species. 6 Encyclopedic, vol. ii (1806), p. 584 ; a Mediterranean and temperate Asiatic species. 7 Diagnoses, vol. i, pt. 3 (1843), p. 25. 8 Flore de France, no. 3381 (1805); Prodromus, vol. iv (1830), p. 608. The plant is some- times treated as a variety of G. Aparine , Linn., var. Vaillantii, Koch, or a subspecies of G. Aparine , Linn.: see Flooker’s Stud. FI., ed. 3 (1884), p. 194. 2 06 Take da. — Some Points in the Morphology of the ‘ leaves’ at each of the first few nodes, or often for several nodes. Meri- carpaea vaillantioides , Boiss.,1 a small annual plant from Assyria, is described by some authors as having- four ‘ leaves ’ at each node,2 and by others six ‘ leaves’.3 This discrepancy is due to the fact that the plant actually bears four ‘ leaves ’ at each of the lower nodes, and six ‘leaves ’ to the node in the upper region of the stem. A highly interesting feature is furnished by a specimen of Asperula arvensis , Linn., which usually produces four ‘ leaves ’ at the first node, and five or six members at the second. The specimen under consideration is one of those distributed under Siehe’s Botanische Reise nach Cilicien, No. 144. One of the two stipules at the first node 20. 21. 22. Fig. 19. Double stipule of Asperula arvensis, Linn, x 5. Fig. 20. Forked stipule of Asperu la sherardioides , Jaub. et Spach. (seen from under surface, showing forked midrib and recurved margin of lamina), x 5. Fig. 21. Double stipule of Galium leiophyllum , Boiss. et Hohenack. x 5. Fig. 22. Ditto of Asperula aspera, Boiss. x 5. in this particular specimen is provided with two midribs and a notched lamina (Fig. 19), evidently showing a transition to the five-membered whorl at the next node. A specimen of Asperula sherardioides (Boiss.), Jaub. et Spach.,4 has also been found to exhibit a similar phenomenon. This Persian plant is a small annual and does not attain more than 10 cm. in height. The first node usually possesses four, but very rarely six ‘ leaves ’, while the second node has five or six leaves, or occasionally four. 1 Diagnoses, vol. i, pt. 3 (1843), p. 52. Cf. Hook. f. in Benth. et Hook., Gen. PL, vol. ii (i873), P- 149- 2 Cf. Hook. f. in Benth. et Hook., Gen. PL, vol. ii (1873), p. 149. 3 Cf. Boissier, FI. Orient., vol. iii (1875), p. 83. 4 lllustr. PL Orient., vol. i (1843) p. 153, tab. 83 ; Boissier, FI. Orient., vol. iii (1875), p. 29. The plant was originally described under the name of Crucianella sherardioides , Boiss. : Diagnoses, vol. i, pt. 3 (1843), p. 24. Stipules in the Stellatae, with special reference to Galium . 207 The particular specimen referred to above was distributed under No. 7093 of J. Bornmiiller, Iter Persicum alterum. It measures about 4 cm. high, and possesses four ‘ leaves ’ at each of the first two nodes, while the third node, which is situated immediately below the inflorescence, has six ‘ leaves ’. One of the two stipules of the second node has the lamina and midrib forked (Fig. 20), while the other is normal. It is therefore evident that, so far as our knowledge goes, in Galium and in some allied genera, the species with several to many ‘ leaves ’ at each node generally start in the seedling with a four-membered whorl, which is succeeded by whorls consisting of a larger number of ‘ leaves Correlating these facts, it appears probable that in the Stellatae the four- membered whorl (composed of two true leaves and an interfoliar stipule on each side of the stem) represents the more primitive type, while the whorls with more than four members (i. e. with more than two stipules) represent a derived type. In connexion with the arrangement of ‘ leaves ’ in the seedlings, it may be of no little interest to examine the manner in which double stipules (i. e. those with a forked or double midrib) are distributed on the stem of adult plants, and their relation to the ‘ leaves ’ at the neighbouring nodes. Among the three ‘ four-leaved ’ species of Galium above mentioned, G. gracile , in which double stipules are frequently met with, produces the stipules of this nature generally in the middle and upper regions, and only rarely near the base of the stem. If a double stipule is present at a node, it is often found that a few succeeding nodes also bear one or two stipules of the same character. In some cases, double stipules may be found at several nodes on a stem, while in the more usual cases such stipules are produced only at one or two nodes. Asperula trijida , Makino,1 furnishes us with a more interesting case. This rare species is a perennial, and occurs in certain mountainous districts of Japan. The plant was described as having ‘leaves four — rarely five — verticillate’, but it occasionally produces as many as six ‘ leaves ’ at a node. The specimens examined were gathered on Mount Ishidzuchi, in the Province of Iyo, in August, 1888, and were sent to Kew from the Tokyo Imperial University. They bear four, more usually five, ‘ leaves ’ at each of the lower nodes, and often six ‘leaves’ to the node in the middle region of the stem (vide infra). Towards the upper part of the stem, a six-membered whorl is as a rule suddenly succeeded by four-membered whorls. Those whorls which are situated in close proximity to the inflorescence have the stipules much reduced in size.2 * The arrangement of the ‘ leaves 5 in one of the specimens 1 Illustr. FI. Japan, vol. i, No. n (1891), p. 2, tab. 68; and in Tokyo Bot. Mag., vol. xvii (1903), p. 72. 2 See the excellent figures by Makino, 1. c. This feature is often met with in the Stellatae. It is especially noticeable in such a species as Asperula asterocephala , Bornm., above referred to. 208 Take da. — Some Points in the Morphology of the examined shows the following sequence : the first node of the aerial stem, which is continuous with a rhizome, has four ‘ leaves ’, the second five, the third four(P),1 the fourth five, the fifth four, the sixth four (P),1 the seventh to ninth six each, the tenth four ‘ leaves one of the two stipules having a forked midrib (Fig. 23) ; the eleventh again bears four ‘ leaves’, and one of the stipules, which are smaller than the true leaves at the same node, has a slight indication of a double nature ; the twelfth, which is giving off a branch terminated with an inflorescence, has four members with the two stipules markedly smaller than the true leaves at the same node ; the thirteenth whorl, which is at the ultimate node on the stem, consists of two true leaves (7 to 8 mm. long), and a small but normal stipule (4 mm. long) on one side, and two minute stipules (2 mm. long) on the other side of the node. From this node three peduncles have sprung, two of which have at their first (and only) node two small true leaves only, stipules being com- pletely suppressed, while the remaining peduncle bears at its lowest node a pair of true leaves and a minute stipule (1 mm. long) on one side of the stem. Another specimen examined shows a somewhat similar feature : at the first and second nodes the ‘ leaves ’ have withered and are torn off ; the third node, which is very similar in appearance to the fifth node of the first specimen above described, has five ‘ leaves ’ ; the fourth node bears four ‘ leaves ’, and one of the stipules is equal to the true leaves, while the other is broader and is provided with two midribs (Fig. 24) ; the fifth to seventh nodes are all six-membered ; the eighth node has four ‘ leaves ’, and one of the stipules of this whorl possesses a forked midrib, and is exactly the same in appearance as that delineated in Fig. 22, while the other is normal ; the ninth node is again four-membered but without any double stipule ; the tenth node is similar in every respect to the eighth node, with this difference, that the midrib of the double stipule is more deeply forked (Fig. 25) ; the eleventh (ultimate) node is also four-membered, with two stipules considerably smaller than the true leaves. At the ultimate node of another specimen the writer found two true leaves 2 and two stipules, the latter being much reduced in size, and one of them being provided with two complete midribs (Fig. 2b).3 As has already been mentioned above, the occurrence of double stipules in the case of nodes with more than four ‘ leaves ’ is very rare.4 The only instance the writer has found is presented by Asperula aspera 1 The whorl has been damaged, and consequently it is very difficult to determine this point with absolute accuracy. 2 One of these leaves is shown in Fig. 27 for comparison. 3 Since this species bears flowers with a usually three-, often four-, and rarely five-lobed corolla, a search was made with a view to detect corolla-lobes with a forked vein. Unfortunately, the result has so far been negative. A propos, it may be mentioned that Galium saxatiie , Linn., also produces pentamerous flowers very frequently. i No case in which a double stipule occurs in a whorl with more than five * leaves’ has been observed or recorded. Stipules in the Slellalae, zvith special reference to Galium. 209 (M. Bieb.), Boiss.1 This plant, which is a native of the Caucasus and Persia, usually has six ‘ leaves ’ to a node. In the upper region of the stem of one of the specimens collected by Szovits in Persia and sent to Kew from the Imperial Botanic Garden in Petrograd , the writer found a whorl consist- ing of two true leaves and three stipules, one of which is provided with two midribs, the lamina being at the same time very shallowly notched at the apex. Unfortunately, the specimen is unsuitable for examining the number of £ leaves ’ at the next nodes above and below the one just mentioned. In this plant, branches are given off from several of the lower nodes. The branches usually start with nodes which bear four ‘ leaves ’ to each, and are gradually succeeded by five- and six-membered Figs. 23-27. Asperula trifida, Makino. 23-26. Double stipules. 27. True leaf from the node at which the double stipule delineated in Fig. 26 is borne. All x 5. whorls, as the branches become elongated.2 In the same series of speci- mens the writer found two cases in which one of the stipules of a five- membered whorl had two midribs. The arrangement of the ‘ leaves ’ on one of the branches is as follows : the first node has two true leaves and two stipules, one of the latter having two midribs ; the second node is four-membered, both of the stipules being normal ; the third node five- membered, one of the stipules again having a double midrib (Fig. 22) ; while the fourth whorl is just sprouting, thus preventing the determination of the number of £ leaves * with accuracy. The examples described above clearly show that double stipules may occur in any part, but more often near the base, and also towards the apex of a stem. It may also be remarked that double stipules fall on the whole into two categories, according as they form a transition towards increase or decrease in the number of £ leaves * in a whorl. In other words, 1 FI. Orient., vol. iii (1875), p. 28. The plant was originally described as Crucianella aspera , M. Bieb. : FI. Tauiico-Caucasica, vol. i (1808), p. 107. 2 A similar feature is also found in G. A pa vine, Linn. In the seedling of this species, and also of others belonging to the Stellatae, branches are very frequently produced in the axil of the cotyledons. 2io Takcda. — Some Points in the Morphology of the double stipules may be produced in the region in which the number of ‘ leaves 5 is undergoing increase, or on the other hand reduction from node to node. It is evident that at the base of a seedling a double stipule repre- sents a stage of progression, while towards the apex of a stem it usually represents a stage of retrogression. It has already been pointed out that in the Stellatae the four-membered whorl seems to represent within the limit of possibility the most primitive type. Hence it follows that the progression generally speaking would have been from the four-membered whorls to those with six or more members by fission of the original two stipules. In the case of any ‘ four-leaved ’ species, such as Galium gracile , the occurrence of double stipules therefore indicates a step towards the production of ‘ five- leaved ’ species. A case presenting some phenomena similar to those described above (though not involving stipules) has been recorded by Groom,1 who found in Lysimachia vulgaris that dimerous nodes occurred at the base, tri- to tetramerous nodes higher up on the stem, and finally in the uppermost region of the stem dimerous nodes reappeared. In this case the transition was often accomplished by one or two double leaves at intermediate nodes. The same author2 found in Rhinanthus Crista-galli that a double leaf stood at the transitional region from cyclic arrangement of leaves to acyclic. It appears certain that in the former case (i. e. Lysimachia ) the double leaves represent on the whole in the lower region of the stem a stage of pro- gression, and in the upper region one of retrogression. In the latter case (i.e. Rhinanthus) the double leaf doubtless represents a stage in the complete replacement of two opposite leaves by a solitary leaf, by means of a con- crescence.3 In dealing with double stipules (and double leaves), it should be borne in mind that the foliar organs of this nature may sometimes be produced without relation to the general tendency of arrangement and distribution of ‘ leaves ’ in the stem. Some of the cases above described may again be referred to. In the case of Galium leiophyllum a four-membered whorl near the base of a stem is succeeded by a five-membered whorl, thus showing a tendency of increase in the number of ‘ leaves ’. At the next node the whorl is again four-membered, but one of the two stipules has tvyo midribs (Fig. 21), apparently indicating a transition towards decrease in number. The next whorl is however again five-membered, and all the nodes (about four have so far developed) succeeding this particular one are six-membered. In the examples of Asperula trifida , which have been described above in detail, the number of members at succeeding nodes does not show a regular 1 Longitudinal Symmetry in Phanerogamia. Phil. Trans. Roy. Soc. Lond., B, vol. cc (1908), p. 84 et seq. 2 Groom, 1. c., p. 106. 3 For further instances of similar phenomena see Worsdell, Principles of Plant-Teratology, vol. i (1915), p. 216 et seq. Stipules in the Stellatae , with special reference to Galium . 21 1 sequence, but repetition of increase and decrease may be seen. In such cases, double stipules may be clearly associated with either increase or decrease, or sometimes a decision on the subject may be impossible.1 At the same time it may be noticed in many cases that in alternate nodes the number of the f leaves ’ and also the nature of the stipules are more closely correlated than in successive ones. Explanations may be sought for partly in some physiological factors which either promote or retard the growth of a plant. Boodle 2 found in the seedlings of the gorse that similar irregularities often occurred regarding the distribution of the trifoliate leaves. He is of opinion that this is at any rate partly due to variation in some physiological conditions during the growth of the seedlings. On the other hand, Groom’s suggestion 3 regarding a similar phenomenon exhibited by Lysimachia vulgaris , that ‘ morpho- genous impulses are transmitted along the orthostichies may give the explanation of the alternating number of members at succeeding nodes described in some of the above examples. Turning to the case of a seven- or eight-membered whorl, it may be held that the middle stipule of the three (on one or both sides of the node) has originated by fission from one of the two adjoining stipules. Since the middle stipule usually takes exactly the median position and assumes the same size as the others, its relationship to the one or the other of the two adjoining stipular members cannot be determined. This occurrence may conveniently be termed a congenital fission (chorisis, dedoublement) in con- trast with the phenomenon known as congenital fusion. If this congenital fission once starts in any species it may be repeated more than once, resulting in the production of as many stipules as space will permit on either side of the node. Hence, we more often find the pleiomerous whorls in the species with narrow ‘ leaves ’ and a comparatively thick stem, such as Galium verum. In the case of the pleiomerous whorls, stipules are known to occur as many as five on either side, and they probably develop from as many primordia as there are stipules, without showing any connexion with one another. A parallel case may be found at the cotyledonary node in many of the Conifers and a few Dicotyledons, in which the two original cotyledons have undergone chorisis, thus giving rise to a polycotyledonous condition. In these cases the increase in the number takes place as a rule without any transitional stages, and double cotyledons are not frequently in evidence. The double cotyledons recorded by Groom 4 for Acer P seudoplatanus 5 and 1 Cf. Worsdell, 1. c., p. 238. 2 On the Trifoliate and other Leaves of the Gorse (UZex europaea, L.). Ann. Bot., vol. xxviii, (193:4), p. 527 et seq. s Groom, 1. c., p. 86. 4 1. c., pp. 102-3. 5 Also see Worsdell, he., p. 215, FI. XIX, Fig. 4 a. 212 Takeda. — Some Points in the Morphology of the Fraxinus excelsior, by Miss Chick 1 for Torreya Myristica, and by Hill and de Fra in e 2 for Cupressus torulosa , Abies sibirica , Finns montana var. gallica , P. contort a var. Murray ana , Araucaria Cunning hamii , &c., are therefore worthy of notice. From the evidence given above, the conjecture may be justified that the direct ancestors of the Stellatae possibly had two stipules at each node, and that each one of these two stipules had been derived by means of congenital concrescence from two separate stipular organs, the earlier ancestors of the Rubiaceae as a whole being assumed to have possessed four stipules at each node. Hence, it follows that the species of Galium (and also of any other genera of the Stellatae) having four-membered whorls would on the whole represent the most primitive type in that particular tribe. Whorls with six foliar organs would thus be regarded as a more advanced type among the Stellatae, but at the same time as representing a reversion to the condition found in the ancestors of the Rubiaceae. It is suggestive that among living Rubiaceae cases of concrescence of stipules occur very fre- quently, and in certain genera, such as Palicourea , Cephaelis , &c., some species have four separate stipules, while others have two either partially or completely fused (connate) ones, thus resulting in the production of four- membered whorls. Regarding the stage of evolution at which the stipules assumed the characters of true leaves in the Stellatae, it is difficult to come to a decision. However, it is beyond all doubt that the original type of the stipules in the Stellatae was scale-like, and that the leaf-like stipules have evolved from that type, probably in relation to certain physiological necessities. It may therefore be considered that in this respect A idymaea mexicana stands nearest the prototype of Stellatae, as Goebel maintains,3 and that Galium paradoxum represents the most primitive species of its genus. As to the question whether Galium ge mini folium and Asperula geminif olia , both of which usually bear two scale-like interfoliar stipules at each node,4 furnish us with further examples of primitive species is rather doubtful. The pro- duction of scale-like stipules in these two species may probably have been brought about by the circumstances of their xerophytic habitat, thus indi- cating reduction rather than a retention of the primitive state. Summary. 1. In Galium and other allied genera, each stipule as a rule develops from a single primordium. 2. Fairly frequently, and particularly in four-membered and rarely in 1 The Seedling of Torreya Myrisiica. New Phytologist, vol. ii (1903), p. 85. 2 Seedling Structure of Gymnosperms, II. Ann. Bot., vol. xxiii (1909), p. 221, PI. XV, Fig. 4 b. 8 Organography of Plants (Engl, ed.), pt. 2 (1905% p. 371. 4 Cf. the present paper, p. 203, foot-note. Stipules in the Stellatae , with special reference to Galium . 213 five-membered whorls, stipules may be found which have been produced as the result of a coalescence of two primordia. Stipules of this kind ( = double stipules) possess either a forked midrib or two separate midribs, the apex of the stipules being at the same time more or less two-lobed. 3. Double stipules may occur near the base, or towards the apex, and more rarely in the middle region of a stem. They may in certain cases represent a transition towards increase in number, in other cases a stage leading towards numerical decrease of the organs. Double stipules forming examples of both of these cases are occasionally found on one and the same stem. Sometimes, however, a decision on this subject is hardly possible. 4. In the seedlings of several species examined of the genera Galium , Asperula , Crucianella , and Mericarpaea, the node or sometimes a few nodes succeeding the cotyledonary node as a rule bear a four-membered whorl, consisting of two true (opposite) leaves and two stipules alternating with the former. In the higher region of the stem the number of members in a whorl may in some of the species examined be gradually increased up to eight. 5. The four-membered whorl is considered to represent the primitive type, at the same time indicating the probable character which prevailed among the direct ancestors of the Stellatae. 6. The six-membered whorl, which probably represents the type that characterized the ancestors of the Rubiaceae, is in the Stellatae regarded as having been derived from a four-membered whorl by complete fission (dedoublement) of the two stipules into four. 7. Whorls with more than six members have no doubt originated by repeated fission of the original two stipules. 8. Didymaea mexicana) Hook, fil ., which bears two opposite leaves, and from two to often six, or rarely seven, scale-like stipules at each node, is presumed to approach the prototype of the Stellatae. And in this species also the four-membered whorl very probably represents the most primitive type. 9. Galium paradoxum , Maxim., which bears two leaves and two scale- like stipules at the lower nodes and two true leaves and two leaf-like stipules in the higher region of the stem, is believed to be the most primitive species of the genus in this respect. The present investigation has been carried out in the Herbarium, Royal Botanic Gardens, Kew. All the specimens, except those of Galimm saxatile and G. Apariney used for the investigation are preserved in the Herbarium. The writer has great pleasure in expressing his sincere thanks to Sir David Prain, C. M. G., C. I. E., for the privilege of working in the Herbarium and also using the Library. The writer also takes this opportunity of thanking Mr. L. A. Boodle for the constant encouragement, valuable help, and the interest he has taken during the progress of the investigation. Q 2i4 Takeda. — Morphology of the Stipules in the Stellatae. Postscript. After the above had been written, some of the seedlings of several members of the Stellatae raised in a hot-pit in the Royal Botanic Gardens, Kevv, were found ready for an investigation. The result of an examination regarding the number of members at lower nodes in the stem is as follows : 1. Galium Aparine , Linn. The seedlings bore a four-membered whorl at the first node succeeding the cotyledons. Only one specimen was found to have produced a five-membered whorl at the first node. Thus, the result corresponds to what has been observed in the field, as described in the fore- going pages. 2. Galium Mollugo , Linn, x G. verum , Linn. So far, the seedlings have produced three nodes, at each of which four ‘ leaves ’, consisting of two true leaves and two stipules, are present. The stipules are as a rule similar to the true leaves, but a scale-like stipule occasionally stood opposite a normal, leaf-like stipule. 3. Asperula galioides, Bieb.1 So far, four nodes have developed in the seedlings examined. Four ‘ leaves' were present at each of the four nodes. The stipules are sometimes smaller than the true leaves in the same whorl. 4. Asperula tinctoria , Linn.2 So far, three nodes have been produced, each of which is four-membered. The stipules are distinctly smaller than the true leaves at the same node, often assuming a scale-like appearance. 1 According to de Candolle (Prodr. Reg. Veg., vol. iv, p. 585) this species bears six to eight ‘ leaves ’ to the node. 2 According to de Candolle (1. c., p. 582) the whorls in the lower region of the stem are six- membered, those in the middle region are four-membered, while those in the apical region are two- membered. Studies in Seed Germination. The Genus Marah (Megarrhiza), Cucurbitaceae. BY ARTHUR W. HILL. M.A., F.L.S., Assistant Director , Royal Gardens , Kew. With Plate V and two Figures in the Text. SOME dry prickly fruits, which appeared to belong to a species of Echinocystis, were received at Kew in 1908 from Mr. F. R. S. Bal- four of Dawyck, having been collected by him in the arid Sierra Nevada region of California. The seeds, which were unlike those of any known species of Echinocystis, quickly germinated and were found to exhibit a type of germination similar to that of the Californian ‘ Big root Megar- rhiza calif ornicas described and figured by Asa Gray1 and Darwin.2 A further supply of seeds of this plant and also seeds of allied species from California were obtained through the kindness of Mr. Balfour and other correspondents, and an examination of their mode of germination has yielded some results of interest. In order to ascertain the botanical identity of the seeds originally received, the specimens of Echinocystis and allied genera were obtained on loan from the Smithsonian Institution and examined by Mr. S. T* Dunn.3 The seeds were found to belong to the genus Marah , Kellog {Megar- rhiza (Ton*.), S. Wats.), which is morphologically quite distinct from Echinocystis , Coignaux. In this remarkable genus the underground tubers are said often to reach the size of a man’s body.4 It is of interest to find that the systematic grounds on which Marah was separated from Echinocystis are supported and strengthened by morpho- logical considerations, since not only is the underground tuber a feature peculiar to Marah , but the type of germination is essentially geophilous, 1 Asa Gray : American Journal of Science, vol. xiv, 1877, pp. 21-4, and the Botanical Text Book, Pt. I, Structural Botany, ed. 6, 1879, pp. 20, 21. 2 Darwin, C. : The Power of Movements in Plants. 1880, pp. 81-3. 3 Dunn, S. T. : The Genus Marah , in Kew Bulletin, 1913, pp. 145-53, and p. 238. 4 Hall, H. M., in Univ. Calif. Publ. Botany, vol. i, 1902, p. 19. [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] Q a Hill. — Studies in Seed Germination. 216 resembling that of some Monocotyledons, while In Echinocystis the cotyle- dons are epigeal and the germination of the seed is of the usual cucurbita- ceous type. Dunn, in his revision of the genus Mar ah, enumerates eleven species, and seeds of five of these have been received at Kew.1 The general type of germination appears to be similar throughout the genus, though differences in detail may occur. The least specialized type of germination has been found to occur in M. fabaceus , and M. horridus , the species first examined at Kew, is the most highly developed, whilst M. macrocarpus, M. muricatus , and M. micranthus more nearly approach the condition shown by M. horridus. The plant figured by Asa Gray and Darwin under the name Megar- rhiza catifornica is probably a seedling of Mar ah macrocarpus, Greene, though the herbarium specimens referred to Megarrhiza catifornica , Torn and S. Wats., have been found to belong to Marah fabaceus , Greene. Asa Gray 2 was the first to observe and describe the peculiar germi- nation of the seeds of Megarrhiza catifornica, and subsequently Darwin 3 made them the subject of careful study, and added several important details to the description given by Gray. Darwin’s account is detailed and exhaustive, and the only excuse for the present contribution lies in the fact that the germination of the seeds of other species has been studied and has rendered possible a comparative account of seed germination in the genus. Marah fabaceus, Greene. — In this species the two cotyledons are fused together towards the base and never leave the seed, and their petioles are united to form a short tube. On germination, the petiole tube elongates, and the plumule and radicle are carried out of the seed. The tube is only about 6 mm. long, but on emerging from the micropyle it bends downwards, and the radicle then breaks through the tip and pene- trates deeply into the soil, leaving the outer portions of the end of the petiole tube as a frayed edge, much in the same way as occurs with the hypocotyl of the Radish ( Raphanus ) (PI. V, Fig. 1). When the root has become well established by sending out lateral roots, the plumule breaks through the petiolar tube opposite the point of emergence of the root, and in time appears above the surface of the soil. At the same time a hypocotyledonary tuber begins to be formed by the re-storage of part of the reserve materials contained in the fleshy cotyledons (Fig. 2). This underground tuber increases in size as a result 1 Seeds were kindly sent by Mr. H. F. Shorting, Huntington Beach, California, Professor H. M. Hall, College of Agriculture, Berkeley, California, Mr. W. W. Whitney, San Diego, Cali- fornia, Dr. J. N. Rose, Smithsonian Institution, Washington, and Mr. F. R. S. Balfour of Dawyck. 2 Amer. Journ. Science, vol. xiv (1877), pp. 21-24, with Figs., and the Botanical Text Book, ed. 6, Pt. I, Structural Botany (1879), PP* 2°> 2T> Figs. 43, 44. 3 Darwin : Power of Movement in Plants, 1880, pp. 81-3, and Fig. 58 A. Hill. — Studies in Seed Germination. 2 1 7 of the metabolic activity of the plumule and radicle, and eventually assumes very large proportions. Darwin describes the occurrence of absorbent hairs on the petiolar tube of Megarrhiza californica , and they have been noticed in all those species with a well-developed cotyledonary tube, but in Marah fabaceus no such absorbent hairs are present. This is no doubt due to the shortness of the tube and to the rapidity with which the root bearing its own root-hairs develops. Except for the fact that the petioles of the cotyledons are fused to form a short tube, the type of germination exhibited by this species may be compared with that of Aesculus, where a certain elongation of the cotyledonary petioles takes place to carry the radicle and plumule out of the seed. M. micranthus and M. macrocarpus. — The germination of the seeds of M. muricatus , Greene, and M. micranthus , Dunn, is so similar to that of M. macrocarpus (M egarrhiza californica ) that there is little to add to Darwin’s account. In M. micranthus and M. macrocarpus the seeds are orbicular or ovoid, •about i*5 cm. long by 6 to 8 mm. broad, and the cotyledons are united towards the base. The petiolar tube in M. micranthus reached a length of 2 to 3 cm., and split for some distance on the emergence of the plumule. The hypocotyledonary tuber was ovoid, and quickly emitted roots. In M. macrocarpus the petiole tube grows as much as 6 cm. vertically into the ground, and it is probable that in its native home it may attain a greater length and carry the plumule deep into the soil as a protection against drought (PI. V, Fig. 9), Under artificial greenhouse conditions the plumule bursts through the tube soon after the root has become well established, but under natural conditions it is probable that there may be a considerable resting period after the plumule has been safely deposited deep down in the earth, whilst the tuber is developed at the expense of the reserves .stored in the seed, and until conditions become favourable for the development of the climbing shoot. M. muricatus. — A number of seeds of M. muricatus were received from California, but only one germinated. In this species, the seed is flattened and somewhat orbicular-rhomboid in outline, i*8 to 2*o cm. in diameter. The cotyledons are only 1 to 2 mm. thick, and free except, as in other species, near the base, where they are definitely fused together. A long petiolar tube furnished with absorbent hairs is formed, and the plumule eventually escapes owing to the splitting apart of the component members of the tube at its base. The tuber in this species appears to be partly epicotyledonary in origin at the commencement, though no doubt later the hypocotyl shares in its development. Owing to lack of material, the tuber development could not be followed. 218 Hill. — Studies in Seed Germination. M. horridus. — The species whose germination has been most carefully studied is M. horridus , Dunn, seeds of which were collected in the Kaweah River Valley, California, at 6,000 ft., by Mr. F. R. S. Balfour. Of the region it inhabits, Mr. Balfour writes: ‘ The district for five or six months of the year is under very deep snow, but from May to October there is cloudless sunshine without a drop of rain.’ Nothing is known of the germination of the seed under natural conditions, but it is probable that the seed remains buried under the snow, and may be carried into the ground during the winter, germination taking place in the late spring. Judging from observations made under greenhouse conditions, germination when started would proceed rapidly, and the plumular shoot would no doubt soon appear above ground. This shoot, however, in its first year would probably only have a comparatively short existence. The seeds in M. horridus are ovoid with a thick testa, and measure about 3 75 cm. long. At the broad end they are about 1 cm. in diameter across the cotyledons, and at the narrow end towards the apex the dia- meter is about i*8 cm. The cotyledons are thick and fleshy, somewhat hollowed and fused together at the base. Each cotyledon is about 4 mm. thick, and is stored with aleurone and oil. On germination, the testa bursts by a longitudinal fissure commencing at the micropyle and gradually extending along the pronounced ridge encircling the seed, until by the swelling of the cotyledons it is split into two portions. The petiolar tube carrying the plumule and radicle at its apex quickly elongates and bends over, growing vertically downwards into the soil (PL V, Fig. 3). The tube is covered with a mat of fine, some- what woolly, unicellular hairs, to which particles of soil adhere after the manner of root-hairs ; these hairs in fact, until the radicle develops from the end of the petiolar tube, apparently perform the function of root- hairs and absorb the moisture required for the development of the embryo (Figs. 3-7). The tip of the petiolar tube ends abruptly, in the early stages of germination, in a short brown conical tip which is surrounded by a loose and more or less detached flange or root-cap. The flange may in part represent some portion of the inner seed-coat carried away by the tip of the petiolar tube, and is also no doubt part of the cap of the radicle itself whose tip forms the actual apex of the cotyledonary tube (PL V, Figs. 3, 4, and 6). While the petioles are elongating and carrying the plumule and radicle into the ground, very little development of either organ takes place (Fig. 5), but after an appropriate depth has been reached the radicle begins to elongate, develops root-hairs, and quickly grows down into the soil and sends out lateral roots, anchoring the plumule firmly in the ground (PL V, Fig. 7). About this time a slight swelling becomes noticeable at the point Hill . — Studies in Seed Germination. 219 where the root and the tube of the petioles are united, and is found to indicate the position of the plumule and the commencing tuber. With the development of the root, the petiolar absorbing hairs dry up and cease to function. The tube of the petioles in the specimen grown at Kew was from 5 to 7*5 cm. in length, but under natural conditions it would no doubt penetrate to a greater depth in the soil. The plumule now commences to elongate and breaks through the petiolar tube near the base, the tube splitting under the pressure into its component halves. The split- ting apart of the petioles extends gradually upwards towards the cotyledons, until at length only a small portion of the tube is left intact (cf. Fig. 8). The apex of the plumule is sharply bent over at the tip, and this strong curvature of the shoot apex is maintained by the mature shoot and may perhaps be attributed to the fact that the plant is a scrambling climber aver and among bushes. With the growth of the plumule and its appearance above ground, accompanied by the production of green leaves, the tuber begins to increase in size, and in particular to broaden at the apex. As a result the cotyledon petioles each become split into three portions which, owing to the continued expansion of the tuber, gradually get separated somewhat widely apart, and the young tuber when dug up is seen to be suspended from the seed by six separate strands, the split petioles of the two cotyledons (PL V, Fig. 9). The base of each strand is swollen out and passes gradually into the tuber, which in some cases may appear to be composed of six semi-independent portions. The young tuber has also been found to be deeply cleft into two from below, the two halves roughly corresponding to the two cotyledons and having been supplied by the materials stored in each (cf. Fig. 9). Within the bases of the cotyledon petioles, the tuber develops a some- what flattened-oval area with the shoot rising from the midst, and on this flattened surface several dormant buds are developed, which no doubt would give rise to aerial shoots should occasion require. Internal Structure of Petioles. An examination of the petiole tube in transverse section shows usually six vascular bundles, three belonging to each petiole, and if the section be of a petiole tube which has already split into its component portions it is noticeable that the cotyledon petiole shows grooves which correspond to the spaces between the vascular bundles (Text-fig. 1). As the tuber in- creases in size, these grooves are seen to be the planes of weakness along which the petiole will split into strands, as already described, so that each strand will contain a single vascular bundle. It is of interest to notice that owing to the splitting apart of the petiole tube into its component halves, the edges formerly in union must, when free, be furnished with a ‘ false ’ epidermis. No new 1 false ’ epidermis,. 220 Hill ’ — Studies in Seed Germination. however, appears to be formed on the splitting up of the petioles into strands, but apparently a kind of absciss layer of suberized tissue is laid down, along which the separation takes place. With the friction of the Text-fig. i. Marali horridus. Petiole oi one cotyledon from the split portion of the petiole tube in transverse section, showing the commencement of the further splitting into three strands, each of which will have a vascular bundle. The outer cortical tissues have been worn away. Text-fig. 2. M. horridus. Unit strand of split petiole with active pericycle. soil and increasing age of the seedling, the outer tissues of the petioles cease to function, and become somewhat disorganized and form an irregular layer of suberized cells. Hill. — Studies in Seed Germination. 221 When the petioles have become split into their several strands, it is found that the vascular bundle of each is surrounded by a band of peri- cyclic tissue of recent formation, and in an active state of division (Text- fig. 2). This pericyclic periderm gives rise to suberized tissue on the outer side and cortical tissue on its inner side, and the line of communication between the seed and the young plant is thus maintained for a considerable time during the seedling’s development. The vascular bundle contains a small number of lignified elements, but consists mainly of small sieve-tubes and phloem parenchyma. Summary. 1. On the germination of the seed in Mar ah, the petioles, which are fused together to form a tube, grow out and carry the plumule and radicle into the ground. 2. In M. fabaceus the tube is very short and the germination is almost normal, but in M. horridus and other species the petiole tube elongates considerably and is furnished with absorbent hairs. 3. The radicle breaks through at the base, and later the plumule penetrates the side of the petiole tube and grows above the soil. Even- tually a hypocotyledonary tuber is formed, which may become very large. 4. In M . horridus the petiole tube first splits into its two component halves and then, owing to the growth of the tuber, into six separate strands, each of which is furnished with a vascular bundle. EXPLANATION OF PLATE V. Figs. 1 and 2, Marah fabaceus. Figs. 3-7 and 9, M. horridus. Fig. 8, M. macrocarpus. Fig. 1. Marah fabaceus. Early stage in the germination of the seed, showing the short petiole tube (/.), with its tissues frayed and the radicle bent downwards. The plumule has not yet broken through the petiole tube. t. testa ; c. cotyledons. Fig. 2. An older seedling. The plumule (j1.) has broken through the tube at the point of curva- ture of the tube and radicle. Fig. 3. M. horridus. A seed with a portion of the testa removed, showing the petiole tube with a cap ( 'x .) at its apex. The tube measured 1*5 cm. in length and was furnished with a mass of absorbent hairs. (Slightly mag.) Fig. 4. The flange-like cap ( x .) at the apex of the petiole tube. (Mag.) Fig. 5. The apex of the petiole tube in diagrammatic section, showing the undeveloped plumule (j.), radicle ( V .), and the epidermal hairs. Fig. 6. The apex of the tube, showing the cap. (Mag.) Fig. 7. An older seedling in which the petiole tube has elongated and the root developed. The point of junction of petiole tube and root is marked by a slight swelling. The testa has been removed. The whole length of tube and root is 7-5 cm., the tube measuring 3*5 cm. (Nat. size.) 222 Hill. — Studies in Seed Germination. Fig. 8. Jlf. macrocarpus , Dunn. A seedling in which the plumule ( s .) has burst through the elongated petiole tube and is pushing up to the surface of the soil with tip bent over. The com- mencement of the hypocotyledonary tuber (b.) is noticeable. The petiole tube is covered by fine hairs. The testa has been removed to show the two partially fused cotyledons (c ). (Nat. size.) Fig. 9. M. korrtdus. An old seedling with well-developed plumular shoot (s.) and hypocoty- ledonary tuber which is still in connexion with the seed by means of the petiole tube. The tube has split first into its two component halves, and each petiole has then split into three strands. The tuber is deeply lobed below and on the upper surface bears adventitious buds. (Nat. size.) Studies in Permeability. III. The Absorption of Acids by Plant Tissue. BY MILDRED HIND. With eleven Figures in the Text. IN the first of these studies it has been shown that cells of potato tuber absorb hydrogen ions from aqueous solutions of hydrochloric acid with considerable rapidity. In this paper an account is given of more detailed experiments made to determine to what extent this rapid absorption is a characteristic of acids in general. Earlier work (4) appears to indicate that with strong acids it is the concentration of the hydrogen ion which is the determining factor of the action of acid on plants, while with organic acids the anion is stated to influence the action (6). So far, the methods which have been used have given little information as to the extent to which acids are actually absorbed by plant tissue, nor have these methods yielded results of any great accuracy. In the experiments recorded in this paper the methods of physical chemistry advocated and described in previous papers of this series have been used. Again, so far there has not been produced much evidence as to the substances which are responsible for the absorption of acids. For instance, Czapek (1, 2, 3) seems to be of the opinion that the action of acids on the fat emulsions of the plasma is responsible for the effects produced ; on the other hand, Loeb (5, 7), Pauli (11), and Osterhout (9) think that protein substances are active in many of the phenomena of permeability. Some experiments have accordingly been made in the course of this work in regard to the action of acids on various plant substances, but however suggestive the results of these experiments may be, it is of course impossible to draw from a few isolated experiments any definite conclusions in regard to the structure and function of cell membranes. This work was undertaken at the suggestion of Mr. Stiles and Mr. Jorgensen, whom I would here thank for their advice during its progress. [Annals of Botany, Vol. XXX. No. CXV1II. April, 1916.] 224 Hind. — Studies in Permeability . III. Methods. 1. Electrical Conductivity Methods. — By measuring the change in electrical conductivity of a solution surrounding plant tissue, some idea of the effect of the solution on plant cells can be obtained. If the substance in solution enters without harming the cell and causing an exosmosis of the cell-contents, a decrease in conductivity will be the result ; whereas if the plasma-membrane is affected in such a way that the substances within the cell diffuse out, the conductivity of the external solution will be increased on account of this. The change in conductivity in such cases will therefore be the resultant of the diffusion of the external solute into the cell and of exosmosis of electrolytes from the cell. These experiments were carried out in the same way as the electrical conductivity experiments described in the first of these studies. Twenty discs of potato tuber were put in a stoppered bottle containing icoc.c. acid, the conductivity of which had been measured. The conductivity of the acid solution was then taken at intervals. Dilute solutions of each acid in various concentrations were used. The experiments were all done in duplicate; they were carried out at constant temperature — i8°C. — as it has already been shown that temperature exerts a considerable influence on the rate of absorption (13). The curves were obtained by plotting the increase or decrease in electrical conductivity against the time. 2. Volumetric Analytical Methods. — In order to measure the acidity of the external solution after it had been in contact with the potato discs for different times, the solutions were poured off and titrated with standard alkali. No reliable results, however, were obtained, owing to the great dilution of the solutions, and possibly to the diffusion out from the cells of substances which interfered with the indicators. 3. Electrometric Method. — The method of experimentation involving the use of electrical conductivity can only give a rough approximation to the actual course of absorption of electrolytes, owing to the diffusion out of electrolytes from the cell, which takes place with most substances and even to some extent in distilled water. By measuring the concentration of the hydrogen ions in acid solutions before and after they have been in contact with potato discs, the actual rate of absorption of the hydrogen ion can be measured. Experiments were carried out in the way described in the second paper of this series (13). The hydrogen-ion concentration was calculated in terms of the original pure acid solution. The curves were obtained by plotting the relative hydrogen-ion concentration against the time. The bottles containing the discs and acid were kept at a temperature of 180 C. throughout the experiments. Hind. — Studies in Permeability. III. 225 This method gives no definite information as to the absorption of the anion. Whether the acid is absorbed as such, or whether the kation and anion enter the cell at different rates, is a question at present unsolved. The work of Pantanelli and Sella (10) indicates the possibility of this latter alternative with its resulting complications. The Results. I. The Absorption of Acids by Potato Tissue. Hydrocldoric Acid. Series 1. — The rapid absorption by plant cells of the hydrogen ion of N hydrochloric acid in a concentration of has been indicated in previous 1000 r Time in Hours Fig. 1. Potato in Hydrochloric Acid. papers (12, 13). In the present investigation a wide range of acid con- N q N centrations was employed, varying from to 1 1000 50000 The curves shown in Fig. 1 indicate that a marked decrease in the electrical conductivity of the acid solution occurs with all strengths of acid except the very dilute ones. In all cases examined, the conductivity rose after a time. These curves indicate that in all cases a rapid absorption of acid by the potato 226 Hind.— Studies in Permeability. III. cells takes place, but in the case of the lowest concentrations of acid the increase of conductivity due to the electrolytic exudate from the cells more than counterbalances the decrease, owing to removal of hydrogen ions by the potato tissue. This influence of the exudate on the conductivity curves makes itself obvious after a time in the case of the higher strengths of acid. Time in Hours Fig. 2. Potato in Nitric Acid. chloric acid was obtained by the electrometric measurement of the change in hydrogen-ion concentration of the solution. Experiments were carried N N N out with acid of concentrations In all cases there was IOOO 2COO 5000 a rapid absorption of hydrogen ions, as already indicated for the case of N . . r , . acid in previous papers of this series. IOOO 227 Hind.— Studies in Permeability. Ill . Nitric Acid. Series 3.— Conductivity measurements were made of the external solution, as in Series i, when potato discs were immersed in nitric acid N N N N solutions of concentrations , , , . 1000 2000 3000 10000 The accompanying- figure (Fig. 2) shows that the result with nitric acid is similar to that obtained with hydrochloric acid. Series 4.-— Experiments made with the electrometric method of measurement, to determine the rate of absorption of the hydrogen ion from N Fig. 3. Potato in Nitric Acid. 1000 nitric acid solutions, show that the kation is absorbed rapidly, as in the case of hydrochloric acid. The curve shown in Fig. 3 indicates the rate of absorption of the hydrogen ion when N nitric acid was used. 1000 228 Hind . — Studies in Permeability . III. Sidphuric Acid. Series 5. — Similar experiments were made with sulphuric acid as with hydrochloric and nitric acids. The change in conductivity of the external N N N N N solution, when solutions of strengths — , — , — , — , and were used, are shown graphically in Fig. 4. It will be observed that the general results are similar to those given with hydrochloric and nitric acids. Oxatic Acid. Series 6. — The absorption of oxalic acid by potato tissue was investi- N N N gated in the cases of solutions of concentrations , , . The 53 500 1000 5000 change which takes place in the conductivity of the external solution is Hind. — Studies in Permeability. IIL 229 similar to that in the case of the strong mineral acids, but as Fig. 5 indicates the decrease in conductivity is less, which may be due to a smaller absorp- tion of hydrogen ions, or to the greater exosmosis of electrolytes. > =3 T3 C o O S-oooi L. O _a > LtJ (D C/5 CO CD £_ O c •0002 •0003 TA. 2000 vy. \000_ ty 600 Time in Hours Fig. 5. Potato in Oxalic Acid. Formic Acid. Series 7 and 8. — In the case of formic acid considerably different results were obtained. Conductivity measurements of the external solution N N N were made with acid of original concentration — , , . In all & 500 1000 ’ 2000 cases the conductivity showed at first a slight decrease, which speedily gave place to a rapid increase, this increase being greater the higher the con- centration of the acid (cf. Fig. 6). Measurements of the actual concentration of the hydrogen ion were N made in the case of acid. 1000 Fig. 7 indicates the course of thfe absorption. It will be observed that during the first few hours the hydrogen ion is absorbed at almost the same rate as the hydrogen ions of nitric acid. After a time, however, a remark- able change occurs. This consists in the concentration of the hydrogen ion of the external solution increasing, a phenomenon which was never observed in the case of the inorganic acids. R 2 30 Hind . — Studies in Permeability. Ill . o co co CM Time in Hours Time in Hours Fig. 6. Potato in Formic Acid. . N _ . . Fig. 7. Potato in — — Formic Acid. 1000 Hind . — Studies in Permeability . III. 231 Acetic Acid. Series 9 and 10. — The next acid of the series, acetic acid, was used in N N N N concentrations — , , , , It will be observed from Fig. 8 500 1000 2000 5000 & that the conductivity of the external solution changes in the same way as in the case of formic acid. In Fig. 9 is shown the rate at which the hydrogen ion is absorbed by N potato tissue from acetic acid. The curious rise in concentration of 1000 the hydrogen ion, after a few hours of continuous decrease, is shown again here, as in the case of formic acid. R 2 232 Hind.— Studies in Permeability. HI. The experiments described above show that considerable difference exists between acids, even in dilute solution, as regards their action on plant cells. The mineral acids on the one hand cause much less exosmosis, and presumably therefore produce much less injury than the two fatty acids employed on the other. Oxalic acid appears to occupy an intermediate position. The earlier work referred to in the introduction to this paper is not therefore absolutely confirmed. Support is given to the contention that with organic acids the anion influences the action on plant cells, but, having regard to the different results with hydrochloric and nitric acids, it must be supposed that even with these the anion also has an effect. However, the acids containing ‘ nutrient ’ ions such as S04 and N03, as well as hydro- chloric acid, are sharply marked off from the fatty acids in respect of their power of producing exosmosis from the tissues. A further conclusion which may be drawn from the curves given in this section, is that the rate of absorption of an acid depends on its Hind. — Studies in Permeability. III. 233 concentration, as has already been shown in the second paper of this series (13). The experiments recorded in which the electrometric method of measurement was used show that an actual absorption of hydrogen ions does take place with all the acids examined. With the mineral acids on the one hand, and the fatty acids on the other, there is, however, a very noteworthy difference. In the case of the former there was a continuous absorption of acid as far as the experiments were carried ; with the latter this continued reduction of the acidity of the external solution gave place after a time to a marked increase in the acidity. A discussion of this result follows later in this paper. II. Experiments with Living Plants of Vicia Faba. In order to determine whether the roots of living plants absorb acids in a similar way to potato cells, some experiments were made with Bean plants ( Vicia Faba). Seeds were germinated in sawdust, and the young seedlings transferred to water-culture solutions, so that the roots of the experimental plants should be uninjured. Before an experiment the roots of the plants were well washed with distilled water, then the plants were placed with their roots in bottles containing 100 c.c. of acid. N Series 11. In this series, Bean plants were used with nitric acid. r 1000 Both the conductivity and the hydrogen-ion concentrations of solutions were measured after various intervals. Fig. 10 shows the change in conductivity, and Fig. 11 the rate of absorption of the hydrogen ion. These results indicate that the absorption of acid by living Bean plants takes place in exactly the same way as the absorption by discs of potato tuber. III. On the Part played by proteins in the Absorption of Acids by Potato Cells. In order to obtain some information as to which of the various sub- stances present in the cells of the potato are responsible for this rapid absorption of hydrogen ions, potatoes were ground up with sand and subjected to pressure. The expressed liquid so obtained was filtered off. The filtrate contained sugar and proteins, but no starch. The hydrogen- ion concentration of the liquid was measured, and was found to be only slight. N There was now added to 100 c.c. of an acid solution 1 c.c. of the 1000 filtrate, and the hydrogen-ion concentration of the resulting liquid measured. The hydrogen-ion concentration had decreased considerably. An attempt 234 Hind. — Studies in Permeability . Ill . was made at a partial separation of the colloidal and crystalloidal substances in the expressed liquid, by filtering under pressure in Bechold’s ultra- filtration apparatus through a 3 per cent, collodion filter. The filtrate so obtained would contain the crystalloids, but a large part of the colloids would be left behind on the filter. When 1 c.c. of the filtered solution was now added to 100 c.c. acid, there was produced a much smaller decrease in the hydrogen-ion concentration of the acid ; thus it appears that the decrease in hydrogen-ion concentration, produced by the crude expressed liquid, is probably due to the action of the colloidal substances in the 1000 potato extract. Formic, acetic, and nitric acids wrere used in these experi- ments. The proteins in some of the solution obtained from the potato were precipitated by means of colloidal ferric hydroxide, and the solution then filtered. 1 c.c. of the filtrate added to 100 c.c. acid produced no effect on the hydrogen-ion concentration of the acid. As the removal of protein from the solution also removes its power of reducing the hydrogen-ion concentration of acid solutions, it suggests that the absorbing substances of the living cells may be proteins. It is known that proteins react chemically with acids, and will therefore affect their hydrogen-ion concentration. A solution of peptone in water was therefore prepared, and 1 c.c. of it 235 Hind . — St tidies in Permeability . ///, N added to 100 c.c. of acid. There was immediately produced a marked lOOO x decrease in the hydrogen-ion concentration of the acid, very similar to that produced by i c.c. of potato extract. To obtain further proof that in the case of the liquid obtained from the potato it is the protein that affects the acid, the proteins of the potato were extracted. These were found to act in the same way, in regard to the acid, as did the peptone solution and also the liquid expressed from potatoes. It has been thought by some investigators that lipoid substances play an important part in the absorption of substances by the cell, and lecithin has been suggested as particularly active in this regard. A colloidal solution of commercial lecithin in water was prepared, and i c.c. of this was added to ioo c.c. of N acid ; measurements before and after the addition IOOO 236 Hind. — Studies in Permeability. III. of the lecithin showed that it had little effect on the hydrogen-ion concentration of the acid. Discussion. The results of the experiments described in this paper lead one to conclude that acids in dilute solution, or at any rate their hydrogen ions, readily enter plant cells. The curves obtained in the electrical conductivity experiments show that plant cells are affected in different ways by different acids, and also give an idea of the rate and the manner in which the acids affect the cells ; but considering the complexity of the system, and the number of actions possible, it cannot be assumed that they give a definite measure of either the rate of absorption of the acid, or of the exosmosis of the substances in solution within the cell. They do suggest, however, that in dilute solution some acids can penetrate into the cells without doing much injury to them for some time, while others produce such an effect that some of the substances within the cell pass out almost as soon as the potato discs are immersed in the acid. The hydrogen-ion concentration measurements show the actual rate of absorption of the hydrogen ions by the cells. From the curves, it can be seen that the hydrogen ions of various acids are absorbed in the same way for some time, but that, after some hours, acids such as formic acid produce results different from those produced by the mineral acids. The anions have obviously an effect, and determine whether the acid is toxic or com- paratively non-toxic, toxicity being roughly indicated by the rate of exosmosis of electrolytes from the cell, as indicated by a comparison of the conductivity curves and those of hydrogen-ion concentration. The increase in the hydrogen- ion concentration of the external solution which takes place after some hours in the case of formic and acetic acid is a curious fact which requires explanation. It may be supposed that in penetrating into the cell, these acids, which are both chemically active, may react with some of the cell contents and produce other acids, probably organic acids which cannot normally pass through the limiting layer of the protoplast. But in entering the cell, the acid will have probably reacted with the outer layers of the protoplasm, and will have so altered their chemical constitution that the acids produced in the cell are able to diffuse out, and so increase the hydrogen-ion con- centration of the external solution. Considerable interest attaches to the mechanism by which acids enter the cell so rapidly. So far, opinion has been divided as to whether it is the protein or lipoid substances which are active in absorption. The experi- ments described in the third section of the experimental part of this paper all appear to suggest that, in regard to absorption of acids, it is the protein Hind. — Studies in Permeability. III. 237 substances of the cell which are responsible. The only lipoid substance examined, crude lecithin, gave no result similar to that obtained with living cells, whereas the resemblance between the action of living cells on acids and that of both peptone and the proteins extracted from potato is striking. Perhaps too much stress should not be laid on this isolated experiment, for commercial lecithin is a very impure product, and may contain as much as 50 per cent, impurities (8). Nevertheless, even in such a case one would expect a reduction in the hydrogen-ion concentration of the acid, if lecithin absorbed acid in the same way as the cell. Finally, it should be mentioned that the decrease in the hydrogen-ion concentration of the solutions containing plant tissue cannot be explained by the diffusion out from the cell of protein substances which then react with the acid in the external solution. Tests made for proteins in the external solution failed to reveal their presence there, even after forty-eight hours’ immersion in dilute acid. Moreover, it would be extremely unlikely that complex compounds like proteins would diffuse out during the first few hours of immersion in dilute acid. Summary. 1. The hydrogen ions of all acids examined are rapidly absorbed by plant tissue from dilute solutions. 2. The anion of the acid plays a large part in determining the effect of the acid on the cell, the fatty acids standing in strong contrast to the mineral acids. In the case of the mineral acids the exosmosis of electrolytes produced is considerably less than in the case of formic and acetic acids. 3. Some evidence is brought forward which suggests that proteins may play an essential part in the absorption of acids by plants. No evidence has been obtained suggesting that lecithin is at all active in this regard. Botany Department, The University, Leeds. Literature cited, 1. CzAPEK, F. : Versuche liber Exosmose aus Pflanzenzellen. Berichte d. dent. Bot. Ges., vot. xxviii, 1910, pp. 159-69. 2. : Uber die Oberflachenspannung und den Lipoidgehalt der Plasmahaut in lebenden Pflanzenzellen. Berichte d. deut. Bot. Ges., vol. xxviii, 1910, pp. 480-7. 3. : Uber eine Methode zur direkten Bestimmung der Oberflachenspannung der Plasmahaut von Pflanzenzellen. Jena, 1911. 238 Hind . — Studies in Permeability . IIP 4. Kahlenburg. L , and True, R. H. On the Toxic Action of Dissolved Salts and their Electro- lytic Dissociation Bot. Gaz., vol. xxii, 1896, p. 81. 5. Loeb, J. : The Dynamics of Living Matter. New York, 1906. 6. : Chemische Konstitution und physiologische Wirksamkeit der S'auren. Bioch. Zeitsch., vol. xv, 1909, pp. 254-271. 7. : The Role of Salts in the Preservation of Life, in 1 The Mechanistic Conception of Life : Biological Essays Chicago, 1912. 8. MacLean, H. : The Composition of ‘ Lecithin together with Observations on the Distribution of Phosphatides in the Tissues, and Methods for their Extraction and Purification. Bioch. Journ., vol. ix, 1915, pp. 351-78. 9. Osterhout, W. J. V. : Some Quantitative Researches on the Permeability of Plant Cells. The Plant World, vol. xvi, 1913, pp. 129-44. 10. Pantanelli, E., and Sella, M. : Assorbimento elettivo di ioni nelle radici. Rend, dell’ Accad. dei Lincei, Roma, ser. 5% vol. xviii, 1909, pp. 481-8. 11. Pauli, W. : Ueber physikalisch-chemische Methoden und Probleme in der Medizin. Wien, 1900. 12. Stiles, W., and J0rgensen, I. Studies in Permeability. I. The Exosmosis of Electrolytes as a Criterion of Antagonistic Ion-Action. Annals of Botany, vol. xxix, 1915, pp. 347-67. 13. Studies in Permeability. II. The Effect of Temperature on the Rate of Absorption of Hydrogen Ions by Plant Cells. Annals of Botany, vol. xxix, 1915, pp. 611-18. The Morphology and Anatomy of the Genus Statice as represented at Blakeney Point.1 Part /. Statice binervosa, G. E. Smith , and S. beliidifolia, D.C. (•-- S. reticulata ).2 BY E. de FRAINE, D.Sc., F.L.S., Westfield College , University of London. WITH SYSTEMATIC AND ECOLOGICAL NOTES BY E. J. SALISBURY, D.Sc., F.L.S. With Plate VI, twenty-eight Text Figures, and four Tables* Contents, PAGE PAGE I. Habitats of the Plants ex- 5. The Inflorescence Axis . 273 AMINED, AND CONSIDERATION Statice beliidifolia : of the Factors involved . 240 1. The Root 276 II. Description of the Species 246 2. The Stem ..... 277 III. Anatomy 250 3. The Leaf ...... 279 The Glands ....... 251 4. The Inflorescence Axis . 279 Statice binervosa : IV. Comparison of the Anatomy of 1. The Seedling . . . . 257 the Broad-leaved S. binervosa 2. The Root 258 and its Possible Parents . 279 3. The Stem ..... 263 V. Summary . . . 280 4. The Leaf ...... 266 IN the summer of 1912, at the suggestion of Professor F. W. Oliver, an investigation of the various species of Statice which occur in such variety and profusion at Blakeney Point, Norfolk, was begun ; the area shares with the neighbouring Burnham-Brancaster system the distinction of possessing every British species of the genus with the exception of N. Dodartii (Gri.). The present paper is concerned with the forms which are more specially related to the shingle banks and lows, namely Statice binervosa , G. E. Smith, and S„ beliidifolia , D.C,, while the species more particularly characteristic of the salt marsh will be dealt with in a later communication. The floral morphology and ecology of the genus have been investi- gated by Dr. E. j. Salisbury, to whom the author is entirely indebted 1 Blakeney Point Publication, Number 15. 2 The nomenclature adopted is that used by Mr. C. E. Salmon, Journ. Bot., vol. li, p. 92 et seq. [Annals of Botany, Vol. XXX. No. CXVIIX. April, 1916.] 240 de Fr aine. — The Morphology and Anatomy of the for this part of the paper. It is a great pleasure to acknowledge also the help given me by him in the identifications, in supplying plants raised from seed, and for his assistance in the preparation of the paper.1 I. Habitats of the Plants examined and Consideration of the Factors involved. Since the topography of Blakeney Point has been fully illustrated and described, it will not be further considered here.2 Statice binervosa. This species usually occupies a very definite habitat. It is most commonly found forming a zone upon the flanks of the lateral shingle banks (Oliver and Salisbury, loc. cit., p. 30 of reprint). In this situation it is the most characteristic species, and occurs in company with Frankenia laevis , Armeria maritima) Plantago Coronopus , and Glyceria maritima . These lateral shingle banks occur more or less at right angles to the shore, they are sheltered from wave impact, and are in a state of dormancy ; their shingle is stabilized. The soil of this zone is typically bare shingle in which the interspaces are completely filled with sandy mud. The habitat of vS\ binervosa is therefore only reached by the very highest tides, so that this species is the least maritime of any of the Statices found upon the area. It is in harmony with this fact that Statice binervosa will flourish and flower for years in normal garden soil. The rare occurrence of this species on the Main bank, in which habitat the mobile pebbles of the beach are gradually encroaching on the marsh, indicates that conditions of stability are a neces- sity for its existence. Oliver and Salisbury record, in this connexion, that S. binervosa only occurs on the Main bank sparsely near the crest, where the bank is broad and therefore less mobile, or else where dunes on the sea- ward face make stability more comparable with that on the laterals ; further^ on the ‘ Yankee bank ’ of the Long Hills the binervosa zone is discontinuous, and the breaks correspond with the places where the shingle is unprotected and therefore unstable. In obvious relation to this first requisite of the plant for a fairly stable habitat there is a remarkable difference between its underground parts and those of such typical mobile shingle Main bank plants as Suaeda fructicosa , Silene maritima , and Arenaria peploides. In these latter, rejuvenation by vigorous budding from the prostrated shoots or from the rhizome occurs abundantly, and there is ample evidence to prove that the more the shingling the greater the response of the plant to 1 Text-figs. 2, 3, 4 and 13, and Photographs 1 and 2 were kindly given by Dr. Salisbury. 2 Oliver, F. W., and Salisbury, E. J. : Topography and Vegetation of Blakeney Point, Norfolk. Trans. Norf. and Norw. Nat. Soc., vol. ix, 1913, p. 485 ; also reprinted with separate Paging- Genus Statue as represented at Blakeney Point . /. 241 it.1 2 The underground parts of S. binervosa consist of a short stout stem, which branches profusely at the apex, and which below passes over into a long, stout root of wire-like consistency. The plant has a rosette habit, and shows no signs of any rejuvenation by budding. A slightly different habitat, in which S. binervosa often occurs abun- dantly, is the shingle low, a depression left between closely juxtaposed banks ; it occurs especially at the convergence of laterals near their junction with the Main bank, and being accessible to the highest tides and at the same time secluded and tranquil, a covering of mud of greater or less thickness becomes deposited over the shingle. The lower part of such shingle lows is the main habitat of 6'. bellidifolia , while on the higher margins 5. binervosa occurs. The lows are frequently very muddy, but in some parts, where the tide rarely floods the low, and where, owing to the near presence of dunes, abundance of sand is available, the lows may be distinctly sandy. Finally, if the tide is cut off entirely, the binervosa zone may gradually spread centri- petally amongst the Suae da hummocks. Accumulation of sand, however, soon results in the extermination of the Statice and its replacement by the dune flora. Speaking generally, though the flanks of the lateral banks form the main habitat of the plant, yet the most luxuriant specimens are those from the crest. Here, since the fixity of the shingle is increased, the rate of accretion of the soil is accelerated, and the growth of S. biner- vosa is favoured, until, with the establishment of other plants, the factor of competition comes into play and the binervosa zone is driven lower down the flanks, and only relict plants are able to survive on the crest. In this connexion Salisbury gives some interesting data as to the average height and general vigour both of S. binervosa and also of the somewhat parallel case of Plant ago Coronopus. For example, the average height of crest plants of S', binervosa is 8*5" as compared with 4*1 8/r for plants from the flanks. The explanation given by Salisbury is as follows : 2 e under the rigorous conditions of the sloping flanks, the Statice abounds through the absence of its less hardy competitors, but with the accretion of soil the limiting factor for these latter is removed and only the more robust of the Statices survive the ensuing struggle : these from their perennial char- acter may remain for a considerable number of years, and by virtue of the 2" of soil in which they grow, they will be better nourished and therefore larger than those of the flanks which are rooted in bare or nearly bare shingle.’ Relict plants of S . binervosa also occur near the base of a lateral where they become buried by the advancing shingle fans of the Main bank. Such 1 Oliver, F. W., and Salisbury, E. J. : Vegetation and Mobile Ground as illustrated by Suaeda fructicosa on Shingle. Journ. of Ecol., vol. i, 1913, p. 249. 2 Loc. cit., reprint, p. 38. 2\2 de Frame. — 7' he Morphology and Anatomy of the plants afford a striking instance of the luxuriance due to the temporarily improved conditions. Salisbury suggests that the remarkable increase in size of these over- whelmed plants is probably due to one of two causes, or to a combination of both. He attributes it to the freedom from competition which is pro- cured by the onflow of the pebbles, since the shingle kills most of the plants it covers, or else to the mulch action of the shingle on the buried soil, tending to increase its water-retaining power. The shingle, as such, is obviously not of value, for the lateral roots of the plants when dug up are found to be practically restricted to the soil beneath the covering layer, though they may also occur in the top layers where the soil and humus collect. In order to determine more exactly the effect of shingling on S. bmervosa , four experi- mental areas were started on the crest of a lateral shingle bank on the Marams in November, 1912. Patches of the bank about one yard in diameter were covered with loose shingle from the Main bank to a depth of 5 cm. In August, 1913, the areas were examined and in every case the characteristic plants had grown through the shingle covering, viz. Statice binervosa , Armenia maritima , Plant ago Coronopns , Obione portn- lacoides , and, in two of the patches only, Frankenia laevis. The plants of the covered patches showed a remarkable increase in luxuri- ance as compared with those of the surrounding bank ; not only were the rosettes more spreading, but the size of the individual leaves was much greater, the inflorescence axis much larger, and the flowering heads were distinctly more luxuriant. These features occurred not only in the case of Statice , but also in plants of Armenia and Plantago , and to a much less marked extent in Obione and Frankenia. Further, while practically all the plants of S. binervosa on the bank showed distinctly red coloration of the leaves, those of the covered patches were vividly green, with no trace of red. On removing plants of Statice and Armenia for examination, it was seen that all the leaves of the old rosette had died, and had been re- placed by a vigorous output of new leaves from one of the shoots of the rosette. After examination of the areas, three of the experimental patches Text-fig. i. Plant of Statice binervosa after shingling. x §. a, old dead rosette of leaves ; b, new rosette. Genus Statice as represented at Blakeney Point. /. 243 (1, 3 and 4) were provided with a further 5 cm. coating of loose shingle, while the fourth area (2) was left untouched. In August, 1914, the patches were again examined, with the following results : Area 2. — Coated with 5 cm. of shingle in 1912 and untouched after- wards. Only one isolated Statice had survived ; it was distinctly more luxuriant than the plants on the untouched bank around, and its leaves were not only longer but greener. The inflorescence was distinctly more branched and luxuriant, and two spikes of flowers were present instead of the one usual in the uncovered plants. The length of the inflores- cences was 12 cm. and 13 cm. respectively, while the average length of the inflorescence in the uncovered plants was only 5 to 7 cm. Areas 1, 3 and 4. — Covered with 5 cm. shingle in 1912 and a further 5 cm. in 1913. The additional 5 cm. of shingle had in every case proved fatal to all the plants of the area, except a single runner of Agropyron sp., which had survived the additional shingling. The margins of the experimental areas had escaped the second shingling, and here the plants were covered with only the original 5 cm. of pebbles added in 1912. In this region the plants were numerous, and as regards their inflorescences much more luxuriant. The Obione , Plan- tago, and Armenia were, on the whole, distinctly more flourishing than the plants of the uncovered bank, and Obione seedlings were numerous. The Statice plants had very few leaves, and though the inflorescences were more vigorous, the plants had the appearance of making a final effort before death ensued. The average length of the inflorescence axis of Statice in the un- covered area surrounding the patch was 6 cm., in the covered patch the average height was 18 cm., while in one case the length attained was 35 cm. An examination of the ground, on a hot afternoon after a day of brilliant sunshine, revealed the fact that in the experimental areas, though the surface shingle was dry, 2-5 cm. below the surface the stones were quite damp, an observation which is in agreement with the results of Oliver and of Hill and Hanley.1 Beneath the coating of shingle the ground was quite dry, and on the adjacent untreated lateral both the surface and 2*5 cm. below were absolutely dry. In the experimental areas the effect of the loose shingle was similar upon all the species covered except Agropyron. It appears as though the 1 Oliver, F. W. : The Shingle Beach as a Plant Habitat. New Phytologist, vol. xi, 1914, p. 98. Hill, T. G., and Hanley, J. A. : The Structure and Water-Content of Shingle Beaches. (Blak. Point Publ., No. 11), Jour. Ecol., vol. ii, 1914, p. 35. 244 de Fr aine. — The Morphology and Anatomy of the increased luxuriance is due in these cases largely to the mulch action of the advancing fan, which acts by retaining a damp zone around the growing points of the plants, enabling them to grow for a time with greater vigour. Eventually, it would appear that the rapid raising of the surface level pro- duced by still further additions of shingle proves disastrous for a plant whose only method of combating the attack is to increase the length of the petiole in the effort to bring the leaves well above the surface again. It would thus appear that the restriction of .S', binervosa to its very limited habitat depends entirely on the ill-equipment of its underground parts. It is unable to colonize mobile shingle, not because of the arid conditions and lack of humus, but because of its inability to rejuvenate when covered, and because, as will be seen later, the structure of the stem is unsuited to vigorous branching accompanied by rapid growth, and on the other hand it cannot compete with other colonizers of the crest of the stable laterals, since the centrifugal extension of the turf of the various grasses, or of rhizomatous plants, or of cushion plants subjugates it, hence it is practically restricted to the narrow belt of ‘ No Man’s Land ’ on the sloping flanks of the laterals, where the conditions are probably more rigorous in many respects than in any other part of the area, and where it reigns supreme. Here the water-supply is extremely limited on account of the slope, the food-supply is small, for the accumulation of humus is com- paratively slight, and exceptionally high tides may deposit a coating of mud over the plants ; as will be shown later, the anatomy of the plant is well fitted to enable it to endure all these adverse conditions. Statice beltidifolia (= reticulata ). This species invariably occurs in situations reached by all but the lower tides, and especially in the muddy shingle lows already described, which remain moist throughout the inter-tidal period. The conditions of life in these lows is thus very different from those on the sloping flanks of a lateral, and approximate at certain times to those of a salt marsh, where, too, the plant is sometimes to be found. After a very high tide the salt water stands in the lows, converting them into small lagoons from which the water escapes only very slowly by percolation ;s the tide is left in them long after the Pelvetia and Aster marshes become fully exposed again. If the tides are exceptionally high and evaporation low, they may not become empty between two successive tides. It is clear that after a period of such unusually high tide, the salt water will remain standing in the lows for some time, so that 5. beltidifolia will have to face concentra- tions of salt water, and will thus be exposed at more or less regular intervals 1 My thanks are due to Dr. Sarah Baker, who kindly made these observations for me during a visit to Blakeney in November, 1913, in order to study the especially high November tides of that year. Genus Staticc as represented at Biakeney Point. I. 245 to prolonged periods of salt marsh conditions, while at other times they will approximate more nearly to those of the flanks of the laterals. It is in accordance with this unusual combination of factors that one meets in the anatomy a mixture of characters in many respects recalling, on the one hand, those of the closely related S. binervosa , and on the other, those of the typical salt marsh forms such as S. Limonium . It is of interest to note that Long 1 gives the habitat of 5. bellidifolia in the adjacent marshes at Wells as rolling sand-covered knolls 2 to 3 ft. above the highest spring tides. The much larger size of these specimens here and at Burnham Overy is probably due to the great depth of the sandy mud in which they grow as compared with the shallow mud overlying the shingle in the Biakeney ‘ lows \ In the region of the Long Hills and on the shingle plateau on the north-west side of the upper reaches of the Pelvetia marsh, where it ap- proaches the main shingle bank, the localities of binervosa and bellidifolia overlap to a certain extent, and in these regions a form of S. binervosa occurs which differs in certain respects from the typical one. In many features, both morphological and anatomical, this form, distinguished as the ‘ broad-leaved binervosa ’ in contradistinction to the typical narrow-leaved plant, is intermediate between .S', binervosa and .S', bellidifolia .2 There is a possibility that we are dealing with a hybrid between the two, though attempts to raise the plant from seed have so far failed. The broad-leaved type is entirely absent from the region of the Marams, where also only .S', binervosa occurs, and is restricted to those regions where the possible parent species are both present. In addition to the plants whose habitats have been briefly described above, plants from experimental garden cultures started by Dr. E. J. Salis- bury have also been examined, more particularly .S', binervosa plants grown in a cold greenhouse, and plants raised from seed. A summary of the chief forms of S', binervosa examined is as follows : 1. Tall Form — luxuriant plants growing on the edge of a shingle fan, termed £ Main bank plants \ 2. Dwarf Form — (PI. VI, Photo 1) — (a) Typical S', binervosa plants from the lateral shingle banks, termed ‘ narrow-leaved binervosa ’. (b) Plants from the margin of the muddy shingle lows, termed ‘ mud plants ’. ( P- 237« Genus Statice as represented at Blakeney Point . /. 253 As has been noted by Wilson, the gland cells are very rich in protoplasm, finely granular in appearance, and they contain a very large nucleus. The mucilage secreted by the hairs appears to be associated with tannin, a phe- nomenon which does sometimes occur in the case of mucilage secreting hairs ; 1 moreover tannin is commonly abundant throughout the plant. The physiological significance of tannin in the life of the plant is a difficult ques- tion, and its function appears to vary in different species.2 Sachs concluded that tannin resulted from intense metabolism such as occurs in rapid tissue formation, in vegetative apices and in association with secreting organs, and its presence in the mucilage glands of Statice may be accounted for in this Text-fig. 6. Mucilage gland from the base of the leaf sheath of S. binervosa. x 290. A. = gland in surface view ; B. = gland in transverse section. c. = cuticle ; s. - stalk-cells ; b. = basal cells ; n. — nucleus. way. The question of the further distribution of the tannin throughout the plant will be dealt with later. Mettenius or Licopoli Glands? The second type of gland occurs on both surfaces of the leaf and on the inflorescence axis of all members of the order. The structure of the gland was incorrectly described by Mettenius4 in 1 Haas, P., and Hill, 1. G. : An Introduction to the Chemistry of Plant Products. Longmans, Green & Co., 1913, p. 125. d he following microchemical reactions invariably gave good results: (a) A strong aqueous solution ot potassium bichromate gave a brownish coloured precipitate. (A) A neutral solution of ferric chloride gave a blue-black coloration. 2 Loc. cit., pp. 2 14-2 2 1. These glands aie commonly called chalk glands, since in some species calcium carbonate is excreted by them ; this does not occur in any of the British species of Statice . 4 Mettenius : Filiees Hort. Bot. Lips., 1856, p. 10. (Not consulted.) 254 de Fr aine. — The Morphology and Anatomy of the 1856 as consisting of a group of four cells, a mistake in which he was followed by the subsequent workers, Licopoli 1 and Maury.2 The glands were described by de Bary 3 as arising from epidermal cells which were rounded quadrate in surface view. Two intersecting walls at right angles to one another, and perpendicular to the surface, divided each of the gland-mother-cells into four ; each of the cells thus formed is then divided by a vertical wall into a narrow' inner, and a peripheral cell. The eight cells thus produced constitute the gland proper, the walls between the cells are extremely thin, and the contents are composed of very finely granular protoplasm. Volkens 4 states that the walls which limit the gland towards the inner tissue of the leaf are somewhat thickened, and are distin- guished by the fact that they do not swell up or dissolve under the action of concentrated sulphuric acid. He further describes the occurrence of special ‘ nebenzellcn 5 outside the gland cells ; these ‘ nebenzellen which are not mentioned in de Bary’s description, may be level with the epidermis, or maybe deeper and appear as half-moon shaped appendages of the gland elements ; he regards them as epidermal cells which were displaced from their original level on the formation of the gland. Vuillemin5 also describes the glands as composed of eight thin walled secreting cells, surrounded by four subsidiary cells whose walls are not dis- solved by treatment with boiling potash ; the margins of these cells are marked by cutinized attachments which join them to the base of the gland. He further states that ‘ ces aretes sont legerement carenees et pourvues de deux expansions laterales, exactement appliquees sur la commissure qui separe les cellules annexes ’, so that the latter form an uninterrupted barrier between the parenchyma on the one hand and the epidermis on the other, nothing passing from one to the other except through these sub- sidiary cells. He further states only four of the eight glandular cells are secretory, though exchanges can readily take place between all of them on account of their thin walls. The description of the glands given by Solereder6 differs from Vuille- min’s account in that the walls of the glandular cells separating the internal surface of the gland from the neighbouring tissues are described as being suberized, and the subsidiary cells are stated to have a double cap instead of a single one. 1 Licopoli : Gli stomi e le glandole. Atti R. Accad. d. Sc. Fis. e Mat., vol. viii, 1879. 2 Maury, P. : Etudes sur l’organisation et la distribution geographique des Plombaginacees. Ann. Sci. Nat., Pot., ser. 7, t. 4, 1886, p. 1. * 3 de Pary : Vergl. Anat., 1877, p. 113. 4 Volkens, G. : Die Kalkdriisen der Plumbagineen. Per. deutsch. bot. Gesell., 1884, Pd. 2, P. 334- 5 Vuillemin, P. : Reeherches sur quelcjues glandes epidermiques. Ann. Sci. Nat., Pot., 1887, ser. 7, t. 5, p. 152. 6 Solereder, PI. : Systematic Anatomy of the Dicotyledons. Oxford, Clarendon Press, 1898, p. 496. Genus Statice as represented at Blakeney Point. /. 255 The structure of the glands of the British species of Statice agree in most respects with the description given by Vuillemin, for they are invariably composed of a group of eight gland cells surrounded by four subsidiary cells (Text-fig. 7, A, B and c). In all the forms examined the whole of the walls which separated the subsidiary cells from the gland cells was cutinized, and further the cuticle was continuous across the external surface of the gland ; thus the entire gland was encased in a cuticularized layer (this is well shown in Text- fig. 7, B),and treatment with concentrated sulphuric acid left this case intact, but dissolved away the rest of the gland. It follows from this that escape of fluid from the gland will be accomplished with extreme difficulty, a fact which must be of considerable advantage to plants living under conditions where the water available for absorption is very limited in amount, e. g. S. Text-fig. 7. Mettenius gland from the leaf of S. binervosa. A. = gland in transverse section ; B. — gland in transverse section showing the cuticle raised ; c. — gland in surface view. c. — cuticle ; s. = subsidiary cells. All x circa 150. binervosa , or where absorption is rendered difficult on account of the presence of salt in the water, e. g. S. Limonium. As has already been noted by Wilson, glands showing every stage 01 transition from mucilage to Mettenian glands are met with in abundance in passing upwards from the leaf sheath to the blade, and Wilson regards both these glands as having the same origin, the Mettenian being the primordial form. He considers that the Mettenian glands were probably mucilage secreting organs, the chalk secretion which occurs in some species of Statice having been acquired later. Moreover he states that the con- tents of the Mettenian glands are always of a mucilaginous nature, even when the gland functions as a chalk-secreting organ. It appears more probable that the mucilage gland, which represents a multicellular tri- chome consisting of a secreting head borne on a short stalk, is the primi- 256 de Fr nine. — The Morphology and Anatomy of the tive structure. From such a form, by the embedding of the gland in the leaf tissue and the consequent necessary modification of the stalk- cells, such a type of mucilage gland as occurs in Aegialitis annulata 1 is arrived at. By the limitation of the number of secreting cells to eight, and the consequent modification and reduction of the stalk cells to the four subsidiary cells,2 the type of Mettenian gland characteristic of S. binervosa , &c., is obtained, in which the secreted substance is water and not mucilage. Finally, from such glands, the various modifications found in the chalk glands of such species as Limoniastrum monopetalum can be readily derived. Two views have been put forward as to the working mechanism of the Mettenian gland. According to Licopoli and Maury the product of secretion is amassed in a space which results from the separation of the four internal glandular cells, and it is rejected by the tension of the cells, which, however, always remain joined by their lower parts. This view appears to be incompatible with the structure of the gland as described by later workers. de Bary, Volkens, and Woronin all agree that the case is one of simple osmotic phenomena, and Volkens states that the glands act as valves which become efficaceous as soon as the transpiration of the aerial organs is in excess of the absorption of water by the roots. The number of glands present on the leaves of the various forms is shown in Table I. The portion of the leaf chosen was, in each case, the broadest region of the blade, and the numbers in each case repre- sent the average of a number of counts. Table I. Average Number of Mettenian Glands per scp mm. of Surface. Upper surface Lower surface of leaf. of leaf. S. bellidifolia 9-6 8*3 S. binervosa , ‘ broad-leaved ’ . . . 8*3 9*6 Dwarf Form. S. binervosa : /mud plant 7-5 12*4 sand plant 9*6 11*7 culture from seed 6-9 9-6 narrow-leaved lateral plant . . 6*2 8-9 ■< experimental plant 6*2 6-9 binervosa zone of a lateral . . 4*8 6*9 (one year in greenhouse) crest of a lateral 4*1 4-1 \ (one year in greenhouse) Tall Form. S. binervosa , Main bank plant . . . 5*5 5*5 The glands, speaking generally, are somewhat more numerous on the under than on the upper side of the leaf. It is interesting to note that in the case of the various forms of ther dwarf binervosa , in which the amount of water available for absorption would be expected to be 1 Solereder, Fig. 113, d-f, p. 497. 2 Or possibly by the elimination of the stalk cells, and the modification of neighbouring cells to form the subsidiary cells. 257 Genus Statice as represented at Blakeney Point. /. slightly more than in the typical narrow-leaved plant, the number of glands is slightly greater ; compare, for example, the plants from the muddy and sandy lows, and the culture plants, with the type form. Those plants which have been removed from their original home on the lateral shingle bank and have been for a year in cultivation under dis- tinctly more favourable conditions for root absorption show a smaller number of glands per unit area than in the case of the typical form — a condition to be expected, since any increase in the size of the leaf originally laid down would merely tend to spread out the glands, and would not lead to the development of new glandular structures in a leaf already developed. The same explanation possibly accounts for the difference between the gland counts in the case of the tall form from the Main bank and the type dwarf form from the lateral bank. Finally, it is interesting to note that though the number of glands per unit area in the case of the broad-leaved S. binervosa is intermediate, for the upper surface of the leaf, between that of its possible parent forms, no such relation obtains for the lower surface. Statice binervosa. i. The Seedling. The seeds of 5. binervosa are probably often distributed by very high tides, and seedlings obtained from Blakeney in 1913 1 were found growing in mud among a tangled mass of Rhizoclonium filaments. Since seedling plants are not very commonly met with the species may be a ‘ shy seeder ’, or else it is only when an exceptionally high tide occurs at the season for seed dispersal that seeds are scattered high enough up on the flanks of the laterals for a safe and stable place of germination to be obtained. This is the more probable in consideration of the facts stated with regard to the spread of binervosa near the Pelvetia marsh.2 The seedlings would appear to grow with extreme slowness, for seedlings collected in August, 1913, showed comparatively little advance on those collected in April of the same year. The seedlings of S. binervosa may be found in considerable numbers amongst the pebbles in the binervosa zone during March. Each possesses a pair of narrow, entire, and slightly spathulate cotyledons (Text-fig. 8), with very blunt apices about 4 to 5 mm. in length and from 1 to 1*5 mm. broad in the widest part. In colour the cotyledons are sometimes green, but are more usually of a deep crimson or port-wine colour, which is often shared by the first formed leaves. There is no external indication of a midrib. The cotyledons fuse at the base to form a very short tube. 1 1913 seemed to be an exceptionally good year for seedling binervosa. 2 Report on Blakeney Point for 1914, in Trans. Norf. and Norw. Nat. Soc., vol. x, p. 65. 258 de Frciine. — The Morphology and Anatomy of the The hypocotyl, like the cotyledons, is red in colour and varies in length from 3 to 5 mm. It exhibits a marked contraction at the collet where it passes into the root. Occasionally tricotylous or even monocotylous seedlings are met with. The first pair of leaves are shorter and broader than the cotyledons and have very acute apices. It is not usually, however, until the second pair of foliage leaves is produced that, distinct excurrent veins arc de- veloped at the tips. The cotyledons have each a very small endarch collateral bundle throughout their main portion, but just at the base two very small lateral ones appear. The two main cotyle- donary strands enter the axis as en- darch collateral structures, the lateral strands also enter and those from the opposing cotyledons fuse, pass towards the centre of the axis, and sooner or later die out. The two cotyledonarystrands organize a diarch root according to van Tieghem’s Type 3, but no regular ‘ bifurcation’ and ‘ rotation 5 of the bundles occur, and the change from stem to root structure takes place in an ill-defined manner and very rapidly. Mucilage glands are numerous in the axils of the cotyledons, and Mettenius glands are also present on their surface. Root hairs occur, but are not numerous. 2. The Root. The root system of S. binervosa usually varies in size more or less in proportion to that of the aerial organs. Thus the average height of plants taken from the crest of the laterals and the binervosa zone were respectively 22 cm. and 10-5 cm., the average rooting depths being 8-5 cm. and 4-5 cm. Tall plants of the Main bank had a rooting depth of from 18 to 26 cm. There is a main taproot which usually grows obliquely downwards. Large branches are few, but numerous fine laterals are developed which, like the main root, are characterized by their wire-like consistency. In the stele of the root of the young seedling the xylem occurs in the form of a diarch or triarch plate, with two or three alternating phloem groups. The pericycle consists of a single layer of thin walled cells, while the endodermis is well marked, with thickened inner and radial walls. The cortex is distinguished by its possession of a single ring of Text-fig. 8. Seedling plants of S. binervosa. x 2. Genus Statice as represented at Blakeney Point . /. 259 very large thin-walled cells (/), and the radial and outer walls of the piliferous layer (p.l) are thickened and suberized. The structure of a typical seedling root is shown in Text-fig. 9. At a very early age, while the seedling is still in the cotyledon stage, secondary growth begins, and secondary xylem, consisting of pitted vessels in which, in longitudinal section, the cross-walls are often persis- tent, and lignified fibres are produced, resulting in the production of a strongly lignified core. No parenchyma elements are present in the wood of a typical binervosa root, and some idea of the nature of the xylem may be obtained from Text-fig. 10. Medullary rays are absent except Text-fig. 9. Transverse section of a root of S. binervosa. pxy. = protoxylem ; ph . = phloem ; per. = pericycle ; end. — endodermis ; l. = large cells of the cortex ; p.l. = suberized piliferous layer, x 300. Text-fig. 10. Transverse section of part of the xylem of the root of S. binervosa. Main bank plant, x 236. v. = vessels ; f. = fibres. in the region of the exit of a root-trace, where a multiseriate ray is developed. Annual rings are usually well marked (cf. Text-fig. n). The secon- dary phloem is small in amount, and shows no features of special interest. Simultaneously with the beginning of secondary growth, the pericyclic cells become meristematic, and the cambium thus produced gives rise to a periderm on its outer margin, thus cutting off the endodermis and cortex. This periderm is composed of three to four layers ; of thickish walled cells with dense homogeneous contents. The contents are of the nature of tannin, and may possibly be a Phlobaphene — a decomposition product of plumbagin — the substance stated to occur in the roots of the Plumbagi- naceae. Practically no information could be obtained as to the nature of plumbagin, which is described by Tunmann1 as ‘ ein wenig erforschtey 1 Tunmann, O. : Pflanzenmikrochemie, Berlin, 1913. 260 de Fr nine. — The Morphology and Anatomy of the Pflanzenstoff, der moglicherweise in der Zelle in glykosidischer Bindung auftritt and which he states occurs ‘ im Zellinhalte in alien parenchy- matischen Zellen (auch Markstrahlen, Phloemparenchyma) der Wurzel- rinde ’. Examination of the root for the presence of tannin indicates that a substance of this nature occurs in practically all the parenchyma cells of the cortex, medullary rays, and phloem, as well as in the cork : hence it is concluded that this substance is probably identical with the plumbagin of Tunmann, more especially since ‘ at one time it was thought that tannins were substances of a glucosidic nature and occurred in the plant in combination with a carbohydrate complex such as glucose ; . . . . this is undoubtedly so in some cases \l The cork cambium gives rise internally to a zone of secondary cortex. This cortex is com- posed of thin walled parenchyma, the majority of the cells of which contain plumbagin ; immedi- ately beyond the phloem lies a broad almost con- tinuous band of sclerenchyma fibres (s.f.) , while groups of sclerenchymatous cells or sclereidcs of varying sizes (w. = cortical parenchyma ; fer. — periderm. Observations which are of some interest in this connexion have been made by van Ufford during the course of his investigation of the Flora of the Pierriers ( Moraine talus) of the High Calcareous Alps of the Canton de Vaud.1 The most striking feature in the anatomy of these plants was the development of collenchyma in the peripheral regions of the underground axes. Pie regards this collenchyma as acting in resisting the thrust of the mobile stones while at the same time permitting without rupture the neces- sary stretching needed on account of the mobility of the substratum. In the case of .S. binervosa , since there is no need to provide a mechanical tissue which will allow of further growth in the organ, as the substratum is not mobile, the collenchyma is replaced by the mechanically more efficient sclereides. Examining the relations between the thickness of the collenchyma and 1 van Ufford, Q. : Etudes ecologiques de la flore des Pierriers. These. Montreux, 1909. T 262 de Fr aine. — The Morphology and Anatomy of the the radius of the stele, van Ufford invariably found that in closely related plants the more mobile the station the greater the proportion of collenchyma developed. Two examples will suffice to indicate this fact : Habitat. Plant. Thickness of collenchyma radius of the stele. ^Mobile stones Galium helveticum 1 i*5 ,, Less mobile earth ,, rotundifolium 1 4*3 Fixed earth . . ,, verum 1 6*o Mobile stones . . Cerastium lati folium 1 i*9 2.- Less mobile earth ,, arvense 1 2*5 Fixed earth . . ,, cuneifolium I 5*° He further found that only species which have abundant stereome can survive the thrust of the stones ; those which do not possess it die out. Starr 1 in an examination of the anatomy of dune plants came to somewhat similar conclusions, for she found that comparing the stem of a plant from a mesophytic situation with one of the same species from the dune, the following characters occurred in the dune form : 1. The vessels were more numerous but smaller, the total area being, however, larger. 2. The lumen of the fibres was smaller. 3. The walls of the vessels and fibres were heavier. 4. More sclerenchyma and collenchyma were developed. 5. Slight increase in cork formation occurred. Finally, Haberlandt thinks that mechanical influences, if they do not pass beyond a certain limit, act on the stereome as a stimulus for further building it up. An examination of the roots of the various forms of S', binervosa from the different habitats shows certain constant differences to obtain : Annual growth rings. The annual growth rings in the root of the Main bank plant are always wider than in those of a lateral plant ; this is probably in accordance with the less rigorous conditions to which the Main bank plant is subjected. Starr also found a similar relation be- tween plants growing in a mesophytic situation and those growing on a sand dune, and she states, ‘a majority of xerophytic forms have more growth rings to the given diameter than the mesophytic forms, showing slower growth under the more adverse conditions.’ Wood parenchyma. Development of wood parenchyma is entirely absent in the plants from the Main bank and the narrow and the broad 1 Starr, A. M. : Comparative Anatomy of Dune Plants. Bot. Gaz., 1912, vol. liv, p. 263. Genus Statice as represented at Blakeney Point . /. 263 leaved lateral plants, except in connexion with the exit of a root-trace. In the plants from the muddy lows the beginning of a year’s growth is marked by the formation of wood parenchyma instead of fibres, so that the annual rings are sharply marked off from one another. In the plants cultivated from seed the xylem shows abundance of parenchyma and very few fibres. Sclerenchyma and sclereides. The difference in the development of the sclerenchyma ring and the sclereide groups has already been referred to, and is briefly summarized in Table II. Table II. Distribution of the Sclerenchyma Ring and Sclereides in the Various Forms of S. binervosa. Form. Plant. Tall Dwarf ? Hybrid Main bank Narrow-leaved ’ lateral Muddy low Sandy-mud low k Culture from seed ‘ Broad-leaved ’ lateral Sclerenchyma ring. Broad and almost continuous Sclereide group. Very numerous ?? 9? Narrow and almost continuous Narrow and discontinuous A few isolated fibres only Broad and continuous, but not quite so broad as in lateral form •>1 >> Not numerous Few and small Very few Numerous Vessels. An examination of the vessels in the roots of the various forms shows in every case a gradual increase in size from the primary elements through each successive annual ring (Table III) ; moreover, in those of the Blakeney forms, in which absorption is presumably more difficult, there is a diminution in the size of the vessels : this is in agreement with the observations of Starr already quoted. The diameter of the vessels in the plants cultivated from seed does not fall into line with the other measurements, and it seems difficult to explain the apparent anomaly. Table III. Average Internal Diameter in mm. of the Largest Vessels in the Root and Stem of the Various Forms of V. binervosa. Dwarf form. Xylem. ? Hybrid form. Broad-leaved. Tall form. Alain bank. Narrow- leaved. Sandy. Muddy. Culture. ^Primary o*oi 2 0*010 0*010 0*006 0*006 0*006 Secondary : Annual ring i 0*024 0*018 0*014 0*018 0*008 0*010 1 „ „ 2 0*032 0*026 0*020 0*020 0*01 2 0*014 I n > > 3 °*°44 0*030 0*024 0*024 0*016 — V » }} 4 — 0*036 0*032 0*028 0*020 ) Primary 0*006 o-ooS 0*006 0*008 0*006 0*006 1 Secondary 0*020 0*020 0*016 0*020 0*012 0*010 3- The Stem. The stem of S. binervosa is very short and entirely subterranean, with a much branched crown, each branch bearing a rosette at the ground level. The first inflorescence is terminal and thus brings to an end the growing axis. Growth in the following season is continued by one or more buds developed in the axils of the apical leaves of the T 2 264 dc Fraine. — The Morphology and Anatomy of the rosette (Text-fig. 13, a). In this way the richly branched cushion of the adult becomes formed so that the final condition is an aggregate of closely approximating rosettes. The richly branched crown serves as a trap for sand and silt in which it becomes completely embedded. As the plant increases in size and vigour it will be noticed that from the centre of some of the rosettes more than a single scape is produced. Sometimes this appears to be the result of forking of the scape from its extreme base, but more usually is due to the precocious development of one or sometimes two axillary buds of the terminal leaves (Text-fig. 13, b). In these circumstances, if no additional axillary buds develop, the life of the branch ends with the withering of its leaves. The main stem is usually small in diameter, though in a good-sized Main bank plant it may reach as much as 1 cm. across. Text- 14, A, shows the arrangement of the various tissues in the stem of a tall form plant. The pith ( p.) is large, and composed of thin-walled paren- chyma, with practically no intercellular spaces ; numerous groups of sclereides {self varying very much in size, occur in it. Medullary rays {m.r.) of thin walled cells break up the vascular ring into numerous segmen t; the rays vary considerably in width, being often very broad where a leaf trace is making its exit. A group of sclereides is fre- quently found in the broad rays accom- panying the outgoing trace. The primary vascular strands are very numerous and of small size: secondary growth sets in early, and a narrow zone of secondary wood (xy.) composed of vessels embedded in fibrous tissue is produced ; annual rings cannot be distinguished in the stem as they can in the root. The ele- ments of the protoxylem have spiral markings, and the rest of the xylem is composed of vessels with simple pits on their walls. The secondary phloem is small in amount (pk.)> In the pericycle, masses of sclerenchyma fibres (sc.) occur opposite to each vascular wedge ; no endodermis can be distinguished. The cortex (co.) is made up of small celled, rounded parenchyma, with little intercellular space system ; very numerous groups of sclereides (set.) occur in it, their walls are strongly thickened, and they have many branched pits. A broad zone of periderm (per.) limits the stem. As in the root, he cells of the periderm and many of the thin-walled cortical cells have contents which are of the nature of tannin (cf. p. 359) ; this substance is Text-fig. i 3. Diagram of the longi- tudinal section through the apex of rosettes of S. binervosa. Explanation in text. Genus Statice as represented at Blakeney Point. I. 265 especially abundant just below the insertion of a leaf, where a double layer of irregular cells indicating the line of union of the broad leaf sheath and the stem is very rich in it. Text-fig. 14. Diagram of part of stem of S. binervosa. A = Main bank plant ; B = typical dwarf form ; c = broad-leaved form ; l) — muddy low plant; E = culture plant, per. — periderm ; co. ~ cortex ; scl. = sclereides ; ph. = secondary phloem ; sc. — sclerenchyma fibres ; m.r. — medul- lary rays; xy. — secondary xylem ; p. — pith. The dotted line indicates the limit of the pith, x 30. The stems of plants from the various habitats are in essential features similar to that of the tall form, and only differ in those characters such as have been described for the root, viz. in the proportion of the mechanical tissue present and in the size of the vessels. 266 de Fraine. — The Morphology and Anatomy of the A comparison of A, B, C, D and E in Text-fig. 14 shows that the amount of sclerenchyma and the number of sclereide groups diminish considerably in the mud and culture plants as compared with the main bank form and the narrow-leaved lateral one, while the broad-leaved plant shows less marked differences, though the stereome is distinctly reduced. In the xylem of the sand and mud plants there is a little less fibrous tissue, and slightly more wood parenchyma is developed. In the culture plant no fibres occurred in the xylem of a two-year old plant, and in a narrow-leaved lateral of a similar age the wood was also free from fibres ; but in the lateral plant the wood was much more com- pact than in the culture form, in which considerable wood parenchyma was present. Culture, as in the case of the root, has increased the relative width of the phloem zone. Finally, differences occur in the size of the vessels (see Table II), but since only two-year-old culture plants were obtainable the measurement of the secondary vessels of this stem should not be compared with those of other plants, for increasing age probably means increasing diameter of the vessels up to a certain extent. These variations in the different forms are in entire agreement with those found in the root, and since the short stem is completely subterranean they are explicable on the same grounds. 1 ' 4. The Leaf. The general distribution of the tissues in the leaf of the narrow-leaved binervosa is shown in Text-fig. 15, A. Numerous small vascular bundles are present, and the mid-vein is encircled by a sclerenchyma sheath (se.) ; a large number of sclereide groups varying in size occur, and form a charac- teristic feature of the leaf. Comparison of the proportion of stereome present in the leaf blade of a narrow-leaved plant from the lateral banks and plants from the other habitats shows that in the sand form the sclerenchyma sheath round the mid-vein is much less developed, and the number of sclereide groups distinctly fewer (Text-fig. 15, B) ; the mud form shows similar features (Text-fig. 15, c) ; and they are still further emphasized in the plants from the experimental area (Text-fig. 15, D). The broad- leaved form has the sclerenchyma sheath confined to a cap above and below the main veins, and the sclereide groups are very few in number (Text- fig. 15, e), features which are interesting in comparison with the leaf of N. bellidifolia (Text-fig. 15, F), in which also there are only sclerenchyma caps to the main vein, and no sclereide groups at all. Comparison of Text-fig. 15, A, E and F, clearly demonstrates the intermediate character of the putative hybrid and the two suggested parents. Genus Statice as represented at Blakeney Point. /. 267 An examination of the stereome in the petiolar region further illus- trates the effect of environmental conditions on its development (Text- fig. 16). The petiole of the broad-leaved form (b) is clearly intermediate Text-fig. 15. Diagram of part of the leaf blade of S. binervosa (a-e) and .S’, bellidifolia (f) in its middle region, showing the distribution of the stereome. x 30. a = narrow-leaved plant ; b = sand plant ; c — mud plant ; D = plant from experimental area ; E — broad-leaved form. vb. = mid- vein ; sc. — sclerenchyma sheath ; sc. 1. — sclereides. Vascular bundles dotted ; stereome black. between the narrow-leaved plant (A) and S. bellidifolia (d). In the ex- perimental plant there is little difference in the proportion present, but, as in the case of the root and stem, culture greatly diminishes the development of mechanical tissue (c). Not only does culture with its more favourable conditions diminish the proportion of stereome present, 268 de Frame.' — The Morphology and Anatomy of the but it also has a marked effect on the thickness of the walls of the ele- ments, as can be seen on comparison of A with C and B with D in Text- fig. 17. This again is in entire agreement with the results obtained by Starr in her comparison of mesophytic and dune forms (see p. 262). In the leaf sheath of all the forms the stereome increases very con- siderably in amount ; the walls of the elements become much thicker, and their length increases greatly ; the sclereides become long, fibre-like elements (cf. Text-fig. 17, B and I) = typical petiole sclereides, with Text-fig. 18, A and B = typical sheath sclereides). Text-fig. 16. Diagram of the transverse section ot the middle region of the petiole of S. binervosa (a-c) and S. bellidifolia (d). a = narrow-leaved plant ; l) = broad-leaved plant; c = narrow- leaved plant cultivated from seed. x 30. Stereome in black ; vascular bundles dotted : sc. 1. — sclereides; sc. = sclerenchyma sheath. The detailed structure of a narrow-leaved form growing on a lateral is shown in Text-fig. 19, A: on the upper surface is a well-marked palisade layer two or three cells deep ; these pass over gradually into the rather shorter palisade-like cells of the lower surface ; practically no intercellular spaces are developed. Chlorophyll occurs in all the cells except those of the epidermis : where red coloration is present it is due to red cell- sap. The leaf of a Main bank plant is precisely similar in all respects to that of a narrow-leaved lateral one. The leaf of the mud plants is on an average slightly thicker than that of the typical form, the increased thickness being due to elongation Genus Statice as represented at Blakeney Point. I. 269 of the palisade cells ; the average thickness is 0*488 mm. as compared with 0*448 mm. in the narrow-leaved lateral plant. The sand plants show slight differences in the development of the palisade, only two layers of which are present on the upper surface, and usually only one on the lower, but the individual elements are deeper than in the typical form (Text-fig. 19, B). Culture exercises a marked influence on leaf structure, for the plants Text-fig. i 7. Selereides from the petiole of S. binervosa. A and C from culture plant ; B and D from a narrow- leaved lateral plant. A and B = transverse sections ; C and D = longi- tudinal sections, x 170. Text-fig. iS. Selereides from the leaf sheath of S.binervosa, narrow- leaved plant. A, transverse section ; B, longitudinal section, x 220. grown from seed have a bifacial leaf, while those at Blakeney are iso- bilateral. As is shown in Text-fig. 19, c, the palisade on the upper surface is only two layers deep, and the spongy tissue with large intercellular spaces is well developed. This development of the spongy mesophyll in the culture form is undoubtedly to be related to the absence of the complex of factors acting on the shingle plants. The chief of these are the heat radiation from the hot surface of the shingle during the summer months, and the restricted water-supply of the habitat ; the removal of these allows of the production 270 de Fr nine. — The Morphology and Anatomy of the of a typical bifacial leaf form. Also under cultural conditions the leaves tend to be more erect in position. Text-fig. 19. Transverse sections of part of the leaf blade of S. binervosa (a-p. x 235), and of S. bellidifolia (e. x 220). A, narrow-leaved lateral plant ; B, sand plant; C, culture from seed plant ; D, broad-leaved form ; e, .S’, bellidifolia. Comparison of Text-fig. 1 9, A. D and E, again brings out the intermediate character of the leaf of the broad-leaved plant, for both it and bellidifolia are bifacial in type. Genus Statice as represented at Blakeney Point. /. 271 The margin of the leaf of the various forms shows very little varia- tion, for it is always two cells in thickness ; the cells have very thick walls, and the angles between them are filled with a substance which gives the reactions of cuticle (Text-fig. 20, B, c). Comparison of Text-fig. 20, A, Band C, shows that the margin of the broad-leaved form is thicker than that of the narrow-leaved, though otherwise it resembles it, and is different from the somewhat blunt margin of bettidifolia. This difference between the broad and narrow leaved forms is obviously merely related to the difference in leaf thickness. In the broad-leaved form the average thickness of the middle region of the leaf blade is 0-464 mm. as com- pared with 0 448 mm. in the narrow-leaved. The structure of the epidermis shows differences in the different forms, not only in the thickness of the walls but also in the degree of development of the cuticle and in its striation. Text-fig. 21 shows the structure of a typical epidermal cell in each of the forms examined. Com- parison of Text-fig. 21, B, C, E and G, brings out the fact that in the forms where the conditions of life as regards water- supply are most severe, viz. in the narrow (B) and broad (g) leaved lateral plants, in the sand plant (c), and in the experimental plant (e), the epidermal cells are larger and thick walled, and have a heavy ridged cuticle, while a well-marked internal cuticle Text-fig. 20. Transverse section • , 11 • . , 1 1 t of the leaf-margin of S. binervosa. is piesent as well in the narrow and broad Narrow.leaved form (a), broad-leaved leaved plants and in the sand plant, but form (b), and bellidifolia (c) ; 0 cuticle. X 2 ~ O is absent in the experimental form, where the conditions are probably a little more favourable. The epidermal cells of the mud form (Text-fig. 21, d) were small, heavily cuticularized, and an internal cuticle was also present. Here again adequate water-supply is not available for the roots, owing to the difficulty of water absorption in the salt soil and the open character of the vegetation, which renders the air in the region where the leaves are expanded com- paratively dry, and this may account for the epidermal structure.1 The Main bank plant (A) shows a markedly thinner cuticle than the lateral form, and the internal cuticle is absent. This is possibly in relation to the more adequate water-supply which the plant possesses (cf. p. 244). In 1 A characteristic feature in the leaves of nearly all salt marsh plants is the very slight develop- ment of cuticle, possibly largely owing to the fact that their aerial parts are expanded in a distinctly humid atmosphere. 272 de Fraine . — The Morphology and Anatomy of the the culture plant (f) the cuticle is markedly thinner than in the other forms ; the epidermal cells are also smaller, and no trace of internal cuticle was seen. The cells of the lower epidermis show characters essentially similar to those of the upper, except that the ridging of the cuticle is slightly more pronounced in the case of the Main bank, narrow-leaved lateral, and mud forms, while in the sand plant the ridging is much more marked (Text- fig. 21, K), a fact which is probably to be correlated with the reflection from the sandy surface. The striation of the cuticle as seen in surface view is shown in Text- fig. 21. It varies a little in the different forms, and it is significant that practically none occurs in the culture form. Text- fig. 21. Upper epidermal cells of S. binervosa (a-g), and o f .S', bellidifolia (h). a-h and K in transverse section, x 220; ai-ei, gi, hi, in surface view, showing striation of the cuticle, x 16S. A, Ai = Main bank plant; B, Bi — narrow-lenved lateral plant ; c, Cl = sand plant; D, D 1 — mud plant ; E, El — experimental plant; F = culture plant; G, gi = broad-leaved plant; it, hi = S. bellidifolia ; K — lower epidermis of a ‘sand’ plant of S. binervosa . c. = cuticle; i.c. — internal cuticle. Stomata occur on both surfaces of the leaf ; they show the constant character of being surrounded by three subsidiary cells (often four in S. bellidifolia ) (cf. Text-fig. 24, A and c). In all the forms examined they appear essentially similar in structure, and only differ in the fact that those of the Main bank plant are larger than in any other form, though those from the experimental area are nearly equal to them in size : the stomata of the mud form are the smallest, those of all the other plants are approxi- mately equal to those of the narrow-leaved plants. They usually occur on the level of the epidermal cells, though they may be very slightly raised above them. The number of stomata on the two sides of the leaf is approximately the same, but differences occur in the different forms, as is shown in Genus Statice as represented at Blakeney Point. /. 273 Table IV ; comparison with Table I shows that, generally speaking, increase in the number of glands corresponds with increase in the number of stomata : — Table IV. Average Number of Stomata on the Leaf Blade in S. bellidifolia and the Various Forms of V. binervosa, per sq. mm. of Area. Form. Plant . S. bellidifolia 1 Hybrid Broad-leaved binervosa . . . S. bmervosa : muddy low sandy-mud culture from seed .... -r. c narrow-leaved lateral . . Dwarf {experimental binervosa zone of a lateral (one year in greenhouse) crest of a lateral (one year in greenhouse) Tall S. bmervosa , Main bank plant Upper surface of Lower surface of leaf blade. leaf blade . 73*3 35*9 40*1 35*2 63*6 63-6 56-0 51-2 37’3 35*2 33*2 26*3 36*6 44*2 32*5 35*2 16*6 23*5 34*6 31*8 The structure and distribution of the mucilage and Mettenian glands has already been fully considered (see pp. 231-7). The Mettenian are generally sunken slightly below the epidermal level, while the mucilage glands are raised slightly above it. In the leaves of the culture plant, however, the Mettenian glands are level with the surface. In the broad-leaved form the epidermal cells surrounding a gland are some- what enlarged, a feature which recalls S. bellidifolia , where they are enormous when compared with the remaining epidermal cells. The distribution of tannin in the leaf blade is similar in all the forms. The layers of cells immediately adjacent to the epidermis, both upper and under, are very rich in a substance giving the reactions of tannin (see p. 259), and these two layers contain very little starch ; the two layers immediately internal to them are very rich in starch and contain practically no tannin. Many of the remaining mesophyll cells have tanniferous contents, more especially those in the neighbourhood of the vascular bundle of the midrib. In the petiole this substance occurs in the subepidermal layer and in numerous isolated parenchyma cells. Enormous quantities are present in the mud plant.1 5. The Inflorescence Axis. The distribution of the tissues in the inflorescence axis is shown in Text-fig. 22. The pith (/.) is surrounded by a ring of large, closed, collateral bundles (yb. 1). In the lower parts of the axis in the pith one or more bundles of sclereides may occur, but they usually die out before the insertion of the first scale leaf. The number of vascular bundles varies greatly ; it depends on the size of the inflorescence and the degree of branching ; the number decreases after the emission of each branch. The 1 For the probable function of the tannin see pp. 253, 259-60. 274 de Frame. — The Morphology and Anatomy of the vascular strands are embedded in a mass of lignified tissue (sc.), the outer layers of which are sclerized (sc.t.) in the stouter inflorescences. In the smaller axes this sclerized zone becomes reduced and finally disappears, and the inner ring of bundles are no longer completely embedded in fibres. Immediately beyond the sclerized zone is a ring of very small bundles (vb. 2) ; they lie at the margin of the assimilating region of the axis (a.s.). Below the level of insertion of the first scale leaf these small bundles join on to the internal ring of larger strands. The assimilating layer is com- posed of about four rows of small, rounded, parenchyma cells, but above sVlr.l ' as 1 Text-fig. 22. Diagram of part of the inflorescence axis of S. binervosa. x 70. a.s. = assimi- lating cortex ; vb. 1 = inner ring of bundles; vb. 2 = outer ring of bundles; sc.l. *= sclerized zone; sc. — lignified tissue ; pli. — phloem ; xy. = xylem. The white line indicates the limit of the sclerized zone. the insertion of the scale leaf the outermost layer of these cells becomes palisade-like (p., Text-fig. 23, A), while in the ultimate branchlets two layers of palisade cells are present. The epidermal cells have very thick walls, especially on the outer surface ; the external cuticle (o.c.) is heavily developed, with a ridged outer surface and plugs between the cells ; a well-marked internal cuticle also occurs (i.c., Text-fig. 23, A). Stomata and Mettenian glands are both present on the inflorescence axis. The general structure of the inflorescence axis is the same in all the forms of vS. binervosa , and it is only in the broad-leaved plant that any noteworthy variations occur. The palisade layer of the flowTering spikes of Genus St at ice as represented at Blakeney Point. I. 275 this plant consist of one long and one shorter layer of cells (Text-fig. 23, b)5 and it is intermediate in this respect between its possible parents (cf. Text- fig. 23, A, B and c). It is in the surface of the axis, however, that the chief difference occurs. In the tall and dwarf forms of binervosa it is quite smooth, except in the ultimate branchings. In the ? hybrid all the branches are slightly rough, but in bellidifotia the scape is distinctly granulated throughout. The rough Text-fig. 23. Part of the inflorescence axis of S. bellidifolia (c), narrow-leaved binervosa (a), broad-leaved binervosa (b) (taken from comparable levels). x 347. 0 c. = external cuticle ; i.c. — internal cuticle ; p. = pali- sade ; a.s. = assimilating layer. Text-fig. 24. a. Gland in surface view from the inflorescence axis of S. bellidifolia. x 293. B. Part of the epidermis showing one of the rosette cells. x 220. r. = rosette of enlarged cells; g. = gland ; s. — stoma with four subsidiary cells; c. Stoma of S. binervosa with its three subsidiary cells, x 165. appearance is due to the fact that the epidermal cells surrounding a gland are considerably enlarged and papillate, resulting in the sinking of the gland in a depression, while the glands in surface view appear as rosette-like structures. This is clearly shown in Text-fig. 24, A. The slight wartiness of the axis of the broad-leaved form is due to a similar, though less pronounced, arrangement of the cells surrounding the gland (cf. A and B, Text-fig. 25). In all the other forms of binervosa , though an occasional enlargement 2 j6 de Frame. — The Morphology and Anatomy of the of the cells may occur in the ultimate branches of the scape, it is never developed as a constant character. In the barren spikes of the broad-leaved S. binervosa the assimilating tissue is much more developed ; the palisade cells are longer and are three I ext-fig. 25. Transverse section of part of the inflorescence axis of broad-leaved S. binervosa (a), and of S. bellidifolia (b), showing the difference in the projection of the gland areas, x 167. g. = gland ; r. = ‘ rosette ’ cell. deep, a feature which recalls the barren scapes ot S. bellidifolia. The vascular bundles are also small, and the sclerized zone is only feebly developed. Statice bellidifolia. The habitat of this plant at Blakeney Point has been described on p. 244. Only plants from the lows were available for investigation. 1. The Root. The structure of the root differs essentially from that of S. binervosa , and much more closely resembles that of the salt marsh forms such as .S'. Limonium ; it is much less wiry and distinctly more fleshy. The dis- tribution of the tissues in the root is shown in Text-fig. 26, A ; a comparison with Text-fig. 12, A, shows the very different proportions which obtain between the mechanical tissues in bellidifolia , which is essentially a salt marsh plant, and binervosa , which inhabits shingle. The primary root has a diarch or tetrarch plate of xylem ; secondary growth sets in early and produces considerable secondary wood (xy., Text- fig. 2 6, a) surrounded by a broad zone of secondary phloem ( ph .). The secondary xylem is composed of comparatively few large vessels embedded in xylem parenchyma ; occasional wood fibres occur, but the wood is dis- tinctly ‘ soft * in type, and very different in nature from that of binervosa (cf. Text-fig. 27 and Text-fig. 10). Broad primary medullary rays occur opposite the protoxylem groups, and wide secondary rays split up the wood into segments and still further increase its parenchymatous nature. A broad zone of secondary phloem (/A, Text-fig. 26, a) is produced ; the bulk of it is phloem parenchyma. The secondary cortex is a wide zone of Genus Statice as represented at Blakeney Point. I. 277 aerating tissue (a.co.), precisely similar to that of Salicornia } or of the salt marsh species of Statice. This aerenchymatous cortex is in obvious relation to the habitat of the plant in lows, the soil of which must often be water- logged for considerable periods. A broad zone of periderm encircles the cortex. Throughout the greater part of the root no sclerenchyma or sclereides occur, but near the junction with the stem nests of sclereides develop mostly in a single series immediately beyond the phloem zone. (set. , Text-fig. 26, b). Text-FIG. 26. Diagram of a transverse section of a root of S. bellidifolia. A = young root, x 43 ; B = old root near its junction with the stem. x 60 ; pxy . = protoxylem ; xy. — secondary xyltm \mph. — secondary phloem ; a.co. = aerating cortex ; m.r. — medullary ray ; set. — sclereide nest ; per. — periderm. 2. The Stem . The stem of S. bellidifolia is short and more completely embedded in the mud than in 5. binervosa . The distribution of the tissues is shown in Text-fig. 28. The pith (/.) is large, and contains nests of sclereides ( self though they are less abundant than in binervosa (cf. Text-fig. 14, B). The vascular bundles are broken up into a few large wedges by very wide medullary rays ( m.r .), which only occur at the point of exit of leaf-traces (/./.). The xylem contains comparatively few vessels, but numerous fibres ; the size of the vessels and their markings are the same as those of the Main bank 5. binervosa (see Table III, p. 263). The zone of secondary phloem ( ph .) is wide, and consists chiefly of phloem parenchyma. The cortex, in which, as in the pith, the intercellular space system is well developed, shows no sclerenchyma and only a single series of sclereide groups, and the zone 1 de Fraine, E. : The Anatomy of the Genus Salicornia. Journ. Linn. Soc. Bot., 1913, vol. xli, P- 337- U 278 de Fr aine. — The Morphology and Anatomy of the C Text-fig. 27. Detail of part of xylem in root of S. bellidifolici. c. = cambium ; par. = wood parenchyma ; v. — vessel, x 290. Text-fig. 28. Transverse section of the stem of S. bellidifolia. x 40. p. = pith ; xy. = secondary xylem ; ph. = secondary phloem ; m r. = medullary ray ; l.l. — leaf- trace ; scl. = sclereide group ; a.co . = aerating cortex ; per. = periderm. of periderm is very wide- In the nature and proportion of the stereome present, the stem approximates very closely to the stem of the plants of binervosa cultivated from seed (cf. Text-fig. 28 and Text-fig. 14, E) ; indeed Genus S/a/ice as represented at Btakeney Point. /. 279 practically the only difference between the two lies in the greater develop- ment of air-spaces in bellidifolia and the greater width of the cork layer. The distribution of tannin is similar in the species to that of 5. binervosa. 3. The Leaf (cf. p. 249). The general structure of the leaf is shown in Text-fig. 15, F ; it is dis- tinctly fleshier than in any of the forms of S. binervosa , and is at once distinguished from them by the absence of sclereides in the blade and petiole, though they occur in the leaf sheath ; sclerenchyma fibres are also very few in number, and are present only as a small cap above and below the main veins. The details of the leaf and petiole structure are shown in Text-figs. 15 D, 19 E, 20 C, 21 H, and 21 H 1 ; they have been described on pp. 266-73 in comparison with the leaves of S. binervosa . 4. The Inflorescence Axis (cf. p. 25c). The general structure of the inflorescence axis resembles that of 5. binervosa (see pp. 273-6), but the number of bundles in the inner ring is usually much fewer. As in binervosa , frequent anastomoses take place between the small cortical strands and also between the bundles of the inner and outer ring. The main bundles of the scale leaves are furnished from one of the inner ring of bundles, and the remaining bundles are provided by the cortical strands. Details of the structure of the axis are given on pp. 273-6, and are illustrated in Text-figs. 23 C, 24 A and B, and 25 B. IV. Comparison of the Anatomy of the Broad-leaved S. BINERVOSA AND ITS POSSIBLE PARENTS. Narrow-leaved S. binervosa — A. S. bellidifolia = B. ? Hybrid , broad-leaved S. binervosa — C. Root : Type Periderm Cortex Sclerenchyma and sclereides Phloem (12 A : 10) Shingle Present Not aerating Abundant Very narrow zone (26) Marsh Wider zone than A Aerating Absent throughout the greater length of root Wide zone Shingle Present Not aerating Less abundant than in A Intermediate between A and B Xylem: ' Fibres Wood parenchyma 1 Primary vessels k Secondary vessels Numerous Absent | o*oio mm. diameter 1 0*020 mm. diameter Few Abundant 0*006 mm. diameter 0*028 mm. diameter Numerous Absent 0*012 mm. diameter 0*032 mm. diameter U 2 280 de Fraine. — The Morphology and Anatomy of the A 7 arrow - leaved S. binervosa = A. S. bellidifolia ~ B. ? Hybrid, broad-leaved S. binervosa = c. Stem : (14 ») (27) (14 c) Type Short and subterranean Short and subterranean Short and subterranean Periderm Wide Wider than a Wide Cortex Not aeiating Slightly aerating Not aerating Sclerenchyma and sclereides Very abundant Very few Abundant Phloem Narrow zone Wide zone Narrow zone Medullary rays Narrow Wide Narrow Xylem : ( Fibres Numerous Numerous Numerous J Primary vessels 0*006 mm. in diameter 0*006 mm. in diameter 0.008 mm. in diameter I Secondary vessels o*oi 6 mm. in diameter 0*020 mm. in diameter 0*020 mm. in diameter Pith Many sclereides Few sclereides Intermediate between A and B Leaf'. Type Isobilateral (19 a) Bifacial (19 e) Bifacial (15 n) Thickness Thin. Average = 0*448 mm. Thicker than C Intermediate between A and B. Average = 0*^64 mm. Sclereides in blade Abundant (k a) Absent (r5 f) Few (15 f.) Sclerenchyma Much Little Intermediate between A and B Margin Long and narrow (20 a) Short and broad (20 c) lake a, but thicker (20 b) Cuticle : (21 B, B 1) Thick, ridged, plugs be- (21 H, H 1) Thick, less ridged, no (21 G, G 1) Thick, ridged, plugs ( External tween the cells plugs present l Internal Present Traces only. Epidermal cells smaller than in A and c Present Glands : 1 Mucilage Abundant Less abundant Abundant •1 Mettenian Uppersurface,6*2 persq.mm 9*6 8*3 1 Lower ,, 8-9 „ ,, 8*3 9*6 Stomata Upper surface, 33*2 per sq. mm. 73M 40* r Lower surface, 26*3 per sq. mm. 35*9 35*2 I'etiole : Type Flat (i6j\) Channelled (16 d) Flat (16 b) Sclerenchyma Abundant Little Less abundant than in A Sclereides Numerous Absent, except in the Less numerous than sheath in A Inflorescence axis : Surface Smooth Very granulated (24 a, b ; 25 B) Two layers (23 c) Slightly granulated (23, A) Palisade layer One layer (23 a) Two layers (23 B) (The numbers and letters in round brackets indicate the diagram illustrating the character in question.) V. Summary. i. The morphology and anatomy of the forms of binervosa and *S. bellidifolia which occur at Blakeney Point, Norfolk, is described. Three main forms of binervosa are distinguished : (a) The tall form which grows on the edge of shingle fans on the Main bank (pp. 241 and 244). Genus Statice as represented at Blakeney Point. I. 281 (b) The dwarf form characteristic of stable lateral banks (pp. 240-1) (the typical narrow-leaved form), but which also occurs on the margin of muddy lows (mud plants), and of sandy-mud lows (sandy plants). (c) The ? hybrid form between .S', binervosa and 5. bellidifolia (p. 245). The habitats of all the forms are described, and the ecological factors involved and their possible effects on the plants are considered. Experi- mental evidence is brought forward to show the temporarily stimulating effect of shingle on growth — the effect is due to mulch action (pp. 240-6). 2. Glands (pp. 251-7). The structure, number, and distribution of the two kinds of glands characteristic of the order are described. The mucilage glands function in preventing desiccation of the apex ; the Mettenian glands (the so-called chalk glands) excrete water, but probably only function when the amount absorbed by the roots is greater than the rate of transpiration. Some explanation of the numerical differences which occur is given. 3. The Seedling (pp. 257-8). The morphology and the seedling structure of S', binervosa is de- scribed ; the method of the transition from root to stem follows van Tieghem’s Type 3. 4. Root (pp. 258-63 and 276-7). The structure of the root in all the forms, and of the plants cultivated from seed, is described. In S. binervosa the structure is in every way adapted to withstand the pressure of shingle and is admirably suited, both internally and externally, to life in an habitat characterized by scarcity of water. The root of S', bellidifolia , on the other hand, resembles that of many salt marsh plants, and shows none of the characters of stabilized shingle plants. Differences between the roots of the two species are chiefly seen in the following features : (a) The proportion and distribution of the stereome-sclerenchyma and sclereides. (b) The nature of the xylem elements. (c) The character of the cortex. The effects of the differences in the habitats of the various forms of S. binervosa are considered ; the chief variations occur in : (a) The annual growth rings. (b) The development of wood parenchyma. (< c ) The proportion of sclerenchyma and sclereides, (d) The size of the vessels. 5. Stem (pp. 263-6 and 277-9). The stem in both species is short and subterranean ; its outstanding characters are the abundance of fibres in the wood and the nests of sclereides in the pith and cortex. The differences obtaining in thq various 2 &2 de Froine . — Genus Statice as represented at Blakeney Point . /. forms of 5. binervosa are considered ; they are of a similar nature to those present in the root and are due to similar causes. 6. Leaf (pp. 256-73 and The structure of the leaf is bifacial in .S'. bellidifolia and in the hybrid S. binervosa , but is isobilateral in all the other forms. Differences occur in the abundance of the sclereides present ; they are numerous in the typical binervosa , absent except in the leaf sheath in bellidifolia , and few in the ? hybrid. The effects of the various habitats on their production is ex- amined, and the influence of culture under favourable conditions in diminishing mechanical tissue (as in the root and. stem also) is pointed out. The effect of culture on the development of the intercellular space system, and on the production of cuticle, indicates a definite reaction of the plants to the external conditions. The degree of development of the cuticle, both internal and external, appears to depend on the water relations of the habitat in all the forms examined. The stomata have three subsidiary cells in .S', binervosa and four in S. bellidifolia ; they occur on both surfaces of the leaf, and details of their distribution are given. 7. Inflorescence axis (pp. 27 3-6 and 279). An inner series of large, collateral bundles and an outer series of small cortical ones occur in the axis. The inner bundles are embedded in lignified fibres and are surrounded by a zone of sclereides. The cor- tical bundles lie at the margin of a narrow assimilating zone, at the outer edge of which is a palisade layer of one (dwarf forms of binervosa) or two cells (? hybrid binervosa and bellidifolia). In the sterile branches of bellidifolia and the ? hybrid binervosa the palisade cells become deeper and the number of layers is increased. The proportion of stereome diminishes in the smaller axes. The surface of the inflorescence axis is smooth in .S', binervosa , slightly rough in the ? hybrid form, and distinctly scabrid in 5. bellidi- folia ; the roughness is due to the enlargement of the cells surrounding the Mettenian glands. 8. The anatomical characters of the ? hybrid and its two parent forms are summarized (pp. 279-80). 9. The floral morphology of 5. binervosa and the ? hybrid form, and of S. bellidifolia , is fully described (pp. 246-50).1 It is a great pleasure to acknowledge my indebtedness to Professor F. W. Oliver, both for the help he has given me in connexion with the experiments and also for specimens of plants collected at various times.2 1 The author is entirely indebted to Dr. E. J. Salisbury for this section of the paper. 2 The investigation was partly carried out in the Ecological Laboratory at Blakeney Boint, Norfolk. * Studies of Protoplasmic Permeability by Measurement of Rate of Shrinkage of Turgid Tissues. I. The Influence of Temperature on the Permeability of Protoplasm to Water. E. MARION DELE, Yarrow Fellow , Girton College , Cambridge. With seventeen Figures and five Tables in the Text. Table of Contents. PAGE Section I. Introduction . . . 283 ,, II. Apparatus and Regula- tion of Temperature . 284 ,, III. Procedure in a Typical Experiment . . . 286 „ IV. Characteristics of the Material used . . 2S9 A Onion Leaves . .289 B. Dandelion Scapes . 292 ,, V. Choice of a Solution for investigating Rate of Plasmolytic Shrinkage page Section VI. Rate of Plasmolytic Shrinkage in a Subtonic Solution .... 295 A. Leaf of Onion . . 295 B. Scape of Dandelion 300 VII. Critical Consideration of the Relation between Permeability and Tem- perature put forward BY F. VAN RYSSELBERGHE 305 VIII. Summary and Conclu- sions .... 308 55 293 Introduction. THE effect of temperature on the permeability of protoplasm to water has much interest biologically, but no critical work dealing with it has hitherto been published. It was known that, generally speaking, root absorption is retarded by low soil temperatures, and Wieler states that the bleeding of vine stems is accelerated eight times by a rise in temperature from 8° to 420 C. The first attempt to measure the tempera- ture effect was the work of Krabbe in 1895, and was based on the plasmolysis of cylinders of turgescent pith tissues at different tempera- tures, the time required for the total contraction in any given solution varying with the temperature. This method was adopted by van RysseE berghe in 1901, and extended to the consideration of the passage of water through the protoplast both in plasmolysis and deplasmolysis. According [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] 284 Delf. — Studies of Protoplasmic Permeability by to van Rysselberghe, the effect of temperature is accelerating only up to 200 C., but this conclusion is open to criticism both on experimental and on theoretical grounds, as will be shown in the course of this paper. It was suggested to me by Dr. F. F. Blackman that the plasmo- lytic method of investigating the effect of temperature on the permea- bility of protoplasm to water could be made more effective by the use of seme magnifying arrangement enabling the observer to follow the process in all its stages. This has been achieved by means of a particular form of optical lever devised by him, which gives a magnification of 350 diameters on changes in length of short strips of tissue, which though fixed in position are not subjected to any strain. I am indebted to Dr. Blackman not only for the use of this apparatus, but for much help and suggestion throughout the research. The work was carried out in the Cambridge Botany School in the year 1915, while holding a Yarrow Research Fellowship at Girton College. Section II. Apparatus and Regulation of Temperature. The method of measuring the rate of tissue-shrinkage is based upon the great magnification of alterations of length of strips of tissue that can be obtained by the principle of the ‘ optical lever V The short lever C, which the plant actually displaces, carries a tiny mirror on which the image of a cross wire is projected by a Nernst burner (v, Fig. 1) in a tube. From the mirror the image is reflected on to a millimetre scale W, and a magnification of x 350 is thus attained. The fulcrum of the optical lever is a small aluminium cylinder free to rotate on a horizontal axis, the bearings being steel points working in agate cups. From the cylinder adjustable wires project at right angles to the axis for a few mm., and to these the threads connecting to the plant tissue below on one side and to the counterpoise D below on the opposite side are attached. The counterpoise is a few mm. of fine wire and so adjusted that it will just assure a downward movement of its own side when the thread connecting to the plant on the other side is cut. There is thus always a* minute extension-strain on the plant tissue. The clamp for the plant below is carried on the same ‘ invar ’ rod, E, as the lever, and the whole is supported from above independently of the rest of the apparatus. The clamp is designed to hold a tubular structure such as part of an inflorescence axis or a cylindrical leaf, so that water or solution may be passed through it continuously, and bathe the delicate pith cells which line the interior. The plant cylinder A (Fig. 1) is firmly fixed to a glass nozzle G of narrow bore, by winding it round with cotton fastened below to a small rectangular block of cork through which the nozzle passes. The 1 The principle of the optical lever has been previously used by Professor Bose for records of minute movements of plants; see J. C. Bose: Plant Response, 1906. Measurement of Rate of Shrinkage of Turgid Tissues . /. 285 cork Is gripped by a metal clamp, F, sliding horizontally in and out of a horizontal arm of the vertical support E. The nozzle and plant cylinder are suspended vertically in a conical glass chamber formed of an inverted Erlenmeyer flask with the base cut off. To the upper end of the cylinder is attached a slender glass hook B, the base of which pierces the cylinder horizontally, while the shank stands Fig. i. vertically and centrally above it. The upper end of the hook carries a loop of very fine cotton, waxed to prevent torsions, and attached at its upper end to the projecting wire arm of the optical lever. The upper end of the glass hook is bent sharply in a plane perpendicular to that of the piercing base to avoid slipping. The glass nozzle can be adapted by means of rubber tubing to take tissue-cylinders of 1 to 8 mm. diameter. The horizontal part of the nozzle is connected by rubber tubing with a glass spiral, H, contained in the flask chamber. Three tubes pass through 286 Del/. — Studies of Protoplasmic Permeability by a cork in the neck of the flask, two of which are governed by the three-way tap at K, whilst the third, I, is an outflow, the level of which at M determines the rate of flow of the liquid through the chamber and coil ; the side tube O is an additional outlet for more rapidly emptying the chamber. By means of the tap K a liquid entering through z can be diverted at will to the tube leading to the coil H or to the tube J which acts as a by- pass ; or the tap can be turned so that no liquid can enter from Z. The whole flask is packed in cotton-wool or ice for high and low temperatures respectively, and is encased in a wooden box to diminish further loss of heat by radiation. The box is raised on a wooden block P during an experiment when the plant cylinder is immersed in water in the conical flask. The block can be removed by a handle when the chamber is to be lowered ; the nozzle can then be adjusted or removed, together with its clamp, to receive a new piece of tissue. When an experiment is being made, a thermometer is also suspended in the chamber so that its bulb stands close by the plant cylinder. (In the figure this is shown on the opposite side.) The flow of liquid entering z is determined by two three-way taps placed one on either arm of a double Y-tube and connected with delivery tubes x and Y leading from the solutions. X and Y are further connected with long flat coils of glass tubing (not shown in the diagram), and both are surrounded with a large water-bath, Q, kept at constant temperature by means of a thermostat. A current of water or solution passes from the aspirators along one of the coils where it acquires the temperature of the water-bath. It then enters the tube X or Y and can be diverted at will to either the outer tubes T or s leading to the chamber, or to the inner tubes U or R leading to the outflow N. It is therefore possible to have two streams of liquid running at the same rate, traversing an equal distance through the same water-bath and hence acquiring the same temperature ; and either of these liquids can be turned through the plant cylinder immediately without any perceptible change in temperature. Section III. Procedure in a Typical Experiment. The optical lever was first adjusted relative to the source of light so that a sharply defined cross wire in a bright field of light was thrown upon the distant millimetre scale ; the lever, light-source, and scale remained in these positions throughout the experiments and together gave a magni- fication of nearly 350 diameters. This was estimated directly by measur- ing the deflexion of a spot of light on the scale when a vertical thread attached to the middle arm of the optical lever was depressed (by hang- ing on it a very small piece of fine wire) through a distance of 1 mm. as measured by a vertical scale immediately behind it. A small counterpoise attached to the free arm of the optical lever was adjusted by trial until the tension in the plant cylinder, when fixed Measurement of Rate of Shrinkage of Turgid Tissues. /. 287 in position, was reduced to a minimum. That this was accomplished was evidenced by the fact that even the most flaccid tissues when left attached to the lever showed no sign of the slow creeping extension that an appre- ciable elongating strain must produce. Whilst very sensitive to any change in the length of the plant cylinder, the apparatus would give the same reading for many hours when a turgid piece was kept at constant temperature and freely supplied with water. In all the experiments, the freshly gathered material was kept in slowly running tap-water for some hours before use in order to ensure that the material was turgid. In ordinary cases one or two hours in water was sufficient for this purpose, but in very dry weather six hours or even more would be necessary. This previous immersion appeared to have no effect on subsequent treatment with sugar. When first placed on the apparatus distilled water was passed through the plant cylinder to see whether any further intake of water was occurring ; the sugar solution was only passed through when the length of tissue showed a practically constant reading with the distilled water. For an experiment at any temperature other than that of the laboratory the water-bath and chamber were previously adjusted so that the latter was at the desired temperature with distilled water flowing through the apparatus. Several trials were made beforehand, and it was found that, apart from variations in the temperature of the laboratory itself, the temperature of the water-bath had to be well above that desired in the chamber, e. g. from 30 at lower temperatures to 1 50 C. at highest temperatures. The control of the temperature was effected by (a) cooling or heating the water-bath, and (b) regulating the rate of flow of liquid through the apparatus. A temperature of 50 to 6° C. was produced in the chamber by packing the water- bath with ice and salt, and keeping melting ice in the chamber ; by this means a constant temperature could be maintained for hours. A temperature of 8° to ]o° C. was obtained by keeping the water- bath cool with ice and salt, but leaving the chamber to establish its own equilibrium. At high temperatures the water-bath was heated by means of a gas flame regulated by a thermostat, and the rate of flow of the two solutions to be used was carefully adjusted beforehand, so that at a known rate for each solution the temperature of the chamber would be maintained. The regulation of the flow of liquid was achieved by adjusting the pressure under which the liquid was driven through the apparatus, aided by the various taps, which could be turned partly off if necessary. The sup- plies of liquid were contained in ‘ aspirators ’ — Marriotte bottles — which give a constant flow of liquid ; so adjusting the height of the exit-tube, M, of the chamber determined the rate at which, other things being equal, the liquid would flow through the apparatus. With practice it became possible to keep the temperature of the chamber constant during 288 Del/,— Studies of Protoplasmic Permeability by an experiment within i° C including the change from water to the plasmolysing solution. When the chamber temperature was constant, the tap K (Fig. i) admitting the water current was turned off, the outlet M of the chamber blocked, and the chamber lowered by removing the block of wood P upon which it stood. The metal clamp containing the glass nozzle was removed from its socket, and the piece of tissue was then fitted to the nozzle and bound tightly. The nozzle was then inserted in the clamp, the clamp fixed in position again so that the material was vertical, and the glass hook inserted so that the spot of light from the mirror occu- pied a convenient position on the scale. The chamber was then raised, and the tissue therefore at once submerged in the water of the chamber at the required temperature. The tap K was at once turned to admit water, the exit-tube M unblocked, and a reading of the scale taken. The whole process of fixing on the plant after the temperature was estab- lished was a matter of about one minute, so that there was no unneces- sary preliminary exposure of the tissue before the beginning of the observations. During the course of an experiment, readings of the scale were made at frequent intervals, concurrently with readings of the thermometer in the chamber. At high temperatures the readings were often taken every half-minute when first the plasmolysing solution was turned through the apparatus. Just before the sugar solution was to be given to the plant, and whilst the latter was still supplied with water, the sugar current was circulated through the path Y, R, N (Fig. i) and out to a waste receiver. The rate of this current was adjusted until it was equal to that of the water current through the chamber, and in a few minutes its tempera- ture was that of the water leaving the water-bath. The water current was then diverted from the plant to the ‘ by-pass ’ J (Fig. i) by the three-way tap K and the taps of X and Y (Fig. i) reversed so that the paths of the sugar and water currents were interchanged. The sugar reached the chamber tube z in thirty seconds, and the tap K was turned so that the sugar passed up the coil and through the plant. Immediately after turning the tap K a reading was made of the scale, and in about thirty seconds, when the denser sugar solution could be detected stream- ing over the edge of the cylinder, a second reading was made. During the replacement of water by solution there was no variation in the position of the spot of light on the scale. In some cases after plasmolysis had been observed, the water current was again turned through the plant, in order to observe the course of deplasmolysis, Measurement of Rate of Shrinkage of Turgid Tissues . I. 289 Section IV. Characteristics of the Material used. The leaves of onions and the scapes of dandelions were used through- out the experiments, since these are readily obtainable, of suitable diameter, fairly uniform in structure, and not too rigid. In order to ensure as far as possible the supply of comparable material, a number of scapes and leaves were marked and kept under observation. An examination was also made of their structure to see what tissues were mainly concerned in the plasmo- lytic shrinking. It was clear that in both cases the cavity of the interior is well lined with thin-walled living cells which would be freely exposed to the action of any solution flowing through it. A. Onion Leaves. Onion leaves were used from plants in their second season of growth, i. e. from full-grown bulbs which had been planted some weeks previously. The leaves grow most actively soon after they have burst through the Fig. 2. Diagrammatic representation of cross section of middle region of leaf of onion, pal. palisade tissue ; cp. colourless parenchyma ; v.b, vascular bundle. x 7. Fig. 3. Transverse section of middle region of onion leaf, drawn with Zeiss D.D. and 5*5 objective (Beck). sheath which at first surrounds them. At this time the growth involves the entire length of the leaf, but after four or five days it is practically restricted to the basal region. This basal region continues to grow slowly for a week or more, according to the age of the plant. When all growth in length has .ceased, a leaf generally falls over, making a sharp bend in the middle or near the base, and though such leaves may live for many days they were generally more or less flaccid and were avoided for experimental purposes. The onion leaf is a hollow cylindrical structure for the greater part of its length, but the central cavity is filled with a delicate parenchymatous 290 Delf. — Studies of Protoplasmic Permeability by tissue at the extreme tip and the base. The leaf has. no stereome and is sup- ported by the turgor of the thin-walled cells, more or less aided by the framework of slender vascular bundles. The epidermis is cuticularized and has particles of wax embedded in it which prevent the leaf from being easily wetted externally, even when kept under water for many hours. Imme- diately beneath it are two layers of typical palisade cells, arranged with their long axes perpendicular to the length of the leaf and occupying more than half the thickness of the tissue. Beneath these and at right angles to them is a single layer (or more in the region of the bundles) of shortly rectangular cells (a, A, Figs. 3, 4), containing fewer chloroplasts than the palisade cells, and within these are two or three rows of much more elongated cells (b, b; in Figs. 3, 4), the innermost of which are colourless Fig. 4. Longitudinal section of middle region of onion leaf, taken between the bundles ; drawn with Zeiss D.D. and 5-5 objective (Beck). and contain only a watery plasma with a few degenerated plastids. On the interior of these cells are the collapsed and broken remains of the paren- chymatous cells which once occupied the central cavity. The mean isosmotic equivalent of the cell-sap was determined by applying the tissue- tension method of de Vries to the tissues previously well soaked in water, and was found to vary from 0*20 to 0*23 grm. M. cane- sugar, according to the age of the leaf, the older leaves giving the higher values. By repeated microscopic measurement it was found that at these strengths the only cells which show any sign of plasmolysis were the vertically elongated cells internal to the palisade (A and B in Fig. 4). For this purpose a longitudinal section was cut from a turgid piece of leaf mounted in water, and an uninjured cell of the interior selected and measured by a micrometer scale and a high-power lens. The section was supported under the cover-glass by two strips of paper, and was irrigated with solution of any concentration by means of a narrow strip of linen dipping into a beaker and resting on the stage of the microscope. Measurement of Rate of Shrinkage of Turgid Tissues . T 291 Measurements were made at intervals, and when a contraction seemed com- plete a somewhat stronger solution was substituted. By this means it was found that the cells of the same section behaved differently in the same solution, and some light was also thrown on the behaviour of the tissues with hypertonic solutions. In one experiment, in which the section was irrigated successively with solutions of 0-33, 0*25, and 0*30 grm. M. sugar, cells corresponding in position to B, B (Figs. 3 and 4) were observed, and each passed through the phases illustrated in Fig. 5. Now the cells of which this cell is typical are just plasmolysed by the solution 0-23 grm. M. and have no longer therefore any turgor, yet on the addition of stronger solutions they undergo further shrinkage. Their walls In water After 2 hours •23gm M.sugar After 1% hours •25gm. M.sugar After 2 hours •30gm. M.sugar Fig. 5. A single cell from a longitudinal section of a turgid onion leaf, showing stages in plasmolysis with successive sugar solutions, drawn to scale, (r) In water. (2) Alter 2 hours 0*23 grm. M. sugar. (3) After hours 0*25 grm. M. sugar. (4) After 2 hours 0-30 grm. M. sugar. first showed signs of crumpling after some hours in 0-3 grm. M. sugar, and with still stronger solutions the walls collapse altogether upon the shrunken protoplast. This complete collapse of the inner cells appears to be due in the first instance to the constraining effect of the palisade cells, which were all plasmolysed by the 0-3 grm. M. solution, but it also probably indicates an imperfect permeability of the cell-walls for sugar molecules. In any case, the cell-wall does not always remain completely extended after the proto- plast is withdrawn fiom it, as is often assumed to occur in the plasmolysis of plant cells, but with hypertonic solutions other factors cause its further contraction. This leads to a prolonged shrinking of the whole tissue with hypertonic solutions, which is not primarily a question of plasmolysis, but is a sort of mechanical ‘ settling down ’ of the cellulose walls. 292 Delf — Studies of Protoplasmic Permeability by B. Dandelion Scapes. Pith Cavity Endodermis Epidermis Fig. 6. Diagrammatic representation of transverse section of middle region of dandelion scape at flowering period, x 10. It was found that a dandelion scape finishes its most active growth in length just before the opening of the inflorescence. In the cases observed, the flowers remained open for two days, and during this time there is slight growth in length confined to the uppermost tapering part of the scape. After flowering, the bracts close over the inflorescence, and in about a fortnight they reopen for the dispersal of the fruits. During this fort- night there is active growth in length in the basal region of the scape, and a certain amount of increase in dia- meter is also attained — by tangential stretching rather than by meristematic ac- tivity. In collecting material, lengths were always cut from the basal region of a straight healthy scape, bearing un faded flowers. They were transferred to a beaker of water on the spot, and were supported vertically to avoid geotropic effects during the in- terval between collecting the material and making an experiment. The scape has a very simple structure and, excepting for the epidermis and the l ing of vascular bundles, is entirely made up of parenchymatous cells, with numerous intercellular spaces between them. The epidermal cells are thinly cuticularized. Two or sometimes three layers of hypodermal cells are narrow radially, elongated vertically, and thickened with cellulose on their tangential walls. The remaining cortical cells are thin-walled and rather longer than broad ; the pith cells are rect- angular with square ends and at least twice as long as broad. The pith cavity is clearly formed by the rupture of the central cells at an early stage, and the remains of their torn walls adhere to the inner living pith cells. The outer cortical cells contain numerous scattered chloroplasts, but starch Fig. 7. Parc of trans- verse section of middle region of dandelion scape taken be- tween the larger bundles ; drawn with Zeiss D. D. and 5*5 abjective (Beck). Measurement of Rate of Shrinkage of Turgid Tissues. I. 293 was never found in any abundance except in the endodermis. In an old scape bearing the nearly ripened fruits, no starch at all could be detected except in the endodermis in the uppermost region of the scape. There is a ring of vascular bundles embedded in the ground tissue, and laticiferous elements occur in the pericycle, but no interfascicular cambium develops, and for the purposes of this research these tissues can be disregarded. Thin strips of turgid dandelion scapes show well-marked tissue tensions. By the method of de Vries the mean isosmotic strength of the cell-sap was found to be equivalent to a solution of 0*43 grm. M. cane-sugar at the beginning of the flowering season (April in 1915), and 0-53 at the end (late June). This was confirmed by microscopic observation, but no exact measurement of single cells was made as for the onion. Section V. Choice of a Solution for investigating Rate of Plasmolytic Shrinkage. For the plasmolysing solution it is desirable to use one which neither injures nor penetrates the protoplast. The solution which has most frequently been used is cane-sugar, which fulfils both these conditions more or less satisfactorily, and can be obtained readily in a pure form. The choice of a suitable concentration is a matter of more difficulty. The simplest course seemed to be to find the isosmotic equivalent of the cell-sap and to use a solution which was j ust hypertonic to this. By examining thin strips of dandelion scapes and onion leaves after the manner of de Vries, it was found that a sugar solution of 0-5 grm. M. was j ust hyper- tonic to the cell-sap of the dandelion, and 0-3 grm. M. was hypertonic to that of the onion. These solutions were first used for the various tempera- tures, but owing to the difficulty of deciding when the contractions had ended, and to the variable forms of the curves obtained, both stronger and weaker solutions were subsequently experimented with. The solutions thus investigated were respectively 0-731, 1 * 0-5, 0-3, and 0-18 grm. M., all being made up to * weight-normal 5 standards. The whole contraction produced by these solutions was measured by a microscope on a micrometer screw travelling horizontally, as well as by the optical lever apparatus, and was found to vafy somewhat with the age of the material, but approximate values are given in Table I. Sugar solution. 0-18 grm. M. 0-30 grm. M. 0-731 grm. M. Table I. Dandelion. Contraction % C ontradion of {scale divisions'). original length. 15-40 c* 2-0-4 % 2CC-25O 2-2*5 % Onion. Contraction {scale divisions'). 20-50 200-350 % Contraction of original length. °-3-°\5 % 1 This strong solution (25 %) was largely used by van Rysselberghe X 3% 2Q4 Delf.- — Studies of Protoplasmic Permeability by It will be seen that there is a much greater shrinkage with strong than with weak solutions, but that the contraction is easily measurable with the weakest solution employed. In dealing with plasmolysis rates at different temperatures, it is necessary to have some definite standard of comparison. The length of the pieces used varied from 28 to 30 mm. — being usually nearer 29 mm. ; the absolute contraction, therefore, was necessarily also variable. It seemed best to relate the total linear contraction observed in each ex- periment, therefore, to a fixed length which was taken as 100, the observed contractions at any point being thus expressed as percentages of the total contraction. The ordinates in Figs. 11, 13, 14, are of this type. Fig. 8. Curves comparing effect of 018 and 0-731 grm. M. concentration of sugar on the plasmo- ly lie contraction of onion leaves, plotted to same scale. Ordinates are divisions of scale. For this method it was necessary to know when the plasmolytic contraction in any solution had ceased, a thing almost impossible to judge in the case of strong solutions which gave a continuous slow con- traction towards the end, for many hours. In Fig. 8 the actual course of contraction is shown of dandelion scapes under identical conditions at 34° C. but at different concentrations of sugar ; curve B shows part of this slow final shrinkage, and curve A on the same scale the slight but definitely ending contraction with a subtonic solution. In addition to this difficulty, with strong solutions there is often a slower rate of plasmolysis at the beginning of an experiment than after the first half-hour. This can only be due to some secondary factor disturbing the normal course of contrac- Measurement of Rale of Shrinkage of Turgid Tissues. I. 295 tion. for the difference in concentration between the cell-sap and the outer solution is then at its greatest and, other things being equal, must give the fastest rate of plasmolysis in the first phase of an experiment. For these reasons it is hardly possible to estimate closely the effect of temperature on the course of plasmolysis produced by means of markedly hypertonic solutions, and in all the critical experiments on temperature effects in this research subtonic solutions alone were employed. Section VI. Rate of Plasmolytic Shrinkage in a Subtonic Solution. It was necessary to make trial experiments in order to ascertain which strength of subtonic solution would be most suitable for the experiments at different temperatures. The general considerations were that the solution should be weak enough to give a definite end-point in not more than two or three hours at ordinary temperatures, and not too weak to give a contraction of less than about 20 divisions on the scale. In practice, it was found that solutions which gave contractions varying from 30 to 50 divisions on the scale gave a definite ending in about two hours at ordinary temperatures, and also gave a curve of nearly logarithmic form. It can be shown, on theoretical grounds, that this is the form of curve which would be expected with dilute solutions apart from any secondary disturbing causes. It was found that the best results were given with solutions of 0-18 grm. M. sugar for the onion and 0-3 grm. M. for the dandelion ; these two cases will be considered separately. A. Leaf of Onion. The mean osmotic pressure of the cell-sap of onion leaves was found to vary with the age of the leaves, those which had only just finished their growth being isotonic with a sugar solution of 0-20 or 0-2 1 grm. M., and older ones with a solution of 0-23 grm. M. (i. e. four to five atmospheres pressure). The original solutions tried were 0-25, 0*20, 0-18, and 0*15 grm. M., but the last was at once rejected on account of the small contractions given with some leaves. Even with great care in the choice of material and in repeating the same experimental conditions, there was a certain amount of variation in the rate of plasmolysis at the same temperature. At medium tempera- tures, therefore, it was usual to perform several experiments under the same conditions, and to select the ones which gave the curve of approximately logarithmic form (e. g. A, in Fig. 13). It was found that at temperatures above that of the laboratory there was a short period of expansion of the tissues when first the material was fixed in place. At temperatures not higher than 30° C. this was 296 Delf . — Studies of Protoplasmic Permeability. followed by a long period of constant length. In experiments with sugar, therefore, at these temperatures the solution was not applied until the initial expansion had ended. At temperatures above 30° C. the preliminary expansion is maintained for a time, but is soon followed by a considerable contraction, probably owing to an escape of cell-sap caused by the increased permeability for salts at high temperatures. This contraction appeared in about two hours in an experiment at 36° C. (Fig. 9, a), and since the plasmolytic shrinkage is completed within half an hour at that temperature, no correction for the temperature effect is needed. Fig. 9. Curves showing effect of temperature on onion in distilled water at 36° C. (a) and plasmolytic contraction at the same temperature, (b) with 0*18 grm. M. sugar. Ordinates are observed scale readings. The contraction at that temperature (Fig. 9, b) was followed by an expansion indicated at E, X (Fig. 9, B), which is almost certainly due to the continued entry of the sugar molecules into the cells, but the expansion was not followed to its conclusion. At temperatures above 36° C. there was a marked contraction appearing after half an hour or even less of exposure to the water-current (Fig. 10, A). It therefore seemed advisable to apply the solution as soon as the initial expansion had ended, to avoid any unnecessary temperature effect. The plasmolytic contraction was finished in ten minutes or less, and the temperature contraction was then again evident (Fig. 10, B, C, D). In all the curves at these high temperatures there appeared to be an antagonism between the tendency to expand with entry of sugar, and to contract as the effect of the prolonged high temperature. An analysis of these secondary Observed Contractions Fig. io. Curves showing effect of temperature on onion in distilled water (a), and effect of subtonic sugar solution (o*i8 grm. M.) (b, c, d). Ordinates are readings on scale. 298 Del/. — Studies of Protoplasmic Permeability by effects has not been made, but by using the first period of contraction immediately after the application of the sugar current, and by correcting for the concurrent temperature effect curves, of plasmolytic shrinkage from 370 to 420 C. were obtained which were of similar form to those at lower temperatures, but correspondingly steeper. The method of correcting the curves is shown in Fig. 10, where the oblique line Y, Z is taken as the base line of the curve. This is, in effect, to subtract from the observed ccn- Fig. 11. Chart ot the course of the shrinkage-time curves of onion leaf at different tempera- tures. All the shrinkages were carried out in subtonic sugar solution (o’iS grm. M. cane-sugar). The individual curves are representative ones from a group taken at each temperature, and in all cases the absolute shrinkage is brought to a standard amount of 100 units as represented by the ordinates. traction at any point the temperature effect, estimated from the curve A previously obtained with distilled water only. It was not easy to repeat exactly the conditions of a high temperature experiment, but close temperature intervals of successive experiments were selected in order to make the results as representative as possible. After thus selecting a suitable curve for each of a number of tempera- tures, these were all plotted to the same scale, the contractions at any point being expressed as percentages of the total contraction in that experiment. These temperature curves for the onion are reproduced in Fig. 11, and the actual contractions observed are given in Table II. Measurement of Rale of Shrinkage of Turgid Tissues . I. 299 Fig. 12. Curve showing the rates of shrinkage of tissue of onion leaf under uniform external osmotic compression but at different temperatures. This curve exhibits the alteration of protoplasmic permeability brought about by temperature. The actual ordinates are the values of the tangents of the shrinkage curves where the 50 % shrinkage line cuts them in series. Table II. Temperature. Observed contraction Time taken to in scale divisions. complete contraction , Min. 5° C. 19 95 i4° C. 2o°5 85 1 90 C. 41 100 26° C. 42'5 100 30° C. 34 35 35° C. 36-5 25 37° C. 20*8 1 2 S8° C. 33 10 *42° C. 28 10 300 Delf. — Studies of Protoplasmic Permeability by The tangents of these curves at any stage in the contraction gives a measure of the rate of the contraction at that stage, and these values found for different curves at the same stage and plotted to the corresponding temperatures as abscissae give the relative rate of plasmolysis. The values of these tangents for the onion are given in Table III. Table III. Rate of plasmolysis at 30 %, 50 %, and 70 % contraction, in onion, irrigated with 0*18 grm. M. sugar. Temperature . Amount of contraction completed. 30 %• 50 %• 7° %• 5° C. 2*8 2*3 o*77 i4° C. 3*4 2*0 i*4 1 90 C. 4-5 3*4 2 *3 26° C. 6*2 5-o 2*7 30° C. 14*3 10*0 f * o 33° C. 26*6 I 2*9 8*5 3/°C. 25*4 19*5 1 8*3 3^° C. 23 18*4 ii*9 420 C. 51-8 36*2 12*1 From this table it can be seen that at all stages of the plasmolytic shrinkage produced by these subtonic solutions, there is a considerable increase of the rate of shrinkage with rise of temperature. The temperature effect seems to be much more marked towards the end of the contraction, but as the endings of experiments were always more liable to possible errors of interpretation than any other part it seems better to take the values at mid-plasmolysis, and these are plotted to the corresponding temperatures in Fig. 12, The coefficients of increase of permeability deduced from the curve are : Temperature range . 5°-i5° 1 0°— 20° i 5s- 2 50 20°-3O° 25°-3r9 30°-40° Coefficient of increase for io° C. i-4 i’5 2*0 2*6 2?9 3*o B. Scape of Dandelion . In order to confirm these results, similar experiments were carried out with the scapes of dandelions, which have a more homogeneous structure than onion leaves. The solution employed was 0*3 grm. M., and was found by trial as for the onion. It was afterwards found that the dilute solutions used for each plant bore the same ratio to the corresponding isotonic solution (1 : 1*4), which is in itself a somewhat striking testimony to the trustworthiness of the apparently arbitrary method of selection of these subtonic solutions. Measurement of Rate of Shrinkage of Turgid Tissues. I. 301 The osmotic pressure of the cell-sap of dandelion scapes was found to vary with the advance of the flowering season, being isotonic with a solu- tion of 0-42 grm. M. at the beginning (i. e. early April in 1915), and with a solution of 0-53 grm. M. at the end (i. e. late June) : this represents a vary- ing pressure of about 9 to 11 atmospheres. During the final experiments on this material, which were made in the latter part of June, the contractions given by the 0*3 grm. M. solution were reduced to only 8 or 10 divisions, and in order to obtain one of the high temperature results it was necessary to increase the strength of the sugar to 0-40 grm. M., when a contraction of 23 divisions was obtained at a temperature of 42° C. Dandelion Plasmolybic Contraction % at 19° C. Fig. 13. Curves showing percentage contraction of dandelions at iq° C. with subtonic solution (0-3 grm. M. sugar). Ordinates are percentages of total contraction. As before, a number of experiments were made at each tempera- ture, and the most typical curve selected to represent each temperature. As a rule the separate individual curves fell very close together when plotted to ordinates of percentage contraction. In Fig. 13 is illustrated the greatest divergence found for any one temperature. In this set are however included curves from material previously long irrigated with distilled water as well as those for short soaking in tap-water. From such a set, the curve A from material soaked two hours in tap-water was selected as the one falling with the simplest logarithmic regularity throughout. 302 Delf. — Studies of Protoplasmic Permeability by In experiments with dandelion scapes it was often noticed that after all plasmolytic contraction had ceased a gradual re-expansion of the tissues occurred, while the tissues were still surrounded with the same sugar solution. This naturally suggested a slow entry of the sugar molecules into the cells, causing an extension of the protoplast again. In order to test this hypothesis, the water-current was turned on again, and in the recovery of the tissues which ensued the extension was always slightly greater than the initial contraction had been. Since every care had been taken to ensure an initial turgidity of the tissues, the only explanation appears to be that there had been a penetration of the sugar molecules, which had therefore raised the osmotic pressure of the cells, and enabled them to take up more water at the same temperature than they could formerly have done. This point seems worth emphasizing, since sugar has always been regarded as the most impenetrable of substances with regard to the protoplast. The experiments at high temperatures are attended with the same difficulties as in the onion, and the curves obtained at temperatures above 35° C. were corrected in the same way for the temperature effect. The collective curves for all the temperatures investigated are shown in Fig. 14, and the corresponding contractions and times are given in Table IV. It can Measurement of Rate of Shrinkage of Turgid Tissues. I. 303 be seen that the general distribution of the curves with respect to tempera- ture is very similar to that for the onion excepting at low temperatures. Table IV. Table of contractions at different temperatures with 0*3 grm. M. sugar. Contraction as Temperature. divisions on Time. scale. 8° C. 3h'5 3 hr. 8 min. 1 4° C. 34 4 hr. 45 min. 190 C. 36 1 hr. 27 min. 240 C. 39 1 hr. 5 min. 26° C. 15 45 min. 290 C. i («) !(*) 26-5 35 „ 320 C. j 3 2 27 -S 28 „ 2( „ 34° C. 29 18 „ 33° C. 8*6 l6 „ 36° C. 1 d)' 1 T 24-2 H „ 42°C. 2 2°5 27 10 *4 >> The tangents of these curves at different stages in plasmolytic contraction are given in Table V, and the values at mid-plasmolysis plotted as before to the corresponding temperatures give the curve repro- duced in Fig. 15. Table V. Temperature. Rates of plasmolytic contraction at 3° to 50 % 7° /o of total contraction. 8° C. !*4 1*0 079 1 4° C. 2*8 1*7 0*96 3 90 c. 4‘3 27 i*9 2 4° c. 5*1 i*9 26° c. 7-2 4"5 2*4 29° c. j («) id) 167 9-8 3*5 32° c. L5*5 i6»6 io*3 I 2-2 3*9 5*2 33° c. 23 ]3'6 9*i 34° c. 15*1 14-9 67 36° c. i 0) Id) 28*9 15*6 Sc9 42° c. 35*9 31’2 22*8 237 J5°5 *77 From this curve for 50 per cent, contraction, the following temperature coefficients may be derived : Temperature Coefficients of increase range. for io° C. 304 Del/. — Studies of Protoplasmic Permeability by Comparison of the Two Plants. Although the two types of material selected are not physiologically or morphologically very similar, their general behaviour in relation to temperature is obviously in agreement. In Figs. 12 and 1 5 the two curves have the same general relations. On closer comparison it will be seen that Fig. 15. Curve of relation to temperature of the rate of shrinkage of tissue of dandelion scape. Interpretation as in Fig. 12 (p. 299). for temperatures of 250 C. and upwards the values of the relative per- meability are practically identical, while below 250 the dandelion values decline faster than those of the onion. It cannot yet be said whether this is related to the fact that the dandelion is a summer plant, the experiments being done in April to June, while the onion leaves had lived through the winter and were investigated from January to March. The agreement and divergence between the two plants is best brought Measurement of Rate of Shrinkage of Ttirgid Tissues. /. 305 out by taking a common relative value of unity for their permeability at the middle point 250 C. We then have the appended series of values : ..o 5 . 10. 1 5°. 20°. 25° 30°. 35° 40°. Onion 0-36 0*44 0*50 o-66 1-0 i-7 2-9 5*° Dandelion — 0*22 0-30 0-5° 1*0 1-9 3-0 5'° Section VII. Critical Consideration OF THE Relation between Permeability and Temperature put forward by F. van Rysselberghe. The only other known published results on the exact relation between permeability of protoplasm to water and temperature are those brought forward by F. van Rysselberghe 1 in 1901. As these results differ com- pletely as to the effect of temperatures above 20° C. they must now be care- fully examined. Studies were made of the time taken to plasmolyse and to deplasmolyse different kinds of material, such as cylinders of elder pith, epidermal cells of Tradescantia , and Spirogyra. Each kind was examined at the temperatures set out below, and harmonious ratios for the relative rates of the observed processes at different temperatures were obtained. The ratios are appended, the rate at o° C. being taken as unity (van Ryssel- berghe, p. 189) : Temperature o°C. 6° C. 120 C. 160 C. 20° C. 250 C. 30° C. Rate 1 i*9 4-5 6-25 7*15 7-5 8 It will be seen that above 20° C. there is practically no further increase in permeability, and the graphic record Fig. 1 6 copied from van Rysselberghes paper shows how this curve agrees broadly with our present results up to 20°, but differs completely in type afterwards. Many of his types of measurement must have been difficult to carry out with precision ; such as the exact time for complete plasmolysis of a cell, and still more so the exact time taken by such a cell to recover its normal state. The data provided for this type of observation do not lend themselves to critical analysis, but the measurements of the shortening of pith are recorded at several successive points of time, and it is proposed to consider them carefully. Long cylinders of elder pith were plasmolysed in strong sugar solution. The procedure was as follows : the cylinders were freshly cut from the plant, freed from traces of wood, and soaked for six hours in water at 160 C., during which time they expanded from 100 mm. to 1 14 mm. The cylinders were then cut longitudinally into two half-cylinders, as equal as possible. One half was placed in sugar solution at o° C. as a control, and the other kept in sugar solution at one of the temperatures specified above. The cane-sugar solution used for plasmolysis was 25 per cent. (0-731 grm. M.). At all temperatures the half-cylinders shrank 1 F. van Rysselberghe : Influence de la temperature sur la perm^abilite du protoplasme vivant pour l’eau et les substances dissoutes. Bull. Acad. Roy. de Belgique, 1901, pp. 175-221. 306 Delf. — Studies of Protoplasmic Permeability by a great deal in length, ultimately 37 per cent., but did not alter appreciably in diameter. At intervals of two hours the lengths of the half-cylinders were measured, and the amount of shrinkage in mm. is recorded (loc. cit., p 184). In Fig. 17 the data are presented in graphic form, the dots being the original data and the broken lines joining them being an attempt to recon- struct the course of the curves over the gaps in the record. In this form the results can be easily compared with the curves set out in *;the body of this paper. Fig. 16. Relation of temperature and permeability of protoplasm. From F. van Rysselberghe (loc. cit., p. 190). It will be seen that/as is to be expected, each curve begins with a more rapid fall, and the rate of plasmolysis gradually declines to zero when the shrinkage is complete at the pith-length of 7 2 mm. F01 the low tempcia- tures the ends of the curves are beyond the twenty-four hours at which the records cease, while for the high temperatures the contraction is practically complete when the first record is taken at two hours. The fiist matter for criticism is van Rysselberghe s method of exti acting a tempei atui e- relation from these data. He takes what may be called a vertical section of the chart ; that is to say, the amount of contraction shown by the respec- tive curves at two hours of time is adopted, the latios being those set out on p. 305. Now it should be clear that the only vertical comparison of the curves which has any real signi ficance is the impossible one that should be taken an instant of time only after the beginning. As each cui ve begins with the same big osmotic pressure force (the difference of concentration of Measurement of Rate of Shrinkage of Turgid Tissues . I. 307 the cell-sap and of the bath of cane-sugar), at that time the ratios of the amounts of contraction would be significant, but as these initially equal forces die away quickly where the curve is steep, and slowly where it slopes gently, they henceforth have all different forces at any point of time, and the effect of such vertical comparisons becomes more and more remote from any significance as time proceeds. It is obvious that at twenty-four hours the vertical comparison would indicate the same contraction for all temperatures of 12° C. and upwards. Fig. 17. Course of plasmolysis of elder pith in 0*731 grin. M. saccharose solution at different temperatures : the data taken from F. van Rysselberghe’s table. The ordinates are actual lengths in mm. To extract the real temperature relation from such a set of curves a horizontal section must be made through the curves, and comparison thus made of relative rates, because where any one horizontal line cuts each curve there the force at work is equal throughout the set. Tangents measured at these points will give a true index of the effects of the respective temperatures. Such treatment must give a different picture of the temperature effect from that presented by van Rysselberghe. The data on the time-relations of deplasmolysis are curiously imperfect and difficult to analyse. The conclusions drawn from them suffer from the same defect as those on plasmolysis. It becomes therefore clear that van Rysselberghe’s conclusions lack any precise experimental foundation. There is a further disturbing matter that shows itself very clearly when these data are presented in a graphic form. It is thereby made obvious that 308 Delf. — Studies of Protoplasmic Permeability by whereas the curves from o° to i6° C. are well separated by temperature effect, the curves for 20°, 250, and 30° C. are practically identical. This van Rysselberghe regards as a true index of the temperature effect. While this physiological view is entirely different from that substantiated in this paper, it is interesting that an appearance of the same result can be got for dandelion scapes by our method, but only by using very strong plasmoly sing solutions . The interpretation of this phenomenon seems to be that there is an absolute upper limit to the rate at which the mechanical tissue-system of cell-walls can collapse and shrink. If so, then however fast and suddenly the protoplasts are crushed by the strong osmotic pressure of 25 per cent, cane-sugar, the cell-wall reticulum — which is what is measured — cannot follow at a rate above a certain limit. In van Rysselberghe’s experiments that limit seems to be reached with 0-731 grm. M. cane-sugar at 20° C. A great part of each contraction curve must consist of the ‘ settling down ’ phase, and over this permeability of protoplasm has no control. The moral to be drawn from these considerations is that, for demon- strating the full accelerating effect of temperature, it is essential to employ plasmolysing solutions which are subtonic , so that the force exerted on the cell is slight. At low temperatures the absolute rate of contraction will be very slow, but it will then be possible to demonstrate the continued accelerating effect of high temperatures without the records being cut off by some mechanical limiting factor which has nothing to do with permeability of protoplasm. Section VIII. Summary and Conclusions. The method of following the rate of shrinkage of strips of tissue bathed in solutions that tend to plasmolyse them has been developed as a means of measuring the permeability of protoplasm to water. Hollow leaves of onion and hollow scapes of dandelion have been selected for investigation because they present on their inside a natural surface of uncuticularized thin-walled cells. Apparatus has been employed by which a current of liquid flows continuously through a short cylinder (30 mm.) of the material under investigation. By suitable arrangements a solution of sugar can be instan- taneously substituted for the water-current and vice versa without any alteration of temperature. An ‘ optical lever ’ is connected to the upper end of the tissue cylinder and alterations in length of 0-003 mm. (one part in ten thousand) are made evident. The change of protoplasmic permeability with temperature between 50 C. and 420 C. has been carefully studied by this technique. Such experiments cannot be satisfactorily carried out with strong plasmolysing solutions, as the shrinkage is then very quick at even moderate Measurement of Rate of Shrinkage of Turgid Tissues. /. 309 temperatures, and there seem to be mechanical limitations to a quicker shrinkage being manifested at higher temperatures. With the weak subtonic solutions adopted, 0-18 grm. M. for onion and 03 grm. M. for dandelion, plasmolysis is not reached, contraction is comparatively slow, and the experimentation is confined to the strictly normal state of falling turgor, without actual separation of the protoplast from the cell-wall. A large number of records with such subtonic sugar solutions have been made at different temperatures. The agreement between individual records at the same temperature is generally good, though the absolute shrinkage dealt with is only o-ia mm. Fig. 13 illustrates the most extreme divergence observed at any one temperature. Sufficient records have been taken to establish that the most regular curves follow a nearly logarithmic course. The most regular medium curve of each temperature group has been selected as representative, and these are charted together — onion, Fig. 11, dandelion, Fig. 14 — all the contractions being brought to the uniform scale of 100 units. These curves show a continuously increasing rapidity of shrinkage with rise of temperature over the whole scale. Up to 350 C. it is found that, when perfused with water alone, the tissue length remains constant for a long period of time. Above 350 C. some shrinkage takes place in water alone, so that the further shrinkage when perfused with sugar is not quite a simple effect. The curves charted for these high temperatures represent sugar-shrinkage less the amount of water-shrinkage during the corresponding time. Such a corrected curve errs, if at all, on the side of being not steep enough. To obtain the true measure of temperature-effect from these curves in the charts, it must be remembered that the rate of shrinkage, as expressed by the slope of the curve at any point, is a product of two factors. Of these, one is the compressing force acting on the protoplast and causing exudation of water (the osmotic force of the perfused solution, plus the elastic pressure of the cell-wall, less the osmotic pressure of the cell-sap) ; and the other, the permeability of the protoplast for water at the temperature prevailing. The influence of the permeability factor can only be obtained from a set of temperature shrinkage-curves by comparing them at really corresponding points, that is at points where the compressing force is equal throughout the set. Such corresponding points are those lying on any one horizontal line on the chart, parallel to the abscissa-axis. For the three lines corresponding to 30 per cent., 30 per cent., and 70 per cent, of the total contraction, respectively, the exact slope of every curve has been ascertained by measuring the values of the tangents at these points on a large-scale diagram. The values of these tangents, repre- senting the relative rates of contraction, are set out for the three series Y 310 Delf. — Shi dies of Protoplasmic Permeability. I. on PP- 3°°) 3°3, and the most acceptable series for the two plants (the 50 per cent, series) are plotted against temperature in Figs. 12 and 1 5. As the force at work was, by selection, equal in all these cases, the differences of rate shown are true measures of the state of the per- meability. It is evident that the augmenting effect of temperature increases to the highest temperature investigated. These curves are in opposition to the generally accepted curve for this relation put forward by F. van Rysselberghe in 1901 (see Fig. 16, p. 306), which indicates hardly any rise of permeability after 20° C. In Section VII van Rysselberghe’s data are critically examined, and it is pointed out that his method of deriving a temperature-relation from his data (plasmolysis of cylinders of pith in 25 per cent. (0-731 grin. M.) cane-sugar) cannot give a true relation. The data themselves are also open to some criticism on account of the very strong plasmolysing solution used. Here the effective force at work is so great that the shrinkage is very quick even at medium temperatures, and there seem to be mechanical limitations to any quicker shrinkage at high temperatures even though the protoplasm becomes more permeable. Further consideration will be devoted to the theoretical significance of the experimental results now brought forward in a later paper of this series. It will suffice at present to give the quantitative conclusions as follows : 1. The permeability of protoplasm for water, as measured by the rate of tissue-shrinkage in a dilute sugar solution, is increased continuously by temperature up to the highest temperature investigated — 420 C. 2. Taking the value of the permeability at 250 as unity, the following relative values are established : Temperature 5° 10° 15 20° 2 5° cu 0 0 35° 4°° Onion 0-36 0-44 0-50 o*66 1*0 i*7 2-9 5*° Dandelion — 0*2 2 0-30 °'5° 1*0 i-9 3*° 5*o 3. An approximate measure of this increase is given by the following sets of temperature-coefficients for increase of permeability with rise Temperature. Onion Dandelion m O _ .O leaf. scape. 5 -J5 i-4 — I0°-20° 1-5 2*3 T5°-2r° 2*0 3*3 20°-30° 2*6 3-8 25°-35c 2-9 3*° 30°-40° 3*0 2-6 A Note on the Vegetative Anatomy of Pherosphaera Fitzgeraldi, F. v. M. BY PERCY GROOM, M.A., D.Sc., Professor of the Technology of Woods and Fibres , Ini ferial College , London. With one Figure in the Text. HEROSPHAERA is a coniferous genus separated by Archer from 1 Dacrydium , but by some botanists placed within this genus. It includes only two species, both Australian— P. hookeriana , Archer, restricted to alpine Tasmania, and P . Fitzgeraldi , F. v. M., only known to occur on the Blue Mountains of New South Wales. The latter species is a dense, prostrate little shrub, whose slender twigs are clothed with numerous narrow, keeled, leaves, which are about 3 mm. in length. In their book ‘ The Pines of Australia Baker and Smith state that the plant is ‘ found at the base of most of the chief falls and add that their material (used in the present investigation) was obtained at the Lower Falls, Leura, in which site the shrub ‘ only grows where it can catch the drips from the falls ’. The remarkable nature of the habitat, recalling that of certain Hymenophyllaceae in tropical forests, caused me to examine the structure of the wood and leaves, in the hope of discovering a conifer truly hygro- phytic in structure. a. The Wood. In its construction the wood of the stem by no means suggests a hygrophyte. In the thickest stem available the diameter was 4*5 mm., yet the number of annual rings present was thirteen. The radial thickness of each annual ring was therefore 0*173 mm., and thus represented 144 annual rings to the inch radius. The thickness of the annual rings increased outwards from the first to the tenth, but decreased in the remaining three rings ; it is therefore probable that the stem had passed its maximum rate of growth in radial thickness. All the tracheides of the secondary wood are thick-walled, and recall those forming the summer-wood (autumn-wood) of such a conifer as Pinus [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] 312 Groom . — A Note on the Vegetative sylvestris , especially as their lumina are narrow. But the inmost tracheides, often exactly one series, of the annual ring show wider cavities than else- where, and thus represent a microscopically thin £ pore-zone 1 of spring- wood. The bordered pits of the tracheides are mainly on the radial walls, where they are by no means abundant, being uniseriate even in the spring- wood ; their fusiform apertures are spirally directed, and in the summer-wood exceed in length the diameter of the chamber. Bordered pits also occur on the tangential walls, more especially of tracheides forming the outer boundary of the annual ring. The medullary rays are all uniseriate and remarkable for their shallow- ness, being usually only i to 4, occasionally 5, cells in height. They consist solely of thin-walled parenchyma, in the lateral radial walls of which the pits are nearly equal to the height of the cell, or each pit of such a kind is replaced by two superposed ones. The Leaf. The leaf is xerophytic in structure. In the first place the epidermis has a thick cuticle and otherwise thick walls, while the stomata are sunken. Secondly, there is a single layer of thick-walled hypoderma, which is continuous except within the stomatic apertures. Contrasting with this tegumentary system is the very loose green tissue lying within and simulating palisade-parenchyma though excavated by a large intercellular system. The centre of the leaf is occupied by a strand that is denser in structure, except that it includes a large median ventral resin-duct. The feature worthy of special note is the extensive development of the transfusion-tissue in comparison with the puny xylem proper. The stomata are restricted to the upper face of the leaf, where they are ranged in several longitudinal rows. In each row the successive stomata arc mostly separated merely by a single short epidermal cell, but nearer the ends of such a row these single short cells become replaced by longer ones, or by only two or several cells. The longitudinal rows of stomata are separated laterally from one another by only two lines of epidermal cells. The communication of the stoma with the internal atmosphere of the leaf is sharply limited. For within each stoma the gap not only in the thick-walled hypoderma, but also in the mesophyll lying within, is very small ; indeed the interruption in the latter tissue is limited to the local separation of either two or three cells, which, in surface section, bound an intercellular space whose oval outline approximately coincides with the periphery of the guard-cells. The hypoderma at certain isolated spots, in transverse section, is as much as three cells in thickness. Anatomy of Pherosphaera Fitzgeraldi , F. v. M. 313 The green tissue immediately within the hypoderma deviates from typical palisade-parenchyma as regards both shape and orientation of the cells and dimensions of the intercellular spaces. The general design of this loose tissue is that of two layers of elongated cells, simulating palisade-cells, and stretching between the hypoderma and the central strand of the leaf. These cells, and particularly those forming the outer layer, are mainly perpendicular to the leaf-surface, but those radiating from the central strand towards the lateral angles (in transverse section) deviate from this design so greatly that in the angles they are actually tangential to the local curved surface. The cells forming the outer green layer frequently are lobed either Transverse section of leaf of Pherosphaera Fitzgeraldi (slightly diagrammatic). laterally or basally, or have dilated bases ; this is especially the case within the upper face of the leaf, so that here in transverse section a number of structures may be seen that apparently are short cells but really are the basal portions of long cells. The green cells of the outer series attain their maximum length towards the upper face of the leaf. The cells forming the inner layer of palisade-like tissue on reaching the central strand may bend so that their inner portions run parallel to the length of the leaf and thus contribute to the outermost layer of the central strand : such cells consequently lose their palisade-like form, and in transverse section are liable to be mistaken for tissue confined to the central strand. The cells forming the outermost layer of the central strand contain chlorophyll, are elongated in the direction of the leaf-axis, and may emit 3 H Groom. — Vegetative Anatomy of Pherosphaera Fitzgeratdi. lateral lobes that serve to connect them with the surrounding tissue. Thus there is no absolute distinction between the cells of this layer and of the inner palisade-like layer. The remainder of the central strand is composed of : (i) parenchyma ; (f) the single median ventral resin-duct, showing a double epithelium ; (3) phloem ; (4) xylem, with transfusion tissue. It was impossible to observe the minute details in the histology on the phloem in the dried herbarium material alone available. The transfusion-tissue (tt. in the illustration) in transverse section extends from the feeble xylem in the form of two wings, which ascend towards the upper face and sometimes curl inwards at the outer edges, which may almost meet in the middle line. The transfusion cells may extend to the outermost layer of the strand. In width of lumen they increase from within outwards ; while in length they vary from long narrow tubes to short cells whose length scarcely exceeds the breadth. Conclusions. Pherosphaera Fitzgeratdi recalls familiar European shrubs and trees growing in peat-bogs, at alpine altitudes, or in arctic regions, both as regards the construction of its leaves, including the xerophytic epidermis and hypoderma associated with very loose mesophyll, and as regards the narrowness of the annual rings. While the dominance of thick-walled tracheides in the stem is paralleled by similar tendencies in species of Pinus , Larix , and Picea when grown at considerable altitudes in Europe, yet the structure of the leaves is not widely different in design from that displayed by various species of Pinus and Juniper us of more low-lying sites. The cause of these anatomical features of Pherosphaera Fitzgeratdi , which can grow in a soaking habitat, demands local investigation. Al- though the Tasmanian species is described as an alpine shrub, such does not appear to be the case with this species. For Dr. Stapf points out that the highest parts of the Blue Mountains, where the plant grows, are clad with Eucalypti, and that the Katoomba plateau (where the. specimen described in Hooker’s ‘ leones Plantarum ’ was obtained) has a general elevation of 3,000 to 3,500 ft. The Morphology of Phylloglossum Drummondii, Kunze. BY K. SAMPSON, B.Sc., Post-graduate Student , Royal Holloway College , University of London. With five Figures in the Text. Contents. PAGE I. Introduction . . . .315 II. Macroscopic Structure . .316 III. General Morphology of the Tuber ..... 318 IV. Vascular Anatomy of Fertile Plants 318 (a) Stelar Anatomy of a large Fertile Plant . . .318 PAGE (/;) Origin, Course, and Form of the Tuber Stele in Fertile Plants .... 332 (V) Leaves connected with the New Tuber 326 V. Vascular Anatomy of Sterile Plants 327 VI. Branching in Phylloglossum 328 VII. Conclusion .... 330 VIII. Summary 33i I. Introduction. JpHYLLOGLOSSUM Drummondii was described for the first time in 1843 by Kunze, who regarded the genus as occupying a position intermediate between the Lycopodiaceae and the Ophioglossaceae, its supposed affinity with the latter family being founded on a superficial resemblance in habit to Ophioglossum Bergianiuni} The following year Roeper associated P/iylloglossum with the Lycopodiaceae, and in that family it has been placed by all later writers.2 While the anatomy and general structure have been dealt with by Mettenius,3 Bertrand,4 and Bower,5 no satisfactory conclusion has been reached in regard to the morphology of its annual tuber. Mettenius com- pared the yearly tuber with that of Orchis and the dropper of Tulipa , but, on the ground that these organs are structurally different, he maintained that the comparison was of little use. Bertrand described the anatomy of Phylloglossum in great detail, but he left the problem of the morphology of its tuber unsolved. Bower followed Treub in comparing the yearly 1 Kunze : Bot. Zeit., 1843. 2 Roeper: Zur Flora Mecklenburgs, ii, 1844. 3 Mettenius : Bot. Zeit., 1867. 4 Bertrand : Areh. bot. du Nord de la France, Nos. 30-34, 1884. 5 Bower : Phil. Trans., 1886. [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] 3 1 6 Sampson. — The Morphology of growth of Phyllogl os sum with the embryonic form of Lycopodium ccrnuum , noting a resemblance between the tuber of the former, and the ‘ proto- corm ’ of the latter.1 His conception of Phylloglossum as ‘a permanently embryonic form of Lycopod ’ received wide acceptance.2 Wernham, drawing an analogy between the tuber of Phylloglossum and the droppers of certain Monocotyledonous plants, suggested that like these it is partly foliar and partly axial. His investigation, however, was limited to two specimens.3 The morphology of the annual tuber of Phylloglossum is, then, an open question. Moreover, such solutions of the problem as have been put forward are hypothetical, and based on analogy rather than on a strict examination of anatomical structure. Since ‘ the question of the position of Phylloglossum chiefly turns upon the view we take of its annual tuber and its protophylls/ 4 it is of exceptional importance that its morphology be understood. When, therefore, by the kindness of Professor Benson a quan- tity of material was placed at my disposal, I decided to discover if a re-investigation of its anatomy might not throw light on the problem of its morphology. The result of this work is set forth in the present paper. II. Macroscopic Structure. The general features of a fertile plant of I^hylloglossum, at a stage previous to spore-dispersal, may be seen from Fig. i. This figure was drawn from a specimen about i-|" high, bearing above the level of the ground five long tapering leaves, and a smooth cylindrical axis (p) terminating in a compact strobilus (s) ; while unbranched, horizontally running roots (r) and two tuberous bodies are buried in the substratum. The shorter and more slender tuber, t, belongs to the growth of the present year, the other ; T, dates from the previous season. The former, at the age figured, appears as a smooth exogenous outgrowth about f long, and slightly swollen at the free end, in which may be seen the outline of an enclosed bud, b. The tuber is an organ of perennation ; it is set free at the end of the season by the decay of other parts of the plant, and after a period of rest germinates, the activities of the enclosed bud producing roots, leaves, frequently a cone, and always another tuber. These organs constitute the yearly growth of the plant. All that remains of the growth of the previous season is the torn sheath, sk, of the old tuber. The general structure of Phylloglossum will be clearer, if reference be made here to the research of Professor Bower on the germination of 1 Treub : Ann. Jard. Buit., vol. 8, 1889-90. 2 For Bower’s more recent views, see Presidential Address to Section K (Botany), British Associa- tion, Australia, 1914. 3 Wernham : Ann. of Bot., vol. xxiv, 1912, p. 335. 4 Bower : loc. cit., 1914, p. 6. Phy l log loss tint Drummondii , Kunze. 317 perennating tubers, by means of which he was able to trace the development of a new tuber, and the origin of the bud enclosed in its free end.1 A new tuber appears first as a projecting mass of tissue with the growing point situated in a depression in the centre. The growing point, however, does not long maintain its central position, becoming first sunk, and then inverted owing to the rapid and unequal growth of surrounding tissue. Finally, it takes the form of a conical mass of meristematic tissue at the base of a narrow channel. This channel, although less distinct in older specimens, never becomes completely obliterated. Its position can generally be recog- nized, at least in transverse sections of a tuber, by the smaller cells which surround it. It had earlier been observed by Bertrand, and called the e canal de Braun \2 This protected position of the bud, and the fact that by intercalary growth it is carried some distance into the substratum, are fea- tures showing a marked specialization for a geophytic habit. The tuber can, however, be regarded also as a means of vegetative propagation, since more than one may be formed during the season.3 In the course of this investigation two different collections of material were ex- amined— the one which I owe to the kindness of Professor F. E. Weiss was obtained from New Zealand, the other was made in South- r West Australia by Professor Benson in mid- sn winter, August 3, 1914. This collection consisted of over one hundred plants, about one-third of which were fertile, showing Fig. i. features similar to those of the specimen figured. Many of the sterile plants were very young, bearing only one or two leaves, but, like the fertile plants, each possessed one new tuber. The New Zealand specimens had evidently been collected later in the season, and, moreover, they were on the whole better developed plants. Five were exceptionally interesting in that each possessed two new tubers. These plants had been set aside by Professor Weiss for further investigation, and were with great generosity handed over to me. 1 Bower : loc. cit, 1886. 2 Bertrand : loc. cit., 1884. 3 Thomas : Proc. Roy. Soc., 1901-2, p. 290. Sampson,— The Morphology of III. General Morphology of the Tuber. The tuber of Phylloglossum has in turn been compared with the adventitious buds borne on the roots of Ophioglossum } the tuber of Orchis ,2 the droppers of certain Monocotyledons,^ and the ‘protocorm’ of L. cernuum .4 The present paper seeks to bring forward anatomical data which throw new light upon the morphology of this plant. It is suggested that Phyllo- glossum is derived from some Lycopod possessing a branched vegetative system, which has become reduced and specialized in the adoption of a geophytic habit. The sharply defined pedicillate cone and the specialized storage tuber are advanced characters, and the juxtaposition of these with the frequent occurrence of feebly developed leaves and the absence of phloem from the stem is evidence that Phylloglossum is a reduced form. Examples of branching in Phylloglossum have hitherto been limited to a few isolated cases of dichotomy of the strobilus, but, as a result of an anatomical study, it is believed that branching occurs at least once in the yearly growth of every fertile plant, namely, on the formation of its annual tuber. The details in vascular anatomy, upon which this view is based, will be dealt with under the following heads : (a) Stelar anatomy of a large fertile plant. (b) Origin, form, and course of the tuber stele in fertile plants. (< c ) Leaves connected with the tuber stele. IV. Vascular Anatomy of Fertile Plants. (a) The Stelar Anatomy of a large Fertile Plant . Considered generally, the most striking features in the internal anatomy of Phylloglossum are the frequently medullated protostele, the mesarch position of the protoxylem, and the marked degradation of vascular tissue. The last point is illustrated by the almost complete absence of phloem, the breakdown of the protoxylem in the leaf-traces and in the stem, and the frequent occurrence of weakly formed tracheides. In Phylloglossum the plan of the stele depends largely on the size of the plant and the number and position of tubers, leaves, and roots, and it is therefore not easy to give a general description. If Phyllo- glossum be a form which has suffered reduction relatively recent in descent, the largest and best developed specimens would be those in which the more primitive features might be expected to occur. It is 1 van Tieghem : Recherches sur la symetrie de structure des plantes vasculaires. 1871. 2 Mettenius : loc. eit., 1867. 3 Wernham : loc. cit. , 1912. 4 Treub : loc. cit., 1889-90. Phy lloglossu m Drummondii , Kunze. 319 for this reason that a description of the best differentiated plant avail- able is given first. The plant in question is shown in Fig. 2, A. It bears on one side a large new tuber above which is situated a leaf, si, somewhat shorter than the normal leaves of the plant (see p. 326). On the opposite side, two small protuberances may be distinguished, one of which represents an abortive tuber, t' , the other, rl, a much reduced leaf. The fact that there is in this plant an attempt to form a second new tuber will be dealt with later (p. 328) ; the importance of the plant in the present con- nexion lies in the anatomy of the well-developed tuber and its relation to the stele of the main axis. Before dealing with this, the general plan of the stele must be mentioned. At the base of a fertile plant the entering root-strands unite to form a medullated protostele, which, interrupted by frequent small gaps, Fig. 2. passes into the peduncle of the cone, and by repeated division supplies the sporophylls. The vascular supply of the new tuber is given off but a short distance above the level of the roots. The leaf-traces, which consist of a single mesarch strand of xylem, are often given off so near the base of the plant that they are in direct connexion with the root- strands. The structure of the present specimen will be clearer if sections are described, proceeding upwards from the base of the plant. In Fig. 3, sects. 9, 8, and 7 (drawn with the Abbe camera), the entering root- strands are seen to unite to form the main stele, which consists, at this level, of anastomosing meristeles arranged round a central mass of paren- chyma. In sect. 6, in position corresponding to the large central stele of previous sections, two steles are cut across, one consisting of a number of meristeles arranged roughly along two sides of a square, the other, a compact, medullated mass of xylem, with a well-marked gap, g, on the inner side. Both steles are associated with the exit of leaf-traces ; in 3 20 Sampson. — The Morphology of the smaller stele one passes off from each edge of the gap. This stele is also connected with the vascular supply of a root. In the same sec- tion (sect. 6) the strand of the tuber, t, is seen, closely associated with the compact central stele. Higher up the two are united, forming a solid mass of xylene which is surrounded by the three leaf-traces, which emerged lower down, and a fourth, which is just being given off (sect. 5). Pky l log! os sum Drummondii , Kunze, 321 In higher sections this solid mass of xylem decreases rapidly in size (sect. 4), and finally dies out (sects. 3 and 2). The bearing of these facts on the relation of the tuber to the main axis is clear from the vertical plan of the stele given in Fig. 4, A, which shows the relative position of the sections just described. On the right- hand side of the diagram is a large tuber, the stele of which makes a sharp upward bend before passing out and down the shaft of the tuber. The result is that sections below the bend cut the tuber stele in two places (sect. 6), while sections at a higher level cut the bend itself, and, therefore, show but a single mass of xylem (sects. 5 and 4). The collec- tion of isolated tracheides in sect. 3 represents the highest part of the bend. The most important features in the stelar anatomy of the plant considered may be summed up as follows : At the base of the plant 322 Sampson . — The Morphology of a medullated stele divides, each daughter stele showing a gap on the inner surface.1 The smaller daughter stele, making a sharp bend, sup- plies the new tuber. The original gap is not continued over the bend, but it appears as the stele of the tuber passes out, giving it a charac- teristic U-f°rrrL From the tuber stele several leaf-traces pass out, con- structed as those which arise from the main axis, and supplying leaves which number among those of the yearly growth. The origin of the tuber stele from the stem, its structure, and its connexion with leaf-traces, point to one conclusion, namely, that it is morphologically a branch. In the following pages it will be seen if these features are sufficiently constant in fertile plants to justify this view. (b) The Origin, Course , and Form of the Tuber Stele in Fertile Plants. In the preceding section of this paper a plant was described, which is to a certain extent unique, at least in the material upon which this investi- gation was made. Its singularity consists, not so much in any particular feature, as in the fact that it combines several features of supreme impor- tance in a morphological study. The aim of the present section is an examination of these features as they occur in other fertile plants. Owing to the diversity in stelar anatomy which fertile plants show, it seems best to begin with a brief description of the tuber stele and its relation to the main axis in three distinct cases. Fig. 5 gives a series of transverse sections of a large fertile plant with a single new tuber. Sect, i, through the base of the peduncle, shows the stele as an almost continuous ring of xylem surrounding a relatively large pith. In sect. 2 this ring possesses a very distinct break at one side, the xylem being somewhat horseshoe-shaped, with cortex and medulla con- tinuous through the gap. In the cortex facing this gap are three leaf- traces, which later become connected by tracheides with one another and with the edges of the gap (sect. 3), with the result that a tube of xylem is again formed of diameter nearly twice that of the ring seen in a higher section. This tubular condition lasts for one or two sections only, as the stele divides to form two arcs of xylem, one of which supplies the tuber, while the other breaks up to form the root-strands (sects. 4, 5, and 6). It is important to note that the arc which passes out into the tuber is that connected above with the three leaf- traces already mentioned. The origin of the vascular supply to the tuber in this plant resembles, in several respects, that of the stunted tuber now to be described. We find, as before, at the base of the peduncle a ring of xylem, in which only small gaps occur (Fig. 3, sect. 1). Lower down a break appears on one side of the stele, giving it a horseshoe form in cross-section (sect. 2). Later 1 Cf. the gaps occurring at each dichotomy of the axis in members of the Lepidodendreae, which possess medullated steles. Phylloglossum Drummondn , Kunze . 323 Fig. 5. 324 Sampson. — The Morphology of this gap is closed, in this case by xylem, which appears between the edges of the gap and connects the stele with a single leaf-trace (sects. 3 and 4). It is this part of the stele which supplies the strand for the new tuber. The difference between the tuber strands, as they leave the stele in the two plants, depends partly on the fact that the tuber is far better developed in the one case than in the other. Whereas, in the plant with a single tuber, the stele divides almost equally, in the plant now under consideration only a relatively small mass of xylem passes out to the stunted tuber. A difference in form may also be noted, the stele of the stunted tuber being from the first a compact strand of xylem with a small core of enclosed parenchyma. The stele of one more fertile plant has yet to be described, since it is, perhaps, more typical of the smaller fertile plants in which the vascular tissue is less well developed. The plant in question possessed a cone, six leaves, and a somewhat slender new tuber. At the base of the peduncle two separate strands of xylem are found instead of a hollow cylinder as in larger plants. Lower down these increase in size, and a third strand appears, connected with a leaf-trace which joins the stele relatively early. The stele now shows a form roughly comparable with the horseshoe- shaped stele found at a corresponding level in larger plants, but it is less striking since the stele in the peduncle is already interrupted by large gaps.1 In this plant no leaf-traces are connected with the vascular supply of the tuber. It arises from the medullated stele as a small arc of xylem, and, as it passes down the shaft of the tuber, it early becomes a slender cylindrical strand. In two of the three plants just described, a distinct break in the stele of the main axis was observed immediately above the point at which the strand of the new tuber is given off. Though less conspicuous in the smaller fertile plants, owing to the weaker development of the stele, this break was found in eleven of the twelve cases examined, while the one specimen in which a gap was not found showed a thinning of the xylem in a corresponding position. Although this gap generally extends for a very short distance, the sections in which it does occur present a most charac- teristic appearance, and one which has been noted by previous workers. Thus Wernham concludes that the ‘noticeable (J-form of the upper part of the stem is, perhaps, the most striking feature’,2 while Jeffrey describes Phylloglossum as having ‘ a tubular stele, which, in the lower tuberous portion of the stem, constitutes in cross-section an almost continuous horse- shoe of xylem \ He states also that ‘the opening in the horseshoe corre- sponds to the outgoing strand, which passes into the resting tuber’.3 1 The stelar gaps of Phylloglossum are not regarded as foliar gaps,' though sometimes found in apparent connexion with leaf-traces (Jeffrey: Bot. Gaz., vol. xlvi, 1908, p. 245). 2 Wernham : loc. cit., p. 33S. 3 Jeffrey : loc. cit. , p. 244. Phylloglossum Drummondii , Kunze . 325 The fact that a gap is caused in the main stele by the exit of the vascular strand of the tuber has been used as evidence for regarding the tuber of fertile plants as morphologically a branch (p. 322). It is therefore of the greatest importance to find that the gap is such a characteristic feature in the anatomy of fertile plants.1 Other evidence was found in the fact that the tuber stele of the plant first described makes a sharp bend as it passes out, thus suggesting such a change in the direction of growth of the organ as the branch theory of its morphology would involve (Fig. 4, A, /). In the majority of fertile plants this is not the case, for the tuber strand follows, almost as soon as it is free, a direct downward course, but stelar tissue is sometimes found above the level at which the tuber stele is given off (Fig. 5, sects. 2 and 3), and doubtless represents the definite bend seen in a better-differentiated plant. That the difference is chiefly one of degree may be seen by comparing the large tuber of Fig. 4, A, with the left-hand tuber of Fig. 4, B. In this connexion it is also important to note in Fig. 5 how the form of the tuber stele changes as it passes out. When first free, it consists in cross-section of a horseshoe-shaped mass of xylem, with the arms of the horseshoe turned towards the main stele (sect. 4), but by the time it enters the shaft of the tuber (sect. 6) the curvature is in the opposite direction. The two positions of the gap in this single strand correspond to the two gaps seen in Fig. 3, sect. 6, where the stele of the tuber is cut in two places below the bend. We have, therefore, additional evidence that the tuber is an organ which, in the course of specialization, has made a definite change in its direction of growth. It is interesting that features similar to the above are shown by the steles of several other well-developed tubers. Before leaving the subject of the course of the tuber stele, reference should be made to its behaviour as it traverses the shaft of the tuber. Not infrequently the tuber stele diminishes rapidly to a slender strand of tracheides, and dies out before the swollen end of the tuber is reached. In the larger specimens, on the other hand, the horseshoe form of the stele may be retained until it nears the swollen region, where it breaks up to form several separate strands. By further subdivision the tuber stele forms an irregular sheath of tracheides in the tissue surrounding the bud of the storage tuber.2 These tracheides, which are little lignified and difficult to distinguish from the cells of the ground tissue, die out before the base of the bud is reached. Their presence in the largest tubers is 1 The possibility of the gap being foliar, which was suggested by Wernham, is rendered still less probable by recent work, which dissociates very completely the Ophioglossaceae from the Lycopods. Tniesipteris , even if the stelar gaps are foliar, which is doubtful, is more commonly regarded now as showing Sphenopsid rather than Lycopsid alliance. Moreover, the * organe de Mettenius which Wernham suggests may be the vestige of a megaphyllous leaf is in reality a stage in the reduction of one of the normal leaves of the plant (p. 326). 2 This sheath of tracheides was described and figured by Bertrand, loc. cit., 1885. Z 326 Sampson . — The Morphology of doubtless connected with the need for a more abundant water-supply. The more primitive features of the tuber would not be looked for in the long shaft of the tuber, which is clearly an adaptation to a geophytic habit, but in that part of the tuber most closely connected with the main axis. Such features have already been found in the origin of the tuber stele, and others will be dealt with in the following section. c. Leaves connected with the New Tuber. Mettenius, in his memoir of Phylloglossum , describes a small tongue- shaped body, frequently found above the new tuber. This structure he regards as an atrophied leaf, since, in his material, it was connected by transitional forms with the normal leaves of the plant.1 A similar body was observed by Bertrand, and named the ‘ organe de Mettenius but his material showed none of the stages which led Mettenius to his conclusion.2 Bower, however, states that, in several of the mature plants he examined, a leaf placed above the new tuber was frequently smaller than the others, and he suggests that it corresponds to those structures observed in a similar position by earlier writers.3 The aim of this section of the paper is twofold ; in the first place, to confirm the work of Mettenius by showing that the organ which bears his name is indeed a reduced leaf, and, secondly, to show how some of the lower leaves of the plant are connected with the new tuber. The general structure of the tuber, the bud of which is sunk at the base of a narrow channel, has already been described (p. 317). The mouth of this channel is situated on the outer surface of the tuber, often at the base of a small hump of tissue borne in the angle between the new tuber and the peduncle of the cone. In some cases this hump of tissue can only be observed in a microscopic examination (Fig. 4, B, rdf in others it may be seen with the naked eye or with a lens, and recalls then the tongue- shaped body described by Mettenius and Bertrand (Fig. 2, rd.). A vascular supply is generally found, consisting of a slender strand of tracheides, which dies out before the tip of the organ is reached (Fig. 5, sect. 1, rd,, and Fig. 3, sect. 1). Occasionally, in place of this tongue-shaped structure a small but otherwise normal leaf is found, the ‘ supernumerary leaf’ of Professor Bower. Such a leaf is shown in Fig. 2, a, sd ., and appears in section in Fig. 3, sect. 1. A complete series of transitional forms connects such leaves with the various structures described above, and there is no doubt that all are merely stages in the reduction of one of the normal leaves of the plant. 1 Mettenius : loc. cit., P.-99. 2 Bertrand: loc. cit. , 1885. 3 Bower : loc. cit., 1886, p. 670. Phylloglossmn Drummondii , Kunze. 327 The occurrence of a stunted leaf in such a position is explained by its connexion with the tuber. Earlier in this paper (p. 324) reference was made to the fact that a leaf-trace, travelling in the cortex opposite the ramular gap, was connected with the stele of the tuber. Such a leaf-trace may belong to a normal leaf, or to a leaf in one of the stages of reduction just mentioned. So constant is this feature in fertile plants, that of fifteen tubers examined, only four were found unconnected with any leaf, and these were borne by small plants with relatively few leaves. Moreover, a tuber is frequently found bearing more than one leaf, as in the plant of Fig. 5, where the steles of two normal leaves and one which is much stunted are connected with the stele of the new tuber. Plant b, sketched in Fig. 2, was conspicuous both for the number of its leaves and for the fact that they were arranged unequally round the axis of the cone. It was found that from one tuber five leaf-traces were given off, from the other only two. The leaf-traces are generally given off from the base of the tuber stele, that is, from the part near the parent axis, where the more primitive characters are likely to occur (p. 326), but in one case a small vestigial strand passed off from the tuber stele at some distance down the shaft (Fig. 4, B, r./.). The fact that leaf-strands pass off from the stele of the storage tuber, some of them supplying leaves which are quite indistinguishable from those borne by the main axis, confirms the view that the tuber of Phylloglossu m is a specialized branch. Moreover, the stunted growth of certain leaves may be correlated with this specialization, since imperfectly developed leaves are always those connected with a storage tuber. V. Vascular Anatomy of Sterile Plants. The resting tubers of Phylloglossmn always produce, on germination, one or more roots, a tuft of the characteristic cuneiform leaves, and a new storage tuber, but the spore-producing part of the plant may be absent. In these plants some of the leaves on the side from which the tuber is given off may be reduced, but the different stages in reduction observed in fertile plants were not all found. No sterile plant with a second new tuber has been recorded. Serial transverse sections of a sterile plant show, in general, the following stelar structure : At the base of the plant the entering root- strands unite with the small strand of the tuber to form a stele, consisting of a core of xylem, which seldom shows a definite medulla. From this stele leaf-traces are given off, each consisting, as in fertile plants, of a single mesarch strand of xylem. As the leaf-traces pass into the cortex the stem stele diminishes in size, breaks up, and, before the leaf-bases are free, 3 28 Sampson . — The Morphology of completely disappears. This apparent disintegration and dying out of the upper part of the stele in sterile plants is the most striking feature in their anatomy, and one demanding an explanation. Earlier in this paper (p. 321) a plant was described in which the stele of a storage tuber made a sharp upward bend as it passed out, and this bend was taken as indicating a change in direction of growth. It may be noted that sections through this bend correspond very closely with transverse sections taken at a certain level in sterile plants, and the apparent dying out of the upper part of the stele in these plants may be due to a sharp bend in the axis. If this be so, sterile plants must consist of a slender unbranched axis, which bends over and forms the annual storage tuber. In support of this is the fact that Professor Bower, working on the ontogeny of the yearly growth of sterile tubers, identified the growing point of the new tuber with the apex of the stem itself.1 In the largest sterile plant examined, the course of the stele is some- what different from that described above, the chief difference being that it is for a short distance medullated, as are the steles of fertile plants, and the tuber strand, as it passes out, assumes the characteristic U-form. More- over, the medullated stele shows a break which, corresponding in position to the tuber stele, recalls the ramular gap of fertile plants. The upper part of the stele is, however, apparently lost among the leaf-traces as in other sterile plants. The U-form of the tuber stele, and the break which occurs in the stem stele above the level of its exit, are difficult to bring into line with the con- ception of sterile plants suggested above. They are explained, however, if, on analogy with fertile plants, it be assumed that branching has occurred on the formation of the new tuber, but that here the fertile branch has been arrested early in its development. It seems probable, therefore, that two conditions may exist in sterile plants : that a sterile plant may consist either of a simple axis, concerned only with the formation of a storage organ, or of an axis which has divided, one branch forming a tuber, the other being completely abortive, thus representing a condition intermediate between the small sterile and the simple 2 fertile plants. In either case the anatomy of sterile plants bears out the conclusion that the tuber of Phyllo- glossum is the specialized terminal part of a leafy axis. VI. Branching in Phylloglossum. Phylloglossinn has hitherto been regarded as a typically unbranched form of Lycopod, the rare cases of branching which have been recorded being restricted to the cone-bearing axis. Professor Thomas states that 1 Bower : loc. cit., 1886. 2 ‘ Simple ’ here refers to fertile plants with a single new tuber in which the axis is, therefore, only once branched. Phylloglossum Drummondii , Kunze . 329 about one plant in 2,000 may possess a forked strobilus, the two arms being equally developed.1 If, however, we hold that the tuber is morphologically a branch, branching in Phylloglossum can no longer be regarded as a rare occurrence, since it happens at least once in the yearly growth of every fertile plant. Moreover, the New Zealand form not infrequently produces more than one new tuber during the season, and in such cases two acts of branching must have occurred.2 When two new tubers are formed they may arise on opposite sides of the plant, with the old tubers between them (Fig. 2, b), or near together on the same side (Fig. 2, c). In this case the habit of the plant suggests that the two tubers are the result of a dichotomy, and this is supported by its stelar structure. A medullated stele is found in the peduncle, and is later interrupted by a gap, below which, on the same side of the plant, a curved band of xylem passes out into the cortex. So far there is close agreement with the stele of a fertile plant with one new tuber: the difference lies in the behaviour of the branch stele. Whereas it normally passes, with some modification in form, down the shaft of a single tuber, in the plant now under discussion it gives rise to two smaller strands, and thus supplies two tubers. The daughter steles are slender strands, showing no characteristic gaps, since the division takes place when the original stele is an irregular curved band of xylem. The duplication of tubers in this way is consistent with the conception of the tuber as a specialized part of the axis, and affords an example of double branching in Phylloglossum. Double branching also occurs in the axis of plants bearing a new tuber on two opposite sides. The anatomy of one of the three plants showing this feature has been fully described (p. 318), and, since the others agree with it in general structure, it is not necessary to do more than refer to the plan of the stele.3 A large stele is formed at the base of the plant by the entering root- strands. This divides, and the smaller product of the division supplies one of the new tubers. The larger of the two steles divides again, a small strand passing out to the second new tuber, while a large medullated stele passes up the peduncle of the cone. The distance between the two points of branching is so short that the tubers may appear to arise at the same level. That this is not the case is shown by the stelar anatomy (Figs. 3 and 4, a). In these plants the new tubers have been formed by two successive 1 Thomas : loc. cit. 2 Thomas states that, while plants with one tuber are still in the majority, two tubers are frequently formed. 3 In one case the axis apparently branches twice, but the stele of one new tuber is unconnected with the vascular tissue of any other part of the plant. This can only be regarded as an anomaly resulting from the reduction which the plant has suffered in the course of adaptation to its environment. 330 Sampson. — The Morphology of acts of branching at the base of the axis, which finally terminates in a cone. In the previous examples the duplication was due to a dichotomy of the branch, which more frequently forms a single tuber. Thus, not only is branching found occasionally in the strobilus of Phylloglossum , but normally in fertile plants on the production of each new tuber. It can, therefore, no longer be said that Phylloglossum is characteristically an unbranched form. VII. Conclusion. The interest which has been manifested in the monotypic genus, Phylloglossum , since its discovery in 1843, has centred round its most characteristic feature, the annual storage tuber. Treub saw in it a resemblance to the organ of L. cernuum , which he designated a ‘ protocorm,’ regarding it as of phylogenetic significance.1 In this way he initiated a confusion which has been at the base of our difficulty in interpreting the true morphology of Phylloglossum . With the increase of detailed knowledge of the embryogeny of Lycopods, the inconstancy of the ‘protocorm ’ has been more fully realized, and opinion as to its primitive nature has been modified, the present tendency being to regard it as an ‘opportunist local swelling’, of physio- logical rather than of phylogenetic significance.2 This new conception of the ‘ protocorm ’ has not, however, helped to elucidate the problem of the morphology of Phylloglossum , if the old com- parison with the embryo of L. cernuum still be made. Except inasmuch as both the ‘ protocorm 5 of L. cernuum and the tuber of Phylloglossimi are manifestations of the tendency to local swellings seen throughout the family, the resemblance between them is purely superficial. The present work, by detailed anatomical investigation, has shown that the tuber of Phylloglossum is a highly specialized leafy axis, the terminal bud of which functions both as a means of vegetative reproduction and as an organ of perennation. Though different in appearance and structure, the tuber of P hylloglossum is comparable with the resting buds of L. inundatum and the ‘tubers ’ of certain Indian species of Selaginellal The position which Phylloglossum has hitherto occupied in the family Lycopodiaceae is rendered less isolated, since it can no longer be regarded as typically unbranched. On the other hand, the marked geophytic specialization, the mesarch character of the xylem, and the medullation of the stele, together with long established custom, justify the retention of separate generic rank. 1 Treub: loc. cit. 2 Bower: loc. cit„, 1914. 3 Bancroft, N. : Ann, of Bot., vpl. xxviii, 1914, p. <585. Phylloglossum Drummondii , Kunze . 33i VIII. Summary. I. That the tuber of Phylloglossum is, in fertile plants, a modified branch is supported by the following facts : 1. A gap is left in the stele of the main axis by the exit of the vascular strand of the tuber. 2. The stele of the tuber often shows a corresponding gap. 3. The tuber bears leaves, some of which are considerably reduced. II. In general, sterile plants consist of a simple axis, the apex of which has formed a storage tuber. It is possible that, in the larger specimens, branching occurs as in fertile plants, but the arm, which in the latter produces a cone, is in sterile plants arrested early in development. III. The tuber of Phylloglossum can no longer be compared with the protocorm of Lycopodium cernuum , but the two genera are brought nearer together, since Phylloglossum has proved to be not characteristically an unbranched form. In conclusion, I wish to thank Professor Benson for suggesting this work, and for her helpful criticism during the course of the investigation. The Physiological Anatomy of Spartina Townsendii. BY GEO. K. SUTHERLAND AND A. EASTWOOD. With seven Figures in the Text. Introduction. ^PARTIN A is a small genus of very characteristic grasses, mainly ^ natives of the Atlantic seaboard of America, where they are to be found abundantly in salt marsh and estuary. Spartina cynosnroides, the freshwater Cord Grass, penetrates inland to the Missouri River, and in the Western States it forms a large part of the grass of sloughs and wet marshes. In Europe only four representatives occur, if we regard S. Townsendii and S. Neyrauti as one species. S', juncea is restricted to the western portion of the Mediterranean, whither it was introduced probably by shipping. Of the other three the oldest known is S. stricta , which Stapf regards as undoubtedly indi- genous. It has the widest distribution of the European forms, but, not- withstanding its long establishment, it is becoming scarce on the south coast of Britain owing to the rapid spread of a later species. S. alternifiora was recorded first from the neighbourhood of Bayonne at the beginning of last century, and later it was discovered at the head of Southampton Water, down which it spread until its progress was checked by the remaining species, S. Townsendii , about whose origin and first appearance considerable uncertainty exists. In his Flora of Hampshire, published in 1883, Townsend gives the first record of this species as 1878, when it was collected in the neighbour- hood of Hythe, Southampton Water, by the brothers Groves, who described it shortly afterwards as a distinct species. But there is no doubt that it existed earlier, although overlooked. A specimen in the Warner Herba- rium at University College, Southampton, collected near Hythe in 1870 and labelled S', stricta , is undoubtedly S'. Townsendii . This carries it definitely [Annals of Botany, Vol. XXX. No. CXVIII. April, 1916.] 334 Sutherland and Eastwood. — The Physiological back to that date, but there is reason to believe that it occurred even before that time. The older accounts of the Spartan grasses vary much, and in Sowerby’s ‘Grasses’ (1861) the opinion is expressed that the plants of S. alterniflora , collected near Southampton, were so like S. stricta that they could not be regarded in any other light than as intermediate varieties. But S', stricta is a fairly constant species showing remarkably few varia- tions. Therefore the probability is that the doubtful specimens were in reality plants of S'. Townsendii , which in many of its characters is inter- mediate between S', stricta and S. alterniflora. Unless the plants were examined carefully in situ, and at the flowering season, the appearance of this new plant might be overlooked for several years, mixed as it was with a very similar species, and growing in places not readily accessible. This view would help us to understand better its present extent and profusion. Townsends Spartan grass has been characterized by its phenomenal success on the shelving mud banks along the entire western shore of Southampton Water, whence it has spread with amazing rapidity over the available and suitable mud flats between Selsey Bill and St. Alban’s Head, which form the natural boundaries of the sunken valley of the old Frome or Solent River. Here the numerous creeks and estuaries, harbours, and salt marshes from Poole to Chichester are well protected, while tertiary forma- tions have supplied abundant mud. Again, the various stages of the pro- gress eastward and westward have been short and favoured by eddying currents, which have helped largely in fruit dispersal. On both sides of these limits, where chalk ridges reach the sea, there are extensive stretches of shingle beach and cliff, broken by few suitable openings, with the result that for the time being the natural spread seems checked. The value of this grass in fixing and in reclaiming shifting and unsightly mud banks has been recognized, and already attempts have been made to utilize it. Plants placed in the Medway have made considerable progress ; others have also become acclimatized at both Blakeney Point and Wells Marsh in Norfolk. More recently the experiment has been extended northwards to the mud flats of the Forth and Don mouth. At the present time it is the dominant species in the south of England. S. stricta survives in a few quiet backwaters, while A. alterniflora is dis- appearing fast before its more vigorous competitor, whose adaptation and success may be gauged by a glance across the Spartan beds from Cadland to Calshot, or from Lymington Harbour to Hurst Castle Bay. It is inevitable that such an extensive and thick vegetation should affect the deposition of silt near river mouths, and hence tend towards a quicker levelling up. At present there exist no reliable data with regard to the rate, but it is hoped that a series of accurate survey measurements may be made at different stations along Southampton Water. 335 Anatomy of Spartina Toivnsendii. Apart from its usefulness, the grass presents interesting ecological features, ~and the present Gstudy was undertaken in the hope that it would throw some light on its adaptation to a life of periodic submersion, as wel as form an introduction to a contemplated series of experiments on its physiology. The distribution and spread of the plant has been treated fully by Stapf, who is still carrying out his observations on its ecology. External Morphology. Rhizome. Spartan grass owes its power of rapid extension mainly to its characteristic rhizomes, which vary in length from a few inches to over a foot, rarely exceeding one-quarter inch in diameter. Their length is dependent for the most part on the type of soil and the available space. They travel horizontally through the mud at a depth of from o, \ to 4 inches, but frequently in the young stages they show positive geotropism and pene- trate downwards for a short distance. Near their point of origin they are firm, with short hollow internodes. The greater portion of their length, however, is soft and flexible, the whole structure being adapted for pene- trating a soft substratum. Short scale leaves are found in the bud stage, but the mature rhizome is invested merely by colourless sheaths devoid of blade and ligule. After proceeding for a short distance, rooting more or less freely at the nodes, the tip turns upwards and gives rise to aerial branches. New rhizomes are produced sympodially, and in the early stages there is nothing to distinguish them from the aerial shoots, except that the latter remain for a longer time within the sheath and are given an upward tendency. The creeping axes are shorter in a stony substratum, and hence it may also be that the question of the available space plays a part in determining the tendency and consequently the ultimate formation of culm and rhizome. Culm. The erect cylindrical aerial axis, covered almost entirely by numerous investing leaf-sheaths which enhance its rigidity, reaches a height of from % to 4 feet. The first shoot arising from the upturned rhizome tip is usually the dominant one. Secondary culms spring distichously from its basal portion. Of these the lower and outer generally grow more quickly, checking the development of the higher ones. Only at the margin of the clump or belt is it possible for plants to develop all or most of their aerial shoots. Leaf. The leaves are also distichous. The first three or four are practically scales. As soon, however, as the axis rises above the ground short green-tipped blades appear, increasing in size as the stem is ascended, until they reach a length of twelve to eighteen inches. The lowest blades are thrown off at an early stage, separating at the articulation. This helps to distinguish this species from N. alterniflora , in which the lower blades are retained longer and wither gradually. A short soft-celled ligule, tipped 336 Sutherland and Eastwood . — The Physiological with hairs, encircles the stem and prevents water or mud from lodging between it and the sheath. The long smooth pale-green sheath clasps the axis firmly, completely surrounding it for the greater part of its length and passing into the blade through the pulvinar articulation. The lower blades form angles of from 450 to 6o° with the stem, while the upper ones are more erect. The angle increases with submersion. During the fall of the tide the leaves swing up and down in the surface film until they are suddenly released and spring into position. The adaxial surface of the blade is increased enormously by from forty to fifty ridges and furrows running from just above the pulvinus to the apex. These provide an increased assimilatory surface in addition to pro- tection for the stomata. The papillae and waxy coating on this surface give it a velvety and glaucous appearance. Root. The roots are adventitious and divided into two distinct sets. One type is long, relatively thick, smooth, and practically devoid of branches. This fixing form grows normally to a depth of six to twelve inches, serving to anchor the plant firmly in the soft mud. While its root- cap is large, root-hairs rarely, if ever, occur. The other type is shorter, thinner, and much branched, possessing a poorly developed root-cap and very sparse root-hairs on the youngest branches. Generally this latter form springs from, and clusters round the basal joints of the aerial axis, forming a densely woven, horizontal matting which aids largely in transforming the soft mud banks into comparatively firm turf. Similar tufts of branched roots spring from the rhizome nodes also. Their function is mainly absorp- tive. Both kinds of roots frequently show negative geotropism, probably an adaptation to the continuous silting. Anatomy. Epidermis. The adaptation of the plant to alternate aerial and aquatic existence is marked by peculiar and special epidermal structures wherein it differs from any other grass genus examined. The two most striking features are the very distinctive hydathodes and the forked accessory stomatal papillae, both of which are undescribed as far as we can ascertain. A distinct waxy coating covers all exposed portions above ground, giving the glaucous appearance. This is most pronounced on the adaxial surface of the leaves, where the epidermis is thinner, shows less cutinization, and is covered by numerous papillae as in Fig. 1, 1. These conical or peg- like projections also show little cutinization. On the narrow cells of the ridges they run along in pairs which frequently become coated over in the old fully developed leaves, as in Fig. 1, /, pd. This adds to the protection of the most exposed portions. The epidermal cells along the sides of the grooves are wider, and the number of papillae consequently greater. These, however, undergo no change. The papillae aid the waxy 337 Anatomy of Spartina Townsendn. coating in preventing water from adhering to, and wetting the upper surface. Their efficiency may be gauged by the fact that leaves, sub- merged in the laboratory for more than twenty-four hours, were quite Fig. i. i. Epidermis of upper surface of leaf-blade, showing simple papillae ( pa and pa) ; hydathode (Jiy) ; and special stomata ( st ). 2. Epidermis of abaxial surface with ordinary stoma; pits (/) ; and hair (h). 3. Epidermis of abaxial surface of leaf-sheath immediately below the articulation : silica-cell (si) ; saddle-cell (sa). 4. Epidermis of articulation region, showing thick- walled pitted cells with very wavy outline. 5. Radial longitudinal section through 3 : nucleus beneath silica body (11). 6. Enlarged surface view of the two types of short cells. dry when taken out and shaken slightly. The normal period of natural submersion rarely exceeds a few hours. The number of variations in the epidermis of different portions of the plant are merely changes rung on an essentially simple ground-plan, either of long cells alone, or more often of long cells alternating with 338 Sutherland and Eastwood . — The Physiological short ones, singly or in pairs. Their radial walls may be straight, but more frequently they present an undulating margin, most pronounced in positions of great strain like the pulvinar region, as in Fig. i, /. The inner surface of the sheath, protected by being pressed firmly against the stem, has thin, straight walls, while the outer epidermis (Fig. i, 3 and/) is coated with a thick development of cuticle. The adaxial surface of the blade is cutinized slightly, whereas the abaxial side has a strong, resistant layer which not only protects the mesophyll, but adds materially to the rigidity of the leaf. The poor development of this coating on the upper surface is compensated by the protection afforded by the fur- rows, and by the curling of the leaves when the water-supply is limited or transpiration excessive. Long cells . These occur alone only in protected regions like the inner epidermis of the sheath and along the sides of the leaf-grooves. In most other parts they alternate with short cells, and possess strongly thickened outer and radial walls with numerous round or elongated pits (Fig. i, 2, y and /). Over the articulation between sheath and blade the thickening is most pronounced, forming strong radial flanges or girders between elongated large pits (Fig. 6). Freedom of movement at this point is facilitated by the shortened cells, whose folded fan-like walls (Fig. i, /) are capable of lengthening with a kind of bellows-action. The most interesting of these long cells are the motor-cells (Fig. 4,2, m.c .), which were first described for grasses by Duval-Jouve, who regarded them mainly as silica-containing cells, to which he gave the name ‘ cellides bidliformes \ In Spartina Townsendii they form belts three cells wide running along the bottom of the furrows of the blade. They are clear and colourless, containing little solid matter but abundant water easily given up. While their inner and radial walls are very thin and collenchymatous, their outer walls differ only slightly from the adjoining epidermal cells. They are well developed towards the middle of the blade, being much deeper than the other cells along the furrows, but towards the margin little difference is seen. This accounts for the leaves rolling up completely only when excessive drying takes place. Short cells . These show greater variation and are of two distinct types, one containing no silica or only traces, the other with a relatively massive, definitely shaped silica body. The former, which occur singly or in pairs between long cells, may be regular, but frequently they become saddle-shaped, cruciform, biscuit-shaped, or even dumb-bell-like (Fig. 1, sa). They invariably show less thickening than the adjacent long cells, and no pits are present on their outer walls. More interesting, however, is the second type containing a distinct silica mass. Each is accompanied by a short cork-cell placed always on the side towards the base of the plant, and forming a kind of saddle, in the hollow of which Anatomy of Spartina Townsendii . 339 the silica-cell lies, partly embraced by the upcurved ends. This is shown in Fig. 1,3, 4, and 6. This type appears wedge-shaped in longitudinal sections of the leaf. The blunt end is outwards (Fig. i,y, si) and covered by a very thin wall, difficult to distinguish owing to its delicate structure and the refraction cf the mass beneath. The silica body occupies the greater part of the cell. It is invested by a thin layer of protoplasm and blocks the entire upper and wider end of the cell cavity by which its outline is determined. It is easily recognizable both by the presence of small included air-bubbles, and by its resistance to stains. These cells develop from ordinary short, rectangular ones, and are most abundant in the upper outer epiderm of thes heath, where, immediately below the articulation, they equal the long cells in number. They cease at the point gc Fig. 2. 1. Surface view of special type of stoma with forked papillae ( fpa ) ; simple papillae {pa). 2. Transverse section through swollen vesicular ends of guard cells. 3. Transverse section near the middle and thickened portion of guard cells ( gc ). where the break appears later. A few occur along the ridges of the blade and on the peduncle, but none have been found on the rhizome, the stem, or the first-formed leaves. Their distribution lends strong support to the view that they add to the rigidity of the plant, which is aided largely by the closely investing leaf-sheaths, on whose outer surface they are most numerous. Stomata , The stomata belong to the characteristic grass type. The walls of the middle portions of the guard cells are so thickened that the cavities connecting the swollen ends are reduced to narrow passages, as in Fig. 2, 3. The slit is slightly longer than these rigid bridges, which are carried bodily apart by the swelling of the vascular thin-walled ends (Fig. 2, 2). The most active stomata are those situated on the adaxial surface of the leaf over the loose chlorenchyma, where they are more numerous than in any other part of the plant. There they occur in two, rarely 34° Sutherland and Eastwood . — The Physiological three, close rows on each side of the laminar furrows. These rows (Fig. 4,2) are about two cell-widths apart, and almost corresponding distances from the motor-cells forming the bottom of the groove, and the line of hyda- thodes nearer the angle of the ridge. The stomata are of exceptional interest on account of the unique structure and placing of the papillae on the subsidiary cells. The papillose epidermis of this surface has been noted already. There are two massive papillae on each subsidiary cell, placed opposite the end of the stomatal slit (Fig. 2, 1 and j). These ex- pand at the top into two, three, or more rarely four, short branches which are strongly lignified, like the thickened walls of the guard cells. Fre- quently there is a simple papilla between them, corresponding to the middle of the pore. All bend over the guard cells, forming a fringe round and over the stoma as in Fig. 2, 1. A small piece of leaf immersed in water showed a tiny air-bubble captured by these furcate papillae. Doubtless when the leaves are sub- merged the entangled air-bubbles prevent the entrance of water through the slits, and in this way the most active stomata are prevented from admitting water into the air-spaces at a time when they are open or partly so. This apparatus, along with the simple papillae and the waxy coating, goes a long way towards an explanation of the plant’s adaptation for its dual existence. The normal type of storim occurs sparsely over all chlorenchyma. They are abundant on the inner epiderm of the sheath, but then it presses so tightly against the stem that no water gains entrance. Although their appearance on the rhizome sheaths is more surprising, the absence of chlo- rophyll prevents the manufacture of osmotic substances, and consequently the guard cells are inert. Hydathodes. Hydathodes of a type apparently hitherto undescribed take the place of the water pores found in many submerged plants, being distributed widely in definite tracts in the active chlorenchyma, usually near large water-storing cells. On the upper surface of the leaf they are arranged in a row (Figs. 1 and 4) along each side of the furrow, about two or three cell-widths from the ridge angle, and one or two from the upper line of stomata. Here the epiderm consists of long cells alone, and the hydathodes are placed between every two or three of these in longitudinal series. On the abaxial surface of the leaf-blade and sheath they form a similar line in the large-celled tissue between the sclerenehyma bands over the bundles. They are absent from the rhizome, the invested portion of the stem, and the inner surface of the sheath, but occur abundantly on peduncles, and even glumes in the more-pronounced assimilatory bands, although always near fairly large cells with watery content. They are placed invariably between the ends of two long cells which become slightly narrower as they approach one another (Fig. 3, /). However, instead Anatomy of Spar tin a Townsendi 7. 34i of meeting their corners project outwards again and meet the walls of the lateral cells, forming a cylindrical cavity whose rim in surface view is very thick, with the inner side smooth and circular, while the outer is wavy, varied, and even pitted (Fig. 3, 1). The hydathodal space is bounded therefore partly by four cells, along whose lines of contact are four vertical ridges or flanges, with the intermediate portion facing each cell thinner walled. A transverse section of the leaf shows a flask-shaped organ, as in Fig. 3, 2, with the neck projecting into the epidermal cavity described, and the swollen basal portion embedded in the underlying tissue ; a radial longitudinal section differs inasmuch as the basal part is elongated and boat-shaped, as in Fig. 3, y The cap portion of the hydathode has a distinctly stratified, mucilaginous wall, fitting the cavity tightly yet free from its walls. Although in the adult plant it is usually on a level with the surface or slightly below it, in the very young stages it projects a little distance beyond. It rests on a strongly lignified and cutinized collar (Fig. 3, 2, c) which marks the region where the thin-walled swollen base abuts on the retreating sloped lower corners of the adjacent epidermal cells. This collar is so strongly developed as to give the impression of a thick r t , ,, N r . . . . Fig. 3. 1. Surface view of hydathode (Jiy), from partition at this point. this is abaxial surface of leaf. 2. Transverse section of leaf accentuated by a slight projecting showing hydathode : («) nucleus; _ (0 collar. / & r J ° 3. Radial longitudinal section of abaxial surlace ot ridge on its inner upper surface, leaf, showing hydathode. The basal portion of the hyda- thode is elongated parallel to the axis of the plant and has pointed pyramidal ends. It is thin-walled and densely filled with protoplasm, in the centre of which lies a relatively large nucleus. In addition to shape, it is sharply marked off from the surrounding cells by the absence of chloroplasts. The protoplasmic content of the cap is also dense, and stained sections give the impression of its being connected through the narrow neck by numerous strands of protoplasm to that in the lower part. A nucleus has also been observed in the cap, placed sometimes near the tip, sometimes partly hidden by the collar. The more frequent occurrence of nuclei in the cap portion in younger material would strengthen the view that the hydathode consists of two cells whose common wall has been resorbed at an early stage. 34 2 Sutherland and Eastwood . — ■ The Physiological In very young leaves the tip of the hydathode projects a short dis- tance above the surface like a swollen glandular hair. This suggests a possible theory as to its ontogeny. The hairs so abundant in many land grasses would be useless to the plant when submerged. This is partly borne out by the fact that many hairs drop off before the leaves unroll and open. Then the hairs in Spartina correspond in position to the short cells between two long ones, just as the hydathodes do, and occa- sionally they have been observed occurring along the same lines as the latter in young plants. It is possible, therefore, that these excretory organs are really hairs modified to meet a new set of conditions. An extensive examination of the genus is necessary before a definite statement can be made with regard to the point. These hydathodes differ from any described forms, and are certainly unique in grasses. The nearest approach to them are the secretory cells, discovered by Sauvageau in aquatic Monocotyledons, and described by him in various papers on the structure of the leaves of these. In Cymadocea aequora , for example, they are distributed for the greater part along the margins of the leaves ; in others they may be scattered irregu- larly over the surface. These, however, are merely larger epidermal cells whose outer walls remain thin, and become distended and convex, while the inner portions penetrate slightly into the mesophyll. They resemble those described above in function, but lack their more definite structure. The Spartina type of hydathode may undoubtedly be regarded as a kind of safety-valve for getting rid easily and quickly of excess water and mineral salts, both of which, in abundance, are accessible, as a rule, to the plant. Thus fairly rapid loss of these is not dangerous to it, a fact which helps to explain the absence of any kind of epithem acting as a filtration-tissue, as in so many Dicotyledons. This want, common curiously enough to most Monocotyledons, facilitates rapid exudation. Their activity in this respect may be demonstrated very simply by cutting some plants and placing the cut stems in water under a bell-jar. In a few hours an immense number of drops may be seen on blade and sheath. The hydathodes forming the lines along the furrows function so quickly that tiny sparkling drops may be detected in less than an hour. These are much more active than those on the abaxial surface, owing doubtless to their greater proximity both to the large water-storing cells and the special assimilating tissue. Large quantities of salts, especially sodium chloride, are present in the excreted water. These may be detected by chemical methods in the drops given off, but a more striking demonstration of their presence can be seen in nature. While examining some plants, whose upper leaves and peduncles had been exposed continuously for some days of neap tides to a fairly dry atmosphere, our attention was directed to numerous, small, white, worm-like Anatomy of Spartina Townsendii . 343 castings scattered over these parts of the plants. These were often from x to 2 mm. long, occurring over the hydathodal lines and only on the living parts. On examination they proved to be heaps of cubical crystals, mainly of sodium chloride. Rapid evaporation after exudation had enabled the salts to crystalize out, and their accumulation in such castings had been favoured by a continuation of dry weather and the failure of the neap tides to submerge the upper portions of the plants for a few days. The worm- casting-like form of the heaps of crystals, their distribution on the plant, and their absence from withered or dead portions prove conclusively that they were products of excretory activity, not the remains after evaporation of clinging drops of water or spray. What is most surprising is the amount of salts excreted by these small structures. A shoot of Spartina with several leaves, when cut off under water and then transferred to a bell-jar with the cut end dipping in water, shows practically as great an excretion as potted plants placed in similar atmo- spheric conditions. The cut ends of the vessels are immersed freely in the water, continuity with the supply being secured without the intervention of an active cortical tissue as in the roots. The position of the most active of all the hydathodes in the stellate spongy chlorenchyma, with a thin layer of the same between them and the water-storing cells, precludes the possibility of much local pressure outside the hydathodes themselves. This would seem to throw the onus of their activity upon their own structure and contents, inducing the belief that the excretion is due mainly to some form of protoplasmic activity. When the plants are submerged, some such system is necessary to get rid of excess water and salts, and prevent flooding of the air-spaces. A simi- lar danger must be met even when they are exposed by the fall of the tide, for the cushion of air, held by the dense vegetation belt over the moist substratum, is more saturated than the atmosphere immediately above, with the result that transpiration is lessened and an auxiliary method of water-excretion rendered necessary. This is found in the system of hydathodes, which also enables the plant to throw off immense stores of salts. Rhizome and Culm . Both the underground and aerial shoots conform to the usual grass type. One distinctive feature, associated with the habit of the plant, is the development of air-passages which in the former are large, and separated by radial parenchyma plates only one or two cells in thickness, while in the latter they are smaller and separated by much thicker radial walls. Usually they arise by the development of stellate tissue, and its subsequent disruption ; sometimes large rounded cells break down without any appearance of stellate cells. Leaf. The leaf also shows few variations from the normal furrowed grass type, except in the epidermal characters already described. Forty to A a 2 344 Sutherland and Eastwoods — - The Physiological fifty fibro-vascular bundles, one for each ridge, run up through sheath and blade. They are of two kinds (Fig. 4. 1 and 2), alternating with one another. Round each bundle are two rings of cells. The inner consists of smaller cells, regular and very strongly lignified in the larger type, less regular and Fig. 4. 1. Transverse section through leaf-blade, showing the distribution of hydathodes, air- passages, and sclerenchyma. 2. Portion of 1 very much enlarged : (pa) papilla ; (st) stoma ; (sc) sclerenchyma; (hy) hydathode; (ch) chlorenchyma ; (me), motor cell; (c) partition cell; (iv) water-storing envelope;