Cornell Lab of Ornithology Library at Sapsucker Woods llustration of Bank Swallow by Louis Agassiz Fuertcs Cornell University Library QH 105.M5A21 C.3 An ecological .fflev oj, ft^^^^^^^ 3 1924 022 555 662 The original of tiiis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924022555662 AN ECOLOGICAL SURVEY OF ISLE ROYALE, LAKE SUPERIOR PKEPARED UNDER THE DIRECTION OF CHAS. C. ADAMS. A Report from the University of Micliigan Museum, publislied by tlic State Biological Survey, as a part of the Report of the Board of the Geological Survey for 1908. LANSING, MICHIGAN WYNKOOP HALLENBECK CRAWFORD CO., STATE PRINTERS 1909 AN ECOLOGICAL SURVEY OF ISLE ROY ALE, LAKE SUPERIOR PREPARED UNDER THE DIRECTION OF CHAS. C. ADAMS. A Report from the University of Michigan Museum, published by the State Biological Survey, as a part of the Report of the Board of the Geological Survey for 1908. LANSING, MICHIGAN WYNKOOP HALLENBECK CRAWFORD CO., STATE PRINTERS 1909 TABLE OF CONTENTS. Introductory Note and Acknowledgments, by Dr. Charles C. Adams. PART I. ECOLOGICAL PAPERS. 1. Isle Royale as a Biotic Environment, by Dr. Charles C. Adams 1 I. Introduction 1 1. Itinerary and Personnel of ther Party 1 2. Aims and Method of Work 2 3. Previous Biological Investigations upon Isle Royale 4 4. Historical Note.. , 5 5. Available Maps of Isle Royale 5 II. The Biota Considered by Stations 6 1. Location of Field Stations in 1905 6 2. General Characteristics of the Stations 8 III. The Evolution of the Gross Environment 31 * 1 . Geological Succession 31 2. Topography and its Origin 39 3. Atmospheric Influences and their Evolution - 41 a. Chmate 41 b. Seiches 44 c. Climatic Succession 45' d. Lake Storms and their Influence 46 4. Surface Currents of Lake Superior 47 5. Origin of the Habitats 50 2. The Ecological Relations of the Invertebrate Fauna of Isle Royale, Michigan, by Dr. H. A. Gleason 57 I. Introduction 57 II. The Lake 58 III. The Inland Lakes 63 IV. The Tamarack Swamp and the Arbor Vitae Swamps 6,5 V. The Gravel and Sand Beaches 66 VI. The Rock Beach .• 69 VII. The Cladonia Clearings and Jack Pine Ridges 71 VIII. The Balsam-Spruce Forest : 75 IX. Artificial Clearings 77 X. Summary 77 3. The Ecological Distribution of the Birds of Isle Royale, Lake Superior, by Otto McCreary 81 I. Introduction 81 II. Light-house Peninsula, (between Rock Harbor and the Head of Conglom- erate Bay) 81 1. Lake Superior and Beach (Station I, Sub. 1.) 81 2. Spruce and Balsam Forest (Station I, Sub. 2 and 3) 82 3. Tamarack and Arbor Vitae Swamps (Station I, Sub. 4) 84 4. Jack Pine Ridge (Station I, Sub. 5) 84 5. Sphagnum and Spruce Bog (Station I, Sub. 6) 85 6. Valley at the Head of Conglomerate Bay (Station I, Sub. 1) 85 III. Trail to McCargoe Sove 85 1. Ransom Clearing (Station II, Sub. 1) 85 2. Benson Brook (Station II, Sub. 1) 86 3. Spruce and Tamarack Swamps (Station II, Sub. 2 and 5) 87 4. Rock Ridge Clearings (Station II, Sub. 3) 87 CONTENTS. Page. IV. Western End of Rock Harbor and Trail to Sumner Lake 87 1. Harbor. (Vicinity of Station III, Sub. 2) 88 2. Small Islands (Station III, Sub. 1) 88 3. Bulrush Zone and Delta (Station III, Sub. 3) 89 4. Trail to Sumner Lake (Station III, Sub. 4) 89 a. Birch Forest 89 b. Birch and Coniferous Forest 89 5. Sumner Lake (Station III, Sub. 5) 89 V. Siskowit Lake Region 90 1. Siskowit Bay and Shore (Station V, Sub. 1) 90 2. Trail to Siskowit Lake (Station V, Sub. 4) 91 3. Siskowit Lake (Station V, Sub. 6 and vicinity) 91 4. Burning West of Outlet to Siskowit Lake (Station V, vicinity of Sub. 9). 92 5. Long and Menagerie Islands (Station V, Sub. 10) 92 VI, Summary 93 1. Water Birds 93 2. Shore Birds 94 3. Birds Frequenting Swamps 94 4. Birds of Clearings and Partial Clearings 95 5. Birds Frequenting the Forests 95 The Fall Migration of Birds at Washington Harbor, Isle Royale, Lake Superior, by Max Minor Feet 97 I. Introduction 97 II. The Environment 97 1. The Clearing 97 2. The Fol-est 98 3. Food 99 III. The Weather Conditions and Migrants 99 1. Weather Conditions 99 2. The Bird Migrants 100 a. Warblers 100 b. Sparrows 102 c. Hawks 103 d. Owls 104 e.' Thrushes 104 f. Other Birds 104 IV. Large Bird Waves 105 1. First Wave 106 2. Second Wave 106 3. Third Wave 107 4. Fourth Wave 107 5. Fifth Wave 109 6. Sixth Wave 109 V. The Relation of Weather to Migration 109 1. Influence of Wind ' 110 2. Influence of Temperature Ill 3. Influence of Barometric Pressure Ill 4. Condition of the Sky Ill 5. Summary and Conclusion Ill VI. The Routes of Migration 112 VII. The Perils of Migration 113 1. Fatigue 113 2. Natural Enemies 115 3. Blunders and Fatalities 116 The Ecological Succession of Birds, by Dr. Chas. C. Adams 121 I. Introduction ' 121 II. Representative Literature on Habitats and Succession 123 1 . Habitat Preference 1 23 2. Succession 126 III. The Major Avian Environments 128 IV. Minor Avian Environments and their Associations 133 V. Avian Succession * 134 1. General Remarks 134 2. Succession on Isle Royale 134 a. The Aquatic Association and Habitat 136 r LETTER OF TRANSMITTAL. To the Honorable the Board of Geological Survey of the State of Mich- igan : Gov. Fred M. Warner, President. Hon. D. ;M. Ferry, Jr., Vice-President. Hon. L. L. Wright, Secretai-y. Gentlemen : — I beg to present herewith for printing, a report by l^r. Chas. 0. Adams on the ecology, that is the natural history, of Isle Eoyale. This comes to us with the apjjroval of Dr. A. G. Ruthven, our Chief Field Naturalist, and our Board of Scientific Advisers, and is a continuation of the work published in our annual report for 1905. This contribution to the Biological Survey of the State, which the legislature authorized me to supervise by Act No. 250 of the session of 1005, comes from the University Museum. The explorations Avere made without expense to the State Survey by means of contributions from friends of the Museum. As this work is in harmony with the aims of the Biological Survey we are fortunate in securing such co-operation. The reports on the Porcupine Mountains and Isle Royale at the north end of the state complement the work on Walnut Lake, Oakland county, and that iti Huron and Tuscola counties. I trust that the present report will be of service to the schools of the state. Very respectful! v, ALFRED C. LANE, State Geologist. ILLUSTRATIONS. Page Fig. 1. The Light-house at Rock Harbor, Isle Royale 16 2. Cliffs between Tonkin and Conglomerate Bays 16 3. Summer, storm waves upon the beach (I. 1) at the head of Tonkin Bay, 16 south of the Light-house 4. Sand beach at the head of Conglomerate Bay (I, 1) 16 5. Beach pool (I, 1) near Tonkin Bay 16 6. Transition from the beach (I, 1) to rock clearing (I, 2), south of the Light- house 16 7. Natural rock opening (I, 2) or avenue, farther up the slope than in Fig. 6. 16 8. Natural rock clearing or opening (I, 2) north of the Light-house at Rock Harbor 16 9. Natural rock clearing (I, 2) south of the Light-house 16 10. Arbor Vitae bog (I, 4) near Tonkin Bay 16 11. View from the Jack Pine ridge (I, 5), looking toward the head of Con- glomerate Bay 16 12. Second growth of White Birch on the Trail to the Jack Pine Ridge (I, 5), Conglomerate Bay 16 13. Jack Pine ridge (I, 5) Conglomerate Bay 16 14. Sphagnum-Black Spruce bog (I, 6) near the Jack Pine ridge 16 15. Small island near the head of Rock Harbor (HI, 1) 16 16. Bulrush zone and delta at the head of Rock Harbor (III, 3) 16 17. Exposed section of spit formed as the water level has lowered in Rock Har- bor, near the beginning of the trail to Sumner Lake (III, 4) 16 18. Sumner Lake (III, 5) 16 19. Western end of Sumner Lake (III, 5) 16 20. Northeastern margin of Sumner Lake (III, 5) 16 21. Southeastern corner of Sumner Lake (III, 5) 16 22. Western end of Sumner Lake (III, 5) 16 23. Northern shore of Sumner Lake (III, 5) 16 24. Rock opening about Camp on Siskowit Bay (V, 3) 16 25. Rock opening at the Siskowit camp (V, 3) 16 26. Rock opening on Siskowit Bay (V, 3) 16 27. Border of the opening about the Siskowit Camp (V, 3), near the beginning of the trail to Siskowit Lake (V, 4) 16 28. Ant nest in the opening at the Siskowit Camp (V, 3) 16 29. General character of the south shore, near the eastern entrance to Siskowit Bay (V, 2) 16 30. Rock pool on the beach (V, 2), where a variety of invertebrates was secured 16 31. Saxifraga aizoon on beach (V, 2) .- 16 32. General view along the shore at V, 2 16 33. Farther up the same slope as Fig. 32 and adjacent to it 16 34. Still farther up the slope and adjacent to Fig. 33 16 35. Looking up the slope on the western portion of Station V, 2 16 36. Upper portion of western part of Station V, 2 16 37. De'tail of western part of Station V, 2. .•. 16 38. Character of ground cover in parts of the Balsam-Spruce forest (V, 4) 16 39. Open space in the Balsam-Birch forest (V, 4) 16 40. Open space in the Balsam-Birch forest (V, 4) 16 41. Tamarack Swamp (V, 5) 16 42. Spruce margin of Station V, 5 16 43. Black Spruce margin of Station V, 5 16 ILLUSTRATIONS. Page. 44. Bog margin of Station V,5 16 45. Long Island gull rookery, (V, 10) 16 46. Pond in Tamarack — Black Spruce swamp (V, 11) 16 47. Margin of Lily pond (V, 11) 16 48. Black Spruce in Cassandra zone of Station (V, 11) 16 49. Maple forest on the Desor trail (III, '04) 16 50. Forest along Washington Brook (IV, ' 04) 16 51. Showing origin of the glacial Great Lakes, their relation to the ice sheet and their Mississippi drainage. (After Taylor and Leverett). Cham- beriin and Salisbury, Geology, III, p. 396. Fig. 516 33 52. Showing the Algonquin stage of the Great Lakes; a water barrier to north- ward dispersal of the land biota. (After Taylor and Leverett). Cham- berlain and Salisbury, Geology, III, p. 401. Fig. 521 34 53. Contour Map of Isle Royale, Michigan. Contour interval 100 feet. Pre- pared by Dr. A. C. Lane 34 54. The Nipissing Great Lakes; showing the fresh-water highway or barrier in the west and the sea barrier in the east, (After Taylor). Chamber- 36 liu and Salisbury, Geology, III, p. 404. Fig. 522 55. Surface currents of Lake Superior. To show their possible influence on the origin of the biota. (Drawn by Hall, after Harrington.) 56 56. "The Wendigo Road", from the clearing at the club-house to Wendigo, Washington Harbor 56 57. Long Island (V, 10), Siskowit Bay, looking toward Isle Royale Light-House. 56 58. Gull rookery on Long Island (V, 10) 56 59. Gull rookery on Long Island (V, 10). .' 56 60. Eagle nest at Tobin Harbor (IV, 8) 56 61. Variations in the shell width of Polygyra albolabris 298 62. Variation in the shell height of Polygyra albolabris 298 63. Lymnaea stagnalis varieties from Isle Royale 298 CONTENTS. vii Page. b. The Shore and.Marsh Association and Habitat 138 c. Bog-forest Association and Habitat 139 d. Aspen-birch Association and Habitat 139 e. The Climax Association or Formation and Habitat 141 3. Internal Factors 142 4. Environmental Factors 143 5. Environmental and Associational Convergence 144 6. Succession and Environmental Evolution 1 45 7. The Relation of Succession to Organic Evolution 145 VI. Some Principles of Succession 146 VII. Some Advantages of a Knowledge of the Laws of Succession ' 149 6. The Coleoptera of Isle Royale, Lake Superior, and their Relation to the North American Centers of Dispersal, by Dr. Chas. C. Adams 157 I. Introductory Note 157 II. Notes on the Habitat Relations of Beetles 158 III. The Succession of Beetle Associations.. 160 1. The Lake Shore and its Beetle Associations 160 2. Rock Openings and Associated Beetles 160 3. Lake, Pond and Bog Habitats and Associations 161 4. The Forests and their Beetle Associations 161 IV. The General Characteristics of the North American Beetle Fauna 163 1. Compiled Generalizations on the Fauna 163 2. Comments on the Preceding Generalizations, and on the Literature of Geographic Distribution 182 V. The Present Centers of Dispersal of the Beetle Fauna 183 VI. The General Characteristics and Affinities of the Isle Royale Fauna 190 1. Faunal Characteristics 190 2. Miscellaneous Notes on the Fauna 191 VII. Lists of Isle Royale Beetles 192 1. List of Species Collected in 1905 ,.- 192 2. Supplementary List of Isle Royale Beetles, by A. B. Wolcott 204 PART II. ANNOTATED LISTS. 1. Notes on the Vegetation of Isle Royale, Michigan, by W. P. Holt 217 I. General Observations on the Plant Societies 217 1 . Bog Societies 217 2. Shore Vegetation 222 3. Forests , 224 4. Burnings ■. 225 11. Annotated List of Plants 227 2. Annotated List of Certain Isle Royale Invertebrates, by Dr. Chas. C. Adams. . . . 249 3. Annotated List of the MoUusca of Isle Royale, Michigan, by Bryant Walker 281 I . Introduction 281 II. Faunal Affinities 281 III. Annotated List 283 4. Report on the Isle Royale Orthoptera of the 1905 Expedition to. Isle Royale, by A. P.Morse. : 299 I. General Remarks 299 II. Annotated List of Species 502 III. Station List, 1905 Collections.. 5. Neuropteroid Insects from Isle Royale, Michigan, by Dr. James G. Needham. . . . 305 6. Diptera of the 1905 University Museum Expedition to Isle Royale, by Prof. James S. Hine 308 7. Annotated List of Isle Royale Hymenoptera, by E. S. Titus 317 8. The Ants of Isle Royale, Michigan, by Dr. Wilham Morton Wheeler 325 9. The Cold Blooded Vertebrates of Isle Royale, by Dr. A, G. Ruthven 329 10. Annotated List of the Birds of Isle Royale, by Max Minor Peet 337 I. Introduction 337 II. Classified List of Birds Observed in 1905 339 1 . Summer Residents 339 2. Migrants. . 339 3. Winter Residents (migrants from the north) 340 4. Permanent Residents... 340 viii CONTENTS. Page. III. Annotated List 340' 11. Notes on Isle Royale Mammals and their Ecological Relations, by Dr. Chas. C. Adams 389 I. Introduction 389 II. Mammal Successions ^ 390 1. Lake-Pond-Swamp Series 391 2. The Land Series 392 III. Faunal Affinities and Migrations •. 393 1. The Geographic Affinities of the Fauna .' 393 2. Post-Glacial Origin of the Fauna 394 IV. Annotated List 396 INTRODUCTORY NOTE. Tlirough the generosity of ^Nlr. Bryant Walker, of Detroit, Hon. Peter White* and ilr. H. M. Kanfman of Marquette, the University Museum of the University of Michigan was able, in the summer of 1904, to send a party to the Porcupine ^Mountains and to Isle Royale, Michigan. The aim of the party -was to collect specimens for the museum and to make an ecological suiTey of the regions visited. The party was only able to sjjend a few weeks on Jsle Royale at that time, but through the continued generosity of ^Mr. White and Mr. AValker, the survey was continued dur- ing the summer of 1905. The present volume on the natural history of the island ha.s resulted from these surveys. To Ml-. White and Mr. Walker the Museum is under special obligations for their hearty and substantial support, not only in the funds provided, but also for their aid in securing the transportation of the par-ty. Many other individuals also ai-sisted in various ways. Those who aided the party in the matter of transportation were: Mr. Henry Russel, of the Michigan Central Railway; ilr. Geo. T. Arnold, of the Union Ticket Office and Dock of Mackinac Island; Mr. H. H. Brigham, of the U. S. and Dominion Transportation Company ("Booth Line") ; Mr. Henry Meyering, of the Graham and Morton Line; Mr. M. Adson, of the Duluth, South Shore and Atlantic Railway. The suiwey is furthermore indebted to Section Director C. F. Schneider of the Michigan Section of the Clim- atological Service of the U. S. Weather Bureau, for the loan of meteoro- logical instruments ; to Major Lansing H. Beach, Detroit, of the Light House Establishment, for permission and suggestions as to camping in the abandoned Light-house at Rock Harbor; to Mr. Geo. C. Stone, Sec- retary of the Washington Club of Duluth, Minn., for the use of their grounds and many favors from their care-takers, Mr. Chas. Preulx and Mr. Michael HoUinger; to Mr. K. Neutson, of Park Place ("Neutson's Resort"), Rock Harbor, Isle Royale, for many favors during the stay upon the island; to Mr. J. H. Malone, Keeper of the Isle Royale Light, and to his sons, particularly to the Assistant Keeper, Mr. J. A. Malone, for many favors and for their hospitality. It is a pleasure to have this opportunity of thanking these persons for their cooperation. On the return of the party from the field, work was at once begun upon the collections, and in this a large number of specialists have aided by the determination of the specimens. Acknowledgements are made to such persons throughout the report and will not be repeated here. Those who were not members of the party, but who have prepared papers are: Mr. Bryant Walker, of Detroit, Michigan, Dr. W. M. Wheeler, American Museum of Natural History ; Mr. A. P. Morse, Research Assistant of the Carnegie .Institution, and Wellesley College; Dr. Jas. G. Needham, Cor- ,* Recently deceased. xiv INTRODUCTION. nell University; Prof. J. S. Hine, Ohio State University; Prof. E. S. Titus, Utah Agricultural Experiment Station ; Dr. A. G. Ruthven, Univer- sity Museum, University of MicJiigan, and Mr. A. B. Wolcott, Field Museum of Natural History. The volunteer members of the Museum party should be mentioned in this connection : Dr. K. A. Brown, Dr. H. A. Gleason, Mr. W. P. Holt, Mr. Max Minor Peet, Mr. Otto McOrearj^, and the writer. It will be evident that the volunteer work of this report comprises the major part of it. Personally the writer wishes to express his appreciation of the assist- ance of Mr. Walker and Mr. White; of the cooperation of the members of the party and the many specialists who have examined the specimens; and of the valuable suggestions and assistance of : Mr. Norman B. Conger, Inspector U. S. Weather Bureau, Detroit; Dr. Glover M. Allen, Boston Society of Natural History; Mr. Frank Leverett and Mr. P. B. Taylor. of the U. S. Geological Survey; Prof. H. F. Wickham, State University of Iowa; and to Mr. A. B. Wolcott, of the Field Museum of Natural History. Also to Dr. A. C. Lane of the Michigan Geological Survey for many favors and courtesies, including the preparation of the topographic map, and to Dr. A. G. Ruthven, Chief Field Naturalist of the Survey, for assistance in the publication of the report. The shortcomings of this report will be no more evident to any one than to the writer. If, however, with its defects, it preserves some "van- ishing data," and presents suggestions for the improvement of such ecological surveys, it will have served the purpose for which it was intended. CHAS. C. ADAMS. July 23, 1908. Hull Zoological Laboratory, University of Chicago. ERRATA ISLE ROYALE REPORT. Page 2, line 29, foi- and the read for the. Page 5, line 30, for sources read resources. Page 11, line 43, for larger read large. Page 18, line 11, for White Spruce read Black Spruce. Page 14, line 7, for has lieen read TkmJ Beeii. Page 15, line 16, for Cicada read TiVicem. Page 16, line 40, f or. anti-ZJow read ant-lion. Page 19, line 28, for hardwood read hardwoods. Page 21, line 24, for HibWscus read Hippiscus (and elsewliere in the report). Page 21, line 25, for versicolor read plckeringi. Page 21, line 48, for Limnaea read Lymiiaea (and elsewhere in the report). Page 22, line 26, for 2F read 2/). Page 22, line 43, for Aechna read Aeschna. Page 26, line 21, for Grophaena read Gurophaena. Page 27, line 2, for Wllow read Mllowp. Page 29, line 14, for J'/o. Jp read Fifls. .{5 and 57. Page 46, line 21, for the hearing of the latter read their. Page 47, line 46, for e read i. Page 48, line 14, for Fig. 53 read /''iff. 55. Page 50, line 33, for f read 5. Page 61, line 28, dele (Fig. 29). Page 63, line 10, dele Fig. SO. Page 63, line 16, for Lake read Lakes. Page 64, line 3, for Fig. Z2 read Fig. «6. Page 65, lines 48 and 49, for Formica adamsii read Formica adomsi. PsLge 65, lines 50 and 51, dele No. llJt. Page 77, line 27, for Grophaena read Gyrophacna. Page 93, line 25, for XI read YI. Page 110, line 6, for condition of weather read condition of sky. Page 135, line 31, for fostered read forested. Page 152, line 29, for Burns, F. Z. read Burns, F. L. Page 3 58, line 43, transpose Buprestids and Trichias. Page 159, line 9, for Grophoema read Gyrophacna. Page 161, line 44, for Their read tlietr. Page 188, line 28, dele William. Page 205, line 46, for BoUtoWas read Boletohius. Page 205, line 46, for Ney Jersey read A'cio Jersey. Page 257, line 17, read Saltieidae=AtUdae. Page 261, line 43, for Jassidaeae read Jassidae. Page 284, line 28, for 6^-62 read ii'tiys. 61-6$. Page 306, line 26, dele S. Page 306, line 28, add 3. Page 306, line 29, add S. elongata. Page 342, line 23, for Fig. 1,5 read J'ijy. 57. Page 350, line 35, add Fig. 60. Page 354, line 17, add Fig. 11 ; line 2/5, dele Fig. 17. Page 393, line 15, for influences read inferences. Page 397, line 7, for Oontoa read Caton. Page 407, line 26, for J. N. Malone read J. II. Malone. Page 419, line 14, for Hoops .read Hoopes. ISLE EOYALE AS A BIOT.IC EN\'IKONMENT. BY Dlt. CHARLES C. ADAJJS. I. INTKOnrCTIOX. 1. Itinerary and Personnel of the Paiii/. The University Museum party left Ann Arbor, Michigan, June 21), and reached the abandoned light-house at Rock Harbor, Isle Royale, on tlic morning' of July 5, 1905. The party was composed of the fcdlowing: Is'. A. Wood, Dr. E. A. Brown, Dr. H. A. Gleason, ^V. P. Holt, Otto Mcf'rcary, a camp hand, B. F. Sa.very, and the writer. In general, the duties of the various members were as follows : ^Ir. Wood, the ^Inseum taxidermist, looked after the trapping of mammals and the preparation of bird and mammal skins. He a\ as assisted by Dr. Brown, who gave most of his at- tention to the stvidy of the, bird life, and ^^'ho remained with the party until July 25. In the study bf bird life. Dr. Brown, Mr. Wood and Mr. McCreary co-operated, the latter devoting his entire time to the ecolog- ical phase of the work. Dr. Gleason devoted his attention to the collec- tion and ecological study of invertebi-ates, particularly molluscs and in- ' sects, and most of the photographs were taken by him. In collecting in- sects about the camps, he was assisted by B. F. Savery. Jlr. Holt's time was devoted to the study of the vegetation. The writer, who was in charge of the expedition, gave special attention to the environmental dynamics, biotic succession, and the general correlation of the work of the various members of the party. During the stay at Rock Harbor, Pie;. J, the following localities were examined : The shore, from the light-house south to the head of Conglom- erate Bay; the region about the head of Rock Hai'bor and Sumner lake; a. line from the mouth of Benson brook to Sargent Lake and Mc- Oargoe Cove.; and the vicinity of Tobin Harbor; in other words, the localities included in Stations I-IV. The party remained at Rock Harbor fi'om July ~) to August 1, and then moved to Siskowit Bay. Here Mr. Max M. Peet joined the party on August 8, and devoted his attention to collecting birds f nd mammals. He also took a number of the photographs. While at the Siskowit Camp, the bay and lake of that name were examined, and also the Hay- town trail and the islands near the Isle Royale Light. All of these lo- calities are included in Station V. On August 17 the party moved to Washington Harbor, and was then partially disbanded. The members who remained made their headquar- ters on the grounds of the Washing-ton Club, at the head of 'S'i'ashington Harbor. After September .5 ^Ir. Peet alone remained there until the 22nd, in order that he might continue the study of the fall migration 2 MICHIGAN SURVEY, 1908. of the birds. He returned on the last boat of the season for Duluth, Minn. During the previous (1904) season, the Museum party had made n general examination of the vicinity about Washington Harbor, so that it was now thought desirable to devote more time to other localities. With the exception of bird migration, no detailed work was done in this vicin- ity in 1905. In addition to the region about the head of the harbor, Lake Desor was also reached from this point by means of the road along the crest of the Greenstone Eange. 2. The Aim and Methods of TFor/i;. The field work was conducted on the same general plan as that pursued during the previous season in the Porcupine Mounitainis and a.t Washinglton Harbor. Much more ground was covered, however, because it seemed improbable that a third trip could be made to the same region. It therefore seemed desirable to gain some idea of the biota of the island as a whole, because of its- Canadian character. Even then, the survey was confined almost exclu- sively to the region south of the Greenstone Range. As mentioned in the report for 1904 (Ruthven, '06, pp. 11-12), the aim of the work was not simply to collect specimens, but also to study the relations of the plant and animal life (the biota) to their surround- ings. The environment as well as the biota was considered from a dyn- amic standpoint, and aU effort was made to analyze the environment in order that the dominant conditions and processes of which it is com- posed might be recognized, and their laws of change be perceived and formulated. To resolve such a problem as this must of iiccessitr require more time and detailed investigation than the possibilities of a few months work will permit, and yet it is equally evident that prelimi- nary work should be carried on from a genetic standpoint, because such a method determines upon what facts emphasis should be placed, and th'* broader and more general relations, as well as the details, are equally subject to a genetic and dynamic treatment. In preliminary work of this character, it is considered of special importance to discover, if pos- sible, the order of the major biotic successions, because these succes- sions must be clearly perceived before their causes can undergo ade- quate analysis. Our knowledge of causes generally lags far behind our recognition of successions. Thus throughout the study of the Isle Royale biota a special effort was made to investigate the genesis or successions of events. The en- vironment has not been considered as limited to habitats alone, but also to include that greater unit, the geographic. To ignore this is to overlook the real background. It is believed that certain advantages are derived from this method of work, which, although they may be recognized from other points of view, are likely to be subordinated to other facts. It should not for a moment be thought that this method is considered as the only one of approach, but it appears to have certain advantages which seem to justify its adoption. Nor should it be inferred that the genesis of the biota and the habitat is all that should be in- cluded in an ecological survey. The problem of succession is only one of several which clearly show the intimate relations and responses between organisms and their environment ; others tliat remain, to be investigated involve physiological and structural changes, and various modifications of habit and behaAior of both plants and animals. ECOLOGY OF, ISLE ROY ALE. 3 The ecological relations in the north are so different, in some respects, from those farther south, that one may easily form an erroneous con- ception of the conditions under which such a preliminary investigation may be made. A very favorable condition for the work was the fact that practically all the time was devoted to it, instead of only occasional trips being made for the purpose, as is necessary with those busy with other duties. There are also certain advantages in being able to be in the field continuously, as a certain familiarity with conditions is ac- quired in the beginning, especially where the variety of forms is limited, which otherwise would involve time tipon each visit. Although most of the members of our party were upon the island only during July and August, yet at this time those seasonal phenomena were concen- trated which require much more, or several times that amount of time for their development farther south. The seasonal contrast is well il- lustrated when the summer season at the other extreme of the State of Michigan — 500 miles away — is contrasted with that of Isle Eoyale. Such relations are further reinforced by the fact that the species and societies which are dominant in the various northern habitats are very much smaller in number than farther south. This necessarily sim- plifies the j)i'oblem, and to a corresponding degree reduces the chances of error in anticipating biotic changes which are correlated with those of the environment. This is a relation of much importance in the study ■of succession. The writer was especially impressed with the relative simpUeiti/ of the problem of environmental relations and of the biotic succession upon Isle Eoyale, and has received further confirmation of the opinion that a tendency to exaggerate the complexity of the environ- ment is prevalent. An important aid in environmental analysis has been received from the effort to distinguish the major or geographic features of the gross environment from the minor habitat unit.S which make up the mosaic or complex, although their mutual and genetic relations were not overlooked. Some of these relations have been well expressed by Montgomery in his comment on distribution ('06, p. 6) as follows: "And, as is always the case Avhen. the method has been consistent and scientific, the factors of distribution and the meaning of it will ulti- mately be stated in ^'ery simple form. These factors appear to us now to be enormously complex, but this is because we have hardly com- menced to analyze them." At this point it should be mentioned that there are certain dif- ficulties which tend to confuse the field worker, which, if clearly understood, will often be of aid in ecological studies. In pursuing field studies, in addition to a knowledge of the species, one of the first essentials is a familiarity with the habits and habitat preferences of organisms; and further, there should be the ability to recognize how the dominance of one society is transformed into that of another. The lack of a sufficient power of constructive imagination makes the detection of such transformations very difficult, perhaps even impossible to some. This limitation almost completely restricts such a student to the purely descriptive phases of field ecology, because the explanatory phase lies beyond his grasp, although there remains for him a large 4 MICHIGAN SURVEY, 1908. field for useful and valuable activity in the study of habitat preference, and the mutual relations of the associated species in given habitats. A familiarity with the forms studied, under diverse circumstances, de- velops a certain perspective which is a great help in preventing confusion caused by minor and relatively insignificant details. The limited time spent in the present investigation did not permit detailed studies of the interrelations of the organisms within the habitat, either in their relation to the environment or to each other. In local studies attention is usually given to detailed life histories rather than to a deliberate study of their interrelations as members of a society. The emphasis which is sometimes placed upon individual life histories would lead one to expect that such histories could be assembled and would give us the same result as if they had been studied as a society. But the points of view are so different that such a result, although theoretically possible, is unlikely to be obtained. At this time we only wish to emphasize the fact that both methods should be used to secure the best results. For example, in applying these principles to the study of birds, the life histories of the dominant species of a society might first be given special attention. Then the relations of the dominant species to others of the association and to the environment may be deter- mined and later on the subordinate kinds considered. This will involve prolonged study in the field (and laboratory) of the habits of nesting, feeding, rearing of young, etc., as influenced not only by other mem- bex's of -the same species, but also by other species in the same habitat. The same general method is applicable to other groups of organisms. 3. Previous Biological Investigations upon Isle Royale. Previous to the investigations by the Museum party in 190-4, (Euthven and others, '06) there seems to have been very little study of the Isle Royale biota. Several collections of plants and animals have been made, but very little has been published about them. In 1848 W.. D. Whitney was "ornithologist and botanist" for the government geological survey parties, and he pub- lished a list of the plants found. (Foster and Whitney, '51, pp. 359-381). Incidental mention is also made in these geological reports of the collec- tions of animals (Foster and Whitney, '50, pp. 17, 51, 201; Jackson, '49. pp. 423, 440, 441,) ; but, so far as known to the writer, no detailed reports were published. So far as the vegetation is concerned, the most important source of information is the maps of the Ives Linear Survey. Here the general character of the forest, the extent of the swamps, and the underbrush are indicated. Mr. Henry Gilman ('73), of Detroit, made two visits to the island (one of which was in 1873), and his botanical and ethno- logical collections were presented to Columbia and Harvard Colleges. In 1890 Mr. F. E. Wood made a collection of plants from the vicinity- of Rock Harbor and presented them to the herbarium of the Botanical Department of the University of ]\Iichigan ; and in 1901 W. A. Wheeler ('01) published a short paper on some plants taken on the northeast end of the island. The invertebrate fauna found in the deep water off Isle Royale was examined by Smith in 1871, and .a list of Coleoptera from Isle RoyaJe was published by Hubbard and Schwarz ('78). The writer has recently published a paper on certain phases of the problem of succession, as ECOLOGY OF ISLE ROYALS. 5 illustrated bj' the birds upon Isle Royale. This paper, with some ad- ditions, is included in this volume. Detailed references to these papers will be found in the accompanying bibliographies. From the above remarks, is it quite evident that very little attention has been given to the biological conditions of the island, and much remains to be done. In all probability other naturalists have visited the Isle, but I have not learned of their results. 4. Historical Note. The history of Isle Royale, since its cession by the Chippewa Indians in 1843, is, in brief, one of prospecting, mining explorations, fishing, summer resorts, and scientific surveys of the topography, hydrology, geology, and biology. A general historical account is given in Lane's report ('98) on the geol- ogy of the island, and need be mentioned here only in outline. There is abundant evidence that in prehistoric times the Indians mined copper on the island. Within three or four years after the cession of the island to the United States, it was invaded by prospectors and explorers, so that by 1847, according to Lane, "the island presented perhaps as lively a scene as ever in its history." At this time the Linear Survey was made by William Ives. But this period of activity was only of short duration, for the decline was almost as rapid as had been the ascent, and by 1855 the "island was a desert once more, with no permanent in- habitants." (Lane). This passive condition of affairs lasted until the Lake Survey engineers arrived in 1867. This survey continued until 1871, during which interval explorations were somewhat revived, and continued for several years, only to be followed by another relapse and still another ascent in 1891, when a number of careful and detailed explorations were made for copper by means of the diam^ond drill. But this activity also ceased about 1892. A year later, and again in 1895, Dr. Lane visited the island for geological investigations. The mineral sources are thus seen to have been the main attraction. The forest growth is too stunted and inaccessible to have merited the attention of lumbermen, although several timber prospectors were pre- sent during the summer of 1905. During more recent years the fish- ing and summer resort business have attracted some attention to the island, and have made it accessible during the summer through regular steamboat service. The climate, sceuery and the fishing make the island very attractive as a summer resort, but it should be recognized that if the scenery is to be preserved the forests must be protected from fires, because reforestation is exceedingly slow on land with such a shallow soil. It is to be hoped that the geographic isolation may be a protection from such devastation, beca,use the cool summer climate, the rocky coast, the forests, the picturesque scenery, and the surrounding Lake Superior, are natural features which should long- remain attractive to summer visitors. If the interest in copper should revive permanentlj^, the biota Avill become greatly modified, in which case some conception of the conditions upon the island in 1904 and 1905 will be preserved by these records. [It may be of interest to note here that 86,000 acres of the island were for sale in the winter of 1908 for $150,000. Lane.] 5. Available Maps of Isle Boyale. The available maps are not generally known to the public, and are therefore listed here, especially those which are of value from a biological standpoint. e MICHIGAN SURVEY, 1908. 1. The Ives Linear Survey Maps. Because of their large size (2 inches to the mile), and the details concerning the character of the swamps, the forest and the soil, this is the most useful map for the field. Photolithographic copies of the township maps, of which there are eighteen, may be secured for 25 cents each from the General Land OflSce at Washington, D. 0. 2. The U. S. Lake Survey Chart of Isle Eoyale, (Catalog No. Sh.). This is very useful because it indicates the topography, in part by hachures, and gives the details of the coast, including soundings and the character of the bottom. A large tract of the interior, between lakes Desor and Chickenbone, is unmapped. This map may be secui'ed for 25 cents from the Lake Survey Ofiflces at Detroit and Duliith. An excellent chart of the entire Lake Superior basin may be secured from the same source. 3. Lane's Geological Map. Published by the Michigan State Geo- logical Survey. It accompanies Lane's report ('98) on the geology of the island, and is on a scale of % of an inch to the mile. 4. Passage Island Topographic Sheet. This is the only sheet pub- lished by the U. S. Geological Survey which includes any part of Isle Koyale, and it covers only the extreme northeastern end of the island. This may be secured from the Survey for 5 cents. The contour inter- val is 20 feet, and the scale one inch to the mile. 5. An English land company is said to own much of the island, and has published a map on a scale of "'/^ of an inch to the mile. The agent for this company is R. E. Goodell, Houghton, Michigan. II. THE BIOTA CONSIDERED BY STATIONS. 1,, Tile Location of Field Stations in 1905. As a detailed survey of the entire island was impossible, it was necessary to select representa- tive localities and conditions, or habitats, and to devote to these alj available time for study and collecting. In order to make sure that these conditions were representative, considerable care was necessary in locating these stations. In general a Station, in the strict sense, stands for a region, while a Substation refers to a particular habitat, usually of relatively limited extent. The character and extent of a Substation, (or, as it is generally called, for the sake of brevity, a "station,") was determined primarily by the relatively homogeneous character of the conditions. Thus a "station," as the Balsam-Spruce forest (V, 4) for example, varied somewhat in its extent with different groups of organisms. In the case of birds it included a greater area than was necessary for many invertebrates, such as land snails, but in every case such a "station" is intended to enable one to determine wiiat organisms were dominant and characteristic of such a sample situation. Some such system of sampling is generally advantageous or necessary, and this is particularly essential in the case of a surveying party, in order to give deflniteness and co-ordinated activity to their work, parti- cularly if the results are to be made at all comparable. Of course some individual judgment is necessary in applying such a plan to diflferent groups, but no more perhaps than is necessary to carry out any other comprehensive plan. ECOLOGY OF ISLE ROY ALE. 7 1. Location of Field Stations, 1005. Station I. Light-house Peninsula, between Rock Harbor and the head of Conglomerate Bav, Sec. 26 and N. B. y^ Sec. 34, T. 66 N., E. 34 W. Sub. 1. Lake and Bay Beaches. Sub. 2. Natural Rock Clearings, N. E. % Sec. 26.. Sub. 3. Balsam-Spruce Forest, N. E. % Sec. 26. Sub. 4. Tamarack, and Arbor Vitae Swamps, Sec. 26. Sub. 5. Jack Pine Ridge, S. W. % Sec. 26 and S. E. % Sec. 27. Sub. 6. Sphagnum-Spruce Bog, S. W. % Sec. 26 and S. E. % Sec. 27. Sub. 7. Light-house Clearing,"^. W. % Sec. 26. Station II. Rock Harbor and McCargoe Cove Trail, Sec. 27, 22, 21., 20, 29, 30, T. 66 N., R. 34 W., and Sec. 25 and 26, R. 35 W., T. 66 N. Sub. 1. Benson Brook and Ransom Clearing (outlet of Benson Lake), N. E. % Sec. 27 and S. E. % Sec. 22, T. 66 N., E. 34 W. Sub. 2. Tamarack Swamp, S. W. 14 Sec. 22 and S. E. % Sec. 21, T. 66 N., R. 34 W. Sub. 3. Rock Ridge Clearings (burned over). Sec. 21 and 20, T. 66 N., R. 34 W. Sub. 4. McCargoe Cove, at end of Trail, X. E. 14 Sec. 20, T. 66 N.. R. 35 W. Sub. 5. Forbes Lake, N. E. y^ Sec. 28, T. 66 N., R. 34 W. Sta.tion III. Western End of Rock Harbor, Sec. 28, 33 and 32, T. 66 N., R. 34 W., and Sec. 5 and 4, T. 65 N., R. 34 W. Sub. 1. Small Island, S. E. 14 Sec. 32. Sub. 2. In Harbor at West end of Island, Sub. 1. . Sub. 3. Bulrush Zone and Delta, Sec. 32, T. 66 N., R. 34 W. Sub. 4. Trail to Sumner Lake, Sec. 33, T. 66 N., R. 34 W. Sub. 5. Sumner Lake, Sec. 33 and 34, T. 66 N., R. 34 W. Sub. 6. Southwest Coves of Rock Harbor, Sec. 5 and 4, T. 65 N., R. 34 W. Station IV. Tobin Harbor and Vicinity, T. 66 and 67 N., R. 33 W. Sub. 1. Scovill Point, Sec. 26 and 35, T. 67 N., R. 33 W. Sub. 2. Island No. 14, Sec. 26, T. 67 N., R. 33 W. Sub. 3. Bavou, North of Monument Rock Trail, N. W. % Sec. 34, T. 67 N., R. 33 W. Sub. 4. Trail to Monument Rock, N. W. % Sec. 34, T. 67 N., R. 33 W. Sub. 5. Clearing at Neutson's Resort (Park Place), Sec. 4, T. 66 N., R. 33 W. Sub. 6. Small island in Tobin Harbor, Sec. 5, T. 66 N., R. 33 W. Sub. 7. Head of Tobin Harbor, Sec. 7, T. 66 N., E. 33 W. Sub. 8. Trail to Greenstone Range, Sec. 7, T. 66 N., R. 33 W., and Sec. 12,- T. 66 N., R. 34 W. Sub. 9. Mountain Top, Sec. 12, T. 66 N., R. 34 W. Station V. Siskowit Bay, Lake and Vicinity. Sub. 1. The Beach, (at camp). Sec, 32, T. 65 N., R. 35 W. Sub. 2. Heath Zone and Beach, Sec. 33, T. 65 N., R. 35 W. Sub. 3. Rock Clearing (at camp). Sec. 32, T. 65 N., R. 35 W. 8 MICHIGAN SURVEY, 1908. Sub. 4. Trail through Balsam-Birch Forest, Sec. 32 and 31, T. (!o X., E. 35 AV. Sub. 5. Tamarack Swajiip, N. W. i/i Sec. 32, T. 65 N., E. 35 W Sub. 6. South Shore of Siskowit Lake, Sec. 31 and 32, T. 65 N., R. 35 W. Sub. 7. Havtown Trail, from Siskowit Lake, West Line of Sec. 2i, across Sec. 13, T. C5 N., R. 36 W., cf. Lane, '98, ])1. XL Sub. S. Arbor Yitae Swamp, at end of Haytown Trail, N. W. U Sec. 13, T. 65 N., E. 36 W. Sub. 0. Outlet of Siskowit Lake, N. W. V^ Sec. 36, T. 65 N., R. 36 W., and Sec. 31, T. 65 N., E. 35 W. Sub. 10. Louo- Island (lull Rookery and Menagerie Island, T. 61 X., R. .!.-. A\'. Sub. 11. Tamaiack-Spruce Swamp, Sec. 33, T. 65 X., R. 35 W. The following stations were examined by the Museum party during the season of 1904. Tart of these Stations were re-examined and will be referred to by Station number and date, thus : Sta. I, '04. Sration I, '04. Clearing on the Shore of Washington Harbor, Sec. 21), T. 64 X., R. 38 W. Station II, '04. Washington Cvpck, Sec. 29, T. 64 X., R. 38 W. Station III, '04. Trail along the top of Greenstone Range (Desor Trail). T. 64 X., R. :!7, 38 W. Station TV, '04. Washington Brook, Sees. 2S and 32, T. 64 X., R. 38 W. Station Y, '04. Tamarack Swamp, Sec. 20, T. 64 X., R. 38 W. Station VI, '04. South of Greenstone Range, Sec. 32, T. 64 X., R. 38 W. Station YII, '04. Lake Desor, T. 64 X., R. 32 W. Station YIII, '04. Western end of Siskowit Bay, Sees. 27 and 28. T. 64 X., R. 37 W. Station IX, '04. Southwestern end of Jlinong Trap Range, Sec. 30, T. 64 X., R. 39 W. Station X, '04. Washington Hai'bor. T. 64 X., R. 38 W. 2. General Characteristics of the ^tation^H. In this section, I do not aim to give a completely correlated account of the biota of each sta- tion, but to present a general idea of the main characteristics of the various situations examined, and some of their common and represen- tative plants and animals. Photographs illustrating the charac- teristics of the various "stations" will accompany this section, and should be consulted in connection with the text. Station I, Huhstation 1. The Lake and Bay Beaches. This "station"' includes the shore line from Rock Harbor, near the light-house, Fie/. ] . to the head of Conglomerate Bay. The entire shore was not studied in detail, as most of the time was devoted to the beaches which are being formed at the heads of the coves and bays. Quite a variety of condi- tions are represented along this shore, due not only to the degree of exposure to the Avaves of Lake Superior, but also to the character of the rocky coast itself. All degrees of shore and beach are developed, from overhanging and vertical cliffs. Fig. 2, with bases strewn with large blocks lowered by sapping, to a shore line with a low angle strewn with shingle and gravel, and a sandy beach, as found at the head of Conglom- erate Bay. In harmony with the dip of the rocks and the effect of the ECOLOGY OF ISLE ROYALE. 9 glacial ice movement upon the valley slopes, whicli tend to be gentle on the southeastern side, the corresponding shores of the bays and coves are usually at a low angle, except possibly where faulting has taken place, or a wave cut terrace has been developed. The northern sides of the bays are comparatively abrupt, and there is thus a tendency for the cliffs to occur mainly upon the northern slopes and shores. The larger bays are the submerged portions of the valleys, mark the location of the less resistant rocks, and are inherited topogi'aphic fea- tures; but many of the minor coves and the rocky headlands have been carved by the activity of the present lake. The beaches are only de- veloped at the heads of the coves and bays, and are very largely com- posed of shingle and gravel. The only extensive sand beacli seen was at the head of Conglomerate Bay. The character of the material com- posing these beaches clearly shows its local origin, and emphasizes the isolation which prevents long shore transpoi'tation of such material. Thus only floating material is liable to extensive long shore dispersal, a signifi- cant fact that bears upon .the dispersal of the snail life along the shore. During severe storms, the wave action upon this coast is quite in- tense and even the waves of the summer storms are quite active, as may be seen by referring to Fig. 3. The blue deep lake water comes close up to the shore, so that generally no breaker line is developed off shore. In several places there are numerous reefs or islands (usually the iso- lated continuations of the rock ridges), which tend to bi"eak the force of the waves rolling in from the open lake. No effort was made to study the life of the open lake, only the shallow water of the bays and coves being examined. The major environmental features of the coast are the Lower, Middle and Upper Beaches; but these are only differentiated clearly at the heads of the coves and harbors. The Lower and Middle Beaches are only seasonal expressions of the same phenomena, but ecologically they are fairly distinct. The Loiccr Beach. This beach extends from the shallow water to the upper limit of the summer waves. The submerged portion is not sharply defined above on account of the changes in level of the water surface, due to waves, the periodical and seasonal fluctuations, and the atmospheric pressure (seiches). In time there has been a downward migration of the entire beach zone, a tendency which is in part counter- acted by the northward elevation of the land. This is the zone domi- nated by water, ice, and wave action. It is certainly a sharply defined tension line upon an exposed coast which clearly suggests that it is not probable that many forms of animals have made the transition from fresh water to the land under such conditions. If we consider the shore habitats as including all stages from a rock cliff to the sand beach, the lower beach and the protected shores are the most favorable aquatic habitats upon such shores. Upon the sloping rock, shingle, gravel and sand beaches is found a varied fauna. In winter, when the bays are frozen over, a calm is pro- duced which must be favorable to the preservation of the aquatic life upon this stormy coast. The general character of the sandy beach at the head of Conglomerate Bay is shown in Fig. Jj. The life of the submerged portion of the shore is quite limited, except on the beaches and protected portions. The vege- 2 10 MICHIGAN SURVEY, 1908. tation consists of algae, which grows in moderate abundance, though not luxuriantly, as found about the Gull Eookery (V, 10), or at the fishermen's camp at Rock Harbor, a fact which suggests that the aibun- dance of suitable nitrogenous material is much greater in such places than in the open lake water. With the development of the fall storms, Mr. J. A. Malone states that these rocks (V. 10) are washed free of the algae, thus evidentlr necessitating a repopulation of these surfaces each season. The characteristic fauna secured in the shallow water shore margins were the snails, Limnaea stagnalis. L. emarginata, and Physa sayii. A small fish, the Miller's Thumb, Z' ran idea franklini, is also fairly abun- dant and characteristic of this shore. Upon low rocky shores beach pools, Fig. 5, are occasionally found which, when favorably located, are supplied with water by the ordinary summjer waves, otherwise by storm waves and rains. The precarious existence of life in such places is indicated by the general type of the fauna, which shows exceptional power of locomotion, usually coupled with a short life cycle. The immature stages of insects are rather characteristic, as shown by nymphs of the water boat- men, Corixa, dragonflies and Caddis fly larvae. Water beetles were represented by Rhantus Mnotatus, and the snails by Limnaea emarginata and Planorbis parvus. The Gulls and Spotted Sandpipers should be mentioned as birds which frequent these conditions. The Middle Beach. This beach occupies the strip of shore over which the winter waves retreat as they fall to the upper summer storm limit. It is thus seen that the Middle Beach is only a temporary or summer aban- donment of part of the upper shore, which is repeatedly claimed by the winter waves. In summer this strip is exposed to denudation; in the fall amd early winter, to the fury of the waves, and, later, it is covered with ice. Driftwood and debris tend to lodge here and to accumulate. It is an important region of biotic invasion for land forms. Beach pools are also developed in this area, upon the abandoned wave cut ter- races of earlier lake levels. Upon the cliff faces, sloping rock shores and shingle beaches, little is found that is favorable to life, but upon the protected sand of the Middle Beach, relatively favorable conditions for many organisms are found during its period of exposure. The character of the substratum of the Middle Beach varies from rock to shingle, gravel and sand. The characteristic features of the vegetation, where the wave action is not too severe, are the fruits which are washed ashore by the waves, together with certain annuals and lichens. The fauna varies with the character of the conditions. The open character of this beach and the relative abundance of animal food makes such situations favorable for spiders of the genus- Pardosa. The same open character makes the shores a favorable patrol for certain butterflies, particularly Basilar- cMa arthemis. Insects and snails washed ashore by the waves also characterize this habitat. The Upper Beach. This part of the beach is beyond the reach of the waves, and forms the transition between the open beach area and the inland forests. The width of this belt varies greatly with the gradient of the shore. Where the beach is continuous with a more or less bare rock ECOLOGY OF ISLE ROYALE. 11 ridge, this habitat may be rather extensive and ill defined, as at the ridge south of the light-house (I, 2), but when it borders a depression, as at the head of the rockbound coves, or where a beach is well developed, this transitional zone is more clearly defined and limited. When this beach is wide and grades into the rock openings, as in F'if/arrs 6 and 7. the crustaceous and foliaceous lichens grow upon the rocks; but if soil accumulates, as is shown in Fig. 6, the Cladonia — Bearberry society becomes established, and includes some annuals, svich as Solidago. A limited variety of insects, especially ants, characterize such conditions. When adjacent to the forests, in depressions, this beach is generally bor- dered by alders, some aspens and young trees. The fauna consists largely of insects, such as butterflies, certain dra- gonfiies and Hymenoptera, which freqvient the open places on wing. Station I, Substation 2. Natural Rode Clearings. This Station con- sists of two small rock openings, one just north of the light-house, and the other south of it, on the north side of the entrance to Tonkin Bay. only a short distance from the light-house. They were both park-like ave- nues extending along the ridges, largely bordered hy the Balsam-Spruce forest. The north ridge will first be considered. The general character of the opening is well shown in Fig. S. The White Spruce, Balsam, Paper Birch and Arbor Vitae bound the ridge on either side, within which there is a distinct heath zone of Bearberry and patches of Cladonia, while along the central aisle there is a shallow residual and humic soil on the almost bare rock. The south slope is rather gradual, but the north slope and the end of the ridge at the shore form a cliff. The fauna of this location was limited. Snails were found among the Cladonia, such as Vertigo, Zonitoides arhorea and Pyramidula crorikheitei anthonyi. This was also a runway for Hares. The south opening or clearing is situated on a low sandstone ridge which slopes down to the beach, and is thus in marked contrast to the north clearing, which ended in a cliff. This gradual slope beautifully illustrates the transition from the bare rock beach, through the moss and lichen zone, to the Cladonia, Bearberry and Solidago flora, (Figs. 6 and 7), and on to the crest of the ridge, Fig. 9, with its dominance of Cladonia and Bearberry. The severity of the conditions is furthered by the weathering of the sandstone into thin scale like layers, about % of an inch thick, which become loosened and slide down the slope. Thvis a vegetation may become fixed to the rock surface,, but not permanently to the slope. These scale like fragments are shown in Fig. 6. That a greater amount of vegetation would grow here, if the soil were allowed . to accumulate, is shown in Fig. 6, where such conditions have been produced by the presence of a larger boulder. The Cladonda-BeaTbevry avenue extends along the crest of the ridge. Fig. 9. This is bounded by large Jack Pines near the beach, and farther from the shore by the Balsam-Birch forest. The zonal distribution on the ridges is quite marked ; the central strip is composed of Cladonia, Bearberry, Solidago, and Linne-a iorealis; while this is bordered by a shrub zone composed of Juniperus nana, alder, Arbor Vitae and young Balsams, and a bordering tree zone is com- posed primarily of Balsam. When once the shade of the forest, es- 12 MICHIGAN SURVEY, 1908. specially that of the Balsams, encroaches upon the Cladonia society, the Bearberry first becomes reduced in number, and is then replaced by Aster macrophyllus, and a moss from the forest floor. The former is perhaps the most striking and characteristic shade plant upon Isle Koyale. The succession, or order of invasion on the ridge, from the Cla- donia to the Juniper and into the BaJsam foi-est, is thus briefly shown in the transverse section from the central ridge to its margin. This zonal phenomenon, as will be seen later, is only an expression of the relative rates of invasion, and is not a phenomenon separate from the normal succession. The soil upon the top of the ridge is about two inches deep. It is residual, supplemented by the humus from a now extinct crustaceous lichen society (that of the Cladoni a -Bearhervy), and at its margins by the Juniper, Balsam, Birch and Jack Pine leaves and debris and further, to an important degi'ee, by the excrement of the numerous Varying Hares which frequent the rock ridges. In the case of rock ridges which entend down to the beach and are thus in direct communication with the shore drift, conditions exist which show how such ridges may have been invaded by lichens from two sources — the shore drift and the exposed beach itself — because of the continuity of the rock habitat. Of course possibly another origin is to be found in the fact that this ridge was itself once a beach. Ants, grasshoppers and a few other Insects characterize this fauna, which is limited in variety, but fairly abundant in individuals. The Hares are abundant and form distinct paths or runways, as shown in Fig. 9. Station I, Substation 3. Balsanv-Wliite Spruce Fwest. This station included the forest traversed by a blazed trail from near the south- eastern part of Sta. I, 2, and extended northward to the clearing about the light-house (I, 7), and beyond it to the north rock clearing (I, 2). Most of the region occupied by the forest is of low relief, with an occa- sional low rock ridge or hill. The dominant tree was the Balsam Fir, with much Paper Birch and White Spruce. Where the forest was very dense, especially if due to the number of Balsams, the ground was densely shaded and there was almost no herbaceous ground cover ; but wherever there was a small opening, due to a fallen tree, or where one had been cut down, there was an abundant growth of Large-leaved Aster and White-flowering Raspberry; and it was in the midst of such conditions that young Balsams abounded. These were very characteristic plants in such conditions. In msost cases ai thick layer of humus covered the ground, but the tree growth was of small size. The common size of the Balsam was about 4 inches, the larger ones reaching 8 to 10 inches. The Birches averaged lai'ger, usually about 6 inches. No evidence of burns were seen, but probably many trees have been cut from this vicinity, because of its proximity to the lightrhouse, and the former Indian camp- ground now occupied by the fishermen. The Balsam appeared to become dominant at this place, as more young trees of this species were seen than of any other. The fauna found in this forest was rather limited, and doubtless great numbers of the insects which were taken in the clearing about the light-house (I, 7), bred in the adjacent forests. This is particularly true of the Cerambycids and other wood infesting beetles, the wood-boring ECOLOGY OF ISLE ROYALE. 13 Hymenoptera {Uroceriis) , and their parasites. In addition to such si^e- cies as feed npon Balsam, White Spruce and Paper Birch and their asso- ciated vegetation, there were those animals Avhich are dependent upon the shade, moisture, soil, decaying logs and other features associated with forests. To this class belong certain insects which frequent decay- ing timber or the fungi growing \ipon them, and the earthworms of the soil, the ground beetles or Carabids, and the ground-inhabiting spiders, Lycosids. Some of the birds found were: Chickadee. Red-breasted Nuthatch, Golden-crowned Kinglet, Whitewinged Crossbill and Purple Finch. Station I, 4- Tamarack and Arl:)or Yitae or White Spruce Sivanvps. This swamp is located in one of the valleys near the head of Tonkin Bay, and extends back from the bay about one-fourth of a mile. It begins just back of the beach and is bordered by a strip of Alders, Paper Birch, Mountain Ash, young Balsams and White Spi'uces. The rock walls of this valley are about 75 or 100 feet apart and are well shaded and covered by lichens and mosses, the south surface largely by lichens alone. Back of the marginal beach strip above mentioned, comes the dense growth of very large Arbor Vitae trees, intermingled with nu- merous large fallen trunks, partially decayed and covered with a dense growth of mosses. In the dryer places the ground is covered with a dense litter, and a thick damp or wet mass of mosses, but no pools of water. The undergrowth is composed of young Balsams, Birch and Ground Hemlock, Fig. 10. Proceeding farther up the valley, the Arbor Vitae is replaced by Balsams and Paper Birch; the forest is more open, and the amount of moss on the ground is greatly reduced, and is replaced by a growth of Large-leaved Aster and large quantities of Ground Hemlock — all of this vegetation being indicative of mesophytic conditions. In this re- gion there are scattered pockets or small pools of water containing dogwoods. Still farther up the valley the Balsams and Arbor Vitae con- tinue and Tamaracks are added, but no standing water was found. The valley turns, and returns to the bay on the north side of the ridge which bounds the Arbor Vitae swamp on the north; the entire basin is thus somewhat horseshoe shaped. The returning sedtion becomes almost pure Tamarack and contains numerous small pools of water. The conspicu- ous feature of this environment is its jungle-like character, the rapid ac- cumulation of litter and humus, and the damp substratum. The fauna of such a bog is surprisingly limited in variety and amount. A few shells were found, as Pi/inmidula cronkhcitei anthoni/i and, ii^ the small pools, Pisidium. The large numbers of Mosquitoes and Black Flies made up for all deficiencies, and were almost intolerable. The birds frequenting this forest were the Eed-breasted Nuthatch, Black- throated Green Warbler and Chickadee. Station I, 5. The Jack Pine Bidge. This ridge is located near the mouth of Conglomerate Bay, on the north shore. Some general idea of the location is given in IPig. 11, which is a view looking toward the head of Conglomerate Bay. Just back of the beach, on an outcrop of conglomerate, was a small rock clearing, with Cladonki. Juniperus nana, and a wild rose. From here the trail extended through a narrow strip of forest, composed of Balsams, White Spruce and Arbor Vitae, with an 14 MICHIGAN SURVEY, 1908. undergrowth of Balsam, Mountain Alder, and a ground cover of Large- leaved Aster, and passed on through a belt of j'oung growth of Birch, with the usual White-flowering Raspberry and Large-leaved Aster, Ficj. 12, and up the face of an escarpment to the crest of the ridge, Avhich had a height of about 100 feet above the lake level. From the abundance and characteristic growth of Jack Pines on this ridge, the station takes its name. Part of the ridge has been burned over, as was shown by the burned and fallen timber, but the part to which our attention was given was apparently an original growth. The Jack Pine was scattered, and largely occupied the depressions and the larger crevices. The ridge is fairly flat topped, but is occasionally broken by transverse gullies, which contain Aspens, Birches, etc. The surface of the lava has weathered but little in some places, the original roche moiitancos surface being very clearly preserved, and the planed glacial surface but little eroded. Near the escarpment, however, disintegration and decomposi- tion have been much more active, probably influenced in part by lake waves at former levels, thereby developing a talus slope, composed of an- gular blocks, and in some places forming a stony soil. All intermediate stages are found between these two extremes. In addition to the large amount of bare rock surface, and that covered by only a thin layer of soil and vegetation, the shallowness of the soil is further evidenced by over- turned trees. Fig. 13. This soil is of residual and organic origin, the crustaceous lichens and the Cladonia-Beavberry society, and later the Jack Pines, having contributed much to its formation. The excrement of the Hares has also been an important factor in soil formation, and that of the Lynx also, though to a much less degree. The process of weathering must be relatively rapid on this ridge, because it is exposed to the winds at all -seasons of the year, and to the marked seasonal and daily changes of temperature. The heat of the noonday sun is excessive, and the radiation from the nearly bare rock must be rapid, as it also is at night, so that the various influences con- sequent to temperature changes are allowed full play. Weathering is further favored by the irregularities of the surface, and the crevices, which allow the accumulation and downward conduction of this moisture, thus permitting the prying action of ice. In general, the succession of plant societiesi on this ridge appears to be about as follows: Lichens are the pioneers on the rock surface, and these may be of several species, Umhilicaria. and the crustaceous and foliaceO'US forms. As a soil develops in the crevices or on the sur- faces, these are followed bv Cladonin. Bearberry, Sibhaldiopsis tridentata, Solidago, Diervilla diervilla (Bush Honeysuckle) ; and later, when the soil becomes deeper, by Amelanehicr, Pruints pcnnsylraiiica (probably dispersed to these ridges by birds) and Jumperws nana.. The presence of the Small-toothed Aspen, willow and an occasional Birch probably in- dicates the next society. In the shade of the Birches and Jack Pines Solidago and Aster maorophyllu.'i occur, if suflicient soil is developed. Prom the character of the vegetation in the ravines which traversed the ridge, and upon the talus slope toward the bay, it is apparent that the next society tends to be that of Birch and Aspen with some Balsam, Pennsylvania Cherry. Mt. Alder; and a ground co\er of Large-leaved Aster, Large-flowering Baspberi-y, Ground Cornel and Lycopodiiiw. It ECOLOGY OP ISLE ROYALB. 15 is clearly seen that among these there are several elements of the Balsam, White Spiruce and Birch forest society, which tends to ulti- mately possess the ridge. The fauna of the ridge is quite diversified, and there is a general faunal correlation corresponding with these successions of the vegetation. Thus during the Lichen-Heath stage, ants and spiders, certain shells, and grasshoppers are abundant. As the soil becomes thicker or the crevices deepen, a subterranean fauna, consisting of myriapods, earth- worms, etc., develops. As shrubs and trees encroach in patches, the animals frequenting the open tend to perpetuate themselves mainly at the open margins. Prom this condition on, so far as the fauna is con- cerned, it is largely a question of an ''opening" or a forest environment. So long as this habitat remains open, the grasshoppers, ants, spiders, butterflies, flies, and certain Hymenoptera, Hares and Bats are character- istic, and this condition tends to continue as long as the trees are scattered. The Cicada is very characteristic of the Jack Pine stage, and although it occurs elsewhere in young Birches it is not so character- istic as on these hot ridges. With the advent of the Balsam-Birch society, which is slowly encroaching upon the ridges, the forms frequenting the open will disappear, or linger in the open spots where local conditions have retarded the advance of the forest. Only a few birds were seen here, but Hares had been numerous, as was shown by the large amount of excrement, and there was similar evidence of the occurrence of the Lynx. A bat was flushed from under a stone at the edge of the escarp- ment. Station I, 6. Tamarack-Spruce Bog. This is a very small bog located ai the base of the north slope of the Jack Pine Eidge (I, 5), and roughly estimated as about 250 by 300 feet in extent. The central part is covered with sphagnum, Cassandra, and a scattered growth of Labrador Tea. Widely scattered throughout the bog occur Tamaracks and Black Spruces, small Birches, Dwarf Cranberry, Cotton Grass and alders. No standing open water was found in this area, nor was the bottom quaking. Bordering the sphagnum zone is one of alders, willows, and a tall grass which merged into a zone of Tamaracks, willows, alders, Cassandra and Balsam:, Fig. 14. Along the western end a narrow strip of water, a few inches deep, was found, which flowed through a ravine across the ridge. Along this outlet the deeper soil and moisture has permitted tl>e development of Balsam, Birch, Small-toothed Aspen, Mt. Maple, Ground Hemlock, Ground Cornel, Large-leaved Aster, and a few Black Ash trees. The fauna, like the vegetation, was not studied in detail, but the fol- lowing general relations were observed. In the open central Sphagnum- Cassaudra society were numerous large aiut nests. A Toad was ob- served here; and the following birds: Golden-crowned Kinglet, White- throated Sparrow, Cedar Waxwing, and Black-throated Green War- bler. Station I, 7. Light-house Clearing. This was a small clearing which has been made about the Light-house: it connects by a path to the fish- ing camp on Bock Harbor. It covers about half an acre, and Avas origi- nally, in all probability, a Balsam and Spruce forest like the surround- ing forest. A sod covered much of the ground, and there were numerous 16 MICHIGAN SURVEY, 1908. weeds, of which the Cow Parsnip umbels furnished excellent places for collecting Syrphid flies, Cerambycid beetles and Hymenoptera. The fauna of this clearing consisted largely of insect? which fre- quent flowers, and butterflies which fly in open places ; but a few animals were found about the Light-house itself. The Chipping Sparrow bred in this clearing. Fi(j. 1. Station II. This station included the clearing at the mouth of the stream which drained Lake Benson, and which we called Benson Brook, and followed the blazed trail to Sargent Lake, and on to McCargoe Cove. The clearing at the beginning of the trail at Rock Harbor marks the site of the former settlement called Ransom on the old maps. Station II, Suistation 1. Ransom Clearing and Benson Brook. The clearing was occupied by scattered Small-toothed Aspens and Birches, and was well sodded with grass and Red Clover. Our attention was called to this locality because of the great number of Garter Snakes (ThauinopMs sirtalis) which were found there. These snakes were very abundant in a small area east of the mouth of the brook, in a rank growth of grass and among some rails. The brook contained but little life, although it was carefully examined near its mouth and farther back where the trail crosses the brook. Only a few dead Fliysa were found, and a young fish, at the moutli of the brook. Station II, Substation 2. Tamarack Swamp. This is a long swamp which is crossed by the trail, and which contains a scattered tree growth of Tamaracks, Black Spruces and Arbor Vitae, a dense shrub growth of Cassandra and Labrador Tea, and a ground cover of Sphagnum and Pitcher Plants. While no water was seen on the surface, it was a wet swamp. This locality was only examined for birds and mammals. Station II, Substation o. Rock Ridges. This station number is given to the open rock ridges which were crossed by the trail between II, 2 and Sargent Lake. These ridges have been burned over and are largely destitute of soil and the Cladonia growth usually found on other rock ridges. Small-toothed Aspens generally border these ridges which have a northeasterly southwesterly direction. The heat during the middle of the day is excessive. The scant vegetation which grows in some crevices and depressions in the rock leaves an open area which is decided- ly favorable for grasshoppers. In some places they were exceedingly abundant and many ridges were examined almost solely for their grass- hopper fauna. In the dry soil on one ridge an anti-lion larva was found in the dust at the base of its funnel, and a .large Oarter Snake was taken on another. The grasshoppers found here were Chloealtis conspcrsd and abdoininalis. Circotettix vcrnicnlatiis. Mclanoplus alaskanns and fasciatus. Station II. Substation //. McCargoe Cove. This station simply marks the location of the end of the trail, and the cove where a few molluscs were found. There were dead shells of Anodonta grandis footiana. which were abundant at the edge of the water. Here upon the low rocky shore were also found specimens of Llmnaea stagnalis. Station II, Sutstation 5. Forltes Lake. The examination of this small lake was mainly confined to the north shore, as the south shore Geological Survey of Michigan. Annual Report for 1908. FIG. 1. THE LIGHT-HOUSE AT ROCK HARBOR, ISLE EOYALE. FIG. 2. CLIFFS BETWEEN TONKIN AND CONGLOMERATE BATS. Geological Survey of Micliigan. Annual Report for 1908. ^^•^Vf ^"J^- 5W FIG. 3. SUMMKR STORM WAVES UPON THK BIOACH (I, 1) AT THE HEAD OP TONKIN BAY, SOUTH OF THE LIGHT-HOI'SE. FIG. 4. SAND BEACH AT TUB HEAD OP CONGLOMERATE BAY (I, 1). Geological Survey of Michigan. Annuai Report for 1008. FIG, 5. BEACH POOL (I, 1) NEAE TONKIN BAY. FIG. 6. TBANS'ITION FROM THE BEACH (I, 1) TO ROCK CLEARING (I, 2), SOUTH OF THE LIGHT-IIODSB. Geological Survey of Michigan. Annual Report for 1908. FIG. 7. NATURAL ROCK OPENING (I, 2) OR AVISNUE, FARTHER UP THE SLOPE THAN IN FIG. 6. Geological Survey of Michigan. Annual Report for 1908. FIG. 8. NATURAL ROCK CLEARING OR OPENING (I, 2) NORTH OF THE LIGHT- HOUSE AT ROCK HARBOR. FIG. 9. NATURAL ROCK CLEARING (I, 2) SOUTH OF THE LIGHT-HOUSE, ADJACENT TO FIG. 7. Geological Survey of Michigan. Annual Report for 1908. FIG. 10. ARBOR VITAE BOG (I, 4) NEAR TONKIN BAY. FIG. 11. VIEW FROM THE JACK PINE RIDGE (I, 5), LOOKING TOWARD THE HEAD OF CONGLOMERATE BAY. Geological Survey of Mlclalgan. Annual Report for 1908. FIG. 12. SECOND GROWTH OF WHITE BIRCH ON THE TRAIL TO THE JACK PINE RIDGE (I, 5) CONGLOMERATE BAY. ^^^; FIG. 13. JACK PINE RIDGE (I, 5) CONGLOMERATE BAY. Geological Survey of Michigan. Annual Report for 1908. FIG. 14. SPHAGNUM-BLACK SPEUCE BOG (I, «) NEAR THE JACK PINE RIDGE. FIG. 15. SMALL ISLANDS NEAR THE HEAD OF ROCK HARBOR (III, 1). Geological Survey of Michigan. Annual Report for 1908. FIG. 16. BULRUSH ZONE AND DELTA AT THE HEAD OP ROCK HARBOR (III, 3). Geological Survey of Micliigan. Annual Report for 1908. FIG. IT. EXPOSED SECTION OF SPIT FORMED AS THE WATER LEVEL HAS LOWERED IN ROCK HARBOR, NEAR THE BEGINNING OF THE TRAIL TO SUMNER LAKE (III, 4). Geological Survey of Michigan. Annual Report for 1908. FIG. 18. SUMNER LAKE (III, .j), EASTERN END. ^ j»T#" i 1».1*J S^ te ,,, H^ j j^S » I HHj ■ 1 j^*.,^ ,. ■ 1 FIG. 19. WESTERN END OF SUMNER LAKE (III, 5). Geological Survey of Michigan. Annual Report for 1908. FIG. 20. NORTHEASTERN MARGIN OF SUMNER LAKE (III, 5). PIG. 21. SOUTHEASTERN CORNER OP SUMNEE LAKE (III, S). Geological Survey of Michigan. Annual Report for 1908. fig: 22. WESTERN END OF SUMNER LAKE (III, 5). FIG. 23. NORTHERN SHORE OF SUMNER LAKE (III, 5). Geological Survey of Michigan. Annual Report for 1908. FIG. 24. ROCK OPENING ABOUT CAMP ON SISKOWIT BAY (V, 3). FIG. 25. ROCK OPENING AT SISKOWIT CAMP (V, 3). Geological Survey of Michigan. Annual Report for 1908. FIG. 26. BOCK OPENING ON SISKOWIT BAY (V, 3). J. FIG. 27. BOEDER OF THE OPENING ABOUT THE SISKOWIT CAMP (V, 3), NEAR THE BEGINNING OF THE TRAIL TO SISKOWIT LAKE (V, 4). Geological Survey of Michigan. Annual Report for 1908. PIG. 28. ANT NEST IN THE OPENING AT THE SISKOWIT CAMP (V, 3). FIG. 29. GENERAL CHARACTER OF THE SOUTH SHORE, NEAR THE EASTERN' ENTRANCE TO SISKOWIT BAY (V, 2). Geological Survey of Michigan. Annual Report for 1908. FIG. 30. ROCK POOL ON THE BEACH (V, 2), WHERE A VARIETY OF INVERTE BEATES WAS SECURED. Geological Survey of Michigan. Annvial Report for 1008. FIG. 31. SAXIFRAGA AIZOON ON BEACH (V, 2). Geological Survey of Michigan. Annual Report for 1908. FIG. 32. GENERAL VIEW ALONG THE SHORE AT V, 2. Geological Survey of Michigan. Annual Report for 1908. FIG. 33. PARTHEE UP THE SAME SLOPE AS IN FIG. 32 AND ADJACENT TO IT. Geological Survey of Michigan. Annual Report for 1908. FIG. 34. STILL FARTHER UP THE SLOPE AND ADJACENT TO PIG. 33. ■Geological Survey of Michigan. Annual Report for 1908. FIG. 35. L00K:ING UP THE SLOPE ON THE WESTERN PORTION OF STATION V, 2. Geological Survey of Michigan. Annual Report for 1908. FIG. 36. UPPER PORTION OF WESTERN PART OF STATION V, FIG. 37. DETAIL OF WESTERN PART OF STATION V, 2. Geological Survey of Michigan. Annual Report for 1908. i^z FIG. 38. CHARACTER OF GROUND COVER IN PARTS OF THE BALSAM-SPRUCE FOREST (V, 4). Geological Survey of Micliigan, Annual Eeport for 1908. FIG. 39. OPEN SPACE IN THE BALSAM-BIRCH FOREST (V, 4). Geological Survey of Michigan. Annual Report for 1908. FIG. 40. OPEN SPACE IN THE BALSAM-BIRCH FOREST (V, 4). i J .^ '^*'^ 1 ' ^" 'i ..^.fa., ji ^ igy W^^ ^i>^|^H 1 m H |H iV ■ ^H ■ 1 Wtk mM n 1 i ■ L m FIG. 41. TAMARACK SWAMP (V, 5). Geological Survey of Michigan. Annual Report for 1908. FIG. 42. SPRUCE MARGIN OF STATION V, 5. Geological Survey of Michigan. Annual Report for 1908. FIG. 43. BLACK SPRUCE MARGIN OP STATION V, 5. FIG. 44. BOG MARGIN OF STATION V, 5. Geological Survey of Michigan. Annual Report for 1908. if- FIG. 45. LONG ISLAND GULL EGGKEEY, (V, 10). FIG. 46. POND IN TAMARACK — BLACK SPRUCE SWAMP (V, 11). Geological Survey of Micliigan. Annual Report for 1008. FIG. 47. MARGIN OP LILY POND (V, 11). FIG. 48. BLACK SPRUCE IN CASSANDRA ZONE OP STATION (V, 11). Geological Survey of Michigan. Annual Report for 1908. FIG. 49. MAPLE FOKEST ON THE DBSOE TRAIL (III, '04). Geological Survey of Michigan. Annual Report for 1908. '^ w^BM^EE^ML," LJrflifc ' ^-^F^ -^i^--* ^^BBS^S^^^S -j^-; flfllBlBI^^MHl^^^^wi^^^^^vli^S^^'^^^^^v^ ^ -i/.---r;i.*^.i:' ^^fe-iSs^^^^^^^ ^'^^S^V^tJ^S l^^^^^^^^^^^^^gS^^^^i'JWMr*' *^«' ~' f ^^^^^^^^g| ^^^^^^^^l^j Bgri»^^Syi!^^!ijB^a3^^^\'~^cSBg?^y«'^^3MPfc. ^v^ aL \^^^^Hw-' 3 jk^R M^^j^^KWEBBj^^jfeC^^Rw^ ' ' '^^^^V.s^IH ^^^^^^^^^^^\/^^'rflPA|P ^SB^I^Bt^fcr'j!^I^^^P^mS0^?j?«f .ASj^^^^^ tt^HI vt ^^^m llj Fig. 52. — Showing the Algonquin stage of the Great Lakes, dispersal of the land bioto. A water barrier to northward Fig. 53. At this time, Fig. 52, the ice sheets had retreated far enough to the northeast that the climate of the Superior basin must have been so greaitly ameliorated that animal life could have lived in its water. This inference seems probable because fossil shells have been found in the beach lines of the same lake farther to the south by Lane and Walker (Lane '00, pp. 248-252), and at Port Huron, Michigan by the writer in company with Dr. J. W. Goldthwait and Dr. A. G. Euthven (Goldthwait, '07, p. 118). Here were found an abundance of Gonio'basis livesceiis, occa- sional valves of Sphaerium striatimtm Lam. and Unionid fragments, a fauna like that of the present beaches. It is therefore not improbable that this fauna invaded the Great Lakes drainage from the Mississippi during the early stages of the great glacial lakes, when they still over- flowed into the Mississippi drainage. Latitude 48° Nurth Longitude 89° West Note: Contoxira for lOO-foot intervals are ahoum in continuous brown lines, and show where the take shore would be were the island, for instance, suddenly de- pressed 100, mi, SOO etc., feet, respectively. The Roman numerals and heavy dotted lines indicate the stations visited in the course of the biological survey. The stations established in IMUare marked wUha prime — ' The others are located in 190S. latitude 48° North Longitude 880 West To"w^irsH:iPS xobth TOL. VI PLATE r. GBOLOGICAL SURVEY OF MICHIGAN L.L. HUBBARD STATE GKOLOOIST i. ECOLOGY OF ISLE ROYALE. 57 THE ECOLOGICAL RELATIONS OP THE INVERTEBRATE FAUNA OP ISLE ROYALE, MICHIGAN. BY DR. HENEY ALLAN GLEASON. I. Introduction. The most recently emerged portions of Isle Royale are the rock and gravel beaches which together constitute virtually the entire shore of the island. Animal life is found upon them almost to the edge of the water, and well within the limits of wave action. The physiographic succession in the island is such that the areas originally occupied by beach pass through a series of changes in the physical factors, a series which is accompanied, sometimes hastened, sometimes retarded, by cor- responding vegetational successions, and which culminates in the final pr climax plant association of balsam and sjiruce forest. The detail of this physiographic and vegetational succession is by no means uniform ; it may proceed along either of two well-marked lines, depending on the immediate physical and biotic conditions, certain intermediate stages may ibe prolonged or omitted entirely, and various other deviations may occur. Nevertheless the final stage is always the same. Accompanying the changes in physiography and vegetation is a similar and dependent change in the fauna, so that there is a corresponding series of animal associations, beginning on the beaches and developing in the same di- rection, with the same deviations or omissions, to the, final or climax association in the balsam-spruce forest. The preceding general statement rests on the assumption that the areas now occupied by the climax biotic associations have developed from the, beach associations through a series of stages intermediate in time corresponding to those associations which now stand intermediate in space between the two extremes. Or briefly, as some ecologists have expressed it, the lateral distribution in space recapitulates the vertical distribution in time. Such an assumption is evidently closely akin to the recapitulation theory of the evolutionists, and just as that so-called biogenetic law has been accredited with more than its true value, so has this ecological dictum possibly much less importance than has been usually supposed. The weakness lies in too little consideration of the time element. It is certain that the higher land in Isle Royale has been submerged. This is shown by the old beach marks now many feet above the present level of the lake. Consequently by the gradual emer- gence all of the island has passed through a beach stage. But it is un- warranted to conclude from this that the faunal or floral associations of the former beach were similar to those of the present, or that in the intermediate stages the biota resembled that which now occupies the area between the ancient beach and the present shore. While it is like- wise certain that with a continued subsidence of the lake level the pre- sent beaches will eventually be left far above the water, it must not therefore be assumed that their biota will show the same successions or reach the same climax as those of the past. Changes in the tempera- 8 58 MICHIGAN SURVEY, 1908. ture or rainfall may certainly keep pace with the changes in lake level, or even be caused by it, and in either case they would exert a profound, influence on the biota. Migration of species is still taking place among both plants and animals, and may introduce new or even dominating species ani,ong the present forms. The so-called equatorial pressure of southern species is fully as s-trong now as it was directly after the close of the glacial epoch. Lastly, and most important of all, the influence of the biota itself is always to be reckoned with. Both plants and animals are continually becoming more plastic, adapting themselves to new conditions, and extending their habitats into new associations. They push forward more rapidly than the changes in physiography, some- times hastening and sometimes retarding physiographical action, and at all times greatly influencing the subsequent successions. A biotic association may develop into another by a mere re-arrange- ment of the interrelations, numerical or otherwise, of the component species, without the necessary loss of some or addition of others. But such cases are rare, and the Isle Eoyale observations show that no two associations have exactly the same species, and that with each pro- gression there has been an addition of certain forms which become the most characteristic types. The first bit of beach formed was occupied by an association possibly not unlike that of the present beaches. All the species must evidently have immigrated from beyond the island. When the soil deposits on the beach were sufficient to support a second as- sociation its species were derived partly from the beach itself and partly from new immigrants. The further development of biotic associations on the beach was then possible not only from immigrants, but also from the two associations already present. Similarly at the present time each association on the island is constantly being invaded by species from all the others, and many of them are actually able to establish themselves. This tends toward a homogeneity in the biota hardly in full accordance with the recapitulation idea. Indeed, it is very probable that independently of all physiographic agencies the whole surface of the island would eventually be occupied by the balsam-spruce forest and its attendant fa unal association. In many cases it is virtually certain that the lateral succession does faithfully repeat the vertical, and the zonation of plants around a pond may be taken as an example, but the filling of a pond is only a single step in the genetic development of the biota of an island. With this preliminary note of warning, the truth of this recapitulation theory will be assumed for the island, and the discussion of the insect and molluscan fauna will follow the genetic lines indicated in the fii*st paragraph. The relationship of the various physiographic types on the island to each other may conveniently be expressed by a diagram, (see end of paper), indicating the direction of the devolpment by arrows. It must be reiuembered that pra.ctically any one of the intermediate stages may be omitted. II. The Lake. The lake (Superior) must obviously be regarded as the first stage in the genetic development of the faunal associations. Broadly speaking, the lake fauna is divisible into two main groups. The first is pelagic in ECOLOGY OF ISLE ROY ALE. 59 character and includes those species wliose distribution is entirely inde- pendent of the shore, for example, most of the species of fish. The second group is littoral; the species occur along the shore in compara- tively shallow water, and are to a greater or less extent dependent upon the land in its relation to the character and slope of the bottom and to the motion of the water. Members of the latter group only are con- sidered here. The two dynamic factors just mentioned are the most important ones that influence the biot.a of the lake. There are no currents of sufficient rapidity to affect the animal life. The direction of the wind, whether oflf-shore or on-shore, may respectively lower or raise the level a few centimeters, especially when the wind blows lengthwise of the long narrow inlets, such as Conglomerate Bay {Fig. 11). Some fixed or slow-moving species may accordingly be alternately submerged and ex- posed, while motile forms can at once adjust themselves to any change of level. Of far greater importance is the motion of the water caused by wave action. It is only on rare occasions that the lake is quiet. Gentle waves come in nearly all the time, and after storms become of great violence. Wave action is of itself suflBcient to inhibit the growth of shells along the exposed shores, where they might easily be torn loose and crushed against the rocks. Such forms are consequently re- stricted to the shores of the smaller bays or to the lee side of islands. Wave action is of importance further in determining the character of the bottom. Where the shore is exposed directly to the lake it is usually of massive rock, all the fragments having been washed down to deep water. In small shallow coves, where the waves break always in one direction there is usually a sloping beach of gravel extending across the end perpendicular to the direction of the waves. Every breaker sorts over this gravel so that it is nearly impossible for a fauna to develop. In larger coves or bays, where the violence of the wave action is reduced by distance, the gravel is finer or even a beach of sand may rarely be formed. Along the steep or clifE-like sides of these coves the bottom is frequently covered with angular rock fragments too large to be moved by the water. These are frequently inhabited by shells. In general the development of a free littoral fauna demands quiet water where the animals will not be dashed on the rocks or stranded on the shore, and for attached ispecies there is required either quiet water or a firm bottom which will not be dislodged by the waves. A more de- tailed discussion of this as afl'ecting the distribution of shells will be given later. In the larger inland lakes, of which Siskowit Lake, the only one of the class studied, may be taken as an example, essentially the- same con- ditions obtain as on Lake Superior itself. The difference in temperature and content of the water seems to be of minor importance. The waves in the larger lake can naturally reach a larger size, and their influence is felt far into the bays. Thus at the head of Rock Harbor, about six kilometers from the lake proper, the distribution of shells and the al- most total absence of free forms indicate that even there wave action is of importance. In Siskowit Lake, although larger than Rock Harbor the force of the waves is so reduced by every headland or island that on the quiet water in their shelter a rich fauna of such free forms as 60 MICHIGAN SURVEY, 1908. water-striders and whirligig beetles is found on the surface, while numerous mussel shells live on the silt or sand bottom. In Sumner Lake and others of limited area the motion of the water has no measurable effect on the biota, and they will therefore be treated under a separate heading. The distribution of shells along the shore, particularly species of Limnaea and Physa, is of especial interest. Having relatively low motility they are correspondingly limited in their distribution and the factors governing it are more readily determined. These will perhaps be made clearer by concrete illustrations. Tonkin Bay is a small inlet about half a kilometer long, opening to the east upon the lake, and with steep, approximately parallel sides. It is narrowed half way up by two beaches lying perpendicular to its length. By this the wave action on the upper part is reduced, but still may sometimes be sufficient to wash heavy drift-wood upon the beach. In the outer half the wave action is but slightly less than on the lake itself, and no shells are found. In the inner or upper half Limnaea stagnalis L. (Nos. 50, 54, 57), Limnaea emarginata Say (Nos. 50, 57), and Physa sayii (Tap.) (Nos. 50, 57), live along both sides where the bottom is rock, but not across the ends. They live only on a rock substratum, which may be either horizontal or vertical, and in water up to 45 cm. in depth. The larger species, Limnaea stagnalis, is more abundant in the deeper water, and only the smaller species live at a depth less than 15 cm. They then prefer the vertical walls to the hori- zontal or flat bottom. Conglomerate Bay is a rocky inlet {Fig. 11) similar to the one jvist described and about 1.6 km. long. Being wider at its mouth than Tonkin Bay the force of the wave action is felt farther up the bay. Near the end the waves have little effect, as is evidenced by a sandy beach {Fig. Jf), almost without driftwood. At the upper end of this bay along the north side Limnaea emarginata Say (Nos. 118, 125) andP/iysft sayii Tap. (Nos. 118, 125) are found in water 1.5 — 45 cm. deep, in the deeper water on the tops of flat rocks, in the shallower water, also on the verti- cal sides and in small crevices. They never occur on the sand or gravel deposited around the rocks, as is frequently the case near the sand beach at the upper end of the bay. The distance to which they extend from shore is greatest opposite the concavities of the shore line and least opposite the small rocky headlands. Their distribution in both Tonkin Bay and Conglomerate Bay seems to be regulated mostly by the wave action, since they seek the most protected places, avoid the shallow water where the wavies would strike them most, and do not live, on loose or small rocks, gravel, or sand, which would easily be dislodged. The fact that the smaller shells are found at the least depth, while the larger Limnaea emarginata inhabits the deeper water, would indicate that the small size of the former renders them less easily dislodged by the waves. Again their greater abundance on the ,north side suggests the possibility of a light relation. ^ Siskowit Lake, with its rocky shores and large area, offers essentially the same condition as Lake Superior itself, and the shells have the same general distribution. Along the very gently sloping rocky shore near the outlet Limnaea stagnalis occurs in abundance, always at a depth ECOLOGY OF ISLE ROYALE. 61 of 10-40 cm. Along the south side of a large island near the south shore, where they are sheltered from waves in every direction, the same spe- cies is abundamt. They live on rocks in the full sun in water 10 — 45 cnn deep, Avith the optimum depth at 20 — 25 cm. They may occur on the tops or sides of rocks, but never on the sand between them. Associated with the Limnaeu, but much less abundant, are Planorhis hicarinatus royalensis Walker (No. 210), P. campanulatus Say (Nos. 210, 211), Lampsilis luteolus (Lam.) (Nos. 210, 211), Anodonta nmrginata Say (No. 210), and Anodonta grandis footiana Lea (Nos. 210, 211). Opportunity was given to observe the behavior of Limnaea stagnalis (No. 217) in waves of some size near a small circular island half a kilo- meter out in the lake. The bottom was gently sloping, and either of solid rock or of large rounded fragments. There were no overhanging trees, so the shells were found in uniform abundance in the usual depth of water on all sides of the island. At the time the island was visited a strong wind was blowing, and the waves were probably nearly as high as they ever become on Siskowit Lake. One or two shells were seen which had been washed loose, and of course would be unable to re- attach themselves until the waves abated. It would be expected that in such cases the shells might be crushed or broken or the animal killed. That such may happen was evidenced by finding a few live shells which had been cracked and then healed, leaving an irregular surface. Their occurrence here and elsewhere only upon rocks of considerable size shows that they require a firm substratum, and where the rocks are free from any coating of slime they can certainly endure higher waves.' Around the island under discussion the rocks were washed perfectly clean. The beach in front of the camp at Siskowit Bay {Fig. 29) was inhabited by large numbers of (No. 200) Physa sayii Tapp., PJiysa sp., Limnaea stagnalis L., and Limnaea emarginata Say, so that more detailed obser- vations of them could be made, and a few experiments carried out to show their sensitiveness to the depth, or bathytropism, as it has been termed. The beach here is of rock with a gentle slope of about one in five, corresponding to the dip, except where blocks have worn off, leaving low vertical walls. The wave action here is very light, its force being cut off by a series of islands lying between the beach and the main body of Siskowit Bay. This was well shown by the conditions on August 2, when there was scarcely a ripple inside the islands, although the bay outside was covered with whitecaps. The beach is covered with a thin coat of slime formed mostly of excrement from the snails. On such a beach snails may live close to the edge of the water, but the larger Limnaeas still occupy their usual depth of 1.5 to 4.5 deci- meters. About 10 A. M., on August 3, all the shells to a depth of about 1 decimeter were gathered from a strip of the beach about 10 meters long. They were comprised in the following species: (No. 200) Limnaea stagnalis L., Limnaea emarginata Say, Physa sayii Tapp., and Physa sp. The smaller Physas were especially abundant and about 200 of them were taken. Four hours later, at 2 P. M., 60 shells, all of the smaller spe- cies, had migrated upon the same strip. The only evidence concerning the way that they came is that one shell of Limnaea stagnalis was seen to drift up over a low wall into the shallow zone. This method could 62 MICHIGAN SURVEY. 1908. hardly account for 60 of the smaller ones, however, appearing in so short a space of time. It may be taken as indicating a general and continued migration in all directions within their bathytropic limits. It was noticeable that the large Limnaea emarginata and Limnaea stagnalis, aside from the one specimen mentioned above, live at an aver- age depth of 3 dm. and never deeper than 4.5 or 5 dm. To test their bathy- tropism six of them were picked out of the deeper water by hand and held in contact with the bottom in the shallow zone until they extended their feet and attached themselves. At this time the water was very quiet, moving just enough to cause a faint sound on the beach. But the size of the shell of the two Limnacas is so large that they offer consider able surface to the water and are consequently easily washed loose. Two of the six swung a little from side to side and were then washed off and carried by the undertow into water 3 dm. deep, where they again at- tached themselves. A third, without being shaken by the waves, clun."^ to the rock for some time, then suddenly let go its hold and drifted over a low ledge into the deeper water. Two others immediately started to crawl down the slope, and one in about fifteen minutes, the other in about half an hour, had crawled over the ledge into water 3 dm. deep, where they both remained stationary. The sixth remained attached, and in three hours had crawled 2 dm. parallel to the shore, keeping at the same depth. The next morning, twelve hours later, it had disappeared, and of course could not be recognized in the deeper water. On August 4 two shells of Limnaea appeared in the shallow zone, but it is not known whether they drifted or crawled up. They were there at least three hours. After they were last observed a fresh breeze sprang up from the east and the slight wave action caused by it prob- ably washed them down. The level of the lake varies somewhat with the direction and inten- sity of the wind, so that in front of the camp a strip of beach up to 5 dm. in width may or may not be covered with water. The smaller shells, Physa sayii Tapp. and Physa sp., live in this zone in spite of the fact that they are sometimes out of water. So far as observed they are never exposed for any considerable length of time, so that. they do not become dry. Then again the weathering of the rock has left bowl-shaped hollows a centimeter or so across and about the same depth, and the snails usually get into them. To summarize, the known facts bearing on the distribution of these four species are as follows: 1. Their lower limit is 4.5 to 5 dm. depth of water, governed possibly by the water-pressure or the food supply. 2. The upper limit is for Limnaea stagnalis and Limnaea emarginata 1.5 dm. of water, for Physa sayii and Physa sp. the shore-line. The cleaner the rock and the less the wave action the shallower the water which they may inhabit. 3. Their horizontal distribution is controlled by (a) full exposure to the sun; (b) a rock bottom; (c) a certain minimum of wave action. But two species of insects were collected which should properly be considered here, caddice flies and stone flies. The larva cases of the caddice flies were collected only in the outlet of a small stream emptying into Rock Harbor, in, 1 — 1.5 m. of water (No. 163 or ECOLOGY OF ISLE ROYALE. 63 164), but the imagos were common all along the shore of the lake, es- pecially on the gravel beaches. One (No. 192) was taken on the boat about 2 km. off the south shore of the island. Stone flies were also fre- quently collected along the beaches, where they came up to breed. They were most numerous, however, on steep or even vertical cliffs with south- ern exposure (Nos. 24, 80). Near the entrance to Conglomerate Bay (Fig. 2) they were seen collected in such a place by thousands. The water there was at least 4 m. deep. A few hair-worms, Gordms aquaticus (L.) (No. 207), were collected in 2 to 3 dm. of water on the rock beach {Fig. SO) in front of the camp on Siskowit Bay. The various mussels collected in Siskowit Lake and elsewhere, even though sometimes associated with Limnaea stagnalis, belong rather to the associations of the smaller inland lakes. III. The Inland LaTce. The smaller lakes are mainly surrounded by tamarack swamps, with the vegetation showing the characteristic zones, certain ones of which, as the rushes, water-lilies and pond weeds, live in the lake itself. The bot- tom is covered with peaty mud or with slime, and the wave action is never severe enough to interfere with the growth of either fauna or flora. In many of the smaller lakes, in fact, the water lily zone is so wide and the open water so restricted that there is practically no wave action at all {Fig. Jf6). Accordingly both fauna and flora are richly developed both in species and individuals. The fauna may be roughly classified into several groups according to their habitat in order to facilitate description. The interrelations of the different spec- cies are complex in the extreme, and of course could not be properly worked out in such a short time as the lakes were under observation. a. The Fauna of the Bottom. In Sumner Lake (III, 5) {Figs. 18-22' and in sheltered places in Siskowit Lake several species of shells live on ihe bottom in sand or mud and at a depth of from 3 dm. to 1 or 2 m. Planorhis trivolvis Say (No. 135) lives in the shallower water, prefer- ably in mud. It is nowhere abundant, but was collected in both lakes. One specimen only was found in Sumner Lake in a little pool with mud l)ottom. Shells were commoner on the shoreward side of an island in Siskowit Lake, on a bottom composed of sand and mud. They were well buried under the sand and the majority of the shells were dead. Mussel shells, especially Anodonta marginata Say and Anodonta grandis Lea, were common in all the smaller lakes and at the upper end of Rock Harbor. They were most abundant in the deeper water with a sand bottom, particularly where there was comparatively little vegetation. In certain sheltered bays at the upper end of Sisko- wit Lake they were especially numerous. Muskrats carry them to the shore to eat, and leave the empty shells in heaps, which were conspicu- ous sights along most of the lake shores. In Sumner Lake live shells were very scarce, but the piles of dead ones on the bank testified to their former abundance. At the upper end of Eock Harbor some small shells, PlanorMs hicari- natus Say (89), PlanorMs exacutus say (89), Plwnoi'his parvus Say {89, 163, 164), Yalvata tricarinata Say (89, 163), Yalvata sincera ny- 64 MICHIGAN SURVEY, 1908. landeri Ball (80, 163, 164), Amnicola lustrica Pils, (89, 163, 164), and Pisidium sp. (163, 164), and Amphipods were dredged from a depth of 1.5 to 2 meters near the mouth of a small stream {Fig. 22) where the bottom was thickly covered with small twigs and other coarse vegetable debris. From the same place the caddice fly larvae were obtained, as mentioned previously. The same fauna was collected in the stream it- self, but only near the mouth, where the water was deep, the cvirrent slow, and the conditions in general much like those of a lake. May flies probably breed in similar places. No larvae were seen, but a few images were collected (No. 178). The fauna of the bottom shows a connection through the presence of Pisidium sp. in the last case with that of the small streams in the tamarack swamps and with that of the brooks, like the outlet of Siskowit Lake. The accumulation of vegetable debris and the more restricted amount of water are both approaches toward the conditions in the former places. In Siskowit Lake, where Planorhis campanulatus, Phmorhis hicarinatus ropalensis and Anodonta grandis footiana were associated with Limnaea stagnalis, another transition was shown between the faunas of the inland lakes and the larger lakes as typified by Lake Superior itself. b. The Free Fauna of the Water. No species were observed except fishes and leeches. The latter were abundant in Sumner Lake, especially among the water lilies and in the shallow water along the shore. c. The Fauna of the Surface. Hardly belonging properly to this groiip were the small shells, Limnaea catascopium Say (220), Physa sp. (220, 221), Valvata sincera nylanderi Ball (220), and Amnicola Unwsa Say (220), found abundantly on the under side of water lily leaves. Their distribution is directly controlled by that of the water lilies, that is, near the shore, and in the larger lakes only in the sheltered bays. Pro- bably a third of the leaves had one or sometimes two shells attached to them. Water stridors, Gerris remigis Say (No. 96), were abundant, usually near shore in the water lily zone, but occasionally out in the open water. Whirligig beetles, Gyrimis minuttis Fabr. (No. 219) were also common, but not abundant on the smaller lakes. In the sheltered bays of Siskowit Lake they collected in immense swarms, keeping mostly near the shore among the water lilies and under overhanging brush. Donacia proxima (Nos. 171, 184) and Donacia cincticornis (Nos. 171, 175) were abundant on Sumner Lake, resting on the water lily leaves. When alarmed they would fly a short distance close to the water, making a little trail behind them, and alight on another leaf. d. The Free Aerial Fauna. Bragonflies of several species are abun- dant along all of the lakes. They usually keep close inshore or over the water lilies, and fly regularly in patrols around the lake, searching all the time for insects but keeping up a uniform rate of speed. Aeschna sp. was probably the most abundant, and associated with it were Enallagma hageni Walsh and L&ucorhinia proxima Anth. The butterfly Argynnis atlantis Edw. also occurs (No. 169). The inland lakes may be regarded as small detached portions of the main lake, cut off from it by the lowering of the level of the latter. Since they are composed of stagnant water with little or no wave action, where organic material may accumulate in quantity, they support a ECOLOGY OF ISLE ROYALE. 65 different fauna and their genetic development is along a different line, culminating however in the climax type or balsam-spruce forest. The only intervening stage is the tamarack swamp. IV. The Tamarack and Arhor Vitae Swamps. Nearly every inland lake in the Isle Royale region is wholly or partly surrounded by tamarack swamps, (Figs. 14, 19, 22, 41, 41, 48). It is not necessary to discuss the general structure of the vegetation, since that is described elsewhere in this report, but it may be indi- cated here that the ground cover is a spongy mass of sphagnum covered with a dense growth of ericaceous shrubs, such as Cassandra and Ledum, and that the trees are almost entirely tamarack and black spruce. The forest cover is open enough to allow ample illumination. Tamarack swamps may be found of all ages, from those developing at the edge of a lake to those which have completely covered the lake and are now dying as an association. Their surface is generally level, the older parts being successively somewhat higher as they are built up by the accumu- lations of peat. When the level is nearly that of the lake the beds of sphagnum are interspersed by little streams or pools of water, some of them being merely extensions of the lake itself, or some of them serving as inlet or outlet. The smaller ones have no bottom except the sphagnum itself, while the larger have a loose incoherent bottom of slime. In the larger of these streams are found small bivalve shells, Pisidium sp., embedded in the slime at the bottom (No. 230; V-5), and other material; and the beetles Haliplus rtificolUs DeG., Hydroporus tristis Payk, and Agahus congener Payk. (No. 230, V-5). In the smaller ones, which are fre- quently only a decimeter or two wide and half as deep, there is no difference in the vegetation except for a little Utricularia in the bottom. Animal life is there very scarce (No. 237, V-5), but included Pisidium sp. As the swamps become older the water is limited to small shallow pools, seldom more than one decimeter deep or three or four decimeters wide. Their bottoms are covered with dead leaves and sphagnum, and they are usually densely shaded by the forest growth above. In them are found small bivalves, Pisidium affine Sterki (11 A, 79A), P. suhrotundumi Sterki (116, 181, 182, 237), P. suhrotundnm Prime (116, 237), and water beetles, Haliplus ruficollis Deg. (No. 116, 1-4) and Scutopterus hornii Cr. (No. 181, 144). The latter is restricted, so far as observed, to this single habitat in the pools in tamarack and arbor vitae swamps. Dragon- flies are the principal aerial insects, but are not abundant. A fly (No. 240, V-5) was taken on the flowers of Solidaga neglecta. In still drier swamps, where there is no longer any standing water, (Fig. 14). ants are a characteristic feature of the fauna. They build huge dome-shaped nests, 4 to 7 dm. high, composed within of sphagnum and other vegetable debris, and smoothly covered on the outside with leaves of Cassandra, doubtless to prevent drying. Formica adamsii Wheeler (No. 115, 1-6) seems to be the only species concerned, and a nest from which the collection was made was photographed. No. 114, taken at the same time from a similar nest, has been identified 9 66 MICHIGAN SURVEY, 1908. as Formica dryas Wheeler, suggesting a possible confusion of the num- bers. No other insects were observed except the omnipresent black-flies and mosquitoes. At the head of the numerous fjord-like inlets along the shore there is usually a swamp tract extending for some distance inland in the same direction as the inlet itself. The level is but little above the lake itself, but there is no permanent standing water or lakes as in the tamarack swamps. The standing water is limited to small scattered pools, seldom more than a meter across, and the forest cover is pre- vailingly of arbor vitae. The shade is exceedingly dense, and the ground " is covered with tangles of underbrush and fallen logs. The fauna is accordingly reduced to a minimum, and the few forms col- lected were all dredged from the leaf-covered bottoms of the small pools, and included bivalve shells, PyramidiiU striatella (Anth.), and Pisidkim stihrotvrulum Sterk. (No. 182), and water beetles, Scutopterus homii Or. (Nq. 182). The latter were very scarce. Faunistically the arbor vitae swamp is very closely related to the later stages of the tamarack swamp, as a comparison of the species will show. At the ends and around the sides the swamp grades imper- ceptibly into the balsam-spruce forest. In connection with the swamps must be mentioned the fauna of the small rapidly flowing streams leading out of the inland lakes. The bottom is usually rock or gravel, and the swift current prevents the accumulation of organic debris. In Benson Brook on the north side of Rock Harbor in still, deeply shaded places were dredged up (No. 149) Pallifera dorsalis (Binn.), Pyrarnidula altern-ata (Say), Pyramididn striatella (Anth.), Zonitoides exiginis (Stimp.) and Physa sp. In the outlet from Siskowit Lake, in small pools 5-15 cm. deep with a bottom of slime covered with loose pebbles, were collected several shells (No. 238), Physa sp., PisMinm mediamim Sterki, P- siibrotundum Sterki, and Miisculin.m securis (Prime). The current where these were col- lected was very slow. In the more swiftly flowing water nothing could be found. Owing to the peculiar geological structure of the island the swamps have a generally oblong form with approximately parallel sides. Along the sides the swamps grade imperceptibly into the balsam-spruce forest {Fig. 43), and on the ends as well, though there the transition is more gradual and the facies are usually separated by an intermediate zone marked by dense thickets of alder. V. The Gravel and Sand Beaches. The gravel beaches are found in but certain places along the shore {Fig. 1), where the slope of the banks and the action of the waves permit the formation of the , gravel deposits. Optimal conditions are found at the heads of the numerous inlets or coves ; such as Conglomerate Bay {Fig. 4), and Tonkin Bay, already described, and many other similar places. They also occur, however, along the shore of the lake itself, where the wave action is at its minimum. Their distribution appears to be controlled principally by the slope of the bottom, since the gravel could not be piled up on slopes of too steep pitch, and they are almost ECOLOGY OP ISLE ROYALE. 67 mvariably in locations so bounded by rocks or shore that the waves strike them always in one direction. An instance of this was seen near the light-house. A small inlet about 5 m. in length and width opened towards an island. Waves struck it in two directions, both ■diagonally, but rebounding from the rocks continued into the inlet in one direction. At its back was a small but typical beach, the only one in the immediate vicinity and likewise the only spot where the waves always came in the same direction. As a consequence of this directive action the beaches always lie at right angles to the direction of wave action. The gravel of which they are composed varies in size from fragments as large as one's flst to mere sand, but the biota of the sand beaches is so diffei-ent that it requires separate discussion. There is no vege- tation, but the beaches are frequently strewn with dry drift wood in which several kinds of fruits, dead insects and shells may be found. The gravel is dry on top, but is always moist at a depth of one or two decimeters or even less. The broader beaches have full exposure to ithe sun, but the narrower are shaded, and all are bounded at the rear by a narrow but dense zone of alder. The fauna of these beaches is limited in species, probably owing to ■the lack of food, although the number of individuals is relatively large. Caddice flies are rather common running about over the finer gravel just above the reach of the waves, or sometimes taking short flights (No. 10). Stoneflies are associated with them; they crawl about actively over the wet gravel near the water's edge and do not attempt to fly. They are frequently struck by waves which merely wash them a little farther up the bank. A few species of ants are also common, running over and through the gravel (No. 38). They prey on dead caddice flies or even on live ones when they succeed in capturing them. The most characteristic group, however, consists of several species of spiders, which fire found in great abundance on the coarser gravel in the sun (Nos. 16, 25, 38, 39, 60), Lyco^a pratensis Emer., Pardosa lapidicina Emer., Pardosa groenlandioa Thor., Eho latkhorax Keys. They run Avith great rapidity and at the least alarm crawl under the rocks, where it is almost impossible to find them. After the first alarm they usually show themselves in 10 to 15 seconds, but being frightened again, they crawl for some distance under the gravel and are lost permanently. Many of them carry egg cases, and if forced to drop them they spin a web which they follow back in a short time. These spiders are very numerous, probably 10 or 12 to every square metre over all "the gravel beaches. Other insects observed were, a small beetle (38) crawling over the ■sandiest part of the beach; two species of small beetles (89) crawling through the coarse sand and fine gravel at the water's edge; a click "beetle, Gorynibites medianus Germ. (41) crawling over sand in a shaded place near a rock cliff; a Scarabaeid, Serica vespertina Gill. (43) ; a heetle, Macropogon rufipes Horn (60). Some fish worms (40) were also found buried 3 dm. deep in moist coarse sand under the gravel "beach in front of the light-house. They were above the level of the -ground water. Butterflies and wasps, which were so abundant on the 68 MICHIGAN SURVEY, 1908. sand beaches, were collected but once. The butterfly, Pyrameis cardm Linn. (39) flew out of the woods, rested a moment on the gravel, and then visited a dogwood flower. The single wasp (41), AmmopMla sp., was seen flying low over a small area of sand near the water's edge on a gravelly beach. Some fossil beaches were observed, rising several meters above the lake. The gravel was then thinly covered with lichens, and in some cases even supported a scanty growth of flowering plants. A beetle, (37) Leptura chrysocoma Kby., was collected on a ro^e in such a place. The contents of the drift washed up on the beaches is of some inter- est as indicating a possible way in which new forms might reach the island. Here were found Lvmnaea stagnalis (19) ; a dead butterfly, Avosia plexippiis Linn., (19) ; some dead ladybugs, Anatis IS-punctata Oliv. (21) ; shells (21) ; butterflies (21) ; one snail shell, Polygyra albo- Idbris (Say), badly broken but still containing part of the body (39). The vegetable drift (18, 21) included cones or fruits of jack pi'ne, balsam, arbor vitae, and alder. Sand beaches are formed in the same way and under the same condi- tions as the gravel beaches already mentioned, but only where the wave action is much reduced by distance from the lake. The principal ecologi- cal difference between the two lies in the presence of the sand, affording a fairly uniform surface, and a finer substratum in which various species may live protected from predaceous ants and spiders. The principal beach studied was at the head of Conglomerate Bay, {Fig. 4), and may be described in some detail. The beach was more than 100 meters long, and divided at the middle by a small stream running through it into the bay. One portion was only 2-6 m. wide, and over- hung by alders. There the sand was always moist, and the fauna very scanty. The other portion was 10-20 m. wide, fully exposed to the sun, and sloping very gently back to the usual zone of alders. There was some drift wood scattered about over it. A warm sunny open place like this attracts many casual visitors from the neighboring woods. Three species of butterflies were especially char- acteristic. Papilio turnus (No. 29) was the most abundant. Tbey flew back and forth along the beach at a general height of 2-3 meters, occa- sionally flying out over the water and dipping into it now and then. They very seldom alighted on the sand. The red butterflies (No. 29) hovered low over the sand but when they alighted chose grass or low shrubs along the margin. No. 29 includes Pyrameis hunteri Fabr., Pyrameis carckii Linn, and Basilarchia wrthemis Dru. The black butterflies were not common (No. 29). They flew rapidly and irregularly over the sand and the edge of the water at a height of 1-3 m. and very rarely alighted. Two other casual visitors were ob- served but not caught; a red winged grasshopper which flew over the sand at a height of 2 m., and dragonflies which hovered over the small stream. Both came from, and returned to, the woods. Peculiar to the beach were small blue butterflies, Phyciodes tharos Dru. (No. 29), and two or three species of sand-wasps (No. 31), includ,- ing Diodontus n. sp., AmmopMla sp., and Xanthosarus latimaniiK Say, which flew rapidly over the surface at a height of about 1 dm. but very rarely alighted. When dead they were preyed upon by ants. One ECOLOGY OF ISLE ROYALE. 69 or two species of flies (No. 31) (Cynomyia caclavcrlna Desv.) were also ■common. Crawling over the sand wei'e ants (No. 30), spiders with eggs cases, Pardosa groenlandica Thor. (No. 30), and beetles, Bemhiditmi cannula •Chaud. (No. 30). The latter were very numerous, and included two species. They ran rapidly and irregularly over the sand, and especially the fine gravel just back of the wet margin. When alarmed they try to hide under small pebbles, or sometimes fly a short distance. A dead shell of Limnaea stagnalis (No. 32) was found on the beach, and a dead Polygyra alholabris in the small stream (32). VI. The Rode Beach. Where the slope of the shore is steep or the action of the waves severe, gravel or sand cannot accumulate, and the bare rock is left exposed. The ecological conditions affecting animal life here are so different from those of the gravel beaches that they require especial mention. Rising directly from the water they are naturally exposed to the full force of the waves, .{Fig. 3), which dash upon them to a considerable beight, washing away all loose particles and effectually preventing even the most meager formation of soil. Beyond the reach of the waves, rains and drainage water act with greater or less effect in the same way. The vegetation is therefore limited to various species of crustaceous or foliaceous lichens, which are true lithophytes. Even they are absent from the lower portions where the wave action is more continued^ and •especially where the ice may scrape them off. Higher up the procumbent juniper and Cladonia appear and the whole eventually merges into the Gladoma clearing to be described next. Some idea of the zonal succes- sion of the different jjlants may be gained from the following table, showing the heights of the different zones on a rock beach near the Hock Harbor light-house, Figs. 6 and 7. Zone. Height— feet. Total height. Crustaceous lichens .... 7 ft. 7 in. 4 ft. 9 in. 4 ft. in. 6 ft. 7 in. 4 ft. 1 in. 7 ft. 7 in Foliaceous lichens ... 12 ft. 4 in. 16 ft. 4 in 22 ft. 11 in. Forest 27 ft 6 in The first two zones, to the height of twelve feet above the lake, are included here in the rock beach. Naturally these levels may vary with different localities, being lower in more sheltered places. Over the lower portion of the beach the fauna is practically without shelter or protection, and in the zone of foliaceous lichens shelter is afforded only to very minute species. There are sometimes small fis- sures in the rock, but only two species were observed to enter them. During all or part of the day the beaches are exposed to the direct rays of the sun, and the rock consequently reaches a temperature ■ far above that ever reached by the air. 70 MICHIGAN SURVEY, 1908. The temperatures observed on July 11 may be given as an example. Time. Air at 4 ft. in sun. Roclc surface. Rocli. 51 ° Fhr. 58 " " 56 " " 56° " 68 (sun) 61 (shade) 88 95 100 2 p. m 94 (sun) 90 The absence of plant growth also tends to limit the number and char- acter of species to predatory forms, and the number of individuals is small. A small rock beach jutted into the lake near the light-house, and was at most but one meter high. Although sheltered from the waves by an island, it was still completely flooded by even moderate waves. Most of the surface was accordingly without vegetation, but besides the crustaceous lichens there was one species of moss, a few plants of harebell, and several tufts of grass. Five species of insects were found on this beach, four of which were merely casual visitors. Some spiders (No. 46), Pardosa groenlandica Thor., wandered upon the rock from the neighboring gravel beach, but finding no rocks to hide under they soon left. Ants (No. 46), Formica dry as Wheeler, were rather common, but it was easy to see that they came from, and returned to, the gravel beach. The only food they obtained appeared to be the remains of dead caddice flies. A species of fly, Hydrophorus philomirius Wheeler (No. 46), was very common on those parts of the rock which were constantly wet by the waves. They were seldom seen over the dry. portions, but re- mained resting on the wet rocks. This fly was of common occurrence in the uplands and will be mentioned also under other headings. A few stoneflies (No. 46) were found on the wet rocks where the waves ' struck. The only species confined to the beach was one species of beetle, Bembidiv/in grapei, which ran over the surface, hiding from time to time in tufts of moss. On a smaller rock beach exposed to the full force of the waves wero' collected a spider (No. 47) and an ant, Formwa dryas Wheeler (No. 47) ; a butterfly (No. il) , Basilarchia arthemis Dru., was also taken, while hovering over the beach. On a larger beach near by, the elevations of which were given in a preceding paragraph, the faufta was better developed. A jumping spider was fairly abundant, and was a fine example of protective coloration,, being almost invisible against the gi'ay rock background. Another spider (No. 48) and red mites (No. 48) hid under the foliaceous lichens. A small beetle (No. 48) was abundant, running rapidly over the rock, never- attempting to fly, but hiding in the crevices. A brightly colored red and black beetle was common. It ran rather slowly but flew easily. No ants were seen. Besides the forms just mentioned, which may be con- sidered normal members of the rock beach association, there was collected a caddice fly (No. 48) and a running spider (No. 48), undoubtedly a straggler from the Cladonia zone above. ECOLOGY OF ISLE ROYALE, 71 At otlaer times were collected on rock beaches ants, (No. 15) Gam- ponotiis herculeanus L., carrying away dead caddice flies, and as acci- dental visitors a Ciinhex americann- I^ach (No. 106). a butterfly (No. 107), Basilarchia artheinis Dru., and a running spider (No. 103), Lycosa pratensis Emer. In connection with the rock beaches may be mentioned the beach pools {Fig. 5), which are depressions in the rock filled with water by high waves. They are naturally most abundant on flat or gently sloping beaches, and their permanency varies with their size and depth, affecting evaporation, and with their height above the lake, affecting the frequency with which they are filled. In those which are permanent are found shells, Limnaea emarginata Say (No. 58), and Plan-orhis parvus Say (No. 59), and a few insects, Rhantus Mnotatus Harr. and Gorixa sp. (73, 74, 75). The water beetles and water boatman are strongly stereo- tropic, staying on the bottom or in, crevices, and leaving it only to dart quickly to the surface for air. YII. The Cladonia Clearing and Jack Pine Ridges. The elevated position of the rock ridges and their physiographic rela- tion to the uplands are the two chief factors determining the succession of biota upon them. In response to the rapidity of drainage, and the slowness of soil formation the first plant life to invade the rock beaches is a lichen association composed to a large extent of Cladonia rangiferina, which carpets the rock to a thickness of 1 to 3 dm. With it are associated various xerophilous shrubs and herbs, but no trees. Consequently the insolation is strong, and after rains that water not removed by surface drainage is soon evaporated. The soil consists only of those thin de- posits formed by the disintegration of the underlying rock and the decay of the vegetation, and is held in place by the tufts of lichens. Such natural clearings in the forest are frequent near the lake {Figs. 6, 7, 9), either on gentle slopes but little above the lake and consequently of late origin, or upon the devated rock ridges {Figs. 8, 25, 26), where they are of much greater age. Their shape and size varies naturally with the topography. In these Cladonia clearings has been developed a very characteristic faunal .association, rich in species and in individuals, and especially dis- tinct in the number and variety of insects. The fauna may be con- veniently divided for discussion into three groups, aerial, terrestrial, and subterranean. Since the latter is the most nearly fixed in habit, it may be described first. 1. Subterranean Fauna. In the shallow depressions and crevices of the roc\i}{Figs.7,25,26),Sire thin soil deposits supporting a dense growth of various plants, especially the Cladonia lichens, the bearberry, and dwarf juniper. Ants are frequent, running over the surface and ex- cavating below it, but they make their nests only in the deeper crevices or under the densest growth of plants Avhere the depth of soil is suffi- cient to allow them to make their excavations and to conserve the mois- ture supply. In the crevices they are usvially 1 dm. or more below the surface. Camponatus herculeanus L. (22), Myrmica rubra L. (61), and Leptothoraoo canadensis Prov. (63) are the species generally represented. The nests are more frequent near the margin of the rock clearings, 72 MICHIGAN SURVEY, 1908. where the soil is better shaded. A nest of, Formica sanguinea Latr. (No. 72) was placed under a decaying limb, and the soil beneath it was largely composed of minute fragments of rotten wood. This ant has two sorts of pupa cases. Another colony, Leptothoraw caruidensis Prov. (No. 77), was also collected in Cladonia clearings. The largest species of ant (No. 62), Camponotus herculearms L., is found always singly, and no nests were ever observed. Spiders also occur in the looser soil deposits, but most of them prob- ably belong to the surface, such as (No. 71) Lycosa Icochii Keys, which had an egg case attached, although buried under two cm. of soil. The largest spider, (No. 67) Coelotes sp. of which only one specimen was observed, is apparently entirely subterranean. It spins a pocket just about large enough for its own body, and when uncovered does not attempt to run, but buries itself in the soil or in crevices. A third species was a mite (No. 64), Bhyncholophus simplex Bks. Other species are found in fewer numbers, such as the flshworm (No. 70), in soil under bearberry at a depth of 5 cm.; a shell, Zonitoides arboreus Say (No. 65) ; myriapods (No. 64), and a few other insects, including beetles, beetle larvae, and one Jassid (No. 64). 2. Terrestrial Fauna. Aside from the ants, which I have included in the first group, shells, spiders and grasshoppers are the most import- ant members of this fauna. Of the former but one species is included, Polygyra albolalris Say. It was not seen alive, but their dead shells are abundant on nearly every Cladonia clearing as well as the drier forest covered ridges (Nos. 20, 33, 88, 93, 138, 145, 174, 197). The live ones are also found in damper places or even in swamps (No. 113). Spiders were numerous especially in the clumps of Cladonia, where they crawled over and under the mats, frequently carrying egg cases. Three species were observed, Gnaphosa hrumalis Th., Pardosa sternalis Th., and Lycosa kochii Keys, (all No. 22). During the first part of July grasshoppers were infrequent, except the wingless stages, but during the last part of the month and in August they were extremely abundant. They are not confined to clearings with a copious growth of Cladonia or other vegetation, but are equally abundant on the most barren rock-ridges. Immature specimens of Chloealtis con- spersa Harr. (No. 22) were hopping over the lichens on July 6. Mature forms of Melanoplus huroni Blatchl. and Circotettiw verrucu- latusKby. (No 44, 35, 108, 131, 132), were very abundant. They fly well, making a clicking noise the while, and very rarely leave the sunny open ridge. Chloealtis conspersa Harr. (Nos. 143, 144) was collected in similar places from Prmius pennsylvanica, Diervilla, and Coptis tri- foUa, and the grasshopper Melanoplus alaskanus Scudd, (Nos. 146, 147) was taken on Gnaphalium, Diervilla and ^ass. 3. Aerial fauna. The light and warmth of the Cladonia clearing attracted many flying species, including the cicada, TiMcen riuiosa, Say, var. (44, 108, 111) ; bees, Monumetfm albifrons Kby. (68), Xantlws- arus latimanus Say (68, 108), X. nielanophea 8m. (108); the dragon- flies, Aeschna (No. 09), Ophiogoniphus cohibrUius and Tetragoneurin spinigera Say (132) ; the butterflies, PapiUo turnus Linn. (97), Basil- arohia arthcmis Dru. {97),Argynnis myrina Cramer (97), and Argynnis atlantis Edw. (32), and hosts of blackflies, Simulium vemistum, Say. ECOLOGY OF ISLE ROY ALE. 73 The butterflies, Basilarchia arthciuis Dru. and Argynnis athtnUs Edw., are so characteristic of these clearings that we knew them by the com- mon name of "clearing" butterflies. The blackflies are abundant, and are preyed upon by dragonflies, probably the chief reason for the occur- rence of them so far from the swamps. Of particular interest was the small fly, Hi/drophoriis phUomhrhi^ Wheeler, mentioned before in connection with the rock beaches. They were numerous over all the clearings, but they settled in especial abund- ance on the moist newly exposed soil which I uncovered. It is probable that they do this only for the moisture or coolness, but in one case a number of them swar*med over the pupa case of an ant, (No. C6). Of especial interest was the fauna of the large complex of Cladonia clearings just behind the camp at Siskowit Bay. (V, 3), Figs. 2//, 25, 26. There was a unifonm gentle slope from the margin of the bay back some distance inland, on which large areas were occupied by the usual growth of Cladonia, juniper and bearberry. The whole was surrounded and intersected by balsam and spruce forest. Shells were quite rare, although a few of the usual species, Polygi/ra alholahris (Say) (233), were collected. The subterranean species of ants so common about Rock Harbor were not observed. They were replaced by another species, Formica- fusca L. (223, 224, 226, 227), which built large circular flat-topped nests (Fig. 28), 5 to 8 dm. in diameter, composed of earth and vegetable debris and covered with debris of balsam and spruce needles. Two sizes, a larger (223) and a smaller (224), were sometimes associated in the same nest. Many nests had been almost completely destroyed by the pileated woodpeckers. Spiders, Pardosa sternall.'i Th. (No. 225), were frequently seen crawling over the ant's nests. Other spiders crawl over and through the Cladonia, dragging egg cases behind them, and crawl- ing into holes and crevices. Grasshoppers were abundant, as usual. Some short winged nymphs of Melanoplus fasciatus Barnst-VValk., (No. 208) were taken in thickets of Juniperus nana. They usually hide down in the juniper and will not jump out if frightened, but crawl down close to the ground, so that they are practically invisible. When once seen they can be picked up with the fingers. Sometimes they leave the clumps of juniper and jump or fly out over the Cladonia and rocks. These flights seldom exceed 1-2 m. in length, but on one occasion one flew 6 m. high and disappeared among the balsam trees. The adults of the same species (193, 201, 208, 214), with full length of wings, fly long distances at a height of 3-7 m. or more, making the usual clicking noise. They alight only on the bare roQk or on short Cladonia, avoiding the other vegetation. One fiddling grasshopper, Camnnla pellucida Scudd. (No. 228), was also taken from mats of the juniper. Bumblebees, particularly Bomhtts terricola Kby. (208) , visited the flowers of Diervilla and Melampyrum. Other bees, including Tenthredopsis nehclloidcs McOHl, Coelioxys moesta Cr., XantUosarus mclanopli€a> Sm., and X. latimanus Say, visited the same plants. A small carabid beetle, Carabus serratus Kby. (No. 208), crawls over and through the Cladonia, foraging. Leptura chrysoconia Kby. (208) was taken in the same locality. 10 i 74 MICHIGAN SURVEY, 1908. The yellow clearing butterfly, Basilarchia arthcmis Dru. (208), is very common, flying in regular paths up and down the clearing at a height of about one meter, sometimes alighting on the ground and some- times on the flowers of Opulaster. Vrocerus flavicornis Pabr. and U. fiavipennis Kby. (208, 209, 228) Avere especially common. They fly low, usually 2-3 feet above the ground with a moderate but uniform velocity. They are searching for balsam trees in which they deposit their eggs, and were sometimes taken crawling over the trunks. A small brown wasp flies low over the ground like an asilid. Asilid flies, Asilus annulaki.s Will. (208), fly low, 1-2 ft. above the ground, alight on tufts of grass or Cladonia and crawl down into it. It could not be determined what they were hunting. Three species were taken on the flowers- of the harebell, Campanula rotundifolia. They were Coelioxys nivcsta Cr., Xanthosarus melatioplwea Sm. and X. latimanns Say. Insects were more numerous on the flowers of Opulaster, from which were collected Tenthredopsis nrhelloides Mc- nm, ^Prosopis sp., Argynnis atlantis Edw., Eristalis dimidiatiis Wied., Phormia terraeiiovac Desv., P. regina Meis, and Hyetodesmia serva Meis. A wasp, Euttypus americanus Cress. (235), was found backing over the ground dragging a spider, Lycosa JcochU Keys. At brief intervals it dropped the spider and ran rapidly back and forth looking for the hole to which it was taking its capture. It seemed to have a general idea of its location, but had to crawl always exactly to it. Having found it, a similar search was begun for the spider, and then the journey was resumed in a direct line toward the hole. The typical Cladonia clearings just described were almost invariably on the lower ridges or gentler slopes. They were surrounded, and eventually entirely covered, by the balsam-spruce forest. On certain of the higher or steeper ridges, there was another intermediate stage in which the clearings were covered with jack pine. This was due ap- parently to their position; the formation of soil was slower and the drainage better, so that, even with a considerable depth of soil they were still too dry for balsam or spruce, and were accordingly occupied by the xerophile jack pine. In general ecological conditions they were but little different from the treeless associations. The ground vegetation was, as usual, Cladonia or bearberry, and the forest cover was scarcely- heavy enough to make much shade. But the mere presence of trees indicates that there was a greater deposit of the soil. Under the bear- berry and Cladonia, the soil was quite thin, but there were more loose rocks, and larger and deeper fissures, which were filled with soil. The effect on the fauna was to increase the number of subterranean species a,nd diminish the number of aerial forms. In the soil deposits up to 5 cm. deep there is practically no animal life, although ants crawl over the surface. Nests of Lasius niger L. are common in crevices and under loose stones at a depth of 1 dm. or more (Nos. 79, 82). A nest of Lasius niger L. (No. 83) was excavated under and at the side of a large stone. The stone formed the roof of shallow excavations where the pupae were stored, and the vertical wall of earth at the side was honeycombed with rounded passages 1-2 cm. high, 2-4 cm. broad, and separated by thin partitions. Under larger stones their ECOLOGY OF ISLE ROYALE. 75 nests may be built at less depth, as one of Formica fusca L. (No. 100) at a depth of 4 cm. These loose rocks tend to conserve the moisture just as do the crevices. Beetle lai-vae are rarely found, owing to the abundance of ants which feed upon them. TTiey occur under rocks or in the deepest soil deposits where the moisture is conserved. (Nos. 80, 82, 102.) No. 102 contains two species of larvae, one a Cistelid, the other Drasterius sp. The latter when collected had been captut-ed by an ant, Formica fusca L. (No. 102). A dead beetle, Dipolataxis liberta (102), was collected under aflat rock. Spiders are abundant, especially Drassus neglectus Keys (No. 101), Cicurina arcuata Keys (No. 102), and Lycosa pratensis Emer. (103). The former builds a small pocket-like web 2 by 3cm. in cavities under rocks, at a depth of about 1 dm. Spider egg cases were frequently found under stones or in rotten wood (No. 102). Myriapods were rarely seen. They seem to have regular runaways excavated thi*ough the wood or soil (No. 103). A dead caterpillar was also een: proceeding with comparative rapidity. They may be due to wave action, drainage, elevation, rock disintegra- tion or soil formation as direct agents, or to changes in the soil com- position, soil moisture, light, or heat through the indirect agency of the vegetation. Corresponding to the wide diversity in physical con- ditions there has been developed a fauna of many species adapted to many different, modes of living. Through the agency of the forest cover the light is reduced to a constant minimum, the temperature is made more uniform, the soil becomes of uniform character throughout, and the moisture is kept nearly constant. Indirectly the diffuse light is normally too weak to allow the growth of a ground cover of herba- ceous plants so that the variety of food supply is reduced. In short, the change is from heterogeneity of ecological conditions to homo- geneity, and the number of species varies directly with the heterogeneity of the habitat. This is true not only for Isle Royale. but for any biotic association. Here, however, the homogeneity is especially marked, because two species alone, the balsam fir and the white spruce, are dominant throughout. The soil in the balsam-spruce forest is a damp closely packed leaf mold, sometimes deep, sometimes shallow over the rocks, and com- posed of decaying balsam and spruce needles, mixed with decaying sticks and interwoven with fungus myceliura,. 'S'S'hen the forest is not so dense aspen and birch trees may be growing, and their leaves also mix in the mould. In such places there may be a very thin ground cover of Aster viacrophyllus, Liimaea amcricana and Pyrola chlorantha; otherwise the soil is without cover. Above this rises the dense growth of trees, the younger ones and the lower branches stunted or dead from lack of sufficient light. The insect fauna is composed almost entirely of subterranean species, all few in number, and mostly colorless. A few species of spiders are seen, and a minute CoUembolan, Tomocerns nig^r Bourl. (No. 140). Two species of myriapods (No. 140) are rather abundant in the mould, one other larger species was seen once (No. 140), and an Enchi/traid earthworm (No. 140). A few species of small shells are rarely found at depths of about 5 cm. or sometimes on the surface. They are Pyra- midula stritttvlki (Anth.), Zonitoidcs arboreus (Say), Yitrea l)inncyana (Nise), and Ancijclus sp. (140). There are no ants except a large black species which forages singly over the surface, Cainponotus hc.rciilcanus L. (No. 140). A single black Carabid, Valthvs gregarimt^ Say (No. 140, 236), also runs over the surface and hides under old balsam cones. A few species of flying insects occur, especially mosquitoes, and when ECOLOGY OF ISLE ROYALE. 77 the fresli mould is turned over a few of the moisture-lovino; flies, Hijdrophorus philomhrius ^A'heeler, appear and rest on the moist ex- posed surface. The trees' themselves shelter a more varied population. Most of the Buprestids and Cerambycids caught in the tent probably came from the forest. The dead trees of balsam or spruce are attacked by wood- boring larvae, which construct a network of chambers just between the wood and bark. Some of these turn into the wood and extend to the center, following a longitudinal or tangential path for most of the way. These holes may be filled with dust part of the way, but the greatest portion is em,pty. They are about 3 by 5 cm. in diameter, of an elliptical shape, but at the ends sometimes widen out into cham- bers a couple" of centimeters broad. Two species of larvae occupy these burrows (No. 205), and in one was found a small s])ider, Amatirohhis hennetti P>lk. (No. 205). Under the loose bark of trees which have decayed further spiders, Aiiianrohutfi hennetti Blk. (No. 205), frequently build their webs. A beetle, Caltithun adveiiu Le C. (No. 142), forages here for food, and in one case a sliell (142) was taken. A nest of Formica sangidiieu Latr. (No. 78) was found in the rotten wood of a fallen tree, but the ants probably foraged over a rock clearing near by rather than in the forest. In prostrate decaying logs the fauna is not. different from that of the leaf mold, and the same species were collected. A number of the mushrooms of the genus PJeurotus were collected on dead trees and they were inhabited by large numbers of beetles (229) Tn'toiiKi thovucicu Say, T. nuicra Lee, Boletohiiis einrticollis Say, and Grophacna S]). /J. Artificiul Clearings. The clearing about the camps both at Eock Harbor and at Siskowit Bay attracted many species of insects, p.articularly strong fliers, such as Hymemoptera, Lepidoptera and Diptera. At Rock Harbor Cow- parsnip, Heracleimi lanatum,, introduced in the island some way, was in bloom and it attracted a number of species of bees and flies. In a similar clearing on the north side of Eock Harbor a number of shells were collected from the under side of dead logs (150). Some of these were observed at no other place. They included Polygyra albolahHs (Say), Acantltintila harpa (Say), Bifidaria tappaniana (C. B. Adams), Zonitoides arljorea (Sa>), Pyranddula cronkheiti anthmuji Pils., Coehlieopu luhrica (Miill), and TaUonia eostata (Miiller). A plant of Opulaster blooming in the same clearing attracted a multitude of insects (148), including the flies Platycliirus pcltatiis ileigen, l^i/rpltus zenitulis Williston, Hphaerophoria eijlindricu Say, Eristalis diinidiatus Weed and Temnostomu acqitalig Loew; the bees Halictns rersans Lowell, Xanthosanis latiinaniis Say, and Bomhus tenicola Kby; the beetle Lcptiira ehrysocoiaa Kby. and the lepidopteron Cnpido scpio- lus Bd. X. Suiiimary. From the lake, i*epresenting the most primitive habitat, there are three lines of development culminaKng in the climax association; first, 78 MICHIGAN SURVEY, 1908. through the tamarack swamp and peat bog; second, through the gravel beach and arbor vitae swamp; third, through the rock beach and Gladonia clearings. Physiographic forces have some direct part in causing the successive changes in ecological factors, but most of them are due to the retroaction of the vegetation upon the habitat. The first stages of the series are marked by a severity of conditions which limit the fauna to a few well adapted species. The intermediate stages have generally a wide variety of conditions, leading to the development of a varied fauna. The most noteworthy in this respect is the fauna of the Gladonia clearings. The ultimate or climax stage is homogeneous because of the dominance of a few species, and the fauna is again lim- ited to a few well adapted species. , i^ke. }. Original Inland Lake Roek Be ash. Hook Clearing Jaok Pine Birch - Aspen ^ Beach Pool Grovel Beach Tamarack-Arbo Vltao Swamp Intermediate SaiJ.eainfSpruoe Climax ECOLOGY OF ISLE ROYALE. SI THE ECOLOGICAL DISTEIBUTION OF THE BIRDS OF ISLE KOYALE, LAKE SUPERIOR. OTTO ll'CUEARY, AGItlCULTUKAI, EXPERI.MEXT STATIOX. GENEVA, N. Y. I. IXTRODUCTIOX. In this report I shall discuss the habits of the different birds and their relation to their environment as found vipon Isle Royale. The different localities visited will be described, the birds listed as found in each locality, and the details of their habits and distribution de- scribed. On account of the limited time, I was unable to examine a large part of the island, but representative localities were visited,, so that a general idea of the bird life of the island can be gained from this re- port. For example, a nvimber of tamarack swamps were visited and certain birds were fotind in each of these; it therefore seems reasonable to infer that these birds are found in the many other tamarack swamps which were not visited. Observations were made in five different localities by members of the Museum party, but only those visited by the writer will be described. In connection Avith this paper the ''Annotated List of Birds" should be consulted. These localities will be taken up in the following order: 1. Light-house Peninsula. 2. Trail to McCargoe Cove. 3. West End of Rock Harbor and Trail to Summer Lake. 4. Siskowit Bay Region. II. LIGHT-HOUSE PENINSULA. This station included the land between Conglomerate Bay and Rock Harbor. The conditions in this small strip of country varied very much, and on this account it will be divided into a number of stations as fol- lows : 1. Lake Superior and Beach (Station I, 1). 2 Spruce and Balsam Forest (Station I, 2 and 3). 8. Tamarack and Arbor Vitae Swamps (Station I, 4). 4. Jack Pine Ridge (Station I, 5). 5. Sphagnum and Spruce Bog (Station I, 6). 6. Valley at Head of Conglomerate Bay (Station I, 1). 1. Lake Superior and Beach (Station I, 1). This station included the whole of Tonkin and Conglomerate Bays and tliat portion of Lake Superior and Roclr Harbor which could be seen from the light-house. The water Avas deep, cold and contained very little vegetation. The shore bordering the lake was composed of jagged, desolate, wave-washed rocks (Figs 2, 5), and only in the 11 82 MICHIGAN SURVEY, 1908. most protected portions of the bays and harbors were trees found growing near the edge of the water (Fig 4). This accounts for the fact that no shore birds or vegetable feeding water fowl were ob- served here. The birds seen in this station were as follows : Herring Gull, Loon, American Merganser, Hooded Merganser, Spotted Sandpiper, Song Spar- row, Myrtle Warbler, Olive-bacJced Thrush, Crow and Osprey. The Herring Gull was the only bird seen on the water in large num- bers. At almost any time of the day there wore fifteen or twenty in sight, and sometimes they came in large flocks to eat the refuse thrown -along the shore of Rock Harbor by the fishermen. Seventy-seven were once counted, and occasionally the number was greater. When not feeding on the water they passed the time soaring in the air or resting on the bare rocks. They seemed to prefer soaring during windy weather. With the head toward the wind they would move slowly upward and forward for some time, then turn suddenly and soar away with the wind at a rapid rate, then swing around in a graceful curve and again mount upward. The American Merganser, Loon and Hooded Merganser were occa- sionally sieen on the water. On July 27 and 28 a female Hooded Mer- ganser and six young were observed. These ducklings were yet small and could be overtaken with a row boat, but when pursued they escaped by diving. Thus it will be seen that, excepting the Gulls, water birds were scarce and the shore birds nearly lacking. Only one shore bird, the Spotted Sandpiper, was seen and that was observed two or three times; this Avap probably a migrant. The other birds seen on the shore, were the Crow, Myrtle Warbler, Song Sparrow, and Olive-backed Thrush. They occasionally came from the bushes and forests to feed there. Of these birds the Song Sparrow Avas seen the most often, and almost every morning could be heard singing on the small rocky islets partially covered with bushes, which lay just east of the light-house. 2. fiprucG and Balsam. Forest {Station I, 2-3). In this forest of spruce, balsam and birches, there were many low rock ridges whose tops were almost destitute of soil and trees, thus forming a long, narrow, natural clearing of not more than two hundred yards in length and from thirty to sixty yards in width {Fig. 8). Near the light-house there were five of these ridges from thirty to two hundred yards apart; while farther to the west there were more of them, but they were farther apart. . On account of these openings in the forest, there were many birds here that frequented partial clearings, yet no birds that inhabit large tracts of cleared land, except the Chipping Sparrow, which occurred in the small clearing at the light-house. The birds found under these conditions were as follows : Red-breast- ed Nuthatch, Chipping Sparrow, Nashville Warbler, Black-throated Blue Wai'bler, Black-throated Green Warbler, Chickadee, Flicker, Gold- en-ciH)wned Kinglet, Bay-breasted Warbler, Crow, Myrtle Warbler, Sparrow Hawk, Magnolia Warbler, AVilson's Thrush, Olive-backed ECOLOGY OF ISLE ROYALE. 83 Thrush, Pine Siskin, Purple Finch, White throated Sparrow, Tree Swal- low, Barn Swallow, Sharp-shinned Hawk and White-winged Crossbill. The Purple Finch, Pine Siskin, Sparrow Hawk, Sharp-shinned Hawk, Bay-breasted Warbler, Black-throated Blue Warbler, Tl-ee Swallow and Barn Swallow were only oceasionallj- seen. Some of the Warblers were common, and it was interesting to note the difference in the localities which they frequented. The Myrtle Warbler was most frequently seen near the shore. The Magnolia Warb- ler frequented the small spruce and balsam trees but was not seen on the sliore. The Black-throated Green Wai'bler was always observed in that part of the forest where there were many birch trees, and the Nashville kept near the partial clearings. < Nests of the Myrtle Warbler, Chickadee, Golden-crowned Kinglet, Olive-backed Thrush and Chipping Sparrow were found in this locality. The nest of the Myrtle Warbler was found July 7, on a small jack pine standing near the edge of a rocky cliflf, which rose perpendicularly from the water to a height of about twenty feet. It was composed of small twigs, dried grass and pine needles, and contained four young about a week old. While we were near, the old bird approached the nest very cautiously. It would fly from tree to tree until within about fifty feet of the nest and then drop down near the ground and fly low until below the nest; when leaving it flew along the edge of the cliflf. On the same day a Golden-crowned Kinglet was seen to take a bit of moss and fly into a clump of stunted spruce trees on a rock ridge. The tops of the spruce were so thick and bushy that it was impossible to see the nest from the ground, although the tree was not more than twenty-five feet high. On climbing the tree a half finished nest was found built mostly of green moss. By July 21 the nest was finished and contained eight small eggs. This beautiful mossy cup was about four inches in diameter and of the same depth, but the cavity contain- ing the eggs was still smaller, as the wall of the nest was about two and a half inches thick and lined with hare fur. In front of the light-house at the edge of the beach, stood a small spruce about twenty feet high, on a horizontial limb of which was the nest of a Chipping Sparrow, composed entirely of grass. When the nest was found on July 5 it contained four young that had evidently just hatched. An Olive-backed Thrush's nest was found July 8, in a low limb of a spruce that stood near the shore. It was found five feet from the ground, composed of grass and moss, and contained three very young birds. The old bird would not approach while I was near the nest and was so shy that the true owner of the nest was diflflcult to determine. Probably more nests would have been found had we arrived upon the island earlier, as many young were able to fly when we came, and several immature Magnolia Warblers were found at that time in the bushes near the light-house. Birds were more abundant in this locality than in any other of the same size. Why this was true, I did not determine. 84 MICHIGAN SURVEY, 1908. 3. The Tamarack and Arior Vitae Swainps {Station I, -^). This almost impenetrable swamp of cedar and tamarack, situated at the head of Tonkin Bay, extended back about a quarter of a mile toward the southwest. To cross this swamp was difficult, on account of the fallen trees and numerous low branches, but a rock ridge extended from the bay through the middle of the swamp, almost to its western end, and furnished a convenient route into it. This ridge in- fluenced the bird life of the vicinity because of its different ecological conditions. It was bare in places, but most of it was partially covered with birch, spruce and balsam. The birds seen in this swamp habitat were as follows: Nashville Warbler, Red-breasted Nuthatch, Chickadee, Black-throated Green Warb- ler, Eaven, Brown Creeper, Yellow-bellied Flycatcher, Hairy Wood- pecker, Winter Wren, Black-throated Blue Warbler, Flicker and Canada Jay. On July 11, nests of the Black-throated Green and Nashville Warblers Mere found on the north slope of the ridge within twenty-five yards of each other. The nests of the Black-throated Green was in a cedar tree about twenty feet from the gi^ound. It was composed of grass, moss and twigs and contained young. The nest of the Nashville Warbler was in a cavity in a thick bed of moss which covered the face of a small cliff five or six feet high. Here, in a soft nest composed of lichens and lined with grass, were found five young in the down. • These two birds acted very differently when one was near their nest. The Black-throated Green would come within less than ten feet of the observer and scold while moving restlessly about among the branches. The Nashville Warbler was not as bold, for it remained up in the tree tops. It would hop on a branch, turn around a few times, turn anxious- ly toward the nest and then repeat the performance ', but it never uttered a sound. Here was one of the diflSculties in judging what localities birds preferred. These two birds nested on the slope of a rock ridge and fed in a cedar and tamarack swamp. To which did they belong? How- ever, judging from other observations, I would say that if the natural clearing had not been here the Nashville Warbler would not have been found, while the Black-throated Green might have been. The Black-throated Green, Black-throated Blue and Nashville Warb- lers, Chickadee, and Eed-breasted Nuthatch were nearly ah^ays found in this swamp, and these were in the more open parts where the trees were not so close together. I \'isited the thickest part of the swamp many times without seeing a single bird. ^. Jack Pine Ridge {Station I, 5). This habitat was on the north side of Conglomerate Bay and-composed a portion of the south side and the top of a hill about 100 feet high. The side of the hill was dry and rocky, and was partially covered with scattered aspens and clumps of jack pines {Fig. IS). Where there were no trees the ground was partially covered with mosses, lichens, bearberries, golden rods, etc. The top of the hill was bare rock with jack pines and a few plants growing in the crevices. Occasionally there was a small gullv with other trees growing in it. ECOLOGY OF ISLE ROYALE. 85 On account of tlie desolate character of this locality few birds were found here. A Cedar Waxwing's nest containing five eggs was found July 10. Juncos and White-throated t:?parrows were occasionally heard singing among the jack pines. •5. Sphagnum and Spruce Bog. {Station I, 6). This small bog, situated on top of the hill north of Conglomerate Bay, was covered with sphagnum moss and bushes with several black spruce trees scattered over it. There were also several tamaracks and spruce at the edge of the bog. (Fig. lIi). The birds seen, here were: Golden-crowned Kinglet, White-throated Sparrow, Cedar Waxwing, and Black-throated Green Wai'bler. The Oven Bird and Wilson's Thrush were heard in the forest near by. 6. Valley at Head of Conglomerate Bag {Yicinlty of Station I, 1). This location included the alders and the partial clearing at the mouth of the brook that emptied into the head of Conglomerate Bay. The partial clearing, evidently due to flre, as blackened logs were still lying around on the ground, was covered with weeds, raspberry bushes, dogwoods and clumps of small birches. The birds seen here were : White-throated Sparrow, Canadian Warb- ler, Redstart, Flicker, Winter Wren, Chickadee, Nashville Warbler, Mag- nolia Warbler, Olive-sided Flycatcher, Olive-backed Thrush, Sparrow Hawk and Cedar Waxwing. The Redstart and Magnolia Warbler set nied to be restricted to certain parts of this locality. The Redstart was always seen among the alders, while the Magnolia Warbler kept among a patch of evergreens at the foot of the hill on the north side of the habitat. III. TRIAL TO lICCAnOOE COVE. This station included the country along the trail which ran from Rock Harbor to McOargoe Cove. This trail started on the north side of the harbor at the mouth of Benson Brook wliicli it followed nearly to Lake' Benson, then it crossed tlie liills to Sargent Lake and from there it went to ^tlcCargoe Cove. As I did not make any observations north of the Greenstone Ridge, I will only describe that portion of the^country betwefen Rock Harbor and the top of the Ridge. In this portion there were several different conditions which will be described in the follow- ing order : 1. Ransom Clearing (Station II, 1). 2. Benson Brook (Station II, 1). 8. Spruce and Tamarack Swamps (Station II, 2 and 5). 4. Rock Ridge Clearings (Station II, 3). ]. Ransom Clearing (Station II, 1) . This small clearing on the lowland at the mouth of Benson Brook was covered with grass and large clumps of alders, birches and aspens. These bushes scattered through the clearing formed an excellent habitat for birds, and, althougli the clearing was small, thirteen species were observed here. They were as follows: Black-billed Cuckoo, Canada Jay, 86 MICHIGAN SURVEY, 1908. Song Sparrow, Alder Flycatcher, Wliite-throated Sparrow, Redstart, Eed- eyed Vireo, Cedar Waxwing, "^Vilson's Thrush, Olive-backed Thrush, Sparrow Hawk, Purple Pinch and Pine Siskin. Every time this station was visited there were one or two Alder Flycatchers among the alder bushes, sometimes on top of the highest bush and sometimes near the ground. They seemed to be always on the lookout for insects, and every few minutes they would fly several feet into the air and a snap of the bill told that some insect had been caught. They could often be located by their "pep" of alarm, and in the morning I frequently heard them sing a short song. The Redstart and Nashville Warbler were often seen among the alders also. Both were always on the move. The Redstart kept flitting from branch to branch, only pausing an instant at each one to look for insects, while the Nashville Warbler would light on a limb and start to hop toward the top, looking an instant at each leaf as it passed. • 2. Benson Brook {Station II, 1). The conditions along this little brook are difficult to describe in a general way because they were so diverse; every few rods there was a change. The little stream meandered through dense forests of cedar, spruce and birch; through thickets of alders, dogwoods and small maples; rushed through narrow ravines between bare topped ridges, over rocks, through forests of birch and aspen until it finally reached the harbor at Ransom clearing. The birds found along this brook were the White-throated Sparrow, Redstart, Winter Wren, Red-eyed Vireo, Cedar Wasrwing, Oven Bird, Sparrow Hawk, Wilson's Thrush, Olive-backed Thrush, Blue Jay, Canada Jay, Crow, Purple Finch, Sharp-tailed Grouse, Grinnell's Water Thrush, Flicker, Magnolia Warbler, Hairy Woodpecker, Nashville Warbler, Red- breasted Nuthatch, Golden-crowned Kinglet and Chickadee. The Spar- row Hawk, Blue Jay, Flicker, Sharp-tailed Grouse, Cedar Waxwing and Purple Finch were seen more often in the clearings where there were berries, grasshoppers and other insects. The Winter Wren and Water Thrush were always seen near the brook. The former frequented places wheje the undergrowth was thick. It was often observed flying along the brook and stopping every few yards to look under the leaves and logs for insects, and one was shot with a spider {Amaurohius 'bemietti Blk.) and two mosquitos in its mouth. Sometimes this shy bird would venture away from its damp retreat, perch upon the top of a tree and pour forth a melody that rivalled any song heard in these woods. The Oven Bird and Red-eyed Vireo were nearly always found among the birches and aspens. The former very frequently was flushed from among the honey-suckle bushes on the gi'ound, but the Vireo was always in the trees. The Magnolia Warbler, Red-breasted Nuthatch, and Golden- crowned Kinglet were always seen in that part of the forest where there Avere several spruce or cedar trees. A large number of different species of birds was observed in this habitat, but that was because it was so large. In reality the country was rather desolate, for with the exception of some damp places aftng the brook, the original forest has all been burnt off and was only partial- ly replaced by a second growth of birch and aspen. 11 ECOLOGY OF ISLE ROYALE. 87 3. Tuinamck and Spruce Stvainps {Station II, 2 and 5). About a quarter of a mile north of Benson Brook there was a swamp similar to I, 5, except that it was laroer and had more spruce and tamarack trees scattered through it. The ground was covered with sphagnum, Labrador tea, pitcher plants, etc., but apparently nothing that would attract birds except the trees, The birds seen here were the Eed-breasted Nuthatch, Marsh Hawk, Junco, Canada Jar, Black-throated Green Warbler, Black-throated Blue Warbler, Chickadee, Golden-crowned Kinglet, White-winged Crossbill, Yellow-bellied Flycatcher, and White-throated Sparrow. The Junco probably strayed here from a large rocky clearing near by, as only one was seen in the swamp, but it was heard in the clearing every time I visited it. About a quarter of a mile further on toward Greenstone Ridge, the trail crossed another swamp similar to this one, though it was somewhat longer. Since the conditions were the same in these two places, many of the same birds would be expected to occur in each, and this was the case as will be seen by comparing the list given above with the following : Olive-sided Flycatcher, Red-breasted Nuthatch, Nashville Warbler, Canada Jay, Chickadee, White-winged Crossbill and Golden-crowned Kinglet. Near Forbes Lake there were two other swamps and in these the fol- lowing birds were seen: White-throated Sparrow, Canada Jay, Cedar Waxwing, White-winged Crossbill, Red-breasted Nuthatch, Golden-crown- ed Kinglet, Chickadee, Nashville Warbler and Flicker. All these were found in both swamps with the exception of the Nashville Warbler and Flicker. There is a marked similarity in the lists of birds seen in each of these five swamps, and five of the species were found in all of them. Jf. Bock Ridge Clearings {Station II, 3). This habitat consists of all the rock ridges which were crossed by the trail after it left Benson Brook. These ridges were nearly all bare on the top, owing to the absence of soil. They had been burnt over several years ago and the stumps that are left show that they were originally almost if not entirely covered with forests. The trees that were found in places where there was a little soil were almost entirely aspen and birch. The birds found in this habitat were the Cedar Wax- wing, Junco, Bay-breasted Warbler, Mourning Warbler, Robin, White- throated Sparrow, Olive-backed Thrush, Sparrow Hawk and Red-eyed Vireo. Very few birds were seen in the clearings, probably because the heat of the sun drove them to the shade, as miost of the birds were observed at the edge of the clearings, in places where the ground was partially covered with trees. IV. WESTERN EXD OP EOCIC HARBOR AND TRAIL TO SUMNER LAKE. Thjs station comprised the western end of Rock Harbor and a por- tion of the adjoining land. It was divided into five habitats. 88 MICHIGAN SURVEY, 1908. 1. Harbor (A^icinitv of Station Til, 2). ■2. Small Islands (Station III, 1). 3. Bulrush Zone and Delta (Station III, 8). 4. Trail to Sumner Lake (Station III, 4). a. Birch Forest. b. Birch and Coniferous Forest. 5. Sumner Lake (Station III, 5). /. The Harhor iViciiiify of t^tation TIT, 2). In this habitat the following list of fiSh-eating birds were found: Loon, American Merganser, Herring Gull, Kingfisher and Bald Eagle. An adult American Merganser and a number of young were observed about the middle of July, and about a week later another adult female and twenty-three young were seen. Although the young birds were quite small they were good swimmers, and it was imi)Ossible to get near them in a row boat, except by cornering them in a small bay or in the end of the harbor. The Loon was often seen and heard here, and, once seven were seen together? Occasionally one of the flock would swim around and around in a circle as fast as it could, splashing the water so that it could be heard for at least half a mile. It was impossible to get near these birds, not even close enough to shoot them with a shot gun, for as soon as they thought it was da.ugerous the.A- ^^'0uld dive, to appear after a few minutes very much farther away. It is very difficult for the Loon to rise from the water, as it must fly a long distance flapping its wings and pushing the water with its feet before it can get into the air. The Eagle was seen on a tree at the edge of the water. 2. SmaU Island.^ {Station ITT, 1). Near the west end of the harbor there were two small islands partially covered with stunted cedar, spruce and birch trees, where many birds nested. The probable reason for this was that no squirrels inhabited the islands. On one island three or four I'ods long were found the nests of four Cedar Waxwings, two Myrtle Warblers, a White-thl-oated Spar- row and a Song Sparrow, and on the other island which was somewhat smaller, were a number of Cedar "SVaxwing's nests, three containing egg-s or young, and the remainder being empty, most of them last year's nests. The Waxwing's nests were from tliree to fifteen feet from the ground and were composed entirely of lichens (Vsiiea). These birds do not get excited as do many birds when their nests are disturbed. When I looked into these nests I did not hear a scolding note, although some of the owners were sitting on a tree not far away. Four Myrtle Warbler's nests, two old and two new, were found. These nests were placed on spruce and cedar trees, from six to ten feet from the ground, and were composed of small twigs and grasses with a lining of feathers. One nest contained small young, July 21, and the other contained nearly fully fledged young. The White-throated Sparrow's nest was made of small sticks and grasses with a lining composed entirely of gi'ass. It was on some bushes about a foot and a half above the ground, and contained one egg. ECOLOGY OF ISLE ROY ALE. 89 3. Bitli-tish Zone and Delta (Citation III. 3). Tliis small grass and sedge covered marsli was too smiall to attract many marsh birds, and a pair of R«amp Sparrows with, two young, a pair of Kingfishers and Song Sparrows, a Red-winged Blackbird and the Lesser Yellow Legs Avere the only birds observed here. The last two were only observed once, and no doubt they were only stragglers here. This small marsh was surrounded by a forest of spruce, birch and balsam, and here the Golden-crowned Kinglet, Magnolia Warbler, Chick- adee and Red-breasted Nuthatch wei'e found. //. Trail to Humnrr Lal-e {mation III, J-i). Starting from the harbor this trail first went up a hill through a birch forest, then across a narrow cedar swamp into a birch, spruce and balsam forest and down the hill to Sumner Lake. As the birds found in the birch forest were not the same as those found in the birch, spruce and balsam forest, the habitats will be distinguished. The cedar swamp was too small to be of any importance, and the birds in it were nearly the same as in the birch, spruce and balsam forest of which it will be considered a -part. a. Birch Forest. Judging from what had been observed before these birch woods were visited, I expected to find the Oven Bird and Red-eyed Vireo, and upon investigation, many of both kinds were found. A family of Black-throated Grreen Warblers were also seen. Several Cedar ^'S'axwings and White- throated Sparrows were observed along the edge of Rock Harbor near the trail, but they occurred almost everywhere along the edge of the Harbor irrespective of the kind of trees. In rowing along the shore these birds were seen very much more often than any other. l). Birch and Coniferous Forest. This habitat was frequented by the Chickadee, Golden-crowned King- let, and Red-breasted Nuthatch, the three most common birds in all the coniferous forests that were visited. The Winter Wren was heard in the cedar swamp. 5. Sumner Lalce (Station III. 5). This habitat included Sumner Lake and the grassy marsh whicb sur- rounded it. Everywhere in the marsh the ground was soft, and the thick mat of grass sank under the weight of the body until the water poured into the shoe tops. The line dividing the grass and sedges from the forest was very distinct, but there were several stunted tamairacks and alders growing out in the marsh (Figs. 18-23). Many White-throated Sparrows were heard in the forest near the marsh, and at the foot of one of the alder bushes near the edge a nest was found hidden in a bunch of grass growing around the bush. Here in a well built nest of grass were two nearly fledged young (July 18). On the same day another nest of this bird was found on the other side of the lake, in a position similar to the one described above, but instead 12 90 MICHIGAN SURVEY, 1908. of young it contained four bluish white eggs densely and irregularly variegated with brown. Out in the marsh a Bittern was flushed from the grass, and near by a deserted nest containing a bad egg and the bones of two young was found. This nest was only a depression in the tangled mat of grass in which it was situated. Two Loons were seen on the Lake many times, and these two birds were much 'tamer than Loons usually are, for they swam very close to the bank where I was standing. As soon as they saw me one of them gave a weird and rapid "ha! ha! ha!" and on being imitated it would reply every time. .V Hooded Merganser, another fish-eating bird, was also observed here. V. SISKOWIT BAY REGION. When I arrived heie in August the breeding season was practically over. Many young birds could fly almost as well as the adults, and families were roving about the forests. Sandpipers were probably mi- grating then, and although many were seen here it cannot be said that they bred. In two weeks other birds began to come from the north in large flocks, so that most observations were on habits of birds during migration. Another evidence that the breeding season was over was the decrease in the amount of singing. This was first noticed on July 20, and in the next few days some species were heard for the last time. The follow- ing is a list of birds with the last date upon which they were heard singing : Nashville Warbler, July 24 ; Myrtle Warbler and Olive-backed Thrush, July 25; Wilson's Thrush, July 26; Magnolia Warbler, Black and White Warbler and Redstart, Aug. 4; Winter Wren, Aug. 8. Although birds are more apt to be found in all kinds of conditions during migration, yet many of them showed a preference for certain localities, so the -localities in which the birds were seen will be given. This station has been, subdivided into the following habitats : 1. Siskowit Bay and Shore (Station V, 1). 2. Trail to Siskowit Lake (Station V, 4). 3. Siskowit Lake (Station V, 6 and vicinity). 4. Burning West of Outlet to Siskowit Lake (Station Y, vicinitv of 9). - 5. Long and Menagerie Islands (Station V, 10). 1. Siskowit Bay and Shore (Station V, 1). The conditions at this place were about the same as those at Rock Harbor, and almost the same species of birds were seen. Those seen here were: Herring Gull, Loon, Scaup Duck, Solitary Sandpiper, Spotted Sandpiper, Kingfisher, Amei'ican Merganser and Osprey. I cannot say with any certainty how many of these birds bred in this vicinity, but the Gull and Merganser did, as a female Merganser with a flock of very small young was seen several times, and the Her- ring Gulls bred on the Islands south of the bay. The Loon, Kingfisher and Spotted Sandpiper were observed nearly every day. The Solitary Sandpiper was seen only once, on August 16. On August 8 four young Gulls were obtained from a fisherman, and - ECOLOGY OF ISLE ROYALE. 91 Ave had an opportunity to study the habits of these birds. One was nearly full-grown, while the other three were just getting their wing feathers. All were quite tame and the oldest would eat from the hand and allow itself to be picked up. We were surprised to find how clean these young Gulls were, for the nests were as filthy as those of the domestic Pigeon. They all seem very fond of bathing, and the largest one took a bath several times a day. It A^ould swim out into the bay, splash water over itself with its head and wings, dip its head under water, then shake itself; after repeating these performances several times it would come to the shore, flap the wings and jump as if try- ing to fly. They were very particular about keeping their bills cleau, for after eating they would walk to the water, immerse the bill and shake the head. 2. Trail to Sislcoirif Lake {Station V, J,). This habitat included all the forest along the trail between Siskowit Bay and Siskowit Lake. If it had been in the breeding season it might have been divided into two or three different habitats, but the migrating birds did not seem to show any preference for a particular forest. The birds seen at this station were as follows : Golden-croA'STie^l Kinglet, Chickadee, Raven, Pigeon Hawk. Winter Wren, Red-breasted Xutliatch, Bay-breasted Warbler, Red-eyed Vireo, Hairy Woodpecker, Magnolia Warbler, Black-throated Green Warbler, Brown Creeper, White-throated Sparrow, Tennessee Warbler, Flicker, Canada Jay, Junco, Blue Jay, Pileated Woodpecker, Nashville Warbler, Sparrow Hawk, Chipping Sparrow, Grinnell's Water Thrush, Purple Finch, Pine Grosbeak, Sharp-shinned Hawk, Myrtle Warbler, Black-throated Blue Warbler, Olive-backed Thrush, Downy Woodpecker, Yellow-bellied Fly- catcher, and Cape May Warbler. The nests of only two birds were found here, the Chickadee and Gold- en-crowned -Kinglet. The Chickadee's nest was in a dead birch tree about ten feet from the ground, and contained four young which were able to leave the nest August 11. The Kinglet's nest was in a spruce tree about thirty feet from the ground. Both old birds were observed carrying insects into the tree, but the top was so thick that the nest could not be seen from the ground. On August 16 the young birds werfe still in the nest. The Nashville Warbler, Olive-backed Thrush, Junco, White-throated Sparrow and Chipping Sparrow frequented partial clearings or clear- ings in the breeding season but were found in the forests in the second week in August. On August 11 a flock of birds were seen feeding iu the top of a tall tamarack. They were mistaken for warblers but on shooting one to identify it, it was found to be a Chipping Sparrow. 3. Siskowit Lake {Station V. 6). This Lake was six miles long and about two miles wide at the widest part. The shores were mostly rocky, and trees grew down nearly to the waters edge. The birds found here were : Herring Gull, Osprey, Eagle, Spotted Sandpiper, American JMerganser, Loon, and Kingfisher. 92 MICHIGAN SURVEY, 1908. The Song Sparrow and Grinnel's Water Thrusli were also seen along the shore. The American Merganser, Loon, and Eagle nested in the neighborhood. Three different families of Mergansers Avere seen on the lake. One consist- ed of a female and three young, but I did not get close enough to the other two flocks to count them. When first observed, these tvvo flocks were to- gether, but they separated when we rowed toward them. Two young Loons in the down were seen August 30. An Eagle's nest composed of sticks was found about 125 yards north of the lake, on top of a dead pine which was at least sixty feet high. The nest was four feet in diameter, and contained one young bird nearly ready to fly. .'i. Biiriiiiig West of Outlet to ^iisl-oirit Lake {Station T. 9). Here the original forest had all been bui'nt away and was only partially replaced by a second growth of birch, mountain ash, aspen, wild cherry, June berry, and northern maple. Between the trees the ground was covered with grass, currants, fire weed and other plants. The stream that formed the outlet of Siskowit Lake formed the east- ern boundary of the burning. The birds found in this partial clear- ing were as follows: Purple Pinch, Cedar Waxwing, Hawk Owl, White- throated Sparrow, Chickadee, Eedstart, Myrtle 'S^'arbler, Flicker, Eed- eyed Vireo, Black and White Warbler, Xash^ille "\^'arbler. Sharp-tailed Grouse, W^ater Thrush, Olive-sided Flycatcher, ('hipping Sparrow and Song Sparrow. The Hawk Owl bred some place near here, as a. young bird with only down on its head was taken August 4. This owl was seen flying around the clearing in the middle of the day and in the bright sunlight. The young bird was quite tame, or rather it was ignorant of the ways of man. It flew fromi one dead stub to another uttering a peculiar screech a.« it flew. The old bird was seen about a quarter of a mile away on the top of a dead tree, but was wary and flew away. Along the stream tliere were several dead trees still standing, and on these trees eight to ten or more Myrtle Warblers were seen many times. These warblers sat om the limbs and watched for flies like flycatchers, and ever\' few minutes the snap of a bill sounded the death note of some unfortunate insect. They did not sit in one • place as long as a. flycatcher does, but on the other hand they were not constantly in motion like most warblers. Very little can be said about the other birds that were seen here. The Purple Finch and Cedar ^^'axwing fed on ' the berries here, and a Grouse was taken with berries and grasshoppers in its crop. The Water Thrush was seen near the lake and stream. 5. Long cmd Menagerie Islands {Station V, 10). These two long narrow rocky islands were on the south side of Sis- kowit Bay about three miles from the mainland. Long Island was covered with trees except for a wide belt along the shore which was washed clean by the waters. Menagerie Island, on which the light- house was situated, had very few trees on it, as the top was barely out of the reach of the waves in severe storms. ECOLOGY OF- ISLE ROY ALE. 93 Menagerie Island was visited twice, oq August 6 and 10. Tbe birds seen here were: Song Sparrow, Barn Swallow, Tree Swallow, Herring Crull, Spotted Sandpiper, and Humming Bird. The Barn Swallow built in the boat-house and under the cliffs aloug the shore. On August 16 the nests under the cliffs contained young nearly ready to fly. These cup shaped homes were comjposed of moss and mud, lined with feathers, and placed on small projections of the rock. The light-house keeper, Mr. J. A. Malone, told us that the Tree Swal- low built in the tower; but at this time the young were probably gone as none were observed entering the light-house, although many were flying around. Long Island was visited on August 6, but no observations were made on any birds except the Gulls. These birds nested hei*e by the thou- sands. The nests were among the rocks, some being just beyond the reach of the waves of ordinary storms, and others back among the bushes. They were from one to two inches thick, and composed of grasses, sticks or moss, depending on which of these materials was found near. Most of the nests were on the south side of the island, and only a few were found on the north shore. At the approach of the boat the young Gvills that could not fly swam out into the water or hid in the bushes, while the old birds flew around overhead utter- ing their weird notes of alarm. XI. Suminary. This brief review of the birds found in each of the habitats studied on Isle Royale will give an idea of the birds that should be expected to occur in similar habitats of the island which were not visited. Of course only the common birds will be mentioned, because preference cannot be de- termined by a few observations. The habitats of this rugged and hilly island presented a variety of conditions. There Avere bays, lakes and harbors, with rocky shores, wave-beaten and desolate. There were swamps that were covered with sphagnum moss and low bushes witli here and there a black spriice or tamarack tree, other swamps that were covered with a dense forest of cedar and tamarack. There were clear- ings and partial clearings, forests of birch, containing scattered bal- sams and spruce, and still other forests of spruce and balsam con- taining a few birch trees. The characteristic birds of each of these habitats will be discussed in the order just given. 1. Water Birds. The water birds found on the harbors and small . lakes were the Herring Gull, Loon, American Merganser, and Hooded Merganser. Of these birds the Herring Gull was the most abundant species and could always be seen on Lake Superior and quite often on the smaller lakes on the island. The American Merganser pro- bably ranked second in abundance. The Loon was quite numerous, and at first it seemed as if theA- were more abundant that the Merganser, but in time it became evident that the Merganser was the more numer- ous, though much less conspicuous, as they did not make any noise, while the Loon is very noisy and can often be heard a mile away. The Mer- ganser frequented the bays, harbors and larger inland lakes. The Loon was seen very often on the larger bodies of water, but seemed to 94 MICHIGAN SURVEY, 1908. prefer the smaller lakes more than the other water birds, as every little lake contained a pair of Loons. Young Mergansers and Gulls were often seen, but, strange as it may appear, young Loons were only seen once, August 10. The Osprey, Eagle and Kingtisher were also seen several times, but only the latter was seen around any of the smaller lakes, and it was not often seen. These lakes abounded in small fish and would have been a good feeding ground for Kingfishers, but there were no sand banks around the small lakes AAhere it could have nested, and this may have been the reason for its absence. There were two sand banks along the shores of Bock Harbor, and these were used as nesting sites. 2. ^Jwrr Birds. The Solitary and Spotted Sandpipers were seen along the shore, but these were probably migrants as only one or two Spotted Sandpipers were seen before August 1. Although they were not shore birds the Cedar Waxwing, Winter Wren and White-throated Sparrows were often seen and heard while rowing along the shore. The Cedar Waxwing would sit on the tops of the dead trees and every few minutes would fly out over the water after insects. Herring Gulls nested on the shores of the smaller islands in large numbers but very few nested on the maini island. There is a reason why they choose the smaller islands instead of the mainland, and it is probably because there are nO' minks, lynx or other carnivors on these small islands. The Gull seems to place its nest on the shore at random, without any view to protection or secrecy, and if there were mink or lynx about the young would soon all be killed by these animals. The Barn Swallow nested underneath the cliflfs along the shore at Menagei'ie Island and at Srovill Point. The Song Sparrow and Myrtle Warbler were often seen feeding on the shore, and both were found breeding near it. The Song Sparrow frequented the small rocky islands in front of the light-house, one of the islands in the west end of ■Rock Harbor, and also Bansom Clearing on the north side of the Har- bor. Even in this clearing it was never seem far from the water. The Myrtle Warbler was found breeding on the north shore of Tonkin Bay, and on an island in the west end of Bock Harbor. 3. Birds Frequenting Swamps. The characteristic birds of the tamarack-spruce swamps were the Cedar Waxwing, Chickadee, Red- breasted Nuthatch, Golden-crowned Kinglet, White- winged Crossbill, Canada Jay, Nashville Warbler and White-throated Sparrow. Pro- bably none of these birds were found here simply because it was a swamp, for all frequented other localities. The White-throated Spar- row, Cedar Waxwing, and Nashville Warbler are chai-acteristic of par- tial clearings, and this was really a partial clearing because the trees were so far apart. The White-winged Crossbill, Bed-breasted Nuthatch and Golden-crowned Kinglet ate characteristic of coniferous forests, and as the trees in the swamp were nearly all coniferous trees, this would therefore be their natural habitat. The White-winged Crossbill feeds on the seeds of the tamarack trees, and during the first few weeks of July it was only seen where there were tamarack trees; dur- ing the latter part of July, when the seeds of the spruce became more matute, they were seen many times in the spruce and balsam forests. ECOLOGY OF ISLE ROY ALE. 95 The seeds of tlie tamarack mature quicker than the spruce, hence the Crossbills would prefer tlie tamai-acks during the earlier part of the summer. The Black-throated Green Warbler was characteristic of forests where there were a number of large birch trees, and this bird was only seen in those swamps which had several of these trees around the edge. Indeed the only true swami> bird seen here was the Marsh Hawk, and that was only seen once. In the thickest part of the cedar swamp only a iew birds were seen, and these were the Winter Wren, Chickadee, Red-breasted Nuthatch, and Bi'bwn Creeper. The Canada Jay, Nashville Warbler, Black-throated Blue Warbler and Black-throated Green Warbler Avere seen where the trees were tall and farther apart. 4. Birds of Clearinf/s and Partial Clearings, The characteristic birds of the clearings were the Chipping Sparrow, Junco, White-throat- ed Sparrow, Flicker, Cedar Waxwing, Purple Finch and Sharp-tailed Grouse. The Cedar Waxwing and Purple Finch were often seen feeding ou berries in the clearing-s, and a Flicker was observed scratching in an ant's nest and eating the ants. iUany ants nests were found scratched to pieces, probably by these birds. The characteristic birds of the partial clearings were the White- throated Sparrow, Cedar N'^'axwing, Chickadee, Olive-backed Thrush, Wilson's Thrush and Nashville Warbler. 5. Birds Frequ&iiting the Forests. In the forests of birch or aspen the Eed-eyed Vireo and Oven Bird were quite abundant, and in many small tracts of birch and aspens these were the only birds seen. Other birds seen many times in these forests were Wilson's Thrush, Chickadee, Black-throated Green Warbler and Canada Jay. The characteristic birds of the spruce and balsam forests were the Chickadee, Eed-breasted Nuthatch, Golden-crowned Kinglet, Magnolia Warbler, Canada Jay and Wilson's Thrush. The Magnolia Warbler seemed to prefer places where the trees were not very high, for on the small rocky knolls which were covered with stunted spruce and balsam, this bird was more numerous than elsewhere. ECOLOGY OF ISLE ROYALE. 97 THE FALL MIGRATION OF BIRDS AT WASHINGTON HARBOR, ISLE ROYALE, IN 1905. BY MAX MINOR PEET. I. INTEODUCTIOX. Our observaitions of the fall migrations of birdf? at Washington Har- bor extended over the period from August 18 to September 22. A hasty examination was made of the bird life here before migration had really set in (August 5 to 8), and the observations gathered at this time, together with the records obtained the previous year, gave us an insight into the conditions existing thei*e. This was important, as migration had commenced while we were still at Siskowit Bay. Isle Royale is situated about fifteen miles from the north shore of Lake Superior, and lies nearly northeast by southwest. Situated as it is several miles from the north shore and with an unbroken stretch of water 100 miles across lying south of it, the island makes an ex- cellent point for the migrants to stop before crossing the lake. The birds seemed to center at Washington Harbor as if focussed there from the north shore, and in all probability the birds observed there repre- sented the avian life of many square miles on the mainland. Records were kept of the species seen each day and are given in tabulated form at the end of this paper. II. THE ENVIEONIXENT. 1. The Clearing. Under the head of clearings, we include the three artificial clearings and the narrow roads connecting them. The first of these was situated on the shore of Washington Harbor, near its head and close to the mouth of Washington River. The trees had been en- tirely cleared away over an area of several acres, making a rectangular clearing which had been seeded to timothy and short grasses. The waves had cut away the soil along the shore leaving a nearly vertical bank two or three feet high, in some places overhanging the water. The land gradually rises from the water's edge, more rapidly at the south- ern end where a low bluff is formed. On this bluff the Club-house stands, and beloAA', nearer the lake, is a little group of four small houses, the largest of which we used as a camp. Other buildings were also located in this clearing. Part of the clearing was overgrown with brush and small trees. These had been burned and the debris left where it fell. Many small bushes, weeds, and vines sprang up among the fallen logs and branches, forming on ideal retreat for the smaller birds such as warblers and sparrows. The rank growth of the vegeta- tion made it almlost impossible to penetrate any distance into it. Here the Lincoln Sparrows were most abundant during their migration. As the soil was very shallow, the timothy gi-ew short and scattering, and probably furnished little protection for the birds, as it was cut 13 98 MICHIGAN SURVEY, 1908. about the middle of August. Xear the road leading to the second clearing to the north was a small spot cleared for a garden. This bare ground was the favorite feeding place of the Horned Larks. On the short, steep slope which skirted the road to the second clearing, thirty or forty stumps had been left. These were the favorite perches for the Sparrow and Sharp-shinned Hawks, and the tops of many of them were covered with the harder portions of grasshoppers, these insects forming one of the principal foods of these birds. Three nar- row roads or trails left this clearing, one to the second clearing, one to the head of Siskowit Bay, and the third to Lake Desor. The first of these was kept open and had originallj' been much wider than at ])resent, being narrowed by a fringe of alders, birches, and small bushes together with young balsams and spruces. The second clearing, consisting of 3 acres, was divided into two parts,, a grassy tract and a garden in which potatoes, carrots, etc., were grown. From this a road (Fig. 56) led to the third clearing, called Wendigo, which was about the same size, and contained two old log houses and two or three decaying sheds. Few of the stumps had been removed and hawks used them as per- ches. The ground was overgrown with short, nearly dead grass. White-footed mice were abundant in these clearings after nightfall,, and many Northern Hares were seen along the roads just at dusk. A narrow road wound past the clearing and off along the base of the bluffs for a mile or more to several abandoned cuts made by the old mining compa^iy. It was along these roads, which ran approximately . north-east and south-west, that the bulk of the migrants passed. Even during the heavy migration comparatively few birds were observed in the denise forests, although many passed along the river. It has been generally noticed that many birds, the smaller migrants in particular, as the sparrows and warblers, prefer the borders of clearings, and a long narrow road thl'ough heavy timber and bordered by bushes and small trees, appeared to be an ideal place for them. All the clearings were surrounded by the dense, coniferous forest except the first which fronted on the lake. 2. The Forest. The forest may be considered to consist of all that portion which has not been entirely cleared of the native trees. The major part consists of balsam and spruce with a heavy undergrowth of ground hemlock, and in places along the river there are dense thickets of alder. The soil in the depressions is damp, with small pools of water standing on the decayed leaves. Washington River flows through the lower portion of tiie forest. It is a stream sixty or seventy feet wide near its mouth, but it rapidly diminishes in size, so that near Wendigo is not more than five or six feet across. However it becomes much more rapid and the banks are covered with refuse and fallen logs and branches. Few resident birds were found in the dense, dark forest, and still fewer migrants were found there. During very severe weather the Chipping Sparrow sometimes retreated to the protection of the balsams, but it never Avandered far from the open. The White-throated Spar- row was quite common, breeding in the forest along tlie river, and even during migration it was found most abundant in, the underbrush. The ECOLOGY OP ISLE ROYALE. 99 Sharp-shinned and Sparrow Hawks rarely remained here, except dur- ing the night, or in very stormy weather. The warblers were scarcely ever fonnd in the heavy timber, but along the more open part of the river and in the alder thickets they were abundant. By far the most common warbler along the river was Grinnell's Water Thrush. This bird was coniined alm,ost entirely to the forest, and especially to that poi-tion. bordering Ihe stream Avhere fallen logs and rubbish furnished their favorite haunt. They seemed to be migrating in pairs, but no immature birds were seen with them. During the stormy period last- ing from September 2 to 5, the Water Thrush came out into the road and clearings. The Wilson and Olive-backed Thrushes bred in the forest, but during migration they preferred the open and were only occasionally found in the heavy timber. The path skirting the river was also a favorite route for them. The maple brush which bordered the forest in many places was the favorite habitat of the Hermit Thrush. This and the diminutive Winter Wren were sometimes met with among the very densest conifers. Among the other birds occurring here were the Brown Creeper, Golden-crowned Kinglet, and Bed-breasted Nuthatch. Chickadees were nearly always present. This habitat was chosen by nearly all of these birds during migration, probably because it furnished the right kind of food and excellent protection. ;Many other species were occasionally met with, but they were only wanderers and no particular significance can be attributed to tJieir occurrence here. 3. Food. The clearing afforded abundant food for nearly all species; The grassy meadows and dry hillsides were infested with great swarms of grasslio])pers which rose up befoi-e one as a buzzing cloud. Nearly all the birds taken, among which might be mentioned the Sharp-shinned and Sparrow Hawks, Thick-billed Redwing, Busty Grackle, Flicker and Nighthawk, fed to a, greater or less extent on these pests, as was shown by an examination of their stomachs. Many other species of insects were abundant, blackflies, deer flies, and "no-see-ums" being; at times almost unbearable. The Deer Mouse was very plentiful; and also the Xorthern Hare, as many as twenty or thirty of the latter being seen at one time feeding in the road between the first and second clearings. These animals together with the large number of Red Squirrels found along the edge of the road furnished abundant food for the owls and migrating hawks. Seeds were plentiful and constituted the principal food of the Savannah and other sparrows. Wild red raspberry bushes were common and these berries together with several other kinds were greedily eaten by many of the birds, especially the Cedar Waxwings. A few wild flowers grew in the clearing and these were occasionally visited by the Ruby-throated Humming Bird. Insect life characteris- tic of coniferous forests was probably abundant because the Brown Creeper, Chickadee, and Golden-crowned Kinglet fed here almost ex- clusively; otherwise this habitat did not appear to furnish much food for the migrants. III. THE WEATHER CONDITION'S AND MIGRANTS. 1. Weather Conditions. Throughout the period of thirty-flve days during which obsei-vations were made on migration at Washington Har- 100 MICHIGAN SURVEY, 1908. bor, a daily record was kept of the direction of the wind, temperature, and the general weather conditions. The barometric readings, taken at INirt Arthur, thirty-five miles nearly due north from the Harbor, are from the daily weather maps. A comparison of my observations ou the weather with those from Port Arthur shows that the conditions at the two places were much the same, so I feel safe in assuming the barometric pressure at the island to be approximately that recorded just to the north. An examination of the daily weather maps for this period shows that the same isotherms and isobars include both localities. Un- fortunately Port Arthur is the most northern station on the daily Aveather map that could have any appreciable effect on the bird life