EXPERIMENTAL POLLINATION AN OUTLINE OF THE ECOLOGY OF FLOWERS AND INSECTS BY Frederic E. Clements and Frances L. Long QK926 | Q £ Published by the Carnegie Institution op Washington Washington, 1923 Oty* B. 31 Mi ffitbrara Nortlt (Carolina &tatp Inineraitij olina State Library ^["CAROLINA STATE UNIVERSITY L THIS BOOK IS DUE ON THE DATE INDICATED BELOW AND IS SUB- IECT TO AN OVERDUE FINE AS POSTED AT THE CIRCULATION DESK. 19 1980 MAY /yyy CLEMENTS & LONG Edith S. Clements I Painted flowers of Aconitum and Delphinium, showing the method of experiment. North Ca(J$i* S'aLe Library # Raleigh EXPERIMENTAL POLLINATION AN OUTLINE OF THE ECOLOGY OF FLOWERS AND INSECTS BY .aca Frederic E. Clements and Frances L. Long North Q£M!ina State Librarv Published by the Carnegie Institution of Washington Washington, 1923 CARNEGIE INSTITUTION OF WASHINGTON Publication No. 336 Copies of this feoek NOV 6 1923 CONTENTS. 1. Introduction and Methods Introduction Objectives Methods General principles Normal pollination Experimental pollination Organization of experiments Change of position or place Concealing or disguising flowers. . Mutilation Artificial and painted flowers .... Addition of parts or substances . . Competition Manipulation of insects Life-history methods and records . Life-history record of a represen- tative species 2. Normal and Experimental Pol- lination Aconitum columbianum Page 3 3 4 4 4 6 6 7 7 Normal pollination 15 Structure Behavior Experiments Change of position Horizontal racemes Racemes inverted Mutilations 15 15 1(3 16 16 16 17 Cotton plugs 17 Stamens removed 17 Hood split 17 Hood removed 17 Hood and nectaries removed 17 Lower sepals removed 17 Side petals removed 18 Competitive relations 18 Artificial and painted flowers. . 18 Normal colors 18 Artificial flowers 19 Painted flowers 20 Addition of honey and odor. . . 21 Honey 21 Perfumes 22 Summary 22 Delphinium scopulorum 22 Normal pollination 22 Habit and structure 22 Behavior 22 Experiments 23 Change of position 23 Inverted racemes 23 Horizontal racemes 23 Mutilation 23 Cotton plugs 23 Petals removed 23 Spur removed 23 Landing-platform removed. . 24 Artificial and painted flowers . . 24 Crepe-paper corollas 24 Crepe-paper corollas with spurs 24 Painted flowers 25 2. Normal and Experimental Pol- lination {Continued). Delphinium scopulorum (Continued). Addition of odor Perfume Summary Rubus deliciosus Normal pollination Habit and structure Behavior Variation in visits Experiments Mutilation Petals split or shortened .... Stamens covered Artificial and painted flowers . . Crepe-paper corollas Painted corollas Addition of honey or odor .... Honey Summary Rubus strigosus Normal pollination Habit and structure Behavior Experiments Mutilation Floral envelopes or stamens removed Rosa acicularis Normal pollination Habit and structure Behavior Experiments Mutilation Corolla shortened Stamens masked Artificial and painted flowers . . Crepe-paper corollas Addition of honey and odor. . . Honey Honey and talcum powder. . Camphor Summary Geranium caespitosum Normal pollination Habit and structure Behavior Calendars Experiments Change of position Flowers vertical or inverted Mutilation Cotton over nectaries Excision Artificial and painted flowers. . Crepe-paper corollas Addition of honey and odor. . . Honey Summary Chamaenerium angustifolium Normal pollination Habit and structure 25 25 26 26 26 26 26 27 29 29 29 29 30 30 30 31 31 31 32 32 32 32 33 33 33 33 33 33 33 34 34 34 35 35 35 36 IV CONTENTS Page 2. Normal and Experimental Pol- lination (Continued). Chamaenerium angustifolium (Cont.). Behavior 45 Experiments 48 Change of position 48 Racemes inverted 48 Mutilation 48 Floral envelopes removed. . . 48 Upper petals and the in- cluded sepal removed .... 49 Stamens and style removed . 49 Artificial and painted flowers. . 51 Crepe-paper corollas 51 Painted corollas 51 Addition of honey and odor. . . 52 Honey 52 Odor 52 Perfumes 53 Flavoring extracts 53 Summary 54 Pachylophus caespitosus 54 Normal pollination 54 Habit and structure 54 Behavior 55 Experiments 56 Artificial flowers 56 Crepe-paper flowers 56 Petals obscured 56 Mentzelia multiflora 56 Normal pollination 56 Habit and structure 56 Experiments 56 Crepe-paper flowers and com- petition 57 Crepe-paper corollas 57 Summary 58 Frasera speciosa 58 Normal pollination 58 Habit and structure 58 Behavior 58 Experiments 59 Mutilation 59 Types 59 Artificial and painted flowers. . 59 Crepe-paper corollas 59 Painted corollas 60 False corollas 61 Addition of nectar and odor. . . 61 Nectar 61 Summary 61 Mertensia pratensis 62 Normal pollination 62 Behavior 62 Experiments 63 Mutilation 63 Changes of corolla 63 Castilleia miniata 63 Normal pollination 63 Habit and structure 63 Behavior 63 Experiments 64 Mutilation 64 Upper lip removed, spike in- verted 64 Painted flowers 64 Painted bracts 64 Page 2. Normal and Experimental Pol- lination (Continued). Pentstemon glaber 65 Normal pollination 65 Habit and structure 65 Behavior 65 Calendars 66 Experiments 68 Change of position 68 Racemes inverted 68 Racemes horizontal 70 Mutilation 70 Cotton at the corolla mouth. 70 Styles, stamens, and stami- node removed 70 Corolla split 70 Corolla lips separated and staminode raised 71 Upper lip removed at the throat 71 Lower lip shortened half. ... 71 Lower lip removed 71 Lower lip and part of the tube removed 71 Lips removed except lower lobe 71 Corolla tube shortened half. 71 Petals separated 71 Lower lip split into three petals 71 Tube end of a corolla slipped over the staminode 72 Comparative relations 72 Artificial and painted flowers. . . 72 Crepe-paper corollas 72 Corolla painted with water- colors 72 Honey and odor 72 Sirup added 72 Pentstemon gracilis 73 Normal pollination 73 Habit and structure 73 Behavior 73 Calendars 74 Experiments 75 Change of position 75 Racemes inverted 75 Racemes horizontal 75 Mutilation 75 Landing-platform removed. . 75 Brush of staminode removed 75 Anthers and recurved por- tion of upper lip removed 75 Upper lip partly removed. . 75 Lobes of upper lip separated 75 Petals split to the base 76 Pentstemon glaucus 76 Normal pollination 76 Normal behavior 76 Experiments 76 Mutilation 76 Upper lobes split to base ... 76 Lower lip removed 76 Outer lobes of lower lip re- moved 76 Staminode removed 76 CONTENTS Page 2. Normal and Experimental Pol- lination (Continued). Pentstemon glaucus (Cont.). Middle lobe of the lower lip removed 77 Upper lip removed 77 Pentstemon secundiflorus 77 Normal pollination 77 Behavior 77 Calendar 77 Experiments 77 Odor 77 Powders and extracts 77 Cotton wads sprinkled with peppermint 78 Summary 78 Monarda fistulosa 79 Normal pollination 79 Habit and structure 79 Behavior 79 Experiments 80 Mutilation and competition. . . 80 Plan 80 Summary 88 Resume 89 Variation in number of visits. ... 89 Changes of position 90 Masking with cotton 90 Mutilation 91 Artificial flowers 92 Painted flowers 92 Honey and odor 93 3. Competition and Constancy 94 Significance 94 Competition 94 General plan 94 Rubus strigosus 95 General relations 95 Rubus strigosus and Rubus deliciosus 95 Comparison 95 Experiments 95 Summary 97 Rubus strigosus and Rubus deliciosus 97 Comparison 99 Summary 99 Competition of Rubus with Frasera, Cleome, etc 99 Comparison 97 Experiments 99 Summary 99 Rubus strigosus and two or more competitors 99 Comparison 99 Summary 100 Rubus deliciosus 101 Experiments 101 Summary 101 Rosa acicularis 101 Comparison 101 Experiments 102 Summary 102 Competition with normal and mutilated Aquilegia 105 Geranium 105 Comparison 105 Experiments 106 Summary 109 Page 3. Competition and Constancy (Con- tinued) . Chamaenerium angustifolium 110 Comparison HO Experiments HO Summary HO Pentstemon HO Comparison HO Experiments 113 Summary 113 Monarda fistulosa 115 Comparison H5 Experiments 115 Summary 115 Weight and composition of pollen loads 119 Value and methods 119 Discussion 127 Constancy 128 Definitions 128 Early observations of constancy 129 Bennett's studies of constancy 130 Christy's studies of methodic habits 130 Mueller's results 130 Bulman's studies 131 Ord's conclusions 131. Plateau and Perez 131 Constancy in Bombus 132 Lovell's conclusions 132 Kranichfeld's observations 133 Origin of oligotropism 133 Resume 134 Experimental results in compe- tition 134 Effects of competition 134 Constancy as shown by pollen loads 135 4. Principles and Conclusions 136 Introductory 136 Early experiments of Plateau and others 136 Artificial flowers 136 Nectaries 137 Color sense of bees 138 Response to detached petals. ... 139 Color preferences of nocturnal moths 140 Response to color without anten- nae '. 140 Perception of form 140 Response of wasps to color 143 Response of wasps to odor 145 Main researches of Plateau 145 Masked flowers 145 Removal of corolla 147 Response to different colors 148 Addition of honey to vivid nectar- less flowers 149 Anemophilous flowers 150 Entomophilous flowers of dull color 150 Artificial flowers, second series . . . 150 Artificial flowers of green leaves . . 151 Conclusions as to artificial flowers 152 General summary 152 Role of vexillary organs 153 VI CONTENTS Page 4. Principles and Conclusions {Con- tinued), Main researches of Plateau (Con- tinued). Choice of colors by insects 154 Errors made by Anthidium 155 Admiration of syrphids for bright flowers 156 Attraction of colored cloths and brilliant objects 156 Constancy among bees 157 Mistakes made by bees 157 Removal of the antennae of bum- ble-bees 158 Evidence of the attractive role of odor 159 Decorollate poppies and insect visits 159 New experiments with artificial flowers 160 Conclusions as to artificial flowers 162 Macroglossa and false flowers. ... 164 Entomophilous flowers little vis- ited by insects 165 Related studies and critiques 166 Comparative importance of odor and color 166 Perez's critique of Plateau's work 167 Effect of colors at the hive 168 Critiques of Kienitz-Gerloff 168 Knuth's critique 169 Reeker's experiments with arti- ficial flowers 171 Decorollate poppies 171 Response to color and odor by a hawk-moth 175 Vexillary nature of the plume in Muscari 175 Forel's experiments with covered dahlias and with artefacts .... 176 Response of Syritta 178 Andreae's experiments with arti- ficial flowers 178 Andreae's conclusions 181 Andreae's criticisms of Plateau's work 182 Plateau's criticisms of Andreae's work 182 Wery's experiments with decorol- late and artificial flowers 183 Plateau's criticisms of Wery's experiments 185 Experiments of Weismann and Errera 186 Orientation of the honey-bee at flowers of the same species .... 186 Discrimination between similar species of flowers 187 Orientation of the bee within the flower 188 Chance observations of visits to imitations 189 Knoll's critique of Plateau's study of Macroglossa 193 Recent investigations 193 The color sense of the honey-bee 193 Can bees distinguish colors? .... 194 Page 4. Principles and Conclusions (Con- tinued) . Recent investigations (Continued). The pollination of green flowers . . 195 Conspicuous flowers rarely visited by insects 196 Response of honey-bees to colored artefacts 197 Pattern vision in the honey-bee. . 198 Experiments with cotton blos- soms 198 Color sense and memory in the honey-bee 199 Frisch's researches — sense of color and form in the honey-bee .... 200 The supposed color sense of the honey-bee 201 Seasonal changes in response to honey 202 The sense of smell in the honey- bee 203 Bombylius and the colors of flowers 205 Vision and flower behavior of Macroglossa stellatarum 206 Response of bees to spectral bands 207 Evaluation of Plateau's researches 208 Forel's estimate 208 Contradictory nature of Plateau's later conclusions 212 Conclusions as to Plateau's views 213 Senses of insects 213 Sight 213 The mosaic theory 213 Criticisms of Plateau's views as to vision 215 Sensibility to color 215 Perception of form and move- ment by insects 215 Discrimination of form 216 Vision in honey-bees 216 Vision in ants 217 The role of ultra-violet in at- traction 217 The homing faculty in bees and wasps 217 Fabre's experiment with bees and wasps 217 Lubbock's and Forel's critique of Fabre's conclusions 218 Homing faculty of bees 220 Sense of direction in ground- wasps 220 Disturbance of memory in wasps 221 Observations on the homing of Bembex and Pampilus 221 Memory of place in Osmia. . . . 222 The field and nest flights of the bumble-bee 223 The homing of the mud-dauber wasp 223 Experiments on the orientation of bees in homing 224 Homing abilitv in Polistes .... 224 Smell 225 CONTENTS Vll Page 4. Principles and Conclusions (Con- tinued). Smell (Continued) . Sense of smell in insects de- prived of antennae 225 Forel's criticisms of Graber's results 226 Hauser's experiments 227 Olfactory pores 228 Experiments with antennse removed, mutilated, or coated 228 Experiments with wings, legs, and stings mutilated 229 Mclndoo's experiments with deantennate insects 229 Present status as to the seat of the olfactory sense 230 Intelligence 232 Relation between the senses and mental faculties of in- sects 232 Memory and general intelli- gence of wasps 233 Memory of place in bees 233 Memory of time and asso- ciation of impressions. ..... 235 Memory of time and memory association in honey-bees. . . 236 Intelligence of honey-bees. . . . 236 The psychic powers of insects 237 Page 4. Principles and Conclusions (Con- tinued). General resumS 238 Treatment 238 Attraction 238 Color 238 Mutilation 238 Artificial flowers 239 Painted flowers 240 Inclosing flowers in glass 241 Green flowers and showy nec- tarless flowers 241 Color preference 241 Odor 244 Masking or covering flowers to conceal color 244 Odor of honey 244 Effect of added odors 245 Relative value of color and odor 245 Form 245 Distinction and role 245 Attraction at a distance and near at hand 246 Learning and habit 247 Memory and intelligence 247 5. Pollinators and Flowers Visited 249 6. Flowers and their Visitors 256 Bibliography 268 Index 273 Description of Plates 276 EXPERIMENTAL POLLINATION AN OUTLINE OF THE ECOLOGY OF FLOWERS AND INSECTS BY Frederic E. Clements and Frances L. Long LIST OF PLATES. Plate Page 1. Painted flowers of Aeonitum and Delphinium Frontispiece 2. Life-history of the flowers of Aeonitum columbianum and Delphinium scopulorum, 272 3. Life-history of Rubus strigosus and Potentilla gracilis 272 4. Life-history of Heracleum lanatum and Sedum stenopelatum 272 5. Life-history of Galium boreale and Saxifraga bronchialis 272 6. Life-history of Campanula rotundifolia and Erysimum asperum 272 7. Life-history of Geranium caespitosum and Dodecatheon meadia 272 8. Life-history of Chamaenerium angustijolium 273 9. Life-history of Pachylophus caespitosus 273 10. Life-history of Pirola elliptica, Frasera speciosa, and Gentiana amarella 273 11. Life-history of Gilia aggregata and pinnatifida, Mertensia sibirica, and Lithosper- mum multiflorum 273 12. Life-history of Pentstemon glaber 273 13. Life-history of Castilleia miniata and Monarda fistulosa 273 14. Life-history of Allium cernuum and Zygadenus elegans 274 15. Mutilated and inverted flowers of Aeonitum 274 16. Mutilated flowers of Aeonitum, Delphinium, and Monarda 274 17. Mutilated flowers of Geranium, Chamaenerium, and Pentstemon glaber 274 2 1. INTRODUCTION AND METHODS. INTRODUCTION. The original plan for the development of quantitative ecology con- templated two series of monographs, one dealing with the plant in its vegetative relations, the other in its reproductive aspects. The first has received attention in a number of monographs, some of which deal with the individual and some with the community, but the factors and processes involved in the behavior of the flower were necessarily somewhat neglected for a time. Studies of the life-histories of flowers were first begun in 1910 and these were followed by observations and experiments upon the relations of flowers and insects. These served to disclose the nature of the problem and to indicate the methods needed for a comprehensive experimental attack upon it. The main investigation was begun in 1918 and has been carried on actively during the succeeding summers. As a consequence the limits of the field have expanded greatly and the present treatment is to be regarded as a preliminary endeavor to organize it upon an adequate experimental and quantitative basis in nature. The study of the relations between flowers and insects, begun effectively by Kolreuter (1761) and Sprengel (1793), underwent an enormous expansion at the hands of Delpino (1867), Hildebrand (1867), Mueller (1873), Darwin (1876), Kerner (1876), and Knuth (1894) without becoming experimental in even a small degree. The first investigator to recognize that the great mass of observational results needed to be refined by means of experimental methods was Plateau (1877, 1895), and practically all other experimental studies have been a direct or indirect consequence of his work.1 The present investigation constitutes the exception, as it was begun with a different objective and with contacts sufficiently slight to permit an independent development of methods. This was of especial importance in view of the comprehensive nature of the plan, as well as in affording a detached view of the methods already employed. This appears to have been justified by the outcome, not only in the matter of methods but also with respect to results and conclusions. Moreover, the difference in objectives has made it possible to plan experiments and interpret results without a bias in favor of Plateau's views or those of his critics. While the relative values of color and odor in attraction constitute one of the most interesting phases of the general problem, they have too long held the center of the stage as a consequence of Plateau's dramatic challenge. However, even in the present case, the adequate analysis and evaluation of the work that has been done make it necessary to treat attraction as the paramount theme. This investigation has been carried out at the Alpine Laboratory, which is situated at an altitude of 8,500 feet in the montane forest climax on Pike's Peak. It is perhaps unique in dealing with an insect fauna in exclu- sive contact with a native flora, though some of the experiments have 1 Since this was written, the admirable monographs of Frisch and of Knoll have become available. These deal almost exclusively with intensive control researches into the response of one or two species to color, odor, or form, and are indispensable to students of experimental pollination. While it is now impossible to abstract them as fully as they deserve, a general account of the methods and results will be found in Chapter 4. 3 4 EXPERIMENTAL POLLINATION. concerned species that are probably outside the normal experience of the visitors. Some studies have also been made in the plains grassland at the base of the range, and in the alpine meadows on the summit of Pike's Peak, as well as a few preliminary ones in Arizona and California, but these are reserved for the most part for a later treatment. For the flowers the nomenclature is that employed in Clements and Clements' "Rocky Mountain Flowers," while the names of the Lepidoptera are those found in Holland's "Butterfly Book" and "Moth Book." In the other groups the determinations have been made by the following specialists, to whom grateful acknowledgment is made: Dr. J. M. Aldrich, Bureau of Entomology, Washington, D. C; Professor T. D. A. Cockerell, University of Colorado, Boulder, Colorado; Professor C. Howard Curran, University of Kansas, Lawrence, Kansas; Dr. F. E. Lutz, American Museum of Natural History, New York; Professor S. A. Rohwer, U. S. National Museum, Washington, D. C; Professor 0. W. Oestlund, University of Minnesota, Minneapolis, Minnesota; Professor M. W. Swenk, University of Nebraska, Lincoln, Nebraska; Professor C. T. Vorhies, University of Arizona, Tucson, Arizona; Professor H. L. Viereck, Biological Survey, Washington, D. C. Objectives. — As already indicated, the primary object of the present investigation is to place the study of the flower in relation to its environ- ment on the basis of experiment and measurement. At the same time it is intended to give the fullest value to the synthetic nature of the problem by placing the chief emphasis upon the mutual relations of flowers and insects. The life-history of the flower has been given greater attention than heretofore and it is felt that the life-history of the insect in the broader ecological sense must receive similar study. The central theme is attraction and the behavior of the insect at the flower, and in spite of the work already done, this offers an enormous opportunity for quantitative research. The efficiency of both flower and insect is susceptible of much greater accuracy of measurement and it appears probable that this will disclose a new field of correlations. In this connection the experimental study of competition is especially significant and promises to throw a flood of light upon reciprocal adaptation, as is likewise true of mutilation experiments. This leads to the consideration of the evolution and phylogeny of flowering plants under the influence of insect and wind pollination, as well as a re-examination of the effectiveness of cross and self pollination. With respect to the insects the major queries concern the respective roles of the senses, the acquisition and fixity of habits, and the relation between instinct and intelligence. METHODS. General principles.— An endeavor has been made to develop a compre- hensive system of research, based primarily upon experiment and measure- ment. This has been made as complete as possible, though it is fully recognized that the further development of this great field will reveal new objectives and new methods of attack. The great majority of the methods have been tested in the present study, others are now being used in the work under way, and a few are still to be applied. Most of the latter have been employed by other investigators and hence are considered for the sake of INTRODUCTION AND METHODS. 5 completeness. The observational method has necessarily been continued in connection with normal pollination and the life-history of the flower, but it has been made as quantitative as possible and has frequently been supplemented by experiment. The experimental methods have been made as simple and direct as possible and have been applied chiefly to attraction, behavior at the flower, and competition between flowers in this first series. Most of them have been subjected to repeated check and the results are thought to be fairly conclusive for the region and the time concerned. The conditions have differed materially from those obtaining in practically all other studies in this field, in that both flowers and insects were in the natural relations that have existed for a long period. While gardens yield behavior results of as much interest as those of natural habitats, it is obvious that their essentially exotic nature renders them of little value in questions of adaptation and evolution. This difference is also to be taken into account in other respects; for example, artificial flowers appear to be much more readily visited in gardens than in nature, as would be expected from the difference in the habits of the insects. Moreover, it is thought that the experimental investigation of a native insect fauna in the midst of its natural floral environment, and the con- verse, furnishes a norm to which can be referred other studies that are artificial in some degree. Some such standard appears indispensable, since it is evident that many of the contradictions and discrepancies in the results of various investigators are to be explained by differences in con- ditions and setting rather than by faults of method or observation. This is certainly true of many of the points at issue between Plateau and his critics. The effects of time, place, weather, grouping, etc., are often decisive, as shown in the next chapter, and they need always to be checked by actual trial, or, much better, eliminated by simultaneous observations in contiguous areas in so far as possible. The time of day, week, or season not only has an effect due to lapsed time, but also one especially of differ- ence in sun, cloud, temperature, wind, condition of flowers, habit of insects, etc. Even the weather of the previous day may have a profound effect, if it has been rainy, unusually cold or warm, or windy. Differences of location and particularly region usually produce decided effects, owing to changes of conditions as well as of flower and insect populations. In fact, marked differences of behavior have been noted in spots a yard apart where no differences of sunlight, temperature, or wind were demonstrable at the time, but where the behavior of the insects had been determined by earlier shade, exposure, by nearness to their nests, etc. The kinds, number, and grouping of the species and individual flowers naturally have a pro- nounced effect, as do also the time of blooming, the position in the flowering period, the nectar flow, and the rate of nectar production. It is equally evident that the composition of the insect fauna as to orders and species, the number of individuals, the sexes, social habits, age, etc., will greatly affect the results. Finally, it has been found that the position and nearness of the observer, as well as his clothing (cf. Lovell, 1914:407), produce effects to be taken into account, while an increase in the number of observers in the same spot may completely change the response. Still other factors enter into the problem, thus completing the certainty that researches 6 EXPERIMENTAL POLLINATION. widely removed in time and space will differ much in detail and often in principle as well. Normal pollination. — This is primarily a matter of observation and is important in experimental studies chiefly because it furnishes a back- ground of normal behavior against which departures may be measured. The methods are simple and have long been exemplified in the works of Delpino, Hildebrand, Axell, Mueller, Darwin, Kerner, Knuth, Loew, MacLeod, Robertson, and others. For the more exact results needed in connection with experiments, it has proved necessary to enter into finer details as to behavior, to deal with a larger number of individuals, and especially to record the number of visitors of each species, as well as the number of flowers visited by each individual. The failure to note the number of visitors of each species deprives practically all of the observa- tional studies of any quantitative value, in spite of the contention of Knuth (1906:195) that the "statistical" method of Mueller has greater possi- bilities than one would be inclined to believe and that the reproach of affording an inaccurate idea of the number of pollinators because it counts the visits of species and not of individuals is of no importance. This is contradicted by the earlier statement that 'a disadvantage is involved, though one that can hardly be avoided, as it is almost impossible to count all the individual visits to a conspicuous flower.' In a critique of Knuth's "Bliitenbiologie," Roberston (1922:148) states that anthecological data "are lists of insect visitors made to show the species, their frequency, their efficiency as pollinators, and the possibility of their having some influence in determining the characters of the flowers. Mueller's lists show these details. In the case of the bees he indicated the sexes, and whether they were sucking nectar or collecting pollen. To note the sexes is important, because female bees fly longer than males and are more likely to make repeated visits. To note the fact of pollen- collecting is also important. A female bee will carry pollen all day from flowers on which the male rarely occurs. From observations at Carlinville the females of nest-making bees average 20.6 visits to the males 10.3. The inquiline bees show females 8.8 to males 8.0. In anthecology Mueller's lists are valuable as regards species and visits, but they fail to indicate the frequency. In 1908 I rejected Mueller's method and adopted the practice of capturing the individuals as they came, noting species and counting individuals It is impossible to indicate the importance of insects to flowers by lists of species, because efforts to increase the lists involve an exaggera- tion of the importance of rare and exceptional cases." Experimental pollination. — This includes practically all the experi- mental methods that deal with the relations of flowers and insects, though competitive methods are considered separately for the sake of convenience. No sharp line can be drawn between these and the methods that make use of colored objects or odorous substances, but it seems better to consider the latter in a separate section. Experiments may be devised to show the role of different parts in attraction, landing, or guidance, the behavior of insects in securing nectar or pollen, or their efficiency in the transfer of pollen. However, in many cases two or more of these processes are affected INTRODUCTION AND METHODS. 7 by the same change, and for this reason the various experimental methods are organized with respect to the change concerned. There is almost no limit to the number of changes and combinations that ingenuity can devise, but the following discussion is restricted essentially to changes that have been used or are now in process of being used. Organization of experiments. — All experiments have been made in the field under natural conditions. Controls have been regularly employed and the results checked by repetition. Comparative values have been secured by using a definite number of flowers, an equal area, or an equal number of them. The latter method alone seems to be entirely without error and has completely replaced the others, since it yields exact and directly comparable expressions of choice. For the same reasons experi- ments should be observed at similar times and for periods of the same length in so far as possible. For the best results it is often necessary to have two or three observers, so that the time difference can be eliminated and the same group of competing insects followed under the same condi- tions. The grouping of the plants and the relative position of the flowers are matters of much importance and must be considered with reference to the habits of the insects and the type of behavior to be tested. All experi- ments carried on with a plant or group that insects have been in the habit of visiting must reckon with the effect of the habit itself. While such results are dependable as to behavior, they do not permit an exact analysis of the factors entering into it. On the other hand, advantage may be taken of habit to insure certainty in regard to a response, as when artificial or mutilated flowers are alternated in the cluster with normal ones. It must also be recognized that the reactions of mature insects contain a large element of habit and that the real response to certain stimuli can be obtained only by using young ones that have just emerged. Moreover, it has become more and more desirable to deal adequately with individual behavior, and this can be done only by working with marked bees. Plateau has more than once pointed out that experiments with flowers and insects demand the greatest patience and almost unlimited time, and that one must expect to have many of them rendered incomplete by changes of weather and other hazards. The problem is the same as in all experi- mentation in nature and must be met in so far as possible in the choice of the region, the detailed organization of plans, and the introduction of the maximum degree of control. The modifications brought about for experimental purposes may be grouped as follows: (1) changes of place, grouping, or time; (2) concealing or disguising flowers or clusters; (3) removal or mutilation of flowers, parts, markings, etc.; (4) artificial or painted flowers or parts; (5) addition of flowers, parts, substances, etc. ; (6) combinations of two or more changes. The distinctions between the different types of modification are far from absolute and certain changes might well be placed in another group. Change of position or place. — These may concern the plants or flowers of one species, or of two or more species. In the latter case they have to do primarily with competition and are considered later under that heading. Changes that have to do with the time of blooming or the grouping may be 8 EXPERIMENTAL POLLINATION. utilized to hasten or retard flowering or to increase the total attraction of a group of individuals. Changes of position are especially valuable in connection with the study of normal behavior and in determining the correlation between habit and intelligence in different species and indi- viduals. They may deal with the cluster, the flower, or the flower part, but in the last case the change is essentially a mutilation and is considered as such. The simplest method is to change the entire cluster in position by 90 or 180 degrees, either by bending and fastening it in the position desired, or by cutting it off, placing the end in a vial or in wet cotton, and attaching it alongside a normal inflorescence. The use of single flowers permits a wider range of changes, as these can be turned through two circles with the assumption of very different positions. Such changes not only affect the appearance of the flower with respect to attraction, but they necessitate a different behavior in one or more of the successive processes of landing, guidance, obtaining nectar, collecting pollen, and departure. A completed visit thus becomes a new problem in the solution of which species and individuals exhibit striking differences. Concealing or disguising flowers. — This may operate upon the plant or cluster, a single flower, or a part of it, such as the petals or stamen-mass. When the nectary is obstructed by a cotton plug, the effect is essentially one of mutilation, while the use of green leaves to mask the corolla or the disguising of the rays of one species with those of another produces an artificial flower in effect. Masking is primarily a device to conceal the color and thus permit the determination of the role of odor, but it must always be done with the effect of habit in mind. Clusters or single flowers may be covered with pots, boxes, wire cages, or other objects, or they may be variously disguised by means of paper, cloth, leaves, etc. The perianth may be similarly covered on either one or both faces, one or more of the petals covered, the anthers or nectaries masked with paper, cotton, or foreign petals; in short, any part or parts may be concealed in any manner that seems desirable. This is similarly true of the rays and disks of com- posites, and of all vexillary organs, such as spathes, colored bracts, etc. Painting the petals or other parts is also a type of masking. Finally, odor may also be masked by means of glass globes, tubes, etc., permitting color and form to act alone as attractive factors. Mutilation. — By this is understood the removal of flowers or parts, the splitting or cutting of parts, and such changes of position as result in a different form. The removal of flowers is chiefly significant in such definite inflorescences as the umbel and head with show-flowers or ray- flowers, in which all or part of either kind of flower may be cut out to dis- close the role of the other. With respect to other changes a radiate head resembles a single flower to a considerable degree. In regular flowers the most important mutilations arise from the removal of corolla or perianth in whole or in part, the shortening of the petals, or splitting them into parts of various forms. The mutilation of irregular flowers may also be made to throw light upon attraction, but it is particularly valuable in revealing the role of the specialized parts, as in the larkspur, peas, and mints. Hoods, spurs, standards, keels, and lips may be removed wholly or partly, or they INTRODUCTION AND METHODS. 9 may be variously split to increase the attractive surface. In the interior of the flower any or all of the stamens, staminodes, or pistils may be removed or modified, one or more of the nectaries excised or otherwise changed, and the protective hairs of various sorts trimmed or cut as desired. All vexillary organs external to the flower may be treated in the same manner as petals. Decisive changes in form may be effected by bringing petals together or turning them back, or by treating them to produce an artificial zygomorphy, while the internal arrangement of the flower may be modified by changing the position of stamens, staminode, style, scales, etc. It is obvious that the nectar may be completely removed and the odor also modified in various ways. Two or more mutilations may be combined in the same flower or progressive mutilation may be carried out in a series, ranging from normal flowers through those with more and more parts removed until the pedicel alone is left. Finally, mutilation may be applied to the guide lines, stripes, grooves, etc., but these are usually best modified by masking them with water-colors. Artificial and painted flowers. — Between the purely artificial flower at one end of the series and the painted natural flower at the other lie many forms, which differ chiefly in the degree to which artificial or foreign materials are used. Artificial flowers proper may be made of paper, cloth, wax, or other materials, and may be either crude or accurate copies of natural flowers, according to the purpose intended. Plateau has raised many objections to those used by his critics (p. 163), but these seem to have little weight (p. 239). Flowers with one or more artificial parts are termed composites and usually consist of the natural center of a flower or head supplied with artificial petals or rays. In some cases the entire flower or head is used and accessory colored parts added. Artificial stamens, staminodes, or pistils may be added in special cases to replace the natural ones, but such uses are limited. Imitations may be made of green leaves, with or without natural centers, and they are also fashioned by using the centers or disks of one species with the petals or rays of another. One modification of particular value consists in replacing the nectaries or anthers of one species with those of another. The best results have been obtained with natural flowers painted with water-colors, since these are artificial only in color. Such paints may also be employed to mask stripes and spots or to supply new markings to test the directive value of the guide lines. In certain cases natural flowers may be killed by the vapors of osmic acid or otherwise, or they may be used in the dried form when the petals or rays are papery in texture. Bits of colored paper or cloth, or detached petals, have something of the value of artificial flowers, but belong properly in the category of colored objects that can be employed to test color vision. Addition of parts or substances. — Additional parts, such as petals, rays, stamens, etc., may be supplied from flowers of the same species or from those of different species. Perhaps the most interesting change of this kind is where the number of nectaries is doubled, and especially when those of another species are alternated. In flowers where the nectar accumulates in considerable amounts in tube or spur, it may be withdrawn and exchanged with that of a different species. Pollen may similarly be 10 EXPERIMENTAL POLLINATION. transposed or the same result obtained by the exchange of stamens. In the great majority of cases, however, addition deals with honey or sugar solutions on the one hand or odorous substances on the other in order to determine the role in attraction. The results with odors depend largely upon whether these are natural ones to which the insects are accustomed to respond and the best method is to employ fragrant flowers or parts regularly visited. Competition. — Competition is regarded as natural when plants of two or more species grow so close or intermingled that their flowers compete for the same group of visitors. It is brought about artificially when plants, clusters, or flowers are transferred in such manner as to result in com- petition. The distinction practically disappears when individuals are transplanted or seeds sown in such a way as to form a competition group. Natural competition groups are also constituted when the flowering period of one species is retarded by pruning or cutting back, or accelerated in various ways to cause it to overlap in some degree the period of an associated species. However, when one or more species are transplanted to a different climax or region, the resulting group is more or less artificial in its relations. As a rule, the simplest method is by the transfer of inflorescences or single flowers, which are kept fresh in bottles of water or by means of wet cotton. Clusters have the advantage in saving time and effort and in exerting a stronger attraction for visitors. The best results are obtained when two species are employed reciprocally as bouquet and plant at the same time, but this demands two observers. Mixed bouquets of two or more species or separate bouquets of the same often give good results with a single observer. When visitors are not too abundant, as many as a half-dozen species may be followed at one time if the flowers are close together and not too numerous. Most mutilation experiments are essentially studies of competition between normal and mutilated flowers and it is often profitable to combine these with competition tests between normal flowers of several species. It is especially desirable to have the number of flowers or heads the same for each competitor, and this is secured by basing the number in bouquet or cluster upon that in the group to be used, or by removing flowers to the number desired. Since the standard or species in the natural position is regularly favored in consequence of the habit of the visitors, it is desirable to scatter the competitors through the group, as they may other- wise remain unnoticed. Manipulation of insects. — As has been indicated, it is felt that the greatest advance in the study of insect behavior can now be made by dealing with individuals. This not only permits greater accuracy in organizing the results for orders, genera, and species, but it also opens up a new and fertile field scarcely touched as yet. A prerequisite for such work is a simple and rapid method of catching and marking individuals, such as the one devised by Giltay (1906:468). While this will demand still more time and patience, the gain in detail and accuracy over present methods will be as great as that secured by replacing lists of species by a record of visitors and visits. In fact, the actual number of visitors, espe- cially in terms of flight from the nest or hive, can be determined in no other INTRODUCTION AND METHODS. 11 way. Many new facts are revealed as to flight, speed of working, constancy, seasonal adjustment in relation to changing maxima of flowering, and so forth. The analysis of behavior rests upon three factors — instinct, habit, and individual adjustment. Instinct is here regarded as fixed habit, and can be largely evaluated by investigating the comparative behavior of related species and genera. It is probable that habit is constantly passing over into instinct, as seems well illustrated by the perennial adjustment made by groups of species and individuals in the particular floral environment to which their round of activities is restricted. As both observation and experiment have shown, visits to flowers are largely determined by habit, and it is impossible to secure conclusive evidence as to the senses and mental powers of insects without eliminating this actor. The best, though hardly the simplest, way of doing this is to base all studies of attraction, for example, upon the use of individuals that have just hatched and hence have had no opportunity to form habits. This demands the location and control of nests and the marking of individuals as they emerge. It can be done in a less exact manner by transferring nests or hives to a different climax, as from the plains at the foot of Pike's Peak to the montane or alpine zone, but even here marking or cages must be employed for accurate results, except where a species peculiar to the plains is used. Pollination cages afford the best means of complete control, but those so far employed have separated nest and flower group, with the result that the caged bees finally became panic-stricken. Cages several meters long and high enough to accommodate an observer when seated, into which nests are introduced before the young emerge, furnish an almost ideal installation for the study of initial responses to color, form, and odor and the gradual fixation of habits. These can then be removed with their occupants to a totally different group of species, or the flowers to which they are accus- tomed can be mutilated to call forth new responses and the consequent adjustment of habits to new conditions. Here, as in all experimental ecology, the basic problem is to secure laboratory control under field conditions, and the pollination cage appears much the best solution. However, there will always remain certain experiments that can be carried on best or solely in the laboratory, where maximum control, uniformity, and accuracy can be secured. The requisite technique has been so carefully developed by Frisch (1914, 1919) and Knoll (1921, 1922) that their methods will serve as the point of departure for all such work in this field (cf. also Porsch, 1922:485). The most conclusive evidence as to the role of color in attraction has been furnished by insects with the antennae coated or amputated, even Plateau admitting its cogency. The relatively small number of experi- ments made with anthophilous insects indicate the desirability of extending such work, and the questions raised by Mclndoo's researches render this imperative. Owing to the injury usually caused by the amputation or excision of the antennae or other organs, the chief task is to discover a substance that will coat them with litt'.e or no injury and that can not be readily removed. All substances that contain alcohol, turpentine, essential oils, etc., must be avoided and the preference given to mixtures of paraffin 12 EXPERIMENTAL POLLINATION. and vaselin with the lowest possible melting-point and corresponding penetration. The individuals must be marked and observed from day to day, and only those utilized which are essentially normal. The tests should be made under natural conditions with materials to which the insect has been accustomed, and should meet the requirement laid down by Forel, namely, that the insect recognize a certain substance and dis- tinguish it from others in a constant and indubitable manner when normal and not when mutilated. The application of such tests to normal insects, those with the antennae coated and those with the olfactory pores cov- ered, should be decisive. The results can be rendered even more decisive by covering the insects' eyes and contrasting the response of insects with either the antennae or the olfactory pores coated to fragrant flowers that are habitually visited. Many other modifications of insects are possible and can be developed as need arises, such as removing the scopa, attach- ing artificial ones, filling the corbiculse with wax, etc. In connection with determinations of efficiency, constancy, etc., it is helpful to stain the pollen of various species with particular dyes and thus simplify reading the pollen record of its behavior. One of the most interesting series of experiments contemplated deals with the reversal of the characteristic habits of diurnal and nocturnal pollinators. Life-history methods and records. — In the endeavor to determine the exact relation of the flower and its behavior to the habitat, simple methods have been devised for following and recording all changes in minute detail. Quite apart from yielding a complete account of the development of the flower, such records have proved indispensable in correlating floral changes with physical factors and insect behavior, as well as in connection with competition and autogamy. The methods are essentially observational, though the subject affords an increasing opportunity for the use of experi- ments in the correlation of flowers or parts. The essential features are: (1) labeling flowers in the order of development; (2) visiting the plants sufficiently often to obtain a detailed record; (3) recording changes on a tabular form that permits ready checking against the preceding observation. The usual plan has been to mark two or three adjacent plants and to follow the development of 10 flowers on each simultaneously. This furnishes an adequate check on individual behavior, and it is practically impossible to follow a larger group when a number of species is concerned. When the buds are sufficiently large, a label is attached to each and the flowers are numbered in the order of their appearance. Ordinary price-tags are employed and the size determined with respect to the flower. In the case of minute flowers, especially those of umbellifers and grasses, the smallest tags are too large for individual flowers, and other devices must be employed. Tags may be placed at every third or fifth flower in large umbels or at corresponding spikelets in panicled grasses, but in the smaller inflorescences this often produces great distortion. Diagrams with the flowers numbered sometimes afford the best solution, while with the smaller radiate heads of composites the rays may be numbered in ink and thus furnish divisions that enable one to follow the disk-flowers accurately. In large heads and umbels and such spikes as those of Phleum, the inflorescence is labeled and INTRODUCTION AND METHODS. 13 threads are used to mark divisions sufficiently small to permit following the florets with accuracy. Visits are regularly made once each day, except during the rapid develop- ment of the warmest days, when morning and afternoon visits are often necessary. Once or twice during each series, records are taken in the early morning, at noon, and in the evening in order to obtain the finest details in the changes. Similarly, one or two visits at sunrise and sunset are necessary to determine the times of opening and closing of many species. The regular visit is made in the morning, preferably at the same hour, though this must often be modified as a result of weather or by other duties. The use of two persons, one to observe and the other to record, effects a great saving of time, but when this is impossible, the entire record should be made by the same individual. The record sheet for each species is ruled to hold the entries for 20 flowers, extra sheets being employed when all the flowers of a head or umbel are to be taken into account. The same form is used for both the field and the final typewritten record and an endeavor is made to enter the observations so that the field sheet can be copied directly. This necessitates a fixed set of abbreviations in order to save both time and space. The entries for each visit are made beneath the preceding one, the space being left blank when no change has occurred, as this permits ready comparison with the last condition. The date and the hour are entered in the first column for each species, the round of visits always being made in the same order to allow the same interval, especially when two or three visits are made in one day. A record is kept of periods of cloudiness and rain, in addition to the usual records of temperature and humidity. The striking differences in the rate of floral development in sun and shade ecads of the same species have been the subject of a special study, in which temperature and humidity were also determined under the forest canopy. Life-history record of a representative species. — The detailed life- histories have been recorded for about 100 species of the Pike's Peak region. For the majority of these this was first done in the summer of 1912 and has been repeated in 1921 and 1922, two simultaneous sets of readings being taken in 1921 by different observers. Because of the limitations of space, the detailed table (table 1) is given for a single species only, and this is restricted to 10 flowers on two different plants, taken from the 1921 observa- tions. The main features of the life-history of 26 species are illustrated in plates 2 to 14, and the stages described in the corresponding legends. 14 EXPERIMENTAL POLLINATION. Table 1. — Aconitum columbianum. Plant I. Plant II. Date. 1 2 3 4 1 5 6 7 1 2 3 July 23, 3 p.m. 02 24, 10 a.m. 0 4 4 p.m. 0 6 25, 10 a.m. op 4 p.m. 1 26,10 a.m. 1 1 4 p.m. . . . 27, 11 a.m. 1 6 5 p.m. . . . 28, 5 p.m. . . . 29, 11 a.m. . . . 30, 11 a.m. . . . 4 p.m. 6 18 31, 11 a.m. 18 4 4 p.m. 4 8t mm 0 2 mm mm 0 4 mm mm 0 6 mm en open as 2 a s ad 2 a d as 3 a s ... 3ad ... 2 as ad 2 a d as las ... lad 4 a s ... 4ad ... 10 a s ... 10 ad ... 8 as ad 8 a d as st rec ad 0 2 mm 0 2 mm 0 2 mm 0 2 mm 0 2 mm open open open open 4 a s 0 4 mm open 0 4 mm 0 6 mm 0 4 mm 0 6 mm 0 4 mm 0 6 mm 2 as 2 ad 10 as 10 ad 8 as Sad 4 a 8 4ad 12 a a 12 ad st rec p falls las lad 9 a s 9ad 2 a s 2 ad 8 as Sad 8 a s 10 a d st rec p falls 8 a s 8 a d 4 a s 4 ad 4 a s 4 ad 4 a s 4ad 8 as Sad st rec p falls 2 a s open 4ad 4 a s 4ad 12 a s 12 a d 10 a s 10 ad st rec 2 ad 2 a s 2 ad 4 a s 4 a d 19 a s 19 a d 6 a s 6 ad las 1 ad 2 a s 2ad 2 a s c enl open open 4 a s 4 a d 4 a s 2 a s 4ad 10 as 10 a d 4 a s 4 ad 3 a s 3 ad las 1 a d 4 a s 4 a d st rec 2 ad 2 a s 2 ad 2 a s 2 ad 10 a s 10 a d 6 a s ad c enl c enl 2, 10 a.m. c < ;nl c brwn ... p falls c brwn p falls 3, 6 a.m. c b c brwn 11 a.m. p f alls p falls Gad 8 a s st rec 4 a s 4 a d 26 a s aborts c brwn p falls pfllg o open, a anther. st stigma. c carpels. Contractions. shedding. rec receptive. brwn turning brownish. d shed. enl enlarging. fllg falling. 2. NORMAL AND EXPERIMENTAL POLLINATION. Treatment. — In the following treatment the species have been arranged in general accordance with their phylogenetic sequence, beginning with the buttercups and terminating with the mints. The discussion of each is divided into two sections, the first dealing with normal pollination, the second with experimental pollination. The experiments are considered under four headings: (1) change of position; (2) mutilation; (3) artificial and painted flowers ; (4) addition of honey and odor. Those that deal with competition and constancy are reserved for the following chapter. Refer- ences are given to the European and American observations on the pol- lination of the same or related species, accompanied by a brief abstract where it seems warranted. Each table or group of tables is summarized in detail and a rSsume' of the general results is given at the end of the chapter. These are further discussed in connection with the conclusions of other investigators in the final resume1 at the end of the fourth chapter. ACONITUM COLUMBIANUM. NORMAL POLLINATION. Structure. — The hood in Aconitum is formed of two colored sepals united, the other two sepals making a landing-platform for insect visitors. Two petals are modified into nectaries, while two form the sides of the hood. The sepals and petals are colored alike, increasing the amount of color in the flowers and making them more easily seen and attractive. The sepals that constitute the landing-platform are smaller than the other sepals and petals, but large enough to support the weight of the visitors. The side petals arch above the stamens and protect them to some extent from the rain, but do not interfere with the access of pollinators. However, they hide the stamens from view when the flowers are seen from certain positions. The nectaries are long and stalked, with a crested hood at the top. To secure nectar, the visitors must have a proboscis 10 mm. long in order to reach through the stalk to the hooded portion containing the nectar (plate 2). Behavior. — The most frequent visitors to Aconitum are Bombus juxtus and bifarius, of which the former is far more frequent, evidently because this species is more numerous in the region. It is larger than bifarius and covers more of the flower. It lands on the two lower sepals, with its head toward the base of the nectaries, and the hind legs curve around the small sepals, while the front pair grasp the side ones. The under part of the thorax rubs back and forth against the anthers and stigmas as the bee sucks nectar. It pushes the proboscis into both nectaries and often stops to brush the pollen from its sides on to its legs, as it leaves the flower. B. juxtus visits the flower for nectar, but in getting this, pollen is brushed from its hairy thorax on the stigmas, and at the same time pollen is dusted on the bee. When sucking nectar, the tip of the abdomen reaches to the three styles. The bee scrapes pollen from its head parts as it leaves the flower. It usually goes from the lower flowers on the raceme to the upper ones and then down again. 15 16 NORMAL AND EXPERIMENTAL POLLINATION. Bombus bifarius lands on the sepal platform with its head above the group of stamens and pointing toward the base of the nectaries. It then moves up to such a position that its thorax is above the stamens and its head is at the base of the nectaries, the hind legs resting upon the two front sepals. As it sucks nectar, its body moves back and forth, thus rubbing the lower side of the thorax and abdomen against the anthers and stigmatic surfaces. None of the North American species of Aconitum has previously been studied with respect to its pollination. In Europe, A. napellus and lycoctonum have received the most attention, and Kronfeld has pointed out that the species of this genus are almost exclusively bumble-bee flowers, the areal limits of Aconitum and Bombus coinciding in a remarkable way. In addition to Bombus, Knuth cites only visits by Macroglossa and Lycaena (1908:50); in the Pike's Peak region the sole visitors are bumble-bees. EXPERIMENTS. CHANGE OF POSITION. Horizontal racemes. — Racemes were placed in a horizontal position and attached to the normal ones by thread. Since the flowers of Aconitum are fastened at various points around the stem, the tip of the hood on some pointed sidewise and in others it pointed up or down. Both Bombus juxtus and bifarius easily went into flowers with the hood pointing down- ward, using the side petals instead of the lower sepals as a landing plat- form. They then secured nectar without taking an uncomfortable position. B. juxtus hovered over the hood in flowers where it pointed sidewise. Finally, it went to the open end where the stamens were exposed, pushed its proboscis about as if exploring, and then found the nectary. The next flower had the tip of the hood pointing up, and B. juxtus landed at once on the side petals and took nectar without any inconvenience. In another case where the hood pointed sidewise, the bee hovered over the spur first, apparently trying to find a place to land. It then went to the side of the flower and attempted to land, but slipped off the edge of the side petal, succeeding only on the second attempt. The next flower was horizontal with the hood directed upward. The bee hovered at the spur, went to the other end, and landed easily. It did not learn by one experience that it could find the nectar at this end when the flower was in the horizontal position. In each case it hovered where it would normally expect to alight, before going to the place where landing was possible. Racemes inverted. — Racemes were cut off and tied to the plant in an inverted position and a piece of moist absorbent cotton was placed around the cut end to prevent wilting. Some individuals of both Bombus juxtus and bifarius found the nectar readily, while in other cases they were frightened away or gave up too soon. The former passed over some flowers without attempting to land, and merely hovered near others. A few individuals started to alight and then flew away as if bothered by the change. Some bees landed at the lower sepals, which now pointed up and occupied the position usually taken by the hood, turned around, and quickly walked into the flower upside down. The next flower was normal and B. juxtus went to this as usual. The third was inverted and the bee ^^ ;. ~ ACONITUM COLUMBIANUM. 17 proceeded exactly as in the first inverted flower, apparently noticing the change of position before landing. One B. bifarius mastered the situation readily and went to five inverted flowers in succession, turning upside down just after alighting at each one. Another B. juxtus landed three times at a group of inverted flowers, tried to push its proboscis into the hood without turning upside down, and in each case failed to find the nectary. MUTILATION. Cotton plugs. — When absorbent cotton was placed in the nectary, Bombus juxtus landed, pushed out its proboscis to find the opening, and tried repeatedly to make a way through the cotton into the nectary. At the next flower it hovered but did not land. When the styles and stamens were hidden by a cotton wad, B. bifarius made the same unsuccessful effort to find the nectary opening. Another individual jerked back as it was about to land, and then quickly flew away to the next flower. A third hovered above the flower and departed without landing, as did several others of both species. Stamens removed. — Bombus juxtus noticed the change as readily as when the flowers were covered with cotton wads, while B. bifarius either hovered above the flower before alighting, or landed directly and flew away at once without projecting the ligule or trying to find the nectar. Hood split. — The hood was split longitudinally in some flowers, thus making them more conspicuous, as it then exhibited a pair of wings ascending behind the nectaries. In some cases this change had no effect upon landing, as it appeared neither to frighten the bees nor to attract them in increased numbers, but in others it greatly increased the attraction. They landed in tHe usual manner, but had a very hard time hanging on to the stamens. Because they were unable to find a suitable position, they often went away without getting nectar. Hood removed. — The hood was removed, leaving the two nectaries to project above the remaining flower-parts. This made the nectaries con- spicuous and changed the general aspect of the flower very much. The effect on the visiting bees was not uniform. Bombus juxtus sometimes hovered above the flowers, but did not land, or stopped without pushing out its ligule. Some exposed the ligule, then noticed the change, and flew away. Others were more persistent, for they landed, at once found the slit down which the nectar runs, and emptied each of the nectaries, repeating this performance at the next flower. Hood and nectaries removed. — An individual of Bombus juxtus alighted as usual, and explored in all directions with its tongue, in the unsuccessful endeavor to find the opening. One adventurous individual crawled between the side petals as if expecting to find the nectary. When it got through, it turned around, crawled over the stamens, and flew away. Lower sepals removed. — Since this is the part of the flower on which Bombus usually rests its hind legs, it found difficulty in balancing properly while seeking the nectary, but finally succeeded in reaching the nectar. 18 NORMAL AND EXPERIMENTAL POLLINATION. Side petals removed. — Some bees evidently noticed the changed appear- ance brought about when these were removed, and hovered above the flowers without landing. However, most of them seemed to observe no change, for they landed and sucked nectar as if the flowers were normal, and even came back a second time. The second pair of legs in these cases was crowded closer to the base of the stamens and rested there. Flowers with side petals removed were also placed in the horizontal position with the hood pointing upward. One bifarius treated these flowers as it did Rosa and tumbled about on the stamens, collecting pollen and making no effort to get nectar. Another followed and did the same thing. This was in marked contrast to all other observations on normal Aconitum, in which Bombus had never made any effort to collect pollen. Ordinarily, the pollen was collected accidentally as the body moved back and forth across the stamens in the act of gathering nectar. Competitive relations. — The five types of mutilation were represented by 5 flowers each, and these were arranged with 25 normal ones (plates 15 and 16). Table 2 gives the results; mere inspections are indicated by an i. Table 2. — Visits to normal and mutilated flowers. Date. Time. Species. Normal . Side petals off. Nectary and hood off. Hood off. Hood split. Cotton over stamens. July 27 July 28 3 to 5 9 to 11 Bombus juxtus. . . . Bombus bifarius. . . Bombus juxtus. . . . 5 35 o 0 2: 16 i 2:3i a: Hi 3 6 14 14 8 32 0 1 i li Total 67 9 9:30i 23 54 2i :95 The number of visits to the mutilated flowers was nearly a half greater than those to the normal, due chiefly to the response to the flowers with the hood split. This greatly increased the extent of the color surface and the attraction even to a greater degree, as each flower received 4 times as many visits as a normal one. Since those with the hood off averaged nearly twice as many visits as the latter, the exposure of the nectaries evidently played a part in the attraction. The other mutilations were visited about the same as the normal flowers, except where cotton was present, this change being noted even in rapid flight. Individual differences in behavior were especially noted in the case of the flowers with the hood and nectaries removed. Three bees inspected these flowers to one that landed, and of the latter some flew away at once, while others extended the ligule several times in the endeavor to find the opening to the nectary. ARTIFICIAL AND PAINTED FLOWERS. Normal colors. — The flowers of Aconitum columbianum are either blue-purple, or white. The plants studied grew along brook-banks and were lighted by sunflecks, which made the purple flowers more conspicuous than the white ones, though the two were equally numerous. The response ACONITUM COLUMBIANUM. 19 of Bombus to the two colors is shown by table 3, which records the number and order of visits, each time given representing a different bee. Table 3. — Visits to white and purple flowers. Time. Speoies. Visits. llh00m a.m 3 W, G P. 3 W. G W. 4 P. 2 W, 1 P, 3 W, 14 P, 3 W, 3 P. 2 P, 2 P, 2 P. 2 W, 9 P, 2 W. 4 W. 11 10