HARVARD UNIVERSITY. LIBRARY OF THE MUSEUM OF COMPARATIVE ZOOLOGY 77/0 a VUQ^-' JQ/yuUXAa^ If, \^i(d- iSxttrlu. ^ (T, iQlT- Bulletin OF THE 1 iLi.raiRv Illinois State Laboratory OF Natural History Urbana, Illinois, U. S. A. VOLUME XII 1915, 1916, 1917 Contributions to the Natural History Survey of Illinois made under the direction of Stephen A. Forbes 1918. c/ ^ ScHNEPP & Barnes. State Printers Springfield, III. 1918. 8040—600 CONTENTS ARTICLE I. THE RELATION OF EVAPORATION AND SOIL MOIS- TURE TO PLANT SUCCESSION IN A RAVINE. BY FRED page THEODORE ULLRICH. (1 Map, 18 Plates) September, 1915 1-16 Introduction 1 Physiography of the McLeish ravine 1 Succession of vegetation 3 Plants of the ravine and the oak uplands 4 Location and description of the stations 6 Measurement of the evaporative power of the air 6 Interpretation and discussion of results 8 Summary and conclusion 12 Measurement of soil moisture 13 Interpretation and discussion of results 14 Conclusion 15 Acknowledgments 16 Literature cited 16 ARTICLE II. A CLASSIFICATION OF THE LEPIDOPTERA BASED ON CHARACTERS OF THE PUPA. BY EDNA MOSHER, PhD. (9 Plates) March, 1916 17-159 Introduction 17 Changes preceding pupation 18 External morphology 22 Classification 30 Analytical table of superfamilies 30 Pupae with functional mandibles 34 Micropterygoidea '. 35 Eriocraniidse 35 Pupae without functional mandibles 37 Generalized pupae without maxillary palpi 37 Hepialoidea 37 Hepialidae 38 Cossoidea 38 Psychidse 39 Cossidae 40 Kucleoidea 41 MegalopygidEe 42 Eucleidse 43 Pyromorphidas 44 Generalized pupae with maxillary palpi 44 Tineoidea 44 Prodoxidse 45 Acrolophidae 46 Tineidffi 47 Heliodinidse 47 Aegerioidea 48 Aegeriidse 49 Tortricoidea 51 Epiblemidae 52 Olethreutidse 54 TortricidEe 56 Sparganothidse 57 Gracilarioidea 58 Nepticulidse 61 Heliozelidse 62 Tischeriidae 63 Bucculatrigidse 64 Lyonetiidae 64 Gracilariidae 65 Phyllocnistidae 68 Specialized pupae with pilif ers 69 Pyralidoidea 69 Pterophoridae 70 Attevidae 71 Pyralididae 72 Papilionoidea 78 Megathymidae 79 Hesperiidae 80 Lycaenidse 83 Papilionid* 85 Pieridfe 87 Nymphalidae 88 Specialized pupae without pilifers 95 Yponomeutoidea 96 Epermeniidae 96 Yponomeutidae 97 Coleophoridse 98 Gelechioidea 98 Lavernidae 99 Scythridaj 100 PAGE Gelechiidae 101 Chrysopeleiidae 104. Oecophoridae 104 Stenomidae 105 Cosmopterygidae 106 Elacliistidae 106 Noctuoidea , 107 Noctuidse 107 Arctiidae H^ Liparidse 121 Bombycoidea 123 Lasiocampidae 123 Bombycidae 124 Notodontoidea 125 Geometridas 126 Notodontidae 132 Dioptidae 134 Sphingoidea 135 Sphingidae 135 Saturnioidea 140 Hemileucidse 142 Ceratocampidae 143 SaturniidaB 144 Phylogeny 147 Acknowledgments 150 Bibliography 152 Plates 153 ARTICLE III. A PRELIMINARY CLASSIFICATION OF DIPTERA, EXCLUSIVE OF PUPIPARA, BASED UPON LARVAL AND PUPAL CHARACTERS, WITH KEYS TO IMAGINES IN CERTAIN FAM- ILIES. PART I. BY J. R. MALLOCH. (30 Plates) March, 1917.161-409 Introduction 161 Acknowledgments 163 Habits and habitats of species 163 Methods of collection and preservation 164 Economic importance of the order 165 Arrangement of families 168 Scope of work 173 Characters of the larvae 173 Suborder Orthorrhapha 173 Division Nematocera 173 Division Brachycera 174 VI PACK Suborder Cyclorrhapha 175 Division Acroptera 175 Division Aschiza 175 Division Schizophora 176 Characters of the pupae 177 Suborder Orthorrhapha 177 Division Nematocera 177 Division Bracliycera 178 Suborder Cyclorrhapha 178 Divisions Acroptera, Ascliiza, and Schizophora 178 Keys to suborders: — Larvas 179 Pupas 179 Imagines 180 Suborder Orthorrhapha 180 Keys to divisions: — Larvee 180 Pupae 180 Imagines 181 Division Nematocera 182 Tabular arrangement of families 182 Keys to families: — Larvae 183 Pupae 185 Imagines 188 Tribe Polyneura 190 Tipuloidea 190 Tipulidce 191 Keys to subfamilies 193 Ctenophorinae 194 Tipulinse 195 Principal papers on North American Tipulidaj 206 Limnobiidae 207 Keys to subfamilies 208 Cylindrotominae 210 Limnobiinge 212 Pediciinge 216 Limnophilinae 220 Rhamphidiinae 226 Eriopterin^ 227 Hexatominae 232 Trichocerinae 234 Papers on the Biology of North American Limnobiidae 238 Ptychopteridae 238 Rhyphidse 241 Vll PV6E Tribe Eucepliala 245 Mycetophiloidea 246 Bolitophilidae 247 Mycetophilidse 248 Keys to subfamilies 251 Mycetophilinse 251 Sciophilinffi 255 Sciaridse 258 Macroceridae 260 Platyuridae 260 Principal papers on North American Mycetophiloidea 262 Culicoide'a 263 Psychodidag • 264 Papers dealing with biology of Psychodidse 273 Blepharoceridas 274 Principal papers on North American Blepharoceridas 276 Culicidas 276 Keys to subfamilies 278 Some of the more important works on North American Culicidse 279 Dixidse 279 Paper containing account of North American Dixidas 280 Chironomoidea 280 CeratopogonidPB 281 Chironomidse 284 Keys to subfamilies 286 Tanypinse 287 Chironominas 287 Important papers on North American Chironomoidea 290 Orphnephilidse 290 Tribe Oligoneura 291 Cecidomyioidea 292 Cecidomyiidffi 293 Important papers on North American Cecidomyiidse 297 Bibionoidea 297 Bibionidffi 298 Important papers on biology of North American species. . . . 300 Scatopsidse 300 Paper on North American Scatopsidae 302 Simuliidae 302 Principal papers on North American Simuliidse 304 Addenda to Nematocera 305 Trichocerinae 306 Division Brachycera 307 Tabular arrangement of families 308 Vlll PAGK Keys to families: — Larvse 308 Pupge 310 Imagines 312 Tribe Platygenya 314 Stratiomyioidea 314 Stratiomyiidag 315 Key to subfamilies 316 Stratiomyiinae 317 Clitellariinas 322 Beridinse 331 Geosarginse 331 Pachygasterin^ ' 334 Xylomyiinae 340 Principal papers on North American Stratiomyiidse 346 Xylophagidse 346 Coenomyiidse 351 Acanthomeridae 354 Tabanoidea 354 Tabanidae 355 Principal papers dealing with the biology of North Amer- ican Tabanidae 361 Leptidae 362 Cyrtoidea 368 Nemestrinidae 368 Cyrtidae 368 Papers on North American Cyrtidae 369 Asiloidea 369 Mydaidae 370 Apioceridse 373 Asilidse 373 Bombyliidae 389 References to Descriptions of larvae and pupae of North American Bombyliidge 395 Therevoidea 396 Therevidse 396 Scenopinidas , 398 Tribe Orthogenya 399 Empididoidea 399 Empididae 400 Papers on the biology of North American Empididae 403 Dolichopodldae 403 References to papers on the biology of North American and European Dolichopodldae • 406 IX ARTICLE IV. THE ZYGOPTERA, OR DAMSEL-FLIES, OF ILLINOIS. page BY PHILIP GARMAN, Ph.D. (16 Plates) June, 1917 411-587 Introduction 411 Acknowledgments 413 Morphology 413 Nymph 413 Adult 423 Life history and habits 438 Egg 438 Nymph 439 Adult 444 History of the Zygoptera 446 Paleontology 446 Ontogeny 450 Phylogenetic comparison of Zygoptera and Anisoptera 453 Egg 454 Nymph 455 Adult 457 Classification 464 ■ Agrionidffi 466 Agrioninee 466 Coenagrionidae 476 Key to subfamilies 476 Lestinae 477 Coenagrioninse 499 Bibliography .... 577 Nymphs 578 Adults 580 Index to genera and species 584 Abbreviations used in lettering plates 586 ERRATA AND ADDENDA Page 5, first column: line 2, for Sheperdia read SJiepherdia ; line 11, for americana read americanus. . Page 9, line 5 from bottom, for XVII read X. Page 42, line 4 from bottom of key, for Pyromorphidae read Eucleidae. Page 73, line 7 from bottom of key, for or read a7Z(Z seldom. Page 100, just below key, insert as follows: — The following species were examined: Lophoptilus eloisella Clemens Laverna hrevivittella Clemens Page 107, lines 8 and 9 from bottom, page 108, line 10 from bottonr, and page 110, line 10, for CuculUanae read CuculUinae. Page 110, line 8 from bottom, dele Polia. Page 112, line 19 from bottom, for Metathoracic read Mesothoracic. Page 129, line 8, for never read sometimes. Page 131, at end of second line insert Paleacrita Riley. Page 158: first column, after Paleacrita, 127, add 131; second column, after Polia dele 110. Page 170, line 4, for Strayiomyiklae read Stratiomyiidae. Page 243, line 2, for alternata read alternatus. Page 307: line 5 from bottom, for wit?i read and; line 16 from bottom, for Homeodactyla read Homoeodactyla. Page 314, line 15 from bottom, for Cecidomyiidae read Coenomyiidae. Page 321, line 12 from bottom, for Stratomyia read Stratiomyia. Page 324, line 6, for pantherina read pantherinus. Page 370, line 18, for Empidoidea read Empididoidea. Page 428, line 8 from bottom, for Mesnotum read Mesonotum. Page 461, line 10 from bottom, for Aeshnide read AesJinidae. Page 478, line 13 from bottom, for vigilas read vigilax. Page 519, line 2, dele side. Page 528, line 10, for caruncluatum read carunculatum. "X^vo Bulletin OF THE Illinois State Laboratory ov Natural History Urbana, Illinois, U. S. A. STEPHEN A. FORBES, Ph.D., L.L.D., Director Vol. XIL September, 1915 Article I. THE RELATION OF EVAPORATION AND SOIL MOISTURE TO PLANT SUCCESSION IN A RAVINE BY Fred Theodore Ullrich Article I. — The Relation of Evaporation and Soil Moisture to Plant Succession in a Ravine. By Fre;d The;odore U1.1.RICH.* Introduction For some time geologists have surmised that differences in the evaporative power of the air and the soil-moisture content account for the succession of vegetation that accompanies the physiographic changes in the development of a ravine. In order to determine whether or not this surmise can be supported by experimental data, this study was made during the summer of 19 13. The ravine selected for investigation was the McLeish ravine, which lies in the south- eastern part of the beautiful village of Glencoe, about sixteen miles north of Chicago. It is mapped in the Chicago folio of the United States Geologi:al Survey as a part of the Evanston-Waukegan re- gion. The general geographic features of this region are the moraine plain or rolling upland, the present shore, and the lake plain with its associated beach ridges. The rolling upland which constitutes the larger part of this area is glacial in origin, consisting of the usual glacial drift, clay, sand, gravel, and boulders, and rises more than sixty feet above the level of the lake. Since the ice sheet retreated, some of the glacial material has been reworked by rivers, waves, and winds, and thus is stratified. The heterogeneity in the composition of the drift and the resulting differences in the resistance to the forces of corrasion, and the great elevation of the upland above the level of the lake have furnished and are furnishing excellent conditions for the development of ravines. Of the many that have been carved in this upland, the McLeish ravine, although one of the smaller, may be considered a good representative. Physiography of the McLeish Ravine (Plate xvni) It would be foreign to this study to give a complete exposition of the geology and physiography of the ravines of this upland region. Such a presentation is given by Atwood and Goldthwait ('08). How- ever, an intelligent appreciation of the method and results of this study requires a brief description of the location and direction of the *Accepterl by the University of Chicago for the degree of Master of Arts in Botany. courses of the ravine and its tributaries, the probable origin of the ravine, and its present physiography. Tlie McLeish ravine is formed by the confluence of two gulhes, the heads of which, if measured in straight distance, are each about 2000 feet inland. From the junc- tion of the two gullies the ravine has an almost easterly direction for about 1500 feet, until it reaches its main tributary from the south, v/hen it continues in a northeasterly direction for about 800 feet to the point where it empties into the lake. From the south, in addition to the main tributary, the ravine receives a number of small gullies. Physiographers have summarized under three heads the succes- sive stages in the cycle of the development of such a ravine as the one here considered : the V-shaped valley, with slopes normally con- vex; the U-shaped valley, a stage of development where detritus descending the slopes is not all carried away by the stream, and con- sequently the valley is being widened faster than it is deepened ; and, finally, the valley with a broad bottom, in which transformation is affected partly by erosion and partly by deposition in the ravine. The tributaries of the McLeish ravine are mostly in the first stage of the cycle, while the main course oi the ravine shows that it is passing from the V-shaped stage into the U-shaped stage. It is altogether likely that this ravine had its infant development accelerated by swampy depressions in the upland near to the lake. Such swamps or marshes within reasonable distance from the head of the cliff of the lake shore would, through seepage, make materials between them and the lake shore so mobile that the head of the gully would work inland with unusual rapidity. The influence of such de- pressions in ravine formation can at present be seen in the vicinity of Glencoe. In the weekly visits to this ravine, the physiographic changes due to weathering, wash, and lateral corrasion were decidedly noticeable. Some of the records made at the time of observation read as follows : ( I ) — during a heavy rain — small streams of clay are flowing leisurely to the bottom of the stream ; ( 2 ) it will be only a matter of a rela- tively short time when two large hard-maple trees will become so undermined by the streams caused by heavy rains as to fall across the channel; (3) a large slump of clay has been deposited at the bottom of the ravine during the preceding week, due to the undermining action of the stream and to the percolating of the water through the soil above the cavity; (4) in following the ravine, from the point where the main tributary enters, to its mouth, a gradation of detritus ranging from boulders and cobblestones to sand is passed (PI. I, Fig. [) ; (5) the waves of the lake usually maintain a bar of shingle and sand across the mouth of the ravine, which results in the formation of a pool of water which gradually filters into the lake as more water is brought down the ravine; (6) after heavy rains the channel is invariably reopened by sweeping the materials which form the barrier into the lake. This gives some conception and appreciation of the force of the current as an active agent in modifying the physiography of the ravine. These statements indicate the dynamic character of the McLeish ravine. Succession of Vegetation After this brief consideration of the geographical and physio- graphical features, we may now turn our attention to the vegetation. Cowles ('oi), in his "Plant Societies of Chicago and Vicinity," calls attention to the fact that the slopes of the embryonic V-shaped ravine, after they attain sufficient stability, develop a carpet of luxuriant vegetation. He says : "In a comparatively few years the vegetation leaps, as it were, by bounds through the herbaceous and shrubby stages into a mesophytic forest, and that, too, a maple forest, the highest type found in our region." This quotation describes admi- rably what has taken place in the greater portion of this ravine and its tributaries. A somewhat detailed examination of the summer and fall flora of the ravine showed that the slopes of the gullies leading into the ravine (PI. I, Fig. 2) were nearly or quite devoid of vege- tation, with the exception of some mosses on the north-facing slopes. A wash or gully near the mouth of the ravine, comparable to the head of a ravine working inland, had rather severe conditions for plant growth in the upper half of its slope. Owing to the extreme exposure to light, wind, and heat, and to the absence of humus in the soil, only a few of the xerophytic pioneers succeeded in establishing themselves. Near the foot of this wash a more or less stable minia- ture plateau (PI. II, Fig. 3) was formed. On this, during the latter part of the summer, there was a great variety and profusion of asters, — one of the most pleasing sights in the ravine at that time. The slopes with mesophytic forest vegetation, with its usual undergrowth, constitute the most general feature of the ravine. This mesophytism in a ravine is regarded by ecologists as only temporary. After the vertical cutting has reached base-level, and as lateral cutting reduces the inclination of the slope, removing the humus and increasing the exposure, there is retrogression from mesophytism to xerophytism. This tendency towards xerophytism is checked through the develop- ment of vegetation, which finally results in the establishment of the permanent mesophytic climax of the region. In the ravine under con- sideration one of the strongest features showing the significance of exposure in the gradual retrogression from mesophytism to xero- phytism was the paucity of tender mesophytic forms and the less luxuriant vegetation in general on the south-facing slopes. In the early spring the north-facing slopes of the ravine were covered more abundantly with the various plants of the vernal flora and certain mosses than the south-facing slopes. Plants of the Ravine and the Oak Uplands The plants of the ravine are recorded under two more or less distinct physiographic areas; the xeromesophytic slopes of the gul- lies (Region i of the table) and the mesophytic slopes of the ravine (Region 2). Some of the plants of the oak uplands (Region 3) are listed, to emphasize the differences between it and the ravine. The vegetation of the oak uplands may be thought of as either antedating or succeeding that of the ravine. The vegetation in these three suc- cessional regions is not entirely distinct, but even a superficial ex- amination of the following list shows striking differences. As this study covers only the period from June 21 to Octt^ber 18, 191 3, much of the vernal vegetation is omitted. In the examination of this list of plants it should be borne in mind that the vegetation of the slopes of the gullies and ravine is noted much more completely than that of the oak uplands ; that the in- dividuals of certain species were generally more numerous on the slopes of the ravine than on the slopes of the gullies ; and that in certain localities of the ravine such plants as Osuiorhiza longistylis (PL III, Fig. 5), Amphicarpa monoica (PI. II, Fig. 4), and Aralia niidicaidis (PI. IV, Fig. 6) had developed pure growths to the ex- clusion of all other species. Furthermore, and most important of all from the standpoint of this study, the list of plants for any of the three regions must not be considered as static but as dynamic in char- ticter. As already emphasized in an earlier ]jart of this paper, with the changes in the physiography of the ravine the vegetation of its slopes tends to assume a xerophytic aspect, but only until stability of topography favors the succession of forest types, whose climax in the Evanston-Waukegan region is essentially the same as that of the mesophytic slopes of the ravine, the hard-maple type. An investigation which is to determine whether or not the differ- ences in the evaporative power of the air and in the soil-moisture content account for these successions and the minor differences within each of the areas where plants have been listed, involves, as the first step, the selection of certain typical stations — stations significant from the standpoint of exposure and ravine development ; as the second step, the measurement, at regular intervals, of the evaporative power 1— 1 <0 W Ph M Ph M w [Trees and shnibs] Saponaria officinalis Polygala Senega X X Jiiniperus comimi n is X Aster la' vis X Sheperdia canadensis X Aster multiflorus X Thuja occidentalis X Gentiana quin qu e folia X Populus grandidcnidia X Polytrichum comm^me X Prunus serotina X HeliantMis divaricatus X Salix glaucophylla X Asclepias syriaca X Eld'agnvs argent ea X Desmodium rigidum X Ostrya virginiana X Monarda fistulosu X X X Lonicera Stillivantii X Caiherinea undxdota X Tilia americana X X Mnitim sp. X Ceanothus americana X X Anemonella thalictroides X X Acer saccharum X Adianium pedaium X Carpinus caroliniana X Botnjcliium virginianum X Juniperus virginiana X Asclepias phytolaccoides X Fraxinus americana X Thalictrum dioicum X Hamamelis virginiana X Uvularia grandiflora X Viiurnum Opulus X Miiella diphylla X Cornus alternifolia X Hepatica acutiloba X Cornus panictilata X Trillium declinalum X Cornus circinata X Viola cucrdlata X Bides Cynoshati X Galium triflorum X Psedera cpiinquefolia X Anemone virginiana X Menispermum canadense X Cryp tota'uia can a dc n ns X Celastrus scandens X X X Osmorhisa longistylis X Bhus toxicodendron X X X A mphicarp a mon o ica X Quercus alba X Impatiens biflora X Quercus ruira X Allium tricoccum X Querents velutina X Smilax herbacea X Quercus macrocarpa X A r alia nudicaulis X Carya ovata X Aralia racemora X Corylus americana X Desmodium grandiflorum Polygonaium cnmmiita- X X [Herbaceous plants] tum Snnicula marilandica X Budhecl'ta hiria X Keracleum lanatum X Fragaria americana X Solidaqo canadensis X Melilotus alba X Convolvulus sepium X Equisetum arvense X Sanguinaria canadensis X Fquisetum hyemale X Helianthus Mrsutus X Pedicularis canadensis X Erig eron philadelp h id i s X Echhwcysiis lobata X Eriqeron ramosus X Krigia amplexicaulis X of the air and the soil-moisture content at the stations selected; and finally, as the third and last step, the interpretation of results. Location and Description of the Stations In this study seventeen different stations were established ; fifteen in the ravine, one in the oak uplands, station i6, and still another in the open uplands, station 17 (PI. IX, Fig. 15), the vegetable garden of the McLeish estate. The fifteen stations in the ravine were lo- cated, speaking generally, as follows: Station i (PI. I, Fig. 2) in an embryonic ravine; station 2, in a portion of a ravine, with lateral cutting delayed, forming a miniature clay caflon; stations 3 (PI. I, Fig. I), 4 (PI. Ill, Fig. 5), 5 (PI. IV, Fig. 6), 6 (PI. V, Fig. 8). and 7 (PI. VI, Fig. 9), in the part of the ravine with the mesophytic forest vegetation on the slopes; stations 8 (PI. IV, Fig. 7), 9 (PI. VII, Fig. 10), 10 (PI. VII, Fig. 11), II (PI. II, Fig. 4), and 12 (PI. VIII, Fig. 12) in a portion slightly less mesophytic than the preceding; and stations 13 (PI. VIII, Fig. 13), 14 (PI. II, Fig. 3), and 15 (PI. IX, Fig. 14), on the slopes of a wash or gully near the mouth of the ravine. The specific geographical location of each of these stations in the ravine and its tributaries and also of those in the oak and open uplands is shown on the surface map (PI. XVIII), while the angle of slope, direction of exposure, and soil and vegetation char- acteristics are enumerated in the following tabular summary. The "wilting coefficient" — a term used by Briggs and Shantz ('12 : 9) to represent the moisture content of a soil corresponding to the wilting point of a plant — for various soils in the McLeish ravine was correlated with the soil-moisture equivalent by an indirect method devised by these authors ('12:56-57). This coefficient indicates the limit of soil-water content above which growth must occur, al- though plants will live and continue to draw water below this point. In listing the dominant vegetation of each station the emphasis was placed on the herbaceous forms and the seedlings of shrubs and trees because they are a much better index of present ecological con- ditions than the mature trees. Measurement oe the Evaporative Power of the Air The instrument that was used to measure the evaporative power of the air in the different stations was the Livingston ('10) porous cup. The structure and operation of this simple instrument have been so well set forth by the inventor and others that have used it for scientific purposes that an explanation of its management would be o e5 w o « o OO r- iH C^J to lO -* «o >o « o e5 o e3 o tJ « u CO TtH CO CO CO o t~ Ol 05 «5 cc m t~ OO 'El ft 02 S §5 ;S c =5 e -f^ e •S ^ o S 1 s § *- s i ^ s o c e c • 1 =c S ■f* S l^§ o ^ '^fen i !S > •5^ o S e o (a to .9 CO 0 0 PI (>;. S a 03 l^co 0 PI Is -2 "^ S !h 0 03 CO CO cS (I) to ® >-> a 03 02 0 03 02 0 02 'S 0 Vi m 7; 03 « 0 fM 02 C/J 7i w CD H ^' ^ f4 H y, ^ ^' r^' ^' N ^' !5 ^ 02 w ^ ^ ^ a ® 0 ft ej W to 10 ^ 1 « ^ 0 OJ P 0 =H S m +j 0 ^-^ ^ ^ g- h:; <3 < h '^ 10 CO t> 00 OS 0 iH N CO -* in CO T-\ r-l i-H iH iH iH iH 8 superfluous. The writer is aware that the instrument does not register with complete accuracy the evaporative power of the air when there has been precipitation during the period for which the measurement is made. During a rain, indeed, there may be a flow of water into the bottle. Drizzling rain for a long period will cause more water to enter the cup than will the same aggregate precipitation if due to a heavy rain extending over a short period. Since the results in this study have only a comparative value, the entrance of water into the bottles does not vitiate the results. This fact was first pointed out by Brown ('lo) and later emphasized by Fuller ('ii). In the handling of the instruments for scientific accuracy, in the selection of the intervals of time for reading, and in plotting the results, the writer had the constant advice and suggestion of Dr. Fuller, of the Department of Botany of the University of Chicago. Without his help many of the difficulties would not have been so readily over- come. All the instruments were standardized before being placed in the respective stations, and restandardized at intervals of six to eight weeks. Bv the coefficients thus obtained all the readings were re- duced to a common unit. The first reading was made on July _S- 1913. and weeklv readings were made thereafter until October 18, when the last record was made. Thus the investigation extended over a period of 112 davs. Very few of the instruments were molested during the entire season, and hence the regularity of the results for each of the stations is rarely interrupted. These results are expressed as the average daily loss for the weekly interval between the readings. These results are graphically represented with the weekly intervals as abscissae, and the amount of daily loss by the standard atmometer in cubic centimeters as ordinates. INTERPRETATIOX .\XD DISCUSSIOX OF RESULTS I. The graphs in Figure 16 (PI. X) show that the average daily rate of evaporation for the weekly intervals was almost uniformly greater in the embryonic ravine than in either the narrow or broad portions, and greater in the broad than in the narrow. The average amounts of evaporation per day for the entire season at stations i. 2, and 3 were 5.44 c.c, 3.27 c.c, and 4.75 c.c. respectively; that is 2.17 c.c. per day greater in the embryonic ravine than in the narrow portion, .69 c.c. greater in the embryonic ravine than in the broad portion, and 1.48 c.c. greater in the Inroad than in the narrow portion. In terms of percentage the rate of evaporation was 66.3% greater in the embrvonic than in the narrow portion, 14.?^^ greater in the embryonic than in the broad, and 45.2 9f greater in the broad than 9 in the narrow.* Since all of these stations are located near the floor and represent siiccessional stages in ravine development, the data suggest a movement from moderate moisture conditions in the em- bryonic ravine to greater moisture in the narrow portion, with a reversion to slightly drier conditions as the ravine changes from the \'-shaped to the U-shaped type. 2. The graphs in Figure 17 (PI. X) show that the rate of evapo- ration was only about two-thirds as great at the foot of the northwest- facing slope as near the top of the same slope, and that there was little difference in the rate of evaporation for the greater part of the season between the foot and the middle of this slope. The average amounts of evaporation per day at the foot, middle, and near the top of this slope were 4.75 c.c, 4.28 c.c, and 6.67 c.c. respectively. The graphs in Figure 22 (PI. XIII) show similar results; the amount of evaporation per day was 2.33 c.c. or 63.4%, greater near the top than at the foot of the north-facing slope. The rate of evaporation at the middle of this slope was, for about the first two-thirds of the season, nearly the same as near the top. The graphs in Figure 19 (PI. XI) show simi- lar results for the opposing slope, the average amount of evapo- ration per day being 4.75 c.c, 5.51 c.c, and 7.6 c.c. respectively for the bottom, middle, and top of the southeast-facing slope — 2.85 c.c, or 60%, greater near its top than at its foot. A similar series of stations in a less mesophytic portion of the ravine exhibit a similar set of con- ditions to those found in the more mesophytic portion. Here the rates of evaporation at the bottom, middle, and top of the north-facing slope (PI. XIII, Fig. 22) were respectively 3.67 c.c, 6.36 c.c, and 6.0 c.c, while for the south-facing slope (PI. XIV, Fig. 24) they were 3.67 c.c, 5.74 c.c, and 7.93 c.c. ; that is, the average amount of evaporation per day was 4.26 c.c, or 11.6% greater near the top of the south- facing slope than at its foot. The graph for the daily rate of evapo- ration for weekly intervals at the middle of the southeast-facing slope (PI. XI, Fig. 19) is diagrammatically between the graphs for the evap- oration at the foot of the northwest- and near the top of the south- east-facing slopes, and likewise the graph for the rate of evaporation of the middle of the south-facing slope (PI. XIV, Fig. 24) is between the graphs for the foot of the north- and near the top of the south- facing slopes. As already pointed out. the rates of evaporation for the middle of the northwest- and north-facing slopes (PI. XVII, Fig. 17, and PI. XIII, Fig. 22) were not intermediary between those at the foot and near the top of these slopes; the coincidences in the *In the computation of the percentage rates the smaller of the two numbers is always used as the base. 10 rates of evaporation at the middle and foot of the northwest-facing slope (Fig. 17) during the early part of the season might be attrib- uted to the fact that the instrument at the middle of the slope was almost surrounded by a dense growth of Osmorhisa longistylis (PI. Ill, Fig. 5) ; and the slightly higher rate of evaporation at the middle than at the top of the north-facing slope during the latter part of the season (PI. XIII, Fig. 22) might have been due to the unusual expos- ure of the middle of the slope compared with that at the top (PI. VII, Figs. 10 and 11). Further, the differences in evaporation on different parts of the northwest- and north-facing slopes (Figs. 17 and 22), are less than on corresponding parts on the opposite slopes, and thus a station between two other stations would give less striking differ- ences. In summary, the graphs in figures 17, 22, 19, and 24 show that as the northwest-, southeast-, and south-facing slopes are as- cended there is a graded rise in the evaporative power of the air, the southeast- and south-facing slopes showing this more decidedly than the northwest- and north-facing slopes. 3. The graphs in Figure 18 (PI. XI) show that the rate of evap- oration was generally greater near the top of the southeast-facing slope than near the top of the opposite slope, the average amount per day for the entire series being nearly i c.c, or 13.9%, greater; the graphs in Figure 20 (PI. XII) show that the rate of evaporation was almost uniformly greater at the middle of the southeast-facing slope than at the middle of the northwest-facing one, the average amount of evaporation per day being 1.23 c.c, or 28.7%, greater; the graphs in Figure 23 (PI. XIII) show that the rate of evaporation for every weekly interval was greater near the top of the south-facing slope than near the top of the opposite north-facing one, being 1.93 c.c, or 32.1%, greater; while the graphs in Figure 25 (PI. XIV) show re- sults at variance with those in Figure 20 (PI. XII) — a greater rate of evaporation at the middle of the north-facing slope than at the middle of the opposite south-facing one. This was undoubtedly due, as be- fore suggested, to the unusual exposure of the middle of the north- facing slope, which was subject to less shade than any of the other stations in the immediate vicinity excepting the one near the floor of the ravine. The results shown by figures 18, 20, and 23 indicate a greater loss of water and consequently drier conditions on the south- east-facing slope than on the northwest-facing slope, and likewise greater loss of water and drier conditions on the south-facing slope than on the north-facing one. This difference in evaporation is un- doubtedly due to the fact that the incident rays of the sun strike the southeast- and south-facing slopes more nearly at right angles, which means greater absorption and radiation of energy on the southeast- 11 and south-facing slopes than on the northwest- and north-facing slopes. Furthermore, this difference tends to explain the presence of tender mesophytic forms on the north-facing slopes of a ravine dur- ing the early part of spring, before the development of leaves on the trees, and their absence on the south-facing slopes. 4. The graphs in Figure 21 (PI. XII) show that the rate of evaporation was greater throughout the season in the oak uplands than near the top of the northwest- and of the southeast-facing slopes, being 2. 11 c.c, or 31.6%, greater per day for the entire season in the oak uplands than near the top of the northwest- facing slope, and 1. 18 c.c, or 15.5%, greater than near the top of the southeast- facing slope. The graphs in Figure 26 ( PI. XV) show similarly that the rate of evaporation in the oak uplands was generally greater than near the top of the north- and south- facing slopes, the average daily evaporation' for the entire period being 2.78 c.c, or 46.3%, greater in the oak uplands than near the top of the north-facing slope, and .85 c.c. per day, or about 10%, greater than near the top of the south- facing slope. This indicates that the upper portions of the south- facing ravine slopes have an evaporation very close to that of the oak uplands. Indeed one might expect a greater evaporation near the top of the south-facing slopes than in the oak uplands, due to the in- fluence of slope exposure. Further, the differences in the average amount of evaporation per day between the oak uplands and near the floor of portions of the ravine were 4.03 c.c. and 5. 11 c.c. In summary. all of the preceding statements show greater evaporation in the oak uplands than in any part of the mesophytic portion of the ravine. These differences may account for the more xerophytic aspect of the vegetation in the oak uplands than of that in the ravine (see p. 5). The graphs in Figure 21 (PI. XII) and Figure 26 (PI. XV) further show that the rate of evaporation in the open up- lands, the strawberry bed. was nearly twice as great as in the oak up- lands or in the most xeroph)^tic portion of the ravine, the average daily rate for the entire period near the top of the northwest- south- east-, north-, and south-facing slopes, the oak uplands, and the open uplands being 6.67 c.c, 7.6 c.c, 6 c.c, 7.93 c.c, 8.78 c.c, and 16.39 c.c. respectively. The high rate of evaporation in open uplands clearly shows the effect of shade in retarding evaporation. This gives some conception of the relatively high rate of transpiration of the plants that constitute the various crops as compared with the vegetation that springs up in the forests, it having been clearly shown by Livingston ('10) that there is a definite relation between rate of evaporation of water from a free surface and the rate of transpiration in a leaf. 12 5- The graphs in Figure 2y (PI. XV) show relatively high rates of evaporation when compared with the rates in other portions of the ravine. The average amounts of evaporation per day at the foot, middle, and head of the gully, or wash, were 8.93 c.c, 8.31 c.c., and 9.32 c.c. respectively. The evaporation on the gully slope shows less gradation than on the mesophytic slopes of the ravine, the differ- ence between the average at the foot and the head of the gully being only .39 c.c. The evaporation on the gully slope was about the same as that in the oak uplands, which was 8.78 c.c. per day for the season. There is no question that during certain parts of the day the evapo- ration was greater on the gully slope than in the oak uplands. These high rates of evaporation on this slope suggest a vegetation with a xerophytic aspect (see p. 5). In the early part of the summer the gully slope was nearer devoid of vegetation than during the latter part of the season. Reference has been made in an earlier part of tiiis paper to the luxuriance of the late-summer flora of the slope (PI. II, Fig. 3). These changes in the vegetative conditions M'Ould tend to account for the special variations in the graphs for the gully slope, the presence of abundant vegetation about the atmom- eter tending to check the evaporation and its absence tending to in- crease the evaporation relatively. 6. A comparison of all the graphs, with the exception of those for the station in the open uplands and the stations on the gully slope, shows a remarkable similarity in the variation of the average daily loss for the weekly intervals, the minimum evaporation occur- ring during the week preceding August 16 and the maximum during the week preceding September 12. The graphs for the station in the open uplands and the stations on the gully slope show that the lowest rate of evaporation was during the same week as for the other sta- tions, while the highest rate for the open uplands was during the week preceding August 2, and for the stations on the gully slope high rates occurred during the weeks preceding July 12 and 26, and August 2 and 31, as well as during the week preceding September 12. SUMMARY AND C0NCI2v^ '^S ^ ^ !"" 5sS3^xv L ^ 5 ^ ''^V /fj'^^\/^ ^"//ly^^y^ >w yl^|>i»»^^\ y,/ '%/ ^'*'»^ \ j^ j^ \. y/^ ^-'^OSl ^/ \ 1 N^^ ^L^^ \ ««!!'"' vO >r 1 Fig-. 16. Evaporation rates in an embryonic ravine (1), in a portion of ravine with lateral cutting delayed (2), and at the bottom of the most mesophytic portion of ravine (3). stations 1, 2, and 3. JULY AUGUST SEPTEMBER OCTOBER l\) 1 n i \ ( ) y \ ~ t V K. /\ ^, / .A ^ \ k / \ y \ /A > ^ A K ^ / ^ ^H \. / \ / ^ n i^ // \ > y 'X ^ / 1 '/ \ 7 y > V ^^^ iii.^ ^ ^* *ti\i // / p- ^ Y 5^ ^^ ^ '^J '4^ vv f ^ -4^** - 3 n \ ^ _i Fie. n. Evaporation rates at the foot (3), middle (4), and top (5) of the north- west-facing slope of the most mesophytic portion of the ravine. Stations 3, 4, and 5. Plate XI JULY AUGUST SEPTEMBER OCTOBER - / 4 /" N, / 3\ \ ,^ V J f . u s ' ^' / f . y \ 0 X / \ / \ r \ / 1 /^ s. -4 r \ \ / y V» ^ ~~ — c \ 1 " » f 1* s / Y V \^ ^^ ^ ^ ^ > s^ 3 ^ V £^ ^ r y V> ^ ^ s •^ \ y >/* f Fig-. 18. Evaporation rates at bottom (3) and near the top of the northwest (5) and the southeast-facing (7) slopes of the ravine. Stations 3, 5, and 7 JULY AUGUST SEPTEMBER OCTOBER 2 n 1 5 ? A K 1 1 / A / \ <« r - — r A^ V / \ I ^ ^ \ ^ '^X \ ^ V f \ i r \ k T ' / V \ / 6 V j / J'S, > y j t s, f \ "^ ^ V V / "'Z 5 4 A \ 1 / \. h. 7 \ / V ^ f N \ / A y ^ / ^ ^ I ^ > \ \ 1^ V- ^ ^ V J y ^ r \ ^ ** \ A V > r \ r kAl ^ 7 ^ / i- Fig. 19. Evaporation rates at foot of the northwest facing slope (3), and middle (6) ana top (7) of the southeast-facing slope of the most niesophytic portion of the ravine. Stations 3, b, and 7 PivATK XII JULY AUGUST SEPTEMBER OCTOBER 1 9n fc: 10 b \ ,> / i\ , <" / / V L / / v s // •». / > \ / \ y r~- p; V / V / i Y V Y /^ \ y / X ^ \ Sj y r y, > p* A Y ^ / -*' ^ r ^ ^ _ __ 7 L— Fig. 20. Evaporation rates at the middle of both the northwest (4) and the Southeast (o) facing- slopes of the most raesophytic portion of the ravine. Sta- tions 4 and 6. JULY AUGUST SEPTEMBER OCTOBER r~ 1 / \ / \ i r \ / ^ I .0— / . \ ;^ ^ ^/ \ \ . 1 > , / - \ / 0 \ \ 1 I 16 / ^ \ / \ > V \ / L f A \ / \ /' \ ^ \ n / \ \ ./ \ 1 V \ ' V / ^ k l\ 7 \ / ji\ / ^* A > / X \i >s y ^%\ 7 N't \ V ^ ^1 r ) \^ i / v \ / f V / V. ^ / \ J \ / \ / ^ a > ^ ' f \\ f > V y^ ■^ ' / (J / _ _L _ Fig-. 21 Evaporation rates near the top of both norhwest (5) and southeast (7) facing slopes of the ravine, the oak uplands (16), and the open uplands (17). Sta- tions 5, 7, 16 and 17, PI.ATE XIII JULY AUGUST SEPTEMBER OCTOBER on 10 9 A in A .^ «>' >=: », / A ^ / / ( ) J f \ / > / s L-*" \ 1 t y'' ^ k / *■ ^ / /^ > y / \ y \ \^ X r > 5 y V \ ^ / \ \ > r r s / s / ^ \ ^ X ^ ^ y r \ 1 r / \ \ / / 0 yMI r s X X* k^ / y r ^ ~~ \ y r _^ . ^ r ^, Fig. 22. Evaporation rates at the foot (8), middle (9), and near the top (10) of north-facing slope of ravine. Stations 8, 9, and 10. JULY AUGUST SEPTEMBER OCTOBER ?0 — 1 15 \ / 10 12 / /\ / ,y ^ \ / - — / A r\ ^ y J' '"n ^ v. / V / k i ■* y \ y f 0- \ / / \ \ / / ^ > f / \ f0* V > f 5^ — / (. > 1 A / / " 1 / / s^ / / \ \ / y' J*^ \ / ^rfi^ \ ^ / \ / Vy >^ N 0 ^ r > r V r „ y \ \ I / s ^ y*" r N X > \ r 4 f — r^ \/ ... J ^^ Y1 ^. L L Fig. 23. Evaporation rates at the bottom (8) and near the top of the north (10) and of the south (12) facing slopes of the ravine. Stations 8, 10, and 12. Plate XIV JULY AUGUST SEPTEMBER OCTOBER 1 lot 12 A 12 / ^ / V /I / ■*^ f / V ^ L / J \ y / f / r \ \ > / / ^i / \ 4 / / /' ■■"«*< f \ \ V / / / y y 11 \ y r\ y '.. / V \ / V J y ; AiN / ^ / / r s. / / \ L / s. s {^ / ^ ^^ y " y 1 r r r \ / \ / \ .' y ^ s ^N ./ ^ ^ _ ^ \ j / / s ^ y r s > A L u f \ w > n ^, Fig-. 24. Evaporation rates at the foot (8), at the middle (11), and near the top (12) of the south-facing- slope. Stations 8, 11, and 12. JULY AUGUST SEPTEMBER OCTOBER 1 90 |f\ 10 10 / i Y ^ /*■ • »-_ f/ ' V A / 7 \ /, \ A jA y f I N ^ \ A y R ' / / \ / ^ ^ f ^ ^ > Y \ / y '" I A r ^ ^ / r \ 4 V\ f V _ _ ^ Fiitr. 25. Evaporation rates at the middle of both the north (9) and the south (11) facing- slopes of the ravine. Stations 9 and 11. Plate XV JULY AUGUST SEPTEMBER OCTOBER 1 1 V / \ y > y / \ / > ^ s / ^ \ y ^ / V \ > 1 i I 1 \ \ 1 \ \ J \ f ' i 'i^ y R / /J^ \ --/ V- A 1 \ / > f K / 0^ \ / / S s.^ /1?\ 1 / / L / ^ / V' ^ 1 / V \ J. k / i \ \ 1 y k^' i *■ \ Y -^ /^ ^. / r X^ \ ,/ "^ / 7 ^1 / > N / / A\ ^ / / 1 / s. \ ^ s. \ / <> / / 1 > o y / "*■ 1 / \ / f s s y 0 \ / /' / Sv / /" V '/ t r \ / \ y \ ^ ^ y / / \ ' ^ y \ / / y /" r- \ y y ■^ y/ ' _ ^ ^^ IL ^^ wm ^ ^ ^ ^^ ,^ ^ J Fig. 26. Evaporation rates at the foot f8), near tlie top of the north (10) and of the south (12) facing- slopes of the ravine, in the oak uplands (lb), and in the open uplands (17). Stations 8, 10, 12, 16, and 17. JULY AUGUST SEPTEMBER OCTOBER 1 2C 15 1i \ n\ 13 . y \ / ^ _ / \ \ / y ^ / \ \ / f ^ \ ' 1 i 'l^\ IC 14 y / / \ . \ h y — A y \ 5>>s, L y \ / \ N \ '/ ** y A / \ \ ,V ^ s. J y \ V y / \ \ — ^ / 5 \ 1 JL \ y r \ s , 1- f "^^ ^ ^ X V Fig. 27. Evaporation rates at the foot (13), middle (14). and top (IS) of the wash or gully slope. Stations 13, 14, and 15. Plate XVI 17 16 15 14 13 12 10 9 8 7 6 5 4 3 2 ] ^5 ^™ . ^^ ^" Fig-. 28. Diagram siiowing- tlie comparative evaporation rates at the different statioas on the basis of the maximum average amount per day between July 5 and October 18, 1913. JULY AUGUST SEPTEMBER OCTOBER "~a- "■ IT 2 90 \ / \ ^ \ J^ X / \ x" '>^ /'S. / \ A<^ ii^ ^^ ^^ ^^^\ Z'^S/^^ ^^ TS^" "^---"^ \^^V^^^^ -^"^ ^4Srmitrmi^ n 1 MmIh - r/ 10 Pig 29. Graphs showing the range of soil moisture in the narrow ravine, Sta- tion 2, at depths of 7.5 cm. (a) and 15 cm. (b). Broken lines show the wilting coef- ficients for ihe soil at depths of 7.5 cm. (c) and 15 cm (d). Plate XVII JULY AUGUST SEPTEMBER OCTOBER t V \ 1 v i ^n >■ i^ y \^ y > / \ /\\ y > \ / V / \ / y s y > y ^ V / / y \ s ^^ '" \ \ \ > / t V / \ / > y 0C\ ^ y >«. \ / A /u 1 \ / ^ N r - .>" > ^L / N: ».^ y '^ ■ i— m^ r V "v X" ^ "" ^ in ' ~ "•" "■ ■ > — ._ ~" ^ ^ "" ~ • ■ "^ ~ ^ •~ -" " "~ •~ _ _ ■^ _ Fig. 30. Graphs showing the rang-e of soil moisture at the bottom of the me- sophylic ravine, Station 3, at depths of 7.5 cm. (a) and 15 cm. (b). Broken lines show the wilting coefficients for the scil at depths of 7.5 cm. (c) and 15 cm. (d). JULY AUGUST SEPTEMBER OCTOBER 1 ^ ~_\ ^n / / y > / / r / »» s \, / / ?o 1 s. f /' \ , \ / \ s y --t < ^ f \ / , s ^ X t^ \ / \ ^^ y / T S ^ f ^^ S*'' 10S: ' .- ., N 'CZ _ ,,^ . ■ — ^ _ ._ _ ^ ^, 'd' ^ ^ "~ ■"" ■^ ~ ~" • " ~ ^ ^ ^ _ _ _ ^ Fig. 31. Graphs showing the range of soil moisture at the foot of the wash. Station 13, at depths of 7.5 cm. (a) and 15 cm. (b). Broken lines show the wilting coefficients for the soil at depths of 7.5 cm. (c) and 15 cm. (d ). PI.ATI3 XVIII Map of McLeish Ravine, Gleiicoe, Illinois. Scale about 1-5 miles to the inch. AXINKTIA (IKV 'SIONI'I'II 'MOnNI'IT' APR 24 19:6 Bulletin OF THE Illinois State Laboratory Natural History Urbana, Illinois, U. S. A, STEPHEN A. FORBES, Ph.D., LL.D. Director Vol. XII. March, 1916 Article II. A CLASSIFICATION OF THE LEPIDOPTERA BASED ON CHARACTERS OF THE PUPA BY Edna Mosher, Ph.D. > APR 24 1.9 :t. Bulletin OF THE Illinois State Laboratory Natural History Urbana, Ii^linois, U. S. A. STEPHEN A. FORBES, Ph.D., LL.D. Director Vol. XII. March, 1916 Article II. A classification of the lepidoptera based on CHARACTERS OF THE PUPA BY Edna Mosher, Ph.D. CONTENTS PAGK Introduction 17 Changes preceding pupation 18 External morphology 22 Classification 30 Analytical table of superfamilies 30 Pupae with functional mandibles 34 Micropterygoidea So Eriocraniidae So Pupae without functional inanilibles 37 Generalized pupae without maxillary palpi 37 Hepialoidea 37 Hepialidae 38 Cossoidea 38 Psychidae 39 Cossidae 40 Eucleoidea 41 Megalopygidae 42 Eucleidae 43 Pyromorphidae 44 Generalized pupae with maxillary palpi 44 Tineoidea 44 Prodoxidae 45 Aerolophidae 4G Tineidae 47 Heliodinidae 47 Aegerioidea 48 Aegeriidae 49 Tortricoidea 51 Epiblemidae 52 Olethreutidae 54 Tortricidae 56 Sparganothidae 57 Gracilarioidea 58 Nepticulidae 61 Heliozelidae 62 Tischeriidae ; 63 Bucculatrigidae 64 Lyonetiidae 64 Gracilariidae 65 Phyllocnistidae 68 PAGE Specialized pupae with pilifers 69 Pyralidoidea 69 Pterophoridae 70 Attevidae 71 Pyralididae 72 Papilionoidea 78 Megathymidae 79 Hesperiidae 80 Lyeaenidae 83 Papilionidae 85 Pieridae 87 Nymphalidae 88 Specialized pupae without pilifers 95 Yponomeutoidea 96 Epernieniidae 96 Yponomeutidae 97 Coleophoridae 98 Geleehioidea 98 Lavernidae 99 Seythrididae 100 Gelechiidae 101 Chrysopeleiidae 104 Oecophoridae 104 Stenomidae 105 Cosmopterygidae 106 Elachistidae 106 Noctuoidea 107 Noctuidae 107 " Arctiidae 119 Liparidae 121 Bombycoidea 123 Lasiocampidae 123 Bombycidae 124 Notodontoidea 125 Geometridae 126 Notodontidae 132 Dioptidae 134 Sphingoidea 135 Sphingidae 135 Saturnioidea 140 Hemileucidae 142 Ceratocampidae 143 Saturniidae 144 Phylogeny 147 Acknowledgments 150 Bibliography 152 Plates 153 Article II. — A Classification of the Lcpidoptera based on Char- acters of the Pupa."^ By Edna Mosher, Ph.D. Introduction It is within comparatively recent times that the immature stages of insects have been considered of any taxonomic value. The economic entomologist early realized the value of being able to recognize the immature stages, for in many orders of insects the larval stages alone were responsible for many ravages upon crops and orchards. Still the matter was not taken up by the systematists, and the workers in the field of economic entomology contented themselves by rearing the adult to determine the species, and then describing, perhaps all the stages, or more probably the larval and adult stages as being those of economic importance. Nowadays we are beginning to see that it is impossible to construct an adequate classification of any group of insects unless we use every bit of information obtainable on their life historv and habits. It is possible to multiply instances of the value of the larval stages in classification, so that one scarcely needs to cite examples ; but the pupae have been less frequently used. There are cases, however, in which the only good taxonomic characters available are found in the pupal stage of the insect. Such instances are found among the nema- tocerous Diptera, particularly in the family Chironomidae. Scudder ('89) was the first to attempt a classification of lepidopterous pupae, but his keys to the chrysalids were based, not on structural characters, but on the various projections from the body, the cuticular append- ages, the coloration, and the mode of suspension. Among the Lepidopteraa great deal of work has been done towards the classification of the larvae, but until 1893 nothing of importance had been done towards a study of the pupae. In this year Dr. T. A. Chapman, in a paper entitled "Some Neglected Points in the Pupae of Heterocerous Lepidoptera," called attention to the fact that the pupae possessed some remarkable taxonomic characters which might he used to clear up many of the disputed points in the classification •Contribution from the Entomological Laboratories of the University of Illi- nois, No. 48. 18 of the order. This he endeavored to do for the groups in which material was availahle for study, and he has since pubHshed other articles as additional material was obtained. However, Dr. Chapman attempted no classification of the Lepidoptera on this basis, merely pointing out the pupal characters of the major groups and calling attention to instances in which a study of these characters would ap- parently alter the existing schemes of classification. The attention of American entomologists w^as called to this subject by Dr. A. S. Packard ('95) in a paper entitled "Attempt at a New Classification of the Lepidoptera." He made a new grouping of the order based upon pupal characters and figured a large number of species. His determinations of the homology of the various parts of the pupae studied were far from correct, and this, of course, in- validated many of his conclusions. Since that time nothing has been done in America towards a classi- fication of the Lepidoptera based on pupal characters. The purpose of the present investigation is to present such a classification as far as material has been available for study. There is also an attempt to throw some light on the relationships existing between the different groups. Changes Preceding Pupation The person who begins the study of pupae with the preconceived notion that the pupal stage is an interpolated one in the insect's life and that a pupa bears little or no resemblance to either larva or adult, will probably find abundant cause for a change of mind before his study is completed. In the case of Lepidoptera one is apt to think that no similarities could possibly exist between any of the three stages of the insect's development after it leaves the egg. After care- ful study, however, one is surprised with the resemblance between the stages, for it is of the highest importance in the study of any group to be able to homologize larval, pupal, and imaginal characters. This has been done to some extent in certain orders of insects, par- ticularly in those groups w^here the resemblance between the larva and adult is more striking than in the case of the Lepidoptera. Attempts have been made, however, even in this order, to homologize the mouth-parts of the larva and adult, and some of the larval structures have been homologized with certain structures in the pupa ; but appar- ently the idea that all three stages should be studied has been left for other minds to entertain. The first striking difference between larva and pupa is that of size. This difference is easily explained by the great difference in the size of the alimentary canal. Another striking difference is that the 19 pupa apparently lacks legs and prolegs. As will be shown later, the legs are always present, but folded and not in use, while the scars of all the prolegs remain to show their location and are very easily identified in the majority of cases. Many lepidopterous larvae possess striking tubercles and warts, and usually an abundance of setae. All larvae possess setae, but they are often inconspicuous. On the ex- posed portions of the body surface, in so far as observed, the pupa always retains the scars of these warts and tubercles, and the pupal body possesses setae arranged in most cases in the exact order in which they occurred in the larva. ]\Iany other structures of the larva can be easily identified in the pupa, and these will be discussed later. In the case of insects with complete metamorphosis the name pupa is applied to the stage of the insect in which it is more or less quies- cent while undergoing the changes which are necessary to fit it for its adult life. This word pupa, from the Latin meaning baby, was applied to this stage by Linnaeus from the resemblance of certain pupae to a baby which has been swathed or bound up, as was the custom in many parts of Europe at that time. This name was per- haps more appropriate for the pupae of the Lepidoptera than for those of any other order of insects because the appendages are usually all soldered to the thorax. The change from larva to pupa in the Lepidoptera has been observed by many workers and is full of surprises for the amateur who wishes to breed these insects. The caterpillar when ready to pupate stops feeding, and in many instances leaves the food plant and wanders about, often apparently in the greatest of haste. Many are then seen on sidewalks, garden paths, and other traveled places, especially during the autumn months, when the majority of larvae are seeking a place to spend the winter. These larvae, if confined, will refuse food and many of them spin silk threads which are used to suspend the pupa or to form a cocoon. The alimentary canal is always freed of any food materials. The larval skin at this time loses its luster and becomes more and more wrinkled ; and the body becomes shorter and shorter and appears swollen, which is due to the molting- fluid glands pouring their secretions between the outer and inner layers of cuticle. Some drops of a yellowish or reddish fluid are usuallv found in the place where larvae are confined and this, together with their peculiar appearance, often leads the amateur breeder of Lepidoptera to think that decomposition is taking place, and results in the hasty disposal of the now helpless insect. In the case of larvae which spin a cocoon these changes are not so easily observed, unless the cocoon is a very frail one, because most of the changes described take place inside of the cocoon. These changes may occupy but a few 20. hours, or may last for nearly a week. In the case of the common tomato-worm, Protoparce Carolina, the transformation process usually requires five days ; certain species of Papilio observed took but three days, but the time varies much with different individuals and the con- ditions under which they live. When the molting fluid has done its work in loosening the larval cuticle, this splits along the meson of the thorax, and is gradually worked to the caudal end of the body, liberating the enclosed pupa. The liberated pupa is covered with a more or less transparent cuticle and resembles the pupa of the more generalized Neuroptera, Trichop- tera, and Coleoptera. In all of these orders, the insects on casting their larval skins show the first resemblance to the adult insect. In the Neuroptera, Trichoptera, and Coleoptera, the appendages, as well as the body, are encased in a pupal skin, are free from each other and the body, and together with the body segments possess considerable freedom of motion. This does not mean that the pupae, have any power of locomotion ; on the contrary they are quite helpless, and for this reason are frequently — in common with the great majority of pupae — protected by some sort of a cocoon, or earthen cell. The lepidopterous genus Micropteryx, which is supposed by many to be the most generalized of its order, retains freedom of motion in all the appendages and in all but the fixed caudal segments of the abdomen. This freedom of motion is gradually lost in lepidopterous pupae as specialization advances, and the adult appendages are not fully de- veloped when the pupal stage is assumed, although the cases of the appendages of the pupa are fully formed. Specialization in the pupa consists also in the hardening of the exposed parts of the cuticle through the deposition of chitin, and in the soldering of the appen- dages to each other and to the body of the pupa. In the generalized families the appendages are soldered to each other but often remain free from the body surface; later the wings become attached to the body surface, but any parts of the antennae, legs, or maxillae extend- ing beyond their caudal margins remain free. The tips of these ap- pendages are provided for in various ways in the higher families, but are always found soldered firmly to the surface of the body of the pupa. Proceeding hand in hand with the soldering down of the ap- pendages is the loss of motion in the abdominal segments. Among certain families there is motion between all of the adjacent segments. There is, however, a successive loss of motion between segments, until the conjunctiva between all but two of the segments is inflexible in some forms, and even in some of the Lepidoptera, entire freedom of motion has been lost in all of the segments. 21 Among generalized forms where the appendages are soldered to- gether, the cuticle of the exposed parts of the body contains but very little chitin, and is but slightly differentiated in texture from the cuticle of the hidden surfaces. When the imago emerges, or even before that time if the body is slightly pressed, the appendages separate very readily from each other, and are not torn upon the emergence of the insect, so that the pupal skin often remains complete except for the slit on the dorso-meson through which the imago emerged. A very differ- ent condition exists, however, among highly specialized forms. Here the exposed portions of cuticle become very hard and firm, while those which are not exposed are very thin and delicate and are almost en- tirely destroyed at the emergence of the imago. The outer covering, of course, being so firmly soldered together, remains in one piece and is apparently complete except for the slit through which the insect emerged. This has led many to think that this outer chitinized por- tion was the entire pupal skin and that it was a structure, analogous perhaps to an egg shell, in which the pupa had been enclosed. Another remarkable difference between the generalized and the highly specialized Lepidoptera lies in the fact that in the latter the ap- pendages are not fully formed when pupation takes place, but con- sist of the transparent cuticular coverings through which one or more slender tracheae may be seen. The duration of the pupal stage doubt- less influences this, there being a stronger tendency among highly specialized forms to hibernate as pupae. During the life of the pupa the adult parts are developing, and be- fore it is time for the imago to emerge, the cuticular parts of the adult are fully formed. In the generalized families previously mentioned and in some specialized forms where the pupal cuticle remains more or less transparent, one is able to see a part of the development taking place, especially in the case of the appendages. The scales appear on the legs and wings and the color pattern may often be easily traced on the latter several days before the emergence of the insect. This stage of the insect, after the cuticular parts are fully formed, and while it still retains its pupal skin, has been designated as the preimago*. If the pupal skin is not already dark in color, it grows considerably darker in the last few days before the insect emerges, and one is thus able to determine when the preimago stage is reached. *Packard applied the term subimago to the corresponding stage in certain Hymenoptera. This is an unfortunate use of the term as subimago had already been applied to the first winged stage of the Ephemeridae. 22 ExTERNAi, Morphology The most important work on pupal morphology has been done by E. B. Poulton and Dr. T. A. Chapman. Poulton ('91) in his paper on the "External Morphology of the Lepidopteroiis Pupa" discusses a few pupal structures but does not attempt to name all of the parts or to locate any of them. So far as known he was the first to point out that the pupal structures were more than cases for the imaginal structures and objected to the terms pterothecae, ophthalmothecae, etc., as applied to pupae. Believing that Poulton's theory is correct, such terms have not been used in this discussion, nor the terms wing cases, antennal cases, leg cases, etc., but these are spoken of simply as wings, antennae, maxillae, etc. Chapman's papers, already referred to, dis- cuss very fully some of the structures and describe their exact loca- tion; but as they include only a few figures one is left very much in doubt as to the identity of many of the structures and their location. W. Hatchett Jackson ('91) published a very valuable paper on the "Morphology of the Lepidoptera" in which he discussed the external determination of sex in the pupa. A short discussion of the chrysalis was published by Dr. S. H. Scudder ('89), and some of the parts were named. In a paper previously cited. Dr. A. S. Packard ('95) gives many figures of pupae and names the parts, but his homologies are far from correct. It seems necessary, therefore, before proceeding further, to discuss the principal pupal structures and indicate their location by means of figures. The homologies given in this paper were determined by a series of dissections of pupae in various stages of development, the preimago being found most valuable for this purpose. Pupae of nearly every family mentioned in this discussion have been studied in this way, be- ginning with the Micropterygoidea and extending through the He- pialoidea, Cossoidea, and other generalized families, including the Saturnioidea which are believed to be the most specialized of lepidop- terous pupae. The change from larva to pupa has been watched in many species and the subsequent folding and soldering down of the appendages carefully noted. A large number of species have been bred and a study of the method of dehiscence, as shown bv the pupal skin, has thrown considerable light on many instances where there was doubt as to the number of free abdominal segments, or where a suture was obscured by folds or other modifications of the integument. The three regions of the body — head, thorax, and abdomen — are easily recognized, and each will be discussed in turn. There occur, on all of the regions of the body, in different families prominent pro- 23 jections and ridges of various types especially in the Papilionoidea. These projections have no morphological significance. THK HEAD The usual sclerites found in the head of generalized insects may be located in lepidopterous pupae. The sutures are distinct in gen- eralized pupae, but are obliterated in the more specialized groups. Vertex. — This is an area found on the dorsum of the head. It reaches its highest development in the Gracilarioidea, but is usually distinct in all generalized pupae. It is bounded cephalad by the Y- shaped epicranial suture (es), and may be seen in Figures 3, 25, 29, 33' 49' 53' and 56; v. This area was referred to by Chapman and Packard as the dorsal head-piece. Front. — The front is the sclerite to which the antennae are at- tached. It is bounded by the epicranial suture on the dorsal surface, and on the ventral surface by the fronto-clypeal suture, which nor- mallv extends for a short distance caudad from the base of each an- tenna and then transversely to the median line. In some pupae where there is a "shoving back" of the head parts as in the Pyraustidae (Fig. 76) and Sphingidae, the front is located on the dorsum of the head. The fronto-clypeal suture is usually not distinct except in very gen- eralized forms. The superfamily Gelechioidea, however, shows it very distinctly. It is indicated in Figures i, 8, 26, 30, and 36; f. In generalized pupae the front bears two setae on each side of the meson which are often very conspicuous. Genae. — These sclerites are distinctly bounded in Eriocraniidae and Hepialidae (Figs, i and 8, g). They are found laterad of the front and clypeus, and mesad of the glazed eye. The mandibles are alwavs adjacent to the genae at their lateral margins. Clypeus. — Remarkably few pupae have the clypeus definitely bounded. The suture between the clypeus and labrum is seldom pres- ent, although it is often indicated by a furrow. It is then impossible to determine accurately as to its presence, but it has been considered as if it were present. The boundaries of the clypeus are shown very distinctly in Figure i, cl. In the Hepialidae (Fig. 8) there is no clypeo-labral suture present although all the other head sutures are distinct. The clypeus can usually be identified by the presence of the invaginations for the anterior arms of the tentorium, which are asso- ciated with its lateral margins. This sclerite often bears prominent setae, and in the pupae of species whose larvae are borers it has often a distinct cutting plate or ridge. 24 Tentorium. — The invaginations for the anterior arms of the ten- torium are very distinct and are either small pores or slit-like openings. They are associated with the lateral margins of the clypeus and are distinct in most pupae (Figs, i, 14, 19, 30; at). Labrum. — The labrum is usually distinct along its lateral and dis- tal margins, but seldom separated from the clypeus by a distinct su- ture. Like the clypeus it usually bears setae which are especially con- spicuous in the Eriocraniidae (Figs, i and 2, lb). A peculiar de- velopment occurs in the Heliozelidae and some other families where the labrum extends caudad over the appendages (Fig. 50). Pilifers. — This term is applied to the caudo-lateral projections of the labrum, which are so well developed in many Lepidoptera. They are very large in certain superfamilies, notably the Pyralidoidea and the Papilionoidea, and their presence is easily detected by the lobes which are adjacent to the caudo-lateral angles of the labrum and often approximate, or meet on the meson caudad of it, or are separated by a narrow piece of the labial palpi (Figs. 70, 72, 74, 76, 78, 79; pf). The mandibles figured by Scudder ('89, Vol. 3, PI. 87, Fig. 25) are the pilifers. There are often well-developed pilifers present, how- ever, when there are no external indications of their presence. Mandibles. — The mandibles are always located adjacent to the caudo-lateral angles of the labrum. They are not functional except in the Micropterygoidea. In this superfamily, as shown in a pupa of the Eriocraniidae (Fig. i, md), the mandibles are very large and used by the pupa to cut its way out of its cocoon and in working its way to the surface of the ground. In the Hepialidae (Fig. 8, md) and in some other families (Fig. 11, md) the mandibular area is def- initely bounded. In still other families the area is distinctly elevated and usually rugose, as in the Eucleidae and Aegeriidae (Figs. 19 and 36, md). This type of mandibular area is observed in many of the Sphingidae. In the majority of pupae, however, the mandibles are represented by a smooth area situated in the position indicated above. Bye-pieces. — These are situated laterad of the genae and mesad of the antennae. There are always two regions to be noted : a smooth mesal portion, sometimes only a narrow band but often a wider lunate piece, called the glazed eye-piece; and the larger lateral portion, the sculptured eye-piece. The latter is so called bcause it is always sculp- tured like the adjacent parts of the thorax. The sculpturing on the head is seldom like that found on the thorax and abdomen, but, strange to say, that on the sculptured eye-piece is always like that on the thorax, although the eye-piece is probably an extension of the ver- tex. On the dehiscence of most generalized pupae the eye-pieces are 25 separated from the face-parts and remain attached to the conjunctiva which joins the vertex to the prothorax (Fig. 43). In the specialized forms they remain attached to the face-parts. A pecuHar modification is found in the Eucleoidea (Figs. 17, 19, 23; se, ge) in which the eye- pieces form movable flaps seemingly to protect and to cover the meso- thoracic spiracles which lie underneath. The glazed eye-piece prob- ably represents the pupal. eye. Antennae. — These are always attached to the front and extend laterad, curving to the ventral surface of the body mesad of the meso- thoracic wings. They may always be identified without any trouble (Figs. I, 8, II, 15, 28; a). In pupae with broadly pectinate antennae, as the Saturniidae, the mesal portion is frequently elevated and has been referred to as the "stem of the flagellum" of the antennae. Labial Palpi. — These appendages lie adjacent on the meson caudad of the labrum except in the Eriocraniidae (Fig. i, Ip). They are visible in the majority of pupae (Figs. 8. 15, 28, 45, 61). They are frequently overlaid and concealed by the maxillae at their proximal end as in Figures 61 and 67, Ip. Often they are entirely concealed by the maxillae with the exception of a small V-shaped piece just caudad of the labrum (Fig. 72). This was thought by Scudder to be a special piece for covering the base of the tongue. Maxillae. — Where labial palpi are visible they occupy a mesal po- sition, caudad of the labrum, with the maxillae laterad of them. When they are invisible and apparently absent, the maxillae lie adjacent on the meson, often overlying and concealing the proximal ends of the labial palpi as mentioned above. The maxillae (Figs, i, 8, 17, 24, 28; mx) vary greatly in length but are never entirely lacking or con- cealed in the pupa. They often extend beyond the caudal margin of the wings, being sometimes free and sometimes soldered to other ap- pendages. The greatest development is found in certain of the Sphingidae where the maxillae do not extend beyond the caudal mar- gin of the wings, but the extra length is taken up in a loop at the proxi- mal end which forms the so-called "jug handle" of Sphinx pupae. The maxillae are always measured on the meson from the caudal margin of the labrum to their distal end and are usually compared in length with the wings, which are measured from the caudal margin of the labrum to their caudal margin on the meson. Maxillary Palpi. — Each palpus is represented on each side by a subrectangular or triangular area caudad of the eye-pieces and lying along the cephalic margins of the prothoracic and mesothoracic legs, frequentlv reaching as far mesad as the proximo-lateral angle of each maxilla. The normal position of these appendages is discussed under 26 the family Eriocraniidae and shown in part in Figure i, mp. They may also be seen in Figures 28, 30, 32, 36, 38 ; mp. Structures which may be maxillary palpi are found in the genus Gracilaria (Fig. 47). The peculiar extensions of the maxillae in the Cossoidea and Eu- cleoidea are not considered as maxillary palpi (Figs. 15, 19, 23). THE THORAX The three segments of the thorax are always distinct. They are only visible on the dorsum, because the ventral and lateral surfaces are covered by the appendages. Prothorax. — This segment probably varies more in size and shape than any of the others. There are some forms, as in the Gracilarioi- dea, Yponomeutoidea, and others, where the prothorax is very short on the meson (Figs. 53, 56, 58; p) or even invisible (Fig. 54), but is very wide at each lateral margin. It is longer in the Galleridae and certain families of Noctuoidea than in any other pupae examined. ProtJioracic Legs. — These lie adjacent to the maxillae at their proximal end. The coxae are frequently exposed, especially in gen- eralized pupae where the appendages are free (Figs, i, 11, 19; cxi), and dissection frequently showed a segment cephalad of the coxa, the trochantin, although there was no distinct suture indicated on the exterior, this being covered bv the mouth-parts. The trochanter is a very small segment usually found at the caudal end of the femur when the leg is folded and is therefore generally concealed by the tibia and tarsus. The femur extends from the trochanter cephalad to the caudal margin of the head. It is frequently concealed by the tibia and tarsus which are the only portions of the prothoracic leg always visible, but they are often shoved slightly laterad so that a portion of the femur is exposed (Figs, i, 8, 24, 32, 36; fi). The tibia and tarsus are seldom divided bv a suture except in generalized pupae, where all the segments, even of the tarsi, are readily distinguished (Fig. i, li). Mesothorax. — The mesothorax is usually considerably longer than the other segments in specialized forms, but in generalized pupae all the segments are more nearly equal. Mesotlwracic Spiracle. — This is usually located on the dorsum be- tween the prothorax and mesothorax, sunk deep in the conjunctiva between the segments with an opening adjacent to the caudo-lateral angles of the prothorax. Its primitive position appears to have been much farther ventrad (Fig. 2, msp) and it is found in this position in the specialized Trichoptera. It retains this primitive position in the superfamily Eucleoidea and in the family Nepticulidae. The caudal margin possesses curious modifications in different families in 27 the way of elevated ridges, tubercles, setae, etc., and in some of the Papilionoidea, partictdarly in the families Hesperiidae and Lycaeni- dae, there seems to be a definite external closing apparatus in many of the genera. Sometimes there is a tuft of setae ; in others, a plug or plate of somewhat honeycombed appearance. Mesothoracic Legs. — These are folded in exactly the same man- ner as the prothoracic legs and the femora are very seldom exposed, but may be seen in Figure i, f2. The coxae are frecjuently visible ( Figs. I and 48, cx2 ) . The mesothoracic legs are usually longer than those of the prothorax. The tibia and tarsus of each leg are always exposed (Figs, i, 8, 23, 30; U). They lie on the venter, between the prothoracic legs and the antennae. MesotJioracic Wings. — The wings of the mesothorax almost con- ceal those of the metathorax, except in the most generalized forms where the appendages are free. In most families they are the only wings visible on the ventral surface (Figs, i, 8, 36, 41 ; wi). Tegnlae. — The tegulae are the large lobes which, in the adult, cover the proximal end of the wing. They do not form separate pupal pieces but are indicated in some pupae (Fig. 2, t). The tegulae are referred to by many authors as the patagia. The patagia are lobes of the pronotum which project over the mesonotum. Alar Furrozvs. — The furrows along each lateral margin of the mesonotum are designated as the alar furrows. They are best devel- oped in the Aegerioidea (Fig. 37, af) although there are distinct de- pressions in many families. Axillary Tubercles. — In the genera Tropaea and Telea of the Saturniidae, there is found a large tubercle at the base of each wing, with sometimes an additional smaller one. The edges of these tubercles are strongly chitinized and somewhat roughened, and serve to cut the cocoon for the emergence of the moth. Thev are probably assisted in this by the peculiar development of the wing sclerites of the preimago, which protrude into these tubercles and are sometimes found to have cut the pupal skin at the apex of the tubercle. Metathorax. — This segment is longest in generalized forms, where its length is nearly equal to that of the mesothorax. Metathoracic Legs. — The tibiae and tarsi of the metathoracic legs are never normally exposed for their entire length but are concealed by the other appendages excepting at their distal end. Only a small portion is visible in specialized pupae, and the appendages are often wholly concealed (Figs. 2, 8, 36, 45; lo). Metathoracic Wings. — These are usually covered by the meso- thoracic wings except for a narrow strip along their dorsal margin. 28 In a few families a narrow strip of the metathoracic wings is visible on the ventral surface caiidad of the mesothoracic wings (Figs, i, 8, 19; W2). THE ABDOMEN The abdomen consists of ten segments, of which three, segments 8-10, are always "fixed"; that is, they possess no power of indepen- dent motion. In the generalized forms motion is possible between all of the other segments. A segment is said to be movable when there is movement between its caudal margin and the segment caudad of it. In many pupae the movable segments are capable of being telescoped so that only their caudal margins are visible. When the cephalic margin of a movable segment is referred to, it includes the rounded part of the segment which is covered by the transverse conjunctiva when the segments are retracted, or telescoped. Proleg Scars. — The scars of the larval prolegs are found on the ventral surface near the meson (Fig. 11, psc) and are often conspic- uous. Tubercle Scars. — Those families in which the larvae have promi- nent tubercles show very definite scars in the pupae. These are espe- cially noticeable in the Saturniidae. Setae. — There are usually setae present on the abdomen, and they are arranged much as are those of the larvae. They are often very in- conspicuous, otherwise they might furnish good taxonomic characters. There is often a dense covering of secondary setae over the entire sur- face, as in some gelechiids and lasiocampids. The Pterophoridae re- tain a spiny armature similar to that found in the larvae. Spines. — These are found covering the dorsum of the abdomen in generalized pupae (Fig. 49), and larger ones are also found at the caudal end of the body (Figs. 27, 31 ). They are arranged in rows on the segments in Tineoidea and Tortricoidea (Figs. 27, 31, 39, 41). Flanged Plates. — The flanged plates are best developed in the pupae of borers, but are found in other pupae as well. Figure 9 shows them well developed on the dorsum and also shows a well-developed ventral plate on the seventh segment. They are usually developed along the cephalic margin of the segment and prevent the telescoping of the segments. Genital Openings. — In the male the genital opening is situated on the ventro-meson of the ninth abdominal segment. It is usually either a mere slit-like opening as in Figure 5, without any adjacent eleva- tions, or it has a distinctly elevated tubercle on each side as in Figure 8, go, and occasionally is situated in a slight depression. In the fe- males there are two openings which may or may not become con- 29 fluent. These may be mere rounded pores or slit-like openings, and are associated apparently with the eighth and ninth segments. The boundary lines between segments 8 and 9, and 9 and 10 are rarely distinct on the meson, and where they are distinct it seems as if the caudal opening were associated with the tenth segment. In the more specialized pupae the caudal margins of the eighth and ninth segments are more strongly curved cephalad near the meson than in the male (Figs. 34 and 44, go) and the segments are dovetailed together. The presence of the two openings apparently represents the more general- ized condition (Figs. 7, 17, 28; go). They are confluent in Podosesia syringae (Fig. 36) and Arc hips argyrospila (Fig. 44). Anal Opening. — This is always situated on the meson near the caudal margin of the tenth segment. It sometimes shows as a cir- cular opening (Fig. 7, ao) but is usually slit-like (Figs. 8, 14, 17). It is usually surrounded on each side with several prominent wrinkles or folds. Anal Rise. — The anal opening is frequently situated on the summit of a mound-like elevation known as the anal rise. The setae on this rise are very conspicuous in certain families of Tortricoidea (Fig. 38, ar). Abdominal Spiracles. — Spiracles are always present on abdominal segments 1-8. The spiracles of the first segment are covered, so far as observed, by the wings, except in the superfamily Eucleoidea and the family Nepticulidae. The spiracles of the eighth segment are never functional and show no distinct opening. Spiracidar Furroivs. — On the cephalic margin of the movable seg- ments cephalad of the spiracles are found furrows which frequently extend almost to the meson on both dorsal and ventral aspects. They occur in several families, as the Liparidae and Geometridae, but are best developed in the Sphingidae, where they are lacking in but a few genera. They are usually separated by sharply carinate ridges and are of various types, but their function is unknown. Cremaster. — The cremaster is a prolongation of the tenth segment and is not found in the more generalized pupae. It was homologized by C. V. Riley with the suranal plate of the larva. It is of various lengths and shapes and often bears setae at the distal end. Two types of cremaster are shown in Figures 41 and 65, cr. Its length is meas- ured on the ventral surface from its junction with the curve of the ventral surface of the body, as in Figure 44, where ab represents the cremastral length. 30 ClvASSlFlCATlOX As no extensive classification of the Lepidoptera based on pupal characters has been attempted hitherto, little has been done to deter- mine what characters are of value in defining superfamilies, families, and genera. It has been necessary, therefore, to base specific, generic, and other distinctions on those characters found in such material as could be secured. The present investigation has been limited by the difficulty in obtaining representatives of many groups, and it is not expected that the tables and descriptions given will do more than furnish a basis for later work upon the subject. It is hoped, however, that they will call the attention of entomologists to the vast possibili- ties opened up by the use of the taxonomic characters available in pupae, and that further studies of the dififerent groups will make it possible to identify an insect in one more stage of its life cycle — which can not fail to be of importance in the case of our economic species. Analytical Table of Superfamilies a. Mandibles present, large, functional, decussating, and extending be- yond the lateral margins of the body. mICROPTERYGOIDEA aa. Mandibles, if present, never large, parallel or subparallel, and usually represented by small elevated tubercles. b. Movable abdominal segments present cephalad of the fourth, or if no segments are movable cephalad of the fourth then the ap- pendages free from each other and never soldered to the body wall, and the vertex longer than the prothorax measured on the meson. c. True maxillary palpi never present, but sometimes lateral exten- sions of the maxillae (Figs. 15 and 19). d. Body heavily chitinized and bearing transverse rows of spines or setae on the abdominal segments; spiracles never visible on the first abdominal segment. e. Mesothorax never more than twice the length of the meta- thorax; seventh abdominal segment with a large flanged plate on the ventral surface ; antennae filiform, short, only reaching caudad to the proximal end of the mesothoracic legs; head sutures all present except the elypeo-labral. HEPIALOIDEA. ee. Mesothorax always more than twice the length of the metathorax ; seventh abdominal segment never with a large 31 flanged plate on the ventral surface ; antennae, if present, pectinate and reaching farther caudad than the proximal end of the mesothoracic legs; none of the head sutures distinct for the whole length. ^^oioi^tt^t-, * COSSOIDEA. dd. Body never heavily chitinized, and never bearing the spines or setae on the abdominal segments in rows ; spiracles always visible on the first abdominal segment. EUCLEOIDEA. cc. True maxillary palpi usually present ; if absent, then the ap- pendages free from each other, or the vertex longer than the prothorax on the meson, or the body possessing a distinct cre- master. d. Dorsum of abdomen with a covering of small spines, usually over the entire length of the segment and not arranged in distinct rows ; if spines are arranged in rows then the maxil- lary palpi are absent; vertex always longer than the pro- thorax on the meson. ' GRACILARIOIDEA. dd. Dorsum of the abdomen with a distinct row of spines along the cephalic margin of the segment, with or without a caudal row ; spines seldom found elsewhere on the segment but, if present, then the maxillary palpi present and well developed. e. Caudal row of spines never present on the dorsum of the abdominal segments ; maxillary palpi always present. TINEOIDEA ee. Caudal row of spines always present on the dorsum of the abdominal segments; maxillarj^ palpi usually present. f. Distinct cremaster never present; setae never present on the anal rise ; wings narrow and pointed ; large spines always present on the venter of the tenth abdominal seg- "'^''*- AEGERIOIDEA. ff. Distinct cremaster usually present, if not, then setae pres- ent on the anal rise ; wings broad and never pointed ; large spines never present on the venter of the tenth abdominal segn.ent. TORTRICOIDEA. bb. iNIovable abdominal segments never present cephalad of the fourth; appendages never free from each other and usually soldered to the body wall. 32 c. Lobes indicating the presence of pilifers always present except in Gallerinae (Fig. 69) and Oeneinae (Fig. 80). d. Maxillary palpi usually present, if absent, then abdominal segment seven is movable in the male, the body covered with a spiny armature, and both prothoracic and mesothoracic legs extending cephalad between the eye-pieces and the antennae, the former reaching nearly to the cephalic margin of the glazed eye, or a deep furrow lined with setae present on the dorsum between the ninth and tenth abdominal segments; antennae never clubbed at the distal end ; femora of the pro- thoracic legs usually visible ; labial palpi very seldom visible except as a small triangular or polygonal area caudad of the labrum and between the pilifers. pyrat IDOIDEA dd. Maxillary palpi never present; antennae always clubbed at the distal end ; femora of the prothoracic legs never visible ; a deep furrow lined with setae never present on the dorsum between the ninth and tenth abdominal segments; labial palpi never visible except as small triangular or polygonal areas caudad of the labrum between the pilifers and often entirely concealed. PAPILIONOIDEA. cc. Lobes indicating the presence of pilifers never present. d. Mesothoracic wings on the ventral surface at meson usually extending considerably beyond the caudal margin of the fourth abdominal segment, if not, then the body depressed, mostly in the thoracic region, the incisions between the mov- able segments very deep on the dorsum and venter and less deep at the lateral margins, and the caudal part of the an- tennae always adjacent on the meson for a considerable dis- tance ; abdominal segments 1-4 usually longer than the otlier segments ; epicranial suture always present. e. Maxillary and labial palpi present and well developed and a large portion of the prothoracic femora always exposed ; if maxillary palpi are not present then the fronto-clypeal suture never visible ; prothorax distinctly shorter on the meson than at each side, so that each half is triangular in outline; appendages soldered to each other but not to the body wall ; fronto-clypeal suture never visible ; antennae with the caudal portion very rarely touching and not moniliform in appearance. " yPONOMEUTOIDEA. ee. Maxillary palpi usually present, but labial palpi and pro- thoracic femora seldom visible, if visible, then the fronto- clypeal suture distinct ; prothorax usually the same length 33 on the meson as at each side so that each half is subquad- rangular in outline ; appendages usually soldered firmly to each -other and to the body wall ; body usually ovate in out- line, broadest in the thoracic region and usually strongly depressed ; f ronto-clypeal suture usually visible ; antennae usually moniliform in appearance, the caudal portion al- ways adjacent on the meson, usually for some distance ; if only touching, then the f ronto-clypeal suture is distinct. GELECHIOIDEA. dd. Mesothoracic wings on the ventral surface of the body at meson rarely extending beyond the caudal margin of the fourth abdominal segment, if beyond, then maxillary palpi never present ; abdominal segments 1-4 or 1-6 rarely longer than the other segments; epicranial suture seldom visible. e. Labial palpi usually present and well developed and from one fourth to one fifth the length of the wings, if not vis- ible then the body usually shaped as in Figure 104 ; the abdomen with setae arranged around the scars of larval verrucae, and usually flanged plates on the abdomen and a cremaster present ; or with a more or less dense covering of setae never arranged around larval verrucae, the body never of the shape in Figure 104 and flanged plates never present on the abdomen nor a distinct cremaster. f. Labial palpi usually present and well developed and the prothoracic femora usually exposed, or if not, then both prothoracic and mesothoracic legs reaching cephalad to the eye-pieces ; if both labial palpi and prothoracic fem- ora are wanting then the body of the type in Figure 104 ; the abdomen with setae arranged around the scars of larval verrucae, flanged plates usually present on the abdominal segments and a distinct cremaster often present; body never with a more or less dense cover- ing of setae, except arranged as mentioned above ; max- illary palpi occasionally present. NOCTUOIDEA ff. Labial palpi sometimes present, body never with a cre- master and always with a more or less dense covering of setae which are never arranged around larval ver- rucae; prothoracic femora never exposed; maxillary palpi never present. BOMBYCOIDEA. ee. Labial palpi never visible, unless represented by small tri- angular or polygonal areas caudad of the labrum; body very seldom with visible setae. 34 f. Suture adjacent to the proximal ends of the antennae and separating the clypeus and front always present and very distinct; antennae never broadly pectinate so that the width is one fifth of the length; spiracular furrows often present. g. Antennae usually considerably broader near the prox- imal end, their greatest width usually greater than that of the prothoracic legs; antennae usually more than three fourths the length of the wings, if not, then the epicranial suture is present, or the cremaster is wanting, or if present, bifurcate at the distal end or bearing hooked setae ; dorsum of the abdomen usually with a deep furrow between the ninth and tenth seg- ments; scar of a caudal horn never present on the dorsum of the eighth abdominal segment; labial pal- pi sometimes visible as small triangular or polygonal areas caudad of the labrum. nqTODONTOIDEA. gg. Antennae rarely very much broader near the proxi- mal end, usually filiform, their greatest width seldom greater than that of the prothoracic legs, if greater then the cremaster is never wanting, nor bifurcate, nor with hooked setae ; antennae never more than three fourths the length of the wings; epicranial suture never present; dorsum of the abdomen never with a deep furrow between the ninth and tenth seg- ments; scar of a caudal horn usually present on the dorsum of the eighth abdominal segment ; labial pal- pi never visible. SPHINGOIDEA. ff. Suture adjacent to the proximal ends of the antennae and separating the front and clypeus obsolete for the greater part of its length ; antennae always broadly pectinate and the width at least one fifth of the length and often wider; spiracular furrows seldom present. SATURNIOIDEA. PUPAE WITH FUNCTIONAL MANDIBLES Among the Trichoptera, from which the Lepidoptera are supposed to have descended and to which they are known to be very closely related, there are many pupae which have functional mandibles. They function, though, merely to assist the pupa to escape from the cocoon. Among the generalized Lepidoptera the pupae of one super- 35 family, the Micropterygoidea, have large mandibles which serve the same purpose as in the Trichoptera. SuPERFAMiLY MICROPTERYGOIDEA The most generalized lepidopterous pupae known belong to the superfamily Micropterygoidea, which includes two families, the Micropterygidae and the Eriocraniidae, characterized by the presence of functional mandibles. Except for a description of the fragments of the head by Chapman, no pupa of the Micropterygidae has been described, but this family is undoubtedly the most generalized, because the adults possess functional mandibles. The first complete life history of any American species of Erio- craniidae was worked out by Busck and Boeving and published in the Proceedings of the Entomological Society of Washington in 1914 (Vol. XVI, pp. 151-163). These authors gave a short description of the pupa and included some excellent figures. A more detailed de- scription is given here as this species furnishes a working basis for the study of all other lepidopterous pupae. This was made possible by the generosity of Dr. L. O. Howard, Honorary Curator of Insects, U. S. National Museum, who donated some excellent material of Mnemonica auricyanea collected this vear by Mr. August Busck at Falls Church, Va. The pupae of this species (Figs, i, 2, 3) are very small, averaging 3 mm. in length in the males and 4 mm. in the females. The body is covered by a thin transparent cuticle, which shows all the imaginal parts in mature pupae, making it exceedingly difficult to distinguish pupal structures from similar structures in the adult. It is also very difficult to determine the number and position of the setae. The head shows all the sutures usually present in generalized in- sects. The vertex is short, the epicranial suture fairly distinct and extending to the lateral margins of the head. The fronto-clypeal suture extends transversely between the caudo-lateral angles of the an- tennae. The front bears two long" straight setae on each side of the meson about half-way between the antennae and the cephalic margin of the head. In the middle of the cephalic aspect, between the an- tennae, arises a long, fleshy, beak-like projection which contains the long tuft of hairs present in the adult. Just caudad of the front is the clypeus and laterad of these are the genae in the usual position for the Lepidoptera. The suture between the clypeus and labrum is broad and somewhat chitinized, and closely appressed to its ental sur- face is the tentorium, to which the mandibles are attached. The 36 labrum is a large fleshy projection bearing on its ectal surface six pairs of very long setae which extend beyond the lateral margin of the body, and on the ental surface two groups of much shorter setae, which project slightly beyond its caudal margin. All of the appendages of the head are free. The labial palpi are rather short, with three seg- ments, and are somewhat enlarged and blunt at the distal end. The mandibles are exceedingly large and are attached to the ental surface of the clypeus, extending beyond the lateral margin of the body. They are heavily chitinized and serrate along the cephalo-lateral margin. The distal end is broadened and thickened, somewhat circular in out- line, concave and strongly toothed. The maxillae are short and the halves are widely separated. Each half is strongly bent near the dis- tal end, which is directed cephalad and mesad. The maxillary palpi are long, apparently with six segments, and pass from the mouth dorsad and then out towards the lateral margin of the head, making a series of curves which finally bring them between the eyes and the antennae. The distal end is folded close to the body and lies just caudad of the eye. The antennae show a long pedicel with many shorter segments and extend for more than half the length of the wings. The thoracic segments are all more or less movable. The thorax is short, strongly elevated, and moves freely, the greater part of its exposed portion being conjunctiva. The mesothorax and metathorax are nearly equal in length, but seem to possess little power of inde- pendent motion. On the dorso-meson of these two thoracic segments and the first abdominal segment is found a strap-like cuticular thick- ening which is apparently for strengthening the thorax. The tegulae are indicated by the dotted lines in Figure 2 because they do not seem to be distinct pupal structures. The thoracic appendages are also free. All of the coxae are visible and usually the femora of the prothoracic and mesothoracic legs. The metathoracic legs are usually hidden be- neath the wings except at the distal end, which normally curves around the caudal end of the body. The wings never extend to the caudal margin of the body. The first seven abdominal segments are movable in both sexes. The remaining segments are not distinctly sutured and possess no power of independent movement. The genital openings are rather difficult to locate. That of the male is found as a slit-like opening on the ventro-meson of the ninth segment (Fig. 4). There are two openings in the female (Fig. 7, go), apparently located on the ventro- meson of the eighth and ninth abdominal segments. 37 The tenth segment is longer in the female than in the male, pre- sumably on account of the ovipositor. The females always have the eighth, ninth, and tenth segments curved ventrad and closely ap- pressed to the ventral surface of the body. This is shown where the caudal segments are slightly separated from the body, in Figure 2. Figures 6 and 7 give dorsal and ventral views of these caudal seg- ments, and Figures 4 and 5 show the same segments of the male. The anal opening in both sexes is found near the caudal end of the body on the tenth segment. The spiracles are small, circular, and not pro- duced. The mesothoracic spiracle is situated in the conjunctiva con- necting the prothorax and mesothorax. Functional abdominal spira- cles are visible on segments 2-7. The dorsum of the abdomen is practically covered by very minute spines arranged in groups. The following species was examined : Mnemonica aiiricyanea Walsingham. PUPAE WITHOUT FUNCTIONAL MANDIBLES This group includes all the superfamilies of Lepidoptera known, except the Micropterygoidea. In many of the other families the pupae possess mandibles, but they are functionless, and only indi- cated as small parallel tubercles or lobes. Generalized pupae without maxillary palpi The Hepialoidea, together with the Cossoidea and Eucleoidea, differ from all other generalized pupae possessing free abdominal seg- ments cephalad of the fourth, because of the absence of the maxillary palpi. Some of the families included here possess lateral prolonga- tions of tlie maxillae which resemble maxillary palpi (Figs. 15 and 19) and have been considered as such by some authors. These prolonga- tions never separate from the maxillae at dehiscence, and dissection has failed to find any maxillary palpi present in the mature pupae. None of these superfamilies possess all of the sutures found in the head of the generalized type, and none of them show the long, seg- mented antennae present in the Eriocraniidae. SUPERFAMILY HEPIALOIDEA This. includes a single family, Hepialidae, of which the known larvae are borers. The species in this country are of rare occurrence. Their larvae are borers in the stems of shrubs or trees. In Europe some of the species are abundant and injurious. The specimens of 3& Sthenopis thide were obtained through the courtesy of Mr. J. M. Swaine, of the Canadian Department of Agriculture, who obtained them from the stems of willow at MacDonald College, Quebec. Family Hepialidae The pupae of this family are very generalized as to the number of sutures present in the head, the number and arrangement of append- ages, the comparative length of the mesothorax and metathorax, and the nearly equal length of the first seven abdominal segments. These characters are easily seen in Figures 8-10 and need no further de- scription. The pupae are, however, exceedingly specialized as to the chitinization of the body, the spines, toothed ridges and cutting plates on the abdominal segments, and, more than all, in the soldering down of all the appendages to each other and to the body, exactly as in the most specialized of pupae. The head, thoracic segments, and the first two abdominal segments are firmly soldered together, but abdominal segments 2-y are free in the male and 2-6 in the female. The only consolidation of the head parts is that of the clypeus and labrum, between which the suture has been lost. The antennae, as well as all the other appendages, are very short in comparison with the length of the body. These pupae are of considerable size, that of Sthenopis tJiulc being about 30 mm. in length. The larvae as far as known are borers, and their pupae have special adaptations for cut- ting their way to the surface. The most peculiar of these adaptations is the ventral plate on the seventh abdominal segment (best seen in Figure 9), which has not been found in any other pupae examined. The sharp ventral projections on the front also serve as cutting sur- faces, but similar projections are found in many pupae, particularly among other species whose larvae are borers and in very many of the leaf-mining species. The opening of the mesothoracic spiracle has reached the normal position for most lepidopterous pupae, being be- tween the prothorax and the mesothorax at each caudo-lateral angle of the former. The genital opening is found in the male on the meson of the ninth segment lietween two slightly elevated tubercles. In the female there is a single opening apparently on the eighth segment. The following species was examined : Sthenopis thnle Strecker. SUPERFAMILY COSSOIDEA The pupae of this superfamily are less generalized as to head parts than the Hepialoidea, but nevertheless resemble them very closely in size, shape and arrangement of the appendages, in the num- 39 ber of free segments, and in the fact that all the appendages are firmly soldered to each other and to the body. The antennae, how- ever, are of a different type, being pectinate in the Cossoidea. The metathorax varies considerably from the generalized type, being very much shorter, so that the mesothorax is about four times its length, measured on the median Hne. Many of the species in this superfamily have larvae which are borers, and many of the pupae are fitted to work their way to the surface of a burrow. However, the ventral plate of the seventh abdominal segment, which is so distinct in the Hepialoi- dea, is not present. All of the pupae have some of the body segments armed with spines and strongly toothed chitinized ridges, and a strong ridge or projection is generally present on the head. The families may be separated as follows : a. Abdominal segments 2-6 movable in the female and 2-7 in the male; dorsum of abdominal segments armed with a row of sharp spines on the cephalic margin, and a row of setae, which are directed cephalad, on the transverse conjunctiva at the caudal margin ; females without wings and antennae and larva-like in appearance. PSYCHIDAE. aa. Abdominal segments 3-6 movable in the female and 3-7 in the male ; dorsum of abdominal segments armed with a toothed ridge along each margin ; sexes similar Cossidae. Family Psychidae In this family there are no sutures apparent on the head except between the clypeus and labrum and the mandibles. The antennae are short and pectinate. The prothorax is longer and the meta- thorax much shorter than in the Hepialoidea, to which, however, these pupae show many resemblances. The dorsum of the abdomen has toothed chitinized ridges along the cephalic margin of some of the segments and rows of setae along the caudal margin on the trans- verse conjunctiva. The caudal end of the body bears two large, strong hooks directed ventrad. This description applies mostly to tlic males (Figs. 11-13), as the females are quite different as seen in Figure 14. The females never leave the cocoon during their entire life and have no provision for locomotion, even in the adult. It is an aston- ishing fact that no pupal wings are developed, because in all other families where the adult females are apterous the pupal wings are always developed, sometimes as much as in the males. Neither are there any pupal antennae present, no eye-pieces, nor traces of maxillae. The labrum and mandibles show very distinctly, both being consid- 40 erably elevated. The legs are scarcely developed, being represented by transverse chitinized elevations on the venter of the thoracic seg- ments. The abdominal segments are much as in the male. They show on the venter the proleg scars, on the dorsum the rows of toothed chitinized ridges and setae, but the body setae are much smaller and difficult to distinguish and are not represented in the figure. A single genital opening is found in the female, on the eighth abdominal segment. No hooks are present at the caudal end of the body. The abdominal spiracles are present on the first eight abdom- inal segments, but there is no visible opening for the mesothoracic spiracle in either sex. The only genera available for study were Thyridopteryx and Oiketicus. These resemble each other very closely and the difference between the pupae can hardly be considered as generic. The pupae of Oiketicus are larger and stouter, the males examined averaging i8 mm. in length, while those of Thyridopteryx were slenderer and only 15 mm. in length. The two genera may be separated thus : a. Abdominal segments 2-6 with a caudal row of setae, the row on the the sixth interrupted and shorter than the other rows ; caudal spines stout and simple; spiracles scarcely produced beyond the surface of the body except at their cephalic margins. Thyridopteryx Stephens. aa. Abdominal segments 2-5 with a caudal row of setae, no row on the sixth ever present ; caudal spines slender and with a distinct tooth ; spiracles distinctly produced beyond the surface of the body. Oiketicus Gruilding. The following species were examined : Thyridopteryx eplieineraefonnis Haworth Oiketicus abbotii Grote Family Cossidae The Cossidae are borers in the larval stage and seem to be very closely related to the Hepialidae, although they resemble them less than do the Psychidae. This family has segments 3-7 of the abdomen free in the male and 3-6 in the female. There is another sexual dif- ference to be noted, viz., the presence of an extra row of spines on the abdomen of the male. In this sex the seventh segment has two rows of spines and the succeeding segments one row; in the female the sixth is the last segment with two rows, the remaining caudal seg- ments having but one row. The epicranial suture is not distinct in any species, but at dehiscence Prionoxystus robiniae shows a small piece of the vertex on each side of the meson, and this with the con- junctiva bears the eye-pieces. The lateral part of the fronto-clypeal suture is distinct and the clypeo-labral suture is always visible. 41 This family is usually divided into subfamilies of which two, the Cossinae and Zeuzerinae, are discussed here. Figure 15 shows the ventral surface of the head and its appendages in a member of the Cossinae, the arrangement of the other parts being the same as in the Zeuzerinae (Fig. 16). The maxillae have prominent lateral projections in Cossinae which resemble maxillary palpi. These always adhere to it at dehiscence and are not found in Zeuzerinae. Only one genus of each subfamily was studied. The pupae are very large, those of Prionoxystus rohin- iae and Zcu::cra pyrina being respectively 45-50 mm. and 30-35 mm. in length. The two genera studied may be separated as follows : a. Head without a prominent cephalic projection ; maxillae with an apparently segmented lateral projection on each side resembling a maxillary palpus, but adhering to the maxillae at dehiscence; an- tennae more than half the length of the wings and gradually taper- ing ; abdominal segments with the cephalic ridges much larger than the caudal ones and armed with long even teeth. Prionoxystus Grote. aa. Head with a prominent cephalic projection, maxillae never with an apparently segmented lateral projection on each side ; antennae less than half the length of the wings and narrowed abruptly near the middle; abdominal segments with the cephalic and caudal ridges similar, the teeth short and uneven Zeuzera Latreille. The following species were examined : Prionoxystus robiniae Peck Zeuzera pyrina Linnaeus SUPERFAMILY EUCLEOIDEA The pupae of this superfamily are cjuite specialized as to the head parts, the epicranial suture being the only one visible in all the families. They have followed a very different line of development from the Cossoidea and Hepialoidea, because all of the generalized families retain freedom of motion between all the segments except those fixed at the caudal end of the abdomen, and between all of the appendages. The cuticle is very thin and transparent in almost all genera and the dorsum of the abdominal segments in all of them has a covering of small spines over the greater part of the segment. All of the families show the spiracles distinctly on the first abdominal segment. The only other family in which this was observed, the Nepticulidae, has a well-developed maxillary palpus. The meso- thoracic spiracle of each side is in a rather unusual position in this 42 superfamily. The opening is on the dorsum in the normal position, and is very large, with a strongly arched cephalic margin ; but the spiracle is on the ventral surface directly under the sculptured eye- piece in Megalopygidae and Eucleidae, and a little farther laterad in Pyromorphidae so that it comes partly under the antennae. The spiracle, with the adjoining parts slightly pushed aside to show their relation, is seen in Figure 21. The family Pyromorphidae being more specialized than the other two families differs from them considerably, but its relationship to them is evident. The three families included here may be separated as follows : a. Dorsum of abdominal segments Avith spines on the cephalic part and a covering of coarse setae on each eaudo-lateral part which does not usually extend to the meson; maxillae simple quadrangular pieces, without any lateral prolongations; a large conical tubercle caudad of each abdominal spiracle on segments 2-6 ; mesothorax never ex- tending caudad to the first abdominal segment Megalopygidae. aa. Dorsum of abdominal segments with short spines, but never with a covering of coarse setae on any part ; tubercles never present caudad of any of the abdominal spiracles. b. Labial palpi present; mesothorax with a long mesal lobe reaching on to the first abdominal segment ; maxillae always less than half the length of the wings Pyromorphidae. bb. Labial palpi absent; mesothorax never with a long mesal lobe reaching on to the first abdominal segment ; maxillae more than half the length of the wings Pyromorphidae, Family Megalopygidae The Megalopygidae have the head and thoracic segments free, also abdominal segments 1-7 in the male and 1-6 in the female. The ap- pendages are entirely free from each other and from the body wall. The body is soft and covered with a thin, delicate, transparent cuticle which is slightly chitinized. There are always setae on the dorsum of the abdominal segments as well as spines. The setae are found on each side of the meson on the caudal half of all the segments. The epi- cranial suture is distinct but all the other head sutures are obliterated. The front has a distinct projection and the mandibles show as dis- tinctly elevated tubercles. The size and arrangement of the parts may be seen in Figures 17 and 18. This family together with the Eucleidae possesses a very peculiar eye-piece. Chapman ('94, p. 349) called attention to this structure and spoke of it as the "eye-flange". This eye-piece, in reality the sculptured portion, is free along its lateral and caudal margins and extends well out on to the antennae. 43 It is, however, much more wrinkled and sculptured than any other portion of the body. These eye-pieces move up and down in living pupae during respiration and allow one to see the mesothoracic spiracle underneath. The mesothorax possesses some well-defined alar ridges and its caudal margin extends in a broad curve nearly to the caudal margin of the metathorax. The large conical tubercles are found caudad of the spiracles on abdominal segments 2-6. The body of Lagoa crispata Packard, the only species studied, is strongly arched on the dorsum of the abdomen and is short and thick-set. Its length is about 18 mm. and the greatest breadth 10 mm. The following species was examined : Lagoa crispata Packard. Family Eucleidae The Eucleidae retain the same movable segments as the family Megalopygidae, which they strongly resemble. The pupae of Eucle- idae, however, are usually only half the size of the latter, averaging about 9 mm. in length. They also retain the same head sutures, but, as in Prolimacodes, they often show a distinct furrow marking the position of the lateral part of the fronto-clypeal suture. The eye- pieces are identical with those described for Megalopygidae. The size and arrangement of parts may be seen in Figures 19, 20, and 23. In two of the genera studied, Sibine and Euclea, the maxillae, in ad- dition to the usual cephalo-lateral extension found throughout the family (Fig. 23), have peculiar modifications in the form of long lateral prolongations extending to the antennae. Usually only the dis- tal end of this prolongation is seen between the eye-piece and the an- tennae, as in Figure 19, the dotted line showing the connecting part. These two genera also have a distinct groove in each half of the max- illae, into the caudal part of which the femur of the prothoracic leg is fitted. The cephalic margin of the pronotum has a distinct median notch, which makes it appear bilobed, and each lobe is prolonged cephalad over the caudal margin of the head (Fig. 22). The meso- notum is prolonged into a rounded or pointed lobe which reaches on to the first abdominal segment. Only three genera were available for study. These may be separated by the following table : a. Maxillae never with a lateral projection reaching to the antennae; mesothorax with a strongly carinate median line ; caudal lobe of the mesonotmn broadly rounded Prolimacodes Schans. aa. Maxillae with lateral projections reaching to the antennae ; meso- thorax never with a strongly carinate median line. b. Mesonotmn with the caudal lobe pointed Euclea Hiibner. bb. Mesonotum with the caudal lobe broadly rounded, almost trun- cate Sihine Herricli-Schaeffer, 44 The following species were examined : Prolimacodes scapha Harris Bitclea dclphmii Boisduval, chloris Herrich-Schaeffer Sihine stimiilea Clemens Family Pyromorphidae This family is much more specialized than either the Megalopy- gidae or the Eucleidae and resembles them but little. The body is flattened and has lost the power of motion except in the abdomen. Abdominal segments 2-7 are free in the male and 2-6 in the female. The appendages are also very slightly soldered together. The pres- ence of, spines on the abdominal segment, together with the absence of maxillary palpi, is considered sufficient evidence that it belongs to the superfamily Eucleoidea. Figures 24 and 25 show the essential points of its structure. The only genus available for study was Harrisina. The following species was examined : J-Jarrisina americana Guerin-Meneville. Generalized pupae with maxillary palpi The remaining pupae which retain either free segments cephalad of the fourth abdominal segment or free appendages, have followed two distinct lines of development. In the first group the generalized condition of the body found in the Eriocraniidae has been retained as to comparative length of segments and the covering of the dorsum of the abdomen with fine spines. The metathorax is nearly always more than half the length of the mesothorax, while the prothorax tends to become shorter at the meson and broader at the lateral margins, so that each half appears triangular. In the second group, the covering of spines on the dorsum of the abdomen has been gradually changed and there is one very well-developed row of spines at the cephalic mar- gin of each segment, with or without a similar caudal row. In this group the prothorax is longer and somewhat quadrangular in shape and the metathorax is relatively shorter. This group includes the superfamilies Tineoidea and Tortricoidea, and being much smaller than the other will be considered first. Superfamily TINEOIDEA The families included here possess one row of spines along the cephalic margin of the dorsum of the abdominal segments, and well- developed maxillary palpi. In one family, Prodoxidae, the primitive 45 covering of fine spines has been retained, but it may be easily differ- entiated from all other pupae bearing spines of two sizes in a similar position on account of the large maxillary palpi. The family Heliodinidae is included here for the sake of con- venience as it possesses only the cephalic row of spines on the dorsum of the abdominal segments. It is, however, much more nearly related to the Tortricoidea. The families Prodoxidae and Acrolophidae are more nearly related to the Tineidae. Of these the Prodoxidae are undoubtedly the most generalized, retaining more head sutures and a greater number of free segments, in addition to the spines mentioned above. The Acrolophidae are more generalized than the Tineidae in the matter of free segments, but have the appendages firmly soldered to each other and to the body wall. This is probably due to the fact that tlie larvae are sod-borers and that the pupa works its way to the surface. The families may be separated as follows : a. Mesonotum not produced into a long caudal lobe ; mesotliorax seldom more than twice the mesal length of the metathorax. b. Abdominal segments 2-7 movable ; dorsum of abdominal segments with a covering of spines on the caudal part ; maxillae more than twice as long as the labial palpi Prodoxidae. bb. Abdominal segments 3-7 movable ; dorsum of abdominal segments never with a covering of spines on the caudal part ; maxillae shorter than the labial palpi. c. Antennae never extending to the caudal margin of the wings; wings broadly rounded; appendages firmly soldered to each other and to the body ; a lateral projection never present on each side of the tenth abdominal segment Acrolophidae. cc. Antennae extending beyond the caudal margin of the wings ; wings pointed; appendages only slightly soldered together and separating at the slightest touch; tenth abdominal seg- ment with a prominent lateral projection on each side end- ing in a spine Tineidae. aa. Mesonotum produced into a long caudal lobe ; metathorax never more than one fourth the mesal length of the mesothorax. Heliodinidae, Family Prodoxidae In this family abdominal segments 2-7 are free in both sexes. The head shows the epicranial suture plainly, and dehiscence always takes place on the front of the head along what is apparently the fronto- clypeal suture, at least for a part of the distance, as shown in Figure 26. The front bears a prominent chitinized projection armed with two stout teeth. The lateral margin of the eye-piece extends on to the antenna for a very short distance. The appendages are very 46 slightly soldered to each other, but scarcely to the body wall, and sep- arate very easily. The lateral view, Figure 27, shows the relative length of the segments. The abdominal segments, although they have developed a prominent cephaHc row of spines on the dorsum, still re- tain the covering of very fine spines on the remainder of the seg- ment. The eighth abdominal segment bears a pair of very stout hooks at the apices of rounded tubercles (Fig. 27a). The pupae examined measured about 10 mm. in length. The following species was examined : Prodoxiis quinqucpiinctella Chambers. Family Acrolophidae In this family segments 3—7 of the abdomen are movable in both sexes, but the appendages are quite firmly soldered to each other and to the body wall so that they do not readily separate even at dehiscence. There is probably also dorsal movement of the second segment, as the conjunctiva is well developed and both the first and second segments separate at dehiscence. The larvae of members of this family are sod- borers and it seems quite natural that pupae with this mode of life should have their appendages soldered down at a much earlier stage than those of the leaf-miners, for instance, or of the pupae that live in cocoons. There are none of the small spines of the generalized type present on the dorsum of the abdomen in this family, but a well- developed row of spines at the cephalic margin of the segments. There are also short lateral and dorsal projections of the tenth segment, with very sharp chitinized edges, which are evidently to aid the pupa in working its way to the surface. The head bears a strongly chitinized transverse ridge near the cephalic margin of the ventral surface. Figures 28 and 29 show the arrangement of parts in a pupa of this family in which there is a remarkable develo])ment of the labial palpi. The pupae are from T5-20 mm. in length. The genera may be sepa- rated as follows : a. Labial palpi never with distinct cutting plates near their proximal margin, the palpi not extending much over half the distance to the distal ends of the prothoracic legs ; two pairs of coxae visible ; spines of the abdominal segments long and narrow. Hypocolpus Walsingham. aa. Labial palpi with distinct cutting plates near their proximal margin, the palpi extending as far as the distal ends of the prothoracic legs ; a single pair of coxae visible ; spines of the abdominal segments triangular PseudanapJwra "Walsingham. 47 The following species were examined : Hypocolpus mortipennelhis Grote Pseudanaphora arcanella Clemens Family Tineidae In this family the free abdominal segments are 3-7 in the male; no females were available for examination. Segments 1-3 separate dor- sally at dehiscence. The appendages are very slightly soldered to- gether and all separate readily except the metathoracic legs and an- tennae, which extend beyond the caudal margin of the wings and are ciuite firmly fastened together, the legs being underneath the antennae. The appendages are also slightly soldered to the body as far as the third abdominal segment. The arrangement of parts is shown in Figures 30 and 31. The fronto-clypeal suture is indicated by a clear line in the otherwise fairly well-chitinized cuticle. Segments 3-8 of the abdomen bear a cephalic row of spines on the dorsum directed caudad, while the ninth segment bears an interrupted group of spines directed cephalad. There are none of the fine spines of the generalized type of pupa present in this family. The tenth abdominal segment shows a prominent lateral projection on each side, ending in a spine. The setae of the body are very conspicuous. The pupae are about 4 mm. in length. The following species was examined : Tinea pellionella Linnaeus. Family Heliodinidae This family has usually been associated with the Yponomeutidae, but it seems from pupal characters to be more closely related to the tortricids. It is very similiar to these in arrangement of parts; the Heliodinidae, however, have longer maxillae and they plainly show that dorsal motion is possible between the second and third abdominal segments. There are also curved setae at the caudal end of the body in the genus Brenthia (Figs. 32 and 33) strongly resembling those found in the Epiblemidae. Choreutis (Figs. 34 and 35) has a small dorsal plate on the tenth segment with a strong seta at each end which appears to represent an early state in the development of a cremaster. The possession of a single row of dorsal spines on the abdominal seg- ments, however, is like the remainder of the Tineoidea, and it is easier to classify them as such. They differ from the remainder of the super- family in having one more free segment in the male, abdominal seg- ments 3-7 being free in the male and 3-6 in the female. The thorax 48 differs markedly, too, the prothorax and metathorax being much shorter. The mesonotum has its caudal margin produced into a long lobe, while in the other families the caudal margin of the mesonotum is very slightly curved. The appendages are very slightly soldered to each other and to the body, and the wings reach on to the fourth ab- dominal segment. The spiracles are small, circular, and very slightly produced. The pupae are from 6-8 mm. in length. The genera may be separated as follows : a. Body setae longer than the abdominal segments, heavily chitinized and forked at the end ; maxillae, measured on the meson, about half the length of the wings ; abdominal segments without deep punctures along the cephalic margin, but with a row of sharp triangular spines. Breniliia Clemens. aa. Body with very short inconspicuous setae ; maxillae extending to the caudal margin of the wings ; abdominal segments 2-6 with a row of deep punctures along the cephalic margin, and with a row of sharp triangular spines just cephalad of the punctures. Clioreutis Hiibner. The following species were examined : Brcnthia pavonacella Clemens Choreutis inflatella Clemens, gnaphiclla Kearfott SuPERFAMiLY AEGERIOIDEA ■ The Aegerioidea, together with the Tortricoidea retain freedom of movement in abdominal segments 3-7 in the male and 3-6 in the female. The appendages are soldered to the body so that there is no ventral movement possible between the first two abdominal segments ; but there is undoubtedly dorsal movement, and at dehiscence these segments separate very distinctly from each other and the thorax, indi- cating that they have only recently lost their power of motion. In this superfamily is included the one family Aegeriidae. They form a very compact group in which it is hard to find satisfactory characters differ- entiating the genera. Moreover, pupae in good condition are difficult to obtain ; but it is hoped that the characters used here in separating the genera and in defining the superfamily will hold good for those groups to which they are applied. The sexes vary considerably and it has not been possible in all cases to obtain both male and female. This superfamily has most often been associated with the Tineoidea, but pupal characters indicate a much closer relationship to the Tortri- coidea. It is apparently somewhat nearer to the primitive families Eriocraniidae and Nepticulidae than the Tortricoidea, owing to the 49 fact that a very large maxillary palpus is present in all genera, and that spines which reach well around to the ventral surface are found on abdominal segments 2-10, especially on the tenth segment. There are no setae yet developed on the anal rise, and there is not as much consolidation of the fixed caudal abdominal segments. The seventh segment in the female seems but recently to have lost power of motion. The abdominal segments are more nearly equal in length than in the Tortricoidea. Family Aegeriidae The pupae of this family vary considerably in size, from the genus Aegeria, with species varying from 8-16 mm. in length, to the genera Memythrus and Bembecia, containing the largest species examined, varying from 20-25 mm. They are all provided with various forms of cutting plates for working their way to the surface, most of these being on the head, which is heavily chitinized at the cephalic end and usually has many ridges and projections, making it difficult to deter- mine the sutures. The clypeus often bears a sharp transverse ridge, sometimes toothed, which undoubtedly serves the same purpose. The body is elongate, cylindrical, with the abdominal segments approxi- mately equal in length, showing a generalized condition. The arrange- ment of parts in a pupa of this family is shown in Figures 36 and '}^'/. It will be noted that the maxillary palpi are very large, and they remain uniformly so throughout the family. The appendages extend beyond the wings in most of the genera, but the caudal parts of these are not soldered to the body wall. The fronto-clypeal suture is always dis- tinct along the lateral margins of the front from the proximal ends of the antennae almost to the invaginations for the anterior arms of the tentorium, but only shows transversely as a paler band of color in the strongly chitinized cuticle as indicated by the dotted line in Figure 36. Dehiscence invariably follows the course of this suture, and the front with the antennae are separated from the rest of the head parts. The epicranial suture is often obscured by the numerous elevations of the vertex and front, but it is always present. The antennae are always enlarged at the proximal end and again at the distal end, where they are somewhat club-shaped, thus differing again from the Tortricoidea. The mandibles are distinctly elevated in most genera. The wings are narrow and pointed, differing markedly from those of the Tortri- coidea. They are not soldered to the body wall at their distal end. The thorax always has a carinate median ridge, which may be distinct on all the segments, and is always distinct on the prothorax and the cephalic half of the mesothorax. The alar furrows are very deep, and one edge, usually the mesal one, is sharp and heavily chitinized. There are always two rows of spines on the dorsum of some of the abdom- 50 inal segments, which extend around to the ventral surface. These rows of spines are always present on segments 3-6, the number vary- ing on segments 2 and 7, while there is always one row on segments 8-10. The number of rows of spines on segment 7 differs in the sexes, there being two rows in the male and only one in the female. The spines on the tenth segment are very broad, and this row extends nearly to the ventro-meson. Each of the spines has a seta inserted near its tip, which is not heavily chitinized and therefore easily broken. There are never any setae present on the anal rise. The genera of Aegeriidae may be separated as follows : a. Maxillae always more than half the length of the wings, generally nearly or quite equaling their length ; coxae of mesothoracic legs never adjacent on the meson below the maxillae, b. Clypeus with a prominent elevation near its caudal margin, bearing a heavily chitinized transverse ridge or series of projections which are probably to assist the pupa in cutting its way out of the burrow, c. Clypeus with the chitinized transverse ridge produced into a dis- tinct point on each side of the meson, d. Mesothorax with the median carinate ridge extending nearly its whole length, very distinct on the metathorax ; second ab- dominal segment with the two rows of spines distinct in both sexes ; maxillae never reaching the caudal margin of the wings and ending opposite to the antennae ; pupae averaging 20-25 mm. in length Sanninoidea Beutenmiiller. dd. Mesothorax with the median carinate ridge usually extending only along the cephalic half, never distinct on the meta- thorax ; second abdominal segment never with two distinct rows of spines in either sex; maxillae reaching the caudal margin of the wings and ending opposite the mesothoracic legs; pupa usually 8-15 mm. in length. S'yiuintliedon Hiibner. cc. Clypeus with a transverse row of separate projections. ParJiarmonia Beutenmiiller. bb. Clypeus not prominently elevated at its caudal margin and never bearing ridges or projections which could be used in cutting, c. Tenth abdominal segment with eight spines in a row ; caudal end of body just cephalad of the anal opening without setae. Podosesia Moschler. cc. Tenth abdominal segment with ten large spines in a row and two smaller ones, one on each side of the meson ; caudal end of body just cephalad of the anal opening with a row of four setae which are inserted under small projections. MemytJirus Newman, aa. Maxillae about two fifths the length of the wings; coxae of meso- thoracic legs and their tarsi adjacent on the meson caudad of the maxillae Bemhecia Hiibner. 51 The following species were examined : Sanninoidca exitiosa Say Synanthedon tipuliformis Clerck, accrni Clemens, piciipes Grote and Robinson, pyri Harris, scihda Harris ParJiarmonia pini Kellicott Podoscsia syringae Harris Memythrus asilipertnis Boisduval, dollii Neumoegen Bembecia inarginata Harris SuPERFAMiLY TORTRICOIDEA This superfamily, like the Aegerioidea, is distinguished by the pres- ence of two rows of spines on the dorsum of most of the abdominal segments. The Tortricoidea form a more compact group than the Aegerioidea in regard to the arrangement of appendages, which varies so little throughout the families that any member of the superfamily may be easily recognized by this arrangement, together with the pres- ence of spines on the abdominal segments. This characteristic arrange- ment is shown in Figures 38, 40, 41, and 44. There are often projec- tions from the head, much as in the Aegerioidea, but there are never as many head sutures present. The thorax shows the alar furrows in many instances but they are never as well developed as in the preced- ing superfamily, and never have sharp chitinized edges. The abdomen also shows a greater degree of specialization and its fixed caudal seg- ments are much more strongly consolidated, the sutures being very difficult to determine in many cases. The seventh segment has also become firmly fixed in the female. It was found impossible to group the pupae of this superfamily according to any of the schemes of classification now in use. The four groups into which the Tortricoidea discussed in the following pages have been divided are designated as Epiblemidae, Olethreutidae, Tor- tricidae, and Sparganothidae. These names, however, are without any significance whatever as far as previous classifications are concerned, and are merely used as a matter of convenience. Lack of material has prevented further study in this group at present, so it has been impossible to determine the correct family names. No attempt has been made to bring the nomenclature up to date. The generic names used by Meyrick and Walsingham have been followed as nearly as possible. The four groups or families of Tortricoidea must have had a common ancestor, but owing to the development of the maxillary pal- pus within the groups it would be impossible to consider one as derived from another. The line of development appears to have 52 been towards ( i ) a reduction of the spines on the dorsum of the ab- dominal segments, these disappearing first from the tenth segment and then from the segments cephalad of it; (2) the loss of setae on the anal rise; and (3) the development of a long cremaster. The families of Tortricoidea may be separated by the following table : a. Body without a distinct cremaster; setae always present on the anal rise Epiblemidae. aa. Body with a well-developed cremaster. b. Ninth abdominal segment always with a distinct row of spines, especially in the males; tenth abdominal segment sometimes possessing spines ; cremaster broader than long ; setae always pres- ent on the anal rise, c. Cremaster never curved ventrad, the caudo-lateral angles not pro- duced into prominent hooks ; the caudal margin usually showing three short lobe-like projections; second abdominal segment with the cephalic row of spines present and the caudal row well developed; setae of the anal rise always laterad of the anal opening Olethreutidae. ec. Cremaster curved ventrad, the caudo-lateral angles produced into prominent hooks; second abdominal segment lacking the cephalic row of spines and the caudal row poorly developed; setae of the anal rise always on the caudal part of the eleva- tion Tortricidae. bb. Ninth abdominal segment lacking a distinct row of spines, although a few spines are sometimes present in the males; setae never present on the anal rise; cremaster nearly always longer than broad ; tenth abdominal segment never possessing spines. Sparganothidae. Family Epiblemidae The pupae belonging to this family (Figs. 38, 39) have no cre- master and there are always setae present on the anal rise. They are usually less than 10 mm. in length and slender, tapering gradually from the thoracic region to the somewhat blunt end of the body. The genus Carpocapsa is sometimes an exception as the body is often very stout, and the genus Eucosma has a cylindrical body strongly resem- bling the pupae of the Aegeriidae. The maxillary palpi usually extend to the proximo-lateral angles of the maxillae ; only Epinotia and Enarmonia of the genera studied had shorter palpi. The maxillae are about two fifths the length of the wings, and the labial palpi are usually half the length of the maxillae. The rows of spines on the dorsum vary somewhat in the different genera. All have two ro\vs present on abdominal segments 2—7 although the cephalic row of segment two is weak in Eucosma, Hemimene, and some species of Ancylis. Occa- 53 sionally the caudal row of segment seven is weak in the females of some species. As a general rule the eighth segment has but one row of spines, the cephalic one, but two rows have been found in the species of Epinotia and Eucosma, usually in the males. There is, in most genera, considerable difference between the sexes. The antennae do not vary as greatly in this family as in some others, but there is a great variation in the rows of spines, these being usually smaller and less numerous on the caudal row of segment seven in the male and on segment eight in the female. The genus Mellisopus does not show characters of sufficient importance to allow of its retention as a sepa- rate genus, and it is therefore included with Carpocapsa. The only points of difference between that genus and Carpocapsa pomonella, its nearest ally, are that the spiracles are oval, somewhat rectangular and slightly produced, while in Mellisopus latiferreamis they are large and circular but not strongly produced. There is however consider- able variation. Mellisopus shows a slight carinate ridge on the meta- thorax, but this, again, is extremely variable. The phylogeny of the group is extremely doubtful. If the spines on the dorsum of the abdominal segments in Eucosma were homol- ogous with those found in the generalized pupae of Nepticulidae and others, it would certainly be the most generalized form and the others would probably follow the secjuence of the table to genera. The spines, however, are much broader than any observed in the general- ized types. Eucosma also shows a remarkable resemblance to the ;\egeriidae, so it is probable that it is the most generalized of the Tortricoidea examined. The genera of Epiblemidae may be separated a=; follows : A,. With two long distinct setae present on each side of the anal rise. b. Caudal end of body with one row of long, heavily chitinized, flat- tened setae inserted along the line of the row of spines on the tenth abdominal segment. c. Dorsal surface of abdominal segments between the cephalic and caudal rows of spines covered, at least in part, with short trian- gular spines; cephalic row of spines composed of alternating large and small spines Eucosma Hiibner. cc. Dorsal surface of abdominal segments between the caudal and cephalic rows of spines always smooth ; cephalic row of spines of approximately the same size. d. Portion of first coxae exposed on the meson below the maxillae more than half the length of the second coxae; body often stout, with the length scarcely three times the greatest width, but extremely variable ; length averaging 10 mm. Carpocapsa Treitschke. 54 dd. Portion of first coxae exposed on the meson below the maxillae never more than half the length of the second coxae ; body always slender, with the length about four times the greatest width ; average length 8 mm Tmetocera Lederer. bb. Caudal end of body with two rows of setae showing, one row of four setae inserted along the line of the row of spines on the tenth abdominal segment and another row at the caudal margin of the body. c. Caudal row consisting of four setae ; maxillary palpi always touching the proximo-lateral angles of the maxillae, d. Caudal row of setae of two kinds, the mesal ones much slen- derer than the lateral ones ; maxillae less than one third the length of the wings ; labial palpi about two thirds the length of the maxillae ; maxillary palpus touched by the prothoracic leg Hemimene Hiibner. dd. Caudal row of setae similar ; maxillae never less than one third the length of the wings; labial palpi about half the length of the maxillae ; maxillary palpus touched by both prothoracic and mesothoracic legs Ancylis Hiibner. cc. Caudal row consisting of two setae ; maxillary palpi seldom reaching the proximo-lateral angles of the maxillae. Epinotia Hiibner. aa. Never with two long, distinct setae on each side of the anal rise. b. Lateral spines of the tenth row noticeably larger than the others; setae at the caudal end of the body very short and slender, not heavily chitinized ; setae of the anal rise very small and difficult to locate, usually two present on each side Thiodia Hiibner. bb. Lateral spines of the tenth row not noticeably larger than the others ; setae at caudal end of body long and heavily chitinized ; one seta on a distinct papilla on each side of the anal rise. Enarmonia Hiibner. Tlie following species were examined : Bucosnia strenuana Walker, scudderiana Clemens Carpocapsa pomonclla Linnaeus, saltitans Westwood, latifcrreamis Walsingham Tmetocera ocellana Schiffermueller H einunene incanana Clemens Ancylis comptana Frolich, platanana Clemens, diminutana Kearfott Bpiiiotia saliciana Clemens, piccafoUana Kearfott Thiodia signatana Clemens Enarmonia fana Kearfott Family Olethreutidae The Olethreutidae (Fig. 40) include those species which possess a well-developed cremaster, usually broader than long and somewhat 55 flattened, bearing eight strongly chitinized, flattened, hooked setae ; and usually having similar but smaller setae on the anal rise. Exceptions to this latter character are found in the genus Polychrosis, and in Hxartema fcrrifcraninn, which does not agree with the remainder of the genus in this respect. The group is further characterized by the presence of a well-developed row of spines on the ninth abdominal segment in all the males examined and in nearly all of the females, the exceptions being in the genus Exartema, where the spines were smaller and fewer in number. In most genera this row of spines has several additional spines on each side, usually near the meson. The only other species of the superfamily which resemble the members of this group are the species of Archips in group (b), but these have no setae on the anal rise, and very seldom have spines present on the ninth abdominal segment. Hxartema ferriferanum is the only species among those examined which might be confused, as the row of spines on the ninth segment of the female is not well developed, while the males of Archips cerasivorana sometimes have a few spines present. This particular species of Exartema, however, has a prominent cephalic projection, ending in a point, directed ventrad, while the species of Archips are blunt at the cephalic end, and the bodies are usually larger and prominently enlarged in the region of the thorax. The antennae show marked sexual differences, being much longer in the males. The rows of spines on the dorsum of the abdominal segments also vary in tl)e sexes, the caudal row of segment eight being poorly developed or lacking in many females, though well developed in the males. The row on the ninth segment is much better developed in the males. The genus Polychrosis shows a peculiar development of the spiracles. Instead of the small, produced tubular spiracles common to the Tortri- coidea it appears to have them very much enlarged. This prominent enlargement around the spiracle has a deeply concave surface, and the very small tubular spiracle in the center is about one sixth of its width. A similar condition, but not so well developed, is found in Exartema sciotoaniim. The maxillary palpi are well developed and reach the proximo-lateral angles of the maxillae in Olethreutes (b) and Poly- chrosis, but in Episimus, Olethreutes (a), and Exartema they are not well developed. The genera of Olethreutidae may be separated as follows : a. Tenth abdominal segment with spines, usually three or four rows closely approximated, seldom with a single row. b. Long chitinized setae present on the anal rise, usually slightly shorter and narrower than those of the cremaster. c. With two setae on each side of the anal rise, very similar to those on the cremaster Episimus Walsingham. 56 cc. "With one seta on each side of the anal rise, smaller than those of the cremaster OletJireutes (a) Hiibner. bb. Setae never present on the anal rise Polyclirosis Ragonot. aa. Tenth abdominal segment without spines. b. Well-developed setae present on each side of the anal rise. e. Maxillary palpi well developed, reaching the proximo-lateral angles of the maxillae OletJireutes (b) Hiibner. cc. Maxillary palpi short, never reaching the proximo-lateral angles of the maxillae Exartema (a) Clemens. bb. Without setae on the anal rise Exartema (b) Clemens. The following species were examined : Bpisimus argiitanus Clemens Olethretttcs (a) niveiguttana Grote, (b) iiiolachitana Zeller Polychrosis sling crlandmia Kearfott, vitcana Clemens, hotrana Schif- fermueller Bxartema (a) sciotoannm Kearfott, concinnanuni Clemens, nigra- num Kearfott, inornatum Clemens, pcrnuindamun Clemens Bxar tenia (b) ferriferanuni Walker Family Tortricidae This group is distinguished by its peculiar type of cremaster and the presence of setae on the anal rise. The maxillary palpi are not present in Peronea but are found in Argyrotoxa, where they are shorter in the male than in the female. The maxillae are usually about two fifths the length of the wings, the labial palpi nearly half the length of the maxillae. There are no spines present on the tenth ab- dominal segment, and they are not well developed on the second and third segments. There is always a well-developed cephalic row on the dorsum of the tenth segment, but the caudal one does not extend as far laterad in the male and is usually lacking in the female. In Argyrotoxa the cephalic row of spines on the eighth and ninth segments is on a prominent ridge which can be plainly seen on the lateral margin in dorsal view. There are always two setae present on each side of the anal rise and these are always on the caudal part of the elevation. Figures 41 and 42 show the arrangement of parts in this family and Figure 43 the dehiscence of part of the head, showing the eye-piece. The genera of Tortricidae may be separated by the following table : a. Maxillary palpi present in both sexes ; spines of the cephalic row on abdominal segments 7-9 on distinct ridges which show plainly on the lateral margins of the body, the spines extending laterad beyond the spiracles in some segments; setae on the ventral side of cremaster and between eremastral hooks not heavily chitinized and resembling the ordinary body setae Argyrotoxa Stephens. 57 aa. Maxillary palpi not present in either sex ; spines of the cephalic row on abdominal segments 7-9 not on distinct ridges, the spines never extending laterad beyond the spiracles ; setae on the ventral side of cremaster heavily chitinized, and usually not extending far beyond the caudal margin of the body Peronea Curtis. The following species were examined : Argyrotoxa albicomana Clemens, hcrgnianniana Linnaeus Peronea sp., minuta Robinson, logiana Schiffermueller, var. vihiirnana Clemens Family Sparganothidae This family (Fig. 44) includes the species in which the cremaster is well developed and much longer than broad, except in Archips (b) and Phaecasiophora. The cremaster in nearly all species bears eight strong hooked setae which are usually not much flattened except in the genera mentioned above. There are never any setae present on the anal rise, and most of the species have no spines present on the ninth abdominal segment and none of them a well-developed row. The caudal row of the eighth segment is often lacking in the female and is poorly developed in the male. The females of Platynota flavedana have no cephalic row on the second abdominal segment. The members of this group include the largest of the Tortricoidea examined, most of them considerably over 10 mm. in length; the thoracic region usually appears considerably enlarged, and the abdomen is long and tapering. The vertex is shorter than in the other groups. The max- illary palpi do not reach the proximo-lateral angles of the maxillae in the males, but sometimes do so in the females. In Platynota flavedana the palpi appear to extend only along the cephalic margin of the pro- thoracic leg. The setae of the body are usually very long and promi- nent in this group. Sexual differences are noticed in the length of the maxillary palpi and antennae and in the development of the rows of spines on the dorsum of the abdominal segments. There are no available characters by which all the species of the genus Archips can be associated in a single group and it undoubtedly represents two genera, because there are two distinct types of cremaster present. It is also difficult to find good structural characters to separate the genera Harmologa and Archips (a). The color markings are very distinct in Harmologa, and the body is also very noticeably enlarged in the region of the first three abdominal segments, so that in ventral view the lateral margins of the wings appear curved, instead of approximately parallel as in Archips (a). The genera Epagoge and Platynota are also closely related and are grouped together by some writers. The genera of Sparganothidae may be separated as follows : 58 a. Transverse conjunctiva showing prominent dark brown spines scat- tered over a lighter brown surface, b. Cremaster much longer than broad, not flattened. c. With four setae inserted at the caudal end of the cremaster. d. Dorsum of second abdominal segment showing a slightly crenu- late, chitinized cephalic margin, the cephalic row of spines on this segment not well developed in the males and wanting in females; head never with a cephalic projection; abdomen never with prominent cavities on the dorsum of the second and third segments ; dorsum of abdomen always with darker transverse bands and spots of color Harmologa Meyrick. dd. Dorsum of second abdominal segment without a crenulate cephalic margin ; cephalic row of spines Avell developed on this segment in both sexes ; head often with a cephalic pro- jection, or if not, then prominent cavities are present on the dorsum of the second and third abdominal segments; body of uniform color Arcliips (a) Hiibner. cc. With two setae inserted at the caudal end of the cremaster. Cenopis Zeller. bb. Cremaster broader than long, flattened PJuiecasioplwra Grote. aa. Transverse conjunctiva never showing prominent dark brown spines, surface of uniform color, b. Cremaster longer than broad, not flattened; labial palpi always considerably more than half the length of the maxillae, c. Cephalic row of spines on second al)dominal segment lacking in the female ; cremastral setae noticeably flattened. Platynota Clemens, cc. Cephalic row of spines on the second abdominal segment present in the female ; cremastral setae not noticeably flattened. Epagoge Hiibner. bb. Cremaster broader than long, distinctly flattened ; labial palpi not more than half the length of the maxillae. .Arcliips (b) Hiibner. The following species were examined : Harmologa fumiferana Clemens Arc/lips (a) argyrospila Walker, magnoliana Fe:nm\d,parallcla Robin- son, obsolctana Walker, rosacea iia Harris Arcliips (b) cerasivorana Fitch, fcrvidana Clemens Cenopis chambersana Kearfott Phaecasiophora coiifixana Walker Platynota flared ana Clemens Bpagoge sidfiireana Clemens SuPERFAMiLY GRACILARIOIDEA This superfamily name is given to a number of families apparently of common origin, w^hich have proceeded along similar lines of devel- 59 opment. The species are all leaf-miners and are very small, the pupae of the largest species examined being 7 mm. in length. Very few of the generalized families have been available for study, so that it is exceedingly difficult to trace the relationships existing between the more specialized families without first having carefully studied a num- ber of more generalized forms. There is included in this group the Nepticulidae, in many respects the most generalized pupae studied, next to the Eriocraniidae, and certainly resembling the latter more than any of the other generalized forms examined. It is just at this point in our investigation that more material is needed to clear up the relationships of the groups which have apparently branched off here and have had a common ancestor with the Nepticulidae. From all the evidence at hand it seems probable that development has proceeded along two well-defined lines, the first, represented by the superfamily Gracilarioidea, having early lost the maxillary palpi while still retain- ing the covering of spines on the dorsum of the abdominal segments, and having developed the triangular type of prothorax ; the second having retained the maxillary palpi for a much longer time, but having lost the covering of spines, while 'developing the same type of pro- thorax. Of the second type no material has been examined which would show any intermediate stages between the families Nepticulidae and Epermeniidae. The latter family has apparently continued the line of development begun in the Gracilarioidea as it still retains the seventh abdominal segment free in the male though it is fixed in the female. The presence of the maxillary palpi precludes its derivation from the Gracilarioidea and would lead us back to some point below the Heliozelidae because this family also has lost them. As we have only the Nepticulidae for comparison, it has been assumed that this branch has arisen coordinately with them. In the superfamily Gracilarioidea, with the exception of the family Lyonetiidae, all the pupae have free appendages, the cuticle is very slightly chitinized, and the dorsum of the abdomen is covered, in part at least, with fine spines. There is a tendency in some genera, as Lithocolletis and Ornix, towards the development of a single row of spines, so that there is often one or more rows of larger spines at either the cephalic or caudal margin, or at both margins, of the seg- ment. This seems to indicate the way in which the rows of spines were developed in the Tineoidea and Tortricoidea. The characters which are common to all the members of the superfamily are the long vertex, which is always longer than the prothorax at the median line, scarcely ever less than twice its length and often much longer, and the long 60 metathorax, with the loss of a well-developed maxillary palpus in families above the Nepticulidae. Chapman described the genus Graci- laria as possessing maxillary palpi, and in two species, sassafrasella and negimdella, a structure has been found (Figs. 45, 47) which may be the maxillary palpus ; but there never is a distinct, oblong piece lying caudad of the eye-piece as is usually the case when the maxillary palpi are present, and of all the species of the superfamily examined these two were the only ones in which there was any doubt as to its absence. The head is in most families either produced into a prominent projec- tion or there is a heavily chitinized cutting plate near the cephalic mar- gin on the ventral surface. The prothorax has a tendency to become shorter on the median line and longer on its lateral margins, so that each half is triangular. In such cases the length along the lateral margin is about four titiies the length on the median line. In the more generalized forms the prothorax is more like the rectangular type found in the Tineoidea, but it is depressed or sunken, giving it a neck- like appearance. The metathorax still retains its primitive condition, and isusually more than half the length of the mesothorax. In nearly all of the families the wings are long in proportion to the body, and in the majority they are about two thirds its length. The bodies of most of the families included here retain the generalized type found in the Eriocraniidae with a slight depression near each lateral margin in the region of the spiracles. The spiracles are usually small, circular, and slightly produced, appearing tubular. The Lyonetiidae seem to be an exception to almost every rule. They have no free segments, the appendages are all soldered to the body, and there are no spines visible on the abdomen. They seem to be more nearly related to the Buc- culatrigidae than to any other family, although there are strong rea- sons for considering them, related to the Phyllocnistidae. The follow- ing table will serve to separate the families of Gracilarioidea : a. Maxillary palpi well developed and extending along the caudal margin of the ej^e ; spiracles visible on the first abdominal segment. Nepticulidae. aa. Maxillary palpi never well developed, and if present never extending as an oblong piece along the caudal margin of the eye. b. Antennae never extending more than half the length of the wings ; labrum very long and lobe-like, extending down over the labial palpi for about one fourth of their length ; spines on the dorsum of the abdomen very fine and not easily distinguished. Heliozelidae. bb. Antennae always extending at least three fourths the length of the wings, and usually equaling them in length or extending beyond their caudal margin ; labrum never long and lobe-like and never 61 extending down over the labial palpi for one fourth of their length, e. Appendages free, never firmly soldered to the body wall ; abdomen always with some of the segments movable ; dorsum of the abdo- men always with spines, d. Abdominal segments 3-7 movable in the male, 3-6 in the female ; antennae and metathoracic legs not approximately equal in length and both seldom extending beyond the caudal margin of the wings. e. Labial palpi present ; caudal end of body ending in two stout spines directed dorsad; abdominal segments 3-6 with the two setae nearest the meson on the cephalic half of the seg- ment so closely approximated that their bases touch. TiSCHERIIDAE. ee. Labial palpi never visible ; caudal end of body never with curved hooks, but the tenth abdominal segment with a prominent lateral projection on each side ending in a stout straight spine; abdominal segments 3-6 with the setae nearest the meson at the cephalic end of the segment never closely approximated Bucculatrigidae. dd. Abdominal segments 4-7 movable in the male, 4-6 in the female ; antennae and metathoracic legs approximately of equal length and both always extending beyond caudal mar- gin of the wings, e. Abdominal segments 3-7 never with two deep punctures or pits with heavily chitinized edges on the meson at the cephalic margin with one or more heavily chitinized spines adjacent; the length of abdominal segments 8-10 always greater than that of segment 7 Gracilariidae. ee. Abdominal segments 3-7 with two deep punctures on the meson at the cephalic margin wdth one to three heavily chitinized spines adjacent ; abdominal segments 8-10 never showing distinct segmentation, their total length less than that of the seventh segment Phyllocnistidae. ee. Appendages always firmly soldered to the body ; dorsum of abdo- men without visible spines ; abdomen without any movable seg- ments Lyonetiidae. Family Nepticulidae These tiny species of leaf-miners average 2 mm. in length in the females and 1.5 mm. in the males. The body is flattened, with a trans- parent, slightly chitinized cuticle, and is white in color until the adult scales are formed. Although their size makes it difficult to determine the number of free segments, it is believed that there is some degree of motion between all of the abdominal segments except the fixed 62 caudal ones. There is some degree of movement between the seventh and eighth abdominal segments in both sexes, but it is apparently- greater in the male. The arrangement of parts may be seen in Figures 48 and 49. The head does not show all of the sutures found in the Eriocraniidae, but the epicranial and fronto-clypeal sutures are always present. The appendages are all free and segmented as in the Eriocraniidae, and the thoracic appendages are widely separated to show all the coxae. There is a strong resemblance between this family and the more generalized members of the Eucleoidea, but the presence of the large maxillary palpi prevents their being included in that super- family. The spiracles are visible on the first abdominal segment, and the length of the thoracic segments indicates a very generalized condi- tion. The genital opening of the male is located as shown in Figure 48. In the females there is an area covered with setae on the venter of the eighth segment, as in the Eriocraniidae, but no openings could be accurately determined. The following species were examined : Nepticiila nyssacfoliclla Chambers, platancUa Clemens. Family Heliozelidae This family includes some very small pupae which measure only 2-3 mm. in length. They have all the appendages free and widely separated. The cuticle is transparent and the body white in color, with the conjunctiva so little differentiated that it was impossible to deter- mine the number of free segments with accuracy. Segments 2-7 in tlie male and 2-8 in the female have some power of motion, but whether this is movement of the whole segment in the case of the second and third, or merely dorsal movement, was not determined. The family (Fig. 50) is characterized by its short antennae, and its long labrum which projects down over the labial palpi. They also have shorter appendages than any of the other families with transpar- ent cuticle and white bodies, because in all others the metathoracic legs and antennae extend considerably beyond the caudal margin of the wings and are often longer than the body. The epicranial suture is near the cephalic margin of the head. While this family may have retained more free segments than the Gracilariidae it is undoubtedly more specialized than some of the genera in that family. The pro- thorax is much longer at its lateral margins than on the meson; there is no trace of maxillary palpi, and the labial palpi are not so well developed as in the generalized Gracilariidae. The genera included in this family have long been associated with the Elachistidae, but the pupae show no resemblance whatever to this 63 family. The name Heliodinidae has been applied by ;some authors to the genera included in this family, but Meyrick, in Lepidoptorum Catalogus, Part 13, uses this name to include the genera Brenthia, Choreutis, etc. The name Heliozelidae is used by Spuler (Die Schmet- terlinge Europas, 19 10) and this name has been adopted there. The genera may be separated as follows : a. Abdomen with one or two prominent lateral setae on each side of the tenth abdominal segment ; mesonotum not produced into a prominent lobe extending down on the metathorax Aniispila Hiibner. aa. Abdomen never with prominent lateral setae on each side of the tenth abdominal segment; mesonotum produced into a prominent lobe extending down on the metathorax Coptodisca Walsingham. The following species were examined : Antispila ampelopsisella Chambers, cornifoliella Clemens Coptodisca jiiglandiella Chambers, splcndiforella Clemens Family Tischeriidae These pupae are from 3.5-6 mm. in length and have abdominal segments 3-7 free in the male and 3-6 in the female. They are always considerably chitinized, so that the pupae vary in color from yellow to brown. The spines on the dorsum of the abdominal segments are very distinct and in some species they are of two sizes. All of the species examined except Tisclieria hcliopsisclla (Fig. 54) had certain of the body setae very long, heavily chitinized, and forked at the end. These setae vary in length, but the shortest are nearly as long as the abdominal segments and are very conspicuous. The dorso-mesal setae nearest the cephalic margin were closely approximated on segments 3-6 or 7 of the abdomen so that their bases were in contact. The caudal end of the abdomen is bifurcate, and ends in two heavily chitin- ized hooks which are directed dorsad. The arrangement of parts may be seen in Figures 51—54. These pupae have become more specialized in certain respects than many of the Gracilariidae, although they re- tain one more free segment. This is noticeable in the development of the prothorax, in the distinct rows of larger spines on many of the segments, and in the strong caudal hooks. The fronto-clypeal suture shows as a clear area, indicated by the dotted line in Figure 51. This family includes two genera, Coptotriche and Tischeria, with no well- defined characters for separating them. The two species of Tischeria, acnea from blackberry and nialifoliella from apple, at one time consid- ered identical, show distinct differences in the pupae and both species resemble Coptotriche, while hcliopsisclla is very different from all the 64 rest. Dyar's list names but one species of Coptotriche, but three dis- tinct types of pupae have been obtained from mines in oak leaves. Unfortunately no adults have yet emerged from these, so the species can not be determined. The genera may be separated as follows : a. Caudal margin of the dorsum of the second abdominal segment heavily chitinized and toothed, the teeth being larger than the adjoining spines CoptotricJie Walsingham. aa. Caudal margin of the dorsum of the second abdominal segment not heavily chitinized or toothed Tischeria Zeller. The following species were examined : Coptotriche ccUericlla Clemens Tischeria aenea Frey and Boll, malifoliella Clemens, hcliopsisella Chambers Family Bucculatrigidae This family, Bucculatrigidae, including the single genus Buccu- latrix, has been placed in various positions by different authors. It is quite evident that it is more specialized than most other families of the Gracilarioidea in the loss of the labial palpi and that it has pro- ceeded along a different line of development. Nevertheless, no one can fail to see the relationship between the pupae of the Bucculatrigidae and the other members of this superfamily, particularly to some of the species of Cameraria where there is a lateral projection from each side of the tenth segment and a distinct row of larger spines on the dorsum of the abdominal segments. The lack of labial palpi, together with the spines on the abdominal segments, is sufficient to distinguish the family from all the others included in the superfamily. The arrange- ment of parts may be seen in Figures 55 and 56. The pupae examined had an average length of 3 mm. The following species were examined : Bucciilatrix sp., pomifoliella Clemens, trifasciella Clemens. Family Lyonetiidae This family is a very difficult one to place satisfactorily by pupal characters alone, as it has completely lost the power of motion in the abdominal segments and all the appendages are soldered down. This is another of the families which has been a source of anxiety to many lepidopterists. The shape of the prothorax, the length of the vertex, together with that of the wnngs and appendages as compared with the body, the small tubular spiracles, and the absence of maxillary palpi seem without any doubt to indicate its relationship to the members of the superfamily Gracilarioidea and consequently it is included here. 65 From a careful study of the pupal characters available it seems to be more nearly related to the Bucculatrigidae than to any other family. A comparison of Figure 57 or 59 with 67 will show that the develop- ment in the Lyonetiidae has not been towards the shortening of the segments and the consolidation of abdominal segments 8-10 as in the Phyllocnistidae. Moreover, it still retains the generalized type of body found in the Nepticulidae, while the Phyllocnistidae have developed the cylindrical type. The shape of the maxillae and the position of the femur of the prothoracic leg are as in the Bucculatrig- idae, and like them the Lyonetiidae have no labial palpi visible. The Lyonetiidae do not spend their pupal life within the mine, nor in a cocoon, but are exposed and fixed by the caudal end to some cross threads on the under surface of the leaf (Clemens, Tineina of N. America, 1872, pp. 189-191). The soldering down of the appendages and the loss of motion of the abdominal segments seems to be a modi- fication to suit the new conditions of life, and is analogous to the con- dition found in certain families of Papilionoidea and the species of the genus Elachista, in which all power of motion is lost. Bedellia has developed certain ridges and projections, similar to those found in the Papilionoidea, which seem to be correlated with this manner of pupal life. Only two genera of Lyonetiidae were studied. These were from 4-6 mm. in length and may be separated as follows : a. Head blunt, without a prominent projection ; antennae and meta- thoracie legs equal in length ; caudal end of body with a few straight spines on the dorsal surface of the tenth abdominal segment: body withoat prominent ridges Proleucoptera Busck. aa. Head with a long projection; antennae much longer than the meta- thoracic legs ; caudal end of body with hooked setae ; body with prominent ridges Bedellia Stainton. The following species were examined : Proleucoptera suiilaciclla Busck Bedellia somnidentclla Zeller Family Gracilarhdae This large family includes those pupae with free appendages and with abdominal segments 4-7 free in the male and 4-6 in the female. The antennae and metathoracic legs are of approximately the same length and both are longer than the wings. The most nearly related family, the Phyllocnistidae, differs in having on the dorsum of each abdominal segment two prominent pits or punctures with heavily chitinized edges associated with some large curved spines, in having 66 a much more cylindrical body with large deep furrows between the segments, and in having the fixed caudal segments very short. The genus Gracilaria is undoubtedly the most generalized, if we consider the peculiar structures (Fig. 47) found in some species to be maxillary palpi. The tendency in the Gracilariidae is toward the loss of the maxillary palpi, and the development of the triangular type of prothorax, which usually is elevated on the median line. There is also a shortening of the maxillae and labial palpi and of all the ap- pendages in relation to the rest of the body, and a stronger chitiniza- tion of the surface of the body tending to a soldering down of the appendages. There is also taking place the development of two sizes of spines on the dorsum of the abdominal segments and the formation of single rows of larger spines. Finally there is the development of the cremaster. There are two distinct divisions of the Gracilariidae to which subfamily names have been given. These may be separated as follows : a. Prothorax depressed' and neck-like, somewhat quadrangiilar in out- line, the length at the lateral margin never more than twice the mesal length Gracilariinae. aa. Prothorax usually with an elevated ridge on the meson, triangular in outline, the length at the lateral margin about four times the mesal length Lithocolletinae. Subfamily Gracilariinae The Gracilariinae (Figs. 45 and 46) include all the genera in which the generalized quadrangular type of prothorax has been re- tained. In all the genera the caudal end of the body is blunt and the tenth segment bears a row of 6 or 8 spines, larger than those on the other body segments. The labial palpi are always long and never covered by the maxillae at their proximal end. The following table will serve to separate the genera of Gracilariinae : a. Dorsum of abdomen sparsely covered with very coarse spines, some- times with additional fine spines. b. Head with a cutting plate on the ventral surface near the cephalic margin, which is usually serrate ; maxillae as long as the mesotho- racic legs Gracilaria Haworth. bb. Head with a prominent projection at the cephalic end, not a distinct plate; maxillae never as long as the mesothoracic legs. Ornix Treitschke. aa. Dorsum of abdomen thickly covered with very fine spines which are almost invisible Parectopa Clemens. 67 Subfamily Lithocolletinae In the Lithocolletinae all the genera but Acrocercops have a strongly elevated median ridge on the prothorax, and in all but Acro- cercops and Marmara the proximal part of the labial palpi is covered by the maxillae so that the lateral margin can not be traced cephalad to the labrum. The genus Lithocolletis, which seems very distinct from other genera in the subfamily, includes two distinct types of larvae. On this basis the genus was divided into two groups designated as the "flat-larval group" and the "cylindrical-larval group." Dr. Chap- man in 1902 (Entomologist, Vol. 35, p. 141) proposed the name Cameraria for the flat-larval group, and this name is used here as our investigation shows that the cremaster is a decided genus character, and, furthermore, that members of the same genus have the same type of cremaster. It is therefore deemed impossible, from a study of the pupal characters, that one genus could include both forms with and without a cremaster. The pupae of the cylindrical-larval group stud- ied, moreover, showed two distinct types of cremaster, L. hicidicos- tella (Fig. 66a) having a rather broad cremaster with curved setae, while in L. tiliacella and L. argentinotclla the cremaster is long and slender (Figs. 66b and 64) and the setae are T-shaped, the former having one such seta and the latter two. From the standpoint of pupal characters these would properly form three genera. It is in- teresting to note that Meyrick (Genera Insectorum, Part 128) places these in different sections of the genus, and that Miss A. F. Braun in her work on the "Development of the Color Pattern in Lithocol- letis" (Journ. Acad. Nat. Sci. Phila., Vol. 16, Series 2, 1914) also includes them as members of different groups in her phylogenetic tree. The genera of Lithocolletinae may be separated as follows : a. Dorsum of abdominal segments with spines of the same size ; caudal margins of abdominal segments never distinctly elevated, b. Dorsum of each abdominal segment covered with spines for its entire length. c. Maxillae at least seven eighths the length of the wings; labial palpi almost half the length of the maxillae, their proximal end not covered by the maxillae; spines on dorsum of abdo- men very small and inconspicuous except a row of six spines on the tenth segment Acrocercops Wallengren. cc. Maxillae not more than one third the length of the wings ; labial palpi about one third the length of the maxillae, their proximal end covered by the maxillae ; spines on the dorsum of the abdomen small but distinct, with a few larger ones on the tenth segment Leucaniliiza Clemens. 68 bb. Dorsum of each abdominal segment covered with spines for about one fourth its length, c. Head without a prominent pointed projection; maxillae more than half the length of the wings and longer than the pro- thoracic legs; proximal part of the labial palpi not covered by the maxillae Marmara Clemens. cc. Head with a prominent projection ; maxillae never half the length of the wdngs nor as long as the prothoracic legs ; labial palpi covered by the maxillae at the proximal end. Cremastohomhycia Braun. aa. Dorsum of abdominal segments covered with spines of two sizes, the caudal margins of the segments usually distinctly elevated ; labial palpi always covered by the maxillae at the proximal end. b. Caudal end of body never with a distinct cremaster. Cameraria Chapman, bb. Caudal end of body always with a distinct cremaster. LitJiocolletis Hiibner. Species of Gracilariidae examined : Subfamily Gracilariinae Gracilaria negiindella Chambers, sassafrasclla Chambers, violacclla Clemens Ornix priiniz'orclla Chambers, crataegifoliella Clemens, conspic- uclla Dietz Parectopa salicifoliella Chambers, Icspedezacfoliella Clemens Subfamily Lithocolletinae Acroccrcops vcnustclla Clemens Lencantliiza amphicarpcacfoliclla Clemens, ostcnsackcnclla Fitch Marmara salictella Clemens Cremastohomhycia solid aginis Frey and Boll Cameraria hamadryadella Clemens, ostryella Chambers, tuhiferella Clemens Lithocolletis lucidicostclla Clem., argentinotclla Clemens, tiliacella Chambers Family Phyllocnistidae This family is very nearly related to the Gracilariidae, and the principal characters used to distinguish it are given under that family. The arrangement of parts may be seen in Figure 67. It will be noted that Phyllocnistis has long heavily-chitinized setae much as in the Tisclieriidae except that they are not forked at the tip. There is a fleshy prolongation on each side of the tenth abdominal segment. This family shows a somewhat higher degree of specialization in the pro- thorax and labial palpi than most of the Gracilariidae. It is, however, not as much specialized as the species of Lithocolletis which have de- 09 velopcd a cremaster, but is more like Cameraria. It may have been developed from the same stem as Cameraria, but its development is more likely to have been parallel with that of the family Gracilariidae. The body is considerably more chitinized, however, than in any mem- ber of that family. This family includes a single genus, Phyllocnistis Zeller, in which the pupae are from 3-4 mm. in length. The following species were examined : Phyllocnistis ainpclopsisclla Chambers, imignis Frev and Boll. Specialized pupae with pilifers There are tw^o superfamilies of Lepidoptera, the Pyralidoidea and the Papilionoidea, in which the pilifers are enormously developed, and their presence is indicated in the pupa by lobes which extend from the caudo-lateral angles of the labrum towards the meson and in many instances are adjacent on the meson (Figs. 70, 72, 74, 76, 79; pf). Besides the presence of these lobes there are many other points of resemblance which would seem to indicate that these two superfamilies had a common ancestor. SuPERFAMiLY PYRALIDOIDEA This superfamily includes all those pupae which possess lobes in- dicating the presence of well-developed pilifers and which do not pos- sess clubbed antennae. This comprises the family Pterophoridae, the family Attevidae, previously included in the Yponomeutidae, and probably all of the subfamilies of Pyralididae, although pupae of only six of these were examined. The Gallerinae do not possess the lobes indicating the presence of pilifers, and differ in many other respects from most other pyralids. The antennae are long, at least five sixths the length of the wings, and in some instances extend beyond them. The maxillae and meso- thoracic legs are both long and extend to the caudal margin of the wings in most genera. The femora of the prothoracic legs are visible except in some genera of Pterophoridae. In all of these families the appendages are soldered to each other and to the body wall, but in the Pterophoridae they are very slightly soldered and separate readily. The seventh abdominal segment is free in the males of Pterophoridae and Attevidae but fixed in the female. In the Pyralididae it is fixed in both sexes. The families of Pyralidoidea may be separated as follows : a. Maxillaiy palpi never present; the prothoracic and mesotlioracic legs always extending cephalad between the sculptured eye-piece and the antennae ; body always roughened with short spines or with small , 70 groups of long barbed spines and setae arising from small eleva- tions; dorsum of abdomen never with a deep furrow between the ninth and tenth segments Pterophoridae. aa. Maxillary palpi usually present, if not, then the dorsum of the abdo- men with a deep furrow between the ninth and tenth segments; body surface seldom roughened with spines or setae. b. Epicranial suture never present; fronto-clypeal suture visible for about half the distance between the proximal end of the antennae and the meson ; seventh abdominal segment free in the male and fixed in the female ; dorsum of abdomen never with a furrow between segments nine and ten Attevidae. bb. Epicranial suture usually present, if not, then the dorsum of the abdomen with a deep furrow between the ninth and tenth seg- ments ; fronto-clypeal suture never indicated ; seventh abdominal segment fixed in both sexes Pyralididae. Family Pterophoridae This family possesses a curious combination of generalized and specialized characters which make its position rather difficult to deter- mine. It has lost the maxillary palpi, the femora of the prothoracic legs are seldom visible, and the epicranial suture is present in but one genus, Pterophorus, where only a small portion of it is visible. On the other hand the seventh abdominal segment is free in the male and fixed in the female. This is clearly seen at dehiscence, for none of the abdominal segments possess much power of motion. The appendages (Fig. 70) are only slightly soldered to each other and to the body wall, and generally separate very readily. The wings are slender and pointed and, together with the other appendages, project slightly beyond the margin of the fourth abdominal segment. The clypeus, labrum, and sculptured eye-piece each bear tw^o prominent setae in Pterophorus and Oxyptilus but in Platyptilia they are very small. There is usually a seta near the caudal margin of each gena. The proximal portion of each antenna is usually considerably widened and ridged and in Pterophorus and Oxyptilus bears long spines. The prothoracic legs are exceptionally long in this family and reach nearly to the caudal margin of the wings. The maxillae are often overlaid by the prothoracic legs for a part of their length, and sometimes are only visible for a short distance at their proximal and distal ends, the entire mesal portion being concealed. The location of the genital openings is unusual, appearing to be always on the tenth abdominal segment, which extends very far cephalad and forms a sort of ventral plate on the fixed caudal segments. In Platyptilia the plate is not so prominent and the dividing sutures between the segments may be dis- 71 tinguished. At the cephalic margin of this plate is a large group of hooked setae in Pterophorus and Oxyptilus, and in Platyptilia a rounded tubercle bearing four hooked setae. The abdominal spiracles are slightly produced. The mesothoracic spiracle is situated mesad of its usual position. It is also slightly produced. The peculiar spiny armature of most of the genera makes them very easy to distinguish from all other pupae. They are always found exposed, attached by the cremaster, and vary in length from 8-15 mm. The genera may be separated thus : a. Body with long, prominent barbed spines and setae arising mostly from dorsal and lateral elevations; tenth segment with a mass of hooked setae at its cephalic margin, b. Femora of the prothoracic legs exposed; dorsal and lateral eleva- tions with barbed spines of varying lengths. Pteropliorus Cleoffroy. bb. Femora of the prothoracic legs never exposed ; dorsal and lateral elevations usually with two barbed spines which are very broad at base and on the side of each is inserted a stout straight seta, Oxyptilus Zeller. aa. Body without any long barbed spines or setae, but with short, widely separated triangular projections on most of the abdominal seg- ments ; tenth segment with a rounded prominence near the cephalic margin bearing about four hooked setae Platyptilia Hiibner. The following species were examined : Pterophorus paleaceus Zeller Oxyptilus temiidactyhis Fitch Platyptilia carduidactyla Riley Family Attevidae The genus Atteva (Figs. y2, 73), formerly included in the family Yponomeutidae, was found to differ in all its important characters from the members of that family and to be closely allied to the Pyralid- idae. It retains the same arrangement of setae on the clypeus and labrum as that in the Yponomeutidae. The setae at the caudal end of the body are also similar in arrangement to those in. theYponomeu- tidae, but the subfamily Phycitinae also have setae arranged in this way. It seems very probable that the Attevidae and Yponomeutidae arose from a common stock, but that the former branched off before motion was lost in the seventh segment of the male. In the Attevidae there is a narrow conjunctiva between the seventh and eighth segments in the male and there is slight motion possible. The eighth, ninth, and tenth segments are unusually long and distinctly segmented. There is 72 no epicranial suture present, and at dehiscence the eye-pieces are not separated from the other face-parts, which indicates a high degree of speciaHzation. The maxillary palpi are present, but not as well de- veloped as in most pyralids. The labial palpi are represented by a small polygonal area caudad of the lobes indicating the presence of pilifers which meet on the meson. The fronto-clypeal suture is pres- ent for about half the distance between the proximal ends of the an- tennae and the meson and it dehisces for this distance at the emergence of the imago. This family includes the single genus Atteva. The pupae of this family are from 15-20 mm. in length. The following species was examined : Atteva aiirca Fitch. Family Pyralididae This family (Figs. 74, 75, 76) includes a number of subfamilies, of which only six are discussed here. The epicranial suture is present in all of these except the Epipaschiinae and a few genera of Phycitinae but the vertex is very short in all of the others, and often represented by a small triangular area, adjacent to each antenna, which does not reach to the meson. The antennae are long, at least seven eighths the length of the wings and often much longer, and the distal ends never meet on the meson. The labial palpi are visible only as small triangu- lar or polygonal areas except in the Crambinae, which often show a large portion between the halves of the maxillae. The maxillae are always long except in the Gallerinae, usually reaching the caudal mar- gin of the wings and sometimes extending beyond them. The maxil- lary palpi are present in all subfamilies except the Epipaschiinae. Each prothoracic leg is from one half to three fourths the length of the wings and its femur is always exposed. The mesothoracic legs generally extend to the caudal margin of the wings. The abdominal segments never possess spines except in the Gallerinae but are smooth or punctate. The spiracles are of different types, some being slightly produced. The location of the mesothoracic spiracle is difficult to determine in most species, there being no visible opening. The ap- pendages are always firmly soldered to each other and the body wall. The pupae vary in length from 8-20 mm. The following table will serve to separate the subfamilies of Pyralididae : a. Maxillary palpi always present; epicranial suture usually distinct, at least for a part of its length. b. Maxillae never more than three fifths the length of the wings; dorsum of thorax and abdomen with a prominent median ridge and the segments covered with small spines Gallerinae. 73 bb. Maxillae always more than three fifths the length of the wings ; dorsum of thorax and abdomen never with a median ridge or with small spines on the segments, c. C'remaster absent or never long and well developed ; furrows usually present on the dorsum between the ninth and tenth abdominal segments, or on the lateral part of the tenth seg- ment; head usually rounded; body never with a shouldered appearance ; labrum in its normal position, d. Caudal end of body with all the setae straight and very short; cremaster short and blunt; lateral margins of the dorsum of tenth abdominal segment with prominent deep furrows extending caudad to the proximal end of the cremaster; a large portion of the labial palpi often ex- posed Crambinae. dd. Caudal end of body never with all the setae straight, but usually long and hooked ; lateral margins of the tenth ab- dominal segment never with deep furrows unless they are extensions of the dorsal furrow between the ninth and tenth segments ; labial palpi never w ith more than a very small triangular or polygonal area exposed, e. Dorsal furrow, if present between the ninth and tenth abdominal segments, with a crenulate margin; meso- thoracic spiracles never tubular, but slit-like and not plainly indicated; caudal end of body without a cre- master, and bearing a transverse row of six or eight slender hooked setae Pyralinae. ee. Dorsal furrow usually present between the ninth and tenth abdominal segments but never with a crenulate margin, if the dorsal furrow is absent then the meso- thoracic spiracles produced and tubular; cremaster w^anting or very short Phycitinae. cc. Cremaster always present and well developed ; dorsal furrows never present between the ninth and tenth abdominal seg- ments or on the lateral part of the tenth segment ; head blunt ; body with a distinctly shouldered appearance ; labrum always cephalad of its normal position Pyraustinae. aa. Maxillary palpi never present ; epicranial suture never visible for any part of its length ; dorsal furrow between the ninth and tenth ab- dominal segments strongly curved caudad and apparently lined with a fringe of short setae Epipaschiinae. Subfamily Gallerinae The pupae of this subfamily are very different from most pyralids and there is some doubt as to whether they should be included with 74 this family. The lobes which indicate the presence of pilifers are not well developed and the maxillae are short (Fig. 69). The body is short and thick and covered on the dorsum of the thorax and abdo- men with short spines. A strongly elevated median ridge is also present on the thorax and on the first eight abdominal segments. The prothorax is very long, at least half the length of the mesothorax. The following species was examined : Galleria melonella Linnaeus. Subfamily Cramhinac The pupae of the Crambinae are easily recognized by the peculiar form of the short, blunt cremaster and the deep lateral grooves on the tenth abdominal segment. Some of the species show a large portion of the labial palpi, indicating that this subfamily is one of the more generalized. The maxillae reach almost to the caudal margin of the wings and the tips of the mesothoracic legs meet on the meson just caudad of the maxillae. The segments are almost smooth, never punctate. The following species were examined : Crambiis vnlgivagellus Clemens, trisectus Walker, caliginosclliis Clem- ens. Subfamily Pyralinae The species of this subfamily resemble closely the genera Plodia and Ephestia of the Phycitinae. There scarcely seems to be more than generic differences between them. The epicranial suture is present and the vertex always extends to the meson. The maxillary palpi are well developed and usually reach the proximo-lateral angles of the maxillae. There is never a cremaster present, but a transverse row of hooked setae at the caudal end of the body. The two genera studied may be separated as follows : a. Dorsum of abdomen with a furrow between the ninth and tenth seg- ments, the caudal margin of the furrow distinctly erenulate. Pyralis Linnaeus. aa. Dorsum of abdomen without a furrow between the ninth and tenth segments Hypsopygia Hiibner. The following species were examined : Pyralis farinalis Linnaeus Hypsopygia costalis Fabricius 75 Subfamily Phycitinae This group is, for the most part, easily distinguished from other pyralids by the presence of the suture on the dorsum of the abdomen bet^^■een the ninth and tenth segments, the presence of maxillary palpi, and, usually, of the epicranial suture. Of the genera examined, Ephestia and Plodia alone were without this dorsal furrow, and they possess tubular spiracles on the mesothorax. These two genera seem rather more closely related in many respects to the Pyralinae than to the Phycitinae. The maxillary palpi always extend to the proximo- lateral angles of the maxillae, and the epicranial suture is present in all genera examined but Pinipestis and Mineola, though it is very near to the suture between the head and prothorax. The vertex is usually represented by a small triangular area adjacent to each antenna. The lobes enclosing the pilifers meet on the meson in some genera. The genera of Phycitinae may be separated as follows : a. Dorsal surface without a prominent furrow separating the ninth and tenth abdominal segments ; mesothoracie spii-acles tubular, b. Abdominal segments punctate; maxillae reaching the caudal mar- gin of the wings Ephestia Guenee. bb. Abdominal segments smooth ; maxillae never reaching the caudal margin of the wings Plodia Guenee. aa. Dorsal surface with a prominent furrow separating the ninth and tenth abdominal segments; mesothoracie spiracles never tubular, their exact position usually difficult to deteraiine. b. Body depressed ; tenth abdominal segment with the caudal end dis- tinctly margined and wdth six straight setae inserted on the ven- tral side of the margin Acrohasis Zeller. bb. Body never depressed ; tenth abdominal segment never with the caudal end distinctly margined or with, setae inserted on the ventral side of the margin, c. Caudal end of body wdth four long hooked setae, and on each side of these a short spine or hooked seta extending laterad. d. Tenth abdominal segment with lateral spines very different from the caudal setae, e. Ninth abdominal segment with a lateral spine on each side similar to those on the tenth segment, and two hooked setae on the dorsum adjacent to the caudal mar- gin; caudal hooked setae equidistant. . .Mineola Hulst. ee. Ninth abdominal segment without lateral spines or hooked setae on the dorsum adjacent to the caudal margin, f. Caudal spines not adjacent; equidistant; of equal length Meropiera Grote. 76 ff. Caudal spines adjacent ; two of them shorter than the other two Psorosina Dyar. dd. Tenth abdominal segment with lateral hooked setae similar to the caudal setae Canarsia Hulst. cc. Caudal end of body with a transverse row of six long hooked setae of equal length; epicranial suture never present; head wdth a prominent pointed cephalic projection. Pinipestis Grote. The following species were examined : Plodia inter punctella Hiibner Bphestia kuehniella Zeller Acrobasis ruhrifasciella Packard Mincold indigindla Zeller Meroptera pravella Grote Psorosina hammondi Riley Canarsia nhniarrosorclla Clemens Pinipestis zimmcrmani Grote Subfamily Pyraiistinae This group is distinguished by the peculiar "shouldered" appear- ance of the body, caused by the great width of the thorax as compared with the head and by the position of the labrum, which is always cephalad of its normal position and often located near the cephalic end of the body. There is never a furrow on the dorsum between the ninth and tenth abdominal segments. The maxillary palpi are always present and only a very small portion of the labial palpi is exposed. The epicranial suture is present in all genera. The mesothoracic legs and antennae, together with the metathoracic legs which are hidden by them, usually extend beyond the caudal margin of the wings. The mesothoracic spiracles often have peculiar ridges along their caudal margin which are sometimes covered with setae. Similar ridges are found in certain families of Papilionoidea and Notodontoidea. The shape of body and arrangement of parts in the Pyraustinae resembles that of certain Sphingidae, and would seem to indicate that the Pyraustinae are not as closely related to the Phycitinae as the other subfamilies, which all show a very close relationship. The genus Pyrausta as understood at present probably does not represent a natural group. Of the species studied P. fissalis and P. illihalis have long narrow cremasters of similar type, while P. fntilalis and P. in- sequalis have short broad cremasters of rather different types. There is also great variation in the length of the appendages, but this is not such a decided generic character as the form of the cremaster. This 77 subfamily includes the largest pyralids examined. The genera of Pyraustinae may be separated by the following table : a. Setae of the cremaster always hooked and equal in length to the ere- master or sometimes longer; the other appendages never extending beyond the caudal margin of the wings, b. Setae of the thorax and abdomen very long, heavily chitinized, and forked at the distal end, usually much longer than the seg- ments; mesothorax and metathorax having a deep oblong pit with strongly chitinized edges at the base of each wing. PJilyctaenia Hiibner. bb. Setae of the thorax and abdomen never prominent, scarcely ever visible ; mesothorax and metathorax never having a deep oblong pit at the base of each wing. c. Mesothoracic spiracle with a prominent elevation adjacent to the caudal margin, which bears several ridges fringed with setae ; front with a distinct tubercle or small ridge at the base of each antenna Desmia Westwood. cc. Mesothoracic spiracle without any prominent elevation ad- jacent to the caudal margin ; front without a tubercle or ridge at the base of each antenna Pantagraplia Lederer. aa. Setae of the cremaster either straight and equal in length to the cremaster, or hooked and much shorter than the cremaster ; the other appendages often extending beyond the caudal margin of the wings. b. Prothorax with a distinct tubercle on each side of the meson ; cre- mastral setae straight and spread out fan-like. TJioleria Hiibner. bb. Prothorax without a distinct tubercle on each side of the meson ; cremastral setae hooked, and not spread out fan-like. Pyrausta Schrank. The following species were examined : Phlyctaenia ferriigalis Hiibner Desmia funcralis Hiibner Pantagraplia limata Grote and Robinson TJioleria reversalis Guenee Pyrausta fissalis Grote, illihalis Hiibner, futilalis Lederer, inseqiialis Guenee Subfamily Bpipaschiinae Only one species of this group has been examined, so no very definite statements can be made regarding it. The species examined seems to differ mainly in the absence of the maxillary palpi, which are present in all of the other subfamilies. The epicranial suture is not visible and the labrum is slightly cephalad of its normal position. The -dorsum of the abdomen shows a decided furrow between the ninth 78 and tenth segments which is strongly curved caudad and apparently covered with some fine whitish setae, but this appearance may be due to the fine striations present. The caudal margin of the furrow is crenu- late. There is a very short cremaster present, bearing a small group of hooked setae, which are more than half as long as the tenth segment. The following species was examined : Lanthape platanella Clemens. SUPERFAMILY PAPILIONOIDEA The members of this superfamily are distinguished by the pos- session of lobes indicating the presence of well-developed pilifers and by their distinctly clubbed antennae. The genus Oeneis is an excep- tion, however, in not having the lobes well developed ; but this is probably due to specialization, as it seems very closely allied to the Satyrinae; especially in the length of the prothoracic legs. Many of the Papilionoidea have prominent ridges and tubercles on the surface of the body, but there are also many genera in which the body surface is quite smooth and destitute of tubercles and ridges. The epicranial suture is present in three families, Megathymidae, Hesperiidae, and Lycaenidae. There has been a great deal of discussion and disagree- ment over the arrangement and subdivision of the families of the Papilionoidea. Some have divided it into two superfamilies, Hes- perioidea and Papilionoidea, but the pupae show no characters to war- rant such a division. The family Lycaenidae has been considered by many as the most specialized, or among the most specialized, of the families, yet it still retains the epicranial suture. In this family, how- ever, the labial palpi are entirely concealed except in the case of the aberrant genus Feniseca, and the shortening of the prothoracic legs is similar to the condition found in the Nymphalidae. It is impossible without further study of existing forms and a larger series of species to discuss fully the relationships between the different families. It is sufficient for the present to state that the Lycaenidae seem more nearly related to the generalized Hesperiidae, but have developed in a similar manner to the Nymphalidae, and that the Pieridae, Papilionidae, and Nymphalidae seem very closely related. The families of Papilionoidea may be separated as follows : a. Proximo-lateral angles of the maxillae extending laterad to the eye- pieces, b. Maxillae never reaching the caudal margin of the wings ; wings adjacent on the meson caudad of the maxillae. .Megathymidae. 79 bb. Maxillae always extending to the caudal margin of the wings and sometimes beyond; wings never adjacent on the meson. Hesperiidae. aa. Proximo-lateral angles of the maxillae never extending laterad to the eye-pieces, b. Mesothoracic legs never extending cephalad to the eye-pieces. c. Epicranial suture always present; head without projections; exposed part of maxillae never as long as the wings. Lycaenidae. ec. Epicranial suture never present ; head always with prominent projections; exposed part of maxillae usually as long as the wings. d. Head with two prominent projections, one at each cephalo- lateral angle ; metathoracic wings visible in ventral view. Papilionidae. dd. Head with a median projection ; metathoracic wings not visible in ventral view Pieridae. bb. Mesothoracic legs extending cephalad to the eye-pieces and for a short distance between the sculptured eye-piece and the antenna. Nymphalidae. Family Megathymidae The Megathymidae, or giant skippers, are evidently the most gen- eralized of the Papilionoidea although they differ but little from the more generalized Hesperiidae, some doubt existing as to whether they show enough difference to warrant their being considered as a dis- tinct family. However, as only one pupa of this family has been seen by the writer and as that had lost some of the face-parts so that a complete description can not be given, no very definite stand can be taken here as to its position in the superfamily. The members of the superfamily Papilionoidea, as a general rule, possess but little free- dom of motion in the free segments and these are rarely capable of being telescoped. In the Megathymidae not only are the free seg- ments capable of a great deal of motion and of being telescoped, but there appears to be dorsal motion possible between the third and fourth abdominal segments, and the seventh abdominal segment ap- pears to possess freedom of motion in the male. The abdominal seg- ments are of nearly equal length, and the eighth, ninth, and tenth are distinctly segmented. These characters, however, appear to be re- tained in such generalized Hesperiidae as the genera Calpodes and Amblyscirtes, in which if all the above-mentioned segments do not re- tain freedom of motion they have certainly but recently lost it. In placing this family in the Papilionoidea it has been assumed that the pupae possess lobes indicating the presence of pilifers, but these parts 80 were absent in the pupa examined. The labial palpi are represented by a small triangular area, but it is not known whether or not maxil- lary palpi are retained. The maxillae are much shorter than in the Hes- periidae, being only about two thirds the length of the wings, but this indicates nothing as to their position, as both generalized and special-^ ized pupae possess short maxillae. None of the other appendages are longer than the maxillae except the wings, which lie adjacent on the meson caudad of the maxillae. The epicranial suture is present and the vertex is of equal length throughout, being about one fifth the length of the prothorax measured on the meson. The entire body sur- face is covered with a whitish bloom, and on the dorsum of abdominal segments 7-10 there is in addition a dense covering of rather coarse setae. The pupa examined was 40 mm. in length and about 10 mm. in breadth, and belonged to the genus Megathymus. The following species was examined : Megathymus yiiccae Boisduv^al and Le Conte. Family Hesperiidae The Hesperiidae retain considerable freedom of motion of the ab- dominal segments, and in many genera it would seem that dorsal motion is possible between the third and fourth abdominal segments and that the seventh abdominal segment is free in the male, or at least that they have only recently lost the power of motion. The epicranial suture is present in all genera, and the vertex is about one fifth the length of the prothorax measured on the meson, while the lateral mar- gins are considerably longer. The labrum in most genera is consid- erably cephalad of its normal position. The antennae never reach to the caudal margin of the wings but are from two thirds to three fourths of their length. The prothoracic legs are about half the length of the wings, the mesothoracic usually two thirds of the same, while the metathoracic pair is seldom visible. The maxillae always extend to the caudal margins of the wings and frequently consider- ably beyond. The mesothoracic spiracles usually have a peculiar kind of plug or plate which seems to form an external closing apparatus or guard, while some have prominent tubercles caudad of the opening, usually with a dense covering of setae. The thorax and abdomen usually have a more or less dense covering of setae, and some of the species have the entire body covered with a whitish bloom, which is of comparatively rare occurrence among lepidopterous pupae. The abdomen frequently has a furrow on the dorsum between the ninth and tenth segments, similar to the furrows found in the Pyralididae but never so deep. The cremaster in all genera is more or less trian- 81 gular, with hooked setae on the distal end, and frequently has an im- pressed triangular area on the dorsum. The classification of the Hes- periidae has long been in dispute, and with the limited amount of material available for examination it is impossible to state just how a classification of the pupae would agree with any of the proposed schemes. It is believed, however, that Scudder's arrangement would probably be followed, as the material available falls readily into his groups. As to the relationship between these groups there might be some difference of opinion. The pupae at first sight are readily divid- ed into two groups, one with the abdominal segments caudad of the fourth considerably shortened, possessing narrow flanged plates on the movable segments which prevent the telescoping of the body, and with the segmentation indistinct between the fixed caudal segments (Fig. yj). This group also has the body prominently convex on the dorsum of the mesothorax and on the entire ventral surface of the thorax and abdomen. The labrum is cephalic in position. The other group possesses abdominal segments of more nearly ecjual length, hav- ing distinct sutures between the fixed caudal segments and the mov- able segments capable of being telescoped. This group has apparently just recently lost the power of motion in the seventh abdominal seg- ment of the male and dorsal motion between the third and fourth abdominal segments. The body is shaped like the majority of lepidop- terous pupae, and the labrum never quite reaches the cephalic margin of the body. Of this group, the genera possessing maxillae extending beyond the caudal margin of the wings, Calpodes (Fig. 78) and Am- blyscirtes, are undoubtedly more generalized, not on account of the maxillae, but because in all the other members of the group there is considerably more consolidation of the caudal abdominal segments, so that they seem intermediate in position between the genera men- tioned above and the first group. The following table will serve to separate the genera of Hesperiidae : a. Abdominal segments 5-7 never with an elevated ridge or flanged plate along the cephalic margin and always capable of being tele- scoped; body never with a prominent convexity on the ventral sur- face in the region of abdominal segments 1-4. b. Maxillae extending free for a considerable distance beyond the caudal margin of the wings. c. Maxjllae extending beyond the caudal margin of the body ; head with a long cephalic projection Calpodes Hiibner. cc. Maxillae never extending beyond the caudal margin of the body ; head never with a long cephalic projection. Amhlyscirtes Scudder. 82 bb. Maxillae never extending free beyond the caudal margin of the wings. c. Body with a dense covering of long setae and whitish bloom; mesothoracic spiracle with a strongly elevated oval area adja- cent to its caudal margin, this area chitinized in the center and surrounded by a dense band of short setae with a longitudinally striate chitinized rim forming an outer margin. PJiolisora Scudder. cc. Body sparsely covered with short inconspicuous setae ; dense whitish bloom never present ; mesothoracic spiracle with a some- what circular elevation adjacent to its caudal margin, which is entirely covered with setae TJianaos Boisduval. aa. Abdominal segments 5-7 with an elevated ridge or flanged plate along the cephalic margin which prevents their being telescoped; body with a prominent convexity on the ventral surface in the region of abdominal segments 1-4. b. Mandibular area with distinct tubercles which are usually black and bearing stout setae; head slightly narrower than the meso- thorax. c. Mesothoracic spiracle semicircular in outline, the opening circu- lar, a broad thick band of setae around the caudal half, d. Dorsal furrow or depression on the ninth abdominal segment with its caudal margin distinctly crenulate; ventral surface of cremaster with an elongate furrow broadened out at the distal end of the cremaster Tlioryhes Scudder. dd. Dorsal furrow or depression on the ninth abdominal segment not distinctly crenulate ; ventral surface of cremaster with a triangular depression broad at the proximal end and nar- rowed to the distal end of the cremaster. Epargyrcus Hiibner. cc. Mesothoracic spiracle semicircular in outline, the opening cir- cular and surrounded by a broad thick band of setae, and caudad of this a distinctly elevated chitinized ridge forming an outer margin Cocceius Godman and Salvin. bb. Mandibular area smooth, without tubercles ; head as broad as the mesothorax Eudamus Swainson. The following species were examined : Calpodes ethlins Cramer Amblyscirtes vialis Edwards Pholisora catullus Fabricius Thanaos hrizo Boisduval and Le Conte, luciliiis Lintner Thorybes daunus Cramer Bpargyreus tityrus Fabricius Cocceius pylades Scudder Budamus proteus Linnaeus 83 Family Lycaenidae The Lycaenidae are small pupae, between 8 and 15 mm. in length, which have the general shape of arctians although they are generally less curved on the ventral surface (Fig. 79). They retain a small portion of the vertex on each side and the epicranial suture usually touches the caudal margin of the head at the meson, making each half of the vertex triangular. The lobes indicating the presence of pilifers always meet on the meson except in the genus Feniseca. The antennae always extend to the caudal margin of the wings and lie adjacent on the meson, concealing the distal ends of the maxillae. The prothoracic legs are shorter than usual, varying from two fifths to one third the length of the wings. The mesothoracic legs are about half the length of the wings and the metathoracic pair are never visible. The body is usually quite free from projections or elevations, Feniseca being the only exception known, and it bears small rounded tubercles on its dorsal surface. The head is limited to the ventral surface of the body, and the suture between it and the prothorax is located on the cephalic margin of the body, sometimes forming a slight ridge. The prothorax is longer than is usual in Papilionoidea, being about half as long as the mesothorax. There is little, if any, motion possible be- tween any of the abdominal segments, and they fit together so as to form a smooth surface. Even the pupal skin after dehiscence shows no separation of the abdominal segments. The surface of the thorax and abdomen is covered with a reticulation of fine elevated lines with small papillae at their intersections and usually in the spaces between. These papillae usually bear cuticular appendages, of various types, the most peculiar being the fungiform type of the genera Chrysopha- nus and Heodes. There is no cremaster present in any member of the family. The ventral surface of the abdomen frequently bears groups of small hooked setae. The genital openings are usually ob- scured. The anal opening is peculiar, in many forms being transverse instead of longitudinal. The mesothoracic spiracles are closed by a plug or plate which fills up the opening and usually presents a honey- combedn appearance. The following table will serve to separate the genera of Lycaenidae : a. Exposed portion of maxillae never more than three fifths the length of the wings ; cuticular appendages of the body never fungiform. b. Ventral surface of the body never with hooked setae eaudad of the anal opening; thorax and abdomen usually densely covered with spiculate cuticular appendages ; exposed portion of maxillae scarcely more than half the length of the wings. 84 c. Ventral surface of ninth abdominal segment with a group of hooked setae on each side of the meson. d. Thorax with the median line slightly elevated; raised lines of the reticulations very prominent ; papillae short, cylin- drical hicisalia Minot. dd. Thorax with the median line never elevated ; raised lines of reticulations not prominent ; papillae conical. Uranotes Scudder. cc. Ventral surface of ninth abdominal segment never with hooked setae on each side of the meson. d. Ventral surface of ninth abdominal segment with a group of straight cuticular appendages on each side of the meson ; setae on bodj^ very dense and about one fourth the length of the abdominal segments ; papillae conical. Mitura Scudder. dd. Ventral surface of ninth abdominal segment without cutic- ular appendages of any kind; setae of body not dense, but long, usually half the length of the segment ; papillae short, cylindrical Thecla Fabricius. bb. Ventral surface of body with hooked setae caudad of the anal opening; papillae most numerous in the spiracular region; thorax and abdomen sparsely covered with short cuticular ap- pendages with very minute spicules; exposed portion of max- illae three fifths the length of the wings. c. Papillae of the spiracular region confined to an area caudad of the spiracle ; spiracle of the second abdominal segment not adjacent to the wing Cyaniris Dalman. cc. Papillae of the spiracular region surrounding the spiracle ; spiracle of the second abdominal segment adjacent to the wing, the distance between them less than the width of the spiracle Rusticus Hiibner. aa. Exposed portion of the maxillae more than three fifths the length of the wings. b. Body of typical lycaenid shape, never flattened at the caudal end ; cuticular appendages fungiform. c. Fungiform cuticular appendages small and inconspicuous, not visible with a low-power lens; color light yellowish brown, not spotted Clirysoplianus Hiibner. cc. Fungiform cuticular appendages large and conspicuous, easily visible with a low-power lens; color dark brown with black spots Heodes Dalman. bb. Body with the caudal end flattened and curved slightly ventrad. Feniseca Grote. The following species were examined : Incisalia niphon Hiibner Uranotes melinus Hiibner 85 Mitura danion Cramer Thecla acadica Edwards, calaniis Hiibner, liparops Boisdiival and Le Conte Cyaniris ladon Cramer Rusticus scudderi Edwards Chrysophaiiits t/ioe Boisduval Heodes Jiypophleas Boisduval Feniseca tarquinius Fabricius Family Papilionidae The pupae of this family are usually long and slender, tapering gradually to the pointed caudal end, which is called the cremaster although it seldom resembles a true cremaster, and extends very little beyond the anal opening. The body always has two prominent cephalic projections, one at each cephalo-lateral angle of the head, a less promi- nent lateral projection on each side the metathorax at the base of each wing, and a low median carinate ridge which extends along the pro- thorax on to the mesothorax for about half its length, where it forms a more or less prominent projection. From this prominence the ridge divides, the remainder of its course being on the metathorax, the divisions sometimes extending to the abdomen to form the dorso- lateral abdominal ridges. There is also usually present a lateral ridge on each side. These four ridges are continued to the end of the body, and are often present on the tenth segment or on the cremaster when absent from the remainder of the abdomen. On the ventral surface there is usually a ridge on each side of the face-parts, beginning at the cephalic projections and extending to the proximo-lateral angles of the maxillae. The labruni is in its normal position and the lobes which indicate the presence of pilifers seldom meet on the meson, but are separated by a small portion of the labial palpi. The epicranial suture is never present, and the proximal ends of the antennae approach each other very closely on the dorsal surface of the head. The antennae never extend as far caudad as the wings. The wings are usually somewhat pointed on the ventral surface near the meson and the metathoracic wing is always visible here, extending for a considerable distance caudad of the mesothoracic wings. The maxillae always extend to the caudal margin of the wings. The legs are of the length usual in lep- idopterous pupae with the exception of the genus Iphiclides, in which the prothoracic legs only are about the usual length, the mesothoracic legs ending before the former — a very rare occurrence in this order. The genital openings are located in the usual positions, those of the 86 female being confluent on the ventro-meson of the eighth and ninth abdominal segments. Just caudad of the genital openings, at the cephalic margin of the tenth abdominal segment is a small tubercle on each side of the meson closely appressed to the surface of the body. The caudal end of the body bears a mass of very short hooked setae. The fourth, fifth, and sixth segments are movable, although they fit closely together to form an even surface and are not capable of telescoping. At dehiscence they separate, showing deep incisions. The genera of Papilionidae may be separated as follows : a. Body surface without distinctly carinate ridges; dorsal surface of abdomen always with a row of small rounded tubercles on each side of the meson, the largest on segments 4-7, and usually a row of smaller tubercles on each side of the spiracles ; ventral surface with two distinct tubercles on each leg, a transverse row near the caudal margin of the mesothoracic wings, and often on the veins near the margin Papilio Linnaeus. aa. Body surface with distinctly carinate ridges, but never with small rounded tubercles on the abdomen, occasionally with very minute tubercles on the wings and other appendages, b. Body without prominent lateral expansions of the abdominal seg- ments or dorsal carinate ridges, c. Body with a very low dorso-mesal ridge on the thorax with a small mesothoracic elevation; a prominent carinate ridge at each lateral margin of the body and no dorso-lateral ridge; cephalic projections large and prominent ; body very strongly convex on the ventral surface in the region of the wings. Euplioeades Hiibner. ce. Body with a low dorso-mesal ridge on the thorax ending in a very prominent mesothoracic elevation ; a very low dorso- lateral and lateral ridge on each side of the abdomen ; cephalic projections not very large ; body never strongly convex on the ventral surface IpliicUdes Hiibner bb. Body with prominent lateral expansions of the fii'st four abdom- inal segments, making this the widest part of the body ; abdom- inal segments with dorsal carinate ridges on each side of the meson, most prominent on segments 5-7, which are highest in the middle of each segment and curve to each margin, giving it a scalloped appearance in lateral view Laertias Hiibner. The following species were examined : Laertias philenor Linnaeus IpliicUdes ajax Linnaeus Biiphoeadcs troilns Linnaeus, palamedes Drury Papilio daunus Boisduval, eiirymedon Boisduval, rutuhis Boisduval, glaucus Linnaeus, polyxenes Fabricius, thoas Linnaeus, machaon Linnaeus, zolicaon Boisduval, brevicauda Saunders 87 Family Pieridae The pupae of this family resemble the Papilionidae very strongly as to the general shape of the body and arrangement of ridges and projections. They are much smaller, however, and are easily recog- nized by the fact that they possess a single median cephalic projection instead of two cephalo-lateral projections as do the Papilionidae. The epicranial suture is never present. The labrum is usually slightly cephalad of its normal position and a small portion of the labial palpi is always exposed. The maxillae vary in length from two thirds the length of the wings to extending slightly beyond their caudal margin. The legs are of normal length. The antennae are more distinctly clubbed than in the Papilionidae and sometimes reach the caudal mar- gin of the wings. The caudal end of the body is very like that of the PapiHonidae except that the four ridges are seldom present, and the hooked setae are inserted in a slight concavity. The genital openings are in the usual positions. On the ventral surface of the tenth abdom- inal segment there is a low ridge, circular in outline, which encloses the anal opening and terminates at its cephalic end in a small tubercle on each side of the meson. These tubercles are located just caudad ■oi the genital openings. Similar tubercles and ridges are found in the Nymphalidae, but they are rather more prominent in that family. Tubercles without the ridges occur in the Papilionidae. There is very little question as to whether or not the Pieridae and Papilionidae are related, but which is the more specialized seems to be a questionable point with all workers in the group. Aside from the question of prominences or projections, which, after all, seems a matter of small importance, there is little of fundamental difference between the two families excepting the length of the thoracic seg- ments, which are more generalized in the Papilionidae, and the ridges and tubercles just mentioned on the ventral surface of the Pieridae, which resemble the Nymphalidae. The Nymphalidae certainly seem to be the most specialized of the Papilionoidea, although this is an- other much debated question. The two families have undoubtedly been developed from a common ancestor and represent parallel lines of development. The genera of Pieridae may be separated by the following table : a. Distance from the cephalic margin of the prothorax to the distal end of the cephalic projection much less than the length of the pro- thorax ; ventral line of body often convex but never forming a prom- inent angle. b. Thorax with a strongly earinate median ridge, highest at the mid- dle of the mesothorax, and forming a prominent projection; ab- domen with a lateral carinate ridge on each side, forming two prominent projections on the second and third segments, the latter more prominent ; a median carinate ridge from the fourth abdominal segment to the caudal end of the body ; ventral line of body practically straight Pontia Fabricius. bb. Thorax without a strongly carinate median ridge, either without a median ridge or with one of equal height throughout, c. Ventral surface of body convex, but without any prominent rounded projection; a low lateral ridge present along the wings, extending on the abdomen to the caudal end of the body. Eiirymns Swainson. cc. Ventral surface of body produced into a prominent rounded projection which, near the caudal margin of the wings, is as wide as the body just caudad of the wings ; body without any prominent ridges, a lateral ridge present along the wings but scarcely indicated on any of the abdominal segments except the tenth Eurema Hiibner. aa. Distance from the cephalic margin of the prothorax to the distal end of the cephalic projection about equal in length to the thorax ; ven- tral line of body forming a prominent obtuse angle at a point about equidistant from the cephalic and caudal ends . . SyncJiloe Hiibner. The following species were examined : Pontia protodicc Boisduval and Le Conte, rapac Linnaeus Eurymits pJiilodice Godart Burenna nicippe Cramer SyncJiloe genntia Fabricius Family Nymphalidae The members of this family liave been variously subdivided. Some writers make several families of the species included here, while others divide them into subfamilies and tribes. At present no good charac- ters are known for the division of this group into families, but it must be admitted that the same difficulties lie in the way of dividing it into the subfamilies and tribes proposed by Scudder; consequently, several subfamily names have been introduced here to facilitate the grouping of the species. The Nymphalidae are distinguished from all other families lacking the epicranial suture by the fact that both prothoracic and mesothoracic legs extend cephalad to the eye-pieces and the meso- thoracic legs extend for a short distance between the sculptured eye- pieces and the antennae. The prothoracic legs are very short, rarely more than one third the length of the wings. The antennae and maxillae, except in a few instances, reach to the caudal margin of the 89 wings. The proximal ends of the antennae extend ahnost to the meson on the dorsum of the head. The labial palpi are represented by a very small portion caudad of the labrum and in many cases are entirely concealed. With the exception of Anaea andria the meta- thoracic wings are not visible on the ventral surface. The genital openings are in the usual position. The circular furrow enclosing the anal opening with the small tubercles caudad of the genital openings, is present in nearly all genera. When tubercles are present on the sur- face of the body they are usually on the dorsum of the abdomen and are arranged in seven rows, as follows: a dorso-mesal row (Fig. 8i, dint), a dorso-lateral row on each side about half-way between the meson and the spiracles (Fig. 8i, dlt), and a dorsal (Fig. 8i, dst), and ventral row (Fig. 8i, vst) on each side of the abdominal spiracles. The subfamilies of Nymphalidae may be separated as follows : a. With prominent tubercles on the dorsal surface of the body, or at least on the abdomen, a dorso-mesal row, a dorso-lateral row on each side, and a row dorsad and ventrad of each row of spiracles. Nymphalinae. aa. Without prominent tubercles on the dorsal surface, at least not ar- ranged in rows as above, b. Second abdominal segment with a prominent carinate median elevation, somewhat constricted at its base Basilarchinae. bb. Second abdominal segment without a prominent, carinate median elevation, c. Body compressed, with a distinct dorso-mesal carina on the thorax and abdomen; ventral surface of ninth and tenth ab- dominal segments with hooked setae, the tubercles on the ninth segment covered with very short hooked setae .... Apaturinae. cc. Body not compressed ; dorso-mesal carina never present on both thorax and abdomen, d. Abdominal segments caudad of the wings rapidly tapering and forming a sort of hemisphere; dorsum of abdomen with a prominent transverse ridge, e. Head with a prominent transverse ridge, extending along the middle of the eye-pieces and the lateral margin of the body; second, third, and fourth abdominal segments of approximately the same length ; cremaster directed ven- trad ; transverse ridge on the fourth abdominal segment. Anaeinae. ee. Head without a transverse ridge, but with two prominent tubercles; second abdominal segment longer than any of the others ; cremaster directed caudad ; transverse ridge on the third abdominal segment tuberculate for its entire length Euploeinae. 90 dd. Abdominal segments caudad of the wings not rapidly taper- ing to form a hemisphere ; dorsum of abdomen never with a transverse ridge, e. Mesothorax prominently elevated ; head with a transverse ridge forming slightly produced cephalo-lateral angles; cremaster with hooked setae Satyrinae, ee. Mesothorax not prominently elevated ; head never with a tranverse ridge ; caudal end of body without hooked setae ; cremaster never present Oeneinae. Subfamily Nymphalinae This subfamily includes all the genera with prominent tubercles on the surface of the body. There are usually seven rows of these, mostly on the dorsal surface of the abdomen, as follows : a dorso- mesal row; on each side of this a dorso-lateral row; and a row^ dorsad and ventrad of the abdominal spiracles on each side (Fig. 8i). The majority of species have a cephalo-lateral projection on each side of the head ; in some these are very prominent ; in others, reduced to small rounded tubercles or wanting. The body is usually strongly con- vex near the caudal margin of the wings on the ventral surface, and the cremaster is curved ventrad. The cremaster is more prominent in this subfamily than in the family Papilionidae and bears a mass of short hooked setae at its distal end. The species of Nymphalinae have been grouped into three tribes, mostly according to the size and arrangement of the tubercles. These three tribes may be separated as follows : a. Dorso-mesal tubercles smaller than those of the dorso-lateral rows; cremaster never with prominent lateral projections at the base, b. Cremaster longer than broad ; dorso-mesal tubercles always pres- ent on abdominal segments 3-8 and usually on the second seg- ment Vanessini. bb. Cremaster usually broader than long ; dorso-mesal tubercles often wanting and never present cephalad of the fourth segment. Argynnini. aa. Dorso-mesal tubercles equal in size to those of the dorso-lateral rows ; cremaster always with a lateral projection on each side at the base. Melitaeini. Tribe VANESSINI This tribe includes the species wnth all the rows of tubercles repre- sented and most of them complete. The tubercles of the dorso-mesal row are considerably smaller than those of the dorso-lateral row, which are usually very prominent. The rows of tubercles on either 91 side of the spiracles are always very small. The cremaster is long and never has prominent lateral tubercles at its proximal end. The genera of Vanessini may be separated by the following table : a. Cephalic prominences conical, with length and breadth approximately equal ; dorso-lateral tubercles on the fourth abdominal segment al- ways larger than the others. b. Dorso-lateral tubercles on abdominal segments 2-7 long and sharp, spine-like, the length considerably greater than the breadth; dorso-mesal tubercle absent on the second abdominal segment. Euvancssa Scudder. bb. Dorso-lateral tubercles on abdominal segments 2-7 not sharp and spine-like ; the length scarcely, if any, greater than the breadth ; dorso-mesal tubercle present on the second abdominal segment. c. Dorso-lateral tubercles on fourth abdominal segment at least twice the size of the others ; median elevation of the mesothorax a compressed carinate ridge and usually very prominent. Polygonia Hiibner. cc. Dorso-lateral tubercles on fourth abdominal segment very little larger than the others ; median elevation of the mesothorax pyramidal and not very prominent Aglais Dalman. aa. Cephalic prominences usually blunt, the length less than the breadth ; dorso-lateral tubercles on the fourth abdominal segment never larger than the others. b. Cephalic prominences broadly rounded; scarcely elevated beyond the outline of the body; no distinct prominence on the median line of the mesothorax Junonia Hiibner. bb. Cephalic prominences distinctly elevated beyond the outline of the body ; a distinct prominence on the median line of the meso- thorax Vanessa Fabricius. The following species were examined : Euvancssa antiopa Linnaeus Polygonia intcrrogationis Fabricius, couinia Harris, fauniis Edwards, prognc Cramer Aglais milberti Godart Junonia coenia Hiibner Vanessa atalanta Linnaeus, huntcra Fabricius, cardiii Linnaeus Tribe AEGYNNINI The species included here resemble those of the preceding tribe, excepting that the dorso-mesal row of tubercles is only present on a few segments or is entirely wanting. The cremaster is short and broad and never has a prominent projection on each side at the prox- imal end. The genera of Argynnini may be separated by the following table : 92 a. Dorso-lateral row of tubercles of approximately equal size. b. Dorso-mesal row of tubercles present on abdominal segments 4-7 r dorso-lateral tubercles much larger than the stigmatal rows ; cari- nate ridge present on mesothorax Argynnis Fabricius. bb. Dorso-mesal row of tubercles absent on all segments ; dorso-lat- eral tubercles about equal in size to the dorsal stigmatal row ; mesothorax without a carinate ridge Euptoieta Doubleday. aa. Dorso-lateral row of tubercles of different sizes, the largest on the third abdominal segment, b. Body with a very strong ventral curve opposite the third and fourth abdominal segments ; mesothorax wdth a strongly ele- vated median ridge througliout its length; head projections very prominent, irregularly bilobed. Agraulis Boisduval and Le Conte. bb. Body without a strong ventral curve ; mesothorax with a small ridge on the caudal half; head projections short, pointed. Brenfliis Hiibner. The following species were examined : Argynnis cybele Fabricius Euptoieta claudia Cramer Agraulis vaniUac Linnaeus Brcnthis niyrina Cramer Tribe MELITAEINI The species in this group have the dorso-mesal and dorso-lateral tubercles of approximately equal size, but none of them are very large and they are usually rounded. The cremaster always has a prominent projection on each side at its proximal end. The genera of Melitaeini may be separated as follows : a. Tubercles of the dorsal spiracular row not present on the second ab- dominal segment ; no tubercles present on the eighth al)dominal seg- ment. b. Dorsum of abdomen with a distinct tranvcrse ridge on the fourth segment ; dorsal spiracular row of tubercles not distinct on any of the segments PJiyciodes Hiibner. bb. Dorsum of abdomen without a transverse ridge on the fourth seg- ment ; dorsal spiracular row of tubercles very large on the third and fourth abdominal segments Cliaridryas Scudder. aa. Tubercles of the dorsal spiracular row present on the second abdom- inal segment ; tubercles present on the eighth abdominal segment, b. Tubercles of the eighth abdominal segment nearly as large as the others, the abdominal tubercles all broadly rounded and never longer than broad ; cremaster with a deep depression on the dor- sal surface and a circular depression on the ventral surface, the lateral tubercles very prominent, rounded, smooth and polished. EupJiydryas Scudder. 93 bb. Tubercles of the eighth abdominal segment much smaller than the others, the abdominal tubercles all somewhat pointed and longer than broad ; cremaster never with a deep depression on the dor- sal surface, but with a long narrow furrow on the ventral sur- face, the lateral tubercles somewhat triangular in outline and slightly rugose like the cremaster Cinclidia Hiibner. The following species were examined : Fhyciodes tharos Drury Charidryas nycteis Doubleday and Hewitson Buphydryas phaeton Drury Cinclidia harrisii Scudder Subfamily BasilarcJiinae The genus Basilarchia differs from all the genera of Nymphalinae^ with which it is generally included, on account of the absence of the rows of tubercles. There are two cephalic projections, as in many Nymphalinae, and on the dorsum of the second abdominal segment there is a very large carinate projection. This is somewhat oval in outline as seen in lateral view, being constricted at the base. Tlie body is not prominently excurved in the region of the appendages as in the Nymphalinae, but is of the same general shape. It agrees with the Nymphalinae only in the characters common to all members of the subfamily and is therefore placed in a separate subfamily. The following species were examined : Basilarchia astyanax Fabricius, artheniis Drury, archippus Cramer. Subfamily Apaturinae The species of this subfamily, included by Scudder in the Nym- phalinae, show no characters which unite them with that subfamily. The group, according to Scudder, included the genera Chlorippe and Anaea, which differ so widely in the pupae that they could not well be combined in the same subfamily. The name Apaturinae has been re- tained for the genus Chlorippe. These pupae are strongly compressed, with a prominent median dorsal carinate ridge. There are two ce- phalic projections and the ventral surface of the body forms a straight line while the dorsum is strongly arched. The antennae are slightly elevated and tuberculate. The genital openings are sunken and almost concealed. On either side of the anal opening near its cephalic end there is a small tubercle covered with hooked setae. The cremaster is short and triangular, and the hooked setae are nearly all on the ventral surface. 94 The following species were examined : Chlorippe celtis Boisduval and Le Conte, clyton. Boisduval and Le Conte. Subfamily Anacinae This subfamily includes a single genus, Anaea, which was included with Chlorippe in the tribe Apaturini by Scudder. The pupae are so different, however, that they have in common only the ordinary nymphalid characters. The body is never compressed, but the abdom- inal segments caudad of the wings taper very rapidly and form a hemisphere. The long cremaster is inserted near the center of the hemisphere and curves ventrad. The fourth segment has a prominent transverse ridge. The ventral surface of the abdominal segments caudad of the wings is very short and the genital openings are con- cealed. The head has a prominent transverse ridge at the cephalic end which extends caudad through the middle of the eye-pieces and along the lateral margin of the body. The metathorax has a rather prominent rounded ridge on the meson. The antennae and maxillae extend to the caudal margin of the wings. The following species was examined : Anaea andria Scudder. Subfamily Buploeiime This subfamily is ecjuivalent to the family Lymnadidae of some authors. It includes two genera, of which only Anosia has been ex- amined. The general shape of the body is very like that of the genus Anaea of the subfamily Anaeinae, but it has the second abdominal segment very long, as well as the thorax, and the cremaster extends caudad. There is never a ridge on the head, but it has a tubercle at each cephalo-lateral margin. The transverse ridge is on the third seg- ment and is tuberculate. The maxillae do not reach the caudal margin of the wings in Anosia and the antennae lie adjacent on the meson caudad of them. The following species was examined : Anosia plcxippns Linnaeus. Subfamily Satyrinae The Satyrinae are similar in shape to the Nymphalinae, but have no tubercles on the surface of the body and but few prominent ridges. The head always has a prominent transverse ridge at the cephalic end, and this often forms slight cephalo-lateral angles. There is also a slightly carinate ridge at each lateral margin of the body extending 95 as far caiidad as the second abdominal segment. The metathorax al- ways has a median elevation, which sometimes forms a prominent angle. The circular ridge surrounding the anal opening is not strongly elevated but the tubercles are prominent on each side of the genital opening on the ninth segment. The genera of this subfamily may be separated by the following table : a. Cremaster broader than long, with hooked setae present on the ventral surface ; genital opening never with a tubercle on each side. Cissia Doubleday. aa. Cremaster longer than broad, the hooks always inserted at the dis- tal end, never on the ventral surface ; genital opening always with a distinct tubercle on each side, b. Mesothoracic elevation rounded ; cremaster concave at tip with the hooked setae inserted in the hollow; body surface with fine indeterminate striations Cercyonis Speyer. bb. Mesothoracic elevation with a distinct angle ; cremaster not con- cave at tip ; body surface smooth Saiyrodes Scudder. The following species were examined : Cissia eurytus Fabricius Cercyonis alope Fabricius Satyrodes canthus Linnaeus Subfamily Ocneinae The genus Oeneis has none of the distinguishing characters of the subfamily Satyrina^ and therefore is not here included with the mem- bers of that group. The body has the general shape of a lycaenid, and the segments seem as devoid of motion (Fig. 80). In other respects it is a typical nymphalid. The antennae do not quite reach the caudal margin of the wings, and overlie the maxillae at their distal end, so that the antennae are adjacent on the meson. Their proximal ends are very near the meson on the dorsal surface of the head. There is no distinct ridge surrounding the anal opening, nor are any tubercles present caudad of the genital openings. There is no cremaster, nor hooked setae at the caudal end of the body. Only one species was examined : Oeneis seniidca Say. Specialized pupae without pilifers This division includes the remaining superfamilies of Lepidoptera. The seventh abdominal segment is fixed in both sexes in all the fami- lies except the Epermeniidae, in which this segment is fixed in the 96 male. None of the species included here have dorsal movement be tween any of the segments cephalad of the fourth. In this they differ from the superfamilies Pyralidoidea and Papilionoidea, some mem- bers of which retain dorsal movement of the third abdominal segment. This group includes all the most specialized families. The origin of most of these is doubtful. The Noctuoidea show the strongest rela- tionship to the Pyralidoidea ; the Notodontoidea, to the Gelechioidea. All the evidence at present points to the fact that the Pyralidoidea and Gelechioidea have descended from a common ancestor closely allied to the Yponomeutoidea. The Sphingoidea and Saturnioidea, which show considerable relationship to each other, seem to have arisen from a common stem with the more generalized Bombycoidea, which in turn seem nearly related to the Noctuoidea and Notodontoidea. SuPERFAMiLY YPONOMEUTOIDEA The families included here show well-developed labial palpi, and have a large portion of the prothoracic femora exposed. All show the maxillary palpi except the Coleophoridae, and the same arrange- ment of parts prevails throughout the super family. The epicranial suture is present in all families. The prothorax is always very short on the meson, but much longer on each lateral margin so that each half is triangular. The appendages always reach beyond the caudal mar- gin of the fourth segment, and in some cases are almost as long as the body. They are soldered firmly to each other but are free from the body wall. Abdominal segments 1-4 are longer than any of the others. There are usually spines or setae present at the caudal end of the body but seldom a cremaster. The pupae are usually less than 10 mm. in length. The families may be separated by the following table : a. Cremaster present, but short, with hooked setae at the distal end ; ninth abdominal segment with a deep lateral cavity on each side; seventh abdominal segment free in the male Epermenhdae. aa. Cremaster absent; ninth abdominal segment never with a deep lat- eral cavity; seventh abdominal segment fixed in both sexes, b. Maxillary palpi present; caudal end of body without lateral pro- longations ending in spines Yponomeutidae. bb. Maxillary palpi never present; caudal end of body with lateral prolongations ending in sharp spines Coleophoridae. Family Epermenhdae This family, which has usually been combined with the Elachisti- dae, or, by some writers, with the Scythridae, is here associated with 97 the Yponomeutidae, the only important differences between the two famihes being the freedom of the seventh abdominal segment in the male and the presence of a very short cremaster in the Epermeniidae. Another difference is that in Epermeniidae the wings and other appen- dages are somewhat elevated at the meson and slope to each lateral margin. A comparison of Figure 68 with Figures 82, 83, and 85 will show the similarity in arrangement of parts in the Yponomeutidae and Epermeniidae. The following species was examined : Bpcrmcnia pimpinclla Murtfeldt. Family Yponomeutidae The genera comprising this family resemble each other very strongly in all important characteristics, but nevertheless possess very clear generic distinctions. They closely resemble certain of the gen- eralized gelechiids (Figs. 88, 89), and many authors have associated those genera with the yponomeutids. The presence of a distinct fronto-clypeal suture and the peculiar arrangement of the antennae in the family Gelechiidae, together with the apparent loss of the labial palpi, seem to separate it very clearly from the Yponomeutidae, in which the fronto-clypeal suture is never distinct and the antennae are never adjacent on the meson except in Zelleria (Fig. 82). The typical arrangement of parts is seen in Figures 82, 83, 84, 85, 86, making further description unnecessary. The abdominal spiracles are all con- siderably produced and tubular, being longest in Plutella. There is no cremaster present in any of the genera. The genera of Yponomeuti- dae may be separated by the following table : a. Mesothoraeic spiracles produced, tube-like ; setae at caudal end of tenth segment hooked; maxillae more than three fourths the length of the wings Plutella Schrank. aa. Mesothoraeic spiracles not produced, slit-like; setae at caudal end of tenth segment straight, or occasionally slightly curved at end; maxillae much less than three fourths the length of the wings, b. Caudal end of tenth segment showing four straight setae, generally two directed laterad and two caudad ; maxillary palpi touched by both prothoracic and rnesothoracic legs. c. Maxillary palpi long, reaching the proximo-lateral angles of the maxillae ; labial palpi never becoming wider than at their prox- imal margin; antennae never touching on the meson. Yponomeuta Latreille. cc. Maxillary palpi short, never reaching the proximo-lateral angles of the maxillae ; labial palpi wider through most of their length than at the proximal margin ; antennae touching on the meson. Zelleria Stainton. 98 bb. Caudal end of tenth segment showing eight setae ; four of these on the ventral surface extending laterad and sometimes slightly curved at the tip ; mesothoracic legs never reaching cephalad to the maxillary palpi ArgyrestMa Hiibner. The following species were examined : Plutella maculipennis Curtis Yponomeuta padellus Linnaeus, malinelhis Zeller Zelleria celastrusella Kearfott Argyresthia freyella Walsingham Family Coleophoridae This family has usually been associated with the Elachistidae, but since the division of that family the name means very little unless we use it to include the genus Elachista and others closely related, which certainly would not include the Coleophoridae. They seem, rather to be more closely allied to the Yponomeutidae and are so considered here. They differ mainly in the loss of the maxillary palpi and in the lateral extensions of the ninth abdominal segment (Fig. 87). The ap- pendages are usually very long, extending nearly to the caudal mar- gin of the body, and often beyond it, when the movable segments are contracted. The abdominal segments from the first to the sixth are very much longer than the remaining segments. The following species were examined : Coleophora caryaefoliella Clemens, vcrnoiiiacdla Chambers, mativo- rella Riley. SuPERFAMiLY GELECHIOIDEA This superfamily includes those pupae which possess a distinct epicranial suture, with the caudal portions of the antennae lying adja- cent on the meson and usually separating at their distal ends to expose the metathoraic legs. The maxillary palpi are usually present, but labial palpi and prothoracic femora are seldom exposed. The body is usually ovate in outline as seen in dorsal or ventral view; widest in the thoracic region and somewhat depressed. The superfamily is closely related to the Yponomeutoidea. It includes here two groups representing two distinct lines of development: the group retaining, the fronto-clypeal suture, including the families Lavernidae, Scythri- dae, Gelechiidae, and Chrysopeleiidae ; and those in which it is not distinct or is absent, including the families Oecophoridae, Stenomidae, Cosmopterygidae, and Elachistidae. The latter group may represent a distinct superfamily when all its allied genera have been studied, but at present there is no evidence to warrant such a conclusion. The fol- lowing table will serve to separate the families of Gelechioidea : 99 a. Fronto-clypeal suture always distinct for its entire length, sometimes forming a prominent curve or angle at the meson, b. Labial palpi exposed for their entire length, c. Femora of the prothoracic legs visible ; maxillary palpi either reaching the proximo-lateral angles of the maxillae or approach- ing them very closely; tenth abdominal segment with stout spines at the caudal end Lavernidae. cc. Femora of the prothoracic legs never visible; maxillary palpi minute; tenth abdominal segment with hooked setae at the caudal end Scythridae. bb. Labial palpi never exposed for their entire length, usually con- cealed, c. Maxillary palpi present and usually reaching the proximo-lateral angles of the maxillae ; antennae usually adjacent on the meson for the caudal two fifths of their length, separating at distal ends to show the metathoracic legs Gelechiidae. cc. Maxillary palpi never present; antennae adjacent on the meson for the caudal two fifths of their length but not separated at their distal ends • Chrysopeleiidae. aa. Fronto-clypeal suture never distinct for its entire length and never reaching the meson. b. Abdominal segments 4-6 movable, with very deep incisions be- tween the segments on the dorsal and ventral surfaces; body depressed. c. Maxillary palpi large, usually reaching the proximo-lateral angles of the maxillae; hooked setae never present on the ven- tral surface of the ninth abdominal segment Oecophoridae. cc. Maxillary palpi minute ; hooked setae always present on the ven- tral surface of the ninth abdominal segment Stenomidae. bb. Abdominal segments 4-6 never all movable ; incisions between the segments of equal depth on all surfaces ; body not depressed. c. Maxillary palpi present ; sixth al)dominal segment movable. Cosmopterygidae. cc. Maxillary palpi absent; no abdominal segments movable. Elachistidae. Family Lavernidae The pupae belonging to this family are more generalized than any other members of the superfamily and closely resemble pupae belong- ing to the family Yponomeutidae. They have been included here on account of the distinct fronto-clypeal suture, present in all except the more specialized Gelechioidea, and also because the prothorax, which is so short on the meson in Yponomeutidae with each half triangular in outline, in this family loses that condition and becomes almost as long on the meson as at the lateral margins, so that each half is sub- 100 quadrangular. This is the only family of Gelechioidea which retains both labial palpi and exposed portions of the prothoracic femora. The appendages have the characteristic arrangement of the superfamily (Fig. 88) except that the antennae do not separate near their distal end to expose a portion of the metathoracic legs, but these are seen caudad of the antennae adjacent on the meson. The wings are long and pointed in Lophoptilus, but rounded in Laverna. The first four abdominal segments are longer than the remaining caudal segments in this family, and the appendages, are soldered to them, but are free for the remainder of their length. This family has been considered as a subfamily of Elachistidae by most authors, and has usually included' Cosmopteryx, which is a much more specialized genus. The following t;ible will serve to separate the genera of Lavernidae : a. Head long, somewhat pointed, the length more than half the greatest width ; f ronto-clypeal suture making an acute angle at meson ; spines of the tenth segment extending dorsad and not visible in ventral view ; exposed part of metathoracic leg about one fifth the length of the portion of the antennae lying adjacent on the meson. LopJwptilus Sircom. aa. Head short, blunt, the length about equal to half the greatest width ; f ronto-clypeal suture making an obtuse angle at meson ; spines of tenth segment extending laterad and visible in ventral view ; ex- posed part of metathoracic leg about equal in length to the portion of the antennae lying adjacent on the meson Laverna Curtis. Family Scythridae The pupae of this family also resemble the Yponomeutidae in some respects, and have been included with them by some authors. Others have associated the family with the Elachistidae, while Stainton in- cluded it with the Gelechiidae as the genus Butalis. The antennae in this family meet on the meson, but do not separate to show the distal ends of the metathoracic legs as is the general rule in this superfamily. Instead, the mesothoracic wings lie adjacent on the meson caudad of the antennae, and the appendages are firmly soldered to each other and to the body (Fig. 89). As the appendages extend caudad for about half the length of the seventh abdominal segment, it follows that there can be no motion possible between any of the abdominal segments, unless there be slight dorsal motion. The prothorax is typically gele- chiid in character. The abdominal spiracles are considerably pro- duced and tubular, their length varying in the different species. The setae of the body are nearly all hooked, and a few longer ones are pres- 101 ent at the caudal end of the body. Only one genus of this family was available for study. The following species were examined : Scythris eboracencis Zeller, iinpositclla Zeller. Family Gelechiidae The pupae of the Gelechiidae never show any portion of the labial palpi, unless it should be a very small triangular area caudad of the labrum, between the halves of the maxillae. The prothoracic femora are never exposed (Figs. 90, 91, 92, 93, 94). The fronto-clypeal su- ture is always distinct and usually e>;tends almost straight across be- tween the proximal ends of the antennae, but occasionally each half is directed cephalad near the meson so that an angle is formed at their junction. The caudal parts of the antennae usually lie adjacent on the meson for about two fifths of their length and usually cover the caudal ends of the maxillae and sometimes of the prothoracic and mesotho- racic legs. They usually separate at their distal ends to show the meta- thoracic legs, or what will be referred to as such in this paper. There was not enough available material in condition for dissection to de- termine whether the maxillae ever reached the caudal margin of the wings and overlaid the metathoracic legs, as might easily be the case. The maxillary palpi are present in all, but do not always reach the proximo-lateral angles of the maxillae. They are always reached by both prothoracic and mesothoracic legs. The wings vary somewhat in length, but are usually firmly soldered down, and the abdominal seg- ments are somewhat depressed on the ventral surface, forming a shallow cavity into which the wings are fitted and, therefore, are not elevated above the surface of the body. There are usually very deep incisions between the segments, especially on the dorsal and ventral surfaces. Many species have the incisions deeper on the ventral sur- face so that the caudal end of the body may be strongly curved ven- trad. The pupae are usually very active, and many of them are able to move after the fashion of click-beetles. The body is entirely cov- ered with setae in some genera, while others have a fringe of setae along the margin of certain slightly projecting ridges and occasional depressions found usually on the seventh abdominal segment. There seems doubt as to the generic standing of the following species : Aris- totelia physalicllo, Gnorinwscliema laz'crnclla, and Rcciirvaria variclla; at least they differ from other species examined in these genera. In the case of Aristotclia it has been impossible to determine which species is the type of the genus. The genera of Gelechiidae may be separated as follows : 102 a. Body setae very long and heavily chitinized, often as long as the seg- ments of the abdomen, and, as seen under high power, usually forked at the apex ; f ronto-clypeal suture curving eephalad from the proxi- mal ends of the antennae to the cephalic margin of the body; ci'e- master always present; dorsal cephalic margin of segments two, three, and four of the abdomen with a slight rounded projection on each side of the meson, edged with a dense fringe of whitish setae di- rected eephalad, caudad of this a prominent elevation bearing a sim- ilar fringe of setae on the summit (Fig. 91), b. Dorsal surface of the cephalic margin of the movable segments with a prominent cavity on the meson having heavily chitinized edges TricliotapJie Clemens. bb. Dorsal surface of cephalic margin of the movable segments with- out any cavity; a strongly chitinized ridge separating the ce- phalic margin from the remainder of the segment, eephalad of this a band of short spines, then a prominent furrow, the furrow and the remainder of the cephalic margin being deeply punctate. Ypsoloplius Fabricius. aa. Body setae never modified ; segments two, three, and four of the ab- domen never as described in a. b. Entire body with a dense covering of whitish setae visible to the unaided eye, giving the pupa a furred appearance, c. Maxillary palpi long, reaching the proximo-lateral angles of the maxillae; antennae lying adjacent on the meson for about two fifths of their length; cremaster short and blunt, the end set with about eight stout curved setae ; length 8-10 mm. Anacaynpsis Curtis. cc. Maxillary palpi short, never reaching the proximo-lateral angles of the maxillae; antennae just meeting on the meson; cremaster with a sharp point curved dorsad, with curved setae at the sides and base ; length 5-6 mm Aristotelia (a)Hiibner, bb. Entire body never covered with setae. c. Seventh abdominal segment with a dense fringe of setae on some portion, d. Fringe of setae confined to the cephalic and lateral edges of a prominent lateral cavity. e. Body smooth, not depressed ; cephalic edge of the lateral cav- ity trilobed (Fig. 93) Evippe Chambers. ee. Body with small spines, strongly depressed ; cephalic edge of the lateral cavity bilobed Telpliusa Chambers. dd. Fringe extending around the segment or nearly so, e. Fringe extending around the segment in a straight line ; body not noticeably depressed, the surface smooth. Recurvaria (a) Haworth. ee. Fringe extending around the segment in a more or less wavy line; body noticeably depressed and very broad in the thoracic region, surface with punctures or spines. 103 f. Abdominal segments 8-10 distinctly tapering to the cau- dal end of the body ; fringe of seventh segment extend- ing in a wavy line around the body and edging two very large lobes on the dorsal surface, Trypanisma Clemens. ff. Abdominal segments 8-10 not tapering but blunt and slightly rounded at the caudal end of the body; fringe extending in a wavy line around the seventh segment, without prominent lobes on the dorsal surface. Gelecliia Hiibner. cc. Seventh abdominal segment without any fringe of setae. d. Caudal end of body with short, stout projecting spines. e. Caudal end of body with one such spine on the dorso-meson, projecting dorsad; fronto-clypeal suture almost straight. PJitliorimaea Meyrick. ee. Caudal end of body with a median spine and one on each lateral margin; fronto-clypeal suture extending cephalad from the base of each antenna to the cephalic margin of the head Sitotroga Heinemann. dd. Caudal end of body with straight or curved setae, e. Hooked setae present at the caudal end of the body. f. Antennae reaching the caudal margin of the wings, their caudal ends separated to show the metathoracic legs. g. Caudal end of body with at least five long hooked setae on each side of the meson; antennae slightly enlarged at their proximal ends, making the cephalic end of the body somewhat truncate Recurvaria (b) Haworth. gg. Caudal end of body never with more than two long curved setae on each side of the meson ; cephalic end of body rounded Aristotelia (b) Hiibner. ff. Antennae never reaching the caudal margin of the wings nor separating at their caudal ends to expose the meta- thoracic legs Gnorimosdiema (a) Busck. ee. Hooked setae never present at the caudal end of the body; a few short, straight setae present. Gnorimoschema (b) Busck. The following species were examined : 7'richotaplie flavocostella Clemens YpsolopJius citrifoliellus Chambers, eupatoriellus Chambers Anacampsis sp., rhoifructella Clemens Aristotelia (a) salicifungiella Clemens Aristotelia (b) physaliella Chambers Evippc priimfoliella Chambers Telp/iiisa qticrcinigracclla Chambers, palliderosacella Chambers 304 Rccurvaria (a) apicitripunctclla Clemens Recnrvaria (b) variella Chambers Trypanisma prudens Clemens Gclcchia ccrccriscUa Chambers, discoocellcUa Chambers, scrotinclla Busck PJithorimaea sp. Sitotroga cerealella Olivier Gnorimoschema (a) lavernella Chambers Gnoriinoschcma (b) gallaesolidagmis Riley Family Chrysopeleiidae This family includes the genus Chrysopeleia which was formerly included with the Elachistidae. It has the same arrangement of parts as that in the genus Elachista but retains the f ronto-clypeal suture and has no cremaster (Fig. 95). Until more is known of the relationships of this genus it seems better to place it in a family by itself. The ap- pendages are slightly elevated and firmly soldered to each other and to the body. They extend well on to the seventh abdominal segment, so that there appears to be no motion possible between any of the body segments. There is no cremaster present and only a few short straight setae at the caudal end of the body. The abdominal spiracles are pro- duced and tubular. The pupae are very small, being only about 3 mm. in length. The following species was examined : Chrysopeleia ostryaeella Chambers. Family Oecophoridae • This family (Figs. 97 and 98) includes those pupae in which the fronto-clypeal suture is not present for its entire length and which have large maxillary palpi, usually reaching the proximo-lateral angles of the maxillae. All the species examined showed the presence of very fine setae arranged in groups over the surface of the abdomen, but these were hard to locate in Psilocorsis. The incisions between the segments are very deep in all members of the family. Of the three genera studied Psilocorsis, Agonopteryx, and Depressaria, the first seemed more nearly related to the Stenomidae, while the other two were typical gelechiids, except that the fronto-clypeal suture was never distinct. It seems probable that a revision of the group might sepa- rate these genera. The table to genera will indicate the principal dif- ferences. The body is usually strongly depressed and 5-10 mm. in length. The following table will serve to separate the genera: 105 a. Femora of the prothoracie legs exposed. b. Wings pointed, extending on to the sixth abdominal segment; proximal end of each antenna elevated and roughened with trans- verse ridges; eremaster present Psilocorsis Clemens. bb. Wings not pointed, and never extending beyond the caudal mar- gin of the fourth abdominal segment ; proximal end of each an- tenna never elevated or roughened with transverse ridges; ere- master never present Depressaria Haworth. aa. Prothoracie femora never exposed Agonopteryx Hiibner. The following species were examined : Psilocorsis ohsolctclla Zeller, qnercicclla Clemens Depressaria heracliana De Geer Agonopteryx ncbulosa Zeller Family Stenomidae The Stenomidae include most of the genera formerly grouped under the name of Xylorictidae. The appendages are arranged as in the Gelechiidae and the maxillary palpi are minute (Fig. 96). A small portion of the labial palpi is usually apparent. The appendages are firmly soldered to each other and to the body wall. They extend slightly beyond the caudal margin of the fourth abdominal segment, so that there are three free segments, the fourth, fifth, and sixth. The margins of these free segments are serrate along the edges of the in- cisions, w'hich are very deep, especially on the ventral surface, and per- mit the caudal end of the body to be very sharply curved ventrad, reaching almost to the caudal margin of the wings (Fig. 96a). The fronto-clypeal suture is visible for a short distance mesad of the prox- imal end of each antenna, but it never reaches the meson. There are many curved setae present on the ventral surface of the ninth abdom- inal segment. The body is always more or less depressed, and in Stenoma is about 8 mm. in length, in Menesta 3 mm. The genus Menesta, formerly included in the Gelechiidae, seems more closely allied to Stenoma and is included here. Stenoma possesses peculiarly modified setae on the body surface. These genera may be separated as follows : a. Antennae modified at their proximal ends, forming an enlarged cor- rugated area ; hooked setae on the ventral part of the ninth abdom- inal segment never on a distinct prolongation Stenoma Zeller. aa. Antennae never modified at the proximal end ; hooked setae on the ventral part of the ninth abdominal segment on a distinct prolon- gation Menesta Clemens. 106 The following species were examined : Stenoma schlaegeri Zeller Menesta albaciliaeella Chambers Family Cosmopterygidae This family name as used by most authors is equivalent to Laver- nidae or Momphidae, and the genera included under all these names are usually associated with each other. The Cosmopterygidae are much more specialized, however, as they retain neither visible labial palpi nor prothoracic femora (Fig. 99). The appendages are firmly soldered to each other and to the body wall as far as the caudal margin of the sixth abdominal segment, which allows freedom of movement to this segment. There are some generalized characters present, however — the length of the first six abdominal segments, which are as long as in Yponomeutidae, and the shape of the pro- thorax, which is shorter on the meson than at each lateral margin. The abdominal spiracles are slightly produced and tubular. There is a very short cremaster present bearing eight hooked setae, of which four are longer than the remainder. The pupae are about 4 mm. in length. The following species was examined : Cosmo pteryx clandcstinclla Busck. Family Elachistidae This family has been variously subdivided in the past few years, for, like the Tineidae, it included a large number of species which did not form a natural group. Some authors do not retain this family name, but as the nomenclature of the group appears to be still in a rather unsettled condition, this name is retained for the present to in- clude the genus Elachista. The appendages are arranged as in other gelechiids, but there is no trace of maxillary palpi (Fig. 100). The surface of the body is covered with large rounded tubercles and the dorsal surface shows three distinct longitudinal elevations or ridges, one on the meson and one near each lateral margin, bearing the spir- acles on the summit. The wings and other appendages are firmly sol- dered to each other and to the body wall, and there appear to be no free segments. The prothorax is typically gelechiid and the meso- thorax shows a decided alar furrow on each side. There is a distinct cremaster present, but it shows no hooked setae. The pupae are sus- pended from a stem or leaf after the manner of some papilionids, with a silken girth around the body. The pupae average 3.5 mm. in length. The following species was examined : Elachista praelineata Braun. 107 SUPERFAMILY NOCTUOIDEA This stiperfamily includes tliree families, Noctuidae, Liparidae, and Arctiidae. The Syntomidae also belong to this group, but as only one species of this family was examined and this showed no charac- ters to separate it from the Arctiidae, the Syntomidae are not dis- cussed as a separate family. The Noctuoidea and Bombycoidea in- clude all the specialized families which retain labial palpi. The fami- lies of Noctuoidea may be separated thus : a. Body seldom with setae arranged around sears of larval verrucae, if present, then the femora of the prothoracic legs exposed, or a long cremaster present bearing hooked setae at the distal end ; prothoracic femora usually visible, if not, then the mesothoracic leg usually ex- tending cephalad to the eye-pieces Noctuidae. aa. Body always with setae arranged around the sears of larval verrucae ; femora of the prothoracic legs never visible. b. Maxillae never more than two fifths the length of the wings ; body setae conspicuous ; labial palpi usually visible Liparidae. bb. Maxillae two thirds the length of the wings, or longer ; body setae inconspicuous ; labial palpi seldom visible Arctiidae. Family Noctuidae This family (Figs. loi, 102, 103), with a few exceptions, is char- acterized by the presence of labial palpi and of maxillae which extend to the caudal margin of the wing, or very closely approximate this length. Very many of the genera have a large portion of the pro- thoracic femora exposed. Those which do not show any portion of the prothoracic femora have the mesothoracic leg extending cephalad to the eye-pieces, with a few exceptions in the genera Homopyralis, Plusiodonta, and Anomis. Those lacking labial palpi have setae ar- ranged around the scars of larval verrucae, as in the Arctiidae. They differ from the Arctiidae in having hooked setae on the cremaster, and in lacking flanged plates on the abdominal segments. Maxillary palpi are found in some members of the subfamilies Agrotinae, Cucul- lianae, and Hypeninae. Since there was not enough material available for study to furnish a basis for subfamily characters, the genera have been grouped as seemed best for purposes of classification. As far as possible the names of subfamilies as used by Hampson in the "Cata- logue of Lepidoptera Phalaenae" have been adopted. This arrange- ment could not be followed throughout, however, and so it must be re- membered that the subfamily names used here are adopted as a matter of convenience and do not stand for the genera which Hampson 108 grouped together. His generic and specific names have been adopted as far as possible. As to the phylogeny of the group, too little ma- terial has been examined to warrant a decided opinion on the subject. It seems probable, however, that most of the subfamilies discussed rep- resent the ends of many lines of development. There are various stages of development found in all groups, and there are some mem- bers of each subfamily studied, except the Phytometrinae and Momi- nae, which show the epicranial suture. The subfamilies mentioned above which retain maxillary palpi are undoubtedly the most generalized, the Mominae, which show neither labial palpi, prothoracic femora, maxillary palpi, nor an epicranial su- ture, are undoubtedly the most specialized, but nothing can be said vx^ith certainty as to the other groups. No attempt has been made to arrange the subfamilies in phylogenetic order either in the tables or in the discussion of subfamilies. The subfamilies of Noctuidae may be separated by the following table: a. Prothoracic legs reaching cephalad to the eye-piece, mesothoracic legs never reaching as far cephalad; prothoracic femora usually visible. b. Cremaster, or caudal end of the body, with all the setae curved or hooked, never with any long straight setae. c. Setae of the cremaster usually all of the same size and length, either slightly curved or hooked; cremastei"' never with six or eight hooked setae of which the mesal ones are larger and longer than the remainder Acronyctinae. cc. Setae of the cremaster or caudal end of the body usually of two sizes, d. Body never with scars of larval verrucae bordered with setae ; labial palpi and prothoracic femora always visible, e. Wings and maxillae produced at meson into a distinct rounded projection extending on to the fifth abdominal seg- ment ; labrum located near the cephalic end of the body ; body never heavily chitinized Phytometrinae. ee. Wings and maxillae never extending beyond the caudal margin of the fourth segment ; labrum in the normal posi- tion ; body always heavily chitinized Cucullianae. dd. Body with rounded or oval areas bordered with setae, as in the Arctiidae, especially noticeable on the dorsum of the abdomen and in the spiracular region ; labial palpi and pro- thoracic femora never visible Mominae. bb. Cremaster or caudal end of the body never with all the setae curved or hooked. c. Cremaster or caudal end of body with at least two long straight stout setae, 1-2 mm. in length, rarely with additional hooked setae Hadeninae, 109 cc. Cremaster or caudal end of body never with long straight stout setae, d. Cremaster bifurcate, narrowed at the caudal end ; dorsum of movable abdominal segments with one or more rows of deep circular pits with dark, chitinized margins Agrotinae. dd. Cremaster short, broader at the distal end, very thin and plate-like, the caudo-lateral angles produced into short rounded lobes, with two or three small rounded projections between often bearing small delicate setae ; dorsum of the movable abdominal segments never with pits ; body strongly rugose, the abdominal segments spinose Agaristinae. aa. Prothoracic and mesothoracic legs both reaching cephalad to the eye- piece or to the maxillary palpus where this is present ; prothoracic femora seldom visible, b. Cremaster usually present; curved or hooked setae always pres- ent at the caudal end of body, usually eight in number. e. Body always covered with a whitish bloom Catocalinae. ec. Body never covered with a Avhitish bloom Hypeninae. bb. Cremaster or setae never present at the caudal end of the body ; fifth abdominal segment with a row of spines along the caudal margin extending from the sculpturing of the dorsum almost to the meson on the ventral surface Sarrothripinae. Subfamily Agrotinae This subfamily, as here considered, includes those pupae with a stout, rugose, more or less bifurcate cremaster, and with a row of large circular pits with heavily chitinized margins along the cephalic margin of some of the abdominal segments, usually between the fourth and seventh (Fig. loi). In these pupae, the prothorax is very long, at least tw^o thirds the length of the mesothorax, and the epicranial suture is sometimes present in the genus Agrotis. Labial palpi are al- ways present and exposed for their entire length, and the prothoracic femora are seen in some of the genera. The mesothoracic legs, an- tennae, and maxillae are of practically the same length and usually ex- tend to the caudal margin of the wings. The metathoracic legs are seldom visible and only the prothoracic legs extend cephalad to the eye- pieces, and these do not separate the sculptured eye-piece and antenna. The body is stout, and wdien retracted the length is about three times the width. The genera of Agrotinae may be separated as follows : a. Femora of the prothoracic legs visible. b. Dorsal cephalic margin of the movable abdominal segments with a single even row of pits, numbering less than fifteen ; epicranial suture present ; cremaster often with a row of four setae near its proximal end Agrotis Ochsenheimer. 110 bb. Dorsal cephalic margin of the movable abdominal segments with an interrupted row of pits numbering more than fifteen ; epicra- nial suture never present; cremaster never with a row of four setae near its proximal end Hapalia Hiibner. aa. Femora of the prothoracic legs never visible Noctua Linnaeus. The following species were examined : A gratis hadinodis Grote, bicarnea Guenee Hapalia incivis Guenee Noctua clandestina Harris Subfamily Cucullianae As only two specimens of one genus, Graptolitha, were available for study, little can be said as to subfamily characters. These speci- mens differ from members of other subfamilies except the Catocalinae in having all the setae at the caudal end of the body hooked. There are two setae at the meson very much larger and more heavily chi- tinized than the remaining setae, which are usually four in number. In other respects, as the length of prothorax, size and shape of body, arrangement of appendages, presence of epicranial suture and labial palpi, exposed femora of the prothoracic legs, and traces of maxillary palpi, they resemble the Agrotinae and especially the Hadeninae as the movable abdominal segments are finely punctate along their ce- phalic margin. The following species were studied : Graptolitha laticinerea Grote, antemmta Walker. Subfamily Hadeninae This subfamily includes pupae having stout straight setae or spines at the caudal end of the body. There are usually two, from 1-2 mm. in length, and they may be inserted in a short cremaster or directly in the caudal end of the body (Fig. 102). One genus, Cirphis, has addi- tional slender hooked setae. The prothorax is very long, as in Agro- tinae, at least two thirds the length of the mesothorax. The epicra- nial suture is present in the genera Polia, Hadena, Lycophotia, and Eriopus. The appendages, which in Agrotinae are of the same length and generally reach the caudal margin of the wings, are in this sub- family unequal in length. The maxillae usually reach the caudal mar- gin of the wings, but the mesothoracic legs are shorter, and the anten- nae in some forms equal these or are very much shorter. Except for Cirphis and Monima the abdominal segments are punctate. These tv.^o genera have the movable abdominal segments pitted as in the Ill Agrotinae. The prothoracic leg extends cephalad to the eye-piece, but there is never a long point extending cephalad between the antennae and the sculptured eye-piece. The genera of Hadeninae may be sep- arated by the following table : a. Dorsum of movable abdominal segments with a row of deep pits along the cephalic margin. b. Caudal end of body with a very short cremaster ending in two straight sharp setae with four slender hooked setae at the prox- imal end, the hooked setae about half the length of the straight ones CirpJiis Walker. bb. Cremaster very short, with a long straight seta inserted in each caudo-lateral angle; hooked setae never present. Monima Hiibner. aa. Dorsum of movable abdominal segments never with a row of deep pits along the cephalic margin. b. Mesothoracic leg never reaching cephalad to the eye-pieces; pro- thoracic femora always exposed, c. Caudal end of the abdomen with four closely approximated spines of the same size and shape RJiodopJiora Guenee. cc. Caudal end of body with only two spines. d. Spiracles noticeably modified, usually with a prominent de- pression near their caudal margin ; spiracular opening di- rected somewhat caudad; cephalic third of the movable abdominal segments usually slightly elevated and densely punctate, e. Cremaster slender, pointed, spines closely approximated ; abdominal spiracles with a darker, elevated, crescent- shaped area almost surrounding the spiracle and with a deep cavity larger than the spiracle directly caudad of it ; mesothoracic spiracles not modified Meliana Curtis. ee. Cremaster short, blunt; area around each abdominal spir- acle slightly elevated and darker in color, a prominent cav- ity caudad of the spiracle with a chitinized ridge along the caudal margin of the spiracle; mesothoracic spiracles also modified, the opening extending half the distance between the antennae and the meson, f . Chitinized ridge along the caudal margin of the abdominal spiracles distinctly serrate ; clypeal region not promi- nently elevated and without deeply impressed lines. LapJiygma Guenee. ff . Chitinized ridge along the caudal margin of the abdominal spiracles not distinctly serrate; clypeal region promi- nently elevated and with deeply impressed lines. Prodenia Guenee. 112 dd. Spiracles normal, without any prominent elevations or de- pressions adjacent; cephalic third of the movable abdominal segments usually not elevated ; segments punctate, e. Epicranial suture present and distinct. f. Cephalic third of movable abdominal segments with fine punctures, the remainder of the segments smooth ; no cremaster present Lycopliotia Hiibner. ff. Abdominal segments all finely punctate and with wavy impressed lines, more densely punctate along cephalic margin of the movable segments ; a short, distinct rugose cremaster present Hadena Schrank. ee. Epicranial suture wanting. f. Abdominal segments 5-7 very finely punctate on the ce- phalic third CJilortdea Westwood. ff. Abdominal segments with very large, shallow punctures, at least on the cephalic third of segments 4-7. ; g. Abdominal segments all punctate ; segments 4-7 having the punctures very large on at least the cephalic two thirds of each segment and fine punctures on the re- mainder Pyrrhia Hiibner. gg. Abdominal segments 1-7 punctate ; segments 4-7 with larger punctures on the cephalic half of each segment and usually with finer punctures on the remainder. Polia Ochsenheimcr. bb. Metathoracic leg reaching to the eye-piece; cephalic margin of tenth abdominal segment with a row of deep oblong pits ; protho- racic femora never exposed Eriopus Treitschke. The following species were examined : Cirphis unipuncta Haworth, phragmitidicola Giienee Rhodophor-a gaurac Smith and Abbot, florida Guenee Meliana alhilinca Hiibner. Laphygma friigipcrda Smith and Abbot Prodenia ornitJiogalli Guenee Lycopliotia inargaritosa Haworth Hadena vulgaris Grote Chloridea obsolcta Fabricius, vircscens Fabricitis Pyrrhia umbra Hiifnagel Polia renigera Stephens, picta Harris, nicditata Grote Briopus floridcnsis Guenee Monima alia Guenee Subfamily Agaristinae The members of this subfamily show remarkable uniformity and it is rather difficult to separate the genera. They have been given family 113 rank by some authors, and while the species inckided here differ from the typical noctuid in many respects, still no structural characters could be found to warrant their separation from the Noctuidae. The entire body surface is very rough and spinose, while the cremaster is short, broad, and decidedly flattened, with its caudo-lateral an- gles produced into rounded lobes, and with the caudal margin often crenulate. There are sometimes short straight setae present along the caudal margin. The antennae, mesothoracic legs, and maxillae usually reach the caudal margin of the wings, or approach it very closely. Each prothoracic leg extends cephalad to the eye-pieces. The epicranial suture is always present. The labial palpi are always visi- ble, but the femora of the prothoracic legs are concealed except in occasional specimens. The abdominal spiracles are somewhat elevated and are surrounded by a heavy, dark, chitinized border. The open- ings appear to be fringed with fine setae. The genera of this sub- family may be separated as follows : a. Antennae always reaching the caudal margin of the wings; the row of spines along the caudal margin of the abdominal segments not larger than the spines of the segment; ninth abdominal segment never with scattered spines larger than those of the other segments. b. Dorsum of the abdominal segments rugose and densely spinose on the first nine segments Euthisanotia Pliibner. bb. Dorsum of abdominal segments rugose and moderately spinose on the first seven segments ; the eighth and ninth segments roughened, but with very few spines Ahjpia Hiibner. aa. Antennae never reaching the caudal margin of the wings ; the row of spines along the caudal margin of the abdominal segments larger than those of the segment ; ninth abdominal segment with scattered spines larger than those of the other segments. Psycliomorplia Harris. The following species were examined ; Buthisanotia grata Fabricius, itiiio Hiibner Alypia octomaculata Fabricius Psycliomorplia epinicnis Drury Subfamily Acronyctinae The subfamily Acronyctinae as typified by the genus Acronycta has little to distinguish it from other subfamilies except that the cre- master is short and usually mound-like and the setae are always of tlie same size and length. With this genus there are here included several others, which probably do not form a natural group, though all possess cremastral setae of the same size and length and in all the 114 members of the group only the prothoracic legs extend cephalad to the eye-pieces. All of these genera except Acronycta have prominent projections on the head and prothorax. The epicranial suture is present only in Eulonche. The labial palpi are well developed in all and the prothoracic femora are visible only in Eulonche, Acronycta, and Achatodes. The maxillae do not reach the caudal margin of the wings in any of the genera. The metathoracic legs are always visible, and in Eulonche the mesothoracic legs are adjacent on the meson. In Plusiodonta they extend to the caudal margin of the wings, but in none of the other genera. The antennae are usually equal in length to the mesothoracic legs and never reach the caudal margin of the wings. The genus Eulonche and a few species of Acronycta are very peculiar in that there are setae on the body arranged as in Arctiidae (Fig. 103). These are easily observed on the mesothorax and the parts of the abdomen where the sculpturing is not so dense. There are also a few short spines present on the tenth segment at the base of the cremaster. The genera of Acronyctinae may be separated as follows : a. Cephalic end of body with two large, rounded, rugose projections on the head, one on each side of the meson. b. Dorsum of abdomen very rugose on segments 1-7 except for a smooth caudal margin; groups of long setae present on thorax and abdomen; epicranial suture present Eulonclie Grote. bb. Dorsum of abdomen never rugose, but with large lunate punc- tures on the cephalic third of segments 3-7 ; groups of long setae never present ; epicranial suture absent. Achatodes Guenee. aa. Cephalic end of body with a single median projection, or without projections. b. Cephalic end of body with a single median projection, either on the head or prothorax. e. A prominent projection on the body near the cephalic margin of the prothorax ; abdominal segments 1-3 with a broad smooth transverse ridge near the caudal margin; movable abdominal segments with thin flanged plates; cremaster very short and broad, with a short, stout, slightly curved seta inserted in each caudo-lateral angle Homopyralis Grote. cc. Projection always present on the head; setae at the caudal end of the body always hooked, d. Distinct cremaster not present; a large sharp point at each caudo-lateral angle of the body, with some smaller points, and three short hooked setae inserted on each side: body dis- tinctly punctate, with a smooth band at the caudal margin 115 of each segment; projection on the head blunt at the end and directed ventrad Plusiodonta Guenee. dd. Distinct cremaster present, its caudo-lateral angles produced and pointed, with two hooked setae inserted on the ventral side of each ; body not punctate, slightly rugose ; projection on the head narrowed to a rounded point. .Anomis Hlibner. bb. Cephalic end of body smooth, cremastral hooks slightly curved. Acronycta Treitschke. The following species were examined : Acronycta americana Harris, popiili Riley, clarescens Guenee, Jiama- melis Guenee Eulonche oblinita Smith and Abbot A chat odes seae Harris H omopyralis discalis Grote Plusiodonta compressipalpis Guenee Anomis erosa Hiibner Subfamily Phytometrinae The members of this subfamily differ markedly from those of the other subfamilies of Noctuidae. The labrum is never in its nor- mal position but is located near the cephalic end of the body, while the wings and maxillae extend beyond the caudal margin of the fourth abdominal segment. The wings are produced into a sharp point near the meson of the ventral surface. The labial palpi are always visi- ble and a large portion of the prothoracic femur is exposed. The dorsal cephalic margin of the movable abdominal segments has a number of prominent furrows with slightly serrate ridges between. The prothorax is not as long as in the previously mentioned sub- families, being only two fifths the length of the mesothorax, and the caudo-lateral angles are somewhat produced. The dorsal surface of the body shows prominent grooves along the caudal margin of the metathorax and the first four abdominal segments. The edges of these grooves are somewhat serrate. The cremaster is somewhat cylindrical and rugose, with two long hooked setae and four shorter ones. The two genera studied are very closely related and may be separated as follows : a. Antenna reaching the caudo-lateral angle of the mesothoracic leg; cremaster with length and breadth approximately equal and longer than its longest setae Pliytometra Haworth. aa. Antenna much shorter than the mesothoracic leg and never reach- ing its caudo-lateral angle; cremaster always longer than broad and about equal in length to its longest setae. Syngraplia Hiibner. 116 The following species were examined : Phytometra brassicae Riley Syngrapha falcigera Kirby Subfamily Mominae The only genus studied in this group was Charadra. This re- sembles the Arctiidae very much more than it does most Noctuidae in the shape of the body, in the presence of setae arranged around the scars of larval verrucae, and in the absence of epicranial suture and visible labial palpi and femora of the prothoracic legs. The an- tennae are broader at the proximal end than is typical in Noctuidae. The appendages are arranged more as they are in Noctuidae and there is a cremaster present, as long as the ninth and tenth abdom- inal segments, which bears hooked setae. The only known Arctiidae which have long cremasters are provided with flanged plates on the movable abdominal segments and the cremastral setae are never hooked. The prothoracic leg extends cephalad between the sculptured eye-piece and the antenna, but the mesothoracic leg never reaches the eye-pieces. The body is slightly punctate along the cephalic mar- gin of the m.ovable abdominal segments. The pupae are found in thin silken cocoons, which differ from those of the species of Arctiidae in that none of the larval hairs are used in their construc- tion. The following species was the only one examined : Cliaradra dcridcns Guenee. Subfamily Hypeninae The only genus available for study of this group as given in Dyar's Hst was Plathypena, but as Balsa possesses the pupal charac- ters which distinguish this from other subfamilies, it is included in the Hypeninae for the present. These characters are the presence of two long and six short hooked setae at the caudal end of the body and the fact that the prothoracic legs and the long point of the meso- thoracic legs extend cephalad to the eye-piece, or, as in Balsa, to the maxillary palpus, which is always present in this genus. Plathypena shows the epicranial suture, but it is not found in Balsa. Both genera have the spiracles slightly produced, and in Plathypena they are on small elevations. The labial palpi are present in both, but the femora of the prothoracic legs are visible only in Balsa. The genera may be separated as follows : 117 a. Dorsum of prothorax, metathorax, and first three abdominal seg- ments with a distinct median ridge ; antennae and mesothoraeic legs reaching the caudal margin of the wings and longer than the maxillae ; caudal margin of mesothorax never with a row of shal- low depressions Plailiypena Grote. aa. Dorsum of body never with a median ridge; antennae and meso- thoraeic legs not reaching the caudal margin of the wings and not longer than the maxillae ; caudal margin of mesothorax with a row of shallow depressions Balsa Walker. The following species were examined : Platliypena scabra Fabricius Balsa vialana Fitch Subfamily Catocalinae This group is distinguished from all other noctuids by the pres- ence of a whitish "bloom" on the surface of the pupa, which is re- tained even in alcoholic specimens. Both prothoracic and meso- thoraeic legs extend cephalad to the eye-pieces. The labial palpi are always present, but the femora of the prothoracic legs are seldom visible. The epicranial suture is found throughout the genera Catoc- ala and Eunetis but is lacking in the remainder of the subfamily. The antennae, mesothoraeic legs, and maxillae either reach the cau- dal margin of the wings or very closely approach it. The cremaster is usually very short or absent, and the setae at the caudal end of the body are usually of two sizes and inserted at different levels except in the genus Eunetis. This generic name is applied to certain species of the genus Catocala of Dyar's list which have a short cremaster, slightly broader at the caudal end, bearing about eight slightly curved setae which are usually directed towards the meson. The fol- lowing table will serve to separate the genera of Catocalinae : a. Epicranial suture present; body tapering rapidly from the fifth ab- dominal segment so that it is more slender in appearance than the typical noctuid, the lateral margins of abdominal segments 8-10 as seen in dorsal view convergent and not strongly convex, b. Cremaster, if present, very short and narrowed at the caudal end, and with eight long hooked setae of two sizes, some larger and more heavily chitinized than the others Catocala Schrank. bb. Cremaster broadened at the caudal end, usually with eight short, stout setae which are slightly curved and usually directed tow- ards the meson Eunetis Hiibner. aa. Epicranial suture not present; body of typical noctuid shape, with the lateral margins of abdominal segments 8-10 distinctly convex as seen in dorsal view. 118 b. Head with a distinct tubercle near the base of each antenna. EupartJienos Grote. bb. Head without a distinct tubercle at the base of each antenna. c. Thorax and appendages with deep indeterminate transverse striations; median caudal spines of cremaster somewhat en- larged near the tip ; metathoracic legs never present. PJieocyma Hiibner. ce. Thorax and appendages approximately smooth ; median caudal spines of cremaster never enlarged near the tip ; metathoracic legs always apparent caudad of the maxillae. d. Cremaster with spines practically all of the same size, no two being larger and longer than the others. .Caenurgia Walker. dd. Cremaster with two spines larger and longer than the others. Zale Hiibner. The following species were examined : Buparthenos nubilis Hiibner Pheocyma lunifera Hiibner Caenurgia erechtea Cramer, crassiuscula Haworth Zale calycanthata Smith and Abbot, lunata Drury Catocala illecta Walker, unijuga Walker, hriseis Edwards, verecunda Hulst, aholihah Strecker, ilia Cramer, inniibens Guenee, neo- gama, Smith and Abbot, pacta Linnaeus, sponsa Linnaeus Bunetis hlandnla Hulst, ultronia Hiibner, grynea Cramer Subfamily Sarrothripinae This group is readily distinguished because it has neither cre- master nor setae at the caudal end of the body, which is probably due to the fact that its members are found in thick cocoons. The dorsal surface of the body is very irregularly rugose with spinous elevations, and there is a distinct row of spines along the caudal mar- gin of the fifth abdominal segment extending from the rugose area on the dorsum around nearly to the meson of the ventral surface. A few spines are present in a similar position on the fourth abdominal segment. The epicranial suture is always present, the labial palpi are visible for their entire length, and only a small portion of the pro- thoracic femur is exposed, or it may be entirely concealed. The maxillae never reach the caudal margin of the wings, being about seven eighths of their length, with the mesothoracic legs meeting just caudad of them. The antennae always reach to the caudal mar- gin of the wrings, while the mesothoracic legs do not, but the latter are slightly longer than the maxillae. Both prothoracic and meso- thoracic legs extend cephalad to the eye-piece, the mesothoracic legs extending between the sculptured eye-piece and the antenna. 119 The members of this subfamily have been treated as the family Nycteolidae by some authors, but there is no evidence in the pupae to separate them from the family Noctuidae. The following species were examined : Sarrothripus revayana Scopoli, proteella Dyar. Family Arctiidae The members of this family all possess distinct setae arranged around the scars of the larval verrucae. These setae are seldom con- spicuous enough to be seen with the naked eye as in Liparidae, but are easily visible with the aid of the microscope. The labial palpi are never visible, unless as small triangular areas caudad of the labrum, except in Halisidota, where they are exposed for their entire length. The femora of the prothoracic legs are never visible. The shape of the body is characteristic, being slightly concave on the dorsum in the region of the metathorax (Fig. 104). Certain genera of Noctuidae, Acronycta, Eulonche, and Charadra, also show setae arranged around the scars of the larval verrucae. In the two genera first named, both labial palpi and prothoracic femora are exposed, while Charadra pos- sesses a long cremaster bearing hooked setae. Those genera of Arctiidae with a cremaster never have hooked setae, but all cremastral setae are flattened at the distal end. The epicranial suture is never present in any member of this family. The prothorax is usually long, often half the length of the mesothorax as in most Noctuidae. The genus Ctenucha, included with the Syntomidae in Dyar's list, shows no characters to distinguish it from the Arctiidae and it is probable that other genera of this family should be included here. The genera of Arctiidae may be separated as follows : a. Abdominal segments 5-7 never with a flanged plate along the ce- phalic margin, or with deep furrows between these segments when the body is retracted ; maxillae nearly as long as the wings ; mesotho- racic wings never meeting on the meson caudad of the appendages, h. Dorsal surface of abdomen flattened; body broadly rounded at caudal end and bearing a row of short setae which are slightly curved at tip ; body brown, concolorous. c. Labial palpi showing for about one sixth of the length of the max- illae; body 18-20 mm. in length; dorsal surface of abdomen never with depressed areas Halisidota Hlibner. cc. Labial palpi never showing except as a small triangular piece about a millimeter in length; body 12-15 mm. in length; dor- sal surface of abdomen with depressed areas on meson of each segment at cephalic margin and one on each side adjacent to the spiracles Euchaetias Lymaa. 120 bb. Dorsal surface of abdomen not flattened ; body tapering at caudal end ; color yellowish or chestnut-brown, strikingly marked with black. c. Distinct cremaster present Haploa Htibner. cc. Distinct cremaster never present, d. Antennae about seven eighths the length of the wings ; maxillae never reaching the caudal margin of the wings. TJietlieisa Hiibner. dd. Antennae as long as the wings ; maxillae always reaching the caudal margin of the wings Ctenucha Kirby. aa. Abdominal segments 5-7 with a flanged plate along the cephalic margin, and with deep furrows between the movable segments when the body is retracted; maxillae never as long as the wings, usually about two thirds the length; mesothoracic wings always meeting on the meson caudad of the appendages, b. Abdominal segments 4-6 with a similar flanged plate adjacent to the caudal margin of the segment. c. Appendages, and usually the thorax, roughened with indetermi- nate transverse striations ; abdominal segments densely, coarsely punctate, d. Head with a small tubercle at the proximal end of each an- . tenna; antennae always much shorter than the prothoracic legs Estigmene Hiibner. dd. Head without a tubercle at the proximal end of each antenna ; antennae as long as, or longer than, the prothoracic legs, e. Body usually 18-20 mm. in length, stout; setae of the cre- master of various sizes and lengths, the shortest ones only about half the length of the longest, and irregularly ar- ranged, always fifteen or more in number. Diacrisia Hiibner. ee. Body seldom over 12 mm. in length, varying from 10-15 mm., rather slender; setae of the cremaster of practically the same size and length, arranged in two transverse rows, with four in the dorsal and eight in the ventral row, never as many as fifteen in number Hypliantria Harris. cc. Thorax and appendages smooth and polished ; body sparsely, finely punctate EcpantJieria Hiibner. bb. Abdominal segments 4-6 never with a flanged plate adjacent to the caudal margin, c. Distinct cremaster always present and long, bearing setae at the caudal end ; antennae about equaling the maxillae in length, usually three fourths the length of the wings. Apantesis Walker, cc. Distinct cremaster never present ; a row of setae at the caudal end of the body ; antennae very much shorter than the max- illae, being about half the length of the wings. . . .Isia Walker. 121 The following species were examined : Halisidota tcssc'llaris Smith and Abbot, caryae Harris Buchaetias egle Drury Haploa clymene Brown Utetheisa bella Linnaeus Ctemicha virginica Charpentier Bstigmene acraea Drury Diacrisia virginica Fabricius Hyphantria cunea Drury Hcpantheria deflorata Fabricius Apantesis virgo Linnaeus, michaho Grote, arge Drury, phyllira Drury, nais Drury hia isabella Smith and Abbot Family Liparidae This family, like the Arctiidae, is characterized by the presence of setae arranged around the scars of the larval verrucae. In the Liparidae the setae are long and coarse, and easily visible to the un- aided eye. With the exception of Porthetria all the genera examined show a characteristic arrangement of appendages (Fig. 105). In Por- thetria the labial palpi were usually concealed by the maxillae, al- tliough a large number of pupae show them present, as in Figure 105. The epicranial suture is never present. The maxillae are always short, never more than two fifths the length of the wings. The legs are usually shorter than in most pupae, the mesothoracic legs never reaching the caudal margin of the wings. The antennae are pectinate and are longer and broader in the male than in the female. The cre- master is always present, smooth, longer than broad, and bears short hooked setae at the distal end. Most of the species examined show a remarkable uniformity of characters, and considerable difficulty was encountered in separating the genera. The difficulty lies in the fact that there is considerable difference between the sexes, not only in the length and breadth of the antennae, but in the size and arrangement of other appendages. In Hemerocampa the adult females are apterous, and the wings in the pupa are not as long as in the males. The wings of the females reach slightly over the cephalic margin of the fourth abdominal seg- ment, while in the male they reach to the caudal margin of that seg- ment. In Hemerocampa the dorsum of the first three abdominal seg- ments is covered on each side of the meson with small vesicles. The following table will serve to separate the genera of Liparidae : 122 a. Spiracular furrows never present on the cephalic margin of the mov- able abdominal segments; labial palpi present and well developed; long setae not present on the face-parts near the caudal margin of the head, on the clypeus, or on the sculptured eye-pieces, b. Labial palpi about equal in width to each maxilla; cremastral setae about one third the length of the cremaster. c. Dorsal surface of abdomen densely covered with long setae ; body brown, concolorous except sometimes for the lighter transverse conjunctiva ; antennae with the stem of the flagellum very broad and distinctly elevated, d. Abdominal segments 8-10, in dorsal view, distinctly narrowed and tapering; clypeus and labrum on a level with the other face-parts; cremaster directed caudad; antennae of male reaching about half the distance along the mesothoracic legs, and distinctly pointed Dasycliira Stephens. dd. Abdominal segments 8-10, in dorsal view, not narrowed and distinctly tapering ; clypeus and labrum both elevated above the level of the other face-parts ; cremaster directed dorsad ; distal half of the antennae of the female distinctly narrowed and pointed, reaching about half the distance along the mesothoracic legs Olene Hiibner. ec. Dorsal surface of abdomen not very densely covered with setae ; body white variously marked with brown, or vice versa ; anten- nae in both sexes with the stem of the flagellum scarcely indi- cated and the distal end always rounded, in the male extend- ing about half the length of the wings and curved mesad so that they often meet, in the female never reaching half the distance along the mesothoracic legs and never meeting on the meson. Hemerocampa Dyar. bb. Labial palpi only half the width of each maxilla; cremastral setae about half the length of the cremaster; antennae in the male broad for almost their entire length and extending about three fourths the length of the wings, adjacent or meeting on the meson, in the female pointed at the distal end and reaching about half the distance along the mesothoracic legs. Euproctis Hiibner. aa. Spiracular furrows always present on the cephalic margin of the movable abdominal segments and extending almost to the meson, five or six in number and separated by sharply carinate ridges; caudal portion of face-parts, clypeus, and sculptured eye-pieces with coarse setae similar to those on the body, labial palpi often con- cealed, but sometimes visible as in the remainder of the family. PortJietria Hiibner. The following species were examined : Dasychira pndibnnda Linnaeus Olene manto Strecker 123 Hemerocampa leucostigma Smith and Abbot Porthetria dispar Linnaeus Buproctis chrysorrhoca Linnaeus SuPERFAMiLY BOMBYCOIDEA This superfamily includes those famihes in which the body is more or less densely covered with setae and which usually retain the labial palpi. They seem to be more nearly related to the Satur- nioidea than to any other superfamily, although the Lasiocampidae show certain points of relationship with the Liparidae. All the members of the superfamily, so far as known, are found in thick silken cocoons, much like those of the Saturniidae. The families of Bombycoidea may be separated as follows : a. Epicranial suture present; labial palpi visible Lasiocampidae. aa. Epicranial suture never present ; labial palpi concealed by the maxil- lae except for a small triangular area at the proximal end. BOMBYCIDAE. Family Lasiocampidae The members of this family usually have mouth parts and appen- dages arranged as in the Liparidae (Fig. io6). The epicranial su- ture is always present and the vertex is longer than that found in any but the more generalized forms. The maxillae are short, never more than one third the length of the wings, and extend very slightly be- yond the distal ends of the labial palpi, or may be shorter than the palpi. The antennae are broad and pectinate, and never extend as far caudad as the prothoracic legs. The coxae of the prothoracic legs and sometimes of the mesothoracic pair are often visible caudad of the maxillae and labial palpi. The prothoracic legs are slightly more than half the length of the wings and the mesothoracic legs never reach the caudal margin of the wings. The setae of the body are very con- spicuous except in Tolype, but are not arranged around the scars of the larval verrucae. The movable segments are capable of being re- tracted till only their caudal margins are visible. There is never a cremaster present, and there are no hooked setae at the caudal end of the body. The body is broadly rounded at the caudal end and the body setae are usually a little longer and coarser in this region. This family is considered by many authors to be more specialized than any of the Saturnioidea, but the presence of the epicranial su- ture and exposed labial palpi shows that they are much more gener- alized. The genera of Lasiocampidae may be separated as follows : 124 a. Entire surface of body except the appendages with a dense covering of fine short setae, giving it a furred appearance ; abdominal seg- ments 8-10 considerably narrower than the remainder of the ab- domen Malacosoma Hiibner. aa. Entire surface of body except appendages never with a dense cover- ing of setae giving it a furred appearance, either with a very sparse covering, or without visible setae ; abdominal segments 8- 10 not noticeably narrower than the remainder of the abdomen, b. Body sparsely covered with very fine short setae except at the cau- dal end, where they are longer and coarser ; tenth segment broadly rounded at the caudal end. c. Lateral surface of abdomen on cither side of the spiracles with setae much thicker than on the dorsum ; dorsal surface of abdo- men very coarsely punctate Lasiocampa Schrank. cc. Lateral surface of abdomen on eitlici" side of the spiracles never with the setae more numerous than on the dorsum ; dorsal sur- face of abdomen very finely punctate. . . .Cosmotriclie Hiibner. bb. Body without any visible covering, of setae ; tenth segment abruptly narrowed at the caudal end, suggesting a cremaster. Tolype Hiibner. The following species were examined : Malacosoma americana Fabricius, disstria Hiibner Lasiocampa quercus Linnaeus (Europe) Cosmotrichc potatoria Linnaeus (Europe) Tolype vclleda Stoll Family Bombycidae This family includes the genus Bombyx, which is domesticated in various parts of the world (Fig. 107). The body is covered with rather coarse short setae which are somew-hat longer at the caudal end of the body. The epicranial suture is never present. The ap- pendages are arranged much as in Lasiocampidae. The labial palpi are almost concealed, being overlaid by large ovate appendages, ap- pearing to be heavily veined, which are presumably the maxillae (no dissections were made of this species). The mandibles are repre- sented by strongly elevated tubercles. The coxae of one pair of legs, probably the mesothoracic, are adjacent on the meson caudad of the maxillae. The legs are short, and both prothoracic and mesothoracic pairs lie adjacent on the meson. The antennae are pectinate, broad at the proximal end and rapidly narrowed, ending in a point oppo- site the distal ends of the prothoracic legs. The mesothoracic wings He adjacent on the meson caudad of the mesothoracic legs. The movable segments may be retracted so that only their caudal margins are visible. There is no cremaster present. The pupa strongly re- 125 sembles those of certain species of Saturniidae and it is quite probable that they had a common ancestor, although the Bombycidae are un- doubtedly more generalized. The only species in America is Bom- by.v niori Linnaeus. SuPERFAMiLY NOTODONTOIDE A The families included here never have the entire labial palpi ex- posed, but a very small triangular or polygonal area is sometimes visi- ble just caudad of the labrum. Some genera of Geometridae have the prothoracic femora exposed, while the epicranial suture is present in some of the members of each family. Although the larvae of Geomet- ridae are easily recognized and are very readily distinguished from those of the Notodontidae, the pupae show much closer relationships, and it is difficult to draw a hard and fast line between the two fami- lies. The three families included here have probably had a common ancestor, and although the Dioptidae retain the epicranial suture they must be considered the most specialized, as both Geometridae and Notodontidae show more generalized characters in some of their gen- era. The following table may be used to separate the families of Notodontoidea: a. Antennae never extending beyond the caudal margin of the wings ; dorsum of abdomen never with a prominent hooked seta on each side of the meson of segments 7-10. b. Maxillae usually more than three fifths the length of the wings, if not, then the caudal end of the body with hooked setae, or the spiracles of the third abdominal segment concealed by the wings and those of the sixth segment farther ventrad than those of the other segments ; prothoracic femora often exposed ; a deep furrow usually present on the dorsum of the abdomen between the ninth and tenth segments ; caudal margin of mesothorax never with a row of deep pits with smooth tubercle-like areas between. Geometridae. bb. Maxillae seldom exceeding three fifths the length of the wings, if so, then the caudal margin of the mesothorax with a row of deep pits with smooth, elevated, quadrangular tubercle-like areas be- tween them ; or with the entire body surface coarsely punctate ; abdominal spiracles of the third segment never concealed by the wings, and those of the sixth never farther ventrad than the re- mainder ; prothoracic femora never exposed ; a furrow never present on the dorsum of the abdomen between segments 8 and 9 except in Datana, where the cremaster is of the type shown in Figure 112 Notodontidae. 126 aa. Antennae extending beyond the caudal margin of the wings ; dorsum of abdomen with a prominent hooked seta on each side of the meson of segments 7-10 Dioptidae. Family Geometridae The pupae in this family (Figs. io8 and 109) never have the labial palpi exposed except as small triangular or polygonal areas caudad of the labrum. The maxillary palpi are never present. The epicranial suture is present in. some of the genera in Groups A and C. Some of the genera have the femora of the prothoracic legs exposed. In all the genera examined, either the prothoracic or mesothoracic leg, or sometimes both, extended cephalad between the sculptured eye-piece and the antenna. The prothoracic and mesothoracic legs are longer than is usual in most families, the former being usually three fourths the length of the wings, while the latter always extend to the caudal margin of the wings. Many of the genera show the fronto-clypeal su- ture extending from the proximal ends of the antennae and directed caudad towards the invaginations for the anterior arms of the ten- torium. The suture is very distinct for the cephalic part and is often indicated by a slight furrow for the remainder of the distance. In the genus Haematopsis there is a prominent cephalic projection bearing hooks which hold the suspensory threads, as this pupa is not found in a cocoon. The antennae vary little throughout the family, and are generally equal in width to, or wider than, the prothoracic legs, usually extending to the caudal margin of the wings. The meso- thoracic wings usually extend farther caudad than in the nearly related families, reaching almost to the caudal margin of the fourth abdom- inal segment, although not visible in ventral view. The mesothorax is very short in some genera, particularly in those of Group D, where it is never twice the length of the prothorax. The mesothoracic spir- acles very often have a decided projection adjacent to their caudal margin, which is usually flattened or tuberculate in form and often covered with very fine short setae. The abdominal spiracles are some- times produced, and in nearly all genera the sixth spiracle, and some- times the seventh also, is considerably farther ventrad than the re- mainder. The abdomen is usually coarsely punctate, sometimes rough- ened with deep transverse striations. A cremaster of some kind is al- ways present. One of the most interesting structures of the abdomen is the dorsal furrow between the ninth and tenth segments. This is usually deep and fringed with very fine setae. This furrow fre- quently projects caudad on the lateral surface of the body and this portion is often separated from the dorsal furrow. This dorsal fur- 127 row is present in all members of Group A except Ennomos and in a few members of Group D. These dorsal furrows are found in other families, notably in the Gelechiidae and some Pyralididae. The attempt to classify the pupae of Geometridae was seriously hampered by lack of material. Reared material was very hard to ob- tain, as the larvae develop slowly and seem to be very susceptible to disease. The available material, moreover, did not seem to fall in with any of the existing schemes of classification, so that the only practical solution of the difficulty was to divide them into groups ac- cording to the pupal characters. These may or may not be natural groups, but they will serve to indicate relationships in some degree. According to the pupal characters there seem to be two large divi- sions, one with hooked setae on the cremaster, the other without them. As the presence of hooked setae is a more generalized character, the groups in which they are present should be the more generalized, and the presence of the epicranial suture strengthens this view. Group A includes representatives of the subfamilies Sterrhinae, En- nominae, and Hydriomeninae as listed by Dyar. Group B includes for the present only the genus Haematopsis. Group C includes the genera Alsophila and Brephos, which are similar in many respects and possess the same type of cremaster. This must also be considered as a generalized group since the epicranial suture is present in Alsophila. Brephos has usually been considered as the most generalized geomet- rid, but no epicranial suture has been located. The maxillae are also much shorter in Alsophila. In Group D the epicranial suture is never present. Spiracular furrows are frequently found in Groups A and D. The adult females of some of the geometrid species are apterous. Although abundant material of one such species, Paleac- rita veriiata, was examined, the pupal wings were found as well developed in the female as in the male. These groups of Geometridae may be separated as follows : a. Cremaster with hooked setae. b. Setae of the cremaster always of two sizes, usually either the two or four farthest eaudad much larger than the others ; dorsum of abdomen usually with a deep furrow between abdominal seg- ments 9 and 10 ; body never with a bifurcate projection at the cephalic end Group A. bb. Setae of the cremaster always of the same size; dorsum of abdo- men never with a furrow between segments 9 and 10 ; body with a long bifurcate projection at the cephalic end, densely covered with hooked setae Group B. 128 aa. Cremaster never with hooked setae. b. Cremaster a stout T-shaped spine Group C. bb. Cremaster always bifurcate Group D. Group A This group includes species in which the cremastral setae are of two sizes, and which generally show the epicranial suture, and a dor- sal furrow between the ninth and tenth abdominal segments (Fig. 109). Spiracular furrows are also frequently present. The genera included in this group may be separated as follows : a. Epicranial suture always present ; prothoracic femora always ex- posed, usually a large portion ; spiracular furrows never present ; dorsum of abdomen with a distinct furrow between segments 9 and 10, and also on the lateral surface of segment 10. b. Dorsum of fifth abdominal segment with a distinct furrow on the cephalic margin ; caudal margin of the furrow between the ninth and tenth abdominal segments with finely serrate edges and with- out setae Hydria Hiibner. bb. Dorsum of fifth abdominal segment never with a distinct furrow on the cephalic mai-gin ; caudal margin of the furrow between the ninth and tenth abdominal segments coarsely serrate. c. Prothoracic and mesothoracic legs extending cephalad between the sculptured eye-piece and the antenna, the mesothoracic leg extending almost as far cephalad as the prothoracic ; caudal margin of the furrow between the ninth and tenth abdominal segments without setae ; cremaster distinctly constricted at the proximal end and circular in outline. . : Eois Hiibner. cc. Prothoracic leg extending cephalad between the sculptured eye- pieces and the antenna, the mesothoracic leg never reaching farther cephalad than the caudal margin of the eye-piece ; cau- dal margin of the furrow between the ninth and tenth ab- dominal segments usually fringed with fine setae ; cremaster not constricted at the proximal end, triangular in outline, d. Cremaster with hooked setae of nearly the same size, the me- dian caudal setae being only slightly larger than the others; three setae inserted along each side of the cremaster, the other two cephalad of these and slightly mesad. Teplirodystis Hiibner. dd. Cremaster with the hooked setae of widely differing sizes, the two median caudal setae being much larger and longer than the others, a smaller seta adjacent on each side, while the other setae are arranged in a tranverse row just cephalad of the caudal setae Cinglis Guenee. 129 aa. Epicranial suture never present. b. Prothoraeic femora exposed, but only a very narrow portion. c. Spiracular furrows present on the fifth abdominal segment ; ven- tral surface of head with a prominent transverse ridge extending from about the middle of the glazed eye-piece ; body setae arising from small dark brown or black papillae. .Philohia Duponchel. cc. Spiracular furrows never present on the fifth abdominal seg- ment; ventral surface of head never with a transverse ridge; body setae arising from small pits Sahulodes Guenee. bb. Prothoraeic femora never exposed, c. Antennae distinctly elevated and covered with five or six rows of small round tubercles; distinct ridges or flanged plates present along both margins of the movable segments, d. Cremaster with the two median caudal spines very much larger than the others ; distinct furrow always present between the ninth and tenth abdominal segments ; color dark brown. Nacopliora Hulst. dd. Cremaster with the four caudal spines about the same size and much larger than the others ; distinct furrow never present between the ninth and tenth abdominal segments; color usually white, sometimes partly brown. Ennomos Treitschke. cc. Antennae elevated but never with distinct rows of tubercles ; ridges or flanged plates never present on the movable segments. d. Maxillae about two thirds the length of the wings ; prothoraeic mesothoracic, and metathoracic legs all meeting on the meson caudad of the maxillae XonfJiotype Warren. dd. Maxillae always more than two thirds the length of the wings ; prothoraeic and mesothoracic legs never meeting on the meson caudad of the maxillae, e. Spiracular furrows present on the cephalic margin of the fifth abdominal segment, with more or less interrupted ridges between ; cremaster with the four caudal setae always larger than the others Ania Stephens. ee. Spiracular furrows never present on the cephalic margin of the fifth abdominal segment, cremaster with two caudal setae larger than the others Cingilia Walker. Group B The pupae of this group are easily distinguished by the long bifur- cate projection at the cephalic end of the body, which is covered with hooked setae. There are never spiracular furrows present on the ce- phalic margin of the movable segments. The mesothoracic legs meet on the meson caudad of the maxillae. The body is very slender and never punctate. This group includes the genus Haematopsis. 130 Group C The species of this group are distinguished by the peculiar T- shaped cremaster. The genera may be separated as follows : a. Epicranial suture present ; prothoracic and mesothoracic legs meeting on the meson caudad of the maxillae Ahopliila Htibner. aa. Epicranial suture never present; prothoracic and mesothoracic legs never meeting on the meson caudad of the maxillae. Breplios Ochsenheimer. Group D This group is characterized by the presence of a bifurcate cre- master. The epicranial suture is never present, but a portion of the prothoracic femora is exposed in many genera. The following table will serve to separate the genera of this group : a. Prothoracic femora visible. b. Cephalic margin of the fifth abdominal segment with one deep spiracular furrow with strongly chitinized edges, just cephalad of each spiracle. c. Mesothoracic spiracle with a broad, very strongly elevated ridge or oval tubercle adjacent to its caudal margin which is covered with fine short setae; surface of spiracular furrow with distinct punctures Pliysostegania Warren. cc. Mesothoracic spiracle with only a very narrow, slightly elevated ridge adjacent to its caudal margin, usually covered with setae ; surface of spiracular furrow without distinct punctures, d. Dorsal surface of abdomen never with a distinct, furrow be- tween the ninth and tenth segments, its caudal margin finely serrate; surface of the spiracular furrow almost smooth. Ectropis Hiibner. d. Dorsal surface of abdomen never with a distinct furrow between the ninth and tenth segments ; surface of the spiracular fur- rows strongly rugose Cleora Curtis. bb. Fifth abdominal segment without any deep spiracular furrow; cephalic margin of the segment deeply punctate, the punctures sometimes confluent; mesothoracic spiracle with a broad strongly elevated ridge or tubercle adjacent to its caudal mar- gin; dorsal surface of abdomen without a distinct furrow be- tween the ninth and tenth segments. . . .Cymatopliora Hiibner. aa. Prothoracic femora never visible; deep spiracular furrows always present on the cephalic margin of the fifth abdominal segment; mesothoracic spiracle always with a prominent elevation adjacent to its caudal margin. b. Dorsal surface of abdomen never with a furrow between the ninth and tenth segments ; abdominal spiracles very strongly produced ; 131 cremaster often showing two lateral setae on each side near the proximal end. bb. Dorsal surface of abdomen with a distinct furrow between the ninth and tenth segments, its caudal margin coarsely serrate ; a prominent lateral depression or furrow present on the lateral surface of the tenth abdominal segment. c. Maxillae four fifths the length of the wings or less ; mesothoracic legs meeting on the meson just caudad of the maxillae; an- tennae more than twice as wide at the proximal as at the distal end Lycia Hiibner. cc. Maxillae always more than four fifths the length of the wings ; mesothoracic legs never meeting on the meson caudad of the maxillae ; antennae of almost the same width throughout, never twice as wide at the proximal as at the distal end. Erannis Hiibner, The following species were examined : Group A Hydria undiilata Linnaeus Bois inductata Guenee Tephroclystis interruptofasciata Packard, absinthiata Clerck Cinglis similaria Walker Philobia enotata Guenee Sabulodes lorata Grote, transversata Drury Nacophora qiierimria Smith and Abbot Bnnomos subsignarius Hiibner, magnarius Guenee Xanthotype crocataria Fabricius Ania limbata Haworth Cingilia catenaria Drury Cosymbia serrnlata Packard Group B Haematopsis grataria Fabricius Group C Alsophila pometaria Harris Brephos infans Moschler Group D Physostegania pustularia Guenee Hc'tropis crepusctdaria Denis and Schiffermueller Cleora pampinaria Guenee Cymatophora ribearia Fitch Palcacrita vernata Peck Lycia cognataria Guenee Brannis tiliaria Harris 132 Family Notodontidae The pupae of this family never show more than a small triangular or polygonal proximal portion of the labial palpi, and maxillary palpi are never present. The femora of the prothoracic legs are never ex- posed. The epicranial suture is present in the genera Apatelodes and Melalopha. The maxillae never reach the caudal margin of the wings. The antennae are always widest at their proximal ends, and there the width exceeds the greatest width of the prothoracic legs. Each an- tenna tapers gradually to a pointed tip and the tips often lie adjacent on the meson caudad of the other appendages. The metathoracic legs are seldom visible. The mesothoracic leg never reaches cephalad to the eye-pieces. The abdomen is always punctate and in most species the punctures are large. A cremaster is usually present and there are various types, as in Figures iii, 112, 113. Packard divided the Noto- dontidae into six subfamilies. The pupae examined show that these subfamilies are well founded, but only tables to genera are given here as so few species of Notodontidae were examined. The genera Schizura and Heterocampa are not w^ell defined and the species are separated with difficulty. The species are listed, however, under the subfamily name. Some authors believe that the genus Apatelodes belongs to the European family Eupterotidae, and is incorrectly listed with the Noto- dontidae. As no pupae of Eupterotidae have been examined, it is im- possible to say whether pupal characters would justify this change. There are, however, no pupal characters as far as observed, which would prevent its being included with the Notodontidae. The two species differ widely and are possibly not congeneric. The following tables will serve to separate the genera of Notodontidae : a. Maxillae one third, or less, the length of the wings ; both prothoracic and mesothoracic legs meeting on the meson caudad of the maxillae ; abdomen very finely punctate. b. Thorax and abdomen thickly covered with very fine, short setae; cremaster a stout spine about one millimeter in length with two short recurving hooks at the tip, each of which bears two or more very fine setae Meldoplia Hiibner. bb. Thorax and abdomen never thickly covered with fine, short setae ; cephalic margin of first abdominal segment without tubercles; cremaster never as described above, sometimes absent, c. Abdominal segments 2-7 with distinct flanged plates at both cephalic and caudal margins, the cephalic plate interrupted by deep pits, giving it the appearance of a row of square tuber- cles; appendages not at all elevated, making a smooth even 133 surface ; cephalic end of body not elevated between the an- tennae ; short eremaster sometimes present. Apatclodes Packard, cc. Abdominal segments 2-7 never with flanged plates; appen- dages distinctly elevated ; cephalic end of body elevated be- tween the antennae ; eremaster never present. Harpyia Ochsenheimer. aa. Maxillae ahvays more than one third the length of the wings; never with both prothoracic and mesothoracic legs meeting on the meson ; abdomen usually rather coarsely punctate, b. Maxillae from one half to three fifths the length of the wings; mesothoracic legs meeting on the meson caudad of the maxillae ; appendages roughened with deep indeterminate striations ; ab- dominal segments coarsely punctate ; a distinct, deep furrow on the dorsum between the ninth and tenth abdominal segments; eremaster short, bifurcate, each half with several short spiny processes, usually directed laterad Batana Walker. bb. Maxillae more than three fifths the length of the wings ; neither prothoracic nor mesothoracic legs meeting on the meson caudad of the maxillae ; appendages usually with shallow indetermi- nate striations; a distinct furrow never present on the dorsum between the ninth and tenth al)dominal segments ; eremaster not as described above, c. Entire body surface with coarse, deep punctures ; cephalic mar- gin of the movable abdominal segments with large lunate punctures and a distinct ridge with a row of large, very dis- tinct punctures just caudad of the ridge; eremaster short, rugose, slightly bifurcate, bearing six long hooked setae ; meso- thorax never with a deeply pitted caudal margin. Synimerisfa Hlibner. cc. Body usually punctate on the abdomen, but not on the appen- dages; movable abdominal segments sometimes with a slight ridge along the cephalic margin, but never with a distinct row of large punctures caudad of the ridge ; eremaster bifur- cate, but never with hooked setae ; mesothorax with a row of deep pits along its caudal margin with smooth quadrangular areas between, and partly covering them. d. First abdominal segment with a small tubercle on each side of the meson at the cephalic margin of the segment ; entire dor- sal surface of the tenth segment distinctly elevated and very rugose ; points of the eremaster divergent. Hyparpax Hlibner. dd. First abdominal segment without tubercles ; entire dorsal surface of the tenth segment not elevated and rugose. e. Wings always touching on the meson ; maxillae never as long as the wings ; cephalic end of body sometimes with two sharp, heavily chitinized spinous projections. Scliizura Doubleday. 134 ee. Wings adjacent on the meson but not touching; maxillae usually as long as the wings ; cephalic end of body without heavily chitinized spinous projections. Heterocampa Doubleday. The following species were examined : Melalophinae Melalopha inclusa Hiibner, apicalis Walker, albosigma Fitch Apatelodinae Apatclodcs torrefacta Smith and Abbot, angelica Grote Cerurinae Harpyia horcalis Boisduval Pygaerinae Datana niiiiistm Drury, modesta Beutenmiiller, angiisii Grote and Robinson, chiriquensis Dyar, contracta Walker, drexclii Hy. Edwards, intcgcrrima Grote and Robinson, major Grote and Robinson, paluiii Beutenmiiller, rohusta Strecker Notodontinae Synnncrista alhifrons Smith and Abbot Heterocampinae Hyparpax aurora Smith and Abbot Schizura ipomoeae Doubleday, concinna Smith and Abbot, unicornis Smith and Abbot Heterocampa giittivitta Walker, bUiucata Packard Family Dioptidae The pupae of this family closely resemble those of the Geomet- ridae, but are more specialized than most of the genera in that fam- ily, although they show the epicranial suture (Fig. 115). The ap- pendages are arranged very much as in the Geometridae, but there is no trace of labial palpi, maxillary palpi, or prothoracic femora (Fig. 114). The antennae are filiform, extending beyond the caudal mar- gin of the wings and about half way down on to the fifth abdominal segment. Each prothoracic leg extends cephalad between the sculp- tured eye-piece and the antenna. The distal ends of the prothoracic and mesothoracic legs and the maxillae are overlaid by the antennae, which lie adjacent on the meson at their distal ends. The abdomen is elevated at the dorso-meson to form a low ridge, and there are promi- nent hooked setae present on segments seven to ten as well as on the cre- -master. This family contains a single American species, Phryganidia calif arnica Packard. The family has usually been placed between the Noctuidae and Notodontidae, and widely separated from the Geomet- ridae. The pupa shows no relationship to the noctuids, and is much more highly specialized than most members of that family. 135 SUPERFAMILY SPHINGOIDEA The members of this superfamily retain but one generahzed char- acter, the presence of exposed portions of the prothoracic femora in some of the more generahzed forms. The shape of the pupa is almost as distinctive as that of the larva, being usually fusiform, often with the head distinctly narrower than the thorax, giving the body a "shouldered" appearance. The epicranial suture is never present, the only distinct head suture remaining being that adjacent to the proxi- mal end of each antenna. The wings and maxillae are unusually long in most members of this superfamily and various means are taken to accomodate the extra length, particularly of the maxillae. The fourth abdominal segment is usually longer on the ventral surface than on the dorsal, and the w'ings are seldom broadly rounded at their caudal margins, but usually somewhat pointed. The position of the head is also changed in many species and found almost, or entirely, on the dorsal surface of the body. The mandibles are often very conspic- uous, being represented by strongly elevated tubercles. The protho- racic legs are usually about half the length of the wings and the meso- thoracic legs three fourths of their length. The metathoracic legs are seldom visible. The antennae are for the most part filiform and vary from two fifths to three fourths the length of the wings. In the genera Smerinthus, Paonias, Marumba, and Cressonia the antennae are con- siderably wider at their proximal end and slightly pectinate, being larger and longer in the male, and the whole appearance of the body reminds one strongly of the Saturniidae. These genera are in many respects the most specialized of the Sphingoidea, and some of them are found in cocoons. It is an interesting fact that the most spe- ciaHzed forms in nearly all of the subfamilies of Sphingidae exam- ined, show relationship to the Saturniidae. This group is therefore considered as related to the Saturnioidea but more generalized. Cer- tain of its members resemble in some respects the Pyralididae and Gelechiidae. A cremaster is always present, usually triangular in out- line and often slightly bifurcate at the distal end. The abdomen often shows three or four transverse depressions on each segment which cor- respond to the annulet-like rings on the body of the larva. Except in rare instances the scar of the caudal horn of the larva is visible on the dorsum of the eighth abdominal segment. This superfamily contains a single family, the Sphingidae. For the most part the genera are easily distinguished, but there were no char- acters found that served to separate the genera Smerinthus and Paonias. The generic names of Dyar's list have been used as far as possible. A monograph of this family, giving tables and descriptions 136 for the identification of subfamilies, genera, and species, has been pre- pared and will be published at an early date. The following table to genera is given simply for the identification of specimens and does not indicate natural relationships. The genera may be separated as fol- lows: a. Spiracular furrows present on all the movable segments, that is, on abdominal segments 5—7. b. Maxillae with a free portion or so-called "raised tongue-case" present — the maxillary loop, c. Maxillary loop with its distal half recurved and touching the proximal portion, forming a prominent loop on the ventral sur- face of the body Herse Oken. cc. Maxillary loop never with the distal half recurved, the distal end touching the ventral surface of the body, d. Maxillary loop strongly arched from the ventral surface of the body, the greatest width of the space between the free por- tion of the maxilla and the ventral surface of the body always greater than the width of the maxillary loop in that region, e. Dorsum of body with a strongly rugose triangular area on each side of the meson of the metathorax and a similar rugose band on abdominal segments 2-8 extending over the cephalic third of the segment; lateral margins of the max- illary loop with a serrate appearance ; the distal end of the loop swollen and bulb-like and at least twice the width of the remaining portion Cocytius Hiibner. ee. Dorsum of body with a strongly elevated transverse ridge on each side of the meson on the metathorax, the ab- dominal segments punctate, but never rugose on the ce- phalic third; lateral margins of the maxillary loop never serrate in appearance ; the distal end of the loop some- what swollen but never twice the width of the remaining portion PlilegctJiontius Hiibner. dd. Maxillary loop never strongly arched, but usually closely applied to the ventral surface of the body, the space be- tween the loop and the ventral surface of the body never so great as the width of the loop in that region, e. Cephalic margin of abdominal segments 5-7 with one deep pocket-like furrow over each spiracle, f. Maxillary loop extending as far caudad as the distal ends of the prothoracic legs and occasionally beyond them. Atreus Grote. ff. Maxillary loop never extending as far caudad as the dis- tal ends of the prothoracic legs Dolha Walker. ee. Cephalic margin of abdominal segments 5-7 with two fur- rows, the ectal furrow shallow, the ental one deep and pocket-like. 137 f. Spiracular furrows 5 mm. or more in transverse length and extending ventrad of the spiracle for a distance equal to the length of the spiracle Clilaeno gramma Smith. ff, Spiracular furrows always less than 5 mm. in transverse length, seldom extending ventrad of the spiracle, if so, then for a distance less than the length of the spiracle. Spliinx Linnaeus. bb. Maxillae of the usual type without a maxillary loop or so-called ' ' raised tongue-case. ' ' c. Cephalic margin of abdominal segments 5-7 always with one deep pocket-like furrow over each spiracle, with or without a shallow ectal one. d. With one deep pocket-like furrow over each spiracle on abdom- inal segments 5-7. e. Surface of body spinose; cremaster broad and truncate, the caudo-lateral angles usually produced into sharp points; caudal abdominal segments flattened on the ventral sur- face, and with distinctly carinate lateral margins. Cressonia Grote and Robinson, ee. Surface of body never spinose ; cremaster pointed, trian- gular in outline, caudal abdominal segments never flat- tened on the ventral surface, nor with distinctly carinate lateral margins, f. Maxillae normally reaching to the caudal margin of the wings, slightly less in some individuals ; mesothoracic wings never meeting on the meson caudad of the max- illae; scar of the caudal horn never present on the dor- sum of the eighth abdominal segment. . .Lapara Walker, ff. Maxillae never more than five sevenths the length of the wings ; mesothoracic wings always meeting on the meson caudad of the maxillae ; scar of the caudal horn always present on the dorsum of the eighth abdominal segment. Daremma Grote. dd. With one deep pocket-like ental furrow and a shallower ectal one ; maxillae about two. thirds the length of the wings. Ceratomia Harris, cc. Cephalic margin of abdominal segments 5-7 with three or four more or less interrupted furrows over ea(^h spiracle, the sur- face of the furrows often punctate like the remainder of the cephalic margin, d. Maxillae never half the length of the wings; average length of maxillae at meson 5-6 mm., sometimes 7 mm. in large speci- mens; dorsal cephalic margin of abdominal segments 5-7 deeply punctate, the punctures adjacent to each other, giving it a honeycombed appearance, the cephalic margin separated 138 from the remainder of the segment by a distinct carinate ridge. \8merintlius Latreille. \Paonias Hiibner. dd. Maxillae half the length of the wings ; average length of maxillae at meson 10 mm., sometimes exceeding this length; dorsal cephalic margin of abdominal segments 5-7 .punc- tate, the punctures somewhat scattered and not presenting a honeycombed appearance ; cephalic margin never sep- arated from the remainder of the segment by a distinct ridge Marumha Moore. aa. Spiracular furrows never present on all of the movable segments, b. Spiracular furrows present on either one or two of the movable segments. c. Cephalic margin of abdominal segments 5 and 6 with three or four furrows over each spiracle, the furrows separated by sharply carinate ridges and extending almost to the meson on both dorsal and ventral surfaces ; cremaster sparsely covered with short curved spines on the dorsal and lateral aspects. Hemaris Dalman. cc. Cephalic margin of fifth abdominal segment with three or more shallow furrows over each spiracle, which never extend on to the dorsal and ventral surfaces; cremaster never with short curved spines on any portion ; labrum always on the dorsal surface of the head, d. Prothoracic femora apparent. e. Mandibular area always elevated and usually forming prom- inent tubercles;. cephalic margin of fifth abdominal segment with one ental furrow over each spiracle, extending for the entire width of the lateral aspect, and five or six interrupted ectal furrows with interrupted carinate ridges between ; body surface dull ; color light brown or coffee-color with darker markings Deilepliila Ochsenhcimer. ee. Mandibular area never elevated; cephalic margin of fifth abdominal segment with either three or four entire fur- rows over each spiracle; body surface highly polished; color black with red markings ; abdominal segments finely punctate on the cephalic half. . . .Dilophonota Burmeister. dd. Prothoracic femora never apparent; cephalic margin 'of fifth abdominal segment with six or more interrupted furrows with short, more or less wavy interrupted ridges between; cephalic portion of maxillae slightly excurved and almost carinate on the meson ; abdominal segments punctate over the entire length Therefrd Hiibner. bb. Spiracular furrows never present on any of the movable abdom- inal segments. c. Pupae always less than two inches in length ; labrum usually at the cephalic end of the body but never on the dorsal surface. 139 d. Pupae with a prominent tubercle on the face-parts, just mesad of each glazed eye, and a prominent tubercle on the labram; maxillae with the proximal portion excurved and carinate on the meson Deidamia Clemens. dd. Pupae never with prominent tubercles on the face-parts or labrum ; maxillae never with the proximal portion excurved and carinate on the meson. e. Abdominal segments 5-7 with one or more interrupted rows of spines along the cephalic margin of the segment, more prominent on the dorsal surface Darapsa Walker, ee. Abdominal segments 5-7 never with spines along the ce- phalic margin of the segments, f. Mandibular area with prominent tubercles. g. Body surface rough, deeply punctate over the entire sur- face of the abdominal segments, especially segments 8-10 ; body tapering rapidly from the fourth abdom- inal segment to the long pointed cremaster ; femora of prothoracic legs never apparent. . . .AmpJiion Hiibner. gg. Body surface smooth and polished, the cephalic portion of the segment with punctures of medium size, the caudal portion finely, sparsely punctate ; body taper- ing gradually from the fourth abdominal segment to cremaster; prothoracic femora apparent. Lepisesia Grote. fp. Mandibular area never elevated or with prominent tuber- cles, g. Dorsum of eighth abdominal segment with a row of large deep punctures or pits along the cephalic mar- gin ; cremaster slender, never more than 2 mm. in length; body always light colored with darker mark- ings Ampelopliaga Bremer and Grey. gg. Dorsum of eighth abdominal segment never with a row of large punctures or pits along the cephalic margin; cremaster always more than 2 mm. in length, the breadth equal to the length ; body always dark brown. SpJiecodina Blanchard. cc. Pupae always more than two inches in length ; labrum either at the cephalic end or on the dorsal surface of the body. d. Color black, often marked with red ; body surface smooth and polished, with a very few small punctures on the abdominal segments; labrum at the cephalic end of the body. PseudospMnx Burmeister. dd. Color uniform dark brown ; body surface roughened and densely, coarsely punctate ; labrum on the dorsal surface of the body, at least a distance equal to its own length away from the cephalic end of the body PJiolus Hiibner. 140 The following species were examined : Herse cingulata Fabricius Cocytius antaeus Drury Phlegethontius quinquemacidata Haworth, scxta Johannsen Atreiis plebeia Fabricius Dolba hylaeiis Drury '' C hlaeno gramma jasmincarmn Boisduval Sphinx kalmiae Smith and Abbot, driipiferariim Smith and Abbot, gordius Stoll, hicitiosa Clemens, chersis Hiibner, cremitiis Hiib- ner Cressonia jiiglandis Smith and Abbot Lapara bombycoides Walker, conifcraruin Smith and Abbot Daremma catalpae Boisduval Ceratomia amyntor Geyer, undidosa Walker Smerinthus jamaicensis Drury, cerysii Kirby Marumba modesta Harris Hcmaris diffinis Boisduval, gracilis Grote and Robinson, thysbe Fabricius Deilephila lineata Fabricius Diloplionota alio Linnaeus, alope Drury Theretra tersa Linnaeus Dcidamia inscription Harris Darapsa pholiis Cramer Amphion nessus Cramer Lepisesia gaurae Smith and Abbot, juanita Strecker Ampelophaga myron Cramer, versicolor Harris Sphecodina abbotti Swainson Pseiidosphinx tetrio Linnaeus Pholus pandorus Hiibner, achemon Drury SuPERFAMiLY SATURNIOIDEA The pupae of this superfamily retain none of the generalized char- acters found in the families previously discussed and all the sutures of the head are obliterated, even those adjacent to the proximal ends of each antenna. The body is usually heavily chitinized, and although there are always a few setae present they are rarely visible to the un- aided eye. The superfamily is distinguished by the presence of broadly pectinate antennae, in the Ceratocampidae for about one third of the length, while in the Hemileucidae and Saturniidae they are broadly pectinate to the distal end, and generally have the stem of the flagel- lum elevated. The greatest width of each antenna is at least one fifth 141 of its length, often much wider, and the antennae seldom extend farther caudad than the prothoracic legs. There is a marked differ- ence in the sexes, the antennae of the male being much broader, some- what longer, and often meet on the meson, covering nearly all of the appendages except the wings. The legs are shorter than in most super- families, the prothoracic and mesothoracic legs usually either meeting or lying adjacent on the meson. The maxillae never reach the caudal margin of the wings, and their greatest length is not more than one third the length of the wings, but they are usually much shorter. The mesothoracic wings always lie adjacent on the meson and the meta- thoracic wings are often visible on the meson in the Saturniidae. The family Ceratocampidae has a row of broad triangular spines set on the edge of a flanged plate along both cephalic and caudal margins of the movable abdominal segments. They usually possess very long cremasters, which are always bifurcate at the distal end. The Hemi- leucidae have short cremasters, while there are none present in the Saturniidae and few of the genera have spines at the caudal end of the body. A paper on this superfamily, giving tables for the identification of families, genera, and species has been prepared and the first part, ^'The Classification of the Pupae of the Ceratocampidae and Hemi- leucidae," was published in the Annals of the Entomological Society of America, Vol. 7, 1914, pp. 277-300. The manuscript for the re- mainder is now in the hands of the editor of the same journal. The following tables of families and genera and the descriptions, with some slight additions and corrections, are taken from the paper mentioned above. The generic names in Dyar's list are used, but the genera are not arranged to indicate natural relationships. The superfamily Saturnioidea may be divided into three families as follows : a. Pupae with the movable segments provided with flange-like plates which prevent their being telescoped, their lateral margins distinctly tapering caudad and each segment noticeably smaller than the seg- ment cephalad of it; wings never elevated dorsad above the surface of the body ; a distinct cremaster always present ; stem of the flagel- lum of the antenna never elevated and distinct. b. Pupae with a distinctly bifurcate cremaster ; body usually rough- ened with spines on the exposed surface of the thorax and abdo- men; metathorax with prominent oblong tubercles on each side of the meson extending one third or more of the distance between the meson and the margin of the first pair of wings ; pupae always found in the ground Ceratocampidae. hb. Pupae without a distinctly bifurcate cremaster; body never roughened with spines on the exposed surface of the thorax and 142 abdomen; metathorax never with prominent oblong tubercles; pupae found either in cocoons or in the ground . . Hemileucidae. aa. Pupae with the movable segments never provided with flange-like plates which prevent their being telescoped, the lateral margins ap- proximately parallel so that the segments appear of equal size and are usually telescoped so that only the caudal margins of the seg- ments are visible ; wings prominently elevated dorsad above the level of the body, the caudal portion of the mesonotum and meta- notum always depressed adjacent to the wings ; a distinct cremaster rarely present; stem of the flagellum of the antenna always ele- vated and distinct Saturniidae. Family Hemileucidae Margins of the free segments never with a row of spines; the body surface never roughened with spines ; antennae with the stem of the flagellum never distinct, the central axis never set with spines, the antennae tapering gradually from the part having the greatest width ; maxillae never more than one sixth the length of the wings ; proleg scars seldom prominent on abdominal segments five and six and rarely with the anal proleg scars visible; mesothoracic wings with the anal angles broadly rounded, usually at the cephalic margin of fourth ab- dominal segment, and usually reaching the caudal margin of the fourth abdominal segment ventrally ; metathoracic wings never produced be- low the anal angles of the first pair of wings and never visible in ventral view ; metathorax never with prominent tubercles ; abdominal segments 5 to 7 with the cephalic margin produced into a thick obliciue flange-like plate directed caudad ; cremaster short, never bifurcate at tip. Although not usually included with the Hemileucidae the genus Automeris is placed in this group owing to the very evident relation of the pupae to those of the genera Hemileuca and Pseudohazis. Morphologically thev seem to be more nearly related to the Hem- ileucidae, but they are found in cocoons like the Saturniidae. Some of the members of this family pupate in the ground. The description of this family is of necessity very incomplete owing to lack of material. According to our available knowledge of the subject the three genera may be separated as follows: a. Cremaster bearing setae arranged in a tranverse row and spreading out fan-like Pseiidoliazis Grote and Robinson. aa. Cremaster never with setae, either with curved spines or without spines or setae of any kind, b. Cephalic part of segment above the flange-like plate either smooth or with fine longitudinal striations; pupae found in ground. Hemileuca Walker. 143 bb. Cephalic part of segment above the flange-like plate with sharp transverse ridges, deep furrows between; pupae found in cocoons Automeris Hiibner. The following species were examined : Pseudohads eglanterina Boisduval Hemileuca maia Drury, burnsi Watson, olivae Cockerell Automeris pamina Neumoegen, io Fabricius, leiicana Hiibner, incar- nata Walker Family Ceratocampidae Body with the margins of the free abdominal segments usually bearing a row of spines, and the exposed surface of the thorax and abdomen usually roughened with spines ; antennae never broadly pec- tinate throughout, but broadly pectinate and almost parallel for about one half the length, then narrowed rapidly to about half the greatest width, tapering gradually to a pointed tip, the stem of the flagellum never distinct, the surface convex and the central axis of the antenna usually bearing one or two rows of small spines; maxillae never less than one fourth the length of the wings ; tips of the mesothoracic tarsi meeting obliquely on the meson, never lying adjacent on the meson; proleg scars very prominent on abdominal segments five and six, the scars for the anal prolegs often very conspicuous ; mesothoracic wings with the anal angles broadly rounded, usually located at the cephalic margin of the fourth abdominal segment and never reaching ventrad to the caudal margin of the fourth segment ; metathoracic wings never produced below anal angle of the mesothoracic wing and never visible in ventral view; metathorax with distinct tubercles, more or less ob- long in outline, on each side the meson and extending more than one third the distance from the meson to the margin of the wing; the suture between the seventh and eighth segments never deep, or with distinct crenulations on its margins ; cremaster always present, usually long and bifurcate at tip. Five genera of this family have been described. One genus, Syssphinx, consisting of three species, was not available for study. The pupae of this family are always found in the ground. The remaining genera of Ceratocampidae can be separated by the following table : a. Surface of pupa never spinose ; cremaster broader than long, broadly and shallowly bifurcate, never over 2 mm. in length. Citlieronia Hiibner. aa. Surface of pupa spinose ; cremaster at least twice as long as broad, bifurcate at tip, always more than 2 mm. in length. 144 b. Thorax rugose with short isolated spines, abdominal segments not spinose, but bearing a row of spines along both cephalic and caudal margins of segments 1-7, the spines along the caudal margin of segments 5-7 much longer than the spines of the cephalic row Basilona Boisduval. bb. Thorax and abdominal segments densely spinose ; abdominal seg- ments 1-7 with a row of spines along both cephalic and caudal margins, the spines in the cephalic row on abdominal segments 5-7 usually much longer than the spines in the caudal row ; maxillae, measured on meson, one fourth the length of the wings, c. Usually with prominent scattered spines on the thoracic segments, at least four times as long as those covering the segments; an- tennae with the central axis bearing a row of prominent spines curved caudad ; if without prominent spines on the thoracic seg- ments and antennae, then the maxillae are one third the length of the wings. d. Eighth abdominal segment never with a prominent transverse ridge in the middle of the segment bearing a row of spines; glazed eye-piece alwaj^s lighter in color than the remaining surface of the body; species always more than an inch in length Adelocepliala Herrich-Schaeffer. dd. Eighth abdominal segment always with a prominent trans- verse ridge in the middle of the segment bearing a row of spines ; glazed eye-piece alwaj^s the same color as the re- maining surface of the body; species never more than an inch in length Dryocampa Harris. cc. Without prominent scattered spines on the thoracic segments, the longest never four times the length of those covering the segments ; antennae with the central axis never bearing prom- inent spines, the spines never curving caudad ; maxillae always one fourth the length of the wings Anisota Hiibner. The following species were examined : atheroma rcgolis Fabricius Basilona impcrialis Drury Adelocepliala hicolor Harris, bisecta Lintner Dryocampa rubicunda Fabricius Anisota virginiensis Drury, stigma Fabricius, senatoria Smith and Abbot, skinneri Biederman, consularis Dyar Family Saturniidae The members of this family have the antennae broadly pectinate throughout, or nearly so, and the stem of the flagellum is usually dis- tinct and raised above the level of the pectinations. The maxillae are always short, never more than one third the length of the wings, and 145 usually very much shorter. The tibiae and tarsi of the prothoracic legs, and the tarsi of the mesothoracic legs lie adjacent on the meson, but never meet obliquely on the meson as they do in the ceratocampids. The mesothoracic wings always have their anal angles broadly rounded and the wings always reach the caudal margin of the fourth abdominal segment on the ventral surface. The metathoracic wings are pro- duced around the anal angles of the mesothoracic wings and usually form prominent angles on the fourth abdominal segment. The meta- thoracic wings always extend for at least a short distance along the caudal margin of the mesothoracic wings on the ventral surface of the body. The metathorax never has distinct oblong tubercles which are one third or more the width of the segment, such as are found in cera- tocampids. The suture between the seventh and eighth abdominal segments is never deep, with distinct crenulations on its margins, and is indistinct in many species. The cremaster, if present, is very short and is never bifurcate at the distal end. The known pupae of members of this family are found in silken cocoons. Some of these are very thick and tough, others thin and pa- pery. Only nine genera of this family have been available for study. These may be separated by the following table : a. Lateral margins of abdominal segments 5-7 never approximately par- allel, but tapering from the cephalic margin of the fifth segment, the lateral margins usually distinctly convex ; caudal end of body usually with stout curved spines. b. Tenth abdominal segment never flattened into a transverse plate with the caudo-lateral angles produced into short lobes. e. Caudal end of body without spines; body surface with slightly wavy, transverse ridges with distinct furrows between ; meso- thorax never with a prominent tubercle at the base of each wing. Copaxa Walker, cc. Caudal end of body with stout curved spines ; body surface never with slightly wavy, transverse ridges with distinct furrows be- tween; mesothorax with a prominent tubercle at the base of each wing, d. Lateral aspects of the cephalic margin of abdominal segments never with spiracular furrows; caudal end of body with an oval area set with slightly curved spines arranged in two groups and nearly all pointing outwards Telea Hiibner. dd. Lateral aspects of the cephalic margin of abdominal seg- ments 5-7 with spiracular furrows separated by slightly wavy carinate ridges; caudal end of body deeply rugose, with a slight concavity containing a circular group of strongly recurved spines Tropaea Hiibner, 146 bb. Tenth abdominal segment, viewed dorsally, in the form of a transverse plate, concave on the caudal margin, the caudo- lateral angles produced into lobes; the other segments strongly concave in ventral view, with five short curved spines inserted close together in the caudo-lateral margin of each lobe. Agapema Neumoegen and Dyar. aa. Lateral margins of abdominal segments 5-7 always approximately parallel, the caudal end of the body never with stout curved spines, b. Maxillae always one fourth, or less, the length of the wings, the proximal two thirds of their margins never strongly concave; mesothoracic wings with their anal angles on the cephalic mar- gin of the fourth abdominal segment or caudad of that portion of the segment, c. Maxillae less than one fifth the length of the wings ; antennae of males with the sides tapering gradually to a pointed tip. d. Both eye-pieces never visible in either sex ; glazed eye-piece seldom visible in the females, never in the males ; caudal end of abdomen never with a band of coarse setae. e. Eye-pieces never visible in either sex; caudal end of abdo- men never with spines or setae Callosamia Packard. ee. Glazed eye-piece visible in the females ; caudal end of abdo- men with a few very short sharp spines. E upackardia Cockerell. dd. Both eye-pieces visible in either sex ; caudal end of abdomen with a transverse band of coarse setae. .EotliscMldia Grote. cc. Maxillae never less than one fifth the length of the wings; an- tennae of males with the sides approximately parallel for the greater part of their length, tapering rapidly to a blunt rounded tip ; a small portion of glazed eye-piece always visible in the females Saviia Hlibner. bb. Maxillae always more than one fourth the length of the wings; the proximal two thirds of the lateral margins of the maxillae concave ; first pair of wings with their anal angles on the third abdominal segment opposite the second pair of abdominal spiracles Pliilosamia Grote. The following species were examined : Copa.va lavcndcra Westwood Tclca polyphciniis Cramer 7'ropaea lima Linnaeus Agapema galhina Clemens Callosamia promethea Drtiry, angulifcra Walker Eupackard'ui callcta Westwood Rothscluldia ori^oba Westwood, joruUa Westwood Samia cccropia Linnaeus, glovcri Strecker, Columbia Smith, rubra Behr Philosamia cynthia Drury 147 PHYI.OGENY The characters used as a basis for determining the phylogeny of the order are primarily: (i) the nmnber of movable segments; (2) the freedom of the appendages; (3) the number of sutures present in the head; (4) the relative length of the body segments; (5) the presence or absence of visible labial palpi and maxillary palpi ; (6) the presence of exposed portions of the prothoracic femora in special- ized pupae; and (7) the method of dehiscence. In the most generalized forms there is complete freedom of mo- tion possible between the head and thorax, and between all the seg- m.ents of the thorax and abdomen with the exception of the eighth, ninth, and tenth abdominal segments, which are always fixed. As specialization proceeds, there is a gradual loss of motion ; first between the head and thorax, then between the segments of the thorax, and last of all between the different segments of the abdomen. The loss of motion in the abdomen begins first at the cephalic end, but by the time that complete motion of the second segment has been lost there begins a loss of motion of the seventh segment. This takes place first in the female, and there is a large series of forms, including the super- families Gracilarioidea, Tortricoidea, and Aegerioidea, which retain freedom of motion in the seventh segment of the male, while there is taking place at the cephalic end of the abdomen the loss of motion of the third abdominal segment. There are, however, a few gen- era of Gracilarioidea which have lost freedom of motion of all the body segments and which form the most specialized end of that series. The pupae which have lost motion of all the abdominal segments except the fourth, fifth, and sixth, are those usually referred to as obtected pupae. There are few pupae more specialized than those of the superfamily Gracilarioidea which retain freedom of motion of the seventh abdominal segment in the male, but there are a few generalized forms both in the Pyralidoidea and Papilionoidea in vvhich this is the case, as it is also in the family Epermeniidae of the Yponomeutoidea. These three superfamilies are usually considered as more specialized than the Gracilarioidea. As the number of mov- able segments determines the position of a superfamily in the series it is readily seen that these superfamilies must be considered as more generalized than those in which motion is lost in the seventh segment ill the male. It will be remembered that a segment is spoken of as movable when motion is possible between its caudal margin and the segment caudad of it. As the appendages become soldered to the body wall on the ventral surface no motion of this part of the segment is possible if the incision between its caudal margin and the next seg- 148 nient is covered by the wings, therefore it can not be considered as a free segment. In many cases, however, dorsal movement of such seg- ments is possible, which gives the segment freedom of movement in certain directions; as, for instance, in curving the caudal end of the body cephalad on the ventral surface, well illustrated in the movements of most tortricids. Such forms must be considered as more general- ized than those which have lost entire motion of the segment, and thus the Pyralidoidea and Papilionoidea must occupy a lower position than the Yponomeutoidea, whose members have lost dorsal motion of the third abdominal segment, while the other two superfamilies mentioned retain it. There are certain specialized forms in other superfamilies in which motion is lost in all the body segments, notably in the family Elachistidae of the Gelechioidea and in certain genera of the family Nymphalidae in the Papilionoidea. There are also many genera in various families which retain movement in only one segment. The appendages of the generalized pupae are entirely free from each other and from the body wall and are often considerably spread out from the surface of the body so that the pupae strongly resemble those of the Trichoptera. In these forms there is but a slight degree of chitinization in any part of the body. The appendages are grad- ually soldered down, however, first to each other, while all remain free from the body w^all, and then there takes place a gradual sol- dering down of the appendages to the body wall, beginning first at the cephalic end of the abdomen. In many pupae the appendages are soldered to two, three, or four abdominal segments while the portion of the appendages caudad of these segments remains free and allows freedom of motion of the abdominal segments underneath. Such a condition exists in many genera of the Aegerioidea and Gelechioidea. The pupae with free appendages could only exist successfully in pro- tected situations from which an easy egress was possible, and so they are only found in cocoons, or in mines in leaves and stems of plants. Pupae with any other environment lost the freedom of the appendages much more rapidly, as in the case of the Lyonetiidae and some of the Papilionoidea. The number and arrangement of the sutures present on the head has already been discussed under the head of external morphology, pages 23 to 25. These sutures are gradually obliterated, beginning with the clypeo-labral, which is lost among very generalized pupae. The epicranial suture is one of the last to disappear, and its presence indicates the degree of specialization in many of the higher forms, as it is retained in some members of many superfamilies which are high in the series. 149 The fronto-clypeal suture is visible for a part of its length in most pupae, and is especially distinct for its entire length in some of the Gelechioidea ; but dehiscence often showed the presence of this suture when it was impossible to locate it on the pupa. The part of this su- ture adjacent to the proximal end of each antenna is the last head su- ture to be obliterated, and it is lacking only in the Saturnioidea. The segments of the body are more nearly of equal length in gen- eralized than in specialized forms, especially in the abdomen. The prothorax is short in the Micropterygoidea and becomes gradually longer in the specialized superfamilies. The metathorax is long in generalized forms and nearly equals the mesothorax in length. As specialization proceeds, the mesothorax becomes longer and the meta- thorax much shorter, so that the comparative length of these two seg- ments furnishes another means of determining the position of a super- family in the series. The abdominal segments also become consoli- dated, first at the caudal end of the body, where they gradually be- come shorter than the cephalic segments. After motion is lost in the cephalic segments, they, too, gradually shorten, until the movable seg- ments are much longer than any of the others. The presence of visible maxillary and labial palpi also furnishes an easy means for the identification of generalized forms. The labial palpi are retained throughout the series, but are gradually overlaid and concealed by the maxillae. The presence or absence of visible labial palpi, however, indicates the degree to which specialization has pro- ceeded along a given line. Labial palpi are visible to some extent in some members of all superfamilies except the Saturnioidea. The maxillary palpi are usually the first to disappear, but these palpi are often present in the pupa, when lacking in the imago. The maxillary palpi in generalized forms reach the proximo-lateral angles of the maxillae, but gradually decrease in length until they are visible only as a small triangular area caudad of the sculptured eye-piece. When the appendages are free their position is considerably laterad of that which they gradually assume as they become soldered to each other. The legs are folded in such a way that in generalized forms almost the entire femur of the prothoracic leg is exposed. Later the tibia and tarsus of this leg are folded so that their position is nearer the meson than formerly and the femur is entirely concealed. The presence of an exposed portion of the prothoracic femur is a general- ized condition which is retained by forms exceedingly specialized in other respects, and is found in some genera of Sphingidae. As to the method of dehiscence, there are several things to be noted, although all too little is known of this interesting phase of pupal 150 life. There is a tendency for the generaHzed forms to emerge from the mine, cocoon, burrow, or other place of protection, as a pupa, and consec|uently the body is provided with some structure which assists in its progress. The appendages and body segments are usually sep- arated from each other at dehiscence and the body splits along the median line of the vertex and thoracic segments, the vertex carrying the sculptured eye-pieces with it. The front, with the antennae, is com- pletely separated from the rest of the head parts in some forms, by a splitting along the epicranial suture on the dorsum, and along the fronto-clypeal suture on the ventral surface. When the f ronto-clypeal suture is not entire it usually splits for a part of its length, thus al- lowing it to be considerably elevated. In specialized forms it is usually the imago which emerges, the pupal skin being left behind in the cocoon or other place of protection. The appendages and body seg- ments remain firmly soldered together and the imago escapes through the opening made by the splitting of the vertex, when present, the pro- thorax, and the mesothorax; or, if this is not sufficient, an irregular opening which does not follow the line of any suture is made in the cephalic end of the body. In these forms the eye-pieces remain at- tached to the other face-parts. The phylogeny of any group is usually determined by the devel- opment of a single character. Many workers have used the venation of the wings to arrange a series of genera or species in phylogenetic order. Others have used the genitalia, or the arrangement of setae. The pupae present many and varied characters which may be used to arrange such a series. In this investigation a series was arranged for each of the characters previously mentioned and the results of these series were combined. These characters have the advantage over those used by previous authors in that they comprise practically all of the important structures of the body and are all present in the same individual. It is quite probable that other characters might be used to indicate the development of the order, such as the number and ar- rangement of the genital apertures, the form of the spiracles, and the arrangement of setae, none of which have been investigated sufficiently to admit of their use in this paper. ACKNOWIvEDGMENTS The subject for this investigation was suggested by Professor A. D. MacGillivray in September, 19 12. Since that time many species of Lepidoptera have been collected, both in the larval and pupal stage, and reared to maturity. The Graduate Scliool of the University of 151 . Illinois has made liberal appropriations for the purchase of material, and a large series of named pupae has been obtained from the follow- ing sources : the American Entomological Company, Ward's Natural Science Establishment, the Kny-Scheerer Company, the New England Entomological Exchange, the New Jersey Entomological Company, and from Mr. Wm. Beutenmiiller, A. H. Manee, of Southern Pines, N. C, E. J. Oslar, of Denver, Colo., and Miss Annette F. Braun, of Cincinnati, Ohio. Dr. L. O. Howard, Chief of the Bureau of Ento- mology of the U. S. Department of Agriculture, Mr. August Busck, of the U. S. National Museum, and Dr. Wm. Barnes, of Decatur, 111., have furnished a number of rare species which contributed in no small measure to the progress of the investigation. The collections of the Illinois State Laboratory of Natural History have been freely available, and for this courtesy Professor S. A. Forbes deserves our hearty thanks as well as his assistants, Mr. J. R. Malloch and Mr. C. A. Hart. Dr. T. H. McDunnough, of Decatur, 111., and Miss A. F. Braun have assisted in the identification of bred material, thus mak- ing available many species that otherwise could not have been used. The collections furnished by Miss Braun and Mr. Beutenmiiller deserve especial mention. Miss Braun collected, bred, and determined more than one hundred species of the so-called Microlopidoptera which really formed the working basis for this part of the investigation. Most of the work done on the Sphingidae has been from specimens furnished by Mr. Beutenmiiller, who has collected a large series of forms. Mr. Samuel Henshaw, of the Museum of Comparative Zool- ogy at Harvard, very kindly loaned a large series of Papilionoidea, most of which were collected by Dr. S. H. Scudder and described in his "Butterflies of the Northern United States and Canada." Profes- sor J. G. Needham, of the Department of Entomology at Cornell University, also loaned a large number of species from the collections of that University, which were especially valuable and could not be obtained elsewhere. It would be impossible in this brief space to enumerate the many ways in which Professor MacGillivray has assisted in the preparation (>f this paper. His interest in the work has been unfailing, and what- ever of value it may contain is due to his inspiration, encouragement, and helpful criticism. 152 BiBUOGRAPHY Chapman, T. A. '93. Some neglected points in the pupae of the heterocerous Lepidoptera. Trans. Ent. Soc. London, 1893:97-119. '93. On a lepidopterotis pupa (Micropteryx purpurella) with functionally active mandibles. Trans. Ent. Soc. London, 1893:255-265. '94. Some notes on the Microlepidoptera whose larvae are ex- ternal feeders. Trans. Ent. Soc. London, 1894:335-350. '96. Notes on pupae — Orneodes, Epermenia, Chrvsocorys and Pterophorus. Trans. Ent. Soc. London, 1896: 129-147. Jackson, W. Hatchett. '91. Morphology of the Lepidoptera. Trans. Linn. Soc. London, Zool., Ser. 2, Vol. 5. Packard, A. S. '95. Attempt at a new classification of the Lepidoptera. Mono- graph of the Bombycine moths of America north of Mexico, Part L ]\Iemoirs of the National Academv of Sciences, 7:56-83. Poulton, E. B. '91. The external morphology of the lepidopterous pupa. Trans. Linn. Soc. London, Zool., Ser. 2, 5 : 245-263. Scudder, S. H. '89. The butterflies of the Eastern L^nited States and Canada. 3 vols. Tutt, J. W. '00. A natural history of the British Lepidoptera, 2: 38-100. 153 Plates The following plates show in outline the principal structures of pupae of many of the families discussed in this paper. No attempt has been made to show all of the setae, spines, or tubercles which may occur, but only those which are most important and are of taxonomic value. The following" abbreviations have been used : a, antennae g, genae al-alO , abdominal segments 1-10 ge, glazed eye-piece af, alar furrow go. genital opening ao, anal opening lb, labrum ar, anal rise 1., prothoracic leg at, invaginations for the anterior 1.,, mesothoracic leg arms of the tentorium I3, metathoracic leg cl. clypeus IP, labial palpi els, clypeo-labral suture md, mandibles cdm, caudal margin of an abdominal mp. maxillary palpus segment ms, mesothorax cm, cephalic margin of an abdominal msp, mesothoracic spiracle segment mt, metathorax cr, cremaster mx, maxilla cs, cremastral setae P. prothorax cxl. coxa of the prothoracic leg Pf, pilifer cx2, coxa of the mesothoracic leg psc, proleg scar cx3, coxa of the metathoracic leg s, spiracle dlt, dorso-lateral row of tubercles se. sculptured eye-piece dmt. dorso-mesal row of tubercles sf. spiracular furrow dst, dorso-spiracular row of tubercles t, tegulae es, epicranial suture ts. tubercle scar f, front V, vertex fcs, fronto-clypeal suture vst, ventro-spiraeular row of tubercles fl femur of the prothoracic leg wl, mesothoracic wing f2, femur of the mesothoracic leg w2, metathoracic wing fp, flanged plate INDEX TO GENERA AND HIGHER GROUPS Achatodes, 114. Aerobasis, 75. Acrocercops, 67. Acrolophidae, 4.5, 46. Aeronycta, 114, 115, 119. Acronyetinae, 108, 113. Adelocephala, 144. Aegeriidae, 24, 48, 49, 52. Aegerioidea, 27, 31, 48-51, 147, 148. Agapema, 146. Agaristinae, 109, 112. Aglais, 91. Agonopteryx, 104, 105. Agraulis, 92. Agrotinae, 107, 109, 110. Agrotis, 10.9. Alsophila, 127, 130. Alypia, 113. Amblyscirtes, 79, 81. Ampelophaga, 139. Amphion, 139. Anaeampsis, 102. Anaea, 89, 93, 94. Anaeinae, 89, 94. Ancylis, 52, Ania, 129. Anisota, 144. Anomis, 107, 115. Anosia, 94. Antispila, 63. Apantesis, 120. Apatelodes, 132, 133. Apatiirinae, 89, 93. Apaturini, 94. Archips, 29, 55, 57, 58. Arctiidae, 107, 116, 119. Argynnini, 90, 91. Argynnis, 92, Argyresthia, 98. Argyrotoxa, 56, Aristotelia, 101, 102, 103, Atreus, 136, Atteva, 71, 72. Attevidae, 69, 70, 71. Antomeris, 143. B Balsa, 116, 117, Basilarchia, 93, Basilarchinae, 89, 93. Basilona, 144, Bedellia, 65, Bembecia, 49, 50. Bombyeidae, 123, 124. Bombyeoidea, 33, 96, 123-125. Bombyx, 124, Brenthia, 47, 48, 63. Brenthis, 92, Brephos, 130, Bucculatrigidae, 61, 64, Bucculatrix, 64. Butalis, 100. C Caenurgia, 118, Callosamia, 146. Cissia, 95, Citheronia, 143. Cleora, 130, Cocceius, 82. Cocytius, 136. Coleophora, 98. Coleophoridae, 96, 98, Copaxa, 145, 156 Coptodisea, 63. Coptotriehe, 63. Cosmopterygidae, 99, 106. Cosmopteryx, 106. Cosmotriche, 124. Cossidae, 39, 40. Cossinae, 41. Cossoidea, 26, 31, 37, 38-41. Crambinae, 72, 73, 74. Cremastobombycia, 68. Cressonia, 135, 137. Ctenucha, 119, 120. Cucullianae," 107, 108, 110. Cyaniris, 84. Oymatophora, 130. Calpodes, 79, 81. Cameraria, 64, 67, 68, 69. Canarsia, 76. Carpocapsa, 52, 53. Catocala, 117. Catocalinae, 109, 110, 117. Cenopis, 58. Ceratocampidae, 140, 141, 143. Ceratomia, 137. Cercyonis, 95. Charadra, 116, 119. Charidryas, 92. Chlaenogramma, 137. Chloridea, 112. Chlorippe, 93. Choreutis, 47, 48, 63. Chrysopeleia, 104. Chrysopeleiidae, 99, 104. Chrysophanus, 83, 84. Cinelidia, 93. Cingilia, 129. Cinglis, 128. CirpMs, 110, 111. D Darapsa, 139. Daremma, 137. Dasychira, 122. Datana, 133. Deidamia, 139. Deilephila, 138. Depressaria, 104, 105. Desmia, 77. Diacrisia, 120. Dilophonota, 138. Dioptidae, 125, 126, 134. Dolba, 136. "Dryocampa, 144. E Ecpantheria, 120. Ectropis, 130. Elachista, 65, 104, 106. Elachistidae, 62, 96, 98, 99, 100, 106. Enannonia, 52, 54. Ennoniinae, 127. Ennomos, 129. Eois,128. Epagoge, 57, 58. Epargyreus, 82. Epermenia, 97. Epermeniidae, 59, 95, 96, 97, 147. Ephestia, 74, 75. Epiblemidae, 47, 52. Epinotia, 52, 53, 54. Epipasehiinae, 72, 73, 77. Episimus, 55. Erannis, 131. Eriocraniidae, 23, 24, 25, 26, 35, 37, 48, 59, 60, 62. Eriopus, 110, 112. Estigmene, 120. Euchaetias, 119. Euclea, 43. Eucleidae, 24, 42, 43, 44. Eucleoidea, 25, 26, 29, 31, 37, 41-44, 62. Eiicosma, 52, 53. Eudamus, 82. Enlonche, 114, 119. Eunetis, 117. Eupackardia, 146. Euparthenos, 118. Euphoeades, 86. Euphydryas, 92. Euploeinae, 89, 94. Euproctis, 122. Euptoieta, 92. Eurema, 88. Eurymus, 88. 157 Euthisanotia, 113. Euvanessa, 91, Evippe, 102. Exartema, 55, 56. F Feniseca, 78, 83, 84. G Galleria, 74. Galleridae, 26. Grallerinae, 72, 73. Gelechia, 103. Gelechiidae, 99, 100, 101, 105, 135. Gelechioidea, 23, 33, 96, 98-106, 148, 149. Geometridae, 29, 125, 126, 134. Gnorimoschema, 101. Gracilaria, 26, 60, 66. Gracilariidae, 61, 62, 65, 69. Graeilariinae, 66. Gracilarioidea, 23, 26, 31, 58-69, 147. Graptolitha, 110. H Hadena, 110, 112. Hadeninae, 108, 110. Haematopsis, 126, 129. Halisidota, 119. Hapalia, 110. Haploa, 120. Harmologa, 57. Harpyia, 133. Harrisina, 44. Heliodinidae, 45, 47, 63. Heliozelidae, 24, 59, 60, 62; Hemaris, 138. Hemerocampa, 121, 122. Hemileuca, 142. Hemileucidae, 140, 141, 142. Hemimene, 52, 54. Heodes, 83, 84. Hepialidae, 23, 24, 37, 38, 40. Hepialoidea, 30, 37-38, 39, 41. Herse, 136. Hesperiidae, 27, 78, 79, 80. Hesperioidea, 78. Heterocampa, 134. Homopyralis, 107, 114. Hydria, 128. Hydriomeninae, 127. Hyparpax, 133. Hypeninae, 107, 109, 116. Hyphantria, 120. Hypocolpus, 46. Hypsopygia, 74. Incisalia, 84. Iphiclides, 86, Isia, 120. J Junonia, 91. L Laertias, 86. Lagoa, 43. Lanthape, 78. Lapara, 137. Laphygma, 111. Lasiocampa, 124 Lasiocampidae, 123, 124. Laverna, 100. Lavernidae, 99, 106. Lepisesia, 139. Leueaiithiza, 67. Liparidae, 29, 107, 119, 121, 123. Lithocolletinae, 66, 67. Lithocolletis, 59, 67, 68. Lophoptilus, 100, Lyeaenidae, 27, 78, 79, 83. Lycia, 131. Lycophotia, 110, 112. Lyonetiidae, 59, 60, 61, 64. M Malacosoma, 124. Marmara, 67, 68. Marumba, 135, 138. Megalopygidae, 42, 43, 44. Megathymidae, 78, 79. Megathymus, 80. Melalopha, 132. Meliana, 111, 158 Melitaeinae, 90, 92. Mellisopus, 53. Memythrus, 49, 50. Menesta, 105. Meroptera, 75. Micropterygoidea, 24, 30, 35-37, 149. Micropteryx, 20. Mineola, 75. Mitura, 84. Mnemonica, 35, 37. Mominae, 108, 116. Momphidae, 106. Monima, 110, 111. N Nacophora, 129. Nepticulidae, 26, 29, 41, 48, 59, 60, 61, 65. Noetiia, 110. Noetuidae, 107, 134. Noctuoidea, 26, 33, 96, 107-123. Notodontidae, 125, 132, 134. Notodontoidea, 34, 76, 96, 125-134. Nycteolidae, 119. Nymphalidae, 78, 79, 87, 88, 148. Nymphalinae, 89, 90, 93. O Oecophoridae, 99, 104. Oeneinae, 90, 95. Oeneis, 78, 95. Oiketicus, 40. Olene, 122. Olethreutes, 55, 56. Olethreutidae, 52, 54. Ornix, 59, 66. Oxyptilus, 70, 71. P Paleaerita, 127. Pantagrapha, 77. Paonias, 135, 138. Papilio, 20, 86. Papilionidae, 78, 79, 85, 87. Papilionoidea, 23, 24, 27, 32, 65, 69, 76, 78-95, 96, 147, 148. Parectopa, 66. Parharmonia, 50. Peronea, 56, 57. Phaecasiophora, 57, 58. Pheoeyma, 118. Philobia, 129. Philosamia, 146. Phlegethontius, 136. Phlyctaenia, 77. Pholisora, 82. Pholus, 139. Phryganidia, 134. Phthorimaea, 103. Phyciodes, 92. Phycitinae, 71, 73, 74, 75. Phyllocnistidae, 60, 61, 65, 68. Phyllocnistis, 68. Physostegania, 130. Phytometra, 116. Phytometrinae, 108, 115. Pieridae, 78, 79, 87. Pinipestis, 75, 76. Plathypena, 116, 117. Platynota, 57, 58. Platyptilia, 70, 71. Plodia, 74, 75. Plusiodonta, 107, 114, 115. Plutella, 97. Podosesia, 29, 50. Polia, 110, 112. Polychrosis, 55, 56. Polygonia, 91. Pontia, 88. Porthetria, 121, 122. Prionoxystus, 40, 41. Prodenia, 111. Prodoxidae, 44, 45. Prodoxus, 46. Proleucoptera, 65. Prolimacodes, 43. Protoparce, 20. Pseudanaphora, 46. Pseudohazis, 142. Pseudosphinx, 139. Psilocorsis, 104, 105. Psorosina, 76. Psychidae, 39, 40. Psychomorpha, 118. 159 Pterophoridae, 28, 69, 70. Pterophorns, 70, 71. Pyralididae, 69, 70, 72, 135. Pyralidoidea, 24, 32, 69-78, 96, 147, 148. Pyralinae, 73, 74. Pyralis, 74. Pyrausta, 76, 77. Pyraustidae, 23. Pyraustinae, 73, 76. Pyromorphidae, 42, 44. Pyrrhia, 112. E Eeeurvaria, 101, 102, 103. Eliodophora, 111. Eothschildia, 146. Kusticus, 84. S Sabulodes, 129. Samia, 146. Sanninoidea, 50. Sarrothripinae, 109, 118. Sarrothripus, 119. Saturniidae, 25, 28, 125, 135, 140, 141, 142, 144. Satiirnioidea, 22, 34, 96, 123, 140-146, 149. Satyrinae, 78, 90, 94, 95. Satyrodes, 95. Sehizura, 133. Scythridae, 96, 99, 100. Scythris, 101. Sibine, 43. Sitotroga, 103. Smerinthus, 135, 138. Sparganothidae, 52, 57. Sphecodina, 139. Sphingidae, 23, 24, 25, 29, 135. Sphingoidea, 34, 96, 135-140. Sphinx, 137. Stenoma, 105. Stenomidae, 99, 104, 105. Sterrhinae, 127. Sthenopis, 38. Symmerista, 133. Synanthedon, 50. Synchloe, 88. Syngrapha, 116. Syntomidae, 119. Syssphinx, 143. T Telea, 27, 145. Telphusa, 102. Tephroclystis, 128. TTianaos, 82. Theela, 84. Theretra, 138. Thiodia, 54. Tholeria, 77. Thorybes, 82. Thyridopteryx, 40. Tinea, 47. Tineidae, 45, 47. Tineoidea, 28, 31, 44-48, 59, 60. Tischeria, 63. Tischeriidae, 61, 63, 68. Tmetocera, 54. Tolype, 123, 124. Tortricidae, 52, 56. Tortricoidea, 28, 29, 31, 44, 45, 48, 49, 51-58, 59, 147. Trichotaphe, 102. Tropaea, 27, 145. Trypanisma, 103. U Uranotes, 84. Utetheisa, 120. V Vanessa, 91. Vanessini, 90. X Xanthotype, 129. Xylorictidae, 105. Y Yponomeuta, 97. Yponomeutidae, 69, 71, 96, 97, 98, 100, 106. Yponomeutoidea, 26, 32, 96-98, 147, 148. Ypsolophus, 102. Z Zale, 118. Zelleria, 97. Zeuzera, 41. Zeuzerinae, 41. Plate XIX Fig. 1. Mnemonica auricyanea, ventral view, male. Fig. 2. Mnemonica auricyanea, lateral view, female. Fig. 3. Mnemonica auricyanea, dorsal view, female. Fig. 4. Mnemonica auricyanea, dorsal view, caudal end of abdomen, male. Fig. 5. Mnemonica auricyanea, ventral view, caudal end of abdomen, male. Fig. 6. Mnemonica auricyanea, dorsal view, caudal end of abdomen, female. Fig. 7. Mnemonica auricyanea, ventral view, caudal end of abdomen, female. Fig. 8. Sthenopis thule, ventral view, male. Fig. 9. Sthenopis thule, lateral view, male. Fig. 10. Sthenopis thule, dorsal view, male. Fig. 11. Thyridopteryx ephemeraeformis, ventral view, male. Fig. 12. Thyridopteryx ephemeraeformis, lateral view, male. Fig. 13. Thyridopteryx ephemeraeformis, dorsal view, male. Fig. 14. Thyridopteryx ephemeraeformis, ventral view, female. Fig. 15. Prionoyxstus robiniae, face-parts, male. Fig. 16. Zeuzera pyrina, ventral view, male. PI.ATK XIX Plate XX Fig. 17. Lagoa erispata, ventral view, female. Fig. 18. Lagoa erispata, dorsal view, female. Fig. 19. Euelea chloris, ventral view, female. Fig. 20. Enelea chloris, dorsal view, male. Fig. 21. Enelea chloris, mesothoracic spiracle and adjacent parts. Fig. 22. Enelea chloris, cephalic view of head and prothorax. Fig. 23. Prolimaeodes scapha, ventral view, male. Fig. 24. Harrisina americana, ventral view, female. Fig. 25. Harrisina americana, dorsal view, female. Fig. 26. Prodoxus qnincinepnnctella, face-parts. Fig. 27. Prodoxus quinquepunctella, lateral view. Fig. 27a. Prodoxus quinquepunctella, tnhei'cle of the eighth abdominal segment, lateral view. Fig. 28. Hypocolpus mortipennellns, ventral view, female. Fig. 29. Hypocolpus mortipennellns, dorsal view, female. Fig. 30. Tinea pellionella, ventral view. mah'. Fig. 31. Tinea ])elli()ne]la, dorsal view. male. Plate XX Plate XXI Fig. 32. Brenthia pavonaeella, ventral view, female. Fig. 33. Brenthia pavonaeella, dorsal view, male. Fig. 34. Choreiitis gnaphiella, ventral view, female. Fig. 35. C'horeutis gnai)hiella, dorsal view, female. Fig. 36. Podosesia syringae, ventral view, female. Fig. 37. Synanthedon pictipes, dorsal view, head, thoi'ax and aljdoiuinal segments 1-3. Fig. 37a. Synanthedon pictipes, dorsal view, al)d(»niinal segments 8-10. Fig. 38. Aneylis comptana, ventral view, male. Fig. 38a. Aneylis comptana, anal rise, lateral view. Fig. 39. Aneylis comptana, dorsal view, male. Fig. 40. Exartema ferriferanum, ventral view, malt'. Fig. 41. Peronea minuta, lateral view, male. Pirate XXI Plati: XXII Fig. 42. Poroiica miiiuta, vcnti'al view, female. Fig. 43. Peronea luiiiiita, showing the connection of the t'ye-i)ieees with the vertex after removal of antennae at dehiscence. Fig. 44. Archips argyrospila, ventral view, female. Fig. 45. Gracilaria negnndella, venti'al view. male. Fig. 46. Gracilaria negundella, terminal segments of antenna. Fig. 47. Gracilaria sassafrasella, cephalic end of prothoracic and meso- thoracic legs, with adjoining ai'ca snp])osed to l)e the location of the maxillary jialpns. Fig. 48. Nepticula platanella, ventral view. male. Fig. 49. Nepticnla i)latanella, dorsal view, nude. Fig. 50. Antispila cornifoliella, ventral view, mal(\ Fig. 51. Goptotriche zelh I'iella, ventral view, male. Fig. 52. (''o])totriche zellerit'lla, dorsal view, male. Fig. 52a. Coptotriche zelleriella, lateral view, caudal end of abdomen. Fig. 521). Goptotriche zelleriella, tips of strongly chitinized setae. Fig. 53. Tischeria malifoliella, dorsal view, female. Plate XXII Plate XXIII Fig. 54. Tisehoi'ia holiopsisella, dorsal view, iiialc. Fig. 54a. Tischcria hcliopsisella, lateral view, caudal end of al)domen. Fig. 55. Bucculatrix sp., ventral view, male. Fig. 56. Bucculatrix sp., dorsal view, female. Fig. 56a. Bucculatrix sj)., lateral view of head. Fig. 57. Bedellia somnulentella, ventral view. Fig. 58. Bedellia somnulentella, dorsal view. Fig. 59. Proleucoptera smilaciella, ventral view. Fig. 60. Proleucoptera smilaciella, dorsal view. Fig. 61. Cameraria hamadrj-adella, ventral view, male. Fig. 62. Cameraria hamadryadella, dorsal view, male. Fig. 62a. Cameraria hamadryadella, lateral view, head. Fig. 62b. Cameraria hanuidi'vadella. dorsal view fifth and sixth abdom- inal segments, female. Fig. 63. Gracilaria negundella, dorsal view, male. Plate XXIII 60 «t Plate XXIV Fig. 64. Lithocolletis argoiitinotella, vinitral view, fomalo. Fig. 65. Lithoeolletis argentiiiotella, dorsal view, male. Fig. 66a. Lithoeolletis Incidieostella, dorsal view, cremaster, Fig. 66b. Lithoeolletis tiliacella, doi'sal view, cremaster. Fig. 67. Phyllocnistis insignis, ventral view. male. Fig. 68. Epcrminia pimpinella, ventral view, male. Fig. 69. Galleria melonella, lateral view, male. Fig. 70. Oxyptilus tennidaetylis. ventral view. Fig. 71. Oxyptilus tennidaetylis, dorsal view. Fig. 72. Atteva anrea, ventral view, male. Fig. 73. Atteva anrea, dorsal view, male. Fig. 74. Ephestia kuehniella, ventral view, female. Fig. 75. ]\Iiii('()]a iiidiginella, doi'sal view. Platk XXIV Plate XXV Fig. 76. Pyrausta futilalis, ventral view, female. Pig. 77. Epargyinais tityrus, lateral vieM'. Fig. 78. Calpodes ethliiis, ventral view. Fig. 79. Cyaniris ladon, ventral view. Fig. 80. Ocneis semidca, ventral view. Fig. 81. Euvanessa antiopa, lateral view, fifth al)clominal segment. Fig. 82. Zelleria eelastrusella. ventral view, male. Fig. 83. Yponomeuta malinelliis, ventral view. Fig. 84. Plutella maciilipennis, dorsal view. Fig. 85. Argyresthia freyella, ventral view, female. Fig. 86. Argyi-esthia freyella, dorsal view, female. Fig. 87. Coleophora malivorella, ventral view. Fig. 88. Lophoptilus eloisella, ventral view, female. Fig. 88a. Lophoptilns eloisella, lateral view, eaudal end of al)domen. Fig. 89. Scythris eboraccnsis, ventral view. Pi. ATE XX\' Plate XXVI Fig. 90. Gc'lcehia serotiiu'lla, vcnti'al view. Fig. 91. Ti"iehtotai)h(> liavoeostella. dorsal view. Fig. 92. Ti'iehtotapho fiavocostolla, ventral view. Fig. 93. Evippe prunifoliella, veiiti-al view. Fig. 93a. Evippe prunifoliella, latei'al view caudal end of a1)domen. Fig. 94. Ypsolophns eiti'ifoliella, ventral view. Fig. 95. ('hryso])eleia ostryaeella, ventral view. Fig. 96. Stenonia sehlaegei'i, venti-al view. Fig. 9(3a. Stenoma sehlaegeri, doi'sal and latei'al views, foni-th alxloni- inal segment. Fig. 961). Stenonia sehlaegeri, ti])s of large lateral setae. Fig. 97. Psiloeorsis (luercieella, venti'al view. Fig. 98. Psiloeorsis (iucrcicella, lateral view. Fig. 99. Cosmoptcryx clandestinella, ventral view. Fig. 100. Elaehista ]))'ae1ineata. ventral view. Plate XXVI 98 u Plate XXVII Fig. 101. Agrotis })icai'nea, dorsal view. Fig. 102. Cirphis iinipuneta, ventral view. Fig. 103. Eulonehe oblinita, ventral view. Fig. 104. Isia Isabella, lateral view. Fig. 105. Hemerocampa leucostigma, ventral view. Fig. 106. Malaeosoma disstria, ventral view. Fig. 107. Bombyx mori, ventral view. Fig. 108. Brephos infans, ventral view. Fig. 109. Hydria undulata, dorsal view, eremastral setae not present. Fig. 110. Cymatophora ribearia, dorsal view, caudal abdominal segments and eremaster. Fig. 111. Tephrocl.ystis absintliiata, doi-sal view, caudal abdominal seg- ments and eremaster. Fig. 112. Melalopha inclusa, dorsal view, abdominal segments and ere- master. Fig. 113. Datana angusii, dorsal view, caudal al)dominal segments and eremaster. Fig. 114. Schizura ipomoeae, dorsal view, caudal abdominal segments and eremaster. Fig. 115. Phryganidia calif ornica, ventral view. Fig. 116. Phryganidia californica, doi'sal view. PLATii XXVII mun i!D lyi/ 77/0 Bulletin OF THE Illinois State Laboratory Natural History Urbana, Illinois, U. S. A, STEPHEN A. FORBES, Ph. D., LL,. D. Director Vol. XII March, 1917 Article III. A PRELIMINARY CLASSIFICATION OP DIPTERA, EXCLUSIVE OF PUPIPARA, BASED UPON LARVAL AND PUPAL CHARACTERS, WITH KEYS TO IMAGINES IN CERTAIN FAMILIES. PART I. BY John R. Malloch Jr MAR 26 1917 Bulletin OF THE Illinois State Laboratory ov Natural History Urbana, Illinois, U. S. A. STEPHEN A. FORBES, Ph. D., LE. D. Director Vol. XII. March, 1917 Article III. A PRELIMINARY CLASSIFICATION OP DIPTERA, EXCLUSIVE OF PUPIPARA, BASED UPON EARVAL AND PUPAL CHARACTERS, WITH KEYS TO IMAGINES IN CERTAIN FAMILIES. PART I. BY John R. Malloch CONTENTS PAGE Introcluction 161-173 Acknowledgments 163 Habits and habitats of species 163 Methods of collection and preservation 164 Economic importance of the order 165 Arrangement of families 168 Scope of work 173 Characters of the larvae 173-177 Suborder Orthorrhapha 173 Division Nematocera 173 Division Brachycera 174 Suborder Cyclorrhapha 175 Division Acroptera 175 Division Aschiza 175 Division Schizophora 1^6 Characters of the pupae 177-179 Suborder Orthorrhapha 177 Division Nematocera 177 Division Brachycera 178 Suborder Cyclorrhapha 178 Divisions Acroptera, Aschiza, and Schizophora 178 Keys to suborders: — Larvae 179 Pupae 179 Imagines l^t* Suborder Orthorrhapha 180-407 Keys to divisions: — Larvae 1^*^ Pupae 1^0 Imagines 1°1 Division Nematocera "• 182-307 Tabular arrangement of families 182 Keys to families: — Larvae 1^^ Pupae 185 Imagines 188 Tribe Polyneura 190-245 Superfamily Tipuloidea 190-245 Family Tipulidae 191-207 Keys to subfamilies 193 Ctenophorinae 194 Tipulinae 195 Principal papers on North American Tipulidae 206 IV PAGE Family Limnobiidae 207-238 Keys to subfamilies 208 Cylindrotominae 210 Limnobiinae 212 Pediciinae 216 Limnophilinae 220 Ehamphidiinae 226 Eriopterinae 227 Hexatominae 232 Tiichocerinae 234 Papers on the Biology of North American Limnobiidae 238 Family Ptychopteridae 238-241 Family Rhyphidae 241-245 Tribe Eucephala 245-291 Superf amily Mycetophiloidea 246-262 Family Bolitophilidae 247-248 Family Mycetophilidae 248-258 Keys to subfamilies 251 Mycetophilinae 251 Sciophiliuae 255 Family Sciaridae 258-260 Family Macroceridae 260 Family Platyuridae 260-262 Principal papers on North American Mycetophiloidea 262 Superfamily Culicoidea 263-280 Family Psychodidae 264-274 Papers dealing with Biology of Psychodidae 273 Family Blepharoceridae 274-276 Principal papers on North American Blepharoceridae 276 Family Culicidae 276-279 Keys to subfamilies 278 Some of the more important works on North American Culicidae. . 279 Family Dixidae 279-280 Paper containing account of North American Dixidae 280 Superfamily Chironomoidea 280-291 Family Ceratopogonidao 281-284 Family Chironomidae 284-290 Keys to subfamilies 286 Tanypinae 287 Chironominae 287 Important papers on North American Chironomoidea 290 Family Orphnephilidae 290-291 Tribe Oligoneura 291-305 Superfamily Cecidomyioidea 292-297 Family Cecidomyiidae 293-297 Important papers on North American Cecidomyiidae 297 Superfamily Bibionoidea 297-305 Family Bibionidae 298-300 Important papers on biology of North American species 300 Family Scatopsidae ' 300-302 Paper on North American Scatopsidae 302 Family Simuliidae 302-305 V PAGE Addenda to Nematocera 305-307 Trichocerinae 306-307 Division Brachycera 307-407 Tabular arrangement of families 308 Keys to families: — Larvae 308 Pupae 310 Imagines 312 Tribe Platygenya 314-399 Superf amily Stratiomyioidea 314^354 Family Stratiomyiidae 315-346 Key to subfamilies 316 Stratiomyiinae 317 Clitellariinae 322 Beridinae 331 Geosarginae 331 Pachygasterinae 334 Xylomyiinae 340 Principal papers on North American Stratiomyiidae 346 Family Xylophagidae 346-351 Family Coenomyiidae 351-354 Family Acanthomeridae 354 Superf amily Tabanoidea 354-367 Family Tabanidae 355-361 Principal papers dealing with the biology of North American Tabanidae 361 Family Leptidae 362-367 Superf amily Cyrtoidea 368-369 Family Nemestrinidae 368 Family Cyrtidae 368-369 Papers on North American Cyrtidae 369 Superf amily Asiloidea 369-396 Family Mydaidae 370-373 Family Apioceridae 373 Family Asilidae 373-389 Family Bombyliidae 389-396 References to descriptions of larvae and pupae of North American Bombyliidae 395 Superfamily Therevoidea 396-399 Family Therevidae 396-398 Family Scenopinidae 398-399 Tribe Orthogeuya 399-407 Superfamily Empididoidea 399-407 Family Empididae 400-403 Papers on the biology of North American Empididae 403 Family Doliehopodidae 403-407 References to papers on the biology of North American and European Doliehopodidae 406-407 Article III. — A Preliniinary Clcissification of Diptcra, exclusive of Pupipara, based upon Larval and Pupal Characters, mith Keys to Imagines in certain Families. Part I. By J. R. M alloc h. Introduction Economic and taxonomic entomologists have long felt the need of a synoptic treatise on the immature stages of Diptera. Owing to the added impetus given to the study of this order by the comparatively recent discovery of the economic importance of certain members of the Diptera, the lack of available literature in English dealing with the larvae and pupae has been considerably emphasized. It is not the ijitention of the writer to attempt a definition of the larval and pupal characters that may be depended upon invariably for the separation of all the families of the order; a simple presentation of the main fea- tures involved in the differentiation of the families, genera, and spe- cies available to him is all that is attempted. It may be that these characters will in the main, with necessary modifications, prove of fundamental importance; but even if later they must needs be rele- gated to the category of things that have been they will at least have served the purpose for which they were intended — the identification of a number of important economic groups — and also constitute step- ping-stones to the firmer ground that can be reached only by a more extended knowledge of the larvae and pupae, and by means of patient and intelligent comparative examination of a larger amount of ma- terial than is now at my command. 1 have been requested by several of mv entomological colleagues to undertake the task of presenting analytical keys to the immature stages of Diptera, but for lack of material that would add to the infor- mation already published I have hitherto refrained from doing so. I have been steadily acquiring material, however, for over two years, which, taken in conjimction with material previously in the collection here and that kindly loaned me, probably includes more examples of immature stages of Diptera as a basis for classification than have been brought together elsewhere in the United States. It is, more- over, my opinion that when any entomologist has a considerable amount of additional information upon any phase of a subject it is 162 incumbent upon him, in the interest of his fellow workers, to publish not only his new data, be they new species or life histories, but inci- dentally to link up with his data such facts already published as have a direct bearing upon the subject in hand. It undoubtedly takes valu- able time to make a comparative description of a new species, and where several species in a genus require to be described it necessarily means considerable w^ork to the describes to make a synoptic key to the whole genus as well as to give a description of each of the new species. But when one considers that the time so spent is infinitesimal as compared with that saved to subsequent workers who would other- wise have to puzzle over descriptions in order to discover the differ- entiating characters, it is evident that synopses and comparative de- scriptions are not only useful but should be regarded as indispensable in scientific work. It is not with the intention of assisting the narrow specialist that this paper is written; the purpose is to enable the observant student of nature and the economic entomologist to recognize those forms that often come to their notice and thus obviate the frequent delays and discouragements in obtaining information through other channels. If this object is attained, even in a small measure, science will have gained some advantage and the author will be satisfied. An effort has been made in the general discussions, and particu- larly in connection with life histories, to avoid the use of pedantic terminology, as the writer is of the opinion that except in formal de- scriptions clarity of expression can be attained and conciseness com- passed by the use of language that is understood by the non-entomo- logical reader, much of the terminologv and phraseology incessantly cropping out in entomological publications being due either to the training or the personal whim of the writers. This paper deals primarily with Illinois species, most of the material used having been collected in the state by the various mem- bers of the staff of the State Entomologist's Office or that of the State Ivaboratory of Natural History. In a few cases, however, I have used cxtralimital material belonging to our collections, and have also bor- rowed examples of either larvae or of pupae, or of both, which were not obtained in Illinois, in order to ascertain by an examination of the specimens information not included in previously published de- scriptions that would enable me to complete, as far as possible, data upon certain genera or species. The classification adopted is essentially that used by Brauer in 1883*; but in detail I have not accepted his arrangement, nor have I *Deiiksehr. k. Akad. Wissensch., Wien, math.-naturw. CI., 1883, pp. 1-100. 163 used his anatomical nomenclature, for, as has already been pointed out by Lundbeck in his "Diptera Danica", Brauer has erred in this paper in the interpretation of different parts of both larvae and pupae. These misinterpretations are not emphasized herein, as the present writer believes that in undertaking original vi^ork such errors are likely to occur, and that in viewing the results of the efforts of authors we ought to adopt a perspective that permits a correct appreciation of the ditliculties under which the work was done. By this means we shall arrive at an estimate of the infinitesimal nature of existing blemishes in comparison with the immense advantages afforded to succeeding workers by a perusal of the author's published results. The present paper is presented without any foolish assumption of infallibility, and, as already stated, with a view to adding to the knowledge we already possess of the early stages of North American Diptera. In order to keep the size of the paper within as small a compass as possible, descriptions of species published by the writer in previous papers are not reproduced, but citations are uniformly given to facili- tate reference to them. ACKNOWLEDGMENTS I have to acknowledge the assistance rendered me in this study by the loan or gift of material as follows: from Dr. E. P. Felt, State Entomologist of New York; from W. L. McAtee, of the U. S. Bureau of Biological Survey; from J. J. Davis and J. A. Hyslop, of the U. S. Bureau of Entomology; from C. W. Johnson, of the Boston Society of Natural Historv; and from Dr. R. D. Glasgow, of the University of Illinois. In 1890-91 Dr. S. A. Forbes obtained a large amount of material from the streams and lakes in Yellowstone National Park, some of which represents genera and species of Diptera not in our collections from Illinois, and these have been used as a basis for con- tributory descriptions. Most of the material from the Illinois River which I have studied in this connection was collected by Mr. C. A. Plart, and some of it was used bv him in preparing his paper on the Entomology of the Illinois River*. Several of the species discussed or described were obtained bv W. P. Flint and D. K. McMillan in the course of their field work for the State Entomologist's office. HABITS AND HABITATS OF SPECIES Under family headings and very frequently in the discussion of genera I have given notes upon the habits of larvae and imagines, *Bull. 111. State Lab. Xat. Hist., Vol. 4, Art. VI. (1895) 164 while in the case of species that I have reared the notes furnish data upon the habitat. This makes unnecessary preliminary details on these points. In fact the Diptera have such a variety of habits in the larval and imaginal stages, and are so generally distributed, that one might be pardoned if he were to dismiss the subject with the laconic remark, "omnivorous and omnipresent". METHODS OF COLLECTION AND PRESERVATION It is necessary, in my opinion, in papers of this nature to give some general directions regarding methods of collecting and, as a part of collecting, rearing species. For the ordinary collecting of forms that frequent manure, de- caying wood, fungi, and mud or comparatively dry earth, as well as phytophagous species, the best temporary receptacle is a small round tin box about an inch and a half in diameter and three fourths of an inch deep. Dealers' in entomological supplies have these for sale, and they may be obtained with a paper-covered lid, upon the surface of which necessary data may be written. Many species may be reared to maturity in these boxes, the principal objection to this course being that it is not possible to learn whether the flies have emerged without opening the box, and frequently the specimen escapes upon the remov- al of the lid. I have had very satisfactory results with rearing-cages consist- ing of Petri dishes, the upper dish fitting over the lower when inverted. These cages, especially in the spring, proved all that were required to produce imagines from the larvae of Empididae, Xylophagidae, Syrphidae, and many other families. Aquatic species may be put in Mason fruit-jars or in small bottles, a convenient size of the latter being two-ounce. If not too tightly corked, specimens may be kept in such receptacles over night, but I find it best to use a cotton stopper instead of a cork unless while carrying the material in from the field. Many species, in fact most of the smaller forms, may be successfully reared in the two-ounce bottles, but I prefer to remove them during the pupal stage, or just before they transform to that stage, to a two-dram vial containing a little water and fitted with a cotton stopper. If transferred before transforming to the pupa the larval skin may be more easily found than in the larger bottle. A mistake frequently made by entomologists in preserving larvae is to put the live specimens into 85% alcohol. This course almost in- variably results in a shrinking of the skin and consequently seriously 165 impairs the value of the specimens. In order to procure the best re- sults in the case of soft-bodied specimens it is especially necessary that the examples be placed at first in water, which should gradually be brought to the boiling point, or near it, and then set to cool. If the water is allowed to boil violently it often results in a distortion or over-expansion of the specimens. Upon removal from the cool water place the specimens in 25% alcohol to remain six or eight hours; next transfer them to 50% strength, in which they should remain twenty- four hours, — after which treatment the larvae may safely be trans- ferred to 85% alcohol, in which they should be kept. Pupal exuvia that have the integument chitinized may be pre- served dry, but even such forms must be placed in alcohol if the imago has not emerged, as they shrink very much when preserved dry. I have succeeded in obtaining presentable specimens from com- pletely dried-out larval and pupal exuvia by boiling them in water. A larva or a pupa after drying out is rarely restored to its original form except by much patient work; but exuvia, even months after they have dried out, invariably recover their form when boiled. The head parts are generally easily traceable in larval exuvia, but in preserved larvae, especially of Brachycera and Cyclorrhapha, in which the head is retracted, dissection must be resorted to in order to get at the internal structures. I have had some success in ascertain- ing these details when I did not wish to cut up the specimen, by boiling it in 10% potash; but this is a tedious operation, especially if the specimen is a large one, and I prefer to expedite matters by cutting off the anterior two thoracic segments — caudad of which the cephalic skeleton does not extend — which may readilv be cleared so that dissec- tion is possible in a few minutes. The larger species have such heavily chitinized opaque heads that they are not good objects for slide mounts, but the smaller ones should invariably be mounted in Canada balsam. It is always necessary that some system of cross-reference be used for slide, vial, and imago in order to facilitate reference. More detailed information upon methods of mounting specimens is given in my article on the Chironomidae of Illinois.* ECONOMIC IMPORTANCE OF THE ORDER No other order of insects equals the Diptera in diversity of habits in larval and imaginal stages. Many of the families are largely bene- ficial, but unfortunately the good done by them is counterbalanced by *Bull. Til. State Lab. Nat. Hist., Vol. 10, Art. VI. (191.5) 166 the injury inflicted by others. The essentially phytophagous families, that is those families of which the great majority of the species feed upon plants, are very greatly outnumbered by those that are scaven- gers or predaceous or parasitic. If we exclude those that are fungiv- orous, only four families remain that can be classed as preponder- atingly phytophagous — Cecidomyiidae, Trypetidae, Agromyzidae, and Chloropidae ; a few species in these families are predaceous. It must be borne in mind that a phytophagous species is not necessarily in- jurious from the economic standpoint, as many species feed upon and keep in check noxious plants and may therefore be regarded as bene- ficial. It is but a step from the phytophagous to the scavenging habit, and in Drosophilidae we find species that may feed upon Cruci ferae, mining the leaves, or in sap exuding from trees and in vegetable refuse. Few other scavenging Diptera feed upon living plants, the only additional exception known to me being those that are fungivo- rous. There are eight families that may be considered as essentially fungivorous — Macroceridae, Bolitophilidae, Platyuridae, Mycetophili- dae, Sciaridae, Platypezidae, Phoridae, and Drosophilidae. Many of the Sciaridae occur in decaying vegetation, while the habits of Phori- dae are remarkably diverse, some being true entoparasites. The scavengers belong to more than a score of families. In Muscidae all the species scavenge ; but in some other families, An- thomyiidae, for example, we find phytophagous and inquiline species, though these are greatly in the minority and the family is essentially one of scavengers. The Sarcophagidae include some species that are true entoparasites. but the great majority are feeders upon decaying animal and vegetable matter. The scavengers are in the great major- ity of cases really beneficial, transforming dead animal and vegetable matter into such forms as can be utilized as food by growing plants. In reducing the bulk of putrefying substances, which, absorbed by the growing larvae, are transformed into the bodies of the resultant imagines, they remove what is noxious to man. It is chiefly when scavengers such as the common house-fly contaminate our food by contact, after feeding on foul substances which are impregnated with disease germs, that there is real danger from these insects. Rarely the screw-worm fly and some of the flesh-flies deposit their eggs or larvae in wounds, either on man or on animals, and in this manner produce serious ulcerations, and the larva of the former has been known to cause the death of persons by penetrating the brain, which it entered by way of the nasal passages. The flesh-flies and some other groups sometimes cause myiasis in man, the larvae finding their way into 167 the stomach with food in which the flies have deposited their eggs or larvae and which has not been prepared for consumption by judicious cooliing, or carefully examined so as to exclude infested portions. We may class as true parasites nine families, some of which, as Tachinidae [setts, lat.), Dexiidae, and Pipunculidae, are highly bene- ficial, and others, as Gastrophilidae, Hippoboscidae, and Oestridae, are distinctly injurious. The parasites of this order destroy many injuri- ous species of insects, and, next to the parasitic Hymenoptera, con- stitute the most important check upon their increase. Another group of highly beneficial species is that containing the predaceous forms. Two of the families which are to some extent beneficial in the larval stage — Tabanidae, and Culicidae in part — are injurious as imagines, turning their attention from insect larvae, on which they chiefly prey in the early stage, and giving it largely to mammals, including man. This radical change of habit is, however, exceptional, as other predaceous families in this and other orders feed upon insects in both the larval and imaginal stages. Many Syrphidae are aphidophagous as larvae, the greater portion of the species being scavengers, while the imagines are flower-frequenters. The aquatic families, with the exception of the Sciomyzidae and Ephydridae, which are in large part aquatic, belong to the Orthorrha- pha. With the exception of the Mycetophiloidea, which contains five families, the Oligoneura, which contains the Cecidomyiidae, and the families Bibionidae and Scatopsidae, all the families in the division Nematocera are aquatic either wholly or in large part. The aquatic species in the division Brachycera are contained in five families — Leptidae, Stratiomyiidae, Tabanidae, Empididae, and Dolichopodidae. As already indicated in the foregoing general discussion, the larvae of some of these families are predaceous and may justly be considered beneficial ; the others feed upon algae and decaying vegetable matter, and while their presence in water that is intended for drinking pur- poses is undesirable it is not necessarily harmful unless the vessel con- taining them is small and they are numerous enough to foul the water, either with excreta or exuvia. With the exception of some Chiro- nomidae and Culicidae there are few species that frequent reservoirs or cisterns, most of them preferring lakes, ponds, or streams. My information regarding the habits of the order in general leads me to the conclusion that as a whole their beneficial and injuri- ous activities practically offset each other. The fact that there are injurious species which cause great recognized damage, such as the malarial and other disease-bearing mosquitoes and the Hessian fly, very largely outweighs in the mind of the uninformed the benefits— 168 few of which are apparent except to a student of the Diptera — that are directly or indirectly due to the presence of other forms. With advance in a knowledge of the biology of the insects of this order will come a realization that their injurious and beneficial effects are practically balanced. ARRANGEMENT OF FAMILIES In agreement with the method generally used by systematists in zoological work, the arrangement of the Diptera is in accordance with the generally accepted theory of evolutionary development, and the families are thus arranged as nearly as possible in their natural sequence from lowest to highest, using as criteria the rather limited data furnished by available life histories, and by a study of imaginal characters ; but in endeavoring to trace affinities the reader must bear in mind that the families included are but the tips of the evolved branches, and not the entire genealogical tree. The hypothetical primi- tive dipterous larva Ms assumed to have had a complete head with horizontally moving mandibles, the head enclosing the first ganglion ; three thoracic segments, the prothoracic with a pair of spiracles; and ten abdominal segments, the anterior seven, or more, with lateral abdominal spiracles. No larva of this order has yet been discovered which possesses true thoracic legs, but there are many species that have pseudopods or sensory organs upon some of the thoracic seg- ments. This anatomical feature is not accounted as pertaining to a consideration of the phylogeny of the group, as pseudopods are gen- erally regarded as of secondary importance, being developed, partially developed, or absent, in species within the same family. The head is the best single unit available under all circumstances for the purposes of classification, as it is wholl}^ or in large part chitinized and its component parts are accessible for examination eith- er as a composite mass or after dissection, even in alcoholic material, whereas the nerve ganglia and even the tracheal system of alcoholic material are often indistinguishable, and in the case of exuvia entirely lost. The modifications exhibited in the head of the different families are remarkable, and though there are rather abrupt breaks in the chain of ascent as we pass from the Nematocera, with their opposed mandi- bles and complete or almost complete head-capsule, to the Brachycera, with their vertical subparallel mandibles and much reduced head- capsule, and, again, from the higher forms of this group to the Cvclorrhapha, and particularly to the Muscidae, one can trace with considerable probability the line of evolution up to the most highly 169 specialized forms of tlie present day. In this paper a large series of figures of heads of larvae belonging to all the principal sections is given in order to exemplify the evolutionary phases and, incidentally, to permit the student to judge as to the correctness of the classifica- tion by a comparison of the available data thus presented in the most readily comprehensible manner. An unillustrated discussion of ana- tomical details affords no check on possible misstatement or misin- terpretation by an author or on misconstruction of his words by the student, and to prevent error from one source or another such dis- cussions should be accompanied by figures. Brauer divided the Orthorrhapha into three tribes, Eucephala, Polyneura, and Oligoneura, using the structure of the head as his primary character in separating the groups, Eucephala having the head entire, the others having the head-capsule incomplete posteriorly. Oligoneura has in addition to an incomplete head-capsule vestigial mandibles, a character which separates the tribe from the other two. Brauer's classification has been generally accepted, though several writers have pointed out what they consider to be errors in grouping that result from the application of his rules, and Sharp has gone so far as to suggest that his system has been influenced by his use of dichotomic tables*. Recently a suggestion has been made that as the group Nema- tocera is apparently an unnatural one, containing, as it does, some widely dissimilar families which fall together according to Brauer's classification, we should attach primary importance, not to the struc- ture of the head but to the respiratory systemf. Here, again, an arbitrary attempt has been made to divide the group Nematocera into two tribes, Oligoneura and Polyneura, using the respiratory system as a basis for the division. In some respects the suggestion is an im- provement upon Brauer's system of classification, but even so, the composition of both groups shows some confusion which to my mind proves that the respiratory system is not an ideal character on which to base tribes. In fact the only point clearly shown is that some species of all of the families of Oligoneura, rarely, all species of some of them, are peripneustic and hence assumably primitive structurally, while all species, so far as is known, in Polyneura are amphipneustic. metapneustic, or apneustic. That there are both metapneustic and apneustic forms in Oligoneura of this latter classification, and that several have the lateral abdominal spiracles functionless and may *Verrairs "British Flies", Vol. 5, p. 31. (1909) tKiiab, in Ann. Ent. Soc. Amer., Vol. 7, 1915, p. 93. 170 therefore be considered as amphipneustic does not seem to have much weight with those who follow this dichotomic arrangement. Perip- neustic larvae occur in Dolichopodidae and vestigial spiracles are pres- ent in some Strayiomyiidae — facts which point, not to the more primi- tive nature of the species possessing these organs but to the persistence of the latter because of their utility in the larval habitat. There must be a further division of the tribes in Diptera, but until we have data upon a larger number of species in the order any proposed subdivision, including that of the present paper, must neces- sarily be merely tentative. I give a figure of a hypothetical genealog- ical tree of Orthorrhapha which illustrates my opinion of the re- lations of the various families (Fig. i). There are defects in the scheme that will probably be obvious to many, and it remains for some future worker to improve on the suggestion here given. The sequence of families in this paper is according to the author's ideas, differing somewhat from that in Williston's "Manual", but subject to amend- ment upon discovery of new data. Culicidae Dixidae Simuliidae Ceratopogonidac Orphiiephilidac ChiroiiOTnidae Platyiiridae Tipulidae Sciaridae Fig. 1. Hypothetical genealogical tree of Nematoeera, illustrating the grouping in this paper. The nervous system as a means of identification may be elimi- nated because of its limited applicability — in preserved material — but we may with little reserve accept as one of the primary indices to affinities the nature of the respiratory system. Insects normally breathe by means of thoracic and abdominal spiracles, and these are found in the adults and in a great majority of the pupae of the Diptera. It is the exception, however, to find them in the larvae, and as a means of 171 identifying those of many of the lower famihes the absence or pres- ence of the lateral abdominal spiracles is of considerable importance. It is not to be expected that a rule will be discovered as to the absence or presence of these spiracles which will be of invariable applicability to all the families — as at present constituted — in the Nematocera, be- cause we find, from data already in hand, that there appear to be several exceptional genera within the group, and also because perip- neustic larvae occur in much higher families. Taken in conjunction with the structure of the head, however, it is quite possible that the absence or presence of the lateral abdominal spiracles may prove to point to the tracheal system as of primary importance in classification. It must be borne in mind that many families as at present constituted include genera whose larvae have widely different habits, both ter- restrial and aquatic forms occurring at times in closely allied genera and, rarely, within the same genera. In this connection it seems perti- nent to indicate that our system of arrangement is based primarily upon the characters possessed by the imagines, but investigators may, without compunction, remedy errors in the system which, by an ex- amination of the immature stages, are shown to exist. One of the principal characters cited by Brauer for the separation of the Orthorrhapha from the Cyclorrhapha is the nature of the split- ting of the larval skin when the pupa or adult emerges. In the former the skin splits on the dorsum in the shape of a T or + ; whereas in the latter — in which the skin hardens in the last stage to form an en- closing puparium — the splitting is usually transversely across the en- tire dorsum and venter, so that the anterior portion of the puparium comes off cap-like; or there may be also lateral splits that cause the upper and lower halves of the anterior portion to separate. The lines of dehiscence in Cyclorrhapha puparia are not accidental, but are clearly provided for by weak parts in the membrane. The emergence of the imago in Orthorrhapha is assisted by a slight swelling in the thoracic region, and by movements of the developing wings and hard- ening legs ; in Cyclorrhapha the operation is facilitated by the dilation, with air, of a bladder-like sac immediately over the antennae, which, when swollen, forces apart the anterior extremity of the puparium along the lines indicated by the thinner membrane. After emergence the ptilinum, as the bladder-like structure is called, is retracted within an aperture above the antennae and forms a pouch-like cavity. The existence or absence of this peculiar structure in the imagines con- stitutes the primary character for the separation of the two subor- ders— Orthorrhapha being without and Cyclorrhapha with a ptilinum, its presence being indicated by what is called the frontal triangle, im- 172 mediately above the antennae. The famiHes Lonchopteridae, Pipiin- cuHdae, and Platypezidae have a very poorly developed ptilintim or it is undeveloped, but other characters appear to align them with Cyclor- rhapha. In the Orthorrhapha, the imagines of the Nematocera have the antennae usually much elongated and consisting of 7 or more distinct joints; the Brachycera have the antennae consisting of 3 joints, the third being either simple or composed of a number of closely fused ring-like joints having the appearance of an elongated single joint with more or less distinct subdivisions, and there is also, sometimes, a terminal or dorsal arista or style. In the Cyclorrhapha the antenna consists of 3 joints and a terminal or dorsal arista, the second joint in some families being small and entirely or almost entirely enclosed within the very large third, so that the antenna appears to consist of only 2 joints. The wing veins offer a good character for the differ- entiation of the families throughout the order, and being readily acces- sible have been made much use of by taxonomists. The lower forms have a much larger number of longitudinal veins than do the higher, and this seems to point to a coincident reduction of antennal segmenta- tion and wing venation in the evolution of the families of the order. An interesting piece of histological work might be undertaken in connection with the development of the members of the Muscidae and allied families, particularly in the observation of the developmental processes of the head-parts, which are so much reduced in the larvae of these forms. I present in this paper a kev to the imagines of the families of North American Diptera based upon the most recent data available. I have made use of many characters that are not in existing text-books, and in doing so I have simply attempted to put into writing data that many specialists merely carry in their minds. The characters used by Schiner and others for the separation of European families and genera are as a general rule applicable to North American Diptera, but there are many intermediate forms here that do not occur in Europe, and as the keys in use here have largely been copied or adapted from those in use in Europe considerable discretion is required in locating mem- bers of many families by these keys. The described dipterous fauna of this country is growing beyond the possibility of competent handling by one individual, and frequently men who are authorities on certain groups are unable to do more than make a guess at the family status of a particular species. I make this explanation in the hope that stu- dents will realize certain facts : that the beginner in the study is facing a stiff task in this order ; that the knowledge we possess concerning 173 the order is comparatively meager ; and that I reaHze the possibiHties of error that beset me in the present attempt. SCOPD OF WORK In order to keep the size of this paper within a reasonable com- pass I have given generic keys to imagines of those families only that are not correctly dealt with in Williston's "Manual of North American Diptera". I have covered the larval and pupal stages of all the groups avail- able to me. depending but rarely upon printed descriptions for differ- entiating characters. In the Orthorrhapha I have succeeded in obtain- ing larvae or pupae, or both, of all the families except Macroceridae, Orphnephilidae, Nemestrinidae, and Apioceridae. The Cyclorrhapha are not so fully represented, but I have a large quantity of material on hand which will serve as a basis for an extensive paper on this group, and this will be published as Part II of the present article. A bibliography of general papers on Diptera will be given at the conclusion of Part II, and papers upon single families will be listed in connection with the text dealing with such families. Characters of the Larvae Suborder ORTHOEHHAPHA Division Nematocera Broadly speaking, the dipterous larvae of Brauer's eucephalous group of the Nematocera may be distinguished from those of other Diptera and also from other orders, by the following characters. Head with opposed, usually toothed, mandibles, often, in the aquatic or semiaquatic forms, with conspicuous brushes on their ex- ternal surfaces; variously constructed antennae which may be barely distinguishable elevations (Mycetophilidae, Sciaridae) or elongated, and consisting of from i to 6 joints and occasionallv retractile within the head (Tanypinae) ; well-developed maxillae with i- or 2-jointed palpi ; no well-developed labial palpi ; a more or less chitinized labial plate, or submentum, which is frequently dentate upon its anterior margin ; eyes, when present, indicated by a single or double pigmented area on each side. True legs never present ; prothoracic segment occa- sionally with a pair of pseudopods which may be entirely or partly fused (Chironomidae, Simuliidae) ; anal segment in some cases with a pair of more or less elongated pseudopods which are sometimes armed with curved retractile claws in 2 or more concentric series; 174 abdominal, and occasionally also thoracic, segments sometimes with locomotor spinules in transverse series on portions of their ventral surface (Mycetophilidae, part) or with pseudopod-like elevations or fusiform transverse locomotor areas; respiratory system consisting, in its highest development, of i pair of prothoracic, 7 pairs of ab- dominal, and I pair of anal spiracles, the abdominal pairs not func- tional in some cases (Simuliidae, and Bibionidae and Scatopsidae ?), while in other families they are absent entirely, respiration being car- ried on either by means of the anal spiracles, or by these and the pro- thoracic pair, or by means of blood-gills on the anal segment which are sometimes retractile (Simuliidae, Ceratopogonidae) or perma- nently exserted {Chironomus, part). The acephalous larvae differ very markedly from those of the other group in the reduction of the head-capsule. In Cecidomyiidae the mandibles are vestigial or absent and the posterior portion of the head is poorly defined and membranous, but the larvae may be distin- guished by the presence of 13 segments, in addition to the head; by the lateral abdominal spiracles; and, usually, in the last instar, by the presence, on the ventral surface of the first and second thoracic seg- ments, of a chitinized plate, generally referred to as the "breast-bone", which is differently shaped in different species and is used by the mature larvae as a means of propulsion in making their leaps after leaving their cells for pupation — in cases where this change is made in the ground. In other larvae of this group there are at most but 12 distinct segments in addition to the head. The reduction of the head-capsule in Tipulidae and Limnobiidae is in the form of a breaking up of the fusion of the component parts posteriorly, the caudal portion of the head thus having the appearance of several slightly diverging rods, the membrane of the prothorax be- ing attached to the head just anterior to the point of divergence. These larvae may be distinguished from those of the higher families of the Brachycera by their opposed, instead of parallel, mandibles. Respira- tion is carried on by means of prothoracic and anal spiracles — the latter supplemented in the ac|uatic forms bv retractile blood-gills on the ventral surface of the last segment — or in some genera by the anal spiracles only. Division Brachycera The larvae of Brachycera are usually readily distinguished by the large maxillae, with their normally conspicuous palpi, between which are the slender labrum and the vertically moving, knife- or sickle-shaped mandibles on each side of it. The antennae are some- times conspicuously elevated but occasionally very short, and the eyes 175 are often readily distinguishable. The head is usually almost fully retractile within the thorax, being permanently exserted in Xylophag- idae, Coenomyiidae, and Stratiomyiidae. Respiration in this group is normally carried on by means of prothoracic and posterior spiracles, the latter being situated upon the ultimate ( Stratiomyiidae, Leptidae, etc.), penultimate (Asilidae, Mydaidae, Bombyliidae), or antepenul- timate (Therevidae, Scenopinidae) segments. Some species of perip- neustic Dolichopodidae are an exception to this rule. Many species of Brachycera have on their ventral surface locomotor organs, which may consist of mere transverse irregular swellings (Asilidae, in part, Mydaidae, Bombyliidae, etc.), paired pseudopods (Leptidae, part), or transverse series of spinules (Xylophagidae). Many of the families have conspicuous bristles on the body surface, especially on the ventral surface of the thoracic segments and the dorsal and apical surfaces of the anal segment (Asilidae, Mydaidae), or on all of the thoracic and abdominal segments (Xylophagidae, Stratiomyiidae). The normal number of segments in this group is 12, exclusive of the head; but in several families — Tabanidae, Stratiomyiidae, Lep- tidae— only 1 1 are distinguishable. Except in the Stratiomyiidae the pupa is free, that is to say, not enclosed in the indurated larval skin. Rarely the pupa is enclosed in a loose cocoon {Medctcrus) or in a very compact one (Drape tis). Suborder CYCLOREHAPHA Division Acroptera This group includes but one family, Lonchopteridae. The sys- tematic position of the group has long been in doubt, but the most recent and comprehensive work upon all stages of the species clearly points to their much closer affinity with the Cyclorrhapha than with the Orthorrhapha. The larvae are distinguishable from those of other Diptera by the fact that thev have but 9 well-defined segments in addi- tion to the head. A full discussion of the characters of this group appears under the family heading on a subsequent page. Division Aschiza This group consists of Syrphidae, Pipunculidae, Platypezidae, and Phoridae, according to Brauer. I have before me larvae and puparia of all but Pipunculidae. The cephalopharyngeal skeleton is better developed in this group than in Schizophora, but it is less per- fect than in the most specialized orthorrhaphid larva. In tiie aphi- 176 dophagous Syrphidae there are upper and lower chitinized mouth- hooks, which consist of two convergent lateral pieces with a V-shaped anterior extremity and correspond to the pair that are present in Schizophora; but in addition to these there are two to four pairs of small chitinized points or hooks which work horizontally and serve the purpose of grasping prey. The mouth-hooks in Schizophora are not V-shaped and the lower one is absent. These chitinized pieces are less readily distinguishable in the aquatic Syrphidae but are present in modified form, while the mouth-margin is specialized. The an- tennae in all species of the latter family are distinct, usually consisting of a single joint, rarely two, with a pair of short apical processes. The body consists of 12 segments, but the segmentation is indistinct be- cause of the presence of numerous transverse wrinkles or folds in the integument ; the surface of the body in most species bears regularly arranged bristles, which are occasionally upon slight elevations and serve to distinguish the different segments. Pseudopod-like structures are often present on ventral abdominal segments, their apices armed with short bristly hairs. Respiration is by means of prothoracic and anal spiracles, the former occasionally doubtfully functional or appar- ently absent, the latter protruding occasionally in a tube-like appen- dage. The larvae of Platypezidae resemble some of the Syrphidae rather closely, and bear a striking resemblance to the anthomyiid subfamily Fanniinae in having fringed projecting processes on the body seg- ments. The more elaborate mouth-parts readily separate them from all Anthomyiidae, and also, in my opinion, associate them more closely with Syrphidae than with Phoridae, though Brauer places them with the latter in the tribe Hypocera. The Phoridae are much simpler in general form than most Syr- phidae and all described Platypezidae, possessing, in as far as they are known, no prominent body appendages. The mouth parts are similar to those of the other families of the group in having anteriorly fused mouth-hooks. The larvae are amphipneustic — possessing prothoracic and anal spiracles. The transformation to the pupa takes place within the last larval skin. Division Schizophora The larvae of this group, which contains a great majority of the members of this suborder, are readily distinguished from those of any other order by the remarkable reduction of the head, which, when seen laterally, consists of a V- or U-shaped chitinized posterior plate, which has two posteriorly divergent dorsal rods, and, attached to the 177 anterior extremity of this plate, another, smaller plate, which is an- teriorly curved downward and consists of two — rarely one — hooks or mandibles which are either simple or toothed and operate vertically, scraping or abrading the surface of the larval pabulum so that the food is drawn downward and inward to the mouth cavity. These mouth-hooks, and the entire cephalopharyngeal skeleton, are retracted when the larva is at rest. The respiratory system consists of pro- thoracic and anal spiracles which are connected by means of two large main tracheae, the latter being normally connected by means of a slen- der transverse trachea just behind the prothoracic spiracle, and each sending out upon each segment a stout downwardly and slightly for- wardly directed stout trachea with numerous branches, and, in addi- tion, a smaller trachea, which is an offshoot from the inner surface and is directed almost straight cephalad. The structure of the spira- cles is of great value as a character in the classification of this group. The anterior spiracles are often cjuestionably functional in the aquatic forms, and are sometimes almost or entirely absent in parasitic species, the connection with the air being maintained entirely by means of the posterior pair. Means of locomotion in this group consist of spines on various portions of the thoracic and abdominal segments, with, occasionally, poorly developed pseudopods on the abdominal segments, particularly on the apical segment. Rarely there are elaborate appen- dages upon the thoracic and abdominal segments (Fanniinae), but more frequently the surface of the segments is bare except for the locomotor organs. The larvae transform to the pupal stage within the indurated last larval skin, which is then referred to as the puparium. In this stage characters are frequently developed that are of value in classification ; these are dealt with in the section upon characters of the pupae. Characters of the Pupae Suborder OETHOEEAPHA Division Xem.\tocera The aquatic members of this group may in the main be separated from those of the Brachycera by their stalked prothoracic respiratory organs, which are occasionally numerously filamented apically. The few species that are recorded as having these organs sessile may be distinguished from brachycerous pupae by the peculiar recurving of the legs against the ventral surface of the. base of the abdomen and the posterior portion of the thorax (Chironomidae). The terrestrial species are distinguishable from brachycerous forms by the very long antennal sheaths which curve over the upper margin of the eyes ; 178 whereas in the Brach}xera they are either short and inconspicuous or project almost in a straight hne across the front of the head, being usually armed with thorns, or they project divergently downward and are frequently armed with strong thorns. The abdomen has usually 7 pairs of lateral spiracles, but in most of the aquatic forms these are not distinguishable. Division Brachycera As mentioned under the previous heading, the pupae of this group very frequently have spines on the antennal sheaths and, in addition, similar spines on protuberances on the face or other portions of the head, and also on the thorax. The prothoracic spiracles are sessile or but slightly elevated except in some Empididae and Dolichopodidae. The abdomen has 7 pairs of lateral spiracles; the segments usually have girdles (1-2) of spines, thorns, or hairs; and the apical segment is usually armed with two or more strong terminal spines or stout processes. Only in the case of the Stratiomyiidae are the pupae en- closed in the last larval skin. The pupae of Coleoptera and Hymenoptera may readily be sepa- rated from those of Diptera by the mandibulate mouth-parts and the presence of four wing-pads, while the latter character will also sepa- rate those of Lepidoptera, though usually the under wings, or posterior wings, are visible only in the form of a narrow strip along the caudal margins of the front pair. Suborder CYCLOREHAPHA Divisions Acroptera, Aschiza, and Schizophora All the divisions of this suborder may be distinguished by the fact that the pupae are enclosed within the indurated last larval skin. The absence of a well-developed head will readily separate the puparia of this suborder from those of Stratiomyiidae; in the case of Lonchop- teridae, which resemble the latter family, there is no distinct head, there are only 9 distinct segments, and on the dorsum of the second ab- dominal segment there are horn-like respiratory organs as in Phoridae. The chitinized structure of the puparia itself will serve to separate them from those of the Cecidomyiidae that pupate vmder similar con- ditions, the integument of the latter being of a rather flimsy nature. In Lonchopteridae, Syrphidae, Phoridae, and many Muscidae and Anthomyiidae a pair of thoracic respiratory organs are developed upon the first or second abdominal segment in the puparia. These organs do not appear until the pupa is formed, and their mushroom-like sprouting comes rather as a surprise to the observer. 179 Many species that ordinarily make primary, if not exclusive, use of the anal respiratory organs while in the larval stage make exclusive use of the prothoracic spiracles in the pupal stage. This is noticeably so in species that live under aquatic or semiaquatic conditions. Keys to Suborders EARVAE 1. Head complete, or the posterior portion with deep longitudinal in- cisions ; mandibles moving horizontallv ORTHORRHAPHA-NEMATOCERA. — Head incomplete, without a strongly developed upper arcuate plate; mcndibles moving vertically 2 2. Maxillae well developed, their palpi distinct ; mandibles normally sickle-shaped, not protruded much beyond the apices of the maxil- lae, often extending less than half-way to their apices ; antennae well developed, situated upon the upper surface of a slightly arcuate chitinized dorsal plate ORTHORRHAPHA-BRACHYCERA. — Maxillae poorly developed, their palpi visible only in a few groups ; mandibles short and hook-like, usually capable of protrusion much beyond apices of maxillae if these are present; antennae poorly developecl or absent, when present situated upon a membranous surface CYCLORRHAPHA. PUPAE 1. Pupa not enclosed within the indurated last larval skin, or if so the head is distinct as in the larva, or the puparium is slightly flat- tened dorso-ventrally, its texture leathery, not chitinous, and the anterior respiratory organs not distinguishable ; imago, or pupa, emerges through a rectangular split on dorsum of larval skin. . . ORTHORRHAPHA — Pupa enclosed within the indurated last larval skin ; head always retracted, the chitinous portion occupying a position on the inner side of the ventral surface of the puparium ; anterior respiratory organs distinct, either protruded from the antero-lateral angles of the cephalic extremity or from dorsum of base of abdomen; imago usually emerges by forcing off the rounded anterior ex- tremity of the puparium in cap-like form, or the dorsal half of the thoracic portion — the lines of cleavage being along the lateral margins to a point at base of abdomen; rarely emergence is through rectangular splitting of the dorsum of the puparium. . . CYCLORRHAPHA. 180 IMAGINES Antenna consists of a 2- jointed scape — the basal joint usually indis- tinct — and a distinctly segmented flagellum of more than 2 joints ORTHORRHAPHA, pt. Antenna consists of apparently 2 or 3 joints and, usually, a ter- minal or dorsal arista 2 Frontal lunule absent ORTHORRHAPHA, pt. Frontal lunule present, or if poorly developed or apparently absent there are no distinct cross-veins on disc of wing CYCLORRHAPHA. Suborder ORTHORRHAPHA Keys to Divisions LARVAE Head complete, or sometimes incomplete posteriorly ; mandibles ab- sent or vestigial in Ceeidomyiidae, but the body consisting of 13 segments in addition to the head, in other families mandibles present, moving horizontally or nearly so, their position when at rest being on a horizontal plane, or varying but slightly from it, their apices opposed ; labium usually well developed, in the form of a flat plate with or without dentate anterior margin ; larvae frequently peripneustic, the aquatic forms usually with protrusive anal blood-gills NEMATOCl^^RA (p. 182). Head always incomplete and partly retracted within the prothorax ; mandibles moving vertically, their position when at rest vertical, the apices parallel and directed downward; labium poorly de- veloped, rarely or never in the form of a flat plate; larvae nor- mally amphipneustic or mctapneustic, very rarely peripneustic, aquatic forms without protrusive anal blood-gills BEACHYOERA (p. 307). PUPAE Head, except in some Ceeidomyiidae and a few Tipulidae, without strong thorns ; antennae much elongated, always very distinctly traceable, normally curved well over upper margin of eyes and extending to or beyond base of wing, in some cases almost to apex of wing; thoracic respiratory organs much elongated or sessile; legs of variable length, often extending to apex of abdomen, rarely slightly longer than wings, in aquatic forms often recurved against base of abdomen ; abdomen in species with short antennae some- times unarmed NEMATOCERA (p. 182) . Free except in Stratiomyiidae ; head in other families usually with 181 strong thorns, or if these are absent the antennae are very short and project downward and outward and do not curve over the upper margin of the eye or reach nearly to base of wing ; thoracic respiratory organs sessile, rarely stalk-like ; legs of variable length ; abdomen usually armed with strong spines or bristles, or if unarmed there are only 4 or 5 distinct pairs of abdominal spira- cles BRACHYCERA (p. 307). IMAGINES 1. Antenna consisting of at least 7 joints, frequently filiform, rarely conspicuously thickened, but if so the joints are distinct and the flagellum does not appear as a single joint with poorly defined subdivisions, and the palpi are pendulous, consisting of 4 or 5 joints; antennae with or without whorls of hair on the joints, never with a terminal arista or style; palpi always pendulous, normally consisting of 4 or 5 joints, very rarely of only 2 NEMATOCERA (p. 182). — Antennae consisting of 3 j"oints, the third occasionally having more or less distinctly annulated subdivisions, but in such cases the palpi are porrect and consist of 1 or 2 joints ; antennae with short pubescence or with thick branches, in some families with either a terminal or subterminal style or arista; palpi projecting forward, consisting of 1 or 2 joints BRACHYCERA (p. 307). 182 Division NEMATOCERA TABULAR ARRANGEMENT OF FAMILIES My present grouping of the Nematocera is as follows. Tribes'^ Polyneura Superfamilies Tipuloidea Families Tipulidae Limnobiidae Ptychopteridae Rhyphidae Eucephala Oligoneura Mycetophiloidea Culicoidea Chironomoidea Cecidomyioidea i [ Bibionoidea Bolitophilidae Mycetophilidae Sciaridae Macroceridae Platyuridae Psychodidae ^ Blepharoceridae Culicidae Dixidae r Ceratopogonidae "1 Chironomidae t Orphnephilidae Cecidomyiidae Bibionidae Scatopsidae Simuliidae The sequence of the families in the keys is not in accordance with the above list, the keys being framed to facilitate identification and not to indicate affinities. *Tribe in this paper does not have the application given to it in contemporary papers, bnt has that which Brauer gave it. He used it to designate his subdivisions of the larger divisions of Nematocera, etc. 183 Keys to Families LARVAE 1. Head incomplete; thorax and abdomen combined consisting of 13 segments; larvae peripneustic ; usually with a chitinized plate (very distinct in mature larvae) on ventral surface of second thoracic segment Cecidomyiidae (p. 293). — Without the above combination of characters 2 2. Head and thoracic and first and second alidominal segments fused ; larvae with minute abdominal spiracles ; abdomen with a ventral longitudinal series of sucker-like discs Blepharoceridae (p. 274) . — Head free, or if retracted within or fused with prothoracic segment the other thoracic segments are distinct 3 3. Head complete, enclosing first ganglion ; mandibles opposed 5 — Head incomplete posteriorly, either with 3 deep wedge-shaped slits, 2 on dorsum and 1 on venter, or the ventral surface very poorly chitinized and the dorsal one posteriorly in the form of 4 slender heavily chitinized rods, with a weakly chitinized divided plate on anterior half of the dorsum 4 4. Apical abdominal segment with 6 radiating protuberances, which are rarely poorly developed but frequently unequally so; body segments with regularly placed bristles, as shown in Figure 1, Plate XXVIII ; head heavily chitinized, dorsally slightly arcuate and with 2 longitudinal slits, ventrally slightly rounded and with a central slit ; antennae longer than maxillary palpi ; labium pointed, not divided into 2 plates, the anterior margin dentate; mandibles very stout, with only 2 teeth (at apex) in species with- out apical processes Tipulidae (p. 191). — Apical abdominal segment with at most 5 radiating teeth, or if 6 are present the labium is subdivided centrally ; body usually with- out regularly placed bristles, frequently with dense surface pilosity; head sometimes weakly chitinized and without distinct labium ; antennae sometimes short and slender and not as long as maxillary lobe (not palpus) ; labium frequently subdivided into 2 plates; mandibles never with only 2 teeth . Limnobiidae (p. 207). 5. Thoracic segments fused and dilated, forming a complex mass .... CULICIDAE (p. 276) . — Thoracic segments not fused, distinct 6 6. Larvae peripneustic, or with at least rudimentary abdominal spiracles 7 — Larvae amphipneustie or metapneustic 12 7. Larvae with rudimentary abdominal spiracles; mouth with a large articulated process on each side which bears a number of long hairs and closes, fan-like, when at rest; posterior abdominal seg- ments dilated, the last one armed on venter with a sucker-like disc which bears concentric series of bristles, by means of which the 184 larvae retain their hold upon rocks, etc., in the streams where they are found Simuliidae (p. 302) . — Larvae with distinct though sometimes small abdominal spiracles; mouth without fan-like processes ; posterior abdominal segments not noticeably dilated, the last without sucker-like disc ; terrestrial species 8 8. Antennae elongate; body armed with conspicuous bristles or hairs ! 9 — Antennae usually short and inconspicuous, sometimes apparently absent; body without conspicuous bristles 10 9. Anal spiracles at the apices of a pair of long stalk-like processes. . . . SCATOPSIDAE (p. 300) . — Anal spiracles not noticeably elevated, situated near base of dorsal surface of apical segment Bibionidae (p. 298). 10. Dorsal, or clypcal, sclerite of head not conspicuously tapered pos- teriorly; antennae well developed Bolitophilidae (p. 247). — Dorsal, or clypeal, sclerite of head conspicuously tapered pos- teriorly; antennae almost indistinguishable 11 11. Lateral sclerites of head meeting on ventral line only for a short space immediately eaudad of mouth-opening, not connected at posterior margin Mycetophilidae (p. 248). — Lateral sclerites of head connected for a very short space behind mouth-opening and again near posterior margin SCIARIDAE (p. 258) 12. Dorsal surface of first and second abdominal segments each with 2 wart-like elevations somewhat resembling pseudopods, the apices of which are armed with numerous small hook-like setae; larvae aquatic, amphipneustie Dixidae (p. 279) . — Dorsal surface of first and second abdominal segments without ele- vated processes 13 13. All or some of the dorsal segments with narrow, chitinized, strap- like transverse bands, or the apical segment in the form of a short chitinized tube ; rarely the ventral abdominal segments bear a central series of sucker-like discs Psychodidae (p. 264). — ■ Dorsum without narrow, chitinized, strap-like bands, apical segment not in the form of a short chitinized tube ; ventral abdominal seg- ments never with sucker-like discs 14 14. Apical abdominal segment with long, slender respiratory tube .... Ptychopteridae (p. 238) . — Apical abdominal segment without long respiratory tube 15 15. Antennae undeveloped, appearing as pale round spots on sides of head ; ventral surface of head with the sclerites contiguous ante- riorly, Avidely separated posteriorly 16 — Antennae pedunculate, usually well developed; ventral surface of head with sclerite contiguous on entire length, not separated widely posteriorly 17 185 16. Head subquadrate ; abdominal segments with a number of rounded transverse ridges Platyuridae (p. 260) . — Head elongate ; abdominal segments without transverse ridges Mycetophilidae (p. 248) . 17. Abdominal segments not subdivided . . . '. 18 — Abdominal segments subdivided by means of transverse constric- tions . . 19 18. Larva very slender, tapering towards the extremities, without thoracic or anal pseudopods or surface hairs except about 8 at apex of abdomen, aquatic in habit; or stout, with well-defined segments which are armed with strong bristles, some of which are lanceolate ; pseudopods present ; terrestrial, living in manure or under bark, etc Ceratopogonidae (p. 281) . — Larva rarely very slender, generally of an almost uniform thickness, rarely with the thoracic segments appreciably swollen but not fused ; abdominal and thoracic segments frequently with rather noticeable soft hairs, the last segment almost invariably with a conspicuous tuft of hairs on dor.sum near apex ; pseudopods al- most always present, sometimes only the thoracic one distinguish- able in terrestrial forms-which are very rare Chironomidae (p. 284) . 1 9. Body slender, tapering ; abdominal segments each with a single con- striction near anterior margin ; apical segment either with 5 short terminal processes or without distinct processes Rhyphidae (p. 241 ) . — Body stout, of uniform diameter; abdominal segments each with 2 distinct constrictions ; apical segment with 4 rather long processes, the low^er pair longer than the upper. .Limnobiidae, pt. (p. 207). pupae 1. Head with several strong thorns in a vertical series on the median line ; pupae enclosed within galls on various parts of plants .... Cecidomyiidae, pt. (p. 293) . — Head Avithout strong thorns, or if at base of each antenna there is a protuberance it is not sharp and thorn-like, and the pupae are not enclosed in galls on living plants 2 2. Pupa enclosed within a tough, parchment-like envelope consisting of the hardened larval skin, which resembles a muscid puparium. . Cecidomyiidae, pt. (p. 293) . — Pupa free, or if enclosed it is within a cocoon which is not parch- ment-like and does not resemble a muscid puparium 3 3. Thoracic respiratory organs sessile ; abdomen without strong thorns or leaf -like elevations ; legs straight 4 — Thoracic respiratory organs stalk-like, or if sessile the abdomen has strong thorns or leaf-like elevations, or the legs are recurved against base of abdomen and apex of thorax, or the coxae do not 186 conceal the sternopleiira and the scape of the antennae is almost globose ; legs sti-aight or recurved 8 4. Legs short, apices of hind tarsi projecting slightly beyond apices of wings ; antennae short, curved across middle of eye BiBIONIDAE (p. 298). — Legs elongate, usually all tarsi projecting for a considerable dis- tance beyond apices of wings ; antennae elongate, extending to or beyond bases of wings 5 5. Antennae almost straight, noticeably flattened, extending to bases of wings ; thorax not much swollen in front, its anterior pro- file not declivitous Platyuridae (p. 260) . — Antennae distinctly curved, not flattened, extending beyond bases of wings 6 6. Thorax conspicuously swollen, almost globose, its anterior profile de- clivitous; sternopleura concealed Mycp:tophilidae (p. 248). — Thorax not conspicuously swollen, the anterior profile not declivi- tous 7 7. Scape of antennae much swollen, globose ; abdominal spiracles small or absent; sternopleura remarkably enlarged, not concealed by fore coxae and femora Chironomidae (p. 284). — Scape of antennae not much swollen ; abdominal spiracles distinct ; sternopleura not visible, concealed by the large coxae and femora of the fore legs Cecidomyiidae, pt. (p. 293) . Sciaridae, pt (p. 258). 8. Thoracic respiratory organs slender, long, and tube-like: legs straight, extending well beyond apices of wings ; body without armature except a pair of hairs on anterior margin of head ; sternopleura concealed Cecidomyiidae, pt. (p. 293). Sciaridae, pt (p. 258). — Species without the above combination of characters, abdomen usually with hairs or spines, or the sternopleura is exposed 9 9. Pupa in a pocket- or slipper-shaped cocoon consisting of coarse threads, from the wide, open extremity of which project the thoracic respiratory organs, each of the latter consisting of 4 to 60 tube-like branches on a common base; rarely the cocoon is a mass of rather loose threads Simuliidae (p. 302). — Pupa free, or if enclosed or partly so the cocoon is not pocket-like and the respiratory organs do not consist of tube-like branches. 10 10. Pupa when seen from above oval in outline, the abdomen at base not conspicuously narrower than thorax, so that the lateral outline is continuous ; dorsal surface with very strong, almost chitinized, membrane 11 — Pupa with the abdomen well differentiated from thorax, the dorsum membranous, or if strong and almost chitinized, then with surface spines 12 187 11. Thoracic respiratory organs lamelliform, consisting of 4 flat plates, the broad sides of which are contiguous Blepharoceridae (p. 274) . — Thoracic respiratory organs simple, tube-like PSYCHODIDAE, pt. (p. 264). 12. Apical abdominal segment terminating in 2 or 4 paddle- or fin- shaped organs which are fringed on all or a part of their outer surfaces with strap-like hairs ; or if the apical segment terminates in 2 long subconical processes the tarsi are, as in the other group, recurved against the ventral surface of the base of abdomen and apex of thorax so that they do not extend beyond the apices of wings 13 — Apical abdominal segment obtuse, armed with short or elongate spines or thorns, or if ending in a pair of long, slender processes these are more or less oval or circular in transverse section and without strap-like hairs ; tarsi generally entirely straight, rarely the apices of the hind pair incurved slightly, but they are never recurved as above 18 13. Thoracic respiratory organs terminating in numerous thread-like filaments Chiroxomidae, pt. (p. 284). — Thoracic respiratory organs consisting of a single stem, in some cases with a few long, or many short, scale-like surface hairs, but never terminating in numerous thread-like filaments; or occa- sionally the thoracic respiratory organs are not elevated 14 14. Thoracic respiratory organs not elevated, sternopleura exposed. . . . Chironomidae, pt. (p. 284) . — Thoracic respiratory organs conspicuously elevated 15 15. Thoracic respiratory organs situated close to anterior margin of thorax ; thorax and abdomen without stellate hairs Chironomidae, pt. (p. 284) . — Thoracic respiratory organs situated close to middle of thoracic dor- sum 16 16. Apical abdominal segment ending in 2 or 4 broad, flat, paddle-like plates Culicidae, pt. (p. 276) . — Apical abdominal segment ending in 2 long subconical processes . . 17 17. Apical processes armed at apices and on middle of their outer mar- gin with short hairs (3:1) Culicidae, pt. (p. 276) . — Apical processes unarmed Dixidae (p. 279). 18. Apices of legs not extending beyond apices of wings 19 — Apices of posterior legs at least extending beyond apices of wings 20 19. Apical abdominal segment ending in 2 conical processes Ceratopogonidae (p. 281) . — Apical abdom.inal segment ending in 2 upper and 2 lower short thorns Psychodidae, pt. (p. 264) . 188 20. Thoracic respiratory organs long, bifid; apical abdominal segment rounded, without processes ; abdominal spiracles pedunculate .... SCATOPSIDAE (p. 300) . — Thoracic respiratory organs simple; apical abdominal segment not rounded, generally armed with protuberances .21 21. Thoracic respiratory organs elevated but little above the level of disc of thorax ; tarsi of the fore legs overlapping those of mid pair, the latter overlapping those of hind pair, all rather closely fused together and to wings Rhyphidae (p. 241). — Thoracic respiratory organs very conspicuously elevated ; legs not as above 22 22. Thoracic respiratory organs equal in length, rarely with one twice as long as the other ; anterior, middle, and posterior tarsi distinct 23 — Thoracic respiratory organs of conspicuously unequal length, one many times as long as the other ; anterior tarsi overlapping mid- dle pair Ptychopteridae (p. 238). 23. Abdominal segments each with 1 transverse row, sometimes 2 such rows, of thorn-like protuberances ; palpi recurved at apices Tipulidae (p. 191). — Abdominal segments rarely with distinct thorn-like protuberances, usually with weak hairs; palpi straight, not recurved at apices. . Limnobiidae (p. 207). IMAGINES 1. Wing with at least 9 veins extending to the margin (exclusive of the anal vein) ; if there are only 8 such veins the radius is 3-branched, the second branch having its base proximad of the radio-medial cross-vein 2 — Wing with loss than 9 veins extending to the margin, or the vena- tion not as above 8 2. Mesonotum with a more or less distinct V-shaped suture; male hypopygium generally very large, chitinous ; female ovipositor conical, chitinized, and generally protruded 3 — Mesonotum without distinct suture, or if there is a poorly defined suture it is not V-shaped 5 3. Wing with 2 anal veins 4 — Wing with 1 anal vein Ptychopteridae (p. 238). 4. Last palpal joint slender, much longer than the combined lengths of the 3 preceding joints ; auxiliary vein terminating in first vein . . . Tipulidae (p. 191) . — Last palpal joint at most but little longer than the combined lengths of the preceding joints; auxiliary vein usually terminating in costa, connected with first vein by a cross-vein Limnobiidae (p. 207). 5. Costa continued around the hind margin of the wing 6 — Costa discontinued at apex of wing Rhyphidae, pt. (p. 241). 189 6. Wing veins without conspicuous scale-like hairs. .Dixidae (p. 279). — Wing veins with conspicuous scale-like hairs 7 7. Wings short and broad, ovate, occasionally pointed apically; tibiae without apical spurs ; small, robust species with rather short densely haired legs Psychodidae (p. 264) . — Wings elongate, narrow ; tibiae with apical spurs ; rather large, slen- der species, with long, slender, usually moderately hairy or scaly legs CuLiciDAE (p. 276). 8. Wing with 2 to 4 distinct longitudinal veins Cecidomyiidae (p. 293) . — Wing Avith 5 or more longitudinal veins ^ 9 9. Wings with a secondary reticulation of fine creases or lines in addi- tion to the veins ; slender tipulid-like species with long slender legs Blepharoceridae (p. 274) . — Wings without a secondary reticulation of fine creases, at most with a longitudinal furcate crease between media and cubitus 10 10. Abdomen in both sexes with a conspicuous flap-like scale at base of dorsal surface which is detached posteriorly and fringed with long hairs Simuliidae (p. 302) . — Abdomen without such basal process 11 11. Second basal cell of wing present 12 — Second basal cell of wing absent 13 12. Antenna consisting of 2 stout joints, and an apical arista-like one composed of 9 or 10 segments Orphnephilidae (p. 290). ■ — Antenna composed of 10-11 joints, the apical portion stout, not dif- ferentiated arista-like Bibionidae (p. 298). 13. Antenna consisting of 10-11 joints in both sexes (10 if scape is re- garded as consisting of 1 joint), the joints of central portion of flagellum shorter than broad ; radius and eosta conspicuous, the other veins indistinct ; at least the mid and hind tibiae without apical spurs Scatopsidae (p. 300) . — Without the above combination of characters 14 14. Coxae unusually elongated 15 — Coxae not unusually elongated 19 15. Radius with 3 branches ; mcdio-cubital cross-vein present 16 — Radius with only 2 branches, or if there are three present the medio- cubital cross-vein is absent Mycetophilidae (p. 248). 16. Radio-medial cross-vein present, causing the base of the first poste- rior cell to be more or less broadly truncate 17 — Radio-medial cross-vein apparently absent, fused with base of third branch of radius so that the base of fii\st posterior cell is acute . 18 17. Medio-cubital cross-vein much proximad of the radio-medial, caus- ing the posterior portion, divided longitudinally by media, to be much shorter than the anterior portion. . Bolitophilidae (p. 247). — Medio-cubital cross-vein almost directly in vertical line with the radio-medial, the 2 cells divided longitudinally by media subequal in length Rhyphidae, pt. (p. 241) . 190 18. Antennae short, thick, and often flattened. . .Platyuridae (p. 260). — Antennae very long, usually exceeding in length that of body, very slender Macroceridae (p. 260) . 19. Mouth parts chitinized, constructed for piercing, not in the form of a long slender process Ceratopogonidae (p. 281). — Mouth parts not chitinized, fleshy, sometimes in the form of a long slender process 20 20. Radius with 2 branches Sciaridae (p. 258). — Radius with 3-4 branches, if less or indistinct the antennal joints with very long plumes in the male and very conspicuous constric- tions between them in the female Chironomidae (p. 284). Tribe Polyneura I have included in this tribe, as I regard it, but one superfamily, Tipuloidea, containing the families Tipulidae, Limnobiidae, Ptychop- teridae, and Rhyphidae. Brauer limited his tribe Polyneura to Tipulidae (inclusive of Limnobiidae), placing the other two families in Eucephala. I have placed this tribe first in my arrangement because I consider the adults much more primitive structurally than the most generalized forms in the other groups. The larvae undoubtedly show more special- ization than do those of Mycetophiloidea both in the structure of the head, if we accept the capsule as the criterion, and in the respira- tory system, but I find that the larvae of many closely allied species in different families show quite striking distinctions even though the adults are almost inseparable, and therefore have decided to consider the tribe as more generalized than the others. The sequence of fami- lies does not show an unbroken line, but, rather, represents a series of divergent lines of varying lengths, no two of which start from a com- mon point. Superfamily Tipuloidea SUPERFAMILY CHARACTERS Larva. — Head incomplete posteriorly, wholly or partly retractile, or if the head is complete the abdomen has the anterior 6 segments sub- divided, or the posterior respiratory tube is much elongated and mem- branous, and distinct paired ventral pseudopods are present on anterior half of body. Head with opposed mandibles ; antennae well developed. Pupa. — Head without conspicuous armature except in some Limnobiidae; antennae elongate, curved over upper margin of eyes. Wings and legs closely fused to thorax, the former very short, the latter never extending to apex of abdomen ; thoracic respiratory 191 organs usually elevated, one of them much longer than the other in Ptychopteridae, the two normally of equal length and usually slender in Tipulidae, ear-like, appearing like vertical plates, in some Limno- biidae, while in Rhyphidae they are but little elevated and rather stout. Abdomen in Rhyphidae and most Tipulidae circular in cross-section, usually with 2 transverse series of more or less leaf-like or thorn-like protuberances on each segment. Imago. — Distinguishable from all other Nematocera by the pres- ence of the discal cell of wing. Mycefobia has no discal cell. This genus has been placed by most writers in the Mycetophiloidea, but lately its affinities with Rhyphidae have been considered closer, and Edwards traces in the presence of a well-defined gular plate a distinct connection with that family, this plate being almost invariably absent in Mycetophilidae, and even when present differing materially from that of Mycetohia. The venation of the Mycetohia wing differs from that of Mycetophilidae in that the second branch of the radius has its base proximad of the radio-medial cross-vein instead of distad of it. The female of Mycetohia has chitinized spermathecae, this character separating it from Mycetophilioidea, no genus of which possesses them. Some Tipulidae have no discal cell, but they all have a distinct V-shaped suture on the dorsum of the thorax — a character which dis- tinguishes them from other Nematocera. Ptychopteridae possesses an incomplete V-shaped or slightly sinuous thoracic suture, and in com- mon with related Nematocera, except Rhyphidae, has very long legs, slender wings, and a long slender body. Allocation of species of the families must be arrived at by using the key to families of the Nematocera on a previous page. Family TIPULWM This family as limited in the present paper contains only three subfamilies : Dolichopezinae, Ctenophorinae, and Tipulinae. Only the two last named are known to me in their immature stages. The number of species of Ctenophorinae in North America is small, but the genus Tipiila, in Tipulinae, contains a very large number of species, the larvae, pupae, and imagines of which in very many cases bear a striking resemblance to each other. With the present collection it is not possible for me to do more than to indicate the principal characters useful in distinguishing the larvae and pupae from each other and from those of other families. 192 FAMILY CHARACTERS Larva. — Head heavily chitinized, retractile within prothorax; posterior portion deeply cleft longitudinally, one incision on each side of dorsum extending to, or almost to, middle, and one in center of venter extending beyond middle. Antennae elongate, consisting of an elevated base, an elongate joint — in Tipula at least four times, in Xiphura not more than twice, as long as broad — and a very short apical process. Front and clypeus fused, the former with a distinct plate on each side, the anterior margin of which is armed with a few bristles and many hairs; labrum fringed with hairs anteriorly; epipharynx with a number of spines or short processes. Maxillae well developed, fringed on the inner and anterior margins with hairs, and sometimes spinose ; palpi small. Mandibles stout, the apex rounded and with 2 teeth more or less equal in size, lower margin with 2 or more teeth, inner upper margin with a fringe of hairs near middle. Labial plate (submentum) in the form of a flat plate, its outline medianly produced into an acute point anteriorly, the margin more or less dis- tinctly dentate. Hypopharynx heavily chitinized, in the form of a flat plate the anterior margin of which is usually dentate, and with a posterior inverted-U-shaped chitinized piece which arches over the oesophageal opening. Body cylindrical ; segments usually with dis- tinct hairs which are situated on certain portions of each segment, their arrangement being uniform throughout the family ; pseudopods present or absent; segments always with transverse linear depressions, most distinct on dorsum ; apical segment with 6 processes, rarely with- out these being well developed ; ventral surface of apical segment in aquatic and semiaquatic species with fringes of soft hairs on apical processes and with slender protrusive blood-gills ; terrestrial forms with the fringes of hairs much reduced or absent, and the slender protrusive blood-gills usually absent, their function being performed by an irregular protrusive membranous organ. Pupa. — The pupae differ from those of Limnol)iidac in minor characters only,- the principal distinction between them and pupae of the genus Lhnnohia and several other genera consisting in the form of the thoracic respiratory organs, those of Tipulidae, exclusive of Ctenophorinae, being long and slender, while those of the other family are stout and resemble a chitinized flattened plate. Manv of the Limnobiidae, however, have slender thoracic respiratory organs, and other characters must l)e depended upon to distinguish them from Tipulidae. A brief summarv of the characters of tip.ulid pupae is as ii'llows: head without projecting chitinized armature; antennae 193 never swollen at bases ; thoracic respiratory organs slender, of mod- erate length, sometimes slightly swollen at apices — except in Cteno- phorinae, in which they are heavily chitinized, flattened, and highly glossy; legs exceeding the wings in length. Abdomen with i or 2 transverse series of short protuberances on each segment except basal and apical. Palpi recurved at apices. Imago. — See synopsis of families. HABITS OF LARVAE Most of the larvae are scavengers, feeding on decaying vegeta- tion, in mud containing vegetable debris, or in rotten wood. Occa- sionally some species of Tipula cause injury to crops such as oats and hay, or to pastures, by feeding upon the roots of the growing plants. Many of the species are aquatic or semiaquatic, living among floating vegetable matter along the margins of ponds or streams. The food consists of algae and various kinds of vegetable matter. HABITS OF IMAGINES The imagines occur commonly in damp situations, especially where there is a rank growth of vegetation. They feed upon nectar of flowers and upon moisture on vegetation and on the ground. Many of the species are readily attracted to lights. Keys to Subfamilies LARVAE 1. Mandibles with 2 teeth ; antennae about twice as long as broad ..... Ctenophorinae. — Mandibles with more than 3 teeth; antennae about 4 times as long as broad Tipulinae. PUPAE 1. Thoracic respiratory organs broad, glossy, and heavily chitinized (PI. XXXII, Fig. 20) Ctenophorinae. — Thoracic respiratory organs slender, opaque, not more heavily chitinized than dorsum of thorax Tipulinae. IMAGINES 1. Legs very long and slender, the tarsi especially so ; anterior branch of second vein absent, indistinguishable, or perpendicular DOLICHOPEZINAE. 194 — Legs long but not very slender ; anterior branch of second vein present, oblique 2 2. Antennae of male pectinate or subpectinate Ctenophorinae. — Antennae not pectinate Tipulinae. Subfamily CTENOPHORINAE I have before me only a part of a larval exuvium — consisting of the head — and the pupal exuvium of one species of this subfamily, and am consequently unable to give a detailed description of the im- mature stages. The specimens, however, present characters that serve to distinguish at least this species from other Tipulidae known to me. As the characters of the species I have may not be in agree- ment with those of other genera in the subfamily, unknown to me, I shall restrict my generalizations to the genus to which it belongs. XiPHURA Brulle GENERIC CHARACTERS Larva. — Head large, heavily chitinized, arcuate on dorsum. Antennae short and stout, longer than maxillary palpi. Mandibles very stout, without teeth along the lower margin. Labial plate heav- ily chitinized, similar in general form to that of Tipula. Maxillae well developed, the palpi short and stout. Hypopharynx similar in form to that of Tipula. Structure of body not known to me. The parts of the exuvium that remain show that there are numerous long surface hairs present as in Tipula, and the anal spiracles are large, slightly elevated, pale brown, with the central opening darker. Pupa. — Head without protuberances between antennae ; bases of antennae slightly swollen; labium rather prominently protruded. Thoracic respiratory organs heavily chitinized, glossy, irregularly and coarsely wrinkled; apices of tarsi except those of the fore pair ex- tending much beyond apices of wings. Abdomen with a single en- circling series of short broad processes on each segment, the apices of which are acutely pointed. HABITS OF EARVAE All of the known larvae of this subfamily live in much-decayed trees, but whether they feed upon the dead wood or some vegetable growth it contains, or upon insect larvae is unrecorded. 195 HABITS OF IMAGINES The imagines feed upon nectar and sap. They are usually rare, and normally occur in proximity to a suitable larval habitat. XiPHURA FUMiPENNis Ostcu Sacken Ctenophora fumipennis Osten Sacken, Proe. Ent. Soc. Phila., 1864, p. 47. Larva. — Head black, heavily chitinized, antenna short and stout, with a very short apical appendage (PI. XXXII, Fig. 2i) ; mandibles very robust, with 2 strong apical teeth (Fig. 25) ; labium dentate along its anterior margin, the central tooth simple; hypopharynx as in Figure 23, Plate XXXII, the anterior margin transverse. Pupa (PI. XXXII, Fig. 24). — Length, 22 mm. Dark brown, slightly shining. Thoracic respiratory organs glossy black. Abdom- inal spines dark castaneous at bases, becoming pale at apices. Head without protuberances, the organs in the same positions and of the same form as in Tipula; antenna extending beyond base of wing. Thoracic respiratory organs robust, about twice as long as their greatest width, heavily chitinized, their margins irregular (Fig. 20). Abdomen with strong leaf-like process at apices of segments, the tips of which are very acute ; lateral margin of segments with the same armature as in Tipula, one simple process before spiracles and another, bifid, behind them; spiracles small but distinct; apical segment elongated, its apical half consisting of 2 long upper, and 2 shorter lower, processes. The foregoing descriptions were made from the larA^al and pupal exuvia of a female specimen reared by Dr. H. Glasgow, June 8. 19 10. The larva was found in a much-decayed chestnut log in the Auger- ville woods, Urbana, 111. Subfamily TIPULINAE SUBFAMILY CHARACTERS Larva. — Head heavily chitinized ; antennae longer than maxil- lary palpi; mandibles stout; labium well developed, usually dentate; hypopharynx large. Body without surface pilosity or with very short and dense pile; arrangement of bristles as shown in Figure i, Plate XXVIII. Apical segment with 6 finger-like processes, sometimes of very unequal length ; spiracles large. Pupa. — Head usually with 2 small membranous protuberances above bases of antennae; antennae not swollen at base, and often with 196 a small thorn ; palpi recurved at apices. Thorax with or without short wart-like protuberances on dorsum ; respiratory organs long and slen- der; halteres visible above upper margin of wing and resembling in certain respects the posterior wings in some lepidopterous pupae ; legs greatly exceeding length of wings. Abdomen with i, or 2, trans- verse series of thorns on ventral segments beyond apices of legs, the anterior series, if both are present, much weaker than the posterior ; dorsal segments usually with a postmarginal series of thorns. HABITS OF LARVAE As far as known the larvae are scavengers, feeding upon decay- ing vegetable matter. Many species are aquatic or subaquatic. HABITS OF IMAGINES The imagines fly most readily in the late afternoon. The species I have observed in nature are flower frequenters. TiPULA Linne and Pachyrrhina Macquart As very few of the species before me have been reared, and are represented onlv by larvae or pupae or, at most, by both, there is at present no possibility of specifically identifying these immature stages. Neither is it possible for me to cite characters for the separation of the larvae of Tipula and Pachyrrhina, the reared material at hand being quite insufficient to justify any attempt at a generalization. I give a synopsis of the characters that appear to me to be of primary importance in the separation of the forms I have studied, but, un- fortunately, I can specifically identify only a very few of them, and as in my opinion detailed descriptions would occupy more space than their possible scientific value will warrant me in taking, only the notes and synopsis are presented. I leave this subfamily in this condition, however, with the hope that the work now being done bv C. P. Alex- ander on the biology of the crane-flies will satisfactorily fill the very large gap in our knowledge of the early stages of the group. A complete study of our material is not at present contemplated, the forms described in this paper being included merely as indices to the range of specific distinctions, and as adjuncts to the synoptic char- acters cited in the key to the families. Keys to Species LARVAE 1. Apical abdominal ventral segment with slender protrusive blood- gills 2 197 — Apical abdominal ventral i^egmcnl with blood-^ills in the form of an irregular protuberance, rarely acute laterally 9 2. Ventral blood-gills not conspicuously longer than the stellate proc- esses on margin of stigmatal field, these processes slender, regu- larly and very conspicuously fringed Tipula sp. 1. — Ventral blood-gills conspicuously longer than stellate processes, or the latter not regularly and conspicuously fringed, or the proc- esses very unequal in length 3 3. Stellate processes subequal in length, very short, their margins with regular fringe of short hairs 5 — Stellate processes very unequal in length, the upper two short, the lateral and lower ones much longer and with isolated groups of long hairs 4 4. Very large species, more than 40 mm. in length ; penultimate abdom- inal segment with a lateral process Tipula sp. 2. — Small species, 20 mm. in length ; penultimate abdominal segment wdthout a lateral process Tipula sp. 2a. 5. Small species, not more than 30 mm. in length 6 — Larger species, more than 40 mm. in length 7 6. Body with dense pubescence which is most conspicuous on 2 narrow longitudinal lines on dorsum, giving the species the appearance of being vittate ; posterior latero-ventral bristles surrounded with stiff upright hairs Tipula eluta. — Body without -dense pubescent vittae; posterior latero-ventral bristles not surrounded with stiff upright hsdrs . Tipula cu7ictans. 7. Body almost regularly cj'lindrical, as in eluta ; posterior latero-ven- tral bristles not on pseudopod-like elevations, their bases sur- rounded with short, stiff, upright hairs ; dorsum not conspicuously vittate Tipula sp. 4. — Body not regularly cylindrical, the segmentation deep, and the pos- terior segments with distinct pseudopod-like elevations ; poste- rior latero-ventral bristles situated on elevations and not sur- rounded with stiff upright hairs; dorsum conspicuously vittate 8 8. The pale dorsal vittae composed of a number of closely placed pale dots; dorsum with many small pale paired spots; hairs not in- serted in dark brown dots Tipula ahdominalis. — The pale dorsal vittae linear ; dorsum with a number of very incon- spicuous pale dots ; hairs inserted in dark brown dots. Tipula sp. 3. 9. Apical abdominal segment -wdth a conical protuberance on each side proximad of the processes on the margin of stigmatal field (PI. XXXI, Figs. 6, 7) Tipula sp. 6. — Apical abdominal segment without a protuberance on sides 10 10. Upper 4 processes on anal segment decurved, long, and pointed, much longer than lower pair, ventral respiratory organs acute laterally (PI. XXXII, Figs. 7, 8) ; prothorax with 2 horny protuberances on its dorsal margin anteriorly Tipula sp. 5. 198 — Upper 4 processes on anal segment not decurved, usually straight; ventral respiratory organ not acute laterally ; prothorax without horny protuberances on its dorsal margin anteriorly 11 11. Upper median pair of processes on apical abdominal segment very acutely pointed, their posterior face glossy brown, lower pair very small and widely separated ; hypopharynx with upper plate more acute than usual (PL XXXI, Fig. 14) Tipula sp. 7. — Upper median pair of processes on apical abdominal segment not acutely pointed, lower pair of moderate size and rather close to- gether 12 12. Upper plate of hypopharynx Avith acute teeth, its anterior outline distinctly convex (PI. XXXII, Fig. 13) Tipnla hicornis. — Upper plate of hypopharynx with rounded teeth, its anterior out- line almost transverse (PI. XXXII. Fig. 15) Pacliyrrliina ferrughiea. PUPAE 1. Thoracic respiratory organs very long and slender, one much longer than the other, the longest one at least half as long as entire body Tipula sp. 1. — Thoracic respiratory organs equal in length, not more than one fourth as long as entire body 2 2. Ventral abdominal segments beyond apices of tarsi with a median and apical transverse series of spines 3 — Ventral abdominal segments with only the apical series of spines. 4 3. Antennae with a distinct but short thorn on outer side at base ; f rons slightly furcate and covered with irregular small wart-like pro- tuberances between bases of antennae Tipula cunctans. — Antennae without thorn at base; frons not as above. . .Tipula eluta. 4. No minute thorns or bristles laterad of the 2 strong thorns on third and fourth ventral abdominal segments ; protuberance between base of antennae and base of thoracic respiratory organs small, forming a distinctly isolated wart-like prominence Tipula trivittata. — One or two closely paired small thorns or 2 slender bristles laterad of the 2 strong thorns on third and fourth ventral abdominal seg- ments ; space between base of antennae and base of respiratory organs filled with a regularly rounded prominence 5 5. Thoracic respiratory organs short, not extending more than half- way from their bases to medio-dorsal protuberances ; 2 weak bris- tles laterad of the ventral series of thorns on abdominal segments ; apical segment of female much elongated (PI. XXVIII, Fig. 8) Tipula serta? — Thoracic respiratory organs long, extending almost or quite to medio-dorsal protuberances; one or two small thorns laterad of 199 the series of thorns on ventral abdominal segments ; apical seg- ment of female not elongate 6 6. Apical abdominal segment drawn out into a long process consisting of an upper and a lower pair of tube-like organs (PI. XXXII, Fig. 18) Tipula sp. 7. — Apical abdominal segment short and stout, not drawn out into a long process 7 7. Apical segment with 4 small but distinct lobes at tip (PI. XXVIII, Fig. 6) PacJiyrrliina ferruginea. — Apical segment with 2 rather large lobes at tip (PI. XXVIII, Fig. 7) Tipula bicornis. Tipula sp. i Larz'a. — Length, 20-22 mm. Dark brown, with an indistinct pale central vitta and slightly paler along sides. Antennae about 4 times as long as their basal width, dis- tinctly tapering apically; apical joint very small; frontal plate with a rather conspicuous tuft of hairs near outer anterior angle; labrum not conspicuously hairy ; mandibles as in Figure 2, Plate XXXII ; hypopharynx more elongate than in other species examined, its an- terior margin wdth 3 rather large teeth in an almost transverse series and a much smaller one at angles (PI. XXXII, Fig. 1 ); labium as in Figure 3, Plate XXXII. Body without surface pilosity ; bristles very weak though long; latero-ventral bristle on posterior portion of each segment usually duplicated, sometimes triplicated; outer bristle of the transverse series on posterior portion of each of the dorsal segments duplicated; lateral bristles weaker than the dorsal and ven- tral series; apical segment as in Figure 8, Plate XXXI. Pupa (PI. XXVIII, Fig. 14).— Length of body, 15 mm.: that of longest respiratory organ, 9 mm. Dark brown, the lateral longi- tudinal elevation along the spiracular region pale ; abdomen with dor- sum indistinctly, and venter distinctly, bivittate. Thoracic respiratory organs very slender, unequal in length, their apices flattened and split longitudinally; apices of fore tarsi falling- short of apex of first abdominal segment beyond apices of wings, mid tarsi extending slightly beyond the apex of that segment, hind tarsi extending to middle of next segment. Exposed ventral abdominal segments, except apical, each with 2 series of thorns on posterior division, the anterior consisting of 2, widely separated, and the posterior of 4 to 14; no thorns on area covered by legs, and the next 2 series slightly interrupted at middle ; apical segment with 4 long up- 200 wardly curved spine-like processes which are armed at apices with several short thorns. The foregoing descriptions are made from a larva and pupa bearing the Laboratory accession number 26281, obtained by Dr. S. A. Forbes in Delavan Lake, Wis., May 25, 1892, taken in an inlet among weeds at the surface, and one larva, accession number 26282, taken hy the same collector at the same place May 26, 1892. The species is undoubtedly truly aquatic, judging from the struc- ture of the apical segment. No means is at hand for associating the early stages with any described imago. T1PUI.A sp. 2 Larva (PI. XXVIII, Fig. 2). — Length, 45-55 mm. Brown, without well-defined vittae (alcoholic specimens). General structure as in above-cited figure. The principal differ- ences in head structure between this species and the preceding lie in the shape of the labium and the hypopharynx, the former ( PI. XXXII, Fig. 5) having a strongly produced central tooth and no distinct laterals, while the latter (PI. XXXI, Fig. 13) has the anterior margin with only 2 weak protuberances and no rounded teeth; frontal plate as in Figure 9, Plate XXXI. Body without close pubescence, the armature as shown in Figure 2, Plate XXVIII; abdominal segments with the posterior lateral bristles on both dorsum and venter situated on pseudopod-like elevations; apical segment as in Figure 3, Plate XXVIIL The foregoing description was made from alcoholic specimens in the Laboratory collection bearing the following data: Accession number 25756, Urbana, May 7, 1888, taken in woods (C. A. Hart) ; and two examples submitted by J. A. Hyslop (ace. 6687) from Hagerstown, Md. TiPUivA sp. 2a This specimen may really be a young example of the foregoing, as it differs only in size (20 mm.) and in having a process on each side of the penultimate abdominal segment. Locality, Blacktail Deer Creek, Yellowstone National Park, August 28, 1890; taken in an aquatic collection among vegetable debris (S. A. Forbes). TiPULA ABDOMINAUS Say Ctenophora abdominalis Say, Jour. Acad. Nat. Sci. Phila., Vol. 3, p. 18. (1823) Tipula abdominalis Say, Needham, Bull. 47, N. Y. State Mus., p. 575. (1901) 201 A larva which was assumed to be of this species was described by Needham, as cited above. The markings given in that description appear to justify me in considering the species following as distinct from abdominalis. TiPUivA sp. 3 Larva. — Length when full-grown, 50 mm. Dark greenish brown, with 2 continuous moderately broad longitudinal vittae on dorsum, a dark brown median vitta, and a number of isolated pale dots; surface hairs set in small blackish or dark brown dots. Head rather small in comparison with size of larva, lateral view as in Figure 4, Plate XXXII ; antennae normal in size ; labial plate (PI. XXXII, Fig. 6) with one very large central tooth, with more or less distinct shoulders, and 3 smaller lateral teeth; hypopharynx (PL XXXI, Fig, 12) with anterior margin of upper plate slightly convex, the central tooth of the five the largest. Body similar in general structure to that of Species 2, but the apical segment very different, quite closely resembling that of Species 4 except that the ventral blood-gills are distinctly shorter and stouter. Described from specimens obtained by Dr. S. A. Forbes in Black- tail Deer Creek, August 28, 1890, and in Slough Creek August 30, 1891 — both in Yellowstone National Park. TiPULA sp. 4 Larva. — Agrees in general appearance and armature of the abdominal segments with eluta, but differs in being much larger (50 mm.) and in having the apical abdominal radiating processes much less acute (PI. XXXI, Fig. 5). The head agrees closely with that of Species 3, the labium (PI. XXXII, Fig. 9) and hypopharynx (PI. XXXI, Fig. 15) being of the same general structure, differing only in having the former narrower and more acute anteriorly ; f ronto- clypeal region as in Figure 4, Plate XXXI ; mandibles as in Figure 26, Plate XXXII. Specimens are in our collection from Ithaca, N. Y. (March 21. 1897). TiPULA sp. 5 Larva (PI. XXIX, Fig. 3). — Length, 27 mm. Separable from other species that do not have the long slender blood-gills on apical abdominal segment by a pair of slight elevations or tubercles on the dorsum of the anterior third of the prothoracic segment, and by the peculiar formation of the apical segment, shown in Figures 7 and 8, 202 Plate XXXII. The hypopharynx and labium agree with those of Pachyrrhina ferruginca. I have before me the specimen from which Mr. Hart drew up his description of his Species (&) in the paper on the "Entomology of the Illinois River and Adjacent Waters"*, and a number of speci- mens sent in by a farmer October 7, 191 5, from an alfalfa field at Towanda, 111. T1PUI.A sp. 6 This is the Species (a) described by C. A. Hart in his paper on Illinois River species. The apical segment differs from that of any allied species known to me, and this character alone should enable one to identify it. As Hart did not figure this species I have pre- pared drawings of the apical segment which are presented herewith (PI. XXXI, Figs. 6, 7). In addition to the specimen previously recorded from Havana, 111., I have before me one taken in a sandy swamp at Grand Crossing, 111., Nov. 7, 1891 (C. A. Hart). TiPuivA sp. 7 Larva. — Length, 25-30 mm. Brown, apical segment yellowish white on the posterior surface, the 4 upper radiating processes con- spicuously blackened posteriorly, or on what is their inner or under surface when incurved, the lower pair with a black spot near apex which, because of the processes being normally curved upward, as in Figure I, Plate XXXI, is not usually visible. Head of normal size and shape, the dorsal and ventral surfaces as shown in Figures 2 and 3, Plate XXXI ; hypopharynx as in Figures 1 1 and 14 of the same plate ; mandibles as in Figure 27, Plate XXXII. Body with weak pilosity, the arrangement of bristles nnd the general structure as in Pachyrrhina ferruginea; apical segment differing as stated in key. Pupa. — Length, 28-30 mm. Brown, slightly shining. Base of antennae with a sharp thorn-like process on anterior side ; a small rounded tubercle above and behind base of antennae. Thoracic respiratory organ about 10 times as long as its greatest width; post- spiracular and medio-dorsal thoracic protuberances large, the latter sharp and not bifiid, posterior protuberance small; legs ending just before apex of third abdominal segment, the apices of fore tarsi fall- *Bull. 111. State Lab. Nat. Hist., Vol. 4, Art. VI, p. 217. 203 ing short of apices of mid pair. Two thorns on each side of third segment in Hne with apices of tarsi ; 6 thorns on other segments, the lateral one of each series weakest; apical segment as in Figure i8, Plate XXXII. This is the species described by C. A. Hart as Species (e) m the paper previously referred to. It occurs in humid earth and especially under logs or leaves. In addition to Hart's material I have before me two specimens from St. Clair Co., 111. (Nov. 26, 1886). TiPULA ELUTA Loew Tiptila eluta Loew, Bull. Ent. Zeitschr., 1863, p. 290. The external characters of this species have been very fully de- scribed by C. A. Hart in his paper on the "Entomology of the Illinois River and Adjacent Waters"*. The details given here are merely supplementary. Larva (PI. XXIX, Fig. i). — Antennae about 4 times as long as basal width, slightly tapering apically ; maxillary palpi longer than broad ; labial plate with a large rounded central tooth, and 2 much smaller lateral teeth and 2 poorly developed protuberances on each side; hypopharynx similar to that of Species 4 (PI. XXXI, Fig. 15). Lateral abdominal bristles as shown in Figure i, Plate XXVIII. Pupa (PI. XXIX, Fig. 2). — Thorax with the postspiracular, medio-dorsal, and postero-dorsal protuberances small but distinct. Legs in female extending to apex of first abdominal segment beyond apices of wings, in male to middle of the next segment. Armature of ventral abdominal segments similar to that of Species i except that the second visible segment has the widely separated anterior pair of thorns reduced to mere hairs, and the other segments have these same thorns simple, with a small hair at base instead of 2 thorns as in that species; apical segment of sexes as in Figures 11, and 13, Plate XXVIII. These descriptions and the figures are made from specimens used by Mr. Hart as a basis for his descriptions of cluta in the paper above referred to. The specimens were obtained from the Illinois River at and near Havana, 111. The larva is usually found burrow- ing in the sand on the shore, but occasionally is found in the water. I have taken the larvae from wet mud and sand along the margin of a small stream at Muncie, 111., and very probably the species occurs in similar situations throue^hout the state. *Bull. 111. state Lab. Nat. Hist., Vol. 4, Art VI, pp. 212-214. (1895) 204 TiPUIvA CUNCTANS Say Tipula cunctarvs Say, Jour. Acad. Nat. Sci. Phila., Vol. 3, p. 23. (1823) Larva. — Similar to the previous species in general appearance and structure, the principal distinctions being as follows : the f ronto- clypeal region is less hairy, the plates on either side having only an isolated tuft of hairs at outer anterior angle and one bristle and a short thorn on inner anterior protuberant area; the labial plate (PI. XXXII, Fig. lo) has 3 large teeth on each side of the central one ; anterior outline of hypopharynx as in Figure 1 1 of same plate ; the body is not densely pubescent, and shows no vittae on dorsum; and the short, stiff, upright hairs that surround the bristles in eluta are wanting. Pupa. — Length, 18-22 mm. Reddish brown, lateral margins yellowish. Slenderer than cliita. Head rather distinctly protuberant be- tween antennae, the surface with coarse rugae and a few small warts; bases of antennae with a small sharp process on anterior surface. Area between base of antennae and respiratory organs regularly rounded; dorsum of thorax without the distinct protuberances so noticeable in eluta ; apices of fore, mid, and hind tarsi not in a transverse line, but ea,ch successively farther beyond apices of wings. All of the exposed abdominal segments with 2 series of spines each, the anterior one con- sisting of 2, widely separated, and the posterior series of 4-12, close- ly placed; the third and fourth segments have the anterior pair of spines reduced to mere hairs ; dorsal segments with only the posterior armature distinct; apical segment of female as in Figure 19, Plate XXXII, that of male as in Figure 22. Described from examples in the collection obtained at Newton and Hillsboro, 111., April 1896. The species is very common through- out the state and, like 7\ hicornis, is sometimes destructive in meadows, pastures, and other grass lands, the larvae feeding on the roots of the plants. This species is usually found in quite different surroundings from those of eluta, being essentially terrestrial in habit and often found in fields which are rather dry and well removed from any body of water. Tipula trivittata Say Tipula trivittata Say, Jour. Acad. Nat. Sci. Phila., Vol. 3, p. 26. (1823) Pupa. — Length, 22-25 ^^- Dark brown, lateral line pale. Bases of antennae with a sharp protuberance on anterior side ; a 205 small but distinct elevated tubercle immediately behind antennal base. The usual 6 small elevations on dorsum of thorax, the medio-dorsal pair duplicated transversely; wings extending to apex of second ab- dominal segment; legs extending beyond apex of third, terminating in a straight transverse line. All dorsal segments except apical one with a transverse series of thorns on or near posterior margin, the series becoming successively stronger from basal segment to apical ; post- notum with 4 very small protuberances; third and fourth ventral abdominal segments each with 2 widely separated thorns, fifth to eighth inclusive with 4 each; prespiracular and postspiracular spines of equal strength, both simple; apical segments of male and female as in Figures 16, and 17, Plate XXXII. The foregoing description was made from exuvia supplied by Dr. E. P. Felt and J. A. Hyslop, the former obtaining his specimens at Albany, N. Y., May 5, 1909, and the latter obtaining his at Wolf- ville, Md., May 21, 1915. The species is represented in the Labora- tory collections by two imagines from Algonquin, 111. TiPULA SERTA Loew? Tipula serta Loew, Bed. Ent. Zeitsckr., 1863, p. 283. Pupa. — Length, 30 mm. Yellowish brown. Robust. Head with a pair of small rounded elevations above and between bases of an- tennae; antennae without a distinct basal process. Thoracic respira- tory organs less than 6 times as long as their greatest width ; post- spiracular protuberances in the form of rounded elevations ; medio- dorsal pair carinated, simple; tarsi terminating in a straight trans- verse line at apex of third abdominal segment. Armature of ventral abdominal segments confined to posterior margins, that of third and fourth segments consisting of 2 thorns, that of remaining segments of 4; laterad of the thorns on each segment are 2 slender bristles; apical segment much elongated (PI. XXVIII, Fig. 8). The foregoing description was made from the pupal exuvium of a female that was reared from a larva found by the writer under a log at White Heath, 111., in March, 1916 (Ace. No. 46302). Owing to the great uncertainty in identifying species of this ge- nus from descriptions, the above name should be regarded as a tenta- tive one. Tipula bicornis Forbes Tipula hicornis Forbes, 16th Eep. State Ent. 111., p. 78. (1888) This species was originally described by Dr. Forbes, as above 206 cited, the description of the imago being given on page 80. The species stands in the Loew collection at Cambridge, Mass., under the name bicornis Loew, but never was described by Loew. Mr. Hart, in his "Entomology of the Illinois River and Ad- jacent Waters" unfortunately entered in his key the larva of cnnctans as that of bicornis. The larva of bicornis is almost inseparable by superficial charac- ters from that of Pachyrrhina ferruginea, but possibly vA'ith better and more material than I have, such separation may be feasible. In examples which I have dissected I find that the labial plate and hypopharynx furnish characters that appear to be of specific value. These differences are shown in Figures 12 and 13, and 14 and 15, Plate XXXII. The pupae of the two species are also very similar in structure, and I find in the shape of the apical segment of the females the only appreciable distinction. This difference is illustrated by Figures 5 and 7, Plate XXVIII. The apical segment of the male pupa may pro- vide characters for the separation of this sex also; I liave no male pupa of bicornis for examination. The species is very widely distributed in this state, and is some- times destructive to pastures, the larvae feeding on the roots of the grasses. Pachyrrhina ferruginka Fabricius Tipula ferruginea Fabricius, Sept. Antl., Species 28. (1805) The larva (PI. XXVIII, Fig. 4) of this species bears a striking resemblance to T. bicornis and is found in the same situations. I have no well-preserved specimens of authenticated bicornis for com- parison of the external characters with those of ferruginea. The frontal plate is as in Figure 10, Plate XXXI, and dissection of the head shows the distinctions mentioned under bicornis (see PI. XXXII, Figs. 14, 15). Apical abdominal segment as in Figure 3, Plate XXX. The pupa is also very similar to that of bicornis, the only charac- ters that appear to be useful in their separation being found in the apical abdominal segment-as stated in key. (See PI. XXVIII, Figs. 5, 10, 12.) The species is common and widely distributed in Illinois. Principal Papers on North American Tipulidae Hart, C. A. '95. On the entomology of the Illinois River and adjacent waters. Bull. 111. State Lab." Nat. Hist., Vol. 4, Art. VI. 207 Needham, J. G. '07. Report of the entomologic field station conducted at Old Forge, N. Y., in the summer of 1905. Bull. N. Y. State Mus., No. 124 : 199-248. Needham, J. G., and Betten, C. '01. Aquatic insects in the Adirondacks. Bull. 47, N. Y. State Mus. Osten Sacken, C. R. '69. The North American Tipulidae. Monographs of North Ameri- can Diptera, Part 4. (A list of additions and corrections appears at end of Part 3, which was published after Part 4.) Family LIMNOBIWAB This family is of much greater extent than Tipulidae and contains a much larger number of genera, none of which in their larval and pupal stages — judging from the data at hand — show the same uni- formity that is found in the genera Tipula and Pachyrrhina. It is difficult to separate the larvae and pupae of the two families, but I believe that the following summary of characters will serve this pur- pose. FAMII^Y CHARACTERS Larva. — Head in all subfamilies but Trichocerinae, Hexatominae, and Eriopterinae very similar to that of Tipulidae except that the antennae are much more slender, and frecjuently they are shorter than the maxillary palpi. The labial plate is often divided longitudinally in the center, each part being furnished with distinct teeth, while in all Tipulidae known to me the labial plate is entire and subtriangular, with a single apical tooth and usually several laterals. In Hexatominae and some Eriopterinae that I have examined the labium is not chitin- ized, and posteriorly the head is composed of 4 or 6 slender chitinized rods connected by weakly chitinized membrane. In Trichocerinae the head is complete, and the prothoracic spiracles are present. The mandibles in the species with tipulid-like head are much more slender than in Tipulidae, and in the latter when the apical segment has no protuberances the mandibles are very stout and have but 2 teeth, both at apex. The apical segment in Limnobiidae is very differently con- structed in the different genera, but as far as I have seen there are never 6 processes, which in Tipulidae is the almost invariable number. Pupa. — The pupae of all genera of this family known to me may be readily separated from those of Tipulidae by the straight palpi, since those organs in Tipulidae have their apices recurved. 208 HABITS OF I,ARVAI5 The larvae have more diversified food-habits and habitats than do those of TipuHdae. A summary of these is given under the differ- ent genera dealt with in the text. HABITS OF IMAGINES The food of the imagines, when any is taken, usually consists of nectar. Keys to Subfamilies LARVAE 1. Thorax and abdomen with numerous long fleshy appendages , Cylindrotominae (p. 210) . — Thorax and abdomen without long fleshy appendages 2 2. Labium not ehitinized ; mandibles long and sickle-shaped, toothed only on basal half; maxillae with a very long membranous lobe at anterior lateral angle; apical segment with 4 processes, which are fringed with long hairs; penultimate segment in preserved larvae much distended Hexatominae (p. 232). — Labium heavilj' ehitinized except in Eriopterinae ; mandibles stout, toothed on apical half ; and otherwise not as above 3 3. Apical abdominal segment terminating in a pair of long tail-like ap- pendages, the spiracles at their bases, above; labial plate divided in center Pediciinae (p. 216). — Apical abdominal segment terminating in 4 or 5 short protuber- ances, or at least not with 2 long terminal appendages 4 4. Apical abdominal segment Avith 5 short terminal protuberances, the central one on upper margin about as large as the others ; body usually pubescent or roughened; head posteriorly consisting of 6 slender rods Eriopterinae (p. 227). — Apical abdominal segment with 4 terminal protuberances or with- out any, or if there are 5 the central one on upper margin is much smaller than the others and the head is not as above 5 5. Head complete ; prothoracic spiracles present Trichocerinae (p. 234) . - — Head incomplete posteriorly; prothoracic spiracles absent 6 6. Body green, with dense groups of fuscous hairs on dorsum of seg- ments, which give it the appearance of being marked with black LiMNOBIINAE, pt. (p. 212). — Body yellowish, whitish, or brownish, without distinct groups of hairs as above 7 7. Segments broader than long, lateral margins of prothorax with 1 strong hair, those of the other segments with 2 such hairs RHAMPHIDnNAE (p. 226). 209 — Segments at least as long as broad, lateral margins without strong hairs 8 8. Body covered with short decumbent hairs, or, if bare, without nar- row stripe-like locomotor areas ; apical segment with 4 or 5 processes Limnophilinae (p. 220) . — Body without decumbent hairs ; locomotor organs consisting of nar- row transverse strips on dorsum and venter which are slightly leathery and armed with short spinules LiMNOBIINAE, pt. (p. 212) . PUPAE 1. Thoracic respiratory organs very short and stout, not more than twice as long as their greatest breadth ; armature of abdominal segments weak, the same on dorsum and venter, consisting of a single narrow, chitinized or setigerous, band 2 — Thoracic respiratory organs very much elongated, usually more than 6 times as long as their greatest breadth or they are knobbed at apices ; or abdominal armature usually strong, thorn-like or spinose, or not in the form of bands 3 2. Thoracic respiratory organs well separated basally LiMNOBIINAE, (p. 212). — Thoracic respiratory organs subcontiguous basally Rhamphidiinae (p. 226) . 3. A pair of large leaf-like projections above bases of antennae Hexatominae (p. 232) . — No leaf-like projections above bases of antennae 4 4. Abdominal segments each with 2 or more very long thorn-like pro- jections on posterior margin, which are in some species armed with small branches; rarely these projections are confined to apical 2 segments Cylindrotominae (p. 210) . — Abdominal segments each with small spines, or if rather large pro- jections are present they are short and leaf-like, rather numerous, and unbranched 5 5. Thoracic respiratory organs stout, knobbed at apices Pedichnae (p. 216). — Thoracic respiratory organs slender, not knobbed at apices 6 6. Thoracic respiratory organs not slender and tube-like, flattened and but little elevated Eriopterinae, pt. (p. 227). — Thoracic respiratory organs long and usually tube-like, much ele- vated 7 7. Abdomen without dorsal or ventral armature; thoracic respiratory organs not longer than width of thorax, slender and tube-like .... Eriopterinae, pt. (p. 227). — Abdomen either with distinct armature on apices of dorsal segments or on both dorsal and ventral ones and with the thoracic respira- 210 tory organs longer than width of thorax ; or if the armature of ab- domen is very weak the thoracic respiratory organs are not tube- like but acute apically Limnophilinae* (p. 220). IMAGINES 1. Only one submarginal wing-cell present 2 — Two submarginal wing-cells present 4 2. Antennae with 14 segments Limnobiinae. — Antennae with 16 segments 3 3. Tibial spurs present Cylindrotominae. — Tibial spurs absent Rhamphidiinae 4. Tibial spurs absent Eriopterinae. — Tibial spurs present 5 5. Subcostal cross-vein proximad of base of second vein . . . Pediciinae. — Subcostal cross- vein distad of base of second vein 6 6. Antennae with 16 or more segments 7 — Antennae with at most 10 segments Hexatominae. 7. Seventh vein short, abruptly deflected towards anal angle Trichocerinae. — Seventh vein normal, not deflected towards anal angle Limnophilinae Subfamily CYLINDROTOMINAE This subfamily includes but four genera, each containing from one to four or five species, and because of the peculiar anatomical characters of the different stages it may yet be considered as entitled to separate family rank. I have treated the subfamily in this paper as belonging to Limnobiidae. SUBFAMII^Y CHARACTERS Larva. — Head similar to that of Tipulidae, the dorsal surface arcuate, heavily chitinized, and with 2 slits which usually extend proxi- mad of the middle. Antennae slender, longer than maxillary palpi. Mandibles stout, with a strong apical tooth, and several smaller teeth along the lower margin. Labial plate similar to that of Tipulidae, the central tooth simple or bifid. Thoracic and abdominal segments with long fleshy spine-like processes which may be either simple or more or less furcate. Apical segment with 4 slender processes; spiracles rather small. Pupa. — Head without projections; antennae curved over eyes. Thorax very short ; respiratory organs long and slender ; wn'ngs ex- *I do not know what characters may serve to separate pupae of Limnophilinae an