PLEASE HANDLE WITH CARE University of Connecticut Libraries =1153 D1E2M77L Digitized by the Internet Archive in 2011 with funding from LYRASIS members and Sloan Foundation http://www.archive.org/details/biologyofbirchleOOfrie Bulletin 288 June, 1927 V7 AM (ttmutrrttrut Agricultural iExurrtmeut sytatum Npw Hftatrnt, (£imti« rtintt C- o N [M 's ^3 The Biology of THE BIRCH LEAF SKELETONIZER Bucculatrix canadensisella, Chambers Roger B. Friend CONTENTS Page Page Introduction 395 Food Plants 445 History 395 Factors Affecting Abundance 448 Systematic Position 309 Geographical Distribution . . . 455 Laboratory Methods 399 Effect of Temperature on De- Morphology 400 velopment 458 Life History and Habits 424 Control 482 Determination of the Number Summary 482 of Instars 442 Bibliography 483 The Bulletins of this Station are mailed free to citizens of Connecticut who apply for them, and to other applicants as far as the editions permit. CONNECTICUT AGRICULTURAL EXPERIMENT STATION OFFICERS AND STAFF as of June 1927 BOARD OF CONTROL His Excellency, John H. Trumbull, ex-officio, President Charles R. Treat, Vice-President Orange George A. Hopson, Secretary Mount Carmel Wm. L. Slate, Director and Treasurer New Haven Joseph W. Alsop Avon Elijah Rogers Southington Edward C. Schneider Middletown Francis F. Lincoln ■ Cheshire STAFF. E. H. Jenkins, Ph.D., Director Emeritus. Wm. L. Slate, B.Sc, Director and Treasurer. Miss L. M- Brautlecht, Bookkeeper and Librarian. Miss J. V. Berger, Stenographer and Bookkeeper. Mrs. R. A. Hunter, Secretary. G. E. Graham, In charge of Buildings and Grounds. E. M. Bailey, Ph.D., Chemist in Charge. C. E. Shepard "J Owen L. Nolan L , _, Harry J. Fisher, A.B. f Assistant Chetntsts. W. T. Math is J Frank C. Sheldon, Laboratory Assistant. V. L. Churchill, Sampling Agent. Miss Mabel Bacon, Stenographer. T. B. Osborne, Ph.D., Chemist in Charge. H. B. Vickery, Ph.D., Biochemist. Miss Helen C. Cannon, B.S., Dietitian. G. P. Clinton, Sc.D., Botanist in Charge. E. M. Stoddard, B.S., Pomologist.. Miss Florence A. McCormick, Ph.D., Pathologist. George L. Zundel, M.S. A., Graduate Assistant. A. D. McDonnell, General Assistant. Mrs. W. W. Kelsey. Secretary. W. E. Britton, Ph.D., Entomologist in Charge; State Entomologist. B. H. Walden, B.Agr. ) M. P. Zappe, B.S. {■ Assistant Entomologists. Philip Garman, Ph.D. ) Roger B. Friend, Ph.D., Graduate Assistant. John T. Ashworth, Deputy in Charge of Gipsy Moth Work. R. C. Botsford, Deputy in Charge of Mosquito Elimination. Miss Grace A. Foote, B.A., Secretary, Walter O. Filley, Forester in Charge. H. W. Hicock, M.F., Assistant Forester. J. E. Riley, Jr., M.F., In charge of Blister Rust Control. H. J. Lutz, M.F., Assistant Forester on Purnell Project. Miss Pauline A. Merchant, Stenographer. Donald F. Jones, S.D., Geneticist in Charge. W. R. Singleton, S.M., Assistant Geneticist. H. R. Murray, B.S., Graduate Assistant. Administration. Chemistry. Analytical Laboratory. Biochemical Laboratory. Botany. Entomology. Forestry. Plant Breeding. Soil Research. Tobacco Sub-station at Windsor. M. F. Morgan, M.S., Investigator. H. G. M. Jacobson, M.S., Assistant. Evelyn M. Gray, Stenographer. Paul J. Anderson, Ph.D., Pathologist in Charge. N. T. Nelson, Ph.D., Assistant Physiologist. T. R. Swanback, B.S., Scientific Assistant. THE TUTTLE, MOREHOUSE * TAYLOR COMPANY The Biology of THE BIRCH LEAF SKELETONIZED Bucculatrix canadensisella, Chambers Roger B. Friend I. Introduction The biology of Bucculatrix canadensisella, or, as it is more commonly called, the birch leaf skeletonizer, is known to only a very slight extent. Not only does the insect have peculiar habits and a specific structure, but its great abundance during certain years, coupled with its habit of feeding on native birches, renders it of interest economically as well as biologically. In the follow- ing pages are the results of investigations, made during the years 1924, 1925, and 1926, into its habits, reactions, distribution, history, and morphology. The work is not complete, but it is intended that the gaps shall be filled, in part at least, in the future. I am indebted to Professor Alexander Petrunkevitch of Yale University and Dr. W. E. Britton of the Connecticut Agricultural Experiment Station for criticism of the work ; to Professor G. C. Crampton of the Massachusetts Agricultural College for assist- ance in certain details of the morphological part ; to Messrs. A. B. Gahan, R. A. Cushman, and C. F. W. Muesebeck of the United States Department of Agriculture for determining the species of parasites; to Dr. Annette F. Braun of the University of Cincin- nati for some notes on the geographical distribution ; to Mr. C. B. Hutchings of the Entomological Branch, Canada, for the use of an unpublished manuscript, and to Mr. B. H. Walden of the Connecticut Agricultural Experiment Station for the photographic work. II. History The earliest reference to the genus Bucculatrix is found in the first volume of de Geer's "Memoires," in which is given the life history of a "little caterpillar with sixteen legs, smooth, green, which feeds on the lower side of the leaves of Frangula." It was the manner in which this caterpillar spun its cocoon which attracted the attention of de Geer, as the following extract from his * This paper is a dissertation presented in partial fulfillment of the requirements for the degree of Doctor of Philosophy at Yale University. 39 6 CONNECTICUT EXPERIMENT STATION BULLETIN 288 "Memoires" shows: "Ouand elles sont parvenues a leur juste gran- deur, ce qui. arrive dans le mois susdit, elles filent contre les feuilles memes de tres-jolies petites coques alongees. qui meritent extremement d'etre connues, a cause de leur figure particuliere. Ce sont ces coques qui m'ont determine a donner l'histoire de ces Chenilles." He gives a detailed description of the manner in which the cocoon is woven, and also gives brief attention to the pupal and adult stages. There is a plate of illustrations of the larva, the structure of the cocoon, the adult, and the injury to the plant. The species described was Bucculatrix frangulella and the host plant, Rhamnns frangula, the buckthorn. In 1832 de Haan published a posthumous volume of Lyonet's works in which there is a description of a "chenille extremement petite, mais qui emploie une adresse inconcevable a se filer une coque cannelee." This description formed part of a letter from Lyonet to Reaumur written December 22, 1744, and was later sent to the president of the Royal Society of London to be pub- lished if the society saw fit to do so. Most of the description is devoted to the details of the structure and weaving of the cocoon. In his illustrations Lyonet figures the larva, cocoon and its struc- ture, and adult. The larvae were found by Lyonet on the leaves of the oak. This species was Bucculatrix ulmella Mann (Zeller). The history of the genus up to 1862 is given by Stainton in his "Natural History of the Tineina." Linnaeus and Fabricius neg- lected it entirely, and in 1783 Goeze, in his "Entomologische Beitrage," gave the name Tinea frangulella to de Geer's species. Neither de Geer nor Lyonet gave names to the species they described. Retzius, writing contemporaneously with Goeze, and Villers six years later, both gave different names to the Tinea frangulella of Goeze. The next person after de Geer to describe a species of this genus was Haworth, who in 1829 in "Lepidoptera Britannica" described Tinea cuculipenella with the varieties beta, gamma, and delta. Stainton notes that although Haworth's descriptions are very vague, beta was probably Bucculatrix boyer- ella, gamma, B. crataegi, and delta, B. ulmella (Lyonet's species). Three years later, in 1832, appeared the posthumous volume of Lyonet's works, in which is described what proved to be Buccu- latrix ulmella, as mentioned above. In 1834 Stephens translated Haworth's description of Tinea cuculipenella without mentioning the varieties gamma and delta. In 1833 Treitschke had rede- scribed de Geer's species as Elachista rhamnifoliella, and a new species, Elachista gnaphaliclla. In 1838 Duponchel figured four species in his "Lepidopteres de France," in the genus Elachista, namely, E. boyerella, E. rhamnifoliella, E. gnaphaliella, and E. hippocastanella. In 1839 Zeller, in "Isis," placed the following species in section A of his genus Eyonetia: L. rhamnifoliella (giv- ing reference to de Geer), L. albedinella (boyerella of Duponchel), BIOLOGY OF BIRCH LEAF SKELETONIZER 397 L. hippocastani (hippocastanella of Dtiponchel), L. cristatella, L. higricomella, L. cidarclla, and L. crataegi. In 1848 Zeller established the genus Bucculatrix with nine spe- cies, the descriptions appearing in "Linnaea Entomologica," volume III. The nine species, with the authors credited by Zeller, were : Bucculatrix cidarclla Tischer ulmella Mann 3. crataegi Zeller 4. boyerella Duponchel 5. gnaphaliclla Treitschke 6. frangulella Goeze hippocastanella Duponchel nigricomella Zeller cristatella F. R. The species gnaphaliella had been previously (1839) placed by Zeller in Litliocollctis. He included Bucculatrix in a group of leaf-mining moths possessing eye-caps. Much of the history of the genus from Zeller on does not concern us here and will be omitted. Stainton, from whose work much of the above infor- mation has been derived, listed in 1862 nineteen species of Buc- culatrix of which he considered fourteen good and five doubtful. The fourteen were known in the larval form and their food plants were given. Twelve of the fourteen are described by Stainton very fully. This work covers practically all that was known of the genus up to the time of writing. For the earliest described American species we must turn to the writing's of Clemens, who. in the Proceedings of the Academy of Natural Sciences, Philadelphia, for i860, published the descrip- tions of four new species of Bucculatrix: B. coronatclla, B. pomi- foliella, B. agnella, and B. trifasciella. These descriptions were again published in 1872 in a posthumous volume of the writings of Clemens, edited by Stainton. This volume also includes a description of the genus by Clemens. Chambers, in the Canadian Entomologist, volume V, 1873, described and mentioned nine American species of this genus and stated these to be all the described American species known to him. These nine are: B. trifasciella Clemens, B. capitcalbclla n. sp., B. pomifoliella Clemens, B. obscurofasciella n. sp. (possibly synonymous with B. corona- tclla Clemens), B. luteella n. sp., B. agnella Clemens, B. packard- ella n. sp., B. coronatclla Clemens, B. thuiella Packard. Although Chambers considered his obscurofasciella possibly synonymous with coronatclla Clemens, Forbes (1923) gives trifasciella Clem- ens and obscurofasciella Chambers synonymous with packardella Chambers. It is not proposed to give a discussion of systematics and synonymy here, however. This briefly concludes the history 39& CONNECTICUT EXPERIMENT STATION BULLETIN 288 of the genus in America up to 1875, when the species canadensis- ella was described. In the Canadian Entomologist, volume VII, 1875, Chambers described Bucculatrix canadensis ell a, having received his speci- men from Canada. This description (see page 401) concerns the adult only and does not mention the larva nor the larval food plants. B. cidarella of Europe Chambers considered close to B. canadensisella, although quite distinct. The larva of the European species demaryella feeds on birch, but according to the descrip- tion given by Stainton (1862) it also is quite distinct from canadensisella. For twelve years after the description by Chambers there occurs no mention of the species, but in 1887 Lintner recorded the occur- rence of the insect in Monroe County, New York, where the larvae were very abundant on the leaves of Betula lutea during the fall of 1886. In 1890 Packard recorded what was in all prob- ability this species on the leaves of the white birch at Brunswick, Maine. Lintner again reported it from New York in 1893, this time as injurious to all the native birches in the region of Ausable Forks during September, 1891. The same year Fletcher stated that all the birches around Ottawa, especially Betula papyrifera, B. lutea, and B. alba (European white birch) were severely injured. From this time on the reports of the insect become more frequent and the injury caused by its larvae more noticed. Hutch- ings published a brief life history in the 56th Annual Report of the Entomological Society of Ontario (1926), and this treats of the insect more fully than any other publication to date. The species is of some economic importance, and most of the literature on it concerns the injury done to the birch trees. Systematically the genus has been neglected, and when men- tioned it is referred to as aberrant. Forbes (1923) published a key to the species found in northeastern United States with descriptions. For descriptions of species discovered in the pres- ent century in America the writings of E. Meyrick, A. F. Braun, and A. Busck should be consulted ; and for Old World species see the writings of E. Meyrick, especially his "Exotic Micro- lepidoptera." The history of the insect is interesting in view of the fact that at frequent intervals it appears in extraordinary numbers and severely attacks birches over wide areas. In 1886 Lintner found it abundant in Monroe County, New York, and in 1887 it was reported as abundant in Massachusetts. During the years 1890, 1891, and 1892 a serious outbreak occurred in Ontario, New York, and New England. In 1901. 1902, and 1903 it was again very abundant and severely attacked birches throughout this same area. In 1907 a small outbreak occurred on Staten Island, New York, and in 1910 the insect was abundant at Kinderhook, New York. BIOLOGY OF BIRCH LEAF SKELETONIZER 399 In 1909 and 1910 birches in Minnesota were extensively skeleton- ized, and the insect's depredations were severe in Ontario in 1910, 191 1, and 1912, and in New England in 1909, 1910, and 191 1, growing" less serious in 1912 and 1913. The third outbreak of this insect thus covered Ontario, Minnesota, and New England between 1909 and 19 12, with small outbreaks in New York in 1907 and 1910. In 1919 the larvae were again beginning to appear in large numbers. This year they were abundant in New Brunswick and were noticed in Connecticut. In 1920 birches were heavily skeletonized and defoliated in Ontario, Quebec, and New Brunswick. In 192 1 the infestation continued in these regions, and larvae were abundant in Minnesota and appeared commonly in Massachusetts. In 1922 the injury to birch trees was conspicuous over the Great Lakes region and in New Eng- land. This last outbreak began to subside in 1924, although the larvae were injurious in Quebec in 1925. Beginning about 1890 there have been four serious outbreaks of this insect, one about every ten years. Some of the possible reasons for this periodic abundance will be considered under the section dealing with pre- daceous and parasitic enemies. III. Systematic Position The genus Bucculatrix was placed by Zeller in a group of minute leaf -mining moths the adult antennae of which possessed eye-caps. Along with Bucculatrix were Lyonetia, Cemiostoma, Nepticula, etc. The first general treatise on Bucculatrix placed the genus in the Tineina (Stainton 1862). It is usually placed in the Lyonetiidae today and is so classified by Forbes (1923). There are, however, differences of opinion as to the classification of Lepidoptera and of this genus in particular. Thus Forbes places Bucculatrix in the family Lyonetiidae of the superfamily Tineoidea, but Mosher (1916) places it in the family Bucculatrig- idae of the superfamily Gracilarioidea, basing her decision on pupal characters; and Fracker (191 5) places it in the family Buc- culatrigidae of the Tineoidea. The grouping of families and genera in the Tineina is still apparently an open question. The genus will here be placed in the Lyonetiidae according to the clas- sification of Forbes and considered as slightly aberrant. For a taxonomic account of the genus and a key to the species of north- eastern United States the work of this author may be consulted. IV. Laboratory Methods The life history data were secured by rearing individual larvae in glass jars or vials, each receptacle containing wet sand and a fresh birch leaf. Observations were made daily. Adults for 400 CONNECTICUT EXPERIMENT STATION BULLETIN 288 oviposition records were caged over a birch twig in a celluloid cylinder with cloth ends. This permitted natural conditions of light and air. The leaves were examined daily with a glass and eggs were marked with a circle of black ink and numbered. Pupae were obtained by simply placing small pieces of heavy cardboard under the plant in a stock rearing cage. The larvae spun their cocoons on the under side of the cardboard. The pupae were kept in a box sunk in the ground until the early sum- mer. Just prior to the period of emergence they were placed singly or in groups of five in glass vials plugged with cotton or in large gelatin capsules, the ends of which were perforated. This made observations on the emergence of adults a simple matter. All life-history studies were made in an out-door insectary. For dissecting fresh material, it was found best to cover the chloro- formed specimen with a drop of thick shellac, add one drop of alcohol, allow to set a few minutes, and then immerse in saline solution. The shellac became pitchy and held the insect firmly, but at the same time it could be easily removed from the chitin. For studying the external morphology the insects were boiled in 10 per cent potassium hydroxide until clear and then stained in tetrabromfluorescic acid twenty-four hours. The chitinized plates stained deeply red, and the membranous cuticle a light pink. The body being clear, the internal skeletal structures were readily observed. For the temperature experiments the larvae were kept singly in glass vials or in test tubes, the receptacle in either case being plugged with cotton. The food material was kept fresh and unwilted. The individual insects in all cases were from miscella- neous field collections made in the vicinity of New Haven except where otherwise noted. V. Morphology The morphological descriptions will be confined to the external appearance of the various stages and certain important anatomical details of the exoskeleton. The genital organs of the adult will be briefly mentioned as they are of considerable interest morpho- logically and have more or less influence on the external form. The internal anatomy is not further described here, but it is intended that a description of the anatomy and histology will be produced later. The original description of the genus by Zeller (1848) is reprinted below. Bucculatrix Zell. Elachista Tr. Lyonetia ex p. Zell. "Caput lanatum, comosum. "Antennae breviusculae, conchula basali parvula instructae. "Palpi nulli ; os squamis epistomii tectum. BIOLOGY OF BIRCH LEAF SKELETONIZES. 4 01 "Alae anteriores caudulatae ; cellula discoidales acuta postice venulas 6 emittit ; vena subcostalis longissime interrupta ; subdorsalis simplex : "posteriores lanceolatae ; vena mediana in 3 ramos divisa, subdorsalis simplex. "Tibiae posticae pilosae. "Larva 16 pes supra epidermidem foliorum vivit; metamorphosis in folli- culo affixo subit." The presence of palpi will be brought out later, and the vein which Zeller calls "mediana" in the hind wing is designated in this paper the radius. The following is the original description of Bucculatrix cana- densis clla by Chambers (1875) : Bucculatrix canadcnsisclla n. sp. "The ornamentation of this species differs from that of any other yet found in this country, and though allied to B. cidarella of Europe, it is still quite distinct. "Head white. Tuft tipped with dark reddish brown, and the face faintly tinged with purplish fuscous. Upper surface of the thorax brown margined all around with white. Base of the fore wings white, followed by an oblique brown fascia, which is nearest the base on the costal margin, and is followed by an oblique white fascia ; all of these are placed before the middle and are followed by a large brown patch which occupies the entire wing to the ciliae, except that it contains a white spot on the middle of the costal margin. The brown patch is margined before on the dorsal margin of the wing by a small tuft of raised brown scales. At the beginning of the dorsal ciliae is a white spot placed a little before, but becomes almost con- fluent with a longer white costal streak. Behind these streaks to the apex the wing is pale brown, with a darker velvety brown apical spot. Ciliae pale yellowish, with a dark brown hinder marginal line before their middle not extending into the costal ciliae. Hind wings pale fuscous. At. ex. $i inch." A. Adult 1. External Appearance As both the above descriptions are rather brief, the external appearance of the adult is here given in a little more detail. By reference to plate XVII and text figure 12 the important mark- ings can be easily followed. Sexual differences are slight and will be referred to in the description. The general appearance of the adult in repose is shown in plate XVII. The head bears a dorsal tuft of rather long hair-like scales, the center of which is brown and the outer parts white. The "face" is covered with gray or brownish scales. When the insect is at rest the head is bent ventrally so that the labium touches the bases of the prothoracic coxae and the short tongue is curled and concealed between the latter. There are no maxillary palpi, and the labial palpi are very small and concealed beneath the head. The eyes are black and partly concealed by the scapes of the antennae which 402 CONNECTICUT EXPERIMENT STATION BULLETIN 21 are expanded to form eye-caps. These eye-caps are white, and from the anterior border of each there extends down in front of the eyes curved slender scales which give the insect the appear- ance of having "shaggy brows." The pedicel of the antenna is short, and the flagellum contains 29 segments, each of which bears two whorls of brownish scales. The proximal segments of the Fig. 12. Adult moth, enlarged about ten diameters. flagellum have no scales on the ventral side. This "nude" area is usually considered sensory. The first segment of the flagellum (third of antenna) is longer than any of the remaining segments. The antenna is about two-thirds the length of the body and fili- form. The ground color of the fore wings is brown, and although typi- cally reddish, it often varies to a yellowish. The wings are marked with transversely diagonal white bars as shown in figure 12. The basal bar is confluent with a white area on the meso- thorax. The second bar forms an angle with the apex directed distally. It sometimes extends completely across the wing and is often interrupted in the center by brown scales. The remaining bars do not cross the wing but terminate near the midline. There are two extending from the costal border and one from the anal border, all three directed slightly apically. Close to the tip of the wing is another white area whose exact size varies somewhat in different individuals. It extends from the costal to the distal border of the wing but does not include the apex, this latter being dark, almost sable, in color. There are two other prominent dark spots on the Aving, one at the anal angle and one at the distal mar- gin of the second transverse white bar. Both of these are always present, and sometimes there are other dark spots on the costal border. Beginning slightly distal from the middle of the costal border a row of gray cilia extends around the wing almost to the proximal end of the anal border. The tuft of raised brown scales on the anal border of the wing as described by Chambers is usually conspicuous. BIOLOGY OF BIRCH LEAF SKELETONIZER 403 The hind wings are gray and their borders are almost com- pletely ciliated. The superficial difference in shape between the fore and hind wings is due to the more extensive development of scales on the former. The scales on the hind wing are less numer- ous and do not project beyond the wing borders. Both the wings are really pointed. The dorsal side of the thorax is brown with white areas later- ally, these latter being confluent with the white basal areas on the wings. Each tegula bears a group of eight to ten bristle-like scales which extend along the costal border of the wing as far as the metathorax when the wings are folded. The pleural and sternal sides of the thorax are silvery-white. The coxae are large and of the same general color as the sternum of the thorax except that the lateral borders are brownish, particularly proxi- mally. The femora and tibiae are brown laterally and white medially, as is the first tarsal joint. The tarsal joints two, three, and four each have a white ring proximally and a brown ring dis- tally. The fifth tarsal joint is white, and its scales almost conceal the tarsal claws. At the posterior border of the mesothoracic tibiae at the distal end is a pair of spurs, and a pair of similar spurs is found at each end of the metathoracic tibiae. There is a pair of spines at the distal end of each of the first four tarsal joints. A row of thickly set long hairs is found on the anterior and posterior border of the metathoracic tibiae. The abdomen is covered with silvery-white scales ventrally and brown scales dorsally. The males have seven segments superfi- cially distinct on the ventral side, the second to the eighth inclu- sive, and the scales from the eighth practically cover the genitalia. The female has six segments superficially distinct ventrally, the second to the seventh inclusive. Scales from the seventh segment conceal the border between the seventh and eighth, and scales from the latter cover the remainder of the abdomen, giving the appearance of one broad segment. The ninth segment in the female is partly retracted within the eighth, and the tip of the ninth projects very slightly beyond the scales of the latter. The terminal fringe of scales on the male abdomen flares slightly but never does so on the female. The female abdomen is slightly larger than the male. On the dorsal side of the abdomen of each sex there are distinctly demarcated eight segments, the first to eighth inclusive. The body length averages about three millimeters and the alar expanse seven millimeters. The sexes are of equal size. 2. Head (Text figure 13) The head is somewhat compressed anterior-posteriorly, and the occipital surface is flat. The antennae are filiform and composed 4°4 CONNECTICUT EXPERIMENT STATION BULLETIN 288 of 31 joints, of which the first or scape is expanded to form the eye-cap. The second joint or pedicel is short and subspherical. The third joint (first of the flagellum) is half again as long- as any of those following'. The length of the antennae compared Fig. 13. Head and base of antenna of adult, ant f, antennal fossa; at, anterior arm of tentorium; bt, base of tentorium, cl, f ronto-clypeus ; dt, dorsal arm of tentorium; epicr, epicranium; gn, gena; gul, gular region; li, labium ; Ip, labial palpus ; m.v, maxilla ; mxp, maxillary palpus ; occip, occiput; ocf, occipital foramen; pg, postgena ; pil, pilif er ; prob, proboscis; pt, posterior arm of tentorium ; vx, vertex. The abbreviations underlined in the figure are on the posterior surface of the head. with the body length is shown in figure 12. The eyes are black and weakly spherical. There are no ocelli. Between the antennal fossae (ant f ) and connecting them is the suture which separates the f ronto-clypeus (cl) from the epicranium (epicr). A suture running between the eyes and through the epicranium divides off the vertex (vx) anteriorly. The vertex bears the forward-project- BIOLOGY OF BIRCH LEAF SKELETONIZER 405 ing hairs of the dorsal tuft and is divided by a median light suture. The posterior part of the epicranium is likewise divided by a median suture and bears the upward- and backward-projecting hairs of the dorsal tuft. The occiput (occip) lies between the epicra- nium and occipital foramen and is not sharply demarcated from the postgenae laterally. The fronto-clypeus appears to extend laterally to the eyes. The labrum is not present as a distinct sclerite and is represented by a pair of pilifers (pil) placed one above each maxilla. There are no mandibles. The proboscis (prob) is reduced, being about the length of the head. Each half of the proboscis (the galea) bears on its anterior surface a row of eleven papillate projections which appear pentagonal in cross section and each of which terminates in a short peg. The particular function of these was not ascertained. Near the base of each half of the proboscis and also on the anterior surface are three or four setae. Near the base of each maxilla and projecting from the lateral side is a small protuberance (mxp) which may represent the rudiment of the maxillary palpus. The bases of the maxillae (rax) are, as usual with Lepidoptera, firmly fixed in the ventral (posterior in this case) side of the head. The labium (li) is a small triangular sclerite, with a forward-point- ing apex, on the ventral side of the head and lies between the maxillae. It bears a pair of one-jointed palpi (lp). Between the labium and the occipital foramen (ocf) lies a gular region (gul) which is bounded laterally by the maxillae. Its separation from the labium is indistinct. The postgenae are separated dor- sally from the genae (gn) by the suture which divides the epi- cranium and ventrally by the sutures connecting the maxillae with the lower border of the eyes. The genal regions are not distinctly separated from the fronto-clypeus. The tentorium is similar to that of other Lepidoptera. The body of the tentorium (bt) separates the occipital foramen into a dorsal and ventral part. The anterior arms (at) come for- ward from the body and then turn ventrally to terminate at the ventro-lateral angles of the fronto-clypeus. The dorsal arms (dt) extend up from the body to the occiput, bordering the foramen laterally. The posterior arms (pt) extend down each side of the krwer part of the foramen to the maxillae. In the figure of the head all the abbreviations of the parts on the posterior (mor- phologically ventral) side of the head are underlined. 3. Cervical Region (Text figure 14) The head is supported by a pair of laterally placed cervical sclerites (cerv) which extend from the prothorax. At the cephalic end they meet the body of the tentorium, and at the posterior end they articulate with the episterna and then curve medially to meet in the midline. 406 CONNECTICUT EXPERIMENT STATION BULLETIN 288 4. Thorax The three thoracic segments are distinct, although the prothorax is much reduced. The mesothorax is the most developed, due to the development of the fore wings and the powers of flight. In the following description the nomenclature of Crampton (1909) has been adhered to as far as possible. Fig. 14. Prothorax of adult, anterior (left) and lateral (right) aspects. bSj basi-sternum ; cerv, cervical sclerite ; cxi, coxa ; epsi, episternum ; ft, furca; lat, precoxal bridge; pat, patagium ; prn, pronotum; spst, spini- sternum ; tri, trochantin. a. Prothorax (Text figures 14 and 15) The tergal region of the prothorax consists of a central triangu- lar pronotum (prn) and two laterally placed and conspicuous patagia (pat). The apex of the pronotum meets the prescutum of the mesothorax in the midline. The pleural region contains one narrow sclerite, the episternum, (eps) which meets the pronotum above and the coxa (ex) below. Anteriorly it supports the cervical sclerites and meets the precoxal bridge (lat) of the sternum. There is a very minute sclerite, the trochantin (tr), at the articulation of the coxa. The epi- meron is obsolete. From the posterior border of the episternum the pleural apodeme extends into the body cavity and meets the arms of the furca (f x ). Of the sternal sclerites the basi-sternum (bs x ) is the larger and extends laterally in the precoxal bridge to Fig. 15. Thorax of adult, dorsal aspect, a, b, c, axillary sclerites; anterior wing process ; cost, costal sclerite ; cu, cubitus ; ex, coxa ; epm, epimeron ; /, furca ; fren, frenulum ; vise, median area of scutum ; pat, patagium ; pph, postphragma ; prn, pronotum ; psc, prescutum ; pstit, pos- terior chitinous plate on metathorax ; r, radius ; teg, tegula ; tg pi, tegular plate ; wp, wing process ; 1, 2, axillary areas ; 2nd a, 3rd a, anal veins. The inferior numbers indicate the thoracic segment to which the part belongs. 408 CONNECTICUT EXPERIMENT STATION BULLETIN 288 meet the episterna. The median part of the basi-sternum is folded and lightly chitinized, extending slightly down the coxa on each side. The f urea-sternum and basi-sternum are not dis- tinctly separated. Caudally from the furca-sternum the narrow spini-sternum (sps^) extends along the ventral midline to meet the presternal sclerite of the mesothorax. What is here called the basi-sternum is apparently the eusternum of other authors. The furca extends up into the body cavity, and from its base a pair of lateral arms extend to meet the base of the episternum on each side. The prothoracic coxa is not divided into eucoxa and meron, although a' faint "suture" extends some distance up from the trochanter along the caudo-lateral surface. The coxa articulates freely with the episternum and trochantin. The trochanter and femur are of the usual type. The tibia bears no spurs. The coxa, femur, and tibia are approximately equal in length, and the tarsus is slightly longer. The basi-tarsus is almost as long as the second and third tarsal joints combined. There are five tarsal joints and each of the first four bears a pair of short spines on the posterior side of the distal end. The tarsus terminates in a pair of claws. b. Mesothorax (Text figures 15 and 16) The tergal part of the mesothorax consists mainly 01 two sclerites, the large scutum (sc 2 ) and the smaller somewhat rhom- boidal scutellum (scl,). The anterior part of the scutum is rolled in, and from its margin is given off the narrow prescutum (psc 2 ). This has been considered a true sclerite by some authors, but according to Crampton it is simply a phragma. At its mid- point the pronotum is attached. At the anterior lateral angles of the scutum are the large tegulae (teg) and beneath each tegula is a smaller tegular plate (tg pi). On each lateral side of the scutum are two pointed wing processes (wp 2 ) which project slightly and which continue medially along the under side of the scutum as chitinous thickenings. These processes help support the wing. The scutellum bears along its anterior and lateral margins a phragma which projects slightly into the body cavity. The posterior lateral angles of the scutellum continue out as the anal borders of the wings. From the posterior margin of the scutellum the large postphragma (pph) projects ventro-caudally into the body cavity. This phragma is made up of a layer from the mesoscutellum and one from the metathorax. The layers are easily separated. There is no true postnotum (or pseudonotum) present as a distinct sclerite. The curved process (pwp 2 ) which supports the anal area of the wing extends out from the lateral BIOLOGY OF BIRCH LEAF SKELETON1ZER 4°9 angles of the scutellum. This is called the posterior wing process by Snodgrass (1909). The plenron of the mesothorax is largely made np of two sclerites, the epimeron (epm 2 ) and episternmn (eps 2 ) separated by the vertical pleural suture. The pleural apodeme extends into the body cavity from this suture. The episternmn is divided into a dorsal anepisternum (aneps,) and a ventral katepisternum (kepSo) separated by a triangular middle area. At its anterior margin the anepisternum rolls in medially. From the dorsal margin of the sclerite the alar process (alp) projects upward and supports the wing, and the tegular arm (tega) extends anteriorly to the anterior lateral angle of the scutum where it abuts against the tegular plate. The tegular arm and alar process together with a ventral projection on the anepisternum appear to form a single anchor-shaped sclerite fused with the latter and separable from it with no great difficulty. The katepisternum meets the sternum ventrally. The epimeron is a single undivided sclerite somewhat membranous dorsally. It meets the posterior wing process and then arches over as a narrow arm to meet the arm of the furca (f 2 ). Just under the anal area of the wing and dorsal to the epimeron is the somewhat elongate costal sclerite (cost,). There is no distinct trochantin but it may be repre- sented by a triangular area just over the coxa. The pleural apodeme widens at this region and forms a support for the coxa. The anterior sclerite of the mesosternum (presternum, pst 2 ) projects forward from the basi-sternum (bs 2 ) to meet the poste- rior sclerite of the prothoracic sternum and extends slightly beyond it into the body cavity. This sclerite widens as it meets the basi-sternum. The latter is triangular, its apex being poste- rior, and is divided by a median longitudinal suture. From this suture and extending into the body cavity is a median chitinous blade (mbl). Posterior to the basi-sternum is the f urea-sternum f fs 2 ) which extends down the medial side of each coxa as a pedal region (pdr 2 ) and holds the coxa rigidly to the body. The furca (f 2 ) arises from the f urea-sternum and sends from its base a short curved process (fpr 2 ) into the body cavity anteriorly. The arms of the furca meet the arms of the epimera dorsally. The latero-sternites extend from the basi-sternum laterally to the pleural suture. The mesothoracic coxa (cx 2 ) is divided into an anterior eucoxa (eucx 2 ) and a posterior meron (mer 2 ) by a vertical suture on the outer side. On the .medial surface of the coxa lies a heavily chitinized angular plate (cs 2 ) which meets the pedal region of the furca-sternum. The leg articulates at the trochanter, the coxa being immovable. The tibia bears at its distal end on the posterior side a pair of spurs of which the outer is longer. The tarsus is 41° CONNECTICUT EXPERIMENT STATION BULLETIN 288 similar to that of the prothoracic leg. The mesothoracic leg is slightly longer than that of the prothorax. The wing venation (figure 17) is much reduced. The subcosta and costa are probably represented by the single costal vein. The radius (r) is rather faint at the base and gives off five branches distally. The median vein has disappeared except for the branches m 1 and m 2 . The cubitus (cu) is single. There is a faint fold (ista) which may represent the first anal vein. The second (2d a) and third (3d a) anals are distinct. There is some variation in the origin of r 4 , as it sometimes branches off distally to the position shown in the figure. The costal vein bears a retinaculum (ret) for the frenulum. The veins named above are according to Forbes (1923). The axial sclerites of the wing are as shown on the right side of figure 15. The sclerites a, b, c, 3, and the small sclerite between 1 and 3 are hard chitinous plates, but those marked 1 and 2 are thickenings of the wing similar to veins. The alar process of the pleuron abuts on 2, as does the anterior of the scutal wing-proc- esses. The posterior of the two scutal wing-processes abuts on 3, and the posterior wing-process supports a. The anal area of the wing folds along the outer border of b. c. Metathorax (Text figures 15 and 16) The scutum (sc 3 ) of the metathorax is divided medially by a triangular area (msc 3 ). This does not appear to be a distinct sclerite but simply a more lightly chitinized region. The post- phragma of the mesoscutellum is attached to the anterior margin of the scutum, its line of attachment extending to the wing pro- cess (awp.,) at the anterior lateral angles. The scutellum (scl 3 ) is a band stretching across the base of the scutum and appears to overlap the latter, due to the presence of a phragma which pro- jects caudo-ventrally into the body cavity. From the posterior border of the scutellum a membrane drops ventrally to meet a chitinous arm which forms a bridge between the ends of the epimera. The center of this bridge bears a chitinous plate (psnt) to which the tergum of the first abdominal segment is attached. This represents the pseudonotum (Snodgrass), al- though much modified from a primitive condition. At its lateral angles the scutellum continues into a narrow posterior wing process (pwp a ) which supports the anal area of the wing. The pleuron of the metathorax resembles that of the meso- thorax. The trochantin area at the head of the coxa is more dis- tinct here, however. Dorsally the alar process (alp) continues directly with the pleural apodeme, and the anepisternum (aneps 3 ) bears another process which also supports the wing. The costal (cost 3 ) sclerite is prolonged anteriorly as a long arm. The BIOLOGY OF BIRCH LEAF SKELETONIZES. 411 epimeron (epm 3 ) extends further posteriorly than does the same sclerite in the mesothorax. The sternum of the metathorax differs markedly from that of the preceding" thoracic segment. From the central basi-sternum (bs 3 ) extend the narrow latero-sternites (not shown in the dia- gram). The basi-sternum extends caudo-ventrally to meet the Fig. 16. Meso- (right) and meta- (left) thorax, lateral aspect; alp, alar process; ancps, anepisternum, bs, basi-sternum; cs, median coxal support; encx, eucoxa; fs, f urea-sternum ; fpr, f ureal process; keps, katepisternum ; mbl, median blade ; mer, meron ; pdr, pedal region ; tega, tegular arm ; sc, scutum; scl, scutellum. For other abbreviations see figure 15. f urea-sternum (fs 3 ) at the coxal support. There is no pedal region of the furca-sternum, but the coxa is held rigidly by this sclerite plus the basi-sternum. The furca-sternum extends as a narrow arm dorsally and then divides into a furca (f 3 ). The anterior f ureal process (fpr 3 ) is very large and the furca is heavy. Dorso-laterally the arms of the furca meet those of the epimera. The meron of the coxa is much reduced and occupies a poste- rior-medial position, only the eucoxa being visible laterally. The tibia bears a pair of spurs on the posterior side of each extremity, and the outer spur of each pair is the longer. The leg is other- wise similar to that of the mesothorax. The wing's (figure 17) show greatly reduced venation. There are, besides the costal, three principal veins, the radius (r), cubitus (cu), and the second anal (2d a), the median being represented 412 CONNECTICUT EXPERIMENT STATION BULLETIN 21 by two branches only. The costal vein probably represents the combined costal and subcostal. The radius is single and from it there branch the two' divisions of the median (m 2 and m 3 ). The cubitus is single. The frenulum (fren) consists of two stout setae that are held in the retinaculum of the fore wine'. Fig. 17. Fore (above) and hind (below) wings; 1st a, 2nd a, 3rd a, anal veins ; at, cubitus ; fren, frenulum ; m, median ; i- 5 , radius ; ret, retinaculum. The axial sclerites of the hind wing are as shown in figure 15. An angular sclerite (a) in the anal region is pivoted on the pos- terior wing process. The anal region folds along the outer side of this sclerite. Two sclerites (b and c) lie between this and the anterior wing process. These three constitute the chitinous axial plates homologous with those of the fore wing. The areas marked 1 and 2 are thickenings of the wing similar to veins and are homologous to the same areas of the fore wing. The sclerite c may correspond to 3 of the fore wing, and the sclerite b to b and c of the fore wing. The alar processes of the pleuron abut on the area marked 2, and the subcostal area ( 1 ) meets the ante- rior wing process and the sclerite marked c. Snodgrass (1909) has described the typical arrangement of the axial sclerites in the wing, but the tracing of these in the wing here described is uncertain, due to the difference in arrangement, and hence the letters and numbers as given here do not cor- respond to those of the above author. 5. Abdomen (Text figures 18 to 20) The abdomen has nine visible segments in the female and ten in the male, although in the latter sex the tenth is reduced to the socii. The apical segments in each sex are modified to form the BIOLOGY OF BIRCH LEAF SKELETONIZER 4'3 external genital apparatus. The first segment has a strongly chitinized tergum, probably a development in accordance with its function of supporting' the abdomen on the thorax. The sternum of the first segment is indistinguishably fused with that of the second and both are quite membranous. The identification of two sterna is furnished by the presence of two spiracles on each side. Fig. 18. "Alluring" organ on abdomen of adult male, expanded above, retracted below. In the female the segments from two to seven inclusive are of the usual unmodified type, but in the male the second segment shows a peculiar sexual dimorphism. On the caudal margin of the tergum of this segment is located a protrusible organ which, for want of a better name, has been termed an alluring gland. Similar organs called alluring glands have been described as occurring on other parts of male Lepidoptera, and until a histo- logical and cytological investigation is made of this particular case, the common term will be used in describing it. In other species of Lepidoptera there is considerable evidence that these organs give off a distinct odor when protruded, but the alluring function of these in a sexual sense is not definitely proved. This "gland" is shown in figure 18 protruded (above) and retracted within the abdomen (below) . When retracted it folds in an eversible sac, and when protruded the entire organ, including the sac, projects out from the body, looking for all the world like a composite flower. 414 CONNECTICUT EXPERIMENT STATION BULLETIN 288 The scales composing it are of two kinds, some pointed and some lobular. This organ is found in all males and never in the females. The remainder of the male abdomen up to and includ- ing the eighth segment is in no wise unusual. In the female the eighth and ninth segments (figure 19) are modified somewhat. On the sternum of the eighth segment is a slight protuberance Fig. 19. Tip of abdomen of adult female, bd, opening of bursa duct. Fig. 20. External genitalia of adult male, ae, aedoeagus ; an, anus ; lip, harpe ; so, socius ; vin, vinculum. which marks the copulatory opening leading into the bursa duct and thence to the bursa copulatrix. At the end of the ninth seg- ment is the external opening of the vagina (ventrally) and the alimentary tract (dorsally). Petersen (1900) has described in some detail the female and male genital organs of Lepidoptera and shows the transitional stages from the type having one genital opening (at the tip of the abdomen) to that having two as here found. The presence of two genital openings is forecast in the pupa. On each side of the copulatory opening is found a tuft of orange-colored scales, and a third tuft is found on the dorsum at the anterior margin of the eighth segment. These three tufts are normally concealed from view under the posterior margin of the seventh segment. The lateral apodemes from the ninth pro- ject back into the eighth segment. The tip of the female abdomen is usually telescoped so that the eighth segment is partly retracted within the seventh, and the ninth is retracted within the eighth. The posterior part of the eighth is membranous. In the illustra- tion the abdomen is shown with these segments drawn out into view. The tenth segment is not developed. BIOLOGY OF BIRCH LEAF SKELETONIZER 415 In the male (figure 20) the ninth and tenth segments are much modified and are usually retracted within the eighth. As shown in the figure, they are drawn out to expose the external genitalia. The nomenclature given is according to Eyer (1924). The ter- gum of the ninth forms a "roof" over the anus (an) which lies just beneath it. It is called the tegumen. Attached to its distal end are the socii (so) which really belong to the tenth segment and form the anal armature. They are paired and bear many short spines and setae. The sternum of the ninth segment con- sists of a narrow chitinous band, the vinculum (vin), which is fused with the tergum on each side. The paired claspers, called harpes (hp), articulate with the vinculum and are appendages of the ninth segment. They also bear many setae and short spines. The cone-like chitinous organ through which the aedoeagus (ae) projects is called the anellus and also probably belongs to the ninth segment as do the rest of the genitalia. The aedoeagus is a heav- ily chitinized tube supported by the anellus and tapering to a point distally. The penis itself is a soft eversible tube contained within the aedoeagus and is protruded from the ventral side of the tip of the latter. There are on each of the segments one to seven inclusive a pair of spiracles, and visible through the ventral wall of the abdomen are the four pigmented abdominal ganglia of the nerve cord. The ganglia of the entire ventral nerve cord of all stages of this insect are deeply pigmented and usually visible externally. In the adult the appendages conceal all but the abdominal, and these are found at the second segment and at the junctions of the third and fourth, fourth and fifth, and fifth and sixth respectively. The last is larger than the others, being a compound ganglion. The third and fourth abdominal ganglia are often contiguous and sometimes are fused to some extent. 6. Genital organs and alimentary tract (Text figures 21 and 22) The internal genital organs of the male and female are diagram- matically illustrated in figures 22 and 21. In the female the bursa copulatrix (bur) is by far the most conspicuous of these organs, and it occupies much of the anterior part of the abdomen, lying in the region of the third segment. It is connected by a duct to the external opening in the sternum of the eighth segment, and from the dorsal side of this duct near its external end there arises the long slender seminal duct which permits the passage of sper- matozoa from the bursa into the oviduct and thence into the semi- nal receptacle (rec sem). The common oviduct divides into two ducts (ovid) from each of which are given off four ovarioles (ov), each of which terminates in a filament. The filaments on each side unite with each other. The ovarioles extend from the 416 CONNECTICUT EXPERIMENT STATION BULLETIN 2.1 oviduct along' each side of the bursa to its anterior end, curve dor- sally and posteriorly, then dorsally and anteriorly to a common point just above the bursa where the group from each side is attached by the filament tips to the dorsal wall of the abdomen. Fig. 21. Reproductive organs of female, bur, bursa copulatrix ; coll gl, colleterial glands ; ov, ovarioles ; ovid, oviduct ; rcc scm, seminal receptacle ; rcct, rectum ; sem duct, seminal duct. The alimentary tract passes ventrally and to the right of the bursa, curves dorsally to pass above the union of the oviducts, then goes over the common oviduct to the tip of the abdomen, the rectum lying above the vagina. The ovarioles are of the polytrophic type, that is, the nutritive cells alternate with the ova. The seminal receptacle is bilobed and is attached to the dorsal wall of the com- mon oviduct. The colleterial glands (coll gl) are paired and BIOLOGY OF BIRCH LEAF SKELETONlZER 4'7 are connected by a common duct to the dorsal wall of the vagina. They secrete the adhesive substance which attaches the egg to the leaf. In the illustration the genital organs are shown spread out and not in their normal positions. /ejac duct Fig. 22. Reproductive organs of male, ac gl, accessory glands ; aed, aedoeagus ; ejac duct, ejaculatory duct; test, testes; vas def, vas deferens; ves scm, seminal vesicle. In the males the testes (test) are united and enclosed in a com- mon scrotum. The vesicula seminales (ves sem) are paired and unite just under the testes. From the vesicula seminales, which are really enlargements of the vasa deferentia (vas def), the latter ducts pass to enlarged chambers which lead to the ejacu- latory duct (ejac duct) which in turn terminates in the aedoeagus (ae). The accessory glands (ac gl) which presumably secrete a substance which mixes with the spermatozoa, are paired and 418 CONNECTICUT EXPERIMENT STATION BULLETIN 288 are connected with the enlargements at the terminations of the vasa deferentia. In some insects these glands secrete a substance which forms the spermatheca (especially Orthoptera). They occur here attached to each other rather loosely. The alimentary tract begins anteriorly in a large muscular pharynx from which two long filamentous salivary glands are given off and extend through the thorax beside the slender oesoph- agus. The latter extends from the pharynx through the thorax to the mid-gut in the abdomen and gives off a slender tube which enlarges to form a rather large crop. The latter lies mainly in the abdomen over the mid-gut. In the female the mid-gut lies under the bursa and extends to a point between the common ovi- duct and the bursa. It is much greater in diameter than the pharynx. At the termination of the mid-gut the malpighian tubes are attached, one on each side. These are slender and each branches once. Of the two branches of each tube one divides once and the other remains single. This gives three branches of each tube which lie in the region of the ovarioles. The hind-gut extends from the origin of the tubules and enlarges at its posterior end to form the rectum which terminates in the anus. The crop and mid-gut were empty in all the specimens dissected, which adds evidence to the belief that the adults do not feed. In the male the alimentary tract passes to the right of the ejaculatory duct. B. The Egg (Text figure 23) When first laid the egg is a translucent white, but it becomes more opaque after a few days. In shape it is flatly ovoid, and the longest diameter is 0.25 mm. It is made to stick to the leaf by an adhesive substance which surrounds it on the leaf surface in a narrow encircling band. The hexagonal sculpturing of the surface is quite distinct, and the micropyle occurs close to one end. When it leaves the oviduct the egg is soft and flexible, and its shape changes from approximately spherical when it is placed on the leaf surface. The eggs are always laid singly and scattered over the leaf, as shown on plate XVII. C. Larva (Text figures 24 and 25) When first hatched the larva is minute (.35 mm. long), trans- lucent, apodous, and flattened, a typical leaf -mining type. When it leaves the mine at the close of the third instar it has assumed a cylindrical form, the head has shifted from its former horizontal plane to a plane nearly at right angles with the body, and all the legs are present and functional. It measures about 2.5 mm. in length. When fully grown (plate XVII) the larva is about 6.0 mm. long and yellowish green in color with the setae on white BIOLOGY OF BIRCH LEAF SKELETONIZER 419 tubercles. The prothoracic shield is not conspicuous. The head is brown and typically that of a leaf -eating- larva. There are the usual prolegs on abdominal segments, three, four, five, six and ten. .i;'A&?- '■■•■A; Fig. 23. Egg, much enlarged. Actual longest diameter .25 mm. Fig. 24. Head of larva of 5th instar, ventral aspect, ant, antenna ; mx, maxilla ; oc, ocellus. The radical change in structure which occurs at the second molt is due to the change in feeding habits and environment of the larva, for the third instar comes out of the mine to the leaf sur- face. This necessitates the acquisition of legs and the shifting of the plane of the head. The dorsal part of the head capsule is longer than the ventral, and this makes the posterior margin of the epicranium in the leaf-mining instars, where the head is hori- zontal, push back into the prothorax. Tragardh (1913) has described structural transitions in several leaf-miners which change their feeding habits. The third instar resembles the fourth and fifth in general, but the setal pattern is somewhat different and there are fewer crochets on the prolegs. The pro- leg's of the abdominal segments three to six inclusive have one transverse row containing two crochets, and the prolegs of the anal segment bear one crochet. The fourth instar is like the fifth except for size and no further mention need be made of it. The fifth instar is a typical cater- pillar with the mouth-parts well developed. These are shown in a ventral view in figure 24. The antennae are minute, and there are only five ocelli (oc) present on each side. They are arranged 420 CONNECTICUT EXPERIMENT STATION BULLETIN -21 A \ (;, N. q_ *= \ »«i fe < BIOLOGY OF BIRCH LEAF SKELETONIZER 42 I in a curved row whose concavity is ventral. The labium is drawn out into a spinneret through which the duct of the silk glands reaches the exterior. The maxillae bear on the inner surface a pair of curved chitinous hooks. Otherwise the mouth-parts are not unusual. On the dorsal surface the adfrontals extend back to the posterior margin of the epicranium and the frons is about one-third the length of the head capsule. The labrum is bilobed and the mandibles bear four "teeth." The prothoracic legs differ from those of the meso- and meta- thorax in that the terminal segment bears one claw only, whereas in the case of the latter two pairs of legs the claw is protected by a pair of terminal lappets. The prolegs of the abdominal seg- ments three to six inclusive bear on the planta two rows of crochets, three per row. The anal prolegs have a single crochet only. The setal pattern of the body of the larva is of some taxonomic importance and is shown in figure 25. In this description the nomenclature of Fracker, though cumbersome, has been followed. The diagrams are made so that the anterior margin of each seg- ment is to the left and the dorsal midline is at the top. Each dia- gram is that of the left half of each segment projected on a flat surface. Certain segments are alike, and these have been repre- sented by one diagram. The setae of the anal segment < do not conform to those of any of the others. D. Pupa (Text figures 26 to 29) The pupa is spindle-shaped, about three millimeters long and brown in color. Many of the adult structures are evident, and the head thorax and abdomen are distinct. In the description here given the parts, although often incompletely developed, are named in accordance with the corresponding parts of the adult. The vertex (vert) occupies most of the dorsal side of the head and is separated from the frons by the Y-shaped epicranial suture, the frons meeting the arms of the suture. The stem of the Y is indistinct. The frons (fr) extends caudally along the ventral side of the head from the epicranial suture to merge into the clypeal region, there being no demarcation between the two. The frons in figure 26 can be easily distinguished by the presence of the pointed cutting plate in its anterior part. This is the so-called "cocoon-breaker" with the aid of which the pupa emerges from the cocoon. The bases of the antennae (ant) are visible on the dorsal side of the head lateral to the epicranium. On the ventral side of the head and lateral to the frons are the eye-pieces (e). The clypeus bears the bilobed labrum, and on each side of the labrum is a small triangular mandibular sclerite. Neither labrum nor mandibles are found in the adult. The labrum bears a pair 422 CONNECTICUT EXPERIMENT STATION BULLETIN 21 of laterally placed setae. The maxillae are prominent and form the pair of medially placed appendages extending- caudally from the lab rum. Neither maxillary nor labial palpi are visible. The antennae extend caudally almost to the tips of the wings. Be- Fig. 26. Female pupa, ventral aspect, much enlarged, ant, an- tenna ; cl, clypeus ; ex, coxa ; e, eye ; fr, f rons ; max, maxilla ; sp, spiracle. Fig. 2J. Pupa, dorsal aspect, much enlarged. tween the antennae and maxillae lie the folded prothoracic and mesothoracic legs. The tarsal regions of the metathoracic legs are visible between the tips of the antennae, most of this pair of appendages being covered by those preceding. A small part of the metathoracic coxae is visible in the midline posterior to the mesothoracic coxae. The metathoracic legs extend slightly be- yond the tips of the wings. The fore wings extend to the seventh segment of the abdomen on the ventral side and conceal the hind wings. The appendages are loosely attached to each other BIOLOGY OF BIRCH LEAF SKELETONIZER 423 and are free from the body wall. They overlap, more or less, and the covered parts are quite membranous. When dissected out, the regions of the coxa, femur, tibia and tarsus are visible, although often not distinctly demarcated, and the tibial spurs are prominent. The two folds in the legs occur between coxa and femur and between femur and tibia. The tibia merges into the tarsus. Fig. 28. Tip of abdomen of male Fig. 20. Section of dorsum of pupa, ventral aspect. abdomen of pupa. The female has evidence of two genital openings on the ventral side, one on the posterior border of the eighth abdominal segment and one on the interior border of the ninth. These are short slits in the integument. The male has evidence of one genital opening only (figure 28), on the ninth segment. This condition in each sex corresponds to that of the adult. On the tip of the tenth abdominal segment is the indentation marking the anal open- ing, and on the lateral sides of this segment are a pair of short strong spines. On the dorsal side of the pupa the prothorax is constricted in the middle and widens laterally. It lies between the epicranium and mesothorax and abuts on the antennae. The epicranial suture ex- tends to the anterior margin of the prothorax. The mesothorax is a large quadrate sclerite separated by distinct sutures from the pro- thorax, metathorax, and fore wings. Along its midline it is raised into a very slight ridge. At the anterior-lateral angles are a pair of setae. The mesothoracic wings extend around the body to the ventral side. The metathorax is not so long as the mesothorax and merges indistinctly into the wings laterally. It bears also a pair of setae at the anterior-lateral angles, but these are more approximated than those of the mesothorax. The metathoracic wings are almost entirely concealed by those of the mesothorax, the bases only being visible. There are visible dorsally ten abdominal segments, of which numbers two to seven inclusive bear a pair of setae at the anterior-lateral angles, and numbers one to seven bear laterally placed spiracles. Segments four to seven inclusive bear also a pair of medially placed setae. The spiracles on the first abdominal segment are concealed by the hind 424 CONNECTICUT EXPERIMENT STATION BULLETIN 288 wing's. At the anterior margin of the tenth segment is a dorsal tubercle bearing a pair of spines. This and the lateral spines on the tenth segment are purely pupal structures. The dorsal sur- face of the abdominal segments is covered with minute spines (figure 29) and on segments two to seven inclusive there is a row of heavy spines along the anterior margin. In the male abdominal segments three to seven are movable, and in the female segments three to six are movable. The terminal segments are immovably united in both sexes. VI. Life History and Habits All the data here given, except for field records and other cases specifically mentioned, were obtained from records of individual insects reared in New Haven, Conn., on the gray birch, Betula populifolia. The field observations in Connecticut and Massa- chusetts are also of insects occurring on Betula populifolia unless otherwise stated. The period during which records were made covers the years 1924, 1925, and 1926. The first adults appear the last of June in the region about New Haven, and the last disappear the last part of July. In 1924 adults were fairly numerous July 7th, and the last were seen July 31st. During 1926 adults were systematically collected with a net and by hand during July in one locality, a group of birches just north of Mt. Carmel, near New Haven, and these collections' indicated a maximum number of adujts were present the fourth week in July. During' 1926 the season was later than usual. This species was abundant during the second and third weeks in July, but from the 27th to the 31st it declined in numbers from about a maximum to disappearance. Pupae were kept at normal temperatures in an out-door screened insectary during 1924, 1925, and 1926. Those of 1924 were collected in the field during the spring of that year; those of 1925 and 1926 were reared in the out-door insectary. Records were kept of the emergence of 104 adults in 1924, 135 in 1925, and 36 in 1926. In 1924 the period of emergence was between June 4th and July 9th (only three emerged before June 23d) ; in 1925 between June 15th and July 19th, and in 1926 between July' 2d and July 21st. The period of maximum emergence during 1924 was between June 25th and July 9th ; in 1925 between June 1 8th and July 10th, and in 1926 between July 2d and July 9th. In 1925 all records obtained after July 1st were of individuals taken from New Haven to Woods Hole, Mass. The early appearance of adults in 1925 may possibly have been due to high temperatures early in June of that year, for, during the first ten days of June, 1925, the mean hourly temperature was 75.2°F. as opposed to 6o.6°F. in 1926, and 6i.2°F. in 1924 (U. S. Weather Bureau, New Haven, Conn., statistics). BIOLOGY OF BIRCH LEAF SKELETONIZER 425 Hutchings (1926) reports that in Ontario the adults are found during - July (up to the 25th) with a maximum emergence from July 6th to July 14th. In New Brunswick, Gorham (1922) reports adults abundant the first two weeks of July. It would seem from these reports that the adult insect appears at approxi- mately the same time of year over much of its entire range, being at a maximum during the first part of July. Fig. 30. Cast pupal skin and cocoon. From the pupae collected in the field in the spring of 1926 there were secured 54 adults of which the sexes were determined. Thirty-three were females and 21 were males. Of 36 adults secured from laboratory-reared larvae, 19 were females and 17 were males. Although the females outnumber the males, the dif- ference is not sufficiently great to warrant an assumption that there is not approximately an equal number of each sex under natural conditions. When ready to emerge from the cocoon the insect is in the pre- imaginal stage, and the scales of the imago are easily seen through the pupal skin. The pre-imago works its way forward, probably with the assistance of the spines on the tenth abdominal segment, and breaks through the anterior end of the cocoon. The cutting plate on the vertex is of material aid in the process. When about three-fourths of the length is exposed, the body is held at an angle of about 40 from the cocoon. The pupal skin splits at the junc- tion of the vertex and the prothorax and also longitudinally through the prothorax and mesothorax. The eye-pieces remain attached to the antennae and mouth parts. Figure 30 shows a cocoon and an empty pupal skin in the characteristic position. 425 CONNECTICUT EXPERIMENT STATION BULLETIN 288 According to Snodgrass (1922), the emergence from the cocoon is greatly assisted in Bucculatrix pomifoliella Clemens by the presence of three "valves" in the anterior end of the cocoon, these acting as an inclined plane to lift the insect. In B.'cana- densisella these structures are not present. The inner part of the cocoon is woven so that its lateral walls curve in at the base, mak- ing a "round corner," and this might conceivably give the pre- imago an upward lift. However, if a pupa is pushed gently forward by hand, it does not emerge at the normal place, but lower down near the base of the cocoon. By simply rolling over in its attempt to extricate itself, the pre-imago would bring the cutting plate in a more dorsal position with respect to the cocoon and thus insure emergence at the normal location. During the daylight hours the adults rest quietly on the surfaces of the lower leaves of birches, or of the leaves of undergrowth under the birches, or on the lower part of the trunks of the trees. They fly very little unless disturbed, and usually remain motionless for long periods of time. The photograph on plate XVII is a three- minute exposure of a live adult which was carried to and from the camera in the position seen. It shows the adult in the char- acteristic position. They can be collected very easily by simply inverting an empty glass vial over.- the quiet insect and then gently tapping the opposite side of the leaf on which it rests. It will quickly fly or run up into the vial. I have collected one hundred in less than an hour from fern fronds in this manner. When dis- turbed, they fly quickly a few feet, very rarely over five or six, and usually much less. If confined in a bottle they run excitedly for a few minutes when disturbed. From field observations it seems that as a rule the moths remain near the ground during the day and go up into the trees at about dusk. I have often failed to get them in a net by sweeping birches over my head, when many were secured by sweeping within four feet of the ground. Yet, in the same location, by sweeping the birches at dusk, several were netted very quickly at a height of about nine feet from the ground. In 1926 in one particular locality, moths were very abundant during the day on the fronds of ferns growing under the birch trees, but none were seen on the leaves of the birches which were not lower than four feet from the ground. An investigation of the birches at night with the aid of an acetylene lamp showed that the moths were all up on the birch leaves, and none were on the ferns below. The next day over one hundred were easily caught on the ferns, but none were on the leaves of the trees. At times, however, I have found moths on the leaves of the trees during the day five feet from the ground. Gorham (1922) reports that in New Brunswick he found moths on the birches in large numbers at all hours of the day, but he gives no further information on the distribution. BIOLOGY OF BIRCH LEAF SKELETONIZER 427 The moths prefer situations out of direct sunlight, and this may account for their position during daylight hours. This nightly migration into the trees may be affected somewhat by conditions of light, moisture, and amount of undergrowth, but it apparently occurs to some extent wherever conditions are normal. The response of these moths to light is not very definite, due to their habit of remaining quiescent in one spot unless greatly dis- turbed. Attempts to make them show either a negative or posi- tive response to daylight in the laboratory gave inconclusive results. Several attempts to attract them to a lighted lantern, an acetylene lamp, and automobile headlights at night in the field failed completely, although they were present in considerable numbers on the surrounding trees and were relatively more active than during daylight. From the above observations it is assumed that oviposition occurs at night. In the insectary eggs Avere secured from three to seven days after the adults emerged, but in view of the fact that the moths are very inconsistent about ovipositing" in captivity, these data may have to be extended under normal conditions. It was necessary to place several males and females in a cage in order to ensure a supply of eggs. Although I have no data on the number of eggs laid by any one female, an examination of the oviducts shows that there may be a considerable number, for sixty- two fully formed eggs were dissected out from one female caught in. the field, and there may have been several laid before she was captured. This number of eggs was never secured from any female in captivity. From four females thirty-four eggs were secured in one day in a cage, and from these same four, fourteen eggs the following day. They died without further oviposition. These females were reared and laid only the eggs recorded. The adults have been kept alive in cages out of doors for twelve days after emergence from the cocoon, but they usually die sooner. They have never been seen to feed, and apparently they did not touch a honey-and-water mixture placed in the cage. The pres- ence of a 25 per cent solution of honey in the cage did not prolong the duration of adult life. Certainly food is not a requisite and is not necessary to oviposition. All the adults collected in the field died in a few days, so twelve days probably is a fairly long period of life. Humidity and temperature have considerable effect on this, and moths caught in the field can be kept alive four to nine days if held at io°-i2°C., whereas they die in one to three days at room temperature. They will remain active after an exposure to 7°C. for twelve hours, but when the air is cooled to 5°C. they very quickly become inactive. The eggs are laid singly on either side of the leaf and on any part of the surface. There is some preference shown for a posi- tion beside the midrib or some other prominent vein of the leaf. 428 CONNECTICUT EXPERIMENT STATION BULLETIN 288 but not to the exclusion of the rest of the leaf (plate XVII). Eggs are laid on leaves on all parts of the gray birch. Insectary records for 1924 showed a period of oviposition lasting from July 5th to July 21st, and field observations the same year showed unhatched eggs up to August 7th. No field observations were made dur- ing 1925, but gray birches sent from Boston y Mass., to Woods Holt, Mass., on July 9th, carried many unhatched eggs on the leaves. In 1926 eggs were found in the field between July 23d and August 3d, but in view of the fact that they were numerous July 23d and that some of those collected that day hatched July 30th, oviposition must have begun as early as July 16th. Many eggs collected August 3d hatched August 15th, so oviposition occurred as late as August 1st. Oviposition takes place usually during the month of July, and unhatched eggs may be found up to the middle of August. In 1926 the incubation periods of 48 eggs from laboratory-reared adults were 15 days on the average, the maximum being 17 days, the minimum 13 days, and the majority (27) taking 14 days. In all but three cases these eggs went through the incubation period between June 25th and July 13th, a month earlier than normal. They were from laboratory- reared adults which emerged earlier than normal. The other three eggs were incubated at the normal time, between July 31st and August 14th, and they took 14 days each, so the figure for all 48 was normal. This agrees with the period given by Hutch- ings in Ontario in 1925. The eggs have a high degree of fer- tility, and those that do not hatch are rare. A short time before the larva leaves the egg it can be seen curled inside (figure 31). When it emerges, the young larva bores through the bottom of the egg into the leaf, and as it feeds it leaves the egg filled with dark excrement. This habit makes it very easy to determine whether or not the eggs are hatched, for after the larva has left, the egg appears brown or black in contrast to its former translucent condition. For several days the young larva mines close to the egg, but it finally straightens its path and mines in a more or less definite direction. The larva completes the first and second instars and most of the third in the mine. I have found a head capsule in a mine only twice, but the measurements of the width of the head capsules (p. 444) and the descriptions of the larvae clearly indicate three mining instars. While in the mine the larva is always oriented dorso-ventrally with the leaf ; that is, its dorsal side is always toward the upper surface of the leaf. For the first week the mine is extended very little and is always close to the egg, giving a blotch appearance, due to continuous turning of the mine in a small area. A mine six days old measured only 1.5 mm. across the mined area. This larva never makes a real blotch mine, but its excavations are always linear and winding, with slightly BIOLOGY OF BIRCH LEAF SKELETONIZER 429 Fig. 31. Larval mine in birch leaf, slightly less than normal size (upper left) ; embryo in egg (upper right) ; larva in molting web (lower left) ; diagram of cross section of cocoon (lower right). enlarged ends. During the last part of its mining life the larva lengthens the mine very rapidly and broadens it somewhat. Most of the mines are about three-fourths of an inch long when finished (figure 31). When ready to emerge from the mine, the larva cuts a crescentic opening in the lower epidermis, an operation taking about fifteen minutes. It then works its way out until, by bending its body ventrally, it can grip the leaf surface with its thoracic feet. It then quickly pulls itself out of the mine, the entire performance consuming about two and one-fourth minutes. While emerging 43° CONNECTICUT EXPERIMENT STATION BULLETIN 288 from the mine, the larva spins a few long- threads of silk in the shape of a figure "8." I was unable to detect any use to which this peculiar habit could be put. The opening through which the larva emerges is always crescentic, always of the same approxi- mate size, and always cut in the lower surface of the leaf regard- less of which surface bears the egg. Hundreds of mines have been examined, and these were made by larvae which hatched from eggs laid on either side of the leaf, and only once an exit was found on the upper surface. It was thought that both light and gravity might cause the larva to orient itself dorso-ventrally with the leaf, and if this were true, the larva might be made to change its orientation by either invert- ing the leaf or changing the incidence of light on it. The upper surface of a leaf containing six mining larvae was covered over with black paper, and a mirror was so placed that daylight was reflected on the lower surface. This caused light to fall on the ventral surface of the larvae instead of on the dorsal, as under normal conditions. After an interval of 15 days all the larvae had emerged from the lower surface of the leaf. The change in incidence of light had no effect. A branch of a gray birch bear- ing several leaves which contained mining larvae was inverted so that the morphologically lower surface was uppermost. This changed the relation of the larvae to gravity. Under natural con- ditions the leaves never hang horizontally, so the orientation of the larvae is only relative at best. The leaves were examined after an interval of 18 days. Approximately 50 larvae had emerged from the mines, and of these only two had emerged through the morphologically upper, now lower, surface. In view of the fact that occasionally a larva emerges from the upper sur- face normally, the emergence of these two is of no significance. It is quite evident that changing the orientation to either light or gravity has no effect on the position of the insect relative to the morphologically upper and lower surfaces of the leaf after the mine is well under way ; nor will the surface of the leaf through which the larva emerges from the mine be changed by any such procedure. In the above cases all the mines were about half finished when the conditions were changed. The orientation of the larva must be determined after it bores through the epidermis of the leaf from the egg and before it mines to any great extent, for all larvae examined were found with the ventral side toward the lower side of the leaf even when the mine was only two or three days old. Where the insect is abundant it is by no means unusual to find 25 to 40 mines in one leaf. The mine shows more clearly through the upper epidermis of the leaf than through the lower, but this may be due to differences in the structure of these parts of the leaf rather than to the nature of the mine. The larvae are disin- BIOLOGY OF BIRCH LEAF SKELETONIZER 43 1 clined to gnaw through large veins, and usually the mine turns aside at these obstructions. The duration of the mining stage varies greatly. The maxi- mum period observed was 50 days, the minimum 13 days. This variation may be due partly to the conditions under which the larvae were reared. In 1924 and 1925 all the larvae were reared at the normal season and fed on leaves in normal condition. Dur- ing these two years the maximum duration of larval life was 37 days, and the minimum 24 days. In 1926 several larvae were reared from eggs laid by laboratory-reared adults a month earlier than normal — in June, in fact — and the foliage of the birches was in a more rapidly growing" and tender condition, especially since these birches were more or less sheltered and were well fertilized and watered. The water content of the leaves must have been greater than larvae would normally meet. All of these larvae but two 1 were in the mines over 32 clays, and the average period of 20 was 44 days. Of the two remaining, one was in the mine 16 days and the other 13 days. These 22 individuals are not included in computing the mining period. All of the other larvae reared during 1926, 35 in number, on which records were kept of the mining period, were in the mines 30 days or less. The dura- tion of the mining stage during 1924, 1925, and 1926 was on the average 22 to 27 days for 50 individuals, not including the 22 mentioned above. This figure is probably correct for normal con- ditions. Hutchings (1926) gives the mining period as seven to eight days in Ontario, but such a short mining period would bring the larvae to a fully grown condition much earlier than they really appear in the field. There is a difference in food plants to be considered, for larvae in New Haven were reared on gray birch, whereas the common birches in Ontario attacked by this insect are the yellow (B. lutea) and white (B. papyrifera). Nevertheless in Connecticut the larvae appear feeding externally on gray and white birches at the same time, which indicates a similar mining period. Mining larvae have been observed in the woods about New Haven as early as August 6th, and eggs collected on leaves have hatched July 30th when brought into the laboratory a week earlier. Larvae have been found out of the mines August 6th, but this is unusually early. The period when mining larvae occur around New Haven lies approximately between August 1st and Septem- ber 15th, with a maximum number present the fourth week in August. The rearing records coincide with these limits. When once free from the mine, the larva wanders over the leaf for a short time, an hour or two, and then spins its first molting web. A larva has never been seen to feed between the emergence from the mine and the spinning of the web. There may be a difference in different species of the genus, for Chambers (1882) 43 2 CONNECTICUT EXPERIMENT STATION BULLETIN 2S8 observed that Bucculatrix ambrosiaefoliella feeds two days between emerging from the mine and molting-. The larva which emerges from the mine is structurally more like the following external feeding instar than the preceding mining instar. This may argue for possible external feeding, but there is no evidence that it occurs at this time. The exact interval of time between emergence from the mine and either beginning or completion of the first molting web was determined for three larvae, all typical cases. One larva emerged from the mine at 8.20 a. m. and com- pleted its web at 10.45 A - M - > one emerged from the mine at 9.40 a. m., began its web at 10.40 a. m., and completed it at 12.10 p. m.; one emerged from the mine at 10.09 A - M -> began its web about 10.42 a. m., and finished it at 12.12 p. m. The larva often selects a position beside a large vein for its web, but it will also spin on the flat upper surface of the leaf. There seems to be a preference for a hollow over which the "roof" of the web may be spun, as the angle between the base and sides of a glass bottle or the hol- low beside the midrib of the leaf. Having selected a suitable location, the larva lays down a thin basal "floor" web on the sur- face of the leaf. This is about 1.5 mm. in diameter. Then it spins another web over this, making long tacks from side to side by swinging the entire thorax and the first two abdominal seg- ments from one side to the other. The body is held facing out and the threads are always straight. In shifting its position the larva swings the abdomen quickly almost 180 degrees. The periph- ery of the web is thus built up first, the center being weak. A series of short tacks is now made over the "frame" of long threads, and the center is strengthened. This is followed by a series of short tacks all around the edge, a proceeding which evi- dently strengthens the web. A hole is quickly made through the web near the center, and the larva crawls in head first between the "floor" and the "roof." In crawling into its molting chamber the larva doubles ventrally so that its back is down on the "floor" and its feet touch the "roof"; that is, it is oriented dorsally to the leaf. Before it is all inside, the larva swings its head to and fro, weaving a mat on the under side of the "roof." Since the diameter of the web is not much more than half the length of the larva, the latter is forced to turn around, and when completely inside, its head almost touches the last abdominal segment, the body being bent in a U shape and to the right or left (figure 31). It is plainly visible through the web. The larva is not content with getting inside, but actually makes a turn around its molting chamber. All this time it swings its head, weaving figure "8" loops, and in due time it incidentally has to cover the hole in the "roof" by which it entered. Although this opening is always cov- ered, the larva seems to make no deliberate attempt to cover it, doing so eventually as it works around inside. Most of the weav- BIOLOGY OF BIRCH LEAF SKELETONIZER 433 ing" is done after the larva is inside. In two instances to which particular attention was given, the time spent weaving prior to entering the web was eight minutes in each case, and the weaving time inside the web was 56 and 52 minutes respectively. In two other instances the larvae were weaving inside the web 60 and 70 minutes respectively. The principal part of the web is woven from the inside and is supported on the lighter structure previ- ously woven from the outside. The entire process of spinning the web takes about one or one and one-half hours, varying" some- what with the larva. The procedure is essentially as described by Snodgrass for Bucculatrix pomifoUclla. When the larvae are numerous, the birch leaves in August and September are spotted on both sides with many white webs. I call these "molting webs" rather than "cocoons," "pseudococoons," or "cocoonets," as termed by others, because I believe the word "cocoon" should be restricted to that structure, in which the pupal stage is passed. Having" completed its web, the larva retracts its appendages somewhat and remains quiescent a day or two. The tarsal claws and the crochets of the prolegs are not attached to the web, the larva lying freely with its ventral side away from the leaf. If the upper part of the web is removed, the larva falls out. Under such conditions it must molt in some sort of a chamber or fall off the leaf, which might be disastrous, for if food is not available after the molt, the larva dies in a few hours. The small size of the web holds the insect tightly, and the strong attachment to the leaf secures the web against being washed off or lightly brushed off. The web also offers protection from such enemies as ants during a period of helplessness. It is not essential to the process of molting, and seems to be an obstacle to quick molting rather than an aid. If removed from the web, the larva molts perfectly normally. Inside the web it has to pull itself around to get clear of its old skin. After one or two days in the web, the larva molts, and in a few hours, sometimes in one hour, it breaks out through the edge, at the junction of "roof" and "floor" (plate XVII). In molting, the head capsule separates from the rest of the old skin and is cast off first anteriorly. The larva then works its way clear of the remaining skin, casting it off the posterior segment of the abdomen. The molted head capsule and skin are left inside the web and separate from each other. The manner of leaving the web shows how precisely instincts can regulate action. After it has molted, the larva normally bites a hole through the side of the web and emerges, but before molt- ing it will not bite through and hence cannot get out even though it so desires. An individual which had just entered the web (in this case the second molting web) was rendered inert by hydro- cyanic acid gas. After four minutes it regained sensibility and for the next 14 minutes made spasmodic movements while recov- 434 CONNECTICUT EXPERIMENT STATION BULLETIN 288 eringj being apparently normal at the end of this time. It then attempted to get out of the web by pushing against the sides, having "forgotten" the reason for its imprisonment. It pushed vigorously back and forth for seven minutes, stretching the sides of the web in its endeavors to escape, but to no avail. Half a minute's work with its mandibles would have set it free, and had it molted, escape would have normally been accomplished in this manner. The instinct to bite its way out was totally lacking. Finally it began to move around inside the web and spin irregu- larly, then it began to weave the normal figure "8" loops, and in 28 minutes the web was finished. The larva cannot use its only means of escape from the web until the act of molting is accom- plished. This individual later molted and developed normally, not being in any way injured by its treatment. If removed from the web before it is finished, or, if it is finished, before the pre-molting quiescent stage begins, the larva will spin another web or as much of another web as is possible and will molt normally. An effort is always made to complete another web, but sometimes lack of the necessary silk, or exhaus- tion, or some other factor, compels the larva to stop after a few strands have been spun, and it then molts in the most convenient place. If it has entered on the quiescent stage prior to molting and has become fixed in the shape of a horseshoe, it does not straighten out when taken from the web, but retains its curved shape until it molts. Because of the fact that the larva leaves its web so soon after molting, the duration of the instars has been calculated to include the time spent in the web made by the particular instar in question. Thus the feeding period plus the subsequent quiescent period spent in the web gives the length of the instar. The time spent in the first molting web is much affected by temperature, and usually varies between one and four days in this climate. Many larvae spend less than 24 hours in this web, but most of the larvae are in it about two days. If this period is added to the days of mining life, we get a period of 24 to 29 days for the first three larval instars. This is not remarkably long when compared to the length of the next two instars, which together total about two weeks. After emerging from its first molting web, the larva feeds from one to nine days, the individuals varying greatly under the same conditions. If food is withheld from the newly molted larva, it dies in a few hours, a much shorter time than if starved after feeding a day or two. This is probably the result of remaining a day or two in the molting web without feeding. During the fourth instar the larvae are restless and wander about more or less. This probably accounts in part for the variation in the length of the instar, for the rapidity of development is much BIOLOGY OF BIRCH LEAF SKELETONIZER 435 dependent on the amount of food eaten. The average duration of the feeding" period for 73 individuals recorded was about four days. In only one instance was the feeding period as short as one day. Temperature affects the duration of this period to some extent, as will be brought out later. The effect of different spe- cies of birch as food will also be discussed in another section of this paper. The feeding occurs normally on the lower side of the leaf, and the veins and the upper epidermis are left intact. The entire leaf is never consumed. It is due to this habit of skeletonizing a leaf that the insect bears its common name. The larvae will eat which- ever surface of the leaf is toward the ground, and normally this is the lower epidermis. A birch leaf was inverted so that the normal lower surface was uppermost and covered with a black paper. A mirror was so placed that it reflected light on the leaf from below. The larvae normally feed on the lower side of the leaf, and under normal conditions this side is not so light as the upper. If the larvae fed on the lower side of this inverted leaf, they would feed on the lighter side and at the same time on the side normally uppermost. The two sides of the leaf differ in physical as well as chemical constitution of the surface. Of ten larvae placed on the upper side of this inverted leaf, four migrated to other parts of the plant (a normal movement), one remained on the upper side and was feeding when examined, and five went to the lower side of the leaf and were feeding. Seventeen hours elapsed between the placing of the larvae on the leaf and the final observation. Larvae were then placed on the uppermost side of an inverted leaf and watched. Usually they wandered about restlessly for a time until they came to the edge of the leaf. They then turned to the side underneath. Light reflected on the lower surface by a mirror seemed to have no effect. At times move- ment to the lower surface was long delayed and at times direct. It very evidently is a reaction to gravity that impels these larvae to feed on the lower leaf surface and not any dislike for the upper surface nor any negative reaction to bright light. What factors developed the habit of feeding on the lower surface only is another matter. The habit of the larva is to feed continuously over a lim- ited area, and it does not wander far unless the food supply gives out. If disturbed, the larva usually drops off the leaf, spinning a long thread as it falls. After falling a few inches it hangs on the end of the thread a moment and then quickly ascends. The thread is spun out the tip of the spinneret, and when the larva stops its descent, it is attached to the end of the thread by means of the spinneret. When it ascends the thread, it moves its head rapidly back and forth and winds the silk on the prothoracic legs which are held forward. If there is too much silk for the pro- thoracic legs, the mesothoracic legs are brought into use. On 43^ CONNECTICUT EXPERIMENT STATION BULLETIN 288 regaining its support, the larva simply drops the bundle of thread and walks away. This performance can be easily watched under the binocular if a larva of the last instar is used. The spinning activities of the larva, the quickness with which it drops from a leaf, and the distance it drops are much greater in the last instar than in the fourth. The speed with which these little insects can spin a thread while falling a few feet is remarkable. If touched, they snap the body back and forth rapidly and thus wriggle off the leaf and drop toward the ground. Yet after they have fallen some distance, they suddenly check their descent and can be seen to be hanging by the end of a thread. The silk of which this thread is formed must be spun from the silk glands and out of the spinneret as rapidly as the larva falls. The act of spinning apparently occurs automatically when the larva is disturbed. Because of their small size and their greenish color, together with the comparatively small amount of leaf tissue eaten, larvae of the fourth instar are not so noticeable as those following. In localities where Bucculatrix is abundant, however, ten to fifteen larvae may often be found on one leaf. Heavily infested birches frequently have 25 larvae of the fourth and fifth instars feeding on each leaf. During the majority of seasons no such number is likely to be present. The fourth instar molts as did the third, in a white silken web. This web is larger than the previous one, being about 2.5 mm. across. The larva builds the web and lies in it as previously described, being clearly visible. There is a slight difference in structure, as this larva weaves an elliptical mat after it is inside the web. This thickened part gives the second molting web a characteristic appearance, as the first molting web has this struc- ture to only a very slight degree. The time spent in this web varies normally from one to three days, the 75 individuals recorded averaging about two days. This is, of course, affected by the temperature, as was mentioned before. When added to the feeding period this figure gives the length of the fourth instar as about six days. The larva molts as before and emerges from the second molting web as from the first. It normally feeds on the under side of the leaf, skeletonizing it (plate XVIII), and in this instar the feeding is much more extensive. The injury to the foliage is most notice- able at this time, usually during the last of August and most of September. If the larvae are present in large numbers, all the parenchymatous tissue is consumed, and the leaf dies and drops from the tree- These larvae show greater spinning activity than those of the former instar and may be seen suspended from the leaves in great numbers in seasons of abundance. They feed from two- to ten days, the period varying with the individual and being affected by climatic conditions, and an average of 48 recorded individuals gives a period of nearly seven days. This BIOLOGY OF BIRCH LEAF SKELETONIZER 437 period includes the time from emergence from the second molting" web to the spinning of the cocoon. Toward the last part of the feeding period the gonads are clearly seen through the dorsal skin of the abdomen of the larva. About twelve hours before the time when the larva will begin to spin its cocoon, it stops feeding. This interval of time varies considerably and may be much less. By this time the larva has turned brown in color, due to the color of the large silk glands which run almost the entire length of the body. When ready to spin the cocoon, the larva drops from the place of feeding to the ground, spinning out a long thread as it goes. Larvae may sometimes be seen suspended from a thread about fifteen feet long. If the trees on which they are feeding- are shaken, these fully grown larvae drop to the ground quickly and in considerable numbers. Having reached the ground, they crawl under a stone, a fallen branch, a leaf, or any other object lying on the ground and spin their cocoons on the under side of this. Sometimes the cocoon is spun on the ground itself. In captivity they will frequently place the cocoons on the sides of the cage close to the base. I have rearea hundreds of larvae, and they all have dropped to the ground or close to it to pupate. Fletcher (1893) mentions finding three cocoons on the twig of a birch, but all the cocoons which I have found in the field have been on fallen leaves or other objects lying on the ground. The manner in which the larva spins its cocoon is characteristic of the genus and quite unique. The earliest description of this process in the genus Bucculatrix is by Lyonet, who wrote to Reaumur, December 22, 1744, concerning the larva of B. ulmella and its cocoon. This description was not published until 1832 and has been referred to in the historical part of this paper (page 396). De Geer, in the first volume of his "Memoires," published in 1752, described the cocoon of B. frangulella (see page 395), and Snod- grass in 1922 likewise described the manner in which B. pomi- foliella Clemens wove its cocoon. These three papers go into the details of the process by which the larva lays down its threads, and from a microscopic examination of the cocoon of B. canaden- sisella it is apparent that this larva weaves its threads in precisely the same manner as does B. pomifoliella. The general process of weaving is similar in all four species, differing only in a few details. Chambers (1882) described briefly the formation of the cocoon by B. ambrosiaefoliella Chambers, and McGregor (1916) gave a brief description of the finished cocoon of B. thurberiella Busck. In 1892 Fletcher briefly described the general procedure of weaving by B. canadensisella Chambers, and in 1893 Lintner mentioned the same subject, but the latter's description is not correct, and Fletcher's description is not detailed. The larva of B. canadensisella Chambers first lays down an oval mat to serve as a base for its cocoon. It does not previously weave a palisade of poles around the site selected, as do many 43& CONNECTICUT EXPERIMENT STATION BULLETIN 288 species of the genus. It then commences at one end of the mat to weave an outer supporting" ridged structure of comparatively coarse threads (about .005 mm. thick), facing the work and back- ing away as the woven structure progresses over the mat in an arch. The ridges are formed by the ends of a series of loops made from one side to the other. The diagram in figure 32 gives Fig. 32. Diagram of method by which larva of Bucculatrix pomifoliella Clemens weaves its cocoon. After Snodgrass. the principle. Between ridges the threads cross diagonally. This figure is from Snodgrass 1(1922) and gives his conception of the actual motions made in weaving. As the cocoon becomes higher, the larva raises the anterior part of its body, and the radius of the structural arch is gauged by the raised part of the body as it swings from side to side, most of the body being fixed in the midline of the oval base. Possibly the prothoracic legs are used in the weaving to aid in guiding the work, as mentioned by Snod- grass and Lyonet. Certainly these legs are held up to the struc- ture. When the cocoon is about two-thirds finished, the larva enters it, turns about, and crawls out until its head reaches the other end of the mat. It now has its anterior end outside of the cocoon but its posterior end in' the cocoon. Beginning to weave exactly as before, the larva builds up the last third of the cocoon to meet the previously formed two-thirds, gradually enclosing itself as it works. When the two sections meet, they are joined by cross threads. The architecture is not perfect, for the ridges of the two sections rarely coincide, and sometimes the heights of the sections are not equal. The result is a break in the continuity of the ridges at the junction and often a sag in the contour of the cocoon. A completed cocoon is shown on plate XVIII. This outer structure is not closely woven and the insect can be seen clearly inside. It is, however, stiff and gives support to the lining which is to be woven. The sides meet the oval base perpen- dicularly. Having completed its superstructure, the larva weaves a closely knit lining of fine threads (about half the thickness of the threads of the supporting structure) all around the inside by swinging its head in figure "8" loops. Where the walls of the superstruc- ture join the base, the cocoon does not follow but makes a round BIOLOGY OF BIRCH LEAF SKELETONIZER 439 corner, as the diagram in figure 31 shows. It is this lining" which makes the cocoon opaque. Snodgrass has described a series of "valves" in the anterior end of the cocoon of B. pomifoliella Clemens, but in the cocoon of B. canadensis ella these are not pres- ent. The cocoon when first finished is almost pure white, but it soons turns brown. This brown color is due not to the pupa inside, for it is present before the prepupa molts, but to a change in the color of the silk when exposed to air. The time necessary to complete a cocoon is from eight to sixteen hours normally. Inside the cocoon the larva remains two or three days before pupating. This prepupal period plus the feeding period makes the fifth instar about nine days long on the average. The larva molts in the cocoon in a manner differing slightly from that which takes place in the molting webs. In the webs the head capsule is cast off entire and anteriorly while the rest of the larval skin is worked posteriorly off the anal segment. In the cocoon the entire larval skin, head capsule included, is worked off posteriorly. The individuals which were reared in the outdoor insectary under normal temperatures in 1924 pupated from September 4th to September 25th; in 1925 from September 8th to September 13th; in 1926 from September nth to September 23d. This does not indicate the time of disappearance of the last larvae in the field. During these three years an examination of birches about New Haven was made in order to determine the normal close of the larval period. In 1924 the last larvae were found October 9th; in 1925, September 19th; and in 1926, October 9th. The early disappearance of larvae in 1925, although not caused by any apparent natural enemy or unusual climatic condition, was excep- tional. It may have been caused in part by an early season start- ing the life cycle earlier. In view of the fact that larvae will feed at 48 to 50°F. and will eat birch leaves until they begin to turn yellow very few are caught before pupation by cold weather or lack of food. The total larval life occupies from 38 to 46 days, as a rule, as the table on oaee 441 indicates. This is not an average of the com- pleted larval life of a number of insects, but an average of the separate stages of many individuals, rather few of which com- pleted the entire larval period while under observation. Nine larvae carried through from egg to pupa in 1926 averaged 41 days, the maximum being 45 days, and the minimum 36 days. This is as close as could be expected to the 38 days given in the summary for 1926. In the table below, the larval life from the hatching of the egg to the spinning of the cocoon is given for the nine individuals mentioned above. Two days as prepupa should be added to the six days of feeding in the fourth instar to give the total larval period of 41 days. It will be noticed that the larva' does not accelerate through one instar if slowed down on a previous 44° CONNECTICUT EXPERIMENT STATION BULLETIN 288 instar, but that any retardation during the growing period is per- manent as regards time. This is borne out by the other records. Seven larvae were reared in the laboratory in vials containing moist sand, and were under identical environmental conditions. The figures for the stages are given on pages 474-478 (larvae 131- 137). The most slowly growing larva was six days in the fourth instar and was feeding six days and fifteen hours in the fifth, while the most quickly growing larva was four days and nine hours in the fourth instar and was feeding four days and nineteen hours in the fifth. It will also be noticed that the quiescent period spent in the molting web is independent of the length of the feed- ing period, and as the feeding period grows shorter, the propor- tion of time spent in the web during one instar grows greater. In larva number 9 in the table below, two-fifths of the fourth instar is quiescent, and in number 1, three-sevenths, but in numbers 2 and 4. only one-fourth of the fourth instar is quiescent. The effect of food and temperature on larval growth will be discussed later. Table i. Complete Larval Pi :riod to £ 1 ™ ■0 % -a bo j= ho c P .a 2 £ "rt T3 £ "a •3 s-.H "(3 a »— 1 1) s £ ffl h- 1 H h H to tn O H I 8-14-26 26 2 28 4 3 7 S 9-23-26 40 2 8- 1 1-26 22 2 24 6 2 8 7 9-19-26 39 3 8- 1 1 -26 22 3 25 4 2 6 6 9-17-26 37 4 8-14-26 20 2 22 6 2 8 6 9-19-26 36 a 8- 7-26 27 2 20 4 2 6 5 9-16-26 40 6 8- 1 1 -26 30 1 31 5 3 8 4 9-23-26 43 7 8- 7-26 27 3 30 5 2 7 6 9-19-26 43 8 8-14-26 20 2 22 S 2 7 6 9-18-26 35 9 8- 7-26 22 1 23 3 2 5 6 9-10-26 34 Aver. 26±.74 7±.22 6±.i6 39^-7 The chart below (text figure 33) gives the periods during which the various stages may be found in the field around New Haven, Connecticut. These limits are computed from field observations and data obtained in the insectary and are broader than actual field observation alone would give. From what notes there are of the occurrence of this insect elsewhere, it seems likely that these periods are approximately correct for the entire region in which the insect is found. BIOLOGY OF BIRCH LEAF SKELETONIZER 441 cq ffi < u 5 _e .5 jo sA"u(j ^ 00 CO JBJSUl JO qjSuai ibjox 00 1 1 00 •+— »-< 4— •Bdndsjd saeq; 0J 4- 4- CM Suipaaj sabq [ CO N>0 1 O VOOO tS. TtVO SIEllpiAipUI jo jsquin^ v-> § JEJSU1 jo qiSuaj isiox IT) in I md qaA\ Suijjoui p& Ul SAEQ CO M M CO M t-t h-l h-t sjenpiAipui jo jaquin^ CO "3 < 30U3SJ3UI3 JO pOIJ3 rq 3 3 cti X r-' U vo S.'S > CM 3 3 CC O a <* S cu ra bo 3 C T3"3 ■2 a .5 o S « ia *! S'~ o "d a u cu c ho c tsfe _ cu cfl '6> to M* to rt-TE-H — bo cu cu 442 CONNECTICUT EXPERIMENT STATION BULLETIN 288 VII. Determination of the Number of Instars It is very difficult to determine the number of instars by exam- ining the mines for head capsules. A large number of mines were examined for this purpose and in two cases one capsule was found. The extent of growth and the morphological changes undergone during the mining period indicated at least two and per- Fig. 33. _ Seasonal occurrence of the various stages of B. canadensisella in the vicinity of New Haven. The larval periods shown should read mining instars, fourth instar and fifth instar. haps three larval stages. A number of larvae were collected in the field and the width of the heads measured. The head is not sub- ject to growth changes during any one instar, and according to Dyar (1890) a constant numerical ratio exists between the widths of the heads of any two successive instars of a larva. If the heads of two successive instars are measured, or if a large number of miscellaneous heads are measured, the ratio for the species can be determined and the possibility of missing an instar removed. Any dimension of the head may be used, but the width is the most convenient. Several embryos which had developed to the stage where they were about to emerge from the egg and where no further growth of the head could be expected were measured. These were all mounted in Canada balsam. As seen by the table on page 444, the average width is .078 mm., and nine of the twelve measured .076 mm., which latter figure may be considered normal. It is to be expected that the measurements for the first instar would con- form to this figure, and of the sixty-one mining larvae measured, eighteen either equal this figure or closely approximate it. All but two of the eighteen equal it. The average width for the first instar is then .077 mm. Thirty-six of the sixty-one measure .114 BIOLOGY OF BIRCH LEAF SKELETONIZER 443 mm. in width or very nearly so, thirty-four measuring" just that figure and the other two measuring .120 mm. The normal and average for this group is .114 mm. The remainder of the mining larvae measured, thirteen, all give a head width of .171 mm. Two larvae were secured just as they left the mine and before they began to weave the molting web, and their heads measured. Both gave a width of .171 mm. These two are marked (ex) in the third column. This checked the group giving this measure- ment as the last mining instar. Also four larvae were found in the process of molting and with the head capsules just far enough off to permit the measurement of both the old capsule and the new head. Two of these gave the width of the old capsule as .076 mm. and the new head as .114 mm., while the other two gave .114 mm. and .171 mm. for the two widths. This gives a check on the three groups. According to Dyar's principle we should expect .ii4_.i7T__ R .077 _ .ii4 _ In this case the ratio "R " is 1.5, and the number of instars in the mine is, as the figures indicate, three. To further check this principle, a number of external feeding larvae, also collected in the field, were measured. I have placed these forty-one larvae in two groups as the table shows. According to the principle used above, the measurements should be ,257 mm. (.171 x 1.5 = .257) and .385 mm. (.257x1.5) for the fourth and fifth instars. (The actual number of externally feeding instars was determined by actual observation, of course.) In the fourth instar the aver- age width was found to be .245 mm. for the nineteen individuals, with a variation between .228 mm. and .257 mm. The last instar, containing twenty-two individuals, gave an average width of •353 nim. with a variation extending from .304 'mm. to .390 mm. It is questionable whether the two larvae whose head widths are .304 mm. belong to the fourth or the fifth instars if one judges by these two measurements alone. The average width is less than that expected in both the external feeding instars, but even so the measurements are sufficiently closely grouped in each case to determine the instar. It is to be expected that the more nearly the larvae approach the fully grown condition, the more widely will they vary in size, for the absolute extent of variation in size under normal conditions increases with age. The change in environment from the mine to the surface of the leaf, with its difference in manner of feeding involved, would also change the shape of the head, because mining larvae have relatively flatter heads.. The actual measurements obtained of the heads of the first three instars is much closer to the ideal than would usually be expected. 444 CONNECTICUT EXPERIMENT STATION BULLETIN 288 Table 3. Head Widths of the Larvae of B. canadensisella Chambers (All dimensions in millimeters) First Second Third Fourth Fifth Embryo instar instar instar instar instar I .076° .076* 114° .i7i°(ex) .257 .323' 2 .076° .076* 114° .171° (ex) .247 .352' 3 .076° .076* 1 14 .171° .247 .323' 4 .076° .076* 120° .171° .228 •304' s .086° .076* 1 14 .171° .247 .371' 6 .076° .076* 114 .171* .238 .380' 7 .086° .076* 1 14 .171* .247 .38o' 8 .082° .076* 114 .171* .247 •300' 9 .076° .076* 1 14 .171* .228 .380' 10 .076° .076* 114* .171' .247 .371' 11 .076° .076* 114* .171 .238 .36l' 12 .076° .076'* 114* .171' .247 .361' 13 .076* 114* .171 .247 .304' 14 .095* 114* .171' .247 .370' 15 .076' 114* .171' .252 .38o' 16 .082] 114* .228 .352' 17 .076' 114* .250 .361' 18 .076' 120* •257 .361' 19 114* .247 ■352' 20 114* .332' 21 114* .380' 22 114* •370' 23 114* 24 114* 25 114* 26 114 27 1 14 28 114° 29 114 30 114' 3i 114' 32 114' 33 114' 34 114' 35 114' 36 114' Theoretical average 114 .171 .257 .385 Average found .078 ± .0008 .o77±.oooi U4±oooi .i7i±o.o •245 ±.OOI3 .353=! Standard deviation .0039 .0045 001 .000 .0085 .025 Greatest deviation from theoretical .008 .008 006 .000 .017 .081 .0036 111 previous descriptions of the genus Bucculatrix it has been tacitly assumed or explicitly stated that the mining" period included one instar only, and that the insect always molted on the surface of the leaf. The only mention I have found of a larva molting" in the mine is in a description of the larva of B. ambro- siaefoliclla Chambers by Chambers (1882) in which he states that the larva in question molts once in the mine, once on the surface of the leaf, and once in the cocoon. It would be well, however, to apply Dyar's principle, at least to the early stages, before mak- ing any definite statements regarding other species of this genus. BIOLOGY OF BIRCH LEAF SKELETONIZER 445 All the measurements given above were made with an ocular micrometer, using' a low power of the microscope. The smallest micrometer scale division was .019 mm., and it was found imprac- ticable to interpolate to less than one-fourth this, a measurement of .005 mm. The embryos measured were all mounted in Canada balsam. The larvae of the first three instars, marked with a small circle, : '°," were also mounted in balsam ; those marked with an asterisk, "*," were mounted in glycerine; and those marked with an apostrophe, '"," were specimens preserved in alcohol. All the fourth and fifth instar larvae were preserved in alcohol after fixation in Gilson-Carnoy's fluid. VIII. Food Plants The plants on which the larvae feed are restricted to the genus Bctula, with the possible exception of the alder, Alnus incana. Johannsen, who reports (1911) the single instance of larvae attacking the alder, has also reported (1910) the presence of larvae on red oak. There are other species of Bncculatrix which feed on oak, one of which is very common in Connecticut, and it is very probable that the larvae referred to by Johannsen were not B. canadensisella. The mines in oak leaves are very similar to those of the birch skeletonizer, but the cocoons are white and are found on the trunk and branches of the trees. I have not bred this species, but Forbes (1923) gives B. ainsliella Murtfeldt and B. packardeila Chambers as indigenous to northeastern United States, and both feed on oak. Alder is closely related to birch, and although the larvae of B. canadensisella did not survive in laboratory tests, through one complete instar on Alnus (rugosaf), under different conditions they may possibly feed on this plant. Of the species of birch on which this insect lives, four are native and one imported from Europe. These are Betula popalifoUa (gray birch), B. papyrifera (paper or white birch), B. lutea (yel- low birch), B. lenta (black birch), and B. alba (European white birch) respectively. The European birch is a common ornamen- tal tree in northeastern United States and southeastern Canada, and varieties are called the cut-leaf or weeping birch. This tree in Canada seems to be a favorite food plant, but in the vicinity of New Haven it is not quite so severely attacked as the gray birch. Of the four native food plants, the black birch seems to suffer least, although Maheux reports (1926) that in Quebec this tree has been heavily skeletonized. Which of the other three is most severely injured seems to depend on which is prevalent in the locality. In most of Connecticut the gray birch is the preferred food plant, but on the shores of Highland Lake, where the white and black birches are the only two species common, the white birches were heavily skeletonized in 1925 ; and in other parts of 44^ CONNECTICUT EXPERIMENT STATION BULLETIN 288 Litchfield County, where yellow birch is quite common, it is a favorite host. In Ontario and throughout the Great Lakes regions, the yellow and white birches are the trees which suffer most. The black birch in Connecticut is very slightly injured and usually is untouched, even though its branches intermingle with those of the white and gray birches when these two bear thousands of caterpillars. In laboratory tests the larvae ate the leaves of the black birch very readily. These larvae were taken from gray birch and fed on black birch during the fifth instar. Five of the ten larvae pupated normally, although the duration of the instar was 173 hours on the average as compared with 117 hours for the control. This delay in maturing was partly due to the delay the larvae experienced in getting accustomed to the new food plant. The red or river birch (Betula nigra) is not a common tree in northern United States and southern Canada, and this may be the reason that it is not reported as being" attacked by this insect. New England is about its northernmost range, and here it is found only in a few scattered places along river banks. . No attempt was made to rear the larvae on the leaves of this tree, as the material is not readily available, and there are no references in the literature to it as a food plant. There are four other genera of plants belonging to the same family as the birches and growing very commonly in the same localities as these trees. These are Ostrya (hop hornbeam), Carpinus (ironwood), Alnus (alder), and Corylus (hazelnut). Under natural conditions I have never observed any of these plants attacked by the larvae of B. canadensisella, although they very frequently intermingle with the birches. In the laboratory the larvae have been forced to eat the leaves of Alnus but could not maintain themselves on these leaves. The larva itself has really very little to do with the choice of food plants, for this is a leaf -mining insect in the early stages, and if the egg is not laid on a leaf in which the larva can live, death results. Even during the external-feeding stages it is very questionable if a larva could survive long enough to travel from an unfavorable to a favorable plant unless the two plants were very close together. In an attempt to secure eggs on the leaves of the alder, I placed two alder twigs, each bearing two or three leaves, in a cage with five males and five females. One of the twigs had been dipped in the distillate from an aqueous extract of birch leaves, and the other was normal. The moths were collected in the field. Two females lived six days, one five days, one three days, and one two days, but no eggs were laid. In another similar trial with one female and six males, the female lived four days but laid no eggs. In view of the fact that the females are loath to lay eggs in cap- tivity, the results are merely indicative and not conclusive. BIOLOGY OF BIRCH LEAF SKELETONIZER 447 An attempt was made to force larvae to eat the leaves of the alder and the black oak. All these larvae were collected in the field on gray birch. Five larvae in the first molting webs were placed in vials with the leaves of each plant. On the alder all five larvae died in three and one-half days or less without feeding. On the oak some feeding- occurred and one larva went through the fourth instar in eight days and then died, starved, in fourteen. Three of the others died of starvation in four and one-half days or less, and one was accidentally killed. Although alder is more closely related to birch than is oak, yet the black oak was preferred as food, though it could not sustain the larvae. Ten larvae were then similarly kept with the leaves of these two plants, but the leaves were previously dipped in a distillate from an aqueous extract of birch leaves. It is sometimes possible to make insect larvae eat materials that have the odor of their food plants. Of the ten larvae used in this case, five were in the fourth instar and five in the fifth. On the alder both instars fed a little. One fifth-instar larva lived ten days, and two fourth-instar larvae lived seven days, but none went through a complete instar. On the oak there was more feeding than on the alder. One fifth-instar larva spun a cocoon after five days," and three others lived between seven and nine and one-half days. One fourth-instar larva lived six- teen and one-half days, molting meanwhile, and three others lived between five and one-half and eight days. In only one instance on the oak was the fourth instar completed. Although the dis- tillate from the birch extract made the alder and oak more attrac- tive to the larvae, and they ate relatively much more of the leaves w T hen so treated, they did not show any growth except in the one instance mentioned above. All but one gradually shrunk in size and finally died of starvation before molting. Control larvae fed on the gray birch were normal in development. On this basis the possibility of larvae under natural conditions living on either oak or alder seems remote, and the reports of feeding on these plants were probably cases of misidentification of the insect in question. Under laboratory conditions the larvae from the gray birch very readily eat leaves of paper and black birch, and larvae from paper birch just as readily eat leaves of gray birch. In all cases the larvae will mature. The trials conducted were not sufficiently extensive to determine whether or not there is a racial difference in the individuals from different host plants. This racial differ- ence would be primarily manifested by the oviposition response of the adult, and difficulties in securing eggs consistently from females have precluded any definite experimental evidence on this matter to date. When the larvae were reared in the laboratory they were placed on the plants under trial, and if they left these plants, they were put back again. This was continued until they 448 CONNECTICUT EXPERIMENT STATION BULLETIN 288 ate the leaves or died. Under normal circumstances, no such condition would be met, and it is conceivable that the larvae might well starve to death in the midst of food which would sustain life, but which, for various reasons, they would not eat. The preference for birch as food, as concerns the larvae, is partly con- trolled by a chemical sense, for they eat oak and alder leaves more readily when these are first dipped in a distillate from an extract of birch leaves. •IX. Factors Affecting Abundance The phenomenon of periodic outbreaks of Bucculatrix has been dealt with historically in previous pages. Some of the factors which have a bearing on the abundance and rate of increase of this insect deserve consideration. These may be grouped under food supply, climate, and natural enemies (including diseases). Man has not as yet played any direct role in the control of this species. There is no scarcity of food plants in the northern United States and southern Canada, and the endemic population of Bucculatrix has no apparent effect on the growth of birch trees. Between outbreaks the larvae are scarcely noticeable. Paper birch forms a great part of the subarctic transcontinental forest and is a very common tree as far south as the Great Lakes and central New England. Gray birch is common farther south, and in New Eng- land and New York it is a weed tree which is constantly encroach- ing on cleared land. These two are the principal food plants and neither is being extensively cut by man. During an outbreak, when the larvae frequently eat all the foliage on the trees over considerable areas, the birches are not killed, even by several attacks in successive years, due to the lateness of the feeding period. The greater amount of feeding occurs during the last of August and September, and at this time of the year the trees have passed through the most active season and are not so severely injured as they would be by a similar attack earlier in the summer. This insect could probably never eliminate its food plant in any given region. It very probably checks the growth of the trees the year after a severe attack, but this check would not be suffi- ciently great to cause a decrease in the available larval food sup- ply. Another factor that sometimes has some effect on the abundance of a particular insect is the competition for food with other species of insects. The defoliation of the birches in any region early in the summer would very obviously affect the sur- vival of Bucculatrix, which feeds late in the season. At present this factor cannot be considered as of much importance. One of the most serious insect enemies of the birch in New England is the saw-fly, Fenusa pumila Klug, whose larvae mine the leaves during the entire summer, as there are several generations. Since BIOLOGY OF BIRCH LEAF SKELETONIZER 449 this insect confines its work entirely to the new terminal growth, while Bucculatrix larvae feed by preference on the older leaves of the tree, the two live together in harmony. There is always the possibility of the last Bucculatrix larvae of the brood not hav- ing sufficient food, because of the work of the earlier developing part of the brood, and hence being unable to survive. The habit of spending" two days in a quiescent state in the molting web increases this danger, for during these two days the foliage on the tree may be entirely consumed. All the observations made in the field indicate, however, that there is no reason to believe that there occurs any decrease in food supply which would have any very important effect in reducing the numbers of this insect even fol- lowing a year when it was abundant. No data have been obtained on the effect of climate on the sur- vival of this species. The greatest danger to an insect is during the hibernating" period, when severely cold weather sometimes kills off much of the population of certain species. It is a well-known fact, however, that insects which hibernate under the snow are better able to survive extremes of cold than species which hiber- nate above the snow line. For this reason a very cold winter would not be expected to have a very great effect on the popula- tion of the birch Bucculatrix. This is an indigenous insect and is inured to the climate of its present geographical range, and the greatest effect of climate on its abundance is probably indirectly through limitations on the distribution of its food plants. It is not inferred that climatic variations have no effect on the population, but rather that climate alone is not responsible for the more or less regular rise and fall in abundance. The parasites and predaceous enemies of this insect probably account for the increase and decrease in its numbers more than any other one factor. Ten species of Ichneumonoidea and Chal- cidoidea have been reared from the larvae and pupae. One of these, Hemiteles, is very probably hyper-parasitic, as Viereck (1916) states that all the species of this genus are probably sec- ondary or hyper-parasites. The 10 species with the stage of the host from which they emerged are listed below : Stage of Host 1. Bitccidatriplcx sccundus Viereck Braconidae pupa 2. Halticliclla xanticle's Walker Chalcididae pupa 3. Gelis urbanus Brues Ichneumonidae pupa 4. Cirrospilus ocellatus Girault Elachertidae larva (ext. feeding) 5. Gelis bucculatricis Ashmead Ichneumonidae pupa 6. Mesochorus sp. Ichneumonidae pupa */. Pleurotropis bucculatricis Gahan Entedontidae .... pupa 8. Closteroceru s (cinctipcimis Ashmead?) Entedontidae larva (mining) 9. Dcrostcnus sp. Entedontidae larva (mining) 10. Hemiteles sp. Ichneumonidae pupa * This is a new species the description of which, by Gahan, is published in Psyche, volume 34, June, iQ-'7. 45° CONNECTICUT EXPERIMENT STATION BULLETIN 288 The family names are those used by Viereck (1916). These species are all small and occur singly in the host. The extent to which they parasitize the host varies, of course, from year to year and in different localities. In the winter and spring of 1924 there were collected 397 cocoons from which there were secured 29 parasites as follows : Gelis bucculatricis 14 specimens Bucculatriplex secundus 7 Haltichella xanticles 6 Hemiielcs sp 2 These cocoons were collected from several localities around New Haven, where the host had been abundant in 1923. A large number of pupae died without metamorphosing, and only 152 adult moths were secured from this lot. In 1925 most of the pupae of which records were kept were from larvae reared in the insectary, and the parasitism, therefore, was abnormally low. Nine individuals of Bucculatriplex secundus and two of Halti- chella xanticles were obtained from 352 cocoons. Conditions dur- ing the 1925 season of emergence were not normal, as the cocoons had to be kept in the laboratory. The records are not comparable to those obtained a year later. In 1926 there were collected dur- ing April and May 209 cocoons in a locality where the larvae had been very abundant the previous season. No collections had been made in this locality during either 1924 or 1925. All these cocoons contained pupae (as later examination showed), and from them were secured 53 parasites and 58 adult moths. The cocoons were kept outdoors in a shaded place until the emergence period was passed, and then those from which no insects had emerged (98 cocoons) were examined. Five contained dead parasites and 93 contained dead pupae. Of the insects which emerged, then, 47.7 per cent were parasites, and of the total num- ber of pupae collected 27.7 per cent were parasitized. The para- sites were of the following species : Bucculatriplex secundus 37 specimens Pleurotropis bucculatricis 12 Haltichella xanticles 1 specimen Gelis urbanus 2 specimens Undetermined (escaped) ............. 1 specimen It is evident that of the insects which emerge from the cocoons the parasites make up a large percentage, and the parasites are better able to survive than the host. Of the 151 non-parasitized pupae, only 58, or 38.4 per cent, produced adults, whereas of the parasitized pupae, 58 in all, 53 or 91.4 per cent produced parasites. The presence of a parasite in a pupa is very easy to determine after three months, as by this time the parasite has consumed most of the host tissue. A parasite could not have been easily over- BIOLOGY OF BIRCH LEAF SKELETONIZER 451 looked in the examination of dead pupae. Since these parasites occur singly in the host, the percentages are comparable. There is, of course, the possibility that some of the parasitized pupae died before the parasites had developed far enough to be observed in a dead and desiccated host. The fact that of the 93 dead pupae above mentioned 44 had reached the pre-imaginal stage before dying indicates that this possibility would have no great bearing on the results obtained, for had any parasite been present in any of these, it would have prevented the host from reaching the con- dition of the pre-imago. It is also true that the parasites are bet- ter able to withstand high and low temperatures during the period of emergence than is the host. Three lots of 20 cocoons each were kept at different temperatures, one at 3i-33°C, one at room temperature which varied between 18 and 26 , and one at 8-1 5 . The cocoons were placed in test tubes (50 cc. capacity), 10 in each tube. To serve as a check on the humidity effect, one of the tubes of each lot contained a piece of wet blotting paper which produced a moisture-saturated atmosphere in that tube. The other tube received nothing. All tubes were kept corked except for an interval of about one minute each day when they were opened in the room in order to renew the air supply. The relative humid- ity of the room averaged 67 per cent, with a variation of 13-14 per cent each side of this for brief intervals of time. The experi- ment began June 2, 1926. From the cocoons held at room tem- perature 18 insects were secured, nine from each tube. This represented a normal emergence. Four of these were parasites, all Biicculatriplex secundus, and 14 were adult moths. From the cocoons held at 8-1 5 °, two parasites only emerged, one from each tube. One was a specimen of Biicculatriplex, and the other was Mesochorus sp. From the cocoons held at 31-33 two parasites only emerged, both from the tube contain- ing room air. Both were Haltichella xanticles. After being examined July 10th, all the cocoons from which no insects had emerged were removed to the outdoor insectary. Eight adult moths subsequently emerged from the tubes that had been held at 8-1 5°, four from each tube. After the emergence period was well passed the remaining cocoons were examined. No dead parasites were found, and most of the dead pupae had reached the pre-imaginal stage. Although the number of insects concerned was not large, the parasites were very evidently better able to withstand the extremes of temperature than was the host, for all the parasites emerged under these conditions, but no moths were obtained. The two parasites which came out of the tubes held at 8-1 5 emerged July 3, and the two from the tubes held at 31- 33 emerged June 5 and June 6 respectively. In the case of the latter two, it might be suspected that the difference in develop- ment between host and parasites enabled the parasites to complete 45 2 CONNECTICUT EXPERIMENT STATION BULLETIN 288 the metamorphosis and emerge when the host could not, for they were exposed to the high temperature only three and four days. However, from a third tube set up the same as the others but con- taining calcium chloride and held likewise at 31-33 , there emerged two parasites only, one specimen of Bucculatriplex secundus on June 23, and one specimen of Pleurotropis buccula- tricis on July 4. The cocoons in this last tube were exposed not only to the high temperature, but also to the desiccating effect of the chloride. No adult moths were secured, and no parasites died before emerging. In addition to this emergence of parasites from pupae, there is sometimes a considerable parasitism of the mining larvae by Clos- teroccrus and Derostenus. When these parasites were first dis- covered, it was thought that they were one and the same species, as they were in the larval stag'e and resembled each other closely. They are therefore grouped together here. If the mines of the Bucculatrix larvae are examined in September, many will be seen to contain the remains of the larva and in addition a very minute parasite larva about .75 mm. in length. September 10 and 14, 1925, there were collected 619 Bucculatrix mines in gray birch leaves. Of these, 522 were vacant and showed by the exit hole that the Bucculatrix larva had emerged normally. The other 97, or 15.7 per cent, contained each the remains of a Bucculatrix larva and one parasite larva belonging to one of the two genera in ques- tion. The only exception to this was one mine which contained two parasite larvae. The first of October, 1926, the same locality was visited and 289 mines were collected. The mines this year were much less abundant than in 1925. Of these 289 mines, 100 had been normally vacated by the Bucculatrix larvae and 58, or 20.1 per cent, contained parasites. The remainder, 131, contained dead Bucculatrix larvae, but the cause of their death could not be determined. It could hardly have been the parasites in ques- tion, for the larz>ac of these two species were found in the other mines. The above figures show that there may be a heavy mortality of the host by the combined attack of the parasites. Of these, Buc- culatriplex secundus is the most commonly found. Only one locality has been examined for Derostenus and Closterocerus, and it is not known just how widely spread these two species are. Pleuro'tropis bucculatricis, Haltichella xanticles, and Gelis buccu- latricis are also rather common. One specimen only has been secured of Mesochorus and Cirrospilus ocellatus. The former emerged from a cocoon in 1926, and the latter was found in the pupal stage in a molting web of Bucculatrix. Hemiteles may be a secondary parasite and hence of no use in checking the repro- duction of Bucculatrix. There is a possibility that some of the others also are secondary parasites. BIOLOGY OF BIRCH LEAF SKELETONIZER 453 The adults of Derostenus appear the last of the summer, but the adults of the other parasites appear about the same time that the host adults appear. This indicates that there may be other hosts for some of the parasites. Bucculatrix sccundus hibernates as a larva in the pupal cuticle of the host. Derostenus and Clos- terocerus kill the host larva before it completes the third instar and hibernate as larvae in the mines of the host. The other spe- cies hibernate in the pupal cuticle of the host, but the hibernating stage of these was not determined. All the parasites are minute. The Ichneumonoidea adults are about 1.75-2.00 mm. in length, and Haltichella xauticlcs is about the same size. Plcurotropis buccidatricis is about 1.5 mm. long;, and Derostenus and Clostero- cerus are each about .60 mm. in length. More important as enemies of the Bucculatrix larvae than any one of the above species of parasites, and perhaps than all of them combined, are the various species of ants and other predaceous insects which capture the larvae when they descend to the ground to pupate. Ants will not only capture the larvae before the cocoon is well begun, but will also pull a larva out of the cocoon in which it is almost entirely enclosed. In 1925 ants destroyed the entire stock of larvae in the insectary. On one occasion the litter on the ground under a birch which had borne hundreds of larvae was very carefully examined for cocoons after all the larvae had disappeared, and not over 25 entire cocoons were found. A large number of the cocoons were partly completed. This tree had been under observation and no extensive mortality of the larvae on the leaves was noticed. There is no question that most of the larvae reached the ground, and most of these fell prey to their insect enemies before they could pupate. In collecting cocoons in the field in localities where there has been an outbreak of larvae and the trees have been practically defoliated, it is surprising to find relatively few cocoons that are entire and contain pupae. Although no detailed observations have been made on the activ- ities of birds, Dr. Britton informs me that he has observed certain warblers apparently feeding extensively on the larvae. While there is no question that birds do have some effect on the abund- ance of these insects, the effect of ants and other predaceous insects seems to be much greater. The interrelations of host, parasites, and predaceous foes have been very clearly described in the case of the fall webworm by Tothill (1922), whose conclusions are here briefly summarized, and many of the reasons for the occurrence of outbreaks and the following decline in numbers of this insect are applic- able to Bucculatrix canadensisella. Under normally balanced natural conditions the parasites are most effective and keep the host in an endemic and harmless state for a number of years. The predaceous enemies are also effective, for without their help 454 CONNECTICUT EXPERIMENT STATION BULLETIN 2CO the host might increase in spite of the parasites. The combined attack of parasites and predaceous foes reduces the numbers of the host, but at the same time the number of parasites is reduced, for a competitive struggle for food occurs among the species of parasites and among the members of one species. During the last few days of its life in the host, the parasite is much more destruc- tive to the host tissue than at any other time, and although several parasites may start life in one host, which is particularly the case when the host becomes scarce, the only individual that survives is the one which first reaches this rapidly destructive stage, the others perishing from lack of food. The predatory enemies apparently do not discriminate in favor of the parasitized larvae, and this also tends to reduce the number of parasites. Some species of parasites may become locally extinct, and not being strong fliers, do not come in again from the surrounding territory for some years. Any environmental change favorable to the host now gives it an opportunity to increase in the absence of a large part of its enemies, and it soon reaches a stage of great abundance. After a period of years the parasites, which have now found themselves provided with an abundant food supply, increase, and finally, with the aid of the predaceous foes, overcome the host and again reduce its numbers to an endemic state. Over a long period of years the result of these opposing factors is a series of outbreaks following each other at more or less regular intervals. When the host begins to decrease markedly, the parasites also begin to decrease, since they have more difficulty in finding the host, so during the decline of the host population there is not nec- essarily an increase in the percentage of parasitism. For exam- ple, during 1925 the parasitism of Bucculatrix canadensisella mining larvae by Closterocerus and Derostenus was 15.7 per cent, and the following year, in the presence of a very marked reduc- tion in the abundance of mines, the parasitism from these two species was increased only 4.4 per cent. When the larvae of Bucculatrix are abundant there may be expected up to 20 per cent parasitism in the mining instars and an equal percentage of parasitized pupae. To this must be added a heavy mortality due to predaceous enemies. There are also certain undetermined factors, possibly both internal and external, which prevent the development of the insect beyond the pupal stage and cause the mortality of a number of pupae. These last factors are more effective on the host than on the parasite. Aside from the effect of parasites, a considerable number of the mining larvae may sometimes succumb from some cause unknown to the writer. All the factors except parasites maintain a constant attack on the various stages of Bucculatrix canadensisella, and when the Bucculatrix population begins to decline, the severity of this attack is more keenly felt. A parasite population fluctuates BIOLOGY OF BIRCH LEAF SKELETONIZER 455 with a host population and has direct bearing on the periodic abundance of the host but cannot entirely eliminate it, as the para- sites decrease when the host decreases. A species of fungus belonging to the genus V erticillium has frequently been found growing on the dead pupae of this insect, and it was thought at first that this might possibly be the cause of these fatalities. Several attempts to inoculate healthy normal pupae with cultures grown on oat agar failed completely. The procedure followed was to make a small opening in the cocoon and expose the pupa within. A drop of water containing a sus- pension of the spores and mycelium was placed on the pupa, which was then set aside in a petri dish for future observation. Although a number of inoculations were made, in not a single case did an infection of the pupa develop, and it was concluded that the fungus concerned is entirely saprophytic. Several species of V erticillium are found on dead insects. I am indebted to Dr. McCormick of the Connecticut Experiment Station for determin- ing this fungus and for carrying out the inoculations. X. Geographical Distribution This insect, as far as reports in the literature and information acquired directly from entomologists indicate, is found only in the northern United States and in Canada. Its southern limit is North Carolina, and in Canada it occurs in New Brunswick, Que- bec, Ontario, Manitoba, Saskatchewan, Alberta, and British Col- umbia. Mr. Hutchings, of the Entomological Branch, Ottawa, informs me that it probably occurs up as far as the Yukon. It is recorded as far west as Minnesota in the United States. In Ontario, Quebec, New Brunswick, the New England States, New York, Michigan. Wisconsin, and Minnesota it is very common and sometimes appears in such numbers that the birches are defoli- ated. On the map (figure 34) is marked with a cross every locality from which I have definite records of the occurrence of the insect. According to data obtained from Sargent's "Silva of North America" (1896), the four native food plants (the paper, gray, yellow, and black birches) of Bucculatrix canadensis ella occur over a much wider area than that from which the insect is reported. The region occupied by these birches is shaded on the map. The paper birch (Bctula papyrifera) is very widespread and is a favorite food plant. It is found almost everywhere within the shaded region on the map, but it is not abundant west of the Rocky Mountains nor south of Minnesota, Wisconsin, Michigan, and New York. The red birch (Betula nigra) is not a common tree in northern United States and I have no records of its being attacked by this insect. Its range extends much further south 45^ CONNECTICUT EXPERIMENT STATION BULLETIN 288 70 J&. /»22<»r^.9'*v^_ <^^ N&. s^ 5\ ^*^i ^^N&. _/-^.rf^ «§t ^\. x-'^ /Ss- <«\-X\ \ >^ /na- /* *L^c\ ^^ /^ / //i^f^^yV \'^\ 3^l/\/ / / <^/^v^S r\0\ / / «7<2snc*i rP^ \ \wtf*"/ vkz?? «j. /\ ^ / /yTV 8 ^ rOMTTA \ \ *U< / / X V ^""/" wra \/ / mWAA ii^/RS?> ' \/ ^jl//// w/Y//Y/Y/w / I ^svtl/? ^wa7/7T~ ^i