HARVARD UNIVERSITY LIBRARY OF THE Museum of Comparative Zoology f^^-u /J-^D UNIVERSITY OF KANSAS DEC2 91958 SCIENCE BULLETIN UNIVERSITY OF KANSAS PUBLICATIONS University of Kansas Science Bulletin - Vol. XXXIX November 18, 1958 Lawrence, Kansas ANNOUNCEMENT The University of Kansas Science Bulletin (continuation of the Kansas University Quarterly) is issued in parts at irregular inter- vals. Each volume contains from 500 to 1,800 pages of reading matter, with necessary illustrations. Exchanges with other institu- tions and learned societies everywhere are solicited. All exchanges should be addressed to The University of Kansas Science Bulletin, Library of the Untverstty of Kansas, Lawtience, Kan, PUBLICATION DATES The actual date of pubHcation ( i. e., maiHng date ) of many of the volumes of the University of Kansas Science Bulletin differs so markedly from the dates borne on the covers of the publication or on the covers of the separata that it seems wise to offer a corrected hst showing the mailing date. The editor has been unable to verify mailing dates earher than 1932. Separata were issued at the same time as the whole volume. Vol. XX— October 1, 1932. XXI— November 27, 1934. XXII— November 15, 1935. XXIII— August 15, 1936. XXIV— February 16, 1938. XXV— July 10, 1939. XXVI— November 27, 1940. XXVII, Ft. I— Dec. 30, 1941. XXVIII, Pt. I— May 15, 1942. Pt. II— Nov. 12, 1942. XXIX, Pt. I— July 15, 1943. Pt. II— Oct. 15, 1943. XXX, Pt. I— June 12, 1944. Pt. II— June 15, 1945. Vol. XXXI, Pt. I— May 1, 1946. Pt. II— Nov. 1, 1947. XXXII— Nov. 25, 1948. XXXIII, Pt. I— April 20, 1949. Pt. II— March 20, 1950. XXXIV, Pt. I— Oct. 1, 1951. Pt. II— Feb. 15, 1952. XXXV, Pt. I— July 1, 1952. Pt. II— Sept. 10, 1953. Pt. Ill— Nov. 20, 1953. XXXVI, Pt. I— June 1, 1954. Pt. II— July 15, 1954. XXXVII, Pt. I— October 15, 1955. Pt. II— June 29, 1956. XXXVIII, Pt. I— Dec. 20, 1956 Pt. II— March 2, 1958 Editor Edward H. Taylor Editor for Vol XXXIX George W. Byers Editorial Board Edvi^ard H. Taylor, Chairman Charles Michener Paul Roofe E. L. Treece WoRTHiE H. Horr Parb:e H. Woodard, Secretary UNIVERSITY OF KANSAS SCIENCE BULLETIN (TO. f!?V5 7(1 DEC 2 9 195; "mir{ DEVOTED TO THE PUBLICATION OF THE RESULTS OF RESEARCH BY MEMBERS OF THE UNIVERSITY OF KANSAS Volume , XXXLX University of Kansas Publications Lawrence, November 18, 1958 PRINTED IN THE STATE PRINTING PLANT TOPEKA. KANSAS I 95B 27-2656 »'^^\ ^~,,^L Contents of Volume XXXIX NUMBEK PAGE 1. Additions to the Known Herpetological Fauna of Costa Rica with Comments on Other Species. No. III. Edward H. Taylor, 3 2. Notes on Costa Rican Centrolenidae with Descriptions of New Forms Edward H. Taylor, 41 3. Observations on the Ethology of Neotropical Anthoph- orine Bees (Hymenoptera: Apoidea). Charles D. Michener and Rudolf B. Lange, 69 4. The Flora and Ecology of the Chautauqua Hills in Kan- sas Earl W. Lathrop, 07 5. A Subgeneric Revision of the Genus Osniia in the West- ern Hemisphere (Hymenoptera: Megachilidae). Rancndra N. Sitiha, 211 6. Notes on Oviposition and the Hatching ot Eggs of Aedes and PsoropJiora Mosquitoes (Diptera: Culicidae). A. Ralph Barr and Ahdulla Al-Azawi, 263 7. A Revision of the Genus Osmia, Subgenus Centrosmia ( Hymenoptera: Megachilidae ) . Ranendra N. Sinha and Charles D. Michener, 21 b 8. The Comparative Morphology, Phylogeny and Higher Classification of the Butterflies (Lepidoptera: Papil- ionoidea) Paul R. Ehrlich, 305 9. The T enagogonus-Limnometra Complex of the Gerridae. H. B. Hungerford and Ryuichi Matsuda, 371 10. An Anatomical Study of a Neotropical Tree Frog, Cen- trolene prosoblepon (Salientia: Centrolenidae). Theodore H. Eaton, Jr., 459 11. Observations on the Behavior of Brasilian Halictid Bees, HI Charles D. Michener and Rudolf B. Lange, 473 12. Catalogue of the Types in the Snow Entomological Mu- seum. Part II (Mallophaga) K. C. Emerson, 507 13. Mitotic Activity in Allium After Treatment with Chromo- some-Breaking Chemicals. R. L. Hancock and M. Hancock, 519 D£C2 9i9^ '*.rt THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] Lawrence, November IS, 1958 [No. 1 Additions to the Known Herpetological Fauna of Costa Rica with Comments on Other Species. No. III. BY Edward H. Taylor Abstract: This paper treats f)f a series of Costa Rican Salientia. Two species, Hyla microcephahi mirrncephala Cope and Eleutherodoctyhts monni- chortim Dunn, are reported from Costa Riea for the first time. Tlie following species are descrihed as new: Dendrohates granuliferus, Eleutheroducttjlus taurus, Uijla legleri and Hyla dulcensis. The species Hijhi monticola Cope, long known only from the type specimen (now lost), is redescribed. Sixteen forms are illustrated with photographs. INTRODUCTION The collections made in Costa Rica after the completion of my work on "The Frogs and Toads of Costa Rica," * inclnde a large number of Amphibia. Certain of these have been studied and the results published in two supplemental works.** Others remain imstudied. In this, a third supplement, the following species are treated: Bufo holdridgci Taylor Bufo leutkenii Boulenger Bufo nichinochlori.s Cope Dcudrolnites atiratus Girard Dcndrohutes typographus iyeferstein Detulrohates grantdiferus sp. nov. Eh'utlierodactyhts monnichorum Dunn Eh'uthcrodactyltts taunts sp. now //(//« zctcki Gaige Hyla houlengeri Cope Hyla microcephala microcephala Cope * The Frogs and Toads of Costa Rica, Univ. Kansas Sci. Bull., vol. 35, pt. 1, no. 5, July 1, 1952. pp. 577-942, figs. 1-69. ** Additions to the Known Herpetological Fauna of Costa Rica with Comments on Other Species. No. I. Univ. Kansas Sci. Bull., vol. 36, pt. 1, no. 9, June 1. 1954. pp. 597-639, figs. 1-12; and Additions to the Known Herpetological Fauna of Costa Rica with Comments on Other Species. No. II. ibid. vol. 37, pt. 1, no. 13. October 15. 1955. pp. 499-575. figs. 1-18. (3) The University Science Bulletin Hyla incadoi Dunn Htjla monticolu Cope Hyla ^(ih})i Copt' Hylu ah'urodtn Taylor Hyla p.scudopunia Cimtlu'i Hyla legleri sp. iion . Hyla dtilcensis sp. ik)\ . Hyla clcachroa Cope TAXONOMIC TREATMENT Biijo hol(lri(l; p =41()6, snout- vent leret'i. 45.4; lower figure, No. •34105, snout-vent length, 45. Both from 15 km. WSW San Isidro del General, San Jose Province, C. R. Herpetological Fauna of Costa Rica 2.3 Hylo microcephala microcepJiala Cope (Fig. 9) Ilyla microcephala Cope, Proc. Aiiur. Pliil. Soc, Apr. 1886, vol. 23, no. 122,, pp. 281-282 (type locality. Department of Chiriqui, Panama, "along a mountain stream"). A large series of specimens of Hylo m. microcephala KUMNH Nos. 32172-32206 was taken at Golfito, Puntarenas Province, and represent, I believe, the first record of the species for Costa Fig. 9. Hyla microcephala microce))hahi Cope. L pper figure, KUMNH No. 32184 A : aetual snout-vent length, 22.2 mm; lower figure. No. 32172 9, length, 28.6. Both from C^)lfito, Puntarenas Province, C. R. 24 The University Science Bulletin Rica. Specimens were breeding in small temporary pools one to five inches deep, September 6-9, 1952. Diagnosis: A diminutive species characterized by a very much widened third finger; a pair of brown lines from anterior part of eye running down middle of back; a narrow dorsolateral brown line; toes four-fifths webbed; outer fingers one-third to two-fifths webbed; head small; vomerine teeth between choanae not or scarcely extending behind their posterior level. Description of KUMNH No. S2172: Head width (8.5 mm.) slightly greater than length (8.2 mm.), not or scarcely greater than body width; width of eyelid (1.9 mm.) smaller than interorbital width (2.6 mm.); canthus rostralis rounded, curving; loreal region slightly concave; distance between eye and nostril (2.2 mm.) equal to distance from nostril to median notch in upper lip, much less than length of eye (3.1 mm.); tympanum distinct its diameter (1.5 mm. ) twice in length of eye-opening; separated from eye by a dis- tance less than its diameter; its upper part concealed by the distinct supratympanic fold. Vomerine teeth abnormal (normally vomerine teeth on two low moundlike swellings between choanae, closer to each other than to choanae; usually smaller than choanae and scarcely reaching their back level); tongue longer than wide, large, flat, narrowly free behind and along sides, with a slight median notch behind; palatal groove distinct, a little closer to anterior level of choanae than to front of palate; (male with a large external vocal sac; vocal slits large). Arm brought forward, the wrist reaches end of snout. A web remnant between two inner fingers; between second and third, two-fifths webbed, to pad of second; between outer fingers about one-third webbed to pad of outer finger. Digital pads of three outer fingers large, subequal, of inner finger smaller; distal sub- articular tubercles single on first and second fingers; lateral folds on digits somewhat thickened; a large flat inner metacarpal tubercle; palmar tubercle smaller, double or bifid; no distinct skin-fold under forearm; seen from above, fourth finger widened. Leg brought forward the tibiotarsal joint reaches the anterior level of eye. Toes almost entirely webbed to digital pads, which are smaller than those on fingers, except pad on inner; inner meta- tarsal tubercle elongate, oval, flattened; outer tubercle if present, indistinct. When legs are folded at right angles the heels overlap three millimeters. Herpetological Fauna of Costa Rica 25 Skin on head, sides, and upper part of limbs smooth; a somewhat thickened, nearly straight supratympanic fold, continuing straight back from eye, can be traced to above arm onto side ( usually cream colored); chin smooth (in males a very ample vocal sac); venter with distinct granules; much of undersurface of thighs granular; a slight fold across breast interrupted mesially; small axillarv web present; area about vent smooth; the muscles surrounding vent give the area an inflated appearance. Color in life: On exposed surfaces, cream with scattered brownish pigment but forming two moderately distinct lines on body from eye to end of rump; a more or less distinct cream line from tip of snout along canthus rostralis, continuing behind eye to near groin; upper part of tympanum pigmented; lip cream, the loreal region brownish; a brown line present below the cream dorsolateral line, becoming obsolete posteriorly; venter and concealed surfaces cream or creamy flesh. Measurements in nun: (KUMNH Nos. 32172$, 32173$, 32176 J , 32180 ^ , respectively ) : Snout to vent, 28, 28.7, 24.8, 25; width of head 8.5, 8.7, 8, 8.1; length of head, 8.2, 8.2, 7.7, 8; arm, 17.2, 18, 14, 13.5; leg, 43, 46, 37.5, 39; tibia, 14.8, 15.3, 12.8, 13; foot and tarsus, 20, 21.5, 17.6, 17.2. Remarks: This species appears to replace Hyla underwoodi in southwestern Costa Rica. The latter species is common farther north in the northern part of Puntarenas and Guanacaste Province, where I obtained large series in 1953 and 1954. It differs from H. nndenvoodi in having a different color pattern, with hand, foot, and digits more slender. Hijla phlebodes, a mem- ber of the same group, known from the eastern Meseta Central drainage areas, is not uncommon at Turrialba, Cartago Province. It is a smaller species with a distinctly broader head and propor- tionally larger eyes. A brown interorbital line is present. The widened fourth finger is a characteristic of this group of Hyla. Ilyla picadoi Dunn Hyla picadoi Dunn, Copeia 1937, no. 3, p. 164 (typi- locality, a little above farm of Manuel Acosta 2140 m., Volcan Barba, Heredia Province). A specimen of this species, KUMNH No. 36127, has been found in the Talamanca range at the Bosque Nacional (Pan American Highway), elevation about 8000 feet. The paired denticulations in the front of the lower jaws, while e\ ident are low. 26 The University Science Bulletin Hyla monticohi Cope (Fig. 10) Ilijla punctariola monticola Cope, Joum. Acad. Nat. Sci. Philadelphia, ser. 2, vol. 8, 1875, p. 106 (type locality "Cordilleras at from 5000-7000 feet" ele- vation, slopes of Pico Blanco, Costa Rica). The type of this species is presumably lost. The original speci- men was probably in bad condition, since it was not figured and the type description is very brief and inadequate. Fig. 10. Iltjlu monticola Cope. KUMNH No. 36764 9, Cocales Creek, Suretka, Linion Province, C. R. Actual snout-vent length, 43 nun. A species of Hylo which I took at Suretka in southeastern Costa Rica, at an elevation of perhaps 300 feet, agrees with the color description of H. monticola in being transversely banded on dorsum, a condition not normally found in other Costa Rican frogs. Other similarities and differences are discussed following the description of this specimen, No. KUMNH No. 86764 $ , collected June 26, Herpetological Fauna of Costa Rica 27 1954, along Cocales Creek, Suretka, Limon Province, Costa Rica, by Edward H. Taylor. Diagnosis: Related to Hijla gahhi. Eye longer than the distance between eye and nostril; width of eyelid less than the interorbital distance; canthns rostralis, distinct, ronnded; greatest diameter of tympannm about one half of the length of eye; skin glassy smooth; tympanum smooth; a supratympanic fold; area about vent smooth; chin indistinctly granular, breast venter and under thighs with strong granules; outer fingers two-thirds webbed; two inner fingers webbed only at base; tibiotarsal articulation reaches one millimeter beyond snout; snout-to-vent length, 43 mm. Whitish on chin; venter, and under thigh, without pigment; under part of foot darkly pigmented; hands below dirty white; brownish gray above, the back with four black bars; the tibia strongly banded, the thighs less distinctly so. Description: Female. The width of head (17 mm.) is greater than length (14.1 mm.); canthns rostralis distinct, rounded, if projected would form an angle with its fellow between nostrils; loreal region slightly concave, sloping rather sharply to lip; snout rounded at tip, scarcely extending beyond the edge of lip; nostril a fourth nearer to the median point on upper lip than to eye; length of eye-opening (4.8 mm.) greater than distance from eye to nostril, (4.15 mm.), shorter than snout length (6 mm.); tympanum, 2.1 mm. in vertical diameter, 1.9 mm. longitudinal diameter; supra- tympanic fold curving up slightly, runs from eye, back to above arm, separated from tympanum posteriorly; tongue about as long as wide, not or scarcely notched behind, without a free posterior or lateral part; choanae large (1.3 mm.); a pair of strongly elevated ridges directed backward, arise from anterior inner edge of choanae extending back beyond the level of the middle choanae, touching mesially, and each bearing at least six teeth posteriorly; palatal glands open separately much anterior to level of choanae, near midline of the palate without forming a transverse groove; openings of Eustachian tubes smaller than choanae. Arms with a small axillary web; upper arm slender, the forearm much heavier, three outer fingers two-thirds webbed; two inner fingers with a small basal web, the digits with slight lateral ridges or fringes to tip; inner metacarpal tubercle elongate, narrow, com- pressed; two small subecjual palmar tubercles; subarticular tubercles small, that on outer distal joint of fourth finger double; palm with numerous tubercles; leg long, the tarsal joint extending about a milli- meter beyond tip of snout when leg is brought forward; toes fully 28 The University Science Bulletin webl)ecl, the web reaching bases of terminal discs, that on fourth toe somewhat narrowed toward the disc; discs on three outer fingers larger than those on toes, that on inner finger smaller than those on toes; metatarsal tubercle elongate, compressed, extending out from edge of foot; no outer tubercle; a slight lateral fringe on outer edge of outer toes; a distinct tarsal fold. Skin abo\e shiny smooth; sides somewhat wrinkled, with granules on lower part; chin nearly smooth; breast indistinctly granular; belly with strong granules; under thigh granules larger, a few with small median dark dots; an area about vent smooth; an indistinct groove from behind vent passing down under thigh; tympanum smooth without granules. Color: Above rather brownish to lead-gray with frontal region somewhat more brownish; a dark band between eyes, one across scapular region, one across the middle of back and one across the sacrum; tympanum brownish; an area below eye with less pig- ment; sides lighter with indication of some lighter spots in the groin; front of thigh nearly uniform light gray; back of thigh similar; top of thigh with some darker markings; tibiae with five blackish bars; foot dark gray above; foot dark below with considerable pigment; hand whitish below. The upper arm is white save for a dim line of grayish pigment on its outer surface. Measurements in mm.: Snout to vent, 43; width of head, 17; length of head, 14.1; arm, 26; leg, 74; tibia, 26; foot, 31.2. Variation: The color as given in the type description follows: "Color light grayish brown with large dark spots forming transverse bars, one between the eyes, one in front of the scapulae, one behind the scapulae, and one at sacrum. Below unspotted white. Limbs with light brown surfaces above; concealed surfaces, pale, un- spotted. No inguinal spots; a few specks of brown on the sides." Other data given on the type are: "Length of head and body .037 m. fingers not entirely free but a web extends between the outer two to the middle of the phalange. The area of the tympanum in the same, is one-fourth that of the orbit. The head is short and wide and the heel extends nearly or (juite to the end of the muzzle." It will be seen that the two specimens disagree in the degree of development of the web on hand, and the absence of the markings on the limbs. One supposes that the specimen was in bad condi- tion perhaps dried or hardened by preservatives so the webbing may have appeared to be absent between some of the fingers. Until specimens of H. monticola are obtained from the type lo- Herpetologicai. Fauna of Costa Rica 29 cality the identity of this specimen may remain somewhat in donbt. The specimen described was obtained from a shrnb in the shallow, fast-flowing Cocales Creek at Suretka. Hyla ^a])])i was breeding in the stream as were certain other species of amphibians. Hyla gabbi Cope (Fig. 11) Hyla fiubhi Cope, Joiirn. Acad. Nat. Sci. Philadelphia, ser. 2, vol. 8, 1875, p. 103 (type locality, near Sipurio, Linion Province, SE Costa Rica). Specimens of this large Hyla are in the collection from the follow- ing localities: KUMNH Nos. 28200-201 San Isidro del General and Fk;. 11. Hyla gabbi Cope. KUMNH No. 28201, San Isidro del General. San Jose Province. C. R. Actual snout-vent length, 49.5 mm. 34869-75 Escazu, both, San Jose Province; No. 37012, lAIA, Tnrri- aJba, Cartago Province; Nos. 36764-78, 36791-93 Cocales Creek, Suretka, Limon Province; Nos. 32994-32999, 34863, 34869-75, 36791-93 near Sarctii, .\lajuela Province. 30 The University Science Bulletin Hylo (ilvaradoi Taylor (Fig. 12) Htjhi (ilvaradoi Taylor, Univ. Kansas Sci. Bull., vol. 35, pt. 1, no. 5, July 1, 1952, pp. 882-885, fig. 63 (type locality KUMNH Nos. 31864-6,5 Moravia de Chirripo, Limon Province, C. R.). Two specimens, KUMNH Nos. 31864-65, were taken very close to the exact type locality. One, a gravid female, was ready to lay eggs, since some eggs were extruded after capture in the collecting Fig. 12. Hyla alvaradoi Taylor. KUMNH No. 31865, topotype. Moravia de Chirripo, Limon Province, C. R. Actual snout-vent length, 42.3 mm. bag. The preserved eggs measure from two and one-half to three millimeters in diameter and are without pigment. Under a lens the dorsum shows a peppering of minute black specks with a few scattered ones on arm and leg. The venter is only faintly granular. The male has a pair of long vocal slits, the Herpetological Fauna of Costa Rica 31 edges of which are whitish. The pigment is sHghtly more evident on the back, definitely so on the Hmbs. The white hne behind the eye is somewhat less distinct. The curious fold, arching above the vent is present in both of the topotypes. Coloration in life of No. 31865: "Bright yellow-green above, light lines yellow becoming cream over eyes; eyes bronze-brown; throat and concealed parts of limbs, bluish-flesh; dorsal part of limbs nearly flesh-color, slightly tinted lemon-yellow; under surfaces of hands and feet canary to lemon-yellow; anal fold cream." The color faded within a few days in preservative, leaving the specimens nearly cream-white. The life colors of No. 31864 varied somewhat from the preceding as follows: "The dorsum is yellowish with a faint wash of green. The head is bluish green to greenish, generally a little darker on the middle and sides; venter with a whitish or yellowish white area; mental gland yellowish; concealed parts of limbs indefinite flesh; knees bluish; under surface of fingers yellowish, of toes whitish flesh without yellow." Taken June 20, 1952. The specimens were taken from shrubs at the edge of a small stream only a few meters from where the type was taken. Hyla pseudopuma Giinther (Figs. 13, 14) Hyla pseudopuma Giinther, Biologia Centrali-Aniericana; Reptilia and Ba- trachia, Sept. 1906, pp. 274-275, pt. 72, figs A. and B. (type locality, La Palma, San Jose Province [1600 m. elev.]). This species presents problems in identification since the color and pattern is extremely variable, the color changing from nearly blackish to a light olive in a few hours. An hourglass-shaped pat- tern on the back, considerably darker than the adjoining back- grovmd, may be present, only to disappear somewhat later. Speci- mens from near the type locality usually have a dark lateral stripe even when the dorsum is a light olive. The females are often considerably larger than the males and the "hourglass" pattern may appear as two separate darker-edged marks on a light, often fawn, ground-color. I am presenting figures of two of these variations. Specimens have been taken from the following localities: KUMNH Nos. 34973-80 southern slopes of Volcan Po:ls, elevation circa 6800 ft., Alajuela Province; Nos. 36583-96 "in bromelias," Vara 32 The University Science Bulletin Blanca, southeastern slopes, Volcan Pons, Heredia Province; Nos. 36597-600, La Palina between Volcan Barba and Volcan Irazu 4800 ft. (in bromelias), San Jose Province; Nos. 34981-84 Finca Montecliristo, Volcan Barba, (south slope), Heredia Province; Nos. 28192, 32276-32279, Cinchona ( Isla Bonita), 4000-5500 ft. elev., Heredia Province; Nos. 32280-81 (large blotches), 32282-32291, 34971-34972, Moravia de Chirripo, Limon Province (elevation circa 2300 ft.). Fig. 13. Hyla pseudopuma Giinther. KUMNH No. 37030 9, Moravia de Chirripo, Limon Province, C. R. Actual snout-vent length, 44 mm. Herpetological Fauna of Costa Rica 33 Fig. 14. Hyla pseiuhpuma Giintlicr. KUMNH No. 32281 9. Monixia df Chirripo, Linion Province, C. R. Actual snout-vent length, 40 nun. Hylo Je^leri sp. nov. (Figs. 15-16) Tijpe: KUMNH No. 32982(5; eollected approximately 15 km., W'SW San Isidro del General, San Jose Province, al night with paratvpes, in a small swamp, July 10, 1952, by Edward H..Ta\lor. P«raf (/pes; KUMNH Nos. 32981, 32983-32992 all males, same data. No. 35171, female same data; all topotypes collected by Edward H. Taylor and John Baker. Diagnosis: A red-eyed species with a inaxiinum known snoiit-\ent length for males of 34, of female 36 millimeters; inner fingers less than one-fourth webbed; outer fingers approximately one-third webbed; toes nearly four-fifths webbed; tarsal fold not or but faintly indicated; male with vocal sac and paired vocal slits; vo- 2—2656 34 The University Science Bulletin merine teeth in two separated groups between choanae; interorbital distance greater than width of an eyeHd; heel to anterior edge of orbit; venter, chin, and breast granular; tympanum about equal to area of largest finger disc; blackish to dark brown above with or without indistinct mottling; on sides the color often limited by a distinct yellow line from upper arm, along side, to groin; a cream Fig. 15. Hyhi Icgleri sp. nov. KUMNH No. 32983 i . 15 km. W SW San Isidro del General, San Jose Province, C. R. Actual snout-\cnt length, 33 mm. line above vent and across heel; a fold above tympanum obscures its upper border. Description of type: A slight depression on the snout behind level of nostrils to level of eyes; head wider than body, its width at tym- panum (11.6 mm.) nearly equal to its length (12 mm.); length of eye-opening (3.4) about equal to distance between eye and nostril; latter a little closer to median point on lip than to eye; interorbital Hebpetological Fauna of Costa Ric;a 35 space (4 mm.) greater than width of an eyeHd (3.2 mm.); greatest diameter of tympanum (1.9 mm.) approximately equal to distance from eye; canthus rostrahs distinct, rather sharp, curving shghtly, the loreal region shallowly concave sloping broadly to lip. Vomerine teeth on two low transverse ridges between choanae, separated from each other by a smaller distance than from choanae; choanae rather small not larger than \omerine tooth-groups; open- ing of palatal glands in a slightly sinuous transverse groove, much closer to front of palate than to upper level of choanae; tongue about as long as broad (6.5 mm. ); a slight posterior notch indicated; vocal slits nearly parallel; vocal sac indicated externally by a trans- verse fold on throat in front of breast. Arm rather short, thick, and cannot be straightened without breaking muscles; first finger distinctly shorter than second, with numerous brown spiny rugosities forming a broadly triangular area reaching to terminal pad; terminal pads of digits widened, those of three outer fingers much larger than that of first; first two fingers less than one-fourth webbed; three outer fingers about one-third webbed; inner metacarpal tubercle elongate; subarticular tubercles small, those on two outer fingers may be double or bifid; a slight fold on undersurface of forearm; toes about four-fifths webbed; inner metatarsal tubercle moderate, somewhat elevated; outer tu- bercle very small, indistinct; pads on toes smaller than those on three outer fingers, larger than pad on inner finger; heel reaches to front of orbit; legs folded at right angles, the heels overlap three millimeters; no tarsal fold evident. Skin almost completely smooth abo\'e, on sides of Neuter, and underside of thighs; less distinct!)' so on breast and chin; some granules, on sides of vent, the posteriormost pustular, much the largest. Color in life: Dark blackish to blackish brown above on head and body; throat grayish white; ^yellowish green on venter and under thigh; heels, soles, and undersurface of toes blackish green; palms greenish yellow; a short line of cream on upper arm continued on to side of body to groin; a narrow yellowish line borders upper lip; back of thighs brownish olive; eye red; a narrow cream line above vent and across elbow and heel; tibia with indistinct darker bands. Measurements in nun. (Type, No. 32981 J' and No. 35171 $ respectively): Snout to vent, 33, 34, 37; width of head, 11.6, 11.9, 36 The University Science Bulletin 12.4; head length, 12, 12, 12; arm, 19, 18, 22.5; leg, 45, 48.5, 56.2; tibia, 17.2, 17.1, 18.7; tarsus and foot, 23, 22.5, 26. Variatkrn: Some of the specimens have the venter and concealed parts lemon-yellow. The cream lateral line is uneven, sometimes broken. When preserved they are dark to blackish brown, with Vu:. 16. Hyla legleri sp. nov. KUMNH No. 35171 9, 15 km. WSW San Isidro del General, San Jose Province, C. R. Actual .snout-vent length, 37 mm. Herpetologkal Fauna of Costa Rk.a 37 some black spots or flecks visible on back; indistinct bands are evident on tibia. The female specimen No. 35171 differs very markedly in color, being fawn with a flecking of brown spots often contiguous forming indefinite spots; fingers and toes with light pigmentation appearing dirty white in preservative. The yellow line, distinct on upper ann, merges with the ventrolateral coloration; yellowish cream on imder surfaces. The specimen is gravid and the eggs are visible. Retmirks: This species would appear to be related to Hijla ni^ripes Cope taken at an elevation of 50(M)-60(K) ft. on Pico Blanco, Costa Rica. The nostril, however, is closer to the median point on lip than to eye; the tympanum is larger, equal to one third or more of the exposed part of eye ball; the foot is about one-third rather than half -webbed; the leg is shorter, the heel reaching only to front edge of orbit rather than beyond the snout. There is usually a cream lateral line rather than white spots in a dark reticulum. The elevation of the type locality is probably not more than 600 ft. above sea level. The female figured looks very different from the males. How- ever, I find no pertinent structural differences. They were taken in the same small swamp at night, in rain. The species is named for John Legler, ('urator in Charge of herpe- tological collections at the Kansas University Museum of Natural History, who has greatly facilitated my study there. HijJa dulcensis sp. nov. (Fig. 17) Type: KUMNH No. 32168; collected, Golfito, Puntarenas Province, Costa Hica, Sept. 7. 1952 by Edward H. Taylor. Pdrattjpes: Kl'MNH Nos. 32166, 32167; collected with the type, same locality, Sept. 7 and 8, 1952 b\ Edward H. Taylor and John Baker. Diapwsis: Web barely indicated between three outer fingers; webs between first and second toes attached only at the base of first toe; other toes about three-fourths webbed; tibiotarsal joint reaches to front edge of eye; a very slight a.xillary web; subarticular tubercles single. Bones blue (in life). Related to Hyla eleachroa but differs in its .somewhat larger si/e (40 mm. for males) smaller finger and toe discs, the obsolete canthus rostralis, the loreal region not concave, and the choanae larger. 38 The University Science Bulletin Description of type: Head somewhat flattened, its length (14 mm.) a little greater than its width at tympanum (12.9 mm.); canthus rostralis not indicated, the loreal region sloping obliquely from top of snout to lip with no (or but the slightest) concavity; nostrils lateral, the region about them swollen, with a depression between them on top of snout; in front of nostril the line of profile Fig. 17. Upper figure, Ihjla dulcensis sp. nov. KUMNH No. 32168, Type; Golfito, Puntarenas Pro\iiice, C. R. Actual snout-vent length, 39.5 mm. Lower figure, Ilula deachroa Cope. KUMNH No. 31804, Turrialba, Cartago Province, C. R. Actual .snout-vent length, 35 mm. Herpetologic:al Fauna of Costa Rica 39 curves somewhat forward then back and down to Hp; snout extended beyond mouth 1.4 milHmeter; length of eye opening (4 mm.) less than distance between eye and nostril (4.5 mm.); interorbital width (3 mm.) less than width of an eyelid (3.6 mm.), greater than dis- tance between nostrils (2.6); tympanum large (2.2 high x 1.9 mm. long), separated from eye by a distance of two millimeters, its upper rim partly concealed by an ill-defined fold from eye which runs diagonally backwards and down to above arm insertion. Vomerine teeth on two small closely approximated moundlike elevations between choanae and separated from them by a distance equal to length of one elevation, the teeth arranged in slightly diagonal lines pointing forward mesially (nearly transverse in a paratype ) ; diameter of a choana in distance between choanae about two times (or slightly more). Openings of the vocal slits large; tongue with only a small free edge, longer than wide, not notched posteriorly; palatal glands open in a sinuous groove, much closer to anterior limit of palate than to anterior level of choanae. Hand with a web-remnant between three outer fingers, the webs continuing somewhat as small lateral fringes or ridges; be- tween first and second fingers only the ridges are evident; discs on fingers rather large, that on third finger as large as tympanum; subarticular tubercles large, single; inner metacarpal tubercle elon- gate, partly covered by the nuptial gland, which seemingly lacks horny spicules on its surface; the palmar tubercle elevated, divided mesially, the inner part a little the longer. Legs rather short; the tibiotarsal articulation reaches the anterior edge of the eye; when legs are folded at right angles to the body the heels overlap a little more than four millimeters. Web between first and second toes slight, attached to first in basal part only; four outer toes about three-fourths to four-fifths webbed; discs a little smaller than on fingers; inner metatarsal tubercle rela- tively small; the outer smaller but e-ntirely distinct; no inner or outer tarsal fold. Skin on dorsal surface of body and head minutely granulate or corrugate (smoother in female paratype), but nearly smooth on arms and upper parts of limbs; sides nearly smooth; skin on chin covering vocal sac much folded; a small curving fold across breast; venter strongly granulate or areolate; anterior ventral surface of thighs, and to a lesser extent the posterior surface, granulate. Color: — In life, grayish brown on dorsal surfaces, the venter white; the throat sac and chin yellow; legs generally greenish yel- low. In preservative, very light brown above, the pigment (in 40 The University Science Bulletin chromatophores ) rather evenly scattered over most of the hmb surfaces ( a narrow ventral area excepted ) , and the upper and under surfaces of hands and feet; upper lip cream with some scattered pigment. The bones are blue or greenish blue in life but this color tends to fade after preservation. Measurements of Hyla dulcensis and HyJa eleachroa H . dulcensis H . eleachroa Number 32166 32167 32168 30340 30313 30326 Sex 9 cf d" 9 d" cf Snout to vent 40 13. o 37 13.5 39 14 34 12 33.5 11.5 33 Head length 11 Head width 13.6 12.7 12.9 11 11 U) 7 Length of eve 4 4 4 3.8 3.8 3.7 Eye to nostril 4.3 24 4.3 22 4.5 24 3.7 20 3.9 19.7 3.8 Arm 19 Leg (to vent ) 61.5 21 60 20.5 61.4 21 54 17 52 18 49 Tibia 17.6 Foot and tarsus . . 26.6 27 28.4 24 24 23 . 5 Renmrks: This species belongs to a hylid group having the web largely eliminated between the first two toes. Other members of this group in Costa Rica are Hyla houlengeri, H. stauffcri and H. eleachroa. It is probably most closely related to eleachroa. The type locality Golfito meaning "little gulf" refers to the Golfo Dulce on which the town is located. The species is named for the Gulf. THE IJxNIYERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 2 Notes on Costa Rican Centrolenidae With Descriptions of New Forms BY Edward H. Taylor Abstract: The Costa R'can species of this taniily are discussed and dis- tributional data hsted. Four species are described as new: CochratwUa chirripoi, Cochranelhi milUpunctata, Cochranella decorata, and CochrancUa reticulata. A key is given. Thirteen of the species are ilKistrated with photo- graphs. INTRODUCTION Four species of diminutive frogs taken since the preparation of my paper "A Review of the Frogs and Toads of Costa Rica" are referred to the family Centrolenichie. Seemingly they are new and are described herein. In order to faciHtate identification of these and other species of this difficult group, I append a brief key to the Costa Rican centrolenids and list such distributional data for the members of this family as are presented by the material in my collections. The Costa Rican species of the Centrolenidae have the following characteristics in common: The astragalus and calcaneum are fused into a single slender element; the eyes are large, usually protruding beyond the edge of the mouth and directed forward to a greater or lesser extent; the head is usualK' but perhaps not invariably wider than the body. The usual dorsal coloration in life is some shade of green, but in preservative the chromataphores become lavender (magenta to purplish). In the dentate species this pig- ment is seemingly less soluble in preservatives than in the species lacking vomerine teeth. In certain forms, and more especially so in males, the forearm is nuich widened in comparison witli the upper arm, and the musculature tends to prevent the straightening of the member unless the muscles or tendons are torn. In lite the (41) 42 The University Science Bulletin arms seem to be less flexed permanently. In many of the species the skin and muscnlature of the venter, often also of back and limbs, are so transparent that the viscera and bones may be seen through the flesh. Several species have the eyeball surrounded by a creamy or whitish membrane which may be seen through the skin of the upper eyelid or seen from within the mouth. In others the eyeball is black. In CochraneUa chirripoi there was seemingly some variation as if this tunic might be retracted to some extent. This I cannot determine. In many species much of th(^ intestine has a cream- white covering. After a specimen has been preserved for some time the brain often can be seen with considerable clarity and certain muscles will appear different in color from other muscles, thus suggesting a color pattern. All of the species have the terminal phalanx of the digits Y-shaped. T-shaped, or straight, and the variation may occur on different digits of the same specimen. They are never clawlike as in the Hylidae, but there is an intercalated cartilage ])etween the two terminal bones of the digits (as is true of the Hylidae and certain other arboreal frogs). Key to Costa Ricax Centrolenidae ( based on preserved specimens ) 1. Vomerine teeth present. In preservatives, considerable diffuse lav- ender pigment (magenta to purple) giving specimens a general lavender color 2 Vomerine teeth absent. Lavender pigment tending to dissolve in preservatives leaving specimens without pigment (except in eye), 6 2. A bony process on humerus in males ( rarely visible in certain fe- males); part of lavender pigment tends to form small punctate spots ( often absent ) ; 28 mm Centrolene prosoblepon No bony process on humerus visible externally 3 3. A pollex rudiment, in males bearing a sharp spine; 20 mm., Teratoliyhi spiiiosa No pollex rudiment, bearing a sharp spine in males. 4. Skin smooth on dorsum; no web-remnant between first two fingers; outer fingers four-fifths webbed; dorsum with minute lavender flecks and fine scattered white punctations; toes entirely webbed; 24 mm CochraneUa ptilverata Skin more or less distinctly granular on dorsimi; size larger 5 5. Dorsum dark lavender abo\e enclosing discrete rounded cream-white spots; this color also on upper part of thighs, tibia, and tarsus, where the cream spots are larger; legs slender, elongate; terminal Costa Rican Centrolenidae 43 pads rathtr widened; eyes very large, liead wider than bocU ; inter- orbital distanee less than widtli of an eyelid; 29.5 mm., CochrancUa albornacukita Dorsum light Ia\ender with numerous (40-50) dark rounded lavender spots, absent on arms, thighs, and tarsus; present on tibia; interor- bital distanee much greater than width of an eyelid; terminal pads on fingers widened; dorsal granules stronger; .30.2 mm., CocliraneUa granulosa 6. Tympanum concealed; areas about nostrils flattened, not swollen or inflated 7 Tympanum visible though co\ered with a thin pigmented skin; areas about nostrils swollen; the snout, sometimes nostrils, extending be- yond mouth level JO 7. A pair of discrete folds, lateral to Nent, nm diagonally back and down on thigh CochrancUa dccorata No pair of distinct folds in region of vent 8 8. Terminal digital pads enlarged, rounded, distinctly wider than ad- joining part of digit; a white eye tunic, 23.8 mm., CochrancUa niiUcpunctata Terminal digital pads truncate or subtruncate, not or only a little wider than adjoining part of digit 9 9. A distinctive white eye tunic, \isible through upper eyelid as a white spot; 25 mm CocJirancUa flcischnianni A black eye tunic, the upper eyelid appearing dark; hand proportion- alK larger; 24 mm CochraneUa chnjsops 10. W'vb between second and third fingers equally as extensive as that between third and fourth. I'sually a white eye tunic; 25 mm., Ct)chrancUa chirripoi Web between si-cond and third fingers not as extensive as that between third and fourth 11 11. A coarse reticulation of la\ender enclosing irregular light areas on dorsum and lim})s; nostril area moderatcK' inflatetl, not extended ((uitc to 1<'\(1 of mouth; vyv tunic dark; txnipanmn large, directed upward; a strong tar.sal fold, and a strong fold from wr'st to elbow; most subarticular tubercles double; 25 mm. CocliraneUa reticuhita No coarse reticulation enclosing light areas; area about nostrils some- what more inflated 12 12. Interorbital distance double width qf eyelid; visible part of tympanum one sixth of eye length; nostrils protruding bexond level of mouth; tibiotarsal articulation reaches tip of snout; no pigment on thighs; 22 mm CochrancUa valerioi Interorbital distance not twice width of eyelid; tympanum larger. . . 13 13. Eye tunic white; a reticulation on dorsum tending to enclose lighter spots in vertical series; area about nostrils inflated l)ut not extend- ing beyond lip; tympanum directed somewhat upwards; width of eyelid in interorbital distance, twice; 22 mm., CocliraneUa tahimancac Eye tunic black, choauac relatively large, interorbital distance one- fourth greater than width of eyelid; 26 mm., CochrancUa colymhiphylluni 44 The University Sc:ience Bulletin TAXONOMIC TREATMENT Centrolene prusoblepon (Boettger) (Fig. 1) This species, with a type locaHty "plantage Cairo" (La Junta) near Liinon, Linion Province, Costa Rica, has a considerable range in Costa Rica and is Hkewise known from Panama and Barro Colo- rado Island. It occurs on lx)th Caribbean and Pacific slopes. It also has a considerable vertical range, occurring from near sea level to elevations of from 4000 to 5500 feet. Fig. 1. Centrolene prosoblepon (Boettger). KUMNH No. 32932 9 (showing traces of the humeral processes). Tiirrialha, Cartage Province, C. R. Actual snout-\ent length, 20 nun. Specimens in the Kansas collection are: Pacific drainage. KUMNH Nos. 32390-93, 12-15 km. WSW of San Isidro del General San Jose Province; Nos. 36874-76. Agua Buena near Cafias Gordas, 2000 -f ft. elev., Puntarenas Province; Nos. 32927-32, 36877, near Sarchi, Alajuela Province. Costa Rk an Centrolemdae 45 Caribbean drainage. KUMNH Nos. 32419-38. 36873, 37016, Cin- chona, sontheastern slope Volcan Pons 5000-5500 ft. elev., Heredia Province; Nos. 32803-05. 32806, Los Diamantes, near Guapiles, Limon Province; Nos. 30403, 32376-81. 32926. Moravia de Chirripo. Limon Province. Nos. 32396-32416. Cariblanco. Limon Province; No. 32.395 Cartago, Cartago Province; Nos. 32-385-89, near Pacayas, Cartago Province; Nos. 32382-84. Santa Cruz, Cartago Province. The lavender pigment in these, once the specimens were pre- served, is practically unchanged after ten years. One female, KUMNH No. 32932. has been found showing a small humeral process, visible externally. Teratolujin spinoso (Taylor) (Fig. 2) Centrolenclla spinosa Taylor, Univ. Kansas Sci. Bull., vol. 33, pt. 1, Apr. 20. 1949, pp. 259-262 (typo locality, Los Diamantes near Guapiles, Limon Province, Costa Rica ) . Three more topotypic specimens, KUMNH Nos. 32933-35. were taken at Los Diamantes. Limon Province, within a few meters of the exact type locality, August 13, 1952. In life the colors were: "Body nearly uniform green above, yellowish on sides of head; ground color flesh; venter greenish or transparent flesh; terminal digital pads yellowish; iris dark without whitish flecks. The bones, visible through the fle.sh. green." Fu;. 2. Teratohyla spinosa (Taylor). KUMNH No. 32934 topotype, Los Diamantes. near Guapiles, Limon Province, C. R. Actual snout-\ent length, 20.2 nun. 46 The University Science Bulletin I have examined a specimen of this species from Barro Colorado Island, Canal Zone, identified as C puhcruta. it ma>' be presnined that it occurs in suitable localities along the eastern lowland ot Costa Rica and into Panama. The lavender pigment, visible after preservation (after the green color disappears), is less soluble than in many species of Coch- ranclla. The types appear practically unchanged after nearly ten years of preservation. CochnineUa pidverata (Peters) (Vi^. 3) Htjld puherutu Peters, Monatsb. Akad. Wiss. Berlin, 187o, j). 614 (type locality "Chiriciiii" Panama). I have traced this species to the north on the Pacific slopes as far as Sarchi, Alajuela. Dunn reports the species in Colombia and Ecuador. He also reports a specimen from Turrialba, Costa Rica. I have not seen this specimen, but I suspect that this may represent the related species CochrancUa alhomacuhita. I have not seen his South American specimens of C. pulvcratu. There are five specimens in our collection: KUMNH Nos. 36871- 72, 32822-23, are from Sarchi, Alajuela Province; No. 32821 from San Isidro del General, San Jose Province. All are dark lavender on their exposed surfaces, with minute light dots indicated. There is a line of pigment on the upperarm and on the dorsal surface of the thigh. Below, the venter and concealed parts of the Hmbs are clear flesh color. The digital tips are bright orange. The vomerine teeth are present in the adults and seemingly absent in the young. The two females measure 24 and 25 millimeters, snout to vent. The largest male is 23.5 millimeters. The relationship of the species is with Cochranclld alhoniacuhita. In the latter the heel extends three millimeters beyond the tip of the snout and the bones are not green in preserved specimens. The species is larger (29 mm.) with the webbing on digits distinctly greater. The toes are wider, with wider terminal pads. The lavender pigment remains practicalK unchanged, after pres- ervation, in both C. alhomaculata and C. pidverata, at least for some years. In life the color was "bright yellow green with maroon or rusty yellow dots. Below greenish white flesh, transparent; under- side of feet and hands light yellow." Costa Rican Centrolenidae 47 Fig. 3. Cochranclla pulverata (Peters). Upper figure KUMNH No. 32822; actual snout-vent length 22 mm. Lower figure, KUMNH No. 32821; actual snout-vent length 24 mm. Both specimens from Sarchi, Alajuela Province, C. R. 48 The University Science Bulletin Cochranella alhomaculato (Taylor) (FiK. 4) CentrolencUa ulhomacuhita Taylor. Univ. Kansas Sci. Bull., vol. 33, pt. 1, Apr. 20, 1949 ( tvpt' locality, Los Dianiantcs, one milf south Guapiles, Limon Province, C.R.). Four topotypic specimens, KUMNH Nos. 32824-27, were taken at Los Diamantes, near Guapiles, Limon Province, and a single female specimen, No. 32820, measuring 29.5 millimeters snout-vent Fig. 4. Codiranclhi olboinacuhita (Ta\'lor). KUMNH No. 32820. lAIA, Turrialba, Cartago Province, C. R. Actual snout-vent length, 29 mm. length, was taken on a banana plant near a tiny stream draining a swampy area at Turrialba. Here many other specimens were heard calling from tall trees. The latter Turrialba specimen is figured, showing the details of the dorsal markings. In life, the color was dark green above with discrete rounded greenish-yellow spots; venter transparent flesh; hands and feet Costa Rican Centbolemdae 49 yellowish to orange yellow. Chin, and concealed part of arms and legs, pale green. The green color has changed to lavender in preservative. This form is most closely related to Cochronella pulverata. Cochranella granulosa (Taylor) (Fig. 5) Centrohmella grauulosa Taylor, Univ. Kansas Sci. Bull. vol. 33, pt. 1, Apr. 20, 1949 (type locality, Los Diamantcs one mile south of Guapiles, Linion Province, C. R. ). A third specimen of this species, KUMNH No. 82808, was taken at Moravia de Chirripo, Limon Province, in 1952. It agrees with the type in all essential characters, save that in preserved specimen the dark lavender spots are larger. The specimen, a gravid female, measures 30 millimeters in snout-vent length. The ovarian eggs show some pigmentation. The specimen was taken on a Chinese ginger plant near a small stream. Its color in life was "dorsal surfaces bright yellowish-green Fig. 5. Cochranella granulosa (Taylor). KUMNH No. 32808, Mora\ia de Chirripo, Limon Province, C. R. Actual snout-vent length, 30 mm. 50 The University Science Bulletin on back; the legs and concealed parts canary yellow, with tips of toes bright orange-yellow; venter transparent flesh with a greenish or yellowish wash; blackish on upper eyelids." When preserved the green disappears and the lavender pigment produces rather large dark lavender spots on a cream background. CoclirancUa decorata sp. nov. (Fifi. 6) Type: KUMNH No. 36896 9, colloctcd at Hda. La Florencia, about 3 miles west of Turrialba, Cartago Province, Costa Rica, Nov., 1952, by Jack Reark. Paratypes: KUMNH Nos. 36883, 36884, (presumably) San Jose, San Jose Province, summer 1954; Prof. Marco Tulio Pacheco, donor. Diag,nosis: A medium-sized species (maximum size known 22.8 mm. snout to vent); head wider than body, the length of eye- opening greater than the length of the snout, nearly equal to the distance of the eye from tip of snout; nostrils not reaching as far forward as edge of mouth; first finger longer than second; outer fingers about half webbed; the web between the second and third fingers one fourth of the length of second finger to pad; that between two inner fingers one fifth length of inner to its terminal pad; tibiotarsal joint reaches at least three millimeters beyond tip of snout; a prominent fold runs from sides of vent back and down on thigh; tongue not or scarcely notched behind; no vomerine teeth; no tympanum visible. Description of type: Width of head (9 mm.) greater than body width and greater than length of head (8 mm.); width of an eyelid (1.3 mm.) twice in interorbital distance (2.6 mm.); nostrils not extending forward to level of mouth; distance between eye and nostril (1.8 mm.) nearly equal to distance between nostril and median point on upper lip (1.85 mm.); areas about nostrils not or scarcely swollen; no canthus rostralis; tympanum not visible; no supratympanic fold or only an indication of a slight thickening. Tongue a little longer than broad, with a faint suggestion of a posterior notch; diameter of a choana (.7 mm.) in distance between choanae (1.8 mm.) about two and one-half times; palatal glands opening in a concave transverse groove between anterior edges of choanae. (Male with a subgular vocal sac, the vocal slits short terminating near edge of tongue. ) Upper arm very slender, unpigmented; forearm widened, the bones of the hand more or less visible externally; digits with terminal pads wider than the fingers; web of two outer fingers equals length of fourth finger to its terminal pad; web between second and third Costa Rican Centrolenidae 51 Fig. 6. Cochranella dccoruta sp. nov. Upper figure, KUMNH No. 36884, paratype, San Jo.se (?), San Jose Pro\-ince, C. R. Actual snout-vent length, 22.5 mm. Lower figure, KUMNH No. 36896, type, Hda. La Florencia, Turrialba, Cartago Province, C. R. Actual snout-vent length, 22.8 mm. 52 The University Science Bulletin fingers about one fourth of the length of second, with a fringe extending along edge of digit to pad, but lacking fringe on inner side of the fourth; a web-remnant between inner fingers; a flattened fold on outer side of hand and forearm. Subarticular tulx^rcles small, distinct; a small inner metacarpal tubercle; a distinct oval or rounded palmar tubercle. Legs elongate the heel reaching at least three millimeters beyond tip of snout; heels touch or barely overlap when folded at right angles; toes about four-fifths webbed; a slender, somewhat elongate inner metatarsal tubercle; no outer tubercle evident; a faint tarsal fold. Skin on head smooth anteriorly, with some granules evident be- hind eye and at angles of the mouth; tympanum hidden imder skin and muscles; under a lens dorsum shows obsolete granulation, as does upper surface of femur, tibia, and tarsus; venter and side granular, the venter tending to show transverse wrinkling; about half of undersurface of thigh smooth; posterior half somewhat granular or areolate; a strong skinfold running back and down on thigh, the anal flap somewhat thickened more or less connecting the two folds. Color: In life, greenish above, with pale cream spots; transparent flesh below and on concealed surfaces of limbs. In preservation, at first, the green disappears leaving the ground-color yellowish white with small punctations of dark lavender. These too have now dis- appeared, leaving the dorsum unicolor. Measurements in )nm. (of type, No. 36384, and 36383 respec- tively): Snout to vent, 22.8, 22.5, 22.5; width of head, 9, 8.8, 8.9; length of head, 8, 7.5, 7.6; arm length, 14.8, 14, 14.2; leg, 39, 34, 35; tibia, 12.8, 11.4, 13; foot and tarsus 17.5, 16, 17.2. Variation: The paratypes (not seen in life) in preservation have a scattering of lavender dots over body and the exposed parts of Hmbs. The skin on the dorsum of the paratypes is smoother than in the type and much of the femur is glassy smooth while the granu- lation is almost completely obsolete. The webbing on the hand is a trifle less than in the type. The donor of these specimens obtained them from his students. He believed them to have been taken at San Jose, San Jose Province. Remarks: The character of the folds in the region of the vent separates this form from other known centrolenids of Costa Rica. A similar decoration appears on Cochranella lutzorum Taylor and Cochran, from Petropolis, Rio de Janeiro. In life there were orange Costa Rican Centrolenidae 53 spots visible within the mouth on the palate, about the internal nares. The type was taken from a leaf of Chinese ginger growing near a small fast-flowing stream. The specific name is derived from the Latin decoratus = orna- mented or decorated, referring to the folds about vent. Cochranella millepunctata sp. nov. (Fig. 7) Tqpe: KUMNH No. 36887: collected at La Palma, San Jose Province, Costa ■ Rica, June 20, 1954 by Edward H. Taylor. I'uTuUjpes: KUMNH Nos. 36883-86; 36888-94, topotypes, same data; No. 36897 Sarchi, Alajuela Pro\ince; No. 23943, (?) Moravia de Chirripo, Cartago Province. Diagnosis: A small species, maximum size ( $ ) 23.8 mm.; nos- trils not reaching forward to level of mouth, the areas surrounding them not elevated or swollen; front profile of snout nearly vertical, tympanum hidden; eye with a white tunic; no vomerine teeth; vocal sac present in males; terminal pads widened, distinctly wider than digits and rounded anteriorly rather than truncate; wide anal flap, the granules below vent "frosted" white; interorbital distance a fifth wider than width of an eyelid; legs folded the heels fail to meet; the tibiotarsal articulation brought forward reaches to tip of snout or slightly beyond; dorsum and head closely granular, the granules, visible to eye; sides finely granular; lateral parts of venter granular, the median area vaguely so; toes about three-fourths webbed; outer fingers two-thirds webbed; less than one-fourth webbed between second and third fingers. Description of type: Head a little broader than body; width of head (8.6 mm.) greater than length (7.3 mm.); eyes relatively small, moderately elevated, directed forward, the length of eye- opening less than its distance to tip of snout, longer than its distance from nostril; front of snout a rounded oval; nostril a little nearer mid-point on lip than to eye; no canthus rostralis; length of snout from front level of eye, two millimeters on mid-line; only a vague trace of a supratympanic fold. Tongue (distorted in type) normally about as long as broad, free behind and on sides for about one seventh of its length; ( male with vocal slits); choanae small, circular; diameter of choanae con- tained in distance between choanae about four times; openings of Eustachian tubes equal to or a little smaller than choanae. Arm well developed, the fingers relatively short, the terminal pads widened, that of third finger largest (1.4 mm.) distinctly 54 The University Science Bulletin wider than smallest width of digit (.9), the pads transversely oval; web between two outer fingers half length of fourth finger; between second and third, web about one fifth length of second finger; no web between the first and second fingers; first finger a little longer than second; fingers with slight lateral fringes except inner side of the third; a slight fold from base of first finger to elbow; no pigment on upper arm; forearm much widened at base; a small outer meta- carpal tubercle; an oval palmar tubercle. Leg long, the tibiotarsal articulation reaching two millimeters beyond snout; when legs are folded, heels touch and overlap (.5 mm.); toes two-thirds to three-fourths webbed, the terminal pads somewhat widened, the largest about as wide as that on first finger, smaller than pads on others; a small, slightly projecting metatarsal tubercle; subarticular tubercles of hand small, none double or bifid; those on toes a little smaller. Skin above nearly smooth (under a lens very dimly granular on head, body and forearm); chin and breast smooth; venter with flat somewhat indistinct granules, more distinct laterally, obsolete on most of underside of thighs; anal flap wide transversely; vent fol- lowed by a quadrangular area of tubercles, frosted white; no anal folds. Color in life: Above generally bluish green, with extremely numerous flecks of silvery white; eyeball with a white tunic; iris of eye whitish to golden, with numerous lavender flecks about rim; pupil longitudinal, with considerable purple flecking around it. Venter transparent flesh, the tips of the toes orange. Intestines, visible through flesh, cream-white; a whitish covering also below the pericardial cavity. In preservation, the green disappears leaving a fine reticulation of lavender or magenta enclosing small white dots, much larger than the minute "frosted" white flecks seen in life. Measurements in mm.: (type, No. 36895 largest male, and No. 36897 largest female, respectively): Snout to vent, 21.4, 21.5, 2-3.8; head width, 8.5, 8.8, 9.2; head length, 7.3, 7.2, 8; arm, 16, 15.5, 14.6; leg, 38.2, 37.5, 37; tibia, 12.3, 12, 12; foot and tarsus, 16.2, 16.3, 16.1. Variation: There is but little variation in the specimens from La Palma. The specimen from Sarchi has a slightly wider head, slightly shorter limbs, and the webs between the toes are less excised. This specimen agrees with the type in coloration. The type locality of Coehranella valerioi (Dunn) is likewise La Palma, San Jose Province. The two forms may be readily distin- guished by color markings, valerioi having a green median stripe; Costa Rican Centrolenidae 55 Fk;. 7. Cochranclla niiUipumtata sp. nov. Upper figiuc, KUMN'II No. 36886, actual snout-vent length. 20 mm. Lower figure, KUMNH No. 36887, type; actual snout-vent length, 21.4 mm. Both specimens from La Palma, San Jose Province, C. R. 56 The University Science Bulletin in the latter the area about the nostrils much swollen and protrud- ing, the snout and nostrils actually projecting in front of mouth; a small tympanum is visible, and there is more webbing on hand. CocJiranella fleischmanni is also known from La Palma. This spe- cies is larger with a white spot on the upper eyelid. The transforming tadpoles of this species were taken with the types. One specimen, with tail completely absorbed, measures 13 mm. This was taken together with other tadpoles, some having tails three or four millimeters in length. The specific name is derived from the Latin, millc = thousand, and piinctiim = small dot or spot; meaning, "thousand spotted," CochraneUa fieischmanni (Boettger) (Fig. 8) The type locality is San Jose, San Jose. Specimens are in the collection from the following localities: KUMNH Nos. 23807-08, 30399 Cartago, Cartago Province ( Caribbean drainage ) ; No. 36895, La Palma. San Jose Province, 4500 ft. elev.; Nos. 30400-02, Rio Segundo, Alajuela Province (Pacific drainage). CochraneUa chrijsops (Cope) (Fig. 9) Hylclla chrysops Cope, Proc. Acad. Nat. Sci. Philadelpliia, 1894, p. 196 (type locality [restricted], San Jose, San Jose Pro\ince, Costa Rica.) Specimens of C. chrysops were found to be plentiful at Sarchi, Alajuela. Although formerly placed in the synonymy of C. fleisch- manni, when directly compared with that species the two are found to be distinctly different; their ranges overlap a considerable extent on the Meseta Central of Costa Rica. This form differs from fleischmanni in lacking a white eye tunic (and the "white spot" on upper eyelid); in having distinctly larger hands and wider digits; the first toe is shorter, the first and second being of equal length; the head is more distinctly set off from the body. The largest male taken measures 24 millimeters in snout-vent length. The smallest fully transformed specimen was 9.2 milli- meters in length. The following localities are represented: KUMNH Nos. 32943-55, 32962-63, 32965-66, 32968-69, near Sarchi, Alajuela Province: No. 32956 ( locality uncertain ) ; Nos. 32957-59. Cartago, Cartago Prov- ince. Costa Rican Centrolenidae 57 Fig. 8. CochrancUa fleischmanni (Boettgcr). Upper figure, KUMNH No. 30401. actual snout-vent length, 23 mm. Lower figure, KUMNH No. 30400, actual snout-vent lengtli, 23.8 mm. Both, Rio Scgundo, Alajuela Province, C R. 58 The University Science Bulletin Fig. 9. CochrarwUa chrysops (Cope). Upper figure, KUMNH No. 34944, Pacayas, San Jose Provinee, C. R. Aetual snout-vent length, 23 mm. Lower figure, KUMNH No. 32958, Cartago, Cartago Province, C. R. Aetual snout-vent length, 24 mm. Costa Rican Centrolenidae 59 Cochranclla chirripoi sp. nov. (Fig. 10) Type: KUMNH No. 36865 9, collected on Cocales Creek, Siiretka, Limon Province C. R., June 29-30, 1954, by Edward H. Taylor. Puratypes: KUMNH Nos. 36862-64, 36866-70, topotypes, same data as type. Diagnosis: A medium-sized species, the largest known, 26 milli- meters in snout-vent length; no vomerine teeth; male with vocal sac; nostrils not extending forward beyond month; no canthus rostralis; tibiotarsal joint reaching two to three millimeters beyond tip ot snout; heels overlap two millimeters when legs are folded at right angles to body; toes nearly fully webbed; web on hand as extensive between second and third fingers as between the third and fourth; tympanum somewhat indistinct, covered with finely granular skin; terminal digital pads on hand rather truncate, wider than narrowest part of finger. A white tunic about eyeball; skin graiuiles distinctly visible to unaided eye. Description of type: Head wider than body, its width (9.5 mm.) greater than its length (7.2 mm.); eye comparatively small, pro- truding but slightly beyond rim of mouth; no canthus rostralis; area about nostrils not or but slightly swollen or elevated (some- what evident in certain paratypes ) ; length of snout from front level of eyes, three millimeters; tympanum partly evident, its size uncer- tain, its surface covered with finely granular skin; width of eyelid (1.8 mm.) 1.55 times in interorbital distance (2.8 mm.); choanae relatively small, the diameter of one in distance between choanae, at least three times; openings of palatal glands in an undulating line, crossing palate closer to choanae than to front of palate; tongue a little wider anteriorly than posteriorly, unnotched on posterior edge, free for less than one sixth of its length; ( a vocal sac present in male, the vocal slits rather far back, reaching to near the angle of the jaw). Upper arm very slender, unpigmeivted, the forearm double width of upper arm; web between second and third fingers almost co- extensive with that between third and fourth fingers; web between first and second finger one third length of second digit to terminal pad. The terminal digital pads one-fourth wider than narrow part of digits; outer metacarpal tubercle small; a somewhat rounded palmar tubercle; subarticular tubercles low, flattened, single; a strong fold on outer edge of hand extending to elbow. Leg brought forward the tibiotarsal articulation reaching two millimeters beyond tip of snout; toes nearly fully webbed, the webs 60 The University Science Bulletin Fig. 10. Cochranella chirripoi sp. no\ . Upper figure, KUMNH No. 36865, typt-; actual snout-vent length, 26 mm. Lower figure, KUMNH No. 36867; actual snout-vent length, 25.7 mm. Both from Suretka, Limon Province, C. R. Costa Rican Centrolenidae 61 reaching the terminal pads (except fourth), the webs slightly excised between toes; digital pads smaller than those on fitigers; inner metatarsal tubercle distinct, somewhat elongate; outer indis- tinct or absent; a distinct tarsal fold. Skin above on exposed surfaces covered with distinct granules visible to unaided eye and present also to some degree on arm; chin and breast smooth; venter strongly granular, the skin on the sides with sliort folds rather than granules; most of the undersurface of thigh granular; below vent an area in which the skin, covered with large granules, appears to be folded vertically; undersurface of tibia and tarsus smooth. Color: Exposed surfaces greenish in life the color extending onto upper arm, and distinctly along upper surface of femur; distally on limbs the color reaches onto outer finger and on the two outer toes for some distance. When preserved for some time the green is replaced, and a sparse peppering of lavender or magenta dots is evident (at least so under a lens). Venter slightly yellowish flesh, transparent; digital tip orange; some pigment below vent. Measurements in mm. (Nos. 36865, 36866, 36869 respectively): Snout-to-vent length, 26, 24, 25; width of head, 9.5, 8.5, 8.9; length of head, 7.2, 8, 8; arm length, 17, 16, 16.3; leg, 44.5, 41, 46; tibia, 14, 13, 14.5; f(K)t and tarsus, 18.3, 17.5, 18.9. Variation: The series of paratypes is fairly uniform save that in certain ones the white eye tunic seems to be retracted but can be seen near the base of upper eyelid; the eyeball seen from the mouth also appears dark over most of its surface. The leg of No. 36869 is proportionally longer, the heel reaching more than three millimeters beyond the tip of the snout. Remarks: The hands of this species have a greater amount of webbing than other species of C'osta Rican centrolenids, and by this character it may be separated from the other known forms. Specimens were discovered by following up their calls. They were found in shrubs and low branches of trees along Cocales Creek, a fast moving, shallow stream ten to fifteen feet wide. The species is named for the local inhabitants of the area, the Chirripo Indians. Cochranella reticulata sp. nov. (Fig. 11) Tifpe: KUMNH No. 32922, collectfcl near hritlgc across Rio Rt-vcntazon at the Inter-Ainerican Institute of Agriculture, Turrialba, Cartago Province, Costa Rica, by Edward H. Taylor. Paratypes: KUMNH Nos. 32921, topotype, same data; No. 32916 Moravia de Chirripo, Limon Province, June 26, 1952; Nos. 32917-20 Cariblanco, 62 The University Science Bulletin Hcredia Pro\-inco, July 20, 1952; Nos. 32923-24 Palmar, Puntarenas Prov- ince, Sept. 3, 1952; No. 32925 Golfito, Puntarenas Province, September 7, 1952 (breeding); No. 36878-82 Suretka, Limon Province, June 30. 1954 ( breeding ) . Diag,nosis: Small, maximum size $ 23 mm., male 22.5 mm.; great- est diameter of tympanum, about two and one-half times in length of eye; interorbital width greater than width of eyelid; no canthus rostralis; nostrils moderately prominent, nearly on a level with the anterior edge of mouth; a slight supratympanic fold not reaching eye; venter and underside of thighs granular; one-half to three-fifths webbed between two outer fingers; one-fourth, or slightly less than one-fourth webbed between second and third; only a web-remnant between inner fingers; toes about four-fifths webbed; a tarsal fold; head a little wider than body; body flesh-green above with a retic- ulum of deep green; eyelids deep green. The tunic of eye dark; venter and underside of limbs transparent flesh; in preservative, a reticulum formed of small groups of lavender dots. Description of the type: Width of head, at tympanum (8 mm.) greater than length (6 mm.); eyes large, strongly elevated, pro- truding beyond edge of mouth; length of eye ( 3 mm. ) greater than length of snout (2 mm.); no canthus rostralis, the loreal region not concave; area about nostrils somewhat swollen, with a slight de- pression anteriorly; tip of snout extending about 0.8 millimeters beyond mouth; tympanum (1.3 mm.) relatively large, contained in length of eye-opening (3.0 mm.) about two and one-half times, directed outward and upward, completely visible from above; a tiny supratympanic fold begins above tympanum and continues to above arm; interorbital width (2.2 mm.) greater than width of eyelid (1.8 mm.). Tongue slightly free behind and slightly notched; no vomerine teeth; choanae subcircular, the diameter of one contained in inter- choanal distance nearly three times; palatal glands open into a transverse groove slightly in advance of level of choanae; vocal slits opening into a large subgular vocal sac. Upper arm very slender, unpigmented; forearm much thickened with a prominent fold from wrist to elbow; inner metacarpal tu- bercle elongate, scarcely elevated; median palmar tubercle rounded, rather prominent; outer subarticular tubercles on first, third, and fourth fingers double or bifid; other small tubercles on palm; fingers broad, not or but slightly widened at tips, terminally subtrimcate. Web between two outer fingers slightly more than half of the length Costa Rican Centrolenidae 63 Fic. 11. CochnmcUa retictihita sp. nov. Upper figure, KUMNIi No. 32921, paratype, actual snout-vent length, 23 mm. Lower figure, KUMNH No. 32922, type, actual .snout-vent lengtli, 22 mm. Both specimens from Inter-American Institute of Agriculture. Tiir- rialba, Cartago Province, C. R. 64 The University Science Bulletin of fourth to terinituil disc; about one-fourth webbed between second and third fingers; only a remnant of web between two inner fingers; a strong fringe to disc on outer side of second and third fingers and on inner side of fourth; a sfight fold on outer side of palm; first finger longer than second; subterminal pads subtriangular. Leg slender, the heel reaching the tip of snout (probably slightly be- yond, normally); when legs are folded, the heels touch; toes four- fifths webbed, the web reaching to near level of the subterminal pads, except on fourth toe. Skin above (under a lens) somewhat granular; a very shallow depression between eyes; venter and underside of thighs granular; a pair of larger rounded areolae or granules near median point under thighs. Color: Above generally greenish flesh with a reticulation of darker green surrounding large circular cream areas; upper surface of hands and feet cream to yellowish; ventral and concealed surfaces transparent flesh. In preservative, cream-white with a reticulum composed of lavender dots surrounding rounded areas. Measurements in mm.: Snout-to-vent length, 22; width of head. 8; length of head, 6; length of arm, 14; leg, 40; tibia, 13; foot and tarsus, 17. Variation: The specimens of this species available show a close similarity in size between the sexes. There is practically no differ- ence in coloration. Specimens in preservative (formalin, transferred to alcohol) tend to have the green color turn to lavender or purple which is slowly dissolved in alcohol. When this happens the color of the muscles may give the appearance of two lighter lateral stripes extending to pelvis with a very slightly darker and wider median strijx". The eyes appear black. After nearly five years in preserva- tive the dorsal pigment marks can still be seen under a lens. The dorsal surface of thigh has scattered pigment but the hands and feet ( except outer toe ) are without pigment. The tongue varies somewhat in shape in preservation, but usually it is a little wider posteriorly than anteriorly and a slight notch is indicated. Remarks: All of the series listed are adult. Most of them were discovered by following calls, the specimens being found on the underside of leaves. Two pairs were found breeding, and in two cases single males (one the type) were observed brooding sets of Costa Rican Centrolenidae 65 eggs under a leaf. The type was taken from a palm-tree leaf at least 25 feet above the ground. It was dislodged by a long bamboo pole. Almost directly under the eggs was a tiny rivulet flowing over rocks, entering the Rio Reventazon at a point some fifty feet away. The call is reminiscent of the call of certain local orthroptera ( katydids ) . Specimens taken at Cariblanco were found at night perched on smooth leaves overhanging a swift-moving irrigation canal. They were not calling. Two females, Nos. 32916 and .32923, are gravid. Cochranella valerioi (Dunn) Centrolenc valerioi Dunn, Occ. Papers Boston Soc. Nat. Hist., vol. 5, .A.iij;. 18, 1931, pp. 397-398. (type locality La Palnia, San Jose Province, C. R. [4500 it. elev.].) No specimens of this species have been taken although search has been made in the type locality. Dunn described the species as follows : Diagnosis: "Centrolene without humeral hooks, no vomerine teeth; tympanum almost aborted; nostrils raised prominently; bones white; color in life white with a dorsal green network. Description: "Tongue circular; head broader than long, semi- circular as seen from above save for projection made by nostrils; eyes directed forward, their diameter greater than their distance from tip of snout; canthus rostralis rounded but distinct; lores con- cave; nostrils protuberant, causing snout to overhang; interorbital space twice as wide as upper eyelid; tympanum barely visible, di- rected upward, 1/6 the diameter of eye; fingers with disks wider than tympanum, truncate; first finger longer than second; web on about /2 of outer fingers, to penultimate phalanx of 3 and 4; inner fingers webbed at base; toes webbed to just short of disks of 3 and 5, not quite to penultimate phalanx of 4; a single weak inner meta- tarsal tubercle; heel reaches to beyond snout; smooth above, belly and thighs rugose; in life white, a narrow green dorsal stripe, green vermiculations on dorsal surface and on shin, thigh unpigmented, white beneath, iris golden, in preservative white, a few faint dark chromatophores where green was in life; length 21 mm., width of head 8; arm 13, leg 39 mm." I examined a specimen presumed to be a paratype, a short de- scription of which was published by me (Univ. Kansas Sci. Hull., 3—2656 66 The University Science Bulletin vol. 35, pt. 1, no. 5, July 1, 1952). This specimen may or may not be of the same species. I have not compared it with the type. The species appears to be related to CochraneUa taluviancae despite obvious differences. Further study is essential before the status of these two forms is fixed. CochraneUa talaniancac Taylor (Fi«. 12) Cuchranella talamancae Taylor, Univ. Kansas Sci. Bull., vol. 35, pt. 1, July 1, 1952, pp. 781-783 (type locality, Moravia de Chirripo, Linion Province, C. R.). This diminutive species ( appro.ximately 22 mm. snout-vent length ) was rediscovered at the type locality, Moravia de Chirripo, Limon Province. The following were taken: KUMNH, Nos. 30887, 32936-38 topo- types. In life the ground color of the specimens was delicate canary yellow with a diffuse green median line joined to a fine greeni.sh reticulation, enclosing lighter areas on dorsum. Chin bluish white, the venter transparent cream-flesh; concealed parts of limbs flesh; upper eyelid dark green over black; iris golden, powdered with deep purple; webs of hands and feet less intense yellow than on the tips of fingers and toes; on dorsal surface of legs there are slightly rounded yellow spots. J This may be the southern representative of CochraneUa valcrioi m (Dunn), which appears to be its nearest relative. I have not found ■ C. valerioi although I searched for it at La Palma, the type locality. CochraneUa cohjnihiphyUiini (Taylor) (Fig. 13) CentroleneUo cohjnihiphtjllum Taylor, Univ. Kansas Sci. Bull., vol. 33, pt. 1, Apr. 20, 1949, pp. 262-263 ( typo locality, American Cinchona plantation [Isla Bonita] Volcan Poas, Heredia Province, Costa Rica). This species is presumably confined to the northern and eastern drainage of the Cordillera Central. The following specimens, other than the types, are in the collection: KUMNH, Nos. 32939-32941 Cariblanco; No. 32942, topotype, Cinchona (Isla Bonita), both from Heredia Province. Costa Ricax (^entrolenidae 6' Fig. 12. Cocliranella tdlmnancae Taylor. I'pper fijiurr, KUMNH No. 32937 topotypc, actual siioiil-vent lengtli. 21 mm. Lower figure, KUMNH No. 32938, topotypc, actual snout-veut length, 21. .5. Roth from Moravia de Chirripo. Ij'moii ProMTice, C. R. 68 The University Science Bulletin Fig. 13. Cochranella colymbiphyUum (Taylor). Upper figure, KUMNH No. 23812, type, actual snout-vent length, 26 mm. Lower figure, KUMNH No. 32942, topotype, actual snout-vent length, 26.2 mm. Both from Isla Bonita [=: Cinchona], Heredia Province, C. R. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 3 Observations on the Ethology of Neotropical Anthophorine Bees (Hymenoptera: Apoidea)' BY Charles D. Michener - and Rudolf B. Lange '' Abstract. Thi.s paprr snininarizes available data on the biology of neo- tropical anthophorine bees. It also presents new information, primarily on nesting behavior, for members of the following genera: Paratetrapedia, Mono- i'ca, Tetrapedia, Melitoma, Ptilnthrix, Peponapis, Melissodes, Melissoptila, Thy^ater, Anthophora. and Ccntris. With few exceptions the biological characteristics of groups set forth by Linsley, MacSwain, and Smith (1955, 1956) are supported. INTRODUCTION Thi.s paper give.s the results of a series of studies of hees of the subfamily Anthophorinae (sense of Michener, 1944) and their nests, carried out principally in the State of Parana, in southern Brasil. One purpose of this paper is to make available for comparative study the rather fragmentary data which we have gathered on the species mentioned below. Another purpose is to summarize the data now a\ailable on the biology of the nonparasitic Neotropical members of this large group of bees, with supplementary or com- parative information on representatives from other regions, to pro- vide a basis for constructive comparative work on the biology of Anthophorinae in the American tropics. \. The preparation of this paper and the fiunres acconipanyinu it was faeilifated by a Krant from the National Seienee Foundation. Thanks are due to Father J. S. Moure for identitieation of the South American liees eoneerned. We are also indebted to Prof. Doniieiano IJias, Escola Superior de AKrieultura "Lniz de Queiroz," Unixcrsidade de Sao Paulo, Piraeieaba, Sao Paulo, Brasil, and to Mr. .Mvaro Wille, Mr. Carl W. Rettenmeyer and Mr. IIowcll \'. Daly of the University of Kansas for data which fhe\' Kalhered and which We have used in this study. 2. Department of Entomology, University of Kansas, Lawrence, Kansas, U. S. A. Field work for this author's part in tfie study was possible thanks to a Jolin Simon Gucilenheim Memorial I'ellow ship and aid kindh made .i\ ,iilal>le b\ the C-'ampanha Naeional de Aper- leiyoameiito de Pessoal de Ni\el Superior, Hio de Janeiro; the Conselho .Naeional d<' Pes- (|uisas, Rio de Janeiro; and th" Rockefeller Fomidation, New York. In particular, thanks are due to Father J. S. Movire for the use of facilities of the Sec(,ao de Zooloilia, Faeiddade de Fiiosofia, Universidade de Parana, Curitiba, Parana, Brasil, durinj: a year of residence in Curitiba. 3. Secgao de Zoologia. M\iseo Paranaens<' and Faculdad<' (^atolica de Fiiosofia, Curitiba, Parana, Brasil. (69) 70 The Umvehshy Sc:ien(:e Bllletix Tribe Exomalopsini Little has been published about the nesting habits of members of this tribe. So far as known, all nest in the ground. A tendency toward communal life exists in this group. In the species discussed below as well as in Ancyloscelis (= Lepter^atis) ( Brethes, 1909; Jor- gensen, 1912a; Michener, 1954) each female presumably makes a separate nest and indeed in Ancyloscelis there seems to be but one cell per burrow, but these may occur in great aggregations. Hicks (1936), however, recorded four pollen-laden females of Exomalop- sis entering a single hole, and Dr. Paulo Nogueira-Neto of Sao Paulo tells us that he observed nmnerous females of another species of Exomalopsis entering a single nest. Moreover, Rozen and Mac- Neill (1958) surmized that various individuals of still another spe- cies were utilizing a single burrow, and they found such burrows in a large aggregation. Claude-Joseph (1926) found various indi- viduals of Tapinotaspis caeridea (Friese)' using a single entrance hole but making separate lateral branches for cells. Linsley, Mac- Swain, and Smith (1954) describe how individuals of Exomalopsis were observed entering probably abandoned nests of other bees. Perhaps this explains how they find nests of their own species to form these common nests. A few tribal characters of the cells, based only on Tapinotaspis, Paratetrapedia, Ancyloscelis, and Exomalopsis are ts follows: Cells in series (Tapinotaspsis and often in Exotnaloi)sis) or not (Paratetrapedia, Ancylosceli ;), not separable from surrounding matrix, lined with waxlike material. Provisions in the form of a firm ball or mass of various shapes [about spherical in Paratetra- pedia and probably in Tapinotaspis, elongate with a depression on one side in Ancyloscelis (Michener, 1954) and elongate and angu- late in Exomalopsis (Rozen and MacNeill, 1958)], not merely packed into the bottom of the cell. Egg laid on top of provisions or, in Ancyloscelis, in a broad concavity along one side of an elongated mass of provisions (Michener, 1954). Paratetrapedia oligotricha (Moure) This is a species of the subgenus Tri^onopedia. An enormous nest aggregation of this species was found in a vertical bank of sandy soil in Floresta de Tijuca, Rio de Janeiro, D. F., Brasil, at an altitude of 500 meters above sea level, on November 28, 1955. 1. This species is placed in a new subgenus of Tapinotaspis, rather than in Exomalopsis, by Michener and Moure (1957). Ethology of Neotropical Anthophorine Bees Fig. 1. Portion of a vertical bank in the Floresta de Tijuca, Rio tie Janeiro, containing an aggregation of nests of Pomtetrapedia oligotricha (Moure). The white mark at the lower right is a ruler about 11 cm. long. There were about ten square meters densely occupied by nests (fig. 1); in one square, 10 cm. on each side, there were 41 nests. There was no evidence of old cells or nests of previous generations in this bank and Dr. Carlos Alberto Campos Seabra who took one of us to this locality had not previously seen this aggregation al- though he is intimately acquainted with the Floresta de Tijuca. Another bank a quarter of a mile away had been similarly densely occupied, almost certainly by this same species, three years pre- viously according to Dr. Seabra's observations but in July, 1955, 72 The University Science Bulletin and again in November, not a bee could be found there, only old nests with cells which contained feces and had apparently produced a crop of bees. It is clear that new aggregations can form in a brief period. Possibly additional individuals are attracted by the odor of those that have already started nesting. On November 28, although the bank was thick with bees in flight, some of which were starting their burrows, and the oldest larvae had not finished the last of their supplies, no males were seen. By January 9, 1956, a new generation had probably been produced, for Dr. Seabra, who visited the bank on that day, observed both sexes in flight and many pairs mating. Often pairs would roll down the bank and complete copulation at the foot of the bank. Dr. Seabra killed four pairs, taken mating, in Kahle's fixative and sent them to us. Three of the females, on later dissection showed no sperm cells in the spermatheca and therefore must have been mating for the first time when captured but the fourth already had a ball of sperms in the spermatheca. There were also sperm cells in the common oviduct. Presumably this individual must have mated pre- viously, as the minute spermathecal entrance as well as our knowl- edge of the situation in Apis indicate that hours must be required for the entrance of a large number of sperm cells into the sperma- theca. That multiple matings are not only common but the rule in Apis is well known (for references to pertinent literature, see Peer, 1956); it is interesting to find evidence, however fragmentary, of such behavior in other bees. On February 22 activity of Faratetrapedia was continuing in the bank according to Dr. Seabra. Although there was no certain evi- dence that a second generation was being reared there, it is very probable. Unfortunately road construction resulted in destruction of the bank so that subsequent observations were impossible. The same species was also observed nesting in the Barigui road- side banks near Curitiba, Parana, Brasil. The number, distribution and orientation of the nests, as well as climatic and soil data, are given by Michener, Lange, Bigarella, and Salamuni (1958). The nests were very few and scattered. Doubtless because of the cool climate of the south Brazilian plateau, the season of activity is some- what different from that of the population at Rio de Janeiro. On October 17, 1955, bees were still prepupae in their cells and on March 11 and 18 nests were being constructed. So far as known each nest is constructed by a single female bee. The nests are more or less winding, unlined, roughly horizontal Ethology of Neotbopical Anthophorine Bees ^•3 burrows 12 to 30 cm. deep, 5-6 mm. in diameter. Descending at intervals from this main burrow are vertical branches 3 to 20 mm. long, each ending in a single cell. The nest is constructed as a short horizontal burrow which turns down near the inner end and Figs. 2 to 4. Nests of Trigonopedia oligotricha (Moure) in various stages of construction. Dotted areas represent pollen halls; lined burrows are filled with earth. In left cell of fig. 3 and right of fig. 4, egg is shown on top of pollen hall. Fig. 5. Entrance of burrow of Tetr(i))cdia inauru Cresson. Figs. 6 and 7. Nests of Tetrapedia mmira Cresson. Fig. 6 shows bur- row without cells but with gray material (lined) narrowing entrance and starting to narrow the burrow in formation of cells. Fig. 7 shows a com- pleted nest. Short lines in various directions represent wood particles; lined area represents hard gray material brought to nests by bees. Pollen (omitted in lower cell ) has dotted lines to show margins of cavities. Eggs are shown in two upper cells. ( Figures 5 to 7 modified from sketches made by Alvaro VVille and Howell \. Daly. ) 74 The University Sc:ience Bulletin ends at the first cell. When this cell is complete and provisioned, the bee continues the horizontal burrow from the point where it turned downward, at the same time filling the vertical burrow, probably with newly excavated material, and soon turns down to make a second vertical burrow and cell. The process continues so that the oldest cell is consistently nearest the entrance, the youngest deepest in the bank ( figs. 2 to 4 ) . The number of cells per nest is unknown but reaches at least four. The main burrow is apparently never closed. The vertical cells are 6 to 6.5 mm. in diameter, 11 to 12 mm. long, lined with shining waxlike material. The provisions are in the form of a rather firm ball occupying the bottom of the cell and with the gently curved egg lying on top. Some cells were found with cocoons but these were probably made by a parasite; we believe that the Paratctropedia makes no cocoon. Monoeca sp. A single female of an unidentified and apparently new species of Monoeca (=: Fiorentinia) was found in a nest in a tiny bare spot among grasses near Sao Jose dos Pinhais, Parana, Brasil, February 21, 1956. The somewhat irregular burrow, with a tumulus of loose soil at the entrance, was 20 cm. deep, 6 mm. in diameter, and ended in a single vertical cell lined with waxlike material and lacking provisions. As the wing margins of the bee were entirely worn off, and her mandibles woni until they did not meet, it was obvious that this was not the first nest of this individual. Apparently Monoeca makes more than one nest. Schrottky (1901) recorded that M. schrottkyi (Friese) made burrows over a meter deep in a bank. Tribe Tetrapediini This tribe was recently separated from the Exomalopsini by Michener and Moure (1957). The biological characters of the tribe, here described, strongly support this separation. The nests are excavated in wood, the vertical cells placed in series in the burrow in the wood. The bees apparently bring in some dark substance to form the cells and to narrow the entrance to the nest. The pollen mass is firm and rather dry, vertically elongate, against one side of the cell, with a concavity which is sometimes divided by a transverse ridge. The egg is placed in a vertical position in the lower part of the concavity. This is suggestive of Ancyloscelis (Michener, 1954). Ethology of Neotropical Anthophorine Bees 75 Tetrapedia maura Cresson The observations recorded below were made five miles north of Guacimo, near Limon, Costa Rica, by Alvaro Wille and Howell V. Daly, both of the University of Kansas, during the period February 17 to 20, 1954. About eleven nests were found in two of the vertical supports (old, long dead trunks of small trees) of an old jungle hut. They were protected from rain by the thatch of the hut. In dark rainy periods there was no activity of the bees. In the morning, sun struck the poles of the hut at 7:27 a.m.; the first Tetra- pedia left their nests at 7:30. Before they left, they stood at the entrances of their nests for a few minutes, then flew off with little or no orientation flight. These bees returned in 10 to 25 minutes without pollen, but by 8:30 bees were returning with pollen. Many of the pollen-collecting trips were very short, only a few minutes long; the bees returned carrying large yellow balls of Ciuurhita pollen on the outside of each hind tibia. In rapid flight the long legs with the pollen are folded near the under side of the body but as the bee approaches a landing they are extended conspicuously downward. Sometimes a dark material, apparently used in nest construction ( see below ) , was carried in the scopa instead of pollen. One bee which must have been involved in excavation was seen several times backing out of its hole with large loads of wood powder. On leaving the nest it hovered and dropped dust by rubbing the legs together in flight. Sometimes a bee was seen to enter a nest backwards. Considerable confusion was noted about recognition of holes; sometimes bees attempted to enter holes much too small; one bee was seen to attempt to enter its hole, then it dragged another bee out by the leg before entering. There was no evidence, however, that more than one bee worked jointly in a single nest, although evidence obtained on opening nests seemed to in- dicate that a hole may be reused. After about 10:30 a. m. activity became less intense, and in the afteYnoon was very slight; the last bee was seen entering its nest at 5:30 p. m. Throughout the day Coelioxys were flying about the poles, sometimes ejitering holes; they may be parasites of Tetrapedia. The nests consist of simple burrows which enter the wood ob- liquely for about a centimeter and then turn down, vertically, paral- lel to the surface of the pole, and extend for 3.5 to about 5 cm. ( figs. 6 to 7). The burrows are oval in cross-section, 3 mm. in short di- ameter, 5 to 6 mm. in the long diameter. The entrance is funnel- shaped due to the placement there of a hard gray material (figs. 76 The University Scienck Hli.letin 5 to 7 ) which narrows the opening. One nest opened had this material at the entrance and only a little along the walls of the burrow (fig. 6). In others the entire burrow or much of it was narrowed by similar material from which the cells were also formed (fig. 7). The cells are about 10 mm. long, 8 mm. wide, end to end in a vertical series. Their inner walls are not smooth and shining, but are made of the gray material mentioned above. All the nests that contained cells had some wood particles occupying a small space at the bottom of the nest and some had thin transverse layers of wood particles between cells. The number of cells per nest, in those that seemed to be completed, was 3 or 4. One nest already abandoned by the mother had the oblique upper part largely filled with gray solid material. The pollen mass is imique among known bees. It is vertically elongate, about 9 mm. long and 5 mm. wide, against one wall of the cell. There is a hollow, sometimes divided by a transverse ridge, in the exposed face of the pollen mass. The egg, which is about 3.5 mm. long, is in a vertical position in the lower part of the hollow (fig. 7). The larvae apparently feed from the surface of the pollen mass, for cells with rather large larvae still have a layer of pollen adhering to the wall of the cell. Tribe Emphorini Some of the biological characteristics of this tribe and of the genera contained in it were recently summarized by Linsley, Mac- Swain and Smith (1956). As always when generalizations are based on few data, as is necessary in such biological studies at the present stage of our knowledge, certain conclusions, useful at the time, prove to be premature. The following comments, then, con- cern additions to the valuable summaries of biological characters listed in tlie above mentioned work. At least some of the South American Ptilothrix nest in vertical banks or in adobe walls, in contrast to North American species (see Strand, 1909; von Ihering, 1904; Ducke, 1901; and species discussed below). P. fructifera places its cells in series, unlike other species of the genus whose biology is known. Diadasia cliilctms (Spinola), according to Claude-Joseph ( 1926 ) , starts its turrets with the first soil excavated from the nest. This species and possibly D. analis (V^ichal) (Bertoni, 1925), and also some Ptilothrix seem fully as gregarious as Melitoma (Strand, 1909; von Ihering, 1904; Ducke, 1901). Ethology of Neotropical Anthophorine Bees 77 Among the most important characters are the form of the pollen mass, which in North American species, as shown by Linsley, Mac- Swain and Smith ( 1956 ) occnpies the bottom of the cell and is not formed into a ball, and the position of the egg, which in these spe- cies is beneath the pollen mass. In Diadama haeri (V^achal) (see Janvier, 1955) and D. chilensis (Spinola) (see Claude-Joseph, 1926), the egg is placed on top of a more or less spherical pollen mass, and Ruiz ( 1940, 1942 ) supports this statement in at least a general way for D. chilensis. In Diadasia ancdis (Vachal) also, the pollen mass is a ball, but the egg is placed at the side and be- neath it, as in North American members of the genus (Janvier, 1955). The observations reported below indicate that the egg of Ptilo- thrix plumata is on the surface of the ball-like pollen mass. As this was based on observation of a single nest it might seem to be an error. Strand (1909). however, also refers to the pollen mass of this species as a ball. Moreover, von Ihering (1904) describes the pollen mass as an oval, and states that the egg is laid after the provi- sioning is completed. From this we believe he meant that the egg is on top of the provisions.' Tribal characters, modified from those listed by Linsley, Mac- Swain, and Smith (1956), are as follows: Nests usually with turrets at entrances; cells arranged serially or singly, urn-shaped, the walls constructed by the bee so that in some genera ( Melitonia ) the cells are easily separable from the matrix; waxlike linings of cells ex- ceedingly thin or perhaps absent; cell cap not lined with waxlike material. Provisions a firm mass usually filling lower part of cell, sometimes ball-like. Egg usually below mass of provisions, some- times on top of it. Feces (or unused pollen) spread as a layer over entire inner surface of cell, no separate pellets visible in this layer but may be visible at top of cell. Cocoon very thin. Melitoma euglossoides Lepeletier and Serville This species has been seen by one of us (C. D. M.) nesting in large aggregations in banks near Yautepec, Morelos, Mexico and in adobe walls near Lima, Peru, and has been recorded nesting in sim- L On reading the manuscript of this paper. Dr. E. G. Linsley of the University of California made some comments that seem important in this connection. He notes that he and his associates have repeatedly observed partial pollen masses of Diaddsia and PtUothrix in cells which were heinj; provisioned. An en^ was never present until the poll' n mass was completed. This indicates that the egK is laid in its position beneath or at the side of the pollen mass after the mass is cr)mplet' })onsa, C. autrani, or C. fureata. Provisions are in the form of a paste occupying the lower half of the cell and sometimes covered with a layer of liquid. The egg is on the surface of the provisions. The cocoon is strong and in the species discussed by Claude-Joseph has a protuberance extending into the hollow pro- jection of the cell cap. Centris lanipes Fabricius Bertoni (1918) mentions finding hundreds of nests of this spe- cies in the mud walls of a house, and a single nest among those of an aggregation of Melitoma. He says that the burrows, one for each female, penetrate horizontally or obliquely for a few centi- meters. In 1929 the same author mentions large mixed colonies of these two species of bees, and niud walls completely mined over a period of years by the Centris burrows so that the holes intercom- municate and the cells touch one another. He also describes turrets at the nest entrances but to us it seems likely that these belonged to intermixed Melitoma nests since turrets are not elsewhere de- scribed for Centris. As there has been taxonomic confusion among relatives of lanipes, he may not have referred to the same species as that discussed below. Nests of this species were found by Father J. S. Moure in old pieces of logs and in wooden buildings near the beach at Caioba, 92 The University Science Bulletin J ■'C"^. Figs. 22 to 24. Nests of Centris lanipes (Fabricius) in holes in old wood. Frass and wood fragments in the holes are represented by short lines vari- ously oriented; the lined areas represent sandy clay perhaps placed in the holes by the bees. Parana, Brasil. He kindly brought one large piece of a long dead tree containing si.x Centris nests to the laboratory in Curitiba. The journey over rough roads apparently destroyed the eggs, so that their position in the pollen mass could not be determined when the nests were opened on December 5, 1955. The nests were all in old holes made by wood boring Coleoptera. These holes were often partially filled by frass and wood fragments evidently left there by the borer. However, the outer several centi- meters were filled by sandy clay, no doubt placed there by the bees, and the cells were constructed in this earth ( figs. 22 to 24 ) . In one case (fig. 22) earth formed a plug, perforated by a hole 5.5 mm. in diameter at the entrance. In this case the large space behind the Ethology of Neotropical Anthophorine Bees 93 plug gave the impression of being a cell under construction; the other cells in the nest were of recent construction as they contained no larvae; probably the bee was still bringing clay to the nest when its work was interrupted. From two to eight cells were formed in each nest, depending partly on the size of the space available for cell construction. The cells occupy all possible positions from horizontal to vertical. The>- are lined with waxlike material; dimensions of the cells are as fol- lows: length 11 to 13 mm.; maximum diameter, 6 to 7 mm.; diam- eter at mouth of cell, 4 to 5 mm. The provisions were semiliquid. Larvae of all sizes, but no prepupae or pupae, were found in the cells. Jorgensen (1912) records very similar nests of Centris nigricentri.s (Burmeister) in an earth filled bamboo stem and (1912a) in aban- doned insect burrows in posts. LITERATURE CITED Behtoni, a. de Winkelried 1918. Notas entomologicas ( Biologicas y Sistematicas), Anal. Cient. Para- guayos, ser. 2, no. 3, pp. 219-231. 1925. Himenopteros nuevos o poco conocidos, Rev. Soc. Cient. Paraguay, vol. 2, no. 1, pp. 74-79. 1929. Nidificacion en colonias de abejas Antoforidas, Rev. Soc. Cient. Paraguay, vol. 2, no. 5, pp. 223-224. Behtoni, A. de W. and C. Schkottky 1910. Beitrag zur Kenntnis der mit Tetralonia verwandten Bienen aus Siidamerika, Zool. Jalirb., abt. f. Syst., vol. 29, pp. 563-596. 1911. Geschlechtsdiniorphisnuis in der Bienengattung Thygater Holmbg. (Hyin.), Deutsche Ent. Zeitschr., pp. 402. BrPhhes, J. 1909. Una Anthuphorina parasita?, An. Mus. Nac. Buenos Aires, ser. 3, vol. 12, pp. 81-83. Cl.AUDK-JoSEPH, F. 1926. Recherches biologique sur les Hynienopteres du Chili, Ann. Sci. Nat., Zool., ser. 10, vol. 9, pp. 113-268. Cho.s, Auguste 1939. Eucera notata Lep. (Syn. Eucera obesa Dours), Verhandlungen \ II Intemat. Kongr. Entom. [Berlin], vol. 2, pp. 1079-1088. Custer, Clarence P. 1928. On the nesting habits of Melissodes Latr. (Hynienop. ), Canadian Ent., vol. 50, pp. 28-31. DOEI-LO-JURADO, M. 1912. Apuntes entoniologicos. Nidificacion y habitos de una abeja silves- tre, la Entechnia, Bol. Soc. Physis [Buenos Aires], vol. 1, pp. 52-56. 94 The University Science Bulletin DucKE, Adolf 1901. Beobachtungen iiber Bliimenbesuch, Erscheinungszeit etc. der bei Para vorkomnienden Biencn, Zeitschr. Syst. Hymenopterologic Dipterologie, vol. 1, pp. 25-32, 49-67. Fahhin(;kh, Josef and Fkanz Tolg 1912. Beitrage zur Kenntnis der Lebensweise und Entwicklungsgeschichtc einiger Hautfliigler, Verhandlungen Naturforschendcn Vereinos in Briinn. vol. 50 (for 1911), pp. 242-269. Frik;se, H. 1919. Die Langhornbiene Eucera difficiUs Duf. Perez iind ilir Nestban bei Arten, Deutsche Ent. Zeitschr., pp. 61-62. 1923. Die europaischen Bienen (Apidae), Walter de Gniyter and Co., Berlin and Leipzig, viii + 456 pp., 33 pis. Herbst, Paul 1923. Beitrage zur Biologic der Chilenischen Arten der Gattung Centris F. (Apidae), Zeitschr. Wissenschaftliche Insektenbiol., vol. 18, pp. 345-350. Hicks, C. U. 1936. Nesting habits of certain western bees, Canadian Knt., \-ol. 68, pp. 47-52. Holmberg, Eduardo Ladislao 1903. Dilectus Hymenopterologicus Argentinus, Anal. Mus. .Nac. Buenos Aires, ser. 3, vol. 2, pp. 377-512. HoppNER, Hans 1901. Weitere Beitriige zur Biologic nordwestdeutscher Hymenopteren, I. Eucera difficilis ( Duf. ) Perez, Allgemeinen Zeitschr. Entom., vol. 6, no. 3, pp. 33-35 (a figure for this paper appears on page 132). Ihering, R. von 1904. Biologia das abelhas solitarias do Brazil, Rev. Mus. Paulisla, vol. 6, pp. 461-481. luGA, Victoria G. 1950. Observatii biologice asupra albinelor solitare miniere: Tctrahnia nificnrnis F. si Eucera chjpeata Er., Buletin Stiintific, Acad. Re- publicii Populare Romane, ser. Geol., Georg., Biol., Stiinte Tehnice si Agr., vol. 2, pp. 251-268. Janvier, Hippolyte 1933. fitude biologique de quelques Hymenopteres du Chili, Ann. Sci. Nat., Zool., ser. 10, vol. 16, pp. 209-356. 1955. Le Nid et la nidification chez quelques abeilles des Andes tropicales, Ann. Sci. Nat., Zool., ser. 11, vol. 17, pp. 311-349. Jensen-Haarup, a. C. 1908. Biological researches amongst the Argentine bees with special refer- ence to flowers they visit. Flora og Fauna, pp. 95-108. Jorgensen, Lavrids 1920. Smaa lagttagelser af nogle danske Biers, Ent. Meddelelser, \ol. 13, pp. 153-159. Jorgensen, P. 1911. Los crisididos y los himenopteros aculeatos de la provincia de Men- doza. Anal. Mus. Nac. Buenos Aires, ser. 3, vol. 15, pp. 267-338. Ethology of Neotropical Axthophorine Bees 95 1912. Beitrag zur Biologic einiger siidamerikanischer Bienen, Zeitschr. Wissenshaftliche Insektenbiol., vol. 8, pp. 268-272. 1912a. Revision der Apiden der Provinz Mendoza, Repiiblica Argentina (Hyni.), Zool. Jahrb., Abt. f. Syst., Geogr., u. Biol, der Ticre, vol. 32, pp. 89-162. LiN.SLEY, E. C, J. W. MacSwain and Ray F. Smith 1954. A note on the nesting habits of Enomalopsis sohnii Coekerell ( Hymenoptera, Anthophoridae), Pan-Pac. Ent., vol. 30, pp. 263- 264. 1955. Biological observations on Xenoglossa fulva Smith with some gen- eralizations on biological characters of other eucerine bees ( Hymen- optera, Anthophoridae), Bull. So. California Acad. Sci., vol. 54, pp. 128-141. 1956. Biological observations on Ptilothrix sumichrasti (Cresson) and some related groups of emphorine bees ( Hymenoptera, Antho- phoridae), Bull. So. California Acad. Sci.. vol. 55, pp. 83-101. MiCHENER, Charles D. 1944. Comparative external morphology, phylogeny, and a classification of the bees (Hymenoptera), Bull. Amer. Mus. Nat. Hist., vol. 82, pp. 151-326. 1954. Bees of Panama, Bull. Amer. Mus. Nat. Hist., vol. 104, pp. 1-155. MiCHENER, Charles D., Rudolf B. Lange, Joao Jose Bigarella and Riad Salamuni 1958. Factors influencing the distribution of bees* nests in earth banks. Ecology, in press. MiCHENER, Charles D. and Jesus S. Moure 1957. A study of the classification of the more primitive nonparasitic anthophorine bees. Bull. Amer. Mus. Nat. Hist., vol. 112, pp. 395-452. Moure, J. S. and C. D. Michener 1955. A contribution toward the classification of neotropical Fuccrini (Hymenoptera, Apoidea), Dusenia, vol. 6, pp. 239-331. Nielsen, Erik Tetens 1941. Nestbau von Eucera longicornis L., Ent. .Meddeleiser, \()1. 22, pp. 142-149. Nielsen, J. C. 1902. Biologiske Studier over danskc enlige Bier og deres Snyltere, Videnskabelige Meddeleiser fra den Naturhistoriske J"orening i. Kjobenhavn, pp. 75-106. Peer, D. F. 1956. Multiple matings in queen honeybees. Jour. Econ. Ent. \nl. 49, pp. 741-743. Pickel, Bento 1928. Contribui^ao para a biologia de Centris sponsa e Acanthopus ex- cellens (Hymen.), Bolctim Biologico [Sao Paulo], vol. 14, pp. 135- 143. Rau, Phil 1929. The biology and behavior of mining bees, Anthophora ahrupia and Entechiiia faiirea, P.syche, vol. 36, pp. 157-181. 96 The University Science Bulletin RozEN, Jerome G. and C. Don MacNeill 1958. Biological observations on Exomalapsii- (Anthophorula) cJiioimra Cockerell, including a comparison of the biology of Exomalopsis with that of other anthophorid groups, Ann. Ent. Soc. Anier., vol. 50, pp. 522-529. Ruiz P., Flaminio 1940. Apidologia chilena. Rev. Chilena Hist. Nat., vol. 44, pp. 281-377. 1942. Notas biologicas de algunos generos de abejas solitarias de Chile, Bol. Sanidad Vegetal [Ministerio de Agricultura, Santiago], vol. 2, no. 1, pp. 8-16. SCHROTTKY, C. 1901. Biologische Notizen Solitiirer Bienen von S. I'aulo, Allg. Zeitschr. Ent., vol. 6, pp. 209-216. 1904. Beitrag zur Kenntnis eniger siidanierikanischer Hynienopteren, Allg. Zeitschr. Ent., vol. 9, pp. 344-349. Su.VESTRI, F. 1903. Contribuzione alia conoscenza dei Terniitidi e 'rerniitofili dell' America nieridionale, Redia, vol. 1, pp. 1-234. Smith, Frederick 1855. Catalogue of Briti.sh Hyinenoptera in the Collection of the British Museum, Part 1, pp. 1-252. Strand, Embrick 1909. Beitrag zur Bienenfauna von Paraguay auf Grund der im Berliner Museum vorhanden Sanimlung von Karl F"iebrig, und der Bestim- niungen von Dr. H. Friese zusammengestellt, Deutsche Ent. Zeitschr., pp. 227-237. Vesey-FitzGerald, D. 1939. Observations on bees (Hym.: Apoidea) in Trinidad, B. W. I., Proc. Royal Ent. Soc. London, ser. A, vol. 14, pp. 107-110. THE UNIVERSITY OP KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 4 The Flora and Ecology of the Chautauqua Hills in Kansas BY Earl W. Lathrop * Abstract. This paper reports 1030 taxa of vascular plants for the Chau- tauqua Hills Physiographic Province in southeastern Kansas. Species percent- age composition for the characteristic plants of the woodland and prairie is included. An account is given of the original vegetation of the area at the time of settlement. The higher plant associations of the province are considered along with de- scriptions of their seasonal aspects. Eleven tables, two maps, and 21 plates, made from photographs by the author, are included in this work. The vege- tation of the Chautauqua Hills in Kansas is most closely related to that of the Texan Biotic Province. TABLE OF CONTENTS PAGE List of Tables 98 Introduction 98 Acknowledgments 101 Climate 102 Original Vegetation 102 Present Vegetation 109 Upland Woods 110 Prairie 113 Flood Plain Woods 135 Prairie Woodlands 135 Ponds and Lakes ? 136 Marshes 136 Gullies and Ravines 136 Invaded Woodlands 137 Roadside Plants 137 Cultivated Fields 137 Abandoned Fields 138 Relationship of the Chautauqua Hills Flora to That of Adjacent Areas 138 * Assistant Professor of Botany, La Sierra College, Arlington, California. (97) 4—26.56 98 The University Science Bulletin PAGE Annotated List 140 Summary 185 Bibliography 186 Plates 189 LIST OF TABLES Table 1. Percentage of Tree Species by Size Class on Hilltops and Slopes of Rolling Upland Woods 114-115 2. Percentage of Tree Species by Size Class in Gullies and Slopes of Upland Woods 116-117 3. Percentages of Tree Species by Size Class in Flood Plains and Border- ing Uplands 118-119 4. Percentages of Tree Species by Size Class in Upland Woods Sur- rounded by Prairie 120 5. Percentages of Tree Species by Size Class in Ravines and Slopes of Upland Woods 121-122 6. Percentage Composition and Basal Area of Grass Species and Forbs in Areas of Prairie Type No. 1 126 7. Percentage Composition and Basal Area of Grass Species and Forbs in Areas of Prairie Type No. 2 127-128 8. Percentage Composition and Basal Area of Grass Species and Forbs in Areas of Prairie Type No, 3 128-130 9. Percentage Composition and Basal Area of Grass Species and Forbs in Areas of Prairie Type No. 4 131-132 10. Percentage Composition and Basal Area of Grass Species and Forbs in Areas of Prairie Type No. 5 133-134 11. Percentage Composition and Basal Area of Grass Species and Forbs in Areas of Prairie Type No. 6 134 INTRODUCTION The name Chautauqua Hills was first appHed by Adams (1899) to that area lying within a triangular belt approximately ten miles wide extending from Yates Center in Woodson County, Kansas, southward to the State line in Montgomery and Chautauqua coun- ties (Maps 1 and 2). Schoewe (1949) describes the area as fol- lows. "The Chautauqua Hills are developed chiefly in the thick sandstone of the Douglas group which farther to the north are re- placed by shales and which there, because of their position between the limestones, produce escarpments." As a result of erosion, the surface of the sandstone belt has been dissected into a series of low hills which are intersected by deep gullies which have been cut by many small streams. The Verdigris, Fall, and Elk rivers cross the area (Map 2) in low narrow valleys. In many places the rivers and valleys are bordered by bluff's which show sandstone outcrops ( Plate 17). Flora and Ecology of Chautauqua Hills 99 This area is described by McGregor ( 1955 ) as being distinguished primarily by sandy soil and sandstone capped hills whose tops and upper slopes are forested (Plates 1 and 2) with Quercus stellata, Q. marilandica, Q. velutina, and Q. prinoides (Table 1) with Q. muehlenbergia, Q. macrocarpa, Canja cordiformis, C. ovata, Ulmus americana, U. rubra, and Celtis occidentalis becoming admixed with MAP I fi GREENWOOD CO. > ^^ YATES CENTER WOODSON CO. FREDONIA ELK CO. CHAUTAUQUA CO.#- \ , ^_ £^ Map 1. The Chautauqua Hills area showing location of upland woods and major cities and highways. Copied from Hale (1950) as taken from aerial photographs, 1954 ( U. S. D. A. Commodity Stabilization Service, In- dex Maps, scale 1:20,000). 100 The University Science Bulletin the above on the lower slopes and along the streams (Tables 2, 3, and 5 ) . Yates Center is the chief town at the north border of the Chau- tauqua Hills area, Toronto, Fall River, Elk River, Sedan, and Elgin mark its west border whereas Tyro, Elk City, New Albany, and Coyville define its east border. MAP 2 ^^f 2. Tracing from the Geologic Map of Kansas, Moore and Landes (1937), showing the area of Douglas sanclstone and major stream courses of the Chautauqua Hills. Flora and Ecology of Chautauqua Hills 101 The Chautauqua Hills area is a physiographic province in Kansas. Actually it is the northern limits of the Texan Biotic Province and is bounded on the east and west by physiographic provinces, the Osage Plains and Flint Hills respectively, and on the north by the Illinoian Biotic Province. From the east there is an influx from the Carolinian and on the west from the Kansan Biotic Provinces. Thus the area is unique in that it is a unit with three boundaries which form transitions. In addition to Schoewe ( 1949 ) and McGregor ( 1955 ) one other publication. Hale (1955), includes mention of the characteristic trees of the Chautauqua Hills as being Quercus marilandica, Q. stel- lata, and Q. vehitina. Miller (1932), Caldwell (1937), other early survey records, and a few historical accounts include mention of the trees of this area. A few publications. Gates (1940), Cockrum (1952), and Fearing (1952) have included discussions and maps which involve the vegetation of this area to some degree. Other publications have been produced which give lists of species found in the various counties of the Chautauqua Hills area. Some of these are Gates (1940), Stevens and Dill (1940), Horr and McGregor (1949, 1951), McGregor (1950, 1956), Humfeld (1951), McGregor and Horr (1949, 1950, 1953, 1955), McGregor and Voile (1950), Yokoyama and McGregor (1951), McGregor and Hartman (1956), McGregor and Lathrop (1957), and Lathrop (1957). Previous to the present study, extensive field work had been done in the area by Dr. Ronald L. McGregor of the Department of Botany, University of Kansas. As a result, many plant specimens are on file at the Univer- sity of Kansas Herbarium which supplemented the writer's collec- tions. Dr. W. H. Horr and Mr. B. L. Wagenknecht have also col- lected numerous specimens from this area. Up to this time, no one has made a study of the vascular flora of the Chautauqua Hills area of Kansas. Since it is in a region of tran- sition of other vegetation areas, it was felt that a floristic and eco- logical study of the Chautauqua Hills would be a worthwhile addi- tion to the knowledge of Kansas plants. ACKNOWLEDGMENTS The writer is grateful to Dr. Ronald L. McGregor for the original suggestion of this problem, guidance, and assistance in the deter- mination of some 3,000 specimen collections. Acknowledgments are also due to Dr. W. H. Horr, University of Kansas, who has given invaluable aid on ecological problems, and Dr. Robert W. Baxter, 102 The University Science Bulletin University of Kansas, for his assistance in this study. Appreciation is given to Mr. Ernest J. Palmer for his determination of many .speci- mens of oaks and to Dr. James C. Mahn for his help in locating his- torical records. The writer is also indebted to Dr. McGregor, Dr. Horr, Dr. Baxter, and also Dr. Henry S. Fitch, Department of Zoology, for generous allotment of their time during the editing of this paper. I especially want to thank my wife whose help made this work possible. This study was aided by a grant from the Sigma Xi-RESA Research Fund and by a grant from the Kansas Academy of Science. CLIMATE Annual temperatures ( ° F. ) for the Chautauqua Hills computed from records covering the years 1898-1942. Mean maximum Normal Mean minimum 68.6 58.4 46.7 Annual precipitation (inches) for the Chautauqua Hills com- puted from records covering the years 1898-1942. Maximtim, ijr. 1915 Normal Minimum, yr. 1936 52.68 37.70 29.52 Average dates of the killing frost and the average growing season for the Chautauqua Hills based on all available records through 1945. Average date of the first killing frost in fall — October 23. Average date of the last killing frost in spring — April 14. Average length of the growing season — 192 days. ORIGINAL VEGETATION A knowledge of the vegetation of the Chautauqua Hills before the time of settlement between the years 1870 and 1880 must be gained largely from surveyors' notes and writings of early residents of the area. These written records include mention of native plants from time to time. However, such mention was usually casual, and the exact species frequently could not be determined. Supple- mental information was derived from oral statements of long time residents of the area concerning the early vegetation and from data of tree ring counts. The latter information was obtained by re- moving small cores of wood from 83 upland trees in the area by means of an increment borer. From these cores it was possible to estimate the approximate age of the tree by counting the annual growth rings. The counts indicate that many upland trees now Flora and Ecology of Chautauqua Hills 103 growing in the area were already there in the early and middle eighteen hundreds before the area was settled by the white man. Six Quercus stellata, ranging in diameter from 12 to 20 inches, dated from the year 1852 to as far back as 1769. One Q. mariJandica, 17.6 inches in diameter dated back to 1851. For 11 trees, 5 Q. stel- lata, 5 Q. marilandica, and 1 Q. velutina, ranging in diameter from 10 to 19 inches, growth rings showed that they began their growth during the 1870's. Additional evidence that the Chautauqua Hills were wooded before and at the time of settlement is to be found in the original surveyors' notes of the area. These records are on file at the State Auditor's Office in the State House and at the Kansas State Histori- cal Society, Topeka, Kansas. Tliese notes, taken in the field during the surveys of township and section lines between the years 1856 and 1877, include mention of the kinds, and in some cases, the diameter of trees on or near the lines. A general description of the land in the area of the surveys was sometimes recorded by the surveyors. Frequent mention is made, in these notes, of blackjack oak (Quercus morilandica) , post oak (Q. stellata), and to a lesser ex- tent, black oak (Q. vehitina) as occurring in the uplands of the Chautauqua Hills. The diameters listed for these species were, on an average, slightly larger than for the same species today, being from 6 to 20 inches. However, several diameters of from 24 to 30 inches were recorded for the above species, indicating that at least some of the trees, growing at the time of settlement, were of even older age than those of the present. Descriptions of the area of the Chautauqua Hills while it was still in undisturbed condition, were included in these Public Land Survey notes. The surveyors described the vegetation and physiog- raphy with respect to specific landmarks that can be easily located at the present time. Their descriptions of the areas surveyed are quoted below, in part. ( The present site of Yates Center, Wodson County, 1856. ) "Tim- ber black jack open woods, land graveley prairie. Land south part of black jack ridge, third rate — north part level first rate timber of black jack, post oak, hickory (Carya), and black walnut (Juglans nigra ) ." (Two miles west of Yates Center, 1856.) "Land very broken with large sandstone on surface. Timber oak (Quercus), hickory, black walnut, black jack, and bur oak (Quercus macrocarpa), un- 104 The University Science Bulletin dergrowth the same." (Chautauqua Hills area between Verdigris and Fall rivers, 1866.) "On either side of the two to three mile wide valley of the river (Verdigris) are broken ridges, the out- crops of rock being chiefly sandstone. Bluff soil poor, valley of Verdigris and smaller valleys have a rich alluvial soil covered with a fine vegetable mold. There is a belt of superior timber along the river and the hills are in several places covered with a young growth of post and black oak . . . Timber is abundant in the valleys and in many places is found in considerable quantities upon the bluff^s . . . Almost entire surface is rolling and broken . . . with sandstone outcrops . . , The naturally rich, well tim- bered, and beautiful valleys of Indian Creek and Fall River . . . timber along Fall River composed of red elm ( Ulmus rubra), white elm (U. americana), bur oak, hackberry (CeJtis), oak, black wal- nut, sycamore (Plat anus occidentalis) , soft maple (Acer sacchari- num), Cottonwood (Populus deltoides), mulberry (Morus), coffee bean (Gymnocladus dioica), and buckeye (Aesculus). Under- growth the same plus willow (Salix) and paw paw (Asimina tri- loba) . . . Along the bluffs and on the ridges there are large bodies of inferior scrub oak (Quercus stellata and Q. marilandica) . . . elsewhere a broken prairie except along the Verdigris and Fall Rivers." In the year 1857 a surveying party, led by Col. Joseph E. John- ston, surveyed the southern boundary of the Kansas Territory. The party crossed the Chautauqua Hills along the southern line of present Chautauqua and Montgomery counties. Descriptions of the area, taken from the private journal of Col. Johnston and re- corded by Miller (1932), are quoted below, in part. (Approaching the Chautauqua Hills area from the east.) "The line today ran parallel to Russell Creek . . . the country gently undulating and soil rich black loam, limestone showing it- self occasionally. Woodlands two or three miles to the southeast on the crest of a ridge beyond the creek." ( On the divide between 12 Mile Creek and an affluent of the Verdigris.) ". . . the latter runs to the southwest in a broad and beautiful valley, the western side of which is abrupt, wood scattered through it." ( Northeast part of present Coffeyville, Montgomery County. ) "An Osage village of 27 huts, a half mile west of the ford ( of the Verdi- gris River), the inhabitants buffalo hunting. . . . Two miles to the south is Nickeokaka (Onion Creek), well wooded. . . . Went to the crest of the ridge this side of the Little Verdigris ( North Flora and Ecology of Chautauqua Hills 105 Caney Creek; Map 2) from which . . . the country appeared to be much more broken and wooded than the east side of the Verdigris ... in 5 miles reached the Little Verdigris . . . moved on to a little south of west to avoid rugged hills. . . . The country, especially to the north, very broken. A good deal of oak in the heights . . . passed over a rugged ridge covered with post oak." (Approximately 2 miles west of Hewins, Chau- tauqua County. ) "Our last camp (near Elgin, Chautauqua County) was just with timbered country. The march today was in prairie. The dividing ridge opposite is a plateau (Fhnt Hills) about 300 feet above this valley; the sides very abrupt and rocky. Limestone near the summit." Similar notes were taken by Hugh Campbell, astronomer on the same survey. His descriptions of the area are recorded by Caldwell (1937) and are quoted below, in part. (Approximately 3 miles southeast of the present Coffeyville, Montgomery County. ) "After crossing Pumpkin Creek we found ourselves ascending beautiful heights or uplands covered with the most luxuriant grass and other vegetation, particularly flowers of various colors ... As the ascent continues, little specks of timber can be seen in all directions . . . on some of these ( small streams ) I found great numbers of wild rose bushes." (Along the march westward.) "The country over which we pursued our course presented a different aspect to that hitherto traversed. Ridges, mounds, and small elevated table- lands, covered with a luxuriant growth of vegetation, intersected with lines of timber marking the courses of gullies or small streams, now occupies the view. The country as far as the eye can see has the same broken and irregular appearance . . . Encamped on a small stream, the banks of which are well wooded with sycamore, Cottonwood, and hackberry." (At Horse Head Creek.) ". . . well timbered on the west bank with oak, cottonwood, sycamore, walnut and cherry (Prumis.) The country on this day's march is still more abrupt and broken . . . passed over many ridges very rocky and covered with dense growth of black jack." (At present Caney City, Montgomery County.) "Passed through a large grove of post oak." (Slightly west of Elgin, Chautauqua County.) ". . . the grass here, as well as on the neighboring highlands is excellent. The principal productions are as follows: Vis. Timber in immediate vicinity of river consists of oak sycamore, and walnut . . . Fish: Cat, trout or bass, buffalo and garr. Game: Deer, antelope, and turkeys are very numerous." 106 The University Science Bulletin Public Land Surveys were not carried out in the southern part of the Chautauqua Hills until approximately 14 years after the northern half was surveyed. Descriptions of the area in survey notes of 1871 are quoted below, in part, (Approximately 6 miles northeast of Sedan, Chautauqua County.) "Timber along stream (Middle Caney Creek) of elm, sycamore, walnut, ash (Fraxinus), and hackberry. Undergrowth of vines . . . Water (of the stream) clear and pure . . . Timber on upland post oak." ( The vicinity of present Elk Falls, Elk county. ) "Creek bottoms . . . are well watered by many fine springs from the bluffs. Clear Creek, Wild Cat Creek, and Elk River are well timbered . . . ash, walnut, oak, hackberry, and sycamore. There are also several groves of upland timber consisting of black jack and post oak. . . ." (General description of the land eight miles north of Sedan, Chautauqua County. ) ". . . mostly high broken prairie covered with sandstone . , . eastern portion covered with black jack and post oak." (T30S, R12E, east central part of Elk County.) "Center of township is rough, rolUng, and broken by ravines . . . the hills covered with a scrubby timber known as post oak and blackjack . . . Prairie covered with sandstone." (T33S, R14E, west central part of Montgomery County.) "Land mostly second rate . . . except in the eastern and western portion of the township where there are high mounds rising from the rolling prairie, composed of sandstone. Timber on mounds post oak and black jack. Along streams elm, sycamore, and Cot- tonwood." The notes of the early surveyors for this area seemed to be con- sistent in mentioning timber for both uplands and bottomland. Upon examination of the notes, however, there are indications that the upland timber was not as dense as it is today. In hsting trees which were marked at section and township corners, their distance from the comer is given and often their space relationship to other trees. Frequently these distances were relatively far and many times mention is made that no other trees were near or available to mark. The uplands were not one continuous woods. The notes include mention of considerable upland hilly prairie over sandstone which was not wooded, just as it is today. Also the relatively large diame- Flora and Ecology of Chautauqua Hills 107 ters given for many of the upland trees indicate that they would have likely been well spaced. Statements from Lockhart (1927), concerning the history of Elk Falls, Elk county, include, "The first settlers arrived over two trails, one from Independence country following the river (Elk) valley, which was covered with bluestem grass higher than the back of an ox. . , . Historical incidents and facts recollected by some of the older residents of the vicinity: Mr. Wert Wicker remembers when places, now covered with a dense growth of timber, could have been mowed, and some, at that time, ( 1877-79 ) actually were hay fields; Mr. George Bennett remembers when the bluestem grass was dense everywhere on the prairie and no country roads were marked, and few fields were broken." General descriptions of the counties in the area of this study written by Andreas ( 1883 ) are quoted below, in part. (Woodson County.) "The county is largely upland . . . Along the streams are belts of timber . . . averaging a half mile in width. In these belts are found oak, elm, hickory, black walnut, hackberry, sycamore, and cottonwood. (Greenwood County.) "It is ... 5 percent of forest and 95 percent of prairie." (Elk County.) "The surface of the county is chiefly high, broken prairie, particularly back from streams, and in the western part the prairies are of increased elevation, rising into what are known as the Flint Ridges (Flint Hills). The uplands . . . although rocky and light of soil, produce abundant grasses . . . being supphed with abundance of clear running water." (Wilson County. ) "Along the streams are timber belts . . . embracing oak, hickory, walnut, hackberry, elm, soft maple, pecan (Carya illinoensis) , sycamore, ash, cherry, basswood (Tilia), and some cedar (Juniperus virginiana)." (Chautauqua County.) "The county is covered with a rich growth of wild grasses . . . and an abundant growth of timber . \ . in some places it covers the hills and draws . . . the timber . . . upon the bluflFy lands is made up of a species of oak vulgarly known as 'black jack.' " (Montgomery County.) "Timber, as in most parts of the state, is scarce, being confined to the belts along the water courses . . . occasional groves are found upon the uplands, but the timber is limited, and of inferior quality." Early township maps of the area often showed the extent of forests. These maps were made up solely on the basis of data 108 The University Science Bulletin written in the early surveyors' notes previously mentioned. These maps show that most of the uplands in the Chautauqua Hills, which are wooded at present, were covered by forests, to a slightly less extent, at the time of the surveys ( 1856 and 1871 ) . Further information concerning the distribution and spread of the forests is afforded by interviews with elderly residents who have lived in the Chautauqua Hills for all or most of their lives. Their recollections of the extent of timber on the uplands is as follows: C. C. Cox of Elk City, Montgomery County, stated that the scrub oak has increased a lot as far back as he can remember (to the 1890's ). What used to be just a few blackjack near his farm is now a forest of the same. He recalled a prairie area that is now a scrub oak woods. He said that some hills were bare when he was young, but that there were scattered trees on most of the hills of the area; Art Oliver of Elk City can recall trees being on the hills and prairie hills as far back as 1897. He said that the oaks scattered into the prairie before hedge rows were planted and that they have spread down over the hills. He said that there is more post oak and black- jack now than 60 years ago, but that there has always been some on these hills. Mr. Oliver said that his father once pointed out a hill to him and stated that when the mature trees there were cut down, the forest that came afterwards was more dense; William Row of Toronto, Woodson county, remembered his father telling him that there was scattered post oak on the hills at the time of settlement. Mr. Row said that there has been an increase of the post and blackjack oak in the area in the last 65 years; Joe Dexter of Longton, Elk county, said that the woods on the hills of the area looked about as they do now as far back as he could remember; Frank Burtenshaw of Elk City, Montgomery county, stated that the scrub oak on the hills covers more area now than it did 50 to 60 years ago and that the trees have spread from the hilltops onto the slopes and into ravines. He stated that the mature trees had been cut in the past and that the growth that replaced it was thicker; Johnny Wilson of Havana, Montgomery county, said that the trees have increased on the hills and have spread. He remembered areas which used to be maintained as hay fields that are now for- ested. He stated that mowing and burning of hay fields and open woods kept down saplings of scrub oak and that the latter grew up as dense timber after mowing and burning ceased, allowing the young growth to continue year after year. Following a study of the list of native trees of the Chautauqua Flora and Ecology of Chautauqua Hills 109 Hills as compiled from historical records, one is impressed by the fact that it corresponds with the common trees of the present wood- lands of the area. It is significant that there is frequent mention of trees on the uplands of the area under consideration because Kan- sas has usually been thought of as being barren of trees except along water courses and valleys. A statement by Fitch and McGregor ( 1956 ) , concerning a wood- land study in northeastern Kansas, applies equally well to this area. "The belief that this and similar areas in northeastern Kansas were virtually treeless at the time of occupation by white settlers is shown to be wholly unfounded by the information obtained from growth rings." For the Chautauqua Hills area, information from survey notes and other historical accounts can be added to the latter evi- dence. PRESENT VEGETATION With the disturbance by the plow, ax, and grazing, the changes in the vegetation of the Chautauqua Hills since settlement has been considerable. These changes have been not so much changes in composition as in extent of area covered. Evidence indicates that there is more forest and much less prairie now than before settle- ment. It is probable that regular burning of prairies kept back the spread of scrub oak as did mowing and burning in prairie wood- lands by the settlers. Much of the upland timber was cut down 60 to 70 years ago by the early settlers for fuel and lumber. Since that time the trees have reproduced from seed, and much second growth has developed from the stumps and suckers from roots. Since fire has been under control, trees have advanced down slopes and into prairie uplands as far as the environment permitted. Most of the lowland and some of the upland prairie was plowed for cultivation. Much of the latter area is in use now for grazing and native hay but is by no means in an undisturbed condition. Yates Center, in Woodson county, is considered the prairie hay capital of the world. Of the 2,678,440 acres in the Chautauqua Hills area, 188,511 acres were in native hay in 1936 and 107,370 acres in 1952. Disturbance has allowed for invasion of plants into fields and overgrazed pastures. Besides invasion of weeds into these places, some trees have become established also. There is great diversity of plant groupings in the Chautauqua Hills ranging from marsh to flood plain woods, prairie, and rocky 110 The University Science Bulletin wooded gullies. The following habitats are typical and representa- tive of the plant communities in the Chautauqua Hills area. Upland woods: These forests exist in continuous to scattered stands (Map 1) on sandstone of the Douglas group (Map 2) throughout the Chautauqua Hills. The forests are characteristically located over low rolling uplands (Plates 1, 3, and 5; Fig. 2) and cap prairie hilltops in numerous places ( Plate 2 ) . However, the aspect of the upland woods changes from place to place. Where these woods border gullies, ravines, and flood plains, the slopes are usually wooded their entire extent (Plates 4 and 5). In prairie areas the upland woods are generally restricted to the upper slopes and high places (Plates 1 and 2). In such areas, Quercus stellata and Q. inarilandica among others (Tables 1 and 4), grow on rough, rocky land where run in water is received from sandstone surfaces and where snow lodges during the winter. These trees are located well above the water table and must withstand varying periods of drought. Seedlings do not become established on lower prairie slopes because the grasses are better equipped to obtain the avail- able moisture. To understand better the composition of the upland woods, the trees in 16 representative areas throughout the Chautauqua Hills were sampled by 45 strip transects. With these, the species and diameter of the first 100 trees occurring in a strip two meters wide were recorded. The data of transects which were taken in areas showing a similar aspect were tabulated and the percentages of the trees by size class were worked out ( Tables 1-5 ) . The distribution of other Quercus stellata — Q. 7narilandica type woodlands are included in the writings of several workers. The oak-hickory forest region, as recognized by Braun (1950) extends from Canada to Texas as the westernmost part of the Eastern De- ciduous Forest. A Q. stellata — Q. marilandica type woods, among others, is listed as occurring in the Piedmont (oak-hickory region of North Carolina ) by Bourdeau ( 1954 ) . His statements are quoted below, in part. "The post oak — black jack oak forest (or the Piedmont) is called precUmax or subclimax by Costing (1942). From a polyclimax point of view ... it would be a physiographic or edaphic climax . . . Whatever the terminology used, post oak — black jack oak forests are restricted, in the North Carolina Piedmont, to dry south facing bluffs, thin rocky soils, strongly eroded soils, or to Flora and Ecology of Chautauqua Hills 111 the Iredell, Orange, and other closely related soil series . . . Trees in these sites are slow growing, rarely exceeding 30 feet in height . . . the stands are poorly stocked and do not form a closed canopy." Braun (1950) lists similar communities of Qiiercus stellata — Q. niarilandica, which exist throughout the Oak-Hickory Forest re- gion, as follows: (1) in the Flatwoods Belt of Alabama and Mis- sissippi, a flat area of heavy soils, water logged during rainy pe- riods, dry and hard at other times; ( 2 ) on the poorest soils in the up- lands of the Coast Plain west of the Mississippi River; (3) on drier ridges with sandy slopes and on steep more or less southerly slopes of the Ozark Plateau, the Boston Mountains, and the Ouachita Mountains; (4) on the open xeric plateau of the western border of the Ozarks; (5) on extensive areas of the old Kansas drift (Mis- sissippi Valley); (6) on the Illinoian drift area (Prairie Peninsula section). The Quercus stellata — Q. marilandica communities are also represented in the Oak-Chestnut, the Western Mesophytic, and the Mixed Mesophytic forest regions. In discussing the Oak-Hickory Association, Costing (1956) stated that throughout the association, various combinations of oak and hickory may occur as preclimax. His illustration was the Quercus stellata — Q. marilandica which occurs widely on poor sites and dry exposures in the association. In Texas and Oklahoma the above species occur in open, savannalike stands, known as Cross Timbers ( Dyksterhuis ) 1948, which mark the transition to grassland in that area. Rice and Penfound (1956) list Quercus stellata and Q. marilan- dica as dominants in the deciduous forest frontier in Oklahoma extending in a strip about 140 miles wide east of U. S. 81 and ex- tending from Kansas to Texas. The upland woods of the Chautauqua Hills extend through Oklahoma and on south to the Cross Timbers in Texas. Several communities of Quercus stellata — Q. marilandica also occur as isolated stands on low sandstone capped hills in outlying areas of the Chautauqua Hills and in limited areas of the extreme southeastern corner of Cherokee County, Kansas. Fitch and Mc- Gregor (1956) reported a few small groves on slopes of a wood- land area in Douglas County, Kansas. McGregor (1948) listed Quercus marilandica as being dominant in some strictly sandy soils of hillsides in the latter county. 112 The University Science Bulletin In the understory of the upland woods (Plate 3; Fig. 2), shrubs such as Symphoricarpos orbiculatus, Rhus copallina, R. glabra, Rtibtis allegheniensis, R. flagellaris, and R. occidentalis grow be- neath the trees and along margins of the woods. Numerous other characteristic species are found and can be listed as spring, sum- mer, and fall aspects. Species in the spring aspect are usually evi- dent because of their coloration. Several species of the woodlands come out early in the spring, have a short growth and die down. Characteristic of such are Hedyotis minima, Androsace occidentalis, Claytonia virginica, Viola pedata, V. papilionacea, Draba reptans, D. brachycarpa, and Sibara virginica. Some of the other common species of the spring aspect in upland woods are Erythroniwn meso- choreum, Chaetopappa asteroides, Viola kitaibeliana var. rafinesquii. Erysimum repandum, Corydalis crytallina, Arenmia patula. Gera- nium carolinianum, Antennaria fallax, Nothoscordium bivalve, Gal- ium aparine, Hypoxis hirsuta, and Fragaria virginiana. Species of sedges which are common at this time are Carex blandu, C. pensylvanica, C. bickneUii, Eleocharis obtusa, and E. engelmanni. Some of the rushes, which appear shortly after the sedges, are J.uncus torreyi, }. marginatus var. setosa, and /. interior. For a period of a few weeks after the peak of the spring aspect is past, few new species appear, after this the summer aspect be- comes evident. It is not as colorful as the spring aspect, but many species are conspicuous for their relative large size. Characteristic of the latter are Tephrosia virginica, Amorpha canescens, A. fruti- cosa, Desmodiwn illinoense, D. sessilifolum, D. glutinosum. Cassia fasciculata, Lespedeza virginica, Apocynum cannabinum, Hiera- cium gronovii, Liatris squarrosa. Aster patens, A. ericoides, Chrys- apsis pilosa, and Solidago ulmifolia. Other less conspicuous species making up the understory of the summer are Plantago virginica, P. purshii, P. aristata, Polytaenia nuttallii, Talinum parviflorum. Poly gala incarnata, P. sanguinea, Lechea tenuifolia, Cyperus fdiculmis, C. ovularis, and Parietaria fioridana. Some of the more common grasses of the summer aspect are Routelous hirsuta, Aiuhlenbergia racemosa, M. brachyphylla, Paspalum ciliatifolium, and Festuca ovina. In the fall the understory of the upland woods is made up of numerous species of grasses as well as several colorful flowering species. A few of the grasses are Andropogon scoparius, Sporo- bolus asper, Aristida dichotoma, A. purpurascens, Muhlenbergia Flora and Ecology of Chautauqua Hills 113 capillaris, Panicum agrostoides, P. spaerocarpon, Leersia oryzoides, and Agrostis hiemalis. Other species of plants characteristic of the fall aspect are Liatris scariosa, Desmodium mariJandicum, Spi- ranthes cerntia, Acalypha gracilens, Solidago missouriensis var. gla- berrina, Conyza canadensis, and Oxalis europaea. Prairie: The prairie which once covered much of the Chautauqua Hills is now confined to a few hay fields, considerable pasture, and to a few relics along railroad right of ways, and roadsides (Plates 1, 6, and 20: Fig. 2). 114 The University Science Bulletin CO O o B CO (D a -a c CO a o o CO CO u N CO V 'o a m ID H CO ID bo a 0) u 11 V ac 6? o 43 CO O) -<-3 03 ti — I ■■■••■•■■ -O -rH ^ (N lO CO ■* b- lO •■■••••• -o -o ■^ ■ .— ( lO lO (N IM (N (N (NC^-H,-HOOOlO.-H^t^ GO 00 ^ CCCD CO I— ii— I.— iCD"— ICO"— I"— ICO ^ (N (N OO -H CO u 111 CO CJ o CO -j^ O "^ e s o 05 'Si se ^ -is ^ fc ?», CO & 00 OO CO cc 2 3 3 3 o cj O cj V. ;.^ *^ t^ « <» « " e o cj ■^ 2 fc 3 00 05 3 3 S 3 3 a : ^ j::e«v'»oj;»:33w333 t; O O O O O O ^ '^ '^s X O-O-O o Flora and Ecology of Chautauqua Hills 115 &5 »— t IC (N O iC O lO J 00 '.- rH rfl Tj( o (Nl^ -« H ftj TS 3 ♦*— ■u O J 1 1-H O bO 1 O -4^ r- 1 w S 2 O O CO 1—1 1 a (U 1—1 a -»^ m 'Oi a D bO ^ lO oo 3 c o i rt Tt< a> I-H Uh c3 O «5 05 o O iC o 55 .5 1 00 CO 1— t 13 .2 C CO 05 tn t« -2 o "o t^ ec^ lOO •M ^ *-H t>'iO c^ E W c o M W) c« lO O -hO lO lO lO 0) i 00 rt -H"oi 00 N C/5 >> 1 U2 V) U 'o V o. C/2 d) s s " fe S S o a. <5 G to o •""O ';:i o o o » 116 The University Science Bulletin 8 C "a a c 03 "3 O a U C/D to "o 0) c &4 n CO e3 •♦J s o3 a o 03 03 tn (U J3 o C tn tn 03 ^ ^ T i I 00 I 1-4 o a oo oooo 00.t o oo 0(N OO 00 03 oo 030 bCg » O >~ 1^ ooo r-^OO oooooo C-i 00 --I oo c^ — < oo r-. -rf oo odd oo die o o — CO oo '-id oo o CO oo o o oo COIN OOOO iM(N O — ' CO CO OOO «0 05 "-I OOOOOO cj m' -H ^' d 00 -S s 1/^ 2 S "» s fe « « cc ;-^0'0''O O o 03 3 O « « " O »5 1 S ^ 5?S 0>iSli) M e - '«* 1 1 C c3 >> e S ■5 CO 00 6" CO 3 h. o3 C O T! a> CO CO 01 j3 t) C to CO 03 (V CO o H I loooo O'OiO'OO'OOOOicoO'OicicioO'O tfjiC Tt<.-H.-ir-i.-H(N-^o6oClC^oi'*iO M i-H ^ CO I 2 I CO in OiO iCiO »ood oo«o lO lO iC lO iCiCO « CD 3 **-• ^ o o o 8 %> OO o3 > e •*^ CO s e:2 oicicoooiooooomo o o M I »Ci lO >o lO 03 I 00 ^ o o lO lO lO >o 00 o lO oo OO tn ^ <^ « so c to fe ^ -: s s o >o oio OiO lO lO O lO »o t— I r-H -^ lO o»o CO O O O lO lO O >0 O lO p iC s e to tJ c! to o 13 -« re e c • !- a .0 •2 S S3 o ao S" CO '2 S £ -^ ~ S 05 -g C 2j S S s CO e o S s u •e |l s S; 2 e 5. ■a j^ iC »C O iC o 10 10 13 V. « « S !~ h eo to 00 S 3 S O U W =* « ^ !0 ■ - S 251 C9 to CO 1 g ^ eoecoococoeo«D cocoeoeocooeocoeo CO — 00 '- CD c^ iCOO bC '53 03 '^ ^* !S "^ 'r^ K rS. -M eo a c 3 •^ s~ S ■ *0 rjo "S C '^ ^ a. « ft S .to s ^ tj eo '^ o" « '^ "* -S e e s ^ O -^ ^ (^ OxS-O"^ ^ CJ u 3 S e 3 2 £ 5 3 ft J3 eo eo eo -^"3 3 eg «3"5,eu 3 3 O 3 O CO ^o-o* Flora and Ecology of Chautauqua Hills 121 65 o H iCiOOO ooooo oooo ooooooooo i-H 060 --HCCC^OCO ^fCOO M O •-I 1-H O Tf O 05 CO C^ t^ (NO (Nl>. --I C ■g « « $^ S K ! § S fc :^ ; fa. S « V. CIS S -o S eg o3 > 5 § e >»i -^ -Q C» S > 65 00 CO CO -as <*-i "^-a W. o o O H 'S >~ I^ Q S » 3 3 S 0) a s - S-B ^ s S CO -rJi 5 C* « S f:=2 •• « S ■^ M CO S *■ " fc. CO CO ■^ ~ 3 2 t~ *. $f <» • ^5 .. fc: e X p g aj 4j -^ ~ ~ fa. c S ■§ ^ s ^- K CO » K ^ CO CO eo cc -^ = 55^ o u o h o CO •- "^ ~ w „ « 3 o CCS V 0»000»OOOiOtCOO'00»00»CO'C»CiO»OW3iCO •rf< 00 --1 <-> c<3 -H Tt* c*j >o CO »-< o ■* CO Tt< Tpcc CO oec -§ 8 o u o o 13 c ca 'a c C > u >, M o a (U V lU (90 8 hi •J CO a 6 a c o -a m cc c3 J2 en 01 XI a C 03 CO I I O I GO I CO lO iC o CO lO lO »o OiO lo lo lO ic»o iC 'O «o lO lO iC »o o IC o iC I o oo Tt< CO OO'COOiOOOOiCOiOiCiOOiO'COiCiCiOO -- CO ^ Tf< i-H 00 CO ^ (N 00 CO CO i-i(N 05 ■* CO CC o 02 s c o > to . e ~ s o » s e «J o « o S.2 ^-g. 00 s s t-r o3 > a u i § ff a x; to w 03 V> 0) to C 'M 00 a a : s •" a. 5 a « to 00 to CO 3 S 3 S V) U CJ CJ 1^ k. I» ^ ' b> P. H. Humfeld 357, 808; R. L. McGregor 3379. Polystichum acrostichoides (Michx.) Schott.: In slightly eroded gulches of wooded hillsides. All counties except Elk. Thelypteris palustris Schott., var. pubescens (Lawson) Fern.: In sand)- soil of wooded hillsides. All counties except Greenwood. Woodsia obtusa (Spreng. ) Torr.: In moist upland woods. Frequent. CUPRESSACEAE Juniperus virginiana L.: Hilltops and slopes of rocky open wocjds. Typhaceae Typha latifolia L.: Growing in dense stands at edges of lakes and marshes. Woodson, Greenwood, and Montgomery counties. Flora and Ecology of Chautauqua Hills 143 ZOSTERACEAE Potamogeton crispus L. : In shallow water of Lake Fegan. Woodson County. Potamogeton diversij alius Raf.: In ponds and streams. Collections by P. A. Rydberg and R. Imler 414, Montgomery County; R. L. McGregor 3203, Woodson County; R. H. Thompson, June 8, 1949, Wilson County. Potomogeton foliosus Raf.: In shallow streams and lakes. Greenwood and Woodson counties. Potomogeton gramineus L.: In marshes and streams. Woodson Coimty. Potamogeton nodosus Poir.: Along the shore of Lake Fegan and in streams of wooded valleys. Elk and Woodson Counties. Najadaceae Najas guadalupensis (Spreng. ) Magnus: In water along the shore of Lake Fegan. Woodson County. Alismaceae Alisma subcordatum Raf.: In marshes and in shallow standing water of waste places. Woodson County. Echinodorus cordif alius (L. ) Griseb.: Reported for Wilson County, Gates (1940). Lophotacarpus calycinus (Engelm. ) J. C. Smith: Reported for Woodson County, Gates (1940). Sagittaria ambigua J. H. Smith: In a shallow water hole of a prairie pasture. Collection by P. A. Rydberg and R. Imler 402, Montgomery County. Sagittaria latifalia Willd., var. latifolia: Borders of lakes and marshes, and in wet roadside ditches. Chautauqua, Montgomery, and Woodson Counties. Sagittaria latifalia Willd., var. obtusa (Muhl. ) Wieg.: Reported for Chautauqua County, Gates (1940). Hydrochartaceae Elodea canadensis Michx.: In marshes and in water holes in prairie pastures and woodlands. Woodson County. Gramfneae Aegilops cylindrica Host.: In a roadside ditch bordering a prairie pasture. Found once, in Woodson County. Agropyran smithii Rydb.: In overgrazed pastures and waste places. Elk, Greenwood, and Woodson counties. Agrostis alba L.: In waste fields and disturbed areas of prairie pastures. Woodson County. Agrostis elliottiana Schultes: Reported for Chautauqua County, Gates (1940). Agrostis hiemalis (Walt.) BSP.: In sandy soil of open wooded hillsides. Woodson County. Agrostis scabra Willd.: In sandy soil of prairie hay fields and open upland woods. Woodson County. Alopecurus carolinianus Walt.: Reported for Chautauqua County, Gates (1940). Andropogon gerardi Vitman: Lower slopes of prairie hay fields and pastures, and prairie openings in woods. Frequent. Andropogon saccharoides Sw.: In prairie hay fields, pastures, and prairie wood- lands. 144 The University Science Bulletin Andropogon scoparius Michx., var. scoparius Prairie hay fields, pastures, and prairie woodlands. Frequent. Andropogon scoparius Michx., var. neo-mexicanus (Nash) Hitchc.: In a rocky prairie pasture. Collection by R. L. McGregor 9807, Greenwood County. Andropogon tcrnarius Michx.: Reported for Wilson County, Gates (1940). Andropogon virginicus L.: Rocky scrub oak wooded hillside. Collection by R. L. McGregor 4181, Chautauqua County. Aristida basiramea Engelm.: Overgrazed prairie pastures and waste places. Collection by B. L. Wagenknecht 1724, Montgomery County. Aristida dichotoma Michx.: Upland prairies and on sandy soil of open scrub oak woods. Montgomery, Wilson, and Woodson counties. Aristida intermedia Scribn. and Ball: Sandy soil of open rocky upland woods. Wilson and Woodson counties. Aristida longespica Poir.: In calcareous soil of a prairie pasture and in sandy soil of open scrub oak woods. Rare. Montgomery and Woodson counties. Aristida oligantha Michx.: In overgrazed prairie pastures and open woods. Aristida purpurascens Poir.: In sandy soil of open wooded slopes. Chautauqua, Montgomery, and Woodson counties. Bouteloua gracilis (KBK.) Lag.: Reported for Greenwood and Woodson counties. Gates ( 1940 ) . Bouteloua hirsuta Lag.: In overgrazed prairie pastures and open grazed woods. Chautauqua, Montgomery, and Woodson counties. Bromtts commutatus Schrad.: Eroded areas in sandy prairie pastures. Col- lections by R. L. McGregor 10319, Greenwood County; 10190, Montgomery County. Bromus inermis Leyss.: Fields, waste places, and overgrazed pastures. Wood- son County. Bromus japonicus Thunb.: Overgrazed prairie pastures and fields. Bromus purgans L.: Prairie pasture gullies and open woods. Greenwood and Woodson counties. Bromus racemosus L.: Fields and waste places. Collections by W. H. Horr, July 11, 1930, Woodson County; July 12, 1930, Wilson County. Bromus secalinus L.: Overgrazed prairie pastures and open sandy woods. Bromus tectorus L.: Low prairie hay fields and pastures. Montgomery and Woodson counties. Buchloe dacttjloides (Nutt. ) Engelm.: Dry overgrazed prairie pastures. Greenwood and Woodson counties. Cenchrus longispinus (Hack.) Fern.: Roadsides, waste places, and disturbed pastures. All counties except Elk and Wilson. Chloris verticillata Nutt.: Waste places and overgrazed prairie pastures. Cinna arundinacea L.: In sandy soil of an open wooded gully. Woodson County. Danthonia spicata (L.) Beauv., var. longipila Scribn. and Merr. : Sandy soil of an open scrub oak woods. Rare. Collected at Woodson County Stale Park, Woodson County. Digitaria sanguinalis ( L. ) Scop. : Waste places, fields, and overgrazed pastures. DipJachne acuminata Nash.: Reported for Greenwood County, Gates (1940). Echinochloa crusgalli (L. ) Beauv.: On banks of drainage ditches in fields and roadsides. Echinochloa pungens (Poir.) Rydb., var. pungens: Roadside waste places and fields. Woodson County. Flora and Ecology of Chautauqua Hills 145 Echinochloa pungens (Poir. ) Rydb., var. microstachya (Wieg. ) Fern, and Grisc: Moist waste places and on banks of streams in open woods. Mont- gomery and Woodson counties. Eleusine indica ( L. ) Gaertn. : Reported of all counties except Greenwood and Wilson, Gates (1940). Elymus canadensis L., var. canadensis: Prairie hay fields and pastures. Chau- tauqua, Elk, and Woodson counties. Elymus canadensis L., var. canadensis, fonna glaucif alius (Muhl. ) Fern.: Prairie hay fields and pastures. Chautauqua and Woodson counties. Elymus villosus Muhl.: On creek banks and in low woods. Collection by R. L. McGregor 10819, Montgomery County. Elymus villosus Muhl., forma arkansanus (Scribn. and Ball) Fern. Reported for Greenwood County, Gates (1940). Elymus virginicus L., var. virginicus: In prairie pastures and borders of woods. Elymus virginicus L., var. virginicus, forma hirsutiglumis (Scribn.) Fern.: In sandy soil of an upland woods. Collected once, in Woodson County. Elymus virginicus L., var. glabrifolius (Vasey) Bush: In upland woods bordered by prairie. Wilson and Woodson Counties. Elymus virginicus L., var. suhmuticus Hook: Reported for Chautauqua and Greenwood counties, Gates ( 1940). Eragrostis capillaris (L. ) Nees: In sandy soil of open grazed upland woods and prairie pastures. Eragrostis intern^edia Hitch. : In prairie hay fields and pastures. Eragrostis megastachya (Koel. ) Link.: Overgrazed pastures and waste places. Eragrostis pectinacea (Michx. ) Scribn.: Rocky prairie pastures and waste places. All counties except Elk and Greenwood. Eragrostis poaeoides Beauv. ex-Roem. and Schult.: Sandy soil of a dry lake bed. Reported for Woodson County, McGregor and Voile ( 1950 ) . Eragrostis reptans (Michx.) Nees: Reported for Chautauqua County, Gates (1940). Eragrostis spectabilis (Pursh) Steud.: Fields and roadsides. Collection by W. H. Horr, July 10, 1930, Woodson County. Eriochloa contractu Hitchc: Reported for all counties except Wilson, Gates (1940). Festuca elatior L.: Low prairie pastures and waste places. Collection by W. H. Horr 207-1, Wilson County. Festuca octoflora Walt.: In moist grassy meadows. Woodson County. Festuca ovina L.: In sandy soil of open^ scrub oak woods. Chautauqua and Woodson counties. Festuca paradoxa Desv. : Roadsides and railroad embankments. Woodson County. Glyceria striata (Lam.) Hitchc: In moist shaded woods. Woodson County. Gymnopogon ambiguus (Michx.) BSP.: In sandy soil of open upland woods. Found twice, in Chautauqua and Montgomery counties. Hordeum pusillum Nutt.: Rocky prairie pastures, wooded gullies, and clear- ings in woods. Greenwood, Montgomery, and Woodson counties. Koeleria cristata ( L. ) Pers. : In prairie and open ground. Collections by W. H. Horr, July 5, 1930, Montgomery County; R. L. McGregor 316-3, Woodson County; 10273, Wilson County. Leersia oryzoides (L. ) Sw.: Sandy soil of open scrub oak woods, Mont- gomery and Woodson counties. 146 The University Science Bulletin Leersia virginica Willd.: Moist woods. Collection by B. L. Wagenkneclit 1772, Chautauqua County. Leptochloa filiformis (Lam.) Beauv.: On an alluvial deposit near Fall River. Collection by R. L. McGregor 5736, Wilson County. Leptoloma cognatum (Schultes) Chase: Sandy prairie pastures. Woodson County. Lolium perenne L.: Reported for Montgomery County, Gates (1940). Melica nitens (Scrlbn.) Nutt.: In open flood plain woods of Fall River. Col- lected once, in Greenwood County. Melica porteri Scribn.: Reported for Montgomery County, Gates (1940). Muhlenbergia brachyphyUa Bush: In creek valley woods. Collection by B. L. Wagenknecht 1695, Montgomery County. Muhlenbergia capillaris (Lam.) Trin.: In sandy soil of open wooded hillsides. Montgomery and Wilson counties. Muhlenbergia cuspidata (Torr. ) Nash: In sandy soil of a rocky upland wooded hillside. Collection by R. L. McGregor 9800, Greenwood County. Muhlenbergia frondosa ( Poir. ) Fern. : Creek banks and wooded gullies. Chau- tauqua County. Muhlenbergia mexicana (L. ) Trin.: Reported for Elk County, Gates (1940). Muhlenbergia racemosa (Michx. ) BSP.: Creek valley woods and upland wooded gullies. Chautauqua, Wilson, and Woodson counties. Muhlenbergia schreberi J. F. Gmel.: Wooded prairie gullies and on a rocky wooded cliff. Woodson County. Muhlenbergia sobolifera (Muhl.) Trin.: In sandy soil of an open wooded gully in upland. Found once, in Woodson County. Panicum agrostoides Spreng.: On sandy open wooded slopes. Montgomery County. Panicum boscii Poir.: In shallow soil of a wooded limestone escarpment. Col- lection by R. L. McGregor 3398, Montgomery County. Panicum capillare L.: Eroded pastures and waste places. Montgomery and Woodson counties. Panicum clandestinum L.: In sandy soil of a shaded bank in an open oak woods. Collection by R. L. McGregor 10850, Montgomery County. Panicum dichotomiflorum Michx.: In prairie hay fields and good grazed pas- tures. Panicum lanuginosum Ell., var. fasciculatum (Torr.) Fern.: Moist low woods and wooded upland gulhes. Chautauqua County. Panicum lanuginosum Ell., var. lindheimeri (Nash) Fern.: In shallow soil over sandstone outcrops of an oak wooded cliff^. Collected once, in Woodson County. Panicum leibergii (Vasey) Scribn.: Reported for Montgomery County, Gates (1940). Panicum oligosanthes Schultes, var. helleri (Nash.) Fern.: Low prairie pastures and at borders of woods. Montgomery and Woodson counties. Panicum oligosanthes Schultes, var. scribnerianum (Nash) Fern.: Reported for all counties except Elk and Woodson, Gates ( 1940). Panicum praecocius Hitchc. and Chase: Reported for Greenwood County, Gates (1940). Panicum spaerocarjwn Ell.: In sandy soil of open upland woods. Mont- gomery, Wilson, and Woodson counties. Flora and Ecology of Chautauqua Hills 147 Panicum virgatum L.: Prairie hay fields, pastures, and in open flats sur- rounded by woods. Frequent. Paspalum ciliatifolium Mich.x., var. ciliatifolium: Prairie hay fields, pastures, and open woods. Montgomery and Woodson counties. Paspalum ciliatifolium Michx., var. muhlenbergii (Nash) Fern.: Rocky prairie pastures and rocky flats surrounded by woods. Paspalum jiuitans (Ell.) Kunth.: On mud banks and extending out on surface of water in streams. Collections by W. H. Horr E173, Wilson County; R. L. McGregor 5731, Wilson County. Paspalum pubiflorum Rupr., var. glahrum Vasey: Low prairie hay fields and pastures. Wilson and Woodson counties. Phlaris canariensis L.: Low rocky prairie pastures. Collections by R. L. Mc- Gregor 10200, Montgomery County; 10247, Wilson County. Phlaris caroliniana Walt.: Low prairie hay fields and pastures. Montgomery and Woodson counties. Phleum pratense L.: Reported for Woodson and Montgomery counties, Gates (1940). Poa annua L.: Reported for Chautauqua and Montgomery counties, Gates (1940). Poa compressa L.: In a moist gulch of a prairie pasture. Found once, in Woodson County. Poa pratensis L.: On banks of gullies in prairie pastures. Woodson County. Poa sylvestris A. Gray: Reported for Montgomery County, Gates ( 1940.) Schedonnardus paniciila us (Nutt. ) Trel.: Prairie pastures, roadsides, and fields. Greenwood, Montgomery and Woodson counties. Setaria faherii Herrm.: Roadsides and fields. Setaria geniculata ( Lam. ) Beauv. : Roadsides, fields, and waste places. Setaria glauca (L. ) Beauv.: Overgrazed pastures, fields, and roadsides. Fre- quent. Setaria viridis ( L. ) Beauv. : Overgrazed pastures and waste places. Sorghastrum nutans (L. ) Nash.: Low prairie hay fields and pastures. Sorghum halepense ( L. ) Pers. : Fields and waste places. Spartina pectinata Link: Low prairie hay fields and pastures. SphenophoHs obtusata (Michx.) Scrib.: Prairie pastures, and prairie wood- lands. Sporobolus asper (Michx.) Kimth.: Rocky prairie hillsides and open wood- lands. Sporobolus prtjamidatus (Lam.) Hitchc. : In sandy soil of an overgrazed prairie pasture. Found once, two miles northeast of Sedan, Chautauqua County. Sporobolus vaginiflorus (Torr. ) Wood.: Reported for Wilson County, Gates (1940). Stipa spartea Trin.: On lower slopes of grazed and ungrazed prairies. All counties except Chautauqua. Triodia elongata (Buckl.) Scribn.: In calcareous soil of a prairie pasture. Found once, seven miles northeast of Elk City, Montgomery County. Triodia fiava (L. ) Smyth: Roadsides, waste places, and overgrazed pastures. Frequent. Triodia jlava (L. ) Smyth, forma cuprea (Jacq. ) Fosberg: In sandy soil along the edge of an open oak woods. Found once, in Woodson County. 148 The University Science Bltlletin Triodia stricta (Nutt.) Benth.: Reported for Montgomery County, Gates (1940). Tripsacum dacttjloidcs L.: Low grazed and ungrazed prairies, and roadsides. Uniola latifolia Michx. : Flood plain woods and wooded gullies in upland. Cyperaceae Bulhostylis capillaris (L.) C. B. Clark, var. cicbra Fern.: Dry exposed lake bed and dry sandy soil of a prairie pasture. Collections by R. L. McGregor 3179, Woodson County; B. L. Wagenknecht 1714, Montgomery County. Carex amphibola Steud., var. turgido Fern.: Reported for Montgomery County, Gates (1940). Carcx artitccta Mack.: Reported for Montgomery County, Gates (1940). Carex bicknellii Britton: Prairie pastures and open woodlands, Woodson County. Carex blanda Dewey: Wooded prairie gullies and flood plain woods. Wood- son County. Carex brevior (Dewey) Mackenz.: In low prairies and woods. Woodson County. Carex busJiii Mackenzie: In prairies and mud banks of water holes. Wood«on County. Carex dnvisii Schwein. and Torr.: On creek banks and in low woodlands. Woodson County. Carex frankii Kunth: On a creek bank and along the shore of a marsh. Col- lections by W. H. Horr, July 5, 1930, Montgomer>- County; July 12, 1930, Wilson County. Carex gravida Bailey, var. gravida: In low prairies and woods, and on banks of streams. Carex gravida Bailey, var. luneUiana (Mack.) F. J. Herm.: Low prairies, woodlands, and on banks of streams. Woodson County. Carex lanuginosa Michx.: Reported for Montgomery County, Gates (1940). Carex leavenworthii Dewey: Reported for Chautauqua County, Gates (1940). Carex hipulina Muhl.: Open woodlands and in gullies of prairie pastures. Woodson County. Carex meadii Dewey: Low woodland prairies. Wilson and Woodson counties. Carex microdonta Torr. and Hook.: Reported for Chautauqua and Montgomery counties, Gates (1940). Carex vidpinoides Michx.: Reported for Greenwood County, Gates (1940). Cyperus acuminatus Torr. and Hook. : On banks of ponds and marshes. Mont- gomery and Woodson counties. Cyperus diandrus Torr.: In sandy soil of an exposed lake bed. Reported for Woodson County, McGregor and Voile ( 1950). Cyperus erythrorhizos Muhl.: In sandy soil of an exposed lake bed. Collection by R. L. McGregor 4108, Woodson County. Cyperus esculentus L.: Moist prairie guHies. Woodson County. Cyperus ferruginescens Boeckl.: Reported for Elk, Montgomery, and Wood- son counties. Gates ( 1940 ) . Cyperus filiculmis Vahl.: Prairie pasture gullies and open woodlands. Cyperus inflexus Muhl.: Prairie gullies, open woodlands, and waste places. Woodson County. Flora and Ecology of Chautauqua Hills 149 Cyperus glohulosiis Aubl., var. robustus (Boeckl. ) Shinners: Open wooded hillsides and along banks of water holes. Cyperus setigerus Torr. and Hook.: Bank of a water hole in a prairie pasture. Collection by P. A. Rydberg and R. Imler 368, Montgomery County. Cyperus strigosus L.: On bank of a creek. Collection by W. H. Horr, July 5, 1930, Montgomery County. Eleocharis acicularis R. and S.: Open woods and low waste places. Woodson County. Eleocharis calva Torr.: Marshes and prairie pasture gullies. Greenwood and Woodson counties. Eleocharis compressa SuUiv.: Reported for Geenwood County, Gates (1940). Eleocharis engelmanni Stued.: Low prairie pastures and wooded gullies. Woodson County. Eleocharis macrostachya Britton: Along the edges of pasture ponds. Eleocharis ohtusa (Willd. ) Schultes: Border of ponds and marshes, and moist prairie gullies. Fimbristylis autumnalis (L. ) R. and S., var. mucronulata (Michx. ) Fern.: In moist sand of an exposed lake bed. Collection by R. L. McGregor 4094, Woodson County. Fimbrisylis baldwiniana (Schultes) Torr.: Muddy edge of a slough in a prairie pastvu-e, and in moist sand of a slight depression on a wooded hill- side. Collections by R. L. McGregor 4403, Chautauqua County; B. L. Wagenknecht 2265, Wilson County. Fimbristylis caroliniana (Lam.) Fern.: In a low moist prairie pasture. Col- lection by R. L. McGregor 4316, Montgomery County. Fimbristylis interior Britton: Sandy soil of a bluestem prairie. Collection by R. L. McGregor 10211, Montgomery County. Scirpus atrovirens Muhl.: Banks of prairie ponds and in wet roadside ditches. Woodson County. Scirpus fluviatilis (Torr.) Gray: Reported for Wilson County, Gates (1940). Scirpus lineatus Michx.: In moist prairie pasture gulches. Collections by R. L. McGregor 10325, Greenwood County; 4318, Montgomery County; 10246, Wilson County. Scirpus validus Vahl., var. creber Fern.: Reported for Wilson County, Gates (1940). Scleria paucijiora Muhl.: In sandy soil of an open oak woods. Collection by R. L. McGregor 3186, Woodson County. Scleria triglomerata Michx.: Prairie hay fields and roadsides. Collections by R. L. McGregor 4312, Montgomery County; W. H. Horr, July 5, 1930, Mont- gomery County. Araceae Arisaema dracontium (L. ) Schott. Lemnaceae Lemna minor L.: Stagnant pond. Collection by R. L. McGregor 2410, Wil- son County. Lemna perpusillaToir. : Woodland marsh. Collection by W. H. Horr, JiJy 11, 1930. Woodson County. Spirodela polyrhiza (L. ) Sclileid.: In marshes and prairie water holes. Wood- son County. 150 The University Science Bulletin COMMELINACEAE CommeUna comi7iunis L.: Borders of creek woods, banks of ditches and fields. Commelina virginica L.: Low woods, thickets, and roadsides. Collections by W. H. Horr, July 5, 1930, Chautauqua County; June 28, 1930, Greenwood County; July 12, 1930, Wilson County. Tradescantia hrac'.eata Small: Gullies in prairie pastures, creek banks, and borders of low woods. Tradescantia ohiensis Raf.: Woodland borders, ditch banks, and fields. JUNCACEAE ]uncus biflorus Ell.: Open woods and moist prairie gullies. Woodson County. Junciis diffiisissimus Buck!.: Open woodland and in sandy soil of an exposed lake bed. Collection by P. A. Rydberg and R. Imler 450, Montgomery County. Reported for Woodson County, McGregor and Voile (1950). Juncus dudleyi Wieg.: Low moist prairie pasture. Collection by P. A. Ryd- berg and R. Imler 450a, Montgomery County. Juncus interior Wieg.: Low prairies and open woodlands. Frequent. Juncus kansanus F. J. Herm.: Reported for Chautauqua County, Gates (1940). Juncus marginatus Rostk., var. marginatus: Along the borders of ponds and marshes, and on muddy banks of prairie gullies. Elk, Montgomery, and Woodson counties. Juncus marginatus Rostk., var. setosus Coville: In sandy soil of open woods. Collected twice in Woodson County. Juncus nodatus Coville: Muddy bank of a run-off stream in a prairie pasture. Collection by W. H. Horr, July 5, 1930, Chautauqua County. Juncus tenuis Willd., var. multicornis E. Mey.: Low moist woods, marshes, and along banks of standing water in prairie gullies. Woodson County. Juncus torretji Coville: In alluvial soil of flood plain woods, prairie gullies, and along borders of fields. Ldliaceae Allium canadense L., var. canadense: Prairies, roadsides, and e.vposed slopes. Allium canadense L., var. fraseri Ownby: Prairie pastures, roadsides, and upper gully slopes. Collections by R. L. McGregor 10310, Greenwood County, 10199, Montgomery County; 10336, Elk County; W. H. Horr, July 5, 1930, Chautauqua County. Allium canadense L., var. lavendulare (Bates) Ownby and H. C. Aase: Prairie pastures, roadsides, and open wooded slopes. Allium stellatum Ker-Gawler: Sandy soil in scrub oak woods, open flood plain woods, roadsides, and prairie pastures. Asparagus officinalis L.: Reported for Elk, Montgomery, and Woodson counties. Gates (1940). Camassia scillioides (Raf.) Cory: Prairie hay fields, good grazed pastures, and open wooded slopes. Elk and Woodson counties. Erijthronium albidum Nutt.: Reported for Wilson County, Gates (1940). Erythronium mesochoreuni Knerr.: Low prairie woodlands and open scrub oak woods. Elk, Greenwood, and Woodson counties. Hemerocallis fulva L.: Border of low woods. Found once, in Woodson County. Nothoscordium bivalve (L. ) Britton: In sandy soil of open oak woods and rocky prairie slopes. Chautauqua and Woodson coimties. Flora and Ecology of Chautauqua Hills 151 Pdygonaium canaUculatum (Muhl.) Pursh.: Along creek banks and in flood plain woods. Smilax bona-nox L.: Thicket of flood plain woods. Collection by B. L. Wagen- knecht 2228, Wilson County. Smilax herbacca L.: Flood plain woods and in a prairie thicket of a prairie gulch. Elk and Woodson counties. Smi/flx tamnoides L., var. hispida (Muhl.) Fern.: Creek banks and woodland borders. All counties except Elk and Woodson. Yucca glauca Nutt., var. mollis Engelm.: Calcareous soil of prairie hillsides and hilltops. Collections by R. L. McGregor 4347, Montgomery County; 4309, Wilson County. Zigadeniis nu tallii Gray: Wooded guUies and prairie woodlands. Amaryllidaceae Cooperia drummondii Herb.: Rocky prairie hillsides. Collections by W. H. Horr E232, Wilson County; W. C. Stevens 8177-1, Wilson County. Hypoxis hirsuta (L.) CoviUe: Sandy soil of open scrub oak woods, and in prairie pastures. Iridaceae Belamcanda chinensis (L.) DC: Escaped from cultivation and established in prairies and along roadsides. Collection by W. H. Horr, July 11, 1930, Woodson County. Nemasttjlis gem,iniflora Nutt.: In prairie woodlands and pastures. Elk and Chautauqua counties. Sisyrinchium campestre Bickn.: Prairie hay fields and pastures, along borders of creek woods, and in sandy soil of upland woods. Chautauqua and Woodson counties. Orchidaceae Corallorhiza wisteriana Conrad.: On a wooded limestone escarpment. Found once by R. L. McGregor 10227, five miles northeast of Elk City, Mont- gomery County. Hexalectris spicata (Walt.) Barhart: Growing through rich leaf mold of a wooded limestone escarpment. Collected once by R. L. McGregor 3393, five miles northeast of Elk City, Montgomery County. Spiranthes cernua (L. ) Richard: In sandy soil of open scrub oak woods. All counties except Elk and Wilson. Spiranthes gracilis (Bigel) Beck: In low, grazed and ungrazed prairies. Col- lected once, in Woodson County. Spiranthes vernalis Engelm. and Gray: Reported for Elk, Montgomery, and Wilson counties. Gates (1940). Salicaceae Populus alba L.: Introduced and naturalized in wasteland woods and stream valleys. Montgomery and Wilson counties. Populus deltoides Marsh: Gullies, streambanks, wasteland woods, and open flood plain woods. Salix amygdaloides Anders.: Creek banks, wood gullies, and lowland woods. Chautauqua, Greenwood, and Woodson counties. Salix caroliniana Michx.: Wooded gullies and stream banks. Chautauqua and Greenwood counties. 152 The University Science Bulletin Salix eriocephala Michx.: Rocky prairie gullies. Rare. Woodson County. Salix humilis Marsh., var. microphylla (Anders.) Fern.: Border of open oak woods. Rare. Montgomery County. Salix humilis Marsh., var. rigidiuscula (Anders.) Rob.: Rocky open hillsides. Woodson County. Salix interior Rowlee: Sandy stream banks. Collections by W. H. Horr 3378, Wilson County; B. L. Wagenknecht 1737; 1834, Chautauqua County. Salix nigra Marsh., var. nigra: Wooded gullies and lowland woods. Salix nigra Marsh., var. lindheimeri Schneider: Wooded gullies and flood plain woods. Collections by W. H. Horr, July 28, 1930, Greenwood County; P. A. Rydberg and R. Imler, Montgomery County. JUGLANDACEAE Carija cordiformis (Wang.) K. Koch., var. cordiformis : Creek banks, flood plain woods, and upland slopes. Carya cordiformis (Wang.) K. Koch., var. latifolia Sarg.: Flood plain woods and stream banks. Collections by W. H. Horr, July 5, 1930, Chautauqua County; July 5, 1930, Montgomery County. Carya illinoensis (Wang.) K. Koch: Open flood plain woods. Wilson and Woodson counties. Carya laciniosa ( Michx. ) Loud. : Rich lowland woods. Collection by W. H. Horr, July 5, 1938, Chautauqua County. Carya ovata (Mill.) K. Koch: Wooded gullies and flood plain woods. Carya texana Buckl.: Reported for Montgomery and Wilson counties. Gates (1940). Carya tomentosa ( Poir. ) : Flood plain woods. Rare. Collection by W. H. Horr, July 5, 1930, Chautauqua County. Juglans nigra L.: Low woodlands and upland wooded slopes. Betulaceae Bettila nigra L.: Base of a wooded slope above a creek. Found once, in Chautauqua County. Corijlus americana Walt.: Reported for Wilson County, Gates (1940). Ostrya virginiana (Mill.) K. Koch: Flood plain woods and slopes of upland woods. Woodson County. Fagaceae Quercus alba L.: Upland scrub oak woods. Found in one location, three miles southwest of Elk City, Montgomery County. Quercus alba L. X Q- muhlenbergii Engelm.: On a wooded limestone escarp- ment. Rare. Collected five miles northeast of Elk City, Montgomery County. X Quercus bushii Sarg.: Along sandstone escarpments, and slopes of upland woods. X Quercus faxoni Trel.: Wooded limestone escarpments, slopes and borders of upland woods. Raie. Two locations in Montgomery County. X Quercus fernouA Trel.: Sandstone escarpments, slopes and borders of open oak woods. Rare. Montgomery and Wilson counties. Quercus imbricaria Michx.: Woods along a small creek. Found once by R. L. McGregor 4464, two miles north of Howard, Elk County. Flora and Ecology of Chautauqua Hills 153 Quercus macrocarpa Michx.: Flood plain woods, wooded upland gullies. Quercus mariliindica Muench.: Hilltops and slopes of upland woods. Frequent. Quercus muehlenbergii Engelm.: Upland gullies, rocky escarpments, and slopes of upland woods. Quercus muehlenbergii Engelm., forma alexanderi (Britt.) Trel.: Upland gullies, rocky escarpments, and slopes of upland woods. Quercus palustris Muench.: Slopes of upland woods and creek valleys. Quercus prinoides Willd.: Rocky escarpments, gullies, and slopes of upland woods. Quercus rubra L. : In gullies and lower slopes of upland woods. Quercus shumardii Buckl., var. shumardii: Slopes of upland woods. Rare. Collections from Greenwood and Montgomery counties. Quercus shumardii Buckl., var. schneckii (Britton) Sarg.: Rocky escarpments, and slopes of upland woods. Quercus stella'a Wang.: Hilltops and slopes of upland woods. Frequent. X Quercus stelloides E. J. Palmer: Upland wooded slopes, and limestone escarpments. Chautauqua, Montgomery, and Wilson covmties. Quercus velutina Lam. : Rocky escarpments, and slopes of upland woods. Ulmaceae Celtis laevigata Willd.: At the base of a scrub oak wooded hillside. Rare. Montgomery County. Celtis occidentalis L.: Flood plain woods, and slopes and hilltops of upland woods. Celis reticulata Torr.: Reported for Wilson County, Gates (1940). Celtis tenuifolia Nutt., var. georgiana (Small) Fern, and Schub.: Flood plain woods, rocky open slopes of uplands. Ulmus americana L.: Flood plain woods, and slopes of uplands. Ulmus rubra Muhl.: Low woodlands, wooded prairie gullies and upland slopes. MORACEAE Madura pomifera (Raf. ) Schneider: Wasteland woods and disturbed areas. Morus alba L. : Flood plains and stream banks. Morus rubra L.: Open flood plain woods and lower upland slopes. C.\NNABINACEAE Cannabis sativa L.: Reported for Chautauqua, Greenwood, and Montgomery counties. Gates (1940). Humulus lupulus L.: Creek banks and flood plain woods. Collections by O. S. Fearing and G. Latham, August 10, 1950, Chautauqua County; W. H. Horr, July 5, 1930, Chautauqua and Montgomery counties; June 28, 1930, Greenwood County. Urticaceae Boehmerirairie woodlands. Baptisia leucophaea Nutt.: Grazed and ungrazed prairies, and prai; ie wood- lands. Baptisia minor Lehm.: Prairie hay fields, pastures, and prairie woodlands. Cassia fascicidata Michx.: Prairie hay fields, pastures, roadsides, and open wooded slopes. Cassia marihindica L.: Rocky prairie gully. Collection by R. L. McGregor 10864, Wilson County. Cassia nictitans L.: In undergrowth of open upland wooded hillside. Col- lection by R. L. McGregor 4087, Woodson County. Cercis canadensis L.: Creek valleys and open wooded slopes. Cercis canadensis L., forma glahrifoha Fern.: Thicket along a creek hank. Rare. Collection by R. L. McGregor 4461, Woodson County. Crotalaria sagittaUs L.: Creek banks, and open woodlands. Montgomery and Woodson counties. Dalea Candida Willd.: Low prairie hay fields and pastures. Dalea laxiflora Pursh: Prairie hay fields and pastures. Collection b\ O. S. Fearing and G. Latham, August 14, 1950, Wilson County. Dalea midtiflora (Nutt.) Shinners: Prairie woodlands. Chautauqua and Woodson comities. Dalea jnirpurea Vent: Grazed and ungrazed prairies, and prairie woodlands. Desman'hus iUinoensis (Michx.) MacM.: Prairie thickets, roadsides, and open woods. Desmodium canadense (L. ) DC: Reported for Chautauqua and .Montgomery counties. Gates ( 1940 ) . Desmodium canescens (L. ) DC: Lowland wooded slope. Collection by B. L. Wagenknecht 1819, Chautauqua County. Desmodium cuspidatum ( Muhl. ex Willd.) Loud., var. hmgifoUum (T. and G. ) Schub.: Reported for Chautauqua County, Gates (1940). 6—2656 162 Thk University Science Bulletin Desmodium f^lutinosuni (Muhl. ex Willd.) Wood: Gully banks, and wooded hillsides. Montgomery and Woodson counties. Dcmioditim iUiiioen.sis Gray: Low woodlands and upland wooded hillsides. Dcs-Jtioclitun nuirilandinim ( L. ) DC.: ()i>en scrub oak woods. Woodson Coinit\'. Desmodium midiftorum ( L. ) DC.: Rocky open woods. Collection by B. L. Wagenknecht 1685, Montgomery Count\ . Desmodium paniculatum ( L. ) DC: Wooded gullies, and open oak woods. Woodson County. Desmodium puucifiorum (Nutt.) DC: Dense upland woods. Collection by B. L. Wagenknecht 2245, Wilson County. Desmodium pcrplexum Scliub.: Reported for Montgomer\' and Woodson counties, Gates ( 1940. ) Desmodium sessilifolium ( Torr. ) T. and C: Lower slopes of upland woods. Montgomery and Woodson counties. Gledi sia triacantlws L.: Creek valley woodlands, and prairie thickets. Ghjcyrrliiza lepidota (Nutt.) Pursh: Low prairies, and waste i^laces. Collec- tion by W. H. Horr, July 5, 1930, Montgomery County. Gymnocladus dioica ( L. ) Koch: Flood plain woods, and slopes of upland woods. Lespedeza capitata Michx.: Low prairie hay field. Collection by W. II. Ilorr, July 11, 1930, Woodson County. Lespedeza repens (L.) Bart.: Reported for Elk and Woodson counties, Gates (1940). Lespedeza striata (Thunb. ) H. and A.: Dr\ exposed lake bed. Collection by R. L. McGregor 4106, Woodson County. Lespedeza stuvei Nutt. : Slopes of open upland woods. Woodson County. Lespedeza violacea (L. ) Pers.: Prairie hay fields, and rocky prairie hillsides. Lespedeza virginica (L.) Britt.: Open wooded hillsides. Medicago lupulimi L.: Roadsides, fields, and waste places. Medicago sativa L.: Roadsides, fields, and waste places. Melilotus alba Desf. : Low prairies, roadsides, and waste places. Melilotus officimdis (L. ) Lam.: Low prairies, roadsides, and waste places. Psoralea argoplujUa Pursh: Slopes and crests of rocky prairie pastures. Col- lections by W. H. Horr, Jul>- 28, 1930, Greenwood County; R. L. McGregor 10347, Elk County. Psoralea esculetita Pursh: Prairie pastures. Woodson Comity. Psoralea tenuiflora Pursh: Prairie hay fields, pastures, and prairie woodlands. Rol)inia pseudo-acacia L.: Open flood plain woods, pasture gullies, and prairie thickets. Schrankia uncinata Willd.: Prairie thickets, roadsides, and rocky pasture hill- sides. Sesbania exalata ( Raf . ) Cory: In al!u\ial soil of an ox bow of Fall River. Col- lections by R. L. McGregor 5728, .Montgomery County; 5729, Wilson County. ^ophora sericea Nutt.: Reported for Woodson County, Gates (1940). ^tropltostyles helvola (L. ) Britton: Sandy soil of open upland woods. Mont- gomery and Woodson counties. Strophostyles leiosperma (T. and G. ) Piper: Sand\' soil of open scrub oak woods. Woodson County. Flora and Ecology of Chautauqua Hills 163 Styloscnithcs bifiora ( L. ) BSP., \;ir. hiflora: Prairie hay fields, pastures, and prairie woodlands. Stylosanthes hiflora ( L. ) BSP., \ ar. hispidiatiinui (Miehx. ) Pollard and Ball: Prairie hay field and pastures. Woodson County. Tephrosia leucosericca ( Rydb. ) Cor\ : Reported for Elk and Chautauqua eounties, Gates (1940). Tephrosia virginianti (L.) Pers.: Prairie lta\ fields, pastures, and prairie woodlands. Trifolium campcstre Sehreb.: Roeky prairie pasture hillsides. Collections by H. L. McGregor 4320, .Montgomery Count) ; 4286, Wilson County. Trifolium pra'en.se L.: Fields, roadsides, and waste places. Trifolium repens L.: Fields, roadsides, and waste places. Virid viUosa Roth. LiNACEAE Linuin sulcatum Riddell: Lo\\- prairie ha>' fields and pastures. Linum usitatissimum L.: Reported for Greenwood County, Gates ( 1940). OXALTOACEAE Oxalis cornicuhita L.: Creek \alle\ woods. Collection by W. H. Horr, July 28, 1930, Greenwood County. Oxalis europaea Jord., \'&t. europaea: Prairies and rocky woodlands. Oxalis ruropca Jord., \ar. europaea, forma pilosella Wieg.: Rocky prairie hill- side ])()rdering an oak woods. One collection from Chautauqua County. Oxalis europaea Jord., ^•ar. hushii (Small) Wieg.: Prairies and open woods. Collections by W. H. Horr. June 28, 1930, Greenwood County; July 5, 1930, Montgomer>- Covmty; R. L. McGregor 3168, Woodson County. Oxalis stricta L.: Rocky pastures, roadsides, and open woodlands. Oxalis violacea L.: Prairie hay fields, pastures and roadsides. Oxalis violacea L., forma alhida Fassett. : In a low prairie hay field. Found once, in Chautauqua Co>mt>'. Gekaniaceae Ocrauium ciiroliiiiinniiii L.: Prairies, roadsides, and open wooded hillsides. Rutaceae I'tclca trifoliata L.: BorcUrs ol Hood plain woods, anil upland wooded hillsidis. Zanthoxt/lum americanum Mill.: Open woodlands, and in shallow soil of a wooded limestone escarpment. Elk, Chautaucpia, and Greenwood counties. Zvcornvi.LACEAE Trihulus terrestria I..: Iloadsides and waste places. SlMARUBACEAE Mlaiifhtis altissiiiia Swingle: Low open woodland. Collections b> \\'. H. Horr, JuK' .5, 1930, Chautan(iua and Montgomery counties. I'OI.YCALACE.AE Folygala incaruuta L.; Prairies and open woodlands. Woodson CJounty. Polygala sanguiuea L.: Prairies and open oak woods. Polygala sanguinca L., fonna albiflora (Wheeloek) Millsp.: Prairie pastures. Rare. Montgomer> and \\'oodson counties. 164 The University Science Bulletin TolijgaUi sanguined L., forma viridcscen.s (L. ) Farw.: Prairies and grassy hanks in optn crct'k woods. ColU'ctions by W. H. Horr, July 11, 1930, Woodson County; P. A. Rydberg 431, Montgomery County. Poltjgala verticillata L., var. isocycla Fern.: Prairie pasture. Collection by P. A. Rydberg and R. Imler 457, Montgomery Cotmty. EUPHOKBIACEAE Acahjpha gracilen.s Gray, var. gracilens: In sand>- soil ot open uplantl woods. Acahjphu gracilens Gray, var. fraseri ( Muell, Arg. ) Weath.: In .shallow soil over sandstone outcrops of a moist gully in a prairie woodland. Collection by R. L. McGregor 10804, Chautauciua County. Acahjpha gracilens Gray, \ar. monococcu Engelm.: In shallow soil under a lime- stone ledge of a wooded hillside, and in sandy soil of an open oak woods. Collections by H. L. McGregor 10823, Chautaucjua Count); 10842, Mont- gomery Comity. Acahjpha rhc>ni})oi(lca Raf.: Roadsides, and border ot a low woodland. Wood- son County. Acahjpha virginica L.: Roadsides, and open flood plain woods. Croton cajntalus Michx.: Roadsides, o\ergrazed pastures, and waste places. Croton glandiilosus L., var. septentrionalis Muell Arg.: Roadsides and waste places. Collections by W. II. Ilorr, July 5, 1930, Chautauciua and Mont- gomery counties. Croton monanthogijnus Michx.: Overgrazed pastures, roadsides, and open low- land woods. Croton texensis (Klotzch) Muell. arg.: Reported for Montgomery County, Gates (1940). Crotonopsis elliptica Willd.: In sandy soil of open upland woods. Collection by R. L. McGregor 10817, Chautauqua County. Ditaxis mercurialina ( Nutt. ) Coult. : Rocky prairie hay fields and pastures. Collections by R. L. McGregor 4342, Montgomery County; 4923, Wilson County. Euphorbia corollala L.: Low prairie hay fields, pastures, and prairie woodlands. Euphorbia dentata Michx.: Open woodlands and wa.ste places. EuphorJna dictyospenna Fisch. and Mey.: Roadsides, waste places, and open flood plain woods. Euphorbia heteropJujlla L. : Sandv soil ot open oak woods, borders of flootl plain woods, fields, and roadsides. Euphorbia hexagona Nutt.: Sandy soil of low open woods, and fields. Collection by B. L. Wagenknecht 1684, Montgomery County. Eupliorbia hinnistrata Engelm.: Rocky prairie pasture. Collection by B. L. Wagenknecht 2183, Wilson County. Euphorbia maculata L.: Overgrazed pastures, rocky open wooded gullies, road- sides, and waste places. Euphorbia margimita Pursh.: In prairie gullies, roadsides, and waste places. Chautaucjua and Woodson counties. Euphorbia missurica Raf., var. viissurica: In an overgrazed pasture, and in rocky waste places. Collections by W. 11. Horr, July 11, 1930, Woodson County; B. L. Wagenknecht 1679, Montgomery County. Euphorbia niissttrica Raf., var. intermedia (Engehn.) L. C. Wheeler: Shallow soil over limestone surfacing in prairie pasture. Collection by B. L. Wagen- knecht 1860, Chautauqua County. FiX)BA AND Ecology of Chautauqua Hills 165 Euphorbia obtusutu Piirsh.: Overgrazed pastures, rocky prairie gullies, and wooded hillsides. Euphorbia serpens HBK.: Roadsides, fields, and waste places. Chautauqua and Woodson counties. StilUngia sylvatica L.: Flood plain woods. Collections by P. A. Rydberg and R. Iniler 394, Montgomerv- County. Tragia urticifolia Michx.: Rocky pastures and moist ravine banks in upland woods. Collections by W. H. Horr, July 5, 1930, Chautauqua County; R. L. McGregor 10367, Elk County; 10829, Chautauqua County; B. L. Wagenknecht 1849, Chautauqua Count>'. Callitrichaceae Callitriche heterophylla Pursh.: In sandy soil of a dry exposed lake bed. Col- lection by R. L. McGregor 3179, Woodson County. Anacardiaceae Rhus uromatica Ait., var. aromatica: Rocky prairie pastures, and wooded upland slopes. Collection by B. L. Wagenknecht 1868, Chautauqua County. Rlius aromatica Ait., var. serotina (Greene) Rehd.: Rocky prairie gullies, and prairie woodlands. Rhus copaUina L.: Prairie thickets, low rocky woods, and open upland wooded slopes. Rhus glabra L.: Prairie thickets, low rocky woodlands, and open oak wooded hillsides. Rhtis radicans L., var. rrjdhergii (Small) Rehd.: Wooded upland gully, creek banks, and low woodlands. Woodson County. Rhm radicans L., var. vulgaris (Michx.) DC. fomia negundo (Greene) Fern.: Wooded gullies, and flood plain woods. Frequent. Aquifoliaceae Ilex dccidua \\a\\..: Reported for Montgomery County, Gates (1940). Celastraceae Celastrus scandens L. : Thickets, wooded cliff, and dense low woods. Euonymus atropurpurcus Jacq.: Wooded stream banks, creek valleys, and wooded hillsides. Elk and Woodson counties. Staphyleaceae Siaphylea trifolia L.: Creek banks, flood plain woods, and upland wooded slopes. Chautauqua and Greenwood counties. ACERACEAE Acer negundo L.: Stream banks, flood plain woods, and prairie thickets. Acer saccharinum L.; Flood plain woods, and lower slopes of upland woods. HiPPOCASTANACEAE Aesculus glabra Willd., var. glabra: Creek bank thicket. Collection by W. H. Horr, July 11, 1930, Woodson County. Aesculus glabra Willd., var. sargentii Rehd.: Creek banks, flood plain woods, lower slopes of rocky wooded hillsides, and wooded gulbes. Chautauqua, Wilson and Woodson cotmties. 166 The University Science Bulletin Sapindaceae Saphidiis drummomli II. and A.: On a wooded limestone plateau, antl in a low- woodland. Chautauqua and Montgomery counties. Balsaminaceae Impatwns pallida Nutt.: Low dense woodlands. Woodson County. Rhamnaceae Ceanothu.s americana L., var. pitcheri T. and C: Banks of streams and lakes, low woods, and gullies in uplands. Ceanothu.s herhaceus Rat'., var. herbaccu.s: Bank ot moist prairie gully. Clia- tautjua County. CeanotJius hcrbaceus Raf., var. puhescens (T. and G.) Shinners: Wooded creek hank hordering upland woods. Collection l)y W. H. Horr, Jime 28, 1930, Greenwood County. Rhamnus lanceolata Pursh.: Reported for Wilson County, Gates (1940). Vitaceae Ampelopsis cordata Michx.: Growing over small trees and shrubs in a low woods, thickets in flood plain woods, and wooded gullies. All counties ex- cept Chautauqua and Elk. Parthenocisms inscrta (Kerner) K. Fritsch: Reported for Wilson County, Gates (1940). Farthenocissus quitiquefolia (L.) Planch.: Dense undergrowth of low woods, and slopes of upland woods. All counties except Elk and Wilson. Vitis cinera Engelm: Growing over and hanging from branches of trees and flood plain woods. Vitis vulpina L.: Growing among trees of flood plain woods. Collections by W. H. Horr, July 11, 1930, Woodson County; E198, July 5, 1930, Mont- gomery County; P. A. Rydberg, and R. Imler 386, Montgomery County. Malvaceae Ahutilun theophrasti Medic: Low overgrazed prairie pasture, in sandy soil of iui exposed lake bed, and waste places. Collections by W. H. Horr, June 28, 1930, Greenwood County; July 11, 1930, Woodson County; B. L. Wagen- knecht 1804, Chautauqua County. Reported for Woodson County by Mc- Gregor and Voile (1950). Callirhoe alcaeoides (Michx.) Gray: Rocky open wooded banks in creek val- leys, and prairie woodlands. Callirhoe involucrata (T. and G.) Gray: Rock> i)rairie hillsitles, and low prairie woodlands. Chautauqua and Woodson counties. Callirhoe leiocarpa Martin: Prairie hillside. Collection by R. L. McGregor 10358, Elk County. Hibiscus militaris Cav.: In alluvial soil along Fall River, flood plain swamp, and on a dam of a small lake. Montgomery, Wilson, and Woodson counties. Hibiscus trionum L.: In sandy soil of an exposed lake bed, roadsides, prairies, and waste places. Malva neglecta Wallr.: Along the edge of a cultivated field. Collected once, in Woodson County. Sida spinosa L.: Rocky prairie gully, open wooded park bordering Neosho River, and waste places. Chautauqua, Wilson, and \\'oodson counties. Flora and Ecology of Chautauqua Hills 167 Sidopsis hispida (Pursh) Kydb., eiiit'nd Kearney: Reported for Wilson County, Gates (1940). GUTTIFERAE Hypericum drummondii (Greb. and Hook.) T. and G.: Prairie pasture gullies, and rocky slopes of upland woods. Chautauqua, Montgomery, and Woodson counties. Hypericum mutihim L.: In sandy soil of a dry exposed lake bed, in shallow soil over sandstone outcrops of stream banks, and on steep slopes of wooded uplands. Montgomery and Woodson counties. Hypericum perforatum L.: Reported for Woodson County, Gates (1940). Hypericum punctatum Lam.: Roadside thickets, creek valley, and in rocky prairie pasture gullies. Chautauqua and Montgomery counties. Hypericum spaerocarpum Michx.: Rocky prairie gullies, rocky wooded hillsides, and roadsides. CiSTACEAE HeUanthemum bicknellii Fern.: Rocky wooded gullies, prairie pasture gulches, and rocky wooded hillsides. Montgomery and Woodson counties. Lechea tenuifolia Michx.: Creek valley thicket, sandy soil o\er surfacing sand- stone of upland wooded slopes, and rocky banks in creek woods. All counties except Elk and Greenwood. ViOLACEAE Hyhanthus linearis (Torr. ) Shinners: Low prairies, banks of prairie gullies, and open woodlands. Elk, Wilson, and Montgomery counties. Viola kitaiheliana Roeni. and Schultz, var. rafitiesquii (Greene) Fern.: Prairie hay fields, pastures, prairie woodlands, and roadsides. Viola papilionacea Pursh: Open flood plain woods, and prairie woodlands. Chautautjua and Woodson counties. Viola pedata L., var. lineariloha DC: In sandy soil of open oak woods, and in shallow soil over surfacing sandstone of prairie pasture slopes. Chautauqua and W^oodson counties. Viola pedatifida G. Don: Rocky sand>' soil of a prairie pasture hillside. Col- lection by R. L. McGregor 4801, Woodson County. Viola pennsylvanica Michx.: Lowland wood, and low prairie woodlands. Wood- son County. Viola sororia Willd. : Along the border of a low woods. Collection by R. L. McGregor 4179, Chautaucjua County. PassiflorXceae Passiftora incarnata L.: In an open flood plain woods. Collections by W. H. Horr, July 11, 1930, E426, Montgomery County. I'ass-ijiora lutea L., \ar. filahri flora Fern.: Lowland woods, and slopes of upland woods. Collections by W. H. Horr, July 21, 1942, E46(), Montgomery County. LOASACEAE Mentzelia oligosperma Nutt.: Sandy soil of open wooded hillsides, prairie hill- tops, and waste places. All counties except Greenwood and Wilson. Cactaceae Mamillaria similis Engclm.: Rocky low pastures, and rock\- slopes of prairie woodlands. Montgomery County. 168 The University Sc:ience Bulletin Opiintia httmifusa Raf.: Rocky overgrazed pastures. Collections by W. H. Horr, July 5. 1930, Chautauqua County; July 11, 1930, Woodson County. Opuniia macrorhiza Engelm.: Rocky gullies, and prairie pastures. Lythhaceae Ammannia coccinea Rottb.: Border of a marsh in a low woods, in sandy soil of oper. oak woods, and on an exposed lake bed. Wilson and Woodson counties. Lythrum calif omicum T. and C: Hilltops and slopes of prairie pastures, and prairie woodlands. Rotala ramosior (L.) Koekne, var. interior Fern, and Grisc: Wet roadside ditch. Collection by W. H. Horr E157, Wilson County. Onagraceae Circaea quadrisulcata (Maxim.) French, and Sav., var. canadensis ( L. ) Hara: Reported for Greenwood County, Gates (1940). Gaxira biennis L., var. pitcheri Pickering: Rocky prairie slopes, and low prairie woodlands. Woodson County. Gaura coccinea Nutt.: Rocky prairie hillsides, and in a low woodland. Wood- son County. Gaura parviflora Dougl.: Rocky overgrazed prairie pasture slopes. All coun- ties except Elk and Wilson. Jussiaea repens L., var. glabrescens Ktze.: In shallow water of streams and woodland ravines. Wilson and Woodson counties. Ludwigia alternifolia L., var. alternifoUa: Rocky banks of prairie pastures. Woodson County. Ltidwigia alternifolia L., var. pubescens Palmer and Steyermark: Prairie pas- tures, and in open woodlands. Collections by W. H. Horr, July 5, 1930, Chautauqua and Montgomery Coimty; R. L. McGregor 10854, Montgomery County; P. A. Rydberg and R. Imler, 416 Montgomery County. Ludwigia palustris ( L. ) Ell., var. americuna (DC.) Fern, and Grisc: In a creek valley grassland. Collection by W. H. Horr E60, Wilson County. Oenothera biennis L., var. biennis: In sandy soil of a dry exposed lake bed, and along a roadside bank in upland woods. Woodson County. Oenothera biennis L., var. canescens T. and G.: In sandy soil of an upland woods. Woodson County. Oenothera laciniata Hill: Open upland woods, prairie pastures, and flood plain woods. Oenothera laciniata Hill, fonna grandifiora (Wats.) Robinson: Reported for Greenwood County, Gates (1940). Oenothera linifolia Nutt.: In sandy soil of an exposed lake bed, in shallow soil over surfacing sandstone of a prairie gully, and in an overgrazed pasture. Oenothera missouriensis Sims, var. missouriemis: Prairie pastures, and rocky prairie guUies. All counties except Greenwood and Montgomery. Oenothera missourienses Sims, var. incana Gray: Reported for Elk, Greenwood, and Montgomery counties. Gates (1940). Oenothera serrulata Nutt.: Rocky prairie pastures and gullies. Oenothera speciosa Nutt.: Prairie pastures, gullies, roadsides, and rocky banks in low prairie woodlands. Flora and Ecology of Chautauqua Hills 169 Oenothera triloba Nutt.: Overgrazed pasture, roadsides, and on a rocky wooded slope. Wilson and Woodson counties. Stenosiphon linifolius ( Nutt. ) Britt. : Reported for Elk, Greenwood, and Wilson counties. Haloragaceae Mtjriophijllum pinnatum (Walt.) BSP.: In a prairie pasture pond, and in a prairie woodland marsh. Montgomery and Woodson counties. Umbelliferae ChaerophijUum procumhens ( L. ) Crantz. : Reported for Montgomery County, Gates (1940). Cfwerophylhtm tainturieri Hook: Roadsides, borders of 1(5W woodlands, fields, and waste places. Cicuta maculata L. : Flood plain woods, in moist soil bordering a marsh, and waste fields. Conium maculatum L.: Roadsides, and low woods. Woodson County. Cryptotaenia candensis ( L. ) DC: Reported for Woodson County, Gates (1940). Dauciis carota L.: Roadsides, fields, and waste places. Woodson County. Eryngium Icavenworthii T. and G.: Overgrazed pastures, prairie gullies, and waste places. Chautauqua, Wilson, and Woodson counties. Eryngium yuccifoUum Michx.: Prairie pastures, hilltops and slopes. All coun- ties except Elk and Montgomery. Lomatiuin foeniculaceum (Nutt.) C. and R.: Rocky prairie pastures, and in sandy soil of roadside banks. Woodson County. Osrnorphiza lougistylis (Torr. ) DC, var. villicaulis P>rn.: Reported for Chau- tauqua and Greenwood counties. Gates ( 1940). Perideridia americana (Nutt.) Reichenb.: Reported for Montgomery County, Gates (1940). Pohjtaenia mittallii DC: Roadsides, rocky overgrazed prairie pastures, and in shallow soil over sandstone outcrops of an upland wooded gully. Ftilimniiun mittallii (DC.) Britt.: Grazed and ungrazed prairies. Chautauqua and Woodson counties. Sanicula canadensis L.: Creek bank, and flood plain woods. Sanicula gregaria Bickn.: Reported for Greenwood County, Gates (1940). Spermolepis inermis (Nutt.) Math, and Const.: Reported for Elk, Greenwood, and Montgomery counties. Thaspium harhinode (Michx.) Nutt.: In a wet roadside ditch in lowland. Woodson County. Torilis japonica (Hout.) DC: Roadside ditch bank, open creek woods, and waste places. Wilson and Woodson counties. yjzia aurea ( L. ) Koch : Creek valley thicket, flood plain woods, wooded prairie gullies, and creek banks. Woodson County. CORNACEAE Cornus drunimondii C A. Meyer: Flood plain woods, wooded gullies, and upland wooded hillsides. Cornus ohliqua Raf.: Open upland woods, wooded gullies, and open flood plain woods. 170 The University Science Bulletin Cornus stolonifera Michx.: In a rock\- woodfd gully in upland, and in flood plain woods. Woodson County. Pbimulaceae Androsace occidentaUs Pursh: In sandy soil of open oak woods, and upland prairie woodlands. Woodson County. Dodecantheon meadia L., var. meadia: In shallow soil over sandstone outcrops of an open oak woods. Foinid once, in Elk County. Dodecatheon meadia L., var. meadia, fomia album Macbr.: In shallow soil over sandstone outcrops of an open oak woods. Found once, in Elk County. Dodecatheon meadia L., \ar. hrachycarpum (Small) Fassett.: In sandy soil of an open scrub oak woods. Rare. Collection by W. H. Ilorr E159, Mont- gomery County. Samolus parviforus Raf. : In moist soil under a sandstone ledge in an open oak woods. Found once, in Chautauqua County. Sapotaceae Bumelia lanuginosa (Michx.) Pers., var. ohkmgifolia (Nutt.) Clark: Rocky prairie woodlands, and rocky gullies of upland woods. Ebenaceae Diospyros virginiana L.: Dry sandy soil in open woods, flood plain woods, creek valley thickets, and low prairie woodlands. Oleaceae Fraxinus americana L. : Flood plain woods, wooded gullies, and upland wooded hillsides. Fraxinus pennsylvanica Marsh., var. subintegerrima ( Vahl. ) Fern.: Flood plain woods, prairie woodlands, and upland wooded hillsides. Fraxinus quadrangulata Michx.: Creek banks, wooded gullies, and lower slopes of upland wooded hillsides. Elk, Chautauqua, and Montgomery counties. Gentianaceae Centattrium texense (Griseb. ) Fern.: Edge of dry limestone ledges in bluestem pastures. Collections by R. L. McGregor 1159, Montgomery County; 10862, WiLson County. Gentiana puberula Michx.: Prairie pasture, and in open prairie woodland. Two collections, from Woodson County. Sabatia campestris Nutt.: Sandy soil of an open wooded hillside and in a prairie woodland. Collections by R. L. McGregor 10830, Chautau(iua County; 3421, Greenwood County; P. A. Rydberg and R. Imler 408, Montgomery County. Apocynaceae Apocymnn cannabimnn L., var. cannabinum: In sandy .soil of open upland wooded hillsides, wet roadside ditches, border of prairie pasture ponds, in sandy soil of a lake shore, and prairie pastures. Apocynum cannabinum L., var. pubescens (Mitchell) DC: In sandy soil of a prairie pasture. Collection b>- W. H. Ilorr, Jvme 6, 1930, Chautauqua County. Apocynum sibiricum Jacq., var. sibiricum: Reported for Chautauqua and Wilson counties. Gates (1940). Flora and Ecology of Chautauqua Hills 171 Apocynum sihiricum Jacq., var. cordigerum (Greene) Fern.: Reported for Elk County, Gates (1940). ASCLEPIADACEAE Asclepias ample xicaulis Sm.: Reported for Chautauqua County, Gates (1940). Asclepias asperula (Dene.) Woodson, subsp. capricornu (Woodson) Wood.son: Prairie pasture. Collection by R. L. McGregor 4295, Wilson County. Asclepias hirtella ( Pennell ) Woodson : Roadside, low prairie pasture, and waste places. Woodson County. Asclepias incarnata L.: Reported for Woodson County, Gates (1940). Asclepias stenophijlla A. Gray: Low prairie pastures, and rocky open hillsides. Elk, Greenwood, and Wilson counties. Asclepias sullivantii Engelm. ex A. Gray: Prairie pastures, prairie woodlands, and open wooded hillsides. Collections by W. H. Horr, July 5, 1930, July 12, 1930, Chautauqua County; W. H. Horr and L. H. Franklin, E261, Wilson County. Asclepias syriaca L.: Prairie pastures, wooded ravines, and border of a low woods. Greenwood and Woodson counties. Asclepias tuhcrosa L.: Low prairie pastures, fields, and roadsides. Asclepias tuhcrosa L., forma hitea Clute: In a low prairie pasture. Found once, in Chautauqua County. Asclepias verticillata L.: Low prairie hay fields, in shallow soil over surfacing limestone of a prairie ravine, prairie pastures, and open upland woods. Asclepias viridiflora Raf., var. virdifiora: Prairie pastures, rocky prairie gullies, and open wooded hillsides. Greenwood and Wilson counties. Asclepias viridiflora Raf., var. lanceolata (Ives) Torr. : Dry bluestem prairie hillside. Collection by R. L. McGregor 10241, Wilson County. Asclepias viridis (Walt.) Gray: Creek valley grassland, prairie pasture, road- sides, rocky banks in prairie woodlands, and in prairie gullies. Greenwood, Wilson, and Woodson counties. Cytunichiwi laeve (Michx.) Pers.: Reported for Chautauqua, Elk, and Wood- son counties, Gates (1940). CONVOLVULACEAE Convolvulus arvensis L.: Roadsides, fields, and wa.ste places. Convolvulus incanus Vahl.: Reported for Woodson County, Gates (1940). Convolvulus pcllittis Ledeb., forma anestius Fern.: Roadside ditch bank. Col- lection by J. L. Wilson, June 8, 1951, Wilson County. Convolvulus sepium L.: Roadsides, fields, banks of prairie gullies, and wa.ste places. Cuscuta cuspidata Engelm.: Reported for Woodson County, Gates (1940). Cuscuta glomerata Choisy: Reported for Elk, Greenwood, Montgomery and Chautauqua counties. Gates (1940). Cuscuta pcnta^ona Engelm.: Growing over low vegetation in a low open wood- land. Collection by P. A. Rydberg and R. Imler 460, Montgomery County. Evolvuhis nuttdtlianus Schultze: In shallow soil on a limestone plateau, and rocky open wooded Iiillsides. Elk, Wilson, and M()ntgomer>- coimties. Ipomoea coccinca L.: In .sandy soil bordering a cultivated field. Collection by W. H. Horr and R. L. McGregor E424, Montgomery County. Ipomoea hederacca (L. ) Jacq.: In sandy soil of a dr\' lake bed, low open wood- lands, roadsides, fields, and waste places. 172 The University Science Bulletin Ipomoea lacunosa L.: In sandy soil ot an exposed lake bed. Collection by K. L. McGregor 3159, Woodson County. Ipomoea pamhiratu (L.) G. F. W. Mey: Creek valley thicket, growing over low shrubs of Hood plain woods, and waste places. Ipomoea purpurea (L. ) Hoth.: Growing over low \egetation ot a field border- ing a Hood plain woods. Woodson County. POLEMONIACEAE Phlox divaricata L., var. hiphamii (Wood) Wherry: Flood plain woods, and open wooded hillsides. Chautauqua and Woodson counties. Phlox divaricata L., var. laphamii (Wood) Wherry, forma Candida Palmer and Steyennark: Bank of a creek below a wooded cliH. Found once, seven miles southwest of Yates Center, Woodson County. Phlox oklahomensis Wherry: In .shallow soil over surfacing limestone in a prairie pasture. Collection by K. L. McGregor 90.51, Chautauqua County, 9052, Elk County. Phlox pilosa L., subsp. pilosa: Prairies, and prairie woodlands. Chautauqua, Montgomery, and Wilson counties. Phlox pilosa L., subsp. ozarkana Wherry: Prairie ha\' fields, pastures, and prairie woodlands. All counties except Elk and Greenwood. Hydhophyllaceae Ellisia nyctelea L.: Flood plain woods, wooded gullies, and upland wooded hillsides. Chautauqua and Woodson counties. Phacelia hirsuta Nutt.: Reported for Chautauqua and Wilson counties. Gates (1940). BOHACINACEAE Cynoglossum cifficiiude L.: Reported for Chautauqua and Greenwood counties. Gates (1940). Hackelia virginiana (L. ) I. M. Johnston: In a prairie thicket. Collection by W. H. Horr, July 11, 1930, Wood.son County. Heliotropium tenellum (Nutt.) Torr.: In rocky prairie gullies, waste places, and prairie pasture hillsides. Lappula echinata Gilib.: Open prairie woodlands. Woodson Countv. Lappula redoivskii (Homem.) Greene: Reported for Wilson County, Gates (1940). Lithospermtim arvensis L.: Along the fence row in a prairie pasture. Woodson County. Lithospemium incistim Lehm.: Prairie pastures, roadsides, and open prairie woodlands. Elk, Chautauqua, and Woodson counties. Myosotis vema Nutt.: Sandy soil of open scrub oak woods, and in prairie wood- lands. Chautauqua, Greenwood, and Wilson counties. Onosmodium hiftpidissimum Mackenz.: Reported for Greenwood County, Gates (1940). Onosmodium occidentale Mackenze: Rocky prairie pastures, and prairie wood- lands. Montgomery and Wilson counties. Verbenaceae Lippia lanceolata Michx., var. reconita Fern, and Grisc: Low prairie pastures, stream and creek banks, and flood plain wot)ds. Verbena blanchardi Moldenke: Rocky overgrazed pastures. Rare. Woodson County. Flora and Ecology of Chautauqua Hills 173 Verbena hmcteosa Michx.: Fields, waste places, and pastures. Woodson County. Verbena canadensis ( L. ) Britton : Rocky prairie pastures, roadsides, open woods, and prairie woodlands. Frequent. Verbena hastata L.: Rocky prairie pasture. Found once, in Woodson County. X Verbena moechina Moldenke: Prairie pasture, and on an open wooded hill- side. Collections by W. H. Horr, July 28, 1930, Greenwood County; P. A. Rydberg and R. Iniler 434, Montgomery County. X Verbena rydbergii Moldenke: Overgrazed prairie pastures, prairie wood- lands, and waste places. Chautauqua and Woodson counties. Verbena simplex Lehni.: Prairie pastures, rocky prairie woodlands, and open oak woods. Verbena stricta Vent.: Rocky prairie gullies, overgrazed pastures, fields, and open flood plain woods. Verbena urticifolia L.: Creek \ alley thickets, low prairie pastures, flood plain woods, and waste places. Labiatae Agastache nepetindes (L. ) Ktze.: Reported for Woodson County, Gates ( 1940). Blephilia hirsuta (Pursh) Benth.: Creek valley thicket. Collection by W. H. Horr, E144, Wilson County. Glechoma hederacea L., \ar. micrantha Moricand: Reported for Chautauqua County, Gates (1940). Hedeoma hispida Pursh: Prairie pasture gullies, low pastures, flood plain woods, and rocky open woods. Hedeoma pulegimdes (L. ) Pers.: Reported for Wilson County, Gates (1940). Isanthus brachiatus ( L. ) BSP.: Creek banks, moist prairie gullies, and flood plain woods. Wilson and Woodson counties. Lamium amplexicaule L.: Roadsides, low prairie woodlands, banks of wooded gullies. Lamium ptirpuretim L.: Low woodland creek banks, roadsides, and borders of fields. Woodson County. Lconunis marrul)iastrum L.: In shallow soil over surfacing limestone in a prairie pasture. Found once, in Woodson County. Lijcopus americanus Muhl., \ar. scabrif alius Fern.: In sandy soil of a dry lake bed, and open woodlands. Collection b\' R. L. McGregor 4110, Woodson County. Marndnum vulgare L.: Flood plain woods. Collection by W. H. Horr, July 11, 1930, Woodson County. Mentha arvensis L., % ar. tillnsa ( Benth. ) S. R. Stewart, fonna glabrata { Benth. ) S. R. Stewart: Reix)rted for Greenwood County, Gates (1940). Monarda citroid(yra Cerv.: Rock\' prairie slopes, and prairie woodlands. Monarda fisiidosa L., var. mollis ( L. ) Benth.: Prairie hay fields, pastures, and prairie woodlands. Monarda russeliana Nutt, ex Sims: Reported for Woodson County, Horr and McGregor (1949). Nepeta cataria L.: Fields and waste places. Collections by W. H. Horr, July 11, 1930, WiM)dson County; July 12, 1930, Wilson County. Phtjsostegia intermedia (Nutt.) Engelm. and Gray: Prairie pasture hillside. Collection by W. H. Horr, July 5, 1930, Chautauqua County. Phtjsostegia virginiana ( L. ) Benth.: In a rocky prairie pasture. Found once, in Woodson Count\. 174 The University Science Bulletin Prunella caroliniunu Mill: Creek \alle\ thicket, creek banks, and flood plain woods. Chautauqua, Montgomery, and Woodson counties. Pycnanthemum tenuifolium Schrad.: Rocky prairie gullies, prairie pasture slopes, and low prairie woodlands. Salvia aziircu Lam., var. grandiflora Benth.: Prairie pastures, rocky prairie woodlands, wooded gullies, and roadsides. Salvia rcfiexa Hornem.: Reported for Greenwood County, Gates (1940). Scutellaria incana Biehler: Prairie hay field. Collection by B. L. W'agenkmcht 1823, Chautauqua County. Scutellaria lateriflora L.: Reported lor Woodson County, Gates ( 1940). Scutellaria ovata Hill: Reported lor Chautauqua County, Gates (1940). Scutellaria parviila Michx., var. parvula: Prairie hay fields, pastures, and open prairie woodlands. Montgomery and Woodson coimties. Scutellaria parvula Michx., var. australis Fassett.: Prairie hay fields and pas- tures. Woodson County. Scutellaria parvula Michx., var. leonardii (Epling) Fern.: Grazed and ungrazed prairies, and rocky prairie woodlands. All counties except Elk and Mont- gomery. Stachtjs palustris L.: On a creek bank, and in flood plain woods. Collections by W. H. Horr, July 5, 1930, Montgomery Coimty; July 12, 1930, Wilson County. Stachijs temdfolia Willd.: On a wooded creek bank. Found once, in Woodson County. Teucrium canadense L., var. virginicum (L.) Eat.: Rocky prairies, low prairie woodlands, roadsides, and border of fields. SOLANACEAE Chamaesaracha conioide.s ( moric. ) Britt.: Reported for Woodson Count)-, Gates (1940). Datura stramonium L.: Roadsides, field, and waste places. Physalis heterophylla Nees: Along the border of a flood plain woods. Wood- .son County. Physalis pubescens L.: Prairie pasture. Collection by R. L. McGregor 10857, Woodson County. Physalis pumila Nutt.: Low prairie pastures, rocky upland open woods, and waste places. Greenwood, Montgomery, and Wilson counties. Physalis virginiuna Mill.: Overgrazed pastures, in sandy .soil of an exposed lake bed, rocky waste places, and roadsides. Solanum americanum Mill.: Overgrazed pastures, fields, and low open woods. Woodson County. Solanum carolinense L.: Waste places, overgrazed pastures, fields, and road- sides. Solanum carolinense L., forma alhiflorum (O. Ktze. ) Benke.: Roadsides, fields, waste places, and overgrazed pastures. Solanum elaeagnifioium Cav.: Reported for Wood.son County, Gates (1940). Solanum interius Rydb.: Reported for Woodson County, Gates ( 1940). Solatium rostratum Dunal: In sandy soil of dry lake bed, roadsides, fields, and overgrazed pastures. Frequent. Flora and Ecology of Chautauqua Hills 175 SCROPHULAKIACEAE Bacopa rotundifolia ( Michx. ) Wettst. : In sandy soil of dry lake bed, in a tem- porary water hole in a prairie pasture, and in moist depressions in fields. Wilson and Woodson counties. Buchnera americana L.: Rocky prairie gullies, and low prairie woodlands. Col- lections by W. H. Horr, July 5, 1930, Montgomery County; July 12, 1930, Wilson County; July 10, 1930, Woodson County; R. L. McGregor 4457, Elk County; 2393, Montgomery County. Castilleja sessiliflora Pursh.: Rocky prairie hay field. Rare. Collection by R. L. McGregor 10259, Wilson County. Collinsia violacea Nutt.: In sandy soil of open scrub oak wooded hillsides, and prairie woodlands. Collections by W. H. Horr E123, Wilson County; R. L. McGregor 9046, Chautauqua County; 9034, Montgomery County; 9030a, Wilson County; 9011, Woodson County. Conobea multifida (Michx.) Benth.: Rocky scrub oak woods, prairie pastures, sandy soil of a dry lake bed, and prairie woodlands. Gerardia aspera Dougl.: Reported for Greenwood, Montgomery, Wilson, and Woodson counties. Gates (1940). Gerardia auriculata Michx.: In a rocky prairie pasture swale. Collection by R. L. McGregor 5237, Montgomery County. Gerardia densiftora (Benth.) Pennell: Reported for Elk and Woodson counties, Gates (1940). Gerardia skinneriana Wood: Prairie pastures, wooded gullies, open woods, and prairie woodlands. Woodson County. Gerardia tenuijolia Vahl., var. parviflora Nutt.: Reported for Woodson County, Gates (1940). Gratiola virgitiiana L.: In a moist prairie gully. Montgomery and Woodson counties. Linaria canadensis ( L. ) Dumont, var. texana ( Scheele ) Pennell : Wooded gully, open upland woods, prairie pastures, and sandy soil of roadsides. Chau- tauqua and Woodson counties. Lindernia anagallidea (Michx.) Pennell: Rocky prairie hillside, and in sandy soil of a dry lake bed. Woodson County. Lindernia duhia (L. ) Pennell: In sandy soil of an ex-posed lake bed, and open upland woods. Collections by W. H. Horr, July 5, 1930, Chautauqua and Montgomery counties; R. L. McGregor 3147, Woodson County. Mimulus alatus Ait.: In moist soil on the bank of an upland wooded gully, creek bottoms, and wet ditch banks. Woodson County. Pedicularis canadensis L.: Reported for Chautauqua County, Gates (1940). Penstemon cohaea Nutt.: l^ocky prairie slopes, and iirairie woodlands. Penstemon digitalis Nutt.: Reported for Montgomery and Wilson counties. Gates (1940). Penstemon tubaeflorus Nutt.: Prairie pastures, rocky wooded gullies, and prairie woodlands. Scrophularia marilandica L.: Prairie pasture. Collection by W. H. Horr, July 12, 1930, Wilson County. Seymeriu macroplujUa Nutt.: Open scrub oak woods, and in sandy soil below a wooded cliff. Woodson County. 176 The University Science Bulletin Verbascwn hlattaria L.: Rocky overgrazed pastures, and prairie woodlands. Verbasciim blatteria L., forma alhiftara (Don) House: In overgrazed pastures. Collections by R. L. McGregor 4323, Montgomery County; 4288, Wilson County. Verbascwn tluipsus L.: Rocky prairie pastures, waste places, and borders of fields. Veronica arvensis L.: Open banks of wooded gullies, and in prairie pastures. Woodson County. Veronica peregriim L., var. peregrina: Banks of prairie pasture ponds, lake shore in upland woods, and in moist gullies. Woodson County. Veronica peregrina L., var. xalapensis (HBK.) St. John and Warren: In open upland woods, in sandy soil of a dry lake bed, and in a wooded prairie gully. Woodson County. Veronica angalis-aquatica L.: Along the edge of a low woodland water hole. Found once, in Woodson Coimty. BiGNONIACEAE Campsis radicans ( L. ) Seem: Growing over low vegetation along a river bank, and in a thicket of a prairie woodland. Montgomery and Woodson counties. Catalpa speciosa Warder: Flood plain woods, fields, and invaded woods of prairies. Martyinaceae Proboscidea louisianica (Mill.) Thell.: Reported for Chatuauqua County, Gates (1940). Orobanchaceae Orobanche ludoviciana Nutt.: Reported for Greenwood County, Gates (1940). Orobanche unijiora L.: Reported for Chautauqua, Montgomery, and Woodson counties. Gates (1940). Lentibulariaceae Utricularia gibba L.: In the water of a low prairie woodland marsh. Woodson County. Acanthaceae Dicliptera brachiata (Pursh) Spreng: Moist soil along a river bank. Collection by R. L. McGregor 5735, Wilson County. Justicia americana ( L. ) Vahl. : In prairie streams, shallow creeks, and moist wooded gullies. All counties except Elk and Greenwood. Ruellia humilis Nutt., var. humilis: Low prairies, prairie woodlands, open oak woods, and open flood plain woods. Ruellia humilis Nutt., var. expansa Fern.: Low prairie hay fields. Woodson County. Ruellia humilis Nutt., var. frondosa Fern.: Prairie gully, and in shallow soil over surfacing limestone in a prairie pasture. Collections by R. L. McGregor 4465, Elk County; 10865, Wilson County. Ruellia humilis Nutt., var. langiflora (Gray) Fern.: Grazed and ungrazed prairies, and open prairie woodlands. Ruellia strepens L. : Flood plain woods. Flora and Ecology of Chautauqua Hills 177 Phrymaceae Phryma leptostachya L. : Along the border of a wooded prairie gully, open Hood plain woods, and lowland thickets. Greenwood, Wilson, and Woodson counties. Plantaginaceae Plantago aristata Michx.: In sandy soil of an exposed lake bed, prairie pastures, open grazed woods, roadsides, prairie woodlands, and waste places. Fre- quent. Plantago lanceolata L.: Roadsides, fields, and waste places. Phintago major L.: Waste places, and borders of fields. Woodson County. Plantago purshii R. and S.: Prairie pastures, and prairie woodlands. Plantago pusilla Nutt.: Open upland woods. Plantago rhodosperrna Dene.: Prairie pastures, lowland woods, and waste places. Greenwood and Woodson counties. Plantago rugelii Dene.: Flood plain woods, creek valley thickets, and roadsides. Collections by W. H. Horr, July 5, 1930, Chautauqua and Montgomery counties; July 11, 1930, Woodson County. Plantago virginica L.: Open upland woods, rocky prairie pastures, prairie gullies, and roadsides. Frequent. Rubiaceae Cephalantlms occidentalis L. : Upland wooded gullies, rocky creek banks, rocky shore of lakes, and open flood plain woods. Diodia teres Walt., var. setifera Fern, and Grisc. : Rocky prairie pastures, fields, waste places, and in sandy soil of an exposed lake bed. Galium aparine L.: Flood plain woods, and lower slopes of upland woods. Chautauqua and Woodson counties. Galium circaezans Michx., var. hypomalacum Fern.: Flood plain woods, collec- tion by W. H. Horr, July 11, 1930, Woodson County. Galium triflorum Michx.: Wooded prairie ravines, and flood plain woods. Greenwood and Woodson counties. Galium virgatum Nutt.: Low woodlands, in shallow soil over surfacing lime- stone of a prairie hay field, and in overgrazed pastures. Collections by R. L. McGregor 10213, 4343, Montgomery County; 4292, 10255, Wilson County. lledyotis minima (Beck) T. and G.: In shallow soil over surfacing sandstone of prairie pastures, and in upland prairie woodlands. lledyotis minima (Beck) T. and G., forma alhifiora Lathrop: In shallow soil over surfacing sandstone in a prairie bordered by scrub oak woods. Collected in one location four miles southwest of Yates Center, Woodson Count)'. lledyotis nigricans (Lam.) Fosberg: Flood plain woods, prairie pastures, and open upland woods. Caprifoliaceae Lonicera dioica L., var. glaucesens ( Rydb. ) Butters: On a wooded limestone escarpment. Rare. Collections by R. L. McGregor 2404, 10841, Mont- gomery County. Lonicera japonica thunb.: Rocky wooded thicket covered hillside near a farm site. Col'ection by R. L. McGregor 4315, Wilson County. 178 The University Science Bulletin Lonicera prolifera ( Kirchn. ) Rehd. : Just below a wooded limestone escarp- ment. Rare. Collection by R. L. McGregor 10230, Montgomery County. Sambucus canadensis L.: Wooded river bank, flood plain woods, and lowland thickets. SympJwricarpos orbiculatus Moench.: Open flood plain woods, and upland scrub oak woods. Frequent. Triosteuni aurantiacum Bickn., var. illnoense (Wieg. ) Palmer and Steyerm.: Reported for Wilson County, Gates ( 1940 ) . Triostcum perfoliaturn L.: Reported for Chautauqua, Greenwood, and Wilson counties. Gates (1940). Viburnum rufiduhim Raf.: Open wooded hillsides, and wooded gullies and ra\ines. Valerianackae Valerianella radiatu (L. ) Dufr., var. radiata: Moist prairie gullies, muddy banks of pasture ponds, wet roadside ditches, and low prairies. Montgomery and Woodson counties. Valerianella radiata ( L. ) Durfr., var. missouriensis Dyal: In a deep depression of a prairie hay field. Collection by R. L. McGregor 10209, Woodson County. Valerianella stenocarpa (Engelm. ) Krock., var. parviflora Dyal: Reported for Wilson County, Gates ( 1940 ) . CUCURBITACEAE Echinocystis lobata ( Michx. ) T. and G.: Reported for Geenwood and Wilson counties. Gates (1940). Sicyos angulatus L.: In rich woodland bordering Neosho River, and in thickets of low woods. Greenwood and Wilson counties. CAMPANtn^ACEAE Campanula americana L., var. illinoensis (Fresn.) Farw.: Flood plain woods, low fields, creek bank thickets, and open woodlands. Chautauqua and Woodson counties. Lobelia cardinalis L.: In moist soil over surfacing sandstone of wooded gullies, and wooded creek banks. Lobelia siphilitia L.: Flood plain woods. Found once, in Woodson County. Lobelia spicata Lam., var. leptostachys (A. DC.) Mackenz. and Bush.: In sandy soil of an e.vposed lake bed, prairie gulches, and moist wooded gullies. Montgomery and Woodson counties. Lobelia splendem \\'illd.: Flood plain woods. Collection by O. S. Fearing and G. Latham, Aug. 14, 1950, Wilson County. Specularia biflora (R. and P.) Fisch. and Mey: Low prairie hay fields and pastures. Woodson County. Specularia perfoliata (L. ) A. DC: R(Kky prairie gullies, prairie pastures, creek banks, and low prairie woodlands. Specularia leptocarpa (Nutt.) Gray: Rocky prairie pastures, in sandy soil of an exposed lake bed, and in prairie woodlands. COMPOSITAE Achillea lanulosa Nutt.: Prairie pastures, and prairie woodlands. Frequent. Actinomeris alternifolia ( L. ) DC. : Moist wooded gullies, and open flood plain woods. Woodson County. Flora and Ecology of Chautauqua Hills 179 Ambrosia artemisiifolia L., var. elatior (L. ) Descourtils: Wooded river bank, in sandy soil of a dry lake bed, fields, and wooded prairie gullies. Chau- tauqua, Wilson, and Woodson counties. Ambrosia bidentatu Michx.: Rocky prairie pasture. Collection by B. L. Wagen- knecht 2239, Wilson County. Ambrosia psilostachija DC, var. coronopifolia (T. and G.) Farvv.: Overgrazed prairie pastures, and prairie woodlands. Ambrosia simulans Shinners: Dry prairie pasture. Collection by W. H. Horr and R. L. McGregor E435, Woodson County. Ambrosia trifida L., var. trifida: Wooded creek banks, waste field bordering river woods, wooded prairie gullies, and sandy soil of an exposed lake bed. Woodson County. Ambrosia trifida L., var. texana Scheele: Dense creek woods. Found once, in Woodson County. Antennaria campestris Rydb.: Reported for Elk and Chautauqua counties. Gates (1940). Antennaria fallax Greene: In sandy soil of open scrub oak hillsides. Chau- tauqua and Woodson counties. Antennaria neglecta Greene, var. neglecta: Sandy soil of open scnib oak woods, prairies, and prairie woodlands. Antennaria neglecta Greene, var. randii (Fern.) Cron.: Sandy .soil of a prairie hay field. Collection by R. L. McGregor 9049, Chautaucjua County. Antennaria plantaginijolia (L. ) Richards: In shallow soil over sandstone out- crops of a scrub oak woods. Greenwood and Wilson counties. Anthemis cottda L.: In sandy soil of a dry lake bed, roadsides, fields, and creek valley woodlands. Collections by W. H. Horr, July 11, 1930, Woodson County; R. L. McGregor 3185, Woodson County. Arctium minus ( I lill ) Bemh. : Reported for Elk, Chautauqua, Greenwood, and Wilson counties. Gates ( 1940). Artemisia htdoviciana Nutt., \ar. ludoviciana: In shallow soil over limestone .surfacing of a prairie pasture. Collection by B. L. Wagenknecht 2191, Wilson County. Artemisia ludoviciana Nutt., var. mexicana (Willd. ) Fern.: In sandy soil of gully bank in a prairie hay field and a prairie pasture. Collections by R. L. McGregor 9818, Woodson County; W. II. Horr and R. L. McGregor E432, Montgomery County. Aster anomalus Engelm.: In calcareous soil of a roadside ditch bordering a prairie pasture. Found once, in Woodson Count>\ Aster azureus Lindl.: Grazed and ungrazed prairies. Wilson and Woodson counties. Aster azureus Lindl., forma laevicaulis Fern.: In sandy-clay soil of a scrub oak wooded hillside. Collection by R. L. McGregor 4085, Wood.son County. Aster ericoides L.: Rocky prairie pasture gulches, prairie hillsides, fields, and roadsides. Aster laevis L.: Prairie hay fields, and on rocky banks in river woodlands. Greenwood and Woodson counties. Aster oblongifolius Nutt.: In shallow soil over surfacing sandstone of open oak woods, and in prairie hay fields. Greenwood, Wilson, and Woodson counties. 180 The University Science Bulletin Aster patens Ait.: Prairie pastures, open oak woods, and rocky prairie gullies. Montgomery and Woodson counties. Aster praealtus Poir.: Prairie pastures. Collections by B. L. Wagenknecht 1728, Montgomen,- County; 2242, Wilson County. Aster sagittifolius Wendenieyer, var. safiittif alius: Open upland wooded hill- side. Collection by B. L. Wagenknecht 2254, Wilson County. Aster sagittifolius Wedemeyer, var. drummondii (Lindl.) Shinners: Prairie pastures, oak wooded hillsides, and rocky prairie gullies. Greenwood and Woodson counties. Aster sericeus Vent.: Reported for Chautauqua, Elk, and Greenwood counties, Gates (1940). Bidens hipinnata L.: Reported for Elk, Wilson, and Montgomery counties, Gates (1940). Bidens cemua L., var. elliptica Wieg.: In sandy soil of an exposed lake bed. Collection by R. L. McGregor 4105, Woodson County. Bidens frondosa L.: Wooded upland gullies, and in sandy soil of a dry lake bed. Wilson and Woodson counties. Bidens poUjlcpis Blake: Flood plain woods, prairie pastures, roadsides, prairie woodlands, and in sandy soil of a dry lake bed. Woodson County. Boltonia latisquama Gray, var. latisquama: In a rocky wooded prairie gully. Collected once, in Woodson Coimty. Boltonia latisquavui Gray, var. recognita Fern, and Grisc. : In a moist prairie gulch. Found once, in Woodson County. Cacalia artiplicifolia L.: Reported for Greenwood and Wilson counties, Gates (1940). Cacalia plantaginea ( Raf . ) Shinners: Prairie pastures, fields, and in shallow soil over limestone outcrops in prairie gullies. Carduus nutans L.: Disturbed area of a prairie pasture. Collection by R. L. McGregor 4462, Elk County. Centaurea cijanus L.: Prairie pasture, and open woodland in the Verdigris River flood plain. Chautauqua and Montgomery counties. Chaetopappa asteroides DC: In shallow soil over surfacing sandstone in open scrub oak woods. Chautauqua, Wilson, and Woodson counties. Chrysanthemum leucantliemum L., var. pinnatifidutn Lecoq. and Lamotte: On a creek bank, and in a low prairie pasture. Collections by W. H. Ilorr, June 28, 1930, Greenwood County; July 11, 1930, Woodson County. Chrysopsis pilosa Nutt.: Light sandy soil of open scrub oak woods and prairie woodlands. Chryopsis villosa (Pursh) Nutt., var. villosa: Light sandy soil along the border of a prairie hay field. Collection by B. L. Wagenknecht 1706, Montgomery County. Chrysopsis villosa (Pursh) Nutt., var. angustifolia ( Rydb. ) Cronq.: In sandy soil of an open scnib oak woods. Found once, in Woodson County. Chrysopsis villosa (Pursh) Nutt., var. hispida (Hook.) and Gray: In sandy soil of an open wooded hillside. Collection by R. L. McGregor 3184, Woodson Covmty. Cichorium intybus L.: Fields, roadsides, and pasture gullies. Wilson and Woodson counties. Flora and Ecology of Chautauqua Hills 181 Cir«um altissimum ( L. ) Spreng.: In sandy soil of open oak woods. Chau- tauqua and Woodson counties. Cirsium undulatum ( Nutt. ) Spreng, var. imduhitum: Rocky prairie pastures, prairie gullies, and fields. Cirs'iuni undulatum (Nutt.) Spreng, var. undulatum, forma album Farwell: In a rocky overgrazed pasture. Collected once, in Woodson County. Cirs-ium undulatum (Nutt.) Spreng, var. megacephalum (Gray) Fern.: Prairie pasture. Collection by W. H. Horr, June 28, 1930, Greenwood County. Cunyza canadensis ( L. ) Cronq.: Border of flood plain woods, sandy soil of an open oak wood hillside, and in waste places. Chautauqua and Woodson counties. Conijza ramosissima Cronq.: Creek valley thicket. Collection by W. H. Horr, July 12, 1930, Wilson County. Coreopsis grandiflora Hogg. : In shallow soil over surfacing SiUidstone in upland prairie woodlands. Chautauqua County. Coreopsis palmata Nutt.: Prairie pasture hillsides, rocky prairie gullies, open oak woods, and roadsides. Montgomery, Wilson, and Woodson counties. Coreopsis tinctoria Nutt.: In sandy soil of open scrub oak woods. Chau- tauqua, Montgomery, and Woodson counties. Dyssodia papposa ( Vent. ) Hichc. : Reported for all counties except Chautau- qua, Gates (1940). Echinacea angustifolia DC: Prairie hay field hillsides, rocky banks of prairie gullies, and roadside banks. Echinacea atrorubens Nutt.: Ungrazed prairies, rocky slopes of prairie wood- land hillsides, and roadside banks. Echinacea pallida Nutt.: Rocky banks in prairie pastures, roadside banks, and low prairie hay fields. Eclipta alba (L. ) Hassk.: In sandy soil of an exposed lake bed. Reported for Woodson County, McGregor and Voile (1950). Erechtites hieracifolia ( L. ) Raf., var. intermedia Fern.: Reported for Wood- son County, Gates ( 1940). Erigeron annuus ( L. ) Pers.: Overgrazed prairie pastures, fields, and prairie woodlands. Wilson and Woodson counties. Erigeron philadelphicus L.: In moist prairie pasture gulches, and wet road- side ditches. Chautauqua and Woodson counties. Erigeron strigosus Muhl. ex Willd., \ar. .strigosus: Prairie pastures, fields, roadsides, prairie woodlands, and waste places. Erigeron strigosiis Muhl. ex Willd., var. heyrichii ( Hsch. and Mey) Gray: Flood plain woods, rocky roadside ditch, fields, rocky prairie pastures, and waste places. Montgomery and Woodson covmties. Eupatorium altissimum L.: Prairie pastures, thickets, and borders of woods. Woodson Counts'. Eupatorium perfoliatum L.: Reported for Chautau(iua C^ounty, Gates (1940). Eupatorium rugosum Houtt.: Reported for all counties except Greenwood and Chautauqua, Gates (1940). F.upatorium rugosum Houtt., forma villicaule Fern.: Flood plain woods, along borders of low woodlands, and wooded gullies. Montgomery and Woodson counties. 182 The University Science Bulletin Eupatorium serotinum Michx.: In low prairie woodlands. Elk and Woodson counties. Euthamia gymnospermoides Greene: Prairie pastures, and prairie woodlands. Woodson County. Caillardia pitlclwlla Foug.: In shallow soil o\'er surfacing limestone in a prairie pasture. Collection by R. L. McGregor 10360, Elk County. CnuplmUum ohtu.sifolium L.: In sandy soil of prairie pastures. Collections by W. H. Horr and R. L. McGregor E421, Montgomery County; R. L. Wagenknecht 2241, Wilson County. Gnaplialium purpureum L.: In sandy soil of a dry lake bed. Collections by R. L. McGregor 3121, Woodson County. Grindclia lanceolata Nutt.: Wasteland, roadsides, and fields. Collection b\' R. L. McGregor 5738, Wilson County. Grindelia squarrosa (Pursh) Dunal.: Rocky prairie banks, and roadsides. Woodson County. Helenium autumiude L., var. autumnale: In a moist wooded gully in a prairie pasture. Woodson County. Helenium autumnale L., var. canaliculatum (Lam.) Torr. and Gray: Specimen in the Gray Herbarium. Hitchcock 737, Chautauqua County. Helenium monatamim Nutt.: Prairie pasture hillside. Collection by W. H. Horr and R. L. McGregor E425, Montgomery County. Helenium tenuifolium Nutt.: Low prairie woodlands, overgrazed pastures, and fields. Helianthus annus L.: Roadside ditch banks, fields, sandy soil of dry lake bed, waste places, and prairie woodlands. Helianthus grosseserratus Martens: Roadside ditch, borders of fields, and waste places. Chautauqua and Woodson counties. Helianthus hirsuta Raf., var. hirsuta: In sandy soil of roadsides, and on a sandy bank in an open oak woods. Chautauqua, Montgomery, and Wood- son Counties. Helianthus hirsiita Raf., \ar. stenophyllus T. and G.: In sandy soil of a bank in an open scrub oak woods. Found once, in Woodson County. Helianthus laetiflorus Pers., var. rigidus (Cass.) Fern.: In sandy soil of a prairie pasture gully, roadsides, railroad embankments, fields, and waste places. Montgomerys Wilson, and \\'oodson counties. Helianthus leptocaulis (S. Wats.) Rlakc: Reported for Montgomer>' County, Gates (1940). Helianthus maximiliani Schrad.: Roadside banks, fields, border of flood plain woods, and in sandy soil of an exposed lake bed. Wilson and Wood- son counties. Helianthus mnllt.y Lam.: Fields, border of low woods, roadsides, prairie gul- lies, waste places, and railroad embankments. Helianthus petiolaris Nutt.: Roadsides, waste places, fields, and railroad em- bankments. Woodson County. Helianthus salicif alius A. Dietr.: In dr>' soil over limestone of pasture and roadside banks. Helianthus tuberosus L.: Roadsides, border of flood plain woods, waste places, and fields. Flora and Ecology of Chautauqua Hills 183 Heliopsis helianthoides ( L. ) Sweet, var. scahra (Dunal) Fern.: Rocky prairie pasture bank, thickets, and borders of woods. Wilson and Woodson counties. Hieracium gronovii L.: In sandy soil of open upland woods, rocky banks of wooded gullies, and in shallow soil over sandstone outcrops. Chautauqua and Woodson counties. Hieracium longipihim Torr. : Low prairies, and prairie gullies. Montgomery, Wilson, and Woodson counties. Hymenopappus corymhosus T. and G.: Moist prairie gulch, and on a rocky prairie hillside. Collections by R. L. McGregor 4326, Montgomery Coimty; 4298, Wilson County. Hymenopappus scabiosaeuus L'Her. : Roadside bordering prairie pasture, flood plain woods, and rocky prairie pastures. Iva ciliata Willd.: On a rocky bank in an open woods, border of woods, and rocky waste places. Woodson County. Krigia dandelion (L.) Nutt.: Rocky prairie woodlands. Chautauqua County. Krigia occidentalis Nutt.: Reported for Chautauqua County, Gates (1940). Kuhniu eupatoriodes L., var. corymbidosa T. and G.: Grazed and imgrazed prairies, and rocky prairie woodlands. Greenwood, Montgomery, and Wilson counties. Lactuca canadensis L. : Reported for all counties except Chautauqua and Wilson, Gates (1940). Lactuca floridaim ( L. ) Gaertn.: Flood plain woods, and in moist soil of a rocky wooded gully. Woodson County. Lactuca ludoviciana (Nutt.) Riddell: Low open woods. Collection 1)\ W. H. Horr, July 5, 1930, Montgomery County. Lactuca scariula L.: Low prairie woodlands, rocky waste places, roadsides, fields, and along the border of an overgrazed pasture. Chautauqua, Wil- son, and Woodson counties. Liatris aspera Michx.: In sandy soil of prairie pastures, roadsides, and prairie woodlands. Liatris punctata Hook., var. punctata: Prairie woodlands, rocky prairie pastures, and roadsides. Woodson County. Liatris punctata Hook., var. nehraskana Gaiser: Roadsides, prairie pastures, and borders of low woods. Woodson County. Liatris pycnostachya Michx.: Prairie pastures, roadsides, prairie gullies, open scrub oak woods, and along edges of fields. Liatris squarrosa ( L. ) Michx., var. glahrata ( Rydb. ) Gaiser: Roadsides, fields, waste places, and in open oak woods. Montgomery and Woodson coimties. Liatris squarrosa (L. ) Michx., var. hirsuta Rydb. ex Gaiser: Upland wooded gully, prairie pastures, roadsides, and in sandy soil of prairie woodlands. Chautaucjua, Montgomery, and Woodson counties. Marshallia caespitosa Nutt.: In a prairie pasture bordering a wooded lime- stone escarpment. Collection b\- R. L. McGregor 3400, Montgomery Coimty. Parthenium hispidum Raf.: Prairie pasture hillsides. Collections b\ \\. L. McGregor 10341, 10366, Elk County. Prenanthes aspera Michx.: Reported for Elk, Montgomery, and \\'()odson counties. Gates (1940). 184 The University Science Bulletin Prionopsis ciliata Nutt.: In clay soil along the fence row of a cultivated field. Collection by B. L. Wagenknecht 1810, Chautauqua County. Pyrrhopappus carolimanus (Walt.) DC: Low prairie woodlands, roadsides, open scrub oak woods, in sandv' soil of an exposed lake bed, creek banks, and flood plain woods. Rutibida columnifera (Nutt.) \\'. and S.: Prairie pasture hillsides, roadsides and prairie gullies. All counties except Greenwood and Wilson. Ratihida pinnata (Vent.) Barnh.: Rocky prairie gullies, and low prairie woodlands. Wilson and Woodson counties. Rudhcckia amplexicaulis Vahl. : Flood plain woods, and low prairie wood- lands. All counties except Elk and Woodson. Rudhcckia laciniata L.: Border of a Hood plain woods. Collection by R. L. McGregor 5734, Wilson County. Rudhcckia seratina Nutt.: Rocky prairie pastures, along the edge of flood plain woods, and roadsides. Chautauqua, Montgomery, and Woodson counties. Rudhcckia triloha L.: Flood plain woods, and in moist .soil at the border of a lake in upland woods. Woodson County. Senecio ohovatus Muhl., var. rotundus Britt.: In sandy soil of an oak wooded hillside. Found once, in Chautauqua County. Senccio plattcnsis Nutt.: In rocky prairie gulches, prairie pastures, rocky wooded ravines in upland woods, and prairie woodlands. Scrina oppositifolia ( Raf. ) Ktze. : Roadside rocky banks, prairie pastures, moist prairie gulches, along the edge of pasture ponds, and gullies in woodlands. Woodson County. Silphium lacimatum L.: Roadsides, fields, waste places, and along the border of flood plain woods. Silphium pcrfuliantum L.: Along the edge of a flood plain woods. Collected once, in Woodson County. Silphium speciostim Nutt.: Roadsides, rocky banks in wasteland, borders of low woods, and in rocky prairie gullies. Solidago altissima L.: Roadsides, borders of woodlands, prairie pastures, and rocky prairie gullies. Solidago flexicaidis L.: In moist soil of a bank in a flood plain woods. Col- lection by B. L. Wagenknecht 1815, Chautauqua County. Solidago gigantca Ait., \ar. Iciophtjlla Fern.: Along the edge of flood plain woods. Chautauqua and Woodson counties. Solidago litidheimeriami Scheele: Reported for Elk County, Gates (1940). Solidago missouricims Nutt., var. glahcrrina (Martens) Rosendahl.: Prairie pastures, roadsides, prairie woodlands, k)vv prairie hay fields, and open scrub oak woods. Solidago petiolaris Ait.: In a rocky gully of a prairie hay field. Collection by R. L. .McGregor 9802, Greenwood County. Solidago rigida L.: In sandy .soil of prairie woodlands, rocky prairie gulches, roadsides, and waste places. Solidago spcciosa Nutt., var. angustata T. and G.: Moist prairie pasture gulch. Collection by W. H. Horr and R. L. McGregor E430, Chautauqua County. Solidago ulmifolia Mvihl., \ar. ulmifloia: In sandy soil of open scrub oak woods, prairies, wooded gullies, and roadsides. Frequent. Flora and Ecology of Chautauqua Hills 185 Solidugo ulmifolia Muhl., var. microphylla Gray: In sandy soil of an open scnib oak woods. Collection by R. L. McGregor 10858, Montgomery County. Sonchus asper ( L. ) Hill : Roadsides, fields, and waste places. Taraxacum erythrospermum Andrz. : Roadside ditch banks, prairie gullies, overgrazed pasture, and fields. Taraxacum officiiuile Wiggers: In moist prairie gullies, borders of woods, roadside banks and fields. Theleaperma intermedium Rydb.: Prairie pastures, rtK'ky prairie gullies, in sandy soil of open oak woods, and rocky prairie woodlands. Tragopogon pratensis L.: Overgrazed prairie pasture. Montgomer>' and Woodson counties. Verhesina virginica L. : Flood plain woods, and in sandy soil of an open scrub oak woods. Chautauqua and Woodson counties. Vernormi haldwini Torr., var. interior (Small) Schub.: Fields, overgrazed pastures, waste places, grazed prairie woodlands, and roadsides. Vemonia baldwini Torr., var. interior (Small) Schub. X V. crinita Raf.: ReE>orted for Chautauqua, Montgomery, and Woodson counties. Gates (1940). Vernonia crinata Raf.: L'. Fig. 2. Closer view of a wooded prairie hilltop of the Chautauqua Hills. Three miles north of New Albany, Wilson County. Flora and Ecology of Chautauqua Hills 191 PLATE 3 .•!t^^j«3»^V-v >a«*¥^^S%^*i; Fig. 1. Open scrub oak woods. Eight miles northeast of Sedan, Chautauqua County. 4^^". Fig. 2. Interior of the above woods. 192 The University Science Bulletin PLATE 4 Fig. 1. Upper slope ot an upland woods. Five miles west, four miles south of Yates Center, Woodson County. Fig. 2. Lower part of the same slope. Flora and Ecology of Chautauqua Hills 193 PLATE 5 Fig. 1. Wooded hillside over limestone outside the border ot the Chautauqua Hills. Five miles west of Syraeuse, Montgomery County. Fig. 2. Wooded hillside over sandstone in the Chautaufiua Hills. Tliree miles southeast of Sedan, Cliautamiua County. 7—2656 194 The University Science Bulletin PLATE 6 KiG. 1. I'ruirif in tlif Ciuiutaiuiua Hills. Iwu links sontli, tlirr east of Toronto, Woodson Connty. I' nnu-s Fig. 2. Prairie slope. Four miles east of Sedan, Chautauqua County. Flora and Ecology of Chautauqua Hills 195 PLATE 7 F.c;. 1. L^ngrazecl prairit- in good toiulitioii, sliowing nearly continuous grass cover. Six miles north of Longton, Elk Connty. Fig. 2. Clipped vegetation of the above prairie showing extent of ground co\er. 196 The University Science Bulletin PLATE 8 ,^. t . Fic. 1. Grazed prairie in tair condition sliovNing abundance ot i(>rhs. Five miles east of Sedan, Chautauqua Count)'. Fig. 2. Clipped vegetation of the above prairie showing e.xtent of ground cover. Flora and Ecology of Chautauqua Hills PLATE 9 1- ^5(il \(ij;etation ol tin- iIjom- prairie sliuvsuig extent of ground cover. 198 The University Sc;ienc:e Bulletin PLATE 10 Fic. 1. Flood plain woods ot Fall river. One mile east of Fall rixcr, Greenwood Countv. Fk;. 2. Interior of flood plain woods near the bank of F'all river. Five miles west of Fredonia, Wilson Coimty. J Flora and Ecology of Chautauqua Hills 199 PLATE 11 Fig. 1. Prairie woodland. Two miles north oi iNew All)any, Wilson County -'■S^:f-' -... .. j/». '\^% M^:?5te^'^ Fk;. 2. I'rairic woodland showing a stand of (.^litTtiiv stellata and p. marihindica. 200 The University Science Bulletin PLATE 12 wm 1:^. -'/'it, m. 1^ Fk;. 1. \'iew ot a small lake in the Cliautaiuiua Hills. Five miles southwest of Elk City, Montgomery County. -^-*J*i In.. 2. \\ ooded banks ol Fall river. Five miles west ot F^redonia, \\'ilson County. Flora and Ecology of Chautauqua Hills PLATE 13 201 Fig. 1. Woodland marsh showing a colony of Potamogeton gramineus var. ))iyri()))Jii/IIus in the left foreground. One of the insect catching plants, Utricularia gihha, forms a dense growth imder the water. Three miles east of Yates Center, Woodson County. m^'imm Kv Fk;. 2. Marsh in a low prauu' with a colony of Typlui hitifolia in iiic .uiddlc foreground. Two miles cast of Yates Center, Woodson Count\-. 202 The University Sc ience Bulletin PLATE 14 Fig. 1. Temporary water hole in a low woodland with a colony of fruiting Fotamogeton nodosus rising above the water. Three miles east of Busby, Elk County. 5*K ..* ".-■■ P Fig. 2. View of a prairie water hole .showing an extensive growth of Spirodela polyrhizxi and other water plants. Seven miles east of Neosho Falls, Woodson County. Flora and Ecology of Chautauqua Hills PLATE 15 2();- ..■'■* ^ it. ,' •j*', > Fig. 1. Upland wooded gully. Two miles northeast of Sedan, Chautauqua County. Fk;. 2. TemporarN w atcr hole in an upland j^ully. Two miles northeast of .Sedan, Chatitancpia Count> . 204 The University Scienc:e Bulletin PLATE 16 'M?«^ "^A*. !'> '^.^i.f • '■-■a -■ •*■ -•>,?» J i\. kri^^^'^ Fig. 1. Gully leading into a scrub oak woods. Two miles west ot Yates Center, \\'oods()n Count\'. Sk- v.A.K §^ •'^':^tv:'-^ Ns^^r-^ Fig. 2. Thf sauif gull> in tlie winter time I Flora and Ecology of Chautauqua Hills 205 PLATE 17 :«a«f«a Fk.. 1. Cliii al)()\t' a crt'ek showing sandstone outcrops and boulders. Six miles southwest of Yates Center, Woodson Countv. Fig. 2. Same cliff in the winter. 206 The University Science Bulletin PLATE 18 Fk;. 1. Open upland wooded slope showing suilaemg sandstone. Two miles northeast of Sedan, Chautau(]ua Count\-. ■■;'* , ■ • Fig. 2. Sandstone outcrops of a wooded gully. Four miles southwest of Yates Center, Woodson County. Flora and Ecology op^ Chautauqua Hills 207 PLATE 19 Kif;. 1. Stream in an npland wooded ravine. Eiylit miles north ot Sedan, Chautauqua Count>'. Fic;. 2. Wooded banks ot a lowland creek. Tvso miles south ot Toronto, Wood.son County. 208 The University Sc ience Bulletin PLATE 20 "^ i>y /:, 1 «.- l^HV^ -^^f^L^U ~ *"«! ■i^m S^^^ 1 *^ ''^^yjHhj i^' l^^^^H siRffYflm ^ F'«i^f! :J ■^1, ^'^'s^ ^--"■^ ..jiStf^- FiG. 1. Ruadsicle habitat of an upland woods. Tvvo niilis soutli, ii\r miles east of Toronto, Woodson County. Fig. 2. Prairie roadside habitat. Five miles southwest of Yates Center, Woodson Countv. Flora and Ecology of Chautauqua Hills 209 PLATE 21 bra*. y.^M':'; Fig. 1. Abandoned field showing weedy growth ot species ot Chenopudiutn, Ambrosia, and Lactuca. Two miles south of Toronto, Woodson County. wt M» Fig. 2. Pasture Held showing characteristics of waste places. Two miles nortli of Hale, Chautauqua County. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 5 A Subgeneric Revision of the Genus Osmia in the Western Hemisphere ( Hymenoptera: Megachilidae)^ BY Ranendra N. Sinha " Abstkaci . In this paper the species of Osmia found in the Western Henii- siihi-n* are placed in eleven subgenera and a key to the subgenera is given, to- gether with descriptions of them. The group conimonl>' placed in a separate genus, Diceratosmia, is included in Osmia. The following new subgeneric names are proposed: Chenosmia, Euthosmia, Trichimxsmia. From an exten- sive analysis of characters, it is concluded that in many features Diceratosmia is more primitive than other subgenera and that most of the subgenera nuist have arisen from Diceratosmia-\\\ic ancestors. INTRODUCTION This paper is a systematic revision, to the subgeneric level, of megachiline bees of the genus Osmia of the Western Hemisphere. The purpose of the study was to determine the natural groups and their limits and interrelationships within the genus. Of approximately 213 species ( appro.ximately 25,000 specimens) studied, 140 occur in the New World. The rest are from the Old World and only a small fraction of the latter belong within the genus Osmia as defined by American authors (see discussion of the Old World Species). On the basis of the American species, the genus has been divided into eleven subgeuera. In selecting the subgeneric characters, each of the specific characters used by previous authors, as well as new ones foimd in the present study, has been examined and evaluated with respect to all species. An effort has been made to detect variations (specific, subspecific and individual) of each character chosen. 1. Contribution number 978 from the Department of Entomologv' of the University of (Kansas, Lawrence. 2. Present address: Department of Zoology. McGill University. Montreal, Quebec, Canada. (211) 212 The University Science Bulletin Sandhouse ( 1939 ) gave a comprehensive subgeneric classification of American forms of the genus while both American and Old World subgenera were briefly treated by Michener (1941). Sand- house recognized six subgenera comprising 130 New World species. Nothosmia Ashmead, the largest of her subgenera, included 73 spe- cies. It was an artificial category containing a number of distinct groups of species. Other workers (Hurd and Michener, 1955:218) have recognized that a further revision was needed in order properly to treat the large number of species. I have distinguished si.x dif- ferent subgenera among the species classified under Nothosmia by Sandhouse. Three American species assigned to the Old World subgenus Melanosmia Schmiedeknecht by Sandhouse (1939:33) have been removed from that subgenus. Of these three species, O. bucephala is the type species of the re-established subgenus Cen- trosmia Robertson to which O. nigriventris also belongs and in which fall various species of Sandhouse's Nothosmia. The other species of Sandhouse's Melanosmia, O. inerm.s, has been transferred to the subgenus Chcnosmia, from typical members of which it differs only by nonmetallic coloration. The subgenus Melanosmia, thus, does not occur in the New World. The genus Diceratosmia of authors (Robertson, 1903:166, Mich- ener, 1949b: 258) has been reduced to a subgenus of Osmia because of the intergradation of characters found in O. (Nothosmia) mar- ginata and other species. Details regarding this change are dis- cussed elsewhere in this paper. DESCRIPTIVE METHODOLOGY The morphological terminology used is mainly from Michener (1944). Specialized expressions are explained as follows: The impunctate band of the clypeal margin is termed "broad" when half or more than half as wide as the maximum width of the scape, and "narrow", when less than half as wide as the maximum width of the scape. These measurements are taken in the lateral part of the clypeal margin, not in the center. Punctations of the clypeus, when described, are those of the upper part of the clypeus. The width of the eye includes the darker area on the outer side and the inner carina on the mesal side measured at the maximum width of the eye. The width of the genal area is also taken at the point of maxi- mum breadth. The measurements of the width of the eye and the genal area are taken from separate lateral views to give the maxi- Genus Osmia in Western Hemisphere 213 mum width of each. The mandible of the male bears two teeth, the upper and the lower. In the female there are three or four teeth. These are numbered in the following order: tooth 1, the uppermost; tooth 2; tooth 3; tooth 4, the lowermost. The second tooth is often small, rudimentary, almost confluent with the first tooth, or even absent in certain subgenera. When this tooth is absent, the tooth following the first tooth is termed the third tooth. The distances be- tween the first tooth and the third and between the third and the fourth teeth are taken from the tip of one tooth to the tip of the other. A few of the mandibular drawings, such as Osmio fi,laiica, are based on worn mandibles. In such cases the supposed lengths of the original teeth have been shown by dotted lines. The inner anterior angle of the hypostomal carina is often broadly rounded. In the descriptions of the carina the expression "at the angle" means the hypothetical apex of the angle that would be made by extending the straight portions of the carina and "behind angle" means anywhere posterior to the apex (real or imaginary) of the angle. In the forewing of the male vein A is that part of the vein M between the first transverse cubital and the first recurrent vein, and vein B is that part of vein M between the second recurrent and the second transverse cubital. The width of the hind basitarsus is meas- ured at the point of the maximum width. The first metasomal ter- gum is the second abdominal segment, i. c, the one posterior to the propodeum; other terga and sterna are numbered accordingly. In the figures the characters of one or two representative species are usually drawn for each subgenus. Also, certain characters have been drawn for a few species which differ from the representative species. In subgeneric descriptions, when a particular character is found to be unique for a group and not to occur in any other subgenus, it is italicized and not mentioned in the descriptions of the other subgenera. In the synonymies, references to the following works have been omitted: Dalla Torre ( 1896), Michener ( 1951, in Muesebeck, et al. ) and Sandhouse (1943). The drawings are all diagrams; they omit many details of no im- portance at the subgeneric level. It should be pointed out that Sandhouse (1939) gives many excellent drawings, for example, of genitalia and sterna of males, which should be used in conjunction with the present work. 214 The University Science Bulletin PHYLOGENY In the discussion of relationships that follows, the words "primi- tive" and "derived" or "modified" are used to describe the condition of the morphological features. The words "generalized" and "spe- cialized" are used to describe the condition of species or groups of species judged by an average of all of the features studied. For instance, a subgenus is generalized if it has a large number of prim- itive characters and specialized if it has a large nmnber of derived characters. Table I was prepared to contrast primitive and derived characters useful among Osniia species, primitiveness being judged by simi- larity of the character to features found among other genera and among families of more wasplike bees. Often two or more derived conditions exist in relation to a single primitive one. In postulating the phylogenetic arrangement of the genera of OsDiiini,'- Michener (1941:151) rightly assiuned that those forms having characteristics most nearly similar to the various short- tongued families of bees and to wasps were the most primitive. On the basis of this assumption, Chelostoma and Prochelostoma were considered to approach most nearly the primitive type because they have the longest parapsidal lines found in the Osmiini, an elongate thorax with the anterior part of the propodeum horizontal, and large pterostigmata. By contrast, Ostnia, having punctiform parap- sidal lines, is one of the most specialized genera of the group. Vari- ous other characters support this conclusion. Diceratosmia (a genus according to his classification) has the shortest parapsidal lines of all the more generalized genera and thus was placed imme- diately below Osmia in the phylogenetic tree. An examination has revealed extremely short parapsidal lines in Osmia marginata Michener, a desert species of Nothosniia. This species seems to be a connecting link between the generalized Diceratosmia and the slightly more specialized Nothosmia. All species of Diceratosmia possess a distinct carina along the inner ventral angle of each posterior coxa. Such a carina is represented in marginata and O. (Chenosmia) catdicola in the form of a weak line. This carina prob- ably indicates primitiveness within the genus Osmia. It exists in stronger form in some of the most generalized genera of the Osmiini such as Prochelostoma, Heriades, Noteriades, and Proteriades. The 2. The name Osmiini is not used in recent classifications (Michener, 1944:256-268; Hiird and Michener, 1955:6-10) in which the groups are included in the tribe Megachilini. Howe\er, the name denotes a natural assemblage of nonparasitic megachilids having arolia. Genus Osmia in Western Hemisphere 215 Table I. — Primitive and Derived Characters in the genus Osmia Primitive Size small Pubescence white or pale, of moderate length Mandible of female with three teeth Mandible of male with upper tooth right angular or acute Mandibular constriction moderately wide Mandible of female lacking transverse ridge or protuberance near base Malar space absent Genal area of female lacking depres- sion immediately below eye Genal area of female equal to width of eye Inner margins of eyes converging an- teriorly Clypeus of female lacking swollen or modified apical margin Clypeal punctations of female nearly confluent Clypeal margin of male having im- punctate band of moderate width llypostomal carina of female low Antennal socket having margin uni- form throughout Flagellar segments in male umodified (cylindrical, about 1.5 times as long as wide) Hind coxa with a \'entral carina Mediotarsal segments of middle leg of male not swollen Strigilis of male having malar spine not greatly produced Basitarsus of hind leg of male lacking tooth Forewing with hairs of cells short Middle femur of male lacking a pro- jection on lower side Hind basitarsus of male parallel-sided Metasomal terga with moderately wide impunctate bands Metasomal sternum 2 of male lacking posterior tooth Derived Size large Pubescence dark or black, long or short Mandible of female with four teeth Mandible of male with upper tooth obtuse Mandibular constriction narrow, or absent and entire mandible broad Mandible of female having transverse basal ridge or protuberance Malar space present Genal area of female with a roimd or elongate depression immediately below e\e Genal area of female significantly wider or narrower * than e> c Inner margins of eyes parallel or di- verging anteriorly Clypeus of female having apical margin swollen or otherwise modi- fied Clypeal punctations of female widely separated or confluent Clypeal margin of male without or with narrow impunctate band H>postomal carina of female high, abruptly reduced behind angle, forming a tooth Antennal socket having superior-mesal margin higher than remaining mar- gin Flagellar segments in male modified ( moniliform, or last segment com- pressed and expanded, or each seg- ment twice as long as wide) Hind coxa without a ventral carina Mediotarsal segments of middle leg of male greatly swollen Strigilis of male having malar spine greatly produced Basitarsus of hind leg of male having tooth on inner side Forewing with hairs of cells long Middle femur of male ha\ing a jiro- jection on lower side Hind basitarsus of male modified (wider towards base, at apex, or in middle ) Metasomal terga without or with nar- row or with wide impunctate bands Metasomal sternum 2 of male having posterior tooth Possibly the primitive condition. 216 The Universiti' Science Bulletin Table I. — Primitive and Derived Characters in the genus O.tmia (Concluded) Primitive Derived Gonocoxite of male having weak sub- Gonocoxitt- of male lacking subapical apical swelling swelling Subapical hairs on gonocoxite of male Subapical hairs on gonocoxite of male diffuse forming a tuft Gonocoxite of male lacking a subapical Gonocoxite of male having a sub- process '»pical process Penis valve of male lacking long ven- Penis valve of male having long, ven- tral longitudinal groove tral longitudinal groove carina is absent in HopUtis, but manifests itself in all forms of inter- gradation from a fairly conspicuous carina to a feeble line in the genus Anthocopa, which is comparatively less primitive in other characters than the previously mentioned genera. O. (Diceratosmia) suhmicans Morawitz (southern Europe, north Africa, etc.) strik- ingly resembles O. (Nothosmia) marginata from California in having extremely short parapsidal lines and weak or evanescent carinae on the hind coxae. Diceratosmia is judged to be the most general- ized group of the genus Osmia and Nothosmia is a close relative which is also generalized and probably derived from the former earlv in the historv of the genus. In considering relations among subgenera, identical derived char- acters in some but not all species of each of two subgenera are re- garded as an indication of probable relationship between the two subgenera concerned, except in instances where it is judged that there has been parallel evolution. Diceratosmia retains twenty-eight of the twenty-nine primitive characters listed in Table I. The only modification is the quadri- dentate condition of the mandible of the female. Nothosmia retains twenty-seven primitive characters and is specialized in two, namely, in having a tooth on the hind basitarsus in some males and in having the impunctate apical bands on the metasomal terga of the male narrow."* A common tendency towards mandibular constric- tion among the females of a few species of each subgenus supports the idea of close relationship of these two generalized subgenera. The third most generalized subgenus is C}iem)smia; it retains the same number of primitive characters as Nothosmia, but shares with Diceratosmia the deri\ed quadridentate condition of the mandible in the female. The presence of a tooth on the hind basitarsus of the males in Nothosmia and the absence of such a tooth in some species of Chenosmia, together with the modification of the apical areas of 3. This and the following discussion suggests that there are few subgeneric characters. As the systematic section shows, there are some other characters which could not be in- cluded in Table I because of lack of information as to what alternatives are primitive. Genus Osmia in Western Hemisphere 217 the terga (impunctate bands narrow or absent) in Nothosmia sug- gest that Chenosmia did not arise from Nothosmia. CJienosmia probably arose from some Diceratosniia-\ike form near the point of divergence of Nothosmia. The fourth generahzed subgenus is Etithosmia; it retains twenty- five primitive characters of the twenty-nine Hsted in Table I and is specialized in four characters, namely, the sparseness of punctations on the clypeus of the female, greatly produced malar spine of che strigilis, modified hind basitarsus of the male, and absence of the carina on the hind coxa. In the first two characters, some species of Diceratosmia are likewise specialized. In the last two characters some species of Nothosmia are similarly specialized. These special- izations bring EutJiosmia equally close to Diceratosmia and to Nothosmia. Euthosmia was probably derived from Diceratosmia independently of other subgenera, Acanthosmioides has twelve derived characters, and seems to have arisen from N othosmia rather than from Diceratosmia (Dicera- tosmia differs from Acanthosmioides in thirteen characters) or Che- nosmia (Chenosmia differs from Acanthosmioides in eleven charac- ters). From Nothosmia, Acanthosmioides differs in only nine char- acters, and there are no modifications in Notlwsinia that indicate that Acanthosmioides was not derived from that subgenus. More- over, the tooth on the hind basitarsus of the male, which in itself is a significant modification for a generalized group like Nothosmia, occurs in Acanthosmioides as well. Chalcosmia and Centrosmia ha\'e an equal number of primitive characters (see table II) and seem to be closely related to each other. Each has the same modifications, and in addition in vome species of each subgenus the inner margins of the eyes dixerge anteriorly. Because of these common characteristics, the two sub- genera seem to be more closely related than either subgenus is to any other. Centrosmia is the more specialized of the two, as indi- cated by the modification of at least four structures ( in some of its species ) in which the species of CJialcosmia are primitive. Also, a few primitive characters of Chalcosmia, such as the abundance of pale-white or yellowish hairs on the body and the simple medio- tarsal segments of the middle leg of the male, are probably qualita- tively even more significant than the derived characters in relating the subgenus to the generalized subgenera, such as Diceratosmia and Nothosmia. Centrosmia and Chalcosmia have more characters in common than either has witli Nothosmia, EutJiosmia or Chenosmia. The generalized subgenus Diceratosmia could have given rise to 218 The University Science Bulletin Table II. — Number of primitive and derived alternatives of twenty-nine char- acters among subgenera of Osmia Subgenus Alternatives Primitive Partly derived* Derived Diceratosrnia 22 21 23 24 17 11 13 16 13 18 13 6 6 4 1 2 8 4 6 7 0 4 1 Xothosniia 2 Chciiosntia 2 Kuihosniia 4 Chalrosniia 10 ('rrttrosntia 10 Acdnthosniioides Monilosniid 12 7 Cephnlosmia 9 Ti ichinosmia 11 Osmia s. sir 12 * Partly derived means specialized in some species, not in others. ChaJcosniia. Howe\er, the narrowness or absence of the apical impunctate hands on the metasomal terga of Nothosjnia is a modi- fication not present in Chalcosmia. Modification of structnres in Euthosniia indicate that it was not ancestral to Chalcosmia although the presence of well-separated piuictations on the clypeus of the female of ChaJcosmia is a modification suggesting Euthosmia. There- fore, it is logical to suppose that the common ancestor of Chalcosmia and Centrosmia arose from Diceratosmia (or the Diceratosrnia stock) somewhere near the point of divergence of Euthosmia from the Diceratosmia stock. This ancestral line, in turn, split into Chalcosmia and Centrosmia. In all of its derived characters {e. ^., the abundance of black pubescence on the body, the quadridentate mandible of the female, and the absence of a carina on each posterior hind coxa), the sub- genus Chenosmia closely resembles the more specialized subgenus Monilosmia. The direction of specialization of some species of Chenosmia, such as from small to large size and from low to high hypostomal carinae in the female, indicates possible relations of Chenosmia to Monilosmia, in which these characters are strongly developed. Moreover, MonilostJiia shares sixteen primitive char- acters with Chenosmia, thirteen with Nothosmia and only twelve with Diceratosmia. Monilosmia shares no derived character (in all its species) with Diceratosmia or with Nothosmia. The close- ness of the relationship between Monilosmia and Chenosmia is further attested by the males of those two subgenera being almost indistinguishable. It is concluded that Monilosmia arose from Genus Osmia in Western Hemisphere 219 Chenosmia which, as previously stated, probably arose from the Diceratos'inia line early in its history. The subgenus Osmia s. sir. preserves the minimum number of primitive characters. The absence of a tooth on the hind basitarsus of the male in Osmia is a primitive condition whereas Nothosmia is modified in this respect. At least some species of Diceratosmia and Chenosmia have this modification. The directions of modifica- tion within the subgenus are unexpected if judged by the characters of Nothosmia, Diceratosmia, Chenosmia or Euthosmia. Certain modifications (already discussed) of each of the other subgenera, except Diceratosmia and Chenosmia, similarly seem to pre\ent any one of them being considered as ancestral stock for the siibgenus Osmia. It resembles Chenosmia in dark hair on the body, black scopa in the female and the quadridentate mandibles of the female (the latter condition is a modification also in the female of Dicera- tosmia). The broadened apical impunctate band on the clypeus of the male in some species of both Chenosmia and Osmia s. str. also suggests a possible relationship between these two subgenera. It is impossible to judge from these facts which, if any, of the generalized subgenera is directly ancestral to the subgenus Osmia. In addition, it was found impossible adequately to compare the subgenus Osmia with certain of other more specialized subgenera, namely Centrosmia, Cephalosmia and Acanthosmioides, because it was not possible logically to assume that any one of the three is either more generalized or more specialized than the subgenus Osmia. All that can be said is that Osmia possibly arose from Diceratosmia or Chenosmia somewhere near the point of divergence of these two subgenera. More probably the subgenus Osmia arose from an extinct ancestor of Dicerato-wiia and Chenosmia. Trichinosmia is another subgenus of uncertain taxonomic posi- tion. With Chenosmia, Trichinosmia shares two modifications and with Nothosmia one. TricJiinosmia is monotypic, and therefore, it cannot have some species with derived characters and some species with primiti\e characters. As in Osmia s. str., the modifica- tions seem to indicate relation to either Diceratosmia or Chenosmia instead of to any one of the specialized subgenera. Cepfialosmia preserved, among some species, more primitive characters than either Chalcosmia or Centrosmia (see table II) and shares more primitive characters with Euthosmia than with any other generalized subgenus. The derived characters so shared are the well-separated punctations on the cKpeus of the female and the long malar spine of the strigilis of the male. Cephalosmia re- 220 The UNn^ERSixY Science Bulletin senibles Chcnosmia in fourteen characters of which two are de- ri\"ed (in which two derived characters Cephalosmia also resembles Monilosmia). Considering the characters in which only some species are specialized, Cephalosmia is equally related to Chenosmia, Di- ceratosmia and Euihosmia. But there is no modification in Euthos- mia that indicates that Cephalosmia was not derived from that subgenus (such modifications occur in Diceratosmia and Chenos- mia). Thus Euthosmia is perhaps ancestral to Cephalosmia. Considering the problem of phylogeny, not from the standpoint of relative specialization but from that of intergradations between groups, we can note that Dieeratosmia almost grades into Nothosmia and Notliosmia into Acanthosmioides; Nothosmia also almost grades into Chenosmia and Chenosmia into Monilosmia and Monilosmia into Centrosmia. Except for Diceratosmia this group of subgenera is almost exclusively American and it seems likely that Trichinosmia and Euthosmia and perhaps Cephalosmia are part of the same assemblage. The subgenera Chalcosmia and Osmia s. str. are more distinct, being separated by large gaps from each other and from any other subgenera. Both are primarily Eurasian with but few American species. It will be noted that the most striking difference between the relationships indicated in the preceding paragraph and those out- lined in earlier paragraphs concern Centrosmia, which was earlier associated with Chalcosmia and here with MoJiilosmia. It seems likely that the latter relationship is more correct. Parallelisms Numerous examples of parallel evolution have been observed in the independent acquisitions of similar characters among related subgenera. The most striking instances are: 1. The compressed last segment of the flagellum in the males of some species of Acanthosmioides and Centrosmia. 2. The thickened clypeal truncation in the females of Centrosmia and Monilosmia (some species). 3. The mesally directed, long and stiff hairs on the hypostomal area of the female in Eiitliosmia and Monilosmia (some species). 4. The sharp tooth near the angle of the hypostomal carina in the females of Monilosmia, Chalcosmia, etc. 5. The parallel or anteriorly dixerging eyes in the females of some species of Osmia, Cephalosmia, Acanthosmioides, Clmlcosmia and Centrosmia. Genus Osmia in Western Hemisphere 221 6. The absence or narrowness of the impunctate bands on the metasomal terga in the males of Trichinosmki, Osmia, NutJiostiiia and EittJiosmia (some species). 7. The absence of the mid-apical emargination of the sixth metasomal tergum in the Chalcosmia-Centr osmia stock, in Acan- tho&mioides and in Osmia. 8. A subapical process on the gonocoxite of the males of Acan- thosmioides (most species) and Osmia (one species). Michener ( 1949c: 140-141) has discussed the possible reasons for parallelisms in evolution; these reasons may apply equally well in the evolution of Osmia. DISTRIBUTION The genus Osmia is known from the Palearctic, Nearctic and Ethiopian regions. It is absent from the Indo-Australian and Neo- tropical regions. The subgenus Dicerafosmia is known from south- ern Europe, Asia Minor, North Africa and probably ranges widely across Asia. In the New World this subgenus is widespread in the southern United States and extends southward to Costa Rica. This extensive distribution seems compatible with the generalized nature of the subgenus. Diceratosmia is found principally in the warmer parts of the north temperate regions. Hence it likely arose in such climates, perhaps when they extended into northern regions now much cooler. Perhaps it has been displaced from cooler cli- mates by competition with more specialized subgenera, all of which are more boreal (except T richinosmia, of which there is only one species). This is in agreement with Matthew's (1915) hypothesis that the most advanced groups (he was dealing mainly with higher taxa) should be nearest to the center of dispersal, and the most conservative groups farthest from it. Because the center of dis- persal was assumed to be holoarctic, the generalized forms would be expected to be of southern occurrence now. No species of Nothosrnia is known from the Old World. In North America that subgenus occurs from coast to coast but not in cold, montane climates of the west. Species of group I of the subgenus are restricted to the eastern and central parts of North America. Species of group II occur in the desert areas from Cali- fornia to New Mexico. AcantJiosmioides, a specialized subgenus probably derived from Nothosmia, is absent in the Old World and is restricted to mountainous and boreal regions of western North America. Females of Acanthosmioidcs closely resemble the species 222 The University Science Bulletin group II of NotJiosniia. White (1952) included five of the six species of group II of Nothosmia as an aberrant group of Aconthos- mioidcs, not having seen the males of any of the species. In the New World, Diceratosmia probably gave rise to N othosmia which in turn split into two groups; the predominantly eastern group I re- mained somewhat generalized whereas the western group II gave rise to the specialized subgenus Acanthosmioides. Perhaps group II of Notlwsmia has become restricted to the southern deserts be- cause that group could no longer survive elsewhere after the more specialized subgenera arose in country of the kind usually occupied by Osniia. This is in accordance with the view that, next to Dicera- tosmia, NotJwsmia is the most generalized subgenus. Chenosmia is distributed from coast to coast in North America but most of the species are in the mountains or boreal regions of the western half of the continent. Monilosmia, a closely related stock, is restricted to the western half of North America except for one species which ranges eastward to the Atlantic. The inclusion of Osmia inermis extends the known distribution of Chenosmia to Eurasia. Although there are many American species of Chenosmia, only this one Eurasian species is known to me; it is not especially generalized nor ISJ othosmia-\\Ve . It is a species of northern distri- bution and it is likely that it migrated to Eurasia from America dur- ing Pleistocene or post-Pleistocene time. Monilosmia is probably an independent group which arose from Chenosmia and remained restricted primarily to the mountainous regions of the western North America. Chalcosmia is principally Eurasian, having many more species and more diversity there than in America; only one species group (plus one probably introduced European species) occurs in Amer- ica. The comparatively more specialized subgenus Centrosmia is principally North American and probably originated in this con- tinent, but one species, O. nigriventris, occurs in boreal parts of both America and Eurasia. Very likely it extended its range from America to Eurasia in Pleistocene or post-Pleistocene time. Of the highly specialized subgenera, Osmia probably is of Old World origin as it has a larger number of species and exhibits more diversity there; only two species occur in North America. Cephal- osmia is principally from high altitudes of the western half of North America and is unknown in the Old World. Trichinosmia is known from Arizona, Nevada and California in semideserts and in forested areas of low elevation. Genus Osmia in Western Hemisphere 223 FLOWER RELATIONSHIPS In the literature and on the hibels of specimens in collections, there are many records of flowers visited. Unfortunately most of these records do not distinguish between pollen collecting and nectiir gathering visits, and therefore merely show that most species of Osmia can be collected from a wide variety of flowers. Accord- ing to Michener (in litt.) females of Clialcosmia and Cephalosmia use pollen from Compositae in provisioning their nests. The re- maining subgenera use pollen mostly from other families of plants. The subgenus Osmia, and perhaps Notliosmia, uses pollen from many families, perhaps with a preference for Leguminosae. Other subgenera utilize pollen principally from Scrophulariaceae, Legu- minosae and Hydrophyllaceae. No American species of Osmia is known to be restricted in pollen collecting to one genus of plants. OSMIA OF THE OLD WORLD Ducke (1899:12-19) listed 266 species of the genus Osmia from Eurasia and Africa, and Friese ( 1911:35-160) lists 269 from the same continents. Numerous additional forms have been named since that time. Sandhouse (1939:1) and Michener (1951:1162-1172) recorded about 130 species from the Nearctic Region. This flgiue excludes the very numerous species of Ashmcadiella, Proteriades, Anthocopa and Hoplitis that would have been included in Osmia by European students of these bees. No systematic attempt has yet been made to correlate the fauna of the Old World with that of the New World with a view to presenting an all round picture of this large and widely distributed group. Michener (1941:147) has pointed out the complexity of the situation in a generic revision of the American Osmiinae, after studying about 60 Old World spe- cies, representing most of the named genera and subgenera. "In their treatment of the genera related to Osmia, European and Amer- ican students have followed widely different paths. The European workers have placed the entire group in a single genus, Osmia, di- vided into a number of subgenera, some of which were unnatural, and to some of which untenable names were applied. American authors, on the other hand, following the example of Cresson, have usually recognized a considerable number of genera, especially among the forms related to Hoplitis." In his treatment, Michener followed a middle ground, not recognizing the fine divisions of Cresson at the generic level but dividing the unwieldy Osmia of European authors into a few major generic units. 224 The University Science Bulletin Michener (1941:151), in considering the possible phylogenetic relationships, has recognized three major Old World genera, Osmia, Anthocopa, and Hoplitis, each with a number of subgenera. Os77iia is the most distinct of the three genera and is easily recognized by the punctiform parapsidal lines. For reasons mentioned elsewhere in this paper, the previously recognized genus Diceratosmia is treated in the present study as a subgenus of the genus Osmia proper. It is important to mention that a number of the Old World spe- cies of Osmia exist which cannot be satisfactorily classified under any of the three large genera recognized by Michener (1941:159). Perhaps some of these species have no close relatives in the New World and need to be grouped under separate genera, as suggested by Michener (1941:165) when he erected the genus Sfenosmia. In order to facilitate future work I have listed all the Old World species examined by me in three different groups below. Group I includes the species confined to the Old World which can be satis- factorily classified under genus Osmia in the sense of the present paper. Group II includes the species of Osmia which are common to both the New and the Old World. Group III includes the rest of the Old World species studied. Some are Hoplitis, some are Antho- copa, and some are questionable as to the genus to which they be- long. Many of these species were placed by Michener (1941). Group I: Osmia auridcnta (Panzer), cornuta Latreille, decemsig- nata (Radoszkowski), dimidiata Morawitz, cmarginata Lepeletier, fidv'.ventris (Panzer), gallarum Spinola, kohlii Ducke, latrcillei Spinola, leaiana Kirby, Jeucogastra Morawitz, kmgicornis Morawitz, mcdamie Magretti, notata (Fabricius), ruhicola Friese, rufa Lin- naeus, rufigastra Lepeletier, submicans Morawitz, taurus Smith, tri- cornis Latreille, versicolor Latreille, vidua Gerstaecker. Group II: Os7nia coerulescens (Linnaeus), inermis ( Zetterstedt ) , and nigriventris (Zetterstedt). Group III: Osmia adunca (Panzer), anthrcnoides Spinola, atro- caeridea Schilling, halearica Schmiedeknecht, halucha Nurse, hi- color (Schrank), hidemlata Morawitz, calviniae Cockerell, cepha- lotes Morawitz, claviventris Thomson, cremdata Morawitz, croatica Friese, difformis Perez, ferruginca Latreille, fiavicornis Morawitz, glutinosa Giraud, gracilicornis Perez, iacoti Cockerell, jheringii Ducke, leucomelaena (Kirby), Ihotellreiei Perez, ligiirica Morawitz, macroglossa Gerstaecker, maritima Friese, mitis Nylander, monili- ferum Cockerell, morawitzii Gerstaecker, panzeri Morawitz, papa- Genus Osmia in Western Hemisphere 225 veris (Latreille), paradoxa Friese, parviila Diifoiir and Ferris, pili- cornis Smith, quodridentata Perez, ruficoUis Dours, rufohirta La- treille, satindersii Vachal, scutellaris Morawitz, sitigtdaris Moravvitz, spinolae Schenck, spintdosa (Kirby), transcaspica Morawitz, tri- dentata Dufour and Ferris, tuberculata Nylander, imcinata Ger- staecker, and vdlosa Schenck. SYSTEMATIC TREATMENT The genus can be recognized by the following characters: Body robust, coloration metallic blue to metallic green (dull and non- metallic in a few species). Farapsidal lines reduced, punctiform or very short, each represented by a small, shining, often elevated area, sometimes no larger than one or two punctures, sometimes no more conspicuous than other unusually large interspaces between punctures, although usually definitely different from, and larger than, the latter; metanotum and the propodeum vertical or nearly so, propodeum without horizontal basal zone. First metasomal tergum with anterior face broadly concave, with a distinct line or carina separating this concavity from dorsal surface of tergum; male with seven exposed metasomal terga and tergum six not dentate laterally. Key to the Subgenera of Osmia in America Since several of the subgenera are most strikingly different from their relatives because of combinations of male and female char- acteristics, difficulties will be found in using the following key in which the sexes are separated. Most of the difficulties are indicated by qualifying words or by mention of exceptions, but caution must be used in placing a species, especially if it is available in only one sex. Males 1. Posterior coxa having a distinct carina on inner ventral angle; mesoscutum with parapsidal line short linear, not punctiform, Diceratosmia Posterior coxa lacking a distinct carina on inner ventral angle; mesoscutum with parapsidal line punctiform or nearly so 2 2. ( 1 ) Metasomal sternum 2 with apical margin truncate or subtruncate, so that median length is not much greater than length measured laterally 3 Metasomal sternum 2 with apical margin convex, largely covering sternum 3; median length greater than length measured lat- erally 4 8—2656 226 The University Science Bulletin 3. (2) Genal area wider than greatest width of eye; pubescence white and black intermixed; hypostomal carina high; antennal socket with upper mesal margin more strongly developed than rest of margin Cepluilosmia Genal area narrower than greatest width of eye; pubescence white; hypostomal carina low; antennal socket with margin uniformly developed throughout Etithosmia 4. (2) Malar space distinct; middle leg having tarsal segments not swol- len, middle femur having weak or strong projection on lower surface Osmia s. str. Malar space absent or narrow, wide only in bucephala; middle leg having tarsal segments swollen or not; middle femur lack- ing projection on lower margin 5 5. (4) Metasomal sternum 4 with apical margin thickened, transversely sulcate in lateral portions. Metasomal tcrga 1 to 5 each having a subapical band of plumose hair Chalcosinia Metasomal sternum 4 with apical margin not modified as above. Metasomal terga 1 to 5 lacking distinct subapical bands of plumose hair 6 6. (5) Middle leg with tarsal segments 2 and 3 greatly swollen (except in bukeri and nigriventris) ; metasomal tergum 6 without or with very faint and undefined emargination (except in nigri- ventris which is black); flagellar segment 11 unmodified (ex- cept in vandtjkei which has a median basal tuft of long hairs on metasomal tergum 6) Centrosmui Middle leg with tarsal segments 2 and 3 not swollen; if swollen, antennal segments modified; metasomal tergum 6 having shal- low or deep midapical emargination, usually delimited by distinct angles 7 7. (6) Metasomal sternum 2 with protuberance or tooth on median apical or subapical area (except in O. Integra which has me- dian band of unusually long hair on this sternum); hind basi- tarsus usually narrow basally and widened distally, Acanthosmioides Metasomal sternum 2 lacking protuberance or tooth on median apical or subapical area; hind basitarsus variable 8 8. (7) Impunctate bands on terga 1 to 3 narrow or absent, usually not depressed (distinct and depressed in certain eastern species of Nothosmia); pubescence usually entirely white 9 Impunctate bands on terga 1 to 3 wide and distinct, depressed (in seclusa, bands punctate but broadly depressed); pubes- cence often partially black 10 9. (8) Flagellar segments 2 to 11 each twice as long as wide; forewing with hairs about half as long as width of ptcrostigma; hind basitarsus lacking a tooth; metasomal terga 2 to 5 with hairs simple, erect; head and thorax with long-branched hairs, Trichinosmia Flagellar segments 2 to 11 less than twice as long as wide; fore- wings with hairs much less than half as long as width of Genus Osmia in Western Hemisphere 227 pterostigma; hind basitarsus with a tooth; metasomal terga 2 to 5 with some plumose hairs and usually with apressed hairs at least near posterior margins of terga; head and thorax with short-branched hairs Nothosmia 10. (8) Metasomal tergum 7 with deep midapical emargination; strigilis with malar spine long; hind leg with tibial spurs curved api- cally; size usually about 10 mm.; color dark blue, often having greenish tinge Monilosmia Metasomal tergum 7 with shallow midapical emargination; strigilis with malar spine short; hind leg with tibial spurs not curved or weakly curved apically; size usually less than 10 mm.; color metallic green or dark green Chenosmia Females 1. Posterior leg having a distinct carina on inner ventral angle of each hind coxa; mesoscutum with parapsidal line short linear, not punctiform Diceratosmia Posterior leg lacking distinct carina on inner ventral angle of each hind coxa; mesascutum with parapsidal line punctiform, 2 2. (1) Mandible having four distinct teeth (three in zephtjros; second tooth much reduced in cerasi) 3 Mandible three-toothed, having first, third, and fourth teeth, or, if a small second tooth is present, either the pubescence is all white or the mandible is slender medially, two-thirds or one- half apical width 9 3. (2) Malar space wide; a round or elongate depression in genal area immediately below lower posterior angle of eye. Osmia s. str. Malar space in most species narrow or absent, or if wide, having clypeal margin greatly swollen; depression in genal area im- mediately below lower posterior angle of eye absent 4 4. (3) Clypeal tnmcation tridentate with one median and two lateral teeth; hair on head and thorax with long branches; hairs on forewing long, two-thirds as long as width of pterostigma, Trichinosm^ia Clypeal truncation usually simple, if modified, not as above; hair on head and thorax with short branches; hairs on forewing short, less than half as long as width of pterostigma, dense or sparse ,. 5 5. (4) Clypeus with punctations separated by smooth ground or some- times largely absent; antcnnal socket with upper mesal margin more strongly developed than rest of margin; metasomal ter- gum 6 with apical margin not produced to form a flange (ex- cept in calif arnica) Cephalosmia Clypeus with punctations confluent at least on anterior half; an- tennal socket with margin unifomily developed; metasomal tergum 6 with apical margin produced forming a flange 6 6. (5) Mandible with protuberance or transverse ridge on outer surface near base; clypeus with apical margin thickened or with two median sockets just beneath margin and mesal to usual hair tufts, each giving rise to a tuft of hairs 7 228 The University Science Bulletin Mandible without transverse ridge or protuberance near base except in sechisa and rostrata which have a ridge; clypeus with apical margin not thickened (except in rostrata) nor with two median sockets each giving rise to a tuft of hairs 8 7. (6) Clypeal margin with two small median sockets each giving rise to a small tuft of hairs in addition to two large lateral tufts ; metasomal terga 2 to 5 each with a subapical band of plumose hairs, without apical impunctate band Chalcasmia Clypeal margin swollen, without two median sockets each giving rise to a tuft of hair in addition to two large lateral tufts; meta- somal terga 2 to 5 without subapical bands of plumose hairs, with apical impunctate bands Centrosmia 8. (6) Hypostomal carina high, abruptly reduced near angle, forming a tooth; genal area wider than eye; clypeal truncation usually longer than margin from end of truncation to lateral angle of clypeus Monilosniia Hypostomal carina low, not abruptly reduced near angle; genal area as wide as eye; clypeal truncation usually equal to margin from end of truncation to lateral angle of clypeus Chenosmia 9. (2) Tufts of orange hairs beneath clypeal margin absent; strigilis with inner concavity of velum deep, malar spine slender, acuminate; scopal hairs restricted to a subapical band, half as wide as distance from gradulus to sternal margin, on each sternum, Euthosmia Two tufts of orange hairs beneath clypeal margin; strigilis with inner concavity of velum shallow, malar spine greatly widened at base; scopal hairs not restricted to a subapical band on each sternum 10 10. (9) Scopa white to testaceous (black and white intermixed in mar- ginata, black in liogastra); metasomal terga 1 to 3 each with or without a narrow apical impunctate band Nothosmia Scopa black; metasomal terga 1 to 3 each with wide apical im- punctate band (except sedula which has much darker body with more black hairs than Nothosmia) Acanthostnioides Subgenus Diceratosmia Robertson Diceratosmia Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29, pp. 166, 171; Cockerell, 1912, Proc. U. S. Nat. Mus., vol. 42, p. 216; Sandhouse, 1939, Mem. Ent. Soc. Washington, vol. 1, p. 1; Michener, 1941, Amer. Midland Nat., vol. 26, p. 162; Michener, 1949, Ann. Ent. Soc. Amer vol. 42, pp. 258. Type species: Osmia quadridentata Cresson, 1878 (not PhyUotom-a qiiadri- dentata Dumt'ril ) = Osmia conjuncta Cresson, 1874, original designation. This subgenus can be distinguished from all others by the longi- tudinal carina along the inner ventral angle of each posterior coxa and by the slightly elongate parapsidal lines of the mesoscutum. These characters are, as indicated elsewhere, weakly developed in an occasional Nothostnia. Male. — Length of body 6 to 10 mm. Pubescence white; meta- somal terga with subapical fasciae. P.unctation on general body Genus Osmia in Western Hemisphere 229 surface coarse in most American forms. Mandible widest at base, narrower at region of constriction which is nearly as wide as apex, with V-shaped notch separating lower tooth from subtruncate to oblique, acute upper tooth; eye wider than genal area; inner mar- gins of eyes parallel or weakly converging anteriorly; clypeus thickly covered with white hairs arising from coarse and confluent punc- tures, apex weakly wavy, arcuately emarginate, or with median third projecting, with narrow or moderately wide impunctate apical band; scape longer than second to fourth antennal segments com- bined; flagellar segments 2 to 11 unmodified, each less than twice as long as wide; hypostomal carina low, uniform throughout. Middle leg with tarsal segments not swollen; parapsidal lines short linear; posterior coxa with carina along inner ventral angle; basitarsus of hind leg about five times as long as wide, not toothed, parallel sided; strigilis with apical margin of velum truncate or obliquely truncate, inner apical angle of velum narrowly rounded, concavity on inner margin of velum shallow, apex of mains produced into a long and moderately narrow spine. Forewing cells with short, moderately sparse hair, distal portion finely papillate. Metasomal terga 2 to 5 without impunctate bands (except narrow bands on posterior terga of coniiincta) ; apical margin of tergum 6 with greatly widened emargination, median third of base of emargination produced, with or without emargination or subtruncation at apex of production, lateral portion turned up, forming a flange; tergum 7 with apex produced, midapex with a deep emargination, forming two lobes or teeth. Metasomal sternum 1 broadly rounded with shallow midapical emargination; sternum 2 with apical margin produced and broadly rounded, covering most of sternum 3, apex subtruncate or with a shallow emargination; sternum 3 with a wide, deep emar- gination; sternum 4 with apical margin broadly convex; sternum 6 with a median apical produced portion the apex of which has a shallow, wide emargination; sternum 8 with apical half broadly convex or narrowly triangular. Gonocoxite broad subapically, abruptly narrowed to a short process. Female. — Length of body 7.5 to 12 mm. Pubescence long or short, branching of hair close and short, white or reddish brown; metasomal terga with subapical fasciae; scopa white or brownish. Mandible moderately broad, constricted distad of base, gradually widened to apex, which is oblique and a little less than twice as wide as constriction, lacking transverse ridge or protuberance at base, with three or four teeth, third and fourth teeth triangular; 230 The University Science Bulletin distance between teeth 1 and 3 equal to or longer than that between teeth 3 and 4; eye about as wide as genal area; inner margins of eyes parallel to slightly converging below clypeus with punctures confluent (separate on sides in azteca), median half or two-thirds of truncation produced, apex of projection truncate, apex of trunca- tion longer than margin from end of truncation to lateral angle of clypeus; hypostomal carina not greatly produced, highest im- mediately behind angle. Farapsidal lines short linear; posterior coxa ivith carina along inner ventral angle; hind leg with basitarsus about four times as long as wide, widest near base and gradually narrowed toward apex, which is subtruncate or rounded; hind tibial spurs straight to weakly curved. In forewing, cells with short ( long in some areas ) , moderately sparse hairs, distal part with fine and dense papillae. Metasomal terga 2 to 5 lacking apical impunctate bands; tergum 6 with apical margin forming a flange. New World species included: Osmia (Diceratosmia) azteca Cresson, hotitena Cockerell, coniuncta Cresson, suhfasciata Cresson. The Old World species included: Osmia (Diceratosmia) gallartim Spinalo, rufigastra Lepeletier, submicans Morawitz, and versicolor Latreille. Species of Diceratosmia are widespread in the Old World; in the New World they range across the southern United States and extend southward to Costa Rica. The American species of the subgenus were revised by Michener ( 1949b ) . The above description is based on New World forms. Subgenus Nothosmia Ashmead Nothos^mia Ashmead, 1899, Trans. Amer. Ent. Soc, vol. 26, p. 75; Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, p. 63. Type species: Osmia distincta Cresson, 1864, original designation and monobasic Leucosmia Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29, pp. 166, 171. Type species: Osmia albiventris Cresson, 1864, original designation and monobasic. Xanthosmia Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29. pp. 166, 171. Type species: Osmia cordata Robertson, 1902, original designation and monobasic. Nothosmia can be recognized by the complete absence of black hairs in both sexes ( except in a few desert species ) , by the absence or narrowness of apical impunctate bands of the metasomal terga, by vein A being at least twice the length of vein B in forewing of the male, by presence of a tooth on the anterior margin of the hind basitars'is of the male and by the subequal lengths of the clypeal truncation and the margin from the end of the truncation to the lateral angles of clypeus in female. Genus Osmia in Western Hemisphere 231 Male. — Length of body 6 to 9 mm. Pubescense white (partially black in prunorti7n) ; metasomal terga often without apical or sub- apical fasciae. Mandible slender, widest at base, narrowed distad of base in some species, gradually widened to apex which is often less than twice as wide as constriction. Genal area as wide as or narrower than eye; inner margins of eyes converging anteriorly; clypeus densely covered with white hairs (most dense in fitusi and prunoriim) arising from fine (finest in tittisi) and confluent punctures, apical margin crenulate, subtruncate or slightly concave, with moderately wide impunctate apical band; scape equal to or slightly longer than second to fourth antennal segments combined; flagellar segments 2 to 11 unmodified, each less than twice as long as wide; hypostomal carina usually low and uniform, in some species gradually raised from posterior end to angle, but never greatly produced. Middle leg with tarsal segments not swollen; hind basitarsus three and one-half to five times as long as wide, toothed, widest at apex, gradually narrowed toward base; strigilis with apical margin of velum truncate, inner apical angle of velum narrowly subtriangular, concavity of inner margin of velum shallow to moderately deep, apex of mains usually produced into a long pointed spine. Forev/ing cells usually with dense, short, hairs (except scattered long hairs in inspergens), apical part papillate, vein A twice as long as vein B. Metasomal terga 2 and 3 with apical impunctate band narrow or absent, that of 4 and 5 some- times wider; midapex of tergum 6 with a shallow or deep emargina- tion; tergum 7 with apex produced, midapex of projection emargi- nate forming two sharp teeth. Metasomal sterna 1 and 2 with apical margins broadly rounded, sternum 2 without ventral midapical projection; sternum 3 with shallow or deep median apical emargi- nation which is thickly fringed with long hair or short hairs; sternum 4 with apical margin subtruncate or broadly rounded; sternum 6 with apex broadly rounded or subtruncate medially; sternum 8 with basal half narrowly triangular, apical half broadly subtriangular, with apex subtruncate, rounded or with a low median projection. Gonocoxite without any process arising before apex, swollen at angulation which bears tuft of bristles. Fenmle. — Length of body 8 to 11 mm. Pubescense not loosely branched, white (partially black in marginata and liogastra); meta- somal terga without apical or subapical fasciae (except in pumila, titusi and marginata); color of scopa white to yellowish (black in liogastra, black and white mixed in marginafa). Mandible slender, constricted moderately or strongly distad of base, gradually wid- 232 The University Science Bulletin ened to apex which is shghtly obHque, Httle wider than to twice as wide as width of constriction, base without transverse ridge or protuberance, apex tridentate, second tooth if present, rudimentary or continuous with first, third and fourth teeth narrowly triangular, distance between teeth 1 and 3 equal to or longer than distance between teeth 3 and 4; genal area narrower than or subequal to eye width; inner margins of eyes converging anteriorly; clypeal truncation simple or with a wide midapical projection, apex of which is truncate, truncation subequal to or shorter than margin from end of truncation to lateral angles of clypeus; hypostomal carina not greatly produced, usually higher than in males, uniform throughout posterior part, shghtly higher than anterior part. Hind basitarsus three to four times as long as wide, of uniform width throughout, or slightly wider near base than elsewhere, apex rounded; hind tibial spurs moderately curved apically or almost straight. Forewing moderately to extremely hyaline, hairs in cells short, dense (cordata with longer and sparse hairs), apex papillate. Metasomal terga 2 to 5 with or without narrow impunctate apical bands; tergum 6 with apical margin forming a flange. Species included: Osmia (Nothosmia) albiventris Cresson, cor- data Robertson, distincta Cresson, inspergens Lovell and Cockerell, liogastra Cockerell, lupinicola Cockerell, marginata Michener, mor- ongana Cockerell, priinorum Cockerell, ptimila Cresson, sand- houseae Mitchell, titiisi Cockerell. The subgenus is here divided into two groups, I (restricted to the eastern half of the United States and Canada) and II (found in the southwestern deserts) in order to show the characteristics of each group of species. Group I {albiventris, cordata, distincta, inspergens, piimila and sandhouseae): These species can be recognized by the following characteristics: In the female there is usually a small, sometimes weak, concavity at each end of the clypeal truncation; each man- dible has four teeth; the apices of the mandibles are much less than twice the widths at the constrictions; the eyes are subequal to or slightly narrower than the genal areas; and the scopa is white. In the male the color of the antennae and tegulae is brown, darker than that of group II; the wings are less hyaline than those of group II; and the metasomal terga usually lack subapical fasciae of simple or plumose hair. Group II (liogastra, lupinicola, marginata, morongana, prttnorum, and titiisi ) : These species can be distinguished by the following characteristics: In the female the apical margin of the clypeus Genus Osmia in Western Hemisphere 233 lacks a concavity at each end of the truncation; the mandibles are tridentate, constricted medially, the apices twice the widths at the constrictions or nearly so; the eyes are wider than the genal areas (subequal in titusi and prunorum); and the color of scopa varies from white, white mixed with black, to black. In the male, the antennae (ventral and lateral aspects only in some) and tegulae are testaceous; the wings are extremely hyaline; and the metasomal terga are provided with weak subapical fasciae of simple hairs. Most of the species of this group were treated by White (1952), who included them in AcantJiosmioides. Chenosmia, new subgenus Type species: Osmia pentstemonis Cockerel!, 1906. This is the largest North American subgenus of Osmia. Females can be distinguished from those of the closely related subgenus Monilosmia by the equal widths of the eyes and genal areas, the equal lengths of clypeal truncation and the margin from the end of the truncation to the lateral angle of the clypeus, and the low hypostomal carinae that are not abruptly reduced and hence not toothed near the angles. Distinctive recognition characters in males are few. Males usually differ from those of Monilosmia in the shorter malar spine of the strigilis, the straight or weakly curved hind tibial spurs, the shallow emargination of metasomal tergum 7, and the toothed basitarsi in a large number of species. Male. — Length of body 5 to 11 mm., usually less than 10 mm. Color of body in some species bright metallic green or blue; when of a dull metallic color, it is commonly green; metallic coloration lacking in inermis. Pubescence white with black intermixed in some areas; metasomal terga (except tergum 1) in most species without apical or subapical fasciae. Mandible slender or moder- ately wide, widest at base, narrower at region of constriction which is subequal to or slightly wider tlran apex, wdth V-shaped notch separating lower tooth from trunctate or oblique, acute to right- angled upper tooth; width of genal area less than eye; inner margins of eyes distinctly converging anteriorly; clypeus with apical margin weakly wavy to serrate, often weakly emarginate, moderately to broadly impunctate; scape subequal to second to fourth antenna! segments combined; flagellar segments 2 to 11 less than twice as long as wide; hypostomal carina not greatly produced, usually uniformly low. Middle leg with tarsal segments not swollen; hind basitarsus three to four times as long as wide, parallel sided or narrower basally and gradually widened towards apex, often with 234 The University Science Bulletin a small tooth on inner margin; strigilis with inner apical angle of velinn rounded, apex of velum truncate or subtruncate; apex of mains produced into short spine; emargination of inner margin of velum shallow to moderately deep. Forewing cells usually with dense and short hairs, distal part with short papillae, vein A usually less than twice as long as B (more than twice in coUinsiae, inde- prensa, ruinula, and proxima). Metasomal terga 2 to 5 with broad apical impunctate bands (narrow on terga 2 and 3 in coUinsiae); tergum 6 with apical emargination which is usually shallower than in subgenus Monilosniia; tergum 7 with apex produced, midapex of projection emarginate, usually forming two teeth. Metasomal sternum 1 with apex broadly convex, often with shallow midapical emargination; sternum 2 large, without apical process or tooth, covering most of third; distal portion of second strongly produced, apex convex, sometimes subtruncate medially or with shallow mid- apical emargination; median apical third of sternum 3 with an emar- gination which is fringed with long hairs; sternum 4 with apical margin convex, or middle third to half weakly to strongly produced, subtruncate to truncate; sternum 6 with a low, wide, rounded me- dian projection, tip of which may be truncate or emarginate; sternum 8 with apical half more broadly triangular than basal half, tip occasionally produced. Gonocoxite simple without a process arising before apex, tapering from base to short distance beyond tip of volsella, then weakly to strongly swollen and angulate, beyond angulation tapering to apex. Female. — Length of body 8 to 14 mm., usually less than 10 mm. Pubescence short and not loosely branched, metasomal terga with- out subapical or apical fasciae; hairs white mixed with black, sometimes largely black. Mandible not broad, slightly constricted distad of base, gradually widened to apex which is much less than twice as wide as width of constriction, lacking transverse ridge or protuberance near base, with four teeth (three in cerasi and zephyros), teeth narrowly to broadly triangular, second tooth often almost continuous with and subequal to first; distance between teeth one and three more than distance between three and four (except in cerasi and zephyros); genal area as wide as eye (slightly wider in dolerosa, malina, and robiistella); inner margins of eyes slightly to distinctly converging; clypeus with punctures confluent, truncation not swollen (except in indeprensa) , usually equal to margin from end of tnmcation to lateral angle of clypeus; malar space absent; hypostomal carina low and uniform, often slightly raised behind angle, any resulting tooth low and rounded. Hind Genus Osmia in Western Hemisphere 235 basitarsus three to nearly four times as long as wide, usually widest in middle, gradually tapering towards base and apex, apex truncate to subtruncate; hind tibial spurs distinctly curved at tips. Forewing with hairs on cells usually dense and short (sparse and longer in briineri, cerasi, and cyonopoda). Metasomal terga 2 to 5 with apical impunctate bands narrow to broad; tergum 6 with apical margin forming a flange. North American species included in this subgenus: Osmia ( Chenosmia ) atriventris Cresson, hruneri Cockerell, calla Cockerell, caulicola Cockerell, cerasi Cockerell, clarescens Cockerell, cobaltina Cresson, coUinsiae Robertson, cijanopoda Cockerell, dolerosa Sand- house, exiqua Cresson, illinoensis Robertson, indeprensa Sandhouse, inermis (Zetterstedt), kincaidii Cockerell, laefa Sandhouse, malina Cockerell, nanuJa Cockerell, pentstemonis Cockerell, pingreeana Michener, pagoso Sandhouse, potentiUae Michener, proxima Cres- son, regulina Cockerell, sanctae-rosae Cockerell, sequoiae Michener, tersula Cockerell, tokopahensis Michener, trevoris Cockerell, tristella Cockerell, zephijros Sandhouse. This subgenus is represented in the Old World so far as known by only one species, Osmia (Chenosmia) inermis (Zetterstedt), which also occurs in boreal North America. Euthosmia, new subgenus Type species: Heriades glaucum Fowler, 1899. Sandhouse (1939:64) included all the species of Euthosmia under the subgenus Nothosmia, but they seem different enough to be recognized as the members of a distinct subgenus. A combina- tion of the following characters separates this group from all other subgenera of Osmia. In the male the apical margin of the second sternum is subtruncate and does not overlap the third to any extent; the apical margin of sternum 3 is without or with only a shallow median emargination; the apical margin of sternum 6 is broadly subtriangular; the basal half of sternum 8 is narrowly triangular and the distal half bears a median projection (subtriangular in claremontensis); the apex of the velum is truncate; the spine of the malus is slender and usually long; and the inner margin of the velum is deeply emarginate. In the female the bases of the maxil- lary stipites and also the ventral area of the head have long curled hairs; the scopa is pale; the tridentate mandible is slightly narrower at the constriction than at the apex; the strigilis has a long sharp apical malar spine projecting downward, the apex of the velum is truncate and the inner margin of the velum deeply emarginate. 236 The University Science Bulletin Because no females of nemoris and claremontensis have been de- scribed, the female characters cited here are of glauca only. Male. — Length of body 4 to 11 mm. Pubescence white. Mandible slender, widest at base, constricted just distad of base, lateral mar- gins from constriction to apex subparallel, with V-shaped notch separating long slender lower tooth from truncate or oblique upper tooth; eye wider than genal area, inner margins of eyes distinctly converging anteriorly; clypeus covered with erect, white hairs aris- ing from fine contiguous punctures, apex truncate, weakly wavy to crenulate with narrow to moderately wide impunctate apical band; scape subequal to or slightly shorter than second to fourth antennal segments combined; flagellar segments 2 to 11 each less than twice as long as wide (in glauca, nearly twice as long as wide); hyjio- stomal carina low and of uniform height anterior and posterior to angle. Second and third tarsal segments of middle leg not swollen; hind basitarsus three to four times as long as wide, not toothed, of uniform width throughout or gradually widened towards apex; strigilis with apical margin of velum truncate, inner apical and basal angles acute, pointed; concavity at inner margin of velum deep; spine of mains long (short in daremontensis) . Forewing with vein A twice to less than twice as long as vein B; papillae on distal part fine, dense; hairs on cells short, dense. Metasomal terga 2 to 5 with moderate to broad impunctate bands, with subapical fasciae of hairs ( except in glauca ) ; apical margin of tergum 6 with or without shallow poorly defined emargination at midapex; apex of tergum 7 produced, midapex of projection emarginate, forming two, usually sharp teeth. Metasomal sterna 2 and 3 with apical margins truncate to weakly convex; sternum 4 weakly to strongly convex; sternum 5 broadly rounded with or without shallow median emargination; sternum 6 with apical margin broadly subtriangular; sternum 8 with basal half triangular and base strongly acuminate, distal half with median projection ( more broadly triangular in clare- montensis). Gonocoxite tapering slightly from base to angulation (strongly angulate in claremontensis) . Female. — Length of body 6 to 8 mm. Scopa white. Metasomal terga without apical or subapical fasciae. Mandible wide, slightly constricted distad of base, gradually widened to apex, which is a little wider than at constriction; protuberance or transverse ridge near base of mandible absent; apex tridentate with teeth blunt, short, two upper teeth almost continuous, lower tooth triangular; distance between teeth one and three slightly longer than between three and four; width of genal area subequal to width of eye; inner Genus Osmia in Western Hemisphere 237 margins of eyes converging anteriorly; clypeus with punctures sep- arate, margin not swollen, apical truncation slightly longer than margin from end of truncation to lateral angle of clypeus; tttfts of orange hair, arising beneath chjpeal margin absent. Hypostomal carina uniform, slightly higher behind angle than anteriorly; hy- postomal area tvith medialhj directed long stiff bent hairs. Basi- tarsus of hind leg over three times as long as wide, subparallel sided, apex rounded; hind tibial spurs almost straight at apices; strigilis with apical spine of malus longer than in male, inner margin of velum strongly emarginate. Forewing with hairs in cells fine, dense. Metasomal terga 2 to 5 with narrow apical impunctate band; ter- gum 6 with apical margin forming flange. Scopa restricted to apical halves of metasomal sterna 2 to 5. Species included: Osmia (Etithosmia) claremontensis Michener, glatica (Fowler) and nemoris Sandhouse. This subgenus is restricted to the Pacific Coast of North Amer- ica. The possibility exists that glauca ( = exilis Sandhouse ) is not closely related to the other species; this is perhaps suggested by the medially produced sixth tergum of males of claremontensis and nemoris. Until females of these species are known, it will be dif- ficult to decide this matter. Subgenus Acanthosmioides Ashmead Acanthosmioides Ashmead, 1899, Trans. Amer. Ent. Soc, vol. 26, p. 76; Cock- erell, 1912, Proc. U. S. Nat. Mus., vol. 42, p. 216; Sandhouse, 1939, Mem. Ent. Soc. Washington, vol. 1, p. 39; White, 1952, Univ. Kansas Science Bulletin, vol. 35, p. 219. Type species: Osmia odontogaster Cockerell, 1897, original designation and monobasic. White (1952) revised the subgenus in detail and included twenty- five species. Five desert species (liogastra, lupinicola, moron- gana, prunorum, and titusi ) grouped under the subgenus by White have been excluded in this revision and transferred to the subgenus Nothosmia. The main recognition characters of this subgenus are the constriction of the mandible distal to the base so that in the female the apex is about twice, and in the male at least one and one-half times, the width of the constriction. In females this char- acter is shared by the western group of Nothosmia. In males, metasomal sternum 2 usually has a midapical tooth or projection and the gonocoxite is provided with a subapical process. Male. — Length of body 8 to 14 mm. Pubescence white with black intermixed; metasomal terga without apical or subapical fasciae. Mandible slender, widest at apex which is at least one and one half times as wide as constriction; apex with V-shaped notch separating 238 The University Science Bulletin long slender, lower tooth from truncate or oblique, obtuse or right angular upper tooth; eye slightly to moderately wider than genal area; inner margins of eyes subparallel to slightly converging; cly- peus very thickly covered with erect, white hairs arising from extra- ordinarily fine and confluent punctures, apex produced, subtrun- cate with narrow, thickened, impunctate margin, median portion dentate; scape usually equal to antennal segments 2 to 4 combined; flagellar segments 2 to 11 each less than twice as long as wide, last segment sometimes compressed and expanded or flagellar seg- ments moniliform; hypostomal carina not greatly produced, usually low and gradually raised from posterior end to angle. Middle leg with tarsal segments usually not swollen, but in longula mediotarsal segments of fore and middle legs much enlarged; hind basitarsus about three times as long as wide, not toothed, of uniform width or widened distally; strigilis with inner apical angle of velum rounded; apex of velum subtruncate; apex of malus produced into a short to slightly longer sharp spine; concavity on inner margin of velum usually deep. Forewing cells with short hairs, distal portion papillate, vein A usually twice as long as vein B. Metasomal terga 2 to 5 with wide apical impunctate bands, usually with sub- apical fimbriae of simple hairs; tergum 6 with apical margin with or without a weak and poorly defined apical emargination, forming two lobes or teeth. Metasomal sternum 1 with apical margin sub- truncate, convex or narrowly subtiiangular, with median emargina- tion; sternum 2 large, largely covering third, ventral surface with a median apical projection or protuberance on each side of which the apical margin is fringed with long hairs (except Integra which has no projection or protuberance, but has a median band of long hairs); sternum 3 with a shallow midapical emargination which is fringed with long hairs; sternum 4 witli a median apical patch of hairs bent mesad at tips, surrounding an impunctate portion of varying shape; sternum 6 with a median apical produced portion with apex either truncate or slightly emarginate; sternum 8 much reduced with apical half either triangular or produced into a sub- triangular process, basal half wide to middle, then abruptly nar- rowed to a slender tip (except longula where it is broadly subtri- angular). Gonocoxite with a process arising before its apex, apical half of the process usually spatulate, distal portion of gonocoxite abruptly narrowed, slender and hairless. Female. — Length of body, 8 to 15 mm. Pubescence usually not long and loosely branched, scopa black. Mandible slender, con- Genus Osmia in Western Hemisphere 239 stricted just distal to base, gradually widened to apex, which is usually twice as wide as width of constriction, lacking transverse ridge or protuberance near base, ti'identate (rudiment of second tooth absent in most species ) ; distance between teeth 1 and 3 equal to or slightly shorter than distance between 3 and 4; eye usually narrower than genal area but often subequal to it; inner margins of eyes slightly converging (parallel in calcarata, integra, and longula); clypeus with punctures confluent, with truncation un- thickened, equal to or slightly shorter than distance from end of truncation to lateral angle of clypeus; hypostomal carina not much elevated, highest at or near angle, not forming tooth. Hind basi- tarsus a little wider just before base than elsewhere, apex subtrun- cate to rounded; hind tibial spurs relatively straight, rarely much curved at tips. Forewing hairs in cells short, dense, papillae on distal part sparse, dense in some areas. Metasomal terga 2 to 5 with moderately wide apical impunctate bands ( scattered punctures extend to apical margins in sediila); tergum 6 with apex forming a strong flange. Species included: Osmia (Acanthosmioides) ashmeadii (Titus), calcarata White, dakotensis Michener, francisconis White, giliarum Cockerell, giffardi Sandhouse, Jnirdi White, integra Cresson, ken- oijeri Cockerell, lanei Sandhouse, longula Cresson, nifoata Cockerell, nigrifrons Cresson, nigroharhata Cockerell, obliqua White, odon- togaster Cockerell, phtjsariae Cockerell, sediila Sandhouse, sladeni Sandhouse, trifoliama Sandhouse, unca Michener, watsoni Cockerell. The species of this subgenus are restricted to western North America. Subgenus Monilosmia Robertson Monilosinia Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29, pp. 166, 171. Type species: Osmia canadensis Cresson, 1864 = Osmia simillirna Smith, 1853, by original designation and monobasic. This subgenus resembles Chenosmia in general characters but females of Monilosmia can be distinguished as follows: The clypeal truncation is longer than the distance from the end of the trunca- tion to the lateral angle of the clypeus; the genal areas are wider than the eyes; the hypostomal carina is high, abruptly reduced near the angle, forming a tooth (often rounded). Species of this subgenus are larger than most Chenosmia (usually about 10 mm. long) and generally darker blue. In the male, there is no single diagnostic character for the subgenus which is applicable to all the species. Nevertheless a combination of larger size, the longer 240 The University Science Bulletin malar spine of the strigilis, the curved apices of the tibial spurs, the deeper emargination of the seventh metasomal tergum, the absence of a tooth on the basitarsus of the hind leg, and the sparser hairs in the cells of forewing help to distinguish males of Monilos- mia, in general, from those of Chenosmia. Male. — Length of body 7 to 11 mm. Color of body dark blue to greenish blue. Pubescence black and white; metasomal terga (except tergum 1) without apical or subapical fasciae. Mandible widest at base, narrower at region of constriction which is slightly narrower to wider than apex, with V-shaped notch separating usually long lower tooth from acute to right angular upper tooth; eye equal to or wider than genal area; inner margins of eyes slightly to distinctly converging anteriorly; clypeal margin weakly wavy to strongly crenulate, narrowly or broadly impunctate; scape usually subequal or equal to second, third and fourth antennal segments combined (longer in cyaneUa, shorter in dmiUima); flagellar seg- ments 2 to 11 each less than twice as long as wide (moniliform in simillima); hypostomal carina usually low and uniform. Middle leg with tarsal segments not swollen; hind basitarsus three to four times as long as wide, narrow at base, gradually widened toward apex, which is often considerably wider than base, not toothed; hind tibial spurs distinctly curved at tips (straight in cijanella); strigilis with inner apical angle of velum rounded; apex of velum usually truncate; concavity of inner margin of velum shallow to moderately deep; apex of mains produced into a short spine. Fore- wing cells usually with dense and short hair ( longer in alholateralis ) , distal portion finely papillate, often with hairs similar to those of cells, vein A equal to or less than twice as long as vein B. Meta- somal terga 2 to 5 with apical impunctate bands wide (narrow to moderately wide cijanella, and rostrata); apical margin of tergum 6 with shallow to deep median emargination usually clearly defined; tergum 7 with apex produced, midapex with emargination, form- ing two teeth (two lobes in ftixta). Metasomal sternum 1 with apical margin broadly rounded, midapex not emarginate (except in rostrata); sternum 2 with apex usually strongly convex, often with midapical third or fourth subtruncate; sternum 2 large, largely covering third; median third of sternum 3 with an emargination which is fringed with long hairs; sternum 4 with midapical third produced, subtruncate or rounded, posterior median portion with long hairs which are usually bent mesad, apical margin uniformly convex or with middle third subtruncate; sternum 6 broadly sub- Genus Osmia in Western Hemisphere 241 triangular; sternum 8 with apical half more broadly subtriangular than basal half, tip may be acuminate. Gonocoxite never greatly widened (except at angle in atrocyanea) , without subapical process. Female. — Length of body 10 to 13 mm. Color of body dark blue with purplish or greenish tinge. Pubescence short, not loosely branched, mixed black and white or entirely black; metasomal terga without subapical or apical fasciae (except in seclusa); scopa black (fuscous in rostrata). Mandible broad, weakly or strongly constricted distad of base, gradually widened to apex which is slightly wider than constriction (in cyaneonitens and rawlinsi apex nearly twice as wide as width of constriction), lacking protuber- ance or transverse ridge near base (except in atrocyanea, rostrafa, scuUeni, and seclusa, which have a low transverse ridge like that of some species of Centr osmia), apex with four teeth, tooth 2 often considerably smaller than others; distance between teeth 1 and 3 greater than distance between 3 and 4; eye usually considerably narrower than genal area; inner margins of eyes slightly to moderately converging; clypeus with punctures confluent, trunca- tion not thickened (with small median angle in sculleni, large median projection in rostrata), usually longer than margin from end of truncation to lateral angle of clypeus; malar space absent; hy- postomal carina strongly elevated, abruptly reduced just behind or well behind angle forming a distinct tooth. Hind basitarsus three to four times as long as wide, usually wider in the middle than at the base or apex, apex subtruncate to truncate; hind tibial spurs strongly curved at tips. Forewing cells usually with short, dense hairs, in some species longer hairs scattered in different areas. Metasomal terga 2 to 5 with apical impunctate bands wide (nar- rower in cyanella); tergum 6 with apical margin forming a flange. Species included: Osmia (Monilosmia) alholateralis Cockerell, atrocyanea Gockerell, brevis Cresson, bridivelli Sandhouse, cara Cockerell, cyanella Cockerell, cyaneonitens Cockerell, densa Cres- son, gahrielis Cockerell, hendersoni Gockerell, juxfa Cresson, raw- linsi Sandhouse, rostrata Sandhouse, sculleni Sandhouse, seclusa Sandhouse, simillima Smith. The subgenus Monilosmia is largely restricted to the western half of North America, but O. similliina ranges eastward to the Atlantic in Canada and the northern part of the United States. Unfortimately the type species of the subgenus is quite aberrant, particularly in the male, and is typical of the group only in a nomen- clatorial sense. 242 The University Science Bulletin Subgenus Centrosmia Robertson Centrosinia Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29, pp. 165, 166, 170. Tijpe species: Osmia bticephala Cresson, 1864, original designation and monobasic. Robertson proposed this name in his key to the genera of the subfamily Osmiinae (Family: Megachilidae ) for the single species, Osmia bticephala Cresson. Sandhouse grouped bticephala, nigri- ventris, and O. (Chenosmia) inermls in the subgenus Melanosmia, with the remark "Although this subgenus is based chiefly on differ- ences of color, both of integument and pubescence, and no single structural character is common to all of the included species, it is retained as a convenient group for three species of this fauna, two of which are holarctic." O. bticephala and nigriventris are here removed from Melanosmia because of the pattern of the seventh and eighth metasomal sterna, and genitalia of the male, and be- cause of the swollen apical margin of clypeus, and the presence of a ridge across the base of each mandible in the female. In the above mentioned characters, bucephala and nigriventris agree with a group of species placed in the subgenus Nothostnia by Sand- house, and here united under the name Centrosmia. Male. — Length of body 7 to 17 mm. Pubescence white, usually with black intermixed; metasomal terga without apical or subapical fasciae. Mandible widest at base, narrowest at region of con- striction which is more than half as wide as apex, with V-shaped notch separating long, lower tooth from truncate or oblique, acute, upper tooth; width of genal area equal to or wider than eye; inner margins of eyes slightly to distinctly converging anteriorly; clypeus thickly covered with white hairs arising from confluent punctures, margin weakly wavy with one or two weak median anterior pro- jections, apical margin narrowly impunctate; scape equal to an- tennal segments 2 to 4 combined (except in bucephala and pikei where scape shorter); flagellar segments 2 to 11 each less than twice as long as wide (more than twice as long as wide in bucephala); hypostomal carina not greatly produced, usually low and slightly raised behind angle. Middle leg with second and third (also fourth in bticephala ) tarsal segments greatly swollen ( except in bakeri and nigriventris) ; basitarsus of hind leg three to four or more times as long as wide, parallel sided, broadest medially, or gradually nar- rowed towards apex (in bucepliala and nigriventris widest near apex, narrowed towards base); strigilis with inner apical angle of velum rounded, apex truncate to shghtly oblique; apex of mains Genus Osmia in Western Hemisphere 243 produced to a sharply pointed spine; concavity on inner margin of velum shallow or absent. Forewing with pubescence of cells usually sparse and short (longer in austromaritima), apex papillate, vein A longer than vein B (except in pikei where A may be nearly subequal to B ) . Metasomal terga 2 to 5 with broad apical impunc- tate bands; apical margin of 6 with or without shallow undefined emargination or truncation; apex of 7 produced, midapex of pro- jection emarginate, forming two blunt, wide lobes (except austro- maritima which has two sharp teeth). Metasomal sterna 1 and 2 with posterior margins arcuate ( first with midapical emargination in vandykei and austromaritima); sternum 2 large, largely covering sternum 3; median apical third of sternum 3 with an emargination which is fringed with long hairs; midapical half or third of sternum 4 produced and truncate or rounded, usually with a median apical patch of bristle-like hairs bent mesad at tips; sternum 6 with a mid- apical projection with or without a patch of bristles on projection; sternum 8 subtriangular with apical process usually long and slender of varying length. Gonocoxite without a process arising before apex (bucephala with subapical ventral angle produced); distal portion of gonocoxite abruptly narrowed, slender, hairless; apex of volsella subtriangular (rounded in bakeri). Female. — Length of body 9 to 16 mm. Metasomal terga without apical or subapical fasciae. Scopa black. Mandible gradually widened to apex which is much less than twice as wide as width of constriction; mandible with transverse ridge or protuberance near base; apex with second tooth rudimentary (large in nigriventris); distance between teeth 1 and 3 greater than distance between 3 and 4, eye narrower than genal area, inner margins of eyes slightly converging to diverging; clypeus with punctures confluent, trunca- tion thickened, longer than distance from end of truncation to lat- eral angle of clypeus; malar space absent (present in bucephala); hypostomal carina greatly to moderately elevated, highest just be- hind to well behind angle, lower at angle so that tooth (sometimes low and rounded) is formed behind angle. Hind basitarsus be- tween three and four times as long as wide, almost uniformly wide throughout, apex truncate; hind tibial spurs curved at tips. Fore- wing cells with hairs fine, sparse, dense in certain areas, papillae on distal part fine, dense. Metasomal terga 2 to 5 with apical impunctate bands, finely lined; tergum 6 with apical margin form- ing a flange. Species included: Osmia (Centrosmia) austromaritima Michener, 244 The University Science Bulletin hakcri Sandhouse, bucephaJa Cresson, nigriventris ( Zetterstedt ) , pikei Cockerell, raritaiis Michener, tanneri Sandhouse, thysanisca Michener, vandykci Sandhouse. This subgenus is known only from North America, except for nigriventris which occurs also in northern and alpine Europe. The species have recently been revised by Sinha and Michener (1958). Trichinosmia, new subgenus Type species: Osmia Intisulcata Michener, 1936. Sandhouse (1939) included Osmia htistilcata in the subgenus Nothosmio. This species differs, however, from all other subgenera, including Nothosmia, in having the following combination of char- acters: head and thorax with unusually long, branched hair; terga without impunctate bands; forewings with long, sparse hair; in the male, flagellar segments 2 to 11 each twice as long as wide, scape shorter than antennal segments 2 to 4 combined, gonocoxite and penis valve of uniform width, the former with scanty, sparse hairs; in the female, clypeal margin tridentate (much as in sctilleni), head as long as broad, terga 2 to 6 with apical bands of simple hair, and strigilis with an unusually long apical malar spine ( one-third of entire malus). Male. — Length of body 10 mm. Pubescence white, with black intermixed in a few areas, hairs with long branches on head and thorax; metasomal terga with apical fasciae of simple hair. Mandi- ble moderately wide, widest at base, narrower and uniformly wide at region of constriction and apex, with V-shaped notch separating long lower tooth from acute upper tooth; width of genal area sub- equal to that of eye; inner margins of eyes slightly converging anteriorly; clypeus with margin truncate, weakly wavy, with apical margin moderately broadly impunctate; scape shorter than antennal segments 2 to 4 combined; flagellar segments 2 to 11 each at least twice as long as wide; hypostomal carina low, slightly higher well behind angle. Middle leg with mediotarsal segments not swollen; hind basitarsus about three times as long as wide, not toothed, par- allel sided; strigilis with inner apical angle of velum roimded, apex of malus produced into a long, moderately sharp spine, inner margin of velum deeply emarginate. Forewing cells with sparse and long hair, distal portion with hairs longer than in any other subgenus, vein A much less than twice as long as vein B. Metasomal terga 2 to 5 without impunctate bands; apical margin of tergum 6 with a moderately deep well defined emargination; apex of tergum 7 Genus Osmia in Western Hemisphere 245 produced, midapical production weakly emarginate forming two short teeth. Metasomal sterna 1 and 2 with apical margins broadly rounded, median third of each subtruncate; sternum 3 with median emargination, fimbria extending beyond apex; sternum 4 with apical margin slightly convex, with a short, stout, median projection; sternum 8 with basal half triangular, apical half broadly rounded. Gonocoxite of nearly uniform width throughout length, slightly curved, without subapical swelling, similar to that of subgenus Cephalosmia, apex of volsella subtriangular. Female. — Length of body 12 to 13 mm. Pubescence on head and thorax with long branches, simple hair on metasomal terga, color mixed black and white; scopa black. Mandible moderately broad, weakly constricted distal to base, gradually widened to apex which is strongly oblique and less than twice as wide as width of constric- tion, base of mandible lacking transverse ridge or protuberance, apex with four teeth, teeth 1 and 2 almost confluent, 3 and 4 nar- rowly triangular, distance between 1 and 3 slightly greater than distance between 3 and 4; eye subequal to width of genal area; inner margins of eyes subparallel or weakly converging anteriorly; clypeus with punctures confluent, truncation not thickened, with one median and two lateral teeth, much longer than margin from end of truncation to lateral angle of clypeus; malar space absent; scape shorter than antennal segments 2 to 4 combined; hypostomal carina moderately high, abruptly reduced behind angle forming weak tooth as in some Monilosmia. Hind basitarsus nearly four times as long as wide, subparallel sided; hind tibial spurs almost straight, forewing with hairs on distal part and cells sparse and longer than those of any other subgenus. Metasomal terga 2 to 5 without apical impunctate bands, with simple erect hairs; tergum 6 with apical margin forming a flange. Osmia (Trichinosmia) latisttlcata Michener is the only species included in this subgenus. It is restricted to Western North Amer- ica, Subgenus Cephalosmia Sladen Cephalosmia Sladen, 1916, Canad. Ent., vol. 48, p. 270; Sandhouse, 1939, Mem. Ent. Soc. Washington, vol. 1, p. 21. Type species: Osmia armaticeps Cres- son, 1878 = Osmia montana Cresson, 1864, monobasic. This subgenus is recognized by the following combination of characters: The head is long and the genal area very wide; the hypostomal carina is high, in the female not abruptly reduced behind the angle; the superior mesal portion of the carina border- 246 The University Science Bulletin ing the antennal socket is more strongly developed than the rest of the carina; the female has a carina on the inner surface of the mandible which arises near the superior basal angle and extends obliquely to the inferior margin, where it ends abruptly about halfway between the base and the apex; the metasomal terga of the female are weakly convex and densely punctured, the apical margin of the sixth being strongly curved and without a wide apical flange. In the male sterna 2 and 3 have the apical margins truncate or subtruncate, sternum 2 only slightly overlapping the base of sternum 3; the base of sternum 6 has a deep triangular emargination, the apical margin bears a strong triangular or rounded median projection. Male. — Length of body, 8 to 13 mm. Pubescence black or black and white intermixed. Mandible widest at base, narrower at re- gion of constriction which is slightly wider than apex, with V- shaped notch separating long (short in some species) lower tooth from triangular, acute, upper tooth; eye slightly narrower than genal area; inner margins of eyes slightly to distinctly converging anteri- orly; clypeus covered with white hairs arising from confluent punc- tures, margin produced, often convex, median portion weakly wavy or dentate, with apical impunctate band narrow to moderately wide; scape subequal to or slightly shorter than antennal segments 2 to 4; segments 2 to 11 unmodified, each less than twice as long as wide ( in montana nearly twice as long as wide ) ; superior niesal por- tion of carina bordering antennal foramen more strongly developed than rest of carina; hypostomal carina never high, moderately raised a little behind angle. Middle leg with tarsal segments 2 and 3 not swollen; hind basitarsus four times as long as wide or slightly less, uniformly wide or wider in distal portion, not toothed; strigilis with apical margin of velum truncate or rounded, apex of mains pro- duced into a short to moderately long, narrow spine, concavity on inner margin of velum moderately deep. Forewing cells with long or short, sparse hairs, distal portion of wing finely papillate, vein A usually less than twice as long as vein B. Metasomal terga 2 to 5 with apical impunctate bands wide; tergum 6 with shallow to deep clearly defined emargination; tergum 7 with apex produced, midapex of projection emarginate, forming two triangular teeth. Metasomal sternum 1 with apical margin weakly convex, median portion with a wide, shallow emargination; sterna 2 and 3 with apical margins truncate or subtruncate; sternum 2 not covering sternum 3; sternum 4 with median third of apical margin usually Genus Osmia in Western Hemisphere 247 produced, apex of projection rounded or truncate, fringed with dense, short, black hairs; base of sternum 6 with a deep triangular emargination, apical margin with a strong triangular or rounded median projection, basal half of eighth sternum broadly triangular, distal half with sides subparallel basally, then strongly converging, apex subtriangular and fringed with bristles. Gonocoxitc thick, tapering slightly from base to apex, lateral surface of apical third beset with hairs, some hairs on mesal surface near apex; free por- tion of volsella rather narrow, subtriangular. Female. — Length of body 10 to 17 mm. Pubescence black or mixed black and white; scopa black. Mandible broad, constricted distal to base, gradually widened to apex which is one and one half to two times as wide as width at constriction, base without transverse ridge or protuberance (except in quadriceps), with four distinct teeth, teeth 3 and 4 triangular, 1 and 2 subtriangular to broadly rounded, distance between 1 and 3 much longer than distance between 3 and 4; inner surface of mandible with a carina which arises near upper basal angle and extends obliquely to lower mar- gin, where it ends abruptly about half way between base and apex; genal area much wider than eye which is more than twice as long as wide; inner margins of eyes slightly converging to slightly di- verging; clypeus (shortened and modified in montana and quadri- ceps) with punctures separated or nearly absent, with truncation shorter than margin from end of truncation to lateral clypeal angle, apical margin simple or modified; mesosuperior portion of carina bordering antennal socket more strongly developed than rest of cafina; hypostomal carina low to moderately high, highest immedi- ately or well behind angle where it forms a rounded tooth; hind basitarsus four to five times as long as wide, of uniform width throughout or tapering slightly, apex truncate to broadly roimded; hind tibial spurs straight to weakly curved distally. Forewing with hairs on cells scanty, usually short and sparse, distal part with fine, dense or sparse papillae. Metasomal terga 1 to 5 with apical im- punctate bands moderately wide; tergum 6 not flanged or rarely with a weak flange. Species included: Osmia (Cephalosmia) calif ornica Cresson, grinnelli Cockerell, marginipennis Cresson, mendocinensis Sand- house, montana Cresson, pascoensis Cockerell, quadriceps Cresson, subaustralis Cockerell. This subgenus is restricted to western North America. 248 The University Science Bulletin Subgenus Chalcosmia Schmiedeknecht Osinia (Chalcosmia) Schmiedeknecht, 1886, Apidae Europaeae, vol. 2, p. 886; Ducke 1900, Berichte des naturvvissenschaftHch-medizinischen Vereines in Innsbruck, vol. 25, p. 16; Friese, 1911, Das Tierreich, vol. 28, p. 108; Sand- house, 1939, Mem. Ent. Soc. Washington, vol. 1, p. 13. Type species: Apis fulviventris Panzer, 1798, designation of Sandhouse, 1939. Gtiathosmia Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29, pp. 165 171; Cockerel], 1912, Proc-. U. S. Nat. Mus., vol. 42, p. 216. Type species: Osrnia georgica Cresson, 1878, original designation and monobasic. This subgenus can be recognized by the following characters: In the male the fourth metasomal sternum is thickened at the apical margin, the lateral portions have narrow transverse sulci (visible in posterior view), and each hind basitarsus bears a tooth. In the female the clypeal margin is thickened or otherwise modified and provided with t\vo small sockets, each giving rise to a tuft of hair, mesal to the two large lateral tufts. There is a transverse ridge or a protuberance near the base of the mandible in the female. Male— hength of body 8 to 13 mm. Pubescence white or slighdy yellowish. Mandible moderately wide, widest at base, narrower at region of constriction which is almost as wide as apex, with V- shaped notch separating lower tooth from truncate or oblique, right angular or acute upper tooth; eye equal to or slightly wider than genal area; inner margins of eyes slightly to distinctly con- verging anteriorly; clypeus covered with white hairs arising from confluent punctures; clypeal margin with median portion slightly produced, wavy or dentate; clypeus with apical impunctate band narrow to moderately wide; scape subequal to second, third, and fourth antennal segments combined; flagellar segments 2 to 11 unmodified, each less than twice as long as wide; hypostomal ca- rina low and raised slightly or considerably behind angle; middle leg with tarsal segments not swollen; hind basitarsus of uniform width, usually strongly toothed (tooth smallest in coeritlescens) , parallel sided; strigilis with apical margin oblique; apex of malus produced into log spine which is widest at base and gradually nar- rowed toward apex; concavity on inner margin of velum shallow to moderately deep. Cells of forewing with sparse (dense in some areas ) and short hairs, distal portion finely papillate, vein A usually less than twice as long as vein B (A is equal to B in tcxana). Meta- somal terga 2 to 5 with apical impunctate bands narrow to mod- erately wide, with subapical fasciae; apical margin of tergum 6 usually with rather shallow but clearly defined emargination; apex of tergum 7 produced, apex of projection emarginate, forming two Genus Osmia in Western Hemisphere 249 blunt teeth. Metasomal sternum 1 with apex broadly rounded; sternum 2 large, greatly produced with apical margin convex, cov- ering most of third, usually a shallow emargination at midapex of produced portion; median apical third of third sternum with an emargination which is fringed with long hairs; midapical half or third of sternum 4 produced and subtruncate or rounded, with a median apical patch of long hairs usually not surrounding an im- punctate area; sternum 4 with apical margin thickened, lateral portion of margin thickened with a narrow transverse stdcus vis- ible from behind; sternum 6 with apical margin strongly to weakly produced medially; sternum 8 with base subtriangular, apical por- tion broadly convex. Gonocoxite wide, gradually narrowed toward the tip, bent mesally toward apex but not angulate, with neither swelling before distal portion nor subapical process; penis valve broad at base and gradually narrowed towards apex (in coerules- cens gonocoxite uniformly slender throughout except at base, penis valve broad, of uniform width except at tip). Female. — Length of body 10 to 16 mm. Scopa yellow or black; metasomal terga usually with subapical fasciae of plumose hair. Mandible broad, slightly constricted distal to base, gradually wid- ened to apex, which is much less than twice as wide as width of constriction, with one or two protuberances or a transverse ridge near base of mandible, with four teeth; teeth 1, 3, and 4 triangular; tooth 2 obliquely truncate at apex (in coeridescens teeth one and two continuous); distance between teeth 1 and 3 subequal to or greater than distance between teeth 3 and 4; eye subequal to or much narrower than genal area; inner margins of eyes slightly con- verging to diverging; clypeus shortened and modified in native American species, punctures usually confluent, truncation longer than margin from end of truncation to lateral angle of clypeus, apical margin thickened or otherwise modified, two small round tufts of bristles mesal to usual broad laterattuft; hypostomal carina high and abruptly reduced a short distance behind angle; hind basitarsus at least three times as long as wide, apex rounded ( in coerulescence subtruncate); hind tibial spurs slightly curved distally. Forewing with hairs of cells short, fine, dense, sparse in some areas, papillae on distal part fine, dense; terga 1 to 5 without apical impunctate bands; tergum 6 with apical margin forming a broad flange. North American species included: Osmia (Chalcosmia) chahjbea Smith, coerulescens (Linnaeus), coloradensis Cresson, georgica Cresson, and texana Cresson. 250 The University Science Bulletin Old World species examined and included in this subgenus: Osmia (Chalcosmia) oiirtilenta (Panzer), coerulescens (Linnaeus), decemsiiinata (Radoszkowski), climidiata Morawitz, lofreillei Spin- ola, leiicogastra Morawitz, Icaiayia Kirby, medanae Magretti, notata (Fabricius), rubicola Friese, and vidua Gerstaecker. The four native American species of Chalcosmia seem to be quite distinct from the fifth, coerulescens, which is probably introduced from Europe. Group I {chahjhea, coloradensis, georgica and texana) is char- acterized by the strongly developed upper margins of the two mesal sockets on the apical margin of the clypeus and the strongly developed protuberances near the base of the mandible in the fe- male; and by the distinctly pointed tooth on each hind basitarsus, the thickened apical margin of the fourth metasomal sternum and the broad and apically tapering gonocoxite in the male. Group II {coerulescens) is characterized by the unmodified mar- gins of the two mesal sockets on the apical margin of the clypeus and the lack of or the extremely weak protuberance near the base of the mandible in the female; and by the extremely weak tooth on the hind basitarsus, the scarcely thickened apical margin of the fourth metasomal sternum and the uniformly slender gonocoxite in the male. The Old World species can be placed in several groups. O, auru- lenta, decemsignata, medanae, notata and vidua agree fairly well with coerulescens and may be placed in group TI. The other species listed belong in two or three other groups. Subgenus Osmia Panzer, s. str. Osmia Panzer, 1806, Kritische Revision der Insekten-faune Deutschlands, vol. 2, p. 230; Sandhouse, 1939, Mem. Ent. Soc. Washington, vol. 1, p. 9; Michener, 1941, Amer. Midland Nat., 26: 163. Type species: Anthophora bicornis Linnaeus, 1758, = Apis rufa Linnaeus, 1758, designation of La- treille, 1810. Amhhjs Klug, 1907, Illiger's Mag. f. Insektenkundc, vol. 6, pp. 19S. Type species: Apis bicornis Linnaeus, 1758, = Apis rufa Linnaeus, 1758, t)y designation of Latreille, 1811. Ceratosmia Thomson, 1872, Ilymenoptcra Scandinaviae, vol. 2, p. 232; Schmiedeknecht, 1884-86, Apidae Europaeae, vol. 2, p. 885; Ashmead, 1899, Trans. Amer. Ent. Soc., vol. 26, p. 74; Robertson, 1903, Trans. Amer. Ent. Soc., vol. 29, pp. 165; Cockerell, 1912, Proc. U. S. Nat. Mus., vol. 42, p. 216; Sandhouse, 1939, Mem. Ent. Soc. Washington, vol. 1, p. 9. Type species: Apis bicornis Linnaeus, 1758, = Apis rufa Linnaeus, 1758, by des- ignation of Sandhouse, 1939. Aceratosmia Schmiedeknecht, 1884-86, Apidae Europaeae, vol. 2, p. 885. Type species: Osmia emarginata Lepeletier, 1841, by designation of Sandhouse, 1939. Pachyosmia Ducke, 1900, Berichte des naturwissenschaftlich-medizinischen Vereines in Innsbruck, vol. 25, p. 18. Type species: Apis rufa Linnaeus, 1758, by designation of Sandhouse, 1939. Genus Osmia in Western Hemisphere 251 . The subgenus Osmia can be easily distinguished from all other subgenera in having long and loosely branched pubescence; a dis- tinct malar space, especially in the female; a round or elongate depression in the genal area of the female immediately below the eye; unmodified flagellar segments which, in the male, are twice as long as wide; a projection (weak in ribifloris) on the lower margin of the middle femur of the male; and male genitalia with elongate penis-valves provided with long ventral openings, and gonocoxites lacking hairs on the mesal surfaces. Male. — Length of body 8 to 11 mm. Pubescence white with black intermixed, or largely black. Metasomal terga without apical or subapical fasciae. Mandible widest at base, narrowest at region of constriction which is almost as wide as apex, with V-shaped notch separating lower tooth from acute upper tooth; eye subequal to width of genal area; inner margins of eyes converging anteriorly; clypeus thickly covered with white hairs arising from fine, con- fluent punctures, margin weakly wavy or serrate in median third, with moderately broad apical impunctate band; scape shorter than antennal segments 2 to 4 combined; flagellar segments 2 to 11 un- modified, each over twice as long as wide; hypostomal carina not greatly elevated, uniform or often slightly raised behind angle; middle leg with tarsal segments not swollen, femur with a projection on lower margin (weak in ribifloris); basitarsus of hind leg four to five times as long as wide, not toothed, parallel sided; strigilis with apical margin of velum truncate, inner apical angle of velum rounded, concavity on inner margin of velum deep, apex of malus may be produced into a long, sharp pointed spine. Forewing cells with moderately short, sparse hair, apical portion papillate, vein A is two to three times as long as vein B. Metasomal terga 2 to 5 with impunctate bands, at least on posterior terga, tergum 6 with apical margin lacking cmargination, turned up, forming a flange; tergum 7 with apex produced, midapex with shallow, wide cmargination, forming two short lobes; metasomal sternum 1 broadly rounded without midapical cmargination; sternum 2 with apical margin strongly convex and usually largely concealing sternum 3; sternum 3 with a wide and deep cmargination; sternum 4 with apical mar- gin broadly convex; apical margin of sternum 6 with a wide, median subtriangular production; sternum 8 with apical half narrowly tri- angular or broadly subtriangular. Gonocoxite long, slender, witli no hairs on mesal surface; volsella broadly conical; penis valve long, tapering strongly apically, nearly reaching to exceeding apices 252 The University Science Bulletin of gonocoxites (as in some Chalcosinia) , ventral opening extending nearly to apex. Female. — Length of body, 8 to 13 mm. Pubescence long, loosely branched, mixed black and white; metasomal terga without apical or subapical fimbriae; scopa black. Mandible broad, slightly con- stricted distad of base, gradually widened to apex, which is strongly oblique and twice as broad as constriction (in lignaria) or only slightly oblique and less than twice as wide as constriction (in ribi- floris); eye much narrower than genal area; inner margins of eyes parallel to slightly diverging; clypeus (short and modified in lig- naria) with punctures confluent, truncation simple or much modified, broader than margin from end of truncation to lateral angle of cly- peus; malar space distinct ( see also hucephala ) ; a round or elongate depression in genal area just below eye; hypostomal carina not greatly elevated, highest well behind angle; basitarsus of hind leg three to four times as long as wide, of uniform breadth or a little widened in middle, apical margin truncate to slightly rounded; hind tibial spurs weakly curved. Forewing cells with long and short hairs intermixed, sparse, dense in some areas, distal part with papillae fine, dense. Terga 2 to 5 with apical impunctate bands narrow or absent; tergum 6 with apical margin forming a flange. Species of Osmia proper are widespread in North America and Eurasia. American species: Osmia (Osmia) lignaria Say, ribifloris Cock- erell. The Old World species examined and included in this subgenus: Osmia (Osmia) cormita Latreille, emarginata Lepeletier, kohlii Ducke, longicornis Morawitz, rufa Linnaeus, taurus Smith, tricornis Latreille. American Species of Osmia of Doubtful Position In the course of examining species of the genus Osmia the author came across several which for some reason or other can not be def- initely placed as to subgenus. An annotated list of these species is given below. Specimens of the species marked with asterisks were not seen by the author. Comments about these are based upon the published descriptions. Osmia aglaia * Sandhouse, Chenosmia or Monilosmia? Osmia angustipes * Cockerell, Chenosmia? Osmia cockerelli Sandhouse, Chenosmia? Osmia enixa * Sandhouse. White ( 1951 ) placed the female de- scribed as this species under O. (Acanthosmioides) francisconis, but Genus Osmia in Western Hemisphere 253 excluded the male from Acanthosmioides. The present author agrees with White in this decision. The species may be a Chen- os^mia. Osmia felti Cockerell, Monilosmia? Osmia foxi * Cameron. It is impossible to assign this species to subgenus from the description. Osmia gaudiosa * Cockerell, Chenosmia? Osmia grindeliae Cockerell. Only one specimen was examined by me. It is 10 mm. long and has a moderately sharp tooth on the hypostomal carina. The genal area is wider than the eye. But the clypeal truncation is equal to the lateral margin of the clypeus. Thus it can not be definitely placed under Monilosmia until more specimens are studied. Osmia hemera Sandhouse, Chenosmia? Osmia hesperos Sandhouse, Monilosmia? Osmia intirbana Cresson, Chenosmia? Osmia iridis Cockerell and Titus, Monilosmia? Osmia lactis Sandhouse, Chenosmia? Osmia melanopleura * Cockerell, Chenosmia? Osmia mertensiae Cockerell, Chenosmia? Osmia mixta * Michener, Monilosmia or Chenosmia? Osmia nelsoni * Cockerell, Monilosmia? Osmia nigritida * Friese. It is impossible to assign this species to any subgenus from the description. Osmia novaescotiae * Cockerell, Monilosmia? Osmia pacifica * "Say," Packard, cannot be assigned to any sub- genus. Osmia paradisica Sandhouse, Monilosmia? Osmia phenax * Cockerell, Nothosmia? Osmia pulsatillae Cockerell, Chenosmia? Osmia pusilla Cresson, Chenosmia? Osmia solitaria * Sandhouse, Monilosmia? Osmia stasima * Lovell, Nothosmia? Osmia virga Sandhouse, Chenosmia? Osmia vallicola Cockerell, Chenosmia? ACKNOWLEDGMENTS The author is indebted to Professor Charles D. Michener, De- partment of Entomology, University of Kansas, for guidance in this work and to Dr. Wallace E. LaBerge, Deparment of Entomology, University of Kansas (now of Iowa State College), for helpful criticisms during the preparation of the manuscript. Deep apprecia- 254 The University Science Bulletin tion is also expressed to the Graduate School of the University of Kansas for fellowships which helped the steady progress of the work. Thanks are due to the following collectors and museum curators who have generously lent the material of their own or specimens in their charge: Dr. G. E. Bohart, United States Legume Seed Re- search Laboratory, Logan, Utah; Dr. G. D. Butler, University of Arizona, Tucson, Arizona; Dr. M. A. Cazier, American Museum of Natural History, New York City, New York; Dr. P. D. Hurd, Uni- versity of California, Berkeley, California; Dr. G. F. Knowlton, Utah State College, Logan, Utah; Mr. K. V. Krombein, United States Department of Agriculture, Washington, D. C; Dr. R. Lam- bert, Canadian National Collection, Ottawa, Ontario, Canada; Dr. L. W. Quate, University of Nebraska, Lincoln, Nebraska; Dr. E. S. Ross, California Academy of Sciences, San Francisco, California; Dr. R. C. Smith, Kansas State College, Manhattan, Kansas; Mr. R. E. Snelling, Turlock, California; Prof. P. H. Timberlake, Citrus Experiment Station, Riverside, California. It should be mentioned that while the classification here proposed does not agree with that of Sandhouse (1939), her work was most useful and was an essential basis for completion of the present study. In many cases characters which she used, and her expressions for their description, have been used with little change. LITERATURE CITED Bingham, C. T. 1897. The fauna of British India including Ceylon and Burma, Hymen- optera, vol. 1, .xxix + 579 pp., Taylor and Francis, London. Dalla Torre, C. G. de 1896. Catalogus Hynienopteroruni Leipzig, vol. 10, viii X 643 pp. DucKE, Adolf. 1911. Die Bienengattung Osmia Panz., Berichte des naturwissenschaftlich- medizinischen Vereines in Innsbruck, vol. 25, pp. 1-321. Friese, H. 1911. Megachilidae, Das Tierreich, Lief. 28, xxvi -\- 440 pp. Hurd, Paul D., and Charles D. Michener 1955. The nicgachiline bees of California ( Hymenoptera: Megachilidae), Bull. Cahfornia Insect Survey, vol. 3, pp. 1-247. Matthew, W. D. 1915. Chmate and Evolution, Ann. New York Acad. Sci., vol. 24, p. 171-318. Michener, Charles D. 1936. Some North American Osmiinae (Hymenoptera: Apoidea), Amer. Mus. Novitates, no. 875, pp. 1-30. 1941. A generic revision of the American Osmiinae with descriptive notes on Old World genera (Hymenoptera: Megachilidae), Amer. Midland Nat., vol. 26, pp. 147-176. Genus Osmia in Western Hemisphere 255 1944a. Comparative external morphology, phylogeny, and a classification of the bees ( Hymenoptera ) , Bull. Amer. Mus. Nat. Hist., vol. 82, pp. 157-326. 1944b. The distribution of the osmiine bees of the deserts of North Amer- ica, Amer. Nat., vol. 78, pp. 257-266. 1949a. Records and descriptions of American megachilid bees (Hymen- optera), Jour. Kansas Ent. Soc, vol. 22, pp. 41-59. 1949b. A revision of the American species of Dicenitosmki (Hymenop- tera: Megachilidae), Ann. Ent. Soc. Amer., vol. 42, pp. 258-264. 1949c. Parallelisms in the evolution of the saturniid moths, Evolution, vol. 3, pp. 129-148. 1951. Superfamily Apoidea, in Mucsebeck, C. F. W., Karl V. Krombein and H. K. Townes, Hymenoptera of America north of Mexico, Synoptic Catalog, U. S. Dept. Agric. Monographs, no. 2, pp. 1043- 1255. Robertson, Charles. 1903. Synopsis of Megachilidae and Bombidae, Trans. Amer. Ent. Soc, vol. 29, pp. 163-178. Sandhouse, Grace A. 1924. Bees of the genus Osmia in the collections of the California Academy of Sciences, Proc. California Acad. Sci., ser. 4, vol. 13, pp. 341-372. 1939. The North American bees of the genus Osmia (Hymenoptera: Apoidea), Mem. Ent. Soc. Washington, no. 1, ii -|- 167 pp. 1943. The type species of the genera and subgenera of bees, Proc. U. S. Nat. Mus., vol. 92, pp. 519-619. Sinha, Ranendra N., and Charles D. Michener 1958. A revision of the genus Osmia, subgenus Centrosmia, l^ni\-. Kansas Sci. Bull., vol. 39, pp. 263-303. White, Jimmy R. 1952. A revision of the genus Osmia, subgenus Acanthosmioides (Hy- menoptera: Megachilidae), Univ. Kansas Sci. Bull., vol. 35, pp. 219-307. Yasumatsu, K., and Y. Hirashima 1950. Revision of the genus Osmia of Japan and Korea (Hymenoptera: Megachilidae), Mushi, vol. 21, pp. 1-18. 256 The University Science Bulletin malar spine malus Fig. 1. Dorsal view of male genitalia, O. (Nothosmia) olbiventris. Fig. 2. Strigilis of female, O. (Euthosmia) glauca. Genus Osmia in Western Hemisphere 257 Figs. 3-38. Apex of left mandible, female: 3, O. (Diceratosmia) con- juncta; 4, subfasciata; 5, O. (Nothosmia) albiventris; 6, cerasi; 7, cordata; 8, distincta; 9, inspergens; 10, vmrginata; 11, pumila; 12, sandhouscae; 13, titusi; 14, O. (C/ienosrrjia) ca//a; 15, cauUcola; 16, clarescens; 17, cobaJtina; 18, coUinsiae; 19, cyanopoda; 20, dolerosa; 21, e.tjgnfl; 22, ilUnoensis; 23, (n- deprensa; 24, Zfle^a; 25, malina; 26, namda; 27, pentstemnni.s; 28, pro.tfm«; 29, regulina; 30, sanctae-rosae; 31, tersula; 32, trevoris; 33, zephyros; 34, O. {Eutho.mia) lignaria. Figs. 129-136. Tarsus of left middle leg, male: 129, O. {Centrosmia) austromaritima; 130, bakeri; 131, bucephola; 132, thysanisca; 133, pifcet; 134, vandykei; 135, O. longula; 136, odontogaster. Figs. 137-139. Middle femur of male: 137, O. (Monilosmia) atrocyanea; 138, O. (Osmia) lignaria; 139, ribifloris. Fig. 140. Ventral view of posterior leg showing the carina on coxa, O. ( Diceratosmia ) subfasciata. Figs. 141-151. Posterior basitarsus of male: 141, O. (Nothosmia) albi- ventris; 142, distincta; 143, marginata; 144, pumila; 145, sandhouseae; 146, ftfwsi; 147, O. (Chenosmia) kincaidii; 148, O. (Chalcosmia) coerulescens; 149, coloradensis; 150, georgica; 151, O. (Monilosmia) densa. THE UNIVERSITY OP KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 6 Notes on Oviposition and the Hatching of Eggs of Aedes and Psorophora Mosquitoes (Diptera: Culicidae)' BY A. Ralph Barr and Abdulla Al-Azawi Abstract: This paper describes a technique for inducing oviposition of aedine mosquitoes. A theory concerning oviposition behavior of these mos- quitoes is also presented. Finally, evidence is given to support the belief that "conditioning" and the presence of an adequate stimulus are necessary to pro- mote hatching of the eggs of Aedes and Psorophora mosquitoes. Eggs of Aedes canadensis react differently than do those of the other species tested; they must be subjected to influences which are still unknown before hatcliing will take place. INTRODUCTION Information concerning the laying and hatching of aedine eggs is meager and conflicting. For this reason the senior author has been carrying out incidental studies on the hatching of Aedes eggs for several years. The results presented in this paper are strictly preliminary but give some interesting leads for future work and confirm some of the results of Horsfall (1956). In addition it is hoped that the present work will stimulate others to investigate this phase of mosquito biology. MATERIALS AND METHODS The mosquitoes used in this work were Aedes vexans (Meigen), A. aegtjpti (Linnaeus), A. canadensis (Theobald), A. trivittatus (Coquillett), A. nigromacidis (Ludlow), A. cinereus Meigen, A. triseriatus (Say), and Psorophora longipalpis Roth. Field-collected females of all species (except A. aegypti) were used. All mos- quitoes were from Douglas Co., Kansas unless otherwise stated. 1. Contribution number 971 from the Department of Entomology, University of Kansas, Lawrence, Kansas. (263) 264 The University Science Bulletin Female mosquitoes were captured in test tubes ( 16 mm. diam- eter x 15 cm. length) while feeding in the field. They were dem- onstrated to have been inseminated either by examination of the spermathecae or by hatching some of their eggs (parthenogenesis does not normally occur in mosquitoes although it may be in- duced [Laven, 1956]). The bottom of the test tube used for catching the mosquito was filled with moist cotton and was placed over the female while she was feeding. When she had finished feeding and had flown up into the tube, cheesecloth was used to close the mouth of the tube; it was secured with a rubber band. When a less porous material was used to close the tube moisture condensed in the tube which proved detrimental to the mosquito. The test tubes with their occupants were stored on their sides in the laboratory at room temperature (24 to 26° C); all experiments were carried out at this temperature. Females usually oviposited well after a single feeding; many refed readily through the cheese- cloth and laid additional clutches of eggs. Females have also been kept in lamp-chimneys at times but were usually more reluctant to oviposit in such a container. The smaller size of the test tube prob- ably "forces" oviposition to a certain extent. All eggs laid were kept in contact with water for at least several days to prevent desiccation. They were then stored in an insectary at a relative humidity of about 70-90%. The eggs used in this study varied in age from about a week to two months; most were 2 to 4 weeks old. It would appear that the eggs do not become less viable in this length of time. All eggs were "conditioned" ( vide infra ) for at least four days unless otherwise stated. INSEMINATION In collecting eggs from mosquitoes one should first make certain that the mosquitoes have been inseminated. This can be quickly done by examining the spermathecae for the presence of sperma- tozoa. Throughout this work females have been checked from time to time and it has been found that the Aedes and Psorophora fe- males taken attacking man in the field almost invariably have been inseminated. An example is given from data collected in Minnesota in June and July of 1952. Dissections were made of 24 females of Aedes vexans and 4 A. cinereus. All 28 females had been insemi- nated; in most cases the large and one of the small spermathecae contained spermatozoa. The second small spermatheca was usually Aedes and Psorophora Mosquitoes 265 either empty or contained only a few spermatozoa. (Occasionally a female had spermatozoa only in the large spermatheca or in all three spennathecae. ) All of these females were in practically per- fect condition ( i. e., were almost entirely unrubbed ) indicating that they had emerged recently. In the species the authors have studied it appears that females usually mate within a short period after emergence, and usually before they seek a host. RESULTS OVIPOSITION In early experiments the senior author encountered difficulty in inducing females to oviposit in the laboratory. With Aedcs aegypti the substrate usually provided for oviposition is damp filter paper but most species of Aedes and Pfiorophora appear reluctant to ovi- posit on such a medium. This difficulty was inadvertently over- come when moist cotton was provided to humidify a cage; it was found that females oviposited readily on this medium. It seems likely that females which lay dormant eggs around the edges of ground pools actually insert their eggs into crevices in the ground; this process has been described for Aedes nigromacidis by Hus- bands and Rosay ( 1952 ) . This would explain why such mosquitoes will oviposit readily on rough surfaces but not on smooth ones ( Deckel, 1955 ) ; it would also explain why Aedes polynesiensis pre- fers to lay its eggs in creases in filter paper ( Wallis, 1954 ) . Aedes aegypti females, which frequently oviposit on smooth containers, probably are less fastidious in this respect. The authors have ob- served that A. triseriatus, a treehole mosquito, also prefers to lay its eggs in cracks in wood. If the above theory is correct, it might be predicted that the eggs of ground pool mosquitoes are more sus- ceptible to the eflFects of drying than is true of A. aegypti. This point is being investigated further in this laboratory at the present time. Most of the mosquitoes utilized fed once before depositing a clutch of eggs and laid about a hundred eggs after a period of about 6 days. A good proportion refed and laid a second batch with the technique employed. Oviposition was observed in Aedes vexans which laid about 1 to 2 eggs per minute. The eggs began darken- ing about 57 minutes after laying. "Conditioning" The normal procedure for hatching Aedes eggs is as follows. Eggs must be kept moist for several days after laying. If they are dried during the first few days (48 hours in A. aegypti [MacGregor, 266 The University Science Bulletin 1916] ) they will be killed. After the embryo has developed to a certain point (wliich has not been fully elucidated) the egg be- comes more resistant to drying. The egg can then be "conditioned" by exposing it to air for a few days; "conditioned" eggs hatch readily when placed in water to which a suitable "hatching stimulus" has been added. The "conditioning" process is necessary for inducing uniform hatching. This was shown by an experiment in which 130 eggs of Aedes trivittatus were submerged in water immediately after lay- ing. After 37 days only 15 had hatched. In another experiment involving A. vexans (from Minnesota) 2,350 eggs were submerged in tap water soon after laying. In a period of 9 days 269 larvae (12.7%) hatched. Batches of these eggs were transferred to bowls for further experiments and of 6 batches of 100 eggs, none hatched in 6 to 9 days. Neither ascorbic acid nor glutathione was effective in hatching these unconditioned eggs although exposure to a partial vacuum induced 44 of 100 to hatch. These data and others not presented demonstrate that most eggs which have not been exposed to air ("conditioned") will not hatch, even in the presence of a hatching stimulus. This is true of Pso- t^ophora longipalpis as well as of the two species mentioned above. The Effect of Submersion of Eggs in Water After Conditioning Eggs of Aedes vexans, A. trivittatus, A. canadensis, and Psoroph- ora longipalpis were submerged in 5 cc. tapwater under the follow- ing conditions: 1. Water aerated (by bubbling air through it) a half hour before using. 2. Unaerated water. 3. Water boiled one quarter hour before using (exposed to air during experiment). 4. Water boiled one half hour before using (exposed to air dur- ing experiment). Other groups of eggs were placed in different volumes ( 1, 2, or 3 cc.) of tapwater to test the effect of depth of the water. The re- sults of these tests are given in table 1. It can be seen that tapwater alone is not effective in inducing hatching of Aedes or Psorophora eggs. Eggs hatched in only two tests and then only 7 of 19. This point is illustrated further by an experiment on Psorophora longipalpis eggs. One hundred eighty- eight eggs of this species were submerged in distilled water imme- Aedes and Psorophora Mosquitoes 267 Table 1. — The Effectiveness of Tapwater in Hatching Aedine Eggs.* Species A. vexans A. (rivittatus A. canadensis P. long ipalpis Time (days)t. . aerated 3'^ hr. . . unaerated ..... boiled }4 hr boiled }/2 hr. . . . 1 0/10 0/10 0/10 0/10 0/9 3/9 4/10 1 0/10 0/10 0/10 0/10 0/10 0/10 1 0/5 0/5 0/5 1 0/10 0/8 0/9 0/11 0/13 0/12 0/8 5 0/10 0/8 0/9 0/11 1 cc. volume. . . 2 cc. volume. . . 3 cc. volume. . . 0/10 0/10 0/10 0/10 0/10 0/10 0/5 0/5 0/5 0/13 0/12 0/8 * The values given are the proportions which hatched; thus 0/10 is none of 10. t Length of time after which the observation was made. diately after laying. After 6 to 9 days none had hatched. They were then conditioned for 4 days, split into two parts and half were put into distilled water and half in "grass extract" ( vide infra ) . In 24 hours 40 of 78 had hatched in the grass extract but none of 78 in distilled water. If a hatching stimulant had been added to the batches of eggs tabulated in table 1, most of the eggs of Aedes vexans, A. trivittatus, and Psorophora longipalpis would probably have hatched. The eggs of A. canadensis would not have hatched (vide infra). These experiments actually served as controls for the experiments utilizing corn broth, grass extract, and urine which are discussed below. Plant Materials as Hatching Stimulants Since Horsfall (1956) has indicated that the broth from canned corn was effective in promoting hatching, a series of trials was made with this medium. A can of creamed corn * was opened and used as quickly as possible. The broth was filtered through cheesecloth and the filtrate diluted with tapwater to the concentrations given in table 2; the final volume in all cases was 5 cc. From this table it can be seen that fresh "corn broth" is effective in hatching the eggs of Aedes vexans, A. trivittatus, and Psorophora longipalpis in concentrations as low as .05%. It appears that at the lower concentrations the time required for this response may be lengthened. These results fully confirm the results of similar ex- periments by Horsfall (1956). It will be noticed that none of the eggs of Aedes canadensis hatched under these conditions. These eggs were all laid by a single female. In further experiments a very few eggs hatched, * Del Monte cream style sweet com. 268 The University Science Bulletin demonstrating that the female had been inseminated. This species is commonly found only in the early spring and thus presumably has a single generation a year (although there are anomalous records). The results given above lend support to the idea that this form has an obligatory diapause in the egg stage. Tarle 2. — The Effectiveness of Corn Broth in Hatching Aedine Eggs.* Species A. vexans A. Irivittatus P. longipalpis A. canadensis Time (days) . . . Concentration: (percent) 5.00 4.00 3.00 2.00 1.00 0.50 0.33 0.20 0.10 0.05 0.02 1 9/10 11/12 19/19 11/11 11/11 7/8 8/10 8/10 6/9 5/10 0/11 1 6/10 8/10 7/10 6/10 7/10 5/10 5/10 4/10 3/10 1/11 0/10 5 6/10 8/10 10/10 6/10 9/10 9/10 9/10 8/10 5/10 11/11 2/10 1 9/10 12/13 7/10 7/10 5/10 10/11 10/10 10/11 6/10 2/10 0/11 5 9/10 12/13 7/10 7/10 5/10 10/11 10/10 10/11 8/10 7/10 2/11 1 0/5 0/5 0/6 0/5 0/5 0/5 0/5 0/5 0/5 0/5 0/5 * The controls for these experiments are shown in table 1. Grass Extract There are many reports in the literature indicating that vegeta- tion might promote the hatching of eggs. We therefore used ex- tracts of grass {Setaria glauca [Linnaeus]) in hatching eggs. Ap- proximately 200 cc. of loosely packed grass blades was macerated in 10 cc. tapwater and strained through cheesecloth. The filtrate was then diluted with tapwater to the appropriate concentration; the final volume was 5 cc. in all cases. The results of these tests are given in table 3; all readings were made after 24 hours. As can be seen from the table, grass extract is also effective in promoting hatching of eggs of Aedes trivittatiis and Psorophora longipalpis. In other tests not listed the same was found true for A. vexans, A. triseriatus, and A. aegijpti. A single larva of A. canadensis hatched which shows the eggs of this species were fer- tile but most were not in a "hatchable" condition. Effect of Alfalfa Pellets Commercial rabbit pellets made of dried alfalfa are used in this laboratory for feeding mammals and mosquito larvae. A group of these was crushed and sieved through screen wire. Five grams of Aedes and Psorophora Mosquitoes 269 Table 3. — The EflFectiveness of Grass Extract in Hatching Aedine Eggs Species A. trivittatus P. longipalpis A. canadensis Concentration: (percent) 100.0 0/10 5/10 5/10 3/10 1/10 2/10 1/10* 5.0 4.0 3.0 2.0 1.0 10/10 7/10 7/10 7/10 8/10 7/10 0/5 1/5 0/5 0/5 0/5 0.5 0/5 * An additional larva hatched after 3 days. this material was macerated in tapwater and then strained through cheesecloth. This stock solution was diluted with tapwater to the appropriate concentration. Five cc. of the final concentration was used in each test and readings were taken after 24 hours. The re- sults are given in table 4. Alfalfa pellets also are effective in stimulating the hatching of aedine eggs although possibly to a lesser extent than fresh grass. Similar observations have been made with Aedes aegtjpti in this laboratory. The foregoing tests indicate that a variety of plant materials are capable of inducing hatching of dormant eggs of many species of aedine mosquitoes. Table 4. — The Effectiveness of Dried Alfalfa Pellets in Hatching Aedine Eggs Species A. vexans A. trivittatus P. longipalpis Concentration : (percent) 5.00 4.00 3 . 00 3/10 6/10 8/10 9/10 4/10 10/10 10/10 10/10 9/10 8/10 0/10 0/10 0/10 0/10 0/10 6/10 7/10 4/10 3/10 3/10 0/10 1/10 1/10 0/10 1/10 1/10 0/10 3/10 2/10 6/10 3/10 2/10 6/10 5/10 0/10 0/10 2.00 1.00 .50 .33 .20 .10 .05 .02 .00 270 The University Science Bulletin The Effect of Reduced Oxygen Concentration The dissolved oxygen in water may be easily removed by ex- posure to a vacuum. This fact is utilized in some laboratories for producing larvae of uniform age. The method works well with Aedes aegypti in the experience of the authors. Some of the control eggs from the experiments given above were tested for viability by exposure to a reduced oxygen concentration. The tubes were merely placed in a desiccator which was evacuated with a water aspirator. The equipment used was faulty but served the purpose although the amount of vacuum used could not be measured. Tliis treatment appears to have no deleterious effects on mosquitoes. Of the eggs tested 63 of 78 Psorophora longipalpis hatched, 20 of 30 Aedes trivittatus, and 2 of 35 A. canadensis. Effect of Reducing Agents Since there are indications in the Hterature that substances which lower the oxidation-reduction potential of a medium are useful in promoting hatching of eggs (Gjullin, Hegarty, and Bollen, 1941), a series of reducing agents was tested. These were: 1. hydrogen. 2. metals: iron, copper, tin, and zinc. 3. other chemicals: ammonium iodide and oxalic acid. Hydrogen was bubbled through the water containing eggs; the amount used was not measured nor was there a control to measure the effects of agitation although the gas was bubbled through the top of the water to reduce the effects of agitation. The other re- agents were made up by using 25 to 500 mgm. in 5 cc. of distilled water which gave concentrations of 0.500 to 0.025% for the soluble compounds. The metals were in as finely divided a state as possible ( filings or small particles ) . All tests were done with Aedes vexans. Hydrogen was bubbled through water containing "conditioned" eggs and of the 30 eggs used ( in three tests ) 21 hatched within 15 minutes (5, 7, and 9 of the 10 in each test). Another group of 21 eggs was not "conditioned" but had been continuously submerged since laying (14 days); of 21 such eggs tested only one hatched. Since no catalyst was used in this experiment the hydrogen may have acted not by lowering the oxidation-reduction potential but by displacing a part of the oxygen in solution. Of the metals tested neither tin nor copper induced hatching of eggs; zinc and iron, on the other hand, were effective (table 6) but the larvae were killed by exposure to zinc. Of the other materials tested ammonium iodide did not promote hatching; oxalic acid did Aedes and Psorophora Mosquitoes 271 induce hatching (table 5) but killed the larvae when they were half out of their shells. There was no hatching in distilled water controls but satisfactory hatching in grass extract controls. The reducing agents which were ineffective may have been too weak or may have killed the larvae before they were able to hatch. The former hypothesis is the more likely one but the viability of the unhatched eggs was not determined. A rusting iron pan was tested as a routine egg-hatcher and worked fairly well but slowly. Eggs did not hatch when placed in water in an aluminum pan. Table 5, — EflFectiveness of Reducing Agents in Hatching Eggs of Aedes vexans Reducing Agent Weight used (mgm.) 24 hrs. 48 hrs. Reducing Agent Weight used (mgm.) 24 hrs. 48 hrs. Iron 25 50 100 500 8/10 4/10 6/10 5/10 8/10 8/10 8/10 5/10 Oxalic acid 25 50 100 500 3/10 0/10 5/10 7/10 25 50 100 500 1/10 5/10 7/10 3/10 4/10 5/10 8/10 4/10 Distilled water 0/20 0/5 0/20 Zinc Grass extract 16/30 4/6 18/30 Urine Since Aedes vexans larvae are frequently encountered in huge numbers in water fouled with urine and feces of cows, urine was tested as a possible stimulant. Male human urine was collected and diluted with tap water to the desired concentration; the final Table 6. — The Effectiveness of Urine in Hatching Aedine Eggs Species A. vexans A. trivittatus P. longipalpis* A. canadensis* Concentration : (percent) 10 0 8/10 8/10 6/10 6/10 7/10 7/10 2/lOt 4/10 0/4 9/9 8/10 7/10 7/10 11/11 7/9 7/10 0/5 5 0 0/5 4.0 0/5 3.0 0/5 2.0 0/5 1.0 .5 .1 8/9 9/11 7/11 0/5 0/5 1/5 * No additional larvae after 2 days. \ Seven additional larvae after 2 days. 272 The University Science Bulletin volume was 5 cc. in all cases. Observations were made after 24 hours. The results of using this medium are shown in table 6. These tests indicate that urine is effective in promoting hatching of eggs of Aedes trivittatus, A. vexans, and Psorophora longipalpis but not of A. canadensis. Tests were made of urea in concentra- tions of 5.00 to .02% to see whether or not it was the active ingredient of urine. No hatching occurred in these tests but they were in- conclusive since there was poor hatching in the controls. Aedes nigromacidis A female laid 156 eggs on the surface of tapwater where they re- mained 13 days with none of the eggs hatching. During this time a scum developed on the water. After the above period the eggs were submerged in the water. Hatching began immediately. In 24 hours 139 of 156 had hatched and after 48 hours 141 of 156. The water used in this test developed a considerable amount of bacterial growth although no food had been added. DISCUSSION All evidence indicates that several steps are necessary for uniform hatching of eggs of multivoltine aedine mosquitoes. 1. The eggs must be in contact with free water while undergoing a certain amount of their development. This period has not been extensively studied but appears to be about 48 hours in Aedes aegypti (MacGregor 1916). 2. To become "hatchable" eggs must be "conditioned"; i. e., ex- posed to air for a certain period of time. The duration of this period also has not been extensively investigated. 3. To hatch eggs it is necessary to use a "hatching stimulus" of some type. It appears likely that all of these stimulants lower the oxidation-reduction potential of the medium ( Gjullin, Hegarty, and Bollenl941). In this study tests on Aedes vexans, A. trivittatus, A. aegypti, and Psorophora longipalpis indicated that the above steps are necessary for inducing hatching; one group of Aedes nigromacidis eggs re- acted similarly. Horsfall (1955) has indicated that most multivol- tine aedines correspond to this pattern. There are a large number of aedines, however, which appear to have a single generation a year and whose eggs will not hatch when put through the above regime. Aedes canadensis is probably one of these univoltine forms although there are anomalous findings in its behavior (large fall populations, for example, are frequently encountered). Our re- Aedes and Psorophora Mosquitoes 273 suits with canadensis eggs confirm tliis hypothesis although the eggs were all derived from a single female. The techniques described in this paper should be of value to culicidologists since larvae can be easily secured from biting females of most southern species of Aedes and Psorophora. This is useful in identifying questionable females and in securing larvae where none can be found in the field. LITERATURE CITED Beckel, W. E. 1955. Oviposition site preference of Aedes mosquitoes ( Culicidae ) in the laboratory. Mosquito News, vol. 15, pp. 224-228. GjuLLiN, C. M., C. P. Hegarty and W. B. Bollen. 1941. The necessity of a low oxygen concentration for the hatching of Aedes mosquito eggs. Journ. Cell, and Comp. Physiol., vol. 17, pp. 193-202. HORSFALL, W. R. 1955. Mosquitoes. New York: Ronald Press, pp. viii + 723. 1956. Eggs of floodwater mosquitoes III (Diptera, Cuhcidae). Condi- tioning and hatching of Aedes vexans. Ann. Entom. Soc. Amer., vol. 49, pp. 66-71. Husbands, R. C, and B. Rosay. 1952. A co-operative ecological study of mosquitoes of irrigated pastures. Calif. Mosq. Cont. Assoc, Proc. and Papers 20th Ann. Conf., pp. 17-26. Laven, H. 1956. Induzierte parthenogenese bei Culex pipiens. Naturwissenschaf- ten, vol. 43, pp. 116-117. MacGregor, M. E. 1916. Resistance of the eggs of Stegomyia fasciata (Aedes calopus) to conditions ad\'erse to development. Bull. Entom. Res., vol. 7 pp. 81-85. Wallis, R. C. 1954. Observations on oviposition of two Aedes mosquitoes. Ann. Entom. Soc. Amer., vol. 47, pp. 393-396. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 7 A Revision of the Genus Osmia, Subgenus Centrosmia ( Hymenoptera : Megachilidae ) ^ BY Ranendra N. Sinha - and Charles D. Michener Abstract: This paper consists of a taxonomic revision of the nine species of the subgenus Centrosmia. Keys to the species, detailed descriptions, and distributional data are given. The following new synonymy is indicated: Osmia universitatis Cockerell, integrella Cockerell, amala Cockerell, and metitia Cockerell are all synonyms of O. pikei Cockerell. INTRODUCTION This work is a taxonomic revision of bees of the subgenus Cen- trosmia Robertson of the genus Osmia Panzer. The name Centros- mia was proposed by Robertson ( 1903 ) for Osmia bucephala Cres- son. Sandhouse (1939) placed it as a synonym for Meladosmia Schmiedeknecht. Sinha (1958) recognized Centrosmia as a dis- tinct subgenus and hsted eight more species, most of which Sand- house had i^laced in Nothosmia. As here understood, therefore, the subgenus includes a section of the Nothosmia of Sandhouse's revi- sion, as well as two species which she placed in Melanosmia. Most species of this subgenus are easily distinguishable by the swollen second and third tarsal segments of the middle leg of the male and by a protuberance or ridge near the base of the mandible and a thickened apical truncation of the clypeus in the female. A full description of the subgeneric characters, in comparison to those of other subgenera, will be given in a forthcoming paper (Sinha, 1958). An adequate discussion of its relationships with other subgenera cannot be given here since it would involve groups (subgenera) which will not be named or defined imtil the appear- ance of the paper by Sinha already mentioned. 1. Contribution number 970 from the Department of Entomology, University of Kansas. Lawrence. 2. Present address: Department of Zoology, McGill University, Montreal, Canada. (275) 276 The University Science Bulletin Eight of the species of this subgenus are extremely scarce, and only a few dozen specimens of these species have been seen after examining approximately 25,000 specimens of Osmia collected from the United States and Canada. Specimens of the species bucephala are somewhat less scarce. The greatest difficulty was encountered in the association of the sexes. Although both of the sexes of the type species, bucephala, and of its relative, nigriventris, were known, the seven other species were described from the male only. The present authors were able to recognize a female of iinwersitatis (= pikei). This sex associa- tion was confirmed by comparing the common subgeneric characters with the female of bucephala. Now that the subgeneric characters of the female are better established, it will be easier to associate the females of the remaining species when they are collected. Ocellocular, postocellar and ocelloccipital lines are measured from the center of each ocellus. The morphological terminology used in the descriptions is from Michener ( 1944 ) . References to species in the catalogues and general treatments of Dalla Torre (1896), Cresson (1887), Friese (1911), and Michener (1951) have been omitted in the synonymies, as these are bibliographic treat- ments wliich list or describe all species described before their dates of publication, without addition of any new information. It will be noted that the present paper contains no illustrations. This is because adequate illustrations for understanding the charac- ters of the species of Centrosmia were presented by Sandhouse (1939), Michener (1957), and Sinha (1958). The authors wish to thank Dr. Wallace E. LaBerge, formerly of the University of Kansas but now of Iowa State College, for his constructive criticism in the preparation of the manuscript, and Mr. Karl V. Krombein for allowing them to see the t>'pes and the col- lections of Osmia in the United States National Museum, Washing- ton, D. C. Key to the Species of the Subgenus Centrosmia Males 1. Middle leg with tarsal segments 2 and 3 unmodified; metasomal tergum 6 with apical impunctate band poorly defined, narrow or absent, distinctly different from bands of preceding terga 2 Middle leg with tarsal segments 2 and 3 swollen or otherwise modified, at least slightly broader than basitarsus; metasomal tergum 6 with apical impunctate band wide like that of pre- ceding terga or tergal margin reflexed 3 Revision of Genus Osmia 277 2. ( 1 ) Entirely black; abdomen widest near posterior end; metasomal tergum 6 with apical portion strongly reflexed forming a flange, especially laterally nigriventris Body metallic bluegreen; abdomen widest near middle; meta- somal tergmn 6 not reflexed bakeri 3. (1) Size large (length usually 10 mm. or more); abdomen widest near posterior end; metasomal tergum 6 with apical portion strongly reflexed forming a flange bucephala Size smaller (length usually less than 10 mm.); abdomen widest in the middle; metasomal tergum 6 with apical jwrtion not or but little reflexed 4 4. (3) Last flagellar segment expanded and compressed; metasomal tergum 6 with a median basal tuft of long, posteriorly directed hairs; tergum 7 with greatly thickened lateral margins. vandijkei Last flagellar segment unmodified; metasomal tergum 6 lacking a median basal tuft of long, posteriorly directed hairs; tergum 7 with lateral margins little thickened 5 5. ( 4 ) Hind leg with basitarsus strongly expanded, greatest width about one third of length tanneri Hind leg with basitarsus not or little expanded, greatest width less than one third of length 6 6. (5) Metasomal tergum 7 with a sharply pointed tooth on each side of deep midapical emargination; clypeal margin with median third produced, apex of projection weakly emarginate. aiistromaritima Metasomal tergum 7 with a blunt lobe on each side of shallow midapical emargination; clypeal margin without midapical projection 7 7. (6) Seventh tergum with emargination shallowly arcuate, margins of tergum lateral to teeth not strongly convex; first flagellar seg- ment (on shortest side) distinctly shorter than second; ster- num 4 with apex conspicuously truncate; sternum 6 having apical process with lateral margins diverging posteriorly, apex rounded pikei Seventh tergum with emargination deep, margins of tergum lat- eral to teeth strongly convex; first flagellar segment (on short- est side) scarcely shorter than second; sternum 4 with apex rounded or subtruncate; sternum 6 having apical process parallel sided or narrowing posteriorly, apex not rounded 8 8. (7) Emargination of sternum 3 very shallow, less than one sixth as wide as sternum, fringed with whitish hairs; sternum 4 broadly rounded; sternum 6 having apical process emarginate at tip thysanisca Emargination of sternum 3 deeper, nearly one fourth as wide as sternum, fringed with yellowish hairs; sternum 4 subtruncate; sternum 6 having apical process truncate raritatis 278 The University Science Bulletin Fenwles 1. Second mandibular tooth nearly as large as the others so that the mandibles arc clearly four-toothed; body entirely without metallic coloration nigriventris Second mandibular tooth much smaller than adjacent teeth; body with at least some areas of metallic blue or green color- ation 2 2. ( 1 ) Mandible with a protuberance at upper basal angle, without ridge at lower basal angle; malar space broad; hind tibia with spurs strongly curved near apices bucephala Mandible without protuberance at upper basal angle, with trans- verse ridge near lower basal angle; malar space inconspicuous; hind tibia with spurs almost straight near apices pikei Osmia {Centrosmia) bucephala Cresson Osmia bucephala Cresson, 1864, Proc. Ent. Soc. Philadelphia, vol. 3, p. 17; Cresson, 1875, Rep. U. S. Georgr. Surv. West of 100th Meridian, p. 724; Robertson, 1895, Trans. Amer. Ent. Soc, vol. 22, p. 125; Evans, 1896, Ca- nadian Ent., vol. 28, p. 13; Friese, 1902, Fauna arctica, Bd. 2, p. 482; Lovell and Cockerell, 1907, Psyche, vol. 14, p. 17; Graenicher, 1910, Bull. Public Mus. City Milwaukee, vol. 1, p. 245; Cockerell, 1911, Canadian Ent., vol. 43, p. 391; Gibson, 1915, 46th Ann. Rep. Ent. Soc. Ontario, p 30; Cresson, 1916, Mem. Amer. Ent. Soc, no. 1, p. 113; Sandhouse, 1925 Canadian Ent., vol. 57, p. 36; Leonard, 1926, Cornell Univ. Agr. Expt. Sta. Mem. 101, p. 1030; Cockerell, 1928, Univ. Colorado Studies, vol. 16, p 116, Beck, 1929, Bull. Brooklyn Ent. Soc, vol. 24, p. 304; Cockerell, 1934 Amer. Mus. Nov., no. 697, p. 13; Graenicher 1935, Ann. Ent. Soc. Amer. vol. 28, p. 305; Brimley, 1938, Insects of North Carohna, p. 458; Proctor. 1938, Biol. Surv. Mt. Desert Region, vol, 6, p. 445; Bohart, Knowlton, and Bailey, 1950, Utah State Agr. College, memeo. series no. 371 (female). Osmia (Centrosniia) bucephala, Viereck, 1916, Connecticut Geol. and Nat. Hist. Surv., Bull. 22, p. 750. Osmia (Melanosmia) bucephala, Sandhouse, 1939, Mem. Ent. Soc. Washing- ton, no. 1, p. 37. Osmia megacephala Cresson, 1864, Proc. Ent. Soc. Philadelphia, vol. 3, p. 18; Putnam, 1876, Proc. Davenport Acad. Nat. Sci., vol. 1, p. 195; Cresson, 1876, Proc. Davenport Acad. Nat. Sci., vol. 1, p. 209; Viereck, in Skinner, 1903, Trans. Amer. Ent. Soc, vol. 29, p. 47; Cockerell, 1906, Trans. Amer. Ent. Soc. vol. 32, p. 305; Cresson, 1916, Mem. Amer. Ent. Soc, no. 1, p. 123; Cockerell, 1919, Ent. News, vol. 30, p. 294; Clements and Long, 1923, Experimental Pollination, Carnegie Inst. Washington Publ. 336, p. 253; Cockerell, 1928, Univ. Colorado Studies, vol. 16, p. 122; Cockerell, 1934, Amer. Mus. Nov., no. 697, p. 14 (female). Osmia latitarsis Cresson, 1864, Proc. Ent. Soc. Philadelphia, vol. 3, p. 20; Cres- son, 1878, Trans. Amer. Ent. Soc, vol. 7, p. 105; Provancher, 1888, Ad- ditions et Corrections av volume II de la Faune Entomologique du Canada, p. 327; Cresson, 1916, Mem. Amer. Ent. Soc, no. 1, p. 123 (male). Osmia lignivora Packard, 1867, Amer. Nat., vol. 1, p. 376; Ashmead, 1894, Psyche, vol. 7, p. 39; Washburn, 1919, 17th Report State Entomologist Min- nesota, p. 233; Cockerell, 1907, Univ. Colorado Studies, vol. 5, p. 37 ( female ) . Osmia lignicola Provancher, 1882, Nat. Canad., p. 208; Provancher, 1883, Faune Entomologique du Canada, Hymenopteres, p. 708 (female). Osmia tarsata Provancher, 1888, Additions et Corrections au volume II de la Faune Entomologique du Canada, p. 328 (male). Revision of Genus Osmia 279 Osmia suhornata Cockereli, 1897, Proc. Acad. Nat. Sci. Philadelphia, vol. 49, p. 342; Sandhouse, 1924, Proc. California Acad. Sci., ser. 4, vol. 13, p. 361, 370. Centrosmia bucephala, Robertson, 1903, Trans. Amer. Ent. Soc, vol. 29, p. 170; Titus, 1906, Proc. Ent. Soc. Washington, vol. 7, p. 157; Robertson, 1928, Flowers and Insects, p. 8. Centrosmia tarsata, Titus, 1906, Proc. Ent. Soc. Washington, vol. 7, p. 158. This is the largest and most highly modified species of the sub- genus. It has extensive black, nonmetallic areas, thus approaching nigriventris which is entirely black. The genal areas of the female are more than twice as wide as the eyes. This character alone dis- tinguishes bucephala from all the other species of the subgenus. The male is unique in having the fourth as well as the second and third tarsal segments of the middle leg much enlarged and modi- fied. Male: Length of body, 12 to 14 mm. Coloration: Black with strong greenish blue tints on face and dorsum of head and thorax, more bluish on abdomen; lateral and ventral surfaces of head and thorax usually nonmetallic but sometimes bluish; mandible black; flagellum dark brown above and black below; legs black. Pubes- cence: Hair of body white, mixed with fuscous on vertex and genal area; on terga 3 to 5 black or largely so; fuscous or black hairs present on tibiae and basitarsi and sometimes replacing white; hair on undersides of tarsi golden. Head: Mandible with a narrow V-shaped notch separating short upper tooth from long slender, lower tooth, apical edge above notch convex, superior apical angle forming a right angle; carinae parallel, upper carina wider than lower, space between them narrow; facial length and width as 8:6.6; postocellar line shorter than ocelloccipital line and subequal to ocellocular; clypeus convex, apex weakly wavy, punctures finer than those on rest of face, apical impunctate band over one-third as wide medially as greatest width of scape; scape shorter than an- tennal segments 2 to 4 combined; length and width of eye as 7.5:3.3; eye nearly as wide as genal area; eyes with inner margins weakly converging to parallel; hypostomal carina moderately low, a little raised well behind angle. Mesosoma: Middle leg with tibial spine nearly twice as long as that on foreleg; strigilis with apex of velum convex, inner apical angle broadly rounded, apex of mains produced into a medium-sized, slender, spine that is about one sixth the length of rest of mains; foreleg with segments of mediotarsus broadened and flattened, second tarsal segment about as broad as long, third shorter, fourth still shorter; middle leg with segments of mediotarsus 280 The University Science Bulletin greatly swollen, tarsal segments 2 to 4 as wide as or wider than long with inferior as well as lateral lobes, basitarsus twice as long as wide at its widest part; hind basitarsus narrowest at base, gradually widened towards apex; hind tibial spurs curved, inner much stouter than and nearly twice as long as outer. Forewing with length and breadth as 11:3.5, papillae on apex fine, moderately sparse, with hairs in cells short, sparse, dense in some areas. Propodeal triangle finely lined, shining below, dull and reticulate above. Metasoma: Terga 2 to 5 with apical impunctate bands one-fifth to one-third as wide as punctate portions, weakly lined, shining; tergum 5 with lateral mar- gins reflexed; tergum 6 with long hairs laterally, apical margin re- flexed, broadly convex with middle third subtruncate or gently con- cave; tergum 7 with long hair on sides bending mesad, surface con- vex, median emargination small and shallow. Sternum 1 with mar- gin convex, usually with minute midapical emargination; sternum 2 with margin convex, most strongly so medially although sometimes with small shallow midapical emargination, a subtriangular or rounded tumescence at or just anterior to midapex which bears fuscous or black hairs directed mesad and caudad and contrasting with adjacent white hairs; sternum 3 with midapical concavity deep, wide, fringed with moderately long light golden hairs directed mesad; sternum 4 with margin broadly rounded, one fourth of posteromedian area covered by a nearly semicircular patch of coarse black bristles directed to the rear and mesad; sternum 5 with small V-shaped median emargination; sternum 6 with apical margin sub- truncate laterally, median produced portion with sides parallel and apex emarginate on each side of broadly triangular middle portion; sternum 8 with apical half nearly truncate on each side of long, slender, median projection, basal half subtriangular. Gonocoxite broad at base with a ventral lobe at level of angle, with a few long bristles on mesal surface at point of divergence of this lobe, and a lateral patch of bristles on dorsal surface of angle; free portion of volsella blunt subconical; penis valve reaching beyond angle of gonocoxite, sides nearly parallel, apex slightly wider, tip broadly rounded. Female: Length of body 13 to 16 mm. Coloration: Black, usually with face, thoracic dorsum, and metasomal terga weakly greenish blue. Pubescence: Hair on face, vertex and metasomal tergum 2 black mixed with white; on dorsum and sides of thorax (only a few isolated black hairs), propodeum, metasomal tergum 1 and sometimes anterior half of metasomal tergum 2 white; on under- Revision of Genus Osmia 281 side of foretarsus golden brown (darker in some individuals); on lower surfaces of middle and hind tarsi dark reddish brown; on rest of body black. Head: Mandible with superior apical angle ( tooth 1 ) acute, tooth 2 missing or represented by feeble convexity nearer tooth 3 than 1; length of triangular tooth 3 about one third length of sharp, slender, triangular tooth 4 ( inferior apical angle of mandible); distance between teeth 1 and 3 over twice distance be- tween 3 and 4; mandibular carinae parallel, lower carina wider than upper; carina on upper edge weakly wavy, slightly curving down apically near base of tooth 1; mandible with apex about one and one-third times as wide as constriction which is slightly nar- rower than base; a distinct protuberance, highest near upper basal angle of mandible, occupies nearly three fourths of basal width of mandible and sharply cuts off a narrow unmodified basal zone; labial palpus with segment 1 three-fourths as long as 2; maxillary palpus with segment 3 a little longer than 2 and nearly as long as 4 and 5 together; face as long as wide; postocellar line little over half as long as ocellocular line; ocelloccipital line two to two and one- half times as long as postocellar; clypeus strongly convex basally, with punctures equal to those on anterior half of paraocular area, coarser than those on rest of face, medially longitudinally keeled and closely and rather finely punctate, apex of truncation broadly and strongly thickened and elevated, irregularly coarsely punctate, with distance between lateral angle of apical truncation and lateral angle of clypeus shorter than length of apical truncation; eyes with inner margins diverging below, each a little less than three times as long as wide; base of mandible nearly twice as wide as eye; malar space longer than flagellar width; flagellum with segment 1 a little less than twice as long as pedicel; hypostomal carina moderately low, slightly raised a short distance behind angle. Mesosoma: Foreleg with tibial spine subequal to that of middle leg; longest hair on fore basitarsus half as long as basitar«us; strigilis with apex of velum oblique, merging into malus, malar spine one third as long as length of malus, inner apical angle of velum more nearly an obtuse angle than rounded, inner basal angle rounded and more pointed than inner apical angle; hind tibial spurs long, slender, strongly curved apically. Forewing with length and width as 11:3.5, with papillae on apex fine, sparse, with hair in cells short and dense with few longer and sparse hairs in some areas. Mesoscutum shining, punc- tures fine and confluent anteriorly and laterally, separate in large discal area. Propodeal triangle reticulate in narrow upper portion, 282 The University Science Bulletin aciciilate below. Metasoma: Punctures shallow and widely sepa- rated, impunctate bands on terga 2 to 5 at least one-fourth as wide as punctate portions, sixth tergum rather finely and closely punctate in contrast to preceding ones. Type material: The holotype of bucephala is from Great Slave Lake, North West Territories, Canada; that of megacephala is from the Rocky Mountains in Colorado; and the lectotype of latitarsis is from New York; all of these are in the Academy of Natural Sciences of Philadelphia. The cotypes of lignivora from Lawrence, Massachusetts are in the Museum of Comparative Zoology, Cam- bridge, Massachusetts. The one labeled Type/1/540 is here desig- nated the lectoholotype. The other, Type/2/540, becomes a lecto- paratype. The types of Ugnicola and tarsata, both from Cap Rouge, Quebec, are in the Musee de la Province de Quebec. The holotype of stibornata from Olympia, Washington, is in the United States National Museum. Distribution: This species is widely distributed in North America, from coast to coast and Alaska and the North West Territories to New Mexico, North Carolina, and Oregon, limited to mountains in the southern part of the range. Specimens from the following locali- ties have been examined: Alaska: Matanuska, July 17, 1944 (J. C. Chamberlin); White Horse to Yukon Crossing, May 15 to 20, 1912 (J. M. Jessup). Yukon Territory: Deer Lake, White Pass, May 19, 1916 (J. A. Kusche). Northwest Territories: Fort Resolution, Great Slave Lake, June 26, 1903, on Taraxicum (Mack). British Columbia: Vancouver, June 14, 1902. Alberta: Bilby, June 25, 1924 (G. Salt); Edmonton, May 30, 1924 (G. Salt). Ontario: Merivale, May 12, 1930 (J. J. de Gryse). Quebec: Danford Lake, June 3, 1917 (F. W. L. Sladen). Nova Scotia: Kings County, May 30, 1931 and June 5, 1932 (G. E. Atwood). Washington: Olympia, May 22 and 25, 1894; Pullman, June 7, 1905. Oregon: Mt. Hood, 3000-6000 feet, June 24, 1925 (E. C. Van Dyke); Pamelia Lake, Mt. Jefferson, July 15, 1907 (J. C. Bradley); Steens Mountains (west side, above 8000 feet) July 12, 1927 (H. A. Scullen). Idaho: Franklin, June 1, 1948 (G. E. Bohart). Utah: Curran Creek Valley, Uinta National Forest, 8000 feet altitude, July 4, 1917 (J. Silber); North Fork of Duchesne River, July 13 to 14, 1927; Monte Cristo, July 3, 1950 (G. F. Knowlton and S. L. Wood); La Sal Mountains, Mt. Tukuhni- Revision of Genus Osmia 283 kivatz (V. M. Tanner). Montana: Lake St. Marys, Glacier Na- tional Park, June 1, 1938 (E. C. Van Dyke). Colorado: Elk Creek, Eraser, Grand County, July 7, 1927 (J. C. Bradley), Ute Creek, 9000 feet altitude, July (H. S. Smith); Russell, June 25, 1907 (L. Bruner); Science Lodge, west Boulder County, June 24, 1938 ( Helen Rodeck), July 8, 1936 (Helen Rodeck, U. Lanham), July 22-23, 1922 (L. O. Jackson); Zimmermans, August 26, 1935 (Figgins); Creede, 8844 feet altitude, August, 1914 (S. J. Hunter); Sargent, June 21, 1926 (E. G. Anderson); South Fork, Rio Grande, 37° 36' N, 106° 43' W., 8500 feet altitude, June 17, 1919. Iowa: Ames, May 6, 1897, at violet (E. D. Ball). Indiana: Elkhart. New York: Keeseville, June 2, 1918 (A. K. Fisher); Ithaca (Chittenden); Axton, June 12 to 22, 1901. Maine: Old Orchard, June 6, 1921 (T. B. Mitchell). Massachusetts: Amherst, June 27, 1904. Connecti- cut: Hartford, May 7, 1899. Pennsylvania: Linglestown, June 26, 1917 (W, S. Fisher). Maryland: near Plummers Island, April 25 and 28, 1915, May 3, 1916, on Mertensia virginica and Nepeta hederacea (J. C. Crawford). North Carolina: Durham, May 4, 1904 (J. R. Randall); Bryson City, June 8, 1923, on Tephrosia vir- giniana (J. C. Crawford); Cruso, June 28, 1934; Smokemount, June 30, 1934. The species has also been recorded from New Hampshire, Virginia, Tennessee, Illinois, and New Mexico. As Sandhouse (1939) has pointed out, there is a general ten- dency for western specimens to have more black pubescence than eastern ones and perhaps also less metallic coloration on the body surface. Thus females from the Appalachian region to Great Slave Lake usually have the mesoscutal pubescence entirely pale or nearly so, those from Colorado have a considerable area of mixed pubescence on the mesoscutum, while those from Washington state usually have a discal area of largely black pubescence and also have the pubescence of the face and vertex largely black. Males show similar tendencies. For example, the hair of the third to fifth terga is mixed black and white in the east, nearly all black from Colorado westward. The differences are probably sufficient to justify recog- nition of two subspecies, O. hucephala hucephala Cresson for the subarctic and eastern area and O. hucephala megacephala Cresson for the montane western area, but as adequate series for a study of variation in any one locality are not available, formal recognition seems premature and is not necessary in any event. 284 The University Science Bulletin Osmia (Centrosmia) nigriventris ( Zetterstedt ) Anthophora nigriventris Zetterstedt, 1838, Insecta Lapponica Descripta, vol. 1 p. 465. Osmia nigriventris, Nylander, 1848, Notiser ur Siillsk Fauna Fennica Forh., vol. 1, p. 260. Osmia (Melanosmia) nigriventris, Schmiedeknecht, 1884-86, Apidae Europaeae, vol. 2, p. 79; Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, p. 35. Osmia corticalis Gerstaecker, 1886, Stettiner Ent. Zeitung, vol. 30, p. 331. Osmia frigida Smith, 1853, Catalogue of Hymenopterous Insects in the Collec- tion of the British Museum, pt. 1, p. 142; Cresson, 1864, Proc. Ent. Soc. Philadelphia, vol. 3, p. 26; Evans, 1896, Canadian Ent., vol. 28, p. 13; Friese, 1902, Fauna Arctica, vol. 2, p. 483; Cockerell, 1905, Trans. Amer. Ent. Soc, vol. 31, p. 332. Sandhouse, 1925, Canadian Ent., vol. 57, pp. 36, 65. ( Provancher's [1883] records of this species relate to Megachile ac- cording to Titus, 1905). Osmia hudsonica Cresson, 1864, Proc. Ent. Soc. Philadelphia vol. 3, p. 21; Friese, 1902, Fauna Arctica, vol. 2, p. 483; Gibson, 1916, 47th. Ann. Rep. Ent. Soc! Ontario, p. 26; Cresson, 1916, Mem. Amer. Ent. Soc, no. 1, p. 120. In the above synonymy, no effort has been made to include refer- ences to all European citations of the species; only original descrip- tions and new combinations have been cited. This is the only wholly black, nonmetallic species of Centrosmia. In this respect it resembles only one other American Osmia, O. inermis (Zetter- stedt), which is morphologically like the group included by Sand- house (1939) in Nothosmia. O. nigriventris shows evidence of relationship with bucephala, e. g., the male hind basitarsus taper- ing toward the base, the enormous inner hind tibial spur of the male, the reflexed lateral margins of the fifth tergum and lateral and posterior margins of the sixth, in the male, etc. The unmodi- fied middle tarsi of the male and the presence of a weak transverse ridge rather than a large protuberance at the base of the mandible of the female distinguish this species immediately from bucephala. Male: Length of body, 11 mm. Coloration: Black, flagellum, tegulae, legs, and margins of metasomal segments brownish. Pubes- cence: White, light fuscous on second and following metasomal seg- ments and outer sides of tibiae, yellowish on under sides of tarsi. Head: Mandible with small narrow V-shaped notch separating short upper tooth from longer lower tooth, apical edge above notch con- vex, superior apical angle feebly acute; carinae subparallel, sub- equal in width, space between them narrow; facial length and width as 6:4; postocellar line shorter than ocelloccipital line and subequal to ocellocular; clypeus convex, apex wavy, punctures fine and very dense like those of supraclypeal area and lower part of paraocular area, apical impunctate band nearly two-thirds as wide as greatest Revision of Genus Osmia 285 width of scape; eye nearly as wide as genal area; inner margins of eyes strongly converging below; hypostomal carina low, a little raised behind angle. Mesosoma: Middle leg with tibial spine smaller than that of foreleg; strigilis with apex of velum subtruncate, inner apical angle broadly rounded, apex of mains produced to a medium sized spine about one-sixth as long as entire malus; foreleg with segments of mediotarsus slightly broadened, second and third tarsal segments slightly longer than broad; middle leg with medio- tarsus unmodified, second and third tarsal segments much longer than broad, basitarsus over four times as long as wide; hind basi- tarsus narrowest at base, gradually widened toward parallel sided apical region; hind tibial spurs curved, inner much stouter than outer and about 1.3 times as long as outer. Forewing with length and breadth as 3:1, papillae on apex very fine, moderately sparse, hairs in cells short, dense. Propodeal triangle dull and minutely granular, weakly striate along upper margin. Metasoma: Terga 2 to 5 with apical impunctate shining bands one fifth to one third as wide as punctate portions; tergum 6 with apical margin re- flexed, broadly so laterally, margin broadly convex except for small shallow median emargination; tergum 7 with shallow median emar- gination, somewhat as in pikei. Sterna 1 and 2 with margins broadly convex; sternum 3 with midapical concavity deep, about one-third as wide as sternum; sternum 4 with margin broadly rounded, posterior median area covered by patch of coarse hairs directed to rear, this patch divided by longitudinal median de- pressed hairless line; sternum 5 with shallow median emargination; sternum 6 with apical margin truncate except for median produced portion which is parallel sided, the apex with lateral angles broadly rounded, middle feebly emarginate; sternum 8 with apical projec- tion broadly subtriangular. Gonocoxite broad at base, with dense patches of inner, outer, and ventral bristles at and just distal to angle; free portion of volsella broadly rounded. Female: Length of body 11 to 12 mm. Coloration: Black with flagellum, tegula, legs, and margins of metasomal segments slightly brownish. Pubescence: Hair of head and thorax white with yellowish to light fuscous hairs intermixed on clypeus, vertex, genal and hypostomal areas; tibiae and tarsi with hairs fuscous, rest of legs with white hairs or with fuscous intermixed on hind femora; first metasomal tergum and sides of second with white hair, sixth tergum with short white or light fuscous hair; rest of metasoma with hair black or dark fuscous, sometimes with white intermixed 286 The University Science Bulletin on terga. Head: Mandible with superior apical angle (tooth 1) rounded acute, tooth 2 nearly as large as adjacent teeth; tooth 3 not much shorter than 4 which is short and triangular; teeth evenly spaced; mandibular carinae parallel, lower carina slightly wider than upper; carina on upper edge nearly straight, very slightly curved down apically near base of tooth 1, mandible with apex only slightly wider than constriction which is scarcely narrower than base; base of mandible abruptly depressed so that there appears to be a rounded transverse ridge across the mandible near the base, just distad of the depressed zone (this much as in pikei, but in that species lower part of ridge is elevated); labial palpus with seg- ment 1 five-sevenths as long as 2; maxillary palpus with segment 3 longer than 2 and slightly longer than 4 and 5 together; face about as long as broad; postocellar line about three-fourths as long as ocellocular, the latter about equal to ocelloccipital line; clypeus convex with punctures over most of surface finer and denser than on anterior half of paraocular area and not much if any coarser than those of rest of face, truncation broadly and strongly elevated, slightly convex so that it is ill defined, anterior portion of thickened region smooth, posterior portion strigose punctate, distance between lateral angle of truncation and lateral angle of clypeus about equal to length of truncation; eyes with inner margins converging below, each three times as long as wide; base of mandible wider than eye; flagellum with segment 1 one and one-third times as long as pedicel; hypostomal carina moderately low, slightly raised a short distance behind angle. Alesosoma: Foreleg with tibial spine about half as long as tliat of middle leg; longest hair on fore basitarsus over half length of basitarsus; strigilis with apex of velum oblique, merging into malus, malar spine one-fourth as long as entire length of malus; inner apical angle of velum broadly rounded, inner basal angle rounded but sharper than inner apical angle; hind tibial spurs long, slender, slightly curved apically. Forewing with length and width as 14.5:5, papillae on apex moderately fine, hair in cells short and dense. Mesoscutum rather dull, punctures nearly confluent an- teriorly and laterally, separate in discal area. Propodeal triangle completely dull and granular, upper margin finely rugulose. Meta- soma: Punctures shallow, largely separated by a puncture width or less, impunctate bands on terga 2 to 5 aciculate, one-third to one- sixth as wide as punctate portions; sixth tergum rather finely and closely punctate in contrast to preceding ones. Type material: The type of frigida from "Hudson's Bay" is pre- Revision of Genus Osmia 287 sumably in the British Museum (Natural History), of hudsonica from "Hudson's Bay Territory" is in the Academy of Natural Sciences of Philadelphia. Distribution: Northern Europe, the Alps. In America, Alaska, Washington and Oregon to Hudson Bay and Ontario. Sandhouse records the species from Alberta, Montana, and Colorado in addition to the areas mentioned. Specimens before us are from Emigrant Gulch, Montana, May 12, 1919 (R. Kellog). Osmia (Centrosmia) hakeri Sandhouse Osmia baked Sandhouse, 1924, Proc. California Acad. Sci., ser. 4, vol. 13, p. 345 (male). Osmia (Nothosmia) bakeri, Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, p. 87. This species difiFers from others of the subgenus Centrosmia in the unmodified midtarsal segments on the fore and middle legs combined with metallic blue coloration. It is close to vandykei in the general pattern of the genitalia, the strigilis, the clypeal mar- gin, and the emargination of the sixth metasomal tergum, but differs from the latter in having unmodified flagellar segments and in lack- ing stiff black hairs on the sixth and seventh metasomal terga. Male: Length of body 7 to 9 mm. Coloration: Dorsum of thorax, face and vertex of head greenish blue, pleura and abdomen dark blue; mandible black; flagellum usually Hghter below than above; legs black except for weakly metallic femora. Pubescence: White, that of fourth and following sterna and lower surfaces of tarsal segments reddish brown. Head: Mandible with a wide, deep, V-shaped notch separating upper tooth from only slightly longer lower tooth; margin above notch oblique, unmodified; superior apical angle acute; mandibular carinae subparallel, converging at base of lower tooth, lower carina twice as wide as upper, space between them narrow; labial palpus with segments 1 and 2 as 5:8; maxillary palpus with lengths of segments 2 to 5 as 3:4:2:1; length and width of face at 6:5; postocellar line nearly equal to or slightly longer than ocellocular line; ocelloccipital line shorter than post- ocellar; clypeus convex, apex truncate, a little swollen, sometimes with shallow emargination and midapical triangular projection, with punctures finer than those of surrounding areas, with apical impunctate band about one-third as wide medially as greatest width of scape; scape as long as antennal segments 2 to 4 combined; pedicel half as long as first flagellar segment; last flagellar segment slightly longer than a typical segment, unmodified; length and width 288 The University Science Bulletin of eye as 5.6:2.7; inner margins of eyes distinctly converging below; width of genal area nearly equal to eye width; hypostomal carina low, slightly raised behind angle. Mesosonia: Foreleg with tibial spine shorter than that of middle leg; strigilis with apex of velum truncate, apex of mains produced straight distad into a short, sharp spine, length of spine about one-eleventh of length of mains, inner apical angle of velum rounded, inner basal angle more broadly so, base of velum more than half as wide as mains; foreleg with seg- ments of mediotarsus not swollen, tarsal segment 2 much longer than wide, third and fourth progressively shorter, basitarsus three times as long as wide; middle leg with coxa unmodified, medio- tarsal segments not swollen, longer than wide, basitarsus five times as long as wide; hind basitarsus gradually narrowing towards base and apex, maximum width in middle, apex truncate; tibial spurs slender, almost straight apically. Forewing three times as long as wide, with papillae on apex fine, thick, with hair in cells short and dense. Propodeal triangle reticulate above, aciculate below. Meta- sorna: Terga 2 to 5 with apical impunctate bands about a fourth as wide as punctate portion; tergum 6 with or without wide, shallow undefined emargination; tergum 7 with produced portion gently emarginate and one-third as wide as apical margin, lateral tooth as wide as or wider than emargination and subtruncate. Sternum 1 with apical margin subtruncate; sternum 2 with slight swelling at midapex, apical margin broadly rounded, a fringe of lateral sub- apical, light yellow hairs; sternum 3 with midapical concavity shal- low, only about one-fifth as wide as sternum, fringed with mod- erately long yellowish hairs directed posteriorly, such hairs con- tinuing anteriorly as a median ventral stripe; sternum 4 with apical margin produced, strongly convex, middle third beset with short, rather sparse hairs, terminating in a fimbria of longer reddish brown hairs; sternum 6 with apical margin weakly convex with narrow, median, apically subtruncate projection beset with irregularly ar- ranged short hairs; sternum 8 subtriangular with slender median apical process. Terminalia similar to those of most other members of subgenus Centrosmia, especially vandtjkei, but slender apical portion of gonocoxite shorter and hairs lateral to apical portion shorter; penis valve more strongly curved than vandykei. Type material: Holotype male and two paratype males collected by C. F. Baker at Claremont, California; two paratype males col- lected by E. P. Van Duzee at Keen Camp, San Jacinto Mountains, Riverside County, California, June 6-12, 1917. The holotype is in Revision of Genus Osmia 289 the collections of U. S. National Museum, Washington, D. C. (U. S. N. M. type no. 28117). Distribution: West coast states from Washington to California. In addition to the type material, specimens from the following locali- ties have been examined by the authors: California: Tetleys Mountain Camp, San Bernardino Moun- tains, May 16 (C. D. Michener); Marsh Creek Springs, Contra Costa County, April 24, 1937 (G. Ferguson); Huntington Lake, Fresno County, July 3, 1917 (I. McCracken). Oregon: Steens Mountains, 7000 feet altitude, July 10, 1927 ( H. A. Scullen ) . Wash- ington: Pullman, May 14, 1904 (A. L. Melander). Osmia (Centrosmia) tanneri Sandhouse Osmia tanneri Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, p. 87. This species resembles pikei in the shape of the sixth metasomal sternum and the genitalia but can be easily distinguished by the compressed and expanded hind basitarsus and the deep U-shaped emargination of the third sternum. This species is known from a single male; hence certain details which were not studied are omitted from the description. The female is unknown. Male: Length of body 11 mm. Coloration: Face and dorsum of thorax dark blue, rest of body blue and greenish blue. Pubes- cence: Hair on face, vertex, hypostomal area, dorsum of thorax, base of posterior surface of forefemur, and middle of first meta- somal tergum, white; on inner surface of tarsi dark brown; remain- der black or mostly so. Head: Mandible with a wide, V-shaped notch separating slightly acute upper tooth from longer and nar- rower lower tooth; margin above notch oblique; mandibular carinae subparallel; postocellar, ocellocular and ocelloccipital lines almost equal to one another; clypeus convex; apical impunctate band nar- row; scape as long as antennal segments 2 to 4 combined; flagellar segments 2 to 11 each nearly twice as long as wide, unmodified; eye more than twice as long as wide, slightly narrower than genal area; inner margins of eyes converging anteriorly; hypostomal carina low and uniform. Mesosonm: Strigilis with apex of velum oblique, apex of mains produced into spine, inner apical angle of velum broadly rounded. Foreleg with segments of mediotarsus not swollen, tarsal segment 2 longer than wide, basitarsus compressed, two and one-half times as long as wide; middle leg with coxa un- modified, mediotarsal segments greatly swollen, tarsal segments 2 10—2656 290 The University Science Bulletin and 3 wider than long; hind basitarsus flattened and expanded, httle over twice as long as wide, widest distal of base, gradually narrowed towards apex which is obliquely subtruncate. Forewing with distal portion densely papillate, hairs on cells short and mod- erately sparse. Propodeal triangle with irregular ridges on upper third, aciculate below. Metasoma: Terga 2-5 with apical impunc- tate bands wide; tergum 6 with very weak median emargination; tergum 7 with deep midapical emargination of produced portion forming lobe on each side. Sternum 1 with apical margin weakly convex; sternum 2 with weak and narrow emargination on middle of moderately convex apical margin; sternum 3 with deep U-shaped median emargination, fringed with fine golden yellow hair; sternum 4 with median third produced, apex emarginate, with patch of posteriorly directed black hairs; sternum 5 with apical margin subtruncate; sternum 6 with narrow, median subtruncate produced portion with patch of hairs; sternum 8 triangular at base, truncate at apex except for slender median process. Gonocoxite lacking hairs on mesal margin, having a tuft of anterolaterally directed hairs on bulging ventrolateral surface of angle; penis valve curved mesally near apex; volsella triangular at free posterior end. Type material: The unique specimen was obtained at Rosevere Creek, Raft River Mountains, Utah, by Vasco M. Tanner (U. S. Na- tional Museum type no. 52875) in June, 1928. Osmia (Centrosmia) thysanisca Michener Osmia thysanisca Michener, 1957, Jour. Kansas Ent. Soc, vol. 30, p. 39. Osmia integrella, Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, pp. 84- 85 (male). Males of this species resemble pikel but are easily distinguished by the small, shallow emargination of the third sternum, the rounded apex of the fourth sternum, the emarginate apex of the process of the sixth sternum, and the relatively broad hind basitarsus. Fe- males of O. thysanisca are unknown, Male: Length of body 9 to 10 mm. Coloration: Blue, face and dorsum of head and thorax often greenish; mandible black, base nonmetallic; flagellum brown; legs dark brown, fore and hind femora weakly metallic. Pubescence: Hair on face, hypostomal area, vertex, dorsum of thorax, and most of first metasomal tergum, white; on genal area, pleuron, second (and often third) metasomal tergum and undersides of fore and middle femora white, often mixed with black; midapical emargination of metasomal sternum 3 with Revision of Genus Osmia 291 pale and short hairs extending forward as band to base of sternum; inner surfaces of all tarsal segments with tawny or reddish hair; hairs elsewhere black. Head: Mandible with wide V-shaped notch separating upper tooth from slightly longer lower tooth; upper tooth with acute angle at apex, margin above notch oblique; mandibular carinae parallel throughout most of their lengths but converging at base of lower tooth; carinae subequal in width, space between them narrow; facial length and width as 6:5; inner margins of eyes distinctly converging anteriorly; labial palpus with lengths of seg- ments 1 and 2 as 7:10; maxillary palpus with lengths of segments 2 to 5 as 8:11:6:4, postooellar and ocellocular lines equal, ocelloc- cipital line about two-thirds as long as postocellar line; clypeus con- vex, apical margin weakly wavy, punctures finer than those on sur- rounding areas, extremely dense, apical impunctate band one-third as wide as greatest width of scape; scape slightly shorter than an- tennal segments 2 to 4 combined; eye with length and width as 5.5:2.5; pedicel half as long as first flagellar segment; flagellar seg- ment 1 almost as long as second; last segment slightly longer than a typical segment, unmodified; hypostomal carina moderately high, uniform; eye about as wide as genal area. Mesosoma: Foreleg with tibial spine slightly longer than that of middle leg; strigilis with apex of malus produced into a short, sharp, slender spine, length of spine about one-seventh the length of whole malus (spine included); velum with inner basal and apical angles smoothly rounded; seg- ments of mediotarsus slightly swollen; tarsal segment 2 a little longer than wide; segments 3 and 4 progressively shorter and nar- rower; basitarsus nearly four times as long as wide; middle leg with coxa unmodified, mediotarsal segments much swollen, tarsal seg- ment 2 slightly longer than wide, 3 wider than long, with anterior margin bulging and rounded, fourth unmodified; hind basitarsus widest slightly basal to middle, gradually narrowed towards ends, apex truncate; hind tibial spurs slender and weakly curved at apices, inner one much longer and stouter than outer. Forewing about three times as long as wide, papillae on apex fine, sparse, cells with hairs short, moderately sparse. Propodeal triangle wrinkled above, aciculate below. Metasoma: Terga 2 to 5 with apical impunctate bands about one-fourth as wide as punctate portions, shining; tergum 6 lacking median emargination, with apical margin not reflexcd, broadly impunctate like preceding terga; tergum 7 with midapical emargination of produced portion moderately deep, form- ing a broad, angular lobe on each side of emargination. Metasomal 292 The University Science Bulletin sternum 1 with apex broadly rounded, without midapical emargina- tion; sternum 2 with apex more broadly rounded, with irregular long hairs along apical margin; sternum 3 with midapical concavity very shallow, about one-sixth as wide as sternum, fringed with fine whitish hairs, margin of sternum to each side of concavity with darker hairs; sternum 4 with large medial patch of black, posteriorly directed hairs, apical margin broadly rounded; sternum 5 with apical margin subtruncate; sternum 6 with median produced por- tion narrow, with cordate patch of bristles, apex of projection arcuate-emarginate; sternum 8 narrowly triangular basally, sub- truncate apically with a long, slender, parallel-sided, median process. Genitalia similar to those of pikei. Type material: Holotype male from Wildhorse Canyon, Steens Mountains, Oregon, 4270-6000 feet altitude, July 5, 1927 (H. A. Scullen) is in the collection of the United States National Museum. Distribution: Oregon and Wyoming. Specimens from the follow- ing localities ( in addition to the type locality ) have been examined : Oregon: Fish Lake, Steens Mountains, 7000 feet altitude, July 11, 1927; White Branch Meadow, Three Sisters, Frog Camp, 5500 feet altitude. Wyoming: Jackson Hole, June 20, 1936. Osmia (Centrosmia) raritatis Michener O^mia raritatis Michener, 1957, Jour. Kansas Ent. Soc, vol. 30, p. 40. Osmia (Nothosmia) universitatis, Sandhouse, 1939, Mem. Ent. Soc. Washing- ton, 1:84 (part). This species is intermediate in its main characters between pikei and thysanisca. Thus the apex of the fourth sternum is neither conspicuously rounded as in thysanisca nor truncate as in pikei, but subtruncate; the emargination of the third sternum is much larger than in thysanisca but smaller than in pikei. The process of the sixth sternum is narrow and parallel sided, with the apex trun- cate, more as in tanned than in any other species. The hind basitarsi are slightly broadened but not so much so as in tanneri. Females of O. raritatis are unknown. Male: Body length 8 mm. Coloration- Blue, face and dorsum of head and thorax greenish; mandible black, base nonmetallic; flagellum dark brown; legs dark brown, femora with weakly me- tallic areas. Pubescence: Hair on face, hypostomal area, vertex, dorsum of thorax, mesepisterna, and first metasomal tergum, coxae, trochanters, tibiae, and outer sides of tarsi, white; on genal areas, and femora white with intermixed fuscous hairs; on second and Revision of Genus Osmia 293 following metasomal terga mixed black and white, black predom- inating posteriorly; on metasomal venter fuscous except for yellow- ish white fringe of third sternum and pale short hairs anterior to this fringe. Head: Mandible with wide V-shaped notch separating upper tooth from shghtly longer lower tooth; upper tooth with acute angle at apex, margin above notch oblique; mandibular carinae subequal in width, space between them moderate, carinae converging toward base of lower tooth. Facial length and width as 6:5; inner margins of eyes distinctly converging anteriorly; posto- cellar and ocellocular lines equal, ocelloccipital line about two- thirds as long as postocellar line; clypeus convex, apical margin wavy, punctures finer than those on surrounding areas, extremely dense, apical impunctate band one-half as wide as greatest width of scape; scape slightly shorter than antennal segments 2 to 4 com- bined; eye with length and width as 5.5:2.4; pedicel half as long as first flagellar segment; flagellar segment 1 almost as long as 2; last segment slightly longer than a typical segment; unmodified; hy- postomal carinae moderately high, uniform; eye about as wide as genal area. Mesosoma: Foreleg with tibial spine longer than that of middle leg; strigihs with apex of mains produced into robust sharp spine, length of spine about one-seventh that of whole mains (spine included); velum with inner basal and apical angles rounded; segments of mediotarsus slightly swollen, tarsal segment 2 longer than wide, 2 and 4 progressively shorter and narrower; basitarsus nearly four times as long as wide; middle leg with coxa unmodified, mediotarsal segments much swollen, tarsal segment 2 slightly longer than wide, 3 wider than long with anterior margin bulging and rounded, 4 unmodified; hind basitarsus widest at basal third, gradu- ally narrowed toward apex (sometimes nearly parallel sided), apex truncate; hind tibial spurs slender, weakly curved at apices, inner one longer and stouter than outer. Forewing about three times as long as wide, papillae on apex fine, sparse, cells with hairs short, moderately sparse. Propodeal triangle weakly wrinkled above, aciculate below. Terga 2 to 5 with apical impunctate bands about one-fourth as wide as punctate portions, shining; tergum 6 lacking median emargination, with apical margin not reflexed, nearly as broadly impunctate as preceding terga; tergum 7 with midapical emargination of produced portion moderately deep, forming a broad, angular lobe on each side of emargination. Metasomal sternum 1 with apex broadly rounded, without midapical emargina- tion; sternum 2 with apex more broadly rounded with irregular 294 The University Science Bulletin long hairs along apical margin; sternum 3 with midapical rather shallow concavity less than one-fourth as wide as sternum, fringed with yellowish hairs, margin of sternum on each side of concavity with long dark hairs; sternum 4 with large medial area of posteriorly directed hairs, apical margin subtruncate; sternum 5 with apical margin subtruncate; sternum 6 with median process narrow, parallel- sided, apex truncate; sternum 8 narrowly triangular basally, sub- truncate apically except for long, slender, median process. Geni- talia similar to those of pikei. Type material: Holotype male, from Camp Creek Ranger Station, Colorado, 8700 feet altitude, 41° 0' N, 106° 12' W, June 19, 1920, is in the U. S. National Museum. Distribution: Colorado, Washington, and California. Specimens from the following localities (in addition to the type locality) have been studied: Washington: Selah, May 4, 1919 (H. A. Scullen). California: Myers, 6300 feet altitude, June 9, 1930. Osmia (Centrosmia) pikei Cockerell Osmia pikei Cockerell, 1907, Ann. Mag. Nat. Hist., ser. 7, vol. 19, p. 367; Cockerell, 1907, Univ. Colorado Studies vol. 4, p. 252; Cockerell, 1928, Univ. Colorado Studies, vol. 16, p. 124; Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, p. 94; Bohart, Knowlton, and Bailey, 1950, Utah State Agr. College, Mimeographed series, no. 371 (female). Osmia universitatis Cockerell, 1907, Ann. Mag. Nat. Hist., ser. 7, vol. 19, p. 538, vol. 20, p. 125; Cockerell, 1909, Canadian Ent., vol. 41, p. 131; Sandhouse, 1924, Proc. California Acad. Sci., ser. 4, vol. 13, pp. 358, 369; Sandhouse, 1925, Canadian Ent., vol. 57, pp. 36, 63; Cockerell, 1928, Univ. Colorado Studies, vol. 16, p. 119; Sandhouse, 1939, Mem. Ent. Soc. Washington, no. 1, p. 84; Michener, 1957, Jour. Kansas Ent. Soc, (in press) (male) (new synonymy ) . Osmia integrella Cockerell, 1907, Ann. Mag. Nat. Hist., ser. 7, vol. 20, p. 124; Cockerell, 1909, Canadian Ent., vol. 41, p. 131; PSandhouse, 1925, Canadian Ent., vol. 57, pp. 36, 63; Cockerell, 1928, Univ. Colorado Studies, vol. 16, p. 119 (male) (new synonymy). Osmia amala Cockerell, 1907, Ann. Mag. Nat. Hist., ser. 7, vol. 20, p. 447; Cockerell, 1909, Canadian Ent., vol. 41, p. 131; Cockerell, 1910, Canadian Ent., vol. 42, p. 312; Cockerell, 1928, Univ. Colorado Stud., vol. 16, p. 119 (male) (new synonymy). Osmia metitia Cockerell, 1909, Canadian Ent., vol. 41, p. 130; Cockerell, 1928, Univ. Colorado Stud., vol. 16, p. 119 (male) (new synonymy). Males resemble thysanisca and raritatis Cockerell but differ from both in the shallowly arcuate median emargination in the seventh metasomal tergum, a deeper emargination of the third sternum, a medially truncate and fimbriate apical margin of the fourth sternum, an apically convex median process of the sixth sternum and the shorter first flagellar segment in comparison to tlie second. The Revision of Genus Osmia 295 female is easily distinguished from biwephala Cresson by smaller size, relatively smaller head, the nature of the tumescenses of the apical margin of the clypeus and the bases of the mandibles, the deep and confluent punctures on the vertex, the longer and sinuate upper mandibular carina, the slender segments of the middle tarsus, and the more uniformly metallic coloration. Male: Length of body, 8 to 10 mm. Coloration: Dark blue, face, vertex and dorsum of thorax often greenish; mandible black, base nonmetallic, flagellum dark brown; legs dark brown or black. Pu- bescence: Hairs on face, hypostomal area medially, dorsum of thorax, most of pleuron, propodeum ( except lateral black patches ) , and metasomal tergum 1 white; on metapleuron, posterior femur and metasomal terga 4 to 7 and second and following sterna, black, on lower surfaces of middle and hind tarsi dark brown to black towards proximal ends; elsewhere black with few white mixed, or hairs of legs and genal areas all black (one specimen has hairs of legs nearly all pale ) . Head: Mandible with a wide, deep, V-shaped notch separating upper tooth from long, narrow lower tooth; apical edge above notch oblique, unmodified; superior apical angle acute; mandibular carinae parallel, lower carina slightly wider than upper, space between them narrow; labial palpus with segment 1 equal to segment 2; maxillary palpus with segment 2 subequal to length of segments 4 and 5 combined and shorter than segment 3, segment 4 much longer than segment 5. Facial length and width as 7.3:6.3; postocellar line subequal to ocellocular line; ocelloccipital line two- thirds as long as postocellar; clypeus moderately convex, apex slightly wavy; punctures finer than those on rest of face, with apical impunctate band narrow, being one-fourth as wide medially as greatest width of scape; eye with relative length and width as 6.5:2.8, as wide as genal area; inner margins of eyes converging below; hypostomal carina rather low, decreasing gradually from angle toward rear; scape subequal to combined lengths of antennal segments 2 to 4; pedicel half as long as segment 2; flagellum with segment 1 more than half as long as segment 2, last segment slightly longer than a typical segment, unmodified. Middle leg with tibial spine almost as long as to little longer than that of foreleg; strigilis with apex of velum truncate, inner apical angle almost semicircular; malus with apex produced into a short, slender, slightly downward and outward projecting spine; velum wider at apex than at base. Mesosoma: Foreleg with segments of mediotarsus slightly swollen, tarsal segment 2 slightly longer than wide at apical end, segments 296 The University Science Bulletin 3 and 4 progressively shorter and narrower (anterior lobes more produced than posterior lobes ) ; basitarsus gradually thickened from base to apex where it is about one-fourth as broad as long. Middle leg with coxa unmodified, tarsal segments 2 and 3 swollen, with anteroposterior width more than width of basitarsus; hind basitarsus almost uniformly broad from base to apex, fully five times as long as broad; hind tibial spurs long, slender, inner one longer and less curved apically than outer. Forewing with maximum length and width as 7:2.3, papillae on apex fine and dense, hairs on cells long and sparse. Propodeal triangle shining and wrinkled above and aciculate below. Metasoma: Terga 2 to 6 with apical impunctate bands fully one-seventh as wide as punctate portions and shining; sixth tergum slightly reflexed laterally and apically, midapex with a shallow emargination ( absent in some ) ; produced portion of ter- gum 7 with wide and shallow midapical emargination, on each side of which the projection is short and bluntly right angular. Meta- somal sternum 1 with very shallow midapical emargination; sternum 2 broadly rounded, often straight or feebly emarginate medially, with apical fringe of long black hairs; sternum 3 with midapical emargination deep, nearly one-third as wide as sternum, fringed with long light brown hairs; sternum 4 with median third produced, apex broadly truncate and covered with coarse black hairs; sternum 6 with broad midapical projection, covered with short curved bristles, apex convexly arcuate and wider than base; sternum 8 truncate apically except for long, slender median process which is uniformly wide throughout length, and a little less than one-third total length of sternum. Gonocoxite with a few long hairs mesally, basal to angle, and with tuft of anterolaterally directed medium- sized hairs on ventrolateral surface of angle, apical process slightly pointed, directed mesad; penis-valves slightly arcuate; volsella sub- triangular at free posterior end. Female: Length of body 9 mm. Coloration: Dark blue, dorsum of head and thorax greenish; mandibles, antenna, legs, and center of abdomen black. Pubescence: Hair on upper half of face, vertex, dorsum of thorax, propodeum, and metasomal tergum 2 black and white mixed (white predominant on supra-antennal area and on thorax); metasomal tergum 1, except a few lateral black hairs, white; pubescence otherwise black; anterior margin of foretarsus with black hooked hairs, posterior surface with erect, longer, curved hairs which are increasingly L-shaped towards the apex; hairs on hypostomal area curved at apices, directed anteromesally. Head: Revision of Genus Osmia 297 Mandible with superior apical angle (tooth 1) slightly acute, tooth 2 present, rather small, midway between 1 and 3; teeth 1, 3, and 4 almost equidistant; teeth 3 and 4 triangular, 4 twice as long as 3, slender, separated from tooth 3 by wide notch; mandibular carinae strongly raised, narrow, lower two equal and parallel, upper sinuous, curving down apically and reaching base of tooth 3; mandible with apex slightly less than 1.5 times as wide as constriction, which is scarcely narrower than base; labial palpus with segment 1 almost equal to segment 2. Facial length and width as 7.2:5.8; postocellar line slightly less than ocellocular line; ocelloccipital line nearly 1.5 times as long as postocellar line; clypeus convex, punctures sub- apically equal to those on rest of face, coarser centrally and basally, apical impunctate band almost as wide laterally as greatest width of scape, apical margin thickened and swollen laterally, showing a median depression when viewed from above; eye about three times as long as wide, base of mandible about as wide as eye; inner margins of eyes slightly converging below; flagellar segment 1 twice as long as pedicel; hypostomal carina moderately high, high- est well behind angle, gradually reduced from point of greatest height; hypostomal area convex, punctures coarse. Mesosoma: Foreleg with tibial spine slightly shorter than that on middle leg; forebasitarsus with longest hair little more than half as long as basitarsus; strigilis with apex obliquely truncate, spine short, di- rected straight forward, inner apical angle of velum sharp, inner basal angle rounded; hind tibial spurs long, slender, curved only at tips. Forewing with maximum length and breadth as 7.2:2.5; papillae on apex fine, dense, hairs on cells short and dense. Meta- soma: Terga 1 to 5 with apical impunctate bands broad, finely lined and shining, band on tergum 2 one-third as wide as exposed part. Type material: Holotypes of pikei, from Halfway House, Pike's Peak, Colorado, of tmiversitatis from Boulder, Colorado, of amala from Florissant, Colorado, and of metitia from Boulder, Colorado, are in the United States National Museum. The type of integ,rella from Boulder, Colorado, has not been located. Sandhouse (1939) states that it is in "Cockerell's collection" but Dr. Hugo G. Rodeck of the University of Colorado Museum has informed me that it is not in the collection of that institution. Distribution: British Columbia to California, east to Wyoming and Colorado. Besides the type material, specimens have been studied by us from the following localities: 298 The University Science Bulletin British Columbia: Vernon, August 5, 1907. Washington: Olympia; Seattle. Oregon: high Cascade Mountains, Linn County, July 20, 1909 (J. C. Bridwell); Corvallis, April 26, 1926 (H. A. Scullen), April 17, 1947 (L. E. Wallace). California: Los An- geles (W. H. Ashmead collection; this record is presumably an error but may indicate the existence of the species in the mountains of Los Angeles County). Colorado: Boulder, May 22, 1913; spring, 1931 (P. Lundy); April 15, 1933 (H. W. Campbell); Geneva Park, May 14, 1929 (H. Rodeck); Jefferson, June 27, 1913 (A. K. Fisher ) . Osinia (Centrosmia) austromaritima Michener Osmia austromaritima Michener, 1936, Canadian Ent., vol. 68, p. 43; Bohart, Knowlton, and Bailey, 1950, Utah State Agr. College, Mimeographed ser. no. 371 (male). Osmia (Nothosmia) austromaritima, Sandhouse, 1939, Mem. Ent. Soc. Wash- ington, no. 1, p. 85. This species closely resembles pikei, thysanisca, raritatis, etc. but differs in having more black hairs on the genal and pleural areas, deeper emarginations of the seventh metasomal tergum and the third metasomal sternum, in lacking a median patch of coarse black hairs on the fourth metasomal sternum, in the equal width of the genal area and eye, in the greater length of the tibia in propor- tion to the width, and in the much longer and coarser hairs at the apical angle of the gonocoxite. Male: Length of body 8 to 10 mm. Coloration: Mandible dark brown to black; flagellum dark above, light below, legs dark brown with weakly metallic femora; face, vertex and dorsum of head dark bluish green; genal area, side of thorax, propodeum, metasomal terga deep blue. Pubescence: Hairs on face, vertex, dorsum of thorax, and metasomal tergum 1 (except on side) white, on second tergum, thoracic sternum, and genal area white mixed with black; on hypostomal area dusky; on metasomal sterna (except golden, midapical fringe on third) fuscous; on lower surfaces of tarsal seg- ments reddish brown; on rest of body black. Head: Mandible with a wide, deep, V-shaped notch separating upper tooth from very long, lower tooth; apical edge above notch straight or slightly rounded, superior apical angle slightly obtuse; apical edge above notch at right angles to base of lower tooth; mandibular carinae parallel, lower carina apically at least three times as wide as upper, narrower basally, space between them narrow; labial palpus with relative lengths of segments 1 and 2 as 11:13; maxillary palpus Revision of Genus Osmia 299 with segment 3 longer than segment 2, subequal to combined lengths of segment 4 and 5; segment 5 much shorter and narrower than segment 4; face slightly longer than wide; postocellar line nearly equal to ocellocular line; ocelloccipital line a little more than two-thirds as long as postocellar; clypeus convex, apex with median one-third produced, a shallow median emargination at apical margin of projection, punctures finer than those surround- ing area, apical impunctate band about half as wide medially as greatest width of scape; scape subequal to antennal segments 2 to 4 combined; pedicel half as long as flagellar segment 1; flagellum with last segment slightly longer than any typical segment, un- modified; eye with relative length and width as 5.2:2.7, nearly as wide as genal area; inner margins of eyes converging below; hypo- stomal carina moderately low, highest far behind angle. Mesosoma: Middle leg having tibial spine longer than that of foreleg; strigilis with apex of velum truncate; malus with apex produced into a long sharp spine, which projects slightly downwards, spine one- fifth as long as entire length (including spine); malus with width at base twice the width of velum at base; velum with inner basal angle more narrowly rounded than inner apical angle; foreleg with segments of mediotarsus slightly swollen, segment 2 as wide as long, segments 3 and 4 progressively shorter, narrower, basitarsus over three times as long as wide; middle leg with coxa unmodified, tarsal segments 2 and 3 swollen, scarcely wider than basitarsus, swelling equally produced on each side; tarsal segment 2 slightly longer than wide, segment 3 as long as wide; hind basitarsus widest medially, truncate apically; hind tibial spurs long and slender, apices slightly curved. Forewing with relative maximum length and width as 5.8:2, papillae on apex fine, sparse; hair on cells long and sparse. Propodeal triangle wrinkled above, aciculate below. Metasoma: Terga 2 to 5 with apical impunctate bands about one-third or one-fourth as wide as punctate portions, shining; tergum 6 with equally wide apical impunctate portion slightly turned up, midapex without emargination; tergum 7 with deep, wide emargination at midapex of produced portion, depth nearly equal to width of emargination, teeth on each side slender, pointed. Sternum 1 having shallow, wide midapical emargination, long posteriorly directed black hairs on entire sternum; sternum 2 with broadly rounded apical margin, hairs similar to those of first, re- stricted to subapical lateral areas; sternum 3 with midapical con- cavity deep, wide ( nearly one-third width of sternum ) fringed with 300 The University Science Bulletin long golden hairs, black hairs shorter than those on first and second; sternum 4 with median third produced medially and apically, al- most truncate, median apical patch of golden hairs bent mesad at tips; sternum 5 with apical margin deeply emarginate; basal margin of sternum 6 strongly concave, posteromedian portion with short hamate bristles in a patch which is V-shaped anteriorly, apical margin with middle third rounded and produced, with a V-shaped median emargination; sternum 8 similar to that of thij- sanisca. Gonocoxite broad and of uniform width throughout length, with tufts of long, coarse hairs on inner and outer sides of subapical angulation; penis valve forcepiform; volsella subtriangular; apical process of gonocoxite stout. Type material: Holotype male, obtained by C. D. Michener at San Pedro Hills, Los Angeles County, California, February 22, 1929, is in the Snow Entomological Museum, University of Kansas. Distribution: California to Utah and Colorado. In addition to the type, the following specimens have been studied. California: Viola, May 15, 1909. Utah: Green Canyon, Cache County, April 15, 1943 (G. F. Knowlton, D. R. Maddock, S. L. Wood). The species was recorded by Sandhouse (1939) from Colorado, but she gave no locality data. Therefore, it has not been possible to evaluate or verify the record or get any idea of the part of Colorado in which this species occurs. Osmia (Centrosmio) vandykei Sandhouse Osmia vandykei Sandhouse, 1924, Proc. California Acad. Sci., ser. 4 vol. 13, p. 344; Sandhouse, 1939 Mem. Ent. Soc. Washington, no. 1 p. 86 (male). This species does not closely resemble any other member of the subgenus Centrosmia, although it resembles pikei in the shape of the genitalia, the swollen midtarsal segments of the middle legs, the shallow emargination of the apical margin of the sixth meta- somal tergum, the deep and confluent punctation of the face, vertex and genal area, the elongated malar spine of the strigilis, and the general shape of sixth metasomal sternum. But a large number of characters, including tlie shape of the apex of the clypeus, the flattened expanded apical segment of the flagellum, the concavities on the under side of the swollen midtarsal segments of the middle leg, the long stiff hairs on the expanded sides of the fifth, sixth and seventh metasomal terga, and shape of the posteromedian portions of the sixth and seventh metasomal terga, distinguish this species from all others. Revision of Genus Osmia 301 Male: Length of body, 10 mm. Coloration: Dark blue, some- times greenish; flagelkim dark above, hght brown below; legs dark brown; face, vertex and dorsum of thorax greenish blue. Pubes- cence: Hair on face, vertex, dorsum of thorax, mesepisternum, hy- postomal area, underside of femur and tibia of foreleg, part of lat- eral area of first metasomal tergum, and apical bands of hairs on second and third metasomal terga white; on remaining parts dusky to black, except reddish brown on undersides of fore and middle tarsi. Head: Mandible with narrow, deep, V-shaped notch sepa- rating sharply acute upper tooth from long slender sharp lower tooth, upper apical edge above notch slanting, lower broadly con- cave (widths of upper and lower teeth equal at bases); mandibular carinae convergent towards base of lower tooth; upper carina about one-third as wide as lower carina on apical portion, equal near base, space between them narrow; facial length and width as 6.6:6; post- ocellar line subequal to ocellocular line; ocelloccipital line nearly two-thirds as long as postocellar; clypeus convex, midapex produced into a very small rounded projection anteriorly, punctures finer than those on rest of face, apical impunctate band about three-fourths as wide medially as greatest width of scape; scape nearly as long as combined length of antennal segments 2 to 4; pedicel half as long as flagellar segment 1; flagellum with segments 9 to 10 progessively wider, segment 11 most strongly compressed and widest, outer edge convex, apex rounded; eye more than twice as long as wide; relative widths of eye and genal area as 3:2.2; inner margins of eyes weakly converging below; hypostomal area of uniform height throughout; labial palpus with relative lengths of segments 1 and 2 as 8:10; maxillary palpus with relative lengths of segments 2 to 5 as 5:7:5:3. Mesosoma: Middle leg with tibial spine slightly longer than that on foreleg; strigilis with apex of velum truncate, inner apical angle more broadly rounded than inner basal angle; malus with apex abruptly produced into a thin, sharp spine projecting sharply down- ward; malus about twelve times as long as spine; velum at base one- third as wide as malus at base; foreleg with segments of mediotarsus slightly swollen, subequal in width, longer than wide, progressively shorter from tarsal segment 2 to 4; basitarsus uniformly widened apically, two and one-half times longer than wide; middle leg with coxa unmodified, tarsal segments 2 and 3 greatly swollen posteriorly (inner surfaces deeply concave), as wide as long; tarsal segment 4 not swollen; hind basitarsus almost parallel sided, truncate at apex; hind tibial spurs slender, distinctly curved apically. Forewing three 302 The University Science Bulletin times as long as wide, papillae on apex fine and sparse; hair on apical cells short and moderately dense, on basal cells longer and more sparse. Metasoma: Terga 1 to 5 with apical impunctate bands, each band nearly one-fifth as wide as punctate portion and shining; tergum 6 with median portion and 7 with median-basal and lateral portions covered by very long and erect hairs; tergum 6 with apical margin broadly very shallowly emarginate; tergum 7 with apex greatly thickened laterally, weakly emarginate medially, resultant lobes very obtuse. Sternum 1 with a moderately deep, narrow midapical emargination; sternum 2 with hair long on apical por- tion, apical margin convex; sternum 3 with a deep, wide V-shaped emargination, margin of V lined with a row of golden hairs pro- gressively longer towards apex, long posteriorly directed, black hair on each side of emargination; sternum 4 impunctate and with- out hairs except for a broad subtriangular patch of slightly curled blackish bristles which is divided basally by a longitudinal impunc- tate stripe, apical margin with a median, truncate, produced por- tion which has a fimbria of plumose hair; sternum 6 with a deep basal emargination, posteromedian portion with a broadly sub- triangular patch of coarse hamate bristles, apical margin with middle third truncate and strongly produced; sternum 8 subtrian- gular, with a long, slender, apical process which is as long as trian- gular portion. Gonocoxite with well-marked constriction and a single long bristle on dorsal surface nearly opposite apex of volsella, beyond and lateral to apical process a group of stiff hairs which are directed latero-posteriorly, mesal surface with several long hairs basad of process; volsella with apex broadly subtriangular; penis valve curved near apex. Type material: Holotype male and one paratype male, collected by E. C. Van Dyke at Fremont National Forest, Klamath County, Oregon, on June 18, 1922, are in the California Academy of Sciences, San Francisco, California. Distribution: Washington to southern Oregon. The following specimens have been examined by us: Washington: Pullman. Oregon: Wildhorse Canyon, Andrews, 4270 feet altitude, July 5, 1927 (H. A. Scullen). RE\asiON OF Genus Osmia 303 LITERATURE CITED Cresson, E. T. 1887. Catalogue of the Described Hymenoptera of America, North of Mexico, Trans. Amer. Ent. Soc, Suppl. vol., pp. 15.5-3.50. Dall.a. Torre, C. G. de 1896. Catalogus Hymenopterorum, vol. 10, viii -|- 643 pp. Friese, H. 1911. Apidae I. Megachilinae, Das Tierrcich, Lfg. 28, xxvi + 440 pp. MiCHENER, Charles D. 1944. Comparative e-xteroal morphology, phylogeny, and a classification of the bees (Hymenoptera), Bull. Amer. Mus. Nat. Hist., vol. 82, pp. 157-326. 1951. Apoidea, in C. F. W. Muesebeck, Karl V. Krombein, and H. K. Townes, Hymenoptera of America North of Mexico, Synoptic Cata- log, U. S. Dept. Agric, Agric. Monographs, no. 2, pp. 1043-1255. 1957. The identity of Osmia integrella Cockcrell and its relatives, Jour. Kansas Ent. Soc, vol. 30, pp. 38-40. Robertson, Charles. 1903. Synopsis of Megachilidae and Bombidae, Trans. Amer. Ent. Soc, vol. 29, pp. 163-178. S.-VNDHOUSE, Gr.\CE A. 1924. Bees of the genus Osmia in the collections of the California Academy of Sciences, Proc California Acad. Sci., ser. 4, vol. 13, pp. 341-372. 1939. The North American bees of the genus Osmia (Hymenoptera: Apoidea), Mem. Ent. Soc. Washington, no. 1, ii -j- 167 pp. SiNHA, RaNENDRA N. 1958. A subgeneric revision of the bees of the genus Osmia in America (Hymenoptera: Megachilidae), Univ. Kansas Sci. Bull., vol. 39, pp. 211-261. White, Jimmy R. 1952. A revision of the genus Osmia, subgenus Acanthosmioides (Hy- menoptera: Megachilidae), Univ. Kansas Sci. Bull., vol. 35, pp. 219-307. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 8 The Comparative Morphology, Phylogeny and Higher Clas.sification of the Butterflies (Lepidoptera: Papilionoidea ) ^ BY Paul R. Ehblich TABLE OF CONTENTS PAGE Introduction 307 Systematic Principles 307 Material and Methods 309 Non-papilionoids Studied 310 Papilionidae Studied 311 Pieridae Studied 312 Nymphalidae Studied 312 Libytheidae Studied 314 Lycaenidae Studied 314 Acknowledgments 315 Comparative Morphology 316 Head 316 Cervix 319 Thora.x 319 Prothorax 319 Mesothorax 321 Metathorax 323 Legs 324 Wings .^ 325 Abdomen 325 Pregenital Segments 325 Male Genital Segments 326 Female Genital Segments 326 Phylogeny 326 Primitive and Specialized Characters 326 The Papilionoidea as a Taxon 330 Relationships of the Papibonoidea 331 Interrelationships of the Famihes of the Papilionoidea 333 1. Contribution number 972 from Department of Entomology, University of Kansa.s, Lawrence. (305) 306 The University Science Bulletin PAGE Classification 335 Key to the Families of the Papilionoidea 335 Family Papilionidae 336 Key to the Subfamilies and Tribes of the Papilionidae 338 Papilioninae ?38 Parnassiinae 340 Baroniinae 341 Family Pieridae 342 Key to the Subfamilies of the Pieridae 343 Pseudopontiinae 343 Dismorphiinae 343 Pierinae 344 Cohadinae 344 Family Nymphalidae 345 Key to the Subfamihes of the Nymphahdae 347 Danainae 348 Ithomiinae 349 Satyrinae 350 Morphinae 351 Calinaginae 352 Charaxinae 352 Nymphalinae 353 Acraeinae 354 Family Libytheidae 355 Family Lycaenidae 356 Key to the Subfamilies of the Lycaenidae 357 Styginae 357 Lycaeninae 358 Riodininae 358 Literature Cited 360 Figures 362 Abstract: This paper is the second and final section of a work on the in- tegumental morphology, phylogeny and classification of the butterflies. It is based primarily on a detailed comparative morphological study of some 250 genera of butterflies and a less comprehensive study of representatives of 24 families of moths. The relationship of the butterflies to the remainder of the Lepidoptera is discussed as are the interrelationships of the various families and subfamilies of the butterflies themselves. A relatively conservative higher classification of the butterflies is proposed in which they are divided into the following famihes and subfamilies: Papilionidae ( Papifioninae, Parnassiinae, Baroniinae); Pieridae (Pierinae, Coliadinae, Dismorphiinae, Pseudopontiinae); Nymphalidae (Ithomiinae, Danainae, Satyrinae, Morphinae, Charaxinae, Calinaginae, Nymphalinae, Acraeinae); Libytheidae; Lycaenidae (Lycaeninae, Styginae, Riodininae). Phylogeny and Classification of Butterflies 307 INTRODUCTION This is the second section of a work on the comparative mor- phology, phylogeny and higher classification of the butterflies. The first section (Ehrlich, 1958) dealing with the detailed integumental anatomy of the adult monarch butterfly {Danaus plexipptis L. ) laid the necessary foundation for the comparative morphological studies which are the central theme of the present work. An attempt has been made to integrate the new morphological data resulting from this study with pre-existing data (principally morphological) in order to provide the broadest possible base for the conclusions drawn.* Systematic Principles Complete objectivity in arriving at classifications and phyletic relationships is at present a Utopian concept, although advances are being made in this direction (see Michener and Sokal, 1957). Some major sources of subjective error in taxonomic work are: 1) preconception (to some degree unavoidable when a worker is dealing with a group with which he has long been famihar); 2) unjustified character weighting (especially a tendency to give more weight to characters studied personally); 3) group favoritism (the tendency to consider one's favorite taxonomic group as higher in the hierarchy of classification than equivalent groups); and 4) frankly subjective decisions ("I feel that the Xidae are worthy of family rank" or "Yus is obviously more closely related to Ztis than Xus"). Every attempt has been made to avoid these errors in the present work, but doubtless numbers 1 and 2 have not been com- pletely eliminated. It is hoped that 3 and 4, abundantly repre- sented in the literature, have been entirely excluded. There are those ( e. g., Warren, 1947 ) that claim that higher cate- gories should be based on the distribution of one or two diagnostic characters. This is an unfortunate concept which may easily lead to polyphyletic taxa and erroneous ideas of relationship. It should be pointed out that a character state found in all known members * Characters of the larvae have been considered only rarely in this work. Unfortunately, there is a general lack of systematic information on these forms, especially of the homologies of their structures ( e. g., tubercles on one larva are not necessarily homologous to tubercles on another ) It should be noted that characters of the larvae are neither more nor less significant than those of the adult, and that properly conducted invcstigat.ons of large num- bers of larval characters would provide an interesting independent check of the conclusions arrived at here primarily on the characters of the imagines. 308 The University Science Bulletin of a group (a diagnostic character) is probably of no greater significance from the viewpoint of phylogeny than one found in, say, 97 percent of the known members of a group. Whether the absence of a well developed third vannal vein ( 3 V ) in the hind- wings was a diagnostic character of the Papilionidae hinged on the discovery of Baronia brevicornis, a rare papilionid unique in its family in possessing the vein. Many similar cases could be cited. The repeated failure of systems based on too few characters to stand the test of time is a matter of record and will not be dis- cussed further here. In the present work the number of characters considered has been limited by time and practicality, but it is hoped that the sample has been sufficiently large to avoid major errors. The question of the nomenclatorial status of the various taxa segregated has received considerable attention. Some previous classifications of butterflies, as exemplified by Clark ( 1948 ) have presented entomologists witli a mass of largely undefined families, subfamilies, tribes and subtribes within the superfamily Papilio- noidea. Fortunately, this extreme splitting has been largely ignored. In the present work it has been found that the Papihonoidea divide primarily into five groups. In an attempt to align the classi- fication of butterflies with that of other superfamilies of insects these five groups have been called families (Papilionidae, Pieridae, Nymphalidae, Libytheidae, Lycaenidae). The Apoidea (6 families) show much greater morphological and behavioral diversity than the butterflies. The Sphecoidea (3 to 18 families) show a variety of form and habits not even faintly approached in the Papihonoidea, as do the Fulgoroidea (which some authorities consider to repre- sent a single family, the Fulgoridae ) . The same is true of the Chal- cidoidea, Scaraboidea, Tipuloidea and others. Within these groups the major divisions are considered to be families. While it is difficvilt to compare and equate differences within major groups, it seems evident that tlie recognition of the primary divisions of the Papihonoidea with superfamilial designations such as "family group" would not be in keeping with accepted entomological practice. As far as possible the morphological distinctness of the various taxa has been kept uniform within the next highest taxon. Thus, in order to have all the families in the superfamily more or less equivalent, the long standing "families" into which the nymphalids have been spHt previously must be considered to be subfamilies, since their elevation to family rank would necessitate the raising Phylogeny and Classification of Butterflies 309 to family rank of all the tribes of the Papilionidae, a move which has not been advocated even by the most extreme "splitters." The degree of morphological distinctness of subgroups may vary greatly among higher taxa. By "morphological distinctness" is meant both the actual degree of morphological difference ( e. g,, presence or absence of a structure is ordinarily considered a greater difference than change of size or shape in a structure ) and also the size of the gap between character states ( /. e., whether the variation is essen- tially continuous or, if not, the degree of discontinuity). Thus the subfamilies of the Nymphalidae have smaller gaps separating them and show less morphological diversity than those of the Papilionidae. This difficulty is inherent in the nomenclatorial system as usually interpreted. The only alternative that might help to give sub- families in these two families equivalent rank would be to place all nymphalids in a single subfamily, more or less comparable to the papihonid subfamilies. This subfamily could then be divided into tribes. In view of the size of the Nymphalidae, this heterodox system seems undesirable. Much of the difficulty in butterfly (and other) nomenclature can be traced to workers who, starting at the specific level, have deemed it necessary to recognize nearly every branching of the phylogenetic tree with a supraspecific taxon. Using this system one is not far down the tree when the family level is reached, and the result is a large mass of family names with meaning only to specialists. For the benefit of other biologists it is suggested that conservatism (f, e., "lumping") be the rule at the ordinal, familial and generic levels. The complexities of phylogeny can be shown equally well with the aid of less important categories such as sub- and superfamily tribe, subgenus, species group, etc. The concept of the family Nymphalidae is a meaningful one to almost every entomologist and to many other biologists. It can be used without misgivings in ecological or experimental work, and is easily explained to beginning students. It is doubtful if the family Apaturidae is a meaningful entity to one in a thousand entomolo- gists. Since convenience is the only excuse for nomenclature the conclusions seem obvious. Material and Methods Some 240 genera and 300 species of butterflies were dissected in the course of the present work as well as 41 representatives of 24 families of moths and skippers. Dried specimens were used ex- 310 The University Science Bulletin clusively. The wings were removed and preserved, along witli locality labels (if any), in glassine envelopes. The body was then wetted with 80 percent alcohol and heated in 10 percent KOH until tlie viscera were soft ( semiliquid ) . The specimens were then dissected under water, the scales being removed with brushes and the viscera with watchmakers forceps and pipettes. The dissected specimens were preserved in 80 percent alcohol. Only determined specimens were used. It did not prove prac- tical to verify all specific identifications, but all doubtful generic determinations were checked. The generic nomenclature employed has been made as up to date as possible; however, wherever a group has been commonly considered either a subgenus or a genus ( e. g., Graphium, Zerene ) it has been retained as a genus in order to accent the variety of the sampling. Since this paper is not pri- marily nomenclatorial, in no case does the use of such a name indi- cate endorsement of the usage employed. The genera and species examined are listed below by families, and placed alphabetically within the families. Except for the en- tities marked with an asterisk all were dissected as described above. Those marked with the asterisk were examined for superficial char- acters. Every character has not been examined in every species, since ordinarily only one sex of each species was dissected, and since sometimes parts were missing or damaged. It is hoped that this deficiency has been compensated at least in part by the large number of genera and species examined. Non-papilionoids Studied Hepialidae: Hepialus hiimuli Linnaeus; Megalopygidae: Mega- lopijge opercularis Smith; Zygaenidae: Zygaena minos Fuessly; Eucleidae: Sibine stimulea Clemens; Tineidae: Tinea pellionelki Linnaeus; Gelechiidae: Gelechia scrotinella Busck; Yponomeu- TiDAE: Atteva aiirea Fitch: Aegeriidae: Melittia satyriniformis Hiibner; Olethreutidae: Carpocapsa ponionella Linnaeus, Exar- tenia fasciatonum Clemens; CossmAE: Prionoxystm robiniae Peck; Castniidae: Castnia atymnius Dalman, Castnia licus Fabricius; Thyrididae: Tliyris lugubris Boisduval; Pyralidae: Crambus vul- givagellus Clemens, Desmia funcralis Hiibner, Galleria mcVonella Linnaeus, Loxostcge similalis Guenee, Nymphula icciusalis Walker; Pterophoridae: Oidaematophorus monodactylus Linnaeus; Urani- iDAE: Chrysiridia leilus Linnaeus, Chrysiridia madagascariensis Lesson; Saturniidae: Actios htna Linnaeus, Eacles imperialis Phylogeny and Classification of Butterflies 311 Drury; Lasiocampidae: Tolijpe distincta French; Geometridae: Haematopis grataria Fabricius; Sphingidae: Celerio lineata Fabri- cius; Notodontidae: Datana ministra Drury; Phalaenidae: { = Noctuidae): Agrotis ypsilon Rottemburg, Catocala cerogama Guenee, Peridroma margaritosa Hiibner, Folia adjuncta Boisduval; Amatidae: Ctenticha sp.; Arctiidae: Haploe sp., Utethesia hella Linnaeus; Hesperiidae: Adopaea lineola Ochsenheimer, Calpodes ethlius Cramer, Erynnis juvenalis Fabricius, Megathymus neumoe- geni Edwards, Poanes zabtdon Boisduval and Leconte, Proteides clarus Cramer. Papilionidae Studied Archon apoUimis Herbst, Baronia brevicornis Salvin, Battus poly- damus Linnaeus, Battus philenor Linnaeus, Battus devilliers Go- dart*, Bhutanitis lidderdalei Atkinson, Cressida cressida Fabricius, Euryades duponcheli Lucas, Graphiwn agamemnon Linnaeus*, Graphium agesilaus Guerin and Percheron*, Graphium agates West- wood*, Graphiwn antiphates Cramer, Graphiwn hathycles Zinken*, Graphium celadon Lucas*, Graphium cloanthus Westwood, Graph- ium columhus Kollar*, Graphiwn dolicaon Cramer*, Graphium epi- daus Doubleday*, Graphium eurypylus Linnaeus*, Graphium eve- mon Boisduval*, Graphium harmodius Doubleday*, Graphium leo- nidas Fabricius, Graphium macareus Godart, Graphium marcellus Cramer, Graphium nomius Esper*, Graphiwn pausanias Hewitson, Graphium payeni Boisduval*, Graphium philolaus Boisduval* Graphium policenes Cramer*, Graphium protesilaus Linnaeus*, Graphium sarpedon Linnaeus, Graphium thymhraeus Boisduval, Graphium xenocles Cramer*, Hypermnestra helios Nickerel, Lam- proptera (= Leptocircus) curius Fabricius, Lamproptera meges Zinken*, Leuhdorfia puziloi Erschov, Ornithoptera priamus Lin- naeus, Papilio aegius Donovan*, Papilio alexiares Hopffer*, Papilio anchisiades Esper*, Papilio aristius Cramer*, Papilio aristodemus Esper*, Papilio bianor Cramer*, Papilio castor Westwood*, Papilio chaon Westwood*, Papilio cresphontes Cramer, Papilio cynorta Fabricius, Papilio demetrius Cramer, Papilio demolion Cramer*, Papilio eurymcdon Lucas*, Papilio glaucus Linnaeus, Papilio hec- torides Esper, Papilio indra Reakirt*, Papilio machaon Linnaeus, Papilio memnon Linnaeus*, Papilio montrouzieri Boisduval*, Pa- pilio multicatidatus Kirby*, Papilio paeon Boisdnva*, Papilio pala- medes Drury*, Papilio paris Linnaeus*, Papilio pilumnus Boisdu- val*, Papilio polyxenes Fabricius, Papilio proneus Hiibner*, Papilio rutulus Lucas*, Papilio thoas Linnaeus*, Papilio torquatus Cramer*, 312 The University Science Bulletin Papilio troilus Linnaeus, Papilio xuthus Linnaeus*, Parides ( = Atro- phaneura) arcus Cramer*, Parides aristolochiae Fabricius, Parides coon Fabricius*, Parides latreilli Donovan*, Parides montezuma Westwood*, Parides mylotes Bates, Parides perrhehus Boisduval*, Parides philoxeniis Gray, Parides pohjdorus Linnaeus*, Parides poly- zelus Felder, Parides rhodifer Butler*, Parides sesostri^ Cramer, Parides varuna White*, Parnassius apollo Linnaeus*, Parnassiiis clodius Menetrics, Parnassius eversmanni Menetries*, Parnassiiis mnemosijne Linnaeus*, Parnassius sminthetts Doubleday and Hewit- son, Sericintis telamon Donovan, Teinopalpus imperiaUs Hope, Trogonoptera brookiana Wallace*, Troides Helena Linnaeus, Tro- ides rhadamantus Lucas*, Zerynthia hypermnestra Scopoli, Zeryn- thia rtwiina Linnaeus. Pieridae Studied Anteos chlorinde Godart, Anthocharis midea Hiibner, Aporia crataegi Linnaeus, Appias nephele Hewitson, Archonias tereas Hiibner, Belenois mesentina Cramer, Cepora nadina Lucas, Colias philodice Codart, Colotis achine Cramer, Colotis danae Doubleday and Hewitson, Delias eucharis Drury, Dismorphia nemesis La- treille, Dixeia cehron Ward, Euchloe belia Cramer, Etirema nicippe Cramer, Gonepteryx rhainni Linnaeus, Hebomoia glauc'ppe Lin- naeus, Ifaballia demophile Linnaeus, Ixias pyrene Linnaeus, Kri- cogonia lijside Godart, Leptidea sinapis Linnaeus, Leptophobia aripa Boisduval, Leptosia xiphia Fabricius, Leucidea brephos Hiib- ner, Melete isandra Boisduval, Nathalis iole Boisduval, Neophasia menapia Felder, Pereute callinira Staudinger, Perrhybris pyrrha Cramer, Phoebis sennae Linnaeus, Pieris protodice Boisduval and Leconte, Prioneris thestylis Doubleday, Pseudopieris nehemia Bois- duval, Pseiidopontia paradoxa Felder, Zerene eurydice Boisduval, Zegris fattsti Christopher. Nymphalidae Studied Acraea encedon Linnaeus, Acraea esebria Hewitson, Acraea nata- lica Boisduval, Acraea sp., Actinote carycina Jordan, Acfinote neleus Latreille, Actinote ozomene Godart, Ageronia amphinome Lin- naeus, Agraulis vanillae Linnaeus, Amathusia phidippus Johannson, Amauris ochlea Boisduval, Amauris psyttalea Plotz, Anadebis hinia- chala Moore, Anaea andria Scudder, Anaea appias Hiibner, An- tirrhaca miliiades Fabricius, Apatura iris Linnaeus, Aprotopus aedesia Doubleday and Hewitson, Araschnia levana Linnaeus, As- terocampa celtis Boisduval and Leconte, Bia actorion Linnaeus, Phylogeny and Classification of Butterflies 313 Boloria toddi Holland, Brassolis astyra Godart, BrassoUs sophorae Linnaeus, Byblia ihjthea Drury, CaJigo sp., Calinaga buddha Moore, Callerehia annada Moore, Callicore marchalii Guerin, Callitaera aurora Felder, Callithomia hezia Hewitson, Catagramma maimuna Hewitson, Catonephele ntimilia Cramer, Cattina crithea Drury, Ceratinia titutia Hewitson, Ceratinia vallonia Hewitson, Cethosia chrysippe Felder, Charaxes hrutus Cramer, Charaxes psaphon West- wood, Chlosyne janais Drury, Cirrochroa malaya Felder, Clothilda numida Hiibner, Coenony7npha hero Linnaeus, Coenophlehia archi- dona Hewitson, Corades inyo Hewitson, Ctipha grymanthis Drury, Cynthia arsinoe Cramer, Cyrestis nivalis Felder, Danatis cleona Stoll, Danaus plexippus Linnaeus, Dichorragia nesimachus Bois- duval, Didonis hibJis Felder, Dione jiino Cramer, Dircenna klugii Hiibner, Discophora sondaica Boisduval, Doleschallia bisaltide Cramer, Doxocopa cyane Latreille, Doxocopa laurentia Godart, Dryadtda phaetusa Linnaeus, Dryas iulia Hiibner, Dijnastor darius Felder, Ehjmnias hypcrmnestra Linnaeus, Ehjmnias malelas Hewit- son, Elymniopsis hanimakoo Westwood, Enispe eiithymlus Double- day, Epinephile jurfina Linnaeus, Erebia epipsodea Butler, Ergolis oriadne Linnaeus, Eryphanis aesacus Herrich-Schiiffer, Eueides aliphera Godart, Eueides tholes Cramer, Euphoedra eleus Drury, Euphoedra medon Linnaeus, Euphydryas chalcedona Doubleday and Hewitson, Euphydryas phaeton Drury, Euploea core Cramer, Euploea d'.ocletianus Fabricius, Euptoieta claudia Cramer, Eupty- chia hesione Sulzer, Eiithalia garuda Moore, Faunis assamus West- wood, Faunis canens Hiibner, Gynaecia dirce Linnaeus, Gyrocheilus patrobas Hewitson, Haematera thysbe Doubleday and Hewitson, Haetera piera Linnaeus, Hcliconius charithonia Linnaeus, Heliconius chestertoni Hewitson, Heliconius hortense Guerin, Heterochroa bredowii Geyer, Hirsutis neitha HopfFer, Historis odius Fabricius, Hypna clytemnestra Cramer, Hypoleria andromica Hewitson, Hypo- limnus bolina Linnaeus, Hypolimnus dubia Aurivillius, Idea idea Linnaeus, Ideopsis gaura Horsficld, Ithomia cleora Hewitson, Ituna phcnareta Doubleday, Kallima inachus Boisduval, Lethe eurydice Johannson, Lethe kansa Moore, Lethe sura Hewitson, Limenitis bredowii Geyer, Limenitis populi Linnaeus, Lycorea cleobaea Godart, Marpcsia pet reus Bates, MccJumitis doryssus Bates, Megis- tanis baeotus Doubleday and Hewitson, Melanargia galathea Lin- naeus, Melanitis leda Drury, Melinaea paraiya Beakirt, Melitaea artemis Schiffermuller, Melitaea dymas Edwards, Metamorpha steneles Linnaeus, Minois pegala Fabricius, Morpheis ehrenbcrgi 314 The University Science Bulletin Hubner, Morpho achilles Linnaeus, Morpho aratos Fruhstorfer, Morpho hecuha Linnaeus, Morpho lacrtes Druce, Morpho sulkow- skyi Kollar, Napeogenes thira Hewitson, Norope cyllaharus West- wood, Neope goschkevitschii Menetries, Ncptis vikasi Horsfield, Nessaea obrimis Linnaeus, Nymphalis pohjchloros Linnaeus, Oeneis semidea Say, Opoptera stilcius Staudinger, Opsiphancs invirae Hiibner, Panacea prola Doubleday and Hewitson, Pandita sinoria Felder, Pantoporia opalina Kollar, Pararge megera Linnaeus, Par- dopsis ptmctatissima Boisduval, Parthenos gambrisiiis Fabricius, Perisama honplandii Guerin, Philaethria dido Clerck, Phyciodes tharos Drury, Pierella lamia Sulzer, Planema aganice Hewitson, Precis sp., Prepona chromus Guerin, Pronophila thelebe Doubleday and Hewitson, Pseiidergolis toedah Kollar, Pyrrhogyra typhocus Felder, Ragad'.a crisiJda Hewitson, Sais rosacia Cramer, Salainis cytora Doubleday and Hewitson, Satyrtts circe Fabricius, Satyrus semele Linnaeus, Scada theaphia Bates, Smyrna blomficldia Fabri- cius, Speyeria cybele Fabricius, Stibochiona nicea Gray, Sticho- phthalma camadeva Westwood, Taenaris phorcas Westwood, Tay- getis ypthima Hiibner, Tellervo zoiltis Fabricius, Temenis laothoe Cramer, Thyridia conftisa Butler, Tithorea harmonia Cramer, Yoma sabina Cramer. Libytheidae Studied Libyfhea celtis Fuessly, Libythea geoffroy Godart, Libythea laias Trimen, Libythea myrrha Godart, Libytheana bachmanni Kirtland. Lycaenidae Studied Abisara neophron Hewitson, Ambhjpodia micale Blanchard, Ana- tole zygia Hiibner, Ancyluris inca Saunders, Atiteros carusius West- wood, Apodeniia mormo Felder, Apodemia nais Edwards, Atlides halesus Cramer, Baeotis bacaenis Hewitson, Calephelis iris Staudin- ger, Callictita cyora Bethune-Baker, Callophrys rubi Linnaeus, Candalides dimorphiis Rober, Candalides meeki Bethune-Baker, Caria lampeto Godman and Salvin, Ciipido minima Fuessly, Curetis bulls Doubleday and Hewitson, Devdoryx epijarbas Moore, Dodona durga Kollar, Durbania amakosa Trimen, Echenais aristus Stoll, Elaphrotis telephus Cramer, Euselasia aurantiaca Godman and Salvin, Euselasia eulione Hewitson, Euselasia mys Herrich-Schiiffer, Fenesica tarquinius Fabricius, Helicopis cupido Linnaeus, Hemiar- gus hanno Stoll, Hypochrysops rex Boisduval, Hypolycaena philip- pus Felder, lalmenus evagoras Hiibner, Incisalia augustinus West- wood, Lasaia sessilis Schaus, Leucochimona lagora Herrich-Schaffer, Phylogeny and Classification of Butterflies 315 Lycaena heUoides Boisduval, Lycaenopsis pseudargiolus Boisduval and Leconte, Lymnas iarhas Fabricius, Lysandra coridon Poda, Mampava nigronotata Bethune-Baker, Megalopalpus zymia Double- day and Hewitson, Mesosemia telegone Boisduval, Metacharis Indus Fabricius, Mimocraea dohertyi Rothschild, Mitoura grynens Hiibner, Nemeobhis lucina Linnaeus, Niphanda fusca Bremer and Grey, Nymphidwm cachrus Fabricius, Ogyris oroetes Hewitson, Philiris innotatas Miskin, Plebejus icarioides Boisduval, Poretia hewitsoni Moore, Pseuderesia lihertina Hewitson, Rhetvs dyson'.i Saunders, Riodina lysippus Linnaeus, Satyrium fidiginosa Edwards, Siseme alectro Westwood, Siseme aristoteles Latreille, Stalachtis euterpe Linnaeus, Stalachtis phlegia Cramer, Stlhoges sp., Styx infernalis Staudinger, Syrmatia dorilas Cramer, Taraka hamada Druce, Tel- ipna bimacula Plotz, Teriomima hildegarda Kirby, Tharsalia arota Boisduval, Theope mania Godman and Salvin, Thestor halhis Fabri- cius, Thisbe irenea Stoll, Thysonotas danis Cramer, Thysonotis liy- mettts Felder, Zeltus antifatimis Doubleday and Hewitson, Zemeros fegyas Cramer. To facilitate comparisons all characters in the family diagnoses have been given numbers and all characters in the subfamily diag- noses have been given letters. Thus, in all the families character number one is the shape of the eye, and throughout the subfamilies of the Nymphalidae character "a" is the amount of scaling on the antennae. All illustrations for the comparative section of this work have been drawn so that the same structures of different genera are the same size. In many cases figure citations refer to illustrations of genera other than those under discussion which show the charac- teristics alluded to. Citations of figures followed by "-Pt. I" refer to the illustrations of the monarch butterfly in the first section of this work (Ehrlich, 1958). Acknowledgments I wish to express my deep appreciation to the following indi- viduals and institutions who have generously given or loaned speci- mens for dissection: Dr. W. L. Brown, Jr. (Museum of Compara- tive Zoology, Harvard University); Mr. Harry K. Clench (Carnegie Museum); Prof. Wm. T. M. Forbes (Museum of Comparative Zoology); Dr. John G. Franclemont (Cornell University); Mr. Harold J. Grant, Jr. (Academy of Natural Sciences of Philadelphia); Mr. and Mrs. Floyd W, Preston; Mr. James A. G. Rehn (Academy of 316 The University Science Bulletin Natural Sciences of Philadelphia); Mr. Norman D. Riley (British Museum); and Dr. F. H. Rindge (American Museum of Natural History ) . The author is indebted to Dr. K. C. Doering, Dr. C. D. Michener, and various members of the biosystematics group of the Department of Entomology, University of Kansas, and Dr. H. S. Fitch of the De- partment of Zoology, University of Kansas, for reading portions or all of this manuscript and offering many helpful suggestions. He is also indebted to Prof. Wm. T. M. Forbes of the Museum of Com- parative Zoology for carrying on a lengthy and constructive corre- spondence over many aspects of the work. Thanks go also to Mr. Kent H. Wilson of the Department of Entomology, University of Kansas, for putting at the author's dis- posal his extensive notes on the Family Papilionidae. Finally I would like to thank my wife, Anne H. Ehrlich, for aid in the preparation of the illustrations. COMPARATIVE MORPHOLOGY This section gives a brief summary, arranged by structures rather than by systematic categories, of the morphological variation found within the butterflies. It emphasizes variation which was found to have taxonomic significance, and variation which was relatively easy to describe. Since many characters which were later found to be of little systematic significance were recorded for about the first 40 genera dissected, these genera appear as a disproportionately high number of the examples cited. Time has not permitted detailed studies of any organs as have been done by Jordan (1898) on the antennae and Renter (1897) on the labial palpi. A number of major areas (area of head around and above foramen magnum, axillary sclerites, female genitalia) have been largely ignored because of difficulties in dissecting, de- scribing or comparing them. It is hoped that future stiidies will fill in these gaps. The terminology used here is that of the first section of this work (Ehrlich, 1958). Head With the exception of the majority of the Lycaenidae, the eyes of butterflies are entire. In most lycaenids the eyes are emarginate (i. e., notched opposite the bases of the antennae — see fig. 5), and do not extend caudally as far as those in the other families (figs. 6, 7). Many genera of the Lycaenidae and Nymphalidae have the i Phylogeny and Classification of Butterflies 317 eyes hairy to a greater or lesser degree, while they are bare in the remainder of the butterflies. The structure of the antennae of the butterflies has been the sub- ject of an exhaustive study by Jordan ( 1898 ) . Of most systematic interest is the variation in the amount of scaling (ranging from com- pletely unsealed to completely scaled) and the arrangement of the antennal sulci and carinae.* A sulcus is a depressed line on the ven- tral surface of the antenna formed by a groove or series of pits ( one to a segment). The carinae are ridges between and flanking the sulci. No original work on the antennal structure has been done in connection with this paper. The carinae are present only in the Nymphalidae and Libytheidae, in which they are almost universal. The presence of the sulci is variable in the Papilionidae and they are usually absent in the Lycaenidae. The antennae of the Pieridae always have one or three sulci. The distance between the bases of the antennae is very variable, ranging from much less than one half the width of the scape to more than the width of the scape. The area between the antennae may be concave if the antennae are very close together, and there may be a strengthening inflection between the bases of the antennae (the transfrontal suture of DuPorte, 1956). The presence of this in- flection is at least to some degree a function of the interantennal distance, it is almost never present when the antennae are close together. The position of the laterofacial sutures and the concomitant size of the paraocular areas are quite variable. In general the lycaenids have the laterofacial sutures contiguous or nearly contiguous with the eye margins, with the paraocular areas extremely reduced or absent (fig. 5). In the other families the laterofacial sutures may be rather close to the eye margins and largely parallel to them (many genera, Amathusia, Calinaga, Leptidca, Papilio, Teinopalpus, etc., see fig. 1) or they may be sorpe distance from the eye margins and curved inward dorsally and ventrally (various nymphalids in- cluding Danans, fig. 1 — Pt. I). A complicating factor is that in some genera (e.g., Calinaga, fig. 1) the anterior tentorial pits do not lie at the juncture of the laterofacial and clypeolabral sutures as they do in Danatis, but rather at the free ends of the clypeo- labral suture. In these genera the laterofacials pass laterad of the pits. Detailed studies of a great many different Lepidoptera as well as members of related orders will be necessary before we can Tenninology after Clench, 1955. 318 The University Science Bulletin hope to have a reasonable understanding of the structures of the face of tlie Lepidoptera. In many PapiHonidae and Lycaenidae the face is essentially flat (fig. 7) while in the other families it is at least somewhat pro- tuberant. Teinopalpus imperialis (figs. 2, 3) has the entire fronto- clypeal sclerite expanded into a balloonlike structure which ac- counts for about one half the total length of the head. Certain pierid genera (Antlwcharis, Leptidea, Pseudopontia, etc.) also have very protuberant frontoclypeal sclerites. In most non-lycaenid genera the face is at least somewhat indented near the eye margins lateral to the frontoclypeal protuberance (fig. 2 — Pt. I). In most of the lycaenids the anterior tentorial pits are low on the face, usually about one seventh of the total height of the face from the lower margin of the labrum. In the other families the p'its are somewhat higher. The labrum of the butterflies is often difficult to define because of the obscurity of the clypeolabral suture. The labrum is greatly reduced in many genera (especially in the Lycaenidae — fig. 5). The pilifers are well developed in most groups, but are reduced or absent in some, including the Lycaenidae, Baronia, Dismorphia, Lamproptera, Metamorpha, Pseudopontia, and Zerijnthia. The size and shape of the mandibular rudiment was also very variable, but it was difficult to describe the variation because of the indefinite boundaries of the rudiment. There is some variation in the depth of the proboscidial fossa (it is usually deeper in the Lycaenidae and Pieridae than in the other families), but it has not been systematically studied. The maxillary palpi showed surprising development in some genera (including Baronia, Caligo, Calinaga, Metamorpha and Pseudopontia), being distinctly two-segmented in Baronia. The galeae of the maxillae are fringed with papillae at the distal end in Actinote, Anaea, Apodemia, Atlides, Caligo, Calinaga, Heliconius, Libytheana, Lycaena, Lycaenopsis, Metamorpha, Morpho, Oeneis, Telipna and many others. Numerous genera, principally papilionids, pierids and (to a lesser extent) nymphalids do not show this fring- ing or at least have the papillae greatly reduced. A detailed study of these structures would probably reveal taxonomic characters. The labial palpi have been studied in great detail by Renter (1897). The palps vary in length from as long as the thorax (some libytheids) to less than the length of the head (some papilionids and lycaenids). There is much variation in the proportions of the three segments and in the pattern of scaling and setation. Phylogeny and Classification of Butterflies 319 In the Nymphalidae, Libytheidae and Lycaenidae (except Styx) the labial sclerite is completely sclerotized, both in front of and behind the palpal sockets (figs. 3, 4 — Pt. I). However, in most papilionids there is a loss of sclerotization in front of the sockets and in many pierids there is a loss of sclerotization behind the sockets. On the vertex of the head there is a setiferous patch ( presumably sensory) called the chaetosema. The variation in this organ is described and figured by Jordan ( 1923 ) . The anterior tentorial arms show a great deal of diversity. In some genera {Apatura, Danatts, Heliconius, Historis, Ithomia, Leptidea, etc. ) they are relatively straight and simple ( fig. 15 ) , not greatly enlarged or downcurved anteriorly and not bearing crests. At the opposite extreme are genera such as Lamproptera and Graphhim which bear extremely high dorsal crests (figs. 18, 19). In many genera, especially in the Lycaenidae, the arms are bent strongly downward anteriorly (fig. 17) and/or are more than t\vice as thick anteriorly as posteriorly. Cervix In all of the Papilionidae studied (and in none of the other butterflies) the cervical sclerites were found to be joined beneath the neck by a narrow sclerotic band ( fig. 24 ) which may be faint or slightly interrupted in the middle. Variation was observed in the shape of the cervical sclerites and in the position and shape of the cervical organ, but no study of this variation was made. In Lamproptera curius, Papilio machaon, Parnassius sminthetis, Teinopalpus imperialis and other papilionids as well as in Apatura cyane and Historis odia a small ventral sclerite was observed in the cervix close to the head ( fig. 24 ) . Among others it was absent from the following genera: Actinote, Anaea, Apodemia, Atlides, Baronia, Caligo, Dismorphia, Fenesica, Ithomia, Leptidea, Libytheana, Lycaena, Lycaenopsis, Metamorpha, Pieris, and Phoebis. Thorax Prothorax: In all the butterflies except the Pieridae the lateral plates of the pronotum fuse together to form a triangular or Y- shaped structure (fig. 8 — Pt. I) which articulates with the dorsal plate. No such structure is found in the pierids (fig. 28). The dorsal plate itself may be roughly triangular, T-shaped, Y-shaped or sagittate. Sclerotized patagia of varying size are found in all the Nymph- alidae (fig. 37), the Coliadinae of the Pieridae (fig. 25), various 320 The University Science Bulletin groups of the Papilionidae and the Libytheidae (fig. 26). In the hbytheids and certain groups of the Papilionidae (Cressidini in particular) the sclerotized area is very small. All of the lycaenids, most of the pierids and a great many papilionids have the patagia unsclerotized. In some of these groups, particularly in certain lycaenids, the membranous patagia are rather prominent and pro- tuberant, in others they are essentially indistinguishable from the rest of the membrane connecting the lateral plates of the pronotum with the mesothorax. The Charaxinae of the Nymphalidae are the only butterflies in which sclerotized parapatagia have been found ( fig. 37 ) . The presternum is generally present in the Nymphalidae, Libytheidae and Lycaenidae, although it is sometimes indistinctly separated from the ventrally fused propleura. It is absent in the Papilionidae and Pieridae. The profurcal arms are simple (figs. 32, 33) in the Lycaenidae, Nymphalidae and Libytheidae. They usually have a secondary an- terior lamella or prong in the Papilionidae and Pieridae (figs. 28,29,30,31). The intercoxal lamella is quite prominent ( fig. 29 ) in the Pieridae and some Lycaenidae (AtUdes, Euselasia, Lycaena, Lycaenopsis, Megalopolpits, Telipna, etc. ) . It is not prominent but present ( fig. 33) in most of the Nymphalidae, the Libytheidae, and some Ly- caenidae (Apodemio, Fenesica, etc.). In the Papilionidae the inter- coxal lamella has migrated caudally, where it usually forms a prom- inent semicircular lamella almost between the furcal arms (fig. 30). In many of the papilionids (Lamproptera, PapiJio, Troides, etc. but not Parnassius, Teinopalptis, etc.) and virtually all the pierids the discrimen is represented internally by a small anterior spine or lamella (figs. 29, 30). In all the Papilionidae except Baronia the spinasternum is pro- duced laterally at the spina (figs. 34, 35, 36). The process may be long ( laterally visible, fig. 24 ) as in Lamproptera, Papilio, Parnassius and Zerynthia or short (not laterally visible) as in Teinopalptis and Ornithoptera. In most genera the processes are narrow; in Zeryn- thia they are broadened at the ends. In Papilio (at least some species) and Lamproptera there are areas of light sclerotization in the membrane around the processes, especially between the proc- esses and the mesothoracic pre-episternum (fig. 34). Lateral proc- esses of the spinasternum have not been found elsewhere in the butterflies. In the vast majority of the Papilionoidea the spina is essentially Phylogeny and Classification of Butterflies 321 an invaginated sclerotized strip with membranous sides (figs. 8, 18 — Pt. I ) . However, in a number of papilionids ( e.g., Ornithop- tera priamus), the spina is a tubular apodeme (completely sclero- tized ) . In the Nymphalidae, Libytheidae, and Lycaenidae the spina- sternum is generally a narrow strip, invaginated at the spina, and gradually broadening caudally until it joins the thorax at two points with a membraneous triangle between them (fig. 17 — Pt. I). How- ever, in the Pieridae the spinasternum is widened into a small oval or diamond-shaped plate between the furcasternum and spina (fig. 31). This is faintly reminiscent of the lateral expansion of the papilionids. In the Cressidini and Ornithoptera of the Papilionidae the spina- sternum caudal to the spina is usually broad and platelike (figs. 35,36). Mesothorax: The prescutum of the Libytheidae is vertical to the main axis of the body or has its upper end slightly anterior to its lower end, giving the mesonotum a truncated appearance (fig. 26). This effect is also noticeable in a number of genera of lycaenids {Apodemia, EuseJasia, Fenesica, Lycaena, Ltjcaenopsis, Mega- lopalpus, etc.). The remainder of the butterflies have the lower end of the prescutum anterior to the upper end and lack the truncated aspect ( fig. 25 ) . In Baronia the scuto-scutellar suture is obsolescent, and in other genera such as Leptidea, Ltjcaenopsis and Zerynthia it is incom- plete centrally. The shape of the scuto-scutellar suture, especially the depth and angle of the inverted "V" is very variable but did not appear to have useful characters at the higher taxonomic levels. The adnotale is sagittate in the Libytheidae ( fig. 26 ) ; it is variable in shape but not sagittate in the other butterflies. In general the processes of thq second phragma are well de- veloped. However, in many genera of the Lycaenidae (Atlides, Euselasia, Lycaena, Lycaenopsis, Megalopalpus, etc.) they are re- duced to a greater or lesser degree. In the sternopleural region of the butterflies there is a great amount of variation in the sutures and in the anepisternum.* The * The nomenclature of the sternopleural rejrion employed throuRhoiit this work has, as far as possible, been brought into line with the ideas of Matsiida (1956). Since the katapleuro-coxal muscle is absent from Donaus (and presumably from all butterflies due to the great degree of fusion of the coxae with the thorax proper) it has been impossible to identify with certainty the pleural costa. The choice of position of the pleural costa seems to lie between the pre-episternal and precoxal sutures (if indeed the pleural costa has not disappeared without a trace in the specialized lepidopterous thorax), and Matsuda agrees ( personal communication ) that considering the pre-episternal suture to be equivalent to the pleural costa gives the most satisfactory interpretation. 11—2656 322 The University Science Bulletin ■•:'.' latter sclerite is present as a separate unit (figs. 37, 38, 39) in many papilionids (including Baronia); the Satyrinae, Morphinae, Cal- inaginae and Charaxinae of the Nymphalidae; the Styginae, Rio- dininae and a few Lycaeninae of the Lycaenidae. It is not present as a separate unit in the remainder of the butterflies (fig. 9 — Pt. I). There is a great deal of variation in the size and extent of the pre-episternum and the strength of the pre-episternal suture. The pre-episternum is about as wide as the katepisternum in Anaea, Apodemia, Heliconiiis, Lycoenopsis, Megalopalpus, Speyeria and many others. It is about one-half as wide as the katepisternum in a large number of genera ( Amathusia, Anthocharis, Caligo, Calinaga, Dismorphia, Morpho, Styx, Teinopalpiis [Fig. 38] ) and is merely a narrow lip in a great many others such as Aciinote, Baronia, Danatis, Lamproptera, Papilio, Farnassius, Pieris, Phoebis, Pseudopontia, etc. (fig. 24). The greatest reduction of the pre-episternum is found in the nymphalid subfamilies Ithomiinae and Satyrinae, where it is some- times essentially absent (fig. 39). The pre-episternal suture varies from being absent in the Liby- theidae (fig. 26) to being well developed with a strong internal ridge in a great many genera {Actinote, Leptidea, Papilio, Far- nassius, Pieris, Telipna, etc. ) . The precoxal suture is absent in the Papilioninae and Baroniinae of the Papilionidae and in the Pieridae. It is obsolescent or absent in the Libytheidae and in numerous genera of the Nymphalidae and Lycaenidae (Actinote, Amathusia, Anaea, Apodemia, Atlides, Caligo, Fenesica, Lycaenopsis, Megalopalpus, Telipna, etc.). In some pierids the epimeron is fused to the meron, the suture which normally separates the two being absent anteriorly. The most extreme example of this is found in Pseudopontia. In some genera, especially lycaenids, the marginopleurite (region between the marginopleural suture and eucoxa) is relatively wide (fig. 27). Some butterflies, especially Pieridae, have the discrimen forked (often broadly) at its anterior end. The postcoxal sclerite is very variable in width and length. It is one half or more the width of the visible meron in caudal view (fig. 20 — Pt. I) in Actinote, Euselasia, Heliconius, Caligo, Danaus, Fenesica, Lamproptera, Libytheana, Lycaena, Metamorpha, Morpho, Oeneis, Speyeria and many others. In Dismorphia, Euse- lasia, Heliconius, Megalajjalpus, Ornithoptera, Papilio, Farnassius and others it is much less than one half the width of the visible meron. Phylogeny and Classification of Butterflies 323 In most genera the postcoxal sclerite ends well below the top of the meron (fig. 20 — Pt. I); however, in many nymphalids (some Ithomiinae, some Satyrinae, some Nymphalinae, and all Morphinae, Calinaginae and Charaxinae) it is long (fig. 53), reaching the top of the meron. Coxal sclerules were found only in the Danainae, Hisforis and Ithomia. They were absent in Actinote, Anaea (a trace present), Apodemia, Caligo, Calinaga, Dismorphia, Euselasia, Fenesica, Lamproptera, Leptidea, Libytheana, Lycaenopsis, Megalopalpits, Metamorpha, Ornithoptera, Papilio, Phoebis, Pieris, Telipna and many others. The epimeron was found to be very variable, both in the presence of the pre-epimeron and various secondary sutures and ridges, but also in its height below the subalare in comparison with the epimeron-subalare distance. A systematic study of these features has not been made. Another very variable character is the shape of the subalare. However, since the exact angle at which it is viewed is greatly responsible for its apparent shape, and since its position differs greatly from specimen to specimen this character was not used taxonomically. Internally the principal variation studied was in the form of the lamella of the discrimen. The lamella is continuous with the furca (the dorsum of the lamella being essentially a straight line, fig. 40) in all the butterflies except the Lycaenidae and Baronia of the Papilionidae. In the Lycaenidae the lamella does not reach the furca, but instead curves downward to the base of the furca (fig. 42). The condition of the lamella in Baronia is intermediate (fig. 41). A variable internal feature that was not studied in detail is the ventral process of the postalar portion of the epimeron. It varies both in length and in the shape of the end (conspicuously spatulate or more or less pointed ) . Metathorax: In the Libytheidae the metanotum is almost com- pletely covered by the mesoscutellum (fig. 26), being below it in position. In the other butterflies the metanotum is below and be- hind the mesoscutellum and is not completely covered by it (fig. 24). The third phragma of most of the Papilionidae is in the form of paired simple lobes (fig. 55). In Apatura, Caligo, Danaus, Morpho, Pieris and others the phragma consists of paired stalked lobes (fig. 21 — Pt. I). Most genera appear to be intermediate between the above conditions, having stalks but no lobes on their tips. 324 The University Science Bulletin In the Libytheidae (fig. 26), Fenesica, Zerynthia and a few other forms the anepisternum is pointed ventrally to a greater or lesser degree. In the remainder of the butterflies it is rounded ventrally. The relative size and position of the meron and eucoxa are sub- ject to a great deal of variation. In Dismorphia, Leptidea, Phoebis and other pierids the meron is much larger than the eucoxa and arches high above it. In Actinote, Caligo, Calinaga, Danatis, Historis, Libythea, Oeneis, Tieris, Pseudopontia and many others the meron and eucoxa are of approximately equal size, but the former may arch above the latter to a variable extent. In general in the Lycaenidae the meron is somewhat larger than the eucoxa but does not arch high above it. In the Papilionidae there is a transverse suture, represented in- ternally by a lamella ( the meral suture and meral lamella ) crossing the meron (fig. 24). This suture is absent or indistinct in the rest of the butterflies. The caudal part of the epimeron (beneath the base of the ab- domen) varies from very thin (fig. 54) to very broad (fig. 55). Because it was a difficult character to quantify it was used very little in the taxonomic portion of this work. In the vast majority of the butterflies the lamella of the meta- discrimen curves downward to the base of the furca. In certain Papilioninae, however, the lamella joins the furca above its base, and in Pseudopontia it runs straight into the furca as in the mesothorax. Legs The various stages of reduction in size and fusion of segments in the prothoracic legs of butterflies are well known. It will suffice to say that the foreleg varies from a completely developed, full sized, functional appendage bearing tarsal claws and a tibial epiphysis (in the Papilionidae) to a vestigial appendage lacking both the tarsal claws and the epiphysis, in which the entire tibia and tarsus are reduced to a small ball at the end of the femur (certain Ithomiinae). A rarely noted unique condition of the prothoracic leg of the Riodininae should be emphasized here. In this subfamily the coxa extends below the joint of the trochanter as a cylindroconical pro- tuberance (fig. 57). The pterothoracic tarsal claws of many butterflies are simple and symmetrical. However, they are strongly bifid in all the Phylogeny and Classification of Butterflies 325 Pieridae,* in Lamproptera curius (but not L. raeges) and the payeni group of Graphium in the Papilionidae, in some Acraeinae, and in some Lycaenidae. The tarsal claws also tend to be asymmetrical in the Pamassiinae and Acraeinae. Although most butterflies possess well-developed aroliar pads and pulvilli (the latter often bifid) on the pterothoracic legs, one or both of these are reduced or absent in the Papilionidae, Baltia, Colias, Gonepteryx, Nathalis and Phiilia of the Pieridae, and certain Nymphalidae (Acraea, Actinote, Agraulis, Euptoieta, etc.). Wings Most of the information in this work on wing venation has been taken from Schatz and Rober (1892), who illustrate the venation of some 480 genera of Papilionoidea. The terminology "cubitus apparently trifid (or quadrifid)" has been adopted from Clench (1955). When M^ of the forewing arises distinctly closer to Mg than to Mj and/or the vein connecting Mj and M3 appears to be a continuation of the vein connecting M3 and Cuj, then the cubitus is said to appear "quadrifid" — the four branches being actually Mo, M3, Cuj and Cu2. When M2 arises midway between M^ and M3 or closer to the former, and/or the vein connecting Mo and M3 runs at a distinct angle to the vein connecting M3 and Cuj, then the cubitus is said to appear "trifid" — the three branches being M3, Cu^ and Cuo. It should be noted that the vein called cu-v in the Papilioninae is said to be a basal vestige of the first vannal vein. Abdomen Pregenital segments: The tergum of the first abdominal segment is strongly pouched in most butterflies. However, in some papilio- nids the amount of pouching is reduced. In a few groups (Anaea, Caligo, etc.) the first abdominal tergum is largely membranous. The prespiracular bar is present in all groups except the Pieridae (fig. 25). The postspiracular bar is reduced (does not completely bridge membrane between tergum and sternum) or absent in the Papilionidae (fig. 24), Lycaenidae, and most Nymphalidae. Both bars are complete in some nymphalid genera (Danous, Ithomia, Metamorpha, etc.) and in the Libytheidae (where the postspiracular bar is especially broad, fig. 26). * In the Pieridae and Papilionidae the prothoracic legs share the characters of the pterothoracic legs, as do the prothoracic legs of the females of the Libytheidae and Lycaenidae. 356 ;;:;.= i:;THE University Science Bulletin There is considerable variation in the size and degree of scleroti- zation of the pregenital terga and sterna, but it has not been studied systematically. There are occasionally special organs on the pregenital segments, such as pads bearing specialized scales on segments 4, 5, and 6 in Caligo and a deep pouch in the tergum of the second segment o£ Faunis. Male genital segments: A number of species were found to have pseudovalves derived from the eighth tergum (e. g., Parnassius sminthetis) or the eighth sternum (e. g., Danatis plcxippiis) . Variation of significance in the higher classification of the butter- flies was found in the relative length of tlie tegumen and uncus; the presence or absence of a superuncus (fig. 62); the form of the uncus ( simple or bifid ) ; the form of the gnathos ( complete, in- complete, absent ) , the form of the valvae ( size as compared to the genitalia as a whole, thin or thick, dentate or smooth, etc.) and in the presence or absence of terminal hair brushes. The majority of the Lycaenidae possess genitalia with a more or less characteristic form (uncus not a pointed process, gnathos in the form of curved crossed arms, valvae reduced, etc.) but there was too much varia- tion to merit employing these characters to differentiate the family. Female genital segments: Although much variation is known to occur in these segments, especially in the lamellae ante- and post- vaginalis and the bursa copulatrix, no study of this variation has been undertaken. PHYLOGENY Primitive and Specialized Characters In order to make reasonable estimates of whether characters in a group studied are primitive or specialized it is usually necessary to have some knowledge of the state of the characters in the taxa pre- sumed to be ancestoral to the group under consideration. Since almost certainly the "Protopapilionoidea" have become extinct, leaving no known fossil record, it is necessary to determine the primitive state of butterfly characters by inference from their states in various groups considered to be related to the butterflies. In order to do this a brief survey of the Lepidoptera has been carried out (see "Material and Methods" for list of non-papilionoids studied) with emphasis on groups which have been thought by various authors to be relatively closely allied to the butterflies ( Hesperioidea; Castniidae; Cossidae; Tortricoidea, etc.). The Phylogeny and Classification of Butterflies 327 information obtained in this survey in combination with that gleaned from the hterature (particularly Shepard, 1930, on the pterothorax; Schultz, 1914, on the pronotum and patagia; Weber, 1924, on the thorax; and Forbes, 1923, on venation, early stages, etc.) and from trends in the character within the butterflies themselves has been used to arrive at the judgments presented in Table I. In general any character state which was found to be widely distributed in the moths was considered to be primitive among the butterflies. The development of some of the structures can be traced with ease and clarity from the primitive to the specialized state. The form of the lamella of the mesodiscrimen is such a character. In the neuropteroid insects (including Lepidoptera) there is probably no anterior sternal center of sclerotization in the thoracic segments, the anterior ventral sclerotizations in these segments consisting pre- sumably of downgrowths from the pleural (coxal) regions. The line of fusion of these downgrowths is called the discrimen ( Ferris, 1940). In the mesothorax (and to a lesser degree in the prothorax and metathorax) of the Lepidoptera this downgrowth seems to have continued, causing an invagination at the line of the discrimen and forming a thin internal "lamella of the discrimen." The lamella probably serves as a longitudinal strengthening device in the pterothoracic segments. In the Hepialidae (which show a great many primitive charac- ters— i. e., characters shared by other orders of insects ) the meso- thoracic discriminal lamella is quite small. It is fairly strongly de- veloped in tlie Cossidae and Castniidae and variable but usually rather weak in the Tortricidae, Yponomeutidae, Pyralididae and Thyrididae examined. In most of the higher moths, the skippers, and the lycaenids the lamella has reached the penultimate stage — high and strong but dipping completely to the base of the furca. In the remaining Papilionoids and in Chrysiridia (leihis but not mada- gascariensis) it has reached the. highest stage of development, being complete to and fused with the furca. Weber (1924, fig. 3, a, b, and c) shows the progressive development of the lamella in Hepialus, Zygaena and Papilio. In the above case the decision as to what is the primitive and what is the specialized case is relatively easy because of the mor- phologically logical sequence. In other cases the evidence is only slightly less conclusive. Sclerotized patagia are almost universal in their occurrence in the Lepidoptera; thus their absence in some members of an advanced group such as the papilionoids would 328 The University Science Bulletin Table I — Primitive and specialized characters in the Papilionoidea Primitive Eyes bare Eyes entire Face at most moderately protuberant Laterofacial sutures fairly close to eye margins Anterior tentorial arms simple, straight Antennae approximately one-half to one scape width apart Antennae moderately long Antennae scaled Pilifers conspicuous Mandibular rudiments large * Maxillary palpi prominent, moveable, segmented Labial sclerite sclerotized all around palpal sockets * Labial palps approximately twice length of head Cervical sclerites not joined No sclerotizations in cervix aside from cervical sclerites Patagia well sclerotized and large Parapatagia well sclerotized and large Lateral plates of pronotum fused dorsally into a Y-shaped or trian- gular structure. Presternum absent Spine of prodiscrimen absent Profurcal arms simple Spinastemum not laterally produced Spinastemum essentially an invagin- ated strip Pre-epistemum of mesothorax broad Anepistemum of mesothorax a large, separate sclerite Precoxal suture present Scuto-scutellar suture complete Metatergum not completely covered by mesotergum Lamella of mcsodiscrimen curved downward before furca Processes of second phragma small Lamella of metadiscrimen curving downward to base of furca Third phragma in the form of lobes Specialized Eyes hairy Eyes emarginate Face extremely protuberant Laterofacial sutures far from eye mar- gins or contiguous with them Anterior tentorial arms crested, curved, etc. Antennae much less than one-half or much more than one scape width apart Antennae very short or very long Antennae unsealed Pilifers reduced or absent Mandibular rudiments small Maxillary palpi not prominent, im- movable, unsegmented Labial sclerite membranous either in front or behind sockets Labial palps much longer or shorter than twice length of head Cervical sclerites joined by a ventral sclerotic strip A small anteroventral sclerite in cer- vix. Patagia not well sclerotized or small Parapatagia not well sclerotized or small Lateral plates of pronotum not fused dorsally into a Y-shaped or trian- gular structure Presternum present Spine of prodiscrimen present Profurcal arms with second anterior prong or lamella Spinastemum laterally produced Spinastemum a tubular apodeme Pre-episternum of mesothorax narrow Anepistemum of mesothorax reduced in size or not a separate sclerite Precoxal suture absent Scuto-scutellar suture obsolete cen- trally Metatergum completely covered by mesotergum Lamella of mcsodiscrimen complete to furca Processes of second phragma large Lamella of metadiscrimen not curving downward to base of furca Third phragma in the form of stalks or stalked lobes * Judgment open to some doubt. Phylogeny and Classification of Butterflies 329 TABLE I — Concluded PRIMITrVE Specialized Dorsum of first abdominal tergum not pouched or weakly pouched Prespiracular bar present Postspiracular bar absent Abdominal sclerites relatively large Abdominal sclerites well sclerotized Pseudovalves absent Tegumen well sclerotized Uncus a single, well sclerotized pro- jection Arms of gnathos fused ventrally Valvae large, broad, complete Sphragis absent Prothoracic legs normal size and functional Procoxae not extended to form a spine- like process below articulation with trochanter Protibial epiphyses present Tarsal claws symmetrical Tarsal claws simple Aroliar pad present Pulvilli present Tibial spurs present Radius 5-branched Forewing with 2V and 3V free and running to margin Hindwing with 2V and 3V present Wings with cell closed Wings evenly rounded Size moderate Body stout Fhght powerful and rapid Dull colored Colors pigmentary only Non-mimetic Pupa with partial cocoon or girdle Larvae herbivorous Dorsum of first abdominal tergum strongly pouched Prespiracular bar absent Postspiracular bar present Abdominal sclerites relatively small Abdominal sclerites weakly sclerotized Pseudovalves present Tegumen membranous Uncus bifid or weakly sclerotized Arms of gnathos not fused ventrally Valvae reduced, narrow, simple Sphragis present Prothoracic legs reduced in size, atrophied or not functional Procoxae with spinelike cylindroconi- cal projection below articulation with trochanter Protibial epiphyses absent Tarsal claws asymmetrical Tarsal claws bifid Aroliar pad absent Pulvilli absent Tibial spurs absent Radius 3- or 4-branched Forewing with 3V absent or fused after a short distance with 2V Hindwing with 3V lost Wings with cell open Wings tailed, scalloped, angulate, fal- cate, etc. Size extreme (very small or very large ) Body slender Flight weak, fluttery Brightly colored Colors in part structural Mimetic Pupa without partial cocoon or girdle Larvae carnivorous seem almost certainly to represent a specialized loss rather than a primitive state. The prespiracular bar of the butterflies was thought at first to be homologous with anterolateral apodemes found on the second ab- dominal sternite of many moths ( including the Cossidae and Casti- niidae). The bar (presumably the homologue of the apodeme 330 The University Science Bulletin lying in the membrane) was found in the butterflies, except the pierids where it is presumably lost, skippers, and a number of moths (Thyris, Desmia, Peridroma, etc.). Catocala cerogama shows a condition apparently transitional between apodeme and bar. However, in Zygaena, Peridroma and Archips, among others, both the bars and the apodemes are clearly present, making it cer- tain that the two structures are not homologous. A trace of the apodeme can also be seen in Thyris. Interestingly in Hepialus there is no sign of either the apodeme or the prespiracular bar, but there is an extension of the first abdominal tergum which is very similar to the postspiracular bar. The question of which is primi- tive, bar or apodeme, will not be answered with real assurance until an exhaustive study of the moths is completed. In this work ab- sence of the prespiracular bar is considered primitive in the Lepi- doptera as a whole, but advanced within the Papilionoidea where it is lost in the Pieridae. Similar problems have arisen in connection with other charac- ters. Is the joining of the cervical sclerites of the Papilionidae by a ventral sclerotic band an adaptation from the primitive condition of the sclerites meeting at the center of the prothoracic "sternum" (see Weber, 1924, fig 1, c [Hepialus] and f [Zygaena]), or is it a secondary advanced condition, developed after the sclerites had become free? The latter alternative is chosen here since both sclerites are free in the Cossidae, Castniidae, Hesperiidae, and the majority of the moths. In most cases, however, the decision as to which state of a char- acter was primitive and which state specialized was relatively simple. The Papilionoidea as a Taxon There is little doubt that Papilionoidea is a monophyletic taxon. With the Hesperioidea (except for Eiischemon) the Papilionoidea may be separated from the rest of the Lepidoptera by the com- bined loss of the frenulum and retinaculum and the possession of clubbed or distally swollen antennae. They may be separated from the hesperioids by the form of the head (discussed later). How- ever, these distinctions could be considered somewhat superficial if they were not accompanied by certain trends, which help to characterize the Papilionoidea. It should be noted that these trends are not necessarily universal within the Papilionoidea or absent from all other Lepidoptera. Among the trends are: loss of ocelli; extreme atrophy of maxillary palps; loss of sclerotization of Phylogeny and Classification of Butterflies 331 labial sclerite; loss of sclerotized patagia; loss of sclerotized para- patagia; development of a presternum; reduction in the size of the prothoracic legs; fusion of segments in the prothoracic legs; loss of protibial epiphyses; loss of tibial spurs; simplification of the meso- thoracic sternopleural region with loss of the precoxal suture and reduction of the anepisternum; development of a pair of prominent processes on the second phragma; extreme development of the lamella of the discrimen in the mesothorax and metathorax; de- velopment of stalks alone without lobes or stalked lobes instead of simple lobes of the third phragma; reduction, through fusion (in- cluding "stalking" ) and loss, of the number of wing veins; modifica- tion of wing shape (tails, scalloping, etc.); development of pre- and postspiracular bars; pouching of the first abdominal tergum; reduction and simplification of the valvae; development of rela- tively narrow thorax and long slender abdomen; reduction of scales and hairs (on thorax and antennae especially); upright egg; loss of use of silk at pupation; development of diurnal habits; development of brilliant pigmentary and structural colors. It should be emphasized once again that the above, with very few exceptions, are only trends developed in the Papilionoidea, not diagnostic characters of the group. Relationships of the Papilionoidea Although any definitive statements on the relationships of the Papilionoidea with the remainder of the Lepidoptera will have to await a comprehensive study of the entire order, a few provisional observations are presented here. Not surprisingly the Hesperioidea appear to be the closest living relatives of the papilionoids. With the exception of the male of Euschemon, which has a frenulum, the skippers all share the pa- pilionoid diagnostic characters of the loss of the frenulum and the clubbed antennae. They are most readily separated from the pa- pilionoids by the shape of the head, which is extremely wide in pro- portion to its height, and by the concomitant extreme separation of the bases of the antennae (the interantennal distance being at least twice the width of the scape). With rare exceptions the twelve veins of the primary wing of Hesperiidae (3V is vestigial or fused with 2V ) all arise from the cell or wing base ( i. e., are "un- stalked" ) . The best generalized morphological description of the hespcrioids is that they appear to be papilionoids which possess a great many 332 The University Science Bulletin primitive characters. Aside from the above-mentioned distinctions they differ from the papiHonoids mainly in this concentration of primitive characters * in each species, not in the characters them- selves. For instance all hesperioids examined retain the protibial epiphyses (in the papilionoids retained only in the Papilionidae ) ; both sclerotized patagia and parapatagia (among papilionoids found in certain nymphalids only); lamella of discrimen curved downward before furca (lycaenids only); reduced but still rela- tively prominent mesothoracic anepisternum (retained in various butterfly groups ) and stout body ( retained in relatively few butter- flies). Not surprisingly the hesperioids have acquired some rather ad- vanced characters, such as the extreme shape of the head, the trend towards stalks on the third phragma, the overhanging of the metano- tum by the mesonotum {Calpodes, Proteides), and the "neck" of the larva. It seems advisable at present to retain a separate superfamily for the skippers despite their obvious affinities with the papilionoids. It would be unwise to discard this well-accepted nomenclatorial practice without further study of the hesperioids. The relationships of the butterflies with the various groups of moths is less clear, partially because of the difficulty of recognizing convergence without a more thorough knowledge of the moths. However, it seems evident that the direct ancestors of the Papilio- noidea are no longer in existence. Aside from the Hesperioidea which are quite similar and closely related to the Papilionoidea, there are no groups showing clear evidence of intimate relationships to butterflies. There are, however, some groups which show a hint of papilionoid affinities, perhaps because of distant phyletic rela- tionship. The Castniidae are often mentioned as a possible papilionoid ancestor. It seems likely that these moths are a primitive offshoot of the line of Lepidoptera which leads eventually to the papilio- noids. They possess clubbed antennae; a reasonable start, for so primitive a group, toward a lycaenid-type lamella of the mesodis- crimen; a strong, but low and unarched, lamella of the metadis- crimen; an upright egg; bright colors; and butterflylike habits. Most * According to Yagi ( 1953 ) they are separated from the butterflies by having eyes of the superposition type. This is a most interesting character, if substantiated, since this type of eye is characteristic of nocturnal insects and is designed for maximum utihzation of available light. The skippers are fully as diurnal as the butterflies. However, Yagi further states that the eye is functionally of the apposition type, because the shape of the cr>'stalline cone causes the light passing through it to be concentrated on the end of the rhabdome, rather than allowing it to pass on to other ommatidia. Phylogeny and Classification of Butterflies 333 of the other characters of the group are primitive and not particu- larly associated with the papilionoid line ( e. g., ocelli usually pres- ent; tentorial bridge "inside" of head rather than at the foramen magnum; large mesothoracic anepisternum; venation complex, etc. )• The cossid studied showed a slight reduction in the size of the mesothoracic anepisternum. Aside from this and the possession by the members of the subfamily Cossinae of an upright egg, there is little to connect the Cossidae with the papilionids. None of the species of tortricoids, pyraloids, yponomeutoids, etc. studied showed any combination of characters to suggest that they are crucial to the matter at hand. They are doubtless more closely related to the butterflies than some of the more specialized moths such as the saturnoids, but beyond this little can be said. In summary, the evidence seems to indicate that the butterflies and skippers are highly evolved representatives of a line which is well isolated from all other living Lepidoptera. Interrelationships of the Families of the Papilionoidea Of the five families of butterflies only two pairs can be associated with any degree of certainty as being more closely related to each other than to the other families. These are the Papilionidae and Pieridae and the Nymphalidae and Libytheidae. The classical character which associates the papilionids and pierids is the complete development of the pro thoracic legs. Ac- companying this is the similar trend in the development of the patagia, the absence of the presternum, the loss of the precoxal suture, the similarities in the profurca and prodiscrimen, the general similarity of the head structure ( including a tendency towards loss of sclerotization in the labial sclerite) and the general tendency to- wards broadening of the spinasternum.* The quadrifid cubitus which is characteristic of the Papilionidae is found also in the Dismorphiinae of the Pieridae (ar^d nowhere else in the butterflies). The bifid tarsal claws which are universal in the Pieridae are found also in a few papilionids as well as in certain nymphalids and ly- caenids. Htjpermnestra helios is a papilionid which has facies very similar to those of certain pierids (Eticliloe sp.).\ Some workers have associated the Pieridae with the Lycaenidae rather than the Papilionidae, but the great mass of evidence is against this. The * The above characters are discussed in detail in the section on comparative morphology and in the family diagnoses. \ The investigations of Homma (1954) on the alimentary canals of butterflies, although relatively incomiiltte, tend to support the systematic relationships put forth in the present work. ■ ilomma states, "The externals of the canal of Papilionidae are somewhat similar to those of Pieridae, and those of Nymphalidae closely resemble those of Satyridae." 334 The University Science Bulletin Pieridae differ from the Lycaenidae in almost every character of the head, in the development of the prothoracic legs, in almost every character of the prothorax, in most of the characters of the meso- thorax ( including the form of the lamella of the discrimen ) and in the development of the prespiracular and postspiracular bars and male genitalia. The Libytheidae share so many characters with the Nymphalidae that very serious consideration was given to including them in the Nymphalidae as a subfamily. The libytheids differ from the nymphalids in having almost com- pletely developed prothoracic legs in the females, only small sclero- tic areas on the patagia, and the metanotum almost completely be- neath the mesoscutellum. The long labial palpi usually associated with tlie Libytheidae are short enough in some species to cause over- lap with the length of the palpi of some nymphalid genera. A num- ber of other characters which are typical of the Libytheidae also can be found within the nymphalids so that they cannot be con- sidered diagnostic. One other entity, the Baroniinae, was considered to be of possible family rank. Baronia differs from the rest of the papilionids in (among other characters) not having the spinasternum laterally produced (some other papilionids have it only slightly produced), in having a well-developed second vannal vein in the hindwing ( ab- sent or rudimentary in the other papilionids), and in having the lamella of the mesodiscrimen curved downward before it joins the furca (a condition unique in the butterflies). In this case it was decided to retain the Baroniinae as a subfamily, since these differ- ences did not seem as great as those separating the Nymphalidae and Libytheidae.* As stated earlier it does not seem wise to place labels of "primi- tive" or "specialized" on groups as closely related as the families of butterflies. However, as a matter of interest, the distribution among famihes of 36 characters for which the primitive and specialized states had been hypothesized was studied. Each character was rated for each family on a scale from 0 to 5, 0 indicating that the character was found throughout the family in its most primitive state and 5 indicating that it was found throughout the family in its most specialized state. Both intermediate states of the charac- ters and their distributions within the families were considered in estimating the intermediate values. A family possessing all the * Necessarily a partly subjective decision. Phylogeny and Classification of Butterflies 335 characters in their most specialized state in all its members would then score 180 points. The sums of the figures were determined for each family and then divided by 180 and multiplied by 100 so that they were expressed as a percentage of the maximum score. The results were: Papilionidae — 53%; Pieridae — 42%; Nymphalidae — 42%; Libytheidae — 37%; Lycaenidae — 43%. Considering the crud- ity of the method the only conclusion which is drawn from these data is that there is certainly no indication that the Papilionidae are "primitive" as has been claimed (largely because of the com- pletely developed prothoracic legs and presence of the epiphyses). Indeed this method showed them to have the highest percent of specialized characters of any family. Figure 64 is a diagram giving the author's ideas on the phyletic relationships of the families and subfamilies of the Papilionoidea. The reasons for the arrangement of the various branches are dis- cussed in this section and under the various families in the follow- ing section. The vertical scale is my judgment of what might be called evolutionary distance, being evolutionary rate multiplied by time. It is, of course, impossible to distinguish these two quantities on the basis of neozoological evidence. The horizontal positions of the taxa and spacing of the lines are determined by convenience. CLASSIFICATION Key to the Families of the Papilionoidea* 1. Lamella of mesodiscrimen continuous with furca, dorsum of the lamella essentially straight (somewhat downcurved in Baronia ) ; eyes not emarginate; patagia either sclerotized or unsclerotized 2 Lamella of mesodiscrimen not continuous with furca, curving downward to base of furca; eyes usually emarginate; patagia unsclerotized Lycaenidae 2. Cervical sclerites not joined beneath cervix; epiphyses absent; hindwing with two distinct vannal veins 3 Cervical sclerites joined or nearly joined by a sclerotic band beneath cervix; epiphyses present; hindwing with only one distinct vannal vein (two in Baronia) Papilionidae 3. Prespiracular bar well developed; prothoracic legs atrophied (smaller than the pterothoracic legs) to some degree (only shghtly in females of Libytheidae); tarsal claws very rarely strongly bifid 4 * Since butterfly specimens have always been, and will continue to be identified by comparison of facics and genitalia with illustrations, no attempt has been made to con- struct superficial keys. This key and those that follow are based on what appear to be the most dependable characters, regardless of accessibility. Wh o c O 4- c c := o o e- o C O "O o o o o o o « o ^ . o c .= — o CDQ.Q. 00.00. 2IO(/)SooZ< «> o „ :i 2 .5 .E K ogr2-s~E2 O .■5 § • a> CO «> o o >» -o • •* .^ *• ^^ -1 CC <0 -J UJ < o 2 o o. < o. Ul < £ tu o. Ui a. > UJ < UJ X I- > CD 64 UJ < o z UJ < o >■ Fig. 64. Relationships of the families and subfamilies of the butterflies. THE UNIVEKSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] November 18, 1958 [No. 9 The Tenagogonus-Limnometra Complex of the Gerridae * BY H. B. HuNGERFORD and Ryuichi Matsuda In 1853, Stal described the genus Tenagogonus in Ofvers af K. Vet.-Ak. Forhandl. 10. p. 263 as follows: "Corpus subellipticum Caput subtriangulare, convexiusculum, tuberculis antenniferis parum productis. Antennae corpore longiores, articulis inter se longitudine fere aequalibus, capite nonnihil longioribus; inter articulum 2 et 3 articulus minimus, globosus adest. Rostrum capite ter fere longius, art. 1 capite vix dimidio breviore, 2 hujus dimidia longitudine, 3 capite nonnihil longiore, 4 hoc plus dimidio breviore. Thorax ab- domine multo longior, convexus, lateribus subobliquis. Hemelytra et alae desunt. Abdominis segmentum penultimum utrimque ramu- lum emittens. Pedes posteriores longissimi; tarsi antici breves, crassiusculi, posteriores longiores, graciliores. (spec. 1.)" He did not designate a species but in the introduction to this paper "Nya Genera bland Hemiptera," he states they came from "Cafferlandet." In 1855, in the same publication vol. 12, p. 45 he described: "Ten- agogonus albovittattis: supra brunnescens, subtus dilute sordide flavescens, albidosericeus; capitus vittis tribus tuberculisque antenni- feris, thoracis linea longitudinali posterius evanescente, vittis 2 anticis abbreviatis, 2 lateralibus aliam dilute flavotestaceum in- cludentibus rostrique art, ultimo nigris; antennis pedibusque fusco- testaceis, his nonnihil dilutioribus. Long. 7-9, lat. 3-3/2 millim. — In terra natalensi." Thus the genus Tenagogonus Stal was set up for a species which Stal later described as Tenagogonus albovittatiis. Fortimately, al- though the type of this species is in the Museum of Stockholm, Sweden, we have a good series of specimens from West Africa that Dr. O. Lundblad has compared with the type and says is the same species. These are all apterous but we have five specimens from "Gabon," three of which are winged (one male and two females). * Contribution number 974 from Department of Entomology, University of Kansas. This study was made possible with the aid of a grant from the National Science Foundation. (371) 372 The University Science Bulletin The distinguishing characters of Tenagogonus as set forth by Stal would be: a subelHptical body, a subtriangular head with small antennal tubercles, antennae longer than the body, its segments of nearly equal length, each longer than the head and a small inter- segment between segments two and three. Beak nearly three times as long as head, segment one about half as long as head, segment two very short, three somewhat longer than head and four shorter than half. Thorax much longer than abdomen. Pentdtimate ab- dominal segment with a horn on either side. Hind legs longest. Front tarsus short, its second segment longer than the first. Then in 1865, Mayr described the genus Limnometra in Verhandl. Zool.-bot. Vereins in Wein, Bd. XV, p. 443. This he described as being closest to Hydrometra (now Gerris Fabricius) with slender antennae as long as body. Hind femora very long, as long as body. Intermediate femora bidentate at apices. Front tarsi with first seg- ment as long or a little longer than the second. Then he described L. fetnorata, L. nigripennis, L. pulchra, L. ciliata, L. inermis and L. mimita. Then in Sept. 1865 in Novara-Expedition, Zoologischer Theil, Bd. II, Abth. 1, Hemiptera, p. 174, he gave a key to seven species of Limnometra and rewrote the generic description this time saying: Size of Hydrometra. Antennae filiform, very slender toward the apex, as long as the body or sometimes shorter, basal segment as long as the anterolateral margin of the pronotum. Beak extending to middle of mesosterniim. Eyes distinctly emarginate on inside. Front margin of pronotum straight, posterior part ex- tending over mesonotum, posterior process triangular. Winged. Front tarsi with t\vo claws, the segments subequal in length. Middle femora apically bidentate, hind femora as long as the body. The type of the genus Limnometra is L. femorata Mayr as stated by Dr. Lundblad. If we compare the descriptions of Tenagogonus Stal and Lim- nometra Mayr we find they agree in having long slender antennae, long slender middle and hindlegs and rather long beaks. Possible differences might be that in Tenagogonus Stal "Thorax much longer than the abdomen" and "front tarsus short, its second segment longer than the first." While in Limnometra Mayr "front tarsi have two segments subequal in length" or "first segment a little longer than the second," "Beak extending to middle of mesosternum" and "Middle femora apically bidentate." The last character is not mentioned in the Tenagogonus description nor how far back on the mesosternum the beak extends. If we examine T. albovittatus Stal and L. femorata Mayr side by Tenagogonus-Limnometra Complex 373 side they do not appear congeneric. The first is a small species from Africa semielliptical in shape and the male lacks connexival spines but has two conspicuous processes on the ventrolateral mar- gin of the last abdominal segment (7th abdominal segment). St^l mentions these in his generic description as arising from the penulti- mate abdominal segment which morphologically speaking is not true; moreover, he does not indicate that it is a male character. The abdomen is indeed greatly reduced and much shorter than the thorax, and the enlarged last ventral abdominal segment plus the genital segments in the male are longer than all the preceding ab- dominal segments. However, the hind legs are shorter than the middle legs, not longer as stated by Stal in his generic description. In Limnometra femorata Mayr we have a large elongate species from the Philippines. The abdomen of the male is elongate, as long or longer than the thorax, with well-developed connexival spines, the last ventral abdominal segment being shorter than the preceding segment and lacking protuberances. Indeed the last abdominal segment plus the genitalia are not as long as the two preceding segments. The beak does not extend to the middle of the mesosternum as stated in Mayr's generic description. The middle femur is apically bidentate, a character we will find of little use. The front tarsal segments are subequal in this species. In 1901, Champion described Limnometra opaca from Panama and Limnom,etra qttadrilineata from Mexico. (Biologia Central! Americana, Rhynchota vol. II, pp. 180-181.) In 1909, Kirkaldy and Bueno in their Catalogue of American aquatic and semi- aquatic Hemiptera (Proc. Ent. Soc. Washington, vol. X p. 209) gave Limnometra Mayr 1865 as a synonym of Tenagogonus St^l 1853 and placed opaciis (Champion) and qtiadrilineatus (Cham- pion) under Tenagogonus StM. In 1915, Doctor E. Bergroth in a paper entitled "Some Javanese Hemiptera collected by E. Jacobson and Th. H. Mac Gillavry" (Zoologische Mededeelingen uitgegeven vanwege's Rijks Museum van Natuurlijke Historic te Leiden, Deel 1, aflev. 2. pp. 121-123) describes Tenagogonus pravipes and states that Limnometra Mayr is a synonym of Tenagogonus Stal and that Limnogonus StA,l is not even subgenerically distinct from the above, because "the transi- tions in the mutual length of the two joints of the fore tarsi and in the apical angles of the last abdominal segment are too numerous." However, Limnogonus Stil is still recognized as a genus and useful as such and his submersion of Limnometra St&l is accepted by some and rejected by others. The use of Tenagogonus Stdl for certain new world gerrids now placed in the genus Tachygerris 374 The University Science Bulletin Drake, was following Kirkaldy and Bueno, 1909. Dr. Poisson, in various papers dealing with African gerrids, follows Bergroth. In 1925, Dr. W. E. China described Limnometra gigas from China, thus bringing again to our attention the genus Limnometra Ma>T. (Bull. Brooklyn Entomological Society vol. 20, p. 218.) In 1933, Dr. O. Lundblad under Limnometra Mayr lists seventeen species and describes Limnometra brevis. He states that Limno- metra femorata Mayr and Tenagogonus albomttatus StS.1 are very dissimilar species and that he is not convinced by Bergroth's opin- ion that the genera are identical. (Archiv fiir Hydrobiologie, Suppl. Bd. XII, "Tropische Binnengewiisser 4," 388-392, 1933.) Unfortunately, he does not state how these genera may be sepa- rated. Several years ago the senior author became aware of the need for better definition of genera in the Gerridae when called upon to detennine species from the East Indies. Dr. Lundblad placed the species Tenagogonus pravipes Bergroth, Gerris kampaspe Kirkaldy and L. brevis Lundblad in the genus Limnometra Mayr along with L. femorata Mayr and other typical Limnometra species and the senior author thought that Tenagogonus pravipes Bergroth was correctly named and that G. kampaspe Kirkaldy and L. brevis Lundblad were congeneric with it. This has led us into a detailed study of all species we could se- cure in these two genera. We have re-examined most of the types, made additional descriptive notes, and offer illustrations of them and their structural characteristics. We have studied thirty-seven species from the Old World, nine of them being new. We have placed Limnometra gigas China in the new genus Gigantometra Hungerford and Matsuda and we are raising T enagometrella Pois- son to generic rank to include T. grandiusculus Poisson and T. longi- cornis Poisson. The other thirty-four species have been examined and their position determined on some thirty-five characters. Many of these characters were discarded as generically insignificant when they did not hold for even obviously closely related species. For example, "antennae as long as the body" holds for males of some species but not for their females, while in other very closely related species the male antennae are also shorter than the body. We have made a chart of the more significant characters and indicated, by number, the species that have these characters. By studying this chart it will be observed that certain species are al- ways found in the line describing a character that is typically Tenagogonus, and others are always in a line describing a character that is typically Limnometra. In addition there are a few species Tenagogonus-Limnometra Complex 375 that may possess a combination of characters that make them dif- ficult to place. For example in No. 25 Tenagogonus kuiterti sp. nov., the male has very small but somewhat variable connexival spines, yet it has the abdominal spiracles in the middle of the segments, has the second front tarsal segment much the longer, the male ab- domen is shorter than the mesosternum, and the hind coxae cover the second abdominal segment, all of which relate it to Tenagogo- nus. While all true Tenagogontis have the second tarsal segment of the front leg longer than the first there are some Limnometra, where the segments are subequal or the second segment a little longer. It is no wonder that Dr. Bergroth gave up trying to separate these genera. However, we do not believe in uniting established genera every time an intermediate species is found. For if we were to follow this practice consistently we would have to unite the fami- Hes Chrysomelidae and Cerambycidae in Coleoptera, the orders Homoptera and Hemiptera, as some have done and finally unite the Plant and Animal kingdoms! A genus should include a group of closely related species and it is difficult to see close relationship between T. albovitfatus Stal and Limnometra femorata Mayr. It is true that the generic descriptions were both inadequate and in- accurate and we have struggled with this problem for many, many months. We hope that this contribution will be helpful in the identification of species belonging to this complex. We believe it convenient and useful to retain both generic names and redefine them. Numbers and the Species for Which They Stand in the Chart of Significant Characters 1. albovittatns Stal 20. brevis Lundblad 2. kampaspe Kirkaldy 21. minuta Mayr 3. divergens sp. nov. 22. robustus sp. nov. 4. madagascariensis Hoberlandt 23. pravipes Bergroth 5. dubms Poisson 24. Jipovsky sp. nov. 6. femorata Mayr . 25. hiiterti sp. nov. 7. kirkaldiji Breddin 26. bornecnsis sp. nov. 8. /?f /uf oru m (Fabricius) 27. fijiensis sp. nov. 9. gigas (China)** 28. pravipes bergrothi subsp. nov. 10. octopunctata Hungerford 29. ntidus Poisson * 11. ciliata Mayr 30. hirsiittis Poisson * 12. cursitans (Fahncius) 31. /ant/gint^ws Poisson * 13. pulchra Mayr 32. zambezinus Poisson 14. anntdicornis Breddin 33. rossi sp. nov. 15. nigripennis Mayr 34. insularis sp. nov. 16. inermis Mayr 35. kaUisto (Kirkaldy) 17. vulpina Breddin * 36. longicornis Poisson ** 18. an<7jyomene (Kirkaldy) 37. grandiiisctdus Voisson ** 19. palauamx EsAi 38. euphrasy tie CKlAaldy)** * known to us only by descriptions and drawings in the literature. * * removed to a new genus recently or its subgenus is going to be raised to generic rank. 376 The University Science Bulletin Chart of Significant Characters abdomen c?" < mesosternum 1,2,3,4,5,20?.22,23?,25,27,28,32,38 abdomen cf = mesosternum 18 abdomen d^ > mesosternum 6,7,8,10,11,12,13,14,15,19,24,26,33,34,35 last cT abdominal seg. + gen. = or > than 5 preceding v. abd. segs. 1,2,3,4,5,22.27,28,32,38 last cf abdominal seg. + gen. = 4 preceding v. abd. segs. 10,25,33,(24 a wingless d') last cf abdominal seg. + gen. < 4 preceding v. abd. segs. 6,7,8,11,12,13,14,15,18,19,24,26,34,35 Ist tarsal seg. < 2nd (front leg) 1,2,3,4,5,18,20,21,22,23,25,27,28,32,33,38 1st tarsal seg. = 2nd 10,14 1st tarsal seg. > 2ad 6,8,11,12,13,15,19,24,26,34,35 abd. spiracles in middle 1,2,3,4,5,18,25,27,28,32 abd. spiracles closer to anterior margin 6,7,8,10,11,12,13,14.15,19,24,26,33,34,35 abd. spiracles variable 24,38 cf without connexival spines 1,2,3,4,5,20,22,23?,27,28,32,(24 in wingle.ss cf") cf with very short connex. spine 24,25,38 cf with connexival spines or triangular plate 6,7,8,10,11,12,13,14,15,18,19,21,26,33,34.35 cf with processes on last ventral abd. seg. 1,2,3,4,5 first two abd. segs 1,2,3,4,5,10,20,22,23,25,26,27,32,38 d^ metasternum = first two abd. segs 15,18,19,24 d^ metasternum < first two abd. segs 6,7,8,11,12,13,14,33,34,35 rf' last ventral abd. seg. > preceding seg. 1,2,3,4,5,20,22,237,25,27,28,32,38 (f last ventral abd. seg. = preceding seg. 10,18,19? d^ last ventral abd. seg. < preceding seg. 6,7,8,11,12,13,14,15.24,26,33,34,35 Tenagogonus-Limnometra. Complex 377 Key to TENAGOGONus-LI^^NO^IETRA Complex * 1. Male abdomen reduced. Ventral abdominal segments short, last segment plus the genitalia at least as long as preceding four segments, usually longer. Hind coxae of male nearly reaching or surpassing posterior margin of second ventral abdominal segment. Male without typical conne.xival spines. (Genus Tenagogonus Stal) 2 Male abdomen not reduced. Last ventral abdominal segment plus genital segments shorter than preceding four segments. Hind coxae of male rarely extending beyond middle of second ventral abdominal segment. Male with connexivum pro- duced into a triangular flattened plate or spinelike process la la. Pronotum with median longitudinal white line: Omphalium with transverse channels leading to lateral scent- gland pores, guarded by hairs, (Genus Gigantometra Hungerford and Matsuda) Omphalium, a circular flat inconspicuous pore, farther from the rear margin of metastemum than in Limnometra. No transverse channels ( Genus Tenagometrella Poisson ) Pronotum with median longitudinal black line ( Genus Limnome- tra Mayr) 10 2. The last ventral abdominal segment of male with two hornlike protuberances on its caudal edge or venter Tenagogonus 3 The last ventral abdominal segment of male without such pro- tuberances ' 3. The protuberances arise on the edge of lateral flaps of the last ventral abdominal segment . . . T. madagascariensis Hoberlandt p. 383 The protuberances not arising as above 4 4. The conspicuous protuberances arising just below distal lateral edges of the connexivum in the male 5 The protuberances arising farther ventrad 6 5. Protuberances upcurved and nearly parallel T. zambezinus (Poisson) p. 384 Protuberances nearly straight and diverging T. divergens, sp. nov. p. 385 6. The incurved hornlike protuberances arising laterally from the caudal margin of the last ventral abdominal segment, T. albovittatus St&l p. 382 The short sloping conical protuberances on the ventral surface of the last abdominal segment T. kampaspe (Kirkaldy) p. 386 7. The male genital capsule ( 9th segment ) with lateral hair tufts . 8 The male genital capsule without lateral hair tufts 9 8. The first genital segment of male medianly depressed ventrally with a large V shaped depression in its rear margin, r. robustus sp. nov. p. 388 The first genital not as above T. brevis ( Lundblad ) p. 390 * Omitting Tenagoponus subgen. Tenagomctra Poisson 1948 in which he places T. hirsu- tus Poisson. T. lanuginrus Poisson and T. ntidits Poisson species which we have not seen. We place in the subgenus TcnagOfionus the following: T. albovittatus Stul, T. madagascar- iensis Hoberlandt, T. zambezinus Poisson, T. divergens n. sp. and T. kampaspe (Kiikaldy). 378 The University Science Bulletin 9. The first genital segment of male with a median longitudinal de- pressed line on its venter .. . T. pravipes bergrothi * subsp. nov. p. 393 The first genital segment of male without the depressed line, T. fijienses sp. nov. p. 394 10. Connexivum produced into a triangular flattened plate, Limnometra minuta Mayr p. 397 Connexivum produced into a spinelike process or an acute pro- jection 11 11. Connexival spines shorter than first genital in the male and in the apterous males may be little more than an acute projec- tion 12 Connexival spines longer 14 12. First genital segment of male with its venter normal, rear mar- gin unmodified T. kuiterti sp. nov. p. 395 First genital segment of male with its venter modified 13 13. The first genital segment with a backward pointing protuberance on either side and the venter of the body pale, unspotted, L. lipovskiji sp. nov. p. 399 The first genital segment with its rear ventral margin angularly produced on either side. Eight large black spots on the venter L. octopunctatus Hungerford p. 400 14. Middle coxa with a spinehke projection on the dorsolateral rear margin L. jluviorum (Fabricius) p. 401 Middle coxa without such projection 15 15. Moderately small species. Head width usually under 1.7 mm., even largest males never over 1.76 mm 16 Larger than above. Head width usually over 1.8 mm 20 16. Front femur as slender as middle femur 17 Front femur stouter than middle femur 18 17. Antennal segments uniform in color, moderately strong. Second tarsal segment of front leg longer than first. Connexival spines of female very short, not surpassing abdomen, L. anachjomcne (Kirkaldy) p. 402 Antennal segments three and four and sometimes distal half of second, white, the segments slender and long. First tarsal segment of front leg longer than second. Connexival spines of female long and slender, surpassing abdomen, L. instihris sp. nov. p. 404 18. First tarsal segment of front leg longer than second; wing veins dark brown L. palauana Esaki p. 405 First and second tarsal segments subequal; wing veins not dark brown jg 19. Underside of body pale grayish, including connexivum. The median longitudinal reddish brown line on pronotiim, slender and faintly margined by paler bands; marginal dark line ^^^^ L. borneensis sp. nov. p. 407 * This species comes from the Philippines but it seems to fit the description of T. pravipes Bergroth from Java and Dr. Bergroth may have overlooked the character on the first genital segment of the male. We are unable to locate the Bergroth types. Tenagogonus-Limnometra Complex 379 Underside of body with brown band on front acetabula and ex- tends back on sides of mesothorax; some brown on the meso- acetabula and underside of connexival segments. The median longitudinal black line on pronotum broad and bordered by broader pale bands, marginal black band prominent, L. rossi sp. nov. p. 408 20. Antennae, middle and hind femora uniform in color, the latter not paler near tip. Middle femur of male with conspicuous long cilia 21 Antennae, middle and hind femora not uniform in color, the latter at least shghtly to conspicuously paler near tip. Middle femur of male may have cilia ( but not as long as the diameter of the femur) 22 21. Middle femur with a short black hne near base dorsally. Rear ventral margin of first genital segment of male produced laterally (see pi. 14, fig. 22d.) L. ciliata Mayr p. 409 Middle femur vdthout the black hne. Rear margin of first geni- tal segment of male with faint production laterally, L. cursitans (Fabricius) p. 413 22. Middle femur of male with two rows of pegs and a large curved spine near distal end L. femorata Mayr p. 413 Middle femur of male may have two rows of pegs but vdthout the large curved spine near distal end 23 23. Middle femur of male without definite fringe of long ciha. Hemelytra usually chocolate browm with costal margin yel- low or orange. Mesosternum not ciliated. Second and third antennal segments usually ringed with white and last seg- ment white L. nigripennis Mayr p. 415 Middle femur of male vdth definite fringe of ciha. Hemelytra not as above. Mesosternum of males usually cihated. An- tennae not as above 24 24. Both middle and hind femora of male vdth definite fringe of ciha L. pulchra Mayr p. 418 Only middle femora of male with definite fringe of ciha 25 25. Second antennal segment annulated, shorter or subequal to vddth of head across eyes. First and second tarsal segments of front leg subequal L. annulicornis (Breddin)* p. 420 Second antennal segment not annulated, usually longer than width of head across eyes. Furst tarsal segment of front leg plainly longer than second L. kallisto ( Kirkaldy ) p. 422 * L. vulpina Breddin — The type female from "N. [Matinag Kitte, Suds"ite, 800-1200 m (Sar. )]" we cannot locate. Since Dr. Breddin identified specimens of L. pulchra Mayr as his L. annulicornis and the females of L. pulchra have shorter connexival spines he prob- ably compared his female L. vulpina with a female of L. pulchra Mayr and said its con- nexival spines were longer than L. annulicornis when his L. vulpina was really a female of his L. annulicornis. 380 The University Science Bulletin ACKNOWLEDGEMENTS We wish to express our appreciation for the assistance of the following who have made this study possible: To the National Science Foundation for its financial assistance which has made possible the completion of this paper. To Dr. Max Baier, of the Natural History Museum in Vienna, Austria, who made possible the re-examination of the Mayr types of Limnometra and most generously gave us the information we needed regarding them. To Dr. W. E. China and Mr. Izzard of the British Museum for the loan of material and for information about types. To Dr. O. Lundblad of the Natural History Museum in Stock- holm for comparing specimens with the StS.1 type and for sending a specimen of his L. brevis for our study. To Dr. Ernst Sutter of the Museum of Natural History in Basel, Switzerland, who had included in material sent for determination some Limnometra from Indonesia that added to our records of dis- tribution. To Dr. Raymond Poisson of the University of Rennes, France, who has most kindly helped us locate certain species. To Dr. Haken Lindberg of Helsinki, Finland, for his attempt to locate the Bergroth types. To Dr. Ludvik Hoberlandt of the National Museum at Praha for the gift of paratypes of his Tenagogonus madagascariensis. To Dr. Teiso Esaki of Kyushu University, Fukuoka, Japan, for sending us specimens of his Limnometra palauana. To Dr. Herbert M. Hale of the South Australian Museum for an opportunity to study a long series of one Australian species. To Dr. T. E. Woodward of the University of Queensland, Aus- tralia, for sending specimens desired. To Dr. J. L. Gressitt of the Bemice P. Bishop Museum of Hono- lulu, Hawaii, for the loan of material from some of the Pacific Islands. To Dr. E. S. Ross of the California Academy of Science for the loan of material from the Pacific. To Dr. Robert L. Usinger of the University of California for the loan of specimens from the Pacific. To Dr. Rupert L. Wenzel of the Chicago Natural History Museum for the loan of material from the Phihppines taken by Dr. Hoog- straal, et al. Tenagogonus-Limnometra Complex 381 To Dr. Mont A. Cazier of the American Museum of Natural History, N. Y., for sending material of this group for determination. To Dr. J. Bequaert of the Museum of Comparative Zoology, Harvard College, for sending material for determination. To Dr. Reece Sailor of the U. S. National Museum for the op- portunity to study the undetermined material of that museum. And finally to Louis Kuitert who collected in Burma, and Louis Lipovsky who collected in Guadalcanal, former students of the University of Kansas who found themselves serving in the armed forces during the war. Both of these men sent back species new to science. Systematic Treatment The genera Tenagogonus Stal and Limnometra Mayr have in common: Rather long antennae, slender beaks with third segment reaching onto the mesostemum, long middle and hind legs and venation of hemelytra. The Genus Tenagogonus Stal 1853 Stal, C. Of Vet. Akad. Forh. 10 pp. 263-264. (Described from "Caffer- landet" for one species unnamed. ) 1855 Stal, C. Of Vet. Akad. Forh. 12 (1): 45. (Described Tenagogonus albovittatus, which is the type of the genus. ) 1865 Stal, C. Hemiptera Africana 3: 168-169. (Assigns to Tenagogonus Stal the species T. fluviorum (Fabricius) and T. swakopensis Stal but does not mention his 7'. albovittatus. ) 1909 Kirkaldy, G. W. and Torre Bueno, J. R. de la. Proc. Ent. Soc. Washington, 10: 209. (Give Limnometra Mayr as synonym of Tenagogonus Stal.) 1915 Bergroth, E. Zool. Med. Rijks Mus. Nat. Hist. Leiden, 1 (2): 121-123. (Gives Limnometra Mayr as synonym.) 1916 Bergroth, E. Proc. Nat. xMus. 51, 2150: 237. 1933 Lundblad, O. Archiv fiir Hydrobiologie 1933 Suppl. 12, "Tropische Bin- nengewasser 4," pp. 388-392. 1948 Poisson, R. Mem. Inst. Sci. Madagascar, ser. A, 1 (2): 94-95. (Proposes subgenera of Tenagogonus.) Rather small gerrids, with moderately long antennae, middle and hind legs. Male with last ventral abdominal segment plus genital segments equal to or greater than preceding four ventral segments, usually as long as or longer than preceding five ventral abdominal segments. Front leg with first tarsal segment shorter than second. Abdominal spiracles in middle of segments. Male without or with very short connexival spines. Males with hind coxae covering second ventral abdominal segment. Male metasternum longer than first two ventral abdominal segments. Male last ventral abdominal segment greater than preceding segments. In Tenagogonus the males are usually smaller than the females. 382 The University Science Bulletin Tenagogonus albovittattis Stal (PI. 1, fig. 1; PI. 7, fig. 1) 1853 Tenagogonus. Sta\, C. Of Vet. Acad. Forh., 10: 263-264. (New genus described from "Cafferlandet" for one species unnamed.) 1855 Tenagogonus alhovittatus Stal, C. Of Vet. Akad. Forh., 12 (1): 45. ( Described from Natal. ) 1916 Tenagogonus alhovittatus Bergroth, E. Proc. Nat. Mus. 51 (2150): 237. 1940 Tenagogonus alhovittatus Poisson, R. nee Stal. Bull. Mus. Roy. Hist. Nat. Belg. 16 (40): 5-7 (= T. zamhezinus Poisson.) 1941 Tenagogonus alhovittatus Poisson, R. nee Stal. Rev. Franc. Ent. 8 (12): 77. 1948 Tenagogonus alhovittatus Poisson, R. nee Stal Mem. Inst. Sci. Madagascar 1: 95. 1952 Tenagogonus alhovittatus Poisson, R. nee Stal Res. Nat. Integrale du Mt. Nimba. 13: 282. 1954 Tenagogonus (s. g. Tenagogonus) alhovittatus Poisson, R. Institut d(S Pares Nationaux du Congo Beige. Exploration du Pare Nat. de I'Upemba 31, 3-4. Oiir specimens are represented by one winged male and two winged females, one wingless male and one wingless female from Gabon, French Equatorial Africa; nine wingless females from Sangmelina, West Africa. Size: 6.7 — 7.4 mm. long; width across mesoacetabula 2.4 — 2.6 mm. in wingless male. 7.3 — 7.6 mm. long; width 2.85 — 3.1 mm. in wing- less female. 9.2 mm. long in winged male and 7.4 mm. long in winged female. Color: Dark reddish brown in ground color. Head above with median black longitudinal stripe continuous with black clypeus and a pair of lateral black stripes. Rostrum yellowish brown except for its last segment which is black. Antennae dark reddish brown. Pronotum with median black longitudinal stripe bordered by yel- lowish area and wide black spot on either side of its median longi- tudinal stripe in anterior lobe, lateral and posterior margins also black as shown in figure. Mesonotum on either side of pronotum with a broad white area. Body beneath yellowish. Structural characteristics: Proportional length of antennal seg- ments: 1st: 2nd: 3rd: and 4th:: 52: 55: 60: 65 in one wingless male, and 50: 45: 47: 63 (curved) in one wingless female. Front femur about as thick as middle femur and slightly curved in both sexes. The relative length of leg segments in a wingless male: Femur Front leg 76 Middle leg 183 Hind leg 175 First tarsal Second tarsal Tihia segment segment 67 10 12 150 46 12 80 17 9 Tenagogonus-Limnometra Complex 383 Hind legs much longer than length of body. Structures of seventh and first genital segments are quite characteristic in male as shown in the figures. Winged form: shape of pronotum and hemelytra as shown in the figures. Scs vein joins with R -f M at the point of forking of the latter. Location of types: Naturhistoriska Riksmuseum, Stockholm. Comparative notes: Broad white band on the mesonotum and peculiar structure of the seventh ventral abdominal and first genital segments in the male readily separate this species from the other related species within the genus Tenagogonus. Tenagogonus du- bitis Poisson, according to personal letter from Dr. Poisson, is closely related to T. alhovittattis, if not synonymous. The speci- mens from Congo Beige identified as T. diibiiis by Dr. Poisson shows actually very little difference from the specimens from Cam- eroons, West Africa, identified as T. albovittatus Stal by Dr. Lund- blad. Whether these two species are the same species or not awaits further investigation. Tenagogonus madagascariensis Hoberlandt (PI. 1, fig. 3; PI. 7, fig. 3) 1947 Tenagogonus madagascariensis Hobedandt, Acta Ent. Mus. Nat Prag 25: 105-112. 3 plates. 1948 Tenagogonus madagascariensis Poisson, R. Mem. Inst. Sci. Madagascar, 1: 89. Figs. 5, 6, 7. 1948 Tenagogenus (s. g. Tenagogonella) madagascariensis Poisson, Mem. Inst. Sci. Madagascar, 1: 93-94. Size: 7.3 mm. long; width across mesoacetabulae 2.2 mm. in a wingless male paratype. 8.3-8.9 mm. long; width across mesoacet- abula 2.95 - 3.35 mm. in wingless female paratypes. Color: Yellowish brown to fuscous in general color. Head with usual three black longitudinal stripes on upper surface. Pronotum with median longitudinal black stripe and broad lateral stripes, posterior lobe occasionally nearly black, subject to individual vari- ation in degree of black pigmentation. Mesonotum yellow on median longitudinal axis, sides fuscous; mesopleural region with broad longitudinal fuscous band. Metanotum and basal abdom- inal tergites darker, sometimes entirely black. Body beneath yel- low, last rostral segment black. Structural characteristics: Proportional length of antennal seg- ments: 1st: 2nd: 3rd: 4th:: 52: 52: 62: 83 in one wingless para- type male; 50: 40: 45: 60 (? curved) in one wingless paratype female. 384 The University Science Bulletin The relative length of leg segments in one wingless male: First tarsal Second tarsal Femur Tibia segment segment Front leg 72 60 7 10 Middle leg 163 130 53 12 Hind leg 155 73 18 11 Middle legs longer than length of body. Front femur a little thicker than base of middle femur; middle femur much thicker than hind femur. Pronotimi prolonged to about middle of meso- notum. No winged form is known. Location of types: National Museum of Praha. Three wingless males and three wingless female paratype specimens are at Uni- versity of Kansas. Comparative notes: This species can readily be distinguished from the rest of the species of Tenagogonus s. sir. by the short pronotum. The seventh abdominal and first genital segments are conspicuously modified ventrally, as in other species of Tenagogonus s. str., but they can be recognized from the accompanying figures. Data on distribution: Madagascar: "Vohemar, Ambomja" (type series, Hoberlandt); "Mahilaka (Sambirano); foret des Roussettes (Montagne d'Ambre)" (Poisson); "(Tananarive)" (New record). 28 wingless males and 56 wingless females from Tananarive, Mada- gascar are preserved in the Francis Huntington Snow Entomological Museum, University of Kansas. Tenagogonus zambezinus (Poisson) (PI. 1, fig. 2; PI. 7, fig. 2) 1934 Gerris (Lininoporus) zambezinus Poisson, R. Bull. Soc. Zool. France, 59 ( 1 ) : 92-93. 1940 Tenagogonus albovittatus (? Gerris [Limnoporus] zambezinus Poisson) Poisson, R. Bull. Mus. roy. Hist. nat. Belg., 16 (40): 5-7. 1941 Tenagogonus albovittatus Poisson, R. Rev. Franc. Ent., 8(2): 77. 1948 Tenagogonus albovittatus Poisson, R. Mem. Inst. Sci. Madagascar, 1: 95. 1952 Tenagogonus albovittatus Poisson, R. Res. Nat. Integrale du Mt. Nimba, 13: 282. Dr. W. E. China kindly loaned us one identified wingless male and one female specimen. The following description is based on them. Size: 8.5 mm. long; width across mesoacetabula 2.77 mm. in one wingless male. 9.9 mm. long and 3.75 mm. wide in one wingless female. Color: Pale reddish brown in ground color, with black markings on dorsal surface; black pigmented area densely clothed with short Tenagogonus-Limnometra Complex 385 adpressed grayish hairs. Head with usual three black stripes. Pro- notum with median black longitudinal stripe and lateral narrow marginal black stripes evanescent apically, a pair of broad short black bands on either side of middle of anterior lobe. Mesopleural region with three black stripes which are confluent anteriorly and posteriorly, each abdominal tergite from second segment on with a pair of black markings. Connexivum darker laterally or with a continuous median longitudinal black stripe. Rostrum yellow except the last segment black. Body beneath pale yellow. Structural characteristics: Proportional length of antennal seg- ments: 1st: 2nd: 3rd: 4th:: 24: 23: 27: ? in wingless male and 26: 20: ?: ? in wingless female. The relative length of leg segments in one wingless male is as follows: First tarsal Second tarsal Femur Tibia segment segment Front leg 77 63 10 12 Middle leg 172 147 40 ? Hind leg 165 90 18 10 Seventh segment in male with ventrolateral processes long and they arise more dorsally than in the preceding species. Eighth segment without conspicuous modification ventrally in male. Location of types: Musee Royal d'Histoire Naturelle de Belgique. Comparative note: Relatively large. The shape and location of the process on ventrolateral margin of seventh segment in male and the simple posterior margin of eighth ventral abdominal segment readily separate this species from the rest of Tenagogonus. Data on distribution: Mzingaze, environ de Vila Pery, Bas Sand- gaze; Calamo (Poisson 1934, 1940). The specimens we examined bear the following labels: "S. Africa, G. E. Hutchinson call. B. M. 1928-395." "S. Rhodesia" 1 apterous male (British Mus.). "Zululand, Nagapa Res. Lab. 23-IV-1922 H. H. Curson" 1 apterous female (British Mus.). Tenagogonus divergens n. sp. (PI. 1, fig. 4; PI. 7, fig. 5) Size: Apterous male holotype 7.56 mm. long; width across the mesoacetabula 2.646 mm.; width across the head 1.47 mm. Color: Light brown. Head with two reddish brown spots above the clypeus and a curved line near each eye; pronotum with median longitudinal reddish brown line and another near its margin; dark brown band on propleuron behind the eye and continued as two 13—2656 386 The University Science Bulletin bands to mesopleuron, the upper one ending behind mesoace- tabuhim; two short bands on mesoacetabulum; two longer ones on metaacetabuliim; under side of connexivum dark; remainder of venter pale except for short black spot at anterior end of mesoace- tabular cleft. Structural characteristics: Proportional length of antennal seg- ments: 1st: 2nd: 3rd: 4th:: 96: 86: 126: 140. Total length of antenna 9.4 mm. Beak slender. Front leg segments: femur: tibia: 1st tarsal: 2nd tarsal:: 150: 127: 16: 26. Middle femur 390 spaces or 8.19 mm. long, remaining segments and hind legs missing. Abdominal segments rather short, the first two ventral abdominal segments together shorter than metasternum. Hind coxa surpassing rear margin of second ventral abdominal segment. Last ventral abdominal segment of male slightly shorter than the two preceding segments. Conspicuous diverging protuberances arising from the surface of the last ventral abdominal segment just beneath con- nexivum. Location of Type: Described from a specimen bearing the label "Brook near Bawomataluwo, Sumatra, 9-12-31. v. d. Meer Mohr." (K. U. col.) Comparative notes: This species has the protuberances of the last ventral abdominal segment of the male arising just beneath the connexivum instead of farther ventral as in T. albovittotus Stal. In this regard it is more like T. zamhezinus Poisson. Tenagogonus kampaspe (Kirkaldy) (PI. 1, fig. 5; PI. 7, fig. 4) 1900 Gerris Kampaspe Kirkaldy, G. W. Ann. Mus. Civ. Stor. Nat. Geneva, ser. 2, 20: 804. 1933 Lininometra kampaspe Lundblad, O. Archiv fiir Hydrobiologie Suppl. 12, "Tropische Binnengewasser 4," p. 371 (quotes New Guinea). The original description: "Gerris kampaspe sp. n. Belongs to subg. Limnometra Mayr. Size and general structure of minuta Mayr, but readily distinguished by its much greater breadth. Fulvo- cinereous; a diamond-shaped mark on the head, a median longi- tudinal line on the pronotum, and legs blackish brown. Venter pale cinereous. Elytra dark fulvo-cinereous with blackish ner- vures. Length 5^2-7/2 mm. New Guinea: Kelesi (Loria), Rigo ( Loria ) ." Fortunately, the type series, two males and two females, are in the Kirkaldy collection at the University of Kansas Entomological Museum, otherwise we would never have been able to recognize Tenagogonus-Limnometra Complex 387 it. No specimens are now entire but the following notes and the illustrations should enable one to identify the species. Size: Winged male: 5.33 mm. long; width across humeri 1.43 mm.; width across mesoacetabula 1.7 mm.; width across head 1.16 mm.; wingless male: 4.2 mm. long; width across mesoacetabula 1.64 mm.; width across head 1.09 mm. Wingless female: 6.3 mm. long; width across mesoacetabula 1.64 mm.; width across head 1.43 mm. Color: General color rather light yellowish-brown; the reddish black figures as shown in the illustration; dark brown to black band on propleuron behind the eyes; the two bands on mesopleuron more or less fused; a brown spot on lateral margin of mesosternum may be present. Venter pale. Structural characteristics: One winged male with two antenna] segments: 1st: 2nd:: 57: 50. One wingless female with three antennal segments: 1st: 2nd: 3rd:: 80: 66: 76. One specimen, a male with a beak reaching one third the distance on mesosternum. Front femur of male considerably stouter than middle femur and strongly curved, other legs incomplete. Relative length of leg segments of a winged male: First tarsal Second tarsal Femur Tibia segment segment Front leg 85 78 9 11 Middle leg 190 160 ? ? Hind leg 190 100 21 12 Middle femur not as long as body in either sex. Connexivum rather broad; no connexival spines. Dorsal abdominal segments of male short; last one as long as two preceding. Hind coxa of male reaching penultimate abdominal segment; of female reaching rear margin of second abdominal segment. Abdominal venter of male a little more than half as long as mesosternum, its last ab- dominal segment plus genital segments longer than the rest of abdomen; the venter of last abdominal segment of male with cau- dally directed lateral protuberances. Winged form: Shape of pronotum and venation of hemelytra as shown in illustration. Location of types: Described from two males, one winged, and two wingless females labeled as follows: The winged male: "N. Guinea Mer. Kelesi, Nov. Die. 1890. L. Loria." "Gerris Kampaspe Kirk. Type ^ ." Wingless male: "N. Guinea Mer. Rigo, Luglio 1889. L. Loria." One wingless female: "N. Guinea. Dilo, Loria VI, VII 90." One wingless female: "N. 388 The University Science Bulletin Guinea Mer. Rigo, L. Luglio 1889." This one bearing the same label as the type should be considered the allotype. These are in the Kirkaldy collection of The University of Kansas Snow Ento- mological Museum. Comparative notes: This species is related to the T. albovittatus Stal group. Data on distribution: Known so far only from these types from New Guinea. Tenagogonus robtistus sp. nov. (PI. 1, fig. 6; PI. 8, fig. 9; PI. 16, fig. 30; text fig. 1) Size: Apterous male: 7.56 mm. long; width across the mesoace- tabula 3.15 mm.; width across the head 1.64 mm. Apterous female: 7.35 mm. long; width across the mesoacetabula 3.15 mm.; width across the head 1.57 mm. Color: Yellowish to reddish brown. Head with a short dark median band on its base that splits into two diverging bands that suddenly converge to unite before the base of clypeus, an undulate band near each eye; pronotum with usual median longitudinal line; lateral dark lines near the margin that are broad on the anterior lobe. Abdominal dorsum more or less mottled with darker brown; black band on propleuron behind the eyes and continued as two broad bands separated by a narrow pale line; the upper band broader and continuing onto metaacetabula, the lower one turning down at its end but not connecting with the dark band on meso- acetabula. Venter nearly white; underside of connexivum em- browned. Structural characteristics: Proportional length of the antennal segments of male: 1st: 2nd: 3rd: 4th:: 100: 84: 107: 116. Total length of antenna 8.55 mm. Beak moderately slender, covering less than one third of mesosternum. Front femur slightly, if any shorter, than middle femur. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 157 137 17 30 Middle leg 385 270 114 30 Hind leg 330 200 40 35 Posterior lobe of pronotum in apterous forms rather narrow and exposing mesonotum around its tip. Connexivum rather broad, no connexival spines. Dorsal abdominal segments of male fairly short, last one as long as the two preceding. Hind coxae of male reaching rear margin of third ventral abdominal segment; of female sur- Tenagogonus-Limnometra Complex 389 passing the middle of second ventral abdominal segment. Last ventral abdominal segment of female as long as the two preceding segments. In the male the last ventral abdominal segment as long as the two preceding segments and together with genital segments longer than the rest of abdomen. First genital segment of male medially depressed ventrally, with a large V-shaped notch on its rear margin. Male genital capsule (9th segment) with lateral hair tufts. Location of types: Described from 3 males 3 females from Sumba, five of them bear the label "Prai Jawang, O. Sumba, Rende Wai, 14. 6. 1949, Dr. Biihler, Dr. Stutter." and one with the label "Pogobina 17. 9. 1949, Dr. Biihler, Dr. Stutter." These and one bearing the label "Ost-Java Lavang 1.50(7 Leg. M. E. Walsh 136" were sent to us for determination by the Museum at Basel, Switzerland. The holotype, allotype and two paratypes will be found there and two paratypes in the University of Kansas collection. Through kindness of Mr. A. M. R. Wegner of the Zoological Museum at Bogor, Indonesia, we have recently been loaned an apterous male and female, two winged males and one winged female bearing the label "Coll. F. C. Drescher, Java, Preanger N. O. I. Bandoeng, 750 m. 1937." These were labelled Limnometra brevis Lundblad but not by Lundblad, they are actually our T. rohustus and are made paratypes. Four of these have been returned to the Museum at Bogor. Comparative notes: This species is most closely related to Tena- gogonus brevis (Lundblad), from which it is readily distinguished by the shape of the first genital segment of the male, and by its parameres which are quadrate in shape not pointed as in T. brevis (Lundblad). Data on distribution: Besides the Sumba and Ost-Java records above we have two females that bear the label "Mt. Apo, Mindanao, Philippine Islds. C. F. Clagg, Baroruig Riv. 7.000 ft. Nov. 8." These came to us for determination from the Museum of Comparative Zoology at Harvard College. In size, shape and color pattern they are almost identical with this O. sumba species. However, since they came from the Philippines and have the posterior lobe of the pronotum somewhat broadened at the humeri and have moder- ately long and spine like projections on the connexivum that are entirely lacking in the Sumba species we believe they represent a good subspecies and give them the name Tenagogonus robustus claggi n. subsp. Males, when they are found, may indicate a good species (see pi. 2, fig. 7; pi. 8, fig. 10). One specimen is in M. C. Z. and the other is in the Snow Entomological Collection, University of Kansas. 390 The University Science Bulletin Tenagogonus brevis (Lundblad) (PI. 2,fig. 8;P1. 2, fig. 6) 1933 Limnometra brevis Lundblad, Archiv fur Hydrobiologie 1933, Suppl. 12. Tropische Binnengewiisser 4, pp. 388-392, Taf. 10, fig. 123. Thanks to Doctor Lundblad we have seen the winged allotype of this species and have had it redrawn to conform in magnification with the other species reported in this paper and also have had Text Figure 1. (A) Abdominal venter and first genital segment of male of Limnometra brevis as drawn by Lundblad in his figure 123 D on page 391. (B) Genital capsule of L. brevis Lundblad as shown in above text figure 123 E showing location of paramere or "genital griff el" at p. (C) Paramere of the above enlarged. (D) Paramere of Ten«gogonu5 rofou^fws n. sp., enlarged. (E) Gen- ital capsule of T. robustus from left side showing paramere at p. Tenagogonus-Limnometra Complex 391 his drawing of the apterous female enlarged. His description and illustrations are splendid and adequate. In the allotype the last antennal segment appears nearly white but the middle and hind femora are uniform in color without the distal ends paler, so usual in many species. While this is a small species it is relatively broad. The width of the head being only 1.53 mm. but width across the mesoacetabula more than 2.73 mm. The second antennal segment is shorter than the width across head. The allotype bears the label "Java Thienemann" and is in the museum at Stockholm, Sweden. The apterous paratype female has the hind coxae slightly surpassing second abdominal segment. The male from Sumba Island, of which Dr. Lundblad gave drawings, must be the holotype and belongs to the Bogor Museum. We have not seen this but the drawing of the first genital segment of the male could not have been made from our T. rohustus described above. Compare Lundblad's fig. 123, D ( p. 391 ) and our fig. 9a on plate 8. Also compare the male genital hook or paramere (fig. 123, F, p. 391 ) and the paramere of T. rohustus on text figure 1 in this work. Tenagogonus pravipes Bergroth 1915 Tenagogonus pravipes Bergroth, E. Zool. Med. Rijks Mus. Nat. Hist., Leiden, Deel 1, 1 : 121-122 (described from Java and says that Limnometra Mayr is a synonym of Tenagogonus Stal 1855. 1933 Limnometra pravipes Lundblad O. Archiv fiir Hydrobiologie 1933, Suppl. 12, Tropische Binnengewiisser 4, p. 371 (quotes Java). The types of this species have not been located by us. Doctor Blote says they are not at the Leiden Museum and Doctor Lindberg cannot find them in Helsingfors. We have found no specimens from Java that fit Bergroth's description which follows: "Tenagogonus pravipes n. sp. "Opacus, supra luride ochraceus, subtus pallide testaceus, linea transversa subbasali verticis, vittis duabus e petiolo comnumi ab ilia linea emisso ortis, antrorsum levissime divergentibus, antice conjunctim basin clypei subattingenti- bus, vitta verticis paullo intra oculos, vitta laterali ab oculo ad basin antennae ducta, vitta curvata sublaterali pronoti antice cum vitta alterius lateris anguste cohaerente, linea media percurrente pronoti et processus ejus, linea angusta sub- laterali processus pronoti paullo ante apiccm ejus abrupta, vitta laterali pro- pleurae, vitta lata partis dorsalis meso- et metapleurae, vitta laterali postice abbreviata mesopleurae, vittula ante acetabula media, margine apicali superiore horum, vittula acetabulorum posteriorum, vitta sublaterali connexivi et ventris in maculus subdissoluta articuloque ultimo nitido rostri nigris, metanoto, seg- mentis duobus primis dorsi abdominis suturisque dorsalibus segmentorum ceterorum hujus infuscatis, antennis et pedibus (coxis exceptis) pallide fuscis. Caput latitudine sua paullo longius, supra pilis paucis longis erectis praeditum. 392 The University Science Bulletin fronte sat fortiter declivi, oculis antrorsum levissime convergentibus, antennis corpore longioribus, articulis primo, tertio quartoque subaeque longis, primo basin versus nonnihil curvato, secundo adjacentibus breviore, rostro medium mesostcmi baud attingente. Metastemum segmentis tribus primis ventris unitis fere aeque longum, orificio prope marginem posticum posito. Abdomen ( $ ) parte sua dimidia dorsali coxas posticas superans, segmento ultimo dorsali duo- bus praccedentibus unitis parum longiore, apice truncate, segmento ultimo connexivi ultra ultimum dorsale breviter vix acute retrorsum prominulo, ventre quam dimidio mesostemi paullo longiore, parte sua pone segmentum quartum sita coxas posticas superante, segmento sexto medio duobus praecedentibus con- junctis subaeque longo, segmento genitali dorsali secundo maris perbrevi, nigri- cante, genitali ventrali secundo dorsale superante et quam hoc latiore. Pedum anticorum femora longiuscula, parte plus quam quarta basali nonnihil incrassata, ad apicem partis incrassatae subconstricta et subangulariter fracta, deinde usque ad apicem recta, linearia; tibiae femoribus paullo breviores; articulus secundus tarsonmi primo evidenter longior. Coxae mediae marginem posticum meta- sterni et basin acetabulonmi posticorum attingentes. Femora posteriora corpori subaequa longa. Long. 9 6.8 mm. "Forma aptera: Pronotum a processu suo impressione leni transverse sepa- ratum, processu metanotum attingente, quam pronoto proprio paullo augustiore, usque ad trientem apicalem subparallelo, deinde apicem late rotundatum versus levissime angustato. Magelang (J.). "Extremely similar in colour to the quite inadequately described T. anadyo- mene Kirk., but it is much smaller and comparison with Ceylonese specimens of that species reveals the foUovi'ing important structural difiFerences: The abdomen in pravipes is much shorter and not longitudinally ridged beneath in the middle, the apical angles of its last segment are much less acute, the fore legs and the second male genital segment are quite differently constructed, and the length-relations between the metastemum and the first ventral seg- ments and between the middle coxae and the adjacent parts are different. The macropterous form of pravipes is unknown. In T. anadtjomene, of which almost only the colour-markings have been described by Kirkaldy and Distant, the venter is as long as the meso- and metastemum together and longitudi- nally carinated in the middle, the apical angles of its last segment are very acutely produced, the second male dorsal genital segment is produced beyond the corresponding ventral segment, the fore femora are almost straight and not incrassated at the base, the fore tibiae are as long as the femora, the two joints of the fore tarsi are of equal length, the metastemum is as long as the two first ventral segments together, the middle coxae reach the middle of the meta- stemum but not the base of the hind acetabula, and the hind coxae barely reach the middle of the second ventral segment. "In a paper sent for publication some months ago but not yet printed I have maintained Limnogonus Stal as a genus distinct from Limnometra Mayr, but after the study of further materials I find that Limnogonus cannot be considered even subgenerically distinct, as the transitions in the mutual length of the two joints of the fore tarsi and in the shape of the apical angles of the last abdominal segment are too numerous. Limnometra is a synonym of Tenagogonus Stal, the type of which is T. albovittatus Stal (1855) from Natal, a species totally for- Tenagogonus-Limnometra. Complex 393 gotten by Stal himself in his later writings and also omitted in the Catalogue of Lethierry and Severin. "N. B. — The segment following immediately after the metanotum has by me previously (Ent. Monthly Mag. 1902, p. 259) been called the metaphragma, but I now with Breddin think it more natural to regard it as the first abdom- inal segment." Tenagogontis pravipes bergrothi subsp. nov. (PI. 2, fig. 10; PI. 8, fig. 8) Size: Winged male 7.56 mm. long; width across the mesoacetab- ula 2.42 mm.; width across head 1.36 mm. Small wingless male: 5.04 mm. long; width across mesoacetabula 1.76 mm.; width across head 1.16 mm. Winged female 6.93 mm. long; width across meso- acetabula 2.35 mm.; width across head 1.3 mm.; wingless female: 6.6 mm. long; width across mesoacetabula 2.52 mm.; width across head 1.28 mm. Color: General color light yellowish-brown. Head and pronotum yellowish brown with the dark reddish-brown to black figures as shown in the illustrations. Black band on propleuron behind eye, continued as two broad bands on mesopleuron, the upper one broader, ending on the metaacetabula and the lower one turning down and ending before reaching the dark band on mesoacetabula. Venter light, nearly white except for a dark spot on metasternum and underside of connexicum embrowned. Structural characteristics: Proportional length of antennal seg- ments of winged male: 1st: 2nd: 3rd: 4th:: 85: 83: 106: 108, total length of antenna 8.02 mm. Beak moderately slender, cov- ering less than one third of mesosternum. Front femur slightly, if any, stouter than middle femur. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 130 114 13 22 Middle leg 300 235 108 27 Hind leg 270 145 30 25 In winged form, shape of pronotum and venation of hemelytra as shown in the illustration. In the apterous form the shape of pronotum is shown for a female and is relatively broader than in male. Conncxivum moderately broad, male without connexival spines and female has the outer angle of connexivum produced al- most into a spine. Dorsal abdominal segments of the male fairly short, last one not as long as the two preceding ones. Hind coxae 394 The University Science Bulletin of male surpassing rear margin of second abdominal segment; of the female reaching or surpassing the middle of second ventral abdominal segment. Last ventral abdominal segment of female shorter than the two preceding segments. In the male the last ventral abdominal segment shorter tlian the two preceding seg- ments and together with genital segments almost as long as the rest of the abdomen. The venter of first genital segment of the male with a median longitudinally depressed line. Male genital capsule without lateral hair tufts. Location of tijpes: Described from fifteen males (five winged) and twenty females (one winged) labeled "Observatory Garden, Manila, P. I." and two females labeled "Mainit, Surigao Mindanao, P. I. Ill, 21, 31, A. C. Duyag". The last two belong to Cahfomia Academy of Sciences. Others are from Bueno coll. in the Univer- sity of Kansas collection. Holotype and allotype wingless. Holo- morphotype and allomorphotype are v^nged. Co77iparative notes: As in Tenagogontis pravipes Bergroth the males have the basal third of the front femur somewhat thicker and then curved beyond. This is not true for T. robustus sp. nov. or for T. brevis (Lundblad) which are closely related species. Data on distribution: Known only from the type series above. We have a species from the Philippines that must be very close to Tenagogontis pravipes Bergroth. In some respects it does not agree with Bergroth's description and since he did not mention the striking impressed line on the venter of the first genital segment of the male we are describing it as a subspecies. Tenagogonus fijiensis sp. nov. (PI. 2, fig. 9; PI. 8, fig. 7) Size: Apterous male: 5.25 mm. long; width across mesoacetabula 2.41 mm.; width across head 1.51 mm. Apterous female: 6.72 mm, long; width across mesoacetabula 2.73 mm.; width across head 1.74 mm. Color: Head reddish brown with two longitudinal black lines meeting at both ends, a black band along each eye; pronotum red- dish brown, with a median longitudinal black line, broader on an- terior lobe, lateral black lines near the margin that are very broad on anterior lobe; black band on propleuron behind eye and con- tinued as two bands separated by a pale stripe; the upper band ending on the metaacetabula and the lower one on the mesoace- Tenagogonus-Limnometra Complex 395 tabula. Underside of male connexivum embrowned, venter other- wise pale. Structural characteristics: Proportional length of antennal seg- ments of male: 1st: 2nd: 3rd: 4th:: 92: 64: 62: 70. Total length of antenna 6.05 mm. Beak moderately stout and short for this genus. Front legs stout, femur much thicker than middle femur. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 125 115 14 20 Middle leg 280 262 74 23 Hind leg 290 170 27 17 Pronotum of the apterous form narrowed at caudal end. Con- nexivum of female, erect on basal segments and reflexed on distal segments, with the caudal angles produced and overlapping. Con- nexivum of male semierect, no connexival spines, its dorsal abdom- inal segments short except the last one which is as long as the two preceding segments and as long as the first genital segment. Ven- trally, hind coxae of male surpass the rear margin of the fourth abdominal segment; in female they surpass the rear margin of the second. Ventral abdominal segments of male short except last one, which together with the genital segments is longer than rest of abdomen. The ventral surface of first genital segment normal. Last ventral abdominal segment of female as long as three preceding segments, tubular, ending in a mid-ventral point. Location of types: Described from holotype, allotype and two male paratypes labeled "Fiji Islands, July 1934 R. W. Paine" "Ta- veuni; Ura. Stagnant pool in rocky stream bed of forest 2500 ft." (K.U. coll.) Comparative notes: The shape of the pronotum in these apterous insects and the tubular last ventral abdominal segment of the fe- male are characteristic. Tenagogonus kuiterti sp. nov. (PI. 2, fig. 11; PI. 9, fig. 11) Size: Winged male: 10.37 mm. long; width across humeri 2.1 mm.; width across mesoacetabula 2.83 mm.; width across head 1.64 mm. Apterous male: 8.74 mm. long; width across mesoacetabula 2.77 mm.; width across head 1.55 mm.; apterous female: 8 mm. long; width across mesoacetabula 2.63 mm.; width across head 1.43 396 The University Science Bulletin mm. The size shows considerable variation. Winged males may be only 7.98 mm. long and apterous males only 6.3 mm. Color: General color light brown. Head and pronotum light brown with the dark reddish brown to black figures as shown in the illustration. Note the pale spot on inner margin of hemelytron in winged forms. Dark brown to black band on propleuron behind eye continued as two bands on mesopleuron, the lower one broader and often broadened in one or two places to join the upper band, usually ending free but occasionally joining the upper of two spots in the mesoacetabula. Venter pale except for a black spot at an- terior end of the mesoacetabular cleft and underside of the con- nexivum embrowned. Stnictural characteristics: Proportional length of antennal seg- ments: 1st: 2nd: 3rd: 4th:: 90: 83: 108: 100. Total length of antenna in this wingless male 8 mm. this male being 6.55 mm. long. Beak moderately slender, not covering one third of mesosternum. Front femur of male slightly stouter than middle femur. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 118 100 13 20 Middle leg 290 215 95 30 Hind leg 290 125 24 24 including claws Winged form: shape of pronotum and venation of hemelytron as shown in the illustrations. In apterous male the shape of pro- notum is shown and also in a slightly brachypterous male, the wing pads too short to be exposed. The male with very short connexival spines, especially in apterous forms. Female with broad con- nexivum and short spines. Last dorsal abdominal segment of male not as long as the two preceding segments but longer than the dorsum of the first genital segment. In female last dorsal abdom- inal segment as long as the next preceding segment. Ventrally the male abdomen nearly as long as mesosternum. First genital seg- ment of male with lateral transverse depression, its venter normal. Location of types: Described from eleven males (three winged) and three females from "Mohnyin, Burma X. 27. 1944 L. C. Kuitert" and "Burma, 1944. L. C. Kuitert" in the University of Kansas col- lection. Comparative notes: This species we are placing in Tenagogonus Stal because it has so many characters of that genus. It is the only Tenagogonus-Limnometra Complex 397 species with connexival spines but they are very short and variable, almost absent in some wingless males. Data on distribution: Besides the type series we have 1 male and 4 females bearing the label "Shingbwiyang, Burma XI-15-1944 Capt. L. C. Kuitert". These are all winged specimens and their connexival spines are evenly developed. Genus Limnometra Mayr 1865 Mayr, Gustav L. Verh. Zool-bot. Ges. 15: 443. (New genus for: L. femorata Mayr and six other species.) 1865 Mayr, Gustav L. Reise der osterreichischen Fregatte Novara Expedition urn die Erde. Zool. Theil 2 ( 1 ) : 174 Wien. ( Key to species. ) 1909 Kirkaldy G. W. and Torre Bueno, J. R. de la. Proceedings of Ent. Soc. Washington 10: 209. (Given as synonym of Tenagogonus Stal. ) 1915 Bergroth, E. Zool. Med. Riiks Mus. Nat. Hist., Leiden 1 (2): 121-123. (Gives Limnometra Mayr as synonym of Tenagogonus Stal 1853.) 1925 China, W. E. Bull. Brooklyn Ent. Soc. 20: 218. (Revives Limnometra Mayr for L. gigas.) 1933 Lundblad, O. Archiv fiir Hydrobiologie. Suppl. 12. "Tropische Bin- nengewasser 4," p. 371. Medium to large elongate gerrids with moderately long antennae, middle and hind legs. Male abdomen longer than the mesosternum. Last male ventral abdominal segment plus the genital segments shorter than the preceding four ventral abdominal segments. First tarsal segment of the front leg usually longer than the second. Abdominal spiracles closer to the anterior margin of the segments. Males with connexival spines or witli connexivum produced into a triangular flattened plate. Hind coxae not covering second ventral abdominal segment. Metasternum equal to or shorter than first two ventral abdominal segments. Last ventral abdominal segment of male equal to or shorter than preceding segment. In Limnometra the males are often larger than the females. Limnometra minuta Mayr (PI. 2, fig. 13; PI. 10, fig. 14) 1865 Limnometra minuta Mayr, Gustav L. Verh. Zool. -hot. Vereins, Wien, 75: 444. 1865 Limnometra minuta Mayr, Gustav L. Hemiptera in Novara Expedition, Zoolog. Theil, Wien 2(1): 175, 176-177. figs. 54a, 54b. (Gives addi- tional description and figures. ) 1904 Gerris minuta Distant W. L. The Fauna of British India Rhynchota 2: 181. (Quotes Nicobar Islands.) 1933 Limnometra minuta Lundblad, O. Archiv fiir Hydrobiologie, Suppl. 12. "Tropische Binnengewiisser 4", p. 371, 373. The type of this species is a winged male bearing the label "No- vara Exp. Sambelong". It fits perfectly Mayr's two descriptions and notes. While mostly a color description he gives the length (8.6 mm.), thus smaller than the other species he knew, the nonciliated 398 The University Science Bulletin middle femora and the broadly based triangular connexival spines, which are useful in recognizing this species; yet one hundred years have passed without any one recognizing additional specimens. Surely on some Nicobar Island this species still exists. The follow- ing notes and drawings of the type may be helpful in recovering this much desired little species. Notes on Limnometra minuta Mayr: Size: Type is a winged male, 8.61 mm. long; width across humeri 1.47 mm.; width across mesoacetabula 2.1 mm. Color: Color pattern not striking. Antennae and legs of uniform color with part of tarsi darker. Antennae darker than the legs. Head has a nearly black tylus; just back and laterad of the tylus stripe are light brown lines that converge and unite into a median .stripe at the anterior level of the eye emarginations; there is an- other brown line near the inner margin of eye that unites in front with the mediolateral band, thus producing an M-shaped figure on head; a dark band between the base of antenna and eye. Pronotum has a median longitudinal, nearly black Hne and another separated from the lateral and caudal margins by a pale line; on the anterior lobe, on either side, between the dark lines is a light brown streak that fades out before the level of humeri; a brown streak on side of prothorax behind eye, this is continued as a dorsolateral streak on mesothorax. There is a dark-brown elongate spot on anterior acetabula, two curved ones on middle acetabula and one on the posterior acetabula. There is a faint brown lateral streak on meso- thorax and a black spot at anterior end of mesoacetabular cleft. Venter pale. Last segment of the beak shining, black. Hemelytra brown. Structural characteristics: Antennae slender. Proportional length of antennal segments: 1st: 2nd: 3rd: 4th:: 93: 73: 98: 133. The total length of this antenna is 8.32 mm. Front femur somewhat curved and about twice as thick as tibia. The segmental formula: Femur: tibia: 1st tarsal: 2nd tarsal:: 133: 115: 17: 18. Middle femur measures 6.93 mm. long and has neither ciha nor pegs. Other segments are gone. Metasternum about equal to first two abdom- inal segments. Abdominal venter with faint median carina. Male genital segments about as long as the last two ventral abdominal segments. The abdominal sides parallel as seen from beneath and the connexival spines are very broad at base, triangular in shape and not quite reaching the tip of the body. We have seen no other species with connexival spines like this. Tenagogonus-Limnometra Complex 399 Limno7netra lipovshji sp. nov. (PI. 3, fig. 20; PI. 9, fig. 12) Size: Winged male: 13.44 mm. long; width across humeri 2.47 mm.; width across mesoacetabula 3.78 mm.; width across head 2.14 mm. Winged female: 10.9 mm. long; width across humeri 1.93 mm.; width across mesoacetabula 3 mm.; width across head 1.84 mm. Wingless forms probably smaller, since one wingless male is only 7.35 mm. long. Color: General color pattern quite distinctive, yellowish brown and black. The black figures as shown in the illustration. Dark brown to black band on propleuron behind eye as usual but the upper band on mesopleuron almost obliterated, only the lower one which is dark brown is present and it fades out before reaching mesoacetabula. Males have a longitudinal black line on basal third of middle femur while females have a short black spot near base as do L. ciliata Mayr. Venter is pale. Structural characteristics: No male has an entire antenna. One with three segments gives the following proportional lengths: 1st: 2nd: 3rd: 4th:: 150: 145: 130: ?. A female antenna: 1st: 2nd: 3rd: 4th:: 120: 100: 90: 100. Beak moderately slender but not covering one third of mesosternum. Front femur of male not stouter than middle femur. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 255 230 35 33 Middle leg 710 700 191.9* 34.9* Hind leg 760 432 65 25 The middle femur and tibia of male have, ventrally, a wide brush of scattered cilia. Winged form: Shape of pronotum and venation of hemelytra as shown in illustration. Male with very short con- nexival spines that are probably more reduced in apterous forms; last abdominal tergite longer than preceding one but shorter than the first genital segment; ventrally the male abdomen longer than the mesosternum; last ventral abdominal segment slightly shorter than preceding one; venter of first genital with stout lateral pro- tuberances arising from its venter. Last venter abdominal seg- ments of female as shown in the drawing. Location of types: Described from three winged males and four winged females labeled "Guadalcanal 1944. L. J. Lipovsky" and * Calculated from smaller male. 400 The University Science Bulletin one female "Guadalcanal 1945. P. H. Eschmeier". All of these are in the University of Kansas collection. Comparative notes: This species, like L. kuiterti has very short connexival spines. The protuberances on the venter of the first genital segment of the male distinguish this species. Data on distribution: Known only from Guadalcanal. Limnometra octopunctata Hungerford (PI. 4, fig. 22; PI. 10, fig. 13) 1955 Limnometra octopunctatus Hungerford, H. B. Journal Kansas Ent. Soc, 28: 67-68. Size: Apterous male: 9.24 mm. long; width across the mesoace- tabula, 2.625 mm. Apterous female: 11.25 mm. long; width across the mesoacetabula, 3.25 mm. Color: Reddish brown above, lighter beneath; head with a broad median brown band that is wider in front and bordered by lighter narrower lines; pronotum with a median black band that is broad in front and tapering on the anterior lobe to a slender line that continues on the posterior lobe to its apex; the posterior lobe with a slender lateral submarginal black line to the shoulders where it broadens and is bordered by a pale line around the apical lobe; behind the pronotum is a median brown to black band that con- tinues to abdominal tip; on this dark band from the third abdominal segment to the eighth there is a median series of gray streaks; on each acetabulum, a pair of black spots; between these spots a small creamy white spot on the mesoacetabulum and on metaacetabulum a large triangular creamy white spot. On the venter eight large nearly round black spots that characterize the species and suggest the name. These spots are located as follows: One black spot in front of and one on the venter beneath each mesoacetabulum, one beneath each metaacetabulum and another on either side of the last abdominal segment. On each side of the first genital segment of the male there is another black spot half hidden beneath the last abdominal segment. If such spots are also on the female they are entirely hidden. Structural characteristics: The male has lost the last two seg- ments of its antennae. The length of the first and second segments are respectively 2.6 mm. and 2.73 mm. so the second is longer than the first. The female also is imperfect and only one antenna is represented by three segments that measure: 1st, 2.52 mm.; 2nd, 2.14 mm.; 3rd, 2.18 mm. The beak is moderately long (3.21 mm.). Tenagogonus-Limnometra Complex 401 elongate third segment reaching back some distance on meso- sternum. The relative lengths of the segments of the beak are 1st: 2nd: 3rd: 4th:: 30: 13: 90: 21 ( (? ). Fourth segment black and shining and its mean diameter about half that of the third which is slender and its diameter about half of the first segment. Front legs of both sexes have the basal two thirds of femora thickened. The relative lengths of the segments, femur: tibia: 1st tarsal: 2nd tarsal:: 167: 137: 23: 23 ( ^ ) and 180: 155: 30: 30 ( $ ). Mid- dle legs long, femur being 9.24 mm. long in male and 10.3 mm. in female; the relative lengths of the segments, femur: tibia: tarsus:: 440: 410: 120 ( ^T ) and 490: 466: ? ( $ ); hind legs shorter than middle legs; the hind femur of the male 9.24 mm. long and that of female 10.25 mm.; the relative lengths of the male segments being, femur: tibia: tarsus:: 440: 217: ?; for the female 488: 290: 70. In the male connexival spines reaching about three fourths the length of first genital and in the female surpassing the first genital and extending half the length of the last genital. Ventrally the rear margin of the first genital segment of the male has laterally on each side a projection. Location of types: Described from holotype male and allotype female and one male paratype bearing the label "Dr. B. Hagen, Tandjong Morawa. Serdang. (N.O.Sumatra.)" These are in the Francis Huntington Snow Museum, University of Kansas. Comparative notes: The eight distinct black spots on the venter will identify this species. The lateral angular projections on the rear margin of the first genital segment of the male appear to re- late it to L. ciliatiis, Mayr. It also has the ciliate row on the rear margin of the middle femur but not as conspicuous as in L. ciliata and is a smaller species. Limnornetra fluviorum (Fabricius) (PI. 3, fig. 19; Tl. 10, fig. 15) 1798 Gerris fluviorum Fabricius, J. C. Ent. Syst. Suppl. 543. 2. (Described from Tranquebar in Tanjore Dist. India.) 1803 Hydrometra fluviorum Fabricius, J. C. Syst. Rhyn. p. 257. 1865 Limnornetra fluviorum Mayr, Gustav L. Hemiptera in Novara Expedition Zool. Theil, Wien. 2(1): 175. 1868 Limnometra fluviorum Stal, C. Hemiptera Fabriciana p. 132. 1904 Gerris fluviorum Distant, W. L. The Fauna of British India. Rhynchota 2: 177. (Records: Pondichcrry [coll. sign]). Bombay; Bor. Ghat (Dixon). Trivandrum, ( Brit. Mus. ) also records Java and Bourbon Isl. 1914 Limnometra fluviorum Ilorvath, G. Ann. Mus. Hung. 2: 660. 1933 Limnometra fluviorum Lundblad, O. Archiv. fiir Hydrobiologie Suppl. 12. "Tropische Binnengewasser 4" p. 371. (Gives Vorderindien, Ceylon, Java. ? Reunion. ) We believe the following to refer also to this species: 402 The University Science Bulletin 1840 Gerris armata Spinola, Maximilien, Essai sur les Insectes Hemipteres, p. 65. This is described from Bombay and is charac- terized by the spine on the intermediate coxa in both sexes as is L. fliivioriim. 1904 Gerris armata Distant, W. L. The Fauna of British India Rhynchota 2: 180-181. This species is from 10 to 13 mm. long, rather heavily banded with black. The antennae and legs are not annulated. The front tarsal segments with the second segment longer. The mesocoxal spine will identify this species. Location of type: Daldorff collection at the Zoological Museum, Copenhagen, Denmark. Data on distribution: This was described from Tranquebar, India and Lundblad records Ceylon and Java. We have before us the following: "India, Tanjore Dist. P. S. Nathan". 1 male, 1 female. (K. U.); "India, Tanjore Dist. Nedungadu P. Susai Nathan". 2 males, 2 females. ( Basel Mus. ) ; "India, Coimbatore Dist. P. Susai Nathan". 2 females. (Basel Mus.); "South India. Coimbatore 6. VI 1950, P. S. Nathan". 1 male, 1 female. ( K. U. coll. ) ; "Coimbatore, India". 1 male. ( Calif. Acad. Sci.); "Coromandel, M. Maindron". 3 males, 3 females. (Kirk. coll. K. U.); "Chikkaballapura, S. India T. V. Campbell". 1 female (K. U.); "Kurumbagaram, Karikal Terr. S. India 16-IV-47 coll. P. S. Nathan", (coll. of H. S. Wallace No. 1706.) 1 male. Limnometra anadyomene (Kirkaldy) (PI. 5, fig. 30; PI. 11, fig. 16) 1901 Gerris (Limnometra) anadyomene Kirkaldy, G. W. Entomologist 34: 117. ( Desc. from Punduloya, Ceylon colls. E. E. Green and Kirkaldy. ) 1904 Gerris anadyomene Distant, W. L. The Fauna of British India. Rhyn- chota 2: 177-178. (Records Ceylon; Punduloya [Green]— PhiUppine Is- lands [Simon]). 1915 Tenagogonus anadyomene Bergroth, E. Zool. Med. Rijks Mus, Nat. Hist. Leiden, Deel 1. 2: 122. 1933 Limnometra anadyomene Lundblad, O. Archiv fiir Hydrobiologie Suppl. 12. "Tropische Binnengewiisser 4" p. 371. (Gives Ceylon, Hinterindien, Philippinen. ) The following is Kirkaldy's brief description: "Gerris anadyomene sp. nov. Belongs to subgenus Limnometra Mayr. Flavous, head with an irregular diamond and two lateral stripes, pronotum with a median and two sublateral lines, irregular markings on ambulacra, etc. black. Elytra dark greyish fulvous; nervures fulvous, apically darker. Spines of seventh segment extending beyond apex of abdo- Tenagogonus-Limnometra Complex 403 men, eyltra extending far beyond apex of abdomen. Male: sev- enth segment deeply, roundly emarginate ventrally. Leng. to apex of elytra 11 mm. Ceylon, Punduloya. Collns. E. E. Green and Kirkaldy. Larger and stouter than minuta (Mayr) to which it is somewhat allied." We have before us Kirkaldy's type series bearing the label "Pun- duloya Ceylon. E. E. Green". There are twenty-five specimens, three of them winged. In only one of the eight males do the con- nexival spines slightly surpass the tip of the abdomen. In all fe- males they are shorter. Therefore his statement is misleading. The following brief notes may be helpful in recognizing this species. Size: Length: 9 mm.-14 mm.; width across mesoacetabula: male 2.73 mm., female 3.61 mm.; width across eyes: male 1.64 mm. female 1.72 mm. Color: General color medium to light reddish brown especially in apterous forms. The lateral pronotal dark band continuous over the humeri in apterous forms, but often obliterated at tip. Antennae and legs uniform in color, not annulated. Each dorsal abdominal segment in apterous forms with two more or less, crescent-shaped dark bands, embracing a paler spot. Underside of connexivum usually dark. Venter light. Structural characteristics: Proportional length of antennal seg- ments of male: 1st: 2nd: 3rd: 4th:: 118: 87: 108: 120; of female: 113: 72: 96: 110. Total length about equal to body length in male, less than body length in female. Front femur not thicker than middle femur, the second tarsal segment longer than the first. Mid- dle femur longer than the body in male, about equal in female. Connexival spines of the male usually surpass the first genital seg- ment but seldom surpass the abdominal tip, and never do so in the female. Location of types: The types mentioned above are in the Kirkaldy collection in the Francis Huntington Snow Entomological Museum, University of Kansas. Comparative notes: In this species the females are strikingly broader and larger than the males. Its nearest relative appears to be L. kuiterti which has shorter connexival spines in the male. Data on distribution: Described from Ceylon and Dr. Lundblad records "Philippines" and "Hinterindien". We have before us the following: Ceylon: Kirkaldy's type series "Pundaloya, Ceylon E. E. Green". 8 males, 17 females (Kirk. coll. K. U.); "Siid Ceylon, Mai 1889 H. 404 The University Science Bulletin Fruhstorfer" 1 male, 1 female (K. U.); "Suduganga River, Matale, Ceylon, R. A. Senior-White" 4 males, 6 females, 4 nymphs. ( Bueno coll. K. U.); "Woodside, Ungalla, Ceylon 3.IX.22. 1 male (Brit. Mus.). S. Shan States: "Heho, 3800", Yawagheve, S. Shan States 7. Ill, 17, Gravely. 2 males. ( Bueno coll. K. U. ) Liinnometra instilaris sp. nov. (PI. 4, fig. 21; PI. 11, fig. 17) Size: Winged male type 11.76 mm. long; width across humeri 1.78 mm.; width across mesoacetabula 2.52 mm.; width across head 1.64 mm. Winged female 9.77 mm. long; width across humeri 1.68 mm.; width across mesoacetabula 2.63 mm.; width across head 1.64 mm. Color: Yellowish brown, the dark markings reddish brown to black, and limited as shown in the illustration. Head markings faint, the holotype having a faint M-shaped figure. Median longitudinal pronotal stripe with pale band on either side, last two or three an- tennal segments white. Distal ends of middle and hind femora not pale. Dark band on propleuron close to lateral pronotal band separated by the cream colored marginal band and continued on mesopleuron. One or two brown spots on all three acetabula. Venter pale. Mesosternum partially pubescent. Structural characteristics: Proportional length of antennal seg- ments of the Holotype from N. O. Sumatra: 1st: 2nd: 3rd: 4th:: 125: 103: 133: 155. Total length of antenna 10.8 mm., one male with body length 11.76 mm. Beak moderately slender, covering a little more than one fourth of mesosternum. Front femur of male slender, not thicker than middle femur. Middle femur of male without a definite fringe of cilia beneath but with many minute spicules. Mesosternum of male almost bare. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 172 145 30 27 Middle leg 460 400 130 27 Hind leg 525 260 50 22 Winged form: Shape of pronotum and venation of hemelytra as shown in the illustration, typical for Limnometra species. Con- nexival spines surpassing genital segments in both sexes. Last ven- tral abdominal segment of male plainly shorter than preceding Tenagogonus-Limnometra CoMPLtX 405 segment, in female slightly longer. The median ventral longitu- dinal keel indistinct on last abdominal segment or two. Location of types: Described from holotype male labeled "Dr. B. Hagen, Tandjong Morawa. Serdang, N. O. Sumatra". Allotype female labeled "Ardjeano, Java" and a paratype male bearing the same label, a female labeled "Chan Yoma, Birmanie" and a male from "Pulo Laout," an island near Borneo. Comparative notes: This species looks like L. borneensis sp. nov. but differs greatly in the shape of the front femur of the male, in the proportions of the leg segments and in lacking the fringe of cilia on the middle femur of the male. Data on distribution: The type series includes Sumatra, Java, Burma? and the island of Pulo Laout. Limnometra palauana Esaki (PI. 3, fig. 16; PI. 12, fig. 18) 1925 Limnometra palauana Esaki, T. Philippine Jour. Sci. 26, 1: 57. 1933 Limnometra palauana Lundblad, O. Archiv fiir Hydrobiologie Suppl. 12. "Tropische Binnengewiisser 4" p. 371. (Records Palau Isl. ) 1937 Limnometra palauana Esaki, T. Tenthredo 1, 3: 362. (Species occurs only on the Island group of Truk but not in Palau Islands on which the specific name was erroneously created. ) Original description: "Body dark yellowish brown. Head dark yellowish brown, with a somewhat indistinct black fascia along the inner margin of eyes; apex black and shining; fasciae on vertex very obscure. Antennae dark brown. Pronotum dark yellowish brown, with the central and marginal lines black; margin and the area along the central median black line yellow. Hemelytra dark yel- lowish brown, veins much darker. Legs dark yellowish brown. Undersurface of body pale yellowish brown, with two longitudinal black fasciae on mesopleura. First and fourth joints of antennae subequal in length, second about two thirds of first, third a little shorter than first. Length of body: male 9 mm.; female 11 mm. Holotype (male), allotopotype (female), and paratopotypes col- lected on Jan. 8, 1915, on Palau Island, Micronesia, by S. Fujita; presented by S. Matsumura; in my collection." This species is nearly allied to Limnometra annuUcornis (Breddin) from Celebes in shape and coloration, but is much smaller, and the markings on the vertex are much more obscure. The color de- scription would apply to a half-dozen species, the size is within the range of several species, the antennae are variable in many cases, and that leaves the description of the antennal segments the only structural character. Fortunately, Dr. Esaki kindly sent us three 406 The University Science Bulletin specimens, two females bearing the label "Caroline Islds, Truk: Toloas 21, 1, 1938 Teiso Esaki" and one male "Caroline Islds, Truk: Toloas-Erin-16 XI 1937 Teiso Esaki", all labeled by him "Limno- metra palauana Esaki". The antennae and legs were mostly free in the bottom of the box but we submit what details we can con- cerning this species. The fully winged female is 8.82 mm. long; the short winged female 8.19 mm. The male is also short winged, also with complete venation as in the female but the abdomen was free in the box. Both sexes have well-developed connexival spines reaching approximately to the caudal tip of the body. The hind coxae of the male reaching slightly beyond the rear margin of the first abdominal segment and barely reaching it in the female. The abdominal spiracles are nearer to the anterior than to the posterior margins of the segments. The abdominal venter is longer than the mesosternum and typically that of a Limnometra. The front femur of the male is thicker than that of the female and is hairy beneath. The first tarsal segment is longer than the second in both sexes. The one unbroken antenna that was free in the box has the antennal formula as follows: 1st: 2nd: 3rd: 4th:: 90: 55: 66: 90, its total length 6.3 mm. The antenna is slender and shorter than the insect but over half its length. The front leg of the male has the following measurements: Femur, 155 spaces long, 22 spaces at greatest diameter; tibia 128 spaces; 1st tarsal 20 spaces; 2nd tarsal 16 spaces. The front leg of a female: Femur 140 spaces long, 14 spaces greatest diameter; tibia 120 spaces; 1st tarsal 30 spaces; 2nd tarsal 17 spaces. The middle femur of the fully winged female is a little larger in diameter than the hind femur and measures 315 spaces or 7.6 mm. long. The middle tibia 280 spaces, 1st tarsal 90 spaces, 2nd tarsal 23 spaces. The hind femur is broken but identifiable and measures 340 spaces long, the tibia 180 spaces, 1st tarsal 37 spaces, 2nd tarsal gone. The rather angular, unevenly thickened front femur with its heavy layer of hair beneath will separate the male of this species from others of similar size and general appearance. Its closest relative appears to be L. borneen- sis, from which it differs in having a broader median black line on the pronotum and the middle and hind femora not being paler at the tips. We also have a series labeled "Los Negros, Admirality Islds, 11-8-45 P. T. Richard 174". 9 males, 2 females (U. S. N. M.) that are a little larger, with slightly longer connexival spines and less thickened front femora in the males. This probably represents variation within the species. See pi. 3, fig. 17 and pi. 16, fig. 28. Tenagogonus-Limnometra Complex 407 Limnometra borneensis sp. nov. (PI. 2, fig. 12; PI. 12, fig. 9) Size: Winged male: 10 mm. long; width across humeri 1.89 mm.; width across mesoacetabula 2.6 mm.; width across head 1.47 mm. Winged female 9.78 mm. long; width across humeri 1.89 mm.; width across mesoacetabula 2.64 mm.; width across head 1.38 mm. Color: Yellowish brown, the dark markings dark reddish brown to black, and limited as shown in the illustration. Head with faint line near inner margin of eye, the median longitudinal pronotal stripe slender and the pale band on either side indistinct. Tip of last antennal segment probably white; distal ends of middle and hind femora pale. Dark band on propleurum close to the lateral pronotal band separated by the cream-colored marginal band; practically no color markings on mesopleuron. Venter pale. Structural characteristics: Proportional length of antennal seg- ments of male: 1st: 2nd: 3rd: 4th:: 83: 43: 60: 95. The an- tenna of the female is 5.84 mm. long, her body 9.78 mm. long. Beak moderately slender, rather short, only covering one fourth of meso- stemum. Front femur of the male thicker than middle femur and broadest beyond its middle. Middle femur of male with fringe of short cilia beneath. Mesosternum of male without long cilia, with only short pubescence. The relative length of leg segments: First tarsal Second tarsal Femur Tibia segment segment Front leg 156 130 24 22 Middle leg 415 335 ? ? Hind leg 390 270 82 26 Winged form: Shape of pronotum and venation of hemelytra as shown in the illustration, typical for Limno7netra species. Con- nexival spines surpassing genital segments in both sexes. Last ven- tral abdominal segment of male^ slightly shorter than preceding segment, in female slightly longer. The median ventral longitu- dinal keel prominent on all abdominal segments. Location of types: Described from two males and two females labeled "Mt. Pol, Sarawak, Borneo. Mjoberg collection, W. W. Funge Begnest" in California Academy collection. Comparative notes: The color pattern and the shape of the front femur of the male are the distinctive features of this species. Its closest relatives appear to be L. palauana Esaki and L. rossi sp. nov. 408 The University Science Bulletin Data on distribution: Besides the types we have a pair from "Mt. Murud, Borneo. Mjoberg collection W. W. Fnnge" that are larger, the winged male being 14.3 mm. long but the front femur of the male is like the type. The antennae are entire and more definitely white at tip. Limnometra rossi sp. nov. (PI. 3, fig. 18; PI. 12, fig. 20) Size: Winged male: 6.9 mm. long; width across humeri 1.36 mm.; width across mesoacetabula 1.9 mm.; width across head 1.18 mm. Winged female: 7.56 mm. long; width across humeri 1.5 mm.; width across mesoacetabula 2.23 mm.; width across head 1.26 mm. Color: Rather dark compared with similar species because there are numerous figures of dark brown or black as shown in the illustra- tion. The dark bands on the connexival segments are rather strik- ing. The black band on propleuron is broad and divides on meso- pleuron into two narrow bands, the upper one along the upper edge and ending at the spiracle and the lower one, an undulate line that ends at base of mesoacetabula. The brown band across the margin of the proacetabula is continued on the mesothorax to end before reaching anterior end of mesoacetabular cleft. Two brown spots on meso- and metaacetabula. Underside of connexivum with dark spots often joined into a sohd band. Venter otherwise pale. An- tennae, middle and hind femora not annulated. Structural characteristics: Proportional length of antennal seg- ments of the male holotype: 1st: 2nd: 3rd: 4th:: 70: 45: 65: 100. The total length of the antenna is 5.88 mm. and the insect 6.9 mm. long; of the female: 1st: 2nd: 3rd: 4th:: 73: 38: 60: 103. The total length of antenna is 5.67 mm. and the insect 7.56 mm. long. Beak slender covering not quite one third of mesosternum. Front femur of male much thicker than middle femur and broadest beyond its middle. Middle femur of male with fringe of short cilia beneath. Mesosternum of male with very short pubescence. The relative length of leg segments. First tarsal Second tarsal Femur Tibia segment segment Front leg 90 93 15 16 Middle leg 250 208 90 26 Hind leg 265 127 40 18 Winged form: Shape of pronotum and venation of hemelytra as shown in illustration. Connexival spines barely reaching tip of genital segments. Last ventral abdominal segment of male equal Tenagogonus-Limnometra Complex 409 in length to preceding segment, in female somewhat longer. Median ventral longitudinal keel on abdominal segments present but not prominent. Location of types: Described from ten males, nine females from "San Jose, Mindoro, Philippine Islands, E. S. Ross." These were taken on various dates from January to April, 1945. The holotype male, allotype and some paratypes in the California Academy of Sciences and some paratypes in University of Kansas collection. Comparative notes: The checkered pattern of the connexivum separates this species at a glance from its relative L. borneensis sp. nov. Data on distribution: Known only from the type series from the Philippines. Limnometra ciliata Mayr (PI. 3, fig. 14; PI. 13, fig. 22) 1865 Limnometra ciliata Mayr, Gustav L. Verb, zool-bot. Vereins, Wien, 15: 444. 1865 Limnometra ciliata Mayr, Gustav L. Hemiptera in Novara Expedition Zool. Theil, Wien, 2(1): 174. ( Places in his key to species. ) 1908 Limnometra ciliata Kirkaldy, G. W. Proc. ot Linn. Soc. of N. S. Wales. 33: 367. (Records Viti Levu, in a catalogue of the Hemiptera of Fiji.) 1933 Limnometra ciliatus Lundblad, O. Archiv. fiir Hydrobiologie, Suppl. 12. "Tropische Binnengewiisser 4," p. 371. (Records Java, Fidschi?, Molukken?) Also referring to tbe above species: 1865 Limnometra inermis Mayr, Gustav L. Verb, zool.-bot. Vereins, Wien. 15: 444. (Desc. from Manila but the type is a female, not a male and is the only specimen in tbe Vienna Museum determined by Mayr as inermis. It also iaears the label "Novara Exp. Manila." ( See Plate 3, fig. 15a. ) 1865 Limnometra inermis Mayr, Gustav L. Hemiptera in Novara Expedition Zool. Theil, Wien, 2(1): 174. (Places in his key to Limnometra and gives about the same Latin description as above but records "Manilla auf LuQon" and follows vv-ith a longer description in German. ) 1933 Limnometra inermis Lundblad, O. Archiv fiir Hydrobiologie Suppl. 12. "Tropische Binnengewiisser 4," p. 371. Mayr described this species as follows: "L. ciliata n. sp. ^ long. 19.5 mm. Fulva, pronoto nigro lineato, capite supra linea mediana lata et lineis 2 lateralibus angustis, ace- tabulus pedum intermediorum muticis, pedibus fuscis, femor- ibus intermediis ante apicem absolute inermibus, intus long ciliatis. J" ava. While the above description is short, the large size, the un- armed but densely ciliated middle femur led Kirkaldy 1908 to re- cord the species from "Viti Lavu" in the Fiji Islands. However, since Limnometra cursitans (Fabricius) from Australia is also a large species with ciliated middle femur it is necessary to give additional notes on both species. Fortunately, there are good structural char- acters that separate these species. 410 The University Science Bulletin Notes on Limnometra ciliata Mayr. Size: Mayr give the male as 19.5 mm. long. Since the males are often larger than the females in certain species of Limnometra it is not surprising to find the average size of this species to be less than 19.5 mm. We have examined specimens of this species from the Philippines to the Fijis and found the length varies from the large male type from Java 19.5 mm. to a small male from Fiji only 13 mm. long. We have other males from Fiji 17.5 mm. long. A male from Ternate is 19 mm. long. In the series from Guadalcanal the largest male is 18.9 mm. long, while the largest female is 14.2 mm. long. Color: Typical color pattern is shown on Plate 3. The colors are brown and black varying from forms having pronotum and heme- lytra mostly black to forms where brown is dominant. Venter is pale. Antennae and legs uniformly brown without pale annula- tions. On base of middle femur a short black streak. Structural characteristics: The segmental formula for the antenna is not constant even for individuals from the same place. For ex- ample the antennal measurements for four specimens, taken at the same time and place in Guadalcanal are as follows: 1st: 2nd: 3rd: 4th 245: 193: 145: 125 $ 222: 168: 136: 122 $ 156: 122: 100: 100 9 160: 127: 110: 112 9 Front femur of male is nearly straight and slender, not thicker than middle femur. First tarsal segment plainly longer than second segment. Middle femur of male with long cilia on its ventral margin, and tibia also somewhat ciliate. No pegs present on femur. Hind femur of male with some short but inconspicuous ciha and a few larger hairs. Tibia with some longer scattered hairs. Meta- sternum about as long as first two ventral abdominal segments. Abdominal venter with a moderate median longitudinal carina. Male genital segments shorter than the last two ventral pregenital abdominal segments. Rear margin of venter of the first genital of the male is characteristic of the species. It has two triangular pro- jections. (See drawings on Plate 13, fig. 22d.) This character, combined with the ciliated middle legs and the general color pat- tern, readily identify the males of this species. The latter also iden- tifies the females. The short black stripe on the base of the middle femur is characteristic of both sexes. Tenagogonus-Limnometra Complex 411 Comparative notes: A little smaller, on the average, than L. cursitans (Fabricius), has a black streak on the base of the middle femur that is lacking in all our specimens of L. cursitans. The latter lacks the triangular projections on the rear margin of the first genital male segment. Dr. Max Beier of Vienna kindly sent us a drawing of the male genital segments of the type from Java. Data on distribution: This species was described from Java and reported by Kirkaldy from Fiji. We have examined the following from: Malay Peniiisula: "Lenggong, Malay Peninsula Lea and Party." 1 male (S. Australian Museum). Thailand (Siam): "Tong. Lower Siam. Dr. W. L. Abbott". 1 male (winged) (U. S. N. M.). Sumatra: male and female (Kirkaldy Coll. K. U. ). Lesser Sunda Islds: "Sumbawa". 2 males (winged). (In our Kirkaldy coll.) These are 18 mm. long. One has width across humeri 3.15 mm., the other 2.77 mm.; one with width across the mesoacetabula 5.16 mm., the other 4.62 mm. "O Sumba, Langgai 13.7.1949 Dr. Biihler, Dr. Sutter". 1 male (Basel). "O Sumba Metoto, 4 Mai 1949. Dr. Buhler, Dr. Sutter. 1 male (Basel). "W. Sumba Kodi 3.8.1949. Dr. Biihler, Dr. Sutter. 1 female (Basel). Borneo: "Borneo Multer" 1 female (Kirkaldy coll. K. U.); "Bor- neo, Schaner" 1 male ( Bueno coll. K. U. ) ; "Borneo, Samdahan, C. T. McNamera" 2 females ( S. A. Mus. ) via Hale. Philippines: "Luzon, Manila 1913 Leg. G. Boetteker" 1 male, 1 female (winged) (British Mus.); "Luzon, 11-27-33. Victor Abalos" 1 female (winged); "Luzon, Manila Observatory Gardens" 7 fe- males, 3 nymphs (Bueno coll. at K. U.); "Los Banos. P. L — Baker" 3 males, 9 females (winged) (U. S. N. M. and K. U.). Celebes: "Sumanga, S. Celebes Nov. 1895 H. Fruhstorfer" 1 female (winged). (In Kirkaldy coll. K. U. ) Moluccas: "Temate 94 Kiikenthal". 1 male, 1 female, winged. Male is 19 mm. long. (Bueno collection at K. U.); "Halmaheira, Soaknorra, Kiikenthal". 1 male, winged, 17 mm. long. ( In Kirkaldy coll. K. U.); "Dammer". 2 females (Kirk. coll. K. U.). New Guinea: "N. Guinea, Bisiatabu, Port Moresby. W. N. Lock". 1 male. (S. Australian Museum). New Britain: "Movenafen". 1 male, 1 female (winged) (Basel, Switzerland). 412 The University Science Bulletin Solomon Islands: "Guadalcanal. 1944. L. J. Lipovsky". 8 males, 52 females (all winged); "Guadalcanal Jan. 1945. P. H. Esch- meier". 2 males, 3 females (all winged), the largest male 18.9 mm. long, while the largest female 14.2 mm. long. "Guadalcanal and Florida Is. I-III 1945 J. R. Stutz". 1 male (winged) is 17.25 mm. long. (Calif. Acad. Science); "Solomon Islands. July- Aug. 1909. W. W. Froggatt". 1 female (Bueno coll. K. U.). Fiji Islands: "Viti Levu", 1 male, 1 female (both wingless) (In Kirkaldy coll. K. U.); "Viti Levu, Lami. 1-1951 Pool. N. L. H. Krauss". 3 females; "Coll. Camille, Van Volxem". 1 male (wing- less). Det. as G. ciliatus (Mayr) by Kirkaldy 1898. This male is 15.33 mm. long. ( In Kirkaldy coll. K. U. ) ; "Viti Levu, Lami. 1-1951 Pool. N. L. H. Krauss". 3 females (winged); "Muanicula, Vanua Levu July 1934 P. W. Paine. In small shallow well." 2 males (winged); "Taveuni. Ura. Aug. 1934 P. W. Paine. Fresh water pool above high tide." 3 males, 5 females (wingless except one male); also "in shallow fresh water swamps above high tide". 3 females (wingless); "Taveuni. July 1934 P. W. Paine. Stagnant pool in rocky stream-bed in forest. 1000 ft." 1 female (winged) (the largest specimens taken by Paine are 15 mm. long); "Mvana, Vanua, Mbalavu VIII-938 Seashore. E. C. Zimmerman". 1 female (winged) (B. P. Bishop Mus.); "Bavatu, Vanna Mbalavu VIII-17- 38 E. C. Zimmerman". 2 males (winged), 5 females (winged). (B. P. Bishop Mus.); "Wainiloka, Ovalan VIM 1-38. Along stream. Elev. 200 ft. E. C. Zimmerman". 3 males, 8 females, 3 nymphs. (2 males and 5 females are winged.) (B. P. Bishop Mus.); "Matuku, 7-8-24. E. H. Bryan Jr." 1 male, 2 females; "Fieju" 1 male winged measuring 17.5 mm. in Uhler coll. ( U. S. N. M. ) ; "Rewa, Fiji. Muir 1906". 1 female (Bueno coll. K. U.). Guam: "Merizo R. 10-19-45 D. G. Frey". 2 females, 3 nymphs, one female winged. (U, S. N. M.) Limnometra inermis Mayr (PI. 3. fig. 15) See under Limnometra ciliata Mayr for references. In the same paper and on the same page Mayr described Limnometra ciliata from Java. It was a large male 19.5 mm. long with cilia on middle femur. Then he described Limnometra inermis from Manila that he said was 13.6 mm. long and lacked the cilia on middle femur. He mistakenly thought it was a male. However, the type in Vienna is a female. It has the characteristic black streak on the base Tenagogonus-Limnometra Complex 413 of the middle femur that both sexes of L. ciliata have and is exactly like the females we have from the Philippines with which were taken the males of L. ciliata. The British Museum has a male and female taken in Manilla. The type lacks antennae and front legs entirely. The middle legs have only the femora which are 11.34 mm. long. The hind legs are gone. Actually the type measures 13.02 mm. long, 2.2 mm. across humeri and 3.46 mm. across middle acetabula. Limnometra cursitans ( Fabricius ) (PI. 5, fig. 28; PI. 14, fig. 23) 1794 Gerris cursitans. Fabricius, J. C. Ent. Syst. 4, 192 17. "Habitat in Nova Hollandia." ( Old name for Australia. ) 1803 Hydrometra cursitans Fabricius, J. C. Syst. Rhyn. p. 259. 1868 Hydrometra cursitans Stal, Carl Kongl. Svenska Vet. Akad. Handl. 7: 131-133. (Says belongs to Limnometra Mayr along with L. fluviorum Fabricius. ) 1909 Limnometra cursitans Banks, C. S. Philadelphia Jour. Sci. 4(6): 583. 1933 Limnometra cursitans Lundblad, O. Archiv fiir Hydrobiologie Suppl. 12. "Tropische Binnengewasser 4" p. 371. (Quotes Australia.) This large Australian species measures from 15 to 22 mm. long and may reach 5.68 mm. across the middle acetabula. Like L. ciliata Mayr the males have the middle femora abundantly supplied with long cilia as are the tibiae. The hind femora and tibiae are also ciliated. The middle and hind femora are longer than the body. The front femora stouter. The first tarsal segment plainly longer than the second. The rear margin of the venter of the first genital segment of the male is slightly produced laterally. Location of type: The type is in the British Museum. Data on distribution: We have before us the following: Australia: "Australia, Pr. of Wales Isl. 11-15-1939, R. G. Wind." 12 males, 14 females, all winged. (K. U. col.). New Gui7iea: "Merauke, Dutch New Guinea. III-27-1939. R. G. Wind." 7 winged (3 males, 4 females); 11 wingless (2 males, 9 females ) . Limnometra femorata Mayr (PI. 5, fig. 27; PI. 13, fig. 21) 1865 Limnometra femorata Mayr, Gustav L. Verb. Zool-bot. Vereins, Wien 15: 443. 1865 Limnometra femorata Nhiyr, Gustav L. Ilemiptera in Novara Expedition Zool. Theil. Wien, 2(1): 174. ( Places in his key to species. ) 1909 Limnometra femorata Banks, C. S. Philippine Jour. Sci., 4(6): 582. 1931 Limnometra femorata Esaki, T. Bull. Biogeographical Soc. Japan, 9 (2): 211. 1933 Limnometra femoratus Lundblad, O. Archiv fiir Hydrobiologie. Suppl. 12 "Tropische Binnengewasser 4," p. 371. (Lists Borneo and Philippines.) Also referring to the above species: 414 The University Science Bulletin 1901 Gerris {Limnometra) kirkaldiji Breddin, G. Rev. Ent. 20: 93. (Described from "Ins. Banguey prope Borneo.) In Brcddin's collection. 1905 Gerris kirkahiyi Breddin, G. Mitt. Naturhist. Mus. 22: 214 (records). 1933 Limnometra kirkahiyi Liindblad, O. Archiv fiir Hydrobiologie Suppl. 12. "Tropische Binnengewasser 4," p. 371. Mayr described this large gerrid from the Philippines. The fol- lowing brief description will identify it. Size: 15-23 mm. long. The males usually larger than the females. Color: General color light brown with the usual black stripes. Hemelytra brown. Last antennal segment white. Middle and hind femora pale near tip. Distal half of middle tibia pale. Structural characters: Proportional length of antennal segments: 1st: 2nd: 3rd: 4th:: 200: 145: 206: 208. Total length of antenna 15.94 mm. Length of insect 23 mm. Middle femur of male longer than the body. Front femur slender, not thicker than middle femur, its first tarsal segment a little longer than second. The males with a large, somewhat curved spine near distal end of middle femur and two rows of about fifteen pegs. Connexival spines long and slender. Location of type: In the Natural History Museum in Vienna. Comparative notes: The large spine near the distal end of the middle femur of the male is characteristic of this species and of L. kirkaldiji Breddin. The former was described from the Philippines and the latter from the Island of Banguay, north of Borneo and south of Balbaboe Island of the Philippines. We have a male from Banguey that was labeled by Breddin "Gerris kirkaldyi Bredd," and eight other specimens from Banguey that also probably came from Breddin and are in the Kirkaldyi collection at the University of Kansas. All of these are the same color as the Philippine speci- mens with only a trace of pink on the pronotum and with light brown hemelytra and clear, shghtly darker veins. Since they are also structurally the same as L. femorata Mayr, we must record Gerris kirkaldyi Breddin as a synonym. However, all the specimens we have seen from Borneo are beautiful specimens with light reddish pronotum and bluish-black hemelytra with black veins. They would have been more deserving of a name than L. kirkaldyi (Breddin) which has nothing to distinguish it from L. femorata Mayr. Data on distribution: The Philippines: "Island Samar, Baker." 4 males (U. S. N. M.); "Island Sibuyan, Baker." 1 male (U. S. N. M.); "Iligan, Mindanao, Tenagogonus-Limnometra Complex 415 Baker." 2 males, 1 female (U. S. N. M.); "Zamboanga, Mindanao, Baker." 1 male (U. S. N. M.); "Davao, Mindanao, Baker." 1 fe- male (U. S. N. M.); "C. N. H. M. Philippine Zool. Exp. (1946-47) F. G. Werner Leg. E. Slope Mt. McKinley, Davao Province, Min- danao, stream through original forest." 1 male (Chicago, N. H. M.); "C. N. H. M. Philippine Zool. Exp. (1946-47) VIII, 22, 1946 H. Hoogstraal. E. Slope Mt. McKinley, Davao Prov., Mindanao. Forest stream." 1 female (C. N. H. M.); "Surigao, Mindanao." 1 male (British Mus.); "Philippine Islands, Basilan, Maloong. vend. M. E. Walsh." 1 male (British Mus.). Banguey Island: "Banguey, Borneo." 1 male labeled by Breddin "Gerris kirkaldiji Bredd." and may be the type. (Kirkaldy coll. K. U.); "Banguey Ins. nordl. Borneo. W. Kedengurg det. 20, VII, 1894." 1 male, 6 females. (Kirkaldy coll. K. U.); "Banguey." 1 male labeled femorata but not by Breddin. Borneo: "Nord. Borneo ex. col. Fruhstorfer." 1 female (Kirkaldy coll.); "N. Borneo. Kenabatangan Dist. S. E. end of Devvhurst Bay. C. N. H. M. Borneo Zool. Exp. 1950. R. F. Inger and D. D. Davis leg." 1 male, 1 female (C. N. H. M.); "O. Borneo. Dagavenan, Sangkoelirang Dist. M. E. Walsh." 1 male, 1 female (Basel, Switzer- land); "Muller, Borneo." 1 male (Kirkaldy coll. K. U.); "Kuching." (Beebe) 1 male, 1 female. All of the Borneo specimens have a reddish pronotum and blue- black hemelytra with black veins. Limnometra nigripennis Mayr (PI. 4, fig. 26; PI. 14, fig. 24) 1865 Limnometra nigripennis Mayr, Gustav L. Verb, zool-bot. Vereins, Wien, 15: 443. 1865 Limnometra nigripennis Mayr, Gustav L. Hemiptera in Novara Expedi- tion Zool Theil, Wien, 2(1): 174. ( Places in key to species. ) 1933 Limnometra nigripennis Lundblad, O. Archiv. fUr Hydrobiologie Suppl. 12 "Tropiscbe Binnengewiisser 4", p. .371, Mayr described this species as follows: "L. nigripennis n. sp. Long. 14 mm. Laete fulva, pronoto nigro- lineato, capite supra lineis 2 longitudinalibus antice confluentibus, fuscis, mesosterno utrinque ad pronoti marginem linea angusta nigra, tegminis cinereo-nigris margine externe laete flavo, acetabulis pedum intermediorum muticis, pedibus fulvis aut brunneis, femori- bus intermediis non ciliatis, ante apicem seriebus 2 spinularum min- utissimarum sine spina longa, femoribus posticus apice flavis, tibiis tarsisque pedum posteriorum nigricantibus. Philippinen." 416 The University Science Bulletin The type is a winged male 13.44 mm. long, 2.1 mm. across the head, 2.31 mm. across humeri, 3.53 mm. across mesoacetabula. Its abdomen is somewhat dermestid-eaten beneath. One antenna lacks the last two segments. The right front leg and right middle leg are gone. The right hind leg has lost its tarsus and the left hind leg has only one, the femur. The left middle leg has lost its last tarsal segment. The left middle femur has some short cilia on the margin of the distal one fourth and two rows of pegs, seven pegs in one row and eight in the other, the tibia has a fringe of short cilia throughout its entire length and a row of short, black procumbent spines paralleling it. The antennal formula is, 1st: 2nd: 3rd: 4th:: 145: 103: 130: 95?*. The fourth segment is white. The second antennal segment is equal to head width. This species is quite variable in color and relative length of an- tennal segments. Typically, the winged forms have the costal margin of the hemelytra strikingly lighter than the remainder of the wing but there are specimens that do not show this. Typically the second and third antennal segments have a pale or even white band and the distal segment is white, except its base. All have the last antennal segment white. The antennae are long and slender, as long as the body. In the type and some other specimens the second antennal segment is not longer than the width of the head but in many it is considerably longer and in some specimens the second segment may be equal to head width in one antenna and longer in the other. Front leg with first tarsal segment a little longer than second one. The mesosternum may be bare or par- tially covered with short cilia. In apterous forms typically the abdominal tergites are dark with a more or less apparent median longitudinal lighter stripe on last four or five segments and con- nexivum pale. However, there are some specimens that do not have the striking pattern described above. We have a series of "Island Samar-Baker" 5 males (2 winged), 3 females (apterous) (U. S. N. M. and K. U.). The winged forms have the pale costal margin; wingless forms have black abdominal tergites with paler median stripes on the last 4 or 5, and a pale connexivum. Both sexes have the fringe of short cilia and parallel row of procumbent spines on middle tibia. The middle femur of male may have a band of very short cilia on its rear margin. * Last segment may be broken. Tenagogonus-Limnometra Complex 411 The relative length of antennal segments: 1st: 2nd: 3rd: 4th:: 2nd seg.: head width: w. S 160: 110: 185: 175 5.6: 4.7 w. S 165: 120: 185: ? 5.8: 4.75 a. s 177: 138: 210: 6.8: 4.95 a. $ 210: 142: 215: ? 10.5: 5.35 a. $ 190: 144: 220: 201 7.2: 5.1 a. 9 R. 142: 94: 142: 130? L. 5.3: 4.7 R. 4.7: a. 9 132: 102: 155: 130 5.1: 4.7 a. 9 140: 92: 145: 170 4.6: 4.6 4.65 W. = winged; a. = apterous; R. = right; L. = left. Data on distribution: "Baguio, P. I. W. Robinson". 1 female winged with pale costa and banded antennae, 2nd antennal segment slightly longer than head width. "Los Banos, P. I, Baker". 3 males (winged), 3 females (wing- less) (U. S. N. M, and K. U.). antennae more or less banded, last segment white. 2nd antennal segment from less than to consider- ably more than head width. The winged male with narrow pale costal margins. The apterous with abdominal tergites dark, with more or less paler longitudinal median stripes on last four or five abdominal segments. Connexivum yellow. Philippines: "Island Polillo, Baker". 1 male, 1 female, apterous. (U. S. N. M.); "Kolambugan, Mindanao Baker". 1 winged female. Costal margin not pale. 3rd antennal segment ringed. ( U. S. N. M.); "Laguna, Mt. Maquiling, Luzon. P. L 5,28-30 A. Durjag." 1 male, wingless. (U. S. N. M.); "Majayjay, Laguna, Luzon. P, L R. C. M. C, Gregor." 1 male, 1 female, apterous. (U. S. N. M.); "Guintaboan, Victorias. Occe. Negros." 1 male (U. S. N.M.); "Guimaras P. I." 1 female; "Calian, Davao Prov., Mindanao. Phil. Islds. C. S. Clagg 31, V." 2 males, 2 females. ( M. C. Z. ); "Philip- pines." 3 males apterous. (Kirk. coll. K. U.); "Palawan Is. 750 ft. V-2-47 F. G. Werner" 2 males (winged) (C. N. H. M.); "Busuanga Is. P. I. H. Hoogstraal". 1 female (winged) (C. N. H. M.); "Busu- anga Is. P. I. H. Hoogstraal". 2 males (winged), 2 females (apter- ous) (C. N. H. M.); "Davao Prov., Mindanao sea level. 47 H. Hoogstraal". 1 male, 1 female (apterous); "Mindanao 1500 ft. 47, F, G, Werner", 3 males ( apterous ) , 9 females ( 1 winged ) ( C, N. H, M,); "Davao, Mindanao Baker", 3 females (winged), last ant. 14—2656 418 The University Science Bulletin seg. white. (U. S. D. A.); "Zamboanga, Mindanao Baker". 4 fe- males ( winged ) , last ant. seg. white, 2nd = head width. ( U. S. N. M.); "Davao, Mindanao. R. C. McGregor". 2 females (winged) (U. S.N. M.). In the last three lots above all have only last ant. seg. white, and 2nd seg. short and costal margin of wings not paler. "Luzon. P. I." 1 male, 2 females (apterous), antennae banded. Connexivum pale. (Kirk. coll. K. U. ); "Luzon". 1 male, 1 female ( apterous ) ( Bueno coll. K. U. ) . Lirnnometra ptilchra Mayr (PI. 4, fig. 23; PI. 15, fig. 25) 1865 Lirnnometra pulchra Muijr, Gustav L. Verb. Zool-bot. Vereins, Wien, 15: 443-444. "Java". 1865 Lirnnometra pulchra Mayr, Gustav L. Hemiptera in Novara Expedition, Zool. Theil, Wien, 2(1): 174. ( Places in a key with six other species. ) 1901 Gerris pulchra Breddin, Gustav. Abh. Nat. Ges. Halle, 24: 20, 85. (Quotes "Celebes. [S. Walk., He. VIII p. 168]— Jave, Borneo.") 1901 Gerris pulchra Kirkaldy, G. W. Ann. Mus. Civ. Stor. Nat., Geneva 20 (2): 804 (These are from New Guinea and are not Lirnnometra pulchra Mayr but Lirnnometra kallisto Kirkaldy.) 1903 Gerris pulchra Breddin, Gustav. Abh. Senck Nat. Ges., 25: 164. (Lists Halmahera ( Oba 232). 1920 Lirnnometra pulchra Hor\ath, G. Abh. Senck. Nat. Ges., 35: 313. (Gives distribution as: Java, Borneo, Celebes, Aniboina and New Guinea. New for Aru Island. ) 1933 Lirnnometra pulcher Lundblad, O. Archiv. fiir Hydrobiologie 1933. Suppl. 12. "Tropische Binnengewiisser 4", p. 371. (Lists: Java, Aru-Inseln, Borneo?, Celebes, New Guinea, Molukken. ) Size: Mayr gives 14.5-16 mm. long. However, we have specimens ranging from 12-17 mm. long. The males are larger than the fe- males. Color: A tan colored species. A faint figure M on the head with only the margins next the eyes showing dark brown to black. The usual median longitudinal black line, bordered by pale lines on pronotum. Lateral black lines extending from front margin to the base of humeri. Behind the humeri a dark line parallels the margin which is pale yellowish in color. This yellow stripe lying laterad of the black stripe is continued forward on either side of eye. There is a black stripe on the side of prothorax and another one oil the proacetabulum. An undulating band on mesopleuron ends above the mesocoxa. Another black band on mesoacetabulum and two on metaacetabulum. Abdominal tergites reddish. Antennae uni- formly brown. The ends of middle and hind femora pale. Structural diaract eristics: The antennal formula for the male type: 1st: 2nd: 3rd: 4th:: 160:125:140:170. The first tarsal Tenagogonus-Limnometra Complex 419 segment of the front leg is plainly longer than the second. All males have mesosternum densely covered with long whitish hairs and all three femora are fringed with cilia, that of the hind femur shorter but plainly visible. The middle femur of the male with many little black pegs among the cilia and two rows of eight or more pegs near distal end. The connexival spines of the female are relatively shorter than in the male. The middle and hind femora as long as the body in the male, a little shorter in female. Location of types: In the "Zoologische Abteilung des Natur- historischen Museums in Wien." Four male specimens in Mayr's type series. The male bearing his own label "pulclird' and "Dr. Doleschal 1859, Amboina" may be designated the type. There are two other males with the same locality and collector label. A fourth male labeled "Molukken. Coll. Signoret." These specimens fit his brief description but do not come from "Java" as he records in his description. However, Dr. Beier assures us that these are the types. They came from the Moluccas. Co77iparative notes: Dr. Mayr 1869, published a key to seven species of Limnonietra including pulchra which is useful. There are three very close and often confused species. L. pulchra, in which the males have both the middle and hind femora plainly fringed with cilia and L. annulicornis (Breddin) and L. kallisto Kirkaldy that have only the middle femora fringed with cilia and have the connexival spines a little longer, especially in the females. Data on distribution: Moluccas: "Amboina, Dr. Doleschal 1859" 3 males comprising the type series and 1 male labeled "Molukken. coll. Signoret"; "Amboina v. 1909 coll. F. Muir" 1 male, 1 female. (K. U. ) and several in (U.S. P. A.); "Prov. Ceram. F. Muir, Jan. 1909" 3 males, Febr. 1909, 1 male (Calif. Acad. Sci.); "Forsten, Ceram" 3 males, 1 female (Kirk. coll. K. U.); "Burn" 1 male ( Bueno coll. K. U.) Java: "Java Occident. Mons Cede, 4000 Aug. 1892. F. Fruhstorfer." 1 male, 2 females. (Vienna); in the Bueno collection as part of the type series of G. annulicornie Breddin we find a female with the label just above and Breddin's label Gerris n. spec, nicht pulchra." But it is pulchra. Celebes: Here belong two more specimens from Breddin's type series of G. annulicornis, namcK' "Ilalmaheira Soa komorra, Kiiken- thal" "Gerris annulicornis Bredd." female and "Halmaheira, Oba 94, Kiikenthal" a male typical of L. pulchra Mayr. 420 The University Science Bulletin Limnomctra pulchra tanganyikensis subsp. nov. ' ' (PI. 4, fig. 25) Three specimens labeled "Tanganyika S." in the Kirkaldy collec- tion in the University of Kansas collection are the only true Limno- jnetra that we have seen from Africa. There is one male that has lost its abdomen and two females, none of them has the antennae present. The color pattern is the same as in L. pulchra Mayr and the male has all three femora fringed with cilia and the meso- sternum hairy as in L. pulchra. We are therefore making it a sub- species of Limnometra pulchra Mayr in spite of its distance from the present distribution of L. pulchra as we know it. Size: Winged male: 12.7 mm. long; width across humeri 2.06 mm.; width across mesoacetabula 2.98 mm.; width across head 1.93 mm.; winged female: 12.2 mm. long; width across mesoacetabula 3.05 mm.; width across head 1.85 mm. Color: Color and pattern as in L. pulchra Mayr. See illustrations on Plate 4, figures 23 and 25. Structural characteristics: Antennae are gone. Front femur of the male plainly stouter than middle femur and fringed beneath \yith cilia; first tarsal segment plainly longer than the second. Middle and hind femora fringed with cilia, that of hind femur shorter but plainly visible. The middle femur of the male with many little black pegs among the cilia. The connexival spines of the female are relatively shorter than in the male. Middle femur not quite as long as the body in the male and shorter in the female. Location of types: Described from one male and two females from Tanganyika in the Kirkaldy collection of the Francis Hunt- ington Snow Entomological Museum, University of Kansas. We would like to see other specimens of Limnometra that are taken in Africa. Limnometra annulicornis Breddin (PI. 4, fig. 24; PI. 15, fig. 26) 1901 Cerris anmilicornis Breddin, Gustav. Abh. Nat. Ges., Halle, 24: 20, 83-85. ( Describes from Celebes. ) 1933 Limnometra annulicornis Lundblad, O. Archiv fiir Hydrobiologie. Suppl. 12. "Tropische Binnengewiisser 4," p. 371. (Quotes Celebes.) Evidently, this species has not been seen since it was described over fifty years ago. We have found in the Kirkaldy collection here at the University of Kansas what appears to be Breddin's type series. The labels are in Breddin's handwriting. There are six Tenagogonus-Limnometra Complex 421 specimens. Three labeled "Celebes Sar. Kratertiimpel des Masarang" (one male labeled Gerris annidicornis Bredd. and two females). These fit his description and have the second antennal segment annulated. They are undoubtedly of his species. The other three are L. pulchra Mayr: a female from "Halmaheira Soa konorra. Kiikenthal" labeled by Breddin Gerris annulicornh- Bredd., a male from "Halmaheira, Oha, 94. Kiikenthal" with the following label in Breddin's handwriting: "Gerris olim pulchra Bredd. (Kiikenthal Hem. p. 164) nicht die Mayr sp." and a female labeled "Java Occi- dent. Mons Cede 4000' Aug. 1892 H. Fruhstorfer" labeled by Breddin "Gerris n. spec, {nicht pulchra.). However, it is L. pulchra Mayr and there is one male and two females with the same locality label in the Vienna Museum. We have then only three specimens of L. annulicornis (Breddin): One male, which we consider the type and two females. We supply the following information from these specimens. Size: The male type is 15 mm. long; 3.27 mm. across mesoacCr tabula; 1.89 mm. across head. The females a little smaller. Color: About the same as L. pulchra but a little darker. The figure M on the head is very distinct and the lateral dark bands on the anterior lobe of pronotum are broader. These lateral bands are continuous beneath the humeri in this species but broken in L. pulchra. The second antennal segment has a pale annulation on its distal half. Structural characteristics: The proportional length of antennal segments of male type: 1st: 2nd: 3rd: 4th:: 130: 80: 115: 130. Total length of antenna about 9.56 mm.; the second segment shorter than width of head across the eyes. Front femur plainly stouter than middle femur; the first and second tarsal segments of the front leg subequal. The fringe of cilia on middle femur of male about as long as the diameter of the femur. Only four small pegs in a single row near distal end of femur. The hind femur without a fringe of cilia. The mesosternum not exactly bare but not hairy as in L. pulchra. The middle and hind femora not as long as the body. Connexival spines rather long in both sexes. Location of types: One male and two females bearing the label "Celebes Sar. Kratertiimpel des Masarang" in the Kirkaldy collec- tion at the University of Kansas. Comparative notes: This species was confused with L. pulchra Mayr by Breddin but the males may be easily separated from that 422 The University Science Bulletin species because the hind femur of the male is not fringed with cilia and its second antennal segment is shorter than the width of the head. The connexival spines of female are longer and the second antennal segment is annulated. Distribution: Known only from the types from Celebes. Limnometra vulpina (Breddin) 1901 Gerris vulpina Breddin, Gustav. Abli. Nat. Ges., Halle, 24: 20, 85-85. 1933 Limnometra vulpina ( Breddin ) Lundblad, O. Archiv fiir llydrohiologie Suppl. 12. "Tropische Binnengewiisser 4," p. 371. (Quotes Celebes.) Dr. Breddin described this species from a single female from North Celebes "(Matinang-Kette, Sudseite, 800-1200 m. (Sar.))" and gave its size as 15/2 mm. long and 4 mm. across the middle coxae. We cannot locate this type. He stated that it had longer con- nexival spines than his Gerris annuUcornis. However, since Dr. Breddin identified specimens of L. pulchra Mayr as his G. annuU- cornis and the females of L. pulchra have shorter connexival spines, he probably compared his female G. vulpina with a female of L. pulchra Mayr and said its connexival spines were longer than in G. annuUcornis when his L. vulpina was really a very large pale G. annuUcornis. Until both males and females from North Celebes make this species recognizable we will consider it a synonym of G. annuUcornis. Limnometra kallisto (Kirkaldy) (PI. 5, fig. 31; PI. 16, fig. 27) 1899 Gerris kallisto Kirkaldy, G. W. Ann. Soc. Ent. Belg., 43: 506. (Desc. female from Mysol. [W.] which is west of New Guinea.) 1899 Gerris kallisto Kirkaldy, G. W. Ann. Mus. Civ. Stor. Nat., Geneva, set. 2.a, 20 (40): 804. (Reports female from "New Guinea, Dorei Hum, Tan, 1875 Beccari". ) 1933 Limnometra kallisto (Kirkaldy) Lundblad, O. Archiv fiir Hydrobiologie. Suppl. 12. "Tropische BinnengewJisser 4", p. 371. (No additions.) Kirkaldy described this species from a single female. He says: "Belongs to subgenus Limnometra Mayr. 5 . First antennal seg- ment one-third longer than the second, one-twelfth longer than the third, fourth twice as long as the second. Apex of rostrum reaching to two-fifths of the length of the mesosternum, which is narrowly canaliculate for half its length. Base of pronotum subacutely tri- angular. Mesosternum feebly tuberculate. First segment of an- terior tarsi one-half longer than the second. Sixth segment of abdomen feebly emarginate beneath, a little swollen apically, con- nexival spines prolonged scarcely beyond the apex of the abdomen. Elytra scarcely reaching the apex of the abdomen. Length 11.7 Tenagogonus-Limnometra Complex 423 mm.; width 2.4 mm. Mysol (W. ) Higgins 1869 coll. Van Volxem. Cinnamon; head immaculate, first and second antennal segments pale yellow, third and fourth darker. Pronotum with a central and two lateral narrow and blackish lines, the former margined with pale yellow on each side, reaching to the base of the pronotum, the latter reaching to the base of the median lobe of the pronotum; posterior lobe margined with pale yellow (a little reflexed), sub- margined with brownish. Elytra with darker nervures. Apex of intermediate and posterior femora, apical half of intermediate and posterior tibiae, and all the tarsi entirely (black apical segment of anterior tarsi excepted) pale yellow. Venter obscure yellow," Dr. A. Collart has informed us that this species is no longer to be found in the Royal Museum of Belgium. In 1928 the senior author saw in that museum the type of Velia alhotrimacuJata Kirkaldy which was described in the same paper. What has become of the female type of Gerris kallisto Kirkaldy is a mystery. We found, however, in the Kirkaldy collection a female labeled "Gerris kaUisio Kirk." by Kirkaldy and bearing the label "N. Guinea Dorei Hum. II Beccari 1875". The month does not agree with the one reported in his second paper but it is imdoubtedly the same specimen and in the absence of the type from Mysol we are sub- mitting drawings of this specimen and some descriptive notes, but must point out that while it does not meet the specifications of the description of Gerris kallisto, unfortunately neither do some of Kirkaldy's known types of other species fit the published descrip- tions. Moreover, we find in the Kirkaldy collection specimens he had labeled as G. pulchra (Mayr) yet the females are structurally indistinguishable from the female he had labeled G. kallisto. We find this species to vary in size from 11.7 mm. to 20 mm. long. The males have the middle femur with a row of short cilia less than one-half as long as diameter of the femur and the hind femur with- out cilia. The middle femur may or may not have pegs and seems to be a variable character in this species. The first tarsal segment of the front leg is plainly longer than the second. Distrihtition: Moluccas: "Mysol (W.) Higgins 1869 Coll. Van Volxem" the female type. New Guinea: "N. Guinea Dorei Hum II Beccari 1875. A female labeled by Kirkaldy Gerris kallisto Kirk. (Kirk. coll. K. U.); "N. Guinea, Dilo Loria VI, VII, 90." 1 male, 2 females (Kirk. coll. 424 The University Science Bulletin K. U.); "N. Guinea, S. E. Haveri, Loria VII-XI, 93" 2 males, 4 fe- males (Kirk. coll. K. U.); "Laloki, Papua" 1 male. (Kirk. coll. K. U.); "N. Guinea, R. S. Roberts Siph. 1944." 1 male, 1 female; "N. Guinea Milne B." 1 male (Bueno coll. K. U.); "N. Guinea S. E. Maroka 1300 m. Loria VII-XI. 93" 2 large males, 20 mm. long. (Labeled by Kirkaldy Gerris ptilchra Mayr. ); "Maffin Bay, Dutch N. Guinea. VI-II-44. E. S. Ross" 2 males, 1 female (Calif. Acad. Sci.); "Finschhafen, New Guinea, IV-15-44. E. S. Ross" 1 female. Another female IV-21-44. (Calif. Acad. Sci.); "Wareo, Finsch Haven, New Guinea, Rev. L. Wagner" 3 males, 3 females. ( South Australian Mus. ); "Brena Bay, N. E. Papua, C. T. M. C. Namara" 1 female (South Australian Mus.); "Mt. Lamongton, N. E. Papua. 1300-1500 ft. C. T. McNamera" 3 males, 3 females. (South Aus- tralian Mus.); "Mt. Gyifrie, N. New Guinea, sea level 1000 ft. April 39 L. E. Cheeseman" 10 males, 7 females. The males having long front femora, not stouter than middle femora and second antennal segment longer than head width. With the same label we find 3 males that have stout front femora, second antennal segment not longer than broad width and middle femora with many black pegs among the cilia. Seems to be a variant of the above. Netv Britain: "Movehafen" 1 male (Basel Mus.). Solomon Islands: "Solomon Is. July-Aug. 1900 W. W. Froggatt" 1 male (Bueno coll. K. U.); "Guadalcanal XI-21-1944. L. J. Lipov- sky" 2 males, 3 females. ( K. U. ) ; "Solomon Is. T. H. L. Waterhouse" 2 males, 1 female (South Australian Mus.). 426 The University Science Bulletin PLATE I Fig. 1. Teiwgogonus albovittatus Stal a. Winged female. Gabon b. Antennae of the winged female. Gabon c. Wingless male. Sangmelina Canieroons, West Africa, Dec. 17, 1934, A. I. Good Fig. 2. T enagogonus zamhezinus {Vois&on) a. Wingless male. S. Rhodesia, S. Africa, G. E. Hutchinson b. Wingless female. Nagapa Res. Lab. Zululand, Africa, Apr. 22, 1922, H. H. Curson Fig. 3. Tenagogonus mudagascariensis Hoberlandt a. Wingless male. Environ de Rogez, Madagascar (paratype) b. Wingless female. The same data as above. Fig. 4. Teiiagogonus divergens Hungerford and Matsuda a. Brook near Bawomataluwo, Sumatra. Sept. 12, 1931, v. d. Meer Molxr ( holotype ) Fig. 5. Tenagogonus kampaspe {Kirkaldy) a. Wingless female. Rigo, New Guinea. July, 1889, L. Loria (type series) b. Winged male. Kelsei, New Guinea. Nov. -Dec. 1890, L. Loria (type series ) c. Wingless male. Rigo, New Guinea. July, 1889, L. Loria (type series) Fig. 6. Tenagogonus robustiis Hungerford and Matsuda a. Wingless female. Lavang, East Java, (type series) b. Brachypterous male. Pogobina, West Sumba. Sept. 17, 1949, D. Biihler, Dr. Sutter (type series) Tenagogonus-Limnometra Complex 427 PLATE I a / lb L ToiboviUatus 3 T madagascarjensis 2 T zambezmus 4 Tdivergens 6b 6 T roDjstuS 428 The University Science Bulletin PLATE II Fig. 7. Tenagogonus robustus claggi Hungerford and Matsuda a. Wingless female. Mt. Apo, Mindanao, Philippine Islands. C. F. Clagg (type) Fic. 8. Tenagogonus hrevis CLnndhlad) a. Winged (copied troni Lundblad) b. Wingless female. Java. Thienemann ( allotype ) Fig. 9. Tenagogonus fijieims Hungerford and Matsuda a. Wingless male. Taveuni, Fiji Island. Jtdy 1934. R. W. Paine (type series ) Fig. 10. Tenagogonus pravipes bergrothi Hungerford and Matsuda a. Wingless female. Observatory Garden, Manila, Philippine Islands (type series ) b. Winged male. The same data as above. Fig. 11. Tenagogonus kuiterti Hungerford and Matsuda a. Wingless male. Mohnyin, Burma. Oct. 27, 1944. K. C. Kuitert (type series ) b. Winged male. The same data as above. c. Antenna of winged male. The same data as above. d. Winged male. The same data as above. e. Wingless male. The same data as above. Fig. 12. Limnometra borneensis Hungerford and Matsuda a. Winged male. Tandjong Morawa, Serdang, N. E. Sumatia. Dr. B. Hagen ( type series ) Fig. 13. Limnometra minuta Mayr a. Antenna. Novara E.xpedition, Sambelong. (type) b. Winged male. The same data as above. Tenagogonus-Limnometra Complex 429 PLATE II 7 T roDustus claggi Da 0 T provipes bergrothi 'j 13b 12 L.borneensis 13 L minufo 430 The University Science Bulletin PLATE III Fig. 14. Liiniiomctid ciliata Mayr a. W inpe) Fic;. 16. Limnutnetra pulcnumu Esaki a. Winged female. Toloas, Tnik, Caroline Islands. Jan. 21, 19-38, T. Esaki (type series?) 1). p'ront leg. The same data as above. Fig. 17. Limnometru palauana Esaki? a. Winged )iiale. Los Negros, Admiralty Island Fig. 18. Limnometru rossi Hungerford and Matsnda a. Winged female. Mindoro, Philippine Islands. Jan. 28, 1945, E. S. Ross (type series) Fig. 19. Limnometra fiiivioium ( Fabr. ) a. Winged male. Tanjore Dist. Nedungadu, India. Mar. 12, 1938, P. Susai Nathan Fig. 20. Limnometra lipovskyi Himgerford and Matsnda a. Winged male. Guadalcanal. 1944, L. T. Lipo\ sky (type series) Tenagogonus-Limnometra Complex 431 PLATE III \7a 17 L. palauana? 20. L lipovskyi 432 The University Science Bulletin PLATE IV Fic. 21. Limnometra insularis Hungerford and Matsuda a. Winged malt-. Tandjong Morawa, Serdang, N. O. Sumatra. Dr. B. Hagen ( holotype ) b. Antenna of winged male. The same data as above. Fig. 22. Limnometra octopunctata Hungerford a. Wingless female. Tandjong Morawa, Serdang, N. O. Sumatra. Dr. B. Hagen (allotype) b. Wingless male. The same data as above. ( holotype ) c. Wingless male. The same data as above, (holotype). Ventral side. Fig. 23. Limnometra pulchra Mayr a. Amboina. May, 1909, E. Muir (type) Fig. 24. Limnometra annulicornis Breddin a. Winged female. Kratet Tiimpel der Mararang, Celebes. Breddin (type series ) Fig. 25. Limnometra pulchra tangamjikensia Hungerford and Matsuda a. Winged male. Tanganyika (type) Fig. 26. Limnometra nigripennis Mayr a. Winged male. Samar Island. Baker Tenagogonus-Limnometra Complex 433 PLATE IV 21 L insutaris 22b 22 L octopuctata 23a 26a i 23 Lpulchra 24 L anndicornis 2 5 L p tanganyikensis 26 L niqripennis 434 The University Science Bulletin PLATE V Fig. 27. Limnometra femorata Mayr a. Winged female. E. slope, Mt. McKinley, Davao prov., Mindanao, Philip- pine Islands. 1946-47, F. G. Werner b. Middle femur of female. The same data as above. Fig. 28. Limnometra cursitans a. Wingless female. Merauke, Dutch New Guinea. Mar. 27, 1939, K. G. Wind Fig. 29. Tenagogonus grandiusculus Poisson a. Wingless female. Sangmelina, Cameroon, W. Africa. Apr. 16, 1932, A. I. Good ( Determined by Poisson ) b. Antenna of wingless female. The same data as above. Fig. 30. Limnometra anadtjomene ( Kirkaldy ) a. Wingless male. Suduganga River, Matale, Ceylon. R. A. Senior-White ( type series ) b. Wingless female. Punduloya, Ceylon. E. E. Green (type series) Fig. 31. Limnometra kallisto {KixV.s\dy) a. Winged male. Moroka, 1300 m.. New Guinea. July to November, Loria (our detennination ) Tenagogonus-Limnometra Complex 435 PLATE V \ \ \yKy 27a 27 L. femorota 28 L cursitans 27 b 30 b \ 29 T longicornis 30 L onodyomene 31 L kallislo 436 The University Science Bulletin PLATE VI Fig. 32. Tenagogonus grandiusculus Poisson Wingless male. Sangnielina, Cameroons, W. Africa. Apr. 16, 1932, A. I. Good. This male was determined by us as T. longicornis Poisson. How- e\er, the female accompanying this specimen was determined by Poisson as T. grandiusculus Poisson. Tenagogonus-Limnometra Complex 437 PLATE VI 438 The University Science Bulletin PLATE VII Fig. I. Teuagogomis albovittatus Stal a. Dorsal view of the tip of female abdomen. Sangmelina, Cameroons, W. Africa. Oct. 14, 1934. A. I. Good (Compared with type by Lund- blad.) b. Ventral view of the tip of male abdomen. The same data as above. c. Lateral view of the tip of male abdomen. The same data as above. Fig. 2. Tenagogonus zambezinus Poisson? a. Ventral view of the tip of female abdomen. Nagana Res. Lab. Zuhiland. Apr. 22, 1922, H. H. Curson b. Lateral view of the tip of male abdomen. S. Rhodesia, S. Africa. 1928- 1935, G. E. Hutchinson c. Ventral view of the tip of male abdomen. The same data as b. Fig. 3. Tenagogonus madagascariensis Hoberlandt a. Ventral view of the tip of female abdomen. Tananarive, Madagascar 1937 (paratype) b. Ventral view of the tip of male abdomen. The same data as above. ( paratype ) c. Lateral view of the tip of male abdomen. The same data as alwve. Fig. 4. Tenagogonus kampasjye iKirValdy) a. Ventral view of the tip of female abdom(>n. Rigo, New Guinea mer., July, 1889. L. Loria (type series) b. Ventral view of the tip of male abdomen. The same data as above. c. Lateral view of the tip of male abdomen. The same data as above. Fig. 5. Tenagogonus divergens Hungerford a. Lateral view of the tip of abdomen. Brook near Bawomataluwo, Suma- tra. Dec. 9, 1931, v. d. Meer Mohr (type) Tenagogonus-Limnometra Complex 439 PLATE VII 3 T modagoscoriensis 5. T diverge ns 440 The University Science Bulletin PLATE VIII Fig. 6. Tenagogonus brevis ( Lundblad ) a. Ventral view of the tip of female abdomen. Java, Thienemann ( allo- type) Fig. 7. Tenagogonus fijiensis Hungerford and Matsuda a. Lateral view of the tip of female abdomen. Fiji Island. July, 1934, R. W. Paine (type series) b. Dorsal view of the tip of female abdomen. The same data as above. c. Ventral view of the tip of male abdomen. The same data as above. d. Lateral view of the tip of male abdomen. Fig. 8. Tenagogonus pravipes bergrothi Hungerford and Matsuda a. Ventral view of the tip of male abdomen. Observatory Garden, Manila, Philippine Islands, (type series) b. Dorsal view of the tip of male abdomen. The same data as above. c. Dorsal view of the tip of female abdomen. The same data as above. d. Ventral view of the tip of female abdomen. The same data as above. Fig. 9. Tenagogonus robustus Hungerford and Matsuda a. Ventral view of the tip of male abdomen. Pogobina, W. Siuiiba. Sept. 17, 1949, D. Blihler, Dr. Sutter (type series) b. Lateral view of the tip of male abdomen. Prai Jawang, East Suniba. June 14, 1949, Dr. Buhler, Dr. Sutter. c. Ventral view of the tip of female abdomen. The same data as above. Tenagogonus-Limnometra Complex 441 PLATE VIII 6. T. brevis lO.T.r.clo^i 7 T f ijiensis 8 T pravipes bengrothi 9. T. robustus 442 The University Science Bulletin PLATE IX Fic. 11. Tenagogonus kuiterti Hungerford and Matsuda a. Dorsal %'iew of the tip of wingless male abdomen. Burma. 1944, L. Kuitert ( type series ) h. Ventral view of the tip of wingless male abdomen. The same data as above. c. Lateral view of the tip of wingless male abdomen. The same data as above. d. Lateral view of the tip of winged male abdomen. Mohnylin, Burma, Oct. 27, 1944, L. Kuitert e. Ventral view of the tip of winged male abdomen. The same data as above. f. Dorsal view of the tip of wingless female abdomen. The same data as above. g. Ventral view of the tip of wingless female abdomen. The same data as above. h. Ventral view of the tip of winged female abdomen. The same data as above. P'ic. 12. Limnometra lipovskyi Hungerford and Matsuda a. Ventral view of the tip of female abdomen. Guadalcanal Island. Jan. 1945, L. J. Lipovsky (type series) b. Lateral view of the tip of male abdomen. The same data as above. c. Ventral view of the tip of male abdomen. The same data as above. Tenagogonus-Limnometra Complex 443 PLATE IX 444 The University Science Bulletin PLATE X Fig. 13. Limnometra octopunctata Hungerford a. Ventral view of the tip of male abdomen. Tandjong Morawa, Serdang. N. O. Sumatra (holotype) b. Dorsal view of the tip of male abdomen. The same data as above. c. Dorsal view of tlie tip of female abdomen. The same data as above. ( allotype ) d. Ventral view of the tip of female abdomen. The same data as above. ( allotype ) Fig. 14. Limnometra minuta Mayr a. Ventral view of the tip of male abdomen. Novara Expedition, Sambe- long (type) Fig. 15. Limnometra ftuviurum (Fabricius) a. Ventral view of the tip of female abdomen. Tanjore Dist. India. P. Susai Nathan b. Lateral view of the tip of male abdomen. The same data as above. c. Ventral view of the tip of male abdomen. The same data as above. Tenagogonus-Limnometra Complex 445 PLATE X 15 L fluvorum 14 L minuta 446 The University Science Bulletin PLATE XI Fig. 16. Limnometra anadijomene (Kirkaldy) a. Lateral view of the tip of male abdomen. Punduloya, Ceylon. E. E. Green (type series) b. Ventral view of the tip of female abdomen. The same data as above. c. Dorsal view of the tip of female abdomen. The same data as above. Fig. 17. Limnometra instilaris Hungerford and Matsuda a. Ventral view of the tip of male abdomen. Tandjong Morawa, Serdang, N. E. Sumatra. Dr. B. Hagen (holotype) b. Lateral view of the tip of male abdomen. The same data as above. c. Ventral view of the tip of female abdomen. The same data as above. Tenagogonus-Limnometra Complex 447 PLATE XI 17 L insularis 448 The University Science Bulletin PLATE XII Fig. 18. Limnometra palauana Esaki a. Lateral view of the tip of abdomen. Toloas, Tnik, Caroline Islands. Jan. 21, 1938, T. Esaki (type series?) b. Ventral view of the tip of abdomen. The same data as above. c. Lateral view of the tip of abdomen. The same data as above. d. Ventral view of the tip of abdomen. The same data as above. Fig. 19. Limnometra horneensis Hungerford and Matsuda a. Ventral view of the tip of male abdomen. Mt. Pol. Sarawak, Borneo, (type series) b. Lateral view of the tip of male abdomen. The same data as above. c. Ventral view of the tip of female abdomen. The same data as above. Fig. 20. Limnometra rossi Hungerford and Matsuda a. Lateral view of the tip of male abdomen. San Jose, Mindoro, Philippine Islands. Apr. 1945, E. S. Ross (type series) b. Ventral view of the tip of male abdomen. The same data as above. c. Ventral view of the tip of female abdomen. The same data as above. Tenagoconus-Limnometra Complex 449 PLATE XII 15—2656 450 The University Science Bulletin PLATE XIII Fig. 21. Limnometra femorata Mayr a. Ventral view of the tip of male abdomen. E. slope, Mt. McKinley, Davao province, Philippine Islands, 1946-47, F. G. Werner b. Lateral view of the tip of male abdomen. The same data as above. c. Ventral view of the tip of female abdomen. The same data as above. Fig. 22. Limnometra cilkita {W-Ayr) a. Ventral view of the tip of female abdomen. Samanga, S. Celebes. Nov. 1895, H. Fruhstorfer b. Dorsal view of the tip of male abdomen. Sumbawa c. Lateral view of the tip of male abdomen. The same data as b. d. Ventral view of the tip of male abdomen. The same data as b. Tenagogonus-Limnometra Complex 451 PLATE XIII 452 Thk University Science Bulletin PLATE XI\' Fig. 23. Limnometra cursitans ( Fabr. ) a. Dorsal view of the tip of female abdomen. Merauke, Dutch New Guinea. Mar. 27, 1939, R. G. Wind b. Ventral view of the tip of female abdomen. The same data as above c. Ventral view of the tip of winged male abdomen. The same data as above d. Lateral view of the tip of male abdomen. The same data as abo\e e. Ventral view of the tip of wingless male abdomen. The same data as above Fig. 24. Limnometra nigripennis Mayr a. Ventral view of the tip of female abdomen. Samar Island. Baker b. Dorsal view of the tip of female abdomen. The same data as above c. Ventral view of the tip of male abdomen. Philippine ( type ) Tenacogonus-Limnometra Comple:x 453 PLATE XIV 24 L ni^ipennis 454 The University Science Bulletin PLATE XV Fig. 25. Limnometra pulchra (Mayr) a. Ventral view of the tip of female abdomen. Amboina (type series) b. Lateral view of the tip of male abdomen. The same data as above c. Ventral view of the tip of male abdomen. d. Apical part of male front femur e. Apical part of male middle femur Fig. 26. Limnometra annulicomis Breddin a. Ventral view of the tip of male abdomen. b. Lateral view of the tip of male abdomen. c. Ventral \iew of the tip of female abdomen The same data as above Celebes ( type series ) The same data as above The same data as above Tenagogonus-Limnometra Complex 4.55 PLATE XV annul icornis 456 The University Science Bulletin PLATE XVI Fig. 27. Limnonietra kallisto (KiT]^a.\dy) a. Ventral view of the tip of female abdomen. New Guinea. Beccari, 1875 (Det. by Kirkaldy.) b. Lateral view of the tip of female abdomen. The same data as above Fig. 28. Limnometra palauana Esaki? a. Ventral view of the tip of female abdomen. Los Negros, Admiralty Is- land. Aug. 11, 1945. P. J. Ribert b. Ventral view of the tip of male abdomen. The same data as above c. Lateral view of the tip of male abdomen. The same data as above Fig. 29. Metastemum of Gigantometra gigas (China). Ta Han, Hainan, China. June 24, 1935 Fig. 30. Metasternum of Tenagogonus robustus Hungerford and Matsuda. Prai Jawang, East Sumba. June 14, 1949. Dr. Biihler, Dr. Sutter Fig. 31. Metastemum of Tenagogonus granditisculus Poisson. Sangmelina, Cameroons, W. Africa. Apr. 16, 1932, A. I. Good Tenagogonus-Limnometra Complex 457 PLATE XVI 28 L pobuano? 30.T robustus 29 G qigas 31 T. grandiusculus THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXTX] November 18, 1958 [No. 10 An Anatomical Study of a Neotropical Tree Frog, Centrolene prosohlepon (Salientia: Centrolenidae ) BY Theodore H. Eaton, Jr. Abstract: The osteology and major features of the myology of Centrolene prosohlepon are described, and the position of the Centrolenidae is considered. The family appears to be an offshoot from a Neotropical group of Htjla in which the following characters are present: dorsal color green; forearm relatively large, especially in male; eyes large and directed about 45° forward; bones green; femur with a low posterior crest. To these features the following, distinctive of Centrolenidae, are added as specializations: fusion of tibiale and fibulare into one bone; a basal pad on the palm; T-shaped tenninal phalanges ( the latter independently of their occurrence elsewhere among frogs. ) INTRODUCTION The family Centrolenidae was proposed by Taylor ( 1951 ) for a group of small Neotropical tree frogs in which the tibiale anti fibulare (astragalus and calcaneum) are fused into a single bone. Usually the terminal phalanges are T-shaped, and the palm has a basal pad. In these features the family differs from Hylidae, but it resembles the latter in other respects: intercalary cartilage between terminal and penultimate phalanges; procoelous vertebral column with double condyle on sacrum, dilated sacral diapophyses; arciferal pectoral girdle. Many of the species of Centrolenidae so far known are Costa Rican (Taylor, 1952), but the family extends from Mexico to southern Brazil (Taylor and Cochran, 1953). The general appearance of frogs in this family is shown by Fig. 1, the photograph of a male and female of Centrolene prosohlepon. In life they are green to cream-color dorsally, yellowish or whitish ventrally, but in alcohol the dorsal pigment becomes lavender or pink; the same is true of certain Hylidae {Agahjelmis, Phyllomcdtisa, and a few species of Hyhi, as H. alleei). In both sexes, but most (459) 460 The University Science Bulletin Fig. 1. Male (below) and female (above). This and all fol- lowing figures are of Centrolene prosoblepon (Boettger). Photo- graph by courtesy of Professor E. H. Taylor. obviously in the male, the forearm is massive as compared with the shm upper arm, and in preserved specimens it is difficult to force extension of the elbow without breaking the arm; this peculiarity is seen likewise in the above-mentioned Hylids. A striking feature of the Centrolenids is the nearly transparent skin, through which one may see muscles, nerves, parts of the skull, and Study of a Neotropical Tree Frog 461 even the brain. The bones are often green. (See Fig. 2 for details of dorsal and ventral surfaces. ) Three genera are recognized. Centrolene, normally with vomerine teeth, is peculiar in having a hooklike spine on the lateral face of the humerus in the males (as does the Australian Hyla htimeralis). C. prosohlepon (Boettger) is the only species in Costa Rica. The males of Teratolujia (one species, spinosa) have a sharp spine on the prepollex; this may or may not project through the skin. Vomerine teeth are present but very small in Teratohyla. Of the Costa Rican species of Cochranella, granulosa, aWomactdata and pulverato have vomerine teeth and about as much pigmentation as Centrolene. The remaining species, lacking vomerine teeth, tend also to reduce their pigmentation. The tympanum is still readily visible in C. valerioi, talamancae and colymhiphyUum, but is reduced and largely concealed under the skin in chrysops and fleischmanni. Evidently, then, the major group is the genus CocJi- ratieUa, from which Teratoliyla differs in possessing the prepollical spine, and Centrolene in having a humeral spine, both of these being specializations of the male. Fig. 2. Details ot ventral ( left ) and dorsal ( rij^lil ) surfaces of male, K. U. catalog No. 11202. X 2. 462 The University Science Bulletin ACKNOWLEDGMENTS To Professor Edward H. Taylor I am indebted for the opportunity to study these remarkable little frogs, including examination of all the species named above as well as certain others not yet pub- lished; also for information regarding their habits and appearance in life, and for the use of the photograph. Fig. 1. With his permis- sion and that of John M. Legler, curator of the herpetological col- lections at Kansas University, I have dissected a male and female of Centrolenc prosohlepon and examined skeletons of this and numer- ous species of Central American Hylidae. (Data on the dissected specimens are as follows: Male, catalog No. 37016, coll. by E. H. Taylor, Cinchona, Heredia Prov., Costa Rica, July 12, 1954. Female, catalog No. 32385, coll. by E. H. Taylor, near Pacayas, Costa Rica, July 2, 1952.) I wish to thank Professor E. Raymond Hall for use of the facilities of the Natural History Museum while I was Visiting Professor at Kansas University during the summer of 1957. OSTEOLOGY Skull (Figs. 3, 4). In correlation with the diminutive size of Centrolene prosohlepon (male snout-vent length 24.7 mm.), its skull shows reduction in the number and degree of ossification of bones as compared, for example, with Rami. This is especially evident in the cranium, where the otic capsules, frontoparietals, occipitals and parasphenoid all are fused, with no distinguishable sutures, and parts of the otic and ethmoid regions are cartilaginous. There is a large median dorsal fenestra, occupied by a slightly flex- ible sheet of connective tissue; through this the cerebral hemispheres are visible. Positions of the three semicircular canals show clearly by reduction of the bone enclosing them, so that they are outlined by protruding ridges. In the dry skull the limits of the parasphenoid, frontoparietals, palatines, vomers and ethmoid cannot be distin- guished, but in dissection, using Clorox to clear connective tissue from the surface, it was possible to see some faint sutures or margins of ossification, as shown in the figures. From the lateral edge of the otic capsule the slender squamosal extends forward to the posterior margin of the orbit, and also sends a narrow process down on the side of the quadrate. Between the squamosal and the ossified portion of the otic capsule is a zone of cartilage, not evident in the dry skull. The stapes, ossified super- ficially, lies just ventral to the rim of this cartilage on its posterior side, and is quite firmly fixed in the fenestra ovalis. The distal end Study of a Neotropical Tree Frog 463 clei clav. Fig. 3. Skull and first eight vertebrae of male, skeleton No. 41054, dorsal aspect, X 10. Fig. 4. Same, ventral aspect. Fig. 5. Hyoid apparatus of male, No. 11202, ventral aspect, X 10. Fig. 6. Right half of pectoral girdle, shown as flattened, male. No. 11202, ventral aspect, X 10. of the stapes reaches the dorsal margin of the tympanum rather than the center, and it seems doubtful that this bone can have much value in sound transmission. It may be on the way toward a de- generation which is reflected in the concealment of the tympanum under the skin in some species of CochrancUa. The tympanic ring is weak and comes off with the skin in dissection. There is an ossi- 464 The University Science Bulletin fied operculum, and the white otohth is visible through the posterior wall of the otic capsule. Bones of the upper jaw are supported by the quadrate, only partly ossified, and by the pterygoid, extending forward to meet the maxillary along the lower margin of the orbit. Between the quadrate and the posterior end of the maxillary is a slender quadra- tojugal. In front of the orbit the maxillary sends a preorbital process dorsomedially to meet the nasal and ethmoid. The pre- maxillaries articulate by narrow ascending processes with the nasals and also with the delicate cartilage rims of the ethmoid. Teeth are present as shown in the maxillary and premaxillary; there is some evidence of replacement in the anterior part of the series, where a partial second row can be seen in the bones. The vomer- ine teeth are borne on short ridges medial to the internal nares, and number three or four in each fascicle. The foramina for cranial nerves are similar to those in Rana, ex- cept that a large fenestra incorporates both the optic and trochlear foramina; it is covered largely by connective tissue. The mandible is composed of Meckel's cartilage, enlarged prox- imally to make the articulation with the quadrate, and ossified only near the symphysis, as a small mentomeckelian bone. A thin den- tary covers the cartilage anteriorly, and the angular does so over most of the posterior and medial surfaces, but these bones do not quite meet each other. The hyoid apparatus (Fig. 5) resembles, but is simpler than, that figured by Homer ( 1955, Fig. 143C ) for Leptodactijhis. The anterior horn, exceedingly delicate and flexible, curves forward as far as the posterior edge of the tongue; it then recurves behind the end of the mandible, medial to the depressor mandibulae, and is attached to the lower wall of the otic capsule close to the stapes. The body of the hyoid is a thin cartilage plate receiving the pos- terior ends of muscle strands from the tongue. On its posterior corners are the two bony horns (first branchials). The arytenoid cartilages enclose the larynx on each side, meeting ventrally. The thyroid cartilage is a thin, delicate ventral arch ending laterally in two fairly firm horns attached to the pericardium. There is no cricoid cartilage. (In Leptodactijhis, as cited above, a cricoid is present and the body of the hyoid bears two pairs of lateral proc- esses. ) Vertebral Column (Figs. 3, 4, 7, 8). Little needs to be added to the details figured. There are no neural spines, and the atlas Study of a Neotropical Tree Frog 465 lacks transverse processes. The centra ( Fig. 4 ) are procoelous, the sacrum has two condyles for the coccyx (Fig. 8), and the sacral diapophyses are slightly dilated. In these points and in the partic- ular form and position of the transverse processes the Centrolenid frogs are precisely like the smaller species of Hyla. The figures of the pelvis (Figs. 7, 8) are self-explanatory. Pectoral Girdle (Fig. 6). In this figure the cartilage is stippled. The sternum is a thin, fragile sheet. The cartilaginous medial ends of the coracoids overlap in the usual arciferal manner except at their anterior ends, where they fuse between the clavicles. A zone of cartilage intervenes between the ossified coracoid and scapula at the glenoid fossa, and there is a small glenoid foramen. The suprascapula bears a splinterlike cleithrum on its anterior edge, and has also a narrow irregular zone of endochondral ossification. Dorsally the suprascapula overlaps, and is fastened by connective tissue to, the tip of the first transverse process (second vertebra). This gives the pectoral girdle rigidity in the body of the frog. Arm and Hand (Figs. 9, 10, 11). There is little difference be- tween the male (shown in the figures) and female, except in the humerus. This bears a prominent thin spine on its lateral face in the male, so placed that it opposes the medial face of the thumb when the forearm is flexed and the hand adducted. Presumably the spine is involved mechanically in amplexus, but no observa- tions of this have been made. The spine is occasionally present in females, normally absent. In addition, the humerus in the male has two crests on its posterior face distally (Fig. 9), the more ventral of which is for the origin of the flexor carpi radialis muscle, and the dorsal one, slightly smaller, for that of the extensor carpi radialis. Between the two crests lies the distal portion of the anconeus. These crests are faintly or not at all indicated in the female, and the associated muscles are much smaller. It should be noted that Centrolcne is not unlike many other frogs in regard to the crests; Ritland (1955) describes a similar condition in Ascaphus. On the ventral surface of the humerus there is also a conspicu- ous groove for the tendon of the coracoradialis muscle (Fig. 10). This muscle, with its tendon, is one of the most constant features of both frogs and salamanders. The groove is deepest and broad- est near the head of the humerus; lateral to it, at the base of the spine, is a rounded ridge for the insertion of the more superficial muscles of the pectoral group. 466 The UNrvERsiTY Science Bulletin 13 Fic. 7. Sacrum, coccyx and pelvis, male, skeleton No. 41054, dorsal as- pect, X 5. Fig. 8. Same, ventral aspect. Fig. 9. Right humenis and radioulna, male, skeleton No. 41054, dorsal aspect, X 5- Fig. 10. Same, ventral aspect. Fig. 11. Right carpus, male, No. 11202, dorsal aspect, X 8. Fig. 12. Right femur, male skeleton No. 41054, dorsal aspect, proximal end to left, posterior border down, X 5. (Figs. 13-15 are from same skeleton, same magnification.) Fig. 13. Right tibiofibula, dorsal aspect, proximal end to left. Fig. 14. Right tibiale-fibulare ( astragalo-calcaneum ), postaxial aspect, prox- imal end to left. Fig. 1.5. Same, plantar aspect. F^iG. 16. Right tarsus, male. No. 11202, dorsal aspect, X 8. Study of a Neotropical Tree Frog 467 A special feature of the radioulna is the small open cleft which persists between the shafts of the radius and ulna where they have not fused completely, near the distal end. The skeleton of the carpus (Fig. 11) is essentially as in Rana. Following Ridand ( 1955 ) I am caUing the "thumb" the first digit, and the small bone lying against the base of it the prepollex. The only entirely cartilaginous pieces are two distal carpals, one at the base of the prepollex, the other at the base of the first meta- carpal. There are four other highly irregular and mostly ossified elements. The two proximal ones probably represent (a) ulnare -j- intermedium fused, on the ulnar side, and (b) radiale + cen- trale 4, on the radial side. Distal to the latter is a piece (c) which is probably centrale 1 + 2. The largest and most complex (d) is that which provides a base for metacarpals 2, 3 and 4. It probably comprises centrale 3 + distal carpals 2, 3 + 4. These inferences are based on the positions and shapes of the pieces, but confirma- tion from developmental stages would be desirable. The phalanges, not figured, number 2, 2, 3, 3 in the fingers. They, like the metacarpals, are ossified in a surface layer but apparently not internally, and have cartilaginous tips at each joint. The ter- minal ones have narrow distal extensions giving them a T-shape, unlike those of HyHdae. Each digital pad is supported by, and encloses, such a T-shaped phalanx, and a small round intercalary cartilage fits between that and the next proximal to it, as in Hylids. Leg and Foot (Figs. 12-16). In contrast to the fore limb and hand the hind leg appears long, slender and delicate, and this is shown also in the proportions of the bones. The most distinctive feature of the femur ( Fig. 12 ) is a low, thin crest on the posterior side near the base. This is not the usual crista femoris, which is lacking. The posterior crest has the insertion of the obturator ex- ternus muscle on its ventral surface, and of the quadratus femoris (more posterior) and gemellus (more anterior) on the dorsal. This crest is also present in the following small species of Hyla, of which I have seen skeletons: alleei, elaeochroa, loqttax, phlchodes, psctulopinna, and rufiocuUs, all from Costa Rica, but it is lacking in a number of larger species. The tibiofibula (Fig. 13) is nearly straight and slightly longer than the femur. The union of tibiale and fibularc into one bone (Figs. 14-16), the principal character on which the family Centro- lenidae is based, might be described as a partial fusion in Ccntro- lene, for the shafts are almost separate and retain their individual 468 The University Science Bulletin fcrs ■feu Muscles of male, No. 11202, X 8. Fig. 17. Pectoral region and arm, ventral aspect. Fig. 18. Right knee, dorsal aspect. Fig. 19. Right knee, ventral aspect, superficial. Fig. 20. Same, the sartorius reflected to show tendons. shape, while the heads are more completely fused. There remains in the tarsus proper a minute prehallux, a cartilaginous tarsal ac- companying it, another tarsal also of cartilage which is associated with the first metatarsal, and a narrow crescent-shaped cartilage probably representing a fusion of tarsals 2 + 3 or 2+3 + 4. A Study of a Neotropical Tree Frog 469 ligamentum tarsi supplens is present in the position indicated by a gap in Fig. 16, but it contains no cartilage. The tips of all the metatarsals are of cartilage. The phalangeal formula in the toes is 2, 2, 3, 4, 3. Otherwise all remarks about the phalanges and digits in the hand apply also to those in the foot. MYOLOGY Certain portions of the muscular systems were dissected and il- lustrated, but time did not permit a full description of the muscles. In Fig. 17 a ventral view of the pectoral and fore limb muscles, not including the hand, shows that the humerus lacks muscles on its flexor surface, and that the only direct flexor of the forearm is the coracoradialis, acting by way of its tendon. The muscle itself is located deep on the pectoral girdle, internal to the two supracora- coidei, and covers the coracoid fenestra ventrally. Its fibers origi- nate on the medial, cartilaginous part of the coracoid, and converge to the tendon, which then runs in its groove on the himierus and terminates at the point marked by a dot on the radio-ulna, in Fig. 10. The greatly enlarged flexor carpi radialis superficialis and ex- tensor carpi radialis (on dorsal aspect of forearm) are actually more important as indirect flexors of the forearm than in the func- tions indicated by their names, and obviously much more powerful than the coracoradialis, at least in the male. Their development is no doubt responsible for the two distal crests of the humerus already described. That this circumstance is not limited to Centro- lenids and certain big-armed Hylids is shown by Ritland's ( 1955, p. 239 ) remarks on Ascaphtts ( he uses the term "antibrachial flexors" for the enlarged forearm muscles): "Mature male Ascaphus have tremendously enlarged forearms, primarily a result of the striking expansion of the antibrachial flexors and correlated development of broad distal wings (cristae) on the humerus for their origin. . . . The muscles of females -and immature have exactly the same relative positions as those of fully grown males, but they are smaller, and since distal cristae are lacking, all originate from the humerus proper." In the female of Centrolene the only differences seen in the shoulder and arm were: (1) absence of the spine and distal crests on the humerus; (2) from the ventral side, the more lateral part of the deltoid (acromiohumeralis) could be seen uninterruptedly as it went to its insertion on the shaft of the humerus; (3) the edge of the anconeus was likewise visible without interruption on its way to the proximal end of the radioulna; (4) the two "anti- 470 The University Science Bulletin brachial flexors" were each no wider than the palmaris longus, al- though the forearm still appeared conspicuously bigger than the upper arm. Jones (1933) gives a ventral view of the pectoral muscles of Hyla arhorea; they are like those in the female of Centrolene. Figures 18-20 are intended for comparison with Noble's (1922) illustrations of thigh muscles and their tendons in Salientia. Al- though he does not show any Hylids or Centrolenids, he lists 21 species of Hijla, 5 of PhyUomeduso, one each of Agahjchnis and Ptemohyla, Centrolene geckoidewn and Centrolenella (now Cen- trolene) antioquiensis, among many others, as having the "typical bufonid tendon complex" on the ventral aspect of the knee; this is characteristic of his suborder Procoela ( Bufonidae, Brachcephalidae, Hylidae, plus the recently separated Centrolenidae ) . The semi- tendinosus tendon receives on its upper edge, just before its in- sertion, the sartorius; the gracilis passes internally to these but at its insertion is combined with them. I have compared the thigh muscles and their tendons in Centrolene prosohlepon with those of Hylo iiUeei and find them so much alike that the same illustration would almost serve for both, even though among the various genera which have the "bufonid complex" there are many minor differences of detail. This, then, is to be added to the numerous features in which Centrolenidae appear directly related to Hyh. CONCLUSIONS It seems evident, in comparing the known Centrolenids with one another, that the primitive stock of this family had the following combination of characters: (a) Those not present in Hyla. (1) Partial fusion of tibiale and fibulare; (2) basal pad in the palm; (3) T-shaped terminal pha- langes. (b) Those present in some Neotropical species of Hyla. (1) Eyes large and directed about 45° forward; (2) vomerine teeth present; (3) dorsal color green, with ample pigment, and capable of turning purplish in alcohol; (4) forearm in both sexes relatively large, with distal crests on the humerus but not a lateral spine in the male; ( 5 ) a posterior crest on the femur. ( c ) Those present in Hylidae and allied families. ( 1 ) Procoelous vertebrae; (2) arciferal pectoral girdle; (3) characters of pectoral and thigh muscles; (4) intercalary cartilages in digits, etc. As Centrolene is undoubtedly specialized in the addition of a Study of a Neotropical Tree Frog 471 humeral spine in the male, and Teratohyla in the prepollical horn, this stem-form of Centrolenidae would be a species of Cochranella. The fusion of tibiale and fibulare appears to be the major spe- cialization setting off the Centrolenids, yet it is not actually a radi- cal change, since the two bones are easily distinguishable in Centrolene for most of their length. The pad on the hand and the T-shaped phalanges appear to be specializations of minor im- portance, occurring here independently of their development in other families. These conclusions should be qualified by the state- ment that the writer has not examined the anatomy of other fami- lies of arboreal frogs, such as Hyperoliidae and Rhacophoridae. Explanation of labelling: Roman numerals indicate foramina for cranial nerves as numbered. ach — acromiohumeralis (deltoid) N. t. — tibial nerve am — adductor magnus op — foramen for deep ophthalmic anc — anconeus nerve ang — angular ar — arytenoid cartilage ci,2.3.4 — centrale 1-4 cb — coracobrachialis clav — clavicle clei — cleithrum coi,2 — comua 1, 2 of hyoid cor — coracoid crt — coracoradialis tendon d — dentary dc2.3.4 — distal carpals 2-4 e — ethmoid ecb — extensor cruris brevis eo — external oblique eno — endochondral ossification eph — episternohumeralis ( deltoid ) f — fenestra fcrs — flexor carpi radialis superficialis feu — flexor caqii ulnaris fib — fibulare ( calcaneum ) fp — frontoparietal gm — gluteus magnus gr — gracilis hy — body of hyoid i — intermedium if — ischioflexorius mm — inentonu-ckeliaii bone mx — maxillary n — internal naris na — nasal N. p. — peroneal nerve p — palatine pab — pectoralis abdominalis pal — palmaris longus pec — pectineus per — peroneus pll — plantaris longus pm — premaxillary pp — prepollex ps — parasphenoid pst — pectoralis stemalis pt — pterygoid q — quadrate qj — quadratojugal r — radiale ra — rectus abdominis r-u — radio-ulna s — sterniun sa — sartorius sc — scapula scs — subcoracoscapularis smb — semimembranosus smt — semitendinosus scj — sfjuamosal spcp — supracoracoideus profundus spcs — supracoracoideus superficialis st — stapes ( columella ) tal — tibialis anticus longus th — thyroid cartilage till — tibiale (astragalus) tp — tibialis posticus u — ulnare 472 The University Science Bulletin LITERATURE CITED Jones, E. Idris 1933. Observations on the pectoral musculature of Amphibia Salientia Ann. Mag. Nat. Hist., vol. 12, pp. 403-420, 4 figs. Noble, G. K. 1922. The phylogeny of the Salientia. I. The osteology and the thigh musculature; their bearing on classification and phylogeny. Bull. Amer. Mus. Nat. Hist., vol. 46, pp. 1-87, 23 pis. RiTLAND, Richard M. 1955. Studies on the post-cranial morphology of Ascaphus truei. I. Skeleton and spinal nerves. Jour. Morph., vol. 97, pt 1: pp 119- 174, 19 figs. 1955. Studies on the post-cranial morphology of Ascaphus truei. II. Myology. Jour. Morph., vol. 97, pt. 2: pp. 215-282, 10 figs. RoMER, Alfred S. 1955. The vertebrate bodv. W. B. Saunders Co., Phila. pp i-viii, 1-644 390 figs. Taylor, Edward H. 1951. Two new genera and a new family of tropical frogs. Proc. Biol. Soc. Washington, vol. 64, pp. 33-40. 1952. A review of the frogs and toads of Costa Rica. Univ. Kansas Sci. Bull., vol. 35, pt. 1, no. 5, pp. 577-942, 69 figs. Tayl(jr, Edward H., and Doris M. Cochran 1953. Frogs of the family Centrolenidae from Brasil. Univ. Kansas Sci. Bull., vol. 35, pt. 2, no. 15, pp. 1625-1656, 7 figs. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXLX] November 18, 1958 [No. 11 Observations on the Behavior of Brasilian Halictid Bees, III ' BY Charles D. Michener - and Rudolf B. Lange ^ Abstract: This paper presents data on the nests of several Brasilian halic- tine bees, as follows: Neocorynura polybioides (Ducke), Pseudaugochlora nigromarginata (Spinola), Megommation insigne (Smith), Augochlora semi- ramis (Schrottky) and morrae Strand, Augochlorella michuelis (Vachal), ?ar- oxystoglossa andromache (Schrottky), Ilabralictus canalkulatus Moure, and Caenaugochlora curticeps (Vachal). Virtually all the data were collected on the Southern Brasilian plateau, in the State of Parana. Several of these species exhibit incipient stages in the establishment of social behavior, as discussed in the Conclusions. INTRODUCTION This paper consists of fragmentary, but we believe significant, observations mostly on nesting behavior of various halictid bees. The observations were made on the southern plateau of Brasil, in the state of Parana, except as otherwise noted. The altitude of this area is about 900 meters; some general features of its climate are noted by Michener, Lange, Bigarella and Salamuni (1958) in con- nection with data on bee nesting sites in the Barigui roadside banks near Curitiba. ■ ^^ . . ■ 1. Part I on Pseudagapostemon, Ann. Ent. Soc. Amer., vol. 51, 1958, pp. 155-164. Part II, on Paroxystoglossa jocasta. Jour. Kansas Ent. Soc, vol. .31, 1958, pp. 129-138. The prepaiation of this paper at this time was made possible by a grant from the Na- tional Science Foimdation. Thanks are due to Father J. S. Moure for identification of the bees, for help in the field work, and especially for niakinj; nearly all of the photographs used in this paper. 2. Department of Entomology, University of Kansas, Lawrence, Kansas, U. S. A. Field work for this author's part in the study was made possible thanks to a John Simon Guggen- heim Memorial Foundation Fellowship and aid kindly made available by the Campanha Nacional de Aperfei^oamento de Pessoal de Nivel Superior. Rio de Janeiro; the Conselho Nacional de Pesquisas, Rio de Janeiro, and the Rockc-fellcr Foimdation, New York. In particular, thanks are due to Father J. S. Moure for the use of facilities of the Sec^ao de Zoologia, Faculdade de Filosofia, Universidade do Parana, Curitiba, Parana, Brasil. 3. Secgao de Zoologia, Museu Paranaense and Faculdade Catolica de Filosofia, Curitiba, Parana, Brasil. (473) 474 The University Soience Bulletin The methods of study used include several of those listed by Michener, Cross, Daly, Rettenmeyer and Wille (1955). Ordinarily nests were opened on cold days or at a time of day when all the occupants were presumed to be inside. If this were impractical or if there were time to obtain information on the activities of the various individuals inhabiting a ne.st, the nest was watched for one or more hours before opening it, and the pollen collectors ( or bees exhibiting other behavior) were captured as they returned to the nest and separately preserved. Subsequently, dissections and meas- urements revealed whether such bees differed in ovary size, fertili- zation ( shown by presence of sperm cells in the spermatheca ) , age or previous activities (shown by mandibular and wing wear), and size ( as indicated by measurements of wing length ) from bees found in the nest or exhibiting different behavior. Ovaries were classified in the following categories: very slender, slender, slightly enlarged, enlarged, and much enlarged. Each ovary has three ovarioles in halictine bees, and in the species discussed in the pres- ent paper, enlargement of the ovaries usually involved all or most of the six ovarioles. (This is in contrast to some workers in Lasio- glosstim where one ovariole, or one in each ovary, may become en- larged while the others remain very slender.) Some presumably generic differences in the ovaries were noted, as follows: In Paroxystoglossa andromache (Schrottky), jocasta (Schrottky), seahrai Moure and spiloptera Moure, the anterior parts of the ovaries, where the ovarioles are small, have the ovarioles arranged in the same horizontal plane so that all of the ovarioles are visible in this region in dorsal or ventral view. The same is, to a lesser degree, true of Pseudaugochlora nigromarginata (Spi- nola ) . In other genera studied the three ovarioles of each ovary are crowded together so that if seen in section they form a roughly tri- angular figure. Also, in some genera, such as Paroxystoglossa and Pseudaugochlora, oocytes of moderate size occur anterior to the large oocytes that are nearing the size for laying. The result is rather long ovaries, a mature oocyte being not a great deal over half of the total length of the ovary ( figure 1 ) . In some other gen- era (e. g., Neocorij7Utra), the oocytes anterior to the large posterior ones are small, so that the ovarioles seem to consist largely of one oocyte each, which is much more than half of the total ovary length ( figure 2 ) . Mandibles were classified in the following categories: unworn, slightly worn, well worn, much worn, very much worn. Behavior of Brasilian Halictid Bees 475 Fic. 1. Ovaries of egg laying female of Pseudmigochlora nigromarginata. Fig. 2. Ovaries of egg laying female of Neocorynura pohjhioides. Neocorynura polybioides (Ducke) Seasonal Cycle: Females of this species appear early in the spring. A few were taken on flowers near Curitiba on August 30, 1955 and on October 1 a single female was seen flying about one of the Barigui roadside banks. Males were first seen on December 14, when both sexes were found abundantly on flowers at Xaxim, a suburb of Curitiba. Nests of this species were found on December 21, 1955, in the Barigui roadside banks, in a vertical bank of decomposed gneiss, at a point where it overhangs and is therefore shaded. On this date four nests were opened. Several freshly opened cells, with feces, were found, indicating recent emergence of adults. One male and one female were found in their cells, not yet emerged. Three male and five female pupae were foimd, and one mature larva. No younger stages were found. One unmated, unwoni ( young ) adult female was taken flying along the bank. On January 3, 1956, a bee was found constructing a new nest in the same vicinity. This does not mean that there was synchroniza- tion of establishment of new nests and abandonment of old ones 476 The University Science Bulletin at this time, however, for two nests opened in similar banks at Arau- caria, Parana, on January 13, 1956, were old nests containing half grown larvae, male and female pupae, and adults of both sexes ready to emerge from the cells. A nest from the Barigui roadside banks opened on January 25 contained a half grown larva, four prepupae, three female and one male pupae, and an open cell ap- parently ready to be provisioned. Finally, a nest opened on March 11 contained one female and three male pupae. From these data it would appear likely that this species passes the winter as fertilized females which emerge as early as the end of August, that their progeny reach maturity by mid-December, and that from that time imtil fall (March) new adults are produced more or less continually. Social organization: Females, with worn mandibles or wings, but with slender ovaries and empty spermathecae, were not found. Therefore we judge that there is no worker caste nor approach to it. However Neoconjnura polyhioides is not an entirely solitary bee. One nest opened on December 21 contained two fertilized females with enlarged ovaries in approximately egg-laying condition (larg- est oocytes 1.3 and 1.9 mm. long) and with the crops full of pollen. Their mandibles were slightly worn, and we judge these to be bees that had overwintered and provisioned the twelve cells of the nest. Their progeny were emerging and one young unfertilized female was also in the nest. Although the youngest bee was a mature larva, it seems likely that the old bees would soon have resumed reproductive activity. Another nest opened on the same day also contained two fertilized females with somewhat enlarged ovaries (longest oocytes .6 mm.) but with unworn mandibles. We do not know whether they had overwintered and produced the brood emerging at this time; possibly the mother or mothers had died and these were among the progeny. A young unfertilized female was also in the nest, as was a female parasitized by a strepsipteran. Although evidence such as that mentioned above indicates that at least two egg-laying bees may occupy one nest (and cell cluster) simultaneously, it is also apparent that a single bee may construct a nest. On January 3, 1956, one female was found building a nest; she had constructed a chamber but as yet no cells. A nest with three empty cells that had never been occupied and only four oc- cupied cells, all containing pupae or adults ready to emerge, was found on January 13, suggesting that a single bee made the nest and that she died before we opened it. On January 25 a single female Behavior of Brasilian Halictid Bees 477 ^ was found in a nest containing pupae; she was mated and had much enlarged ovaries (largest oocyte 1.5 mm. long) but her mandibles were not or scarcely worn. We judge that the parental bee or bees had died and that this was one of their progeny nearly ready to start egg laying. On February 21 a fertilized female with much tattered wings, very much worn mandibles, and ovaries moderately enlarged and yellowish (longest oocyte, or chamber, .5 mm.) was found flying along the bank as though lost. Unpublished work on North American halictids has shown that senile bees often behave in this way and we judge that this may have been an overwintering bee nearing the end of its life. Nesis: Both at the Barigui roadside banks and at Araucaria nests of this species were mostly in loose groups, several in one or two square meters (see Michener, Lange, Bigarella and Salamuni, 1958). Each nest consists of a burrow about 3 mm. in diameter, slightly constricted at the entrance, extending horizontally or sometimes slanting upward or downward into the bank of decomposed gneiss. At a distance of 2.5 to 9.5 cm. from the entrance, the burrow opens into a chamber, which is not especially smoothed on the inside. The chamber is rather irregular in shape, generally larger in horizontal diameters (up to 3.5 cm.) than in height (up to 2.7 cm.). There is no burrow extending on into the bank from the chamber. Within the chamber is an earthen cell cluster sometimes sup- ported by a single large central pillar (fig. 34), more often supported by three to five relatively slender pillars extending from the floor of the chamber to the under side of the cell cluster ( fig. 34 ) . Hori- zontal dimensions of the irregularly shaped and slightly rough cell clusters range from 1.8 to 2.7 cm., while vertical dimensions range from 1.2 to 1.7 cm. Apparently completed clusters contain from 7 to 13 cells. Very inconclusive evidence suggests that a single female might produce as many as 7 cells while two females working jointly might pro- duce 13 or so. The cells are usually nearly vertical but most clusters contain one or more slanting or even nearly horizontal cells. The walls around the cells and the earth of the cluster be- tween cells is rather thick, in contrast to Paroxijstoglossa and some others. The cells are flatter on one side than the other, as usual in halictines. In slanting or horizontal cells the flatter side is lower- most, the more concave side uppermost. Vertical cells have the same shape. The pollen ball is placed near the lower end of the 478 The University Science Bulletin flatter side, as in other halictines. The mature larva deposits feces on the more convex surface near the lower end of the cell (fig. 3. ). The cells range from 9 to 10.5 mm. in length (from the surface of the cell cluster at the cell opening), 4.5 to 5 mm. in greatest diameter, with the opening 3 to 4 mm. in diameter. They are lined with a waxlike material except in the neck region. We have little data on the manner of nest construction. The one nest excavated that was obviously being built had a small chamber (1 cm. in diameter) with a large mass of soft dirt resting on the center of the chamber floor. From this we assume that as the chamber is dug, some of the dirt from it is left in the center of the chamber and later formed into the firm cell cluster with its cells, Fig. 3. Diagram of nest of Neocorynura pohjbiokles. Heavy lines indicate deposition of fecal material in two cells. The open cell has been used twice. Fig. 4. Diagram of nest of Neocorijrmra pohjhinides, .showing a different cell arrangement. One supporting pillar of the cell cluster, which is more or less horizontal, is hroken; the broken surface is shown as black. Fig. 5. Diagram of cell cluster of nest of 'Neocorynura poJybioidcs. show- ing an unusually irregular arrangement of cells. Figs. 6, 7, 8. Diagrams of nests of Fscudaugochlora nigromarfiinata. Some of the cells have been used twice. Cell clusters are of only moderate size. Fig. 9. Diagram of unusual nest of PseudaugochJora nigromarginata, show- ing large cell clusters. The upper cluster has been much thickened by repeated use; both cell clusters have inacti\e regions, shown by earthfilled cells. The cell clusters are largely or entirely supported by roots. Fig. 10. Sectional view of cell with pollen ball and egg. Behavior of Brasilian Halictid Bees 479 supporting pillars, etc. Cells may be reused at least twice (fig. 34). When this happens considerable soil is placed in the bottom of the cell before it is reformed and lined. Obviously the height of the cell cluster (and probably of the chamber) must be increased at the same time. An interesting nest found on December 21, 1955 was being made by an unworn stylopized bee. Possibly it was quite abnormal. It contained a small mass of very soft, friable soil, in which was a single cell lined with waxlike material (fig. 34). Liiderwaldt (1911) described the nests of another Neocorynura, N. erinmjs Schrottky. They were in rotting wood, the cells closely clustered and rectangular in section, and in other ways also so different from those of N. polybioides as to make one wonder if Liiderwaldt's description applies to bees of the same genus as N. polybioides. Pseudaugochlora nigromarginata (Spinola) Seasonal Cycle: Observations were begun on this species on Sep- tember 29, 1955, when a single nest was found in a Barigui road- side bank and opened. It contained numerous open old cells, and one closed with a pollen ball and egg. We presume that females that had overwintered were at this time provisioning newly refur- bished cells in old nests. The ne.xt occupied nest of this species was found in the same vicinity on November 22. It contained two female pupae, no other immature stages, perhaps because five cells were moldy. On February 2 a male was found in a burrow 10 cm. deep in a bank near Curitiba. (At Fazenda Sao Jose, Guaranesia, Minas Gerais, Dr. Domiciano P. de Souza Dias of Piracicaba, Sao Paulo, found a male in a short burrow in a bank on June 19, 1946. This is an area of milder winters than Curi- tiba. ) Nests found in banks near Sao Jose dos Pinhais, Parana, on February 21, 1956, contained cells being provisioned, cells with eggs, and in the same nests, prepupae and pupae of both sexes. We do not know whether the absence of feeding larval stages in the few cells examined was a happenstance or a significant fact. Females were observed on flowers collecting pollen as late in the fall as March 24 and visiting flowers for nectar on June 28, a warm winter day. The fact that this species flies on warm days in winter suggests that it is active whenever temperatures are suitable and that winter may not cause a basic break in the life cycle as it does in many bees 480 The University Science Bulletin Upical of the southern BrasiHan plateau. P. nigromarginata is basically tropical and it or close relatives range to Mexico and southernmost Texas. The well worn mandibles and much tattered wings of the fertilized egg-laying female that had evidently re- cently started her spring egg laying when the nest was dug on September 29 indicate that she must have been active also the preceding summer or autumn. The individual taken on flowers on June 28 was fresh, unworn, and unfertilized. If winter does not kill all but one age group, it will not have the synchronizing in- fluence that it has for many bees and it is not surprising to find bees in various stages throughout the warmer part of the year. It appears certain, however, that reproduction ceases during the winter months in the vicinity of Curitiba. Bertoni (1911) states that this species passes the winter as adults in the nests and re- ported three individuals in a single nest. Social organization: Our data is entirely too meager to give a clear picture of the social behavior of this bee. Several times we noticed a female plugging the nest entrance with her head, as did Ihering ( 1904 ) , but we do not know whether such guarding occurs principally in nests with more than one bee or whether the guard will turn and plug the entrance with her abdomen, as do most halictines. The nest opened on September 29 contained three females, all fertilized, only one of which was in egg-laying condition ( all ovari- oles enlarged, longest oocyte 2.6 mm.). The two with slender ovaries (longest oocyte .7 mm.) each contained a conopid larva among the abdominal organs so that they cannot be considered as normal. The nest opened on November 22 contained but one female; she was fertilized and had large ovaries (all ovarioles en- larged, longest oocyte 1.9 mm.). Although her nest was old, con- taining abandoned earth-filled cells probably used the previous season, it contained but seven cells used during the spring of 1955; possibly this is a normal number for a lone female to prepare and provision. An old nest opened on February 3 contained no imma- ture stages; the single adult bee was an unworn unfertilized female with feebly enlarged ovaries. On February 21 one old nest was found to contain a fertilized female, rather fresh, that was apparently constructing a new cell in the earth of the old cell cluster. On the same day another nest was found containing three females. One was a fertilized egg- laying individual (all ovarioles enlarged, longest oocyte 2.7 mm.) Behavior of Brasilian Halictid Bees 481 showing slightly worn mandibles and wings. A second was also fertilized but unworn and with very slender ovaries ( longest oocyte .8 mm.); probably it was young and a daughter of this nest. The third individual was workerlike in that, although its mandibles were well worn and its wings showed several nicks torn from the margins, indicating moderate age and considerable activity, it was not fer- tilized and its ovaries were very slender (longest oocyte .5 mm.). This is the individual that was carrying pollen into the nest. Neither of the fertilized ones was seen outside the nest in an hour of observa- tion before we opened it. A third nest opened on the same day contained two females, one an only slightly worn but unfertilized individual with slender ovaries, probably workerlike; the other a not or slightly worn egg layer, fertilized, with large ovaries. From February 26 to March 4 seven females were taken on flowers of Cassia in Curitiba while actively collecting pollen loads. Of these one was workerlike in being unfertilized and having slender ovaries in spite of well worn wings and mandibles. Three had large ovaries, and were fertilized; all three were but little worn. The remaining three were also but little worn; two were unfertilized, one of them with one oocyte in one ovary much enlarged (2 mm. long) but the ovaries otherwise slender, the other with the ovaries slender; the third was fertilized with slender ovaries. It is possible to say from these data that the egg-laying individuals often collect pollen and presumably do the entire work of nest con- struction and provisioning. This may happen at any time during the season of activity of this bee. However, it is also apparent that some individuals (probably a minority) work extensively while not fertilized and probably never do mate or lay eggs. Such indi- viduals have been foimd only after midsummer and may be absent in spring and early summer. These workerlike females occur in nests with egg layers and do some ( or all ) of the foraging. There is a suggestion that an average size diff^erence exists be- tween egg layers and the workerlike individuals. Wing lengths of eight clearly egg-laying bees ranged from 7.8 to 8.8 mm. with a mean of 8.3 mm. while those of four obviously workerlike indi- viduals ranged from 7.6 to 7.8 mm. with a mean of 7.7 mm. Indi- viduals intermediate as to ovary development were found; from the enlargement of only one or two ovarioles we judge these individuals to be actually intermediate and not merely young which would later develop large oocytes in all ovarioles, but we cannot be certain of this. 16—2656 482 The University Science Bulletin Nests: The nests are found singly or in groups of three or four in vertical banks of soil; those which we found were in decomposed gneiss or basalt. The distribution and exposure of those in the Barigui roadside banks is shown by Michener, Lange, Bigarella and Salamuni ( 1958 ) . The nests are often high in banks, so that it may be difficult to open them. At the entrance the burrow is constricted to about 4 mm. in diam- eter by material obviously transported from elsewhere in the nest by bees. Sometimes the entrance is located in the center of a coni- cal depression as much as a centimeter in diameter. This depres- sion is often circularly scratched, perhaps by the mandibles. The diameter of the burrow ranges from 6.5 to 8 mm. The burrow is usually unbranched and extends into the bank. It may be rather straight or sinuous, and often slopes upward, sometimes downward. Figures 6 to 9 illustrate some of the patterns which we have seen. At a depth of 14 to 48 cm. (average of eight, 34.1 cm.), the burrow opens into a large cavity or chamber. Horizontal diameters of the chamber in apparently completed nests range from 4.3 by 3.4 cm., in the case of a small chamber, to 10.5 by 6.0 cm. in the case of the largest chamber we found. Vertical heights of chambers vary from 2.7 to 4.7 cm. The chamber is irregular in shape, not particularly smooth walled. Most of the nests which we opened contained only a single chamber and had no burrow leading on from the chamber. However, several had such a burrow, connecting to almost any part of the chamber, and in two cases it lead on to a second chamber (figs. 7 and 9). Ihering (1904) described (under the name Augo- chlora gramminco Smith) a nest in which there were four chambers connected in sequence. He considered the cell clusters farthest from the surface to be the newest because of their smaller size and less regular shape. In the chamber is a large earthen cell cluster supported by fairly robust pillars of earth. The number of pillars ranges from three to seven or eight. They are generally between the bottom of the cell cluster and the floor of the chamber, but in one nest a very large pillar connected the top of the cell cluster with the roof of the cham- ber (fig. 36). The distance between the wall of the cluster and that of the chamber is 6 or 7 mm. Within the cell cluster the cells are in a generally vertical position. Their openings are rather widely separated, 5 to 8 or more mm. apart, the axes of the cells in general converging below. Cells do not usually closely approach one another nor the lower surface of Behavior of Brasilian Halictid Bees 483 the cell cluster. This is in striking contrast to nests of such forms as Paroxystoglossa jocasta (Schrottky) (see Michener and Lange, 1958a ) which are noteworthy for the thin walls between and around the cells, which diverge slightly from their openings. The manner of construction of the cell cluster is unknown. How- ever, one nest, excavated on November 24, 1955, contained a cham- ber 3 cm. long, in the center of which was a mass of soil so delicate that it fell apart at a touch. It certainly contained no wax-lined cells, and it is not clear whether it was supported by pillars or rested on the floor of the cavit>'. By analogy with Paroxystoglossa ( Mich- ener and Lange. 1958a) it seems probable that the cell cluster is built of friable soil probably removed from the walls of the chamber as it is enlarged and that only later does the cluster become firm. The nest mentioned above was inhabited and doubtless being built by a single female. She was an old bee, with wing margins en- tirely worn away and mandibles well worn. She had been fertilized and had large ovaries, but the longest oocyte was only 1.7 mm. long, not nearly large enough to be laid. The possibility exists that she was senile and her behavior abnormal, although the condition of the ovaries does not support this idea. As already indicated, the size of the cell cluster is highly variable. It is not surprising that the number of cells in it is also variable. We have seen apparently completed clusters with from 7 to 43 cells (average number of cells in nine clusters, 15.5). The larger clusters often have occupied cells in only one area, e. g., one end or much of the lateral margin, the rest of the cluster containing old, empty or earth filled cells which had once been used (fig. 9). The largest number of occupied cells that we found in any cluster was 22 in a nest opened on February 21; this cluster also contained 21 open cells, all old and empty except for one which was being provisioned. The nest was occupied by three bees when it was opened; one was a fertilized egg layer, one an imfertilized worker- like individual, and the third a fertilized but unworn, probably young individual. Although the impression given by the distribution of abandoned and used cells in large clusters is one of cluster (hence chamber) growth during the seasons or years of occupancy, leaving sections of the cluster abandoned, the bees do re-use old portions of the cluster. It is common to find cells that have been used two or three times. In such cases some dirt is placed in the cell and the new wax lining is thus separated, at the base of the cell, by .5 to 5 484 The University Science Bulletin or 6 mm. of soil from the old lining (figs. 6 to 8, 10, 35, and 36). The resnlt is that new cells are at somewhat higher le\els than those they replace, and hence the cell cluster becomes thicker as it is reused. Obviously the height of the chamber must be increased also. The cells are lined with waxlike material except near the open- ings. Their length varies from 13 to 15 mm; maximum diameter, 6 to 7 mm; diameter of cell opening or neck, 4 to 4.5 mm. They are flatter on one side than the other, the flatter side corresponding to the lower surface of horizontal cells. The pollen ball is placed against the flatter side of the cell, near the lower end of the cell (fig. 9). The jjollen ball is about 5 mm. in maximum diameter, little if any over 3 mm. in thickness. The curved white egg, 2.8 to 3 mm. long, lies on the free side of the pollen ball, in contact with it at both ends. The egg tapers distinctly posteriorly; near the an- terior end it is .7 to .8 mm. in diameter while near the posterior end it is .5 to .6 mm. in diameter. Megommation insignc (Smith) Seasonal cycle: So few nests of this crepuscular or nocturnal species were opened that we can contribute practically nothing concerning the seasonal cycle. The nests were found in the packed soil of small paths through grassy or weedy areas of nearly level ground near forested areas in the vicinity of the Barigui roadside banks, Curitiba, Parana. Jorgensen (1912) published a fairly ex- tensive account of this species, which should be consulted for further information. A nest opened on December 16, 1955, contained an adult female and, in the cells, two female pupae. A nest opened on January 5, 1956, contained two adult females and, in the cells, eggs, young larvae, and pupae. A nest opened on February 16 contained four adult females and, in the cells, eggs. A nest opened on May 4 contained two adult females, but the cells were entireh empt\. Social organization: From the above, it is obvious that more than one female may occupy a single nest. Of the two females in the nest opened January 5, one was fertilized and had large ovaries (longest oocyte 3.5 mm. long) while the other, which was unfer- tilized, had small slender ovaries, with the longest oocyte .6 mm. long. Both had slightly worn mandibles, the unfertilized one if anything showing more wear than the other. Behavior of Brasilian HALicTm Bees 485 Among the four females removed from the nest which was opened on February 16, one, which had been fertihzed, had well-worn mandibles, a few nicks in the wing margins, and enlarged ovaries (although the longest oocyte was only 2 mm. long). Evidently this was the egg layer. The other three, all unfertilized, had un- worn or slightly worn mandibles, undamaged wings, and slender ovaries with the longest oocyte .5 to .8 mm. in length. To judge by these two nests, there must be workerlike individuals and egg layers, much as in PseudaugocJiIora nigroinarginuta. Measurements of wing lengths of two egg layers were 10.1 and 11.0 mm.; of four probably workerlike bees, 9.5, 9.8, 10.1, and 10.2 mm. Of the two bees which may have been preparing to pass the winter when their nest was opened on May 4, one was unworn and unfertilized, the other had much worn mandibles and wings and was probably also unfertilized. Nest structure: This bee is especially remarkable for its beautiful nests, earlier described under the name Megalopto iponweae Schrottky by Jorgensen (1912) and Bertoni (1918). At the en- trance, each nest opening is guarded by an erect turret made of soil. In bare places the turret is usually relatively low (fig. 37), 10 to 13 mm. in height, but in a grassy place a turret which reached nearly to the tops of the grass blades was 47 mm. high. The inside diameter of the turret is about 7 mm.; its walls range up to about 5 mm. in thickness. The burrow below the ground surface enlarges to about 9 mm. in diameter. It descends straight or in sweeping curves to a depth of 31 to 42 cm. below the ground surface. At a depth of 17 to 30 cm. from the surface, a lateral burrow extends horizontally or usually slightly upward to a large, uniformly shaped, subspherical chamber which measures 5 to 6 cm. in any diameter (fig. 11). The inside surfaces of the burrows are beautifully smooth. They have no wax lining but are so smooth that a careful examination had to be made to be certain that there was no lining. The chamber is also very smooth, but one can see the marks of the mandibles on its walls, unlike the walls of the burrows. The chamber contains a cell cluster, the entire surface of which is smooth and polished, but jiot covered with waxlike material. The cell cluster is 28 to 32 mm. long, about 24 mm. wide, and 24 or 25 mm. in height. It is supported well above the floor of the chamber 486 The University Science Bulletin -^ (^-~ 15 r©16 .5cin Fig. 11. Diagram of nest of Megoiniiuitioii insigne. Broken line near bot- tom .shows position of abandoned, earth filled chamber. Figs. 12, 13. Sectional views of cell, showing pollen mass and egg. Figs. 14, 15. Diagrams of nests of Augochlora semiramis (October 25, 1955). Dotted cells in 14 are old and earth filled. Fig. 16. Longitudinal section of cell of Augochlora semiramis, showing pollen mass and egg. Fig. 17. Cross-section of nest shown in figure 15 at level indicated by arrow showing how vertical branch burrows pass close to a cell. BEHAvroR OF Brasilian Halic:tid Bees 487 by four to seven pillars which are thick where attached to the cell cluster and taper to small diameters (1.5 to 4 mm.) where they attach to the floor of the chamber. The pillars extend downward from the cluster or more or less outward from its lower outer mar- gins so that some of them can be seen from above [see especially Jorgensen's (1912) illustration]. They support the cluster at such a height that its flattened upper surface is 32 to 34 mm. above the floor of the chamber. The upper surface of the cluster slopes downward toward the main burrow of the nest. The measure- ments given above, as well as figure 11, show the unusual amount of space around, and especially above, the cluster. The cells are vertical or nearly so, opening on the upper surface of the cluster. They vary from 18 to 20.5 long, 7.5 to 9.5 mm. in maximum diameter, with the entrances 5 to 5.3 mm. in diameter. Only about the lower half of each cell is waxed. The walls of the cluster beneath and lateral to the cells are about 1 mm. thick and convexities indicating the lower ends of the cells on the under surface of the cluster are not or scarcely visible. As usual in halic- tines, each cell has one side flatter than the others (fig. 12), and the pollen mass is placed near the bottom of the flatter side. It is quite soft and there may be a little liquid in the bottom of the cell. The pollen mass is about 7.5 mm. long (vertically), 6.8 mm. wide, and 4 mm. thick. It is slightly rectangular, seen from the side on which the egg is placed (fig. 13), and looks as though it had slumped down a little in the cell. The egg is placed on the exposed subvertical surface of the pollen mass. It is white, 3.8 mm. long, arcuate, .85 to .90 mm. in maximum diameter near the anterior (upper) end, .8 mm. in diameter near the posterior end. One nest had remnants of an old cluster, now abandoned and largely earth filled, below the level of the occupied cluster (fig. 11). This would indicate long continued use of the same burrow. Augochlora seminu)iis (Schrottky) This bee is extremely common, at least at some seasons of the year, in cleared and savanna areas near Curitiba, Parana. Females were common on flowers of Senecio trichocaulon in the savanna near the suburb of Xaxim, Curitiba, in October, 1955. In Decem- ber it was noted that they were much less common in the area, but were gathering pollen. Twelve females taken on flowers of Senecio on October 9 all had slender to very slender ovaries and unworn or scarcely worn 488 The University Science Bulletin mandibles and wings; all but owe had been fertilized. Three of the twelve were collecting pollen loads on the scopa when they were captured. From this we suppose that young adult fertilized females survive the winter and start their nesting activities in the spring ( October ) . The only nests opened were found on October 25. The three nests which we opened on that day had one, three, and eight bees ( females ) in them. These twelve bees all had not or but little worn mandibles and wings and had been fertilized. One bee in each nest had slightly to considerably enlarged ovaries; the others had slender to very slender ovaries. The ones with large ovaries were not the largest of the bees. These meager data show that several bees may occupy a nest, that bees with slender ovaries may collect pollen, but that such bees are not workerlike in the sense of being unfertilized or smaller than the egglaying individual. The social organization, so far as we know it, seemingly resembles that of Aiigochloropsis sparsilis (Vachal), to be described in a subse- quent part of this series of papers. Workers may, however, appear later in the year. The nests, opened on October 25, 1955, were vertical or slanting burrows in firm bare level ground. They were widely scattered, there being no evidence of gregariousness. The diameter of the opening was about 2 mm. and of the burrow elsewhere, about 4 mm. The burrows were unbranched to a depth of four to eight centi- meters, below which there were several branches (figs. 14 and 15). Cells were located in groups of 5 to 11 or more, about as close as they could be placed, along the vertical main burrow, or on a branch. The axes of the cells are approximately horizontal. The cells do not project in all directions from the main burrow, but extend toward one quarter or another forming a vertical series. Commonly one to several of the branch burrows pass vertically close to the group of cells, often meeting one another to form com- plete rings, so that in some cases (not shown in the figures) the group of cells becomes rather well isolated from the surrounding soil. Since the cells are close together and their walls are built by the bees of very thin soil, the group of cells is very delicate. It cannot be removed intact, but can be removed in fragments from the surrounding soil, which then shows a series of concavities corresponding to the cells. Many of the cells were old and earth filled, showing that the nests had been occupied at least the season before. However, cells being constructed, others being provisioned, and others containing Behavior of Brasilia^ Halictid Bees 489 eggs, small larvae, and half grown larvae were found. Most of these cells constituted new or completely reconstructed groups but some were among groups of old cells of the previous season. The cells are about 7 mm. long, 4 mm. or slightly less in greatest diameter, with the diameter of the entrance about 2.2 mm. They are lined with waxlike material. The lower surface is Hatter than the upper (fig. 16). The pollen ball is considerably flattened, the horizontal diameters being 3.1 to 3.4 mm.; the vertical diameter 2.0 to 2.4 mm. The arcuate egg is placed on top of the pollen ball, parallel to the long axis of the cell as in other halictines. It is white, 2.1 to 2.3 mm. long, .52 to .56 mm. wide, only slightly thinner posteriorly than anteriorly. Aiiguclilom Jiiorrae Strand Two nests of this species, as well as a male in a different burrow, were found in a vertical bank near Sao Jose dos Pinhais, Parana, on February 21, 1956. In the nests roughed out but unlined cells, waxed cells ready for use, as well as small larvae and a white female pupa were found. Thus individuals in all stages could be expected at this season (late summer). Outside the region of the southern Brasilian plateau, we have data on one nest in a roadside bank between Boa Vista and Corcovado, Rio de Janeiro, opened February 25, 1956, by Dr. Carlos Alberto Campos Seabra and Father J. S. Moure. This nest contained a prepupa, eight female pupae, and two male pupae. A nest was excavated by us at about sea level in flat ground at Guaruva, Santa Catarina, on October 30, 1955. The burrow was closed at the surface; we foimd it by chance. A single female was in the nest. There were eight cells, those nearest the surface containing pupae (one male, one female), the rest containing larvae of various sizes, the smallest being in the deepest cells. With such scanty data, we call only be indefinite about social relationships. It seems probable that the nest from Guaruva was made and the eight cells provisioned by the single female that was found in it. One of the nests opened February 21 contained only four cells and was occupied by a single fertilized female with well worn mandibles, much tattered wings, and rather slender ovaries, which may account for the fact that of the four cells, only one was or had been occupied (it contained a small larva). A group of old, earth-filled cells indicated that the nest was being reused. The other nest opened on the same day contained about eight 490 The University Science Bulletin Fig. 18. Diagram of nest of Augochlora morrae. Cells indicated by dotted lines were old and earth filled (February 21, 1956). Figs. 19-21. Diagrams of nest of Augochlorella michaelis (Dec. 4, 1955; Jan. 13 and February 21, 1956, respectively). Fig. 19 and 20 are side views of nests with strongly slanting cells. Fig. 21 is side view of a nest with cells but slightly slanting. The marginal holes around the cell cluster (which has two open cells) lead to the passageways behind the cell cluster. Fig. 22. Top \iew of nest shown in fig. 21, showing passageways behind cells. Fig. 23. Front \'iew of the cell cluster shown in fig. 19 (closed cells are shown dotted ) . Fig. 24. Back view of cell cluster of nest shown in figs. 21 and 22. Shaded areas are eartli supports between the passageways behind the cells. Fig. 25. Diagram of nest of Paroxystoglossa anclromache .showing cell cluster in side view in contact with chamber wall and supported by loose dixt (Oct. 23, 1955). Figs. 26-27. Diagrams of nests of Paroxystoglossa anclromache showing cell clusters in front view, in one case' with pillars (Oct. 23 and 24, 1955). Fig. 28. Top view of horizontal section of cell cluster of Paroxystoglossa (itidromache (Oct. 24, 1955). Fig. 29. Side view, vertical section, of same. Behavior of Brasilian Halktio Bees 491 cells, several parasitized by mutillids, others containing small larvae and pupae. There were three unfertilized, iniworn young adult females with slender ovaries in the nest, probably recently emerged, but the bee or bees that made the nest seemingly had not survived. The nest from Rio de Janeiro similarly contained only young un- fertilized females, in this instance four of them. In this case, however, the nest was dug on a warm day without a preliminary period of watching and it is possible that older bees were afield. The nests are burrows about 5 mm. in diameter, narrowed at the entrance, and extending horizontally into a bank or slope slightly downward, or are vertical in flat ground. The burrows are 10 to 20 cm. deep and may be simple or branched. The cells are essentially horizontal, very like those of Chloralictus, but in denser clusters than usual in that group. Among the nests which we studied the cells were in groups of four to nineteen. The cells are lined with waxlike material, the lower sides flatter than the upper. The pollen is in the form of a flattened firm ball, with horizontal diameters 2.9 to 3.2 mm., the vertical diameter 2.1 mni. Au^ochlorella michaelis (Vachal) Only three nests of this species have been studied. One, in the Barigui roadside banks near Curitiba, Parana, was found about November 4, 1955, and a bee was seen going in and out. It was excavated on December 4, at which time the entrance was closed. It contained a single fertilized female with well worn mandibles and wings, and in the six cells, young of various stages from egg to female pupa. Four of the cells eventually produced females, none males. Another nest was found in a bank at Araucaria, Parana, January 13, 1956. It also contained an old fertilized female (mandibles much worn, wings tattered). Its two cells both contained female pupae. The small size of the nest combined with the worn condition of the female suggest that the nest may have been made by a senile bee that had worked elsewhere pre- viously. The third was found in a bank near Sao Jose dos Pinhais, Parana, on February 21, 1956. This last nest contained, in its cells, seven eggs, a half grown larva, four mature lar\ae, three prepupae, fifteen male pupae and six female pupae. Unlike the first two nests mentioned, the last contained several bees. Three were captured as they returned to the nest with pollen loads on the scopa. These had slightly worn mandibles, indicating at least moderate age, but their ovaries were slender ( longest oocyte .2 to .3 mm. ) and the\' were unfertilized. Clearly they were worker- 492 The University Science Bulletin like individuals. Three other females were in the nest when it was opened. One was unfertilized, unworn, with slender ovaries and was doubtless recently emerged. Two were fertilized and had very much worn mandibles, worn wings, and enlarged ovaries (longest oocyte 1.7 mm. in both cases). Clearly these two were egg layers. Their wing lengths were 5 and 5.1 mm.; those of the three workerlike bees were 4.6, 4.7, and 4.8 mm. The social re- lationships seem similar to those of Pseudaiigochlora nigromar- ginatii, in which there is also a suggestion of a size difference be- tween the workerlike and the egg-laying females. The three nests of this species that we studied were very different from one another, so that one is forced to doubt if they were really made by the same species of Augochlorclhi. Each consisted of a nearly horizontal, rather straight, unbranched burrow 7.5 to 10 cm. deep and 4 mm. in diameter, narrowed at the entrance. The bur- row in the Barigui nest ended in a sloping space on one side of which was a cluster of six completed closed cells and three rough, unlined cells (figs. 19, 23.) These cells were built of walls made by the bee, having a thickness of .7 to 1.0 mm. The subcircular cell cluster could be lifted out of its position intact, but there was no space around the cells except for the flat space over the cell entrances. The convexities of the closed ends of the cells were represented by concavities in the unworked earth in which the nest lay. From the space over the cells, a short burrow ex- tended on into the bank. The axes of the cells were at an angle of about 45° to horizontal, as was the plane of the flat space provid- ing access to them. The nest from Araucaria, having but two cells, lacked a space, the slanting cells merely extending downward from the burrow (fig. 20). The nest from Sao Jose dos Pinhais was similar to the Barigui nest in having a flat subcircular cluster or plate of cells, but this cluster was more nearly vertical, so that the cells themselves were more nearly horizontal. There were 38 cells, two of them empty, in the cluster. Around its margins a series of holes extended inward and connected with interconnecting passages between the backs of the cells and the soil, so that the cell cluster was supported by the spaces and pillars among these passages and by the soil between the holes leading to these passages (figs. 21, 22, 24). The passages had been excavated in the unworked soil behind the cells, leaving the convexities of the closed ends of the cells visible. The cell walls were .5 to 1.0 mm. thick throughout. Behavior of Brasilian Halictid Bees 493 Cells, lined with waxlike material, are flatter on the lower sur- faces than the upper, and are 9 to 9.8 mm. long, 4 to 5 mm. wide at the widest point, and 2.5 to 3 mm. wide at the neck. The pollen ball, which is a flattened sphere, about 3.3 mm. in the greatest diameter and about 2 mm. in the smallest ( morphologically vertical ) diameter, is placed on the flatter side of the cell near the closed end, as usual in halictids. The egg is 2.1 mm. long, arcuate, white, and placed on the upper surface of the pollen ball. Paroxijstoglossa andromache (Schrottky) The biology of this species is probably similar to that of P. jocasta (Schrottky) (see Michener and Lange, 1958a). Our first record of it is of a burrow being dug on August 30, 1955 and a few nests were found in various stages of construction during Sep- tember, October, and January, 1956. Cells being provisioned or containing eggs were found on October 14, 23, and 24, and cells with pupae were found on January 27. Unlike P. jocasta, nests are usually found in banks or steeply sloping ground. We found them in the Barigui roadside banks (see Michener, Lange, Biga- rella and Salamuni, 1958) and in a bank at Xaxim near Curitiba, in a bank near Sao Jose dos Pinhais, and in a bank near Campo Largo, all in the state of Parana. On a cold day (September 16, 1955) three males were found, one in a burrow 3 cm. deep in a bank, two in burrows of Augo- chloropsis diversipennis ( Lepeletier ) . Of twelve nests opened, only one contained two adult female bees. As in such cases in P. joca.sta, one of the bees had enlarged ovaries, the other slender ovaries, but both had been fertilized. The nests are similar to those of P. jocasta except that the btirrow, which is 9 to 25 cm. deep, is horizontal or slants downward in- stead of being vertical. The cell cluster, like that of jocasta, consists of very thin-walled cells, usually very few in number (figs. 28, 29). The largest clusters we found contained only four cells. The cell cluster is placed in a chamber in which it may be supported by pillars as in jocasta but usually the lower part of the chamber, between the cluster and the wall of the chamber, is filled with friable soil which supports the cluster (figs. 25 to 27). Pillars may or may not be entirely absent in such cases. The cells are horizontal and lined with waxlike material, and shaped as in P. jocasta. They are 11 to 12 mm. long, 5 to 5.25 mm. wide, the neck about 2.75 mm. wide; the pollen ball is slightly 494 The University Science Bulletin flattened, about 4 mm. in horizontal diameter. The white arcuate egg i.s about 2 mm. long. Hcihralictus canaliciilatiis Moure Seoivnal Cycle: Thi.s i.s a minute and inconspicuous, and appar- ently rather rare bee about which we can contribute but little. Three nests were found on October 23, 1955, in a small moss covered bank about 20 cm. high, completely shaded in the forest near the Barigui roadside banks, Curitiba, Parana. Later these nests could not be located in spite of weekly visits and were evi- dently closed but on January 8, 1956, loose dirt, indicating recent excavation, was noted at the same place and on January 15 one of the nests was dug. Two males and three unworn young females but no immature stages were found in it. It seems very probable that in spring (October) the nests were provisioned. Appiu-ently they then became closed and another brood emerged in January. On Februiuy 29 a nest was found in one of the Barigui roadside banks (see Michener, Lange, Bigarella and Salamuni. 1958). It contained four cells, one of them with a half-grown lar\a, two with eggs, and one being provisioned. On March 11 a nest in the same area, containing but one cell which was being provisioned, was found. On March 16 a nest was found containing two half- grown larvae, numerous mature larvae and prepupae, and twelve male and three female pupae (white). On March 18 another nest with three prepupae in it was found. These data suggest a brood that matures in the fall and provisions cells in the spring. The summer brood from these cells, appearing in January and provisioning cells through February and early March, gives rise to the brood that matures in the fall. The two broods would be comparable to those of F scudagapostcmon ( Mich- ener and Lange, 1958). Social organization: Of the very few nests examined, two were known to be occupied by but one female, and contained few cells, although they were still being enlarged and more cells would prob- ably have been built. The large nest opened on March 16 con- tained 37 cells and a single old female with much worn mandibles. It seems very probable that this one bee did not make and provision so many cells. Probably several bees had li\ed together in this nest. The presence of some old dirt-filled cells in addition to the 37 occupied ones indicated that this nest had been occupied during the previous generation. Behavior of Brasiliax Halictid Bees 495 Fig. 30. Diagram ot nest ot Habralictus canaliculatus (Feb. 29, 1956). Fig. 31. Diagram of nest of same (Mar. 16, 1956). Fig. 32. Diagram of nest of Cacnaugnchlora curticeps (Oct. 17, 1955). Fig. 33. Diagram of cell, with polUn mass and egg, of same. 496 The University Science Bulletin Fig. 34. Cell clusters of Neocorynura pnlybioides. Black lines represent 1 cm. Top three photographs show vertical sections and a top view of the cluster from a nest opened on Dec. 21. The central cell, at upper left, has been used three times. Middle and lower left show side view and a vertical section of a cluster that was unusual in being supported by a single large pillar (Dec. 21). Middle right is a top view of the same cluster. Two cells were closed, but their contents moldy. Lower right is a \ertical section of a cell "cluster" (of one cell) in the nest of an unworn stylopized female (Dec. 21). Behavior of Brasilian Halictid Bees 497 ^ n'f '-^"^ Fig. 35. Cell clusters of Pseudaugochlora nigromarginatu. Black lines rep- resent 1 cm. At left, top and side views of a cluster from a nest opened on Sept. 29. Faintly depressed areas are plugs of provisioned closed cells. At right are vertical sections through a cell cluster. The upper one shows a pro- visioned cell (left) and smooth but unwaxed cells (center) at least one of which had been used previously as shown by old fecal matter below the base of the present cell. The lower one shows the same provisioned cell (left), a shaped but not smoothed cell (center) and a smoothed but imwaxed cell (right). Except for obviously young bees, every female dissected (si.x, including a few captured flying about banks in the Floresta da Tijuca, Rio de Janeiro, January 9, 1956, by Dr. Carlos Allierto Campos Seabra) had been fertilized and had enlarged ovaries with eggs. Thus there is no sugge.-Jtion of a worker caste. When sev- eral females occupy one nest, their interrelationships may be simi- lar to those of Pseudagaposteinon. Nests: Each nest consists of a slender, unbranched burrow ex- tending into the bank and sloping downward. The burrow is 2.5 to 2.8 mm. in diameter, narrowed at the entrance to 1.5 mm. in diameter. These burrows reach a length of at least 32 cm. (large nest opened March 16); smaller and incomplete nests which miglit have been further deepened had burrows 12 and 18 cm. in length. The cells are horizontal or but slightly slanting. 2 to 4 cm. from 498 The University Science Bulletin Fig. 36. Cell clusters of Pseiidattgochlora nigromarginata. Black lines rep- resent 1 cm. Top, side view and vertical section of cluster which had three supporting pillars below (broken surfaces outlined in black in side view) and an unusual pillar from top of clnster to roof of chamber. The section shows a horizontal cell ( lower left ) and an old earth filled slanting cell ( lower right ) (Nest opened Nov. 22). Center, top view of old cell cluster in which one cell ( lower left ) has been prepared for use. Rootlets which supported this cluster can be seen (Nov. 22). Lower left, vertical section of the same, showing fresh cell at left, and other cells, some of which, although abandoned, were closed at the surface. Reuse of several cells is evident. Fecal material is well shown in the third cell from the right. The second cell from the right shows par- ticidarly well how waxlike linings of various cells ha\e been ra.sped away, ap- parently by mandibular action in enlarging the cell entrance. Lower right, top view of small cluster. All cells were open; entrances of some of them had been much enlarged. Behavior of Brasilian Halic:tii> Bees 499 the main burrow. In the smaller nests, being built by lone fe- males, the youngest cell was that farthest from the nest entrance and the oldest cell was nearest the entrance. Each cell had pre- sumably been constructed at the end of a different lateral burrow, but iUl laterals except that leading to the cell being provisioned were so well filled with soil as to be unrecognizable. In the large nest opened on March 16 the younger and older brood was not system- atically arranged, and the cells were grouped in such a way as to suggest that several may have been constructed branching from the same lateral. However, since all the laterals were completely and indistinguishably earth filled, this is not certain. Fic. 37. Turret at nest entrance of Megommation insigne. Scale at right is in centimeters. The cells measure 5 to 6 mm. in length, 2.8 to 3.3 mm. in maximum diameter, with the diameter of the neck about 1.8 mm. The cells are lined with waxlike material except for the necks. The lower surface or floor of a cell is flatter than the upper and the pollen ball lies on the floor near the distal end of the cell. Horizontal diameters of a pollen ball were 2.1 and 2.3 mm., the vertical diameter was 1.9 mm. The curved egg is placed on top of the pollen ball, in con- tact with it at both ends. 500 The University Science Bulletin Caenaugochlora curticcps (Vachal) Seasonal cycle: The few nests of this small species that were studied were scattered along the Barigui roadside banks; their dis- tribution there was indicated by Michener, Lange, Bigarella and Salamuni (1958). The first nest found was on October 14, 1955. At this time both sexes were on the wing. The nest found on October 14 was opened three days later. It consisted of a burrow, with no cells. The fe- males captured at this time were not or only slightly worn. The last spring male was seen on October 23. The next nest found was opened on October 29. It contained provisioned cells with eggs but no larvae. A nest opened on November 16 already had mature larvae, as well as small larvae and eggs. After the later part of November, no individuals of this species were seen until January 21, 1956 when both sexes were noted flying along the banks. On February 29, it was noted that while adults had not been seen for some time, a well worn female was found in a burrow. On March 11 a few females were again seen flying along the banks. The nvimber of observations of this small and uncommon species is small, but we may surmise that the seasonal cycle is similar to that of Hahralictiis canaliculatus (see above) and Fseudag^a- postenion (Michener and Lange, 1958). There seems to be a brood that appears in the fall (March); both sexes survive the winter and the females provision nests in spring (October, No- vember). The progeny of this brood appear and provision nests in summer (January. February). Social organization: Only three nests could be opened com- pletely because of rarity of the bee and the difficulty of following the deep, slender burrows. Two of the three nests contained three females, one contained four. Observations of activities at nest entrances indicated that some nests contained at least five females. In these nests, during warm sunny weather, one or another of the females commonly plugged the entrance with her head. As with Pseudagapostemon (possibly also Hahralictus canalicu- latus) activities seemed well synchronized. The bees taken from any one nest at one time had similarly developed ovaries, similarly worn mandibles, etc. Except for obviously yoimg bees, all females found had been fertilized. There is, thus, no indication of a worker caste. Perhaps the interrelationships among females in a nest are similar to those that occur in Pseudagapostemon. Behavior of Brasilian Halictid Bees 501 Fig. 38. Cell clusters of Megommation insigne. Black lines represent 1 cm. Top left, top view of cluster with one cell open, others closed and with im- mature stages (Nest opened Jan. 5). Top right, bottom view of same cluster. Broken ends of pillars are circled in black. Middle left, side view of .same cluster. Middle riglit, vertical section of same cluster. Bottom, top and side views of cluster from nest opened on Feb. 16 (retouched to empnasize cell openings ) . .502 The University Sc:ience Bulletin Nests: Each nest consists of a simple or branched burrow which extends into the bank, slanting slightly downward for 18 to 39 cm. and then turns more steeply downward. Total depth of tlie bur- row ranges from 65 to 90 cm. In diameter the burrow ranges from 3 to 4 mm., narrowed to about 2.75 mm. at the entrance. The cells are nearly horizontal, sloping slightly downward, and located one to four centimeters from the main burrow. As with HahroUctus camiUculatus, they are often so arranged that each cell must have been constructed at the end of a separate lateral burrow, but sometimes they are in groups suggesting that several cells may have been connected to a single lateral. Also as in the Fig. 39. Cell clusters of MegommatUm insignc. Black lines represent 1 cm. Top, top and side views of cluster from nest opened on Mav 4. Lower left, ohlitine bottom \ie\v of same. Black circles enclo.sc broken surfaces of pillars. A curious canal, bridged in the center, was the outstanding feature of this cluster. Lower right, pupa in cell (Dec. 16). The reticulate pattern of pale hnes m the waxed parts of the cells appears only after drying, and perhaps residts from differential shrinkage (retouched to emphasize cell .shape). Behavior of Brasilian Halictid Bees 503 Habralictus, once a cell is provisioned and an egg laid in it, the lateral burrow is completely filled with dirt so that it is unrecog- nizable. In nests opened during the season of cell construction and provisioning, it was the lowermost one or two lateral burrows that were open, suggesting that the nests are deepened and new laterals constructed at the bottom as the season progresses. The cells measure 8 to 9 mm. long, about 4.4 mm. in maximvnn diameter, about 2 mm. in diameter at the neck. They are lined with a waxlike substance except at the entrance. The pollen ball lies near the closed end of the cell, is nearly spherical, about 2.8 mm. in diameter. The curved white egg, about 1.5 mm. long, is placed on top of the pollen ball. The egg is but little narrower posteriorly than anteriorly. The species of Caetuntgochlora whose nests were described by Claude-Joseph (1926), C. chloris (Spinola), opaciceps (Friese) [^ scitiilus (Vachal)], and rostraticeps (Friese), construct nests with a cell cluster somewhat like that of Atihapman), 1899. Nirmus fusco-marginutua var tuncricunus Kellogg and Chapman, 1899, Occ. Pap. Calif. Acad. Sci., vol. 6, p. 69. Type host: Cohjm- bus nigricoUis californicus ( Heermann ) . j* syntype and 5 syn- type, no. 1330, from type host collected in the Bay of Monterey, California. Present status: Aquiinirmus aniericanus (Kellogg and Chapman), 1899. Nirmus pcicificus Kellogg and Chapman, 1899, Occ. Pap. Calif. Acad. Sci., vol. 6, p. 70. Type host: Lunda cinlmta (Pallas) (error) and Ccpphus colwnha columhu Pallas; from Carriker in lift, j syntype and 5 syntype, no. 5359, from type host collected in the Bay of Nh)nterey, California. Present status: Qiiadraceps pacifi- cus (Kellogg and Chapman), 1899. Nirmus peninsukiris Kellogg, 1899, Occ. Pap. Calif. Acad. Sci., vol. 6, p. 21. Type host: Phainopepla nitens = Phaim)pcpla iiitcns lepidii van Tyne. Two 5 syntypes, no. 5360, from type host col- lected in Baja California. Present status: Briielia peninsularis (Kellogg), 1899. OxyIi])Curus a1)domincdis Carriker, 1956, Florida Ent., vol. 39, p. 129. Type host: Dciulrortyx barhatus Gould. J paratype and 5 para- type, no. 5705, from type host collected June 12, 1947 by Robert J. Newman at Xilitla, San Luis Potosi, Mexico. Present status: Oxylipeurus abdominidis Carriker, 1956. Oxijlipeurus anmimensis Emerson and Elbel, 1957. Proc. Ent. Soc. Wash., 59, p. 238, figs. 2 and 13. Type host: Lophura diardi ( Bonaparte ) . One female paratype, no. 5720, collected Novem- ber 25, 1953 by Robert E. Elbel on Khao Sawan Mountain, Sieo, Loei, Thailand. Present status: Oxylipeurus aiummcnsis Emer- son and Elbel, 1957. Pectinopygus tordoffi Elbel and Emerson, 1956, Ent. News, vol. 67, p. 173. Type host: Pelecanus crythorhyiichus Gmelin. 2 j para- types and 2 paratype, no. 5706, from type host collected October, 1954 by Harrison Tordoff at Lawrence, Kansas; and j paratype from type host collected October, 1873 by F. H. Snow at Law- rence, Kansas. Present status: Pectinopysent status: Picicola rubinu C'arriker. 1956. Rallicola oriygometrae sid)porzanae Emerson, 1957, Proc Ent. Soc. Wash., vol. 59, p. 186. Type host: Porzana Carolina (Linnaeus). J holotype and 4^ allotype, no. 5668, horn t\i)c host collected 518 The University Science Bulletin May, 1909 in Douglas County, Kansas. Present status: Ralli- coJa ortijgometruc suhporzanae Emerson, 1957. Saemimdssoiiia haemustica Carriker, 1956, Florida Ent., vol. 39, p. 31. Type host: Limosa haemastica {l^mw^en^) . 5 paratype, no. 5711, from type host collected May 15, 1947 by George H. Lowrey at Lawrence, Kansas. Present status: Saemiinchsonia haemasiica Carriker, 1956. Saemwuhsonia lunneraUs aincricana Carriker, 1956, Florida Ent., vol. 39, p. 31. Type host: Niiinenius americanus americanus Bechstein. 5 paratype, no. 5712, from type host collected May 25, 1940 by Rollin Baker in Colorado County, Texas. Present status: Saemimdssunia humeralis americtina Carriker, 1956. Strigipliihis kctupae Emerson and Elbel, 1957. Proc. Biol. Soc. Wash., vol. 70, p. 196, figs. 2 and 7. Type host: Ketiipa zeijlon- ensis lesclienaiilt (Temminck). One male and one female para- types, no. 5721, from type host collected December 26, 1952 by Robert E. Elbel and H. G. Deignan at Huai Yang, Prachuap, Khiri Khan, Thailand. Present status: Striifiphiliis kctupae Emerson and Elbel, 1957. Strigiphihis inacrogenitalis Emerson and Elbel, 1957. Proc. Biol. Soc. Wash., vol. 70, p. 197, figs. 4 and 9. Type host: Glaucidium cuculoides briigeli (Parrot). One male and one female para- types, no. 5722, from type host collected January 5, 1952 by Robert E. Elbel at Ban Lat, Ban Kaeng, Khukhieo, Chaiyaphum, Thailand. Present status: StrigipJidus macro genitalis Emerson and Elbel, 1957. Strigiphihis otus Emerson, 1955, Proc. Ent. Soc. Wash., vol. 57, p. 241. Type host: Otus asio gihnani Swarth. 5 paratype, no. 5664, from type host collected April 24, 1937 by A. R. Phillip at Tucson, Arizona. Present status: Strigiphihis otus Emerson, 1955. Strigiphilus viridiciis Carriker, 1954, Florida Ent., vol. 37, p. 195. Type host: Ciccaba virgata centralis Griscom. Two 5 para- types, no. 5613, from type host collected March 12, 1947 by Marcella Newman at Xilitla, San Luis Potosi, Mexico. Present status: Strigiphilus viridictis Carriker, 1954. Sturnidoecus caligineus mexicanus Carriker, 1956, Florida Ent., vol. 39, p. 39. Type host: T urdus infuscatus (Lairesnaye). Two 5 and one nymph paratypes no. 5614, from type host collected May 27, 1947 by Robert J. Newman at Cerro Conejo, San Luis Potosi, Mexico. Present status: Sturnidoecus caligineus mexi- canus Carriker, 1956. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXIX] NovexMber 18, 1958 [No. 13 Mitotic Activity in Alliiim After Treatment With Chromosome-Breaking Chemicals BY R. L. Hancock * and M. Hancock Abstract: A rapid method of mitotic evaluation was used to study the effect of chromosome altering chemicals upon mitotic activity. These chemicals were found to cause radical supression of mitosis in Allium nieristematic cells. It was suggested that since heterochromatin, the nucleic acid regulator, is damaged by these chemicals and nucleic acid quantitatixe clianges effect mitotic rate that this might be the primary mechanism of inhibition of mitosis. INTRODUCTION Recently experiments have been performed to determine the action of radiomimetic chemicals on chromosomal aberrations. Ford ( 1948 ) found that nitrogen mustard caused more chromosomal anomahes in short chromosomes than in long ones. He observed that there were certain segments that seemed to have a greater frequency of these aberrations than others. Revell (1952) suggests that these "breaks" are possibly an interference of synthesis at cer- tain points on the chromosome. He found in Vicia fabo a great ma- jority of the breaks were at the site of the heterochromatic segment of the L chromosome when these roots were treated with nitrogen mustard, diepoxide, and tertiary butyl hydroperoxide. X-rays gave a more random distribution of breaks. McLeish (1952) found that maleic hydrazide suppresses mitosis and causes breaks in chromo- somes which have visible heterochromatin. Fahmy and Bird (1952) studied the effects of mustard gas, diepoxide, and triazine on the dis- tribution of chromosome breaks of salivary gland cells of Drosophila and found two loci of considerable vulnerability. Both segments were heterochromatic in t>'pe. McLeish (1955) found that maleic hydrazide formed ascentric fragments, which carried nucleolar or- * Medical Sfiidi-iit Fellow of the National FoundatiDii for Inf.mtilc Paralysis. (519) 520 The University Science Bulletin ganizers. Cells which contained 3,4, and 5 nucleolar organizers al- lowed fewer breakages than did normal cells with one or two nucleo- lar organizers. Smith and Lotfy (1955) observed the anaphase chromosomes of Vicia had aberrations in the heterochromatic re- gions that appeared to be effects during chromosomal duplication stages from treatment with propiolactone. However propiolactone caused the same frequency of chromosome breaks in Alliuin nieri- stematic tissue, which lacks the visible heterochromatin. Oehlkers ( 1952 ) found that the greatest frequency of breaks was in the SAT- chromosome and within this chromosome the points of greatest susceptibility were the achromatic region separating the satellite from the rest of the chromosome and the centromere. Kilman (1956) showed that 8-eth()xycaffeine induced chromosome struc- tural changes and caused a greater than random frequency of breaks in the attachment threads of satellites of Allium cepa. It was believed of significance to find the effects of these chromo- some breaking chemicals on mitosis. The procedure was made so as to give comparative results and not just an index of suppression or stimulation. METHODS AND MATERIALS Allium cepa bulbs of approximately similar sizes (3 cm. diameter) were placed, root primordia downward, in each of ten waxed paper drinking cups and ten milliliters of the solution to be tested was pipetted into them. A cardboard box was placed over the ten prep- arations and the bulbs were allowed to grow at room temperature. After 4S hours incubation five roots of comparable lengths were removed and placed in a single vial and treated with Feulgen stain using the method of Setterfield, Schreiber. and Woodard (1954). An ocular micrometer was used to obtain the distal two millimeters of the root which had been found to give a high percentage of meristematic cells. One thousand cells were individually observed from each bulb to give a total of ten thousand cells for each treat- ment that had been checked for mitotic activity. 1:2,3:4 — diepoxy butane', caffeine-, adenosine-, adenine-, adenylic acid-, napthalene acetic acid-, and benzidine'' were made as 10 "^M water solutions. Methyl-bis(beta-chloroethyl) amine hydrochlo- ride^ was made as 5 X 10 ^M; tri-chlortriathylamin-chlorhydrat"' as a 2 X 10 'M; maleic hydrazide- as an 8 X 10 '^M; and the dimethyl benz( alpha) antliracene-' was a saturated solution which gives a con- 1. Peinnsvilar Chcmrfseaich, Inc., Gainsvilk', Florida. 2. Xutritional Biochemical Co. 3. Eastman Ornanic Chemicals Distillation Products Industries, Rochester 3 N Y 4. Sharp and Dohme, West Point. Penn. 5. Nordmark-Werke GMBH-Hamburg. Germany. Mitotic: Actimty in Allium 521 centration of much less than 10 '^VI. In preparing the nitrogen mustard solutions, chemical goggles, a mask dipped with a ten percent thiosulfate solution, hood with blower, and rubber gloves were utilized as precautionary measures. A series of photographs was taken at 900X with the aid of an oil emmersion achromatic lens and Wratten filters G-15 and 1^-58. RESULTS Cells from roots grown in distilled water gave no apparent indi- cation of anomalies as evidenced by the 293 mitotic figures which were found in the ten thousand cells observed. Nitrogen mustard, the bis form, on the other hand, caused many anomalies. Only twenty-six mitotic figures were seen of which almost fifty percent were classed as abnormal appearing. ( See plate. ) Examination of cells treated with diepoxy butane showed only seventeen were in mitosis and again over fifty percent of these mitotic figures were in an abnormal state. Maleic acid hydrazide was found to inhibit mitosis almost as much as the tris nitrogen mustard but these cells showed no anomalies. Other substances were tried incidental to the problem. Naptha- lene acetic acid, a plant growth hormone, permitted growth in only three bulbs and the roots that did develop were thickened and short. However, the extent of the necrotic cells present in this preparation did not allow a percentage tabulation of mitotic figures. Several giant nuclei were seen (Fig. 1) in what appeared to be altered meristematic cells. They were approximately twenty times greater in area than normal nuclei as measured from enlarged photographs. The cortical cells were also hypertrophied. Purine derivatives were tried and all were found to be inhibitory. Only an occasional fragmented or undercharged chromosome was seen. Caficine presented many abnormalities. Large nucleoli were ob- served, and cells which c()ntain\'d undercharged (weakly staining), irregularly dispersed anaphase chromosomes were present. Many pycknotic. karyorrhectic, and necrotic cells were found. In one hundred and eight mitotic figines pres(Mit only eight were definitely believed to vary from the normal. Observation of the adenylic acid treated preparation was based upon only one thousand cells because the pectinase did not function proper!) in this case. Many binucleate cells were observed in the benzidine treated prepara- tion and endopolyploidy was thought to have occurred in several instances. While onK four abnormal mitotic figures were seen 522 The University Science Bulletin the number of cells in prophase were unusually high. Treatment with dimethyl benz-anthracene produced relatively large nuclei in many of the cells along with many in which the cytoplasmic volume had also greatly increased. It was estimated that over fifty percent of the mitotic figures were in telophase in this prepara- tion. In a statistical analysis of the raw data an analysis of variance was used to ascertain the significant differences in the mean effects of the treatments. Bartlett's test for homogeneity of variance within each treatment showed significant differences at the O.lVc level. Treatment number six was left out of this test because of its obvious variability. After a square root transformation and a subsequent Bartlett's test, a value of 21 was obtained for Chi square with nine degrees of freedom, which is not significant at the 1% level. From the results (see table) it is seen that all treatments except number Table 1. — Mitotic Activity (Each count expresses number of mitotic figures per thousand cells ) Chemical j'" treatmentXi'" plant 12 3 4 5 6 7 8 9 lO^iXij Control Distilled water 1 24 33 20 31 21 35 34 23 51 21 293 Heterochromatic breakers Nitrogen mustard (bis) 2 1 16 2 3 1 0 1 0 1 1 26 Diepoxybutane 3 2 3 0 3 112 2 2 1 17 Nitrogen mustard (tris) .. 4 6 9 2 3 16 7 2 4 6 4 59 Maleic hydrazide 5 2 12 5 10 4 6 5 0 5 4 53 Carcinogens Bendzidine 6 0 53 0 5 1 33 36 56 13 53 250 Dimethyl-benz anthracene 7 16 0 0 7 13 1 26 0 3 9 75 Purine derivatives Adenine 8 23 30 47 17 26 38 19 15 45 21 286 Adenosine 9 13 21 30 7 13 15 13 22 19 28 181 Adenylic acid 10 10 20 20 30 0 0 0 20 10 0 110 Caffeine 11 16 29 9 14 7 5 8 5 12 3 108 three have wide variability. When an analysis of variance is com- puted for the group (1, 2, 3, 4, 5, 7, 8, 9, 10, 11), there is shown a significant difference between treatments, (Fg.go = 14.98; P < .01). The same test when applied to treatments 2, 3, 4, and 5 as a group (the heterochromatic breakers) gave a significant difference be- tween treatments of this group, (Fs.gs = 42.19; P<.01). Also 8, 9, and 10 show wide variation, (F,,,, = 73.125; P<.01). Mitotic Activity in Allium 523 DISCUSSION The results (see table) show that chemicals which act on hetero- chromatin in other cells depress mitotic activity considerably in Allium ccpa cells. Diepoxybiitane caused the greatest reduction of mitosis and is also the most consistent in the breaking of hetero- chromatin of the chemicals used (Revell 1952). The bis form of nitrogen mustard was more active as a depressant to mitosis than the tris form. Maleic hydrazide and tris nitrogen mustard gave similar results. Benzidine acted very peculiarly. It depressed mitosis drasti- cally in four bulbs yet may have had a stimulating effect on others. The purine derivatives show depression as a gradient of their molecular weight with adenine giving almost no effect. The plant growth hormones were found to depress the cell division of the predominantly meristematic tissue samples, since they act pri- marily as a stimulus to differentiated cortical cells and not to meristematic cells. A major consideration is the correlation between chromosome aberrations and their influence on the role of mitosis. Nitrogen mustard can effect heterochromatin (Revell '52). Heterochromatin, in turn, controls the nucleic acid supply of the cell. ( Schultz 1947, Caspersson 1950). Rate of division can be explained by assuming a quantitative change in nucleic acid synthesis (Koller 1951). This then could be a mechanism for the necessary correlations between chromosome altering chemicals such as those used in this work and the suppressed mitotic effect. This mechanism could be termed "heterochromatic suppression." SUMMARY A recently developed rapid method of mitotic evaluation using pectinase was used to study the effect of groujss of chemicals, which effect cell systems, upon mitotic activity. Chemicals that previously have been shown to disrupt heterochromatic segments were found to depress mitotic activity extensively. A possible mechanism of action is suggested. Purine derivatives, a plant hormone, and car- cinogens were also evaluated for their effects on mitotic activity. 524 The University Science Bulletin ACKNOWLEDGEMENTS I am deepK indebted to the following persons: Dr. P. G. Roofe for his continuous help throughout the experiment. Dr. B. S. Wenger for reviewing tiie manuscript. Dr. W. E. McEwen for com- ments on the preparation and handling of the chemicals. Dr. R. N. Bradt tor tlie formulation of the statistics, Mr. Bradrick for the statistical tabulations, and Dr. I. L. Baird for instruction in photo- micrography and developing procedures. LITERATURE CITED CasperssOiN, T. O. 1950. Cell Growth and Cell Function. W. W. Norton & Co.. Inc., N. Y. Fahmy, O. G. and M. J. Bird 1952. Chromosome Breaks Among Recessive Lethals Induced by Chemical Mutagens in Dro.sophila melanogaster. Suppl. to Heredity, \()1. 6, pp. 149-159. FoKu, C. E. 1949. Chromosome Breakage in Nitrogen Mustard Treated Vicia faha liool-tip Cells. Proc. 8th Intern. Congr. Gen. Lund., p. 570. KiLMAN, B. 1956. Chromosome Breakage in Allitim by 8-ethoxyeaffeine and X-rays. Exp. Cell Res., vol. 8, pp. 345-368. KOLLER, P. C. 1951. The Experimental Modification of Nucleic Acid Systems in tlic Cell. Symp. Soc. Exp. Bio!., xol. 1, pp. 270-290. McLeish, J. 1952. The Action of Maleic Hvdrazide in Vicia. Suppl. to Hereditv, vol. 6, pp. 125-147. 1955. Radiation Sensitivity and the Mitotic Cycle in Vicia faha. Nature, vol. 175, pp. 890-891. Oehlkers, F. 1952. Chromosome Breaks Influenced by Chemicals. Suppl. to Heredity, vol. 6, pp. 95-105. Revell, S. H. 1952. Chromosome Breakage by X-rays and Hadiomimetic Substances in Vicia. Suppl. to Heredity, vol. 6, pp. 107-124. SCHULTZ, J. 1947. The Nature of Heterochromatin. Cold Spring Harbor S>uiposia Quant. Biol., vol. 12, p. 179. Setterfield, G., R. Schreiber, and J. Woodard 1954. Mitotic Frctiuency Determinations and Microphotometric Feulgen Dye Measurements in Hoot tips. Stain Tech., vol. 29, pp. 113-120. Smith, II. H. ;:nd T. .\. Lotfy 1955. EHects of Beta-propiolactone and Ceopryn on Chromosomes of Vicia and .Mlinm. Amer. J. of Bot.. vol. 42. pp. 750-758. Mitotic Activity in Allium 525 • o • Q « s 1 ^. • 0 Fig. 1. A giant nucleus in a cell treated witli napthalenc acetic atid. Feulgen stain. Inset shows normal size of nucleus. (Enlarged from 9()() X-) Fig. 2. A criss cross anaphase bridge in a cell treated u ith nitrogen mustard. Feulgen stain. ( Enlarged from 900 x . ) 27-2656 Publications of the University of Kansas Science Bulletin Kansas University Quarterly Volumes I to X are no longer available for distribution, since the supply is either ex- hausted or the numbers so reduced that they are kept for distribution in whole sets to libraries. University of Kansas Science Bulletin Volume I Nos. 1-4, weight 8 ounces. Nos. 5-9, weight, 12 ounces. Nos. 10-12, weight, 6 ounces. n Nos. 1-3, weight, 20 ounces. Nos. 4-9, weight, 11 ounces. Nos. 10-15. weight, 20 ounces. in Nos. 1-6, weight, 33 ounces. Nos. 7-10, weight, 25 otmces. rV Nos. 1-6, weight, 33 ounces. Nos. 7-20, weight, 28 ounces. V Nos. 1-11, weight, 33 ounces. Nos. 12-21, weight, 27 ounces. VI No. 1, weight, 27 oimces. Nos. 2-7, weight, 19 ounces. Vn Nos. 1-17, weight, 50 ounces. Vin Nos. 1-10, weight, 52 ounces. IX Nos. 1-21, weight, 54 ounces. X Nos. 1-15, weight, 17 oimces. XI No. 1, weight, 20 ounces. Xn Nos. 1-2, weight, 19 ounces. XIII Pt. I, Nos. 1-9, weight, 12 ounces. Pt. II, Nos. 10-15, weight, 10 ounces. XIV Nos. 1-21, weight, 34 ounces. XV Nos. 1-6, weight, 18 ounces. XVI Nos. 1-6, weight, 14 ounces. XVn Pt. I, No. 1, weight, 18 oimces. Pt. II, Nos. 2-7, weight, 8 ounces. XVni Nos. 1-13, weight, 38 ounces. XIX Pt. I, Nos. 1-7, weight, 6 oimces. Pt. II, Nos. 8-14, weight, 16 ounces. XX Pt. I, Nos. 1-6, weight, 11 ounces. Pt. II, Nos. 7-21, weight, 15 ounces. XXI Nos. 1-16, weight, 32 ounces. ^ XXII Nos. 1-18, weight, 32 ounces. XXIII No. 1, weight, 40 ounces. XXrV Nos. 1-21, weight, 38 ounces. XXV Nos. 1-22, weight, 43 ounces. XXVI Nos. 1-15, weight, 40 ounces. XXVII Pt. I, weight, 20 oimces. XXVIII Pt. I, weight, 20 ounces. Pt. II, weight, 20 ounces. XXIX Pt. I, weight, 20 ounces. Pt. II, weight, 20 ounces. XXX Pt. I, weight, 15 ounces. Pt. II, weight, 17 ounces. XXXI Pt. I, weight, 15 ounces. Pt. II, weight, 24 oimces. XXXII Issued as a single volume, weight, 45 ounces. XXXIII Pt. I, weight, 15 ounces. Pt. II, weight, 20 ounces. XXXIV Pt. I, weight, 30 ounces. Pt. II, weight, 25 ounces. XXXV Pt. I, weight, 50 ounces. Pt. II, weight, 50 ounces. R. Ill, weight. 20 ounces. XXXVI Pt. I, weight, 50 ounces. Pt. II, weight, 50 ounces. XXXVn Pt. I. weight, 50 ounces. Pt. 11, weight, 50 ounces. XXXVni Pt. I, weight, 60 ounces. Pt. II, weight, 30 ounces. Volume XXXIX is issued as a whole volume (t. ;., not in Parts I and II). The University of Kansas Science Bulletin will be sent in exchange for publications of like character, or may be sent on receipt of the amount of postage according to the weights mentioned above; or they may be sent express, charges collect. Where only single articles are desired, the separate should be requested, since separata of a large number of articles are available for distribution. Application for these should be made to Exchange Depart- ment, Watson Library, University of Kansas, LaviTence, Kansas. Date Due UP 2 ^ '&6 -F-£i 2 2 4383