: ” 2 " > ie ie & 2 * =<" et at eee as ~ > —_ - ——— UNIVERSITY OF KANSAS SCIENCE BULLETIN DEVOTED TO THE PUBLICATION OF THE RESULTS OF RESEARCH BY MEMBERS OF THE UNIVERSITY OF KANSAS UNIVERSITY OF KANSAS PUBLICATIONS LAWRENCE, DECEMBER 29, 1961 PRINTED BY JEAN M. NEIBARGER, STATE PRINTER TOPEKA, KANSAS 1961 Fas Scere a CARO ES) JeSyweil> “SJ — 28-5840 Contents of Volume XLII NUMBER PAGE 1. Weights of the Ventricular Walls of the Heart in the AGuleWee's, 08 4605. 25 Leena seh ee Homer B. Latimer, 3 2. Studies in Behavior and Phylogeny of Certain New World Jays>(Garralinae) Stay. . Se eeaeee as 2.0: J. W. Hardy, 13 3. Some Morphological and Functional Aspects of Certain 10. 1: Structures of the Middle Ear in Bats and Insectivores. O'Dell W. Hensen, 151 . Some New Species of Rhagovelia from the Philippines ( Veliidae, Heteroptera ). H. B. Hungerford and Ruichi Matsuda, 257 A Revision of the Bees of the Genus Melissodes in North and Central America. Part III. (Hymenoptera, Joy ts (Ve, | Ae Onan etekdtr ch dy Oe Wallace E. LaBerge, 282 The Cranefly Genus Dolichopeza in North America. George W. Byers, 665 . Summary of Fossil Microfloral Investigations in the Bevier, Weir-Pittsburg, Lower Williamsburg and Blue Mound Coals of Eastern Kansas. Sam L. VanLandingham, 925 . Observations on the Biology and Taxonomy of Flies Found Over Swarm Raids of Army Ants (Diptera: Ta- chinidae, Conopidae)............ Carl W. Rettenmyer, 993 An Application of Factor Analysis to Insect Behavior. Robert R. Sokal and Howell V. Daly, 1067 Factor Analytical Procedures in a Biological Model. Robert R. Sokal, Howell V. Daly, and F.. James Rohlf, 1099 The Bionomics of a Primitively Social Bee, Lasioglossum inconspicuum. ..Charles D. Michener and Alvaro Wille, 1123 THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vou. XLIT] DECEMBER 29, 1961 [No. 1 Weights of the Ventricular Walls of the Heart in the Adult Dog BY Homer B. LATIMER Department of Anatomy, University of Kansas, Lawrence Asstract: The body weight, the weights of the entire heart and of the two ventricular walls, from a heterogeneous laboratory series of 15 male and 31 female dogs, have been studied statistically. The entire heart as a percent- age of body weight and each ventricular wall as a percentage of total heart weight are presented. The ratios of the weight of the left ventricular wall to the right ventricular wall range from 1.80 in the females and also in the group of larger dogs to 1.84 in the smaller dogs and 1.87 in the males. These ratios are not statistically different and all of them fall well within the range of the reported ratios for the human heart. INTRODUCTION The ratios between the weights of the two ventricular walls have been a subject of investigation for over three quarters of a century, with most of the ratios based on the human heart. Miller in 1883 weighed and compared the right and left ventricular walls of the human heart, and his work has been followed by that of Lewis (714), Herrmann and Wilson (22), Jones (’53), Latimer (’53), Keen (55) and others. The weight of the entire heart has been studied in the fetal, postnatal and adult periods in man and in animals, but far less attention has been paid to the weights of the ventricular walls. The ratios between the weights of the ventricular walls in two inbred races of rabbits have been reported (Latimer and Sawin, 59) and these ratios are different, with the left wall significantly heavier, relative to the right wall, in the smaller and more active race. There is also a sex difference in these ratios of the ventricular walls in the larger race of rabbits but not in the small race. These differences in the ratios between the ventricular walls in the two fo “e's a x S/o wT (3) “~ > “& 3e < - =~, £ Maer \O ROG O3 had P A LIBR uae) aay SS SSM sy, 4 Tue Unrversiry SCIENCE BULLETIN races of rabbits suggest that similar differences might be found in the dog hearts. This report will show that although the ratios in male and female, small and large dogs do differ slightly, these dif- ferences are not significant. Also, all of these ratios fall within the range of the ratios reported for the human heart. MATERIALS AND METHODS The dogs from which these hearts were removed were a typical laboratory collection which had been used for only short experi- ments in the Physiology laboratory. There were 15 males and 31 females, all adults and ranging in body weight from 3.86 to 10.89 kilograms. The dogs were weighed on a platform balance sensitive to 0.25 pounds by the students as the initial part of the experiment, and these body weights in pounds were changed later to kilograms. The experiments never lasted over two hours and the dogs were sacrificed by additional anesthetic at the end of the experiment. The entire hearts, as soon as removed from the body, were placed in a moist chamber and dissected and weighed as soon as possible. All weights were made in the fresh condition and in the same man- ner as for the rabbit hearts. Slightly different methods of dividing the human hearts are re- ported, but these dog hearts were dissected by one person and using the same methods employed in the study of the guinea pig (Latimer, 52), human (Latimer, 53) and rabbit hearts (Latimer and Sawin, 59). The method of removing the two ventricular walls has been described in detail in the earlier report (Latimer and Sawin, 59) and only the general plan will be described here. The great vessels were severed close to their emergence from the heart. The free portion of the right ventricular wall was removed by an incision started by putting one tip of the scissors into the orifice of the pulmonary artery and carrying this incision toward the apex of the heart along the right side of the anterior longitudinal sulcus and close to the septal wall. This incision was carried around the apex of the heart and then along the posterior longitudinal sulcus as far as the coronary sulcus, following as closely as possible the junction of the interventricular septum and wall of right ventricle. Next the right ventricular wall was removed by making an incision just distal to the coronary sulcus and also distal to the attachments of the atrioventricular valves. Thus the right free wall was removed, leaving the atrioventricular valves attached to the remainder of the heart, and the papillary muscles attached to the free wall. The VENTRICLES OF Doc HEART 5 chordae tendineae were separated from the papillary muscles at their junction. The free wall of the left ventricle was separated in a similar manner. This was a little more difficult for the left wall tended to curve around as it entered the septum. Corrections in this dissection, if necessary, were made after the cavity of the left ventricle had been opened. Then the two atria were opened and all contained blood was washed from all of the cavities with tap water. Next the two ventricular walls and the remainder of the heart, consisting of the interventricular septum and the two atria, were carefully dried on paper toweling. The three parts were weighed all together in glass-stoppered weighing bottles to obtain the weight of the entire heart. Then each ventricular wall was weighed separately while the remaining ventricular wall was kept in a moist chamber. All weights of the heart and the two ven- tricular walls were made on an analytical balance sensitive to 0.1 mg. but the weights were recorded only to the nearest milligram. I wish to thank the Department of Physiology for permission to collect these dog hearts and the medical students who so kindly provided the body weights. TaBLE 1.—Weights, percentage weights and ratios of the right and left ventricular walls in entire group of adult dogs. All weights in grams except body weight in kilograms. Average and Coefficient standard deviation | of variation IBOCVAWeleN Gases as stone cae eave 6.69 = 1:53 22.90 [BICTIR SSE AS rei Cees Canin NE es ee 55.58 + 13.84 24.91 183: We. Saeed SS Ble he ES OMe On Oe a Mee Pa 292 eS B37 25). 02 HP Pawel a eet BL oh nad y,! PD Ni ES HOI 25.01 Rao lira Vig Wise Vid WV ciccins aires yo ate S220 IU 3748; EVE Ar the tOle bs) Wiebke shake dstacncsd eins oa 0.8438 + 0.158 18.87 NV EAMG a Tbe. Seiahaciclse oe faces Aee 220205 164 7.47 NEMEC VS ec/GRHeArt. fh casa ants onmaces cera ee ORAS =e nO 6.49 VENTRICULAR WEIGHTS IN ENTIRE GROUP OF DOGS The body weight, weight of the entire heart and the weights of right and left ventricular walls are shown for the entire group of dogs, of both sexes, in the first 4 lines of table 1. The last 4 lines show the ratio of the weight of the left to the right ventricular wall, the weight of the entire heart as a percentage of body weight and the weights of each ventricular wall as percentages of total heart weight. The second column contains the average weights, the ratio of the two ventricular walls and the percentage weights to- 6 THE UNIVERSITY SCIENCE BULLETIN gether with their standard deviations, and the last column, the co- efficients of variation. The weights in grams of the two ventricular walls are the most variable followed by the weight of the entire heart and then the body weight. The heart as a percentage of body weight is, as expected, less variable than its weight in grams. Body weight is generally more constant than the weights of the individual organs. The lowest coefficients of variation are for the weights of the two ventricular walls expressed as percentages of total heart weight. The left ventricular wall comprises a larger proportion of total heart weight than the right ventricle and it might be expected to be less variable when expressed as a percentage of total heart weight. The ratio of 1.82 between the weights of the two ventricular walls in these dogs is much like the ratios for selected human hearts (see Latimer and Sawin, ’59, table 3). SEX DIFFERENCES The body weight, the weights and percentage weights of the heart and the two ventricular walls, and the ratios between the ventricles are shown in panel A of table 2. The averages and standard deviations for the 15 male dogs are shown in the second column and the coefficients of variation, in the third column. Simi- lar data for the female dogs are found in the fourth and fifth columns, and the last column contains the “t” values of the sex differ- ences. This last column shows but one measurement, the percent- age weight of the entire heart with reference to body weight which is significantly greater at the 5% level in the male dogs. The body weight averages slightly heavier in the females and the weights in grams of the entire heart as well as both ventricular walls are slightly heavier in the male dogs and consequently it is to be ex- pected that the percentage weight of the heart would be larger in the males. This is in agreement with the report of Northup e¢ al. (57), which will be discussed in a later section. The least sex differences are in the body weight and in the weights of the two ventricular walls as percentages of total heart weight, with the percentage weight of the right ventricle manifest- ing the least sex difference of any of the measurements in panel A. Possibly the heterogeneity of this small group of dogs may have something to do with the lack of sex differences in these measure- ments, for the only significant sex difference is in the percentage weight of the entire heart. The differences in the ratios between VENTRICLES OF Doc HEART 7 the two ventricular walls are very slight, but the males do have slightly heavier left ventricular walls compared with the right walls. TABLE 2.—Weights, percentage weights and ratios of ventricular walls in male and female, and in large and in small dogs. Weights in grams except body weight in kilograms. Coefficient Coefficient Average and p Average and ° sag standard deviation oi vars standard deviation C8 are: ies Panel A 15 males 31 females Body. weight....----.. Holl 2s ilesto) 19.70 6ai2) == 165 24.48 0.24 18 (Srine SG AO OPES PO 59.60 + 14.74 24.72 SpiAiBy Es ej 24.64 1.37 TRERVE Wise te:s tide snetend de easy) 23) 8 yaets) 25.74 11.78 += 2.95 25.00 1.37 Ee SVE iotscorstevestisewis:.c/es 23.68 + 6.04 25.50 21.11 = 6.24 24.81 1.47 atio, lhe WE ANGERS Waive 187) = (0.206 10.98 1.80 = 0.204 11735 1.09 Heart % body weight..| 0.913 = 0.184 20.20 0.809 = 0.136 16.75 2.14 Rev We hearth... 22201 s == ele 22 5.02 222 03ie 8-85 8.30 0.04 Teo Vis Wie 7 heart; <2): Br ES AES 7.18 39.34 += 2.45 622) 0.53 Panel B 22 small dogs 24 large dogs Body weight.......... Sp Ale) ORS: 14.49 7.86 = 1.06 13.53 8.83 Pears ericcteycuss cceis. sche. AWfariay “ES 2)aK0) 19.05 62.75 =+ 13.98 22.28 4.27 RAV Wee ts cose bees 10.59 = 2.26 21.33 USjE7fil SS) Sie ls 22.98 3.83 x DVO iar scavensve.siere (or 6; ace 19.01 + 4.07 21.40 24.64 += 5.58 22.65 3.88 atio, Ibe Wa AE/ARS We Mitfoy 128455" 105239 12.95 1e80h == ORL7 9.82 0.62 Heart % body weight..| 0.887 += 0.132 14.90 0.802 = 0.175 21.83 1.84 R. V. Opheanterc ss. PP} OMY 5 PAUL! 9.19 21°86" "= 7 1.24 5.69 0.71 L. V. W. % heart..... 39.69 += 2.64 6.65 39.29 = 2.59 6.59 0.52 * The “t” values in panel A of 2.14 and above are significant at 5% and 2.98 and above are significant at 1% and similar “‘t’’ values in panel B are 2.08 and 2.83 respectively. DIFFERENCES DUE TO BODY SIZE All of the dogs, irrespective of sex, were grouped according to body weight in panel B, table 2. All dogs below average body weight were placed in the first group, and those above, in the sec- ond group and data, arranged as in panel A, are given for these two groups. In these two groups, selected for body size, one would expect that the “t” values of the differences in body weight and in the weight of the heart and its ventricles, would be significantly different, as shown in the first four lines of panel B. The ratios of the weights of the two ventricular walls and the three percentage weights, however, are not significantly different. Thus the size of the dog does not significantly affect the ratio of the weight of the left ventricular wall to that of the right wall. The percentages of total heart weight are not significantly affected by the body weight, but the heart tends to be slightly larger, as a percentage of body weight, in the smaller dogs. Although the ratios between the two ventricular walls are not 8 THE UNIVERSITY SCIENCE BULLETIN significantly different in the two panels of this table, yet the ratios are somewhat larger in the males and in the group of smaller dogs. The size of the dog, however, does not affect the relative weight of the heart nor its proportions as much as does the sex. These dogs were mongrels and possibly if they had represented two pure breeds of differing size and activity, similar to the rabbits, they might have shown differences in the proportions of the heart, espe- cially if the two breeds represented different degrees of activity and corresponding differences in muscular development. DISCUSSION The primary interest in this study is in the weights of the two ventricular walls rather than in the weight of the entire heart. However, a recent report on 346 adult dog hearts by Northup et al. (57) shows that the dog hearts are relatively larger in the males and also in the smaller dogs. The only significant sex difference in panel A, table 2 shows that the heart forms a larger percentage of body weight in the males than in the females, and the hearts, likewise, are relatively larger, although not significantly larger, in the smaller dogs (panel B). Slightly different methods of apportioning the weight of the in- terventricular septum have been employed in some of the studies on the human heart. It has been divided in different ways between the weights of the two ventricular walls (Lewis, "14 and Herrmann and Wilson, ’22); weighed entirely with the left ventricle (Keen, 559); or only the free ventricular walls have been removed for weighing and all of the septum left with the remainder of the heart. The interventricular septum does contribute to the function of both ventricles, but the weights of the free ventricular walls are more easily and exactly determined and they evidently give as accurate a measurement of the relative weights of the two ventricles as the methods which attempt to apportion the weight of the septum between the two ventricles. Especially in these smaller animal hearts, it would be very difficult to accurately apportion the septum into its right and left portions. The method of weighing only the two free ventricular walls seems to be the method of choice, espe- cially for these animal hearts, and it has been used for some of the studies of the human heart (Fulton et al., 52, Jones, 53 and Lati- mer, 53) and for all of the animal hearts. Table 3 of a previous report (Latimer and Sawin, 59) shows the ratios of the weights of the left to the right ventricular wall for a few human and animal hearts. The range of these ratios is from VENTRICLES OF Doc HEART 9 1.52 to 2.27. The ratios of the weights of the ventricular walls in carefully selected human hearts range from 1.74 to 2.00, and within this range fall the ratios for the entire group of this present series of dogs (1.82) and also the ratios for the groups of dogs as shown in table 2 (1.80, 1.84 and 1.87). The average of all of the ratios in this table of the preceding report is 1.92 or but 0.10 greater than the ratios for all of the dog hearts taken together (table 1). The relationships of the weights of the two ventricular walls in this random collection of dogs is very similar to that reported for the guinea pig heart, for the similar ratio for a series of 100 guinea pigs (Latimer, 52) is 1.96 or but slightly higher than that of these dogs. The only known data on the weights of the two ventricles in the dog are presented in a very brief report by Pembrey (’28). He lists the weights of the right and left ventricles in 5 puppies, one day old and all from one litter. Three puppies have heavier left ventricular walls, and in the other two the right ventricle is heavier. Averages of his weights show that the left ventricle weighs 85.4 and the right, 82.4 Cg. The sexes are not given nor the methods of dissection. Evidently these newborn puppy hearts had not at- tained the adult proportions of the two ventricles, and so far as is known the time of this change has not been determined. There is no significant sex difference in the ratios of the ven- tricular walls in these dogs although the ratio is slightly larger in the males. Unlike the dogs, the rabbits have relatively heavier left ventricular walls in the females. In the race of large rabbits the ratios are 1.69 for the females and 1.52 for the males, or signifi- cantly less in the males. The race of small rabbits however manifest no significant sex difference although the females have a slightly larger ratio. The minimum ratio in the table of the preceding report is 1.52 for the males of a race of large and inactive rabbits (race III) and the similar ratio, also in the males, in a race of small and more active rabbits, is 2.03 and this larger ratio is significantly larger than the ratio of 1.52 for the large males. Likewise in panel B, table 2 the ratio for the smaller dogs is slightly but not significantly larger than in the large dogs. In conclusion, these dogs do not show the significant differences in the weights of the two ventricular walls in the two sexes nor in the small and large dogs, which were so evident in the two races of rabbits. Possibly racial differences of body size or of activity are masked in this heterogeneous group of dogs. Differences in pro- 10 THE UNIVERSITY SCIENCE BULLETIN portions between the two ventricular walls similar to those reported for the rabbits might be found in groups of purebred dogs of dif- fering size and degrees of activity, but they are not evident in this present collection of mongrel dogs. SUMMARY Body weight, the weights of the entire heart and of the two ven- tricular walls are presented, together with the ratios between the weights of the two ventricular walls, the weights of the heart ex- pressed as a percentage of body weight and the weights of the ventricular walls as percentages of total heart weight. The weights of the ventricular walls and the total heart weight are the most variable, with the percentage weights of the two ventricular walls the least variable and the ratios between the weights of the ven- tricles slightly more variable. The ratio of the weight of the left ventricular wall to that of the right in all the dogs together is 1.82, which is well within the range of the published ratios for man and animals. The heart as a percentage of body weight is significantly larger in the males and this is the only significant sex difference in these data. The ratio of the ventricular walls is 1.87 in the males and 1.80 in the females. This difference is not significant. Again these dogs were divided, irrespective of sex, into a group of small, and a group of large dogs. Naturally, the observed weights are significantly different, but the ratios and all of the percentages give no evidence of significant differences. The ratios of the weights of the ventricular walls are 1.84 in the small dogs and 1.80 in the large dogs. LITERATURE CITED Futton, R. M., E. C. Hurcuinson and A. MorcGan Jones. 1952. Ventricular weight in cardiac hypertrophy. Brit. Heart J., 14: 413-420. HERRMANN, G. R., and F. H. Witson. 1922. Ventricular hypertrophy. A comparison of electrocardiographic and postmortem observations. Heart, 9: 91-147. Jones, R. S. 1953. The weight of the heart and its chambers in hypertensive cardio- vascular disease with and without failure. Circulation, 7: 357-369. KEEN, E. N. 1955. The postnatal development of the human cardiac ventricles. J. Anat., 89: 484-502. Latimer, H. B. 1952. Weights of the right and left ventricular walls in the guinea pig heart. Anat. Rec., 118: 247-252. VENTRICLES OF Doc HEART 11 1953. The weight and thickness of the ventricular walls in the human heart. Ibid., 117: 713-724. Latmer, H. B., and P. B. Sawin. 1959. Morphogenetic studies of the rabbit. XXIV. The weight and thick- ness of the ventricular walls in the rabbit heart. Ibid., 135:141-147. Lewis, T. 1914. Observations upon ventricular hypertrophy, with especial reference to preponderance of one or other chamber. Heart, 5: 367-403. MULLER, W. 1883. Die Massenverhiltnisse des Menschlichen Herzens. Hamburg und Leipzig. Leopold Voss. 220 pp. (cited by Scammon). Nortuup, D. W., E. J. VAN Liere and J. C. STIcKNey. 1957. The effect of age, sex, and body size on the heart weight-body weight ratio in the dog. Anat. Rec., 128: 411-418. PEMBREY, M. S. 1928. Weights of fetal and newly born animals. J. Physiol., 66: XI-XII (Proceedings ). ScamMmon, R. E. 1927. Studies on the growth and structure of the infant thorax. Radi- ology, 9: 89-103. 4 tegaall cok) 26 pO Bh 0 ee eh ee, Ce neaGratd bie corm _ 1 »s* poet Ty 7 a PEI TEL. “anal huss "es Lea a bin aly a ay? VILF WTR Ti rae % yeT eet fat ae AA. .! hear hk da ed gate veligh inhi bids hae ie ee jatavgsy ‘ ut te i i aise uy jh ve shied byawt fl, Anh oats tem piv >in ew Sibe i Th Pa tele PANY. if Vy i i. ee el ' is r ii file vat Vi a! P . ’ - soll ‘ ih ce : : ‘ J * itewtl lanai dered cat Saeco eres (ben yelvfore Vile ) + { Fi | ; Th y : t | 4 ' a . = f { td . hs "ai > { )'< s 4 ‘ fhe Lis haw re 1 Siam At iva a. | Tian - bys ear’) yUAL, . 4 | ‘ i x ss ie eh * Lad) rr ws ae, | ligt 7 hip A. OUR e eet a¢ at . g Za i f = 1 Ab wiht oN Tag mi a Ol he jeanne Kieg idae aproben! Sil Aaa wey acne BOR-TA Stet; seckeudta Sidi appre Se iar Gbarantt me ia onan wana 8 Gan Welt te tan Tee, » wali toAl Ro vices tah : An ‘tfivtia 7 , Sot) ol ah (e\/ lat a 1 5 oP eed vl oe = Sacre) ct Seiaey: Setiaadh AG aan 23 Balsenesthuildin ge 10.0862 ca Oe Ne sea Tedder, aad oe ee 27 Courtship: in) the: Mexican Jay» ..401..4 a4.n-be dems ote thee 30 Prenestinganocks) 3 swshs Kapsu tpgeiyear “plea e bo eee 30 Courtship, feeding, in the: wild... jaan. a2: Saaaceea eee 30 Courtshipsfeeding in. captivity) 304445 .. oc washct. dg. oe 31 The significance of courtship feeding and its relationship to social OLU ETH a2. Shield Lo aed are f opas os Sila, Ae ee 33 Nestbuilding in courtship. ..4.¢:..... «inch oats 4s. Secs al eee 35 Variation im HOCKS: © ies. cyg ds « Beayeraalncdagusees seald.« ba ee 37 Comparison of Courtship Habits and Pair-bonds .................. ——s- 38 NESTING: A GIIVITIES fw. (AP) ahi or ida kecconei lesen) « oe oro ae cen ae ee 40 Nestbuilding. an. the, BluewJay, .i:.... itiactocd: 4d4en4 aye Ae 40 Nestbuilding and Associated Behavior in the Scrub Jay ............ 42 Division, of duties and. the “Sentry-habit™ ...,.. ......)...." eee 42 Flightsand. flicht-displayy:.%)< .cacueemas sm) J setotie ace eee 44 TM erritOLy abn) Fs oer OR Ee Oh eae Uke EE eee 45 Ternitoriality, sing the; Blue, Jay cris. 14 dnc ode « ok . ee 45 “Lerritoriality aim the; MexicaniiJay,).. saeco «eee eee 49 Territoriality,.in, the Serb, Jay... 4.). i... nk .«d See 51 Ecological bases of reproductive sociality in Aphelocoma ........ 53 Copulation ,and: Related, Activities)... 5.40.1 gaanehes eee cee eee 57 Pre-copulatory: behavior inthe Blue Jay ...... . 0.02... .dpe eee 57 Copulatory and associated behavior in the Blue Jay ............ 58 Behavior in the Blue Jay During Egg-laying and Incubation ......... 59 Transition; from’ mestbuildimg,, ... -,<0.awas) > aie cyt ed ee 59 Barticipation Jf thes SEXES, escuus> Soul aetees altace oy rachis a 59 Behavior in the Mexican Jay During Egg-laying and Incubation ...... 62 Participation: Of: The SEXES acu ee. ngocs sos denned ns nado ods 62 Attentiveness). .e0k.05. da. Sherawat) ae. See ae See 63 Behavior at: the’ nest. .°". . dye V cates ct slctonnd le bn eee 63 Discussion and Summary of Incubatory Habits ................... 64 Carevor young by the Blue Jay . 6 O2%.). o...0 5. as eee 64 Barly nestling stage’ 2... See wilh ons 64 Srupies OF New Wortp JAys 15 PAGE Behavior after the first week of care of young .................. 66 EATS MACS UIMOmStAS CMAN IAC eRe errs a Rete eke 8 ore viet «ce release vores 67 pliora led ayed Ay Pere on hee Oi i Soha ea and 2 Le Cael A On ge a me ate 68 GareioL voune bythe Mexican Jiayer sa5 ve: Aakers. 68 Pianlynestlinestageuee panne eee eNews tt cg as Garsae seeps 69 Middle wmnestlingmstagey.. #45 el cans en een a tec sae eal or 70 iledolimoestace hegre Oe mee: © Crier Rage ok eT ge ig Siege eh | 70 Part II—NoNBREEDING BEHAVIOR BEHAVIOR: OEM © APs DEVE mW) VEINVA TI PASS ee iepine hs Wai Mntrnt tt tad) Faw either eee (2 Behaviorot Young Blue Jays in Captivity .. 02)... 0...es-eeew sn... 72 Stee etc eke. Pm ae ae URS SD Cid VE EA eho 72 CCC eee? APRON PR LCEO LAE DR RARE bl ONO Nie erie oes PURE» 3 73 REA Chon stomwWalere. its Magee et A ema hue ERPS Rin Hy Speer a ee Preening, bill-wiping, and head-scratching ................ EO Wocalizationse iar is pvnit. Meee: Lien 8 Daa enhanc Mabteteg tye ioe | rp 78 Relationships with one another and other birds ............... 79 Nelationships* with enemies sh. eke. 28, Ee 80 Relationships with humans: Fees tes eee a ee, 81 Sleeping terse Slee fee CE A LUTE eis Foy: opener ee iia ND 43 81 Behayior-sof a* Youne Steller’s Jayin Captivity =<... sue ss. 81 Behavior of Young Mexican Jays in Captivity ....0...0..0........... 82 BMOCKING SANDS HORAGING BEHAVIOR == i pyar ni Lesa se aes ee =. 84 Flocking, Movements, and Foraging in the Blue Jay .............. 84 PAuitiimnaralien Ockimn ory 214: WETIRAL Se 215, tah ech Rea ee ee 84 NAVA gay2 Gio(e) Sika\oa kent Seo On a ee a ee en oe ae oe ee 86 POOCEStOUIAC Gras atte aay ake ek eae 4 hea 2) tg ee et ay be 88 Flocking, Movements, and Foraging in the Mexican Jay ............. 89 Pretivatics OLrtne gHOCK) oe 2. ce 2 eta Biter see ae eee OO OO “SCORIIO es ge. Gah are Meee Se pel SB. Les enn eee ES PP rr 95 ANTERSPECIFICMNELATIONSEHEPS (ie hin ioe Uke Rea RE Paytid oh IR Uy 94 Interspecific’ Relationships of the’ Blueway .4)..)025.5. 0. ute e. 94 Relationships jwathimon=predators) Seen S ieees euch ess aed eyes 94 RElaHOnShIpSe wlth PREGALOES, (Eko te cited Pte pl uesscd cad ik ents oe ct aw) OF Interspecific Relationships of the Mexican Jay .................... 96 Relationships) with'non-predators: 5...) 8) 2. 4s. ee ee ke 96 elationshipsiwithy predators | 5.56.) oe oe 2 Ee, 1 te 97 Interspecific Relationships of the’Serub/Jay 2.) .3.. 205. .4 in. 99 SOCIATE ORDERING CAPTIVE NNIEXTGCANTIAYSa eer ee ati: neo) eiunis ue Wee 101 Establishment: of; the)Reck-order ~ 4). cys ca :.:...:.7.7. .2> gee 114 The Case for the Primitiveness of Territoriality ................. 115 Indications from analogy) 7% 155.23... . oats en. 115 Indications ‘within’ Aphelocomaut Aue —i)-¢26%. «2 02 eee 116 Sociality in the family Corvidae ack. 94h 1s -pa~ee ee 17 DISCUSSION, |< Sa (a6 cca ost RUSE ee Se: Oe eee 117 The Case for the Primitiveness of more Social Habits ........... 118 Behavioral and distributional evidence within Aphelocoma ..... 118 Correlation of sociality and age dimorphism of bill color ....... 118 Possible social adaptiveness of age dimorphism in bill color .... 121 Evolutionary significance of bill color and sociality ........... 122 Intra- and interspecific variability in duration of parti-coloredness.. 123 Relationships within the “Ornate line” of New World Jays ..... 126 Additional comments on Cyanolyc@ ..u.45.)- dence sae ee 129 Other Behavioral and Morphological Traits in New World Jays .... 129 Voealizations: 6 te 4ideidec.}. ai cealic 4a we kee ae Se 129 Trends; in, ‘characteristies: of plumage, 5 «..2c...n1st «6055 - 131 Some inter- and intrageneric considerations ................. 133 Discussion: and ,Gonclusionss ..:3 scas:06 1. bea; a4 eee 136 SUMINGARY SDS BAe clits, sheets lS gle 509 oe ca 25S of PONG se A 141 146 ETrERATURE) CIrmeDeh esac oso cncs Sided ee oes dhos « eee INTRODUCTION The studies reported in this paper were conducted with the aim of contributing to the understanding of behavior in New World jays. Particular attention was given to study of social behavior of members of the genera Cyanocitta and Aphelocoma. An attempt was also made to contribute to the concept of phylogeny of these jays through studies of their behavior. There are but two extensive, published accounts concerning behavior of a New World jay, that of Amadon (1944a) on the Florida Scrub Jay (A. c. coerulescens) and that of Gross (1949) on the Mexican Jay (A. ultramarina arizonae). Rand (1937) raised fledgling Blue Jays (C. cristata) and wrote on their behavior. Scattered notes on behavior in jays of the genus Aphelocoma are to be found in Pitelka’s (1951) systematic study of that group. Bent’s (1946) summaries of the life histories of the North Ameri- can species of jays reveal the paucity of information on behavior of jays. Studies of five species of jays were conducted in the course of my work. These species are the Blue Jay (Cyanocitta cristata bromia), Stupies OF NEw Wokrtp JAyYs 7 Steller’s Jay (C. stelleri macrolopha), the Scrub Jay (Aphelocoma coerulescens woodhouseii), the Mexican Jay (A. ultramarina arizonae), and the Magpie-Jay (Calocitta formosa colliei). Be- cause of convenience, most information was gathered concerning the Blue and Mexican jays. The former is common in eastern Kansas where I reside, and the latter is common in the oak wood- lands at the Southwestern Research Station of the American Museum of Natural History, where I spent two months of the summer of 1956 and a week in early spring of 1958. Limited studies were conducted of early phases of the nesting cycle of Scrub Jays in the Sandia Mountains near Albuquerque, New Mexico, and of wintering populations of that species in Oklahoma. Casual obser- vations of Steller’s Jay were made in the Chiricahua Mountains and elsewhere. In addition, observations of Magpie-Jays in Sonora and Sinaloa, Mexico, have provided information that may aid in better understanding the relationships between Aphelocoma and Cyanocitta and between these two genera and other New World jays. The study here reported was divided into two major parts, (1) observation of jays in the wild and (2) observation of captive jays. Captives were confined in outdoor cages at Lawrence, Kansas, and for most of my stay at the Southwestern Research Station birds were kept in small indoor and outdoor cages for various purposes. It does not seem probable that the behavior patterns common to most of the members of the family Corvidae occur fortuitously in sO many species; these patterns may, therefore, be considered homologous characters, correlated with a common ancestry for members of the family. I am concerned primarily with considera- tions of the adaptive value of the behavior of Aphelocoma and Cyanocitta, some characteristics of which may be of systematic importance at the generic level. METHODS OF STUDY An attempt has been made to make all my observations compara- tive. “Knowing” what to look for in a species can, of course, be a hindrance, in that one may see only what he expects to see and remain unaware of habits outside the realm of his previous experi- ence. I have attempted to guard against this pitfall, while com- paring habits of birds that I was observing with those of species that I had studied previously. In the wild, jays are easily observed, their nests found with ease compared to those of many other passerines, and, although 18 Tue UNIVERSITY SCIENCE BULLETIN the birds are often quiet in the breeding season, they are not shy. Therefore, blinds are ordinarily neither necessary nor particularly useful. A blind was employed twice at nest-sites near pathways that persons used; jays at these nests were furtive and seemed to behave unnaturally when I was not concealed. As a rule, I was able to observe what I considered to be normal behavior by remain- ing motionless at the maximum distance that allowed me to watch whatever activity was taking place. Considerable movement on my part was sometimes necessary, because the range of individual birds or flocks is typically great even within a short space of time embracing one activity. When nesting activities of a population of jays were observed, the nests were given designations consisting of N plus a numeral; thus, N-1, N-2, N-3. These and similar designations will be used in this paper where appropriate. Captive jays were marked with colored plastic bands. The approximate age of the captives, sex, breeding condition, and state of molt were noted upon capture. Adults and juveniles were confined in outdoor cages six feet in height by nine feet wide and twelve feet long. One end of each of these cages was protected from rain and wind by canvas siding and woody shrubs placed inside. Roosting and other perching places, food and water were provided. Food consisted of nuts, acorns, sunflower seeds, millet, milo, melon seeds, bread, occasionally fruits such as grapes, and a generous supply of raw ground horse meat or commercial dog food. A regular supply of raw meat insured longer average life of the captives. Jays captured as adults do not adjust well to captivity. Mexican Jays adapted much more readily to cage life than did Blue Jays. In fact, studies of the latter species in captivity were for the most part limited to observations of fledglings. No adult jays ever became so accustomed to captivity that they were not excited by the presence of people. Captives were handled as little as possible because of their excitability. I first made an attempt to “tame” them by remaining quietly in the cage for periods of several hours, but this method did not prove effective. Successful obser- vations could be made only from a distance or, with caution, from a blind. A small blind was erected at the side of each cage. Even with this arrangement, the captives remained uneasy long after the observer had entered the blind. Young jays of three species were raised from middle or late nestling stage to fully grown, flying stage (usually through the Strupies oF NEw Wortp Jays 19 postjuvenal molt). The three species were the Blue, Steller’s and Mexican jays. These birds were retained in small cages approxi- mately two feet long by one foot wide, by three feet tall. The birds were frequently allowed the freedom of closed rooms. Experimentation was not a major part of the study, but oppor- tunities to study the social and individual reactions of jays to artificially imposed stimuli, resembling certain natural stimuli, were utilized. Such experiments were carried out with both wild and captive, adult and young birds. ACKNOWLEDGMENTS I wish to extend thanks to the following persons and organizations who aided me in various ways in carrying out the present work. Especial thanks are due Profs. Richard F. Johnston, Harrison B. Tordoff, and E. Raymond Hall, who advised and directed me while I was a student at the University of Kansas and supervised the com- pletion of research and writing of this dissertation. Prof. Johnston was my principal advisor in the final stages of research and the preparation of the manuscript. I am likewise indebted to Profs. Henry S. Fitch and Ronald L. McGregor of the University of Kansas, who offered valuable advice in the preparation of the manuscript, to Dr. Robert M. Mengel for the excellent drawings of jays found herein, and to Prof. A. Byron Leonard for assistance in photographic work on figures and tables. Of great aid in the section dealing with phylogeny in this paper was the opportunity to examine certain specimens of jays from Mexico, South America, and Florida, which specimens were generously provided by Prof. Frank A. Pitelka of the Museum of Vertebrate Zoology, University of California, Berke- ley. I am mindful of invaluable grants-in-aid received from the Frank M. Chapman Memorial Fund of the American Museum of Natural History and the Kansas Academy of Science. Dr. Dean Amadon, chairman of the Chapman Fund, Dr. Mont Cazier, director of the Southwestern Research Station, and John Anderson, of the latter institution, made it possible for me to do research in Arizona. PART I.—BREEDING BEHAVIOR CourtsHie BEHAVIOR Early Courtship in the Blue Jay Warm days in middle and late February usually are correlated with at least temporary abandonment of the winter habits in the Blue Jay and the assumption of a peculiar kind of flocking that seems to be in part socially and in part reproductively motivated. 20, Tue UNIverRSITY SCIENCE BULLETIN Such flocking is a phenomenon transitional from nonbreeding to breeding phases of the annual cycle. These early courtship flocks are composed of members of the wintering population and typically consist of from three to six birds. Collecting individuals from these groups, which persist into early May, indicates that they are composed principally of adult males. There is usually one female in each flock, pursued by the males. The behavior of the sexes in such groups is distinct enough that one can predict accurately the sex and age of each bird in the flock. In middle and late May when nesting activity is well underway, there are still a few groups of jays flying about, performing in pre- cisely the same manner as the adults in early spring. However, these birds are first-year birds and seem to be nonbreeders. They engage in courtship activities into early summer. The basis for stating that the first-year birds are nonbreeding individuals is that nesting records of Blue Jays at Lawrence have been those of adults (second-year or older) with only one exception; in 1957, I recorded the successful nesting of a pair in which the female was a first-year bird. According to Hickey (1952:119) pairs of Blue Jays both members of which were first-year birds have been observed in the act of copulation. Laskey (1958:213) records a first-year female that nested, using the previous year’s nest of its parents and laying five eggs. The bird nested later in the year than adults. Probably the first-year birds that nest are among those which, like many adults, do not migrate. My record of the first-year female nesting was among the earliest nesting records of the 1957 season at Lawrence; the nest was begun in April at about the time that migrant Blue Jays are beginning to arrive in this region. The young in this nest were about a week old on May 25—a date when flocks of first-year birds were flying about the area in courtship flocks. As Bent (1946:34) has noted, there usually seems to be in court- ship flocks one bird, the female (?), which governs the activity of the group in courtship. When she flies, the others follow her in straggling formation; if she remains still, they do likewise; if, upon alighting, she hops upward among the branches of a tree, the others follow, continually attempting to be nearest her and occasionally even attacking each other. In the early group courtship ritual, bobbing and flight displays are frequent. Bobbing is a distinctive courtship display of this and other species of jays. In bobbing, a bird moves its entire body quickly up and down several times, accomplishing this by flexing Stupies OF NEw Wokrtp JAys 21 and extending the legs at the ankle joint. As Tyler (in Bent, 1946:34) notes, this action is not bowing, since the whole bird moves up and down in approximately the same plane. In group display, all birds bob at once, although not in synchrony. One male seems to initiate such bobbing each time, after which the activity endures among all birds in the group for no more than three seconds. This is an erect display; the tail is fanned, the crest and head erected, and the body held rigidly upright with feathers closely appressed. The birds sit quietly or give the medium intensity jayer call discussed below. They remain in this tense state until sud- denly one gives the “pumphandle” call, cleeop cleeop, or sometimes a rattling call, brrrrr, and bobs vigorously up and down. The other jays in the group then also bob. The cleeop note is bell-like and not unlike a note of the Cowbird ( Molothrus ater) in its liquid quality. It is used in low intensity phases of predator intimidation, in intimidation of fellow jays, as in the present instance, and in territorial skirmishes. It is nearly always accompanied by the bobbing display. The brrrrr call resembles the sound made by a stick caused to vibrate against the edge of a table. It is uttered with the bill slightly open and seems to originate deep in the upper chest cavity. This call seems to be associated with anxiety, but I am not able definitely to categorize it. I have heard the call at all seasons of the year. It is given by both sexes and is often accom- panied by bobbing, as in these courtship flocks. What may be an homologous call in the Brown Jay (Psilorhinus morio) is produced in the perhaps unique furcular pouch of the interclavicular air sac (Sutton and Gilbert, 1942:161). No such pouch is apparent in the Blue Jay. Figure 1 depicts a typical group of Blue Jays in courtship. In flight from one station to another, and as the birds move about through the branches of a tree, the call, wheedle-eee, is often uttered, expressing the anxiety of the males. This “squeaky-gate” call resembles also the sound made by a clothesline on a pulley. It seems always to be associated with uncertainty or suppressed excitement. It is never given in association with the flight display. Flight display (Fig. 3) is frequently engaged in by all members of the group (except when the female takes the lead in flight; then, the other members of the flock forsake display in order quickly to give chase). In full display, wings are held stiffly extended and widely spread, the head directed straight forward (it often is held bill downward otherwise), and tail widely spread. Both the jaaay and the jayer are given by the jays when in flight display. oD, THE UNIVERSITY SCIENCE BULLETIN Both calls are assembly calls that closely resemble the alarm calls; they are, however, less intensely given and do not connote danger. On April 21, 1955, I observed three Blue Jays engaged in display that was slightly different from the above. In this case, the birds alighted in a row on a limb. The two outside birds began bobbing while the bird in the center remained nearly motionless. As the two bobbed, they gave the cleeop cleeop call. After about three min- utes, the birds flew off, giving the medium intensity jaaay. Thus, in some instances not all birds in the group bob. Occasionally, I was not able to detect a leader among the birds, that is, one toward which bobbing was directed or which caused flight of the group. In the majority of cases, however, all birds bob, and one bird seems to be a leader. The area covered by a courtship flock of Bue Jays is about one- quarter mile in diameter. In this area the flock moves freely, sometimes covering the distance from side to side in a single flight and never remaining in one place for over five minutes. Such groups are in evidence only in the morning hours, from daybreak until, usually, about 9:30 a.m., or slightly later on cloudy days. Once a group forms for the morning activity, it seems not to change in size or membership. Group courtship probably functions to synchronize the reproduc- tive states of the several jays in the flock; the birds gradually change from winter habits to intense courtship. Although I have watched these flocks carefully, I have never observed birds in them that I knew to be paired or members of pairs. Occasionally I have seen mated birds investigate these flocks, but they never join in the display. Pair formation also seems to occur in these courtship flocks. Along with bobbing and flight display, there are aggressive move- ments and sometimes fighting among the males in their attempts to be near the female and prevent the other males from being near her. Through aggression and sometimes fighting, one male eliminates the others (or perseveres the longest) and becomes mated with the female member of the flock. The eliminated males then join or form other flocks, group courtship continuing until most adult members of the population are paired. In the subsequent phases of courtship next to be described, it will be noted that later courting groups usually consist of two or three birds—evidence that the number of competing males in a given flock gradually decreases, Stupres OF NEw Wor tp Jays 23 I am convinced that there are birds, probably always adults, that never take part in activities of courtship in a given year. These birds are already mated, or in the late stages of becoming so. Laskey (1958:212) records one pair that remained mated for three years, did not migrate, and remained together winter and summer. In late February, I once observed two birds in company; one was feeding the other, as is characteristic of late phases of courtship. This occurred when little courtship of any kind had been seen in the area. Among first-year birds, courtship activity is restricted, and many of these birds, at least those that are migratory, probably do not nest. Their reproductive states never advance beyond early court- ship. The gonads of these birds are small but not so small as the gonads of Blue Jays in winter in this region. Late Courtship in the Blue Jay Courtship feeding—As previously indicated, late courtship ac- tivities seem to evolve gradually from earlier courtship by flocks. The number of males in the flock is reduced to one, the latter seem- ingly eliminating his competitors by intimidation—aggressive beha- vior and fighting in some cases. The most characteristic activity of late courtship groups is feeding of the female by a male. In rare instances I have observed the stage of courtship imme- diately preceding the final stage, and characterized by the presence of two males and one female in association. When three birds compose a courting group, two are in closer association with each other than is either with the third individual. One of these two regularly feeds the other, and the third is always a “bystander” (but obviously a member of the group). Thus, courtship feeding enters the courtship ritual before the group of jays is reduced to one male and one female. Courtship feeding in groups of three or two birds is the same. The birds involved in late courtship usually feed calmly for as much as 20 minutes, showing no signs of courting. This low in- tensity of display persists in courtship feeding until the time of false nestbuilding. The only notes uttered, typically, in these periods are low conversational ones, kut or kuet, which are low intensity communication signals between the birds. These signals may be uttered softly in series, kut kut kut, when a bird flies or when it approaches its fellow with food. In feeding, the male approaches the female, which may accept the offering with hardlv any display save for a slight crouching. Mildly intense forms of THE UNIVERSITY SCIENCE BULLETIN Fic. 1—A group of Blue Jays in courtship display. Uppermost bird is the female. Fic. 2.—A Blue Jay in aggressive display toward a nonpredator. Strupres oF NEw Wor tp Jays 25 this ritual may involve on the part of the female slight flutterings of the wings and the uttering of the solicitation call—kueu kueu kueukueukueu. This call is given by the female, not only when begging for food, but also at the time of copulation. The call might better be termed a submission vocalization at these times, but it differs in being slightly more intense. The male may utter the same notes as he responds to the female’s begging or submissive behavior. In the display, the female crouches (rarely does not crouch ) with wings slightly drooped, tail and head tilted upward, crest lowered, and bill either slightly open or closed. At the height of this display, the posture changes slightly; the wings are quivered, and the tail is slightly spread (Fig. 5). When there are only two birds involved in this form of courtship, it is obvious that they are male and female, and in most instances adult birds. In the instances where three birds are in company, the third participates only in a casual way, and I am not certain of its status. There has never been any indication that the third bird was another male competing for the female, or whether it was an adult or first-year bird. Perhaps the role of this bird is related to “helpers at the nest” to be discussed under the Mexican Jay. The area of activity of these latter courting groups seems to be about the same size as that of the earlier flocks in courtship, al- though the birds may remain in one place for longer periods of time and seldom fly long distances. There is no anxiety and no chasing of the female; probably this results in the more casual, slower movements of the group. The courtship activity of this phase of the breeding cycle is func- tional in that the ritual actually involves one bird feeding another bird, a habit that will be maintained through all the later phases of the cycle. In nestbuilding, the male will begin by presenting sticks to the female in precisely the same way he presents food; the female’s begging behavior will be the same as before. In lay- ing and incubation, the female will be the only one occupying the nest; the male will feed the female there. After the young hatch, they assume the begging role; both parents will present them with food. Thus, courtship feeding and begging rituals are intimately tied to functional reproductive behavior, sharing what seems to be the same motor patterns, and closely related stimuli and re- sponses. Whereas in the earlier phase of courtship, the groups of birds remain together for only a few hours in the mornings, the two or 26 THE UNIVERSITY SCIENCE BULLETIN Fic. 3 (upper).—A Blue Jay in the Display Flight of courtship. Fic. 4 (lower).—A Blue Jay in aggressive, investigatory posture toward a predator or strange object. Stup1ieEs OF NEw Wortp Jays 27 three birds associating in the phase of courtship feeding remain together constantly whether courting or not. Moreover, courtship feeding is not limited to any particular time of day, occurring with seemingly equal frequency throughout the daylight hours. False nestbuilding—tThe final stage in the courtship of Blue Jays involves a phenomenon heretofore seemingly unreported. ‘This is the building of a false, or symbolic nest immediately prior to the building of the true nest, or sometimes contemporaneously with its early stages. In eastern Kansas, this phase of courtship becomes evident in middle and late April. Like the forms of courtship previously dis- cussed, false nestbuilding occurs mainly in early morning. If the activity is not observed closely over a period of several hours each day for several days, it might be mistaken for true nestbuilding. The habit has been exhibited by every pair of Blue Jays I was able to follow through to building of the true nest. False nest- building arises directly from courtship feeding, involving somewhat similar stimuli and motor responses, and subsequently evolves to nestbuilding, possessing similarities also to that habit. Courtship feeding continues in false nestbuilding behavior, but the male, in addition to feeding the female, also gives sticks to his mate. These sticks in all cases are broken from trees, often near the false nest-site, but sometimes as much as 100 yards away. The sticks are never picked up from the ground. Often, food and twig presentation by the male are alternated, but at the height of inten- sity of the behavior, more twigs may be presented than food. The female crouches on the false nest-site; she may occasionally leave for brief intervals but always returns to it to accept food or a twig from the male. If the male is absent for more than a few minutes, the female utters a mewing cry not unlike that of the Catbird (Dumetella carolinensis) until her mate returns. This call may be correlated specifically with interruption of a sequence of events com- posing a courtship ritual at a time in the sequence when stimulation of the female is great; the call may, thus, be comparable to displace- ment motor responses expressed by animals frustrated in a tendency or drive. Whether the call serves to “summon” or “solicit” the mate is not certain. The female receives twigs and food while in the solicitation display posture (Fig. 5), with wings extended slightly and quivering. She utters begging notes, kueu kueu, during the presentation. Accepting twigs, the female makes an attempt to ar- range them beneath her on the limb. If twigs are dropped in the 28 THE UNIVERSITY SCIENCE BULLETIN process of arrangement, the birds do not attempt to retrieve them. Sticks may accumulate at the false nest-site, but the accumulation never takes the shape or stability of a true nest. Moreover, a false nest never becomes the true nest of its builders, although it is some- times used as a base for a true nest by another pair of jays. Only once have I observed more than two birds seemingly in- volved in false nestbuilding at the same site. One of these three, attended and pursued by the other two, flew into the tree contain- ing a small platform of sticks, went to this false nest and crouched on it. The other two jays remained nearby, and all three were silent. After three minutes one of the two bystanders flew away. Immedi- ately afterwards, the crouched bird hopped up and began uttering the rattling br’r’r’rr call, at the same time bobbing vigorously. The bird (presumably the female) then began hopping upward through the tree, twisting from side to side, with crest raised. The remain- ing bird followed, duplicating the other’s motion. Both were silent and soon departed. A week later only two birds were active at this same false nest. The involvement of three birds in false nest- building is probably rare, but its occasional occurrence is to be expected as a holdover from the feeding phase of courtship. False nestbuilding leads to true nestbuilding, and for a day or so a pair may indulge in both activities alternately. In such instances the material for the true nest is gathered from the ground by both birds for a few minutes; the activity then switches to the false nest. As before, the female crouches at the site and the male brings sticks broken from trees. Some exceptions to the general order of events in false nestbuild- ing may occur. Rarely, the female may be absent from the false nest-site when the male arrives with a stick. In such instances, the male may attempt to place the stick on the site and afterward may crouch there in the manner of the female. The male then may hop away when the female returns, but if he does not the female may go through the action of feeding the male or attempting to do so. The latter in such an instance does not solicit the presentation and exhibits no definite reaction of any kind to the offering by the female. It is generally thought that solicitation display by one bird stimulates another bird to make a presentation. The behavior described above indicates that in false nestbuilding either sex reacts positively to another bird crouched at the false nest-site by coming to that site and giving the presentation display. The latter display in turn is stimulus for a begging response by the female but elicits no positive Fic. 6.—A female Blue Jay in the PreroDU ator display and copulatory pos- ture—a submissive display. 30 THe UNIVERSITY SCIENCE BULLETIN response by the male. Only the male’s failure to give solicitation display prevents courtship behavior from being reversed—all other positions and reactions of the two birds necessary for successful presentation and acceptance seem possible by both sexes. Either bird may occupy the stimulating position, either may react by flying to the crouched bird, and either may give a presentation display. Courtship in the Mexican Jay My observations of the prenesting habits of this species were made in a week spent in the Chiricahua Mountains in early April, 1958. In this period, I found Mexican Jays occupied with several phases of the breeding cycle. Some were not yet nesting, others were building nests, one pair had eggs, and another had young in the nest. Prenesting flocks—The majority of the birds were engaged in activities seemingly without counterparts in the other species here considered in detail. The birds moved about in flocks of four to 20 individuals, precisely as they do in summer. These flocks were not courtship groups like those in the Blue Jay; little display, posturing, or aggressive movements occurred among members of the flock, or were other activities of courtship discernible. Most of the flocks were small, having between four and eight birds. One of these may have been a family group; this flock consisted of approximately five birds, of which I was never able to find more than two, fully black- billed, adults. In this species, a parti-colored bill is characteristic of most races (including arizonae) during a varying number of months or years before the black-billed condition is attained (Pitelka, 1951:317). Whether attainment of a black bill is exactly correlated with sexual maturation is not certain. Birds with parti-colored bills were neither members of pairs that constructed nests nor were they in any way intimately associated with nesting activities, as were the black-billed birds. The parti-color condition of the bill in two Mexican Jays that I kept in captivity from summer 1956, to spring 1958, did not change much in this time; the condition may not be correlated with sexual maturation. The testes of one of these captives measured 15 x 8 mm. on April 30, 1958. Courtship feeding in the wild—In the five days in early April, I did not observe one instance of courtship feeding in Mexican Jays. On only two occasions, did I observe any approach toward this behavior. A small flock of perhaps four birds was feeding along a road bank on the ground. One of these birds, perched in a low Srupres OF NEw Wor.p Jays 31 tree, suddenly flew down near another jay on the ground and gave a low intensity begging display. In this display, it barely quivered its wings and extended its neck in an arched manner. The bill of the bird was not open, nor was any call given. It slowly circled the other jay, which simply turned to face the circling bird; no feeding occurred. The display lasted for less than five seconds, after which both birds flew off with the other two birds in the flock. Both birds involved in the display were adults, with fully black bills. In the other instance of possible courtship feeding, a flock of 12 jays was moving about in a woodland, feeding quietly. When I approached, they grew excited, and six of them flew into the top- most dead limbs of a tree; for a few moments they perched quietly. At least eight, and possibly ten, of the entire flock’s members pos- sessed parti-colored bills. One of these individuals moved close to a black-billed bird, and their bills touched. It seemed that the bird with the parti-colored bill inserted its bill into the gape of the black-billed bird. No special posture was associated with this incident. The contact lasted for approximately two seconds. The bird with the parti-colored bill flapped its wings and jerked its body up and down during the contact. It is possible that this was ag- gressive or combative behavior, for Mexican Jays in captivity dis- play a combative habit in which one bird grasps the bill or some other part of another bird. This habit is discussed under Social Order in captive Mexican Jays. Gross (1949) does not seem to have observed courtship feeding in the wild in Mexican Jays prior to nestbuilding. He did see (1949:244) one feed another sitting on an uncompleted nest. Courtship feeding in captivity—Studies of captive Mexican Jays indicate that courtship feeding may be a habit as well-developed in this species as in other jays, regardless of my failure definitely to observe it in the wild. Mexican Jays placed in capitivity in July, 1956, and maintained in an outdoor cage at Lawrence from August, 1956, until May, 1958, first displayed courtship feeding behavior in April, 1957. Of the captive Mexican Jays, three were males and one was a female. Two of the males were black-billed, but one had a con- siderable amount of white at the angle of the gape. The female had a slight trace of white on the bill. Figure 7 illustrates the heads of these four captives. It will be convenient to refer to the indi- vidual captives by letter symbols that were used to designate them in the study and referred to the color of the band or bands around 32 THE UNIVERSITY SCIENCE BULLETIN the tarsi of each bird. The fully adult birds were color-banded red on the right leg (RR), and white on the left leg (WL), respec- tively. The remaining birds were banded red on the left leg (RL), the female, and white on both legs (WW). Rate of courtship feeding in captive Mexican Jays varies greatly and occurs mainly in the morning. At its greatest frequency, court- ship feeding may average once every five minutes. It never occurs at such a rate for more than about one-half hour. Approximately one-half the time the bird receiving the food does not eat it immedi- ately but stores it in a crevice. Afterward the bird returns to these caches and eats part of the stored food. Sometimes after eating the food immediately upon receiving it, the receiver goes to stored food and eats part of it too, but this is rare. Feeding in each case is accomplished with little display. The recipient sometimes quivers its wings slightly and usually crouches in taking the food. The call given by the recipient, or, rarely, by the feeder, is kwa kwa kwa—a low intensity vocalization given with the bill closed. In a slightly more intense display, the kwa notes may be preceded by a short ree, hence, ree kwa kwa kwa. A conversational note may also be uttered in connection with feeding: rook or ruk, resembling the kut or kuet notes of Blue Jays. No display or special posture accompanies this note. For a week after courtship feeding began in the captive birds, I saw feeding take place only between two of the four birds. The feeder in this ritual did not otherwise seem to be clearly the dominant one among the four birds, although at this time my information was limited. The recipient was at the bottom of the order, however. The two non-courting birds were unconcerned with the courtship taking place. One of these noncourting birds remained relatively solitary throughout its life in captivity. The other eventually assumed the position of feeder in the courtship ritual, to the exclusion of the previous feeder. This change in the right to feed in the courtship ritual was seemingly an expression of an exchange of positions of the two feeders in the social hier- archy and did not arise directly from competition between the two birds for this feeding right alone. When this change occurred, the ritual of courtship feeding gradually became more intense, involving greater vocalization and more definite display. However, less actual feeding seemed to take place in the ritual. In courtship feeding, the recipient follows the feeder about wherever it goes. The recipient lowers its head near the feeder Stupies OF NEw Wokr tp Jays 33 as the latter picks up seeds. The recipient also pecks weakly at the bill of the feeder. This pecking at first elicits no response from the feeder but if continued for several seconds provokes a threat display (see discussion of sociality, p. 101) and, subsequently, a strong peck. The recipient may peck back but not vigorously. When the feeder is opening a seed, the recipient peers over the feeder’s shoulder and then attempts to take the food. The feeder may then give a small particle to the recipient, but this never happens when the latter is vigorously attempting to take the food. The recipient may then attempt to return the food to the feeder, which seems never to accept it. After this display, the recipient usually stores or eats the food in a normal manner. _ Fic. 7.—Heads of captive Mexican Jays. The parti-colored bill is evident in two of these birds. Upper left, WL, male; upper right, RR, male; lower left, RL, female; lower right, WW, male. Courtship feeding declines in frequency approximately the first week after its initiation but continues sparingly through most of the spring. The significance of courtship feeding and its relationship to social order.—Mexican Jays with parti-colored bills engage in courtship feeding activity, a fact that indicates some first-year birds may reach breeding condition in the wild. It will be recalled, how- ever, that Blue Jays engaged in courtship activities in May and 2—5840 34 THE UNIVERSITY SCIENCE BULLETIN June are first-year birds and are probably nonbreeding individuals with only slightly enlarged gonads. I have mentioned previously (p. 30) that the testes of one of the captive Mexican Jays with parti-colored bill reached a size indicative of breeding condition (15x8 mm.). In the captive Mexican Jays, a female (RL) habitually fed a male (WW), reversing the usual relationship in courtship feeding of wild corvids. Such reversal indicates that the factors initiating the behavior are more complex than the term “courtship feeding” can indicate. In such a highly social form as the Mexican Jay of Arizona, the necessarily frequent interactions of individuals seemingly re- quire the establishment of a social hierarchy if the birds are to remain together in a peaceful flock. Thus, in interaction of two individuals, the quick establishment of one as dominant, the other as subordinate is important. The basis of dominance of one bird to another is superior strength; that is, in the initial competitive interaction between two individuals, superior strength in fighting establishes the victor as dominant over his opponent. But in the interests of conservation of energy, preservation of individuals other than dominants as members of the breeding population, and to a certain extent the facilitation of relatively peaceful (noncom- bative) survival of the social group, fighting is impractical. It has been superseded by ritualized postures and aggressive movements involving brief contact or none and accomplishing the same pur- poses as fighting following the initial interaction. Unnatural conditions created by captivity may have been enough to alter the behavior of captives. Yet, the appearance only in captive birds of reversed “courtship feeding” does not reduce the significance of this unnatural behavior. Courtship feeding, which occurs only in the breeding season, may be an expression of dominance of a male to a female, al- though it is not a method of maintaining dominance. Of course, in order for courtship feeding to be considered an expression of dominance in wild birds, it must be assumed that males dominate females at this time of year in all activities. Observations bear out this supposition. There are two categories of self-sufficient individuals of the Mexican Jay—adults (black-billed) and younger birds (parti-colored bill, exclusive of juveniles and nestlings). In the social order of captive birds, black-billed birds dominate birds having parti-colored bills. Birds with a small amount of Stupres OF NEw Wokr tp Jays 35 white on the bill dominate individuals having a large amount of white on the bill. In a given population, amount of white on the bill (degree of parti-coloredness) may be considered a general index of age. Thus, age influences social order, older birds dom- inating younger birds. Since the captive bird fed by two others was also dominated by them, domination and feeding may be correlated and are possibly related expressions of the same attitudes, the feeding action being an expression of dominance modified by sexual tendency. In my captive Mexican Jays, the youngest male was fed by two other birds, one an older female (a slight trace of white on the bill), and the other a fully adult male. It seems unlikely that the two older captives were reacting toward the young male in the role of par- ent, since the latter was at least nearly a year old. Age, then, seems to be more important than sex in determining social order and determines dominance when the two factors conflict. It may be further concluded that old females do not feed younger males in the courting manner in the wild because conflict between the factors of sex and age are reduced. Reduction of this conflict may be correlated with lack of confinement and the presence in the flocks of aggressive, adult males; both factors might discourage adult females from feeding young males. Nestbuilding in courtship—In the Mexican Jay, nestbuilding seems to be an integral part of courtship, and may be more com- parable to false nestbuilding of the Blue Jay. Scott (1886:82) was perhaps the first to publish evidence of the epigamic nature of nestbuilding in this species, although he did not recognize it as part of courtship. Scott discovered these jays mating in late February and found a completed nest on March 16. No eggs were deposited in this nest until April 1. In the period from March 16 to April 1, an adult frequently sat on the nest and also built another nest within a few feet of the first one. Gross (1949:242-6) gives an excellent account of nestbuilding by Mexican Jays; his observations differ slightly from my own. He describes how all the members of a flock visit the nest-site pe- riodically, with one or two of them carrying sticks for the nest. No more than two birds at a time brought material to the nest-site, but other members of the flock stood close by and occasionally flew down to arrange the nest material others had placed there. The visits of the flock to the nest were made as often as five times per hour. The birds made no pretense at secretiveness around the nest, 36 THE UNIVERSITY SCIENCE BULLETIN all coming with loud calls and perching around the nest while mate- rial was added. The flock included both black-billed and parti-color billed birds, although the latter were not seen to help build the nest. The main body of the nest was composed of sticks broken from trees and was lined with rootlets and horsehair. Finally, Gross noted that false nests were built by the flock, and that even the true nests were often completed long before eggs were laid in them. One flock spent nine days building a false nest and then constructed the true nest fifty feet from it. In another instance, a nest was seem- ingly completed thirteen days before the first egg was laid, although a jay was occasionally seen sitting on this nest in the interim. In the Chiricahua Mountains, I discovered two completed or nearly completed nests on April 1 and 2 and made daily observa- tions of them through the morning of April 6. No eggs were pres- ent in either of these nests, and there were times when I was, er- roneously, almost convinced that these nests were inactive. A flock of approximately a dozen birds moved about throughout the day in the vicinity of each nest. The activities of the birds were mainly social, not having to do with breeding. Yet, in the morning hours the flock would move near the nest several times, sitting in and around the tree containing the nest. At these times, two adult birds would leave the flock and go to the nest-site. Typically, the others maintained a distance of ten to 50 feet. One of the two adults then sat on the nest and moved about as if molding the cavity of the nest; the other bird perched close by. On one occasion, the sitting bird left the nest and returned with fine grasses (this activity was also observed at another nest). The other adult usually did not sit on the nest, but examined it closely, sometimes perched on its rim, and occasionally settled tentatively on it. The birds sometimes pulled at twigs near the nest, as if trying to add these to it. These twigs were never added to the nest, although I noted, as did Gross (1949:243), that most of the twigs comprising the nest seemed to have been broken from trees by the birds. The bird that most often sat on the nest usually remained on it for less than five minutes, but sometimes remained for as long as 10 minutes. The birds showed little concern for my presence, and in fact they were noisy in the vicinity of the nest at all times. Vocalizations in- cluded assembly calls, reek reek or ruik reek ruik reek. I found no definite evidence of false nestbuilding in this species as did Gross (1949:242-6), although one flock of jays under ob- servation were active periodically in two adjacent trees each of Srupries OF New Wortp JAys 37 which contained a partially constructed nest. The birds actually visited only one of these two nests, however; the other may have been a false nest of this flock of jays, since it did not seem to be an old nest from the previous year. In one of the above flocks, most of the birds were adults. The other contained only two black-billed birds (those that came to the nest). From the latter flock I shot a member possessing a parti- colored bill. It was a female with ovary slightly enlarged. Adults collected at other nests on the same day showed enlarged gonads. Flocking at the time of nestbuilding, and the fact that the nest- building birds are active members of these flocks are indicative of the strong social bonds in the species. These bonds remain strong throughout the remainder of the breeding cycle; members of the nesting pair participate in care of the young. Variation in flocks—The flocks associated with nests vary in structure. Of three such groups under observation in April, one consisted of six black-billed birds. Two nests of this flock were located within 150 feet of each other in a grove of oaks in a gully on a hillside. On April 5, one nest contained four eggs, the other contained young birds about a week and a half old. In two hours of observation near these nests, I found only two black-billed birds in the vicinity of each. I collected the birds at the nest containing four eggs, and still no other birds appeared there. A second flock consisted of approximately a dozen birds; four or five were black-billed. Only one nest was found in the area of this group. The nest was still under construction. The third flock con- sisted of about eight birds, two of which were black-billed. These birds were associated with the only active nest found in their area. It may be significant that the flock consisting wholly of black- billed birds was much farther along in the breeding cycle than the other two flocks, members of which had parti-colored bills. In most other species of passerine birds, adults begin the breeding cycle before the younger birds; thus, it is reasonable to expect a consid- erable shortening of the time of preliminary activities of breeding— activities such as pair formation and courtship—in flocks composed of adult jays, resulting in relatively earlier breeding. This, of course, is tenable only if birds with parti-colored bills are sexually immature or later to mature. It is possible that a flock might become com- posed entirely of old birds through loss of subadult members, failure to have a successful nesting season several years in a row, or other chance factors. 38 THE UNIVERSITY SCIENCE BULLETIN Comparison of Courtship Habits and Pair-bonds The differences in the patterns of courtship of jays under con- sideration in this paper seem to be direct expressions of the differ- ences in sociality of the several species. The Mexican Jay remains in a highly social flock throughout the period of courtship. If there is at this time competitive, aggressive display between males, it is certainly not common; the flock as a whole does not engage in specific displays and aggression. Within these social flocks some method of pair-formation exists; courtship feeding occurs, and periodic ritualized behavior connected with nestbuilding is unques- tionably important as a late phase of courtship. But the flock’s social bonds are strong; the pair remains in the flock. Or, perhaps better, the flock remains with the pair, being influenced so much by the reproductive behavior of the nesting birds that it forms a social “circle” around them. In any event, the epigamic gestalt of the Mexican Jay is a primary result of the prominent social disposition of the species; whether such sociality facilitates or stimulates breed- ing behavior, as Richard F. Johnston (verbal communication) has suggested, is not known. Pairs in fact carry out breeding activities without attention from other individuals, but it seems likely that reproductive success is greater when the pairs are members of flocks because of the frequency of the habit. Blue Jays cease to flock midway through courtship in most cases. Flocks of Blue Jays in spring may be termed epigamic flocks; they are concerned almost solely with early courtship. In the late phases of courtship of the Blue Jay, false nestbuilding occurs, but this happens after complete dissociation of the pair from other members of the flock. Behavior in the Mexican Jay involves building of true and false nests while the pair retains full membership in the flock and utilization of the true nest as well as the false nest in the court- ship ritual. I have recorded only one instance of there being more than two Blue Jays in association during false nestbuilding. In this instance three birds came to the tree containing a false nest and one went to the structure and sat on it. Thus, although this stage of courtship in the Blue Jay is rarely reached until all but one male and one female are eliminated from active roles, there seems to be a tendency for “extra” birds to be inquisitive about, or to associate, however loosely, in late, intimate phases of courtship. It seems likely, as indicated previously, that there is survival value inherent in the presence of a group of nonbreeding birds living with and assisting the breeding pair in the Mexican Jay. Stupies OF NEw Wokr tp Jays 39 In one pair of Blue Jays observed in this study, the male disap- peared in early incubation and the female acquired a new mate before the young were hatched; the nesting then was successfully completed. The second male associated with this nest was a first- year bird. This may indicate that there are nonbreeding males interested in or at least aware of nests in their foraging areas; these non-breeding birds do not appear at the nest where both members of the pair are active but appear when the male is lost. That the acquisition of the new mate by the above female Blue Jay had survival value is probable. There may well be a common basis for such an acquisition and the phenomenon of helpers at the nest, indicating a degree of relationship between Blue Jays and Mexican Jays. Acquisition of a new mate after nesting has begun is seemingly rare, in passerines at least, but there is ample precedent for this phenomenon in corvids. Shannon (1958:401-2) mentions seven published accounts of late replacement of mates (all females it should be noted) in the Magpie (Pica pica); replacement may occur in that species after the young are hatched, and a new mate may be acquired within seven hours of the loss of the first one. Thus, although there is a diverse social phenomenon surrounding the pair-bond in different corvids, there seems to be an especial, possibly adaptive, flexibility in the relationships of adults and first- year birds, males and females, common at least to Mexican and Blue Jays and Magpies, among New World garrulines. Displays and postures associated with courtship behavior in the Blue Jay show greater ritualization, specialization, and _ intensity compared with homologous patterns observed by me in the Mexi- can Jay. Similarly, the variety and complexity of calls is greater in Blue Jays than in Mexican Jays. In short, the complexity of behavior concerning pair formation and maintenance seems _ to vary inversely with degree of sociality of the two species. The Scrub Jay is the least social of the forms under discussion. Consequently, the dissociation of members of flocks occurs earlier in the breeding cycle, and although my information is limited, it seems that the process of courtship and pair-formation in the species is even less associated with activities of the flock than is true in the Blue Jay. In late March and early April in the Sandia Mountains of New Mexico, the weakly formed winter flocks of Scrub Jays have completely dissociated, nestbuilding has begun, and courtship feeding is a frequent activity of the pairs. 40 Tue UNIVERSITY SCIENCE BULLETIN The Scrub Jay in Florida, Aphelocoma c. coerulescens, is an exception socially among races of the species for which breeding habits are known. Therefore, information concerning epigamic behavior in that race may not be applicable to the species as a whole. Nonetheless, this is the only race of the Scrub Jay concern- ing which I have found a statement about courtship. Bent (1946: 79) states that S. A. Grimes of Jacksonville, Florida, is of the opinion that in the Florida race the birds remain paired through- out the year. Bent concludes from this and the lack of other obser- vations of courtship in Scrub Jays that courtship does not occur in the species. It may instead be an indication, I think, that court- ship does not involve highly overt, complex behavior, as in the Blue Jay. Either explanation is tentative; Bent (1946:34) as a matter of fact concludes “. . . that courtship is not a conspicu- ous feature . . . of the Blue Jays behavior because “A survey of the literature brings little to light . . .” concerning courtship in the species. Nevertheless, Bent does continue by documenting several examples of courtship in the Blue Jay. NESTING ACTIVITIES Nestbuilding in the Blue Jay Both the male and female Blue Jay usually participate in building the nest. In some pairs, both male and female participate about equally in bringing material to the nest, but in others the female does most or all of the work, the male following her about, occa- sionally picking up a piece of string, large stick, or brightly colored piece of paper. The male's effective contribution to construction of the nest may thus be negligible; he may carry fully as much material to the nest-site as does the female or at least make as many visits to the nest. But the male is usually awkward at nestbuilding, and frequently the material that he transports is eventually dropped to the ground, unless the female is present to prevent this. Nest- building occurs primarily in the morning hours, and, as mentioned previously, may be alternated with false nestbuilding activity. I have never seen more than two birds building a single nest. Thus, it seems probable that normally the last remnants of groups with three or more members disappear with the commencement of activities in breeding of a non-courtship nature. Maintenance of the pair-bond, to be sure, involves the same behavioral patterns that are present in courtship: Feeding of the female by the male continues, StupiEs OF New Wokrtp JAys 4} strange jays are chased away by the male, and the members of the pair remain together almost constantly. In the construction of one nest in spring, 1957, the male assisted little; while the female gathered nest material, he usually perched in a conspicuous position nearby. It seemed that the male served as a “lookout” for danger to his mate. The male’s function as look- out seems to be more prevalent in the Scrub Jay, under which I have described the habit more thoroughly. I might have over- looked this habit in the Blue Jay in which it occurs uncommonly had it not been for my observations of it as a functional pattern in the Scrub Jay. Material for the nest is gathered by Blue Jays over a wide area around the nest-site. This area probably corresponds to the home range of a pair. I watched one pair build a nest, for which much of the material was gathered approximately one quarter mile from the site. The nest was located in a tall pine surrounded for some distance by a lawn relatively clear of dead grasses, sticks, paper, and other items commonly used in building nests. When such materials are available close by, the jays seldom fly farther to secure them. Blue Jays will use paper, or other more unusual materials for nest construction; with an ample supply of some item such as crepe paper available, a great part of the nest may be constructed of it. I knew of one nest in a tree near a picnic ground that was visible from a considerable distance because its base was composed largely of pieces of paper napkins. Several behavioral rituals are connected with nest-building, al- though none are confined to that phase of the breeding cycle. They include the solicitation display by the female near or at the nest, feeding of the female by the male, and presentation of nest material to the female by the male. The last of these is rare, the male usually placing the nest material directly on the nest. The birds are relatively quiet while building and remain less vocal through the remainder of the breeding cycle until the young leave the nest. Calls most frequently uttered are those associated with the above rituals—kueu kueu kueu and the conversational notes, kut or kuet. Other calls may be given, but only when the nestbuilding activities are interrupted by activities completely apart from breeding be- havior. Nestbuilding seems never to take more than five days. Usually by the second day of construction, the nest has attained considerable 42 Tue UNrversiry SCIENCE BULLETIN form. It is on the second or third day that the most intense activity in construction takes place. Through the first day and part of the second, time is frequently divided between the false and true nests. By the third day, only occasional visits, if any, are made to the false nest-site. In one pair observed in the first day of nest construction, each bird brought material to the nest eight times in 60 minutes; the female’s efforts contributed the most. The birds worked closely to- gether, on two occasions coming to the nest at the same time with material from the same place. At the beginning of this hour, they made visits to the false nest, where the female perched quietly. The male fed her there on two occasions, but never did so at the true nest during its construction. At another nest on the second day of construction, the female brought material 11 times, the male eight times, in 60 minutes. The male was, again, somewhat less efficient in his work. Both birds frequently sat on the nest after bringing sticks, shaping the cavity with their breasts. The male several times deposited ma- terials beside the female while she was on the nest. Twice he fed her at the nest and once away from the nest. Construction by this pair was begun on April 26 (the last day that false nestbuilding by the two jays was observed) and was completed on April 29, when the birds were bringing fine rootlets and fibers to the nest in the morning. The female followed her visits to the nests with periods of sitting on it, once for seven minutes. Later that day she sat on the nest frequently, and no further nestbuilding activities were observed. The contribution to nestbuilding by members of a pair does not seem to vary much from day to day. That is, if the male contributes to nestbuilding the first day by visiting the nest with material 40 percent as often as does the female, he will do so in about the same proportion on the following days. Nestbuilding and Associated Behavior in the Scrub Jay I observed nestbuilding by a pair of Scrub Jays in the Sandia Mountains of central New Mexico in April, 1956. The activity was interspersed with courtship feeding, territorial displays, and other activities. Division of duties and the “Sentry-Habit.’—I discovered the nest (N-1; see figs. 8 and 13) when it was roughly half completed. An alternation of duties was being practiced by the pair. Both birds remained together constantly. One would go to the ground to StuprEs OF NEw Wor tp Jays 43 secure material such as grasses and sticks. Meanwhile, its mate stationed itself as a “sentry” in a prominent position (Fig. 9), usually the top of a juniper tree, nearby. Occasionally the sentry deserted its post for a few minutes at a time, but always returned in time to accompany its mate to the nest. Sometimes the sentry “anticipated” the flight of its mate to the nest and arrived there first. While its mate arranged the new materials in the nest, the sentry remained close by and not necessarily in a prominent posi- tion. Once the material had been incorporated into the nest, the bird standing by flew to the ground and gathered material while the other assumed a perch high above. Only occasionally the same bird gathered material twice in succession. When the birds flew a distance of more than a few hundred yards for material, both gathered some before either returned to the nest. The sentry is so termed here because it seemingly serves as a look-out for possible danger while its mate is occupied in gathering nesting material. On one occasion when a Red-tailed Hawk ( Buteo jamaicensis) soared over a sentry, it set up a cry of kwesh kwesh kwesh kwesh, flew quickly to its mate on the ground, and with her flitted off through the undergrowth. A bird standing on the ground, in this open, oak-juniper scrub and concentrating on gathering nest material or food is more vulnerable to predation than a bird under- taking the same activity in a woodland with a heavy canopy of branches and leaves overhead. The survival-value of the presence of a sentry seems definite. From the hillside where I was making observations, I found that I could locate other territories of jays by the presence of sentries. The sentry-habit is not confined to the nestbuilding pericd; a bird gathering food in a conspicuous place may also be accompanied by a sentry. The kwesh call mentioned above is one of the most frequent given by this species of jay. The call is usually uttered in flight, and, as with the jayer call of the Blue Jay, it is a low-intensity alarm call that alerts other jays but does not cause them to assemble. Courtship feeding is interspersed with nestbuilding. A bird may fly up to its mate and in spite of holding grasses in its bill, seem to feed the other bird, after which the nest material is carried to the nest. In courtship feeding observed in the nestbuilding phase, there is notably little display by either member of a pair. The feeding is accomplished quickly, with slight or no fluttering of the wings by the female. Usually no vocalizations are given, but occa- sionally a suppressed begging call is uttered by both birds or by 44 THe UNIVERSITY SCIENCE BULLETIN the female. This call, greer greer, is given with the bill closed and is similar in sound and function to the kueu kueu call of the Blue Jay. Associated with nestbuilding, and probably other phases of the life cycle, are two other characteristics that have counterparts in the Blue Jay. These are the rattling brrrrr’ call and the bobbing display. The call is indistinguishable from that of the Blue Jay but is given much more frequently, although often in similar situations of seeming anxiety. It seems more easily elicited from the Scrub Jay. Ihave heard it after courtship feeding, and when the birds are alarmed by a human too near the nest. It is often associated with bobbing, which, as in one type of bobbing of the Blue Jay, seems to be an intention movement of flight. In fact, in the present species, I have not seen it used in association with courtship. (In the Blue Jay, when it is used in courtship it is accentuated to the point that the bird may actually leave the perch in the activity. It seems in other words to be ritualized and exaggerated in intraspecific in- timidation.) The bobbing that I observed in the Scrub Jay re- sembles, then, the flight-intention movement of the Blue Jay, not the bobbing of that species in courtship. Flight and flight-display—tThe Scrub Jay of the subspecies wood- houseii is short-winged in comparison to the other species of jays herein discussed. Its flight is thus more labored than that of Mexi- can, Blue, and Steller’s jays. Immediately upon launching into the air, woodhouseii resembles in flight the thrashers (Toxostoma), the path of flight being up and down, the tail flitting from side to side. In the launching phase, a peculiar sound is made by the wings, resembling the sound made by the wings of a Mourning Dove (Zenaidura macroura) in flight, but somewhat lower in frequency. If the jay flies a long distance, it eventually moves more easily; it may alternately glide and flap its wings. Flight caused by a human that is too close to the bird is a jerky, “thrasher-like” movement for short distances from perch to perch, where bobbing and the brrrrr call are given. This intermittent flight may be a kind of display, although it is not at all like flight display of the Blue Jay. The weak flight of this jay, at least in the launching phase or on short flights, may be another reason for the assumption by the bird of a prominent, usually high, perch when remaining at rest for more than a few seconds. The high position affords the bird an easy, swift take-off, albeit it downward at an angle into the undergrowth. Stupies oF NEw Wortp Jays 45 Territoriality Territoriality in the Blue Jay—Territoriality is exhibited by the Blue Jay for the first time in the nestbuilding phase of the breeding cycle. Although this phenomenon is not strongly expressed in the Blue Jay, it is nevertheless true territoriality. Its expression is oe fi f ee Lg ‘ Bee Fic. 8.—Tree containing nest of a pair of Scrub Jays (N-1) in the Sandia Mountains, New Mexico. Fic. 9.—An adult Scrub Jay acting as a “Sentry” while its mate gathers nest material from the ground. 46 THE UNIVERSITY SCIENCE BULLETIN strongest from the time of nestbuilding until the young leave the nest, but territory is not defended at any time against other jays not in the same general phase of the breeding cycle as the resident pair. Thus, jays incubating eggs or feeding young in the nest will allow another pair with their young out of the nest to wander into the territory—even into the nest tree—in the course of care for these flying young. A strange jay entering a territory may be greeted by one of the resident birds. The resident flies to the interloper and alights nearby. If the strange bird does not flee, it may give an erect display. If the bird is on the ground, the head is raised and held rigidly still, with bill forward toward the resident individual. The belly feathers are fluffed, the body held low, the back swayed, the tail held low and slightly spread, and the crest fully erected (Fig. 2). At the same time, the resident bird of the territory assumes a crouched position with head lowered and begins to hop cautiously in a circle around the interloper. As the resident circles, he makes threatening jabs with his beak at the strange bird, who turns to face the threatening bird as it moves. Such display may then be followed by a brief flurry of combat in which the birds fly at one another. Previous to combat, however, it is clear that the resident is the aggressive, dominant bird, while the interloper is the reticent, subordinate individual. Audubon (1834:12) noted in captive Blue Jays that when a strange jay was introduced into the company of others it fled to a corner and assumed an erect posture; moreover, a stiff, head-up posture is characteristic of subordinate Mexican Jays in interaction with dominant individuals. In the European Jays (Garrulus glandarius and G. lanceolatus) the appeasement display of a subordinate toward a dominant in- dividual is characterized by the body being held low, plumage fluffed, and head up as in the interloping Blue Jay mentioned above. But, in Garrulus the bill is pointed upward, not toward the opponent as in the Blue Jay, according to Goodwin (1952:304-6). The same author describes the display of aggression in Garrulus: it is a “lateral display” (bird turned at right angles to his adversary, crest raised, head held up with bill directed forward, feathers fluffed, and the body slightly crouched on the tarsi). Erect (head-up) postures with bill directed forward or down often are indicative of intense threatening attitude, although threat postures are not all erect postures (see discussion of gulls by Tin- bergen, 1955:119-25). Appeasement displays according to Moyni- Stupies OF New Wortp Jays 47 han (1955:252) are of hostile motivation just as are threat displays, but in appeasement the escape tendency is stronger than in threat. Though both types of behavior may deter attack from the opponent, the appeasement display does so by reducing the strength of the opponent's attack drive while not provoking a fleeing or escape tendency. Posture in appeasement seems to resemble that of threat in many cases but characteristically differs in that the bill and head are turned away from the opponent, either upward or backward. The head-up display of the Blue Jay seems to combine elements of both threat and appeasement postures as described for other species. While the posture in the Blue Jay probably is basically a threat posture (the crest is raised, bill directed forward) it seems to have become ritualized as appeasement posture (the bird never attacks from this position, may actually avoid attack, does not em- ploy the posture within its own territory, and never elicits an escape response from the opponent). The head is never turned away from the opponent by Blue and Mexican Jays in appeasement or other displays whereas such head turning seems to be a prevalent part of appeasement in Garrulus and many other birds. Bobbing is occasionally employed in territorial encounters, espe- cially those that occur above ground. If the resident male is forag- ing away from the nest-tree and a stranger alights there, the male returns, alights near the interloper, gives the cleeop cleeop call, and bobs deeply. If the trespasser does not flee at this intimidation display, the resident male may fly at the bird; a brief skirmish in which there is generally no body contact often ensues. In such encounters, the birds flutter toward each other and just before col- liding mount straight upward, flapping their wings at each other like game cocks. The resident male sometimes fails to drive the stranger out of the territory. When this happens, the male ignores the stranger, which resumes feeding nearby as if nothing has hap- pened. Blue Jays are often unsuccessful and show little persistence in their territorial actions. A jay driven from the territory some- times returns again and is allowed to stay and forage. Usually it is the male that engages in territorial struggles. The behavior of the female varies. She often remains on the nest but sometimes takes an active part in territorial skirmishes, especially if the male is absent when the trespasser appears. Figure 10 is a map of a portion of an open woodland on the campus of the University of Kansas, showing the locations of nests 48 THE UNIVERSITY SCIENCE BULLETIN of Blue Jays in 1957. Also included are the locations of false nests of certain pairs, and the places where nest-material was gathered. The home ranges of most of these pairs broadly overlapped, the banks of the small lake being frequented by birds of N-1, N-5, and N-8 commonly and by birds of other pairs and nonbreeding indi- viduals occasionally. Similarly the recreation area, used by humans for picnics, was common ground for all the jays in the vicinity—not just the pairs in the area under study. [ did not observe terri- torial defense by all pairs, and I am certain that for some of these pairs such defense was infrequent. The pairs of N-2 and N-5 and pairs that had nested in this grove just north of the lake in previous years seemed to have the most territorial squabbles, principally because they often attempted to keep other jays away from a hedge- row along the north bank of the lake. They also seemed to have frequent encounters with birds halfway between their nests and Oread Hall. These encounters were in most instances with birds from nests in the N-3 area, which often flew down the hill into the grove of small oak trees there, to feed. I never saw territorial encounters at the recreational area; N-8 on the map (Fig. 10) represents the closest active nest to that area that I ever found. This is perhaps indicative of the attitude of the jays toward that place, which as previously stated was common ground for many jays. The size of the Blue Jay territory is difficult, if not impossible, to ascertain, because of-the irregularity of defense (both in time and place) of the area by the resident pair. I have never seen terri- torial defense more than 200 feet from the nest, although defense that distance from the nest was as vigorous as that undertaken nearer. Territorial boundaries seem to be determined more often by some distance from the nest rather than by the presence of obvious demarcations in vegetation or physiography. One pair defended several times along a “line” that must have been based upon distance from the nest. Vegetation and physiographic char- acteristics were uniform from the nest considerably beyond the maximum distance at which territoriality was exhibited. According to Hinde (1956:343) there are often three main com- ponents of behavior in the establishment and maintenance of a territory. These are: (a) Restriction of some or all types of behav- ior to a more or less clearly defined area; (b) defense of that area; (c) self-advertisement within the area. Blue Jays exhibit all of these components in their behavior. In formulating a concept of territoriality in Blue Jays the following characteristics of the expres- StupiEs OF NEw Wokrtp Jays 49 sion of the phenomenon in the species should be noted. Both males and females engage in defense of territory; territorial boundaries are not rigidly defined as in such classically territorial species as the Song Sparrow (Melospiza melodia); the adults feed and rest within the territory, only occasionally leaving it to visit some special food supply or to bathe; territorial defense is displayed neither regularly nor persistently. Irregular defense of the territory seems to be based on the “atti- tude” of the interlopers. A jay that flies into the area, lands high in a tree, feeds young out of the nest, or behaves as a member of a courtship flock, will not be bothered by the resident birds. How- ever, one that flies into the territory and begins at once to forage on the ground or in the low branches of trees, or uses the area as a resting and preening site, arouses territorial behavior in the resi- dent birds. The interloper must behave as if he or she “belonged” in this area, before attack by the resident birds is instigated. This explains the irregularity of territorial display against strange jays. But it does not account for the lack of persistence in territorial behavior in the face of certain responses of the intruder, as men- tioned earlier. Territorial display occurs so infrequently in Blue Jays that it is studied with difficulty. Therefore, the frequency of success of resident birds in territorial skirmishes is not satisfactorily computable. Territoriality, in Blue Jays, involves maintenance of the pair-bond (but not pair formation), defense of the nest-site, and reduction of interference in reproductive activities by other members of the species. “Territoriality” in the Mexican Jay.—Territoriality of pairs does not exist in Mexican Jays. The nesting pair may freely allow from one to five individualsk—members of the flock—about the nest and young. Home range of flocks are separate with no overlapping during the time of breeding activity. Although I have never ob- served any behavior of the species that might be termed territorial aggression, the distinctness of the ranges of breeding flocks implies the presence of a spacing mechanism that is the effective equivalent of territoriality. We can only speculate as to the nature of this mechanism. In nonbreeding activity, two flocks previously dis- tinct while breeding typically unite; the smaller, breeding flocks otherwise remain discrete. Thus, the spacing mechanism is prob- ably involved with reproductive factors. Perhaps the areas occupied by flocks have been maintained over 50 THE UNIVERSITY SCIENCE BULLETIN many generations; the flocks are relatively sedentary, the habitat nearly unchanging over the years. The change in individual com- position of the flocks by old birds dying off, young birds taking their place is gradual. The young birds each year learn the boundaries of the home range by following their parents and associated adults of the flock. The boundaries have themselves been established so that sufficient food, shelter, roosting places, water, and other neces- sities are present to allow the existence of the flock, which remains fairly constant in size. If the foregoing is a satisfactory explanation of the spacing mechanism, then seeing the phenomenon at work in a short period of observation in the field is not possible; spacing Eleventh Street Stadium - INS2 discovered Memorial 5-6 Orive U-¢ ° 50 100 150O-yords EE ee SCALE Laas Fic. 10.—Diagram of portion of campus of University of Kansas showing locations of nests, false nests, and other places of activity of Blue Jays in spring and summer, 1957. Key to symbols: @, true nest; O, false nest; X, nest material gathering place; Numeral—numeral, date (month—day); N, nest (followed by numeral indicating nest of specific pair); E, eggs; Y, young; U-C, nest under construction. Stupres OF NEw Wor tp Jays 51 was effected long ago and its maintenance is now apparent only in the failure of flocks to overlap in range. Territoriality in the Scrub Jay—The Scrub Jay is a highly terri- torial species. Territorial boundaries of three pairs that I observed in the Sandia Mountains, New Mexico, met each other broadly (Fig. 13), but on their other sides did not seem to make contact with territories of any other pairs, although other Scrub Jays inhab- ited the general area under observation. It was therefore difficult to determine the extent of these territories in the directions from the nest in which no defense was necessary. However, when the birds left the vicinity of the nest in these directions, they were quiet and exhibited themselves less, at about where I have drawn “terri- torial boundaries.” Most of the activity of the jays was confined to the areas designated as their territories, except for occasional instances of trespass on their neighbors’ areas (which usually re- sulted in conflict), and regular visits to several areas where terri- tories overlapped. The latter included a well-spring, the only source of water nearby, located approximately at the junction of the three territories, and a grove of oaks several hundred yards up an arroyo from territory N-1 (Fig. 9). The former area seemed to be regarded by each pair as part of its territory. Members of different pairs did not often visit the water supply at the same time. Birds of N-1 and N-2 frequently defended the water hole against each other and against birds of N-3. The latter pair came often to the spring but did not defend it. On the other hand, the oak- grove was not defended, the birds typically being quiet and difficult to locate when there, although their flight to that area was easily traced. The birds frequented the grove in stormy weather, in the heat of early afternoon, and occasionally late in the evening. There is a strong possibility that they roosted there in the early part of the nesting cycle. The territorial boundaries of these three pairs of Scrub Jays were located according to physiographic features such as that which separated territories N-1 and N-2 (Fig. 12). Territory N-3 was located downhill from the other two territories, and the boundary between N-1, N-2, and N-3 was marked by a shelf of rock just below a stone shelter. Territory N-1 was 213 yards by 140 yards at its upper end and 100 yards at its lower end, giving an approximate area of 25,560 square yards, or about 5.3 acres. Dimensions of territory N-2 were not measured, but it was approximately the same size as N-1. The 52 THE UNIVERSITY SCIENCE BULLETIN sizes of N-2 and N-3 were not measured because only one side of each was definitely known. Since Scrub Jays do not sing loudly, proclamation of territory is effected by display and, when called for, by defense. Jays infre- quently trespassed on their neighbors’ territories; if the owner of a territory was not in the vicinity when a trespasser arrived, there was nevertheless obvious wariness on the part of the interloper. When a flying jay reached the boundary between two territories, it frequently turned suddenly in midair, as if a high wall existed along this “line,” even in the absence of residents of the adjacent territory. Large boulders and trees along boundaries seemed to be considered by each pair as part of its territory. Disputes over territorial bound- aries usually occurred at these places. Territorial disputes never included physical contact between contestants but consisted of brief chases, flitting displays, and bob- bing displays accompanied by br’rrrr calls. The flitting display consisted of vigorous hopping about with the tail and body being switched from side to side. The head is held erect, the body feathers are closely appressed, and the whole attitude of the bird Fic. 11.—Typical habitat of the Scrub Jay in the Sandia Mountains, New Mexico. Stupies OF NEw Wortp Jays 53 Fic. 12.—The line of boulders, on a slight rise of ground between two arroyos in the photo, separates two territories of Scrub Jays. is one of alertness. A similar display is given by male Blue Jays in group courtship of early spring. In a typical instance of territoriality a bird ventures to the bound- ary of its territory, whereupon a bird from the adjacent territory appears; the two birds perch in the same tree or on the same boulder along the boundary. The encounter may then involve a threat by each bird toward the other (flitting, bobbing, rattling call), but contact never occurs. The encounter is terminated when one bird retreats into its territory; such retreat usually occurs within 10 seconds of the beginning of display. The other bird remains for a minute or more at the boundary giving the brrrr?r’ call and dis- playing, though not so actively as when the opponent is present. Ecological bases of reproductive sociality in Aphelocoma.—In Aphelocoma, we may now consider the ecological basis for the 54 Tue UNIVERSITY SCIENCE BULLETIN existence of territoriality of pairs of the Scrub Jay throughout much of its range and high degree of sociality of the Mexican Jay in much of its range. Ecological factors possibly have been important in the development of territoriality only in the Scrub Jay of the genus. Although the ecological characteristics of the habitat of the Scrub fe) x . ss a= 0) ra ? Sine a ? f---- Stone Shelter DOE : Territory ne Fic. 13.—Diagram showing intersecting boundaries of territories of three pairs of Scrub Jays in the Sandia Mountains, New Mexico, April 2, 1956. Broken lines represent territorial boundaries. Stippled area is where all three tong overlapped at water source. Only one complete territory, of N-1, is shown. Jay and the Mexican Jay are varied, it seems evident, as Pitelka (1951:381) has pointed out, that the habitat preference of the Scrub Jay is far less narrow than that of the Mexican Jay. The latter species is confined for the most part to pine-oak vegetation throughout its range. Within this formation, it characteristically shows a preference for the more arborescent parts of the woodland (Pitelka, 1951:384) (Fig. 14), riparian situations being typical. StupiEs OF NEw Wokr tp Jays 55 It is possible that restriction to specific habitats may logically be correlated with highly social breeding habits. In a given region a population of the species may find only a small area suitable to its needs. In order to maintain the population, aterritoriality of pairs may thus be a necessity. The Scrub Jay seems able to exist in “poorer, more xeric habitat, although it does not always do so, invading richer habitat often, when the latter is not occupied by the Mexican Jay. The tolerance of the Scrub Jay for poor habitat is exhibited well in southeastern Arizona. In the Chiricahua Moun- tains, the species is rare except locally (for example, in the foothills near the town of Paradise). The jay occurs in isolated single pairs high on relatively bare sides of the canyons, and in dry foothills with pinyon pine (Pinus edulis). In these mountains, the Mexican Jay occupies the vegetationally richer canyon floor. Whether the two species occur sympatrically or not, the Mexican Jay seems ob- ligated to the richest, most arborescent portion of its range; the bird’s ecological requirements are thus probably more strict than those of the Scrub Jay. If we combine the last assumption with another—that the species has a minimum population size below which it could not survive, we arrive at a possible reason for the lack of territoriality in the Mexican Jay. Ordinarily, territoriality is thought of as functioning in the survival of a species. If, however, the Mexican Jay were territorial, with the some spatial requirements per pair as the Scrub Jay but with its present ecological requirements, the com- bination of these two sets of limiting factors might prevent the species from maintaining the population above the minimum level necessary for survival. This argument supposes a minimum popula- tion size, N, that in territorial species represents the breeding population. But N in a species such as the Mexican Jay must rep- resent the breeding pairs plus the non-breeding birds that live with the breeders, contributing toward reproductive success by helping to build the nest and feed the young, and probably affording in- creased protection from predators and ensuring success in other ways. We must assume this contribution, since if we do not the prevalence of these assisting birds is difficult to explain as having selective value to the species. The Scrub Jay, on the other hand, exhibits the alternative in its habits—territoriality of pairs with invasion of a wide range of habitats, except for the subspecies in Florida. The habits of the Scrub Jay in Florida are perhaps unique in the species in seeming 56 Tue UNIVERSITY SCIENCE BULLETIN Fic. 14.—Typical habitat of the Mexican Jay in South Fork Canyon, Chiricahua Mountains, Arizona. Note richness of vegeta- tion compared to Scrub Jay habitat, Figure 11. narrowness of habitat preference; this narrowness correlates well with the possible relationship between strictness or narrowness of habitat preference and the existence of highly developed sociality. A. c. coerulescens is restricted to scrub vegetation and does not occur in pine woods, hummocks, or swamps (Pitelka, 1951:316), as do populations of other races of the species. These other habitats are present in the range of the jay in Florida and are seemingly avail- able for occupying by the jays. Because of restriction to scrub vegetation, we might expect the Scrub Jay of Florida to possess a higher degree of sociality than territoriality of pairs, as, indeed, it does. Stupres OF New Wor tp Jays 57 Copulation and Related Activities Copulation in the Blue Jay occurs in the final stage of nestbuilding. It is preceded by a gradual cessation of nestbuilding activities and the appearance of precopulatory posturing and display. Precopulatory behavior in the Blue Jay—A day or so before copulation takes place, the female Blue Jay becomes broody; having brought a piece of material to the nest, she sits on the structure for several minutes at a time. She also solicits food more often and with greater intensity than she did previously. She frequently begs while on or near the nest and almost always begs when the male appears near her or the nest. Courtship feeding serves to stimulate the tendency to behave sexually in both male and female; this be- havior in turn strengthens the pair-bond. (After copulation, court- ship feeding also helps maintain sexual interest and consequently helps maintain the pair-bond through the breeding season.) The tendency to behave epigamically is at its height at the time of copu- lation; not only is courtship feeding most frequent then, but at the peak of sexual excitement the begging behavior of the female is intense to the degree that it may be termed copulation soliciting behavior. Soliciting behavior elicits from the male a response different from that elicited by begging, demonstrating the distinct- ness of these two similar types of behavior in the female. The drive of the male in courtship feeding behavior seems to be consummated by contact of his bill and that of the female; transfer of food does not seem necessary. The copulatory tendency of the male seems to be satisfied, however, in action involving fuller con- tact of the birds, accomplished through mounting. It has been stated that food-begging behavior and copulation soliciting behavior are similar, and that the latter is probably derived from the former. There may be intermediate responses by the male to begging of intermediate intensity by the female. I once observed a pair of Blue Jays, in the late nestbuilding phase, engaged in a peculiar display that I have termed the nudging display (Fig. 16). Nudging may be the means by which the male consummates a con- tact drive of intensity intermediate between copulatory and court- ship feeding contact drives. Although copulation did not take place after nudging, fluffing and preening of the feathers and vigorous feeding—all associated with postcopulatory behavior—were seen. In addition, the nest was nearly completed, the female was broody, and the time for copulation was near for this pair. In the nudging display, the two birds sat side by side on a limb, 58 THE UNIVERSITY SCIENCE BULLETIN the female fluffed and seemingly broody, the male with throat dis- tended and body hunched forward. The male then edged over to the female and gently pushed against her side; she retained her posture. He gradually moved her a short distance along the branch by his pushing action. Neither bird gave any vocalization. After this display, which occupied only a few seconds, both birds flew into a nearby tree and foraged more actively than in normal feeding behavior. The female occasionally interrupted foraging with vigor- ous preening. Copulatory and associated behavior in the Blue Jay—Blue Jays copulate in trees. Just before copulation, the female assumes the copulatory pose (Fig. 6). This posture, as mentioned, is an intense form of the begging posture (Fig. 5) associated with courtship feeding, but here the tail is held angled upward and slightly spread, the wings are extended forward, and fanned at the wrist, and the slightly-opened bill is directed upward toward the male. The posture of the male is stiffly erect, with crest raised, wings closely appressed to the body, and the bill pointed downward toward the female (Fig. 15). Typically, the female retains the copulatory posture and gives soliciting calls—kueu kueu, and as the male mounts, her calls become more intense. Copulation lasts for about four seconds, after which the male departs and the female remains perched for a few seconds, preening her wings. She then may go to the nest and sit briefly, as she does regularly in this period of the breeding cycle. Both birds engage in flufing and preening the feathers and in vigorous foraging as part of postcopulatory activity, as mentioned previously. The act usually occurs near the nest on a large horizontal limb, but on one occasion, I observed copulation between two birds that were strangers in a particular study area, in the territory of another pair. In this instance the resident male three times pre- vented copulation by swooping down upon the strange birds. Copulation was finally achieved and the fourth attack of the resi- dent male caused the copulating male to fly away. The attacks then ceased, and the strange female was allowed to remain perched and preen and fluff for nearly five minutes. It will be recalled that Blue Jays are territorial toward other jays that feed, or are otherwise active, in another jay’s territory as if they “belonged” there. Doubtless, copulatory activity by two birds indicates that they “belong” in the area where they are copu- lating; hence the aggressive reaction of the resident male. Stupies oF NEw Wokr tp Jays 59 Behavior in the Blue Jay During Egg-laying and Incubation Transition from nestbuilding—The transition from nestbuilding to egg-laying and incubation is a gradual one. “Nest-sitting” is perhaps the earliest element in the transition; such sitting by the female Blue Jay begins in the false nestbuilding stage, when only a mere platform of sticks exists. While the nest itself is being con- structed, the female frequently crouches on it, and as the nest nears completion she spends more and more time in this activity. The male, who has been feeding her all along, increases the frequency of feedings as the nest nears completion. Courtship feeding becomes obviously functional now, allowing the female to remain on or near the nest for greater periods of time. This will help assure more nearly continuous incubation and will reduce the possi- bility of loss of eggs to predators. Participation of the sexes——Only the female Blue Jay sits on the nest in this phase of the breeding cycle. This is typical of most, if not all, corvids (Amadon, 1944a:2). In this period the male feeds the female at or near the nest, forages casually in the area, acting as a guard for his mate and the territory against other jays and predators, and stimulates the female to remain on or near the nest by his frequent visits to the nest with food. Incubation is interrupted frequently by the male feeding the female or by the latter leaving the nest to feed, preen, or merely perch in the vicinity of the nest. In most cases, the female leaves the nest to be fed by her mate only early in incubation; she tends to remain on or near the nest to be fed in the later stages. At all times, feeding takes place within a few feet of the nest or in a nearby tree. Afterward, the male nearly always remains near his mate until she returns to the nest. If she chooses to stay off, the male accompanies her. At one nest, the female sat for 55 out of 60 minutes (5:00-6:00 p.m.), leaving the nest only once and being fed twice by her mate while off the nest. The next day in 60 minutes (6:55-7:55 a.m.) she sat for 46 minutes in periods of 8, 25, 12, and 1, leaving the nest for periods of 3, 5, and 6 minutes. The first two times off she was fed by the male. The last time off she joined her mate in chasing a squirrel. These are typical examples of activities early in incubation. From the middle of the second week until hatching, the female sits for longer periods, whether or not she is fed by her mate. At one nest 60 THE UNIVERSITY SCIENCE BULLETIN Fic. 15.—Precopulatory posture of the male Blue Jay. Fic. 16.—Nudging display, possibly precopulatory behavior. Female on left, male on right. Stupres OF New Wor tp Jays 61 in midincubation, the female sat for 59 of 60 minutes, coming off only to preen and stretch near the nest; she was fed twice by her mate. The following day she was observed to sit on the nest for 113 out of 120 minutes, leaving the nest once to stretch and preen. She was fed by her mate once, at the nest. In late stages of incubation the only change is toward longer periods of incubation and a greater reluctance of the female to leave the nest for any reason. For example, in early incubation at one nest, a stuffed Long-eared Owl (Asio otus) placed promi- nently in the nest tree elicited a general alarm bringing the female from the nest, her mate from his activities, and other jays from the surrounding area to mob the owl. In middle and late stages of incubation, the same ow] seldom caused the female to leave the nest, although mobbing action still occurred on the part of other jays. At one nest in a late stage of incubation, in an observation period of 130 minutes, the female was off the nest 2.3 percent of the time (once for 3 minutes) and incubated 97.7 percent of the time (two periods of 90 and 37 minutes, respectively). The following day in 160 minutes of observation, the female was off the nest 3.1 percent of the time (once for approximately 5 minutes) and incu- bated 96.9 percent of the time (two periods of 20 and 135 minutes, respectively). In her times off the nest, she preened, fluffed, stretched, and foraged. This female lost her mate sometime in early incubation and so was never fed at the nest. She foraged for herself, and late in incubation acquired a new mate. Although this latter bird was not collected, it seemed to be a first-year male. The significance of the acquisition of new mates is discussed previ- ously under Comparison of Courtship Habits and Pair-bonds. The female Blue Jay’s periods away from the nest in the phase of incubation are not primarily for purposes of foraging, but are for the most part spent in flying about, stretching, preening, and otherwise rearranging the plumage. Only when the male is not attentive does the female do a greal deal of foraging for herself. Since she is relatively inactive, she probably requires much less food than otherwise, and seems, in fact, to get little food. I noted on several occasions that contact between the female and male of a pair at the nest did not involve feeding at all. It seems that in almost any instance of such feedings there are two functions, one ritualistic or symbolic and involved with crea- tion or maintenance of the pair-bond, the other nutritional in mak- ing food available to the female. Emphasis is on the first function 62 THE UNIVERSITY SCIENCE BULLETIN until nesting begins and on the second one afterwards, but at no time is the act devoid of both. The fact that the ritual may be performed in the period of incubation without actual transfer of food, serves to emphasize the continued presence of the ritualistic basis after pair-formation. Presumably the need for maintaining the pair-bond is still important and perhaps can be effectively accomplished only through the stimulus provided by the presence of mutual interests of the pair—the nest, eggs, and young birds. Particularly does this seem logical in a species where the female performs all the duties of incubation and in which there are no other ritualistic displays that may be connected with maintenance of the pair-bond; it is possible, then, that the male remains an integral component of the breeding cycle at this time only by the continued stimulation he receives from contact with the female. Behavior in the Mexican Jay During Egg-laying and Incubation Two birds were involved in incubatory duties at a nest of the Mexican Jay observed by Gross (1949:246). The flock associated with this pair in nestbuilding maintained interest in the nest in early incubation, occasionally coming there and investigating the eggs. These visitors were often so intent on investigating the nest that they dislodged the female by nudging her and by grasping her beak if she resisted their attempts. Later in incubation, only the presence of the observer at the nest exciting the incubating bird caused the flock to assemble at the nest-site. My own observations of egg-laying and incubatory behavior of this species are limited to those made at one nest in June, 1956. This nest was discovered soon after the clutch of four eggs had been completed. This was an unusually late nesting (June 4), the dry season having set in several weeks before. The remainder of the population of jays in the vicinity of the Research Station had completed breeding or were feeding young out of the nest. As a result, the nesting habits of this pair may not have been typical of nesting habits at the height of the breeding season, as the obser- vations above indicate. Participation of the sexes—From June 4 until June 18, when three of the young hatched, only one bird, presumably the female, was ever seen at the nest. She was never fed by another jay near the nest, and when she left the vicinity of the nest she was not attended by another jay so far as I could determine. Other jays Stupies OF NEw Wor tp Jays 63 were observed to come near the nest three times in the period of incubation. Two of these times were on occasions when I had placed a mounted bird on a limb near the nest. When a mount of a Long-eared Owl was placed there, the female and three other jays (two with parti-colored bills) flew into trees near the nest-site. They did not enter into mobbing of the owl, but merely remained as interested bystanders while the female scolded and dived at the “enemy. When a mount of a Blue Jay was placed near the nest, a single other Mexican Jay investigated it. The third occasion when other birds besides the female appeared at the nest occurred when I once disturbed the female, causing her to call in alarm. Then, another excited jay flew into the nest tree, called reek reek, and investigated me. Attentiveness—The nest was located at the edge of a road used by as many as six cars a day. The incubating female never grew accustomed to the cars; as often as not she bounded from the nest as they passed. My presence on foot in the vicinity also excited her. Any slight movement on my part while she was on the nest caused her to leave. But for these interruptions, her incubation periods would doubtless have been longer. In a 120-minute period of no disturbance on June 14 (1:30 to 3:30 p.m.) she was off the nest 16.7 percent of the time (2 times of 8 and 12 minutes, respec- tively) and incubated 83.3 percent of the time (2 periods of 70 and 30 minutes, respectively). At 7:30 p.m. the same day, the female was on the nest. On June 13, in a period of 132 minutes (1:33 to 3:45 p. m.), she was off the nest 11.4 percent of the time (3 periods of 11, 2, and 12 minutes, respectively ) and incubated 88.6 percent of the time (4 periods of 31, 15, 26, and 35 minutes, respectively ). The two middle periods of incubation probably would have com- prised one period if I had not accidentally frightened the female from the nest. In early morning hours, the female was more reluctant to leave the nest, and her periods of time away from the nest were shorter than in the afternoon. Behavior at the nest—Unlike the Blue Jay, the female Mexican Jay exhibited no stealthiness at the nest. Almost always when arriv- ing at the nest or leaving it, and occasionally while sitting on it, she gave the reek reek call. It was obvious that some of these times she had not been disturbed by anything; these calls may have been assembly calls given to a mate or members of the feeding flock that inhabited the area. Often her calls were answered from some 64 Tue UNIverRsiry SCIENCE BULLETIN distance up or down the canyon, and these answering calls seemed to attract the female. Thus the vocal exchange may have func- tioned in orientation of the female toward other members of the flock. When she left the nest, she did not remain in or near the nest tree, as is often characteristic of female Blue Jays at the time of incubation, but flew out of that vicinity to feed and rest. No marked changes were observed in the attentiveness of the female from early to late incubation, in contrast to the habits of the Blue Jay. Discussion and Summary of Incubatory Habits In the Blue Jay and in the Scrub Jay, according to Amadon (1944b:12), the female is fed at the nest by the male. Such feeding does not always occur in the Mexican Jay, nor is a male always attentive, or even in evidence near the nest, throughout incubation. Amadon (1944b:12) did not observe feeding of the female Scrub Jay by the male away from the nest but believes that it probably occurs. Such feeding does occur in the Blue Jay, particularly in the early stages of incubation. Feeding of the female by the male at any time in the Mexican Jay probably is not common as in the Blue and Scrub Jay. Compared to the Mexican Jay, the Scrub Jay and Blue Jay are highly secretive in the vicinity of the nest. The former frequently calls in a loud voice at the nest, while the other two are silent or utter only low conversational notes at the nest. In behavior at the nest, the Scrub Jay and Blue Jay resemble each other more than either resembles the Mexican Jay. Care of Young by the Blue Jay Early nestling stage—The duration of periods spent by the female Blue Jay on the nest in late stages of incubation increases greatly as the time of hatching nears. Much of the food that the female receives at this time is brought by the male, who proceeds to feed his mate at or near the nest. For the male, feeding the female at the nest is closely linked with feeding of young birds immediately after hatching; the behavior of the male in these two functions differs hardly at all. The principal change in behavior occurs in the female. She continues to beg from the male and receives food precisely as before hatching, but the movements of the young beneath her and their gaping mouths when she arises from the nest are stimuli causing her also to bring food to the nest. Both adults, then, respond to a begging, gaping stimulus, now Stupies oF New Wortp Jays 65 provided primarily by the young birds. It will be recalled from the discussion of courtship feeding that the male did not gape in response to the arrival nearby of his mate, and this failure was seemingly the factor preventing the female, in turn, from attempt- ing to feed her mate. The gaping young provide this stimulus to the female. She still broods a great deal, remaining on the nest almost as much as before hatching, since the nestlings, once fed, do not gape, and offer stimulus (warmth, inactivity) for brooding. The female feeds, preens, and occasionally perches for several minutes in a tree near the nest, while not feeding or brooding the young. The male always remains in the vicinity of the nest, visits it occasionally, and chases away other jays and squirrels. These activities remain fairly constant until after about the first week. The male never broods the young in early stages of their life and rarely ever does so. Each feeding of the nestlings by the female is usually followed by a period of brooding at this stage. The practice of nest-sanitation is begun by the second day. The fecal matter is sometimes eaten at the nest, but later it is carried away; it then may still be eaten, since I did not observe fecal sacs to be dropped by the birds. By the fourth day after hatching, the contribution of the male to feeding of the young increases somewhat. A male at a nest with young four days old was observed to feed the nestlings five times and the female two times in 60 minutes. The female in this time brooded for approximately 29 minutes. The male brooded, too, for eight minutes while the female was away from the nest to get food for the young the first time. The male continues to feed his mate at the nest. Sometimes he fails to deliver to her the food brought, and she is thus stim- ulated to rise from the nest if she has not already done so. After she arises, the male may then feed the young. The male presents food without much stimulus from his mate, as the female’s begging is weak at the nest. While feeding the young, the adults are quiet except for low conversational notes, kut kut, and an occasional begging call given by the female. There are never more than two adults caring for a brood, in contrast to the case in the Mexican Jay. Territoriality continues to be expressed toward interloping jays behaving in the same way as the resident pair (probably in the same phase of the breeding cycle). 3—5840 66 THe UNIVERSITY SCIENCE BULLETIN Behavior after the first week of care of the young.—By the end of the first week, brooding is restricted mainly to nighttime, midday, and midafternoon, except in unseasonably cool weather when brooding may occur at any time and for periods up to two hours in length. The contribution of the male to feeding of the young is gener- ally greater than before this time, and there are occasional days when he assumes a major share of this duty. On the fourteenth day after hatching at one nest, in a period of 180 minutes (6:45- 9:45 a. m.), the male brought food to the young 18 times, the female twice. Moreover, the female begged the male for food al- most every time he visited the nest; she closely followed him about, or sat on the rim of the nest most of the day. She also contributed to care for the young by occasionally removing feces and by giving a part of the food received from the male to the young. Once she took food from a nestling and ate the morsel. Three days later, division of duties was nearly equal at this nest; the male fed the young five times and the female four times, in 140 minutes. The sexes probably participate about equally in feeding the young, but the previously mentioned tendency for irregular con- tribution to care for the young by either sex from day to day seems characteristic of the species. In all cases observed where one or the other adult failed to feed the young or fed them only a few times in a period of observation, the nonfeeding bird contributed to care of the young by performing other tasks, such as removal of feces, guarding of the nest, and brooding. I did not discover the reason for these failures to feed the young by either sex on a given day. Particularly puzzling was the peculiar behavior of the female that almost ceased to feed the nestlings and simultaneously reverted to begging her mate frequently in a manner characteristic of courtship. She was unquestionably a member of the pair owning the nest; no other adult birds appeared at the nest at any time. Her begging behavior was abnormal, of course, only in the fact that it occurred too frequently for this period of the breeding cycle. Such abnormal behavior was otherwise not ob- served in Blue Jays in this study. Resisting the temptation to clas- sify this behavior as neurotic, it is perhaps best to call it displace- ment begging (such reversions to behavior of a previous phase of the cycle are frequently characteristic of displacement behavior). The causative phenomenon might have been conflict between drives to feed the young and to be fed by her mate. Perhaps the SruprEs OF NEw Wor p Jays 67 crucial moment in this conflict occurred only briefly earlier in the day when the drive to be fed by her mate failed to be consum- mated. This failure then could have upset the normal course of behavior for this phase of the life cycle on this one day. It seems probable to me that other irregularities in participation in care of the young by the sexes may be attributable to similar subtle disturb- ances of the delicate balance of interaction between the adults. In early stages of care of the young, most feeding activity occurs in the morning hours. But after a week, feeding sometimes con- tinues even in the hottest parts of the day. On the afternoon of the day on which the male performed most of the feeding, the male fed the young five times and the female brooded once for two minutes in a 60 minute period of observation. Generally, feedings in the afternoon hours average little more than one per hour. Late nestling stage—Frequency of feeding of the young con- tinues at about the same rate in the last week of care of young, until the latter leave the nest (between the 17th and 19th days after hatching). Brooding is confined to night and brief periods at irregular intervals in the day. At one nest, the male fed the young five times, the female four, in 140 minutes of observation on the day that the young began to leave the nest. However, at another nest with young estimated at 16 days of age, the male fed the nestlings four times, the female twice, in 120 minutes. The latter nest was near a walkway and road, and the infrequency of feedings in the two hours may have been a result of the adults being fre- quently frightened by passersby. Throughout the nesting season, but particularly in the last days before the young are fledged, the adults show little timidity when near the nest. Comparative lack of vocalization by the adults in the vicinity of the nest affords protection of the young and nest. The parent birds also readily attack any predator that ventures near the nest. Beyond the age of about 15 days, the young birds become restless and move about in the nest, stand up in it, and even perch on its rim. The female occasionaly attempts to brood them and achieves a sort of halfbrooding position (Fig. 17). This position is seldom maintained for more than five minutes at a time, for the young squirm about, peck at sticks in the nest, and invariably jostle the female loose when her mate arrives with food. By the 17th and 18th day after hatching, the young move to the rim of the nest even when resting, and then begin to venture out onto nearby branches. 68 THe UNIVERSITY SCIENCE BULLETIN The birds have engaged in wing flapping from the second week of life, but only when the adults brought food. Now they sit near or on the rim of the nest and exercise their wings vigorously when the adults are not about. This exercise soon leads to the first at- tempts at flight by the young. Fledgling stage—Adult Blue Jays are cognizant of territorial boundaries until the young leave the nest, as I have previously discussed. But once the young have flown, their aimless rovings govern the movements of the adults, and the family group may range Over an area several times the size of the territory. The first flights of the young are taken by chance at different times and in different directions; brood-mates are thus separated for a short time. After the young attain flying and navigational ability the brood is reformed and seems to remain more or less a unit for the remainder of the summer. Loose family groups are prevalent in August and September before the flocks of autumn are formed. The adults continue to feed the young for several weeks (as long as one or two months, according to Laskey, 1958:213) after they leave the nest, or into the period of postjuvenal molt, when inde- pendent feeding habits begin to develop. The fact that the young usually remain close together improves their chances of being fed regularly; young out of the nest for at least two weeks frequently are seen perched side by side. Broods certainly away from the nest for over a month are frequently seen calling to their parents from the same tree and flying in a loose group from one grove to another following an adult that is searching for food. Adults seemingly continue to feed the young at the same frequency as before the nest was vacated. The lack of any further recognition of territo- riality is important in the wanderings of the family group, because its members may trespass without conflict in the territory of other jays that have eggs or young still in the nest. Care of the Young by the Mexican Jay In care of young Mexican Jays observed by Gross (1949:247) at one nest, only two adults were involved. The young were brooded in the first week and were left alone only infrequently when both adults were searching for food simultaneously. The male and female shared equally in care of the young, which left the nest at approximately 25 days of age. Gross observed other nests located in the vicinity of the one discussed above, two such nests not more than 75 feet away. At these nests each pair seemed independent StuprEs OF NEw Wor tp Jays 69 of the others. Nest sanitation was practiced, the adults swallowing fecal saces when the young were first hatched, and carrying these sacs away after the young were older. Most of my observations of care of young by Mexican Jays were made after the young left the nest. The few observations made in early as well as late stages indicate some striking differences in the behavior of Blue and Mexican Jays, and certain differences in the care of the young as observed by Gross and by me. Early nestling stage-——At the nest where my observations were made through most of the period of incubation, the female, as mentioned, received no help from any other adult, nor did other adults ever appear at the nest. However, on the day of hatching two additional adults appeared in the vicinity of the nest, and one of them made visits to the nest to feed the female there. The latter bird was excited, brooded the three young (and one remain- ing egg) for short periods, and flew about calling excitedly at other times. She displayed the same lack of caution about the nest as in the period of incubation. On several occasions, she flew to the rim of the nest, and then began a display of rapid fluttering of the wings accompanied by begging calls given with the bill open. At these times, another adult followed her, landing a short distance from the nest. The other bird, presumably a male, hopped to the rim of the nest when the female began to beg and seemingly fed the young as the female stood by. Afterwards, the female brooded Fic. 17.—The half-brooding position of a female Blue Jay in late stage of caring for nestlings. 70 Tue UNIVERSITY SCIENCE BULLETIN while the male sat in the top of the nest tree for ten minutes. If another adult came near the nest while the female was away, she always hurriedly returned, and gave the begging display until the other bird came to the nest or flew away. On one occasion, both adults left the nest in one direction, after which another adult ap- peared in the nest tree from a different direction. The female re- turned hurriedly, calling loudly, and both advanced to the rim of the nest, simultaneously. The strange adult seemed to feed the young while the female watched. The latter brooded after this episode. The begging call of the female is a soft kwaaa kwaaa kwaaa. I twice saw a bird feed the female at this nest; on one occasion, she quickly stood up and gave the food to the young, and the other time she ate the food herself. At the above nest, the young died on the second day after hatch- ing. The nest was a late one; the young hatched after the dry season was well advanced. The breeding activities of the Mexican Jays in the area as mentioned previously consisted only of adults feeding young already out of the nest. The adult birds at the present nest gave several indications that they were unable to find proper food for the young, which seems probable, since only small soft-bodied insects would be suitable, and these were certainly at a premium. Two characteristics described above are not found in the Blue Jay. These are: More than two adults attentive at a single nest, and lack of caution of the adults when near the nest. Both char- acteristics persist in Mexican Jays through the remainder of the nesting activities. Middle nestling stage—I made few observations on care of the young in the first or second weeks of their lives. I discovered two nests, one with eggs and one with young, in the Chiricahua Moun- tains, when most of the adults in the vicinity were still building nests. These two nests were within 100 feet of each other. The flock of perhaps eight birds that inhabited the grove containing these nests was composed entirely of adult (black-billed) birds. In a period of observation of 60 minutes at the nest with young birds, feeding occurred three times, the young each time giving the call characteristic of young captive Mexican Jays. This call is harsh and squealing. Fledgling stage—After the young leave the nest, the adults con- tinue to feed them for at least five weeks, a protracted period of postnestling care similar to that in the Blue Jay. As with the lat- Stupres OF NEw Wor tp Jays 71 ter species, independent feeding behavior appears after this time, along with the inception of postjuvenal molt. The number of jays at least a year old that accompanies a group of three or four juveniles soon after they leave the nest is remark- able. The attending birds include some individuals with wholly black bills, others with parti-colored bills; these are, generally, adults and subadults, respectively. One such group contained four juveniles and approximately ten other birds, five of which were never seen to feed the young but nevertheless remained close by, shrieking at and mobbing persons that came close to the young. All of these birds, then, formed a company whose activities centered around the welfare of four young birds. I have recorded as many as four different birds feeding the same young within ten minutes. As in the early stages of care, the birds caring for the young are not cautious, making the flock easy to find. But the flock is not easily watched, because the birds move about considerably, and the young, although they occasionally sit close together, may dis- perse throughout a grove of trees. The young are seldom silent, giving a begging call like that of young Starlings (Sturnus vulgaris) when soliciting food. Both adult and young Mexican Jays give reek reek calls of alarm and assembly, and soft kwot kwot conver- sational calls that are seemingly low intensity forms of the reek call. The attentiveness of the adults toward the young fluctuates greatly. In early morning hours feeding is at its peak, but there is a period of an hour or so before noon when activity is low. At this time all the birds can be found sitting quietly in trees. There is a period of increased activity for two to three hours in the after- noon, but activity after this declines, and the birds seem to disperse after about 3:00 p.m. No further feeding of the young occurs. I have mentioned previously that a late-nesting pair seemed to have difficulty obtaining small soft-bodied insects for their newly- hatched young in the dry season. Adults with young out of the nest at the time the dry season becomes severe are not faced with this problem, for although insects remain scarce until the rains begin in late July, small lizards are plentiful, and these seem to form a large part of the food given to the young. The adults catch the lizards and then tear at least the larger ones into pieces before giving them to the young. fe Tue UNIveRSITY SCIENCE BULLETIN PART Il1—NONBREEDING BEHAVIOR BEHAVIOR OF CAPTIVE JUVENAL JAYS Behavior of Young Blue Jays in Captivity The study of young jays in captivity afforded the opportunity to observe the development and maturation of many behavior pat- terns evident in wild adults of the species. The difficulties ex- perienced in the study of captive adults were nearly absent in studies of young, since the latter were much tamer and in conse- quence behaved much like their counterparts in the wild. In this study, two young Blue Jays were hand-reared from the late nestling stage, one through the postjuvenal molt (or, from about 19 to 62 days of age) and the other to the age of approximately 118 days. The birds were nestmates. No attempt was made to transform either of the young Blue Jays into “household pets.” Rather, an attempt was made to maintain a type of treatment that would cause the birds to remain tame but not friendly to the degree that they “regarded” themselves as humans. Throughout their lives in captivity the birds were fed raw horse meat and a mixture of commercial dog food and vitamin supple- ment. From time to time they were given other foods such as peanuts, grasshoppers, and other insects. The latter three foods were offered especially in that period when the motor patterns of self-feeding and play were becoming evident. The birds were retained in a small cage in my laboratory, but they were given daily freedom of the room for up to two hours at atime. At the approximate age of 54 days, they were transferred to an outdoor cage measuring 8 by 5 by 5 feet. One of the birds was later placed in a 9 by 12 by 6 feet outdoor cage. Flight —The broodmates of the captives left the nest within two days of the time I captured the latter birds. As Rand (1937:30) has suggested, young Blue Jays leave the nest before they can fly. The fledglings begin frequent exercising of their wings in captivity at about 27 days of age, the periods usually occurring in the eve- ning just before dark. Less often, vigorous exercising of the wings occurs in the morning hours an hour or so after the first feeding. These periods may be correlated with renewed energy supplies afforded by the food. Immediately after being fed, the young usually are drowsy for a period of 15 minutes to an hour before becoming active. Before the frequent wing exercises are indulged in regularly, the birds are unable to fly at all well. Upon their StupiEs OF NEw Wor p Jays 73 release from the cage they flutter weakly along a table top or to the floor to cower in a corner. If placed on a perch, they usually are content to remain there with no attempt to fly. In the periods of exercising the wings, the jays become consid- erably wilder than at other times. If the cage door is opened, they attempt to escape, refuse food, avoid being touched, and even threaten an observer. Ability to fly in a partially sustained manner first occurs when the jays are approximately 30 days of age, although each time they are released from captivity previous to this time, they exhibit progressively greater strength so that eventually they do not always flutter downward. At about 30 days, however, young jays exhibit the tendency to escape; if the door to their cage is left open momentarily they attempt to fly out and upward to a perch away from the observer. In these early successful flights, although the forward upward progress of the birds is strong, navigational ability is poor. They usually arrive at a general destination rather than at a specific predetermined point. Rand (1937:30) noted similar cases where the jays, although they flew strongly, perched on precarious and unsatisfactory perches for a time until greater navigational powers had been attained. In these early flights, the birds frequently fly headlong into window panes and even into walls. If they fall to the floor they are unable to take off again but hop rapidly to a place where they can climb higher and launch themselves. The lag of co-ordination behind strength of flight lasts as little as two days; at 31 days, the jays are difficult to catch once they have escaped the cage. Immediately after escape, they usually fly wildly, presumably until their energy has been somewhat dissipated, and then become quiet and select a perch. If not disturbed, they remain perched for several hours without again flying. Seemingly the confinement of the birds increases their urge to fly, but once “free- dom” has been secured, their drive to escape regresses rapidly. Rand (1937:30) noted that his captive young jays always flew to the higher perch-sites in a room. He assumed it to be a behavioral adaptation to keep the birds away from ground predators. Extension and vigorous flapping of the wings strengthens and co-ordinates the musculature. Wing-flapping is a basic intention movement of flight in these young birds, just as bobbing may be an intention movement of flight in adult jays. The young jays never bobbed. Feeding —When first removed from the nest, young Blue Jays will not accept food from the hand. But after several hours in 74 THe UNIVERSITY SCIENCE BULLETIN which morsels are frequently poked at their bills and simultaneous squeaking sounds are made with the lips, the birds finally respond strongly, although irregularly at first, by gaping, quivering their wings, and accepting food from the forceps. Often the birds must be offered food repeatedly over a period of several minutes before they will respond. The irregularity of response is not due to degree of hunger. The difference between a “correct” and an “incorrect” presentation of food is slight indeed. Acceptance of food takes place from almost any angle, with or without accompanying squeaks, at sudden and slow presentations, and often after repeated offer- ings in what seemed to be exactly the same manner each time. Begging and gaping responses to a finger can also be elicited. Once a bird has gaped, a finger poked into his mouth causes swal- lowing action and a slight chewing action just as in the instances when food is offered. The swallowing action of young Blue Jays includes two or three jerks of the head back and forth as the bird gapes and the food is shoved into its throat. The nictitating membranes are not usually closed at the time of feeding, but the eyelids and nictitating mem- branes may be opened and closed several times in the swallowing action immediately afterward. Placing a morsel of meat on the end of the bill of a stuffed Blue Jay and then shoving the head of this mount at the fledglings pro- duces a strong begging reaction. (Young Mexican Jays and Stel- ler’s Jays also accept food empaled on the bill of a stuffed Blue Jay.) The nestlings will occasionally accept food using this method of presentation when they will not respond to any other method. The first sign that self-feeding actions are soon to appear in young Blue Jays comes at an approximate age of 29 days. A young jay first clutches the perch tightly and hammers on it between his toes with the bill, in a manner similar to that employed by adult jays in feeding. Such actions have been classed as “play” in young jays by Rand (1937:38); the behavior is the primordium of a specialized motor action, general in nature at its earliest manifestation and seemingly useless (hence “play”), but it soon becomes functional. Before it becomes valuable in feeding it awaits the development of the ability to hold an object between the toes. With the first indication of self-feeding, it is almost as if an imaginary object were present between the toes. Other indications of foodhandling also appear soon afterward. One of these is manipulation of the food presented by hand to the birds. Particularly if feeding is delayed for a long time, the first few mouthfuls taken by the birds are hur- StrupiEs OF New Wortp Jays 79 riedly consumed as in the earlier stages of development. But sub- sequent offerings soon afterward are accepted in small bits from the forceps and are taken with the bill closed or nearly closed. These daintily-taken bits of food are then placed along the perch or in crevices and then “worried” for several seconds before being devoured. When I first noticed one of my captives hammering at its perch, I placed a peanut under the bird’s claws. The bird made surpris- ingly strong attempts to peck the nut but only got particles from it so long as I held the morsel carefully with my fingers. When I withdrew the fingers, the bird was unable to grasp and hold the nut properly and eventually dropped it. It is not until this holding ability is mastered by the bird that it is able to utilize its hammer- ing ability to feed by itself. Particles of food taken from the for- ceps and placed in a crevice or merely balanced on the perch at first never are handled with the feet but only with the bill. Thus, three different indications of self-feeding appear independently and only later are perfected and then functionally co-ordinated. These three indications are (1) absence of gaping and taking food in small particles, (2) hammering on the perch with the bill, and (3) grasp- ing and holding objects with the feet. The latter ability is the last to mature, and the functional employment of the other two must await the maturation of the third. By approximate age 36 days, Blue Jays pick up in their bills sun- flower seeds, peanuts, and other food particles and carry them to a perch where they attempt to eat them. Only with soft foods do the birds completely succeed in self-feeding. Flexible pieces of food are most often held in the bill and thoroughly whipped against the perch much as a bird does when it is attempting to kill prey such as an insect larva. The meat is then laid on a perch and pulled apart or placed in a crevice and carefully pecked. By approximate age 39 days, the jays have become adept at hold- ing and breaking open the shells of peanuts and sunflower seeds but still remain awkward at getting food from these. By now, the force with which they hammer the food particles is much greater. The birds also possess greater ability to hold with the feet. Rand (1937:54-6) has written concerning the use of the tongue in manipulation of food by young Blue Jays. My observations agree with his; in preparing to eat a food object, a bird utilizes the tongue with great dexterity to roll the morsel out to the end of its bill and to retrieve the food again into the throat. 76 Tue UNIVERSITY SCIENCE BULLETIN Correlated with maturation of self-feeding is the tendency of young Blue Jays to store food and other articles. The storing be- havior is seemingly stimulated by (1) satiation and (2) inability to break or ingest a foreign object mistaken for food. Rand (1937:44) classified objects such as buttons as playthings, but judging from the manipulatory efforts of the jays, I think the birds regard these objects as not different from foods, such as hard-shelled nuts, that are difficult or impossible for them to open. Both food and nonfood articles are approached and handled in the same way. I did not observe as did Rand (1937:41) the ingesting of pebbles, soil, or other foreign objects. The methods of feeding utilized on live insects by young jays are as Rand described. The birds attack the heads or other vital areas of the insects and peck at these until the insects are lifeless or the head or some other structure is pulled off. Reaction to water.—Drinking and bathing reactions appear at about the same time. Rand (1937:35) noted that at age 19 days a young Blue Jay jumped into a container of water and paid no at- tention to it, but at age 25 days the bird drank voluntarily, and on the 28th day it bathed. The delay in these reactions that my young jays exhibited was doubtless due to the fact that I began giving them water with an eyedropper soon after taking them cap- tive and did not offer them a place to bathe until after approximate age 35 days. When first they did bathe, however, the reactions were similar to those observed by Rand. First bathing behavior appeared at approximate age 40 days. This bathing reaction follows drinking behavior. The bird drinks, then hops into the water, flutters tentatively and quickly jumps out. The jay may repeat this initial and weak bathing attempt two to three times in the next few minutes. If another young jay is present, it will watch intently the bathing attempts of its companion and may itself be stimulated to bathe. Young jays usually cease bath- ing if a human appears nearby. I do not know why the birds react in this way, but it might explain my failure to observe bathing by one of my captives until it was approximately 56 days of age. After bathing, young Blue Jays preen and fluff vigorously. Bathing becomes frequent and more prolonged subsequent to the initial attempt, occurring at least once a day. Preening, bill-wiping, and head-scratching—Preening is part of the response to wetting and certain related stimuli involving ex- cessive moisture. Preening by young Blue Jays is, thus, always en- Stupies OF NEw Wor tp Jays a. gaged in after bathing takes place and nearly always when it is raining, when there is a heavy overcast, or when the humidity is high. This is true even though the birds may be inside, completely sheltered from precipitation. Preening is in addition functional in cleaning away sheaths from growing feathers as Rand (1937:53) noted. Preening behavior occurs occassionally after feeding or other intense activity that does not necessarily cause plumage to become deranged, and in these instances may be displacement activity. True preening—activities of plumage arrangement with the bill, beyond mere picking at the feathers or the flesh beneath—of young Blue Jays is highly developed from about the time that flight feathers are half grown and the birds begin to make their initial flights (15 to 22 days). Specialized forms of plumage arrangement, such as running the vanes of remiges between the mandibles with the tip of the bill, are frequent activities of Blue and Steller’s Jays, but not of Mexican Jays (see page 83). Besides preening, the Blue Jays also engage in vigorous shuffling of wings and tail, stretching of the wings, scratching of the head with a foot, and bill-wiping. While preening arranges the barbs of each feather and assists in clearing feathers of basal sheaths, stretching movements assist in arranging the feathers properly in relationship to each other. Both activities are common to many other passerines. Swallows and Robins (Turdus), for example, are frequent plumage “shufHers.” Bill-wiping first appears in captive Blue Jays on the fifteenth day according to Rand (1937:53); this is in general agreement with my observations, although I found that the action becomes much more prevalent with the advent of self-feeding. Adults of all species of jays that I have observed engage in bill-wiping frequently. Al- though it doubtless serves to rid the mandibles and nearby feathers of excess particles or juices of foods, the action is ritualized, since it is indulged in even when seemingly unnecessary. Blue Jays utilize the indirect method of head-scratching (the foot being extended over the top of the wing) as opposed to the direct method (cf. Simmons, 1957:178-9). Simmons states that a given species always utilizes only one of the methods of head- scratching, but Dilger (in Ficken and Ficken, 1958:277) observed a nestling Blue Jay to scratch directly, and other passerines are known to pass through brief, developmental stages of direct head- scratching. Rand (1937:53) and I observed only indirect head- scratching in captive young Blue Jays. I did not note the age at which head-scratching begins, but Rand (1937:53) recorded it 78 THE UNIvERsITY SCIENCE BULLETIN first at age 15 days. Head-scratching and rubbing the head on the perch becomes more frequent at the time of postjuvenal molt, which is heavy on the lores and circumorbital regions in Blue Jays, apparently causing irritation. Rand (1937:53-4) correlated head- rubbing with time of leaving the nest, since both occurred at the age of 19 days, but the correlation, aside from a relationship of time and maturation, seems to be invalid; there seems to be no logical reason why head-rubbing and time of leaving the nest should be directly related. Vocalizations—Among the young of species under consideration herein, Blue Jays exhibit by far the largest variety of call notes. (Adult Blue Jays are similar in this respect in comparison to adults of the other species. ) Newly hatched Blue Jays utter a high-pitched squealing call, squee, squee, when the nest is touched. This call is not harsh in quality but is penetrating at short range. By the end of the first week of life, this squeal has a harsh quality, squrreesh, and the volume of the call is greater. Both of the above calls are food-so- liciting calls given only when the nest is moved slightly as the adults land on it, or when food, an object even vaguely resembling food, or the bills of the parents are directed toward the young birds. For at least the first ten days, no other calls are given regu- larly except an occasional peeping note, which seems to indicate satisfaction or satiation after feeding. From the middle to late nestling life, a rich chortling call is added. It is given at the moment that food is received and is con- tinuous with the squreesh call, hence: squreesh-oritch-oritch-oritch. In addition, a series of twittering notes, peter peet-peeter is uttered after the food has been swallowed. A suppressed version of the squealing or whining note given before food is taken is sometimes uttered after food is swallowed. Water presented to the birds evokes the same calls as does food. The calls described above are retained in post-nestling life as important parts of the repertoire of the birds. Wild juveniles flying about the trees in June and July in Kansas utilize exactly the same calls. With the advent of postjuvenal molt, there are additions to the repertoire of calls and changes in quality of calls already possessed. Soliciting, chortling calls, and calls of satisfaction gradually drop out of use with decreased dependence for food upon adult jays or humans. Additional calls appearing in the repertoire include a variety of Srupies OF New Wortp Jays W2) faint “conversational” notes and a remarkably varied song, given by both sexes. The conversational notes are usually given by two birds at the same time and consist of a nearly steady stream of twitterings, soft chirping notes, and a stuttering, whining call. The song of these birds is their noteworthy vocalization; I did not hear it given by young Mexican or Steller’s Jays. The song is definitely closely related to the conversational notes. While the latter are often given at times of great activity, the song is given only when such activity ceases. The conversational notes often lead gradually into singing; the twitterings and chirpings become more intense and closer together and finally form a rich many-syl- labled song. This song is somewhat like those of the Purple Finch (Carpodacus purpureus) and the Blue Grosbeak (Guiraca caeru- lea) but harsher and more uneven in rendition. The song is given sotto voce, seemingly indicating contentedness, rather then being given as an announcement of territoriality. (Adult Blue Jays occa- sionally engage in such singing in the wild.) Another call, chur chur, is rarely given by young Blue Jays and is uttered at times of fright. From late in the postjuvenal molt on, the development of adult calls gradually takes place. These include only variations of the jay call (of assembly and alarm) at least up to approximately 118 days of age. No indication of the pumphandle or squeaky-gate calls (associated with courtship and suppressed aggression) or other calls of the adults are given in this time; it is probable that these more specialized calls are not given until at least the autumn of the first year of life. The jay call given by these young birds is derived from the squrreesh of begging juveniles and is more nasal in quality than the jay calls of adults, although it is uttered under similar conditions. Begging juveniles flying about in late summer and early autumn can be heard giving calls intermediate between these two calls. Finally, there is a harsh cuz cuz call, which young Blue Jays give only rarely, each time after feeding. It is associated with restless behavior occasionally exhibited just after the last of a supply of food is devoured and before the birds settle down to rest, Relationships with one another and other birds—yYoung Blue Jays seldom quarrel with each other, as Rand (1937:56) observed. Nestmates, in fact, engage in almost all of their activities together, only rarely struggling with each other over morsels of food. Fre- 80 THE UNIvERSITY SCIENCE BULLETIN quently, they do pull at opposite ends of pieces of paper or steal peanuts or pieces of bread from one another, but there is never any resulting threat from the loser after these interactions. If a Steller’s Jay or a Mexican Jay of approximately the same age as young Blue Jays is placed in their cage, they proceed to become acquainted with the stranger by a ritualized method. The Blue Jay approaches and with the bill pulls at the feathers in the alular region of the other bird. Then the Blue Jay pulls gently at the toes, rectrices, nape feathers, and circumorbital feathers of the stranger; the latter remains nearly still during the investigation. The pulling and pecking is never vicious. Young Blue Jays occa- sionally administer such treatment to each other, but less intently. I am not certain of the significance of this behavior, but it is cor- related with the habit of juvenal Blue Jays of investigating all strange objects, animate or inanimate, by lightly pecking or pulling at them. The behavior has a counterpart in aggression of the peck- order in the Mexican Jay (Fig. 19). At approximate age 35 days, the first evidence of the habit of the male of feeding the female appears in Blue Jays. When a young male holds a piece of food in his bill, a young female may beg him with bill open and wings quivering. Rand (1937:56) noted begging of one young jay by another but found that it al- ways elicited a similar begging response from the other bird. Such behavior is more difficult to relate to courtship feeding be- havior. However, in my two birds, the begging by the female usually stimulated the male to feed her or at least go through the motions of feeding her. Such behavior would seem to indicate a close relationship of courtship feeding and feeding of young birds by adults. Only once did I see the male solicit food from the fe- male. The latter did not possess food at that time and did not respond to the behavior of the male. Relationships with enemies—When mounted specimens of the Sharp-shinned Hawk (Accipiter striatus) or the Long-eared Owl (Asio otus) are brought in sight of a young Blue Jay, it reacts by lowering its crest, crouching, and eyeing the object with one eye. When the mount is brought close to the bird, it flies about in wild panic, uttering no calls. The reaction toward the mount of the hawk is less intense than that toward the owl. When a stuffed Blue Jay is presented to juvenal Blue Jays soon after the presenta- tion of the specimens of predators, it evokes no reaction. Neither Blue Jay in this study had ever seen a live hawk or owl. The birds also react toward predators such as coatimundis StuprEs OF NEw Wortp JAys 81 (Nasua narica), several of which came near an outdoor cage con- taining my captives. But in their response to these mammals, the birds were highly vocal as they flew about in the cage, giving a loud whining call that I otherwise never heard them use. Relationships with humans.—Young Blue Jays can be treated in such a way that they regard humans as their fellows. I success- fully discouraged such behavior in the captives as mentioned earlier. Though the birds did not fear me if I sat in the room and watched them or fed them, they were frightened if I attempted to handle them. The only threat reaction they ever exhibited was toward me, and this occurred only when I poked at them in a deliberate attempt to provoke such a reaction. In threat reaction, the crest is raised and the mouth opened toward the aggressor; the body is held erect and no calls are given. The jays had no oppor- tunity to become familiar with other humans. Sleeping.—In the daytime, a young Blue Jay frequently dozes with its bill directed forward, but at night it always sleeps with its head turned posteriorly and its bill inserted among the feathers of its back. In either case, the two feet of the bird clasp the perch and the body is lowered to a crouched position on the tarsi. As the bird grows older, it crouches less while sleeping. Rand’s (1937: 52) observations on sleeping of young Blue Jays agree with mine on the preceding points. Rand states that the young do not tuck their bills in the back feathers in sleeping prior to age 20 days, which in the wild would be the period when the young were in the nest and when adult roosting habits are not developed. Young Blue Jays are easily aroused from sleep and immediately become active in such instances. When allowed to awaken naturally with the coming of daylight, they do so at dawn but do not become active until they are fed, whereupon they may undertake intense excited movements about their confines. Diurnal sleeping decreases gradually and simultaneously with the development of independent feeding, flying ability, and other signs of maturation. Sleeping in daylight hours disappears en- tirely as the postjuvenal molt progresses. Behavior of a Young Steller’s Jay in Captivity I raised a young Steller’s Jay from the late nestling stage (ap- proximate age 15 to 20 days) through the greater part of the post- juvenal molt (approximate age 120 to 125 days). Until approxi- mate age 50 to 55 days the bird was kept in a cage by itself part of the time and in a cage with two young Blue Jays part of the time. §2 THE UNIVERSITY SCIENCE BULLETIN It was frequently kept in the same room with young Mexican Jays in this period. When the bird was able to feed for itself, it was transferred to a large outdoor cage containing a young Blue Jay and a young Mexican Jay of maturity similar to the Steller’s Jay. In the development of flying ability, feeding and drinking methods, manipulatory ability, preening, vocalizations, relationship with other animals, and sleeping habits, young Steller’s Jays are remarkably like young Blue Jays. Minor differences between the two species in these aspects of behavior were probably due to the greater reticence of the Steller’s Jay, at least until the time of late postjuvenal molt. This greater reticence may have been caused by the lack of competition and companionship from another young bird of the same species. At postjuvenal molt the two species of Cyanocitta diverge in behavior most notably in vocalizations. As indicated above, their juvenal calls are similar, but with matura- tion young of each species begin to acquire calls characteristic of adults of their species. Behavior of Young Mexican Jays in Captivity Two young Mexican Jays were raised in this study. One was taken from the nest at approximate age 25 days (its nestmates had already flown). The other was captured when still unable to fly strongly at approximate age 28 days. The two birds were retained in captivity until approximate age 60 and 116 days, respectively. The birds were given care and treatment similar to that given young Blue and Steller’s Jays. I was not able to tame the two Mexican Jays so well as young of Cyanocitta so that behavior of the former for the most part was probably not typical of young of that species in the wild. Young Mexican Jays are similar to young Blue and Steller’s Jays in their behavior patterns and in the development of these pat- terns. Mexican Jays differ from Cyanocitta behaviorally in the following ways. The use of the tongue in feeding and other manipulatory activities is not well-developed. Young Mexican Jays do not drink as much water as do young Blue Jays and probably Steller’s Jays, although independent drink- ing ability appears at about the same time in all three species. It is possible that the failure of young Mexican Jays to drink as much water as the others is correlated with the fact that the Mexi- can Jay lives in a more nearly arid climate. My captive young Stupies OF NEw Wokrtp Jays 83 Mexican Jays did not bathe, either, which might be correlated with the occurrence of the species in arid regions. In additon to drinking little and not bathing, young Mexican Jays do not preen themselves dry when their plumage is thoroughly wet down in the laboratory. Young Blue Jays and Steller’s Jays treated similarly preen vigorously. The preening habit is not totally absent in Mexican Jays, since the adults preen, although they do not do so as often as do Blue and Steller’s Jays. Young Mexican Jays do fluff their plumage in attempts to rearrange the feathers. I know of no published information concerning development of the preening habit in birds. I assume that most birds preen after becoming wet, but possibly those living in drier climates have less well-developed preening habits. Preening, after all, hastens drying, which would hardly seem conducive to the conservation of moisture, which is managed by most organisms inhabiting arid climates. Of im- portance to individual birds, of course, is the function of water on the plumage as an “evaporative cooler” in arid places immediately after a rain. Failure of the birds to preen also would prolong the effects of this cooling system. The vocal repertoire of young Mexican Jays is of lesser variety than that of young Blue or Steller’s Jays. There is no song (in young or adults). The begging call in early nestling life is a high- pitched whistle, shreee, becoming choreee in later nestling life, with an appended itch-eeitch as food is accepted. In post-nestling life one additional begging call, similar to the begging call of young Starlings, is added to the repertoire. The adult conversational notes, kwot kwot or whut whut, and the alarm and assembly calls, reek reek, are a part of the repertoire prior to postjuvenal molt, in con- trast to the failure of adult jay calls to appear in Blue Jays until postjuvenal molt stage of maturation. Moreover, whereas the jay calls of young Blue Jays are easily distinguishable from jay calls of adult Blue Jays, the reek calls of young Mexican Jays are like those of the adults. Juvenal Mexican Jays exhibited little fear response toward a stuffed Long-eared Owl that frightened juvenal Blue Jays nearby. Since adult Mexican Jays in the wild mobbed this same mount, either captivity altered the natural behavior of the young Mexican Jays or fear of predators in the species is learned. However, be- cause of the unusual lack of fear exhibited by adult Scrub Jays in the wild toward mounts of predators, it seems well to mention this similar behavior of the young Mexican Jays. 84 THE UNIVERSITY SCIENCE BULLETIN FLOCKING AND FORAGING BEHAVIOR Jays are generally gregarious. They are not secretive as individ- uals or pairs, although territoriality exists in some species as we have seen. Size and kind of groups vary considerably in different species of jays, even within the same genus. Blue Jays and Steller’s Jays gather in flocks in the nonbreeding season. In the former species, these flocks vary in size from five to fifty or more birds, and the ag- gregations are loosely organized. One seldom sees 20 Blue Jays fly as a group from one tree to another as frequently is observed in Mexican Jays. Unlike the latter species, there seems to be a weak tendency toward unison movements of Blue Jays and toward much interaction of any sort between members of the flock; the birds move and feed singly; there is much leaving and re-entering of the flock by individuals. Steller’s Jays form flocks less frequently than do Blue Jays, but the aggregations are of the same degree of organi- zation. The genus Aphelocoma, however, exhibits at least two de- grees of organization of the flock. Whereas the group is maintained by such a strong bond in Mexican Jays that unison movements in any season (as well as semi-communal nesting) are characteristic, Scrub Jays have a less well-developed flocking tendency, perhaps less than any other species in either Cyanocitta or Aphelocoma. Flocking, Movements and Foraging in the Blue Jay Autumnal flocking—Loose aggregations and the independence of individuals are characteristic features of flocks of Blue Jays in autumn. A flock of 50 of these birds in an open oak-hickory grove may center its activity in one or several large trees, usually those containing a supply of acorns, upon which the jays feed. Jays in these flocks are not noisy but limit their vocalizations to occasional, low-intensity jay calls and the wheedle-eee call. There is a steady but loose stream of birds leaving and entering the trees. Many of the birds take acorns from the tree, remove the husks, and carry them one to three at a time away from the grove. Birds returning presumably have carried acorns away to some distant hiding place. Some members of the flock fly to the ground and bury acorns in the soft earth. A few birds feed on the fruits in the tree, while still others store acorns in crevices high in branches of the trees. Direct relationship between individuals in these fall flocks is fairly common but always involves aggression. This is wholly threat, since there is no real conflict or contact between the birds. Usually such interactions concern a food article. One bird may Stupies oF NEw Wokr tp Jays 85 store an acorn immediately after which another jay, having watched the storing procedure, may take the acorn and flee with it. The first bird, if it is still nearby, may then give chase briefly. Each Blue Jay seems also to carry with him a foraging “territory” to which the term “individual distance” (cf. Hinde, 1956:342) is ap- plicable. When another jay approaches to within a few feet, it is usually chased away to a distance of 20 or 30 feet before normal foraging by both birds continues. If the flock is large and all its members are feeding in the same tree, the size of these “portable territories” is smaller; their boundaries often are trespassed, so that frequent, short chases and skirmishes occur. Aggression is not accompanied by vocalization for the most part. The antisocial tendency of each jay is evidence of the type of flock characteristic of the species in autumn. In autumnal flocks of Blue Jays, reactions toward individuals of other species of animals occasionally occur. In fact, when a preda- tor is discovered the group of birds may become a closely knit organization in expression of mobbing behavior. Postbreeding flocks begin to form late in summer and are com- posed of several family groups of an area. As was discussed ear- lier, feeding of young jays by the parent birds continues for sey- eral weeks, perhaps occasionally for well over a month after the young leave the nest. In this period, the young and adults com- monly move over an area several times the size of the territory, an area comparable to the home range of the parental birds. By early August, this range is expanded still more, and three or four family groups eventually begin to concentrate their movements in the most favorable place where their home ranges overlap. In this way, a loose flock begins to form. These family groups largely maintain their discreetness in late July and August, moving within the same grove, often occupying the same tree. At this time, the young are easily distinguishable by their juvenal plumage and their habit of frequently begging the adults for food. This disbandment of family ties is closely cor- related with the onset of the postjuvenal molt, as indicated previ- ously. This molt begins in middle or late July for some birds that fledged early, but it is not well under way for most of the young in the region including Kansas until August. Young that were dependent upon parents for food acquire independent feeding abilities at the time of this molt. But in spite of the disappearance of family ties, the birds continue to frequent the area resulting in the loose flocks of autumn. 86 THe UNtIversiry SCIENCE BULLETIN Large flocks of Blue Jays are maintained generally from Septem- ber until late November or early December; a diminution in size of the aggregations accompanies the first extended cold period of winter. Blue Jays also flock in autumnal migration. I have little first- hand evidence on the composition of these flocks, since they are usually observed flying high overhead, where their members can neither be aged nor shot. In October, migrating groups of from 15 to 50 Blue Jays are frequently noted near Lawrence; they fly in straggling formation and sometimes land in the tops of large trees. It is characteristic of these migrating birds that they are silent or nearly so, in contrast to courtship flocks of jays in the spring. Adult Blue Jays in eastern Kansas may be nonmigratory, and first-year birds may be entirely migratory; data concerning this are discussed under winter flocking. Winter flocking—Typically in winter, one finds groups of two to four Blue Jays, but seldom large flocks. These small associations are evenly distributed over an area with the birds occupying win- ter ranges that may be “defended.” Continued observation in a given area reveals a small and constant number of individuals to be present. If members of a resident group are trapped and re- moved or shot, they will be replaced by other jays so that the area continues to support the same number of jays through the winter. At least this is true if the area is particularly favorable—containing abundant food and shelter. Probably jays replacing those removed come to the area from a less favorable winter range. Pitelka (1946:82-4) suggested on the basis of his own investiga- tion and those of other workers that subadult Blue Jays may be migratory, though perhaps only in the northern states. He further suggested that adult birds may gradually become sedentary. Of specimens that he examined in the Museum of Comparative Zoology, 29 first-year birds were collected north of the Mason-Dixon Line, and only four of these were taken in December, January, and February. First-year Blue Jays can be distinguished with experi- ence and practice by their tapered rectrices and dull-colored outer primary coverts and alulae (cf. Pitelka, 1948:83). Laskey’s (1958: 211-18) banding studies indicated that first-year birds leave the area of Nashville, Tennessee, in their first winter and that the win- tering grounds of these birds range from Alabama to Mississippi. According to results of my trapping and specimens in the col- lection of the University of Kansas, wintering Blue Jays in Kansas ‘rs StuprES OF NEw Wokrtp JAys 87 are almost all adult birds, whereas beginning in April and May and continuing through September, the numbers of subadult birds equal or surpass the numbers of adults (Fig. 18). Since fall mi- gration of Blue Jays in Kansas occurs in late September and Octo- ber, birds present in Kansas in November are mostly wintering indi- viduals. It is probable, then, that first-year birds are the principal migrants of the population in Kansas. The resident Blue Jays in Kansas in winter begin courting in late February, and daily obser- vations indicate that wintering birds seen courting then are the ones that begin nesting in the same area a month and a half later. Analysis of specimens of Blue Jays taken throughout the year in the lower peninsula of Michigan indicate that there, too, aduits constitute the greater portion of the population outside the breeding season, while the proportion of adults to subadults is approximately equal in the summer months, June through August (Fig. 18). One would logically suspect that in winter in the southern states subadults would far outnumber adults, assuming that the popula- tions of Blue Jays breeding there are more sedentary than in the north, and that the incoming migrants are principally subadult birds. However, analysis of 92 specimens taken in Louisiana and Mississippi indicate that there, at least, no such ratio of subadults to adults is created (Fig. 18). Several possible reasons may explain this fact, as follows: (1) The collection from Louisiana and Mis- sissippi contains more adults, because these are easier to kill (data from the collections made in Michigan and Kansas tend to refute this assumption); (2) subadult birds do not migrate south (but these age groups from Michigan and Kansas must go somewhere in winter, and observations in spring and fall indicate south to north and north to south movements of Blue Jays); (3) subadult birds from the north either spread out over a large area, winter in the southern states but not in or so far south as Louisiana and Missis- sippi (Laskey’s, 1958:215, data contradict this), or disperse over a large area in the south so that a change in age-ratio at a given place or in a given small sample is hardly detectable. The last possibility seems to me to be the most logical; it is evident that birds summer resident, for example, at Nashville, Tennessee, disperse over a wide area in winter, including Alabama and Mis- sissippi (Laskey, 1958:215). Probably birds nesting in Kansas and Michigan do not also migrate in large numbers this far south but replace in winter the summer resident population of Tennessee, Kentucky, and Arkansas. The problem might be easier to inves- 88 Tue UNIversity SCIENCE BULLETIN tigate if each individual specimen from the southern states in winter were determinable to subspecies (either bromia, migrants from the north, or cristata, the resident birds). But a broad cline in mensural characteristics exists between the two “subspecies,” making racial determinations of individuals hazardous where interspersion of the birds occurs in winter. Food-storing—Food-storing is a well-developed habit in the Blue Jay. It seems that sometimes the behavior is not functional in assuring a food reserve, since an occasional jay will carry a small particle of food high into a tree and wedge it into a crevice between a bud and a twig, where the morsel could not possibly stay fol- lowing the first strong wind. Furthermore, storing may take place in spring when food is plentiful. In addition, jays many times store small bits of food and minutes or seconds later return, extract, and eat the particles. They may take into their mouths and gul- lets more than they can possibly eat at once when they are con- fronted with a large supply of food. But they will carry off such surpluses, disgorge the morsels piece by piece, sticking one in one crevice, another some other place, and so on until each piece is stored or eaten. Usually these stores will not be touched again and thus go to waste, judging by observations of captive birds. The latter would seem to have a better chance of remembering or acci- dentally coming upon their stored food supplies than do birds in the wild. It is certain that in the wild, the type of place often chosen for storage is not easily marked or remembered by the birds. I have mentioned already that stored food may be taken imme- diately by a jay other than the one that stores it; this is particularly true in captive birds. It is difficult to believe that the habit of storing food has arisen in jays in direct response to a need for a future supply of food. Possibly there is conflict between stimulus to eat provided by the presence of food and stimulus to discard the food caused by the lack of physiological need for sustenance, and this results in dis- placement storing. Perhaps, on the other hand a more involved, less direct explanation is necessary—that food-storing is an expres- sion of winter territoriality. Food-storing as a possible spacing mechanism in a winter population of a species previously seems not to have been discussed in the literature. Lack (1954:256-7) discusses winter territoriality in birds, pointing out that in some species it seems functionless, in others it may ensure ownership of the area in spring, and that in others there is a suggestion that there Stupies OF New Wokrtp JAys 89 is food value. But under the latter factor, he mentions only actual] feeding (not storing) and involves food in territoriality as a reasor for spacing, not as a means of it. As an example of how this mechanism might function, we may imagine that a supply of seeds is suddenly made available within the wintering area of several Blue Jays. This supply is more than can be eaten at once, yet another similar supply might be unavail- able for weeks. If the resident Blue Jays eat only what they immediately need, they leave “in circulation” an amount of food that if found by intruding jays might encourage the latter to fre- quent that area regularly, not only when surplus food is available, but more important, when it is not. The resident jays by taking out of circulation not only the immediately devoured seeds but also those that are soon stored (making them comparatively un- available, at least) may not ensure a future food supply composed of these stored foods, but may ensure their acquisition of future foods that appear in the area, by having discouraged intruding jays from remaining or from returning. If, of course, the increased food supply is such that the resident jays of the area cannot keep all of it out of circulation, then the carrying capacity of the area is actually increased, and the number of resident Blue Jays can naturally increase. Flocking, Movements, and Foraging in the Mexican Jay I have not had the opportunity to observe flocking behavior in this species in autumn and winter, but presume it to be similar to flocking in early spring and after nesting is completed in late sum- mer. What I have previously said concerning flocking at the time of nestbuilding and during feeding of the young that have left the nest points out the basic differences between the flocking habits of this species and jays of the genus Cyanocitta. Flocks of the Mexican Jay are highly integrated aggregations in which the individual members all move, feed, rest, sleep, mob, and in other ways act alike as a unit. Flocks are probably in all cases family groups that remain together throughout the year, if, again, we may judge from flocks of early spring and late sum- mer. No migration occurs, but the flocks wander to lowland areas occasionally. Activities of the flock—Except in the nesting season, flocks of Mexican Jays wander about the low hillsides, foraging along creek- bottomlands, in groves of oaks, or in the scrubby vegetation con- 90 Tue UNIVERSITY SCIENCE BULLETIN Fic. 18.—Relative abundance of adult and subadult Blue Jays in the lower peninsula of Michigan, Kansas, and Louisiana-Mississippi based on specimens in the collections of the University of Michigan, University of Kansas, and Louisiana State University, respectively. In the Michigan section, the first three columns represent (1) September-October, (2) November-February, (3) March-May. The wintering period there is divided to illustrate the lack of significant changes in the composition of the population taking place from September through May. In the section concerning Kansas, the divisions seem self-explanatory. In the section concerning Louisiana, the divisions are arbitrary, since no significant changes in the composition of the popula- tion were evident. Crosshatched areas of the columns represent the per- centage of the samples composed of adult Blue Jays. Clear areas of the columns represent the percentage of the samples composed of first-year Blue Jays. The actual percentage indicated in each portion of each column is given numerically to the left of the portion of the column to which it refers. The numeral above the top of each column indicates the number of specimens composing the sample. Stupies OF NEw Wor tp Jays I : 2 ] 3 Summering Wintering Population Population MICHIGAN 22 22 Poren So * SSK 86.5 % 62.7% 50.9% 44.4% 13.5 % Post-nesting | Wintering Pre-spring July-Oct. Nov.-Feb. = March-June through Population Migration Migration through Nesting KANSAS LOUISIANA 92, THE UNIVERSITY SCIENCE BULLETIN sisting of scrub oaks, pinyons, junipers, and herbs. The flocks are difficult for an observer or predator to approach without being detected, because all members are wary and ready to give the alarm cry, setting the entire group to flight. Particularly when the flock is feeding in open vegetation where individual birds are easily seen on the ground, one or two members may remain on exposed perches above the flock—seemingly serving as lookouts. The loca- tion of a flock of Mexican Jays may be detected by the presence of these lookouts scattered about a hillside. Although most feeding by a flock is on the ground or in low vegetation, occasionally in early morning hours, aerial feeding may be indulged in by Mexican Jays. In this activity, the jays sit on exposed perches and launch themselves flycatcher-like after passing insects, catch them, and return to their perches. Though not as quick and adept at the process as true flycatchers, the jays never- theless are successful at foraging in this manner. I have never observed other species here treated to feed in such a manner. Characteristically, most or all the members of the flock engage in flycatching at once. Flycatching by the flock seems to accompany another frequent habit—sunning, in which the birds of the flock sit scattered about in the tops of low trees on exposed hillsides for periods up to an hour in length. Sunning is always engaged in at the time when the sun first appears over the hills, and seems to be most frequent in cool weather. A sunning bird sits facing the sun, with plumage fluffed, and neck drawn in, almost as if going to roost. Sunning time is a period of rest for the jays. It is not, however, the only diurnal rest period. On a clear day there is at least one other period of rest, which usually comes around noon. In inclem- ent weather, much of the afternoon may be spent by the flock, sitting in dense low trees in sheltered places. In early April when the temperature is near 40° F, one may find flocks of jays in the afternoon sitting quietly in oak scrub among boulders at the bases of cliffs or in other nooks protected from the weather. In such situations, a flock is difficult to excite, and the presence of a person merely causes the birds to call briefly and move only far enough to be away from the observer, where again the birds become quiet. Though I have never been able to discover a roosting flock, I have followed groups until nearly dark and presumed that they roosted in trees along gullies on hillsides. In my experience, the birds do not cease activity in the evenings or resume them in the morning along the canyon floors but always on the hillsides. StrupiEs oF NEw Wor tp Jays 93 The typical daily routine of a flock of Mexican Jays seems to consist of a regular sequence of events occurring at specific times. Activity of the flock begins about one-half hour after dawn at which time the birds can be heard calling from the hillsides. There is then a brief period of movement in which the birds fly to their regular sunning locale, which for each flock is fairly constant and is an exposed place on an eastwardly facing hillside. Here the birds become quiet and sit about in the low trees, where they remain, with limited movement and feeding, while the sun rises. The birds stay at this place until the sun has been up from one-half hour to a full hour before they begin to forage actively. By 9:00 a.m. the flock is in full activity and moves down the hillside as the air warms, ultimately reaching the canyon floor. Here, the birds forage in the woodland understory and on the ground. This activity of the jays continues until between 11:00 a.m. and 1:00 p.m., whereupon a period of rest occurs, lasting one to two hours. Feeding and other activities of the flock are resumed usually by 2:00 p.m. if the day is fair and continue until 5:00 or 6:00 p.m. After this, the birds begin their movements back up the hillsides along the small ravines and gullies to their roosting places. There may be a short period of flying about in the hours of evening, usually around 6:00 p. m., at which time the flocks are conspicuous. This may be a period of re- organization of the members of the flock in preparation for the movement to the area where they will roost. Food-storing.—F ood-storing is apparently uncommon in Mexican Jays in the wild, although my captive individuals frequently stored food. I watched a flock of Mexican Jays in the Chiricahua Moun- tains engage for several days in carrying small green apples far up the mountain sides from an apple orchard in a canyon. But I was unable to discover what the birds did with these fruits. Some of these small apples evoked no reaction from two Mexican Jays when placed in a cage with them. Gross (1949:241) noted that Mexican Jays in the Santa Rita Mountains of Arizona readily came to food placed on a feeding shelf near his camp. The birds were especially partial toward bread broken in large pieces. These they frequently carried off two or three at a time and stored in crevices in trees. One jay was ob- served to fly to the ground and cover its piece of bread with leaves. Under normal conditions in the wild (no artificial food supply being present) it seems probable that Mexican Jays store food less 94 THe UNIVERSITY SCIENCE BULLETIN frequently than do Blue Jays. It is a common sight to observe the latter species store food, but under similar natural conditions in the wild, one seldom observes food-storing by Mexican Jays. INTERSPECIFIC RELATIONSHIPS Interspecific Relationships of the Blue Jay Blue Jays are typical corvids, bold, inquisitive birds, often in- volved in mobbing of predators or in other strife. They are quick to lead a fight against predators, yet are themselves predators and are generally recognized as such by other species of birds. Relationships with nonpredators—Blue Jays quarrel most fre- quently with Robins (Turdus migratorius). The latter, in the re- gion of Kansas, in suburban habitats nest in close proximity to the jays, often in the same trees. Robins are usually the aggressors in interactions with jays, attacking them on little provacation. The jays seem always nonaggressive toward adult Robins and barely avoid their plunging assaults. I have not seen Blue Jays robbing nests of other species, but an acquaintance brought to me two fledgling Mourning Doves (Zenai- dura macroura) she had rescued from attacks by two Blue Jays. The literature, of course, contains references to the habits of Blue Jays of killing young birds, devouring eggs, and indulging in other acts of predation. The frequent alarm reactions of Robins as well as Tufted Titmice (Parus bicolor), Brown Thrashers (Toxostoma rufum), Baltimore Orioles (Icterus galbula), and other species toward Blue Jays is further evidence of the status of this species in the avian community. Relationships with predators—The arousal of mobbing behavior in the Blue Jay varies considerably in procedure. Often, it develops in stages, each having an associated posture or display and vocaliza- tion. The initial stage is often one of curiosity (marked by erect, investigatory posture, Fig. 4, and the kut kut or kuet kuet call). This may be followed by anxious uncertainty (bobbing display with the wheedle-eee call), developing quickly to the stage of “identifica- tion” of the predator (bobbing with the cleeop call). “Identifica- tion” leads to “realization” of danger (aggressive and intense in- vestigatory postures and displays with the jayer and jaay calls, the latter associated with greater excitement than the first). The jay then sounds the call of alarm (jaay, jaay, jaay, given repeatedly), whereupon a group of jays assembles within minutes. Mobbing Srupies oF NEw Wor tp Jays 95 then takes place, the birds diving at the predator, giving the jeer jeer call, and engaging in bobbing display. Blue Jays mob in groups of three to 12 individuals. Mobbing behavior is characterized by an aggregation of aroused jays dis- persed at distances of 20 to 50 feet from and surrounding the object of attack. The birds usually remain above the predator. If it is the nesting season, and the predator is near the nest of a pair of jays, the pair owning the nest will approach the predator much closer than do other jays. If the predator is reluctant to leave, the attack may reach a point where the two jays dive at the enemy. In my experiments using a stuffed Long-eared Owl and a stuffed Sharp-shinned Hawk, jays always struck the specimens several sharp blows on the nape, in the course of mobbing, but usually made passes to within a few inches of the heads of the mounts. In such attacks, a jay moves to a place above the predator and within approximately ten feet of it. From here the jay drops si- lently, pulls up just short of the predator, and rises to a safe distance screaming, jeer jeer. Such an attack usually provokes the remainder of the mob to loud vocalization, which includes a variety of jay calls plus the brrrr?r call and, less frequently, the cleeop and wheedle- eee calls. The several calls uttered indicate the variety of responses exhibited toward the predator. The increased excitement that occurs when a jay attacks is at least in part due to the resulting movement of the head and body of the predator as it avoids the jay. If the predator remains com- pletely motionless, as was the case with the stuffed specimens (except when the wind caused motion) the intensity of the mobbing reaction will eventually regress. Blue Jays can become conditioned to the stuffed specimens; although the birds will al- ways show initial concern when confronted by mounted predators, the excitement subsides more quickly each time the jays are sub- jected to the artificial stimulus. When a predator flees, Blue Jays follow but do not mob in the air. If a predator alights only a short distance away, the jays con- tinue to mob it. If, however, the predator flies high or leaves the woodland where the mobbing is occurring, the attackers do not follow, and the mobbing reaction subsides quickly. The individual jays then disperse quickly to their respective ranges. Though displacement and redirection behavior are not well- developed phenomena in the Blue Jay, an example of each is oc- casionally seen at the time of mobbing. Displacement bill-wiping 96 THE UNIVERSITY SCIENCE BULLETIN (the autochthonous pattern follows feeding and, sometimes, preen- ing) is undoubtedly indicative of conflict between tendencies to attack and flee. In this behavior, the bird quickly wipes the sides of the bill back and forth several times on the limb or other perch on which he stands. Redirection behavior in the presence of a predator occasionally is directed toward another jay that is not participating in the mobbing. Most frequent victims of the Blue Jay near Lawrence are the woodland hawks (Accipiter spp.), and the Great Horned Owl (Bubo virginianus). The only mammal that I have seen Blue Jays molest is the Fox Squirrel (Sciurus niger), which probably is a predator at the nests of jays. Strangely enough, the reaction toward the squirrel is totally different from the reaction toward the hawks and owls. Blue Jays never mob squirrels and ignore them for the most part outside the nesting season. When a squirrel enters the vicinity of a nest of a jay, the birds fly to the animal and silently follow it. While following and literally “herding” the squirrel, the birds occasionally dive at and snap their bills at the animal, but utter no sound except low intensity kut or kuet notes. Because the attacking birds do not give alarm calls, the assembly of a mob does not occur. Blue Jays show no fear of squirrels. Occasionally they swoop upon a squirrel that is not close to their nest and harass it for no apparent reason. In comparing the mobbing behavior of Blue Jays with those of other species of jays discussed herein, it should be emphasized that Blue Jays typically strike the predator in mobbing; this contact is forceful, terminating a dive by the attacker. In the breeding sea- son, close approach to the predator is made usually by only two birds, the owners of a nearby nest. Interspecific Relationships of the Mexican Jay The Mexican Jay holds the same community status as the Blue Jay, sharing with it certain behavioral characteristics of both pred- ator and prey. Relationships with nonpredators——tThe relationship between the Mexican Jay and the Western Wood Pewee (Contopus richardsonit ) is similar to that between the Blue Jay and the Robin. Interaction between these two common inhabitants of western woodlands oc- curs only in the breeding season and seems to stem from apparent concern on the part of the flycatchers for their nests, eggs, and young. Srupres oF NEw Wonrtp Jays 97 Gross (1949:241) noted that at a feeding tray, Mexican Jays com- pletely dominated other birds that came there except the White- winged Dove (Zenaida asiatica). He also observed possible preda- tion by a flock of jays on the eggs of the Solitary Vireo (Vireo solitarius) and the young of the Blackheaded Grosbeak (Pheucticus melanocephalus). According to my observations, Mexican Jays are less antagonistic toward other species of their community than are Blue Jays. Since the Mexican Jay is intraspecifically social to a high degree, its activities less often give cause for excitement in individuals of other species. Relationships with predators—Mexican Jays are more easily in- cited to mobbing reaction than are Blue Jays. Mexican Jays also remain in an excited state for a greater period, whether or not the object of mobbing is alive, or remains to be mobbed. This greater and more prolonged excitability seems to be a function of greater sociality in the species than in the Blue Jay. The jays tend to stimu- late each other to a greater degree than is possible in less social species. The high degree of sociality also ensures a larger mobbing contingent, also resulting in greater excitement. A stuffed Long-eared Owl placed in a prominent place within the foraging range of a flock of Mexican Jays, stimulates not only the members of the nesting pair, or pairs, to mob, but the entire flock resident in the area as well. It is impossible, after mobbing begins, to distinguish between the members of a pair owning a nearby nest and the other jays engaged in mobbing. All birds mob in approxi- mately equal degree, all come close to the predator, and all give vent to loud calls of alarm—the reek reek and ruik ruik vocalizations. Mexican Jays show less tendency than do Blue Jays to follow a predator if it flies from one part of a woodland to another, unless the predator remains in sight and moves only short distances at a time. Nevertheless, if the predator leaves or is removed (as in the case of stuffed specimens ), the jays remain where the object of mobbing was and continue to call. Mexican Jays also associated me with stuffed mounts of predators and frequently followed me, calling ex- citedly, in the several days after they had observed me carrying one of the specimens that they had been mobbing. The only time in the nesting season when the presence of a pred- ator seems unlikely to cause mobbing by a large number of Mexican Jays is during incubation. At this time, a stuffed specimen of a predator placed near an incubating female excites her, but may 4—5840 98 THE UNIVERSITY SCIENCE BULLETIN not arouse other jays to mob, although they are in the vicinity. Other jays, in such instances, will sometimes approach to nearby trees and give an occasional reek call of alarm, but in my experience will not mob. I have previously stated that a female Mexican Jay that I observed in incubation was nearly deserted by the resident flock. The failure of members of this flock to mob coincides with their failure to exhibit other types of attentive behavior at the nest during incubation. At the same time, the incubating female vigor- ously attacked mounts of predators near the nest. In these attacks she struck the specimens. Behaving much as a Blue Jay does under similar circumstances, she perched above the predator and then dived silently to deliver a sharp jab at the nape, after which, calling loudly, she rose quickly to a safe position. Mobbing reaction by a flock seldom seems to include contact by any jay with the predator. The jays typically perch 10 to 20 feet away from the object of mobbing and scream in alarm. Tail-flitting and bobbing are characteristic displays in mobbing by Mexican Jays. The bobbing action is not the exaggerated performance typical of the Blue Jay in courtship and is closer to a bowing action than is any bobbing display by the Blue Jay. The emotional gamut, and correlated calls, displays, and postures, exhibited in predator recognition and mobbing seems more re- stricted than the corresponding behavioral complex of the Blue Jay. Discovery and recognition of a predator by just one Mexican Jay seldom occurs, since the birds seldom travel alone. The stage of “recognition” of the predator comes about sooner after discovery than it does in the Blue Jay, and the general calls of alarm, reek reek, are given soon after discovery. Variation in calls thereafter is slight, involving mostly change in intensity of delivery. The reek call may give way in greatest excitement to ruik ruik, combination of ruik and reek—truik-reek, ruik-reek (sounding as if the ruik were an inhaled sound, the reek an exhaled one). Mexican Jays exhibit strong displacement behavior when excited by a predator. Jays thus aroused wipe the bill back and forth across the limb on which they are perched when mildly excited (displace- ment bill-wiping) and either hammer on the perch or scatter leaves or food about with the bill when greatly aroused (displacement for- aging). All of these activities are functional in feeding behavior. The birds hammer open hard food-objects, scatter grain or other items of food about in selecting or searching for food, and clean the bill after feeding by wiping it across the perch. The action most Stupres OF NEw Wor tp Jays 99 closely associated with feeding (hammering) is the one employed redirectionally in greatest excitement, while that action employed in the time of mild excitement is the least closely associated with feeding of the three redirection activities. It seems that the more completely aroused and frustrated is the bird in the situation at hand, the more completely is it divorced from appropriate motor patterns and transported into patterns functionally typical of an- other phase of life. Redirection activity under conditions of stress is frequent and like that of the Blue Jay; a jay in the midst of mobbing a predator chases briefly after a fellow jay. Reaction toward predators other than birds of prey is similar to that described above. Captive Mexican Jays reacted quickly to a rubber mold of a rattlesnake (Crotalus sp.) even though it was a pale yellow in color and resembled a snake only in outline and tex- ture. The birds would not react toward a strip of white paper cut along the same outline as the plastic mold. Interspecific Relationships of the Scrub Jay Comments on agonistic behavior of the Scrub Jay will be re- stricted to discussion of the reaction of the species toward predators. Whereas Blue and Mexican jays react violently toward living and stuffed mounts of predators, my limited observations indicate a different behavior of the Scrub Jay toward the mounted specimens. In the Sandia Mountains, New Mexico, in the early phase of the nesting cycle, mounts of a Long-eared Owl and a Sharp-shinned Hawk placed near the nest of a pair of Scrub Jays provoked no mobbing reaction. The jays coming to the nest after either mount was installed on a nearby branch flew directly to the nest tree, ex- amined the mount carefully, and then usually went to the nest. On subsequent visits to the nest, the birds would hop to the predator, give the kwesh kwesh call of alarm occasionally, but then cease call- ing and hop onto the head of the specimen and peck in a calm, investigatory manner at its eyes and feathers. After but a brief time the jays would resume their activities, paying no further attention to the mount. When I placed the anterior half of a dead Scrub Jay be- tween the talons of the owl, the birds again examined the mount carefully. Then one of the jays hopped to a branch beneath the owl, reached up and with its bill, extracted the piece of dead jay, carried it off over a nearby ridge of ground, and dropped it. This behavior would be considered novel under any circum- stances in most species of passerines. But in comparing the Scrub 100 THE UNIVERSITY SCIENCE BULLETIN Jay’s reaction toward these mounted predators with the reaction at the same time of year (the nesting season) of Mexican Jays and Blue Jays, the behavior is even more puzzling. The strictly terri- torial behavior of Scrub Jays compared with the irregular ter- ritoriality of Blue Jays and the lack of territoriality of Mexican Jays probably accounts for the failure of the Scrub Jays to sound loud assembly calls attracting other jays to the locality of the predator. But the failure of the jays to become at all excited under the cir- cumstances is difficult to explain. The attachment of the birds to the nest site was at this time strong, their behavior was not at all vague but was intense expression of nestbuilding and courtship be- havior. The rarity of Sharp-shinned Hawks and Long-eared Owls in the area seems an improbable explanation for relative lack of ex- citement of the jays, since both species of predator represented by the mounts are also uncommon in the Chiricahua Mountains where Mexican Jays react by intense mobbing display when confronted with the same two specimens. In addition, Great Horned Owls and a Hawk similar to the Sharp-shinned Hawk, the Cooper's Hawk ( Accipiter cooperii) would seem to offer ample conditioning to the form of the two mounted birds as predators. Altmann (1956:241-53) studied the reactions of 39 species of birds, including the Scrub Jay, to predators and models of predators. In 19 of these species (not including the Scrub Jay) no mobbing reaction was ever given even under the influence of mobbing reac- tion of other species. Altmann records encounters of Scrub Jays with owls on three occasions, once with a Short-eared Owl (Asio flammeus) model and twice with Screech Owl (Otus asio) models. The birds did not mob the first species and mobbed the second species both times, but only secondarily after mobbing of the owls had been instituted by other species. Where Altmann recorded no mobbing of the models (for example, with towhees and thrashers ), the birds, in addition, gave no indication that they saw the models; thus, he accumulated little data on failure to respond because of the difficulty in determining whether the models had actually been seen. In the case of the Scrub Jays observed by me, of course, the models were seen. Hartley (1950:315) in his experiments used mobbing response as the criterion of recognition of a predator. Under this criterion, Scrub Jays did not recognize as predators models of the Long-eared Owl used in the present study and of the Short-eared Owl em- ployed by Altmann. Dr. Richard F. Johnston and Dr. Henry S. Stupies OF NEw Wor tp Jays 101 Fitch (personal communication ) recall observing Scrub Jays mob- bing Great Horned Owls (Bubo virginianus) in California (where Altmann’s studies were performed). Thus, it seems possible that Scrub Jays give a predator response to Horned Owls because of some conditioning to that species as a predator which conditioning they have not received in experience with the other two species (with which they almost undoubtedly come in contact). Neverthe- less the peculiar behavior of the Scrub Jays studied by me indicate that further investigation of the reaction of the species toward predator species is needed. SOCIAL ORDER IN CAPTIVE MEXICAN JAYS A social peck-order existed in the four captive Mexican Jays studied by me. The expression of this form of sociality was best observed in the evening hours, when the birds were preparing to roost, for it was in conflict for and selection of the most favorable roosting sites in the cage that the clearest expression of dominant and subordinate behavior was expressed by the jays. As Tordoff (1954:350) observed in captive Red Crossbills ( Loxia curvirostra), competition among the jays for roosting sites was severe. Thus, there were more instances of conflict and other interactions, which enabled me to see the social relationships between any two indi- viduals based on many observations in a short time. Increased activity of both Mexican Jays and Red Crossbills at dusk in cap- tivity may be an expression of similar activity in the wild in both species; Tordoff (1954:350) suggested that in the crossbills there is reorganization of the flock after feeding has ceased, in the subse- quent flight to roosting areas. Establishment of Peck-order Peck-order was not established in these Mexican Jays until several months after they had been placed in the cage. Failure to establish a social order immediately was caused by the slowness of the birds to adapt to captivity, resulting in unnatural behavior at all times. Thus, the birds roosted for awhile wherever they chanced to be when darkness fell. Method appeared in their behavior by approxi- mately the fourth month in captivity, and the adoption of daily routines included the appearance of a peck-order, methodical roost- ing, feeding, and other behavior. A rigid peck-order was established by the spring of the birds’ first year in captivity. This order was of the straight-line type with no triangles of dominance. This type of order may be characteristic 102 THe UNIVERSITY SCIENCE BULLETIN of the species, but a larger number of birds in the cage might have introduced complexities not expressed with only four birds. The order, although it went through several changes, always consisted of one dominant bird, two middle birds, whose statuses in the order changed twice and were nearly equal, and a low bird, which re- mained at the bottom of the order throughout this study. The peck-order was maintained by peck-dominance rather than peck- right, interaction between dominant and subordinate individuals always involving aggression on the part of both birds, but always terminating in retreat of the subordinate individual. Aggression involves jabbing with the bill while the mandibles are being snapped together rapidly, pecking at the carpal region of the wing, at the tarsi, and at the head and bill of a defender. Usually, aggressive interaction is given only when a subordinate bird ap- proaches to within a distance enabling the dominant, without moy- ing its feet, to peck the subordinate. No Mexican Jay ever seeks out his social inferior in order to attack it. Interaction between a dominant and a subordinate bird is practically stereotyped and proceeds as follows (see Fig. 19): The two birds perch side by side, the subordinate standing erect, the dominant bird crouched; either bird may initiate the display. The crouched bird usually lunges at the erect bird, mandibles snapping furiously, and pecks at the tarsi of the subordinate. The erect bird gapes, stands even more erect and then feebly returns the aggression by suddenly crouching and pecking weakly at the tarsi of the dominant. The dominant bird does not move its feet at all, but the subordinate may frequently move about on the perch. No calls are given by either bird. The dominant bird then usually pecks at and grasps feathers in the car- pal region of the wing of the subordinate bird. The dominant then pulls at these feathers, and the subordinate draws back, gaping, and fluttering. The subordinate then may again feebly peck at the dominant bird, which in return even more strongly pecks at the tarsi and carpal feathers of the subordinate. If the latter does not then flee, the dominant bird continues pecking at the head, bill, and eventually grabs the bill of the subordinate and pulls until the latter bird flutters free and flees. The above is a description of a complete interaction between a dominant and a subordinate bird. The activity may be terminated at any time by the fleeing of the subordinate. While such activi- ties usually took place at roosting time they were sometimes insti- tuted in conflict over food or water. Stupies OF NEw Wor p Jays 108 Most interactions took place between the highest and lowest members of the order. The lowest bird seemed habitually to want the roosting place of the highest bird in the order. The choice of roosting place of the dominant bird varied somewhat from week to week, depending probably upon weather conditions. His choice affected the choice of roosting place of all his subordinates. In clear, warm weather, the birds roosted in the open on a horizontal bar, except for the lowest bird in the order, which went to roost last. Although ample space for roosting remained on the bar, the lowest bird always roosted where it happened to be when darkness fell. This was usually in a thicket of branches at the back of the cage. Moreover, while his fellows faced toward the east (presum- ably toward the direction of the morning sun), the lowest subor- dinate faced no certain direction—as often as not toward the west. Behavior of the lowest subordinate in roosting was like that of all four birds before they became adjusted to captivity. Thus, per- haps, the basis for the low position of this bird in the peck-order was his failure to adjust to captivity. Final roosting places were usually fixed by approximately fifteen minutes before darkness, but the activities of competition resulting in the final positions usually began approximately one-half hour before sunset. At this time the birds ceased feeding and began to assume roosting positions for periods of a few seconds to a min- ute at a time. The dominate bird was always first to display this behavior and selected what to the other jays became the most desired position for roosting. To my eyes this position was hardly different from many other situations in the cage. The position was usually on a horizontal bar held in place by two upright posts and was against one of these uprights. Three other positions existed on the bar next to one of the uprights. Two of these (on either side of the south upright) were never used. The third, on the north side of the north upright post, was often used by the second bird in the order. Thus, the north upright post was all that sepa- rated the dominant and his immediate subordinate from each other. The third bird in the order usually roosted on the bar to the north of the second bird. When the birds roosted in the west end of the cage beneath a canopy and on the branches of a limb fastened there, their choice of positions was less consistent. The dominant bird might have several interactions with the sec- ond and third birds in the peck-order in one evening. But long before darkness, the first, second, and third birds in the order had 104 THE UNIVERSITY SCIENCE BULLETIN assumed their eventual roosting positions. Each of these birds took its position again and again, only to leave it to fly about for a few minutes and perch in other places. The lowest bird in the order continued until the end of preroosting activity to fly wildly about, landing near each of the other birds, and when they were absent from their roosting sites, settling momentarily at these places. The principal difference between the lowest subordinate and the other birds is that the former never took advantage of opportuni- ties to “better” himself in the social order and as regards favor- able roosting sites. He occasionally succeeded in pushing one of the others from its roost or in getting situated at one of these places when a bird was temporarily absent from the perch, but if not chased away by any of the higher birds in the order, he deserted the place of his own accord. The three dominant birds seemed always less excited and displayed confidence in their activities as compared to this lowest subordinate. Tordoff (1954:353) noted that in crossbills there were fewer encounters between birds adjacent in the peck-order, except that the despot male was active in dominating the second-ranking bird. In Mexican Jays, the despot interacted most with the lowest-ranking bird, while the birds of second and third rank interacted with no definite preference with each other and with the lowest subordinate. Occasionally it seemed that the dominant bird of the flock delib- erately left its perch and then resumed defense of it against the other three birds, one of which (usually the lowest subordinate ) always seized these opportunities to take over this desired roost- ing place immediately. Table 1 is a summary of interactions be- tween members of the captive group of Mexican Jays, observed on evenings in April, May, June, September, and October. The lowest subordinate bird of the flock exhibited frustration by a consistent displacement reaction correlated with his inability to gain a roosting place, food, and water. In this reaction, he re- tired to a perch in a far corner of the cage and hammered with his bill on the perch. When excited by a person, he also engaged in this behavior. The reaction, it will be recalled, is the same as that of wild Mexican Jays when mobbing predators, but never observed by me in other species of jays. The other captive Mexican Jays infrequently indulged in this reaction, which is probably an indi- cation of their greater emotional stability or confidence. StuprEs OF NEw Wokr tp Jays 105 The Peck-order At the beginning of observations concerning peck-order in my captives, the female RL was probably dominant, since she habitu- ally fed the lowest subordinate, WW, a male. RL’s position was taken over within a week after the beginning of courtship feed- ing, however, by RR, a fully adult male, who thereafter, until his death, was the dominant bird among the captives, feeding WW in courtship, and maintaining dominance at the food tray, water container, and most importantly in interaction concerning roost- ing sites. From April, 1957, until June, 1957, WL was the second bird in the social order. In roost selection activities, this second position was much inferior to the dominant position and barely superior to the third position. In feeding and drinking activities, the second position was barely inferior to the first position and clearly above the third position. On June 22, RL assumed the roosting position previously occupied by WL, and when the latter came close, RL pecked at him viciously, causing WL to flee. After RR died in summer, 1957, a prolonged conflict occurred between the second and third subordinates, WL and RL, for the first position among the three remaining birds. This conflict was still taking place on September 26, in the evening of which day WL seemed to dominate RL and finally assumed the roosting place for which they had competed. WW continued to fly about and was pecked several times by both RL and WL. Thus, WW re- mained the lowest bird in the order. To my knowledge, the TasBLE I.—A Record of Observed Conflicts of Peck-dominance in Captive Mexican Jays Based on 65 Encounters Subordinance in an encounter Percent of conflicts Aggressive in which dominance in - my, _|Total instances bird was RR RL WL | WW Lay kaa tea pies an encounter of dominance dominant Ed ere ato yAS cil cv. Ae « 5 5 39 49 98 15 Dia AS eee eae ae a ereee tg 2 7 10 53 NV Layaecrrca tc neidh., cha bem 0 Dehli ce at 2 6 46 VA iceeMtgs abcd o¥ cree 0 0 Ove exe trses 0 00 Total instances of subordinance.... 1 9 a 48 65 (total encounters) 106 THE UNIVERSITY SCIENCE BULLETIN social order of these birds did not change again during the life of the captives. I have previously discussed (p. 34) under the significance of courtship feeding and its relationship to social order, the possible reasons causing feeding of a male by a female in captive Mexican Jays and its meaning in social order among these jays. Fic. 19.—Illustration of phases of conflict between dominant (bird on right in each pair) and subordinate (left). Upper left, subordinate flies to perch beside dominant preparing to roost and assumes erect posture. Upper right, dominant first pecks at tarsi of subordinate, which attempts to back away but then may feebly return peck at dominant. Lower left, dominant then seizes carpal feathers and pulls, causing subordinate to flutter or pull away. Subor- dinate then may again peck feebly at tarsi of dominant or may flee. Lower right, dominant grasps bill of subordinate and pulls strongly, causing subor- dinate to pull away and (usually) flee. StupiEs OF NEw Wortp Jays 107 MISCELLANEOUS BEHAVIOR Investigatory Behavior in the Blue Jay Investigatory behavior in the Blue Jay is displayed toward strange animals, new food supplies, or other objects arousing curiosity on the part of a jay. This behavior is perhaps most often expressed in the investigation of food-objects in the course of foraging be- havior. A jay will fly directly toward a newly-discovered food sup- ply, veer suddenly, and alight a few feet away from it. The bird then assumes the posture of investigation (Fig. 4). The head is moved from side to side as the bird eyes the food first with one eye, then with the other. The jay then hops around the sides of the area containing the food and gradually works its way closer. Any unnatural movement by the object or near it will cause the jay to jump quickly into the air and land several feet further away from the food. The bird will take food eventually if not alarmed, but at first will take small particles quickly and not eat them at the source of supply. Gradually, as the bird gains confidence, it may come to eat at the source of supply instead of first flying to a nearby perch. But the investigatory behavior may be resumed and the air of cau- tion again displayed at the first hint of unnatural circumstances. This cautious attitude is characteristic of the Blue Jay in most of its “vegetative activities.” It contrasts in part with the attitude of the Mexican Jay under similar circumstances. Investigatory Behavior in the Mexican Jay Mexican Jays are less wary than Blue Jays in approaching a strange object. Food placed in an obvious place attracts only one or two cautious Blue Jays, which come to the supply slowly and carefully. But in contrast a flock of from five to ten Mexican Jays may swarm over a like food supply within minutes after its discovery, all the birds beginning to feed or carry off the food after only a brief examination. The posture of investigation (erect, hopping. stiff- legged ) of the Mexican Jay is similar to that of the Blue Jay but less exaggerated. Doubtless, the degree of exaggeration of this posture is a function of the degree of wariness. No attempt is made by any individual Mexican Jay to monopolize the food; each bird selects bits of it, eats on the spot, or carries it a short distance to a perch. In the latter case, however, the food seems not to be taken quickly away in an expression of fear of the place where the food source is located. The comparative lack of caution displayed by this 108 Tue UNIveRSITY SCIENCE BULLETIN species in comparison to the Blue Jay is possibly a function of greater sociality. The few times that I have observed Scrub Jays reacting to strange objects, they also displayed little development of investigatory be- havior and approached a food supply especially provided precisely as did the Mexican Jays, although not in a flock. (It will be re- called that these same Scrub Jays reacted to a stuffed Long-eared Owl with little of the caution displayed by Blue Jays in the investi- gatory behavior. ) Anting in the Blue Jay Anting is commonly indulged in by Blue Jays according to Whit- aker (1957:251) in her thorough analysis and complete bibliography of records of anting. Blue Jays actively perform anting; that is, they pick up ants in their bills and apply them to their plumage, instead of letting them crawl up their legs. Davis (1950:518-9) observed a jay perched in an oak tree, picking ants off the leaves and placing them deeply beneath the contour feathers of the body. According to Laskey (1958:214) anting Blue Jays she observed rested on their tails while applying ants to the primaries; afterward one jay rubbed a dry leaf along the primaries. Ivor (1943:52) believes that ants are al- ways applied to the ventral surface of the outer primaries, from just below the wrist to the tip of the wing. According to Whitaker (1957:249) a pet Orchard Oriole applied ants to several parts of the posterior half of the body, bases of the rectrices, and wings. I have received from Dr. Edwin C. Galbraith an account of an adult Blue Jay that anted in Galbraith’s backyard almost daily for over a week. The bird came to the yard and stood on a small anthill, picked up the ants, and applied them to its plumage. When my two young captive Blue Jays were approximately 40 days of age, I released a host of ants (Pogonomyrmex sp.) on the floor of their cage. The birds at once became so frightened, flying headlong around in their cage, that I feared that they would injure themselves. The ants utilized here were, of course, of an unac- ceptable species, one that possesses a functional sting ( Whitaker, 1957:206). When the young Blue Jays were approximately 57 days of age, I placed a shallow tray containing about 20 ants of the genus Dory- myrmex into their cage. Ants of this genus have been found ac- ceptable for anting purposes (Whitaker, 1957:202), probably be- cause they do not sting and are odoriferous. The species of ant Stupies OF NEw Wor tp Jays 109 utilized in the present experiment was bright lemon yellow in color and was taken from beneath stones on the grounds of the South- western Research Station in southeastern Arizona. A Blue Jay exhibited interest and no fear when the Dorymyrmex were presented. Hopping near the ants crawling on the floor of the cage, the bird maintained a stiff-legged investigatory posture. With neck outstretched and crest erect, he peered at the ants and at the same time uttered a strained weak conversational note. After several false starts, the bird jabbed hard with his bill and secured an ant. The bird sometimes actually ate several ants before anting, being first stimulated to ant when one of the insects escaped from the bill and crawled over the face of the bird. Then in quick suc- cession the jay usually took several ants and tucked them swiftly under the primaries halfway to the tip of a wing. Anting in the Mexican Jay I was not able to induce Mexican Jays to ant in captivity. There are no published records of the species anting, though anting of some type is found generally throughout the family Corvidae ( Whit- aker, 1957:234-5). Release of ants (Dorymyrmex sp.) into a cage containing two juvenal Mexican Jays elicited no response from the birds, although juvenal Blue Jays anted with the same kind of ant. In one experiment, several hundred individuals of a harmless, odorif- erous species of ant (species not as yet determined) were placed in a shallow tray, the edges of which were coated with petroleum jelly to retain them. The tray was then placed on a prominent limb in the flight cage containing the four Mexican Jays. The birds im- mediately showed interest in the ants but did not ant. Instead, they ate all but the few of the insects that escaped and then set about devouring the petroleum jelly! It seems possible that Mexican Jays do not ant on the basis of this experiment and lack of positive evi- dence to the contrary in the literature. Goodwin (1953:149) concludes that the type of anting behavior is probably adaptive to different ecological situations and is not cor- related with phylogeny. Therefore, possibly, the failure of a species to ant cannot be utilized as a valid character in plotting systematic relationships. Of course, the lack of correlation of method of anting and phylogeny may not be the same sort of problem as is involved in comparing two species of birds, one of which ants and one of which does not ant. In the latter problem, anting may have phylo- genetic implications. 110 THE UNIVERSITY SCIENCE BULLETIN Tail-wagging: Preroosting Behavior in the Mexican Jay Preroosting social behavior of Mexican Jays in captivity has been discussed in the previous section concerning social behavior. A certain behavioral trait that I have termed tail-wagging is not connected with sociality in Mexican Jays but is an integral part of preroosting behavior. In tail-wagging, each of my captive Mexican Jays upon settling in its chosen roosting site drew in its neck, fluffed its feathers, rested low on the perch, and then gently wagged its tail quickly from side to side, three to six times. If the bird then proceeded to sleep, it did not wag its tail again, but if it chose to rearrange itself in any way before sleeping, it again per- formed the tail-wagging activity. None of the other jays that I kept in captivity exhibited this behavior, including the young Mexican Jays. I have not observed jays in the wild going to roost. Tail-wagging may be a subtle method of plumage arrangement connected with “getting comfortable” for roosting, or an expression of “contentment” or “satisfaction” connected with the termination of activity and the coming to rest for the night. I have found no references in the literature to similar behavior in other birds. Tail- flicking and wing-flicking, which are considered by Andrews (1957) to be intention movements of flight, would not seem to be homolo- gous to tail-wagging, since certainly the most remotely possible intention of tail-wagging is flight. The fact that I have failed to observe the behavior in other jays might be of significance in sys- tematic considerations. PART III—PHYLOGENY AND SYSTEMATICS SoME ASPECTS OF THE PHYLOGENY AND SYSTEMATICS OF New Wor tp Jays The preceding discussion of the behavior of jays has led to a consideration of phylogeny and systematics of New World jays. The following account is based also on behavioral investigations of other workers and information on morphology and distribution. Existing Classifications New World jays exhibit a marked diversity in morphology, be- havior, geographic and ecologic distribution. Early classifications (cf. Ridgway, 1904; Sharpe, 1909; Hellmayr, 1934) necessarily were based on the superficial aspects of these four factors; a large number of genera and species were proposed for the western hem- isphere. Current views on classification of these birds are but STupies OF NEw Wor tp Jays 11] slight modifications of those of the earlier authors. Detailed study of the osteology and myology of jays provides a picture of their systematics wholly different from that established by the above criteria. Studies by Ashley (1941) and Hudson (1955) emphasize the muscular and skeletal uniformity within the Corvidae. x| oO] x S|) 4 x] x ca = pd bad = o| = a4] E2(h ESI) ESI x Of Of Oo] of Of Of < 0 3c] || Oo) Ol Nelaiaaay ol x! SO] of S| of S| d of oc} Cl ec] ol Ole xj Oo a < 5 al ay w O° SS = (32 o = > S i - ahs Sie 3 Se one °o 2a < ¥ e 22 0 = o = ° z@ “ _ - Sad ate” (OL. = ° © Gueone 5; > s—] = 79) is ae = S =) d a? =) x aa a x 7) 4 ac 4 . oO aq OF 36 oS = > - “<0 2 «2g 2 € Feo = 2 ee < — La) ae for z = Og O ow o ° oe Our es Nie TOS ov > 2 ogc o OF © Clie co iS - Oo =, a = x z ° ae. SS Heme G a = oro Oo ° se —ae ho}. Co of ro) ° uw? 3g 42 x oO = IS) OF SB 5 22 2 a of > 2 ° 24.2235 3 Go po eee 2 lL OH @ Oe eee Ome! Co OO 165) aca’ ad aq<«o 139 STUDIES OF NEw Wor tp Jays *(9x9} 998) «I—Ppipsos,, *n “yw 04 A[dde A[qeqoid myono0d *n ‘YW JOF BOY pajJOU SONsIIaJOVIVYD “GC *G 9}0U-}00F UI pajoU seBe ydvoxe satoads 9Yy} Jo ,sdnois epipiog puv eurlieWRea[y,, 94} JO stoquiout [Tv 0} A[dde A[qeqoid apuozup ‘n “Y IOF a1aY Pa}JOU SOIjsIIoJORIVYD "fF *saroads ay} JO ,.sdnois TrasnoypooAA puke BOTULOFITeD,, 294} JO saovr [[e 0} A[dde A[qeqoid masnoypoom ‘9 “vy 1OF a1aY Pd}OU SONstIo}ORVIeYD “E ‘ONSHo}OVIvVYD VY} Buliwaq satoads auo IOF A1OSZ9}V9 a}yeIedas vB Burjda19 UvYy} JeYy.eI snuas ay} OF po}sI[ Oo SI OJSltoJOvIeYO JY} ‘s1aUBBUOD s}I 0} IRIS st asked YyoORa Ur sor1dods 94} 90UIS ynq ‘Baus asay} JO Satoads auO A[UO UI MODO0 s9oLQde1 paddij-azGAA °F 14971998 DZIOUDAD pue snulysojisg Ul soop ysowye jJusustd yons se ‘s}iede10f 94} JO uUioWed aseuNjd jo aoUapPIAa [[B 9INOsqO O} JYSNOY} SI sWIOF osoy} Ul JUSWUZId O1UR[aUI JO AISUdJUT 9YT, “TL “‘SnUDIIXAW SNUIYLODIS J 1Of “(Z9S'GEGT) YAS s ‘Maysaaq vydopss1y OF (ZOT:9Z6T) PHONG a ‘snaopjo1a *Q IO} ‘(SISZI9IGL) P4oyDa “spouA *D IOFZ ‘(OE LL8T) Wouusga ‘pyoajps ~spouh “DQ IOF “(ZE:PEGL) AAeUaH{o “Wayoip “Dp IOF ‘(9EST:SS6T) e100 :sruiffp xpsooourAD 10F “(06S!SO6T) UY[[Va *(OLISS6T) YAS v ‘uonsenb ut satoads ay} oj }x0} oy} UT ATTRoyIOads 0} patiajor Useq jOU dARY YOIYM SONsMejovIvyo OF (MOTEq) SUOT}e}O 0} JoJoI a[qGe} VY} UT SIoWe] drs -BIdng “oljstiajyovIeYO sTy} spriesor se AjPTIGeUvA IO} VOUBPIAI = suOTeUTqUIOD yons J9y}O IO O/X ‘9oIMOS 9[qRI[eI Iay}O IO dIM}eIo}y] =O IO X ‘UOTRAIOSqO [euostad = Q 10 YX ‘a0MMOS 9JIGeITaI YPM UOTROTUNUIUIOD [BUOSIod JO 9IN}zeIO -}[ Snjd uoyearasqo [euosied WoIZ UONvOIpU =©0 10 X :SMOT[OF SB SUOTIPUOD [efoeds 10 uOTeULIOFUL JO aoINOS JO sodA} oJRoOYpUT syIeUL [RUONIppe pue soy} JO suOT}VUIqUIOS SB [JAM SB (OSHo}ORIeYO JO VoUVSGe = S.QO ‘ONsTIeJORIeYO Jo uotssassod = s x) pedAoydura sfoquids o1seq ay} JO SULIOF JUSIOBIC] *AeyTUNTS JO UONvoIpUr [eloues vB Se ATUO BAIOS Uv 4yNq AT[eIa}P] ATeyosqe pojosds9zur oq jouUvS sadIpUI sy J, ‘setoads jeY} IOF o[quieAe sem aTqQe} BY} JO AIOB9}eO yore oF UOTeULIOFUI VAT}IsOd Vsnvoaq UOSTIedUIOD IOF sIseq ay} Sv pasn sem Avf anfq ey, ‘safoods oy} JoF a[qupeae sem UONRULIOJUT YOIYM UT sat108e}eV9-oNsT -IopORIeYO JO Ioquinu [e}0} 9Y} Aq PapPIAIP Y7N}s1U9 DPIIOUDAD YM sooise YT YOIYM UT sonstajyovreyo fo Joquinu oy} Safoads yors IO; st 9[qe} 9Yy} Jo UUuINTOO WSU ey oy} UT AjIepIUNSs Jo xopur syy, ‘shel pfIOA, MAN eWOS JO soNst19} -ovreyo [eIOIAvyeq pue [evorsojoydiow jo 9[qe} sAeIeduloo yYW—' % ATAV], Tue University SCIENCE BULLETIN 140 usaMjoq jORJUOD AIBpuodasS syuasoides A[qtssod sty} ‘sdnoi3 sorsedsqns om} esau} ‘(_,dnois yseryorums,,) yspsyovwns 9 “y pue (,dnois mesnoypoo,,,,) s1oupho suaosajnia09 “WV UdEeMjJEq UOHeprAsIoWUL OF 9oUBpIAD poHUM] St o19YJ, ‘SoUT[ aseyy AQ Pe}DoUUOD sedvI UBEA\Joq SINI90 uoNeprIs19} -UT [PRIORI JY} 9}VOIpUT souTT AABAA “TUTWIODOTAYdYy equy ‘shel ppIOM MON FO , SUT] 9JeUIOUT,, ot} FO AuasolAyd pasodoig—'cZ “S17 AVP Q1YOM MIN 1V¥LSSONY — YOOLS VWNODO013HdY (960d Bulmojjo} aas) ,aNI7 3LYNYO, YO , 3NI7 SLYNYONI, ee HO9OLS SN3ZOSIINYSOD VWNODOO0 13HdV WOOLS VNIYVWNVYLIN VNOD013HdY VLOWSH / —~ 4K T . .N =: ‘ ‘ inewuriotuns (, 84noa6 asm op Soom ve Sy eee eae SN39S31NY309 uo Da1uVoyI1d9 i-Iyaiauo ‘SIT1O VS ‘1Y3GIM1IOM= SVNOZIYYV Pp 4 i) " Q S,, SN39S31NYI09'V SN39S31NY3A09'V “WHONOID'N'V VNINWVWWYLIN’'Y YOTOSINN'Y Stupies OF NEw Wor tp Jays 14] cristatella. However, the special social nature of Psilorhinus, and the fact that behavior of most species of Cyanocorax is poorly known, makes an actual recommendation for taxonomic alterations of these three genera seem unwise to me at this time. Although I have not placed Cyanolyca in the above tribal classifi- cation, there can be no doubt that it does belong therein and per- haps. unites the two tribes. Its morphological affinities seem to be more with the Aphelocomini than with Cyanocorini, but be- cause of almost complete lack of knowledge of behavior in the genus, there can be no choice but to await further study before assigning it a place in the classification. Figures 25 and 26 depict the hypothetical evolutionary relation- ships of the genera of jays here discussed, as conceived at this time. Footnotes to figure 26 summarize evolutionary events that have possibly occurred in the “Ornate line” based on morphological and behavorial data discussed above. SUMMARY Studies on behavior and phylogeny of New World jays are de- scribed in this paper. Differences in courtship and reproductive behavior of species of jays seem to be correlated with differences in sociality. Courtship in the Blue Jay (Cyanocitta cristata) consists of three stages, group courtship, courtship feeding, and false nestbuilding. The Mexican Jay (Aphelocoma ultramarina arizonae ) is a highly social race, and thus, although courtship feeding and false nestbuilding occur, all the members of a flock participate in these activities. True nest- building is, in this race, a part of courtship. In the Blue Jay, true nestbuilding arises from false nestbuilding; both members of the pair participate in building the true nest. Blue Jays are territorial, the behavior first being expressed after nestbuilding is commenced. Both members of the pair build the nest. The Mexican Jay in Arizona is not territorial and has narrow ecologic requirements; the Scrub Jay (Aphelocoma coerulescens ) throughout most of its range is territorial and can live in much poorer (more xeric) areas than does the Mexican Jay. Nonterri- torial Scrub Jays in Florida have less ecological adaptability than the rest of the species. Lack of territoriality may thus be an adap- tation to restricted ecologic adaptability. In the Blue Jay copulation occurs after Nudging display in the late stage of nestbuilding. Only the female of the pair, there- THE UNIVERSITY SCIENCE BULLETIN 142 9014} SI WieIseIp yUasoid oY, (849440) XVYOIONVAD ‘auty AreuoTNNyoAs pasodoid yove ul spued} [eIOIAvYyoq pue [eorsojoydiow pasoddns ured -x9 JY} SoejOU}OOF BSurpuodsar100 0} JoJo s[eIOUINNY ‘“SUTAT] [[e 91v SOUT] JO spue ye soureu Aq poe}eorpur sdnois yey} os ‘[euoTsusUIp V11aLlvisiys (VONITOUN=) XVYOIONVAD Av? Q14¥OM MAN WYLSSINV fo) Wwuod TVYLSAONV Ol ol SNNIHYO11Sd VHdOTISSI9 (490468 snapiuoso 5, ) XVYOIONVAD ‘tut000uRAT) ‘shel POA, MON JO OUI] 9}eUIO,, IY} JO 901} OTjoUaZO[AYd [eoKey}OdAY W—'gZ “SI (a6od Bulpsaceid aas) wJNI71 JLVNUONI,, VLilIDO01V9 VLILISONVAD 143 StupiEs OF NEw Wor tp Jays "BOLIaULY YNOS UL [eordo.y Bururewoar fuoTeIO[OO antq [eBINJoNI}s Bursoy yOu ‘AjpeyuoAZ ydvoxa ysaio Sutsoy {10[09 [Iq JO ssoupar0[Oo-yivd pasuojoid sSururejat SuMmouyUN AjPTVIOOS Jo 9aIs9qI “ZT *‘jeordoy SuIUIeWaI ‘OOIxaW puRe ekolLIoWy [eRQUa0 OUT YWOU Burpeoids ‘syreda1of JO Utoyed JOOS oIseq Burpwaaer sny} ‘(UMOIG Burwmios9eq) UOTBIO[OD any [BANjoNAYs Bursoy ‘AT[eyUOIF Jdvoxa ysoid Surlsoy f10j09 ][[Iq jo wsrydiowrp ose pasuojoid Bururejor ‘[eoos Burlurewwoy “TT F (OS! FEET ‘1AeuTIPH ) [Izeig jo purypelqe} Suynrqequr ‘paaimooai-ayesuo0[a puUe VAISUd}X9 Y}JOG }saID ‘SULIOF YsNqoL surureyor {(¢) yso, 10[OO ][[Iq jo wsrydiourp ese pasuojoid f‘umouyun AT[RIN0G “OT ‘osje premy}ioU Surpeoids ynq sapnjzjzel x[pprut ur peordo.y Sururewor {Zurjstsied 10;OO |[[Iq jo wisrydiowip ase jyusurMOId ‘[eIoOs BuIuIeWloYy “6 *‘jeordo1} Sururewoi }sa10 aAIsua}xe ‘OYs Sururey -91 SWIOF YsNGOI ‘][Iq JO ssouUpaiO[Oo-HIed pasuojoid Sururejor ‘fAz[eIOOs Burureyoy *g ‘BolIoULy [BIQUeD jo suTe}JUNOW pue OorxeW Teordonqns 0} premyOU Surpeoids {(a[qeuea ATYSIY ySet0) 3sSa10 aAISUd}Xa BUTSOT Sysnqor Sururewier f][Iq JO ssoupai0[oo-H1ed posuojoid Sururejzoi SAPTBIOOS Bururejoy “ZL ‘sapnyye[, e[pprlur ur UoNNIystp ur [eordo1 Suturewear ‘}s910 AAIsUg}xa NG YOYs Bururejot foyeurour ATsANe[eI SBururewiet ‘Ay yeIOOs SBurureyoy “9 *qnios plie-rwias ‘s}so1oy UreyUNOW [RordoIZqns UL OdIxaJ_ Ul YOU pue voTIaWIW YyNOS wiayqyI0U I9AO Burpeaids ‘[][Iq JO ssaupar1O[Oo-HIed [eUuaAN{ysod SurIso, ‘usteyVd A0[OO Burjzsesy -u0d A[PIAIA pue suUlIOF JsOUL UT ysaio yYSsIIdN oa}esuO;a—a}eUulO sIOUL BuUIWIODIg ‘GC *sapnyzyey a[pprur jeordo1 ‘fu0fF ysnqor ‘asie, Bururejol ‘}sa10 JO WUOF [BYS90UB ‘aATSUd}Xa ‘}IOYS ay} Butssassod ‘10[09 ][[Iq Jo wistydiowlp ose pasuojoid Bulurejzel SAYPeIOOS BUIUIeJOY “fF ‘OOIXOWY UOY INOS puB Uso}s9M ‘sUOIZaI [eoIdOoNGNS SunIqeyUr fulI0; Jsnqo1 asiel Sururejor $(0z200uvDADn UvYyY JasU0C]T [[Iq 94} ApIsuL ssaupaito[oo-1z1ed sty} BuUTUTeJoI ynq) ‘][Iq JO ssaupasojoo-Hied [euaan{}sod SuIso[T fAPTBIOOS a}eIPAULIaJUI SBUIUIvJOY “E *(14a72]938 *D) OorxayW{ JO SUTEJUNOW BIA sapnyyne, WiayyNos BuLlayUa A[Lepuodas ‘suOoTZaI [ve10q pue azeioduwia} sulioyua faseumyd paieq suldojaAsp ‘uoF UI YsNqOI ssa_T pue 9ZIS UT JaT[VWIs Burwmooaq ‘][Iq jo ssoupaio[oo-ied [euaAnhsod SuUISO, ‘[BIIOPII9} BuIWMOdsIg °F ‘sordoy jo ynO puv pieMY}IOU Surpeoids $yso10 a}yeZu0[e Surldo[aAap ‘10[09 [Iq JO WistydiowIp ase BUISOT ‘[BIOOS ssa_T BurMOsag *T ‘UOIINGLySIP [eoIdoN 'f% 2sIMsy Ul pazeAsNIt se syiedai0y jo ulayed AO[OO oIseq fpeay Jo suoIsaI [e}IdI090 pue ‘[euOoIOD ‘TeJUOIF Bula -AOD }SaI0 WAOYS f1O[OO [[Iq ur WIsStydiolWIp ase pasuo[oid SwWAOF Ysnqor ‘TeIOOs ATYSIR ‘O 144 THe UNIVERSITY SCIENCE BULLETIN after, occupies the nest in egg-laying and incubation. The male remains close by and occasionally feeds his mate. In the Mexican Jay, the female performs incubation and is at- tended both by her mate and by other members of the flock. Com- pared to Blue Jays, Mexican Jays are not secretive when near the nest. Male and female Blue Jays feed the young. After the young leave the nest, territoriality is abandoned and the adults follow the young wherever they go. Feeding of the young continues until about the time of postjuvenal molt. Young Mexican Jays are cared for by their parents and by other members of the flock. Compared to young Mexican Jays, young Blue Jays and Steller’s Jays have greater manipulatory facility with the tongue, bathe and preen more frequently, and have a greater variety of vocalizations. Young Blue Jays sing frequently after postjuvenal molt com- mences; young Steller’s and Mexican Jays raised in the course of this study did not sing. Although Blue Jays flock in autumn, these flocks are not closely integrated as in Mexican Jays; individual distance is maintained. Migration is characteristic of first-year Blue Jays; many adults do not migrate. Winter territories or feeding areas seem to be main- tained by Blue Jays, and food-storing may function as a spacing mechanism in winter. Mexican Jays do not migrate, the flocks remaining on foraging and breeding areas the year around. Food-storing seems not com- mon in Mexican Jays in the wild, but is frequent in captive indi- viduals. Mobbing by Blue Jays is preceded by several stages of be- havior as follows: curiosity, anxiety, “identification,” and “realiza- tion” of danger. Usually only two birds (members of a pair) attack a predator. Contact with the predator occurs. In the Mexican Jay, all members of the mobbing flock react alike (unless the predator is near a nest). Mexican Jays in captivity form a peck-order. In four captives studied by me, the order was straight line and established by peck- dominance. Males dominate females and black-billed adults domi- nate subadults, which have parti-colored bills. The age relation- ship prevails over the sexual relationship. Thus, an adult female may feed a young male in captivity. A ritualized series of be- havioral events is involved in expression of peck-dominance. These Stupies OF NEw Wortp Jays 145 events include pecking by both birds at the tarsi, wings, and head of the opponent. There seem to be two phylogenetic lines of jays of New World origin; these may be termed the “Ornate line” and the “Inornate line.” Aphelocoma is the sole genus in the latter line, although Cyanolyca may belong therein. It is here proposed that degree of sociality and the characteristics of dimorphism of bill color and rate of sexual maturation are correlated. It is speculated that a high degree of sociality is primitive, since less sociality and loss of prolonged external parti-coloredness of the bill are correlated. In less social species parti-coloredness is evident in postjuvenal birds and is retained on the inside surfaces of the bill thereafter for a variable period. In the “Inornate line” many Mexican races of the Scrub Jay and Mexican Jay and the species A. unicolor are poorly known be- haviorally, but from what is known, it seems that low sociality has arisen independently in the Scrub and Mexican Jays. The lack of strict territoriality and occurrence of helpers at the nest in the isolated Florida Scrub Jay indicate that this relict race may rep- resent the primitive condition of the species in these respects. In addition, in the species A. ultramarina, the races couchii and sordida (but only in part) may be less social than the remainder of that species and A. unicolor. If age dimorphism in bill color is cor- related with sociality, transition from high sociality to low sociality occurs within the race A. u. sordida from southern Hidalgo, Mexico, northward, toward the range of A. uw. couchii. Because of lack of information on behavior, relationships in the “Ornate line” are particularly obscure, but it seems that in habits Psilorhinus is primitive, although it is specialized morphologically and not ancestral to others of the “Ornate line.” A basic color pat- tern of the foreparts is shown to be common to most members of the “Ornate line.” The “Coronideus group” of Cyanocorax serves to “unite” morphologically Psilorhinus and Cissilopha and other species of Cyanocorax. Many behavioral and morphological char- acteristics are shared by Cyanocorax, Calocitta, and Cyanocitta, the latter genus possibly being most closely related to Calocitta among living jays. Two tribes are proposed as useful taxonomic categories in New World jays; these tribes, Aphelocomini and Cyanocorini, corre- spond to the “Inornate line” and the “Ornate line,” respectively. 146 THE UNIVERSITY SCIENCE BULLETIN LITERATURE CITED ALLEN, J. A. 1905. Supplementary notes on birds collected in the Santa Marta Dis- trict, Colombia, by Herbert H. Smith, with descriptions of nests and eggs. Bull. Am. Mus, Nat. Hist., 21:275-295. ALTMANN, S. A. 1956. Avian mobbing behavior and predator recognition. Condor, 58:241- 253, 2 figs., 2 tables. Amapon, D. 1944a. The genera of Corvidae and their relationships. Am. Mus. Novi- tates, No. 1251:1-21, 1 fig., 1 table, January 24, 1944. 1944b. A preliminary life history study of the Florida Jay, Cyanocitta c. coerulescens. Results Archbold Exped. No. 50. Am. Mus. Novi- tates, No. 1252:1-22, 5 tables, January 24, 1944. AMERICAN ORNITHOLOGISTS’ UNION. 1957. Check-list of North American birds. Fifth ed. Baltimore, The Lord Baltimore Press, The American Ornithologists’ Union, iv + 691 pp. ANDREW, R. J. 1956. Intention movements of flight in certain passerines, and their use in systematics. Behaviour, 10:179-204. ASHEEY Jo F's 1941. A study of the structure of the humerus in the Corvidae. Condor, 43:184-195, 7 figs. AvupuBon, J. J. 1834. Omithological Biography. Edinburgh, Adam and Charles Black, xxxii + 588 pp. Bent, A. C. 1946. Life histories of North American jays, crows, and titmice. Bull. U. S. Nat. Mus. 191:xi + 495 pp., 68 pls. BranptT, H. 1940. Texas bird adventures. Bird Research Found., Cleveland, xii + 192 pp., frontispiece, 16 pls. CHERRIE, G. K. 1916. A contribution to the ornithology of the Orinoco region. Sci. Bull. Mus. Brooklyn Inst. Arts and Sciences. 2:133a-374. Davis, D. E. 1942. The phylogeny of social nesting habits in the Crotophaginae. Quart. Rev. Biol., 17:115-134. Davis, M. 1950. A Blue Jay, Cyanocitta cristata, anting. Auk, 67:518-519. Dickey, D. R., and VAN Rosse, A. J. 1938. The birds of El Salvador. Field Mus. Nat. Hist., Publ. 406, zool. ser. 23:1-609, 24 pls., 29 figs., March 21, 1938. Epwarps, E. P., and Lea, R. B. 1955. Birds of the Monserrate area, Chiapas, Mexico. Condor, 57:31-54, 8 figs. StupiEs OF NEw Wor.p Jays 147 FickEN, R. W., and FIcKen, M. S. 1958. Head-scratching in Seiurus (Parulidae) and other passerines, Ibis, 100:277-278. FRIEDMANN, H. 1930. The Sociable Weaver Bird of South Africa. Nat. Hist., 30:205- 212, 10 photogr. Goopwin, D. 195la. Some aspects of the behaviour of the jay Garrulus glandarius. Ibis, 93:414-442, 2 figs. 1951b. Some aspects of the behaviour of the jay Garrulus glandarius. Ibis, 93:602-625, 3 figs. 1952. A comparative study of the voice and some aspects of behaviour in two Old-world jays. Behaviour, 4:293-316, 2 figs. 1953. Interspecific differences in the anting movements of some corvine birds. Ibis, 95:147-149. 1956. Further observations on the behaviour of the jay Garrulus glan- darius. Ibis, 98:219, 9 figs. Grimes, S. A. 1940. Scrub Jay reminiscences. Bird-lore, 42:431-436, 3 photogr. Gross, A. O. 1949. Nesting of the Mexican Jay in the Santa Rita Mountains, Arizona. Condor, 51:241-249, 3 figs., 1 table. Hart ey, P. H. T. 1950. An experimental analysis of interspecific recognition. In Sym- posia of the Soc. Exp. Biol., No. IV:313-336, 1 pl., 4 figs., 2 tables. New York, Academic Press Inc., viii + 484 pp. HELLMaypR, C. E. 1934. Catalogue of birds of the Americas. Part VII. Field Mus. Nat. Hist. Publ. 330, zool. ser., 13:vi + 531, November 15, 1934. Hickey, J. J. 1952. Survival studies of banded birds. Special scientific report: ( Wild- life) No. 15:177, 71 tables. Washington, U. S. Fish and Wildlife Service. HiInpe, R. A. 1956. The biological significance of the territories of birds. Ibis, 98:340-369. Hupson, G. E., and LANnzitxorrT1, P. J. 1955. Gross anatomy of the wing muscles in the family Corvidae. Am. Midl. Nat., 53:1-44, 35 figs. Ivor, H. R. 1943. Further studies of anting by birds. Auk, 60:51-55, 1 pl. Lack, D. 1954. The natural regulation of animal numbers. The Clarendon Press, Oxford, viii + 344 pp., col. frontispiece, 52 figs., 39 tables. Laskey, A. R. 1958. Blue Jays at Nashville, Tennessee movements, nesting, age. Bird- banding, 29:211-218. 148 THe UnNIversiry SCIENCE BULLETIN Mayr, E. 1943. Criteria of subspecies, species and genera in ornithology. Annals New York Acad. Sci., 44:133-139, art. 2. Mitter, A. H., FrrepMaNnn, H., Griscom, L., and Moore, R. T. (“Co-editor Group” ) 1957. Distributional check-list of the birds of Mexico. Part II. Pacific Coast Avif., No. 33:436, 7 pls. Berkeley, California, Cooper Or- nithological Club. Moore, R. T. 1938. Discovery of the nest and eggs of the Tufted Jay. Condor, 40:233- 241, 4 figs. Moyninan, M. 1955. Types of hostile display. Auk, 72:247-259, 2 figs. NicuHo.s, J. T. 1955. A criterion for young-of-the-year in the Blue Jay. Bird-banding, PX RON. PITELKA, F. A. 1945. Pterylography, molt, and age determination of American jays of the genus Aphelocoma. Condor, 47:229-260, 13 figs., 9 tables. 1946. Age in relation to migration in the Blue Jay. Auk, 63:82-84. 1951. Speciation and ecologic distribution in American jays of the genus Aphelocoma. Univ. California Publ. Zool., 50:iv + 195-464, 14 pls., 21 figs., 62 numbered tables, July 20, 1951. PITELKA, F, A., SELANDER, R. K., and ALVAREZ DEL Toro, MIGUEL. 1956. A hybrid jay from Chiapas, Mexico. Condor, 58:98-106, 5 figs., 1 table. RAnp, A. L. 1937. Notes on the development of two young Blue Jays. Proc. Linn. Soc. N. Y., No. 48:27-59, 2 tables, October 31, 1987. Riwpeway, R. 1904. The birds of North and middle America. Part III. Bull. U. S. Nat. Mus., No. 50:xx + 801 pp., 19 pls., 72 figs., 115 text tables. Ritter, W. E. 1938. The California Woodpecker and I. Berkeley, California, Univ. California Press, xvi + 340 pp., col. frontispiece, 1 pl., 28 figs. SALVIN, O., and GopMaAn, F. D. 1879-1904. Biologia centrali-Americana. Aves. Vol. I. London, xliv + 512 pp. Scott, W. E. D. 1886. On the breeding habits of some Arizona birds. Auk, 3:81-86. SENNETT, G. B. 1877. Notes on the ornithology of the lower Rio Grande of Texas, from observations made during the season of 1877. Bull. U. S. Geol. and Geogr. Surv. of the Territories. 4:1-66. SHANNON, G. R. 1958. Magpie’s rapid replacement of dead mate. British Birds, 51:401- 402, SHARPE, R. B. 1909. A hand-list of the genera and species of birds. Vol. V. London, xx + 722 pp. Stupies OF NEw Wortp Jays 149 Srmumons, K. E. L. 1957. The taxonomic significance of the head-scratching methods of birds. Ibis, 99:178-181. Skxutcu, A. F. 1935. Helpers at the nest. Auk, 52:257-273, 1 pl., 1 fig. 1953. The White-throated Magpie-Jay. Wilson Bull., 65:68-74. Sturcis, B. B. 1928. Field book of birds of the Panama Canal Zone. New York, Lon- don, G. P. Putnam’s Sons, xl + 466 pp., col. frontispiece, 22 pls., 107 figs. SUTCLIFFE, A. 1926. Breeding of the Beechey Jay (Xanthura beecheii). Avicult. Mag., 4:102-3, Sutton, G. M. 1951. Mexican birds. Norman, Univ. Oklahoma Press, xvi + 282 pp., 16 col. pls., 65 figs. Sutron, G. M., and GiBert, P. W. 1942. The Brown Jay’s furcular pouch. Condor, 44:160-165, 4 figs. TINBERGEN, N. 1955. Psychology and ethology as supplementary parts of a science of behavior. In Group Processes (pp. 75-167, 19 figs.), Trans. First Conf., 1954. Ed.—Bertram Schaffner, Josiah Macy, Jr. Found., 334 pp., 40 figs. Torvorr, H. B. 1954. Social organization and behavior in a flock of captive, non-breeding Red Crossbills. Condor, 56:346-358, 1 fig., 4 tables. Van TYNE, J., and Sutrron, G. M. 1937. The birds of Brewster County, Texas. Univ. Michigan, Misc. Publ. No. 37:1-119, 5 pls. WETMORE, A. 1926. Observations on the birds of Argentina, Paraguay, Uruguay, and Chile. Bull. U. S. Natl. Mus., 183:iv + 448 pp., 20 pls. WHITAKER, L. M. 1957. A résumé of anting, with particular reference to a captive Orchard Oriole. Wilson Bull., 69:194-262, 1 pl., 5 figs., 3 tables. WirHersy, H. F., Jourpain, F. C. R., Tickuurst, N. F. and Tucker, B. W. 1938. The handbook of British birds. Vol. I. London, H. F. & G. Witherby Ltd., xl + 326 pp., 32 col. pls., unnumbered text figs. YAMASHIMA, M. 1938. A sociable breeding habit among timaliine birds. Proc. 9th Intern. Omni. Cong., pp. 453-456. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vout. XLIT] DECEMBER 29, 1961 [No. 8 Some Morphological and Functional Aspects of Certain Structures of the Middle Ear in Bats and Insectivores ; BY O’DELL W. HENson, JR. Department of Anatomy, University of Kansas Asstract: ‘ _ . ’ . Bs ' 4 £ . . : lV : [ee Fee ‘ he ul 3 ooh ae \ ; i PS Ae A i Ff - cif WAS als Ve r 4 « *.* te a 7 ns i ’ nt iy Wiis te : nd ‘ 7 ok en hod ' 7 dy r : A ! D “san a, 4 ae hee a. ' < i is “a —s ¥ y. geal i a) ya mae | Bs i [> “\ Sh, hats = Ay; © P as Pea Dery calor facs x 7 af P Te _ i T ri ‘Hy al cad A f 4 - ‘ : ; “| 7 . 72 “te by THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vot. XLIT] DECEMBER 29, 1961 [No. 4 Some New Species of Rhagovelia from the Philippines (Veliidae, Heteroptera) * BY HerRBert B. HUNGERFORD AND RyvuricHt MATSUDA In preparing a report upon the aquatic and semiaquatic Hemip- tera taken by the Chicago Natural History Museum during their Philippine Zoological Expedition (1946-47), we have found so many new species of Rhagovelia that we are submitting this as a separate report. Until now the following five species have been described from the Philippines: Rh. minuta Lundblad (1936), Rh. philippina Lundblad (1936), Rh. orientalis Lundblad (1937), Rh. luzonica Lundblad (1937) and Rh. teretis Drake (1948). With Dr. Lundblad’s splendid paper “Die altweltlichen Arten der Veli- idengattungen Rhagovelia und Tetraripis” (Arkiv for Zoologi, Band 28A, No. 21, 1936) in which he illustrated all species and gave keys to all species for both sexes, and his “Einige neue oder wenig bekannte Ostasiatische Rhagovelia Arten” (Entomologisk Tidskrift, Haft 1-2, 1937), also well illustrated, we thought the task of iden- tifying the philippine Rhagovelia would be an easy one. To our disappointment, all eight species are new. In describing them we are mindful of the variations in color and structures in the species of Rhagovelia from all over the World. Three of them belong to the subgenus Neorhagovelia Matsuda not before recorded from the Philippines. Rhagovelia (Rhagovelia) lundbladi n. sp. Gigs 2as toch) Size Length of body Width of head Width of thorax Apterous male... ...°4.%:: 4.2 mm. 0.8 mm. 1.36 mm. Apterous female ........ 4.3 mm. 0.84 mm. 1.47 mm. Macropterous female .... 4.8mm. 0.84 mm. 1.86 mm. Color. Apterous forms: Head, metanotum, abdominal tergites, inner third and outer margin of connexivum dark brown. Anterior * Contribution No. 1078 from the Department of Entomology, University of Kansas. This paper is a by-product of a project supported by a grant from the National Science Foundation. 9—5840 (257 ) 258 THe UNIversIry SCIENCE BULLETIN lobe of pronotum yellowish; anterior half of posterior lobe of pro- notum yellowish; anterior half of posterior lobe of pronotum dark brown with a pale median longitudinal line, posterior half brown to yellowish brown. Venter brown to yellowish brown. Under- side of first antennal segment, underside of femur and tibia of all legs, coxae and trochanters, acetabula, dorsal base of hind femur, underside of connexivum, and a median stripe on dorsal side of connexivum yellow to yellowish brown. Last three antennal seg- ments and dorsal side of legs dark brown. Macropterous forms: Pronotum with more brown pits, hemelytra with black veins and brown membrane. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 50: 25.5: 32: 31. Female antennal formula: Ist: 2nd: 3rd: 4th:: 45: 26: 31.5: 27. Male leg formulae Femur Tibia Tarsus TRSROTEE [See ee a a 60 62 ripest! wing (5 ey [Pes As oes ee ee re be Rg 98 78 ae ivi) (ee at a nae ie 98 85 tee 4 4 Ae 20 Female leg formulae Tape) Gres) ere era 56 58 18(1 tH a + 8) REM em Not 62 oe. ob cess eib ce AE 90 69 br gas [Ty iets Bl toys Bran ee ee es 87 89 10(1 4 oy: 20 Pronotum much longer than an aye, rounded behind, exposing metanotum on sides and behind its rear margin. Apterous male: Anterior trochanter unarmed but somewhat hairy. Front tibia hairy, as broad distally as femur, with a longitudinal comb-like ridge on inner surface near distal end. Hind trochanter armed with a number of short pegs. Hind femur incrassate and hairy, armed with pegs as shown on Fig. 2, b, c. Hind tibia curved as shown on Fig. 2, b. First genital segment shorter than last ab- dominal tergite. Parameres symmetrical, and as shown on Fig. 2, d. Apterous female: Hind femur slightly incrassate and armed as shown on Fig. 2, e. Connexivum erect beyond third abdominal segment and somewhat overlapping seventh abdominal tergite, which is more than twice as long as its basal width. Comparative notes. The shape of parameres are closest to that of Rh. maculata Distant from Nigeria, but Rh. maculata is black and the armature of the hind femur is entirely different from that of Rh. lundbladi. Data on types: Described from apterous male holotype, allotype and paratypes (6¢ ¢, 39 @), and one macropterous female morphotype, all bearing labels “C. N. H. M. Philippine Zool. Exped. SoME New SPECIES OF RHAGOVELIA 959 (1946-47) H. Hoogstraal leg.,” “E. slope, Mt. McKinley, Davao Province, Mindanao 46,” “Stream through original forest” and “Elevation 3000 ft.”; three apterous males (paratypes) labelled “Mainit, E. slope, Mt. Apo, Davao Province, Mindanao, 43800 ft. XI, 46,” and “Stream through original forest.” Holotype, allotype, female morphotype and some paratypes are in the Chicago Natural History Museum. Other paratypes are in the Francis Huntington Snow Museum, University of Kansas. Rhagovelia (Rhagovelia) cotabatoensis n. sp. (Bigs 3,74 to £) Size Length of body Width of head Width of thorax Apterous male ......... 3.15 mm. 0.735 mm. 1.09 mm. Apterous female ........ 3.57 mm. 0.714 mm. 1.26 mm. Color. Body black with gray overcast and scattered golden hairs. Anterior lobe and sometimes posterior margin of pronotum yellow to reddish yellow. Lateral half of connexivum including margin brown. Antennae and legs black except basal half of first antennal segment, all coxae, trochanters, basal half of front femora and basal fourth of hind femora which are pale yellow. Propleura and underside of connexivum pale yellow. V-shaped spot on venter of mesosternum, all of metasternum and abdominal venter except last segment usually dark brown to nearly black with gray overcast. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 35: 29) Female antennal formula: Ist: 2nd: 3rd: 4th:: ia se 2220: Male leg formulae Femur Tibia Tarsus pvontnlecwr:4 fn Aneta as) SN NE 44 46 IS@eEes 3) Mic dewlecgeee wa Ste Veit sree ke aes 75 54 4: 18: 31 isifrrel Tiere Roe eae lk a saan 68 57 5(1+4+ 2): 15 Female leg formulae HEOMEC Car ey ee ner eee te 43 44 eX hte 2+ 3) MMidcledlogy wis (fo sslechr finn elle 72 55 3: 20: 3 iA Baral 1 en ee 63 59 5(142 of Pronotum much longer than an eye, rounded behind exposing metanotum on sides and behind its rear margin. Apterous male: Anterior trochanter unarmed but hairy. Front tibia hairy, as broad distally as diameter of femur, with the usual longitudinal black ridge on inner surface near distal end. Hind trochanter armed with some pegs. Hind femur only moderately incrassate and armed, as shown on Fig. 3, b, c. Parameres slender, as shown on Fig. 3, d. Apterous female: Hind trochanter with one or more 260 THe UNtversiry SCIENCE BULLETIN pegs. Hind femur nearly as incrassate as in male, but its armature is different (Fig. 3, e). Connexivum narrow, obliquely erect in first four segments, then constricted and nearly vertically erect to overlap last two abdominal tergites and with a dense mass of hairs at apex. Seventh abdominal tergite more than twice as long as its width (Fig. 3, f). Comparative notes: This species is near Rh. lundbladi, but smaller. The hind tibia of the male is straight, not curved as in Rh. lundbladi and the parameres are more slender. This species appears gray while Rh. lundbladi is brown. Data on types: Described from apterous male holotype, apterous allotype and eight apterous paratypes (4 ¢ g¢ ,4 2 9 ). All bear the following labels: “C.N.H.M. Philippine Zool. Exped. (1946- 47) H. Hoogstraal,” “Conel Buayan, Cotabato Province, Mindanao 100 ft. XII, 46,” “Stream through grassland.” Holotype, allotype and some paratypes are in the Chicago Natural History Museum, and some paratypes are in the Francis Huntington Snow Museum, University of Kansas. Rhagovelia (Rhagovelia) hoberlandti n. sp. (Fig. 2, g to 1) Size Length of body Width of head Width of thorax Apteroussmale ve ..>...... 3.57 mm. 0.735 mm. 1.26 mm. Apterous female ........ 3.53 mm. 0.735 mm. 1.26 mm. Color. Brown to nearly black above except anterior lobe and rear margin of pronotum and outer half of connexivum which are yellow. Sides and venter brown to nearly black with a gray frosty covering except propleura and ventral margin of connexivum which are yellow. Antennae and dorsal surface of legs dark brown. Basal half of first antennal segment yellow above and almost entirely pale yellow below. Basal half of front femur above and nearly all its venter yellow. Middle femur with a longitudinal yellowish stripe beneath. Hind femur at base dorsally and nearly all its venter pale yellow. All trochanters, coxae and distal ends of acetabula pale yellow. Venter of last abdominal segment in both sexes light brown to nearly yellow. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 41: 21: 25: 24, Female antennal formula: Ist: 2nd: 3rd: 4th:: 37: 20: 22: 21. Male leg formulae Femur Tibia Tarsus Bironievle tan Ar emrg y eee ks oy fo 45 53 16(1 + 2+ 8) Middleleg sc... enw a eI ed he 35 61 5: 23: 34 Hind leg ..... ee Peer oe FOO 76 10(1+ 2): 18 SoME NEw SPECIES OF RHAGOVELIA 261 Female leg formulae Shame cee. a See eee 44 53 TA(Gle= eo = 1=s3)) Mid dleslecmet aie on. sincere: bleh ca: 70 50 4: 18: 34 aml 1G ee ie 60 63 9(1+4 2): 18 Pronotum much longer than an eye, rounded behind exposing metanotum on sides and behind its rear margin. Apterous male: Anterior trochanter hairy. Distal end of hairy tibia about as wide as diameter of femur, with a longitudinal comb-shaped ridge on inner distal edge (one seventh of the length of tibia). Hind tro- chanter with one or two large pegs and several smaller ones. Hind femur incrassate, armed with pegs or spinous pegs as shown on Fig. 2, h, i. Underside of femur hairy. Hind tibia thickened near base of distal third and armed with two rows of stouter pegs; dis- tal end of tibia with two stout pointed spines as shown on Fig. 2, h. Last abdominal tergite shorter than two preceding tergites and its caudal end plainly broader than its base. Venter of first two ab- dominal segments carinate. Venter of last abdominal segment cov- ered with golden hairs, distal third with a median hairy elevation and a depressed area on either side. First genital segment hairy, very short ventrally, and dorsally not as long as last abdominal seg- ment. Parameres as shown on Fig. 2, j. Apterous female: Connexi- vum broad, obliquely raised, outer half of each segment nearly verti- cal beyond third segment; marginal tufts of long hairs on the margin of fourth and sixth connexival segments; only sixth and seventh abdominal tergites a little longer than their basal widths. Distal half of eighth abdominal tergite slightly elevated and hairy. Venter of seventh abdominal segment longer than preceding segment. Hind femur as shown on Fig. 2, 1. Comparative notes. This species, like Rh. lundbladi, is dark brown in color but smaller in size. The last abdominal tergite of the female is relatively shorter and broader than in Rh. lundbladi. The shape of parameres will separate this species from Rh. lund- bladi (see Fig. 2, d and j). Data on types. This species has been described from the male holotype, female allotype and four paratypes (2¢ ¢,29 ¢ ), all apterous and bear the following labels: “C.N.H.M. Philippine Zool. Exped. (1946-47) H. Hoogstraal leg.,” “Dimaniang Busuanga Is. P. I. (Calamianes group) nr. sea level III. 47” and “small pool beside stream.” Holotype, allotype and two paratypes are in the Chicago Natural History Museum. Two paratypes are in the Francis Huntington Snow Museum, University of Kansas. 262 Tue UNIversiry SCIENCE BULLETIN Rhagovelia (Rhagovelia) usingeri n. sp. (Fig. 3, g to 1) Size Length of body Width of head Width of thorax Apterous maleie 5 sane 3.57 mm. 0.756 mm. 1.26 mm. Apterous female ........ 3.36 mm. 0.756 mm. 1.26 mm. Macropterous male ..... 4.20 mm. 0.756 mm. 1.65 mm. Macropterous female .... 3.99 mm. 0.796 mm. 1.47 mm. Color. Body black with a faint gray overcast dorsally, except anterior lobe of pronotum and margins of connexivum yellowish to reddish yellow. Thoracic pleura and abdominal venter black with a gray overcast. Basal half of first antennal segment, coxae, trochanters of all legs, basal half of front femur, and base of hind femur yellow. Hemelytra dark brown to black, with a basal longi- tudinal white band reaching beyond caudal end of pronotum. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 42: 26: 27: 26. Female antennal formula: Ist: 2nd: 3rd: 4th:: 42: 25: 27: 25. Male leg formulae Femur Tibia Tarsus LOO Eg aes ge 50 57 16(14+ 2+ 8) Muarcddlemecitr se S: oot 76a! ls Ne 79 67 4: 21: 87 TCE OWN. Scpi te ee ari, ene h sb comb 80 76 5 (AD) ee Female leg formulae On tMles, op tes - ys Sten gars 49 53 15(1+ 2-43) Niclerleme; ss. ch tee oe eee ee 78 60 4; 22: 37 ARTI ees Fee's 2h E592. Nat) ed ER ee 77 76 6(1-+ 2): 18 Pronotum in apterous forms much longer than an eye, rounded on caudal margin, exposing metanotum on sides and behind its rear margin. In macropterous forms the shoulders of pronotum not prominent and caudal angle blunt. Male: Anterior trochanter unarmed. Front tibia hairy, as broad distally as base of femur, its longitudinal black ridge on rear distal margin about one fourth length of tibia. Hind trochanter with a number of stout pegs. Hind femur incrassate and armed as shown on Fig. 3, h, i. Hind tibia curved. First genital segment a little longer than last abdominal segment and covered with golden hairs. Last abdominal tergite a little longer than the preceding. Last abdominal ventrite longer than first genital segment and with a low longitudinal carina, trans- versely concave on ventral surface of first genital segment. Para- meres rather large and broad as shown on Fig. 2, j. Apterous fe- male: First three connexival segments nearly erect, thereafter erect to slightly overlapping abdominal tergites and first genital which is distinctly depressed. Basal width of last abdominal segment SoME New SPECIES OF RHAGOVELIA 263 (dorsally ) to its length as 1.5: 2. Hind femur much more slender than in male and armed as shown on Fig. 3, k. Comparative notes. This black species with yellow to reddish yellow markings is similar in appearance to several other species. However, parameres are long and broad, unlike those of any spe- cies that have been figured by Dr. Lundblad. It cannot be Rh. teretis Drake, because the species is “neither brownish black” nor “ferrugineous beneath.” Moreover, the venter of the male is not brownish black nor “narrowed and sharply ridged ventrally,” and the connexiva meet above the seventh segment. Data on types. Apterous holotype, apterous allotype, morpho- type (1 ¢,29 2 macropterous) plus apterous paratypes (8 ¢ 3g, 39 9 ) bear following labels: “C. N. H. M. Philippine Zool. Exped. (1946-1947), H. Hoogstraal leg.” “Caburan, Caburan, Davao Prov- ince, Mindanao, sea level 1:47” and “Small pool beside stream.” Also paratypes (73 3, 492 9 apterous), morphotype (3% 2, 52 @ macropterous) bear the labels: “C.N.H.M. Philippine Zool. Exped. (1946-47), F. G. Werner leg.,” “Barungkot Upi, Cota- bato Province, Mindanao, 1500 ft. 47” and “Stream through original forest.” Other apterous paratypes (8 ¢ g¢, 892 @ ) carry the fol- lowing labels: “C.N.H.M. Philippine Zool. Exped. (1946-1947), H. Hoogstraal,” “Mainit, E. slope Mt. Apo, Davao Province, Min- danao, 4300 ft. XI. 46,” and “Stream through original forest.” Rhagovelia (Rhagovelia) mindanaoensis n. sp. (Figs +5. tetol mi) Size Length of body Width of head Width of thorax Apterous male ........ 2.94 mm. 0.756 mm. 1.18 mm. Apterous female ........ 2.94 mm. 0.756 mm, 1.26 mm. Color. Dorsal surface of body black except for a reddish yellow transverse, rectangular spot on anterior margin of pronotum. An- tennae and legs black except base of first antennal segment, distal end of acetabula, coxa, trochanter and basal half of front femur, hind coxa and trochanter which are yellow. Thoracic pleura and abdominal venter black, covered with frosty gray. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 38: 20: 28: 21. Female antennal formula: Ist: 2nd: 8rd: 4th:: 40: 22: 23: 21. Male leg formulae Femur Tibia Tarsus Bromteleo hee beh ie eh re oo tone 45 45 15(1+2-+ 38) Michelle sleet face F cinns. Pee coe stg «Se ees, S85 6 all 57 8: 24: 32 Hinmieswenta eee rt 62 63 9.5(1+ 2): 18 264 Tue UNrversiry SCIENCE BULLETIN Female leg formulae Bronte: See eee 47 46 15(1 4. 2458) Middle legs] -3.: Se, eee ee 74 56 7: 20: 85 HindSleg 24 4.460 EAL! thee eer ae 63 70 3(1+ 2): 17 Pronotum much longer than an eye, rounded on caudal margin, exposing metanotum on sides and behind its rear margin. Apter- ous male: Anterior trochanter unarmed. Front tibia moderately hairy, not as broad distally as diameter of femur, with an unusually long black comb-shaped longitudinal ridge on inner margin near distal end. Hind trochanter with or without some small pegs. Hind femur moderately incrassate and armed as shown on Fig. 5, g, and quite hairy. Venter of last abdominal segment bare, with a median longitudinal ridge and a lateral and caudal hairy margin. Venter of first genital segment with a median longitudinal carina as shown on Fig. 5, h. Apterous female: Hind femur slightly in- crassate with fewer spines than in male. Connexivum flat to ob- liquely raised. Last abdominal (dorsal) segment short, basal width greater than its length. Comparative notes. This is a short and broad species like Rh. minuta Lundblad, but its parameres (Fig. 5, i, j, k) are entirely different, being not slender and not sigmoid in shape as in Rh. minuta (Fig. 1, a). Data on types. Described from apterous male holotype, apterous allotype and 74 paratypes (32 ¢ ¢, 429 ¢@ ) bearing the follow- ing labels: “C.N.H.M. Philippine Zool. Exped. 1946-47 F. G. Werner leg,” “Barungkot, Upi, Cotabato Province, Mindanao 1500 ft. °47” and “Stream through original forest.” Also 17 paratypes bearing the labels: “C. N. H. M. Philippine Zool. Exped. (1946-47), H. Hoogstraal leg.,” “Caburan Caburan, Davao Province, Mindanao, Sea level 1, ’47,” and “Small pool beside stream.” Holotype, allo- type and many paratypes are in the Chicago Natural History Mu- seum. Some paratypes are in Francis Huntington Snow Museum, University of Kansas. Rhagovelia (Neorhagovelia) hoogstraali n. sp. (Fig. 4, a to e) Size Length of body Width of head Width of thorax Apterous male ......... 3.78 mm, 0.945 mm. 1 5 limmire Apterous female ..... . 3.99 mm. 0.945 mm. 1.51 mm. Color. Black above. Pronotum with a transverse reddish yellow band behind interocular space of head. The band covers anterior three fourths of pronotum, and this spot is separated from paler SoME NEw SPECIES OF RHAGOVELIA 265 propleura by black band. Venter frosty gray except for prothorax, last abdominal segment and genital segments which may be yellow to brown. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 55: 28: 29: 28. Female antennal formula: Ist: 2nd: 3rd: 4th:: 48: 23: 26: 25.5. Male leg formulae Femur Tibia Tarsus iain ace ae nee ee 58 62 18(1+2+43) Nivclemegwrist si sistent ho ul. a5 106 82 6.5: 34: 41 isting: leprae ae ee nee mre me 90 98 4(1 +2): 18 Female leg formulae rOntelLe ce tater se eee NEL At 3 3 52 57 I7/(QUSSO BE By) Micdlemlcoat an...) ttamee ss sh! 92 73 Sol: 139 Pimcmleveropiet nes. ie seyeediie © oe 81 82 Dyes (Il ab B))ps BAY) Pronotum at middle as long as or a little shorter than length of an eye, posterior margin slightly concave. Mesonotum exposing metanotum by a wide margin laterally, and a small margin poste- riorly. Apterous male: Anterior trochanter unarmed but hairy. Tibia hairy and a little wider at its tip than base, and with a short longitudinal comb of spines on its inner distal] end. Hind tro- chanter armed with one large blunt peg and several small ones. Hind femur incrassate and armed as shown on Fig. 4, b. Prono- tum wider than long (57:11). Mesonotum wider than long (55: 42). First genital segment about as long dorsally as last abdom- inal segment. Parameres symmetrical and shaped as shown on Fig. 4c. Apterous female: Front tibia not as broad distally as in male, and lacks the longitudinal comb-like ridge of male. Trochanter with a single peg. Hind femur slightly thicker than middle femur, and its row of pegs as shown on Fig. 4, d. Connexiyvum obliquely upturned. Seventh tergite as wide as long, a little longer than sixth tergite at middle. Comparative notes. The only described species of subgenus Neorhagovelia that might some time be taken in the Philippines is Rh. esakii Lundblad from “Ishigakijima, Bannadake, Ryukyu Is- land,” but it is not a black species, male hind femur is not greatly incrassate and its armature is different from that in Rh. hoogstraali. Parameres are more slender and pointed at distal end. Data on types. Described from male holotype, female allotype and 19 paratypes (9¢ ¢,10¢ 9 ) bearing the following labels: “C.N.H.M. Philippine Zool. Exped. (1946-47) H. Hoogstraal,” “E. slope Mt. McKinley Davao Province, Mindanao :46,” “Stream through original forest.” “Elevation 3000 ft.” Besides the above 266 THE UNIverRSITY SCIENCE BULLETIN series there is a female bearing the label “C. N. H. M. Philippine Exp. (1946-47) H. Hoogstraal leg,” “Mainit, E. slope Mt. Apo, Davao Province, Mindanao 4300 ft. XI,” “stream through original forest.” The holotype, allotype and many paratypes are in the Chicago Natural History Museum. Some paratypes are in the Francis Hunt- ington Snow Museum, University of Kansas. Rhagovelia (Neorhagovelia) werneri n. sp. (Fig. 4, f to j) Size Length of body Width of head Width of thorax Apterous male ......... 3.99 mm. 0.945 mm. 1.51 mm. Apterous female ........ 4.20 mm. 0.924 mm. 1.57 mm. Color. Dark brown in dorsal view except anterior two thirds of pronotum and connexivum which are reddish yellow. Antennae and legs dark brown except for basal half of first antennal segment, and bases of front and hind femora which are yellowish to reddish yellow. Venter brown, with a dark brown to nearly black lateral stripe on either side of abdomen which is covered with a pile of silvery hairs. Structural characteristics. Male antennal formula: Ist: 2nd: 3rd: 4th:: 51: 25: 34: 82. Female antennal formula: Ist: 2nd: 8rd: 4th:: 47: 22: 29: 28. Male leg formulae Femur Tibia Tarsus Finonte lemme yeti ae neh mys eee eae 59 58 17(1+ 2+ 8) inaic [3 Net Fey Aaa ree eDOCS > i GD 98 80 6: 31: 88 Ipitavel keer 0 Geeta De reetd ete to i L 82 80 9(1+ 2): 18 Female leg formulae Hrontalecae etinies cso Posen ae 51 50 18(14+ 2+ 3) WMG eMCON es cone. . tie ine eee 83 70 4; 31: $1 rictlesaiee ee 0.5). Te eee 73 74 10(1+4 2): 18 Pronotum (median length) a little shorter than length of an eye, its rear margin slightly undulate. Mesonotum slightly concave on caudal margin exposing metanotum both caudally and laterally. Apterous male: Front trochanter unarmed but hairy. Front tibia hairy, its distal end a little wider than its base. Distal longitudinal comb about three fourths as long as distal tarsal segment. Hind leg armed as shown on Fig. 4, g. Pronotum at middle wider than long (55:16). Mesonotum wider than long (55:30). Genital segments as long as or longer than last abdominal tergite. Para- meres symmetrical, and shaped as shown on Fig. 4, h. Apterous female: Hind femur not incrassate, its armature as shown on Fig. 4, j. Connexivum with first three segments oblique and remainder SoME New SPECIES OF RHAGOVELIA 267 erect. Seventh abdominal tergite with its basal width a little shorter than its length. Comparative notes. Since this is a brown species with short pro- notum we hoped it would prove to be Rh. esakii Lundblad. How- ever, it is larger than that species, the male femur is more incras- sate and differently armed and the shape of the parameres is dif- ferent, much shorter and not pointed as in Rh. esakii. Data on types. Described from the apterous male holotype, allo- type and 13 paratypes (3¢ ¢, 109 92) bearing the following labels: “C.N.H.M. Philippine Zool. Exped. (1946-47) F. G. Werner,” “Meran, E. slope, Mt. Apo, Davao Province, Mindanao, P. I. XI 46,” and “Original forest 6000 ft. XI 46.” Also 83 paratypes (30 g ¢, 539 @ apterous) which bear the following labels: “C.N.H.M. Philippine Zool. Exped. (1946-47), H. Hoogstraal and F. G. Werner leg,” “Meran E. slope Mt. Apo, Davao Province, Mindanao 6000 ft. XI 46” and “Stream through original forest.” Holotype, allotype and numerous paratypes are in the Chicago Natural History Museum. Some paratypes are in the Francis Hunt- ington Snow Museum, University of Kansas. Rhagovelia (Neorhagovelia) minutissima n. sp. (Fig. 5, a to d) Size Length of body Width of head Width of thorax Apterous male ......... 2.1 mm. 0.63 mm. 0.798 mm. Apterous female ........ 2.14 mm. 0.60 mm. 0.84 mm. Color. In dorsal view, head and rest of body appearing gray, being black covered with gray frost except for a reddish yellow band covering most of anterior two thirds of pronotum behind inter- ocular space of head. Antennae and legs black except basal two fifths of first antennal segment, basal half of front femur, basal fourth of hind femur which are pale yellow. Venter frosty gray except trochanter, coxae, and acetabula which are pale yellow. Structural characteristics. Male antennal formula: Ist: 2nd: 83rd: 4th:: 23: 11: 17: 17. Female antennal formula: Ist: 2nd: 8rd: 4th:: 25: 12: 16: 16.5. Male leg formulae Femur Tibia Tarsus EON Tp EOE Mee Fe oye aed on Fh sas 28 28 9(1+ 2+8) iiclelemlecm rete teen! hy) SY. 50 37 Sh al ous: IBiharol- Iheiee™ as oo eee de te ea eee eee 87 34 3(1 += 2): 9 Female leg formulae HTOnmte le Ome ete ee oko crak ws a 30 30 10(1+ 2-+3) Whicldiedlers. 2 ee cn oe SUE. a 49 88 5: 15; 24 linc CME IS ore Notre oe AT 39 3(1+2): 9 268 THE UNIveRsITy SCIENCE BULLETIN Pronotum shorter than length of an eye, its rear margin concave and sinuate. Mesonotum large, broadly rounded and medially slightly concave, exposing a uniform band of metanotum on sides and rear. Metanotum and abdominal tergites with more or less complete transverse rows of black spines which are directed back- ward. Abdominal venter, margin of connexivum and _ thoracic pleura are also provided with these spines. They are more con- spicuous in male than in female. Apterous male: Front trochanter unarmed but with some long hairs. Front tibia not broader at apex than at middle. Hind trochanter with two or three small pegs. Hind femur not much incrassate and armed as shown on Fig. 5, b. Connexivum flat. First genital segment about as long as last abdominal tergite. Apterous female: First abdominal ter- gite roundly and transversely elevated; second steeply declivent caudally, third and fourth depressed, fourth laterally, obliquely excavated; fifth laterally excavated to produce a median longitudinal carina, the caudal end of which reaches the higher level of sixth and seventh tergites. Basal three segments of connexivum broad. First one flat, second one starts upward turn that becomes vertical on fourth connexival segment, the edge of which is a crescent- shaped ridge marking the constriction of abdomen at caudal end of fifth abdominal segment, beyond which connexivum is narrow and nearly vertical (see Fig. 5, c). Comparative notes. Rhagovelia minuta Lundblad was described from a single male specimen from Los Banos, Philippines. The type is 2.8 mm. long. Since Lundblad figured the right paramere which is strikingly different from any other species, we wanted to draw the parameres of this new species. Unhappily the parameres were lost while dissecting for study. However, Rh. minutissima is 2.1 mm. long and belongs to a different subgenus. Our description and figures should enable anyone to recognize Rh. minutissima. Data on types. Described from the male apterous holotype, fe- male apterous allotype and one apterous paratype which bear the following labels: “C.N.H.™M. Philippine Zool. Exped. (1946-1947) H. Hoogstraal leg.,’” Dimaniang Busuanga II. P. I. (Calamianes Group ); nr. sea level, III 47,” and “Stream through forest.” The holotype and allotype are in the Chicago Natural History Museum. One female paratype is in the Francis Huntington Snow Museum, University of Kansas. 10. ite SoME New SpEcrEs OF RHAGOVELIA KrY TO THE SPECIES OF RHAGOVELIA FROM THE PHILIPPINES Apterous males Pronotum long, covering most of mesonotum (subgenus Rhagovelia), Pronotum short, exposing mesonotum (subgenus Neorhagovelia) .. . Sizersmall@lessithan amt ONG: Hyp.ol 2 ec eee ee et Le. Sizesloneritn ame Oslin y. eerie wekbee Reig, da BE ie Stace Sey. ses ow Paramere slender, sigmoid in shape with its distal end slightly en- IKenraceya | (GLE ae) Vea Ok en ne Rh. minuta Lundblad Paramere shorter and broader, its distal end pointed. Rh. mindanaoensis Hungerford and Matsuda Color of posterior lobe of pronotum black, somewhat overcast with Ipluiistr era yan eater Pre eee Og Ce POR bic atte Boe Tet ohn oe: Color of posterior lobe of pronotum brown to very dark brown. If blackmtasrcoveredawathezoldensnainsies ee oe ieee se a ee Connexivum entirely black (fide Drake)... Rh. philippina Lundblad Connexivum at least margined with yellow or brown.............. Last abdominal tergite longer than two preceding segments, but shorter than first genital segment. Paramere large and broad. Rh. usingeri Hungerford and Matsuda Last abdominal tergite shorter than the two preceding segments, but longer than first genital segment. Paramere slender and with a broadened tip.......... Rh. cotabatoensis Hungerford and Matsuda Sizemange wateleast: 4 em; 1ONGs .F cvonck Sap Ado Fe whem terns he Moye ieye Sizessmallerathans 4mm glOm—iiets 4 Soke oc vce se ONS: so kd oh chs Venter of abdomen narrowed and sharply ridged ventrally, very densely and rather widely clothed posteriorly with extremely long golden hairs along each side of median longitudinal ridge. First geni- tal segment of male extremely large, very long, with apical two thirds bentestrongly, idownwardS «30... Jeonse. fovea oo Rh. teretis Drake WentenOlcabdomenmotsas-abOVeljs: ss cxsarseil-rste <2cckd .>... 2 See Size large over o mm. long... . 02.02.20: +s ee ea Seventh abdominal tergite longer than broad. Connexivum vertical above the end of sixth and all of seventh segment. Rh. minuta Lundblad Seventh abdominal tergite broader at base than long. Connexivum only obliquely raised laterally. Rh. mindanaoensis Hungerford and Matsuda Color of posterior lobe of pronotum black, sometimes overcast with bluish: ‘oray:ceeres eae EE oo RS Nei: ae ee Color of posterior lobe of pronotum brown to very dark brown. If black it is covered with golden hairs...) °°): 023.5. 22) eee Connexivum entirely dark................ Rh. philippina Lundblad Connexivum at least margined with yellow or brown.............. Last abdominal tergite rather broad, its basal width to its length as ORE hee nic tc la poe Rh. usingeri Hungerford and Matsuda Last abdominal tergite rather narrow, its basal width to its length as ER 5 eo acne ee ee Rh. cotabatoensis Hungerford and Matsuda Sizeilarge cat Teast 4mm. long? -2 0/0. ..2.2.. 2. eo See Sizev smaller meee his ho ee Connexivum strongly reflexed inward posteriorly, meeting above sev- enth segment, then continuing in contact to near apex, apex acuminate Or projecting posteriorly: ... ek Rh. teretis Drake Connexivum not reflexed to meet above seventh segment........... “Hind femur of female slender, with a few short spines within api- cally” (Drake). Connexivum yellowish, margined with brownish [ieee Oo a ie es A Rh. luzonicus Lundblad * Hind femur with two long spines, the longer one before middle and the other half way between it and distal end of femur, with ten or more short pegs between the longer spine and the distal end of femur and one short spine between the long spine and base of femur. Con- nexivum basally brown, with a median longitudinal band of yellow and brown margin......... Rh. lundbladi Hungerford and Matsuda A yellowish brown spot on anterior lobe of pronotum that does not reach pleura which is gray. Posterior lobe black, covered with golden Banos, N. Luzon. i taken from a river near Baguio, Luzon, and gave a brief description of them. _ wm CO R bo 8 10 * Rh. luzonica Lundblad was described from a single macropterous male from Los In 1948 Dr. Drake received one apterous male and six apterous females Since he had a male his determination should be correct, but he did not mention the length of his specimens which should be shorter than the macropterous type. tergite was not described for the female. The seventh abdominal 1a 12. SOME NEw SPECIES OF RHAGOVELIA CETUS s ait ca & cas SO een AE Rae mane a etme Rh. orientalis Lundblad Anterior lobe of pronotum entirely yellow, joining propleura which is alsonyellowre. seins foc 35: Rh. hoberlandti Hungerford and Matsuda Very small species, not longer than 2.2 mm. Rh. minutissima Hungerford and Matsuda Much larger species, longer than 3.5 mm......................... A black species. Rear margin of mesonotum medially convex. Rh. hoogstraali Hungerford and Matsuda A brown species. Rear margin of mesonotum medially concave. Rh. werneri Hungerford and Matsuda FicurE | a. Right paramere of Rh. minuta Lundblad (1936) b. Left paramere of Rh, luzonica Lundblad (1937) c. Right paramere of Rh. philippina Lundblad (1936) d. Right paramere of Rh. orientalis Lundblad (1937 ) (copied from Lundblad ) 271 72 Tue UNIVERSITY SCIENCE BULLETIN FIGURE 2 Fic. 2. Rhagovelia lundbladi n. sp. a to f. Dorsal view of apterous male. Hind leg of male. Underside of base of hind femur of male. Left paramere. Hind leg of female. Last dorsal abdominal segments of female. Rhagovelia hoberlandti n. sp. g to 1. Dorsal view of apterous male. Hind leg of male, showing the large spines on tibia. Underside of left hind femur of male. Left paramere. Last dorsal abdominal segments of female. Hind leg of female. "2 Ge oo ® SR a te 10—5840 SoME New SPECIES OF RHAGOVELIA FIGURE 2 274 THE UNIVERSITY SCIENCE BULLETIN FIGURE 3 Fic. 3. Rhagovelia cotabatoensis n. sp. a to f. eu Bane s Ge Ee Dorsal view of apterous male. Hind leg of male. Underside of base of hind femur of male. Left paramere. Hind leg of female. Last dorsal abdominal segment of female. Hien usingeri n. sp. g to |. peat aa . Dorsal view of apterous male. . Hind leg of male. Underside of base of hind femur of male. Left paramere. . Hind leg of female. Last dorsal abdominal segments of female. SoME New SPECIES OF RHAGOVELIA a5 276 Tue UNrversiry SCIENCE BULLETIN FIGURE 4 Fic. 4. Rhagovelia (Neorhagovelia) hoogstraali n. sp. a to e. a. Dorsal view of apterous male. b. Hind leg of male. c. Left paramere. d. Hind leg of female. e. Last dorsal abdominal segments of female. Rhagovelia (Neorhagovelia) werneri n. sp. f to j. f. Dorsal view of apterous male., g. Hind leg of male. h. Left paramere. i. Hind leg of female. j. Last dorsal abdominal segments of female. SoME NEw SPECIES OF RHAGOVELIA Oi. FIGuRE 4 278 THe UNrversiry SCIENCE BULLETIN Ficure 5 Fic. 5. Rhagovelia (Neorhagovelia) minutissima n. sp. a to d. a. Dorsal view of apterous male. Note the short spines on the lateral margin of the connexivum and across the abdominal tergites. b. Hind leg of male. c. Dorsal view of apterous female. d. Hind leg of female. Rhagovelia (Rhagovelia) mindanaoensis n. sp. £ to m. f. Dorsal view of apterous male. g. Hind leg of male. h. Venter of last abdominal and genital segments of male. i, j, k. Various views of the right paramere. l. Hind leg of female. m. Last dorsal abdominal segments of female. SOME New SPECIES OF RHAGOVELIA 279 Ficure 5 11—5840 Fic. 1. Melissodes (Eumelissodes) coreopsis Robertson collecting pollen and nectar from Helianthus sp. in Kansas (photograph by Dr. Delma E. Harding, Iowa State University of Science and Technology, Ames, Iowa). THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vou. XLIT] DECEMBER 29, 1961 [No. 5 A Revision of the Bees of the Genus Melisscdes in North and Central America. Part III (Hymenoptera, Apidae) * BY Wa .uace FE... LABERGE Asstract: This is the third and last part of a monographic revision of the bee genus Melissodes in North and Central America. Eighty-five species belonging to two subgenera are described. One subgenus, Callimelissodes, is newly described. Thirty-nine new species are described: Melissodes limbus, M. lustra, M. pilleata, M. elegans, M. tincta, M. subil- lata, M. gelida, M. terminata, M. bicolorata, M. bimatris, M. minuscula, M. hurdi, M. brevipyga, M. relucens, M. exilis, M. vernalis, M. utahensis, M. monoensis, M. nigracauda, M. comata, M. fasciatella, M. cerussata, M. fumosa, M. perpolita, M. paulula, M. expolita, M. pexa, M. micheneri, M. appressa, M. interrupta, M. lutulenta, M. ochraea, M. paucipuncta, M. plumosa, M. clarkiae, M. rufipes, M. pullatella, M. crocina, M. tribas. Fifty-four names are relegated to synonymy and six names remain as nomina dubia. In an addendum, an additional new species, M. haitiensis, belonging to subgenus Ecplectica (treated in the first part of this revision, LaBerge, 1956) is de- scribed, INTRODUCTION This is the third part of a revision of the bees of the genus Melissodes in North and Central America. The key to the sub- genera of the genus Melissodes published in the first part of this revision (LaBerge, 1956) is repeated herein in modified form. A total of 30,095 specimens representing 86 species have been ex- amined for this part of the revision. Explanations of the methods used in describing the species, the meanings of certain descriptive terminology and methods used in taking certain measurements are described in a section on descrip- tive methodology in the first part of this revision (LaBerge, 1956, * Published with the approval of the Director as paper No. 1055, Journal Series, Ne- braska Agricultural Experiment Station, University of Nebraska, Lincoln, Nebraska. This study was begun in the Department of Entomology, The University of Kansas, Lawrence, Kansas. (283 ) 284 Tue UNIversiry SCIENCE BULLETIN p. 919). No important changes in methodology have been intro- duced into this part of the revision. The reader is also referred to the first part of this revision for a complete list of acknowledgements. I wish to thank the Society of the Sigma Xi for a Grant-in-Aid of Research which enabled me to travel in Europe during the summer of 1957 in order to study the type specimens now located in European museums. Certain per- sons have rendered special aid in comparing specimens, in gra- ciously lending types in their care and in generously giving of their time and advice. I wish to especially thank the following persons for these services: Dr. I. H. H. Yarrow of the British Museum (Natural History), London; Mr. Karl V. Krombein of the United States National Museum, Washington, D. C.; Mr. P. H. Timberlake, Citrus Experiment Station, Riverside, California; and Dr. T. B. Mitchell, North Carolina State College, Raleigh, North Carolina. I am grateful to Dr. C. D. Michener of the University of Kansas, Lawrence, Kansas, for his interest and guidance throughout the period of preparation of this revision and for reading and com- menting upon the manuscript. The Research Council of the University of Nebraska, Lincoln, Nebraska, is especially thanked for a grant (1960) which helped to defray the expenses of preparing the drawings and maps and the typing of the manuscript. Thanks are due also to the Kansas University Endowment Association for a grant (1953) which per- mitted the author to visit various museums in the United States for the purpose of studying type specimens and for another grant (1955) which helped defray costs of preparing illustrative material. PHYLOGENETIC RELATIONSHIPS OF THE SUBGENERA A diagram showing the relationships of the various subgenera of the genus Melissodes was published in the first part of this re- vision (LaBerge, 1956, p. 915). That diagram included three sub- genera not now included in the genus Melissodes (Brachymelis- sodes, Epimelissodes, and Idiomelissodes, now in genus Svastra Holmberg). Also, the diagram did not include a subgenus subse- quently recognized and described below (Callimelissodes). New characters have since been discovered which shed new light on the relationships of the subgenera now included in the genus. There- fore, a new analysis of these relationships is presented here. From among the many characters found to be useful in the taxonomy of the genus Melissodes nineteen were chosen on the bases that they show variation among the subgenera and that the BEES OF THE GENUS MELISSODES 285 primitive versus the specialized condition of each could be logically inferred. These characters and their primitive and_ specialized states are listed below. 14. 15. 16. i 18. 19} Primitive Characters Female sixth tergum with la- melliform, toothed lateral parts of gradulus. Male with five-segmented maxil- lary palpi at least occasionally. Male seventh stemum_ with lateral plates small, flat, hairy below. Propodeum of female longer than metanotum medially. Clypeus flat; mouthparts rela- tively short. Male antennae long, first seg- ment at most one-third as long as second segment. Scopal hairs branched. Galeae with hairs straight. Galeae shiny or at most lightly shagreened. Integument of mesoscutum shiny to shagreened. Distal pale bands of terga 2 and 3 not apical or at least not completely so. Male fifth sternum with apical margin straight to concave. Male clypeus yellow or largely yellow or cream-colored. Size of bee medium, Female with distal pale band of tergum 2 present, interrupted or not, if interrupted or absent, band of tergum 4 also inter- rupted or absent. Eyes converge towards mandi- bles. Male flagellar segments without shiny elongate depressions. Gonostylus with short, simple or barbed, diffuse hairs. Male sternum 8 with simple or bidentate ventral tubercle. to On 10. 11. 14. 15%. 16. 17. 18. iS). Specialized Alternatives Lateral parts of gradulus ab- sent or cariniform, Four- or three-segmented max- illary palpi. Male seventh sternum otherwise. Propodeum of female shorter or no longer than metanotum medially. Clypeus more or less bowed out; mouthparts relatively long. Male antennae short; first seg- ment more than one-third of second. Scopal hairs simple or weakly branched. Galeae with hooked hairs. Galeae dulled by dense tessella- tion. Mesoscutum densely _ tessellate. Distal pale bands of terga 2 and 3 apical. Male fifth sternum convex me- dially to produced. Male clypeus black or largely black. Size small or large. Female with distal pale band of tergum 2 absent or broadly in- terrupted while that of tergum 4 uninterrupted. Eyes subparallel to converging towards vertex. Flagellar segments with shiny elongate depressions. Gonostylus with hairs long, sim- ple, more or less concentrated near base on outer surface. Male sternum 8 with strong me- dioventral crest. 286 ‘THE UNIVERSITY SCIENCE BULLETIN All eight subgenera are specialized for all included species in at least two of these and primitive in at least ten of these charac- ters. Each subgenus (except the monotypic Psilomelissodes) may have one or more species which is specialized in one or more char- acters, whereas the other species of the same subgenus are primi- tive for the same characters. Thus, for each character a subgenus may be primitive, partly specialized, or specialized. The subgen- era and the number of the characters listed above are summarized in table I. In this table a specialized condition is represented by a plus sign, a primitive condition by a zero, and a partly specialized character by a dash. The number of each condition is tabulated for each subgenus at the bottom. Melissodes s. str. Ecplectica Apomelissodes 4o04sa3uUD OyH11-D2;429,09q Heliomelissodes Tachymelissodes Eumelissodes Psilomelissodes Callimelissodes Fic. 2. Dendogram showing the relationships of the subgenera of Melis- sodes Latreille. The lengths of the various lines are of no significance. The phylogenetic diagram (Fig. 2) is based upon the above facts plus special consideration of the characters. An example of the latter is the dichotomy shown in the diagram between Ecplectica- Melissodes and the other subgenera. Although Melissodes s. str. appears specialized in more characters and at least as primitive as is the subgenus Eumelissodes, it is shown as having been derived from an Ecplectica-like ancestor independently of the Eumelis- sodes and the other subgenera. This dichotomy can also be seen in the list of characters given above. Melissodes s. str. has certain specialization of the male terminalia (characters 18 and 19) which can be more logically derived from the condition in Ecplectica than 287 BEES OF THE GENUS MELISSODES TABLE I. Nineteen Characters of the Subgenera of Melissodes. SopOssl[oULAYoR J, + | + te eSoimhe | ay ete Cora a Cia Os eric SopOsst[oULO]ISq +/° SC;4+);/S/4);e];e};/e]el;eol;eo|]sece SopoOssl[OULOl[op] +/+ SES ene aS os oS | eprint = | soeposstetuody 46 pat Si+]O] | /efeo] | jele}eo}exww 5 0 5 Sri S Sais o|o|oox RD EE) MOEMEMU NES) ales eat ae | = SaepossIfauINyy | | + SB S| SS Sp akkoes eae SoS Seo |S Noo H SOPOSST[OT, faeae | ap | ae Boryoo [don +\)° S/o }/So |e +/S/S/O}] So | 24% Diliman Sale ea) eal || es a : 5 : ; 5 é ; : 3 : ee ma : SNe al ogre wes cob i to epg | mre Wee me : 2S | Geert eae es Salles eS cease | ere) edo oa ae ace |e yale | Pies aera ell fea EE Bie Weta os SIV eal eet eect ceca ce : Sate ails a Pesca all || eet I : : : : : : : : 3 :| 3 3 A : Seal mace aes alae Be [ec icon ect |i ee woes So : Sha [pect ven] = aks [Ca tees kan cet Wiese |u| ar ee MN | 09 S|) a} N | co | H fas | oy od | o re ul od re re (al ret re ial re 288 THe UNIVERSITY SCIENCE BULLETIN from that in Eumelissodes. The condition of the same two char- acters in Eumelissodes can also be more logically derived from Ecplectica than from Melissodes s. str. Also, Ecplectica and Me- lissodes s. str. share a specialization (character 4) which separates them from the remaining subgenera and, vice versa, the remaining subgenera share a specialization (character 1) not yet acquired by the former two subgenera. In spite of these considerations, the true phylogenetic picture could still be with a Ewumelissodes-like ancestor giving rise to Eumelissodes and its related subgenera on one hand and to Me- lissodes s. str. through Ecplectica on the other hand. This hypoth- esis lacks the merit of having the evidence of primitiveness of the chosen phylogenetic arrangement. That is, the subgenus Ecplec- tica shows the primitive alternative in 15 of the 19 characters, whereas Eumelissodes is primitive in 11 of the 19 characters. KEY TO THE SUBGENERA OF NORTH AND CENTRAL AMERICAN MELISSODES The following key is an adaptation of that given by LaBerge (1956, p. 920). This key was modified by LaBerge (1956, p. 545) to include an additional species in the subgenus Apomelissodes. It is here further modified to omit those subgenera which have been removed from the genus Melissodes (e. g., Epimelissodes, Brachy- melissodes, and Idiomelissodes) and to include the new subgenus described below (Callimelissodes ). The subgenera of Melissodes are extremely difficult to key, espe- cially in the female sex. Therefore, some subgenera appear sev- eral times in the following key in order to include aberrant species. The males key out readily if one uses genitallic characters, other- wise certain subgenera must appear more than once in the key. The subgenera Ecplectica and Melissodes key out together in the female sex, partly because it is difficult to separate these subgenera in that sex and because in the original treatment of the species of these subgenera (LaBerge, 1956) the species are placed in the same key. MALES Me Clypeus protruding beyond eye from % to % width of eye in profile; metasomal terga 2-5 fringed with narrow marginal bands of appressed white pubescence, bands much nar- rower than basal areas; antennae long, minimum length first flagellar segment equals less than 4% maximum length SECONGMSER MIME oo. occ « ccs th Aca cana Apomelissodes Si (7) )e BEES OF THE GENUS MELISSODES Clypeus usually not protruding beyond eye by as much as half width of eye in profile, if protruding by half width of eye or more, then minimum length first flagellar seg- ments equals % or more of maximum length second seg- ment; terga often not fringed by pubescent bands, but bands when present interrupted medially and/or sub- DLCs eee REM a Oy. Foes ces ie oe e EE ned Sey SD Saye 2 Posterior margin of fourth and usually third sternum broadly convex, or produced into a broad, thin, hyaline, colorless LEE D ys apacbd ett ythect eee eS ote a to eA A 8 Callimelissodes Posterior margins of third and fourth sterna straight to slightly concave, never produced into a flap............ 3 Clypeus strongly protruding beyond eye by % or more of width of eye in profile; maximum length first flagellar seg- ment equals 0.4 or more of maximum length second seg- TITS ee 7 eS or UR el eet ee IE) CL Heliomelissodes Clypeus usually protruding half or less of width of eye in profile, if protruding more, then maximum length first flagellar segment less than 0.4 of maximum length sec- Onde ScOMIENiM ee Tk. te ORs vee we, lie) 4 Maximum length first flagellar segment as long as or almost as long as maximum length second segment and longer than third segment (antennae femalelike); clypeus wholly LOL eyelet, z acter Neck oe De Cer tanya te As Palacio CaP ae Psilomelissodes Maximum length first flagellar segment shorter than maxi- mum length second segment and distinctly shorter than third; clypeus usually pale, occasionally partly or wholly 16) eye) a Vas bo sh eee eo ee Se A ni 5 Minimum length first flagellar segment distinctly more than half maximum length second segment; terga 2-5 with pubescent bands apical, subequal in width across each tergum and subequal in width to each other. . Tachymelissodes Minimum length first flagellar segment half of maximum length second segment or less; terga 2-4 with pale pubes- cent bands usually not all apical or subequal in width, often interrupted medially and usually subapical... ..... 6 Characters of genitalia and hidden sterna................ 7 Hxtermalncharacters: Ss ont a eee Peet on EY Wale 9 Median apical plates sternum 7 without hairs on ventral surfaces, usually small, curled ventrally along an oblique axis to form half or more of an oblique cylinder or scroll, but often secondarily flattened and expanded, or second- Ally seKCCMCed ain ASIZONy See) vee se ke Melissodes Median plates sternum 7 thin, hyaline, with short to moder- ately long hairs on ventral surfaces, not curled ventrally, nelitivelyalarr ea went oe ee ye eee dy ee eR ce 8 Gonostylus short, less than half as long as gonocoxite, in lat- eral view twice as broad or more near base as near apex, narrowing abruptly near middle, not capitate; median 289 290 9 (6). 10 (9). LUGO). PACH OP 13(12). 14(10). 2) Tue UNrversiry SCIENCE BULLETIN plates sternum 7 relatively small, with several short hairs memtedlly ey aeqthe ae; 2%. dd sae camel a eee Ecplectica Gonostylus short to long, usually as long as or longer than half length of gonocoxite, in lateral view not twice as broad near base as near apex, often somewhat capitate; median plates sternum 7 large, with abundant short to moderately lonew hams \ventrally®...2 3842934 e a: _.. .Eumelissodes Terga 2-5 without pale pubescent bands or these all inter- rupted medially, if one or two bands are complete, then thoracic hairs mostly black, or black and white mixed, and labrum all or almost all pale-colored........ Melissodes Terga 2-5 usually with complete bands, occasionally one or more absent or interrupted medially, if with only one or two bands complete, then thorax bright ferrugineous hairs and. labrum all or- mostly: all black: =.4.. >. 222 essen 10 Labrum wholly pale; mandibles usually with basal yellow spots; last two metasomal terga with dark brown to black Wairsy be eed Actes oc aidk- eS ee Pee 14 Labrum with at least a dark margin and mandibles often with- out pale basal spots, or last two metasomal terga with only pale-hairs:or both \.-:a0.s0 Has toe: ee eee 11 Galeae smooth and shiny, without tessellation or shagreening except at tips: jas arcisseeadoubioac ne be eee 12 Galeae dulled by tessellation or shagreening at least in apical halbe y e, Pome ok A ee Cae eee eee Eumelissodes Last two terga with dark brown to black hairs............ 18 Last two terga with pale hairs only............. Eumelissodes Margins of terga 2-4 broadly hyaline, colorless or nearly so, or labrum entirely dark or pale spot covers less than half ofsurtacey Onbothinc . Kise Mats anh abe Eumelissodes Margins of terga 2-4 opaque, black to reddish-brown; labrum mostly pale-colored, at most with narrow apical margin brown to ‘black... css:, gu eee ee Melissodes Tergum 2 with distal pale band complete or only narrowly interrupted medially, if broadly interrupted, then basal pale band indistinct, consisting of diffuse pubescence or partly or wholly dark pubescence.............. Melissodes Tergum 2 with distal pale band absent or broadly interrupted medially, each fascia equal to 4 or less width of tergum, with, basal. pale band) distinct.4).-...545022-.52ee Ecplectica FEMALES Scopal hairs simple or weakly branched, if weakly branched, then clypeus in profile protruding beyond eye by % width of eye or more; pygidial plate not narrow... .Apomelissodes Scopal hairs branched, usually abundantly so, if weakly branched, then clypeus in profile not protruding beyond eye by as much as’% width of eye... .... 4.2 0. eee 2 Clypeus protruding anteriorly beyond eye in profile by % to % width of eye; inner orbits of eyes often parallel; inner sur- BEES OF THE GENUS MELISSODES 291 faces hind basitarsi with hairs dark brown to black (scopal hairs highly plumose, often yellowish)... ... Heliomelissodes Clypeus protruding beyond eye by less than half width of eye in profile, if protruding as much as half width of eye, then inner orbits of eyes distinctly converging towards mandibles and/or inner surfaces of hind basitarsi with hairs bright red to yellow (scopal hairs occasionally weakly PDLATICHE Cie: Al arr, Rh nas fete 1S CU bi rk Cody eke on 8 3 (2). Scopal hairs weak, with few branches, not hiding outer sur- faces of hind basitarsi and tibiae; metasomal terga very sparsely and weakly punctate, dulled by dense, fine sha- greening and weakly banded with sparse pubescence and hairs; pygidial plate V-shaped with broadly rounded ACRES eee aeN MELT tee HB pe WS 2 23. ee 5 Psilomelissodes Scopal hairs strongly branched and hiding outer surfaces of basitarsi and tibiae; or, if weak and with few branches, then terga coarsely punctate at least basally, or moderately shiny to shiny and strongly banded with abundant pubes- cence and hairs, or pygidial plate narrowly U-shaped.... 4 4 (3). Scopal hairs with only one or two braches on each side of rachis; pygidial plate narrow, U-shaped... . Callimelissodes Scopal hairs more highly plumose, most hairs with three or more branches on each side of rachis; pygidial plate varia- ble, usually V-shaped with acute or well-rounded apex, 5 5 (4). Terga 2-4 with distal pale pubescent bands reaching apical margins of terga, of about same width across each tergum and subequal in width to each other, not arising from dis- tinct punctures, as narrow or narrower than basal area of darksepubescence’s se laqh A aN see. Sores SLE Tachymelissodes Terga 2-4 with distal pale pubescent bands, or at least tergum 2, not reaching apices of terga, or, if reaching apices of terga, then diffuse over entire tergum or much wider than basal area of dark pubescence and/or not of about the same width across each tergum or subequal in width to each other; tergal punctures variable................... 6 6 (5). Last flagellar segment as long as broad or slightly shorter, about equal to penultimate segment in length, or, small bees, 8-9 mm. in length with galeae with abundant long hooked hairs above or with sparse, extremely short, blunt, straight hairs and regular dense tessellation. . . .Callimelissodes Last flagellar segment longer than broad and longer than pe- nultimate segment; small to large bees, if small, then ga- leae either with abundant long straight hairs, or surface not densely tessellate, never small and with abundant hooked JEU LS OES LTR Sk on Be AE NN ce OO ee 7 7 (6). Sides of thorax and propodeum with black hairs; first tergum with sparse punctures above, medially restricted to basal third of tergum except a few widely scattered punctures; galeae shiny; terga 2 and 3 with interband zones with sur- SACS): LOSE 11(10). THE UNIvERSITY SCIENCE BULLETIN faces dulled by fine, dense, reticulotransverse shagreen- iY oes SANE ey 5 <2 sos swiped PRO aes Callimelissodes Sides of thorax and propodeum with pale hairs, or, if with dark hairs, then first tergum densely punctate above at least in basal half, and/or galeae dulled by shagreening or tessellation; occasionally terga 2 and 3 with interband zones with surfaces shiny, unshagreened............... First metasomal tergum with extremely sparse punctures restricted to basal third of tergum or less medially; either small to medium-sized bees with pygidial plate narrow, subparallel-sided and about twice as long as median width, or large bees with hairs of inner surface hind basitarsi black and of lower-lateral surfaces of thorax black, Callimelissodes First metasomal tergum with punctures usually abundant at least in basal half, if sparse and restricted to basal third or less, then medium-sized bees with pygidial plate broadly V-shaped and much shorter than twice median width and lower-lateral surfaces of thorax without dark Hrairgiy ch, Cee SR. Se BY Oe | ees ye Large bees, wings deep brown, inner surfaces hind basitarsi with hairs dark brown, mesoscutum with large patch of dark brown hairs; head and thorax coarsely punctate, genal area near lateral margin of eyes with punctures round, deep, separated by about one puncture width or less, Callimelissodes Size variable, if large, then wings not deep brown, or hairs of inner surface hind basitarsi yellow to red, or mesoscu- tum with at most small patch of brown hairs; head and thorax less coarsely punctate, genal area near lateral mar- gin of eye with punctures usually minute, separated by much more than one puncture width.................. Metasomal tergum 6 with postgradular carina with lateral parts lamelliform and ending abruptly in a short blunt tooth; pygidial plate narrow in apical half, often with sides subparallel (unless worn); tibial plate short, with edges exposed except anteriorly and on lower, anterior 10 areal a eel De RIERA) ie Rees Callimelissodes Metasomal tergum 6 with postgradular carina with lateral parts absent or at most cariniform, never with tooth; py- gidial plate short, broad, not narrowed in apical half, not subparallel-sided; tibial plate larger, hairs cover edge an- teriorly and at least on anterior half of lower margin CiMESSHCVOTT Ne FS ik he ten en Lateral and ventral surfaces of thorax with dark brown hairs (including propodeum ); terga 2 and 3 with lateral raised areas of interband zones with large, irregular piliferous punctures, surfaces very shiny, with no tessellation or shagreening; supraclypeal area with surface smooth and 11 shiny, unshaereene@ 25.5 .é.. vs ome ed eee Eumcelissodes 13(12). 14(13). 15(14). 16(15). 17(16). LSUT )s BEES OF THE GENUS MELISSODES Lateral surfaces of thorax with pale hairs at least in some restricted area, or terga 2 and 3 with raised areas with surfaces at least delicately shagreened; supraclypeal area often dulled by shagreening or tessellation............. 12 Eye narrower than genal area in profile, widest part of eye equals half or less of length; lateral and ventral surfaces Ofsthoraxewith black thairs). 5 . 9: sso%u.eedeeeeee 51 Flagellar segments 5-10 nodose; minimum length first fla- gellar segment more than maximum length second seg- MMOS = oss scusse Soe eR Ree boars see hurdi, n. sp. Flagellar segments 5-10 not nodose; minimum length first fla- gellar segment equals or less of maximum length second Segments) 44... tas dye, Abeer a oe S. bepl chee eee 51 0). Tergum 2 with depressed area at extreme base (under edge of tergum 1 unless untelescoped) with coarse, round punctures separated mostly by one puncture width or less; antennae never entirely black, usually red at least below. . 52 Tergum 2 with depressed area at extreme base with minute, round punctures separated by more than one puncture width and often by 2 or 3 puncture widths or more; an- tennae often entirely black or dark brown............. 538 Pale head and thoracic hairs white; clypeus with surface dulled by dense transverse shagreening, gelida, n. sp. (in part) Pale head and thoracic hairs ochraceous to pale fulvous; clyp- eus shiny or moderately shiny, shagreening absent or sparsevand: coaTses.a! . esses. len'slateee dl ius aed’. ae 86 Clypeus with surface dulled by dense reticular shagreening; supraclypeal area dulled by dense reticular shagreening; mesepisterna with surfaces usually with delicate, sparse shagreening: Baja California, ...... se ).stesst comata, n. sp. Clypeus and supraclypeal area often with surfaces shiny, un- shagreened; mesepisterna without shagreening; United Statesnandy Mex COf cise Joey ern. ed. Be: confusa Cresson Pygidial plate broader at extreme base than median length; mandible black basally, labrum cream-colored with black apical margin; galeae shiny, unshagreened; wing veins yellow or red; head and thorax without black hairs, brevipyga, n. sp. Pygidial plate as long as width at base or longer, if broader at base than long (rare), then either mandible yellow at base, or labrum all black, or galeae dulled by shagreening, or wing veins brown to black; vertex, mesoscutum and/or Scutellumiotten, with .dank/hairs) dsicc. 3. .cnoxyah bape 55 Labrum largely or wholly white or cream-colored, more than 0.6 times as long as broad; pygidial plate narrow, % or less as broad at base as long; mandible with basal yellow spot; galeae shiny, unshagreened except at tips; wing veins dark 56(55). 57(56). 58(57). 59(56). BEES OF THE GENUS MELISSODES brown; vertex of head, thorax and terga often with dark ROW WaITSHe ae pier: ln he einen as sy aee dw: velutina (Cockerell) Labrum entirely black, or 0.6 times as long as broad or shorter, or pygidial plate more than % as broad as long, or galeae dulled by shagreening, or mandibles without yellow spots, or wing veins pale; head, thorax and metasomal terga with or without dark hairs; not with the combination PIVEN MADOVCIIE Mirae tite! srr SAL hen. t. shea ats des, a 56 Minimum length first flagellar segment half as long as its own maximum length and no longer than pedicel on that side; galeae shiny above, without shagreening except near tips; metasomal tergum 2 with pale pubescent band less than half as wide medially as apical area; terga 2 and 3 with interband zones dulled by dense shagreening, im- punctate or punctures shallow and obscured by shagreen- ing; apical areas terga 2-5 dulled by dense, fine reticular G)aveveaweverot ayer” 1 el Abe Nae peels eke ee ee os ee Minimum length first flagellar segment equals more than half maximum length first segment and longer than pedicel on same side, if short, then either with galeae dulled by sha- greening, or tergum 2 with pale band broader than half of apical area, or terga 2 and 3 with interband zones shiny or moderately so with distinct punctures; apical areas of terga 2-5 usually moderately shiny to shiny ...... 59 Mandible usually with small yellow basal spot; clypeus dalled by transverse shagreening in interpunctural areas, punc- tures small; labrum with pale mediobasal spot, gelida, n. sp. (in part) Mandible without basal yellow spot; clypeus shiny to moder- ately so, with no or little shagreening dulling surface, punctures coarse; labrum with or without pale spot ...... 58 Medium-sized to large bee, 12-14 mm. in length; tergal apices usually yellow; tergum 1 with apical hairs thickly plumose; minimum length first flagellar segment usually equals 17 or more maximum length second segment, submenuacha Cockerell (in part) Small to medium-sized bees, 9-13 mm. in length; tergal apices usually colorless; tergum 1 with apical hairs sparse, not thickly plumose; minimum length first flagellar segment usually equals less than 144 maximum length second seg- THER ty PRA ieee. tw et Spe cecil otis A seal: ss coreopsis Robertson Metasomal terga 2 and 3 with interband zones with deep punctures of irregular size but many as large and almost as deep as mesoscutal punctures; terga 3 and 4 with basal depressed areas with punctures fully as large and deep as mesoscutal punctures; apical areas of terga smooth, shiny, impunctate; last exposed sternum coarsely punctate; man- dible and labrum black; galeae shiny above, unshagreened, ‘ ; perpolita, n. sp. ~l 1 365 366 60(59). 61(60). 62(61). 63(62). 64(63). 65 (64). THE UNIVERSITY SCIENCE BULLETIN Metasomal terga 2 and 3 with interband zones variously punctured but punctures not nearly as large as mesoscutal punctures; terga 3 and 4 with basal areas with punctures much smaller than mesoscutal punctures; apical areas variously punctate; last exposed sternum usually not coarsely punctate; mandibles, labrum and galeae various; not with the combination of characters given above.... . 60 Mandible with basal yellow spot and labrum with large mediobasal pale spot (as large as % area of labrum and ustialiyslareer)y '. PELL o 2 Fhe ee 61 Mandible without yellow basal spot and labrum with or without pale mediobasal spot (rarely both mandible and labrum with pale spots, but then pale labral spot equals less than % of area of labrum and/or mandibular spots minute; also rarely with labrum black and mandible with mnute yellow ‘spot ) of) er Pah eee ee eee es Galeae dulled above by fine reticular shagreening; wing veins red to yellow; hairs of head and thorax usually pale fer- rugineous or at least’ ochraceous 2,-5.2 .072. agilis Cresson Galeae shiny above, not dulled by shagreening except at tips, or, if shagreened, then either wing veins brown to black or head and thoracic hairs white or with some dark brown ‘admixed 20 Oe 4S, CF ee ae eee 62 Flagellar segment 1 with maximum length equal to about or almost 4% maximum length flagellar segment 3, with minimum length segment 1 much longer than pedicel on same side; large bee, 12-15 mm. in length, menuachus Cresson (in part) Flagellar segment 1 with maximum length equal to much less than 4s maximum length flagellar segment 3, with minimum length scarcely, if at all, exceeding length of pedicel on same side; small to medium-sized, 8-13 mm. in length... 63 Tergum 2 with distal pale band as broad or broader than apical area medially; clypeus with large deep punctures, surface usually shiny, unshagreened; thoracic hairs above nsHally very palevochaceous <2 20.7 Uy. eee 64 Tergum 2 with distal pale band narrower than apical area medially, or, if about as broad, then clypeus with shallow punctures and surface dulled by fine shagreening; tho- qaeremhairs usually white’... 2U, San) Bee 66 Pygidial plate as broad at base as long or slightly broader; mesoscutum with posteromedian area with punctures crowded, separated mostly by half a puncture width or KEI 6 pie RE A era Se 8 pexa, n. sp. Pygidial plate narrower than long and/or mesoscutal punc- tures less crowded, separated mostly by more than half a NUneRmeraNIdehs) . Le I ee: SR} 6 Ee a 65 Mandible with basal yellow macula large, almost forming a complete band across base of mandible; tergum 1 with 66( 63). 67( 66). 68 (60). 69(68). 70(68). TUCTO): (PAGO) BEES OF THE GENUS MELISSODES plumose hairs obscuring apical margin, extremely dense and highly branched... ... verbesinarum Cockerell (in part) Mandibular basal yellow spot small, distinctly triangular; ter- gum 1 with apical band less dense and _ hairs less lpranichied [HE SPE Abas hal ee tal) i) es humilior Cockerell Wing veins all pale, yellow to orange; galeae often with delicate, reticular shagreening above....... snowii Cresson Wing veins red to brown (especially forewings), if mostly pale, at least subcostal vein and pterostigma reddish brown; galeae usually shiny, unshagreened above...... . 67 Terga 2-5 with apical areas always colorless, hyaline; terga 6 and 7 often with brown hairs; terga 2-4 often with hairs of interband zones yellow to brown, verbesinarum Cockerell (in part) Terga 2-5 with apical areas often infumate; terga 6 and 7 with hairs white to yellow; terga 2-4 with hairs of inter- Ipanidiezonesiewihite sh) gant noe 2s atc. | nivea Robertson Vertex, tegulae, mesoscutum and scutellum with abundant dark brown hairs, and/or antennae wholly dark brown to ache ee ewe eae eee eich gh RON Te AIL) BRL 69 Tegulae without dark hairs, vertex with few or no dark hairs, mesoscutum and scutellum usually without dark hairs; antennae always at least red below and dark above, often TONE ALGUIEN Sa. MADER, AU LERE TREY 27h een CME ERG LET... 70 Tergum 2 with depressed basal area with large round punc- tures separated mostly by one puncture width or less; flagellar segments always pale at least below, montana Cresson (in part ) Tergum 2 with basal depressed area with punctures minute and separated mostly by 2 or more puncture widths; fla- gellar segments often entirely dark brown or black, confusa Cresson (in part ) Metasomal tergum 1 with apical % to % medially with short, relatively simple, brown to black, appressed to subap- pressed hairs; terga 2-5 often with apices slightly or deeply ininMAater pA wehe mALEt PT She oe. arent.) We. or spall: Metasomal tergum 1 with apical 4 to 4s medially with long, relatively plumose, white or pale ochraceous hairs, or bare; apical areas terga 2-5 usually colorless, rarely slightly Inte wee Petts Seti ity eae eteies ped. oxy et. Cth. « 73 Galeae usually dulled by dense tessellation above; hairs of posterior basitarus along posterior margin much shorter thanstilial spursips shhh. oy. ets ode sa fumosa, n. sp. (in part) Galeae shiny, unshagreened except at tips and laterally; hairs posterior of hind basitarsus often as long as tibial spurs OMVlON Cen een Mess, Meet. Sees) ERLE No ear ry IS ne, Tergum 2 with distal pale band narrowly interrupted medi- ally; hind basitarsus with hairs of posterior margin dis- tinctly longer than tibial spur .. .. interrupta, n. sp. (in part) 368 THE UNIVERSITY SCIENCE BULLETIN Tergum 2 with distal pale band not interrupted medially (unless worn); hairs along posterior margin hind basitar- sus no longer than tibial spur or shorter, montana Cresson (in part ) 73(70). Length forewing plus tegula equals 1.0 cm. or more, large bees; tergum 1 with apical area obscured by several rows of short, highly plumose, appressed hairs; tergum 2 with apical % or more hyaline, yellowish; terga 2-4 with inter- band zones impunctate or punctures minute and obscured by dense shagreening; first flagellar segment equals little more than pedicel on same side; wing veins red, submenuacha Cockerell (in part ) Medium-sized to small bees, length forewing plus tegula less than 1.0 cm.; tergum 1 with apical area not hidden by plumose hairs; tergum 2 usually with apical hyaline area equal to less than % length of tergum, usually colorless; terga 2-4 with interband zones with distinct punctures or wing veins reddish brown to black .................... 74 74(73). Minimum length first flagellar segment no longer (or scarcely 75(74). 76(75). longer) than pedicel on that side, half as long as maximum length and % or less as long as maximum length segment 2; galeae unshagreened above; labrum usually entirely black; tergum 2 with minute punctures separated by 3 or more puncture widths in interband zone; tergum 2 usually with brown hairs in interband zone and often in apical area as well ip aateurneiy ano V2 dtaike: bale hie limbus, n. sp. Minimum length flagellar segment 1 usually distinctly longer than pedicel on that side, usually longer than half maxi- mum length, and greater than & maximum length segment 2; galeae often shagreened above; labrum often with pale median spot; tergum 2 with punctures and hair color vari- able, but not with combination of characters given above. . 75 Tergum 2 with punctures of interband zone minute, sepa- rated mostly by 3 puncture widths or more, scarcely broader than bases of hairs arising from them; terga 3 and 4 with punctures of interband zones minute to small, sepa- rated mostly by about 2 puncture widths (often less on tergum 4), indistinct and shallow .................... 76 Tergum 2 with interband zone punctures larger, separated mostly by 2 or less puncture widths, distinctly broader than bases of hairs arising from them; terga 3 and 4 with large, round, deep punctures separated mostly by 1 puncture width; or if terga 2, 3 and 4 with minute well-separated punctures, then wing veins yellow to red and bee of me- ditmaestze eas 20. ys Ae Sen. Tis Ce ae eee, rae Labrum usually without pale median spot or spot less than \s of area of labrum; galeae dulled by dense reticular sha- greening above; pygidial plate broad, width usually more thany Awlenetincer,......... 292 uae See, ae utahensis, n. sp. 77(75). 78(77). 79(78). 80(79). 81(79). BEES OF THE GENUS MELISSODES Labrum with pale median spot and spot usually equals % or more of area of labrum; galeae shiny above, unshagreened except at tips; pygidial plate usually narrow, width usually equals less than % length and often as little as % of length, vernalis, n. sp. Terga 2-4 with interband zones impunctate or with minute punctures scarcely broader than bases of hairs arising from them, separated by 2 to 4 puncture widths, surface dulled by fine tessellation; wing veins red; galeae shiny above ex- Cepe at tipseee vas tan, ie eeget esha perlusa Cockerell (in part ) Terga 2-4 with interband zones with distinct punctures sep- arated mostly by one puncture width or less (especially on terga 3-4), surface shiny or dulled by reticular sha- greening; wing veins yellow to dark brown; galeae shiny or dull, shagreened or unshagreened.................. 78 Pygidial plate broader at base than median length; wing veins brown to reddish brown; galeae shiny, unshagreened above except at tips; dorsum of thorax with hairs yellow; tergal vestiture ochraceous; tergum 2 with interband zone punctures deep, round, separated mostly by less than one puncture: widths Cease hava Sone wee relucens, n. sp. Pygidial plate at least as long as broad and usually longer, if broader at base than long, then wing veins yellow to red or galeae dulled by fine shagreening above; hairs of thorax and terga various in color; tergum 2 with interband zone punctures usually separated by about one puncture WAGE as atte Ae NEY) MEM, PEL iets. Sie alice a, 79 Flagellar segment 1 with minimum length equal to ‘4 or less of maximum length segment 2 and scarcely longer than pedicel on same side; galeae often shagreened at least on apical halves above; labrum with mediobasal pale spot; terga 2-5 with apices often infumate, yellowish brown to VE lLOWRU SEP MERE Mra Ets OR TT PERS. oF SPURS SM Fut ret Ot Ay>,t, 80 Flagellar segment 1 with minimum length equal to more than % and usually more than 14 maximum length segment 2, usually distinctly longer than pedicel on same side; labrum with or without pale spot; galeae often shiny, unsha- greened; tergal apices usually colorless......... So Tergum 2 with interband zone with erect to caherect! owe WICH ANTSen ey seein: Sete CRN eee elegans, n. sp. (in part) Tergum 2 with interband zone with erect to suberect, short, browmniiairs) “ate, 0 tie 4 ey re fumosa, n. sp. (in part) Terga 2-5 with at least some of the erect hairs of interband zones brown or black and often with some or all of sub- erect hairs of apical areas brown as well; veins of hind wing, as well as fore wing, usually dark reddish brown to black; medium-sized bees, hamuli usually with 12 or more Ingolcst Ereeu th eer Turns 20! montana Cresson (in part) Terga 2-5 with erect hairs of interband zones and suberect 369 370 82(81). 85( 84). 86(52). 87(86). THE UNIVERSITY SCIENCE BULLETIN hairs of apical areas (when present) ochraceous or white; veins of hind wings often red to yellow; small to medium- sized bees, hamuli often with 11 or fewer hooks......... 82 Tergum 1 with apical hairs short, appressed, white, dense and plumose so that a distinct white band is formed across entire tergum; tergum 2 with distal pale band at least as broad as apical apubescent area and usually broader; occasionally with brown hairs on mesoscutum and _ scu- Felling: erste. eersiige: en, mer Og Lee tens pret eee Gee 83 Tergum 1 with apical hairs dense and plumose only laterally, if at all, medially with hairs sparse and/or barbs minute; tergum 2 with distal pale band often narrower than apical area; dorsum of mesoscutum and scutellum without dark MAINS i occu aeetes. Be tepotbe’s | Sather. ane wy - eee eee 84 Scutellum with brown hairs (at least a few) and mesoscu- tum often with brown hairs; galeae often.dulled by deli- cate shagreening above.................. lutulenta, n. sp. Mesoscutum and scutellum without brown hairs; galeae shiny above, not shagreened except at tips ..... appressa, Nn. sp. Head and body hairs long, those of middle of vertex (head in facial view) mostly longer than flagellar segment 3; galeae shiny above, unshagreened except at tips; vestiture Ochraceous:) MieExiIcOneL. Ana se eee ae rufipes, n. sp. Head and body hairs shorter, those of middle of vertex mostly shorter than flagellar segment 3; galeae shiny or dulled by shagreening; vestiture tawny, ochraceous or white; mostly United, States and, Canada’! 525.172...) Aaa 85 Galeae shiny, unshagreened above except at tips; medium- sized bees, about 11 mm. in length; tergum 1 with apical hairs obscuring surface in lateral thirds or slightly less, not medially; vestiture white to extremely pale ochraceous, bicolorata, n. sp. Galeae usually dulled above by shagreening at least in apical halves; small bees, 9 to 11 mm. in length; tergum without plumose apical hairs obscuring surface at extreme sides; vestiture usually ochraceous to yellow... subagilis Cockerell Mesoscutum, scutellum and tegulae without dark hairs; terga 6 and 7 with abundant dark brown hairs; Central Amer- Habart Pets, 24..tt ise: chp e geese... etek ees persimilis Cockerell Mesoscutum and scutellum and usually tegulae and vertex of head with at least a few dark brown hairs, if dark hairs absent, then terga 6 and 7 with hairs all pale as well; United States and Mexico...................... 87 Tergum 2 with distal pale band not markedly arched or notched along posterior margin, of about same length across tergum and subequal in length to apical area me- dially; southeastern Mexico............... floris Cockerell Tergum 2 with distal pale band markedly arched along pos- terior border so as to be distinctly thinned (rarely inter- rupted) and shorter than apical area medially; United States and Northern Mexico ..... montana Cresson (in part) 4 (3). BEES OF THE GENUS MELISSODES FEMALES Mesoscutal and scutellar punctures minute, mesoscutum with large posteromedian area impunctate or with scattered punctures less than % diameter of mesepisternal punctures; mesoscutal punctures anteromesad of parapsidal lines no larger and mostly smaller than mesepisternal punctures and separated by 2 to 3 puncture widths or more, expolita, n. Mesoscutal and scutellar punctures larger or more crowded than deseribed above; or both...................... First three sterna with long apicomedian hairs with curled tips; scopal hairs long, curled near tips, with 2 or 3 branches on each side of rachis; first tergum with sparse minute punctures in basal third or less... paucipuncta, n. s First three sterna with apicomedian hairs not so long, not curled near tips; scopal hairs almost always more abun- dantly branched, with straight tips; first tergum usually with more abundant and larger punctures in basal third LOMEWOSLTAS Me Seen t ee) See ae Thorax with lateral and ventral surfaces, including propo- deum, with dark brown or black hairs; terga 2 and 3 with lateral raised areas of interband zones with large, pilif- erous punctures, surfaces shiny, with no tessellation nor shagreening; supraclypeal area shiny, unshagreened, n p. bo hymenoxidis Cockerell Thorax with entire lateral surface not usually with dark brown to black hairs, if so, then terga 2 and 3 with raised lateral areas of interband zones with surfaces at least slightly dulled by shagreening or tessellation or shiny but with extremely fine shagreening; supraclypeal area VATIOUSLY "SCUIPUUTEG: kt. ke et ee ke ee a Terga without complete pale pubescent bands these being reduced to lateral fasciae or absent completely; or, if one complete band present, then it is band of tergum 2 and lower lateral mesepisternal surfaces with dark brown hairs, Terga with complete pale pubescent bands on more than one tergum, if complete on only one tergum, either not on tergum 1, or lower lateral mesepisternal surfaces without GER Ka AUS eee ee Cn ene re TEN a ee. Scopal hairs fuscous or almost entirely so... . . Scopal hairs pale or mostly pale (those of Hasitarst foften IMOstly? Of Chtltely, Ganky DIOWM) os e825 3. 4. ee Thorax above, including dorsal and posterior faces of propo- deum and often upper parts of mesepisterna, pale ochra- ceous to slightly ferrugineous (fox-red), laterally and ventrally dark brown to black; medium-sized bee, 11 to Stemi im, lengthy unas 5.56 seen ee. ee: | bicolorata, n. Thorax mostly dark brown to black, occasionally with paler hairs peripherally on scutellum, near bases of tegulae and Ol 371 372 8 (5). 10 (9). 13(12). 14(13). T5(ALSHE THE UNIVERSITY SCIENCE BULLETIN on dorsum of propodeum; small bees, 9 to 12 mm. in lemoihton. mere Aeris Ass tls UDO ANE Te Tes on eee ih Wing membranes deeply infumate, dark brown; clypeus with surface dulled by coarse reticular shagreening; Cuba, pullata Cresson Wing membranes only slightly infumate near tips; clypeus shiny, at most with delicate shagreening; Oregon, pullatella, n. sp. Tergum 1 with punctures of anterior half large, rounded, mostly separated by about % a puncture width, surface only delicately shagreened, shiny ........ dentiventris Smith Tergum 1 with punctures of anterior half smaller, usually separated by much more than half a puncture width and often by more than one puncture width, or irregular, sur- face dulled by dense reticulotransverse shagreening which often) obseures; punctures,» 4 Vises. 2201 ee, See 9 Terga 2-4 with apical areas shiny, impunctate, without hairs, slightly translucent, reddish brown in color; clypeus not ateallesprotiidines 2) se eee rustica (Say) (in part) Terga 2-4 with apical areas moderately dulled by coarse sha- greening, or with short, appressed or subappressed hairs, or punctate, or all of these; clypeus occasionally protruding by ‘almost-% width. of eyejin profile. .....4+2.c. saeae eee 10 Mesoscutum with surface distinctly shagreened (reticularly ); small bees, 10-12 mm. in length; thorax with dorsum with most hairs ferrugineous (more rarely ochraceous)... .... . ‘jul Mesoscutum with surface unshagreened; shiny; medium- sized bees, 11-14 mm. in length; thorax with dorsal hairs mostly pale ochraceous, never ferrugineous, bimatris, n. sp. (in part) Galeae shiny, unshagreened except at tips.. bidentis Cockerell Galeae dulled by dense shagreening, trinodis Robertson (in part) Flagellar segment 2 slightly but distinctly longer than wide; hairs of inner surfaces hind basitarsi brown or dark brown, 13 Flagellar segment 2 as wide as long or wider and/or hairs of inner surfaces hind basitarsi yellow to dark red......... 18 Terga 2 and 8 with apical areas with abundant, suberect, dark brown to black hairs; pygidial plate acutely V-shaped, 14 Terga 2 and 3 with apical areas apubescent or with white or ochraceous, suberect hairs; pygidial plate V-shaped, but apex Lcounged, not acute, 64.0 2. ats ce Bertini i dale Galeae dulled by dense tessellation.......... micheneri, n. sp. Galeae shiny, not tessellate, unshagreened or at most only iShivae Ornette it oe Ee ae eae oe exilis, n. sp. Mesoscutum without dark hairs (or extremely few) and often none on scutellum; forewing and tegulae measures 11 mm. Or moreriuplencth..... 00 eee se eee er 16 Mesoscutum with large patch of dark brown hairs, scutellum also with dark hairs; forewing plus tegula measures less than, Ucn similencth’. 7). osc a eee oe eee 17 16(15). 17(15). 18(12). 19(18). 20(18). 21(20). BEES OF THE GENUS MELISSODES Terga 6 and 7 with lateral tufts of pale ochraceous to white hairs; mesepisterna without dark hairs. ..menuachus Cresson Terga 6 and 7 without lateral tufts of pale hairs; mesepisterna with lower-lateral parts with dark brown to black hairs, semilupina Cockerell Posterior pronotal lobes with black hairs mixed with the pale; mesoscutal dark hair patch reaching tegulae laterally or almost so; clypeal punctures large, round, separated by about half a puncture width, surface shiny, unshagreened OLsonly dslichtlydsO se: 2a-H best fail ile Sree. perpolita, n. sp. Posterior pronotal lobes without black hairs; mesoscutum without large brown patch posteromedially; clypeal punctures small and crowded, surface dulled by dense neticulanushasrecminget: notn. seet ons | ba cerussatd, Nn. sp. Scopal hairs with branches gently deflected outward from rachis and hooked downward toward tips so as to appear somewhat S-shaped (especially on upper half of basitarsi and medially on tibiae), branches short, abundant; terga 2 and 3 with apical areas broad, impunctate and apubes- cent; inner surfaces hind basitarsi with hairs dark brown LOMLAC Ke ere ee a, x ARE Fess apa ee 19 Scopal hairs with straight branches originating in a sharp angle to rachis, or if somewhat S-shaped, then either terga 2 and 3 with apical areas punctate or pubescent or with inner surfaces hind basitarsi with yellow to red hairs, 20 Posterior lobes of pronotum almost always with black hairs (at least one present); outer surfaces middle basitarsi with hairs dark brown; small bees, hamuli usually with IE oWhookss rarely: 13s. 2 owas Axcess. AG denticulata Smith Posterior lobes pronotum usually without black hairs; outer surfaces middle basitarsi with hairs cinereous to pale brown, never dark; medium-sized bees, hamuli usually with 13-15 hooks........ -..........vernoniae Robertson Tergum 1 with apical area broadly hyaline, colorless; inner surfaces hind basitarsi with hairs dark brown or black; clypeus with shiny median boss; vestiture ochraceous to Witter fPscip rite teehdur thins. dey a lelihh Ue ae tristis Cockerell Tergum 1 with apical margin usually opaque, if hyaline, only narrowly so or inner surfaces hind basitarsi with hairs yellow to red, or clypeus without shiny median boss; vesti- tusesvaniously colored t.4s:ta5 ees Sei Mile e cali s. . 21 Tergum 3 with apical area with (a) dark hairs across entire tergum or at least across median third or more, or (b) with pale hairs which do not completely obscure surface and which differ from hairs of distal pale band by being less plumose and more erect, or (c) with no hairs or punc- tures and apical apubescent area longer than distal pale bandhof tergum) 2) atleast medially... (04.5. ::2.3)--- 22 Tergum 3 with apical area covered by distal pubescent band which is apical, if with apical area not covered by distal 7) 374 24(23). 25 (24). 28( 27). THE UNIVERSITY SCIENCE BULLETIN pale band, then either (a) apical area apubescent, impunc- tate, and shorter across tergum than distal pale band of tergum 2, or (b) apubescent, impunctate and_ broadly triangular in shape but no wider than width of tergum (distal pubescent band reaching margin of tergum in lat- eral thirds) Mum anes. dentd, Oho AS eee 55 Mesoscutum without patch of dark brown to black hairs posteromedially, or few if any present; clypeus often pro- trudes slightly, in profile by as much as % width of eye... 23 Mesoscutum with distinct patch of dark brown to black hairs posteromedially; clypeus rarely protruding by as much as sewidthwor eye in profile: fz... Axieyt!. wesc. Tee 30 Tergum 1 with impunctate apical area equal to more than half of basal punctate area medially; tergum 3 often without dark hairs: invapical areaws’ <1 Lins) eee 24 Tergum 1 with impunctate apical area equal to half or less of basal punctate area medially; tergum 3 with apical area often with suberect dark hairs................... il Thoracic hairs white; metasomal pubescent bands white; terga 2 and 3 with apical areas with suberect hairs simple, Widmer loyRONMNer, eo kebeuceobctaobcocceuc: snowii Cresson Thoracic hairs pale to bright ferrugineous or ochraceous; metasomal pubescent bands usually yellowish; terga 2 and 3 with apical area hairs golden or brown, often with short: distinetbarbs Js 2, fails ee Se 25 Veins of hind and fore wings dark brown to black; metasomal sternal hairs mostly dark brown to black; tergum 4 with pale apical band notched medially with area of dark brown to black hairs; terga 2 and 3 with apical areas with dark suberect hairs........... trinodis Robertson (in part) Veins of hind and fore wings mostly red; metasomal sternal hairs mostly ferrugineous or paler, occasionally dark brown; terga 2, 3 and 4 with apical areas with no suberect hairs: or thesespalé invcolors 22s): ies. Se 26 Galeae shiny above, unshagreened except at tips; terga 2 and 3 with apical areas with suberect hairs long (of about same length as those of distal pubescent band) and abun- dant (obscuring surface, especially on tergum 3), perlusa Cockerell Galeae dulled above by dense shagreening; terga 2 and 3 with apical areas with suberect hairs shorter and _ less abundant, scarcely obscuring surfaces ..... . agilis Cresson Terga 2 and 3 with apical areas impunctate and without Stiberectehairs fo cL0ca iid a ie. ee ochraea, n. sp. Terga 2 and 3 with apical areas punctate and/or with sub- CREE hanes! id 20st... 243. Peete ath coat senate ee eee 28 Terga 2 and 3 with apical areas with suberect, long, pale hairs; flagella usually entirely dark brown to black, bimatris, n. sp. (in part) Terga 2 and 3 with apical areas with dark brown, suberect hairs; flagella usually red below..................... 29 BEES OF THE GENUS MELISSODES 375 29(28). Terga 2 and 3 with apical areas with suberect hairs arising from small but distinct punctures; galeae usually dulled bynshagreening es i). sek cose. feces AG. clea: hurdi, n. sp. Terga 2 and 3 with apical areas with suberect hairs not aris- ing from distinct punctures; galeae usually shiny, unsha- greened except at tips.............. submenuacha Cockerell 30(22). Galeae dulled by dense regular tessellation, with long hairs some of which are bent or hooked near tips; terga 2 and 3 with apical areas with abundant, long dark brown to black, suberect hairs arising from minute but distinct punctures; vertex of head and tegulae with abundant dark DrovMPMAITSHAy Ee ie tees oie NE acy. ake! oe moorei Cockerell Galeal sculpturing various, without hooked hairs, or, if tes- sellate and with hooked hairs, then terga 2 and 3 with apical areas impunctate or without dark hairs or both; vertex and tegulae often without dark hairs ........... 31 31(30). Pygidial plate acutely V-shaped; medial tibial scopal hairs reddish brown, anterior and posterior hairs golden yel- low; terga 2 and 3 with apical areas with subappressed, dark brown hairs; tegulae testaceous......... crocina, ni. sp. Pygidial plate V-shaped or U-shaped, if V-shaped not acutely so, but apex rounded; medial tibial scopal hairs as pale as other scopal hairs; terga 2 and 3 with apical areas various; tegulae usually piceous.................. + 82 32(31). Small bees, forewing plus tegula measures about 8 mm. in length; vertex of head and tegulae without dark hairs: thoracic pale hairs white or very pale ochraceous; poste- rior pronotal lobes never with dark hairs. humilior Cockerell Usually larger bees, if forewing and tegula measures 8 mm. or less in length, then vertex of head and tegulae with dark hairs, or pale hairs of thorax ferrugineous, or pos- terior pronotal lobes with at least a few dark hairs ..... 3: 33(32). Small bees, forewing plus tegula less than 8 mm. in length: tergum 2 with distal pale band broadly interrupted medi- ally, forming two lateral oblique fasciae tapering mesally and anteriorly, each fascia about as long as distance between their tips, apical area long, glabrous, fasciatella, n. sp. Usually larger bees, if forewing and tegula measures 8 mm. or less, then tergum 2 with distal pale band not broadly interrupted medially, apical area often with hairs... ... 34 34(33). Tergum 3 with apical area with abundant simple white hairs arising from small but distinct punctures (hairs may be cinereous or even pale brown medially ); tergum 2 with apical area similar but less distinctly punctate; metasomal pale hairs and scopal hairs white; posterior pronotal lobes mMever withudark hains,.)..2 3s. 2t onions nivea Robertson Tergum 3 with apical area usually with abundant dark brown hairs, if with simple white hairs, then these not arising from distinct punctures; metasomal pale hairs and scopal G2 376 Tue UNIVERSITY SCIENCE BULLETIN hairs often ochraceous to yellow; posterior pronotal lobes oftenmewith, darkathairs: iT .vlic). wide: deel teres eee 35 35(34). Terga 2 and 3 with apical areas long (usually medial length greater than medial length of pale band of tergum 3), apubescent, and impunctate, surface dulled by fine, dense, reticular shagreening, as dull as interband zone of tergum 2 (a few scattered, dark brown, short, appressed hairs may be present but, if so, separated by more than length ofmGnevorohairs)) , Jonette aiiier. at 8 le a 36 Terga 2 and 3 with apical areas either (a) short (less than length of distal pale band tergum 3), or (b) with abun- dant hairs, pale or dark, or (c) punctate, or (d) shiny and distinctly less dulled by shagreening than interband zone of tergum 2, or some combination of these. ....... 38 36(35). Tergum 4 with pubescence along apical margin dark brown to black, without apicomedian apubescent area, wheeleri Cockerell Tergum 4 with apicomedian apubescent triangular area, often with a few dark brown hairs along apical margin or in virtually apubescent area......................-. 87 37(36). Tergum 4 with apical apubescent area with abundant, minute punctures obscured by dense tessellation; tergum 2 with interband zone with distinct punctures; mesepisterna usu- ally dulled by fine irregular shagreening, illata Lovell and Cockerell Tergum 4 with apical apubescent area impunctate, surface dulled by dense shagreening as in terga 2 and 3, tergum 2 with interband zone virtually impunctate; mesepisterna with surfaces usually shiny, unshagreened... subillata, n. sp. 38(35). Tergum 2 with apical area impunctate and without hairs (a very few short dark appressed hairs may be present near distal pale band), shiny, extremely finely reticulotrans- versely shagreened; tergum 2 with interband zone with surface dulled by dense reticular shagreening contrasting with shiny apical area; vertex of head and tegulae with abundantidark brown to blackthairss. .. 70:2 ../y, 2 39 Tergum 2 with apical area either punctate or with abundant appressed to suberect, pale to dark hairs, often dulled by shagreening, if glabrous, shiny and impunctate, then tergum 2 with interband zone with surface also shiny, not contrasting with apical area; vertex of head and/or tegulae often without dark hairs..................... 40 39(38). Tergum 3 with distal pubescent band usually reaching apical margin at extreme sides; pale mesoscutal hairs anterior to dark patch white to pale ochraceous; scopal hairs ochra- ceous; tergum 1 with punctures small, separated by about half a puncture width but distinct, surface reticularly SHA TeESHEC Ui A iis. situ . dee pallidisignata Cockerell (in part) Tergum 3 with distal pubescent band well-separated from apical margin across entire tergum; pale mesoscutal hairs a BEES OF THE GENUS MELISSODES ST anterior to dark patch often ferrugineous; scopal hairs usu- ally golden yellow; tergum 1 with basal area punctures large, extremely shallow, crowded, obscured by dense, re- ticulareshagreening) te tae nett Arey ae rustica (Say) 40(38). Tergum 2 with interband zone with distinct, round, regular punctures across entire tergum, may be sparser medially than lateraiiyames gre siert: an tihts 2ST a et Bek 0) seas 41 Tergum 2 with interband zone impunctate or punctures indis- tinct or irregular in shape and size, or restricted to lateral TSCMPATEAS Gee ee ee eae er oe eee come ne ne SATA LTE 50 41(40). Tergum 3 and often tergum 2 with apical area distinctly punctate, punctures at least slightly greater in diameter thanvhairs-arising trom them, 22208 Ln) 42 Terga 2 and 3 with apical areas impunctate or with minute punctures scarcely greater in diameter than hairs arising fromePthenmite:. Or a irr tee eet NR IMSS IF AT 42(41). Small bees, forewing plus tegula equals 8 mm. or less, dorsal thoracic hairs (especially pale hairs) blunt, branches extremely short and abundant making hairs ap- pear thick and clipped; clypeus flat, shiny, punctures coarse, shagreening sparse, delicate and irregular, paulula, n. sp. Small to large bees, if forewing plus tegula 8 mm. or less, then dorsal thoracic hairs with longer, sparser branches thus appearing thinner and sharply pointed; clypeus often slightly bowed outwards, punctures variable, surface often dulled by reticular shagreening: ©2020 -20). Seer... 43 43(42). Tergum 1 with basal area punctures large and deep, medially as large or larger than scutellar punctures but not as deep, well-separated and not obscured by shagreening ....... 44 Tergum 1 with basal area punctures small and shallow, medi- ally distinctly smaller than scutellar punctures, crowded and often obscured by dense shagreening............. 45 44(43). Posterior lobes of pronotum usually with at least one long black hair mixed with the pale; galeae dulled by tessel- DePeHO TRONS TF RE SOI ERAS |? opted ty boltoniae Robertson Posterior lobes of pronotum never with black hairs; galeae with at least basal half shiny and not tessellate or sha- greened, apical half or less often dulled by fine reticular SHASKeeMing tests MN) aa REDREL PA NEES ey fumosa, n. sp. 45(43). Galeae shiny, unshagreened above except near tips; tergum 2 with interband zone punctures small, in median third separated mostly by 2 to 3 puncture widths, laterally by 1 to 2 puncture widths, tergum 4 fringed with black AIT Stee eles Ekg oe RAGS TRIP: Feat RCo pilleata, n. sp. Galeae dulled above by fine reticular shagreening at least in apical third to one-half; tergum 2 with interband zone punctures slightly larger, in median third separated mostly by 1 to 2 puncture widths and laterally by 1 puncture 14—5840 378 46(45). 47(41). 48 (47). 49(48). 50(40). 51(50). 52(51). THE UNIVERSITY SCIENCE BULLETIN width or less; tergum 4 with or without fringe of black aise Weer Ceeeee ee PhS wink ede! tye See ed ee 46 Terga 5 and 6 with small lateral tufts of white hairs; meso- scutal dark patch twice as large as scutellar dark patch or almost so (sometimes larger); mesoscutum with pale hairs white to pale ochraceous................ tincta, n. sp. Terga 5 and 6 without lateral pale hair tufts; mesoscutal dark hair patch equals scutellar dark patch or only slightly larger; mesoscutal pale hairs ochraceous to dull ferrugin- CONSE AS. ot Dae aa PA GF manipularis Smith Clypeus with large crowded punctures, surface shiny, un- shagreened; tergum 2 with interband zone punctures regu- lar in size and spacing, not sparser medially; scopal hairs MMe SAT akat ey ee le i ae pexa Clypeus with punctures often small, surface usually some- what dulled by shagreening, if surface shiny and unsha- greened, then tergum 2 with punctures of interband zone somewhat sparser medially than laterally; scopal hairs often ipalevochraceouss. Wien eee eee 48 Galeae shagreened in apical half or more; tergum 2 with interband zone punctures not markedly sparser medially than laterally, with distal pubescent band usually not in- tempted medially e6e Whe, ake a) Scenes elegans, n. sp. Galeae shiny, unshagreened except at tips; tergum 2 with interband zone punctures somewhat sparser medially than laterally, with distal pale band often narrowly interrupted medially jee tl Cae ee 4 SL ee Ce 49 Wing membranes somewhat infumate, yellowish brown; inner surfaces hind basitarsi with hairs orange to dark TEC): duke mG a pede lets ae ae Aa, SER montana Cresson Wing membranes clear, not infumate; inner surfaces hind basitarsi with hairs dark reddish brown to black, coreopsis Robertson Galeae dulled by dense, coarse shagreening or tessellation, robustior Cockerell Galeae shiny, unshagreened except at tips or only extremely delicately iso! i). ehal -aiorac/ Walaa SOME ES be bl Tergum 3 and usually 2 with apical area with hairs white to brown, long and silky, suberect, often curved away from surface; pale body hairs and scopal hairs white; clypeal punctures small, surface dulled by dense reticular sha- BTECHING. a). Sings Aa Be Shee Sa ceetes gelida, n. sp. Terga 2 and 3 with apical area hairs dark brown to black, short, subappressed, straight, not silky but appearing rigid; pale mesoscutum and often scopal hairs ochraceous to pale GGlinaCEOugett 4.5 0%... UE PA eee 52 Small bees, wing plus tegula measures 8 mm. or less in length; tergum 3 with apical area less than half length of distal pale band medially, with short subappressed brown hairs arising from minute punctures.......... limbus, n. sp. BEES OF THE GENUS MELISSODES 379 Larger bees, if wing plus tegula measures 8 mm. or less, then tergum 3 with apical area at least half as long as distal pale band medially or provided with subappressed black hairs ‘but not»punctate at all), ©). 225 ok. 53 53(52). Terga 5 and 6 without lateral tufts of pale hairs; mesepi- sterna and clypeus often with lower parts with dark brown to black hairs; flagellum often dark brown to black FELONY Does tee ee ee TALE PNER V9 Pein AU SN confusa Cresson Tergum 5 and usually tergum 6 with lateral tufts of pale hairs; mesepisterna and clypeus usually without dark hairs: flagellum ‘usually ‘red below.........:.....5.... 54 54(53). Tergum 2 with distal pale band distinctly and abruptly interrupted medially (lateral fasciae not tapered to mid- line); mesepisterna without dark hairs below; inner surfaces hind basitarsi with hairs orange to red, interrupta, n. sp. Tergum 2 with distal pale band not interrupted medially, or if so, then lateral fasciae tapered to midline; mesepi- sterna often with dark brown hairs below, inner surfaces hind basitarsi with hairs dark reddish brown to black, persimilis Cockerell 55(21). Terga 2-4 with apical areas translucent, colorless to pale yellowish brown; inner surfaces hind basitarsi with hairs VOLO WAG sre ie. eres oor i. LEAR ded War LS PEE IE ok, 56 Terga 2-4 with apical areas piceous, if somewhat translu- cent, not pale yellowish brown but dark brown and opaque; inner surfaces hind basitarsi with hairs often dark DrOWile TOM ACKnn sae e ty Lok ee) SI Pane. TNT NS a, 58 56(55). Mesoscutum with no or few dark brown hairs; terga 2 and 3 with apical areas with hairs white; terga 5 and 6 with hairs white to ochraceous............. saponellus Cockerell Mesoscutum with large patch of brown hairs; terga 2 and often 3 with apical area hairs dark brown; terga 5 and Gvartent withi'dark brown hairs) ©. 0.2 7). eo 57(56). Wing membranes somewhat infumate; terga 3-5 with basal areas (e.g. usually telescoped under preceding segment ) with white hairs; clypeus somewhat bowed outwards, velutina (Cockerell) Wing membranes colorless, clear; terga 3-5 with basal area hairs brown; clypeus relatively flat..........vernalis, n. sp. 58(55). Small bees, forewing plus tegula measures 7.0 to 8.5 mm. in length; pygidial plate acutely V-shaped; flagellum dark reddish brown to black except small dark red ventral spot on segments 3 to 10; tergum 2 with distal pubescent band narrowly interrupted medially; inner surfaces hind basi- tarsiwithy dark nas @. 2Ales. Viol eG microsticta Cockerell Small to medium-sized bees; if forewing plus tegula meas- ures 8.5 mm. or less, then pygidial plate not acutely V-shaped but with apex rounded, and/or flagellum dark red below; tergum 2 often with distal pale band uninter- 380 59(58). 60(59). 61( 60). 62(61). 63( 62). 64(63). THE UNIVERSITY SCIENCE BULLETIN rupted; inner surfaces hind basitarsi often with yellow EO STC) MiamrG ee koe os NS pau oA idk ae Ane ee 59 Tergum 2 with interband zone impunctate or virtually so, with dark hairs suberect to erect; inner surfaces hind basi- tarsi with dark reddish brown to black hairs, grindeliae Cockerell Tergum 2 with interband zone with distinct punctures, if almost impunctate, then with hairs pale and/or appressed to subappressed; inner surfaces hind basitarsi often with hairsisvellowsto ned ack Be oe 60 Pygidial plate acutely V-shaped; mesoscutum with postero- median area with punctures round, deep, crowded, separated mostly by half a puncture width... relucens, n. sp. Pygidial plate V-shaped but with rounded apex (sometimes worn to acute point), if acute, then posteromedian area mesoscutum impunctate or with punctures separated mostly by one puncture width or more................. 61 Small bees, forewing plus tegula less than 8 mm. in length; inner surfaces hind basitarsi with hairs yellow to red; galeae dulled by dense tessellation, with long hairs some of which are hooked or bent near tips. melanura (Cockerell ) Larger bees, if forewing plus tegula measures less than 8 mm. in length, then either inner surfaces hind basitarsi with heirs dark brown and/or galeae shiny, not tessellate; galeal hairs often short and never hooked near tips. ..... 62 Medium-sized bees, forewing plus tegula measures 9 to 10 mm. in length; tergum 2 with interband zone with lateral raised areas with distinct punctures of two sizes, one large and seemingly directed somewhat posteriorly, the other minute; inner surfaces hind basitarsi with hairs dark |cVConi Te a ae en pallidisignata Cockerell (in part) Smaller bees, if forewing plus tegula equals 9 mm. or more in length, then tergum 2 with interband zone impunctate or punctures all of same size, or inner surfaces hind basi- tarsi. with: hairs iyellow;. to.xed | »..a:0-0. :o-4:44d aes 63 Medium-sized bees, forewing plus tegula measures more than 8 mm. in length; inner surfaces hind basitarsi with hairs yellow to red; tergum 2 with distal pale band ex- tremely narrowly interrupted medially; mesoscutal dark hair patch about equal to scutellar dark patch in size, rufipes, n. sp. Smaller bees, if forewing plus tegula measures more than 8 mm. in length, then either inner surfaces hind basitarsi with hairs dark, or tergum 2 with distal pubescent band uninterrupted or mesoscutal dark hair patch smaller than scutellar dark patch, or some combination of these... ... 64 Galeae dulled above by dense reticular shagreening; inner surfaces hind basitarsi with hairs usually yellow to red, rarely dark reddish brown; pale vestiture ochraceous to yellow (especially on mesoscutum); mesoscutum usually 65( 64). 66(65). 67(66). 68(65). 69(68). BEES OF THE GENUS MELISSODES without dark hairs or dark patch smaller than scutellar ark. atehiy ciate Piri) sNeee te uA AR he subagilis Cockerell Galeae shiny, not dulled above by dense shagreening except near tips (less than apical half), if dulled by shagreening, then either inner surfaces hind basitarsi with dark hairs, or pale vestiture white, or mesoscutum with large dark hair patch (at least as large as scutellar dark patch) Tergum 2 with interband zone with appressed to subap- pressed, white pubescence, dark spinelike hairs absent, with distal pale band composed of relatively long, overlap- ping plumose hairs, not interrupted medially unless worn. . Tergum 2 with interband zone with appressed to subap- pressed, dark brown to black, spinelike hairs, or if dark hairs absent, then distal pale band medially composed of short, scalelike, discrete (not overlapping except near base of pale band) hairs, often extremely narrowly interrupted eaverc hull binge, chow anole net, ice alr er ie ite eee Pe Galeae dulled above by shagreening at least in apical annals to one-half; mesoscutum usually with dark hairs posterome- clic ly apie neta Wem st ont e 12 UR lan AO. ate HAN CM utahensis, n. Galeae unshagreened except near tips; mesoscutum without Ski al eR Oe RS A ae ee ae ee, See Tergum 2 with apical area apubescent, shiny, subequal to distal pale band in length medially or longer than half length of distal pale band, interband zone with distinct punctures about equal in diameter to punctures of basal area and separated mostly by half to one puncture width, brevipyga, n. s Tergum 2 with apical area obliterated by distal pale band, or, if apubescent, shorter than half length of distal pale band medially, with interband zone with minute punctures dis- tinctly smaller than those of basal area and separated mostly by one or more puncture widths, 65 66 verbesinarum Cockerell Tegulae with abundant dark brown hairs; tergum 1 with apical area with abundant, short, closely appressed, dark browathains)) Mok is PAWS | Se et eee personatella Cockerell Tegulae with hairs white to ochraceous; tergum 1 with apical area apubescent or with appressed hairs white (a few dark brown appressed hairs sometimes present an- tenolaverallya)s (2 Ast uselky. Sue te tte Terga 2 and 3 with distal pubescent bands at peach near nicl line composed of short, scalelike, appressed, plumose hairs which scarcely overlap laterally or do not; tergum 2 with interband zone punctures regular in size and spacing (often slightly sparser medially) ...........appressa, n. Terga 2 and 3 with distal pale bands composed of relatively long plumose hairs overlapping laterally, not seemingly scalelike; tergum 2 often with punctures irregular in size aUGUSDACINE) cByES & FET AP WUE. g18 18. ot. ae Le: 69 sp. 381 382 THe UNIVERSITY SCIENCE BULLETIN 70(69). Mesoscutum with posteromedian area with punctures sepa- rated mostly by less than two puncture widths, monoensis, n. sp. Mesoscutum with posteromedian area punctures separated mostly by more than two puncture widths... .lutulenta, n. sp. Melissodes (Eumelissodes) agilis Cresson Melissodes agilis Cresson, 1878, Proc. Acad. Nat. Sci. Philadelphia, vol. 30, p. 204; Cockerell, 1894, Ent. News, vol. 5, p. 234; Robertson, 1896, Trans. Acad. Sci, St. Louis, vol. 7, p. 176; Cockerell, 1897, Bull. New Mexico Agric. Exp. Sta., No. 24, pp. 23, 24, 28; 1898, Bull. Univ. New Mexico, vol. 1, pp. 67, 73; 1898, Bull. Sci. Labs. Denison Univ., vol. 11, pp. 67, 73; 1898, Zoologist, ser. 4, vol. 2, pp. 312, 314; Bridwell, 1899, Trans. Kansas Acad. Sci., vol. 16, p. 211; Cockerell, 1900, Entomologist, vol. 33, p. 217; 1901, Ann. Mag. Nat. Hist., ser. 7, vol. 7, p. 129; 1901, Ent. News, vol. 12, p. 40; Robertson, 1901, Canadian Ent., vol. 33, p. 231; Pierce, 1904, Univ. Nebraska Studies, vol. 4, p. 174; Cockerell, 1905, Bull. S. California Acad. Sci., vol. 4, pp. 13, 28, 29; Robertson, 1905, Trans. Amer. Ent. Soc., vol. 31, p. 369; Lovell and Cockerell, 1906, Psyche, vol. 13, p. 110; Snow, 1906, Trans. Kansas Acad. Sci., vol. 20, p. 137; Cockerell, 1906, Trans. Amer. Ent. Soc., vol. 32, pp. 31, 77; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309; 1907, Univ. Colorado Studies, vol. 7, p. 128; 1910, Entomologist, vol. 43, p. 90; Graenicher, 1911, Bull. Publ. Mus. Milwaukee, vol. 1, p. 247; Cock- erell, 1914, Ann. Mag. Nat. Hist., series 8, vol. 14, p. 365; Cresson, 1916, Mem. Amer. Ent. Soc., vol. 1, p. 110; Robertson, 1926, Ecology, vol. 7, p. 379; 1928, Flowers and Insects, p. 8; Cockerell, 1928, Psyche, vol. 35, p. 333; Pearson, 1933, Ecol. Monogr., vol. 3, p. 380; Graenicher, 1935, Ann. Ent. Soc. Amer., vol. 28, p. 304; Brimley, 1938, Insects of North Carolina, p. 462; Linsley, 1946, Jour. Econ. Ent., vol. 39, p. 20; Michener, 1947, Amer. Midland Nat., vol. 38, p. 454; Bohart, Knowlton, Bailey, 1950, Utah St. Agric. Coll. Mimeo. Series No. 371, p. 5. Melissodes agilis aurigenia Cresson, 1878, Proc. Acad. Nat. Sci. Philadelphia, vol. 30, p. 212; Robertson, 1894, Trans. Acad. Sci. St. Louis, vol. 6, p. 452; 1897, Trans. Acad. Sci. St. Louis, vol. 7, p. 8354; Cockerell, 1897, Bull. New Mexico Agric. Exp. Sta., No. 24, p. 20; 1898, Zoologist, ser. 4, vol. 2, p. 79; 1898, Bull. Univ. New Mexico, vol. 1, pp. 67, 73; 1898, Bull. Sci. Lab. Denison Univ., vol. 11, pp. 67, 73; Robertson, 1898, Botanical Gazette, vol. 25, p. 244; Cockerell, 1901, Ann. Mag. Nat. Hist., ser. 7, vol. 7, p. 129; 1903, Ann. Mag. Nat. Hist., ser. 7, vol. 12, p. 450; 1905, Bull. S. California Acad. Sci., vol. 4, p. 103; 1906, Trans. Amer. Ent. Soc., vol. 32, pp. 81, 84, 92: Snow, 1906, Trans. Kansas Acad. Sci., vol. 20, p. 136; Hart and Gleason, 1907, Bull. Illinois St. Lab. Nat. Hist., vol. 7, p. 257; Tucker, 1909, Trans. Kansas Acad. Sci., vol. 22, p. 278; Cockerell, 1911, Canadian Ent., vol. 43, p. 390; Gibson, 1914, Forty-fourth Ann. Rept. Ent. Soc. Ontario, 1913, p. 125; Cockerell, 1914, Canadian Ent., vol. 46, p. 409; 1915, Ann. Mag. Nat. Hist., ser. 8, vol. 16, p. 482; 1916, Canadian Ent., vol. 48, p. 77; Viereck, 1916, Bull. Connecticut Geol. Nat. Hist. Surv., vol. 22, p. 733; Clements and Long, 1923, Carnegie Inst. of Washington Publ. No. 336, p. 253; Cockerell, 1923, Ent. News, vol. 34, p. 48; 1928, Psyche, vol. 35, p. 333; 1930, Amer. Mus. Nov. No. 397, p. 1; 1933, Ann. Ent. Soc. Amer., vol. 26, p. 44. Melissodes pennsylvanica, Hendrickson, 1930, Iowa St. Coll. Jour. Sci., vol. 4, p. 164 ( misidentification ). Melissodes agilis is without doubt the most common species of the genus Melissodes in North America. The females are distinctive in the usually red thoracic hairs (without posteromedian brown patch ), the red to reddish brown wing veins, the dulled galeae and BEES OF THE GENUS MELISSODES 383 the ochraceous scopal hairs. The males can be recognized by the hyaline tergal margins, the dulled galeae, the yellow maculae on the mandibles and labrum and the entirely yellow clypeus, the yellow to pale red wing veins, the ochraceous to red thoracic hairs and the short first flagellar segment. Both sexes closely resemble certain other species of the subgenus Eumelissodes in one or more of these characteristics. These are discussed in the diagnosis of the species whose descriptions follow that of agilis. Female. Measurements and ratios: N, 20; length, 10-13 mm.; width 3.5-4.5 mm.; wing length, M = 3.46 + 0.159 mm.; hooks in hamulus, M = 13.45 + 0.223; flagellar segment 1/segment 2, M = 1S S="0.027- Structure and color: Integument black except as follows: apical half of mandible, lower surface of flagellar segments 3-10 (and often apex of segment 2), and distitarsi rufescent; eyes gray to bluish gray, rarely greenish or black with violet reflections; wing membranes hyaline, colorless, veins red to reddish brown, ptero- stigma yellow to red; tibial spurs yellow to red. Clypeus with small round regular punctures separated by half to one puncture width, surface dulled by coarse regular reticular shagreening, often with short median longitudinal carina in apical half, clypeus slightly protruding beyond eyes in profile but by no more than half width of eye; supraclypeal area sculptured as clypeus but often impunctate medially; galeae dulled above by dense, fine tessellation; maxillary palpal segments in ratio of about 4.5:3.5:3.3: 1.0; vertex with flattened lateral areas with small round punctures separated mostly by one to three puncture widths, surface dulled by irregular reticular shagreening. Mesoscutum with deep round punctures separated by half to one puncture width, slightly larger and sparser in posteromedian area, surface shiny, sparsely or not at all shagreened; scutellar punctures similar to mesoscutal but slightly more crowded; metanotum with punctures half diameter of scutel- lar, separated mostly by half to one puncture width, surface dulled by extremely fine, reticular shagreening; propodeum with dorsal surface reticulorugose, coarsely so basally, posterior surface with coarse punctures except upper triangle, lateral surfaces similar to posterior but punctures more crowded, surfaces dulled by dense regular tessellation; mesepisternum with lateral surface with large shallow punctures separated mostly by half a puncture width or less, surface shiny, unshagreened or finely so. Metasomal tergum 1 with basal three-fifths or slightly less punctate (to apex at extreme 384 THe UNIVERSITY SCIENCE BULLETIN sides ), punctures round, shallow, separated mostly by one to two puncture widths, surface dulled by fine tessellation, apical area impunctate with surface shiny, extremely finely reticulotransversely shagreened; tergum 2 with basal zone with minute round punctures separated mostly by half to one puncture width, surface shiny but with fine reticular shagreening, interband zone with small irregular punctures separated by one to three puncture widths, sparser me- dially than in lateral raised areas, surface dulled by reticulotrans- verse shagreening, apical area impunctate or with minute punctures no broader than base of hairs arising from them, separated by two to four puncture widths, surface moderately shiny, with fine reticu- lotransverse shagreening; terga 3 and 4 similar to tergum 2 but punctures of interband zone somewhat more distinct and more abundant and apical zone smaller or absent; pygidial plate broadly V-shaped with rounded apex. Hair: On face and genal areas pale ochraceous to yellow, on ver- tex yellow to bright rufescent, vertex with or without brown hairs (never abundant). Thorax with sides pale ochraceous to pale ru- fescent, above ochraceous to bright rufescent. Metasomal tergum 1 with long basal and lateral hairs ochraceous to yellow or slightly rufescent, apical area glabrous or with sparse, minute, appressed, brown to yellow hairs basally and laterally; tergum 2 with basal pubescence long, white to yellow, interband zone with short; ap- pressed to subappressed, white to pale brown, relatively simple hairs, distal pale band yellow to white, broad laterally (but usually not reaching apical margin) to narrow medially, usually narrowly interrupted medially, apical area with relatively simple, appressed to subappressed, white to yellow hairs obscuring but not completely hiding surface; tergum 3 similar to 2 but basal tomentum dark brown, interband zone hairs dark brown, distal pale band not in- terrupted, apical area shorter; tergum 4 similar to 3 but distal pale band reaches apex across entire tergum, occasionally with minute apicomedial area of brown hairs; terga 5 and 6 dark brown except white to yellow tufts at extreme sides; sterna yellow to reddish brown medially and paler at extreme sides. Legs pale (white to yellow ) except as follows: fore tarsi, often middle basitarsi on outer surfaces, fore and middle tibiae on outer surfaces near apices, hind basitarsi on outer surface at apices, and usually on and surrounding basitibial plates brown; inner surfaces of hind basitarsi with red to reddish brown hairs; scopal hairs ochraceous to yellow. Male. Measurements and ratios: N, 20; length 9-12 mm.; width, 3.0-4.0 mm.; wing length, M = 3.20 + 0.217 mm.; hooks in ees ee BEES OF THE GENUS MELISSODES 385 hamulus, M = 11.65 + 0.274; flagellar segment 2/segment 1, M = 7.37 + 0.164. Structure and color: Integument black except as follows: clypeus and base of mandible yellow; labrum white or cream-colored with brown apical margin (extremely rarely all brown); eyes green to gray or grayish blue; flagellum yellow to red below, dark red to brown above; tegulae usually testaceous, occasionally piceous; wing membranes colorless, veins yellow; apical margins of terga hyaline, colorless to yellow (in some eastern specimens translucent brown); distitarsi rufescent; tibial spurs white to yellow. Clypeus protruding beyond eye in profile by about half width of eye; eyes strongly converging toward mandibles; first flagellar seg- ment with minimum length equal to about one-sixth or one-seventh maximum length of second segment, penultimate segment about three times as long as broad (minimum width, maximum length), flagellar segments without longitudinal lateral depressions; maxil- lary palpal segments in ratio of about 4:3:3:1. Sculpturing as in female except as follows: clypeus with surface often moderately shiny; mesepisternum with surface often somewhat dulled by fine, irregular shagreening; metasomal tergum 1 medially with basal four-fifths to five-sixths with small punctures separated by one to three puncture widths; interband zone of terga 2 and 3 with slightly more abundant and larger punctures, surface moderately shiny, with reticular shagreening; hyaline apical areas of terga 1 to 5 shiny or moderately so, reticulotransverse shagreening extremely fine: sterna with surfaces shiny to moderately so, usually with distinct regular reticular shagreening. Sternum 7 with median plate subtriangular in shape; larger than lateral plate in area, with abundant short hairs ventrally; lateral plate subtriangular, membranous area between plates almost half size of lateral plate in area, narrow; apicomedial margin between median plates with strong curved carinae on each side. Sternum 8 broad near apex; strongly emarginate apicomedially; with ventral tubercle entire or slightly bidentate, not reaching apical margin of sternum; with several to many hairs on apical margin. Gonostylus slender, tapering apically, not distinctly capitate, in length equal to more than half length of gonocoxite, with abundant short hairs laterally on basal half; spatha about three times as broad as long; gonocoxite with spicules of upper inner surface all or mostly pointed or hairlike; penis valve with prominent dorsolateral lamella which ends proximally in an inflected tooth near spatha (Figs. 72-75). 386 Tue UNIVERSITY SCIENCE BULLETIN Hair: Color of vestiture as in female except as follows: generally more males appear pale rather than bright rufescent than females; vertex usually without brown; metasomal tergum 1 with basal four-fifths to five-sixths with long pale hairs and these medially at least long enough to reach apical margin although not abundant enough to completely hide apical area; tergum 2 with distal pale band often not interrupted medially, usually as long as or longer than apical area medially, interband zone usually with abundant long pale suberect hairs; terga 3-5 similar to 2 but basal zone tomentum brown, interband zones usually with scattered pale pubescence as well as suberect hairs and distal pale bands pro- gressively closer to apical margin; terga 6 and 7 with long white to ochraceous or yellow hairs; sterna all pale or yellowish medially; legs white or ochraceous except as follows: inner surfaces of basitarsi and usually distitarsi yellow to reddish yellow. Geographical Variation. Melissodes agilis is distributed through- out the United States (except Florida), southern Canada and north- ern Mexico (Fig. 9). It is remarkably uniform throughout this range; the chief variations being the degree of brightness of the yellow or red vestiture and in size. In neither of these character- istics is there a marked geographical trend. However, male speci- mens from eastern parts of the range tend to have the apical hyaline margins of the metasomal terga darker in color than specimens from elsewhere in the range. Also, specimens, especially females, from the northern prairie states and from the southcentral and south- western provinces of Canada tend to have the vestiture paler or duller in color than elsewhere. In neither of these two cases is there an abrupt step in the clines which must exist and it would be difficult, if not impossible, to delimit clear-cut subspecies. Bionomics. Very little has been published concerning the nest- ing habits of Melissodes agilis, Rau (1922, p. 34) states that agilis was found nesting on a baseball diamond on July 24, 1915. A single female had begun a single burrow. No cells or provisions were yet present. The burrow “. . . was five inches deep, and went downward quite precipitously.” A second female was ob- served by Rau on August 22, 1915 at its horizontal burrow in the face of a clay bank. Custer (1928) reports a female which was possibly M. agilis using the same burrow entrance as eight females of Svastra obliqua (Say) (see LeBerge, 1956, pp. 974, 975). M. agilis is apparently an oligolege of the composite genus Heli- anthus. Robertson (1926, p. 379) lists this species as oligolectic a cee BEES OF THE GENUS MELISSODES 387 on the composite tribes Astereae and Heliantheae. Out of almost 6,000 bees available to the author for study, a total of 2,135 had floral data attached. Out of these 2,135 bees, 1,909 had been taken on some species of the family Compositae and, more specifically, 1,608 had been taken on some species of Helianthus. These data are summarized in Table IV. It is evident that Helianthus pollen plays an overwhelmingly important role in the nutrition of this species. Of other pollens only the Compositae are of much im- portance and no single genus (or even tribe) of the Compositae plays nearly as important a role as does Helianthus. TaBLE IV. Summary of Floral Records for Melissodes agilis. Plant Data | Records of M. agilis Oen nm = es | © hes, an eet 2 iS) an || 0,° O fo) is HS js se es) Ps, Ae) rr 3 2 FAMILY g 5 SES £3 ne ge Ee Z < | Z Zi Z A Compositae (other than Helianthus), 35 | 43 || 144 | 63 255 | 318 Helianthus spp. 1 | 15 || 450 | 428 /1,193 |1,621 Leguminosae Bae] A NUR2 2k) TS aly 556 69 : | | | mei : Brassicaceae 3 AVe|ezler eed. Vie 200 57 Verbenaceae 1 1 1 9 1 By 5 is a I meas | omar a (puro Labiatae Die 23 ||) a3, 2 11 13 Hydrophyllaceae | 2 Wie ees 24 | 3 Others (12) 15 15 18 4 36 40 Totals 64 89 || 670 | 520 |1,645 [2,165 | | | | Type Material. The lectotype male of agilis (No. 2315) with two male paratypes from Texas are in the collection of the Philadelphia Academy of Sciences. The lectotype female of aurigenia (No. 2332) and allotype male (No. 2332.2) both from Colorado are in the col- lection of the Philadelphia Academy of Sciences. The paratypes of aurigenia also in the Philadelphia Academy of Sciences collection include three females from Colorado, Louisiana and Canada, re- spectively, and five males from New York, Virginia, Kansas and Utah. Two female paratypes (Nos. 2332.3 and 2332.5) are not of 388 THE UNIvERSITY SCIENCE BULLETIN the same species but should be referred to Melissodes perlusa Cockerell which is redescribed below. Distribution. The United States (except Florida), southern Can- ada and northern Mexico (Fig. 9). This species has been collected from April until mid-November, but chiefly during the months of July, August and September. In addition to the type specimens, females and males have been examined from the localities listed below (localities reported in the literature are included). ALABAMA: Decatur; Mobile. Arizona: Arlington; Carr Canyon (Huachuca Mts.); Chambers; Douglas (and § miles N. E.); Flag- staff (and 4 miles N., 7 miles S. and 8 miles N. E.); Fredonia; Grand Canyon; Hereford; Mayer; Mesa (6 miles E.); Nogales; Oak Creek Canyon; Palmerlea; Payson; Phoenix; Prescott; Rosemont; San Francisco Mts.; Sedona (15 miles N.); Show Low; Sonoita (10 miles E. and 11 miles W.); Thatcher; Tucson (and 10 miles S.); Turner; Willcox; Williams; Yuma. Ca.irornia: Altadena; Alta- mont (Mt. House Creek); Anaheim; Antioch; Arvin; Bakersfield; Bear Valley; Bishop; Canby; Carbona; Catalina Island; Chino; Clare- mont; Clear Lake (Soda Bay); Clovis; Coalinga; Colton; Corona; Corral Hollow; Costa Mesa; Davis; Dos Palos; Downey; East High- lands; Exchequer; Firebaugh; Fresno; Hemet; Hueneme; Hunting- ton Beach; Huntington Park; Indio; Kingsburg; Lake City; Lake Tahoe; Lancaster; Lemoore; Lone Pine; Los Angeles; Los Banos (5 miles S.); Mira Loma; Modesto; Monrovia Canyon; Mountain View; Oakley; Ontario; Onyx; Orange; Oxalis; Palm Springs; Pasa- dena; Patterson; Pleasanton; Redlands; Redwood City; Reseda; Riverside; Rock Creek; Sacramento; Sacramento Co.; San Bernar- dino Co.; San Jacinto; San Jose; Shafter; Torrance; Tracy; Turlock; Twain Harte; Vernalis; Visalia; Walnut Creek, Contra Costa Co.; Wasco; Westley; Whittier, Wood Lake, Tulare Co.; Woodland Hills. Cotorapo: Antonito; Aurora; Baca Co.; Berkley; Boulder; Brighton; Buckeye; Canfield; Carson Camp; Cimarron; Clear Creek; Colorado Springs; Cortez; Crowley; Denver; Dixon Canyon; Du- rango; Eads; Elbert; Fort Collins; Fruita; Gilpin (Lump Gulch) Co.; Glen Haven; Glen Park; Glenwood Springs; Golden; Golden (Chimney Gulch); Goodview; Grand Junction; Great Sand Dunes, Alamosa Co.; Greeley; Trinidad; Jim Creek (near Boulder ); Jumbo Reservoir; La Junta; Lamar; Leadville; Logan Co.; Longmont; Max- well City; Meeker; Mesa Verde; Midway (5 miles N.); Mt. Manitou, El Paso Co.; Ovid (3 miles E.); Palisade; Pingree Park; Platte Can- yon; Poudre Canyon; Pueblo; Rifle Gap; Rock Creek; Rocky Ford; 389 MELISSODES BEES OF THE GENUS ‘UOSSAID SIIoD (sapossyaun’y ) ‘WW Jo uoRnNqystp uMoUy ay} surmoys dey 6 ‘DIY 390 THE UNIversiry SCIENCE BULLETIN Springfield (3 miles N. in Lone Rock Draw); Sterling; Stratton; Stratton (Landaman Creek); Ten-sheep Ranch; Timpas; Ute Creek (Sage Flats); White Rock (near Boulder); Wray. Connecticut: Colebrook; Storrs; Wallingford; Westville. District or COLUMBIA: Bennings (Eastern Branch); Washington. Gerorcra: Carrollton; Tifton. Ipano: Aberdeen (and 2 and 6 miles N. E.); Bliss; Brun- neau; Buhl; Central Grade, Nez Perce Co.; Coyote Grade, Nez Perce Co.; Downey; Emmett (10 miles E. on Squaw Creek); Fort Hall (near Blackfoot); Franklin; Grandview; Jerome; Lewiston; Midvale (9 miles S. W.); Moscow; Mountain Home; Nampa; Parma; Tuttle; Twin Falls. I_trvors: Algonquin; Ashkcum; Bath; Beards- town; Beverly Hills; Bloomington; Carlinville; Champaign Co.; Charleston; Chicago; Cook Co.; Danville; Decatur; Devil’s Neck (10 miles N. of Havana); Downers Grove; Edgebrook; Evanston; Fulton (3 miles S.); Havana; Hillview; Macoupin Co.; Manitou; McHenry; Meredosia; Mt. Carmel; Oak Park; Seymour; Urbana; Wellington. INprana: LaFayette; Vincennes. Iowa: Ames; Boone (4 miles N. W., 1 mile E. and 5 miles S. E.); Decorah; Ledges State Park; Montpelier; Sergent Bluffs; Sioux City; Vinton. Kansas: Allen Co.; Anderson Co.; Arkansas City; Assaria; Baldwin; Baldwin Junction; Blue Rapids; Burdett; Butler Co.; Caldwell; Chase (5 miles W.); Cheyenne Co.; Clark Co.; Clay Co.; Cloud Co.; Cold- water; Cullison; Decatur Co.; De Soto; Dickinson Co.; Dodge City; Douglas Co.; Edwards Co.; Garden City; Garnett; Great Bend; Harper Co.; Harvey Co.; Hays; Hodgeman Co.; Hoffee; Hutchinson; Jetmore (10 miles S.); Johnson (2 miles N.); Kendall (3 miles E.); Kingman; Kismet; Lake View, Douglas Co.; Lakin; Larned; Law- rence; Liberal; Logan Co.; Lone Star Lake, Douglas Co.; Manhat- tan; Marysville; Mayfield; Meade; Meade Co. St. Park; Medora; Ne- osho Co.; Nickerson; North Topeka; Norton Co.; Olathe; Osborne Co.; Pottawatomie Co.; Reece; Reno; Republic Co.; Richfield (7 miles $.); Riley Co.; Rooks Co.; Russell Co.; Saline Co.; Scott City (8 miles N.); Scott Co.; Sharon Springs; Sherman Co.; Smith Co.; Sunflower; Syracuse; Thomas Co.; Topeka; Tribune (10 miles E.); WaKeeney; Wallace Co.; Wellington; Wichita; Wichita Co.; Wilson Co. LourstaNa: “La.” Marne: Waldoboro. MaryLanp: Cabin John; Chesapeake Beach; Jessups; Lakeland; Plummers Island. Mas- sAcHusETts: Chicopee; Forest Hills; Holden; Lexington; Wellesley; Woods Hole. Micuican: Allegan Co.; Constantine; East Lansing; Grand Rapids; Lewawee Co.; Macomb Co.; Manistee Co.; Midland Co.; Utica. Mrnnesota: Alexandria; Big Stone Co.; Browns Val- BEES OF THE GENUS MELISSODES 391 ley; Evan; Faribault; Grant Co.; Hallock; Hastings; Itasca State Park; Lake City; Lake Vadnais, Ramsey Co.; Mallory; Marshall Co.; Mound Springs State Park, Rock Co.; Muskoda; North Branch; Olmstead Co.; Ortonville; Park Rapids; Pine Co.; Powder Plant Woods, Ramsey Co.; Renville; Rochester; Sedan; St. Anthony Park; St. Cloud; St. Paul; Traverse Co.; Washington Co.; Yellow Medicine Co. Mussissrppr: Camp Shelby (near Hattiesburg); Hattiesburg; West Point. Missourt: Branson; Cameron (10 miles N.); Colum- bia; Holden; Kansas City; Ozark Lake; Sedalia; Shrewsbury; Spring- field; St. Louis. Montana: Fairview; Hamilton; Missoula; White- hall. Nepraska: Agate; Bennet; Bloomington; Bridgeport; Broken Bow; Cambridge; Cedar Bluffs; Collins; Crawford; Fairmont; Ger- ing; Glen; Gordon; Haigler; Halsey; Hamlet; Hardy; Harrison; Hitchcock Co.; Jim Creek, Sioux Co.; Kearney; Kimball; Lincoln; Lodgepole; Long Pine; Malcolm; Maywood; McCook; Mitchell; Monroe Canyon, Sioux Co.; Nebraska City; Neligh, Niobrara; North Platte; Omaha; Oxford; Palisade; Pine Ridge; Sidney; Sioux Co.; Union; Valentine (Valentine Lakes Refuge); Wabash; Warbonnet Canyon, Sioux Co.; West Point. Nervapa: Fallon; Pyramid Lake, Washoe Co.; Reno. New HampsuirE: Hanover. New JERsEy: Gloucester Co.; Monmouth Beach; Orange; Palmyra; Ramsey; Snake Hill. New Mexico: Albuquerque; Alto; Belen; Capitan; Carrizozo; Cuervo; Fort Wingate; Gallup; Jemez Springs; Las Cruces; Las Vegas; Madrid; Magdalena Mts.; Maxwell; Mesilla; Omega; Organ Mts.; Portales; Rio de los Frijoles; Romeroville; Rowe; Sandia Mts.; San Ignacio; San Jon; Santa Fe (and 12 miles S. E.); Sapello; Socorro; Taos; Vaughn. New York: Albany; Astoria (Long Is- land); Bronx Park; Brooklyn; Buffalo; Central Park; Elizabethtown; Gloversville; Great Kills (Staten Island); Hope; Ithaca; Keene Val- ley, Essex Co.; New Baltimore; New Rochelle; New Russia; Orient; Pelham; Wilmington. Norra Caroiina: Aberdeen; Biltmore; Boone; Bostic; Burgaw; Burlington; Newton; Sanford; Smokemont. Nortu Daxora: Bismarck; Clifford; Dickinson; Drake; Edgeley; Fargo; Grafton (4 miles E.); Lakota; Mandan; Marmarth; McKen- zie; Medora; Monango; Mott; New Rockford; Oakdale; Oakes; Pleasant Lake; Rugby; Schafer; Sheldon; Steele; Valley City; Wash- burn; Williston. Onto: Barberton; Columbus; Summit Co.; Tiffin. Orecon: Arlington; Corvallis; Cove; Echo; Hood River; Hunting- ton; Ione; Juntura (6 and 8 miles E.); La Grande; North Powder; Ontario; Oregon City; Silver Lake; Summer Lake; The Dalles (and 14 miles E.); Vale (Malheur River Canyon). PENNSYLVANIA: Al- legheny Co.; Harrisburg; North Braddock; Philadelphia; Pittsburg 392 THE UNIVERSITY SCIENCE BULLETIN Ruope Istanp: Kingston; Providence; Scituate. Sours Daxora: Ardmore; Buffalo (3 miles S.); Cedar Pass (Badlands); Clearfield; Custer; Deadwood (and 10 miles S.); Elk Point; Interior (White River); Jefferson (3 miles S.); Platte. TENNEssEE: Knoxville. Texas: Adrian; Austin; Bexar Co.; Colorado City; Dallas; Del Rio; Fedor, Lee Co.; Galveston; San Angelo. Uran: Avan Canyon; Bear Lake; Bert; Big Cottonwood Canyon (Wasatch Mts. near Fort Bench); Bluff; Bountiful; Cache Junction; Cache Valley; Clo- ver; Collinston; Corinne; Curlew; Delta; Deweyville; Dugway Proving Ground, Tooele Co.; Erda; Eureka; Far West; Ferron; Fill- more; Fort Duchesne; Garfield; Goshen; Grantsville; Green River; Hatton; Helper; Hurricane; Hyrum; Indianola; Jericho; Johnson’s Pass, Tooele Co.; Kaibab Forest; Kelton; Lampo; Lake Point; Lake- town; Lehi; Lincoln; Logan; Logan Canyon; Magna; Moab; Morgan; Mt. Carmel; Mt. Zion National Forest; Myton; Nephi; Oak City; Og- den; Paragonah; Park Valley; Parowan; Penrose; Petersboro; Pintura; Plain City; Pleasantview; Price; Promontory; Providence; Provo; Richfield; River Heights; Rockville; Roosevelt; Rozelle; Salt Lake City; Saltair; Sevier; Silver City; Skull Valley; Smithfield; Spanish Fork Canyon; Springville; Thistle; Timpie; Tooele; Topaz; Utah Lake; Valley City Junction; Washakie; West Utah Lake; Zion Na- tional Park.* VerMont: Townshend; Woodstock. Virceria: Camp Peary; Falls Church; Four-mile Run (Near mouth of). WasxHtnc- TON: Clarkston; Lind; Maryhill; Penawawa; Pullman, Riparia; Sun- nyside; Walla Walla; Wawawai; Wenatchee; Yakima. Wzusconsrn: East Farmington; Hudson; Maiden Rock; Milwaukee; Prescott; Randall, Burnett Co.; Warrens; Yellow River (mouth of), Burnett Co. Wyominc: Casper; Cheyenne; Clifton; Diamond Ranch, Platte Co.; Flat Creek; Grand Teton National Park; Granite Canyon, Laramie Co.; Green River; Laramie (28 miles E.); Sheridan; Wheatland (N. Fork of Green River); Worland; Yellowstone Na- tional Park. Canada. A.perta: Calgary; Lethbridge; Medicine Hat; Morrin; Scandia; Suffield; Taber (Oldman River); Welling; Whitla. British Cotumsra: Ashcroft Manor (3 miles W.); Fair- view; Kamloops; Keremeos; Lillooet; Nicola; Okanagan Falls; Oli- ver; Spencers Bridge (15 miles E.); Summerland; Thompson River; Vernon; Walhackin. Manrrosa: Altona; Aweme; Brandon; Lauder; Treesbank. Norrawest Territory: “N. W. T.” (mislabeled or misinterpreted? ). Ontarro: Guelph; Ottawa; Stroud; Toronto. México. Curmuanua: Allende; Jiménez (10 and 17 miles W.); Ciudad Juarez; Salaices. Coanuia: San Pedro de Colonias. Dv- RANGO: Torreon. BEES OF THE GENUS MELISSODES 393 Flower Records. In this list are included flower records reported in the literature. Abutilon theophrasti, Althaea rosea, Aplopappus spinulosus, Arctium sp., Argemone sp., A. platyceras, Bidens aristosa, B. laevis, Blephilia hirsuta, Brassica juncea, Brauneria pallida, Car- duus crispus, Carya pecan, Cassia sp., C. chamaecrista, C. fascicu- lata, Centromadia pungens, Chrysopsis hispidus, Chrysothamnus sp.., Cirsium sp., C. altissimum, C. discolor, C. lanceolatum, C. undu- latum, Clematis sp., Cleome sp., C. lutea, C. serrulata, Convolvulus sp., Coreopsis sp., C. lanceolata, C. palmata, C. tripteris, Cosmos sp., Datura metaloides, Enceliopsis sp., Engelmannia pinnatifida, Erica- meria palmeri, Eupatorium sp., E. purpureum, Eustoma artemifo- lium, Gaillardia sp., G. cristata, Gilia sp., Grindelia sp., G. squarrosa, Gutierrezia sp., G. sarothrae, Haplopappus sp., Helenium autum- nale, H. laciniatum, Helianthus sp., H. annuus, H. atrorubens, H. bo- landeri, H. ciliaris, H. coronarius, H. divaricatus, H. grosse-serratus, H. laetiflorus, H. lenticularis, H. maximillianus, H. mollis, H. petio- laris, H. pumulis, H. radulus, H. rigidus, H. salicifolius, H. scaberri- mus, H. subrhomboideus, H. tuberosus, Heliopsis sp., Heliotropium sp., Hibiscus sp., Ipomoea sp., Lactuca pulchella, Lepachys pinnata, Liatris pycnostachya, Medicago sativa, Melilotus sp., M. alba, Men- tha canadensis, Monarda fistulosa, Penstemon sp., Pepo sp., Petalo- stemum sp., P. occidentale, P. purpureum, Phacelia sp., Physostegia sp., P. parviflora, Platycodon grandiflorum, Pluchea camphorata, Pri- onopsis sp., Pycnanthemum flexuosum, P. pilosum, Pyrrhopappus multicaulis, Rudbeckia hirta, R. laciniata, R. triloba, Schrankia un- cinata, Senecio sp., Silphium sp., S. integrifolium, S. laciniatum, S. perfoliatum, S. speciosum, Sium cicutaefolium, Solidago sp., S. cana- densis, S. serotina, S. trinervata, Teucrium canadense, Verbena sp.., V. hastata, V. stricta, Verbesina sp., V. encelioides, V. exauriculata, V. occidentalis, Vernonia sp., V. baldwini interior, V. fasciulata, Ve- ronica sp., Vitex agnus-castus, Wislizenia refracta. Melissodes (Eumelissodes) trinodis Robertson Melissodes trinodis Robertson, 1901, Canadian Ent., vol. 33, p. 231; 1905, Trans. Amer. Ent. Soc., vol. 31, p. 369; Graenicher, 1905, Bull. Wisconsin Nat. Hist. Soc., vol. 3, p. 164-165; 1911, Bull. Pub. Mus. Milwaukee, vol. 1, p. 247; Smith, 1910, Ann. Rept. New Jersey State Museum, p. 693; Robertson, 1914, Ent. News, vol. 25, p. 70; 1926, Psyche, vol. 33, p. 119; 1928, Flowers and Insects, p. 8; Pearson, 1933, Ecol. Monogr., vol. 3, p. 381; Graenicher, 1935, Ann. Ent. Soc. Amer., vol. 28, p. 304; Brimley, 1938, Insects of North Caro- lina, p. 463. Moliseodes pennsylvanica, Robertson, 1897, Trans. Acad. Sci. St. Louis, vol. 7, p. 355 (misidentification ). This species is closely related to M. agilis. The female of trinodis is very similar to that of agilis, differing chiefly in the darker color 394 THE UNIveRSITY SCIENCE BULLETIN as described below. The male can be separated from that of agilis by the piceous tergal margins, the black mandibular bases, the darker wing veins and the often less densely shagreened galeae. Female. Measurements and ratios: N, 20; length, 10-12 mm.; width, 3.5-4.0 mm.; wing length, M = 3.45 + 0.094 mm.; hooks in hamulus, M = 12.50 + 0.224, flagellar segment 1/segment 2, NSS Tig == 0025. Structure and color: Integument black except as follows: apical half of mandible, distitarsus usually and flagellar segments 3 to 10 below rufescent; tegulae testaceous to piceous; eyes yellow to dark gray; wing membranes hyaline, veins dark reddish brown to black; tibial spurs yellow. With structural characteristics of agilis except as follows: clypeus protruding beyond eye in profile usually by less than half width of eye, punctures round, regular, separated mostly by about half a puncture width; vertex with flattened lateral areas usually mod- erately shiny; maxillary palpal ratio about 4.25:3.50:3.75:1.00; galeae finely tessellate above; mesoscutum with punctures separated mostly by half a puncture width or less (including posteromedian area); mesepisternal punctures usually extremely shallow, surface dulled by fine, irregular shagreening; metasomal tergum 2 with interband zone punctures shallow, small, often absent at least; medially, sep- arated mostly by more than one puncture width. Hair: Color of vestiture as in agilis except as follows: labrum and apical half of clypeus often reddish brown; vertex usually with abundant dark brown hairs (occasionally all pale); head and thorax bright rufescent (rarely ochraceous); mesoscutum rarely with a few brown hairs in posteromedian area and scutellum occasionally with brown hairs medially; metasomal tergum 2 with interband zone hairs subappressed to erect, dark brown, apical area with short, appressed, relatively simple, dark brown hairs, distal pale band white and interrupted medially; tergum 3 with apical area as in tergum 2, distal pale band white and narrowly to broadly inter- rupted medially, rarely distal pubescent band all brown; tergum 4 brown at least in small apicomedian patch, often across all or most of apical margin of tergum and often interrupting distal pale band medially, occasionally tergum 4 all brown; terga 5 and 6 usually with pale lateral tufts but in darkest specimens all brown; sterna brown to dark brown except pale at extreme sides. Legs with hairs dark brown except as follows: femora yellow to rufescent; fore and middle tibiae and middle basitarsi often pale basally on outer BEES OF THE GENUS MELISSODES 395 surfaces; tibiae with inner surfaces yellow to red; middle and hind basitarsi with inner surfaces dark red to brown; scopal hairs ochraceous to yellow except usually brown on and surrounding basitibial plates and near apices of basitarsi. Male. Measurements and ratios: N, 20; length, 10-11 mm.; width, 3.5-4.0 mm.; wing length, M = 3.24 + 0.152 mm.; hooks in hamulus, M = 11.25 + 0.054; flagellar segment 2/segment 1, M = 9.06 + 0.156. Structure and color: Integment black except as follows: labrum with large mediobasal cream-colored macula; mandible rarely with small basal yellow macula; clypeus yellow; flagellar segments 2-11 red beneath, brown above; eyes green to yellowish brown or gray; wing membranes hyaline, colorless or slightly milky, veins reddish brown to black; tegulae usually piceous, occasionally testaceous; tarsi rufescent; tibial spurs white to yellow; apical margins of terga piceous to dark brown. Structure as in agilis with the following differences: clypeus pro- truding beyond eye in profile by slightly less than half width of eye; minimum length of first flagellar segment equals one-seventh or less (often less than one-eighth) of maximum length of second segment; maxillary palpal ratio about 8:7:6:1. Sculpturing as in female except as follows: galeae above usually finely tessellate in apical half, often slightly so posteriorly, occasionally shiny except at tips; mesoscutal punctures often separated by more than half but less than one puncture width in posteromedian area; mesepisterna punctures usually deep, surface shiny to somewhat dulled by irreg- ular shagreening; metasomal tergum 1 with basal five-sixths (me- dially ) punctate, punctures smaller and sparser apically, in basal half separated mostly by half to one puncture width; tergum 2 with interband punctures distinct, separated by one to two puncture widths, apical area without distinct punctures; terga 3-5 similar to tergum 2 but interband punctures more regular in size, more crowded and smaller. Sternum 7 similar to agilis but with median plate subtriangular, subequal or slightly larger than lateral plate in area, with ventral hairs minute. Sternum 8 as in agilis but ventral tubercle low, rounded, entire, with several short weak hairs at apex on either side of median emargination. Genital capsule as in agilis except as follows: gonostylus slender, indistinctly capitate, equal to about half length of gonocoxite, with short hairs laterally and minute hairs on inner surface (Figs. 76-77). 396 THe UNIVERSITY SCIENCE BULLETIN Hair: Color of vestiture as in female except as follows: without brown on clypeus, labrum or vertex; mesoscutum and scutellum without brown; terga 2-4 with apical areas with suberect brown hairs (often absent due to wear ), distal pale bands usually complete but occasionally interrupted medially at least on tergum 4; tergum 5 often with a complete distal pale band, occasionally absent or broadly interrupted medially; terga 6 and 7 brown to almost black; sterna rufescent or golden medially and ochraceous to white lat- erally; legs with ochraceous hairs except inner surfaces of tarsi and hind tibiae yellowish. Thoracic and head hairs usually bright rufescent above and somewhat duller at sides, occasionally all thoracic and head hairs dull ochraceous. Remarks. This bee resembles M. agilis very closely, especially in the female sex. It is quite possible that some paler females of trinodis have been identified as agilis or vice versa. That trinodis might be merely a variant of agilis appearing sporadically has been seriously considered. This hypothesis has been rejected on the basis that the males of agilis and trinodis are quite distinct and because trinodis does not appear west of the Great Plains where agilis is very abundant. There is no uniform geographical variation in trinodis which could serve for the recognition of subspecies. However, females from the eastern seaboard, particularly from North Carolina, tend to be much darker in color than elsewhere in the range of the spe- cies. Some few females from this area resemble the females of dentiventris Smith and bidentis Cockerell (see below) on this ac- count. Bionomics. The only observations on the nesting habits of this species were published by Graenicher (1905, pp. 164-165). Grae- nicher briefly describes the nest of M. trinodis in connection with his studies of bee parasites and, in particular, in connection with the bee Triepeolus helianthi Robertson which was found parasitizing M. trinodis. Graenicher states, “. . . a ground inhabiting bee digs down perpendicularly to a depth of 8 cm., then turns off obliquely for a short distance, and continues in a perpendicular di- rection. The cells are somewhat thimble-shaped, their walls are formed of hardened clay with a very smooth and polished inner surface. They are filled about one-half with bee-bread.” Grae- nicher observed Triepeolus helianthi entering the M. trinodis nest and he opened the nest on the following day. He observed two cells, one unfinished, the other closed and showing a white, opaque BEES OF THE GENUS MELISSODES 397 egg of the host, 3 mm. in length, on the surface of the bee-bread. Graenicher then describes the activities of the first and second stage larvae of T. helianthi and, although he later dug up an additional nest of M. trinodis, he records no additional observations on this bee. In regard to flower preferences, Robertson (1926) regarded M. trinodis as an oligolege of Aster, Heliantheae and Heleniae. This is very nearly correct, although the data available to the present author suggests that the genus Aster and the tribe Heleniae are not preferred markedly more than many other composites which these bees visit. On the other hand, the tribe Heliantheae, and particu- larly the genus Helianthus, is of the greatest importance as a source of pollen. The bee would be better described as an oligolege of the Compositae and in particular of Helianthus. The data supporting these conclusions is summarized in Table V. TaBLeE V. Summary of Floral Records for Melissodes trinodis. Plant Data Records of M. trinodis % 2 xo) wa) OS | Sa blo g HS nO oy R= a) apes Be ants eS ee | eyes ae 2 go} 4 Si repaile tele rcs alt Med eee = a =n 30 a4 38 62 Z. Z. Z. Z. Zz H | | | | Compositae, other than Helianthus 9 15 28 13 88 LO1 Compositae, Helianthus spp. 1 6 34 30 26 56 Others (6) 6 6 if 4 5 9 Totals 16 27 69 47 119 166 I Type Material. Lectotype female (Robertson No. 9513) col- lected on Helianthus grosse-serratus, September 20, 1890 by Charles Robertson at Carlinville, [linois, and lectoallotype male (Robertson, 8197) taken on Lepachys pinnata, July 25, 1888 by Charles Robert- son at Carlinville are in the collection of the Illinois Natural His- tory Survey, Urbana. In addition nine female and three male paratypes collected by Charles Robertson at Carlinville are in the collections of the Illinois Natural History Survey. Distribution. This species occurs throughout most of eastern United States from Kansas and North Dakota in the west to Maine and Georgia in the east and in southeastern Canada (Fig. 10). Tue UNrversiry SCIENCE BULLETIN 398 ‘aB10ge'] DyDL0]091g (‘q) ‘WW pur ‘esiogey] vyvssniao (“q) “W ‘aplogey] vapiyso0 (“q) “W ‘P49 -yoop vsnpiad (“q) ‘W “UOSaqoYy sipourly (Sapossyawny) ‘WW JO SUOHNGLASTP UMOUDy oy} surmoys dey “OT ‘IY PIPs0j/021q © Wu lw) \- BEES OF THE GENUS MELISSODES 399 It has been collected between July 7 and October 23, but chiefly during August and September. In addition to the type material, a total of 116 females and 206 males from the localities listed be- low (including localities reported in the literature) have been examined, ARKANSAS: Fayetteville. CoNnNectricur: Colebrook; Storrs; West- ville (New Haven). Gerorcra: Atlanta. Itirimors: Beverly Hills; Carlinville; Chicago; Dubois; Evanston; Kankakee; Palos Park; Urbana; West Pullman; Willow Springs. INpraANa: “Ind.” Iowa: Ames (3 miles N. and 6 miles W.); Farragut; Mt. Pleasant (6 miles S. W.); Sioux City. Kansas: Allen Co.; Baldwin; Baldwin Junc- tion; Blue Rapids; De Soto; Douglas Co.; Franklin Co.; Garnett; Lawrence; Manhattan; Marysville; Montgomery Co.; Olathe; Riley Co. Marne: Waldoboro. Massacuusetts: Sherborn. MuicHicaNn: Detroit. Minnesota: Barrett; Big Stone Co.; Dakota Co.; Fair- mont; Faribault; Freeborn Co.; Hayward; Howard Lake; Lake Vadnais, Ramsey Co.; Pine Co.; Powder Plant Woods, Ramsey Co.; St. Paul. Mississrppr: Camp Shelby (near Hattiesburg), Hattiesburg. Missourt: Silver Spring. Nesraska: Lincoln; Louisville; Malcolm; Nebraska City; Neligh; Omaha; West Point. Nrw Jersey: Chester; Ramsey; Salem Co. New York: Brooklyn; Ithaca; New Rochelle; New Baltimore; Quoque (Long Island). Norra Carona: Bostic; Burgaw; Marion; Raleigh. NorrH Dakota: Monango. Onto: Barberton. PENNsyLvANrA: Pittsburgh. TENNESSEE: Knox Co.; Knoxville. Virernra: Falls Church. Wisconsin: East Farmington; Hudson; Milwaukee; Randall; Yellow River (mouth of ), Burnett Co. Flower Records. Arctium sp., Asclepias incarnata, Aster sp.; A. anomalus, A. praeatus, Bidens aristosa, B. laevis, Blephilia hir- suta, Carduus crispus, Cassia chamaecrista, Cirsium sp., C. lanceo- latum, Coreopsis palmata, C. tripteris, Dichophyllum marginatum, Grindelia sp., Helenium altissimum, H. autumnale, Helianthus sp., H. annuus, H. annuus coronarius, H. atrorubens, H. divaricatus, H. grosse-serratus, H. maximillianus, H. mollis, H. salicifolius, H. tuberosus, Heliopsis helianthoides, Lepachys sp., L. pinnata, Liatris sp., Monarda fistulosa, Pepo sp., Petalostemum purpureum, Ratibida columnaris, Rudbeckia sp., R. hirta, R. laciniata, R. subtomentosa, R. triloba, Silphium sp., S. integrifolium, S. laciniatum, S. perfolia- tum, Solidago sp., S. canadensis, S. rupestris, S. ulmifolia, Symphori- carpos sp., Teucrium canadense, Verbena sp., V. hastata, V. stricta, Vernonia sp., V. glauca, V. baldwini interior, Veronica sp. 400 THe UNtIversiry SCIENCE BULLETIN Melissodes (Eumelissodes) bidentis Cockerell Melissodes bidentis Cockerell, 1914, Ann. Mag. Nat. Hist., ser. 8, vol. 14, p. 362; Stevens, 1951, Bull. North Dakota Agric. Exp. Sta., No. 14, p. 31. This small bee is related to both M. agilis and M. trinodis. It is very similar in color to the darker specimens of trinodis. The fe- male of bidentis can be distinguished from both trinodis and agilis by the shiny, unshagreened galeae, and from agilis and most speci- mens of trinodis by the lack of pale bands on the abdomen. The male of bidentis is similar to that of trinodis in having black mandib- ular bases and dark wing veins, but can be separated from the latter by the longer first flagellar segments and by the unshagreened galeae. Female. Measurements and ratios: N, 20; length, 10-12 mm.; width, 4.0-4.5 mm.; wing length, M = 3.36 + 0.130 mm.; hooks in hamulus, M = 12.10 + 0.216; flagellar segment 1/segment 2, Mi erie 00214 Structure and color: Integument black except as follows: apical half of mandible, flagellar segments 3-10 below, distitarsi, occa- sionally basitarsi, tibiae and femora, rufescent; eyes grayish green; wing membranes slightly infumate, yellowish, veins dark reddish brown to black; tegulae piceous; tibial spurs yellow. Structure and sculpturing as in agilis except as follows: clypeus usually with median carina in apical half; supraclypeal area with small, round, scattered punctures, densely tessellate; galeae shiny, unshagreened except at extreme tips; maxillary palpal ratio about 2.67:2.33:2.67:1.00, last segment sometimes shorter; vertex with lateral flattened areas shiny; mesoscutum with round punctures smaller than in agilis, in posteromedian area separated by one to three puncture widths, surface dulled by fine reticular shagreening; scutellum with surface dulled by reticular shagreening; mesepi- sterna with large, shallow punctures separated by half to one punc- ture width, surface shiny; metasomal tergum 1 with basal two- to three-fifths with small shallow punctures separated by one to two puncture widths, apical area impunctate, surface dulled by reticulo- transverse shagreening; terga 2-4 as in agilis but punctures (espe- cially in interband zone) smaller, sparser, and shallower and sha- greening coarser. Hair: Head pale rufescent with abundant dark brown on vertex and labrum and often with clypeus all or partly brown. Thorax above dark ochraceous to bright rufescent, laterally ochraceous to rufescent except dark brown on anterior and lower lateral surfaces ij LCT BEES OF THE GENUS: MELISSODES 401 of mesepisterna. Metasoma dark brown to black except as follows: tergum 1 with long ochraceous hairs basally; tergum 2 with basal zone pubescence ochraceous; often some pale hairs on lateral sur- faces of terga 2 to 4. Legs dark brown except as follows: inner surfaces of basitarsi and often distitarsi dark reddish brown to black; scopal hairs ochraceous except brown on distal part of basi- tarsi and on and surrounding basitibial plates. Male. Measurements and ratios: N, 16; length, 10-11 mm.; width, 3.5-4.0 mm.; wing length, M = 3.08 + 0.1385 mm.; hooks in hamulus, M = 10.81 + 0.136; flagellar segment 2/segment 1, M = 5.21 + 0.093. Structure and color: Integumental color as in trinodis except as follows: flagellar segment 2 usually dark brown; metasomal terga with apical areas infumate but usually slightly translucent, brown. Structure as in trinodis except as follows: minimum length of first flagellar segment equals one-fifth or more of maximum length of second segment; maxillary palpal ratio about 3:3:2:1. Sculptur- ing as in female except as follows: supraclypeal area often impunc- tate but densely tessellate; mesoscutum and scutellum with punc- tures slightly larger, more crowded, surface moderately shiny, less dulled by shagreening; mesepisterna with punctures deeper; meta- somal tergum 1 with basal three- to four-fifths punctate, punctures basally large, distinct, separated by one to three puncture widths; terga 2-5 as in terga 2-4 of female but apical areas impunctate. Sternum 7 as in trinodis but median plate somewhat larger rela- tive to lateral plate. Sternum 8 as in trinodis but medioventral tubercle pointed. Gonostylus short, broad, indistinctly capitate, with sparse minute hairs basolaterally; gonocoxites with mediodor- sal margins forming a semicircle basad of spatha due to somewhat produced tubercle at margin of each gonocoxite just basad of either end of spatha; spatha just or slightly less than twice as broad as long, without well-defined apicomedial emargination; penis valve narrow, with lateral process short and blunt (Figs. 78-79). Hair: Head and thorax ochraceous to bright rufescent, usually rufescent above. Metasomal tergum 1 ochraceous basally and laterally to apical margin, apicomedially with short, suberect, dark brown hairs; tergum 2 with basal pubescence ochraceous to white, distal pale band extremely narrow, usually interrupted medially, not connected at sides to basal pale band, separated from apex of tergum laterally by about length of pale band, interband zone with suberect to erect, short, dark brown hairs, apical area with short, 402 THe UNIVERSITY SCIENCE BULLETIN subappressed to appressed, simple, dark brown hairs (often worn away); terga 3-5 similar to tergum 2 but basal tomentum dark brown and distal pale band often more broadly interrupted (on tergum 5 usually and tergum 4 occasionally distal pubescent band brown); terga 6 and 7 brown, sterna yellow to brown medially, pale laterally; legs with ochraceous to yellow hairs except inner surfaces of basitarsi and often distitari yellow to orange. Remarks. As in the case of most species of Eumelissodes, M. bidentis is an oligolege of the Compositae and seemingly depends primarily upon plants of the tribe Heliantheae. However the data is at present extremely sparse. Fic. 11. Map showing the known distributions of M. (Eumelissodes) bidentis Cockerell and M. (E.) manipularis Smith. BEES OF THE GENUS MELISSODES 403 Type Material. The holotype female of bidentis collected at West Point, Nebraska, September 21, 1903, on Bidens sp. by J. C. Crawford is in the U. S. National Museum (U. S. N. M. Type No. 22913). I have also examined one paratype with the same data as the holotype except that it was collected on September 22, 1903, which is in the American Museum of Natural History, New York City. Distribution. The known range of bidentis extends from North Dakota to Texas and east to western New York State (Fig. 11). This species has been collected from July 16 to October 8. A total of 26 females and 16 males have been examined (including the holotype) from the localities listed below. Iowa: Decorah. Minnesota: Big Stone Co.; Carver Co. (Zum- bro Heights ); Hastings; Plummer. Nersraska: Omaha; South Sioux City; West Point. New York: Ithaca. Norra Daxora: Fargo; Lakota; Sheldon. Sour Dakota: Brookings. TrExas: Paris. Flower Records. Bidens sp., Echinacea pallida, Gossypium herbaceum, Helianthus annuus, H. maximillianus, H. tuberosus, Physostegia parviflora, Rudbeckia sp., R. laciniata, Sonchus arven- sis, Melissodes (Eumelissodes) dentiventris Smith Melissodes dentiventris Smith, 1854, Cat. Hymen. in Coll. British Mus. Part II. Apidae, p. 321; Robertson, 1894, Trans. Acad. Sci. St. Louis, vol. 6, pp. 463, 467, 469-471, 473-476; 1896, Trans. Acad. Sci. St. Louis, vol. 7, pp. 176-178; 1897, Trans. Acad. Sci. St. Louis, vol. 7, p. 355; 1898, Trans. Acad. Sci. St. Louis, vol. 8, p. 53; 1901, Canadian Ent., vol. 33, p. 230; 1902, Canadian Ent., vol. 34, p. 49; Smith, 1910, Ann. Report New Jersey State Mus., 1909, p. 693: Viereck, 1916, Connecticut St. Geol. and Nat. Hist. Surv. Bull. No. 22, p. 732; Cockerell, 1917, Canadian Ent., vol. 9 is oye Ld Melissodes autumnalis Robertson, 1905, Trans. Amer. Ent. Soc., vol. 31, p. 369; Cockerell, 1906, Trans. Amer. Ent. Soc., vol. 32, p. 114; Robertson, 1914, Ent. News, vol. 25, p. 70; 1926, Psyche, vol. 33, p. 119; 1926, Ecology, vol. 7, p. 379; 1928, Flowers and Insects, p. 8; Brimley, 1938, Insects of North Carolina, p. 462; Michener, 1947, Amer. Midl. Nat., vol. 38, p. 453. Melissodes megacerata Cockerell, 1906, Ann. Mag. Nat. Hist., ser. 7, vol. 17, p. 362 (new synonymy ). The female of dentiventris resembles closely the darkest female of trinodis and those of bidentis. The female of dentiventris can be separated from those of the latter two species by its flatter clypeus which is more closely allied with the paraocular carina laterally, by its densely tessellate galeae and by its more coarsely punctured metasoma as described below. The male is distinguished by usually lacking pale pubescent bands on the metasomal terga (a complete band is present only on tergum 2 of some specimens ). 404 THe UNIVERSITY SCIENCE BULLETIN the dulled galeae, the extremely short first flagellar segment, and by the black mandibles and labrum (often posterior part of clypeus is also darkened ). Female, Measurements and ratios: N, 20; length, 12-14 mm.; width, 4.5-6.0 mm.; wing length, M == 4.18 + 0.058 mm.; hooks in hamulus, M = 14.75 + 0.064; flagellar segment 1/segment 2, M = 1-74 == 0015; Structure and color: Integument black except as follows: apical half of mandible, lower surface of flagellar segments 3-10 and distitarsi rufescent; eyes gray to blue; wing membranes slightly infumate, yellow, veins dark reddish brown to black; tegulae pic- eous; tibial spurs usually reddish brown. Clypeus flat, extreme lateral margin separated from eye margin by less than half minimum diameter of first flagellar segment, with coarse round punctures separated by half a puncture width or less, with well-marked median carina in apical half, surface slightly dulled by reticular shagreening; supraclypeal area usually impunc- tate and shiny medially; flattened lateral areas of vertex with round deep punctures separated by half to one puncture width, surface shiny; galeae above dulled by fine tessellation; maxillary palpal ratio about 2.50:2.67:2.50:1.00, rarely with minute fifth segment. Mesoscutum with large, deep, round punctures separated mostly by half a puncture width (by less in posteromedian declivous area), surface shiny; scutellum similar; mesepisternum with punctures similar in size and spacing to middle of mesoscutum, surface shiny; propodeum with dorsal surface reticulorugose basally and punctate apically, lateral and posterior surfaces coarsely punctate, surfaces dulled by dense, coarse tessellation. Metasomal tergum | with basal three-fourths with deep round punctures separated mostly by half to one puncture width, apical area impunctate, surface moderately shiny, with reticulotransverse shagreening (especially basally); tergum 2 with basal area with small deep punctures separated mostly by half a puncture width or less, interband zone with larger, shallow punctures separated mostly by one to two puncture widths, apical area with small punctures two to three times width of hairs arising from them, surface shiny to moderately so, shagreening extremely fine; terga 3 and 4 similar to tergum 2 but punctures of interband zones more crowded; pygidial plate broadly V-shaped with rounded apex. Hair: Head usually ochraceous with abundant dark brown on vertex, often with dark brown mixed with pale on clypeus and on BEES OF THE GENUS MELISSODES 405 frons down to level of antennal fossae, rarely most of head hairs dark. Thorax dark ochraceous to bright rufescent above and on upper half of lateral surfaces, rarely with a few dark brown hairs in posteromedial area of mesoscutum and medially on scutellum, an- teriorly and lower lateral surfaces dark brown; tegulae dark brown at least posteriorly. Metasomal vestiture usually entirely dark brown to black except ochraceous (or ochraceous and dark mixed ) on basal half of tergum 1 and ochraceous pubescence at extreme base of tergum 2; occasionally tergum 2 with thin distal pubescent band also ochraceous or ochraceous laterally; rarely tergum 3 with distal pubescent band light brown or dark ochraceous laterally. Legs dark brown to black except as follows: inner surfaces of fore and middle tarsi and tibiae reddish brown, inner surfaces of hind bastitarsi often dark reddish brown, scopae yellow to ochraceous except brown at apices of basitarsi and on and surrounding basi- tibial plates. Male. Measurements and ratios: N, 20; length, 10-13 mm.; width, 3.5-5.0 mm.; wing length, M = 4.12 + 0.188 mm.; hooks in hamulus, M = 13.60 + 0.234; flagellar segment 2/segment 1, M=9.6]1 + 0.281. Structure and color: As in female except as follows: clypeus yellow, often dark brown along posterior margin, rarely brown ex- cept median one-third; flagellar segments 2-11 yellow to red below; apices of metasomal terga often slightly translucent; tibial spurs usually yellow. Structure as in female with following differences: minimum length of first flagellar segment usually subequal to pedicel and equal to about one-tenth of maximum length of second segment; maxillary segments in ratio of about 2.0:2.5:2.0:1.0, rarely with fifth segment; basitibial plate rounded apically. Sculpturing as in female except as follows: clypeal punctures less distinct; metasomal tergum 1 with basal four-fifths punctate; terga 3-5 similar to tergum 2 but interband zone punctures smaller and more crowded, apical areas less distinctly punctured than in female. Sternum 7 as in agilis but with median plate subequal to slightly larger than lateral plate in area, with abundant minute hairs ven- trally becoming long and coarse basally, and with a few minute, curled hairs directed inwards from dorsum of inner basal angle. Sternum 8 as in agilis but ventral tubercle strong and acute. Genital capsule as in agilis except as follows: gonostylus equals about half of gonocoxite in length, short, thick hairs basally on ventral surface 406 THe UNIveRSITY SCIENCE BULLETIN with tips split into two or three minute tines; gonocoxite with spicules of inner upper surface half hairlike and half short blunt structures (Figs. 80-81). Hair: Head ochraceous to rufescent (especially on vertex). Thorax pale ochraceous to yellow laterally and ochraceous to bright ferrugineous above. Metasomal tergal vestiture dark brown except as follows: basal half or more of tergum 1 with long pale hairs (reaching apical margin at extreme sides); tergum 2 with basal pubescence pale, distal pubescent band narrow, often white at least in lateral thirds and occasionally across entire tergum; tergum 3 with distal pubescent band occasionally pale laterally; terga 2 and 3 and often 4 with pale hairs along extreme sides. Legs pale ochra- ceous to yellow except as follows: inner surfaces of basitarsi dark red; inner surfaces of distitarsi and hind tibiae often yellowish red; basitibial plates often pale brown. Bionomics. This species is, according to Robertson (1926), an oligolege of the composite tribes Astereae and Heliantheae. Accord- ing to the data gathered from specimen labels, dentiventris is dependent primarily upon flowers of the tribe Astereae (and par- ticularly upon the genus Aster) and only secondarily upon other composites. Out of a total of 39 collections (76 bees) with flower data attached, 24 collections (47 bees of which 44 are females) were from some species of the genus Aster. The other 15 collections (29 bees of which only 17 are females) were taken on seven other genera of composites and of these the genus Chrysopsis ( Astereae ) was the most important. Type Material. The holotype male (No. 17.B.834) of dentiventris from Georgia is in the collection of the British Museum (Natural History) in London, England. The lectotype female (here desig- nated) of autumnalis from Carlinville, Illinois, taken on Aster ericoides villosus on September 21, 1895, by Charles Robertson (Robertson No. 17,765), and the lectoallotype male (here desig- nated) collected at Carlinville by Robertson (Robertson No. 18,- 670) are in the collection of the Ilinois Natural History Survey, Urbana, Illinois. In addition, 23 female and 18 male paratypes of autumnalis are also at Urbana. The male holotype of megacerata collected by G. Birkmann on October 13, 1897, at Fedor, Lee Co., Texas, is in the collection of P. H. Timberlake at the Citrus Ex- periment Station, Riverside, California. 407 BEES OF THE GENUS MELISSODES “UOSSIID) umous ( ‘oBIogey siqoug (7) ‘WW pure A) OW “YWutg siquaaijuap (Sapossyawn gy) “FW JO suUOQNqLyAsSIp UMoUYy 94} SuIMoys dey “Zl “OY StsgUoArguap 408 THe UNIVERSITY SCIENCE BULLETIN Distribution. This species ranges from eastern Texas north to South Dakota and east to Georgia and southeastern Canada ( Fig. 12). It has been collected from July 4 to October 27, but chiefly in September and October. In addition to the type material, 270 fe- males and 108 males from the localities listed below have been ex- amined. This list includes records reported in the literature. ALABAMA: Kushla; Mobile; Saraland; Selma. Arkansas: Hope; Hot Springs. Connecticut: Branford; Rockville; Westville. Dis- TRICT OF CoLtuMBiA: Washington. Grorcra: Atlanta; Griffin; Neal Gap; Thomsons Mills; Tifton. Itirors: Bluffs; Carlinville; Elsah, Jersey Co.; Macoupin Co.; Peoria. Inprana: Elkhart; Gibson Co.; Rush Branch. Kansas: Lawrence; Lone Star Lake, Douglas Co. Kentucky: Louisville. MaryLanp: Beltsville; Bethesda; Bladens- burg; Cabin John; Glen Echo; Hyattsville. | MAssacHuseErts: Edgartown; Truro; Woods Hole. Micuican: Hillsdale Co.; Kala- mazoo Co. Mussissrppi: Camp Shelby (near Hattiesburg ); Hatties- burg; Natchez. Missourr: Branson; Columbia; Louisiana; Ozark Lake; St. Louis. Nrw Jersey: Clementon; Da Costa; Gloucester Co.; Iona; Lakewood; Maplewood; Riverton. NEw York: Long Island (Bellmore; East Quoque; Flatbush; Greenport; Montauk; Northwest; Orient). NorrH Carotina: Black Mt.; Bryson City; Burgaw; Crabtree Creek State Park; Davidsons River; Faison; Har- nett Co.; Lake View; Lumberton; Laurinburg; New River; Norlina; Pender Co.; Pikeville; Raleigh; Reidsville; Richmond Co.; Spout Spring; Southern Pines; Tarboro; Umstead State Park; Wake Co.; West Raleigh; Yadkin Co. Onto: Franklin Co. OxLaHoma: Tuskahoma, PENNsyLvanta: Eberlys Mill; Philadelphia. Souru CaroLina: Greenville. TENNEssEE: Maury Co. Texas: Fedor, Lee Co, Virernta: Barcroft; Falls Church; Fort Humphreys; Four- mile Run (near mouth of). Canada. Onrario: Ottawa. QUEBEC: Cap Rouge. Flower Records. Aster sp., A. anomalus, A. dumosus, A. eric- oides, A. ericoides villosus, A. novaeangliae, A. paniculatus, A. sagit- tifolius, A. turbinellus, Bidens aristosa, B. polylepis, Boltonia aster- oides, Coreopsis tripterus, Chrysopsis sp., C. mariana, C. micro- cephala, Eupatorium perfoliatum, E. serotinum, Helianthus sp., H. annuus, H, divaricatus, H. grosse-serratus, H. radula, Isopappus divaricatus, Lespedeza virginica, Lippia lanceolata, Polygonum pennsylvanicum, Solidago canadensis, S. rigida, S. ulmifolia, Ver- bena hastata, Vernonia sp., Veronica sp. BEES OF THE GENUS MELISSODES 409 Melissodes (Eumelissodes) perlusa Cockerell Melissodes semiagilis var. perlusa Cockerell, 1925, Ann. Mag. Nat. Hist., ser. 9, vol. 16, p. 231; 1926, Univ. Colorado Studies, vol. 16, p. 114. The female of perlusa is similar to that of agilis but differs in that the lateral clypeal carina is closer to the eye, the metasomal terga are more finely punctate and the galeae are shinier. The female is also distinctive in having reddish wing veins, very pale vestiture and often red hairs on the inner surfaces of the hind basitarsi. The male of perlusa is similar to that of agilis except for the slightly longer first flagellar segment, the less coarsely punctate metasomal terga, the shiny galeae and the black mandibular bases. Female. Measurements and ratios: N, 20; length, 12-14 mm.; width, 4.5-6.0 mm.; wing length, M = 3.88 + 0.183 mm.; hooks in hamulus, M = 14.50 + 0.212; flagellar segment 1/segment 2, M = HESS == 101029: Structure and color: Integument color as in agilis except as fol- lows: eyes greenish blue; apical area of tergum 1 usually trans- lucent (even narrowly hyaline in some). Sculpturing and structure as in agilis except as follows: clypeal punctures usually slightly smaller, crowded, apicomedial longi- tudinal carina usually present, protruding beyond eye in profile by less than half width of eye, lateral carina separated from eye margin by half or slightly more of minimum diameter of first flagel- lar segment; galeae above shiny, unshagreened or extremely deli- cately so; maxillary palpal ratio about 4.5:3.5:3.0:1.0; vertex with lateral flattened areas with minute punctures separated by one to three puncture widths, surface shiny. Mesoscutum as in agilis but punctures slightly larger, surface not shagreened; scutellar punctures smaller than mesoscutal. Metasomal tergum 1 with basal two-thirds or less punctate, punctures round, small, distinct, sep- arated mostly by one puncture width, surface dulled by reticular shagreening (almost tessellate in appearance), apical area im- punctate, moderately shiny, finely shagreened; tergum 2 with basal zone with minute round punctures separated by one to two punc- ture widths, surface dulled by dense tessellation, interband zone impunctate or with sparse punctures not much larger than base of hairs arising from them, surface dulled by tessellation, apical area impunctate or virtually so, surface dulled by coarse reticulo- transverse shagreening; terga 3 and 4 similar to 2 but apical area of tergum 4 covered by distal pubescent band. Hair: On head pale ochraceous to dark ochraceous, often yellow- 15—5840 410 THE UNIVERSITY SCIENCE BULLETIN ish on vertex and occasionally vertex with sparse brown hairs. Thorax pale ochraceous or white laterally and posteriorly and dull ochraceous to dull rufescent (usually yellowish) above, without brown. Metasomal vestiture as in agilis except as follows: pale vestiture never rufescent, usually white to ochraceous; tergum 1 with apical area glabrous; tergum 2 with pale interband zone hairs never brown, erect to suberect, long, with distal pale band rarely interrupted medially, of about equal length across tergum and al- most as long as apical area medially, with apical area hairs always pale ochraceous or white, suberect, distinctly plumose and more abundant than in agilis; tergum 3 as in tergum 2 but apical area narrower, basal tomentum brown and distal pale pubescence in- vading interband zone; tergum 4 as in agilis but distal pale band broader and never with minute apicomedian brown area; sterna yellow to pale brown medially, paler laterally. Legs pale ochra- ceous to white except as follows: outer surface of fore basitarsi, outer surface of apical area of middle tibiae and on and surround- ing basitibial plates pale brown; inner surfaces of tarsi and hind tibiae yellow to dark reddish brown. Male. Measurements and ratios: N, 20; length, 10-13 mm.; width, 3-4 mm.; wing length, M = 3.49 + 0.197 mm.; hooks in hamulus, M = 12.95 + 0.228; flagellar segment 1/segment 2, M=—5.42 + 0.095. Structure and color: Integumental color as in agilis except as fol- lows: mandibular bases without yellow spots; labrum with medio- basal pale spot present or absent; eyes yellowish green; tergal mar- gins hyaline, colorless. Structure as in agilis except as follows: clypeus protruding be- yond eye in profile by less than half width of eye; first flagellar seg- ment with minimum length equal to one-fifth or slightly more of maximum length of second segment, usually about one and one-half times as long as pedicel on that side; maxillary palpal ratio about 7:4:4:1, last segment often shorter. Sculpturing as in female except as follows: clypeal punctures less distinct; supraclypeal area often shiny; galeae usually unshagreened, or with delicate reticular shagreening. Sternum 7 as in dentiventris but median plate with basoventral hairs more slender and sparser, median plate slightly larger than lateral plate. Sternum 8 as in agilis. Genital capsule as in agilis but hairs of gonostylus and gonocoxite as in dentiventris although sparser and more slender. BEES OF THE GENUS MELISSODES All Hair: Head and thorax white to pale ochraceous, often slightly darker on upper surface of thorax and on vertex. Metasomal terga as in agilis but pale vestiture always white to extremely pale ochra- ceous, vestiture longer, terga 2-4 with suberect hairs of apical areas long and distinctly plumose, without brown. Legs white to pale ochraceous except inner surfaces of tarsi yellow to pale reddish yellow. Bionomics. M. perlusa is probably an oligolege of the genus Helianthus. Too few collections with flower data are available to arrive at a definite conclusion. Of 38 collections (58 bees) avail- able with flower data, 23 collections (12 females and 28 males) were made on some species of Helianthus (the majority on H. petiolaris ). Of the remaining 15 collections (18 bees), 8 (7 females and 3 males ) were made on some other composite and 7 (4 females and 4 males ) were made on either legumes or labiates. Type Material. The holotype male of perlusa from Mesa Verde, Colorado, July 3-7, 1919, is in the collection of the American Museum of Natural History in New York City. Two paratype males with the same locality data are in the collection of P. H. Timberlake at the Citrus Experiment Station, Riverside, California. Distribution. M. perlusa ranges from Arizona and New Mexico north to Alberta and Manitoba in the western prairies and eastern Rocky Mountains (Fig. 10). It has been collected from June 20 to September 26, but mainly in July and August. In addition to the holotype, 48 females and 49 males from the localities listed below have been examined. ARIZONA: Flagstaff. Cotorapo: Alder; Berkley; Boulder; Clear Creek; Cortez; Golden; Larimer Co.; Mesa Verde; Peetz, San Luis Valley; Tobe. Towa: Lyon Co. Nesrasxa: Glen, Sioux Co.; Harrison; Mitchell; Monroe Canyon, Sioux Co. NEw Mexico: Em- budo; Jemez Springs; Las Cruces; Rowe; San Jose; Santa Fe; Vaugn. Nortu Daxora: Beach; Dickinson; Marmarth; Medora; Nicholson; Valley City; Washburn; Williston. Uran: Kanarrville. Wyominc: Torrington; Worland. Canada. Atperta: Lethbridge; Medicine Hat; Suffield; Whitla. Manrropa: Aweme. Flower Records. Biglovia sp., Brauneria pallida, Grindelia sp., Helianthus sp., H. annuus, H. petiolaris, Lepachys sp., Medicago sativa, Mentha canadensis, Petalostemum sp., P. oligophyllum, Rati- bida columnaris. 412, THE UNIVERSITY SCIENCE BULLETIN Melissodes (Eumelissodes) snowii Cresson Melissodes snowii Cresson, 1872, Proc. Acad. Nat. Sci. Philadelphia, vol. 24, p. 211; Robertson, 1898, Trans. Acad. Sci. St. Louis, vol. 8, p. 53; Bridwell, 1899, Trans. Kansas Acad. Sci., vol. 16, p. 211; Cockerell, 1906, Trans. Amer. Ent. Soc., vol. 32, p. 76; Tucker, 1909, Trans. Kansas Acad. Sci., vol. 22, p. 281; Cresson, 1916, Mem. Amer. Ent. Soc., vol. 1, p. 130. This species is extremely close to M. perlusa Cockerell. The fe- male of snowii is similar to that of perlusa in most respects; how- ever, snowii females have slightly more distinct punctures in the interband zone of tergum 2, often darker hairs on the inner surfaces of the hind basitarsi, more distinctly sculptured galeae and shorter and less plumose hairs in the apical areas of terga 2 and 3. The males of snowii are readily distinguished from those of perlusa by the yellow mandibular bases. They can be separated from the palest males of agilis only with difficulty, but the extremely short first flagellar segment and the white vestiture are distinctive. The males also resemble the males of M. nivea Robertson as discussed in the diagnosis of that species. Female. Measurements and ratios: N, 20; length, 12-13 mm.; width, 3.5-4.5 mm.; wing length, M=3.35+ 0.116; hooks in hamulus, M = 12.85 + 0.608; flagellar segment 1/segment 2, Mi 186 =0:022. Structure and color: Integumental color as in perlusa except eyes usually blue, gray or greenish blue. Sculpturing and structure as in perlusa except as follows: lateral clypeal carina separated from eye margin by more than half and usually by one-third to three-fourths minimum diameter of first flagellar segment; galea above dulled by fine, dense tessellation; maxillary palpal ratio about 3.2:3.2:3.2:1.0; mesoscutal punctures often slightly larger and sparser in posteromedian area; metasomal tergum 1 with basal three-fourths or less with punctures separated mostly by half to one puncture width, surface shiny with coarsely reticular shagreen- ing, interband zone with small shallow punctures separated mostly by two to three puncture widths, surface moderately shiny, apical area with minute punctures or impunctate, surface moderately shiny; tergum 3 as in tergum 2 but interband zone punctures more abundant and more distinct. Hair: Head white, a few brown hairs often present on vertex. Thorax white, often extremely pale ochraceous above. Metasomal vestiture as in perlusa but pale vestiture always white, never ochra- ceous, apical areas of terga 2 and 3 with subappressed hairs simple or plumose only at extreme base, short and a few (especially on BEES OF THE GENUS MELISSODES 413 tergum 3) often pale brown, and sterna brown except white later- ally. Legs white except as follows: fore tarsi, outer surfaces of middle tibiae (near apices) and often basitarsi, and on and sur- rounding basitibial plates brown; scopae white to extremely paie ochraceous; inner surfaces of basitarsi and hind tibiae yellow to brownish red. Male. Measurements and ratios: N, 20; length, 10-12 mm; width, 3.5-4.0 mm; wing length, M = 3.23 + 0.087; hooks in hamulus, M = 11.80 + 0.186; flagellar segment 2/segment 1, M = 10.24 + 0.188. Structure and color: Integumental color as in perlusa except as follows: mandibular bases with large yellow spots; labrum white except narrow brown apical margin (more than half of area pale); eyes bluish gray to greenish blue; tegulae usually piceous; wing veins yellow; flagellar segments 2 to 11 yellow below, red above. Structure as in perlusa except as follows: first flagellar segment with minimum length to about one-tenth of maximum length of second segment and subequal to pedicel on same side; maxillary palpal ratio about 3:3:3:1. Sculpturing as in female except as follows: clypeus with punctures less distinct; galeae with tessella- tion finer, often shiny and unshagreened or tessellate except in apical third; mesoscutum with posteromedian area punctures often sep- arated by 3 or more puncture widths; metasomal tergum 1 with basal four-fifths with punctures separated by half to one or slightly more puncture widths, Sternum 7 as in perlusa with median plate with apical margin transverse and ventral hairs minute, sparse and more delicate medio- basally. Sternum 8 as in agilis but apical hairs minute and ventral tubercle usually not bidentate apically, cariniform. Genital capsule as in agilis but inner apical spicules of gonocoxite at least half short and blunt and gonostylus subequal to half of gonocoxite (Figs. 82- 83). Hair: Vestiture entirely white, rarely slightly grayish or yellowish on dorsum of thorax. Tergal vestiture as in perlusa except as fol- lows: terga 2 and 3 with apical areas with suberect hairs long, white, simple except a few plumose at extreme bases; tergum 2 with distal pubescent band usually as long medially as apical area. Legs white except inner surfaces of basitarsi and often distitarsi yellow. Bionomics. This bee is probably oligolectic on Compositae. The females have been collected most often on Helianthus and 414 Tue UNIVERSITY SCIENCE BULLETIN Solidago, but there is not sufficient information for a more adequate statement of flower preferences. The only non-composite on which females have been taken is Medicago sativa. Type Material, (Lectotype male (No. 2330) and one male para- type collected by Snow in Colorado are in the collection of the Philadelphia Academy of Natural Sciences. Distribution. M. snowii ranges over the western parts of the Great Plains from Alberta in the north to New Mexico in the south (Fig. 12). It has been collected from July 10 to September 18, but mostly during August. In addition to the type material, 30 females and 26 males have been examined from the localities listed below. A majority of these 56 specimens are from Halsey, Ne- braska. Cotorapo: Boxelder Creek (E. of Aurora); Denver; Roggen. NesraskA: Dunning; Glen, Sioux Co.; Halsey; Holt Co.; Mitchell; Neligh; Thedford. New Mexico: Albuquerque; Moriarity; Vaughn. NorrH Daxora: Sheldon. Canada. A.sertra: Leth- bridge; Seven-persons. Maniropa: Aweme. Flower Records, Aster sp., Cleome serrulata, Gaillardia sp., Gutierrezia sarothrae, Helianthus sp., H. petiolaris, H. subrhom- boideus, Lacinaria punctata, Medicago sativa, Solidago missourien- sis, S. nemoralis, S. rigida. Melissodes (Eumelissodes) submenuacha Cockerell Melissodes menuacha var. submenuacha Cockerell, 1897, Entomologist, vol. 30, p. 137; 1877, New Mexico Col. Agric. and Mech. Arts, Bull. No. 24, p. 28; 1898, Bull. Sci. Lab. Denison Univ., vol. 11, p. 66; 1898, Bull. Univ. New Mexico, vol. 1, p. 66; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309. Melissodes hewetti Cockerell, 1905, Ann. Mag. Nat. Hist., ser. 7, vol. 15, p. 527 (new synonymy); 1906, Trans. Amer. Ent. Soc., vol. 32, pp. 84, 86; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309. This large pale species is very similar in appearance to M. perlusa Cockerell. The females of swhmenuacha can be separated from those of agilis by the coarser clypeal punctation and by the lateral clypeal carina being closer to the eye margin. They can be sep- arated from the females of perlusa by the slightly larger size, the dark wing veins, the coarser clypeal punctures, the denser punctures of the basal area of tergum 2, and the usually shagreened surface of the mesoscutum. The males of suwbmenuacha are similar to those of perlusa in the dark mandibular bases, the flatter clypeus, the less coarsely punctate terga and the shinier galeae, but differ from those of perlusa by the short first flagellar segment and the sculpturing of the mesoscutum and base of tergum 2 as described below. BEES OF THE GENUS MELISSODES 415 Female. Measurements and ratios: N, 10; length, 13-15 mm.; width, 4.5-6.0 mm.; wing length, M = 4.54 + 0.211 mm.; hooks in hamulus, M = 15.90 + 0.100; flagellar segment 1/segment 2, M = 2.13 + 0.082. Structure and color: Integument black except as follows: mandi- ble with apical half or more, usually apical margin of clypeus, flagel- lar segments 3-10, tarsi and often tibiae at least partly, occasionally apical areas of terga 2-4 rufescent; eyes grayish brown to grayish green; wing membranes colorless, veins dark reddish brown to black; tegulae testaceous, yellow to rufescent; tibial spurs yellow to red. Structure and sculpturing as in perlusa except as follows: clypeus separated from eye margin laterally by about three-fourths minimum diameter of first flagellar segment, with coarse, irregularly round punctures separated mostly by half a puncture width or less, usually with longitudinal median carina in apical half or more, surface shiny, with sparse delicate cross-striations; supraclypeal area sparsely or not punctate medially, shiny, usually with coarse reticular shagreen- ing; galeae usually with coarse reticular shagreening dulling apical half or more above; maxillary palpal segments in ratio of about 4.0:4.0:2.5:1.0; mesoscutum with posteromedian area with surface slightly dulled by coarsely reticular shagreening; metanotum and propodeum with reticulorugosity finer, tessellation coarser; metaso- mal tergum 1 with basal three-fifths with punctures separated by one to three puncture widths, surface with finely reticular shagreening; tergum 2 with basal area with minute punctures separated by half to one puncture width, apical area with surface somewhat dulled by fine reticulotransverse shagreening. Hair: Head pale ochraceous to ochraceous (brighter on vertex ), vertex often with sparse dark brown hairs. Thorax laterally and posteriorly white to pale ochraceous, ochraceous to yellow above, scutellum often with brown medially and mesoscutum rarely with a few brown hairs posteromedially. Metasomal vestiture as in per- lusa except as follows: pale vestiture not usually white but pale ochraceous; tergum 2 with distal pale band much thinner medially than laterally (interrupted when worn) and usually about one-half to three-fourths length of apical area medially; terga 2 and 3 with apical areas with subappressed to suberect hairs short, plumose only in basal half or less (seemingly simple except when highly magnified ), often pale brown partially at least on tergum 3; tergum 4 with distal pale band extremely broad; terga 5 and 6 with median 416 THE UNIVERSITY SCIENCE BULLETIN dark hairs golden brown, yellow laterally; sterna golden medially to pale ochraceous laterally. Legs as in perlusa except scopae pale yellow to pale ochraceous, Male. Measurements and ratios: N, 20; length, 12-14 mm.; width, 4-5 mm.; wing length, M = 4.14 + 0.170 mm.; hooks in hamulus, M = 14.15 + 0.287; flagellar segment 2/segment 1, M = 7.33 + 0.105. Structure and color: Integumental color as in agilis except as fol- lows: mandibular bases without yellow; labrum with or without mediobasal pale spot; first flagellar segment usually wholly dark brown; eyes yellow to yellowish green; tergal margins extremely broadly hyaline (apical third or more of tergum 1 and almost half of terga 2 and 3), colorless or yellow; wing veins reddish brown to dark brown. Structure as in perlusa except as follows: minimum length of first flagellar segment equals less than one-sixth (usually about one-seventh ) of maximum length of second segment, usually longer than pedicel on same side; maxillary palpal segments in ratio of about 8:7:5:1, last segment often twice as long. Sculpturing as in female except as follows: supraclypeal area often unshagreened; mesoscutellar shagreening often absent or extremely fine; tergum 1 with basal five-sixths to four-fifths punctate. Sternum 7 as in snowii but median plate with apical margin in- clined towards midline. Sternum 8 and genital capsule as in snowii. Hair: Head and thorax pale ochraceous to ochraceous, usually somewhat brighter on vertex and dorsum of thorax. Metasomal terga as in perlusa except as follows: vestiture usually pale ochra- ceous rather than white; terga 3-5 with interband zones with abun- dant, long, appressed pubescence (less abundant than in distal pubescent bands, however). Legs pale ochraceous to white except yellow on inner surfaces of tarsi. Type Material. Cockerell did not specify a single holotype for submenuacha in his original description and none has been found in collections in this country. However, four males from “Las Cruces, N. M., 9-5” have been examined. These are presumably part of the original type series. One male has “submenuacha Ck11.” on a second label written in Cockerell’s distinctive handwriting. A second of these has a note in Cockerell’s handwriting which records the characters which Cockerell published in the original description and a second note by Fox agreeing to the notes by BEES OF THE GENUS MELISSODES 417 Cockerell. These males are evidently those cited by Cockerell as being collected by C. H. Townsend. That male labeled subme- nuacha Ck11. is hereby designated as the lectotype of submenuacha. The lectotype is in the collection of the Natural History Museum of the University of Colorado at Boulder. It is interesting that although no males have been seen from Las Cruces collected by Cockerell on September 22nd, as he states in the original description, a single female of swbmenuacha from Las Cruces collected on Helianthus annuus (which is mentioned in the description ) on September 22nd by Cockerell has been examined. Another label on this specimen reads “submenuacha n. sp.” Could Cockerell have recorded this female as a male in writing the descrip- tion for publication? The holotype female of hewetti from Santa Fe, New Mexico, collected by T. D. A. Cockerell on Cleome serrulata in August is in the collection of P. H. Timberlake of the Citrus Ex- periment Station, Riverside, California. Distribution. M. submenuacha is known from Arizona, New Mexico and western Texas (Fig. 13). It has been collected from May 11 to November 10, but chiefly in September. In addition to the holotype, 10 females and 22 males from the localities listed below have been examined. ARIZONA: Cameron (19 miles W.); Cochise Co.; Douglas; Ma- dera Canyon, Santa Rita Mts.; Nicks (Huachuca Mts.); Portal (3, 5 and 10 miles E.); Sabino Canyon, Santa Catalina Mts.; Sedona (and 15 miles S.); Theba; Tucson. New Mexico: Albuquerque; Em- budo; Hurley; Las Cruces; Santa Fe; Wilna. Texas: Big Bend State Park (Hot Springs); El Paso; Hueco Mts. (W. side of), El Paso Co. Flower Records. Aploppapus gracilis, Bidens sp., Cleome serru- lata, Helianthus sp., Hymenothrix wislizeni, Isocoma heterophylla, Medicago sativa, Verbesina encelioides. Although submenuacha is probably an oligolege of the Compositae, there is not sufficient evi- dence to make a more precise statement of its flower preferences at this time. Melissodes (Eumelissodes) menuachus Cresson Melissodes menuachus Cresson, 1868, Trans. Amer. Ent. Soc., vol. 1, p. 388; 1875, in Wheeler, Report Geog. Geol. Surv. west of 100th Meridian, vol. 5, p. 727; 1876, Proc. Davenport Acad. Nat. Sci., vol. 1, p. 209; Cragin, 1886, Bull. Washburn Coll. Lab. Nat. Hist., vol. 1, p. 211; Cockerell, 1893, Trans. Amer. Ent. Soc., vol. 20, p. 338; Fox, 1893, Proc. California Acad. Sci., ser. 2, vol. 4, p. 118; Townsend, 1896, Canadian Ent., vol. 28, p. 139; Cockerell, 1897, Entomologist, vol. 30, p. 138; 1897, Bull. Agric. Exp. Sta. New Mexico Coll. Agric. and Mech. Arts, no. 24, p. 19; 1898, Zoologist, p. 313; 1898, THe UNIversiry SCIENCE BULLETIN 418 \ \ \\ \ \\ \\ \\ semilupina 4, tS e issodes) (Eumel M. Map showing the known distributions of semilupina Cockerell and M. (E.) submenuacha Cockerell. 13. Fic. BEES OF THE GENUS MELISSODES 419 Bull. Univ. New Mexico, vol. 1, pp. 66, 67, 73; 1898, Bull. Sci. Lab. Denison Univ., vol. 11, pp. 66, 67, 73; 1899, Catalogo de las Abejas de Mexico, p. 14; Birkman, 1899, Trans. Kansas Acad. Sci., vol. 16, p. 211; Fowler, 1902, Univ. California Agric. Exp. Sta., p. 322; Cockerell, 1903, Psyche, vol. 10, p. 77; 1903, Ann. Mag. Nat. Hist., ser. 7, vol. 12, p. 449; Viereck, 1905, Canadian Ent., vol. 37, p. 320; Cockerell, 1906, Trans. Amer. Ent. Soc., vol. 32, pp. 77, 86, 92; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309; 1906, Bull. Amer. Mus. Nat. Hist., vol. 22, p. 443; Snow, 1906, Trans. Kansas Acad. Sci., vol. 20, p. 137; Tucker, 1909, Trans. Kansas Acad. Sci., vol. 22, p. 282; Smith, 1910, Ann. Rept. New Jersey State Mus., 1909, p. 693; Cockerell, 1910, Psyche, vol. 17, p. 246; 1910, Ent. News, vol. 10, p. 4; 1912, Proc. U.S. Nat. Mus., vol. 43, p. 271; Cresson, 1916, Mem. Amer. Ent. Soc., vol. 1, p. 123; Bray, 1917, Pomona Jour. Ent. Zool., vol. 9, p. 94. Melissodes mennacus (!) Uhler, 1877, Bull. U. S. Geol., Geog. Surv., vol. 3, p. 783. Melissodes pallida Robertson, 1895, Trans. Amer. Ent. Soc., vol. 22, p. 127 (mew synonymy); 1905, Trans. Amer. Ent. Soc., vol. 31, p. 369; 1928, Flowers and Insects, p. 8. Melissodes mizeae Cockerell, 1905, Ann. Mag. Nat. Hist., ser. 7, vol. 15, p. 522 (new synonymy ); 1906, Trans. Amer. Ent. Soc., vol. 32, p. 86; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309; 1907, Univ. Colorado Studies, vol. 4, p. 255; Hicks, 1926, Univ. Colorado Studies, vol. 15, p. 225; Cockerell, 1933, Ann. Ent. Soc. Amer., vol. 26, p. 44. Melissodes blakei Cockerell, 1905, Ann. Mag. Nat. Hist., ser. 7, vol. 15, p. 523 (new synonymy); 1906, Trans. Amer. Ent. Soc., vol. 32, p. 107; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309; 1927, Ann. Ent. Soc. Amer., vol. 20, p. 396. Melissodes lavata Cockerell, 1924, Pan-Pacific Ent., vol. 1, p. 56 (new synonymy); 1927, Ann. Ent. Soc. Amer., vol. 20, p. 395; 1928, Univ. Colorado Studies, vol. 16, p. 114. Melissodes octobris Cockerell, 1934, Ent. News, vol. 45, p. 30 (new synonymy ). This species is a large pale bee closely related to M. submenuacha Cockerell. The female of menuachus can be distinguished from any of the preceding species by the second flagellar segment being longer than broad and by the color of the vestiture as described below. The male of menuachus is similar in size to that of sub- menuacha, but has pale spots at the base of the mandible, as well as a pale labrum, and has the first flagellar segment longer in re- lation to the second segment. Female. Measurements and ratios: N, 20; length, 13-16 mm.; width, 4.0-5.5 mm.; wing length, M = 4.56 + 0.180 mm.; hooks in hamulus, M = 16.15 + 0.254; flagellar segment 1/segment 2, M = 1.83 + 0.016. Structure and color: Integumental color as in submenuacha ex- cept as follows: eyes grayish blue to greenish gray; wing mem- branes slightly milky, veins red to reddish brown; tegulae usually testaceous. Structure as in agilis except as follows: clypeus flat, protruding beyond eye in profile by one-third or less of eye width, with small, crowded, round punctures, surface dulled by coarse reticular sha- greening, with distinct median longitudinal carina in apical half; 420 THE UNIVERSITY SCIENCE BULLETIN supraclypeal area shiny, often with coarse reticular shagreening but scarcely dulling surface; galeae shiny, with delicate shagreening above in apical half or less; maxillary palpal ratio about 5.0:4.5: 3.0:1.0, vertex with lateral flattened areas with minute sparse punc- tures, shiny; second flagellar segment ventrally slightly longer than broad. Mesoscutum with large deep punctures of rather variable diameter separated mostly by one-half to one puncture width, sur- face shiny; scutellum similar but punctures more crowded; mesepi- sternum with large shallow punctures separated mostly by much less than half a puncture width, surface shiny with extremely sparse and delicate shagreening. Metasomal tergum 1 with impunctate apical area expanded basally at sides to form two small, impunctate lateral lobes. Hair: Head white except usually ochraceous on vertex. Thorax white to pale ochraceous laterally and posteriorly, bright to pale ochraceous above. Metasomal tergum 1 with basal area with long pale ochraceous hairs, apical area glabrous, pale basal hairs not reaching apical margin of tergum except at extreme sides; tergum 2 with basal white pubescence connected with distal pale band at sides, interband area with sparse, subappressed, relatively simple, pale hairs, distal pale band twice as long as apical apubescent area laterally, notched on posterior border at midline where usually half as long as apical area, apical area glabrous; tergum 3 similar to 2 but basal tomentum dark brown, distal pale band separated from apex by a narrow zone of suberect, pale relatively simple hairs, tergum 4 similar to 3 but apical suberect hairs absent; terga 5 and 6 brown except thick lateral white tufts; sterna brown medially to white at extreme sides. Legs white to pale ochraceous except as follows: fore tarsi, outer apex of fore and middle tibiae, on and sur- rounding basitibial plate, and inner surfaces of hind basitarsi brown to dark reddish brown; scopal hairs extremely long, usually pale yellow, occasionally pale ochraceous. Male. Measurements and ratios: N, 20; length, 12-15 mm.; width, 3.5-5.0 mm.; wing length, M = 4.22 + 0.181 mm.; hooks in hamulus, M = 14.25 + 0.239; flagellar segment 2/segment l, M = 5.00 + 0.083. Structure and color: Integument black except as follows: clyp- eus and base of mandible yellow; labrum white except narrow apical margin brown; flagellum yellow to red below and brown to dark red above except first segment often entirely dark; eyes gray- ish yellow to greenish gray; distitarsi rufescent; wing membranes BEES OF THE GENUS MELISSODES 42] hyaline, veins yellow to pale red; tegulae testaceous; apical areas of metasomal terga hyaline, colorless to yellow. Structure as in agilis except as follows: minimum length first flagellar segment equals about two-thirds maximum length of first segment and about one-fifth (or slightly less) maximum length sec- ond segment; maxillary palpal ratio about 2.5:2.3:2.3:1.0; clypeus flat. Sculpturing as in female except as follows: galeae above shiny, with apical half often delicately shagreened; tergum 1 with basal four-fifths punctate, punctures usually somewhat larger, deeper and more crowded than in female; terga 3 and 4 with inter- band zone with small round punctures separated mostly by one to three puncture widths, occasionally mostly by one puncture width or less, surface dulled by coarse, reticular shagreening; tergum 5 similar but punctures more crowded. Sterna 7 and 8 as in submenuacha. Genital capsule as in sub- menuacha but gonostylus with hairs on ventral surface near base short, stout, blunt, sparse, on outer lower surface mostly short, stout and bifid or trifid at apex; gonocoxite with several apical short, stout, blunt spicules on ventral surface below gonostylus in addition to those on inner apical surface (the latter are as in submenuacha). Hairs: Head and thorax white to pale ochraceous, often brighter on vertex and dorsum of thorax. Metasomal tergum | with long white to ochraceous hairs basally, apically with long, appressed to subappressed pale hairs reaching margin of tergum (in fresh speci- mens ) not obscuring apical area of tergum at least medially; tergum 2 with white pubescence basally, suberect, bristlelike, pale hairs in interband zone, white to pale ochraceous distal pubescent band not interrupted medially (unless worn) and separated from apical mar- gin by one-half to one times length of pale band medially; terga 3-5 similar except interband zones with sparse, delicate, white, ap- pressed pubescence in addition to bristlelike hairs and distal bands progressively closer to apical margin; terga 6 and 7 white to yellow- ish; sterna pale ochraceous to reddish medially, white laterally. Legs white to ochraceous except inner surfaces of tarsi golden yellow to pale rufescent. Bionomics. Hicks (1926, p. 225) has recorded a few notes con- cerning the biology of M. menuachus. He discovered two females of menuachus nesting in the ground near Boulder, Colorado. On account of the condition of the soil, Hicks was unable to excavate these nests, although he did observe the females carrying pollen into their burrows. In both instances he observed a female of 429, THE UNIVERSITY SCIENCE BULLETIN Triepeolus occidentalis Cresson enter the burrow while the female Melisssodes was absent. Apparently T. occidentalis is a parasite of M. menuachus in that region. M. menuachus is dependent upon flowers of the family Com- positae and in particular upon the genera Grindelia and Solidago, according to the collection data available at this time. This is clearly shown in Table VI. Type Material, The lectotype male of menuachus Cresson from New Mexico is in the collection of the Philadelphia Academy of TaBLE VI. Summary of Floral Records for Melissodes menuachus. Plant Data Records of M. menuachus CH n o an aS safest |) ae = 2 fe) Gu 0.2 Om fo) & Ae ice Se ul 2 = — @ - 28 > . S /388\) 5s | 52] bm FAMILY & © A = 5 0 & 3 if Ae a8 |25 8] so | g8 lye Sages Bo] o's ww = 5 = 5 S| 6° Z < Z Z, Z = Compositae: Grindelia spp. 1 2 31 39 30 69 : - | : = : Compositae: Solidago spp. 1 Silla 17 6 23 Other Compositae 9 13 18 15 20 35 -- | Leguminosae 3 3 if 3 4 7 Brassicaceae 1 1 6 0) 9 9 = | Others (5) 5 5 8 2 8 10 Totals 20 27 84 76 77 153 Sciences. The holotype female of pallida Robertson, collected by Robertson (Coll. No. 9619) September 26, 1890, on Helianthus grosse-serratus, is in the collection of the Illinois Natural History Survey, Urbana. The holotype female of mizeae Cockerell, col- lected by Mize at Las Vegas, N. Mex., in August on Grindelia inor- nata, is in the collection of P. H. Timberlake, Citrus Experiment Station, Riverside, California. The holotype female of blakei Cockerell, collected at Beulah, N. Mex. in August, is in the collection of the U. S. National Museum (Type No. 40094). The female holotype of lavata Cockerell, collected at Wray, Colo., August 17- 19, 1919, by F. Lutz, is in the collection of the American Museum of Natural History, New York City. The holotype female of BEES OF THE GENUS MELISSODES 423 octobris Cockerell, collected at Hudson, Colo., October 1, 1933, is in the collection of P. H. Timberlake, Citrus Experiment Station, Riverside, California. Distribution. This species is widely distributed from British Columbia, Alberta and North Dakota, east to Illinois and south to north-central Mexico and Texas (Fig. 14). It has been reported from Camden Co., New Jersey, by Smith (1910), but this is prob- ably in error. It is most abundant in the prairie regions of Colo- rado, Nebraska and Kansas. M. menuachus has been taken from July 2 to October 6, but mainly-in August and September. In addi- tion to the types, a total of 245 females and 209 males from the localities listed below have been examined. This list includes localities reported in the literature. ArIzoNA: Bisbee (12 miles W.); Chiricahua Mts.; Flagstaff (and 7 miles S. and 4 miles N.); > Mt. Graham; Grand Canyon; menuachus 120° 100° 95° 90 Fic. 14. Map showing the known distributions of M. (Eumelissodes) menuachus Cresson and M. (E.) expolita LaBerge. 494 THE UNIVERSITY SCIENCE BULLETIN Huachuca Mts.; Lochiel (4 miles E.); Payson; Sedona (10 miles N.); Southwest Research Station (5 miles W. of Portal); Williams. CaLiFoRNIA: Amedee, Lassen Co.; Benton’s Crossing, Mono Co.; Bridgeport, Mono Co.; Grant Lake, Mono Co.; Mono Lake; Owen’s Valley; Riverside. Cotorapo: Boulder; Brighton; Buffalo Creek; Canon City; Chimney Gulch (near Golden); Colorado Springs; Custer Co.; Denver; Elbert; Eldora; Estes Park; Florissant; Fort Collins; Fremont Co.; Garden of the Gods (near Colorado Springs ); Hoehne; Hudson; Jim Creek, Boulder Co.; La Junta; Lamar; Lari- mer Co.; Limon; Pingree Park; Platte Canyon (near Waterton); Poudre Canyon (W. of Fort Collins); Red Wash; Rock Creek (near Colorado Springs); Rocky Ford; Trinidad; West Cliff, Custer Co.,; Wray. IpanHo: Downey; Franklin; Lewiston; Parma. IL.inots: Carlinville. Towa: Ames; Chickasaw Co.; Sioux City. Kansas: Baldwin; Cheyenne Co.; Clark Co.; Douglas Co.; Garden City; Garnett; Grant Co.; Hill City; Johnson (2 miles N.); Lane Co.; Lawrence; Riley Co.; Scott City (5 miles N.); Topeka; Wallace Co.; Wichita Co. Mrynesora: Lyon Co.; Moorhead; Ortonville; Powder Plant Woods, Ramsey Co.; St. Paul. Monrana: “Mon.” Nepraska: Ashland; Box Butte Co.; Cambridge; Cedar Bluffs; Glen, Sioux Co.; Gordon; Haigler; Hardy; Harrison; Kimball; Lin- coln; Lodgepole; Malcolm; McCool; Monroe Canyon, Sioux Co.; Nebraska City; Neligh; North Platte (8 miles W.); Omaha; Sand Hills, Cherry Co.; Sioux Co.; Steele City, Jefferson Co.; South Bend; West Point. Nrvapa: Pyramid Lake. New Mexico: Beu- lah; Embudo; Glorieta; Hurley (5 miles S.); Las Vegas; Maxwell City; Mescalero; Raton; Rincon; Rito de los Frijoles; Roziata; Tularosa Creek; Santa Fé; Sapello; Sapello Canyon. Norru Da- KoTaA: Beach; Dickinson; Fargo; Glenn Ullin; Grafton; Grand Forks; Jamestown; Mandan; Martin; Medora; Minot; Mott; Rugby; Scha- fer; Sentinel Butte; Steele; Valley City; Washburn; Williston. Onrer- con: Echo; Freewater; Hereford. SourH Daxotra: Deadwood; Hot Springs. Texas: Clarendon; Fedor, Lee Co. Uran: Bear River City; East Promontory; Grantsville; Kaysville; Lake Point; Logan; Magna; Ogden; Promontory. Wasuincron: Coulee City. Wzs- CONSIN: Prescott. Wyominc: Albany Co.; Cheyenne; Clifton, Weston Co.; Douglas; Laramie; Torrington; Weston Co. Canada. ALBERTA: Lethbridge; Medicine Hat; Scandia. British COLUMBIA: Penticton; Similkameen. México. CHrauanua: Aguascalientes (Sta. Barbara Dist.); Salaices. Zacatecas: Sain Alto, Flower Records. Argemone sp., A. intermedia, A. platyceras, Aster sp., A. laevis, A. multiflora, Cassia chamaecrista, Chrysopsis BEES OF THE GENUS MELISSODES 495 sp., Chrysothamnus sp., C. graveolus glabrata, Cleome sp., C. serru- lata, Eustoma russellianum, Gaillardia sp., Grindelia sp., G. inornata, G. squarrosa, Gutierrezia sp., G. californicum, G. sarothrae, Helian- thus sp., H. annuus, H. petiolaris, Hymenothrix wislizenia, Medicago sativa, Melilotus sp., M. alba, Petalostemum oligocephalum, Poly- gonum sp., Rudbeckia laciniata, Sidalcea neomexicana, Solidago sp.., S. canadensis, S. rigida, Verbena sp., Viguiera sp., Xanthocephalum gymnospermoides. Melissodes (Eumelissodes) semilupina Cockerell Melissodes menuacha semilupina Cockerell, 1905, Bull. S. California Acad. Sci., vol. 4, p. 29. Melissodes chrysothamni Cockerell, 1905, Ann. Mag. Nat. Hist., ser. 7, vol. 15, p. 524, (new synonymy ); 1906, Trans. Amer. Ent. Soc., vol. 32, p. 85; 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309. Melissodes mizeae, Cockerell, 1912, Ann. Mag. Nat. Hist., ser. 8, vol. 10, p. 448 ( misidentification ). This species is a large bee closely related to M. menuachus Cres- son. The female of semilupina agrees with menuachus in having a relatively long second flagellar segment. It differs, however, in lacking the pale lateral tufts of hairs on terga 6 and 7 and in hay- ing brown hairs on the anterior and lower-lateral parts of the mesepisterna and coxae. The male of semilupina resembles that of menuachus closely but has a distinct band of long, pale, appressed pubescence covering the apical margin of the first tergum. Female. Measurements and ratios: N, 20; length, 13-16 mm.; width, 4.5-6.0 mm.; wing length, M = 4.70 + 0.247 mm.; hooks in hamulus, M = 18.10 + 0.240; flagellar segment 1/segment 2, (19) Nees 01024. Structure and color: Integumental color as in menuachus except as follows: eyes grayish blue; second (often third) flagellar segment totally black, segments 3 to 10 usually dark red below; tegulae piceous; wing veins dark brown; apical areas of terga usually black, occasionally dark reddish brown. Structure and sculpturing as in menuachus except as follows: length of second flagellar segment occasionally equal to width, usually slightly longer; maxillary palpal ratio about 4.5:4.0:3.5:1.0; clypeus relatively flat, lateral angle separated from eye margin by less than half minimum diameter of first flagellar segment, with median boss near apex but not usually carinate; supraclypeal area shiny, unshagreened, impunctate or punctures sparse and minute; lateral flattened areas of vertex with small punctures separated by one or less puncture widths, shiny. Mesoscutum with large, deep. round, variable-sized punctures, in posteromedian area separated 426 THE UNIVERSITY SCIENCE BULLETIN mostly by one puncture width or slightly more, surface often slightly dulled by reticular shagreening; mesepisterna moderately shiny, surface somewhat dulled by coarse, reticular shagreening. Metaso- mal tergum 1 with punctures of basal three-fifths small, shallow, separated mostly by one to one and one-half puncture widths, sur- face dulled by dense, reticulotransverse shagreening; terga 2 and 3 with interband zone punctures scarcely broader than base of hairs arising from them, separated mostly by two to three puncture widths or more, surface dulled by dense reticular shagreening, apical areas moderately shiny, impunctate. Hair: Color as in menuachus except as follows: labrum and man- dible often brown; anterior and lower lateral surfaces of mesepi- sterna dark brown; tergum 2 with interband zone hairs appressed to suberect, long, distal and basal pale bands confluent laterally and medially; tergum 3 with apical area usually as long as half width of distal pale band, with suberect, relatively simple, white or brown hairs; terga 5 and 6 without lateral pale tufts; sterna dark brown; legs as in menuachus but coxae dark brown, usually femora and trochanters dark brown at least below and occasionally entirely so; fore and middle distitarsi and basitarsi dark brown; hind basitarsi with inner surfaces and scopal hairs at apices of outer surfaces dark brown. Male. Measurements and ratios: N, 20; length, 12-15 mm.; width, 3.5-5.5 mm.; wing length, M = 4.36 + 0.212 mm.; hooks in hamulus, M=15.10+ 0.176; flagellar segment 2/segment 1, M = 4.24 + 0.064. Structure and color: Integumental color as in menuachus except as follows: first flagellar segment usually dark brown or black, rarely red or yellow below; eyes brownish yellow to green; tegulae testaceous; wing veins red to reddish brown. Structure as in menuachus except as follows: minimum length first flagellar segment equals one-fifth or slightly more of maximum length second segment, flagellum slightly crenulate near apex in lat- eral view (segments 9 and 10, and often 7 and 8, somewhat con- stricted below near base); maxillary palpal ratio about 3.0:2.7:2.3: 1.0. Sculpturing as in female except as follows: tergum 1 with basal five-sixths with small round punctures mostly separated by one to two puncture widths, surface dulled by dense reticular shagreen- ing, apical area beneath apical pale pubescent band with minute punctures and dulled at least basally; terga 2-4 with interband zone punctures minute, separated mostly by two to three puncture BEES OF THE GENUS MELISSODES 427 widths, surface dulled by dense reticulotransverse shagreening, apical areas moderately shiny, impunctate; terga 5 and 6 similar but punctures more crowded. Sterna 7 and 8 and genital capsule as in menuachus but sternum 7 with median plate with ventral hairs stouter (especially near base of plate). Hair: Vestiture essentially as in menuachus but dull white rather than ochraceous and never bright ochraceous on dorsum of thorax or vertex of head. Bionomics. This species is an oligolege of the composite genus Chrysothamnus. Females have been collected only from flowers of that genus (as far as the collection labels indicate), and males have been collected more often on Chrysothamnus flowers than on all other plants together. Isolated males, however, have been taken on Solidago, Isocoma and Cleome flowers and several were collected on sugar beets (whether on the inflorescence or on the harvested beets is not clear, but I suspect the latter ). Type Material. The male holotype of semilupina collected by Dr. A. Davidson at Los Angeles, California, is in the collection of the Natural History Museum of the University of Colorado at Boulder. The female holotype of chrysothamni collected at Embudo, New Mexico, September 16, 1897, by T. D. A. Cockerell on Bigelovia ( = Chrysothamnus) sp. is in the collection of the U. S. National Museum (Type No. 40093). Distribution. M. semilupina ranges from British Columbia and the Pacific Coast States east to Colorado and New Mexico (Fig. 13). It is most abundant in Oregon and California. This species has been collected from August 5 to October 29, but chiefly in Septem- ber. In addition to the holotype, 22 females and 89 males have been examined from the localities listed below (including records reported in the literature). Arizona: Tucson. Cairornia: Adelanto (8 miles S.), Mohave Desert; Anaheim; Deep Creek; Fort Tejon; La Jolla; Los Angeles; Los Angeles Co.; Morro Bay; Oro Grande; Riverside; San Diego Co.; Standish, Lassen Co.; Truckee (11 miles E. at Boca Dam); Victorville. Cotorapo: Alamosa; Great Sand Dunes National Monument. Nevapa: Sparks, Washoe Co.; Sutcliffe, Washoe Co. New Mexico: Embudo; Abbotts Ranch, Rito de las Frijoles. Ore- con: Arlington; Bend; Echo; Sisters (and 5 miles E.); Tumalo. Urau: Emery Co.; Iosepa; Juab Co.; Lehi; Park Valley. Wasnx- INGTON: Hunts Junction; “Wash. Terr.” Wyominc: Granger 428 Tue UNIVERSITY SCIENCE BULLETIN Canada. British CotumBriaA: Oliver; Vernon; Walhackin. ALBERTA: Magrath. Flower Records. Chrysothamnus sp., C. nauseosus, Cleome sp.., Isocoma sp., Solidago occidentalis. Melissodes (Eumelissodes) ochraea, n. sp. This species is closely related to both menuachus and to semi- lupina. The female is similar to that of menuachus in having tufts of pale hairs laterally on terga 5 and 6 and by having pale hairs on the lower lateral and anterior parts of the mesepisterna. However, the female of ochraea has the second flagellar segment distinctly shorter than broad, has a slightly more densely punctate mesoscutum than either menuachus or semilupina, and lacks suberect hairs in the apical area of tergum 3. The male of ochraea resembles that of semilupina in the form of the first flagellar segment and in having the dense, apical, pale pubescent band on the first tergum. It can be distinguished from semilupina by the more densely punctate mesoscutum. The female of ochraea also resembles the female of swhmenuacha very closely. However, the males of these two species are quite dis- tinct and the resemblance of the females does not, perhaps, indicate close relationship. The female of ochraea can be distinguished from that of submenuacha by the almost total lack of suberect hairs in the apical area of terga 2 and 3, by the dark brown hairs of terga 5 and 6 and of the sterna. The male of ochraea differs from that of sub- menuacha in the much longer first flagellar segment. Female. Measurements and ratios: N, 20; length, 11-15 mm.; width, 3.5-5.0 mm.; wing length, M = 4.33 + 0.171 mm.; hooks in hamulus, M = 16.60 + 0.222; flagellar segment 1/segment 2, M = 2.12 + 0.035. Structure and color: Integument black except as follows: apical half of mandible, often apical margin of clypeus, distitarsi, often basitarsi and occasionally rest of leg, and sterna rufescent; eyes gray; wing membranes colorless or slightly milky, veins reddish brown to black; tegulae piceous; tibial spurs yellow. Clypeus flat, margin separated from eye by less than half mini- mum diameter of first flagellar segment, with median carina in apical half; maxillary palpal ratio about 4.5:2.5:3.0:1.0, (in one paratype from Whitewater, California, a distinct fifth segment equal in length to fourth is present); clypeus with round, shallow, coarse punctures separated mostly by less than half a puncture width, surface (and BEES OF THE GENUS MELISSODES 429 bases of punctures ) somewhat dulled by coarse, reticular shagreen- ing; galeae above shiny, slightly dulled in apical half or less by re- ticular shagreening; vertex with lateral flattened areas with small punctures of irregular size separated mostly by half to two puncture widths, surface shiny. Thoracic sculpturing as in menuachus but mesoscutum with posteromedian area punctures separated mostly by less than one puncture width and often dulled by delicate re- ticular shagreening, scutellum shiny and punctures crowded, mesepi- sterna shiny, shagreening, if present, extremely delicate. Tergal punctation as in semilupina; terga 1-3 with apical areas impunctate, shiny, reticulotransverse shagreening extremely fine. Pygidial plate V-shaped, with apex more pointed than in swbmenuacha, menuachus or semilupina. Hair: Head and thorax ochraceous, paler on clypeus, frons, genal areas, sides of thorax, and propodeum, often much brighter on vertex and dorsum of thorax (even being orangish here in some specimens but dark ochraceous in holotype). Metasomal pale pubescence and hairs pale ochraceous; tergal vestiture as in menuachus except as fol- lows: tergum 3 with distal pale band separated from apical margin across entire tergum, apical area glabrous, without suberect hairs except occasionally a few at extreme base of area near pale band, medially apical area at least as long as pale band. Leg hairs as in menuachus except inner surfaces of hind basitarsi often reddish brown (as in holotype) and occasionally red. Male. Measurements and ratios: N, 14; length, 12-14 mm.; width, 3-4 mm.; wing length, M = 4.06 + 0.216 mm.; hooks in hamulus, M = 13.86 + 0.231; flagellar segment 2/segment 1, (13) M = 4.31 + 0.134. Structure and color: Integument black except as follows: clypeus and base of mandible yellow; labrum cream-colored except brown apical margin; flagellum yellow to red below (except first segment ), reddish brown to black above; eyes gray to green; distitarsi and sterna rufescent; wing membranes colorless, veins red to reddish brown; tegulae testaceous (allotype) to piceous; apical areas of terga hyaline, colorless to slightly yellow. Structure as in menuachus except as follows: minimum length of first flagellar segment equals one-fifth or more (more in allotype) of maximum length second segment; maxillary palpal ratio about 4.0:3.5:3.5:1.0, minute fifth segment sometimes present; clypeus flat. Sculpturing as in female except as follows: metasomal tergum | with basal five-sixths or more punctate, apical area beneath apica! 430 Tue UNIVERSITY SCIENCE BULLETIN pale band with minute punctures; terga 2, 3 and 4 with more distinct interband zone punctures separated mostly by about two puncture widths. Sterna 7 and 8 and genital capsule as in menuachus but hairs on gonostylus shorter, sparser, never bifid or trifid at apex and hairs of ventral surface of median plate of sternum 7 sparse and weak. Hair: Head and thorax pale ochraceous to white, often somewhat darker on vertex and dorsum of thorax (pale in allotype). Meta- somal vestiture as in menuachus except as follows: tergum 1 with apical pale band of long, appressed, plumose hairs hiding surface across entire tergum (unless worn); terga 2-4 with apical areas with- out suberect hairs except one or two rows limited to base of apical area near pale pubescent bands. Legs white to pale ochraceous ex- cept inner surfaces of tarsi yellow. Type Material. The holotype female and allotype male from Whitewater, Riverside Co., California, October 27, 1934, were col- lected by C. D. Michener on Isocoma acradenia. Nineteen female and nine male paratypes from California are as follows: White- water: 1 female with the holotype; 5 females and 1 male on I. acra- denia, October 27, 1934, P. H. Timberlake; 3 females and 2 males, October 27, 1934, A. L. Melander; 2 females on I. acradenia, No- vember 12, 1932, P. H. Timberlake; 1 female, September 8, 1949. Indio: 3 males, October 13, 1935, E. G. Linsley; 1 male on I. acra- denia, October 15, 1947, P. H. Timberlake; 1 male on I. acradenia (2.8 miles S. E.), October 23, 1951, P. H. Timberlake. Morongo Valley: 1 male on Gutierrezia sp., September 26, 1944, P. H. Timber- lake. Vallecito, San Diego Co.: 7 females, September 24, 1936, C. M. Dammus. The holotype and allotype are in the Snow Entomo- logical Collection of the University of Kansas, Lawrence. Paratypes are in the collections of the University of Kansas, P. H. Timberlake of the Citrus Experiment Station, Riverside, California, the Univer- sity of California at Berkeley, Harvard University (Museum of Com- parative Zoology ), and in the author’s collection. Distribution. Southern California and Arizona (Fig. 10). Since only 13 specimens are known in addition to the type material, the data for these are listed below in full. ARIZONA: Madera Canyon, Santa Rita Mts.: 1 male, October 4, 1956, G. D. Butler and F. G. Werner. Rosemont, Pima Co.: 1 fe- male on Baccharis sp., October 9, 1954, F. G. Werner. Sabino Basin, Santa Catalina Mts.: 1 male, September 28, C. H. T. Townsend. Safford: 2 females on yellow composite, September 24, 1956, G. D. BEES OF THE GENUS MELISSODES 431 Butler; 2 females on yellow composite (30 miles S.), September 24, 1956. Tucson: 1 female, October 20, 1919; 1 male, October 1927, J. A. Downes; 1 female, October 8, 1937, R. H. Crandall; 1 female, September 29, 1939, A. S. Rosenberg; 1 female, November 10, 1939, A. S. Rosenberg. Melissodes (Eumelissodes) bimatris, n. sp. This species is highly variable in the color of the vestiture and is closely related to semilupina and ochraea. Females, because of their color variation, are difficult to separate from the latter two species, especially if they have much of the pubescence worn away. The female is like semilupina in having suberect hairs in the apical areas of terga 2 and 3, but the second flagellar segment is usually distinctly broader than long ventrally. The female also differs from semi- lupina females in the sculpturing of the interband zones of terga 2 and 3, as described below, and in the erect hairs in these zones. The female can be readily distinguished from that of ochraea by the punctation of the mesoscutum and the first metasomal tergum, and by the erect hairs of the interband zones of terga 2 and 3. The males are readily distinguished by lacking yellow maculae at the mandibu- lar base, by the labrum being all or mostly black, and by the sparse mesoscutal punctures. The darkest females resemble M. (Calli- melissodes) nigracauda from which they differ by the shiny galeae and pale scopal hairs. Female. Measurements and ratios: N, 20; length, 11-15 mm.; width, 4.0-5.5 mm.; wing length, M = 4.10 + 0.150 mm.; hooks in hamulus, M = 14.75 + 0.298; flagellar segment 1/segment 2, M = 1.96 + 0.022. Structure and color: Integument black except as follows: apical half of mandible, often distitarsi, and flagellar segments 3-10 below rufescent; eyes gray; wing membrane colorless to slightly milky; veins dark reddish brown to black; tegulae piceous; tibial spurs colorless to pale yellow. Apex of tergum 1 narrowly hyaline. Clypeus relatively flat, lateral angle separated from eye margin by half minimum diameter first flagellar segment or less, with median longitudinal carina usually present; second flagellar segment slightly broader at apex than median ventral length; maxillary palpal ratio about 3.0:2.7:2.3:1.0; galeae shiny, unshagreened above except per- haps delicately so in apical half or less. Sculpturing of head and thorax as in ochraea except posteromedian mesoscutal punctures separated mostly by more than one puncture width and often by two 432 Tue UNIVERSITY SCIENCE BULLETIN or more puncture widths, surface unshagreened. Tergal sculptur- ing as in menuachus except as follows: tergum 1 with basal two- thirds or more (medially) punctate, punctures shallow, separated by one-half to one and one-half puncture widths and reaching apical margin at extreme sides, impunctate apical area not expanded bas- ally at sides into impunctate lateral lobes; tergum 2 with interband zone punctures minute, separated by one to three puncture widths; terga 2 and 3 with apical areas with minute punctures at least bas- ally near distal pale bands. Pygidial plate broadly V-shaped with well-rounded apex, usually less than eight-tenths as broad at base as median length. Hair: Head pale ochraceous to entirely black, dark hairs appear first on labrum and mandibles, second on clypeus and vertex from which areas they spread over entire head in darker specimens. Tho- rax pale ochraceous to white laterally and posteriorly and dull ochra- ceous to somewhat rufescent above in palest specimens; in darkest specimens lateral and posterior surfaces, propodeum, tegulae, and anteriorly on mesoscutum dark brown to black, remainder of dorsum pale ochraceous to pale rufescent, posteromedian area of meso- scutum and median area of scutellum without dark hairs; intermed- iate specimens with entire dorsum, entire propodeum and upper lateral surfaces pale. Metasoma of palest specimen as in ochraea except as follows: interband zone of tergum 2 with hairs erect and without suberect or appressed plumose hairs; tergum 2 with distal pale band well separated from basal band except at extreme sides; tergum 4 with distal pale band narrow, medially narrower than basal area of dark hairs; terga 5 and 6 with little or no pale hairs laterally. Darkest specimens with metasomal vestiture entirely dark brown to black; pale hairs appear first at extreme base of tergum 2, second at base of tergum 1 and in distal pubescent band of tergum 2, third on tergum 3, and lastly on succeeding terga. Tergum 1 with long apical hairs of basal area appressed and reaching or overpassing margin across entire tergum whether these dark or pale (often worn away medially, however). Legs as in ochraea in pale specimens; in dark specimens dark brown to black except as follows: scopal hairs (ex- cept surrounding pygidial plate and at apex of basitarsus) ochra- ceous to yellow, inner surface hind tibiae ochraceous to yellow, hind femora often paler above, inner surface hind basitarsus dark brown to black. Male. Measurements and ratios: N, 20; length, 10-14 mm.; width, 3.0-4.0 mm.; wing length, M = 3.88 + 0.167 mm.; hooks in hamulus, BEES OF THE GENUS MELISSODES 433 M = 13.30 + 0.219; flagellar segment 2/segment 1, M = 4.68 + 0.089. Structure and color: Integument as in menuachus except as fol- lows: base of mandible black; labrum entirely black (allotype) or with small mediobasal pale spot; clypeus yellow with apical margin usually piceous; first flagellar segment dark brown; eyes gray to green; wing veins dark reddish brown to black; tegulae piceous. Structure as in menuachus except as follows: minimum length first flagellar segment one-fifth or less (less in allotype ) maximum length second segment, penultimate 3 or 4 segments slightly crenulate (as in semilupina); maxillary palpal segments in ratio of about 4.0:3.5: 3.5:1.0. Sculpturing as in female except as follows: metasomal tergum | with more than basal five-sixths (medially ) punctate; terga 2 and 3 with punctures usually slightly coarser and more crowded. Sterna 7 and 8 and genital capsule as in M. menuachua except gonocoxite without ventral hairs just below gonostylus. Hair: Head and thorax white to pale ochraceous, often somewhat darker ochraceous on vertex of head and dorsum of thorax. Meta- somal vestiture as in ochraea except distal pale bands of terga 2-4 (especially tergum 2) usually narrower than apical apubescent area; terga 2-4 with apical areas with abundant suberect pale hairs; tergum 2 with interband zone hairs erect; tergum 1 with distinct band of pale appressed pubescence obscuring apical margin across entire tergum. Two specimens (one from Santa Ana River, San Bernar- dino County and the other from Long Barn, Tuolumne County, Cali- fornia) with hairs and pubescence yellow-ochre to pale rufescent and brownish red in basal areas of terga 3 to 5. Legs as in ochraea. Remarks. This species is very remarkable because of the marked dimorphism in color of the females. This is not reflected in the males. The females present two distinct color patterns in the vesti- ture. These two extremes are described above. There are inter- mediate specimens, but these are relatively few in number. Out of a total of 173 females, 96 were classified as the darkest form, 73 as the pale form, 7 as almost perfectly intermediate, 14 as intermediate but nearer the dark form, and 9 as intermediate but near the pale form. It seems likely that a single pair of alleles, or a very few loci, are involved in the genetic background of this dimorphism. The inter- mediate types could be explained by microclimate affecting the rate of development and thus affecting melanism deposition during the prepupal or pupal stages of development. Significant in this respect 434 Tue UNIVERSITY SCIENCE BULLETIN is the fact that there are so very few intermediate specimens and the majority of these are more like one or the other of the extremes in color than they are like the few almost exact intermediates. Further- more, the dimorphism follows no apparent geographical pattern. Both forms of females and intermediates are available from such widely separated areas as Utah, Nevada, and northern and southern California. During preliminary studies, the author had segregated the females as two distinct species. There remains a possibility that this is the true situation. However, with the accumulation of ad- ditional specimens intermediate forms have become available. Fur- thermore, the two extreme forms, plus intermediates, have been collected in several instances at the same time, from the same flowers, and with the same males. Bionomics. This species is apparently oligolectic cn Compositae and, in particular, upon the genus Chrysothamnus, as is its close relative M. semilupina. Out of a total of 72 collections (93 females and 51 males) with floral data attached, 53 collections (80 females and 41 males) were made from some species of Chrysothamnus, whereas only 19 collections (13 females and 10 males) were ob- tained from other composites representing 11 genera. Type Material. Holotype female (pale form) and two female paratypes (one an intermediate form) collected by E. G. Linsley on Chrysothamnus nauseosus speciosus, September 7, 1957, from 8 miles S. of Ravendale, Lassen Co., California. Allotype male from 15 miles S. of Ravendale was collected by B. J. Adelson on Chryso- thamnus nauseosus consimilis, September 7, 1957. The holotype and allotype are in the collection of the University of California at Berke- ley. Five male and twenty-seven female paratypes from California are as follows: LASSEN CO.: Depau: 2 females, October 11, 1952, E. I. Schlinger. Hallelujah Junction: 1 female on Chrysothamnus nauseosus consimilis, (2.5 miles S.), September 6, 1957, E. G. Lins- ley; 1 female on C. n. consimilis, (6 miles N.) B. J. Adelson. Janes- ville (1 mile N.): 1 female on C. n. speciosus, September 8, 1957, B. J. Adelson; 3 females on C. n. speciosus, September 8, 1957, E. G. Linsley. Litchfield: 2 males on C. v. viscidiflorus, September 8, 1957, E. G. Linsley. Madeline (8 miles N.): 1 female on C. n. speciosus, September 7, 1957, J. A. Chemsak. Standish (4 miles W.) 1 female on C. v. viscidiflorus, September 7, 1957, E. G. Linsley. MODOC CO.: Alturas (8 miles N.): 1 female and 1 male, September 7, 1957, J. A. Chemsak. Cedar Pass: 8 females and 1 male, October 11, 1952, E. I. Schlinger. Juniper Flat: 1 female, July 1938, J. J. DuBois. BEES OF THE GENUS MELISSODES 435 Mason Creek Railroad Siding: 4 females, October 12, 1952, E. I. Schlinger. NEVADA CO.: Hobart Mills (7 miles N.): 1 female on Chrysothamnus sp., August 26, 1948, P. D. Hurd. Truckee (11 miles E. at Boca Dam): 1 female, September 15, 1957, E. G. Linsley. SIERRA CO.: Sierraville (3 miles N. W.): 1 female and 1 male on Chrysothamnus sp., September 9, 1957, E. G. Linsley. Twenty-nine male and four female paratypes from Washoe County, Nevada, are as follows: Purdy: 1 female on Chrysothamnus sp., September 6, 1957, E. G. Linsley; 1 female on C. n. consimilis, September 6, 1957, B. J. Adelson. Reno: 1 female and 1 male from 2 miles N., Septem- ber 6, 1957, E. G. Linsley; 1 female from 7 miles N. on C. n. con- similis, September 6, 1957, J. A. Chemsak. Sparks: 1 male from 12 miles N. on Chrysothamnus sp., September 2, 1957, E. G. Linsley; 21 males from 17 miles N. on C. n. consimilis, September 2, 1957, E. G. Linsley. Sutcliffe: 2 males on Chrysothamnus sp., 4 males without floral data, September 2, 1957, E. G. Linsley. Paratypes are in the collections of the University of California at Berkeley and at Davis, R. R. Snelling, Turlock, California, the Snow Entomological Museum of the University of Kansas at Lawrence, and in the author’s col- lection. Distribution. M. bimatris ranges from British Columbia south to southern California and east to Colorado and New Mexico (one eastern Colorado male is dubiously identified as bimatris, although it is in very poor condition (Fig. 12). It has been collected from June to November 8, but mainly during September. In addition to the type material, 148 females and 73 males have been examined from the localities listed below. Arizona: Black Mesa (near Kayenta); Pearce; Safford (30 miles S.); Tombstone (E. of); Yuma. Catirornia: Apple Valley, San Bernardino Co.; Barton Flatts; Caliente Mt. (2 miles N. E.), San Luis Obispo Co.; Carbon; Carmel; Deep Creek, Mojave Desert; Democrat Springs, Kern Co.; Gazelle; Helendale; Hesperia, S. Ber- nardino Co.; Imperial Co.; Lancaster (2 miles N.); Little Lake, Inyo Co.; Livermore (20 miles S. at Arroyo Mocho); Long Barn, Tuol- umne Co.; Los Angeles Co.; McArthur; Morongo Valley; Murphys; Olancha (13 miles S.); Oro Grande; Paynes Creek, Tehama Co.; Riverside; Santa Ana River, S. Bernardino Co.; Seven Oaks; Sonora Pass, Mono Co.; South Fork Camp, S. Bernardino Mts.; Tesla; Tur- lock; Victorville, Vincent; Westgard Pass, Inyo Co.; Whitewater. Cotorapo: Berkeley; Cortez. Ipano: Bliss, Conant; Parma; Ridge- dale. Nevapa: Eureka; Walker Lake, Mineral Co. New Mexico: 436 THE UNIVERSITY SCIENCE BULLETIN McCartys, Valencia Co.; Mescalero. Orecon: Abert Lake; Algoma, Klamath Lake; Echo; Redmond. Urau: Arches National Monu- ment; Beryl; Blanding (19 miles W.); Dugway Proving Ground, Tooele Co.; East Promontory; Freemont Pass, Iron Co.; Lehi, Logan; Milford; Pine Valley Mts.; Promontory; Salt Lake; Torrey; Tridell. WASHINGTON: Brewster; Coulee City; Gardena; North Yakima; Pasco; Stratford. Canada. British CoLtumpsia: Nicola; Oliver; Walhackin. Mexico. Sonora: Agua Priete. Flower Records. Artemesia sp., Aster sp., Centromadia pungens, Chaematoris sp., Chrysothamnus sp., C. nauseosus, C. n. consimilis, C. n. gnaphalodes, C. n. mohavensis, C. n. occidentalis, C. n. specio- sus, C. parryi, C. viridulus, C. viscidiflorus viscidiflorus, Ericameria palmeri, Eriogonum sp., Gutierrezia californica, G. lucida, G. saro- thrae, Helianthus sp., Isocoma acradenia, Rhamnus californica, Senecio sp. Melissodes (Eumelissodes) cerussata, n. sp. This species is known only from three females from San Ber- nardino County, California. These females resemble the females of M. menuachus in the shiny galeae, the long first flagellar segment, the dark hairs of the inner surfaces of the hind basitarsi, and in size, but differ in the generally white vestiture, the abundant, mi- nute punctures in the apical area of tergum 3 and of tergum 2 (but less distinct in the latter). The female also resembles the palest female of M. bimatris, but differs from the latter in the punctate tergal apices and the pale lateral tufts of hair on terga 6 and 7. Female. Measurements and ratios: N, 3; length, about 13 mm.; width, about 4.5 mm.; wing length, M = 3.79 + 0.406 mm.; hooks in hamulus, M = 15.67 + 0.882; flagellar segment 1/segment 2, M = 188 = 0.058. Structure and color: Integumental color as in menuachus except as follows: eyes gray; tergum 1 with apical area translucent, red to yellow; flagellar segments 3 to 10 and apex of 2 yellow below. Structure and sculpture as in menuachus except as follows: clypeal punctures small, round, surface dulled by tessellation; supraclypeal area with few punctures medially, surface moderately shiny, tes- sellate; second flagellar segment slightly longer than broad; maxil- lary palpal ratio about 4.0:3.4:2.8:1.0; lateral areas vertex with small round punctures separated by half to one or slightly more puncture widths, surface shiny but somewhat shagreened; meso- scutum with small impunctate posteromedial area, surface with BEES OF THE GENUS MELISSODES 437 fine reticular shagreening scarcely dulled; metasomal tergum 1 with basal three-fourths with round punctures separated mostly by half to one puncture width, apical area impunctate, without antero- lateral impunctate lobes; tergum 2 with basal area punctures sep- arated mostly by half a puncture width or less, interband zone punctures small, separated mostly by one puncture width, surface dulled, apical area with minute but distinct punctures separated mostly by two to three puncture widths, surface moderately shiny; tergum 3 similar to 2 but apical area punctures more distinct and more crowded. Hair: Head and thorax white except lower surfaces mesepisterna pale brown. Vestiture of metasomal terga as in menuachus except as follows: tergum 1 with basal area hairs white, apical area gla- brous, basal hairs often reach apex of tergum medially; tergum 2 with all hairs and pubescence white, apical area hairs abundant, subappressed to suberect, white; tergum 3 like 2 but basal tomen- tum brown, apical area hairs more abundant, distal pale band reaches apex at extreme sides; tergum 4 like 3 but lacking apical area; terga 5 and 6 dark brown with white lateral tufts. Legs white except basitibial plates orange, fore tarsi and inner surfaces middle and hind basitarsi dark brown, and inner surfaces hind tibiae yellow. Type Material. The holotype female from six miles west of Lud- low, San Bernardino Co., California, was collected October 17, 1951, on Geraea sp., by E. G. Linsley. Two female paratypes were collected at the same time and place on the same flower by P. D. Hurd and Ray F. Smith, respectively (see Fig. 10 for distribution ). The holotype and one paratype are in the collection of the Uni- versity of California at Berkeley. The second paratype is in the author’s collection. Melissodes (Eumelissodes) relucens, n. sp. This species is closely related to menuachus and to ochraea. It is similar to menuachus in the color of the vestiture except that the females usually have paler hairs on the inner surfaces of the hind basitarsi. The females can be separated from those of menuachus and ochraea by the coarser punctation of the mesoscutum, the first tergum, and the interband and basal areas of the second ter- gum. The female has the second flagellar segment about as long as broad. The male of relucens has a short first flagellar segment and a long penultimate segment as in menuachus. The punctation of the male is similar to that of the female, but is much coarser 438 Tue UNIversity SCIENCE BULLETIN at the base of the second tergum and on the sterna. In addition, the male is distinctive in having an extremely broad pygidial plate. Female. Measurements and ratios: N, 2; length, about 13 mm.; width, about 4.5 mm.; wing length, 3.74-3.79 mm.; hooks in hamulus, 13; flagellar segment 1/segment 2, 1.82-1.91. Structure and color: Integument black except as follows: apical half of mandible, lower surfaces of flagellar segments except first, distitarsi, and sterna rufescent; eyes green; wing membrane color- less, veins dark brown to black; tegulae piceous; tibial spurs yel- low to pale rufescent. Structure and sculpturing as in menuachus except as follows: clypeal punctures coarser, irregular in size, crowded, with subapical median shiny boss, surface shiny, un- shagreened; supraclypeal area shiny; galeae above shiny, un- shagreened; lateral areas of vertex with minute, sparse punctures, shiny; maxillary palpal ratio about 8:7:5:1. Mesoscutal punctures large, deep, round, crowded, posteromedially largest and separated mostly by half a puncture width or slightly more, surface with fine reticular shagreening but not or scarcely dulled; scutellum similar but punctures smaller; mesepisternal punctures small, round, deep, separated mostly by less than half a puncture width, mostly less in diameter than posteromedian mesoscutal punctures, surface shiny; propodeum with dorsal surface coarsely reticulorugose basally, with irregular punctures apically except medially, posterior surface punctate except upper inverted triangular area, lateral surfaces coarsely punctate, surfaces moderately shiny, upper tri- angle of posterior surfaces with fine, reticular shagreening slightly dulling surface. Tergum 1 with basal three-fifths medially with small round deep punctures separated mostly by half to one punc- ture width, punctures to apex laterally; terga 1-3 with apical areas impunctate, surfaces shiny with extremely fine reticulotransverse shagreening; tergum 2 with basal area with deep, small, round punctures separated mostly by half a puncture width or less, inter- band zone with small, irregular-sized and spaced punctures sep- arated mostly by one puncture width; pygidial plate V-shaped with relatively straight sides and acute apex. Hair: Color of vestiture as in menuachus except as follows: dorsum of thorax bright ochraceous to yellow and hairs short and appressed; metasomal pale pubescent bands white, that on tergum 2 not reaching apical margin and of almost equal width across tergum (not notched posteriorly as in ochraea), that on tergum BEES OF THE GENUS MELISSODES 439 3 reaching apex at extreme sides; terga 2 and 3 with interband zone hairs white, short, appressed (unlike bimatris); sterna red- dish brown except white laterally; legs white except brown on fore tarsi and on basitibial plate, basitarsi with inner surfaces red to reddish brown. Male. Measurements and ratios: N, 1; length, about 11 mm.; width, about 3.5 mm.; wing length, 3.63 mm.; hooks in hamulus, 12; flagellar segment 2/segment 1, 6.36. Structure and color: Integument black except as follows: clypeus yellow except brown apical margin; labrum with minute pale mediobasal spot; base of mandible with minute yellow spot; eyes gray; lower surface of flagellum yellow, upper dark reddish brown, apical half of mandible, distitarsi and sterna rufescent; wing mem- brane colorless, veins dark reddish brown; tegulae piceous; tibial spurs reddish yellow; tergal apices hyaline, yellow; pygidial plate red, Structure as in menuachus except as follows: minimum length first flagellar segment one-sixth of maximum length second segment or slightly less, penultimate segment almost four times as long as broad; maxillary palpal ratio about 14:12:7:1. Sculpturing as in female except as follows: tergum 1 with basal four-fifths punctate; tergum 2 with basal area punctures coarser than in female (in di- ameter almost equal to tergum |] punctures); terga 2 and 3 with interband zone punctures coarser and more abundant; sternal punctures large and crowded. Pygidial plate broader than median length, deeply notched in apical third so that apical portion less than half as broad as broadest width near base. Terminalia much as in agilis but median plates of sternum 7 with sparse hairs, gonostyli with hairs sparse, inner surfaces gono- coxites with blunt spicules sparse, and sternum 8 with apical hairs sparse and short. Hair: Color and form of vestiture as in menuachus with same ex- ceptions as in female. Type Material. Holotype female and allotype male from Dugout Wells, Big Bend National Park, Texas, collected on August 25, 1954, by R. M. Bohart are in the collection of the University of California at Davis. One female paratype from El Paso, Texas, collected by H. V. Daly, September 13, 1950, on Isocoma heterophylla is in the collection of the Snow Entomological Museum of the University of Kansas at Lawrence (Fig. 15). 440 THe UNIversiry SCIENCE BULLETIN perpolita Fic. 15. Map showing the known distributions of M. (Eumelissodes) hymenoxidis Cockerell, M. (E.) perpolita LaBerge, and M. (E.) relucens LaBerge. BEES OF THE GENUS MELISSODES 441 Melissodes (Eumelissodes) bicolorata, n. sp. This species seems to be most closely related to M. bimatris and the females of bicolorata closely resemble the darker females of bi- matris. However, the females of bicolorata have fuscous scopal hairs, unlike those of bimatris. In the dark scopal hairs and general vestitural color, bicolorata is almost identical with M. (Callimelis- sodes) nigracauda from which it differs by the subgeneric charac- teristics and by having shiny, unshagreened galeae. The male of bicolorata is a medium sized, pale bee with pale wing veins not unlike M. agilis from which it differs by the dark labrum and man- dibular bases. The male closely resembles that of bimatris from which it can be distinguished by the lack of a distinct apical pale pubescent band obscuring the apex of the first tergum, by the pale wing veins, and by the flagellum not being crenulate (as it is in both bimatris and semilupina). Female. Measurements and ratios: N, 20; length 10-13 mm.; width, 3.5-4.5 mm.; wing length, M = 3.55 + 0.096 mm.; hooks in hamulus, M = 14.50 + 0.212; flagellar segment 1/segment 2, M = 1.96 + 0.024. Structure and color: Integument black except as follows: apical half of mandible, usually distitarsus and lower surfaces of flagellar segments 3-10 rufescent; eyes dark gray to greenish gray; wing mem- branes slightly milky, veins reddish brown to red; tegulae piceous; tibial spurs yellow; apex of tergum 1 piceous. Structure and sculpturing as in bimatris except as follows: maxil- lary palpal ratio about 4.0:3.0:3.5:1.0; tergum 1 with basal three- fifths or less medially punctate, punctures not extending to apex laterally and impunctate apical area extending anterolaterally as in- distinct lobes; tergum 2 with interband zone punctures minute and sparse; terga 2 and 3 with apical areas with punctures at bases of hairs minute and sparse but distinct and almost reaching apex of tergum at least laterally on tergum 2 and across entire tergum 3; pygidial plate broadly V-shaped with apex rounded, usually more than eight-tenths as broad at base as median length. Hair: Head dark brown except long hairs of vertex and on face above and surrounding antennal fossae often pale ochraceous. Thorax above, including dorsal and posterior surfaces of propodeum and often mesepisterna and metepisterna just below wing bases, pale ochraceous to slightly ferrugineous (fox-red ), laterally and ventrally dark brown to black. Metasoma dark brown to black except as fol- lows: tergum 1 with basal half to three-fifths of dorsal surface pale 16—5840 4492, THE UNIVERSITY SCIENCE BULLETIN ochraceous (holotype ) to slightly rufescent and tergum 2 with basal zone hairs often ochraceous to pale brown (dark in holotype). Legs dark brown except as follows: median scopal hairs of tibiae and occasionally near base of basitarsi usually paler brown to ochraceous. Male. Measurements and ratios: N, 7; length, about 11 mm.; width, about 3.5 mm.; wing length, M = 3.53 + 0.211 mm.; hooks in hamulus, M = 12.86 + 0.509; flagellar segment 2/segment 1, M = 7.30 + 0.351. Structure and color: Integument as in bimatris except as follows: labrum without mediobasal pale spot; first flagellar segment yellow to red below; eyes greenish gray; wing membranes slightly milky, veins yellow to reddish yellow; tergal apices hyaline, colorless. Structure as in bimatris except as follows: minimum length of first flagellar segment equals about one-sixth maximum length of second segment, penultimate 3 or 4 segments not crenulate; maxil- lary palpal segments in ratio of about 4:3:3:1. Sculpturing as in female except as follows: tergum 1 with basal four-fifths punctate; terga 2 and 3 with interband zone punctures slightly larger and more crowded, and with apical area punctures indistinct and sparse; ter- gum 4 similar to tergum 3. Pygidial plate broad but more as in bimatris than as in relucens, width at base subequal to median length and usually slightly less. Sterna 7 and 8 and genital capsule as in M. menuachus except as follows: gonostyli gently curved in or straight, scarcely capitate, with hairs near base sparse; gonocoxite with a few extremely short hairs at apex just below gonostylus on ventral surface; spatha with well-marked, apicomedian, shallow notch; sternum 7 with median plate with apical margin transverse; sternum 8 with ventral tubercle bidentate, with abundant apical hairs. Hair: Vestiture white to extremely pale ochraceous (usually pale ochraceous only on upper surface of thorax). Metasomal vestiture as in bimatris except as follows: tergum 1 without distinct pale pubescent band obscuring apical margin except at extreme sides and these in width less than a third width of tergum; terga 2-4 with apical area hairs suberect and less abundant; tergum 2 with inter- band zone hairs suberect. Type Material. The holotype female, allotype male and 23 fe- male paratypes were collected by C. D. Michener at Dayton, Ne- vada, July 1, 1950 on Penstemon palmeri. Two female paratypes from Nevada are as follows: 1 female on June 20 and 1 female on June 29, 1927, at Nixon by E. P. Van Duzee. In addition, 17 fe- BEES OF THE GENUS MELISSODES 443 male and 5 male paratypes from California are as follows: Halle- lujah Junction, Lassen County: 1 male, July 7, 1949, P. D. Hurd; 1 male, July 13, 1949, P. D. Hurd; 1 male, July 13, 1949, E. I. Schlin- ger; 2 females and 2 males, July 13, 1949, F. Morishita; 11 females, July 13, 1949, on Chrysothamnus sp., P. D. Hurd; 3 females, Au- gust 9, 1949, on Chrysothamnus sp., J. W. MacSwain. Murphys, Calaveras County: 1 female, September 8-19, 1937, F. E. Blaisdell. The holotype and allotype are in the Snow Entomological Collec- tion of the University of Kansas, Lawrence. Paratypes are in the collections of the Snow Entomological Collection, the University of California at Berkeley, the University of California at Davis, the California Academy of Sciences at San Francisco, P. H. Timbler- lake at the Citrus Experiment Station, Riverside, California, the U. S. National Museum, Washington, D. C. and in the author’s collection. Distribution. Northern California, Nevada and Utah (Fig. 10). In addition to the 52 specimens from California and Nevada listed above as type material, 3 specimens have been examined from Utah as follows: Leota: 3 females, July 17, 1952, on sweet clover ( Melilotus sp.), G. F. Knowlton and G. E. Bohart. Melissodes (Eumelissodes) perpolita, n. sp. This is a small coarsely punctate species similar to and related to M. relucens. The female of perpolita differs from that of relucens by the brown scutellar and mesoscutal hairs, the interband zone of tergum 2 being almost bare but with some brown hairs, the puncta- tion of tergum 2 as described below, the longer apical areas of terga 2 and 3 and the presence of a small median glabrous apical area on tergum 4. The male of perpolita can be told from that of relu- cens by the slightly narrower pygidial plate and the less hairy and less punctate (although more coarsely so) interband zone of tergum 2. The allotype is the only known male of perpolita and, since it was not collected with any of the females, the association of the two sexes is tentative. Female. Measurements and ratios: N, 4; length, about 10 mm.; width, about 3.5 mm.; wing length, M = 3.12 + 0.225 mm.; hooks in hamulus, M = 12.00 + 0.435; flagellar segment 1/segment 2, M = 1.84 + 0.034. Structure and color: Integumental color as in relucens except as follows: second flagellar segment dark below; eyes bluish gray (holotype) to gray; tegulae rufescent. Structure and sculpture as 444 Tus UNIVERSITY SCIENCE BULLETIN in menuachus except as follows: clypeal punctures large, separated mostly by less than half a puncture width, surface unshagreened, apicomedian carina weak; supraclypeal area shiny; galeae shiny, unshagreened; lateral areas of vertex with punctures large, irregu- lar, separated mostly by half to one puncture width; maxillary palpal ratio about 2.7:1.7:1.5:1.0; mesoscutal punctures very large, postero- medially larger than mesepisternal punctures and separated by half to two puncture widths (scattered irregularly), surface shiny; scu- tellar punctures slightly smaller, crowded; metasomal tergum 1 with basal two-thirds with round shallow punctures about same size as scutellar and separated mostly by half a puncture width, surface reticularly shagreened but shiny, apical area impunctate, finely sha- greened, shiny, with anterolateral lobes separated from rest of apical area by an uneven double row of coarse punctures; tergum 2 with basal area punctures small, separated mostly by half to one puncture width, surface somewhat dulled by reticular shagreening, inter- band zone punctures irregular in size and distribution, largest as large as mesoscutal punctures, smallest minute, with conspicuous blank spaces between punctures, surface shiny with fine reticulo- transverse shagreening, apical area impunctate, shiny, longer medi- ally than distal pale band; tergum 3 similar to 2 but apical area shorter; tergum 4 similar to 3 but apical area reduced to small me- dian triangle; pygidial plate V-shaped, apex acute. Hair: Head white with brown on vertex. Thorax white with scutellum brown fringed with white and mesoscutum with postero- median brown patch about one and one-half times size of scutellar dark area; tegulae without brown. Tergal vestiture as in relucens except as follows: tergum 2 with distal pale band slightly shorter than apical area medially, interband zone with short, simple sub- appressed to suberect hairs, usually at least partly brown and sparser than in relucens; tergum 3 similar but interband zone narrower and basal tomentum dark brown; tergum 4 with distal pale band inter- rupted apicomedially by triangular apubescent area; hind basitarsi with inner surfaces yellow to dark red; scopae white. Male. Measurements and ratios: N, 1; length, about 11 mm.; width, about 3 mm.; wing length, 3.18 mm.; hooks in hamulus, 12; flagellar segment 2/segment 1, 5.07. Structure and color: Integumental color as in relucens except as follows: labrum and mandibles black; eyes yellowish gray; first flagellar segment slightly less than one-sixth maximum length second segment, penultimate segment one-third as wide as long or slightly BEES OF THE GENUS MELISSODES 445 less; maxillary palpal ratio about 3.0:2.5:2.5:1.0; pygidial plate longer than broad, apicolateral notches deep so that apical part half median width. Sculpture as in female except as follows: clyp- eal punctures smaller; terga 2 and 3 with interband zones with large punctures more abundant and more crowded; terga 4 and 5 similar to 3 but apical areas shorter. Hair: Head and thorax white. Metasomal vestiture as in relu- cens except as follows: tergum 1 with apical area exposed; terga 2 and 3 with interband zones with hair less abundant. Legs white except inner surfaces hind basitarsi yellow. Type Material. The holotype female from Black Mesa (near Kayenta), Arizona, was collected by Isabel McCracken, September 11, 1936, on Chaemataxis sp. The allotype male from Bisbee (10 miles N. W.), Arizona, was collected by T. Cohn, P. Boone and M. Cazier, September 7, 1950. Two paratype females from Kaibab Forest, Utah, (this is probably mislabelled “Utah” and should be Arizona), were collected by I. McCracken, September 21, 1938, on Aster sp. One paratype female from Cortez, Colorado, was col- lected by I. McCracken, September 13, 1938, on Grindelia sp. The holotype and one paratype are in the collection of the California Academy of Sciences, San Francisco. The allotype is in the Ameri- can Museum of Natural History, New York City. Paratypes are in the collection of the Snow Entomological Museum of the Uni- versity of Kansas, Lawrence and in the author’s collection (Fig. 15). Melissodes (Eumelissodes) fasciatella, n. sp. M. fasciatella is a small distinctive bee from Arizona known only in the female sex. It is not closely related to any of the foregoing species but bears some resemblance in the tergal banding and punctation to M. perpolita so is treated here. M. fasciatella is distinctive in that the pale distal band of tergum 2 is reduced to two short lateral fasciae, each about one-third of the tergum in width and tapering sharply mesad, thus forming two short oblique fasciae. The extensive apical area of tergum 2 is impunctate and shiny. The galeae are shiny and hairs of the inner surfaces of the hind basitarsi are dark reddish brown. Female. Measurements and ratios: N, 20; length, 9-10 mm.; width, 3.0-3.5 mm.; wing length, M = 2.71 + 0.067 mm.; hooks in hamulus, M = 10.85 + 0.150; flagellar segment 1/segment 2, M = ST =E7101238: Structure and color: Integument black except as follows: mandi- 446 Tue UNIVERSITY SCIENCE BULLETIN bles and distitarsi rufescent; flagellar segments 3-10 red below; eyes bluish to greenish gray; wing membranes colorless to milky, veins dark reddish brown; tegulae piceous; tibial spurs yellow; tergal apices often slightly rufescent, tergum 1 narrowly hyaline apically. Structure and sculpture as in perpolita except as follows: lateral areas vertex with minute punctures separated by two to four punc- ture widths, surface shiny; maxillary palpal ratio about 2.0:1.7: 1.7:1.0; mesoscutal punctures large but not larger than mesepister- nal, small posteromedial area usually impunctate, elsewhere punc- tures separated mostly by half to one puncture width, surface shiny, often (not in holotype) with fine reticular shagreening; mesepister- nal punctures large, deep, separated by less than half a puncture width, surface shiny; metasomal tergum 1 with basal half or slightly more with large round punctures, not very shallow, separated mostly by half a puncture width or slightly more, surface reticulotrans- versely shagreened but shiny, apical area impunctate, shiny; tergum 2 with basal area punctures large, slightly less in diameter than those of base of tergum 1, separated mostly by half a puncture width, interband zone punctures smaller to slightly larger than those of basal area, separated by half a puncture width laterally to one or two puncture widths in median third, apical area impunctate, highly shiny; tergum 3 similar to 2 but apical area shorter and interband zone punctures small and denser; tergum 4 like 3 but lacking apical area; pygidial plate V-shaped, apex rounded, longer than broad. Hair: Head white to pale ochraceous on vertex, vertex with few or no brown hairs. Thorax white laterally; scutellum dark brown fringed with pale ochraceous; mesoscutum pale ochraceous with large posteromedian dark brown patch usually twice size of scutellar dark area or larger; tegulae without brown; mesoscutal hairs short, blunt-tipped and usually decumbent except peripherally. Metaso- mal tergum | white to pale ochraceous basally and to apical margin at extreme sides, glabrous apicomedially; tergum 2 white basally; distal pale band in form of two lateral fasciae, each one-third or less width of tergum and sharply tapered towards middle of tergum to form short, oblique, lateral fasciae, interband zone with sparse, appressed to subappressed, pale ochraceous pubescence and bristle- like hairs, apical area glabrous; tergum 3 similar to 2 but basal tomentum brown, distal pale band not interrupted medially (al- though posterior margin slants forward to an obtuse point medially ), and apical area shorter; tergum 4 like 3 but distal pale band reaches apical margin across entire tergum; terga 5 and 6 dark brown with BEES OF THE GENUS MELISSODES 447 lateral white tufts at least on 5; sterna brown medially to white laterally. Legs white except as follows: fore tarsi, outer-apical sur- faces fore and middle tibiae, and basitibial plates brown; inner surfaces hind basitarsi dark reddish brown to dark brown. Type Material. The holotype female and one female paratype from Todd’s Lodge, Oak Creek Canyon, Arizona, was collected on September 7, 1948, by Grace H. and John L. Sperry. The Sperrys also collected one female paratype at Todd’s Lodge on September 29, 1948. In addition, 32 female paratypes from Arizona are as follows: Dos Cabezas (16 miles S.): 1 female, September 8, 1950, T. Cohn, P. Boone and M. Cazier. East Verde River: 6 females collected at 4,500 feet altitude. Madera Canyon, Santa Rita Mts.: 1 female on Aplopappus gracile, September 23, 1956, F. G. Werner. Onion Saddle (9 miles W.), Chiricahua Mts.: 2 females, Septem- ber 10, 1954, J. C. Hall; 5 females on A. gracilis, September 10, 1954, coreopsis Fic. 16. Map showing the known distributions of M. (Eumelissodes) fasciatella LaBerge M. (E.) coreopsis Robertson, and M. (E.) pilleata LaBerge. 448 Tue UNIveRSITY SCIENCE BULLETIN P. H. Timberlake. Prescott: 2 females on A. gracilis, September 17, 1953, P. H. Timberlake. Price (2.9 miles N.): 3 females on Erigeron sp., 3 females on A. gracilis, September 17, 1953, P. H. Timberlake. Sedona: 2 females, September 14, 1955, G. D. Butler; 1 female on Viguiera sp. (10 miles N.), September 13, 1955, G. D. Butler. Seligman: 1 female on Gutierrezia sp., August 29, 1931, P. H. Timberlake. Southwest Research Station (5 miles W. of Portal): 2 females, September 8, 1955, W. Gertsch and E. Ordway. S. Arizona: 3 females, August 1902, F. H. Snow. The holotype is in the collection of the Snow Entomological Museum of the University of Kansas, Lawrence. Paratypes are in the col- lections of the Snow Entomological Museum, P. H. Timberlake, Riverside, California, The American Museum of Natural History, New York City, the U. S. National Museum, Washington, D. C., the University of Arizona, Tucson, the University of California at Davis, and in the author's collection (Fig. 16). Melissodes (Eumelissodes) coreopsis Robertson Melissodes coreopsis Robertson, 1905, Trans. Amer. Ent. Soc., vol. 31, p. 368; 1914, Ent. News, vol. 25, p. 69; 1926, Ecology, vol. 7, p. 379; 1928, Flowers and Insects, p. 8. Melissodes agilis semiagilis Cockerell, 1906, Ann. Mag. Nat. Hist., ser. 7, vol. 17, p. 364 (new synonymy); 1907, Univ. Colo. Studies, 4:225; 1014. Canadian Ent., 46:413; 1919, Canadian Ent., 51:27; 1928, Univ. Colo. Studies, 16:114. Melissodes confusiformis Cockerell, 1906, Ann. Mag. Nat. Hist., ser. 7, vol. 17, p. 366 (new synonymy); 1907, Univ. Colorado Studies, vol. 4, p. 255; 1910, Psyche, vol. 17, p. 246; 1914, Canadian Ent., vol. 46, p. 409; 1919, Canadian Ent., vol. 51, p. 272; 1923, Ent. News, vol. 34, p. 47; Bohart, Knowlton and Bailey, 1950, Utah St. Agric. Coll., Mimeo. Ser. No. 371, p. 5. Melissodes helianthophila Cockerell, 1914, Ann. Mag. Nat. Hist., ser. 8, vol. 14, p. 361 (new synonymy ). Melissodes confusa, Robertson (nec Cresson, 1878), 1894, Trans. Acad. Sci. St. Louis, vol. 6, pp. 458-460, 468, 471, 474, 475; 1896, Trams. Acad. Sci. a4 Bais. vol. 7, pp. 175, 176, 178; 1897, Trans. Acad. Sci. St. Louis, vol. Pop eOoos Melissodes coreopsis is the most common species of Eumelissodes of the Great Plains except perhaps M. agilis. It is a medium-sized bee, the female of which has black and white metasomal bands, a large black dorsal thoracic patch, pale scopal hairs with dark hairs on the inner surfaces of the hind basitarsi, and short subappressed dark hairs in the apical areas of terga 2 and 3. It is not closely related to any of the foregoing species but is most similar to M. rustica from which it differs by the paler thoracic and head hairs of both sexes, the less shiny terga of both sexes, the paler scopal hairs of the female, and the hyaline tergal apices and shorter first flagellar segment of the male. A relatively complete description BEES OF THE GENUS MELISSODES 449 is given below after which subsequently described and _ related species are patterned. Female. Measurements and ratios: N, 20; length, 9-14 mm.; width, 3.0-4.5 mm.; wing length, M = 3.38 + 0.196 mm.; hooks in hamulus, M = 12.70 + 0.206; flagellar segment 1/segment 2, M = 1.76 + 0.024. Structure and color: Integument black except as follows: apical half of mandibles and distitarsi (often basitarsi and tibiae as well) rufescent; lower surfaces of flagellar segments 3-10 and often apex of second segment yellow to red; eyes blue to bluish gray or dark gray; wing membranes hyaline, colorless or slightly milky, veins dark brown to black; tegulae piceous, occasionally slightly testa- ceous; tibial spurs white to yellow; tergum 1 usually rufescent in apical third or half and with extremely narrow apical margin hya- line and colorless or yellow. Clypeus flat, oculoclypeal distance half minimum width of first flagellar segment or less, with punctures relatively regular, deep, round, separated mostly by half a puncture width or less, surface shiny, with irregular cross-striations especially near base; supra- clypeal area with punctures sparse or absent medially, usually slightly dulled by fine reticluar shagreening; vertex with flattened lateral areas with small round punctures separated by one to two or more puncture widths, surface usually shiny; galeae above shiny, unshagreened except near tips; maxillary palpal ratio about 9:6:5:1, last segment often almost obliterated. Mesoscutal punctures large, round, deep, posteromedially larger and separated by one to three puncture widths, anteriorly and laterally smaller and separated by half to one puncture width or less; scutellar punctures smaller, medially separated mostly by one to two puncture widths; mes- episterna with lateral surface punctures as large or larger than posteromedian mesoscutal punctures, separated mostly by less than half a puncture width; surfaces of mesoscutum and scutellum usu- ally shiny, often slightly dulled by delicate reticular shagreening; surfaces of mesepisterna usually unshagreened; propodeum with dorsal surface reticulorugose, coarser near base, posterior surface with abundant shallow punctures except in upper inverted triangu- lar area, lateral surfaces densely punctate, surfaces everywhere dulled by fine, dense tessellation. Metasomal tergum | with basal half to three-fifths with small shallow punctures separated mostly by half a puncture width, apical impunctate area extended basally on each side to form indistinct anterolateral lobes, surface dulled 450 THe UNIVERSITY SCIENCE BULLETIN by dense reticulotransverse shagreening; tergum 2 with basal area punctures round, deep, separated by half to one puncture width, larger apically, surface shiny and unshagreened, interband zone punctures larger and shallower, separated by half to one puncture width laterally and mostly by one puncture width medially, surface dulled by dense reticulotransverse shagreening, apical area im- punctate or with minute, widely separated punctures, surface dulled by dense shagreening; tergum 3 similar to 2 but punctures of interband zone smaller and more crowded and apical area with minute punctures somewhat more abundant. Pygidial plate broadly V-shaped with rounded apex, longer than breadth at base. Hair: Head white to pale ochraceous with long brown hairs on vertex. Thorax white to pale ochraceous except scutellum dark brown with pale fringe and mesoscutum with posteromedian dark patch as large and usually larger than scutellar dark area; tegulae usually with brown hairs posteromedially. Metasomal tergum 1 with basal area white to pale ochraceous, apical area glabrous; tergum 2 with basal tomentum white and connected laterally by white pubescence to distal pale band, distal band white, laterally longer than apical area, narrowly interrupted medially, interband zone hairs subappressed to suberect, all or mostly dark brown, apical area with short, subappressed, relatively simple, brown to black hairs usually present except near apex; tergum 3 as in 2 but basal tomentum brown, distal pale band broader and usually uninter- rupted medially, apical area narrower (distal pale band may reach apex at extreme sides), and with more abundant simple, subap- pressed, dark hairs in apical area; tergum 4 with apical pubescent band broad, white, uninterrupted posteromedially; terga 5 and 6 dark brown with pale lateral tufts; sterna brown or reddish brown medially to white laterally. Legs pale ochraceous to white except as follows: distitarsi yellow to reddish brown; fore basitarsi, inner surface middle and hind basitarsi, on and surrounding pygidial plate and often outer surfaces of apices of fore and middle tibiae reddish brown to dark brown; inner surfaces hind tibiae yellow. Male. Measurements and ratios: N, 20; length, 8-12 mm.; width, 2.5-4.0 mm.; wing length, M = 3.40 + 0.126 mm.; hooks in hamulus, M = 12.10 + 0.143; flagellar segment 2/segment 1, M=8.71 + 0.204. Structure and color: Integument black except as follows: clypeus yellow except apical margin testaceous to brown; mandibular bases without yellow maculae; labrum entirely black or with small medio- BEES OF THE GENUS MELISSODES 451 basal pale spot (in about 50 per cent of specimens); eyes yellow- brown, green or bluish green; flagellum yellow below, dark red to brown above; wing membranes colorless to slightly milky, veins dark red to brown; tegulae usually testaceous; apical tergal areas hyaline, colorless or slightly yellow, basal to hyaline area usually rufescent on at least terga 2 and 3; distitarsi and often basitarsi rufescent; tibial spurs white to pale yellow. Clypeus as in female; first flagellar segment with minimum length equal to less than two-thirds maximum length and equal to one-tenth or less maximum length of second segment, penultimate segment more than three times as long as broad, flagellum in repose surpassing pterostigma, segments 4-10 without longitudinal lateral depressions; maxillary palpal ratio about 8:5:4:1, last segment occa- sionally absent. Sculpturing as in female except as follows: meso- scutum with posteromedian area punctures often somewhat more crowded; tergum 1 medially with basal four-fifths punctate, punc- tures separated mostly by one puncture width; terga 2 and 3 inter- band zone punctures more abundant; terga 2-4 with apical areas impuctate or virtually so, dulled, with shagreening often more reticu- lar and less transverse, especially in interband zones; sterna moder- ately shiny, surfaces usually with coarse reticular shagreening. Terminalia as in M. agilis. Hair: White to pale ochraceous; vertex of head and dorsum of thorax usually more ochraceous than elsewhere; metasomal hairs and pubescence entirely pale, as in M. menuachus except as follows: tergum | with apical area with long, subappressed, relatively simple, pale hairs usually present but not forming dense band hiding mar- gin; tergum 2 with distal pale band narrow, usually half to three- fourths as wide as apical area medially, as long as or longer than apical area laterally, occasionally narrowly interrupted medially; terga 2-4 with apical areas progressively shorter, with abundant long subappressed to suberect, relatively simple, pale hairs. Legs white to pale ochraceous except yellow to reddish yellow on inner surfaces of tarsi and hind tibiae. Remarks. Dr. Delma Harding of the Zoology and Entomology Department, Iowa State College, Ames, Iowa, has provided the author with an excellent photograph (Fig. 1) of a female bee visit- ing Helianthus petiolaris in Kansas. This bee is most likely the fe- male of M. coreopsis. Bionomics. Out of 1,986 specimens of M. coreopsis available for study, 1,061 bear flower labels. These data are summarized in 452 THe UNIVERSITY SCIENCE BULLETIN Table VII and indicate that coreopsis is oligolectic on plants of the family Compositae. The bee visits a great variety of composite genera and species for pollen as well as nectar, and shows some preference for the genus Helianthus and related genera. TABLE VII. Summary of Floral Records for Melissodes coreopsis. Plant Data Records of M. coreopsis : ‘ : a. one ooo o x iS) =i i = i + Ho ab) FAMILY E Ee £8 g 5 28 ag fa) Ee) 2S | Bs | Fa) gs Z Z Z ZA Z = Compositae: Helianthus spp. 1 6 86 211 124 339 Gaillardia spp. 1 1 7 54 14 68 Rudbeckia spp. 1 3 16 44 65 109 Echinacea spp. 1 3 18 36 13 49 Grindelia spp. ! 1 15 22 4 26 Solidago spp. fo |tete 15 15 78 93 Aster spp. Yt 2 14 | 38 3 41 Coreopsis spp. lana | P| Ps.) 25 3 28 Other genera 19 23 56 77 82 159 Leguminosae | 8 2 41 32 43 75 Labiatae 3 S |} 13 12 30 42 Other families (12) | ale 15 31 13 23 36 Totals i 7A} 330 | 579 | 482 /1,061 Type Material. The lectotype female, here designated, of coreop- sis, collected by Charles A. Robertson at Carlinville, Illinois, June 14, 1902 on Coreopsis palmata, is in the collection of the Illinois Natural History Survey at Urbana. The holotype female of con- fusiformis from Fedor, Lee County, Texas, May 6, 1902, is in the collection of the Natural History Museum of the University of Colo- rado at Boulder. The holotype male of helianthophila, collected by T. D. A. Cockerell at Boulder, Colorado, June 16, on Helianthus BEES OF THE GENUS MELISSODES 453 lenticularis, is in the collection of P. H. Timberlake at the Citrus Experiment Station, Riverside, California. Distribution. M. coreopsis is distributed from Alberta, North Dakota and Minnesota south to Oaxaca in Mexico, east to Indiana and west to Utah and Arizona (Fig. 16). It has been collected from April 11 to November 6. It seems likely that in Texas where coreopsis is abundant from April until November there are three generations of bees per year. In Kansas where this bee is active from June until mid-October, there are probably two generations and in North Dakota where the season of activity is limited to July, August and September, there is likely to be only one generation. In addition to the type material, a total of 1,121 females and 865 males have been examined from the localities listed below. Of these 1,986 specimens, 1,140 are from the state of Kansas. There- fore, locality records from that state are given below only as counties in order to conserve space. Localities reported in the literature are included in the list. ARIZONA: Douglas; Flagstaff (Walnut Canyon ); Oak Creek Can- yon. ARKANSAS: Desha Co.; Hot Springs; Washington Co. Coto- RADO: Antonito; Beneva Park; Berkeley; Boulder; Boulder Canyon; Brighton; Buckeye (S. at Horsecreek ); Cameron Pass; Canon City; Chimney Gulch; Clear Creek; Colorado Springs; Cory; Cotopaxi; Crook, Logan Co.; Crowley Co.; Denver; Dixon Canyon; Eads; Elbert; Eldora; Estes Park; Fort Collins; Glen Haven; Jim Creek (near Boulder); La Junta; Lamar; Larimer Co.; Limon; Mason- ville; Mesa Verde; Ovid (3 miles E.); Palmer Lake; Pingree Park, Larimer Co.; Platte Canyon; Portland; Prospect, Weld Co.; Puils Creek, Crowley Co.; Rock Creek, Teller Co.; Rocky Ford; Seibert (13 miles E.); Sterling; Stratton; Ten-sheep Ranch; Timpas; Towner, Kiowa Co.; Trimnath; Valmont (Owens Lake); Virginia Dale; White Rocks (near Boulder). Inurors: Carlinville; Ma- coupin Co. Iowa: Ames; Buffalo Center (5 miles N. W.); Dickin- son Co.; Grundy Co.; Lyon Co.; Onawa; Sioux City. Kansas: Counties: Anderson; Barton; Bourbon; Butler; Chase; Chautauqua; Cherokee; Cheyenne; Clark; Cloud; Coffey; Cowley; Dickinson; Douglas; Edwards; Ellis; Finney; Ford; Franklin; Gove; Greeley; Greenwood; Hamilton; Harper; Harvey; Hodgeman; Johnson; Kearny; Kiowa; Labette; Lane; Leavenworth; Logan; McPherson; Marion; Marshall; Meade; Mitchell; Montgomery; Morton; Neosho; Norton; Ottawa; Pawnee; Pottawatomie; Pratt; Reno; Republic; Rice; Riley; Rooks; Rush; Russell; Saline; Scott; Sedgwick; Shaw- 454 Tue UNIVERSITY SCIENCE BULLETIN nee; Sheridan; Sherman; Smith; Stafford; Stanton; Stevens; Sumner; Thomas; Trego; Wallace; Wichita; Woodson. Minnesora: Moor- head; Rock Co.; Roseau Co.; St. Paul; Yellow Medicine Co. Mis- sourt: Branson; Chillicothe (6 miles N.); Columbia; Holden; Ozark Lakes; Warsaw. Nepraska: Agate, Sioux Co.; Alliance; Box Butte Co.; Brown Co.; Cambridge; Carns; Cedar Bluffs; Crawford; Dun- ning; Fairmont; Glen, Sioux Co.; Gordon, Haigler; Halsey; Hamlet, Hays Co.; Hardy; Harrison; Hyannis (9 miles S.); Imperial; Kim- ball; Lincoln; Lodgepole; McCool Je.; Malcolm; Mitchell; Monroe Canyon, Sioux Co.; North Platte (8 miles W.); Omaha; Pine Ridge, Dawes Co.; Sioux Co.; Wabash; War Bonnet Canyon, Sioux Co.; Weeping Water; West Point. New Mexico: Capitan; Carlsbad Caverns; Corrizozo; Grady; Las Vegas; Maxwell; Rowe; San Jose; Santa Fe (35 miles E.); Sapello. Norra Dakota: Amidon, Beach; Belfield; Bismarck; Dickinson; Fargo; Hatton; Jamestown; Oakes; Ravinia; Williston; Valley City. OxtaHoma: Ardmore; Caddo; Lawton; Nowata (5 miles N.); Okmulgee; Quapaw; Wagoner (5 miles N.); Waurika; Vinita. Sourn Daxota: Ardmore; Buffalo; Cedar Pass (Badlands); Custer; Custer Co.; Deadwood (10 miles S.); Deerfield; Edgemont; Hot Springs; Interior; Okaton; Slim Buttes; Stanley Co. Texas: Adrian; Alford; Alpine (20 miles S.); Atascosa Co.; Austin; Bay City; Bexar Co.; Big Bend National Park; Brazos Co.; Brewster Co.; Chisos Mts. (Big Bend. N. Park); College Station; Corpus Christi; Cotulla; Dalhart (Rita Blanca Lake); Dallas; Del Rio; Denton; Devil's River; Dilley; Eastland Co.; El Paso (15 miles N.); Fedor, Lee Co.; Fort Davis; Fredricksburg; Giddings; Goliad (16 miles E.); Greenville; Guthrie; Harper; Hetty; Hillsboro; Jack Co.; Johnson City (6 miles W.); Kerrville; Ladonia; Lee Co.; Lobo, Culberson Co.; Magnolia; Marfa; Matagorda; Palo Duro Canyon, Randall Co.; Paris; Plano; Poteet; Quemado, May- erick Co.; Roanoke; Rock Island; Romero; Stonewall; Terrell; Vic- toria; Waco; Wichita Falls; Willis; Wolfe City. Uran: Lakepoint. Wyominc: Albany Co.; Diamond Ranch, Platte Co.; Grand Teton National Park; Laramie (37 miles E.); Laramie Co.; Summit; Tie Siding. Canada. Avperra: Lethbridge. Mexico. Oaxaca: Nochixtlan (7 miles S. E.). Flower Records. Amphiachyris sp., A. dracunculoides, Amorpha canescens, A. fruticosa, Aster sp., A. ericoides villosis, A. multiflora, A. novaeangliae, A. paniculatus, A. praeatus, Bidens sp., B. invol- ucrata, Boltonia asteroides, Chrysopsis sp., C. angustifolia, Chrys- othamnus graveolus, Cirsium sp., Clematis sp., Cleome serrulata, BEES OF THE GENUS MELISSODES 455 Cooperia pedunculata, Convolvulus sp., Coreopsis sp., C. grandi- florum, C. palmata, C. tinctoria, Cosmos sp., Echinacea sp., E. angustifolia, E. pallida, E. purpurea, Erucastrum pollichii, Eryn- gium sp., E. leavenworthii, Eupatorium altissimum, Euphorbia sp., Eustoma russellianum, Gaillardia sp., G. pulchella, Geranium sp., Gossypium herbaceum, Grindelia sp., G. squarrosa, Gutierrezia sarothrae, Haplopappus sp., Helenium sp., H. autumnale, H. lacinia- tum, H. latifolia, H. nudiflorum, H. tenuifolium, Heterotheca sub- axillaris, Helianthus sp., H. annuus, H. grosse-serratus, H. maxi- millianus, H. petiolaris, H. salicifolius, H. tuberosus, Heliopsis he- lianthoides, Marrubium vulgare, Medicago sativa, Melilotus alba, M. officinalis, Monarda sp., M. citriodora, M. pectinata, M. punctata, Nepeta cataria, Opuntia sp., O. lindheimeri, O. macrorhiza, Parosela sp., Petalostemum sp., P. candidum, P. oligophyllum, P. purpureum, Prionopsis sp., P. ciliata, Psoralea floribunda, Ratibida sp., R. col- umnaris, R. pinnata, Rudbeckia sp., R. amplexicaulis, R. bicolor, R. hirta, R. laciniata, R. triloba, Salsola pestifer, Silphium sp., S. per- foliatum, S. speciosum, Solidago sp., S. canadensis, S. rigida, S. sero- tina, Tetragonotheca ludoviciana, Tetraneuris linearifolia, Trifolium repens, Verbena sp., V. officinalis, V. stricta, Verbesina encelioides, Vernonia sp. Melissodes (Eumelissodes) nivea Robertson Melissodes nivea Robertson, 1895, Trans. Amer. Ent. Soc., vol. 22, p. 127; 1897, Trans. Acad. Sci. St. Louis, vol. 7, p. 354; 1905, Trans. Amer. Ent. Soc., vol. 31, p. 368; Cockerell, 1907, Ann. Mag. Nat. Hist., ser. 7, vol. 20, p. 128; Robertson, 1928, Flowers and Insects, p. 8; Pearson, 1933, Ecol. Monogr., vol. 3, p. 381; Graenicher, 1935, Ann. Ent. Soc. Amer., vol. 28, p. 304; Brimley, 1938, Insects of North Carolina, p. 462. Melissodes nivea is closely related to M. coreopsis. The females of nivea can be distinguished from those of coreopsis primarily by the small but distinct punctures in the apical areas of terga 2 and 3 and by the short, white hairs in the same areas. The male is very similar to the male of M. agilis in the color of the labrum, mandi- bles and wing veins, but is like M. coreopsis in the extremely short first flagellar segments and in the shiny galeae. The pale pubes- cence and hairs of both sexes of nivea tend to be white, rather than dull ochraceous as in coreopsis or rufescent as in agilis. Female. Measurements and ratios: N, 20; length, 9-12 mm.; width, 4.0-4.5 mm.; wing length, M = 2.84 + 0.095 mm.; hooks in hamulus, M = 11.30 + 0.128; flagellar segment 1/segment 2, Mi 1 70)== 0:019) Structure and color: Integumental color as in coreopsis except 456 THE UNIVERSITY SCIENCE BULLETIN as follows: wing veins dark red to reddish brown; tergum | with apical third occasionally rufescent, apical margin narrowly hyaline. Structure and sculpturing as in coreopsis except as follows: clypeal punctures round, separated by less than half a puncture width, surface shiny with sparse striations, apicomedian longi- tudinal carina usually present; supraclypeal area usually dulled by reticular shagreening, with scattered coarse punctures; maxillary palpal ratio about 2.5:2.5:2.0:1.0, last segment sometimes slightly shorter. Mesoscutal punctures round, posteromedially deep, larger, separated mostly by one to two puncture widths, anteriorly and laterally slightly smaller, shallow, separated mostly by half a punc- ture width or less; scutellar punctures similar to posteromedial mesoscutal punctures but more crowded; surfaces of mesoscutum and scutellum shiny, often slightly dulled by fine reticular sha- greening; mesepisternal punctures as large as posteromedian meso- scutal punctures; separated by half a puncture width or less, surface shiny, unshagreened or extremely delicately and irregularly so. Metasomal tergum 1 with basal area punctures larger and more crowded, apical impunctate area moderately shiny to shiny, with delicate reticulotransverse shagreening; tergum 2 with interband zone punctures small, deep, relatively regularly spaced, separated by half to one puncture width, apical area with small distinct punc- tures two to three times as wide as bases of appressed hairs arising from them, surface (especially of apical area) shiny to moderately shiny; tergum 3 similar to 2 but with interband zone and apical area punctures more crowded. Pygidial plate V-shaped with sides diverging posteriorly and apex well rounded, longer than basal breadth. Hair: Head white with abundant long brown hairs on vertex. Thorax white except mesoscutum with dark brown posteromedian patch often extending forward to a transverse line at anterior mar- gins of tegulae and almost reaching tegulae laterally (darkest in eastern specimens ), scutellar hairs dark brown except white fringe, tegulae dark brown, and pale hairs of mesoscutum often pale cine- reous. Metasoma as in coreopsis except as follows: tergum 2 with distal pale band white, as long as or longer than apical area me- dially, interband zone hairs mostly dark brown in specimens from Atlantic states and mostly white in specimens from prairie states, apical area with short, relatively simple, appressed to subappressed, white hairs (occasionally a few brown medially in darkest forms); tergum 3 similar to 2 but apical area shorter and often with median hairs brown; tergum 4 with apical white band never interrupted BEES OF THE GENUS MELISSODES 457 medially nor fringed apically with brown; terga 5 and 6 with con- spicuous lateral white tufts. Legs as in coreopsis but inner sur- faces hind basitarsi more often reddish brown than darker. Male. Measurements and ratios: N, 20; length, 9-12 mm.; width, 3-4 mm.; wing length, M = 2.85 + 0.135 mm.; hooks in hamulus, M = 10.60 + 0.328; flagellar segment 2/segment 1, M = 10.54 + 0.183. Structure and color: Integumental color as in M. agilis except as follows: labral pale spot occasionally reduced but never absent, basal mandibular yellow maculae often reduced in size and rarely absent; tergal apices hyaline and colorless. Structure as in coreopsis except as follows: first flagellar segment with minimum length equal to one-tenth or less of maximum length of second segment, penultimate segment slightly longer than three times minimum width, segments 5-10 slightly crenulate when viewed from below, flagallum in repose just reaching pterostigma; maxillary palpal segments in ratio of about 3.5:4.5:3.0:1.0, last seg- ment often slightly longer. Sculpturing as in female except as fol- lows: mesoscutal punctures smaller, often more crowded postero- medially; tergum 1 medially with basal four-fifths punctate; terga 2 and 3 with apical areas not distinctly punctate, shiny. Terminalia as in agilis, but sternum 8 with apicoventral tubercle pointed, not bidentate, and slightly surpassing apical margin in apicomedian emargination. Hair: White except inner surfaces tarsi yellow and occasionally dorsum of thorax slightly cinereous. Metasomal tergum 2 with distal pale band as wide as apical area medially or wider; terga 2 and 3 with apical areas with abundant, suberect to subappressed, relatively simple, white hairs; tergum 1 with pubescent not form- ing thick apical band hiding margin of tergum. Bionomics, Of 58 specimens representing 29 collections which have floral data attached, 54 specimens (28 collections ) were taken on some species of Compositae. The single collection not from a composite consists of four females taken on Gerardia sp. (Scro- phulariaceae). M. nivea can, accordingly, be considered as an oligolege of composites and has some preference for species of the genera Solidago, Aster, and Liatris in that order. Type Material. Lectotype female, here designated, of nivea, collected by Charles A. Robertson (Coll. No, 3205) at Carlinville, Illinois, September 8, 1886, on Solidago lanceolata is in the collec- tion of the Illinois Natural History Survey at Urbana. The lecto- 458 THe UNIVERSITY SCIENCE BULLETIN allotype male of nivea, here designated, collected by Robertson (Coll. No. 17648) at Carlinville, Illinois, August 21, 1895, on Lepachys pinnata is also in the Tlinois Natural History Survey col- lection, Distribution. M. nivea occurs from Long Island, New York, south to North Carolina and Alabama, and west to Minnesota, Kansas, Arkansas and Mississippi (Fig. 17). It has been collected from July 16 to October 14, but chiefly in September. In addition to the type material, a total of 136 females and 91 males have been examined from the localities listed below. This list includes local- ities reported in the literature. ALABAMA: Mobile. Arkansas: Fort Smith (25 miles N.), Ouach- ita Mts.; Knob Hill Reservation, Ouachita Mts. Duisrricr or Co- LuMBIA: Washington. ILLINots: Carbondale; Carlinville; Macoupin Fic. 17. Map showing the known distributions of M. (Eumelissodes) nivea Robertson and M. (E.) montana Cresson. BEES OF THE GENUS MELISSODES 459 Co.; Manito. Inpr1ana: Bluffton; Gibson Co.; Rush Branch. Kan- sas: Baldwin; Cherryvale (2 miles S.); Douglas Co.; Garnett; Hutchinson; Logan Co.; Reece; Riley Co. Maryianp: Bethesda; Cabin John; Glen Echo; Indian Head. Mrtnnesota: Ortonville. Mississippi: Camp Shelby (near Hattiesburg). Mussourt: Bran- son; Gilmore; Ozark Lake. Nepraska: Lincoln; Malcolm; West Point. New Jersey: Asbury Park; Jamesburg; Lakehurst; Lake- wood. New York: Astoria, Long Island. Norru Caroura: Black Mts. (valley of); Burgaw; Crabtree Meadows Park; Greensboro; Oxford; Raleigh; Swannanoa. Onto: Columbus. PENNSYLVANIA: Darby. TENNESSEE: Maury Co. Vrecinia: Arlington; Barcroft; Camp Peary; Falls Church; Fort Humphreys; Four-mile Run (near mouth of); Glen Carlyn; Mathias Point; Vienna; Virginia Beach. Wisconsin: Milwaukee. Flower Records. Aster sp., A. ericoides, A. sagittifolius, Bidens laevis, Boltonia asteroides, Chrysopsis mariana, Gerardia sp., He- lenium sp., Helianthus sp., H. annuus, H. atrorubens, Lacinaria sp., Lepachys pinnata, Liatris graminifolia, Prionopsis ciliata, Solidago sp., S. canadensis, S. lanceolata, S. rigida, S. serotina, Vernonia sp., V. glauca. Melissodes (Eumelissodes) pilleata, n. sp. This is a medium-sized, black and white bee related both to M. coreopsis and to M. rustica. The female resembles that of coreopsis in sculpturing and in vestiture coloration but can be distinguished by the black fringe of hairs at the apex of tergum 4, the lack of pale lateral tufts on terga 5 and 6, and the broad pygidial plate. The female is readily confused with that of several closely related species whose descriptions follow below. The male is like rustica in having the apical tergal areas piceous, but is like agilis in having pale spots at the bases of the mandibles and on the labrum, and is like coreopsis in the short first flagellar segment and the shiny galeae. Both sexes have the wing membranes slightly infumate. Female. Measurements and ratios: N, 20; length, 10-14 mm.; width, 3.5-4.5 mm.; wing length, M = 3.27 + 0.089 mm.; hooks in hamulus, M = 12.60 + 0.148; flagellar segment 1/segment 2, M = 1.90 + 0.002. Structure and color: Integument black except as follows: apica! half of mandible, lower surfaces flagellar segments 3-10, and disti- tarsi rufescent; eyes grayish blue to dark gray; wing membranes somewhat infumate with slight violaceous reflections, veins dark 460 THE UNIvERSITY SCIENCE BULLETIN brown to black; tegulae piceous; tibial spurs yellow to red; tergum 1 with extremely narrow apical margin hyaline or testaceous. Structure and sculpturing as in coreopsis except as follows: clypeus slightly protuberant, oculoclypeal distance equals 0.50 to (0.75 times minimum diameter first flagellar segment, with regular round punctures separated mostly by half a puncture width, smaller and crowded anteriorly, without distinct longitudinal carina, sur- face moderately shiny, with sparse but distinct striations, supraclyp- eal area usually with several large deep punctures, surface shiny or slightly dulled by sparse reticular shagreening; galeae unsha- greened above except at tips; maxillary palpal ratio about 2.25:2.50: 2.25:1.00, a small fifth segment often present; metasomal tergum 1 with basal three-fifths with relatively large shallow punctures separated mostly by half to one puncture width; tergum 2 with basal area punctures separated mostly by half a puncture width; terga 2 and 3 with apical areas with minute scattered punctures no broader than twice diameter of hairs arising from them; meta- somal terga shagreened as in coreopsis, but usually shinier; pygidial plate broadly V-shaped, rounded apex, about as broad at base as median length. Hair: As in M. nivea except as follows: more abundant black hairs on vertex of head; mesoscutal dark patch extends forward beyond a transverse line at anterior margins of tegulae and to within one or two hairs of tegulae laterally; scutellum dark brown or black except for peripheral one or two hairs; tergum 2 with distal pale band usually narrower than apical area and often interrupted medially; terga 2 and 3 with apical areas with sparse, simple, ap- pressed, dark brown to black hairs; terga 5 and 6 without lateral pale tufts; tergum 4 with apical fringe of dark brown hairs at least in median third; legs with fore and middle tarsi brown, outer surfaces of fore and middle tibiae brown distally, inner surfaces hind basitarsi dark brown to black, and scopal hairs pale ochraceous. Male. Measurements and ratios: N, 15; length, 10-12 mm.; width, 3-4 mm.; wing length, M = 3.12 + 0.154 mm.; hooks in hamulus, M = 11.20 + 0.175; flagellar segment 2/segment 1, M = 10.05 + 0.699. Structure and color: Integument black except as follows: Clyp- eus pale yellow to cream-colored with testaceous apical margin; labrum white with dark brown margin; mandibles with pale basal maculae similar in color to clypeus; eyes grayish blue to gray; wing membranes slightly infumate, veins dark brown; tegulae piceous; BEES OF THE GENUS MELISSODES 461 flagellar segments 2 to 11 rufescent below; distitarsi rufescent; ex- tremely narrow apical margin of tergum | hyaline or testaceous; terga 2-5 with apical areas piceous. Structure as in coreopsis except as follows: first flagellar segment with minimum length equal to one-eighth or slightly less of maxi- mum length second segment, third segment distinctly longer than three times minimum diameter, not crenulate, in repose surpassing pterostigma and even marginal cell; maxillary palpal ratio about 3.00:3.33:2.66:1.00, minute fifth segment often present. Sculptur- ing as in female except as follows: posteromedian mesoscutal punctures more crowded; tergum 1] with basal four-fifths punctate, punctures become progressively smaller and sparser as they ap- proach narrow apical impunctate area; terga 2 and 3 with inter- band zone punctures smaller and sparser, apical areas virtually impunctate, surfaces shiny to moderately so. Pygidial plate slightly longer than broad. Terminalia as in agilis but sternum 8 with ventral tubercle pointed, not bidentate, and weak, and gonostyli with few hairs basally. Hair: White except as follows: vertex of head dark brown; dark brown mesoscutal patch extends forward to a transverse line at anterior margins of tegulae and laterally to within 5 or 6 hair-rows of tegulae; scutellum dark brown except peripherally. Metasomal vestiture as in female except as follows: tergum 1 with apical third with progressively shorter, relatively simple, suberect to subap- pressed, dark brown hairs; terga 2 and 3 with distal pale bands somewhat narrower, rarely interrupted medially; tergum 4 similar to 3 but apical area reduced to narrow fringe of brown; tergum 5 with distal pale band reaching apex medially; terga 6 and 7 with dark brown hairs. Legs white except inner surfaces tarsi and hind tibiae yellow. Type Material. The holotype male collected by T. B. Mitchell at Southern Pines, North Carolina, September 23, 1950, on Kuhnistera sp. is in the collection of T. B. Mitchell, North Carolina State Col- lege, Raleigh. The allotype female from Southern Pines, North Carolina, September 16, 1918, on Gerardia flava is in the collection of the American Museum of Natural History, New York City. Eighteen female and twelve male paratypes from North Carolina and collected by T. B. Mitchell are as follows: Aberdeen: 1 fe- male and 1 male, September 10, 1923; 1 female and 1 male without floral data and 2 females on Gerardia sp., September 26, 1923; 1 fe- male, October 15, 1957. Raleigh: 3 males in August (no further 462. THE UNIVERSITY SCIENCE BULLETIN collection data). Southern Pines: 2 females on Gerardia sp. and 1 female on Liatris sp., September 26, 1923; 1 male on Kuhnistera pinnata, September 15, 1949; 2 males on Kuhnistera sp., September 10, 1950; 1 female on Chrysopsis sp., September 23, 1950; 3 females without floral data, 1 female on Aster sp., and 1 female on Kuhnis- tera sp., September 30, 1951; 1 male, September 13, 1952; 1 female without floral data and 1 female on Kuhnistera sp., September 19, 1953; 2 females on Chrysopsis sp., October 15, 1957. Two female and three male paratypes from Southern Pines, North Carolina, were collected as follows: 2 males on Aster sp., August 20, 1918; 1 male, August 29, 1918; 2 females, September 13, 1918. Paratypes are in the collections of T. B. Mitchell, the American Museum of Natural History, the Snow Entomological Museum of the Univer- sity of Kansas, Lawrence, and in the author’s collection (Fig. 16). Melissodes (Eumelissodes) confusa Cresson Melissodes confusa Cresson, 1878, Proc. Acad. Sci. Philadelphia, vol. 30, p. 205; Cockerell, 1897, New Mexico Coll. Agr. and Mech. Arts, Bull. No. 24, pp. 20, 24; Birkman, 1899, Ent. News, vol. 12, p. 43; Bridwell, 1899, Trans. Kansas Acad. Sci., vol. 16, p. 211; Viereck, 1902 , Trans. Amer. Ent. Soc., vol. 29, p. 46; Cockerell, 1906, Trans. Amer. Ent. ’Soc., vol. 32, pp. 82, 92: 1906, Trans. Amer. Ent. Soc., vol. 32, p. 309; 1906, Bull. Amer. Mus. Nat. Hist., vol. 22, pp. 443, 454; Snow, 1906, ae Kansas Acad. Sci., vol. 20, p. 137; Cockerell, 1910, Psyche, vol. 17, p. 246; 1911, Sea Ent., p. 43, p. 33; Cresson, 1916, Mem. Amer. Ent. Soc., Vol. 1, p. 116; Rau, 1922, Trans. Acad. Sci. St. Louis, vol. 24, p. 34; Cockerell, 1933, Ann. Ent. Soc. Amer. vol. 26, p. 44; oe Knowlton and Bailey, 1950, Utah St. Agric. Coll. Mimeo. Ser. No. ‘aut Melissodes ruidosensis eat 1896, Entomologist, vol. 29, p. 305; 1898, Bull. Sci. Lab. Denison Univ. , vol. 11, p. 66; 1898, Bull. Univ. New Mexico, vol. 1, p. 66; 1901, Ent. News, vol. 12, p. 48; 1901, Ann. Mag. Nat. Hist., ser. 7, vol. 7, p. 180; 1902, Amer. Nat., vol. 36, p. 810; 1903, Ann. Mag. Nat. Hist., ser. 7, vol. 12, p. 450. Melissodes tenuitarsis Gackerell! 1905, Psyche, vol. 12, p. 99 (new synonymy ); 1906, Trans. Amer. Ent. Soc., vol. 32. p. iG: Snow, 1906, Trans. Kansas Acad. Sci., vol. 20, p. 137. Melissodes civica Cockerell, 1910, Ann. Mag. Nat. Hist., ser. 8, vol. 5, p. 258 (new synonymy ). Melissodes atraticornis Cockerell, 1934, Amer. Mus. Nov. No. 697, p. 9 (new synonymy ). This species is highly variable in both sexes, a fact which makes it difficult to identify and has contributed to the synonymy. It is perhaps most closely related to M. coreopsis, but also shows some structural affinity to M. grindeliae and M. rustica. The female of confusa is similar to that of coreopsis but is darker in color as fol- lows: terga 4 and 5 usually without pale lateral hair tufts, sternal hairs usually dark brown, often with lower and anterior surfaces of mesepisterna with brown hairs. In both sexes the antennae are often wholly black, although this is not so frequent in males as in BEES OF THE GENUS MELISSODES 463 females. The female can be distinguished from that of grindeliae by the punctation of the basal area of tergum 2 as described below. The male has the short first flagellar segment of coreopsis, although often slightly longer, but has shorter antennae as a whole. The male clypeus varies in color from entirely yellow except the testaceous apical margin to almost entirely black. The basal area punctures of tergum 2 are sparse as in the female. Female. Measurements and ratios: N, 20; length, 11-13 mm.; width, 3.5-5.0 mm.; wing length, M = 3.54 + 0.211 mm.; hooks in hamulus, M = 12.70 + 0.219; flagellar segment 1/segment 2, M = 1.86 + 0.013. Structure and color: Integument as in coreopsis except as fol- lows: distitarsi black to dark red; eyes gray to dark gray; flagellar segments 3 to 10 dark reddish brown to black below, second segment entirely dark; wing membranes somewhat infumate, brownish yel- low; veins black to dark brown; tibial spurs yellowish to red. Sculpturing and structure of head and thorax as in coreopsis ex- cept as follows: clypeus slightly protruding forward beyond eyes, oculoclypeal distance half to three-fourths minimum width of first flagellar segment, punctures slightly more coarse; supraclypeal area usually with surface shiny, unshagreened; maxillary palpal ratio about 3.0:2.7:2.0:1.0.. Metasomal tergum with basal three-fifths or slightly less with shallow, medium-sized punctures separated mostly by half to two puncture widths, apical zone extended to form anterolateral impunctate lobes; tergum 2 with basal area punctures round, small, deep, separated by one to two puncture widths or slightly more, surface unshagreened or with delicate reticular shagreening, apical area with small punctures about twice diameter of hairs arising from them, surface dulled by reticulotransverse shagreening, but moderately shiny to shiny. Pygidial plate broadly V-shaped with rounded apex, length subequal to basal width to slightly longer. Hair: Head as in coreopsis. Thorax as in coreopsis except as follows: pale hairs ochraceous above, white to pale ochre laterally; mesepisterna with anterior, ventral and lower lateral surfaces usually dark brown; posteromedian dark mesoscutal patch twice as large as scutellar dark patch or larger, usually almost reaching tegulae later- ally and extending forwards beyond a transverse line at anterior margins of tegulae; tegulae with dark hairs. Metasomal tergum 1 with basal area white to ochraceous, often quite yellow, apical area with minute, closely appressed, brown hairs at least 464 THE UNIVERSITY SCIENCE BULLETIN basally; tergum 2 with distal pale pubescent band pale ochraceous to yellowish, usually reaching apex of tergum laterally, shorter medi- ally but rarely interrupted, interband zone with abundant, erect to suberect, dark brown hairs, apical area with abundant, suberect to subappressed, dark brown, relatively simple hairs in basal two- thirds; tergum 3 as tergum 2 but basal area dark brown, distal pale band longer and reaching apex in lateral thirds or more; tergum 4 as in coreopsis; terga 5 and 6 without lateral pale tufts; sterna reddish brown to dark brown. Legs ochraceous except as follows: distitarsi, fore and middle basitarsi brown or largely so; middle basitarsi with inner surface dark reddish; hind basitarsi with inner surfaces red- dish brown to black; scopal hairs often yellowish; basitibial plates, outer apical surfaces fore and middle tibiae brown; hind tibiae with inner surfaces yellow to dark red. Male. Measurements and ratios: N, 20; length, 9-12 mm.; width, 2.5-4.0 mm.; wing length, M = 3.51 + 0.156 mm.; hooks in hamulus, M = 12.10 + 0.169; flagellar segment 2/segment 1, M=6.08 + 0.163. Structure and color: Integumental color as in coreopsis except as follows: clypeus varies from yellow with testaceous apical mar- gin and dark maculae at tentorial pits to entirely black, most often yellow with dark brown apical margin and infuscated along poste- rior margin between and slightly beyond tentorial pits; labrum and mandibles black; eyes dark gray to yellowish green; flagellum varies from red below and dark brown above with first segment entirely dark to entire flagellum dark brown or black, most often with first segment, base of second segment, tip of last segment and upper surfaces dark brown and red below; wing membranes slightly in- fumate, yellowish; veins dark brown to black; metasomal terga with apical areas hyaline, yellowish brown to yellow, not rufescent bas- ally; distitarsi rufescent; tibial spurs yellow. Clypeus much as in female; first flagellar segment with minimum length equal to almost half maximum length and equal to about one-eighth maximum length second segment, penultimate segment one-third as wide as long or slightly broader, just reaching ptero- stigma or slightly less in repose, flagellum somewhat crenulate near apex in lateral view (involving penultimate three to five segments ); maxillary palpal ratio about 2.7:1.7:2.0:1.0. Sculpturing as in fe- male except as follows: mesoscutal punctures often more crowded; metasomal tergum 1 with basal four-fifths punctate; terga 2 and 8 with interband zone punctures more distinct and often slightly BEES OF THE GENUS MELISSODES 465 larger, apical area punctures minute or absent; tergum 2 with basal area punctures minute, separated by two to four puncture widths and mostly by three or four widths; sterna shiny; reticular sha- greening coarse, often absent medially. Terminalia as in M. agilis, but sternum 8 with abundant apical hairs and with apicomedial tuft of hairs just above ventral tubercle, tubercle acute, not biden- tate. Hair: Head and thorax pale to dark ochraceous, paler laterally, head often with brown on vertex, mesoscutum often with dark brown patch posteromedially, scutellum usually with at least a few brown hairs, tegulae usually with brown. Metasomal tergum 1] pale to dark ochraceous except two or three subapical rows of shorter, relatively simple, dark brown hairs; tergum 2 pale ochra- ceous to white basally, distal band ochraceous, arched medially but usually not interrupted, basal and distal bands connected laterally, interband zone hairs suberect to erect, yellow to dark brown, apical area hairs subappressed to suberect, dark brown to ochraceous (usually at least apical few rows dark); terga 3 and 4 similar but basal tomentum brown, interband zone hairs more often and mostly dark brown, apical areas progressively shorter; tergum 5 like tergum 4 but distal pale band apical; terga 6 and 7 entirely dark brown to dark medially and ochraceous to light brown laterally. Legs ochraceous except yellowish orange on inner surfaces tarsi and basitibial plates brown. Bionomics. The floral data for this species are sparse, but they indicate the usual oligolecty of the subgenus Eumelissodes, that is, a preference for flowers of the family Compositae as pollen sources. A wide variety of composites are visited, however, and it is difficult to state any preference on the basis of the present data. However, M. confusa is unsual in that it has not yet been collected visiting flowers of Helianthus. Also, the author has collected males on more than one occasion (especially in Mexico) sleeping in the flowers of Argemone (Papaveraceae ), but no females have been taken from this flower. Table VIII below summarizes the floral data. Type Material. Lectotype female and lectoallotype male of confusa, both from Colorado, are in the Academy of Sciences of Philadelphia, Pennsylvania. Cockerell apparently did not designate a holotype for his M. ruidosensis. However, I have seen several specimens labeled ruidosensis in Cockerell’s handwriting and Lutz and Cockerell synonymize ruidosensis with confusa in their cata- logue (1920). Two males labeled as cotypes of ruidosensis were 466 Tue UNiversiry SCIENCE BULLETIN TABLE VIII. Summary of Floral Records for Melissodes confusa. Plant Data Records of M. confusa © fe) ro) = AMILY 2 S 38 88 3 : 38 Eg | E&| | 5S | be | be | ge | Z Z Zi Za Zi e Compositae: Cirsium spp. 1 2 15 19 27 46 Helenium spp. hema 3 8),| 2206 YzOeaeen Grindelia sp. heal 1 9 11 1 12 Other Genera 14 19 27 39 17 56 Leguminosae lst 4 8 3 8 11 Labiatae a 3 5 BUGS ay 8 | Geraniaceae hal if 4 6 2 8 Other Families (7) laid 7 9 5. Leal 18 Totals 32 40 83 109 | 141 250 | | found in the collection of the U. S. National Museum, Washing- ton, D. C. I hereby designate one of these (Cotype No. 3361) as the lectotype male of ruidosensis. This male also bears the label “Wooton 111.” The holotype male of tenuitarsis from Oak Creek Canyon, Arizona, collected by F. H. Snow, is in the Snow Entomo- logical Museum of the University of Kansas, Lawrence. The fe- male holotype of civica, collected by Farrar at Mexico City, Mexico, is in the Zoologische Museum der Humboldt Universitit, East Ber- lin, Germany. The holotype male of atraticornis from Pingree Park, Colorado, collected by Louise Ireland on August 14, 1933, is in the Academy of Sciences of Philadelphia, Pennsylvania. Distribution. M. confusa ranges from southern Canada to cen- tral Mexico and from California to Wyoming and Minnesota (Fig. 18). It has been taken from June 23 to September 21, but chiefly during July and early August in the United States, and as late as December 11 in central Mexico. oe 623 MOOTLEL acho. Gc eee 645 355 montana. .% 22.165 ee eee 469 648 nigtacauda’ 2.7 h. 4... eee 327 505 MVEA ui. a nod eae 455 445 ochraéa «.¢o«ckiae> 5 eee 428 BEES OF THE GENUS MELISSODES PAGE BOCLOIEISH See cath as Sek ne eek: 419 SG Aer ee ee Pha. ile ae he 419 SaGicinctay «4.26.24. ns ace Aes 473 pallidisipnata.!.:. 3.22%. 5.4.0... 529 Waucipunctay eye ee 650 jena EY ee a a 631 IECOSC HA Me et NY ooo ante 579 *pennsylvanica, :... 0.6. ..0%4- 654 DERMISA Mc eke WS oe Node 409 SVELDIExda et chao cds Yo cl cete 511 SPEEDICKANS) ee se. es hos cs 498 PEEDOMEAN fee) Sl eusae. Soa es 443 PeLsumilise s. Ae ne, f ek ee 498 personatella ys. fo ace. a hoe: 635 DCX AME REN Pt mis Sy onc tirene on: 621 ephiladelphica +{). 2... .o a cwre ss 654 llleataMee ann ee sbe pret cal, 459 PUI OSA ERI ts once tole 308 [Dae] ULE 529 DULEC ye A ee re 496 pullatellayen. 4). Sead ors... 620 TMMCCTIS, eng hese Ais ts 437 LOWUSEIOL Gri idle ate an to-0d. saree Ss 521 TULA PESHMY, yids opt cee Sons Sa 590 SnMiGOSENSIS) <2 44h oss oe ed +s 452 TMISt Ca ume MN yey: Pie eer eee 543 SCaMGHATUIITG ys ae ary eee 469 saponellus *semiagilis semilupina *semitristis *senilis *simillima snowii stearnsi subagilis subillata *tenuitarsis terminata tincta tribas trinodis tristis tuckeri utahensis *vanduzeei velutina verbesinarum vernalis vernoniae *vernoniana wheeleri THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vou. XLIT] DECEMBER 29, 1961 [No. 6 The Crane Fly Genus Dolichopeza in North America BY GrorGE W. BYERS Department of Entomology, The University of Kansas, Lawrence, Kansas * TABLE OF CONTENTS PAGE TELS ROG IGLELONN Sere ee ay eee ee Cn SG epee in sear ie ee eR wee Aha er rae a 666 GAMO ECE MENTS Menee AW eee iors aa ree seeo cde ie naciioest Re ae No ola 668 NE MUETINISE DOU CHONEZA tra bse Ao els clea abs Co i gees Nets ein bees dae 669 Norn American /IONCRODEZ 28. eel bonhd on ees a kee 670 PLISEOLICA Mee Vi © WMP pea eh er ae LE eee eae cn ene eat 672 LeRETISMSEUCICU ote my iis or bee a oo, Hale, crag st cay Ries by! o eS 675 Collection and preparation of material for study ....................- 676 WASTE SMO LEE SEUIClyan Seana Opie sce MN ete me ehhh ys aoe Nowe sea Ravens af beh 680 Admts——external “morphology. (oi. ache be elo ns ee ed ew 682 Amis——imnternal ymorphology,y (5054.06 Moe dukds wiiekoa yak siti ate dilate slo ys 693 PRPEASHECIHIC May ATIATLOM a et Gee ees kta sary tek wag daa tony ed Dyers AC Oe eat spares 703 esta tainalleai SOR tt ie eos ne Ras hts Wiese ads ts As ent oe oN 711 ianinemnehiaViOGpER = Abe. Phe ck ate el ts Nena eR ol re ARMM fe hg oy SE or 723 OVAOSiG OMe eI ce ek eo eae ge aie Lathe eeRDy og ABN ae apostate eho 727 JER EAS . tela eta A DRESS ge Ieee eee ee aire hve RRR dee nas cla cep ate 728 ARV AC MONDO LOC WANE. Abel Geese Went ONELD che Kee ea OA RIS Peale Hee ole ie AL 731 Tentative key to fourth instar larvae of North American Dolichopeza .... 742 aryac——natunalehistonya emir chia ey Waren Sakae Med oaa ects ee ee 745 BU PAE—IMOLWNOlO mys o.2 ee ee es phere eae AEA Shy Liat tres Heda Saud a> 766 Tentative key to pupae of North American Dolichopeza ............... 773 Me NEPAL MISCOFY). Saas stom te uh ee Cie a eames tyne Aes 6S. ea 776 Key to adult males of North American Dolichopeza .................. 782 Key to adult females of North American Dolichopeza ................. 784 SHC ENMSE DON ChONE ZO Hs: efit te) Mer Pe RAINE ARIES ee onc) 3 ocd ee he 786 DOUCROPEZE AM DOUCKHOPEZE) “AMENCONE yar nc. a ial dees els ne 786 Dolichopeza (Dolichopeza) borealis new species .................-+-- 795 SUP SCMISMOLOVELT. ok, ye AEA eRe cee Tae AST ke ste sas 796 Dolichopeza (Oropeza) australis new species .................------ 796 Dolichapezas(Oroveza)) Carolas Pict as 6 eek aes ARS Bs ks onl pt 800 DON CHOpe Za VOLOpen@)\ COTSAUS Ac. sab ea a ey te le ea Ao dons Bod wlenaieee 807 * This work was done primarily while the author was in the Division of Insects, Univer- sity of Michigan Museum of Zoology, Ann Arbor, Michigan. 23—5840 ( 665 ) 666 Tue UNtversiry SCIENCE BULLETIN PAGE Dolichopeza (Oropeza) johnsonella ...: .. 02.256. . + occa. oe nee 814 Dalichonezat(Oropeza) obscura’... 2. hoc ce ee ee ee 822 Dolichaneza(Oropeza) ‘polita polita oo... 6s... 2 on bso ee 1 ee 833 Dohtchopeza (Oropeza) ‘polita pratti £.2...¢. 02 3.20). eee 840 Dolichopeza (Oropeza) polita cornuta new subspecies................. 845 Doltchopeza: (Oropeza.): sayi ok oss oc es an ek ns eee 850 Dolichopeza (Oropeza) similis: ). 172-2592 2 i243. ss eee 858 Dolichopeza (Oropeza) subalbipes...020. 0. te 2s. . os eee 864 Dolichopeza (Oropeza) subvenosa, 0.22286) 2. ont os eee 873 Dolichopeza (@ropeza) tridenticalata.. 9.28 25.1) A a eee 879 Dolichopeza (Oropeza) venosa 2 odo ee 8 ees 3s be a eee 887 Dolichopeza (Oropeza) walleyi- ... 98.22.5020 os ob eee 894 Summary and-conclusions’ >, 4). io. cs os ee ss a ee 905 Meiterature’ Cited :.. 0262 gurewsee ete | ee en eee 920 INTRODUCTION A hundred years have passed since Charles Darwin, in Origin of Species, wrote: “As the species of the same genus usually have, though by no means invariably, much similarity in habits and con- stitution, and always in structure, the struggle will generally be more severe between them, if they come into competition with each other, than between the species of distinct genera.” During that century, repeated observations on various groups of organisms— vertebrate and invertebrate animals, as well as plants—have sup- ported Darwin's belief, and these observations have been sum- marized in statements known as Jordan’s Law and, more recently, Gause’s Principle. Closely related species and races will by reason of their common origin inhabit ranges in proximity to each other. Competition between these species in areas where their ranges overlap is calculated to result in the survival of one and destruction or displacement of the others. The consequences of such competi- tion are described in a statement of Jordan’s Law by Hubbeil (1936: 518-519): “In any group of allied forms, whether species or races, the most closely allied will be found not in the same ter- ritory and habitat, nor in widely separated regions, but either in adjacent areas separated by a barrier of some sort, or if in the same area then in different but not entirely unrelated or dissimilar eco- logical situations.” Because this expectation is so completely reasonable, instances of nonconformity with it are of unusual interest. Such an instance is found in the crane fly genus Dolichopeza, it having been the ex- perience of collectors that wherever they took one species of the genus they nearly always collected at least one other. In fact, at THe CRANE FLy GENus DOLICHOPEZA 667 one time, in a certain favorable habitat, I captured no less than six species of Dolichopeza in a single sweep of my net! As organisms usually conform to our well-founded generaliza- tions, one tends to regard the case of Dolichopeza with suspicion. If we have interpreted the genus correctly, its species must be reasonably closely related; the fact is that certain groups among North American Dolichopeza comprise pairs or larger aggregations of sibling species, the forms so nearly alike that they may be dis- tinguished only by microscopic details, in many cases. But it might be that the species are ecologically separated in some way as to preclude competition. Or, in view of the existence of groups of sympatric “sibling species,” perhaps many of the recognized forms are really not species at all but merely polymorphic pheno- types of one or a few species. The following study was undertaken partly to try to learn which of the described forms actually are species. This required exten- sive laboratory rearing, field observation and detailed study and interpretation of variation. It was the further purpose of this in- vestigation to find how the species are related to each other, struc- turally and ecologically. This involved some study of comparative anatomy of both adults and immature stages, and again detailed field observations. Lastly, an answer was sought to the question: if there are closely related species living together in the same habi- tat, as has appeared to be the situation, how is it that one has not eliminated the others through competition? Earlier North American workers on the Tipulidae have quite naturally concentrated their efforts on exploration of the fauna, which is still not completely known, even on the basis of the adult flies. Because of the relatively minor economic importance of crane flies, there has been no pressing need for studies of their bi- ology; accordingly, the few earlier detailed accounts of North American tipulids have dealt usually with only one species or a scattering of species within a large genus. The following study, in contrast, attempts to present a more coherent picture, bringing together all available data on every species of the genus Dolicho- peza known to occur in North America. In gathering these data, it soon became evident that nearly every aspect of the biology of Dolichopeza was either uncertain or un- known. Geographical ranges were only sketchily outlined, only fragmentary information was available on the immature stages of but a single species, nothing was known of the internal morphology 668 THE UNIversIry SCIENCE BULLETIN of either adults or immatures, and the habitats of these stages had not been specifically defined. This investigation therefore expanded into a search for whatever could be learned about the genus Doli- chopeza, with the idea that answers to the problems posed above might be found in diverse places. ACKNOWLEDGMENTS If I were to place a note of dedication in the front of this work, that dedication would be to the late Professor J. Speed Rogers, former Director of the University of Michigan Museum of Zoology, for it was Dr. Rogers who encouraged me to undertake this study and who got me interested in crane flies in the first place. As a teacher, he was always enthusiastically interested in the progress of his students and ever ready with encouragement. He placed at my disposal his entire accumulation of Dolichopeza, a collection larger than all others in the world combined, but what was more important he impressed me with the desirability—in fact, the need— of supplementing laboratory observations by the study of living crane flies in their natural environments. He was a cautious and careful worker, maybe even meticulous, and he would not like to see a conclusion reached unless every available shred of evidence had been closely examined. While I cannot measure up to his ideals, it is my hope that the work which follows would have pleased him. My sincerest thanks go to Professor Theodore H. Hubbell, now Director of the University of Michigan Museum of Zoology, for reading most of this paper in its early stages and offering many valuable suggestions, often in the form of thought-provoking ques- tions. His advice and encouragement have been of immeasurable help. Dr. Charles P. Alexander, distinguished student of the world fauna of Tipulidae, generously provided me with his collection rec- ords of Dolichopeza and allowed me to inspect, in his personal collection at Amherst, Massachusetts, the types and other specimens of North American and exotic species. I am indebted to Dr. Alex- ander for this invaluable aid and for his having constructed much of the foundation upon which this investigation is based. For permission to do intensive collecting in various state parks in Indiana and for the courteous co-operation shown by park per- sonnel, I wish to thank the Department of Conservation of the state of Indiana. Recognition is also due Dr. Irving Cantrall for per- Tue CRANE Fiy GENUS DOLICHOPEZA 669 mission to do concentrated collecting on the University of Michi- gan’s Edwin S. George Reserve. Data on distribution and intraspecific variation have been con- siderably increased through the generous co-operation of several individuals and institutions. Dr. Joseph Bequaert made available types and other material in the Museum of Comparative Zoology at Harvard University; the collection of the United States National Museum was loaned by Dr. Alan Stone, and Dr. J. R. Vockeroth sent specimens from the National Museum of Canada. Others who made valuable contributions include Dr. Edward L. Kessel of the California Academy of Sciences, Dr. Jean Laffoon of Iowa State College, Mr. Peder Nielsen of Silkeborg, Denmark, Dr. E. F. Cook of the University of Minnesota, Dr. F. N. Young of Indiana Uni- versity, Mr. J. A. Wilcox of the New York State Museum, Dr. W. V. Balduf of the University of Illinois; and the entomological staffs of the Illinois State Natural History Survey, the Academy of Natural Sciences of Philadelphia, the American Museum of Natural History, Pennsylvania State University, the Milwaukee Public Museum, and the University of Kansas, who generously allowed me to examine collections in those institutions during my visits there. I gratefully acknowledge the co-operation of Dr. William C. Steere, who made the more difficult determinations of bryophytes. Other materials incidentally associated with this research were identified by Dr. W. J. Gertsch of the American Museum of Natural History, Mr. Sherman Moore of Detroit, and Dr. C. F. W. Muese- beck, Dr. B. D. Burks, Mr. H. W. Capps and Miss L. M. Walkley of the United States National Museum. To the many others whose help made this work easier I offer my sincerest thanks. Permission to use portions of maps of North America from the Hall’s Outline Series, by Dr. Robert B. Hall, has been granted by the Department of Geography of the University of Michigan. THE GENUS DOLICHOPEZA Including nearly a hundred and seventy recognized forms in nine subgenera, the genus Dolichopeza comprises a varied and ex- tremely widespread assemblage of species, some of which are not markedly different from certain species of other tipuline genera. By reason of possible future changes in generic definition, species or even subgenera of the present-day genus Dolichopeza may even- tually be assigned to other genera. The recognized subgenera, however, are clearly related and, in general, seem to be natural 670 THe UNIVERSITY SCIENCE BULLETIN groups of species. Dolichopeza includes those tipuline crane-flies which have extremely long legs and feet, the wing vein R,,, ordi- narily absent or atrophied, the rostral nasus small or lacking, and the radial sector of the wing usually short and nearly traverse. Cer- tain of these characteristics, of course, are shared by other genera, and some are not true of all Dolichopeza. The major subgenera may be separated by characters set forth in a key by Alexander (1931b). They are: Dolichopeza Curtis, Nesopeza Alexander, Mitopeza Edwards, Oropeza Needham, Trichodolichopeza Alexan- der, and Megistomastix Alexander. In recent years, three additional subgenera, embracing only a few species, have been described by Alexander. These are: Afrodolichopeza, Hovapeza, and Sinoro- peza. Excepting South America and the Antarctic, all the major land masses of the world have some members of Dolichopeza in their insect faunas, and it may be that later exploration along the western highlands of South America will prove the genus to be also present there. The typical subgenus is the most widespread, represented by two species in North America, by four in Europe, five in Africa and Madagascar, eight in southern and eastern Asia, and _ thirty- five species in Australia and New Zealand. The only other subgenus bridging a major geographical gap is Oropeza, of which there are nine species known from eastern Asia (Japan, China, Korea, For- mosa, and India) and thirteen from North America. Sixty-four species of Nesopeza are recognized, of which all are Asiatic and nearly all concentrated in southeastern Asia. There are ten species of Trichodolichopeza known from east-central to south Africa, and eight species of Megistomastix, found only in the West Indies. Afrodolichopeza includes a half-dozen species from east-central Africa; Sinoropeza comprises two species from southeastern Asia; and Hovapeza is known from a single species from Madagascar. NORTH AMERICAN DOLICHOPEZA This study is limited to the species of the subgenera Dolichopeza and Oropeza that occur upon the mainland of North America. Al- though few of the hundreds of species of Tipulidae inhabiting this continent are likely to be confused with Dolichopeza, it would be well, at the outset, to distinguish the genus from its nearer relatives. In having the terminal segment of the maxillary palpus very much elongated, the antennae with twelve or thirteen segments, the vein Sc, normally absent, and the cubitus with a deflected angle at cross- vein m-cu, Dolichopeza is included in the subfamily Tipulinae and Tue CRANE Fiy GENUS DOLICHOPEZA 671 is excluded from the Limoniinae, some of which have a superficially similar general appearance. Among North American Tipulinae, only three genera have extremely long and slender legs, with the tarsus exceeding in length the femur and tibia together. Of these, Brachypremna and Megistocera, both of which are essentially southern in distribution, may at once be distinguished from Doli- chopeza by their having a nasus on the rostrum, the vein R, ,, pres- ent, and the antennae quite short, approximately the length of the head. The species of Tipula and Nephrotoma that resemble Doli- chopeza may likewise be separated by the presence of vein R,,, and, in Tipula, by the obliqueness of the radial sector. The wings of Oropeza, as of the Tipulinae generally, are char- acterized by the presence of the cell lst M, (Fig. 1), while this cell is not found in the two North American species of the subgenus Dolichopeza (Fig. 2). The cross-vein m-cu joins the media be- yond its fork in subgenus Oropeza but before the first fork of the media in Dolichopeza. Other lesser points of difference will be discussed later. Immature stages of both subgenera of Dolichopeza closely re- semble those of Tipula, although both larvae and pupae of Oropeza may be distinguished by the presence on the sides of the eighth abdominal segment of two blunt, conical lobes. I have not made any comparative study of the eggs of various tipuline genera. De- tailed comparisons of the larvae and pupae of Dolichopeza with those of its nearer relatives will be made in the sections on the morphology of those stages. There are at the present time sixteen species and one subspecies of North American Dolichopeza recognized in the literature. They are Dolichopeza ( Dolichopeza) americana Needham, D. ( Oropeza ) carolus Alexander, D. (O.) dakota Alexander, D. (O.) dorsalis (Johnson), D. (O.) dorsalis rogersi (Alexander), D. (O.) johnson- ella (Alexander), D. (O.) obscura (Johnson), D. (O.) polita (Johnson), D. (O.) pratti Alexander, D. (O.) sayi (Johnson), D. (O.) sessilis Alexander, D. (O.) similis (Johnson), D. (O.) subal- bipes (Johnson), D. (O.) subvenosa Alexander, D. (O.) tridenticu- lata Alexander, D. (O.) venosa (Johnson), and D. (O.) walleyi (Alexander ). Of these, thirteen appear to be valid species: americana, carolus, dorsalis, johnsonella, obscura, polita, sayi, similis, subalbipes, sub- venosa, tridenticulata, venosa and walleyi. 1 consider Dolichopeza pratti to be a subspecies of polita, dakota a synonym of walleyi, and 672 Tue University SCIENCE BULLETIN sessilis and dorsalis rogersi synonyms of dorsalis. Support for the opinions expressed here will be found in the treatments of the indi- vidual species involved. One new species in the subgenus Doli- chopeza and a new species and one additional race of polita in Oropeza are described and named in this revision. In addition to the venational distinction between the subgenera mentioned earlier, a further convenient division of the genus may be made on the basis of the gonapophyses of the male hypopygium (Fig. 3). As the male of the one new species of subgenus Dolicho- peza is unknown, D. americana is in a class by itself, in this regard. Five of the other species have the gonopophyses in the form of small knobs, densely covered at the tip by stout, black, curved spines. This group includes johnsonella, obscura, polita, subalbipes, and tridenticulata, and one new species; it will be referred to hereinafter as the obscura group. In the remaining species, the gonopophyses are generally flattened and _ blade-like, variously tipped, and without the abundant black spines of the obscura group. These constitute the sayi group, which includes carolus, dorsalis, sayi, similis, subvenosa, venosa and walleyi. Gonapophyses of representative species of these groups are compared in figures 4 through 7. The genus may also be subdivided into two major color group- ings: those species that are of a dusky brown coloration (ameri- cana, johnsonella, obscura, polita, and tridenticulata) and those which are chiefly yellowish or buffy and marked with darker spots and stripes, especially in the form of prescutal stripes and pleural spots on the thorax and dark annulations on the abdomen. The darkly colored species are most often found associated with deep shade, in either rock-gorge or wooded situations, while the paler forms are more characteristic of less intensely shaded, often sparsely sun-flecked vegetation. These circumstances, of course, invite speculation about the selective advantage of coloration in Dolichopeza. HISTORICAL REVIEW The genus Dolichopeza was established by John Curtis in 1825 to include the one known European species, which he named Dolichopeza sylvicola (Curtis, 1825: Plate 62.)* This species had already been described as Tipula albipes by Strém in 1768. * Curtis’ British Entomology is a_ series of colored plates, 770 in all, accompanied by ee eed Published in London, these plates were issued a few each month from 3 to 2 Tue CRANE FLY GENUS DOLICHOPEZA 673 7 Fic. 1. Wing of Dolichopeza (Oropeza) tridenticulata, showing cell lst Me or discal cell. Fic. 2. Wing of Dolichopeza (Dolichopeza) americana, showing position of m-cu cross-vein relative to first branch- ing of M. Fic. 3. Hypopygium of Dolichopeza (Oropeza) obscura, male, left lateral aspect; gon—gonapophysis. Fic. 4. Gonapophyses of Dolichopeza (Dolichopeza) americana, dorsal aspect. Fic. 5. Same as 4, left lateral aspect. Fic. 6. Gonapophyses of Dolichopeza (Oro- peza) obscura, postero-ventral aspect. Fic. 7. Gonapophyses of Doli- chopeza (Oropeza) sayi, postero-ventral aspect. 674 Tue UNIverRsITy SCIENCE BULLETIN Accordingly, the type species is Dolichopeza albipes (Strom). America’s great pioneer entomologist, Thomas Say, in 1823 de- scribed Tipula annulata (Say, 1823: 25), the description of which is now thought to pertain to a species of Dolichopeza. Unable to locate the specimens upon which Say based his description, I must assume any types were lost at the time most of the Say collection was destroyed by dermestid beetles, in Philadelphia, about 1840 (Weiss and Ziegler, 1931: 203-212. ) Actually, it is uncertain that what Say described was a Dolicho- peza; the description is quite brief and rather general in its wording. However, sometime between 1823 and 1828, specimens thought to be Tipula annulata Say reached C. R. W. Wiedemann in Austria, and in 1828 he noted that both male and female of this species were represented in his collection (Wiedemann, 1828:54). I have examined these specimens (now in the Naturhistorisches Museum in Vienna), which are two males and a female of Dolichopeza (Oro- peza) johnsonella and two males of a species of Tipula. That these were determined by Wiedemann (they are so labelled) suggests they had not been compared with any specimens in America but were merely identified by comparison with Say’s written description of Tipula annulata. Fifty years later, C. R. Osten Sacken (1878: 40) placed Tipula annulata provisionally in Dolichopeza, noting that “. . . the for- ceps of the male has a different structure, etc.” Seven specimens of Dolichopeza spp., labelled “Dolichopeza annulata Say?” (prob- ably in Osten Sacken’s handwriting) are to be found in a tray marked “Loew and Osten Sacken” in the Museum of Comparative Zoology at Harvard University. Osten Sacken’s catalogue indi- cates that specimens of what he called Dolichopeza annulata were located in that museum. It would appear that Osten Sacken, like Wiedemann, decided after reading Say’s description that certain North American tipulids in his collection were the species in ques- tion. It was Osten Sacken, however, who first critically compared North American Dolichopeza with Tipula. Dr. James G. Needham, in 1908, described a new species, Doli- chopeza americana, which he recognized as a true Dolichopeza, and also established the genus Oropeza to include a structurally similar American species in which the cell Ist M, is present in the wing. It seems that Needham, too, made an independent decision as to the identity of Say’s Tipula annulata, for after having collected cer- tain specimens which he so identified, he made Tipula annulata the type species of Oropeza. It was not long after Needham’s work THE CRANE FLY GENUS DOLICHOPEZA 675 that the actual diversity of North American Dolichopeza became known, through the investigations of Mr. Charles W. Johnson of Boston. In his “New and Little Known Tipulidae,” published in 1909, Johnson described seven new forms and redescribed Oropeza annulata, assigning it the name Oropeza sayi. (Say had not been aware that Linnaeus had, in 1758, given the name Tipula annulata to the crane fly now known as Limonia annulata.) Johnson recog- nized as new species: Oropeza similis, O. dorsalis, O. venosa, O. albipes, O. subalbipes, and O. obscura, and a form polita as a variety of obscura. The individual histories of these, as well as of the other species listed in this summary, are treated under their separate titles. All rearrangements and descriptions of species of North American Dolichopeza published since Johnson’s paper have been the work of Dr. Charles P. Alexander. Arranged chronologically, they are: 1922—Oropeza rogersi; 1930—Oropeza johnsonella; 1931—Oropeza walleyi described and Johnson’s variety polita raised to full species rank; 1931—the genera Oropeza Needham and Dolichopeza Curtis merged, together with other genera, into the genus Dolichopeza,* as subgenera (Alexander, 193lb: 269); 1931—Dolichopeza (O.) tridenticulata; 1940—Dolichopeza (O.) subvenosa; 1941—Dolicho- peza (O.) pratti and D. (O.) sessilis; 1942—Dolichopeza (O.) al- bipes renamed D. (O.) carolus, since, through his action of 1931, he had made Oropeza albipes Johnson a homonym of Dolichopeza albipes (Strom), of Europe; 1942—Dolichopeza (O.) rogersi re- duced to subspecific rank; and 1944—Dolichopeza (O.) dakota described. The biology of North American Dolichopeza has been given brief treatment by Alexander in “The Crane Flies of New York” (Part I, 1919; Part II, 1920); and the late Dr. J. Speed Rogers has made several additions to the knowledge of the biology of adult and immature stages in his studies of the Tipulidae of the Cumber- land Plateau (Rogers, 1930), of northern Florida (Rogers, 1933), and of the Edwin S. George Reserve, Michigan ( Rogers, 1942). MATERIALS STUDIED Over 11,300 adult and 600 later stage immature specimens of North American Dolichopeza have been examined in the course * Johnson (1910:708) had earlier indicated this relationship of the two groups by using the heading “Oropeza Needham (Dolichopeza Curtis)”? above a list of species of Oropeza. However, in a later paper (Johnson, 1925:32), he used the name Dolichopeza americana and the generic name Oropeza for several other species. 676 THE UNIVERSITY SCIENCE BULLETIN of this study. Adult flies were distributed among the various species as follows: americana—1016 similis—67 carolus—1150 subalbipes—530 dorsalis—285 subvenosa—138 johnsonella—425 tridenticulata—2617 obscura—2028 venosa—132 polita sspp.—2150 walleyi—558 sdyi—257 new spp.—3l Eggs of all species have been examined, and first instar larvae were obtained in all instances except subalbipes, subvenosa, tri- denticulata and the two new species. Later instar larvae were obtained for all species except carolus, johnsonella, subvenosa and the new species referred to above; and pupae were found for all but johnsonella and the new species. Of the total number of adults, about 2900 were in the Rogers collection at the University of Michigan, and somewhat over 600 were seen in collections from other individuals and _ institutions. More than 7800 adults and all but a dozen or so of the immature stages were collected during my field trips in the summers of 1949- 1953 and 1957-1958. The types of all species were examined, ex- cept americana and sayi, of which species all type material is lost. In the study of the natural history of Dolichopeza, many hun- dreds of samples of mosses and hepatics, representing perhaps 200 species, were collected and examined for larvae or pupae. Most of these contained neither, but abundant sampling was necessary, since all bryophytes were suspected habitats, especially at the out- set of the search. Many actual habitats certainly would have been overlooked had any other procedure been adopted. Incidental observations on and collections of plants and animals ecologically related to Dolichopeza were made. These observa- tions have been interpreted and the collections identified by various specialists whose contributions are mentioned separately later. COLLECTION AND PREPARATION OF MATERIAL FOR STUDY Species of Dolichopeza vary in their response to light but are generally negatively phototropic, seeking the deep shade beneath undercut rocks and banks or the green twilight below lush vegeta- tion of marsh borders and swampwoods. Unlike the more active and rapidly flying species of Tipula and Nephrotoma found in the same general habitats, Dolichopeza spends the daylight hours in shaded seclusion and must be hunted in somewhat the way a mam- malogist might search for bats in the daytime. THE CRANE FiLy GENUS DOLICHOPEZA 677 Aggregations of individuals numbering from a few to over 200 may be found in cool, shaded niches in the more favorable forest and ravine habitats. In such cases, many flies may be collected by a single sweep of the net, or the patient collector desiring mating pairs or undamaged specimens may take flies directly into the cyanide jar. Under such conditions, I have collected mating pairs into large, empty jars in order to obtain eggs or observe mating behavior. For flies resting among the roots of wind-thrown trees or in narrow crevices in rock, collection directly into cyanide vials may be the best method, although an alternative is to stir within the cranny with a branch, collecting the flies into the net as they move out. In collecting those species of Dolichopeza which are most fre- quently associated with vegetation of marsh borders and swamps (especially D. dorsalis and D. sayi), the usual “sweeping” tech- nique is not satisfactory, as it is likely to break the legs off these delicately built crane flies. Rapid trampling of the vegetation will often bring the flies up from their hiding places. In doing. this, one should look well ahead, especially for flies crossing small open spaces in the vegetation as they move ahead of the collector. Stand- ing quietly in one place, or moving very slowly, has also proved a profitable method for detecting the movements of these crane flies beneath the cover of plants, and so capturing them. Dolichopeza sayi, D. similis and D. dorsalis are seldom seen in concentrations and must be taken individually or a few at a time. Other swamp species may be found either dispersed as individuals or gathered into small groups, where some sort of darkened cavity occurs. Dolichopeza carolus, usually seen hanging from leafy vegetation in forested ravine habitats, is easily taken by an upward sweep of the net. To most entomologists, the words “crane flies” at once suggest the legless condition in which these insects are all too often found in collections. That the Tipulidae are fragile certainly cannot be denied, but proper handling will prevent undue loss of legs from most specimens. Correct preparation of specimens begins with their capture. Large cyanide jars are to be avoided; use of smaller vials (30x90 mm. shell vials, taped for reinforcement, are very good) prevents breakage of specimens because fewer are placed in one vial, and they have less room to shake around, during and after the killing process. Very shortly after a crane fly is killed in the cyanide vial, it becomes more fragile than when alive. It remains in this rigid but not brittle condition for two to three hours, after which there occurs a period of secondary relaxation, lasting some- what over an hour. Following this period, dehydration causes 678 THe UNIVERSITY SCIENCE BULLETIN increasing brittleness of the insects. To attempt to disentangle and further prepare specimens either during the earlier period cf rigidity or after dehydration has become advanced results in great loss of legs, especially in such slender-legged flies as Dolichopeza and many species of Limonia. During the period of secondary relaxa- tion, however, even the most fragile species may be processed with some speed and without much damage. The flies are placed in paper envelopes (a convenient size is 2 x 3% inches, No. 2 drug type), upon which the collection data have previously been marked. While still in the relaxed condition, three or four Dolichopeza, each with the long legs compactly arranged beneath the body, may be neatly tucked into such an envelope. Warm, dry heat or a calcium chloride desiccator may be used to dry the insects. Specimens prepared in this way arrive in the laboratory undamaged, ready to mount, and accompanied by field data. Most breakage of tipulids in collections is due to improper mount- ing. I have never seen a crane fly too large to be conveniently mounted on a point of heavy paper. Flies pressed in envelopes are arranged in one plane, hence are seldom broken during the mounting or labelling processes. Utilization of points cut from two-ply bristol-board for all mounting allows the use of uniformly strong pins (No. 4 are good) and thus prevents damage of speci- mens due both to attempts to pin through the insect’s body and to subsequent pin failure. Dried with the legs folded below the body, flies may be fastened to the point by the left side of the thorax, in such a way that only the one side of the thorax is obscured from view, while the legs receive maximum protection trom the point, the pin, and the labels. China cement thinned with ethyl acetate is a very satisfactory mounting material. In the preparation of labels, the guiding principle should be to make all essential information about the specimen immediately available to other workers who may wish to make use of it, be they in another part of the world or in a different generation or century. Adequate locality data consist of first a major geographical division (state, province, territory, or country, if of small size), then a more limiting designation, a county, district, or reference to a sizeable town, etc., such as is likely to be listed in the indexes to commonly accessible maps. Geographic co-ordinates should be used for little-known areas. References to any but the most widely- known physiographic features is inadequate, as these may not be mentioned in atlases or map indexes. Elevations should always be given for regions of high relief, and specific localities should be THe CRANE Fry GENUS DOLICHOPEZA 679 mentioned for areas where other environmental factors (vegetation, moisture, etc.) vary greatly. Exact dates are very important, es- pecially in data on groups such as the Tipulidae that are very sea- sonal in occurrence. Special material (reared specimens, those from which parts have been removed to microscope slides, etc. ) should be appropriately labelled. Confusion is certain to occur, sooner or later, if identified insects are not each labelled with the proper determination, the authority and the date. Collection of larvae and pupae is a tedious and time-consuming business, for the immature stages are usually diffusely distributed within their microhabitats. It is practically impossible to obtain eggs or first instar larvae by field collecting methods. About two months after the local peak of late-summer emergence of adults is a good time to gather larvae, if numbers are desired, for the larvae will have grown large enough to be readily seen but will not have been seriously reduced in numbers by predators or severe weather. I believe that predation on the young of the spring generation pro- gresses more rapidly and that the Dolichopeza population is thus more quickly reduced. If fewer individuals but higher percentages of emergence are desired, larvae or pupae should be gathered shortly before the local time for emergence under natural conditions. Most natural mor- tality will then have occurred, and, given reasonable care in the laboratory, nearly all larvae or pupae will go on to the adult stage. Some hints on the actual places to look for larvae may be found in the discussion of larval habitats. However, an observation of general application is that additional pupae and perhaps larvae are likely to be found in mosses from which pupal skins are pro- jecting, provided that these sites are discovered early in the season of emergence for the species concerned. Immature stages to be preserved are killed in hot water and stored in 80% alcohol. I have not found it necessary to pass speci- mens through increasing concentrations of alcohol before storage. Adults stored in alcohol tend to lose their color, although color pat- terns are often accentuated in the process. Soft structures, both internal and external, retain their shape and are more easily studied in alcoholic material. Large collections of flies may be stored in a small space, in alcohol, although extensive damage to individual specimens will eventually occur, during subsequent handling. Microscopic preparations used in this study were very simply made. Most materials were soaked in 80% alcohol or warm water until soft, or were first macerated to the desired degree in 7-10% 680 Tue UNrversiry SCIENCE BULLETIN potassium hydroxide solution and then rinsed in warm water, then transferred directly into polyvinyl alcohol mounting medium.* This substance has good optical qualities, is very inexpensive, and has a slight clearing action that makes the KOH treatment un- necessary in the case of smaller objects. A further advantage is that mounted specimens may be soaked off in hot water and re- mounted, in case slides become broken or mounts are unsatisfactory. Polyvinyl alcohol, or “PVA” as it is called, has been in use only about fifteen years, so its durability has not been proved.t METHODS OF STUDY Adults——As in any study of a primarily taxonomic nature, this work is based largely upon post-mortem examination of museum specimens, in this case dried and alcohol-preserved flies. However, familiarity with and understanding of a group of organisms come more from studying them in their natural environment. Following insects about to learn what they are up to may seem an idle occu- pation, in these times when laboratory biology is more the vogue, but particularly in studies involving several closely related, sym- patric species, such as is the case with Dolichopeza, field observa- tions lead to a better understanding of individual species and may provide the only answers to certain problems of inter-species rela- tionships. Large numbers of living flies were also caged in the laboratory for such purposes as obtaining eggs, observing mating behavior, noting approximate length of life, and other things that could not easily be studied in the natural habitats. Contraction or deflection of parts of the body in drying, or dis- tention in alcohol, cause inaccuracies and complications in both linear and proportional measurement. For purposes of general description, body length, as used hereinafter, means the straight-line distance from the most anterior part of the head (excluding anten- nae) to the tip of the abdomen, regardless of the position in which the specimen is preserved. Wing length, the most fixed and easily accessible measurement of adult Dolichopeza, is the straight-line distance between the point of attachment and the tip. Even wing length is of little use in describing species, for environmental factors may result in great differences in size within a species. Discussion of preserved adult crane-flies, therefore, will be concerned chiefly with structural characteristics. * Polyvinyl alcohol is a plastic prepared from PVA powder (I use DuPont ‘Elvanol,” grade 71-24): sift 20 grams of the powder into 88 cc. of cold water; heat at 85° F. in water bath, stirring constantly, until a viscous syrup forms; to 56 parts of this syrup, add oe eat of a mixture of equal amounts of lactic acid and liquid phenol, stirring often and well. + Dr. C. P. Alexander, who prefers Canada balsam, pointed out that, in the Baltic Amber, he has 30 or 40 million years’ proof that balsam is a good mounting medium! THe CRANE Fiy GENuS DOLICHOPEZA 681 Larvae and pupae.—lf larvae are taken in the late third or fourth instar, they will need little attention when brought into the labora- tory. Rearing dishes must become neither too warm nor too dry; and, while the minimum lethal temperature is not known for any species, larvae were lost in considerable numbers at times when the room in which they were kept became warmed to temperatures that are only comfortable to humans. It has been stated that larvae of Dolichopeza inhabit dry moss. It is, however, difficult to esti- mate the actual moisture content of such a habitat, and the fact is that the larvae were usually found unable to survive one night in a covered dish containing no apparent moisture. Rearing larvae from eggs is a difficult task, and my attempts to do this have not met with notable success. However, the follow- ing methods have been found workable. Adult females in breeding cages (wide-mouth, half-gallon canning jars were found adequate ) will oviposit on moist pads of paper (commercial cleansing tissues were used) and sometimes on the sides of the jar. At the end of the laying period, these pads may be removed and the eggs con- centrated upon a circular portion of the paper, such as will fit into a Syracuse dish, or the like. Eggs left in the jar may be removed by washing them down with water (at room temperature ) and col- lecting them in a pipette or dropper. Heavier than water, they will collect in the tip of the pipette and may be transferred to the moist paper pad without overly wetting it. When the eggs are about five days old, fragments of moss should be placed beneath the paper, and the larvae, upon hatching, will migrate to this moss. Excessive moisture in the dish is to be avoided, for the eggs will not mature if submerged, and the very young larvae drown readily in even small amounts of free water beneath the paper. No more water should be in the dish than will be retained by the paper, but the dish should be kept damp by the use of a cover and the occa- sional replacement of evaporated moisture. Larvae feeding under the paper may be observed through the bottom of the inverted dish, to which the pad will adhere by its dampness. As the larvae grow larger, increasing amounts of moss should be placed in the dish, but separated leaves and stems should still be used, in order that the activities of the larvae may be watched at any time and head capsules or cast skins may be recovered after each molt. When larvae have at last transformed into the fourth instar, moss clumps in their natural form may be used, for the last larval skin may be recovered from the pupal burrow, and the larvae are large enough to be located and observed easily. 682 Tue UNtversity SCIENCE BULLETIN After formation of the pupa, it becomes necessary to move speci- mens to a larger dish, so that there will be sufficient space for emer- gence of the adult. Unlike some other tipuline pupae, those of Dolichopeza are not very active and will usually remain in any artificial “burrow” that is arranged in the moss for them. It is, of course, possible to bring larvae through all instars and pupation in a large dish, such as is recommended for pupae, but observation is very difficult and recovery of molted skins is practically impos- sible under such conditions. In order to avoid overheating larvae and pupae during micro- scopic examination, I have used a device containing an ordinary 150-watt electric lamp, with a water-filled Florence flask in front of it, serving both to absorb heat and condense the light into a very intense spot. While larvae are seldom found feeding in the light in their natural surroundings, they rapidly accommodate to the in- creased illumination of the laboratory and may continue eating even while being observed in the bright light of the microscope lamp. For studying preserved immature stages, I have used a Stender dish about one-third filled with paraffin, into which were carved such shapes and sizes of holes as were necessary to hold the speci- mens still, under alcohol. It might be appropriate to say a few things about laboratory culture of mosses. To begin with, it is advisable to keep a covered terrarium stocked with food mosses at all times during the rearing period, and to take from such stock supplies whatever moss is needed to maintain adequate food in the rearing dishes. Isolated cushions of moss tend to become dehydrated and overgrown with molds. When larger quantities of moss are grown together, normal nutrient substrate may be retained and the mosses grow more luxuriantly, while molds seem less able to gain a foothold. Varying the depth of damp sand beneath the terrarium mosses provides a moisture gradient, and mosses may be placed according to their natural water requirements. Mosses which begin to mold should be removed from the terrarium. It has been observed that molds are less likely to occur if the terrarium is kept in a cool, though lighted situation. ADULTS—EXTERNAL MORPHOLOGY General description—Smaller than most of the Tipulinae and larger than the average of genera of the Limoniinae, Dolichopeza is medium sized, as crane flies go. The body length from front of head to tip of abdomen varies from about 7 to 15 mm., depending upon the species. Except in one species, Dolichopeza polita (all THE CRANE Friy Genus DOLICHOPEZA 683 its races), males average somewhat smaller in all dimensions than females. Wing length is approximately 8 to 15 mm. and again depends upon the species. Over-all size may vary not only with species but also with ecological fluctuations, as will later be ex- plained in detail. Those characteristics which distinguish Dolicho- peza from other tipuline genera have already been noted. The combined lengths of femur and tibia considerably exceed that of the wing, and the length of these two leg segments together is nearly always equalled or surpassed by that of the tarsus. The genus takes its name from this most striking feature; Greek dolichos = long, peza = foot. (Oropeza means “mountain-foot” and is derived from the Greek word oros = mountain.) This re- markable leg length, coupled with extreme slenderness, gives species of this genus a very characteristic appearance; it is what led one author to describe the legs of Dolichopeza as “excessively long.” Head.—Most of the surface of the head is occupied by the large, dichoptic compound eyes, which are of about equal size in both sexes and of a black color. There are no ocelli, which is charac- teristic of all Tipulidae. Surface areas of the head capsule proper, such as the occiput, vertex, front and genae, are poorly defined, but the clypeus is distinct, forming the upper (anterior) and lateral parts of the rostrum, or proboscis. There is no nasus on the ros- trum in the North American species. The antennae are composed of a rather cylindrical scape about four times as long as its width, a subspherical pedicel, and a flagel- lum of elongate, verticillate, more or less cylindrical segments, each somewhat thicker at the base than at the tip, and each beyond the first ordinarily shorter than the preceding segment (Fig. 10). In the subgenus Dolichopeza, there are ten flagellar segments (cf. Alexander, 1919:853), while in Oropeza there are eleven. The rostrum bears the mouthparts, conspicuous among which are the two labellar lobes (apparently modified labial palps), each nearly covered on its oral surface by a system of pseudotracheae branching from two arcuate channels that convey liquid food to the tip of the hypopharynx, concealed beneath the heavily sclero- tized, triangular labrum (Fig. 9). Peterson (1916:39) states that 5 no piece is present at the base of any generalized pal- pus (of the Diptera) which can be homologized with the palpifer of generalized insects.” He is supported in this opinion by Rees and Ferris (1939:146), Crampton (1942:34) and others. Accord- ingly, the basal portion of the maxillary palp (Fig. 8) is here inter- preted not as a palpifer but as one of five segments, the terminal one 684 ‘THE UNIverRSITY SCIENCE BULLETIN of which is longer than all the others combined, as in al] Tipulinae. The maxillary palp attaches to a sclerotized structure which, al- though it is mostly internal, is here considered to be the stipes of the maxilla, on the basis of musculature (Fig. 9). The structure I interpret to be formed of the partially fused stipites has been called the “apodeme of the cranial flexor muscle of the lacinia” by Rees and Ferris (1939) and the “maxillary apodeme” by Cramp- ton (1942). Thorax.—In either dorsal or lateral aspect (Figs. 14, 15), the external thoracic surface is dominated by the sclerites of the meso- thorax, which contains the flight muscles. Seen from above, the pronotum (pn) is a narrow, convex band, almost concealed beneath the prescutum, while the mesothoracic postnotum completely ob- scures the metanotum. Several clearly defined areas comprise the mesonotum; these are the prescutum (psc), medially divided scutum (sc), scutellum (scl) with a parascutellum (pasc) on each side, and a postscutellum (pscl), or postnotum. The term “thoracic dorsum,” used in descriptive paragraphs of this paper, refers to the mesonotum, which sometimes bears conspicuous color markings, especially on the prescutum. The broadly V-shaped, complete transverse suture between prescutum and scutum is characteristic of the entire family Tipulidae. On each anterolateral margin of the prescutum is a shallow indentation, the pseudosuture or pseu- dosutural fovea (psf ). All three pleural sutures are distinct and the sclerites of the pleural region easy to identify (Fig. 14). There is, however, some difference of opinion as to what these sclerites should be called. Between the pronotum and anterior coxa is the prothoracic epister- num. Just forward of this is a cervical sclerite, and along its poste- rior side is the small epimeron of the prothorax. The correspond- ing parts of the metathorax, located between the third coxa and the base of the haltere, are somewhat less clearly outlined. Ex- tending from the second coxa to the base of the wing is the con- spicuous pleural suture of the mesothorax. Immediately anterior to this is the mesepisternum, divided, sometimes indistinctly, into a dorsal anepisternum and a ventral pre-episternum (called sterno- pleurite by Alexander, 1942:199), which adjoins the coxa. Spots of dark color are found on these two areas, in certain species. Between the second and third coxae and rather closely applied to the former is the meron of the mesothorax (episternum, in Alexander, 1942:199), sometimes darkly marked. Above this is an irregularly shaped mesepimeron, or pteropleurite, which is THe CRANE FLy GENUS DOLICHOPEZA 685 rst 0 0.8m Xp 5 SCALE, FIGS. 8-9 Oo ! 2 3mm. Fic. 8. Mouthparts of Dolichopeza (Oropeza) polita, antero-dorsal as- pect; lab—labellar lobe, Ibr—labrum, mxp 1—first segment of maxillary palp, mxp 5—fifth segment of maxillary palp, rst—rostrum. Fic. 9. Same as 8, postero-ventral aspect; labm—muscle of the labellar lobe (labial palp ), mxpm—muscles of the maxillary palp, pgm—muscles originating on postgena and inserting on prementum, pm—prementum, pstr—pseudo- trachea, sti—partially fused stipites. Fic. 10. Antenna of Dolichopeza (Oropeza) sp. Fic. 11. Wing of Dolichopeza (Oropeza) carolus; A—anal veins, af—anal fold, C—costa, Cu—cubitus, M—media, m—medial cross- vein, m-cu—median-cubital cross-vein, R—radius, Rs—radial sector, Sc— subcosta. Fic. 12. Wing of Dolichopeza (Oropeza) carolus showing the vein Sc; present. Fic. 13. Same as 12, showing the vein R; . 2 present. 686 THe UNIVERSITY SCIENCE BULLETIN weakly divided into an anepimeron and katepimeron and is ordi- narily unicolorous. Between the mesepimeron and the postscutel- lum lies a sclerite known as the pleurotergite, which Crampton (1942:49) states is actually a lateral division of the postscutellum. What I have called the postscutellum is referred to as the medio- tergite of the postscutellum by Crampton. The mesothoracic spiracle is situated in a membranous area at the anterior edge of the segment, between the anepisternum and prescutum, and the metathoracic spiracle lies just in front of the base of the haltere. All species of Dolichopeza in North America have the wings un- marked except for narrow tinged bands along certain veins (es- pecially the cubitus and m-cu cross-vein) and a weakly or strongly colored, brownish stigmal spot. Figure 11 illustrates the typical venation of subgenus Oropeza, in which the costa, subcosta and radius, including all its branches, are considerably heavier veins than any others. The humeral cross-vein is present, and the arculus is complete. Occasionally the vein Sc, is found in the position in- dicated in Figure 12, and rarely R, ,, is present (Fig. 13). Spurious cross-veins have occasionally been observed, especially between veins R, and R,,;. As mentioned earlier, the cell Ist M, (discal cell) does not occur in the subgenus Dolichopeza (compare Figs. 1, 2). Designation of all veins follows the system used by Alex- ander (1942: 201). Both the femora and tibiae are of a length about equal to that of the fly’s abdomen, and the first segment of the tarsus alone is often longer than these two together. Each tarsal segment is of greater length than all those beyond it combined. The tibial spurs are slender and inconspicuous; they arise perpendicular to the axis of the tibia but bend sharply, so that the tip is nearly parallel to the tibia. The shape of these spurs seems to be the same in all species. Likewise, there is no apparent difference, from species to species, in the untoothed tarsal claws. Leg coloration aids in species identification, although any species may be properly iden- tified by the use of other characters. Male abdomen.—There are nine evident segments of the male abdomen (Fig. 16), the last of which, called the hypopygium, bears the genital appendages and other reproductive structures. The first and second terga are more or less fused, and the first sternum is shorter than the corresponding tergum. The third through eighth segments are unmodified. The tenth or anal segment (as) is rep- resented by a membranous projection surrounding the end of the THE CRANE FLY GENUS DOLICHOPEZA rst 14 occ to) pn psc NN scl o 4 “ 3 pscl Fic. 14. Dolichopeza (Oropeza) sp., right lateral aspect of tho- rax; abs—abdominal sternum, aes—anepisternum, ant—antenna, ce—compound eye, cx—coxa, h—haltere, mr—meron of meso- thorax, ms—mesepimeron, mxp—maxillary palp, pasc—parascu- tellum, pes—pre-episternum of mesothorax, plt—pleurotergite, pn—pronotum, psc—prescutum, pscl—postscutellum, psf—pseudo- sutural fovea, pt—paratergite, rst—rostrum, scl—scutellum, sc— scutum, sp—spiracle, tr—trochanter, w—wing. Fic. 15. Same as 14, dorsal aspect; abt—abdominal tergum, occ—occiput, other abbreviations as in 14. 687 688 THE UNIVERSITY SCIENCE BULLETIN digestive tube and situated close beneath the middle of the ninth tergum. It seems probable that in primitive Diptera there were paired gen- ital appendages, each comprising a basal and a distal segment, aris- ing from the ninth abdominal segment in somewhat the arrange- ment found in existing Mecoptera—Panorpa, for example. These appendages appear to be highly developed lateral phallic lobes (Snodgrass, 1957: 44-47), although they have been widely regarded as modified abdominal limbs, or gonopods. In the Tipulinae, the basal portion (basimere, or coxite, of morphologists), which has been called the basistyle (Fig. 18, bs), is completely fused with the ninth sernum (s9). The distal portion (telomere, or stylus) in many crane flies has become divided into two separately articulated ap- pendages (Fig. 17), a dorsal or outer dististyle (od) and a ventral or inner dististyle (id). The shapes of these structures can be used to a limited extent in classification. The ninth tergum in the subgenus Oropeza bears heavily sclero- tized lateral extensions, which may be called tergal arms (ta) and which function to brace the ventral valves of the female ovipositor during copulation. These have characteristic shapes, depending upon the species. While no such structures are found in the sub- genus Dolichopeza, their absence is compensated by other parts of the hypopygium. I have been unable to learn any function of the various forms of the posterior margin of the ninth tergum, but it is a fortunate coincidence that this sclerite offers one of the most reliable means for species determination in males. It may be shal- lowly or deeply emarginate or crenulate, or may bear a median projection with sharp teeth or blunt lobes. Between the basistyles, on either side of the midline of the ninth sternum, are the gonapophyses (gon), the various shapes of which have already been briefly discussed. The term “gonapophyses” has been used in many ways in the several orders of insects, but most authors agree that the gonapophyses of male insects are median projections in proximity to the gonopore. The origin and homology of the gonapophyses in Dolichopeza are unknown; Fig- ure 29, however, indicates the close relationship between them and the adminiculum and basistyles. It seems advisable to retain the term for these particular structures since it does not appear to be misapplied. The male intromittent organ, or penis (Fig. 17, p) is an extensible, slender and rather densely sclerotized tube, straight tipped, or nearly so, in some species and abruptly curved at the tip in others. THe CRANE Fiy GENusS DOLICHOPEZA Ke) as id 1) HAE | 7 . t fi : i 1S, us Z 4 fT WY Ae fey eres co pees t9 17 18 10mm. —eE——————— Fic. 16. Abdomen of male Dolichopeza (Oropeza) tridenticulata, left lateral aspect. Fic. 17. Hypopygium of male Dolichopeza (Oropeza) tridenticulata, dorsal aspect; as—anal (tenth) segment, gon—gonapophysis, id—inner dististyle, od—outer dististyle, p— penis, ta—tergal arm (lateral arm of ninth tergum), t9—ninth ter- gum. Fic. 18. Hypopygium of male Dolichopeza (Oropeza) venosa, left lateral aspect; bs—basistyle, s9—ninth sternum, other abbrevia- tions as in 17. 689 690 THE UNIVERSITY SCIENCE BULLETIN Its method of extension will be explained later. Where the penis emerges from the floor of the genital chamber, it is braced by a cone-shaped, intensely sclerotized structure (Figs. 17, 28) that is actually formed from partial or complete fusion of two elements, one on either side. This has been termed the aedeagus by Alex- ander (1942, etc.) and the adminiculum by Mannheims (1951:10), Snodgrass (1957:45) and others. Its two component parts have been regarded as parameres (Tjeder, 1948), hypomeres (Rees and Ferris, 1939) and mesomeres (Snodgrass, 1957). These parts ex- tend cephalad in the floor of the genital chamber and enter the abdominal cavity as short apodemes supporting the vesica (Figs. 28, 29). The term “aedeagus” has been used synonymously with what I have designated as the penis, also for the penis and its basal structures combined, and in other more restricted senses, almost always, however, as a phallic structure. Accordingly, I have adopted the term “adminiculum,” meaning a support, for the func- tion of this cone-shaped structure in Dolichopeza is strictly one of support and not intromission in any way. Female abdomen.—There are ten clearly defined segments in the female abdomen, of which the eighth through tenth are modi- fied to form the cvipositor (Fig. 19). As in the male, the first and second terga are somewhat fused, and the first sternum is shorter than the corresponding tergum. The third through seventh segments are unmodified. Small spiracles occur in the mem- branous pleura of the first seven segments but are usually concealed by overlap of the sclerites if most of the eggs have been laid. Ex- cept for the first, which has shifted to the forward edge of its seg- ment, the spiracles are all near the midlength of their respective segments. It has been argued by Crampton ( 1942:81 ) that the type of egg laying apparatus found in Dolichopeza and other crane flies “. should be referred to as an oviscapt, or ovicauda, rather than an ‘ovipositor . . . since its parts are not strictly homologous with those of a true ovipositor of the type occurring in orthopteroid insects.” However, as the structure takes its name from its undeni- able function of oviposition rather than from any homology, the term “ovipositor” is retained. In the typical pterygote ovipositor, the gonopore, opening between the eighth and ninth sterna, is flanked by a pair of blade-like valvulae, attached to basal sclerites (first valvifers) belonging to the eighth sternum, and two pairs of similar valvulae belonging to the ninth sternum. The latter pairs are attached to common bases (second valvifers) at either side, and THe CRANE Fry GENUS DOLICHOPEZA 691 0.5mm. Fic. 19. Abdomen of female Dolichopeza (Oropeza) tridenticulata, left lateral aspect. Fic. 20. Same as 19, showing details of ovipositor; cr— cercus, fr—furca, fy—fused valvulae (ninth sternum), hv—hypovalve (ex- tended eighth sternum). Fic. 21. Details of external reproductive struc- tures of female Dolichopeza (Oropeza) similis, postero-ventral aspect; abc—aperture of bursa copulatrix, other abbreviations as in 20. Fic. 22. Details of reproductive structures of female Dolichopeza (Oropeza) triden- ticulata, left lateral aspect, cut-away view; a—anus, abc—aperture of bursa copulatrix, gp—gonopore. Fic. 23. Ovipositor of female Dolichopeza (Oropeza) johnsonella, left lateral aspect, showing sensory structures. 692 Tue UNrversiry SCIENCE BULLETIN all together comprise the ninth sternum. Snodgrass (1935:611) says of these, “The second valvifers usually retain a close connec- tion with the tergum of the ninth segment, but the first valvifers are often more or less dissociated from the eighth segment.” Above and to the rear of the typical ovipositor are the tenth and eleventh abdominal segments, with cerci arising from the membrane between them and with the anal opening situated close beneath the eleventh tergum. The structure found in females of Dolichopeza, as in other tipulids, does not depart from this basic arrangement of parts except in details. Completely encircling the eighth abdominal segment at its anterior end is a rather strongly sclerotized ring which nar- rowly connects the tergum and sternum (Fig. 20). While the tergum is otherwise unmodified, the sternum is projected backward to beneath the tenth tergum, or further, in the form of a pair of egg guides, or hypovalves (hv), which however are fused nearly to their tips in the genus Dolichopeza. These hypovalves are pocketed on their inner surfaces to receive the tips of the inner dististyles of the male, in order to effect a firm attachment in copulation. Rising slightly from the inner floor of the hypovalves is the membranous end of the median or common oviduct, terminating in the gonopore (Fig. 22, gp). There is, in the narrow ninth segment, a modification of the sclerotized anterior ring found in the eighth. Semicircular above, where it corresponds to the anterior margin of the tergum, this band is, at the sides, deflected abruptly caudad and toward the midline, where its ends meet to form a fairly heavily sclerotized, triangular blade, which may be regarded as the fused second val- vulae (Figs. 20, 21, fv), and thus the ninth sternum. Snodgrass (1903: 178) has called these the fused second gonapophyses. Be- tween this reduced ninth sternum and the hypovalves, or eighth sternum, is the genital chamber. Immediately below the ninth sternum is the upper opening of the reproductive system, that which leads into what may be called the bursa copulatrix, since it is into this opening that the penis of the male is inserted in mating (Figs. 21, 22, abc). Thus, while it would seem that the two sepa- rate parts of the reproductive system open independently to the out- side, it may be seen that both in fact enter the genital chamber as above defined. Lying in the membrane beneath the aperture of the bursa copu- latrix is a sclerotized fork, of which one tine extends to either side of the aperture. The base of this fork is in a pouch formed just above the oviduct (Fig. 22), and it sometimes happens that the THe CRANE FLy GENUS DOLICHOPEZA 693 floor of this pouch (that is, the roof of the oviduct) bears small, usually weakly sclerotized areas, closely associated with the base of the fork, as indicated by the broken lines in Figure 21. Students of the Psychodidae (especially Phlebotomus) have called this fork the furca, the name that is adopted here (Figs. 20, 21, fr). Snod- grass (1903: 178) is of the opinion that the furca “. . . may be the fused and rudimentary anterior gonapophyses,” by which he means the first valvulae and their valvifers in the sense used abovye.* This seems a sound interpretation in spite of the great modification of the parts involved. The tenth segment has a saddle-like and strongly sclerotized tergum (Fig. 23) but no sclerotized sternum. Beneath the pos- terior end of the tenth tergum is situated the anal opening. As in most crane flies, the cerci are strongly sclerotized and blade-like in shape, reinforced along their dorsal edges. Minute pegs grouped at the tips of the cerci (Fig. 23) are thought to be sensory in func- tion, and it sometimes seems that females in the process of ovipo- sition obtain impressions about the nature of the substrate by means of probing with the cerci. Inasmuch as the cerci have an important function also in the actual deposition of the eggs, they may be termed the dorsal (or “tergal”) valves of the ovipositor. A small, indistinct sclerite between the bases of the cerci may be interpreted as the eleventh tergum. ADULTS—INTERNAL MORPHOLOGY Head.—Much of the interior of the head is occupied by portions of the central nervous system, particularly the brain and the large optic lobes, the convex apical surfaces of which are in contact with the basement membranes of the compound eyes. The only other conspicuous structures are the anterior parts of the digestive tube (Fig. 26). In the rostrum, the epipharynx (eph) and hypopharynx (hph) are fused basally, forming a partly sclerotized tube, the basipharynx (bph), which can be dilated by means of a thick muscle attached to its roof and originating on the anterior wall of the rostrum, or clypeus. Flexibly joined to the upper end of the basipharynx is the pharyngeal pump, a structure made up of three sclerotized elements connected by elastic membranes to form a tube that is somewhat expanded at each end and roughly triangular in cross section. Some authors have applied the term oesophageal pump to the pharyngeal pump. The arrangement of attached muscles suggests that dilation of the aboral end of the pharyngeal *Jt is inadvertently stated by van Emden and Hennig (1956: 121) that Snodgrass homologized the hypovalves with the anterior gonapophyses. Td ‘snpNOL}UOA—UA “BTTIXBUT jo sadys—ts ‘spur[s AreAtyes— 3s “un “uoutord—uid ‘duind jeosudreyd—dyd ‘sns -vydosso—seo0 ‘uiniqe[—jq] ‘xuAreydoddAy —ydy ‘xudreydidea—yde ‘spurs Areatyes jo yonp uouruI0o—]ssp ‘xudrvydiseq—ydq ‘uon -10d ro1sezue ‘uIdyshs BANSeBIp JO s[teyod “9G ‘org, ‘uattoads poadrosord Ul seynqny uerysid -[vyy jo souvsrvodde Burmoys “ds (pzado1O) zadoyoyodg jo uteysAs dAsosip jo [red ‘CZ ‘Ol_ =“ SMoryUeA—UA ‘sajosnu sip JeonreA—wya “(anssh asodipe YA posed -ul ATIAvey sseur poUlAj1ozUt Soedutoo se uMOYs ) sainjon44s aAtonpoidei—si ‘eypided [eyooi—di ‘duind jeosudreyd—dyd ‘aynqny uviysidyeyy—dur *( saposnut VSI [BonteA popeys 10A0 pesoduriedns vere poprysun sv UMOYS ) SSBUT 9pOSNUT YS [eurpnyrsuo;—wuy] ‘aqoy reT[eqe[— Pl 4n3-pury—sy ‘do1o—do ‘pyojnoyuapiy (pzado1Q) vzadoysyod eeu ype jo Atuoyeur [eUIOJUT SsOILy ‘PE OA THe CRANE Fiy GENUS DOLICHOPEZA 695 pump provides most of its action. A single salivary duct leads from the ventral side of the hypopharynx along the pharynx to the neck region, where the duct divides into two (Fig. 26). There is no tentorium. Thorax.—By far the greatest volume of the thoracic cavity is taken up by the flight muscles: lateral, vertical bands and two large, longitudinal muscle masses, one on either side of the median plane, as outlined in Figure 24. A short length of membranous esophagus connects the oesophageal pump with the ventriculus in the cervical region. There appears to be no proventriculus. The anterior end of the ventriculus has a decidedly lumpy surface, the nodules being largest just behind the neck and diminishing in size until inconspicuous about midway through the thorax. A mem- branous crop, which with its duct-like proximal portion extends nearly the full length of the thorax, is attached ventrally at the junction of the oesophagus with the ventriculus (Fig. 26). Just within the thorax, beneath the oesophagus, lie the two small sali- vary glands, the ducts of which unite quite abruptly. The ventral nerve cord, passing through the thorax from the suboesophageal ganglion, produces three large thoracic ganglia, close below the digestive tube. These give off branches to the thoracic appendages and muscles. Abdomen.—Except for varying amounts of fat tissue and the thin bands of muscles in the body wall, the contents of the abdomen consist almost entirely of the reproductive system and parts of the slender alimentary canal. Pale greenish gray and shining in a freshly killed fly, the reticulated fat tissue, most abundant in males, is dull white in specimens preserved in alcohol. The ventriculus, entering the abdomen along the floor of the first segment, curves up against the body wall to terminate dorsally in the third abdomi- nal segment, at the point where the four Malpighian tubules attach. These excretory outgrowths are about as long as the entire body and are either coiled mostly in the upper part of the abdominal cavity or are intertwined among the other organs. In life, their color is mottled gray, but alcohol preservation usually renders them pale brown or reddish brown; also in life there may be a zone of yellowish coloration about the attachment of the Malpighian tubules. The hindgut is slightly enlarged at its anterior end, taper- ing evenly to the rectum, a thin walled and sac-like structure close beneath the seventh, eighth and ninth terga. Four irregularly ob- long rectal pads (Snodgrass, 1935: 381) of opaque white color lie in the anterior part of the membranous rectal wall. 696 Tue UNrversiry SCIENCE BULLETIN Male reproductive system—A conspicuous arrangement of strongly sclerotized parts, lying primarily in the eighth and ninth abdominal segments, is visible in specimens cleared in KOH for mounting on microscope slides. These structures have been studied, variously named and employed in classification by many investi- gators (Snodgrass, 1904; Tokunaga, 1930; Edwards, 1938; Tjeder, 1948; Mannheims, 1951; Wood, 1952). As described earlier, each of the two elements forming the conical adminiculum is continuous with a flattened, sclerotized rod lying in the membranous floor of the genital chamber. Upon the anterior tips of these rods, which project into the abdominal cavity as apodemes, rests a bulb or cap- sule, which is joined to the base of the penis (Fig. 28). This has been called the spermatophore sac by Tokunaga, and the central vesicle by Snodgrass. Following the terminology used by Edwards (1938: 14), it is here called the vesica. As the vesica is capable of rotation in the sagittal plane, rocking on the tips of the adminicular rods, it is difficult to describe its parts as being dorsal or ventral, anterior or posterior, as can be done for those genera in which this organ is fixed in position. However when the penis is fully re- tracted, the vesica has the position indicated in Figure 28, and the following description is based upon the structure thus oriented. There are two wing-like outgrowths from the anterior surface of the vesica, one arising at either side of the base of the penis and curving down along the side of the bulb to about the point of flexible attachment of the vesica to the adminicular rods. These are the lateral apodemes (lap). From the posteroventral surface of the vesica extend two rather stout rods, diverging backward, which I have called the posterior apodemes (pap). A transverse, fan- shaped and nearly semicircular apodeme arises from the poster- odorsal surface of the vesica; this is the dorsal apodeme (dap). It has a thickened median keel at the base of which is an almost hemi- spherical, heel-like piston that compresses the vesica when muscles between the dorsal and lateral apodemes are contracted. Muscles between the dorsal and posterior apodemes serve to dilate the vesica. From the vesica, the penis curves forward and downward, within a membranous pouch, through the seventh segment to the floor of the eighth and so out through the adminiculum. The mem- branous pouch could be regarded as a narrow but deep continu- ation of the genital chamber, as it is open to the outside at the adminiculum.* The membrane is suspended at either side from the adminicular rods, the anterior end of the pouch being closed *In some species of Tipula I have examined, the opening continues cephalad between what appear to be equivalents of the adminicular rods. Tue CrANE FLY GENuS DOLICHOPEZA 697 ° é 2 5 cS) 2 . = ° 2) S © = 3 ° ” Fic. 27. Details of reproductive system of male Dolichopeza (Oropeza) polita ssp.; acg—accessory gland, adm—adminiculum, cvd—common vas deferens, mbp—membranous pouch enclosing penis, p—penis, sd—seminal uct, sv—seminal vesicle, t—testis, vd—vasa deferentia, vg—vesicular gland. Enlarged insets: a—common vas deferens, b—accessory gland, c—seminal duct, d—vesicular gland. 24—5840 698 Tue UNIVERSITY SCIENCE BULLETIN beneath the vesica and around the base of the penis. Contraction of muscles originating on the ninth sternum and inserting on the anterior faces of the lateral apodemes brings about rotation of the vesica and exsertion of the penis, which action is aided by small muscles connecting the posterior apodemes with the ninth tergum. The nonsclerotized parts of the male reproductive system are usually complexly intertwined and difficult to disentangle. They have rarely been examined in detail in any Tipulidae (see Keuche- nius, 1913) and, so far as I am aware, never in the genus Dolicho- peza. Ihave unraveled this mass of tubes in most species of North American Dolichopeza and have found a rather consistent pattern, illustrated by the example of Dolichopeza (Oropeza) polita ssp., in Figure 27. There are two thin walled, bladder-like testes (t), which usually lie one before the other near the anterior end of the mass of the reproductive system. In a freshly-killed fly, these are usually found full of spermatozoa. The testes empty into the vasa deferentia (vd), which are at first exceedingly slender, then thick- walled and finally merged into a common vas deferens (cvd) with thick, irregular, possibly glandular walls enclosing a more or less central tube (Fig. 27a). The vasa deferentia are hyaline in appear- ance and are easily recognized among the abdominal contents by their glossy sheen. The common vas deferens terminates in the center of an enlarged, ring-shaped, glandular-walled seminal vesicle (sv). That this structure, and not the enlarged parts of the vasa deferentia (cf. Snodgrass, 1935: 568), is actually the seminal vesicle is indicated by its comparatively large lumen, in which masses of spermatozoa have been observed. Two elongate “accessory glands” (acg) also con- verge and join at the seminal vesicle. These tubular glands are of a diameter about equal to that of the Malpighian tubules, with a relatively large lumen and walls of apparently columnar cells (Fig. 27b). Their color in life is pale rusty, with a granular appearance, but preserved in alcohol they are a dull white. The connection of the seminal vesicle to the other organs is not easily traced; however, it appears that it joins them at one end only, the other end being looped around the bases of the vas deferens and the accessory glands until it touches the end with the opening. From the confluence of all these organs, a long and much- coiled conduit leads to the posterior side of the vesica, between the posterior apodemes. This has been called the ejaculatory duct (Keuchenius, 1913, for Tipula sp.), but because I believe that term hardly describes its function, I have called it simply the seminal THE CRANE FLY GENUS DOLICHOPEZA ° 0.5 1LOmm. (he ee ee J Fic. 28. Reproductive structures of male Dolichopeza (Oro- peza venosa; adm—adminiculum, adr—adminicular rods, dap— dorsal apodeme, (compressor apodeme) of vesica, lap—lateral apodeme, pap—posterior apodeme, vs—vesica. Fic. 29. Re- productive structures of male Dolichopeza (Oropeza) polita ssp., dorsal aspect, showing relationship of adminiculum, adminicular rods and gonapophyses; adr—adminicular rod. Fic. 30. Left basistyle and dististyles of male Dolichopeza (Oropeza) walleyi, inner or mesial aspect, showing musculature (cf. Fig. 17). 699 700 THe UNIVERSITY SCIENCE BULLETIN duct (sd). In species of the obscura group, the seminal duct is quite lengthy and is coiled in the manner of a spring, although the direction of coil is sometimes here and there reversed. Laid out more or less straight but with the small coils still in place, the semi- nal duct in a male Dolichopeza (O.) polita ssp. measured 23 mm. In D. americana and the sayi group, this duct is shorter and less tightly coiled, as a rule. In D. (O.) carolus, it is only about one- fourth to one-third as long as in polita but equally coiled and of about the same diameter as in the latter species. In carolus also the basal parts of the accessory glands are widened, rather re- sembling the common vas deferens. Over-all length of the reproductive system from the vesica to the testis is about 45 mm., in larger species of the obscura group. From the seminal vesicle to the testis was 21.2 mm. in obscura; the common vas deferens measured 11.5 mm. in polita. In americana and the sayi group, the total length is about half as great, the difference in length of the seminal duct accounting for most of the decrease. Perhaps most conspicuous among the nonsclerotized male re- productive organs are a pair of large, flattened, broadly reniform glands lying one at each side and somewhat below the vesica. Earlier authors seem not to have noticed these, but I have seen similar glands in Tipula spp. and other crane flies. They seem to reach their greatest development in Dolichopeza. The outer walls of these are smooth, but the inner surfaces have a vertically corded appearance and are made up of large, irregularly polygonal cells with large nuclei (Fig. 27d). The glands possess a lumen of capac- ity at least equalling that of the seminal vesicle and empty by ducts on their upper medial surfaces directly into the vesica. Judging from their size, they must contribute very significantly to the semi- nal fluid. The designation vesicular glands (vg) seems appropriate for these. Female reproductive system.—The reproductive apparatus of the female Dolichopeza is divided into two portions: the bursa copu- latrix and its associated structures, and the oviducts and their ovar- ioles. These two parts have separate external apertures, as dis- cussed earlier under the external morphology of the females, and they seem to have no internal connections. Leading cephalad from the gonopore is a thin-walled but muscu- lar common oviduct (Fig. 31, covd), which in about the middle of the seventh abdominal segment divides into two lateral oviducts (lovd). These organs, because of their membranous structure, are ‘DupIawDd (pzadoyIyod) pzadoyoyod aewey Jo sammjonij3s 9At} -onpoidarl [eutoyUy “Gg ‘SI ‘syonp Jeoouqeutiods jo yUOUTYOR}e SurMoys ‘podse [vsiop ‘xtyejndoo evsinq jo pus LOMOyUv FTE sv ouleg “PE “oly ‘spurys Alossao0v jo [itjap ‘Tg se oles ‘SG “OILY = “snjoupo (vzado1Q)) vzadoya -0q Jo aeooyyeuttods OM], “ZG “OI ‘eooyyeuttads—yjds ‘yonpiAo = [e1o}e] —pao, ‘e1odouos—ds ‘eooyyeutiads jeuonounj—yjdsz ‘JonprAo — uoUTUOD —paoo ‘xrqjepndoo vsInq—oq ‘spurs Alossoo0Rr—Soe ‘xnQejndoo — esinq jo oinjtoade—oqe ‘pyojnoyuapiy (vzad -01Q)) pzadoyoyog a[eula; Jo saanjons}s aAtjonpoider eutoyUy “Tg “Sy 702 THe UNIversiry SCIENCE BULLETIN most easily seen in older females that have laid practically all their eggs. From each lateral oviduct, lying along the floor of the ab- domen, the many ovarioles rise dorsally and bend anteriorly, their filiform tips often attaching to the dorsal or lateral abdominal walls. Within the ovarioles, eggs in increasing stages of develop- ment may be seen, from scarcely discernible specks in the upper ends to well formed but still pale colored eggs toward the lateral oviducts. In a newly-emerged female, the lateral oviducts and lowermost portions of the ovarioles are stretched taut with fully formed eggs, in which the chorion is already black and the terminal filaments (in those species that have them) completely developed. The eggs lie with their posterior ends directed toward the gono- pore, which is also the way they are oriented in the ovarioles. In gravid females, eggs are found as far forward as the first ab- dominal segment, compressing the viscera tightly against the body wall. In copulation, the penis of the male is inserted into the bursa copulatrix (be), and the spermatozoa are presumably stored in the three spermathecae until the time of oviposition. The sperma- thecae (spth) are very small, intensely sclerotized capsules that are the only structures besides eggs to show at all clearly in flies that have been treated with KOH for mounting on microscope slides. They connect to the anterior end of the bursa copulatrix by filamentous ducts, two of which join the bursa dorsolaterally and one ventrally (Fig. 34). Their shape is nearly spherical in most species but flattened on the side of attachment in Dolichopeza carolus (Fig. 32). The spermathecal ducts are of great length and are thrown into long loops within the abdomen. In teneral fe- males, the distal ends of these ducts, that is, the ends nearest the spermathecae, are soft, wrinkled and coiled. In older females, however, the ducts grow firmer, straighten out, and elongate, as the spermathecae move from the anterior end of the abdomen toward the middle or posterior regions. In Dolichopeza americana, the spermathecal ducts are shorter, and the spermathecae are usually found near the caudal end of the abdomen in both teneral and older flies (Fig. 35). In species of the subgenus Oropeza, the bursa copulatrix extends cephalad to the sixth or fifth abdominal segment, while in Doli- chopeza s.s. it is much shorter, scarcely half a millimeter in length (compare Figs. 31 and 35). The bursa has a thin, yellowish lining and a thick, whitish-transparent outer wall. A pair of nodular, whitish accessory glands (acg), the function of which in THE CRANE FLY GENUS DOLICHOPEZA 703 these flies is unknown, open into the bursa copulatrix on its dorsal wall near the aperture (Fig. 33). A relatively large pouch joins the anterior end of the bursa copulatrix, only a short distance from the attachment of the sperma- thecal ducts. It is rather oval in shape, with thin wrinkled walls in species of the obscura group (Fig. 31), thicker-walled and more elongate in the sayi group, and rigidly thick-walled, elongate and curved in Dolichopeza americana (Fig. 35). Its connection to the bursa is somewhat different from species to species in Oropeza, as will be seen on comparison of Figures 31 and 34. Unable to find any structure in any other insect that is homologous with this pouch, and noting the wide opening from the pouch into the bursa copulatrix, I am inclined to regard the pouch as the func- tional spermatheca (fspth) and to suspect that the three small spermathecae are generally nonfunctional in Dolichopeza, and perhaps in other Tipulinae. This supposition finds some support in the fact that oviposition takes place very soon after mating and that the eggs are fertilized at the time of deposition, presumably by a flow of seminal fluid over them sufficient to insure that some spermatozoa will reach the micropyle of each egg; hence the supply of spermatozoa is presumably large and situated rather near to the opening of the oviduct. The spermathecae, on the other hand, are quite small, and their ducts are in some species nearly twice the length of the entire abdomen. It may be, of course, that the large pouch stores sperm for rather immediate use, while the spermathecae are reservoirs for several hours’ on days’ storage. INTRASPECIFIC VARIATION After studying many individuals of a species, it is possible to con- clude, somewhat reliably, what is the normal or usual condition of practically any observable feature of the organism involved. On this basis, it is possible also to visualize an average individual of the species, in which means of dimension, color, shape and so on are combined. Then, all specimens at hand will be seen to deviate, in one way or another and to varying degrees, from the average. Small departures of common occurrence may, I believe, be re- garded as falling within a normal range, while extraordinary in- dividuals may be considered abnormal. It is also found, from time to time, that with respect to one or more characteristics individuals may group around two or more distinct means, in which cases special interpretation is required, for clines or geographic races may be indicated. Some of the individual, intraspecific variation in Dolichopeza has 704 Tue UNtversiry SCIENCE BULLETIN been misinterpreted, as it has also in many other groups of or- ganisms. It therefore has seemed worthwhile to attempt to classify the kinds of variation encountered and to indicate their probable causes. The classification that follows is not intended to be general in scope but applies only to variation observed in Dolichopeza. The kinds of variation include: 1. Seasonal, or ecological, size variation. Color variation. Morphological variation, both A. Growth anomalies (all of which are abnormalities) and B. Inherited differences, which may constitute (1) abnormal morphological variation, or (2) normal morphological variation. oo po Size variation—Over-all size of individual flies appears to be affected markedly by conditions in the larval habitat. Larvae of Dolichopeza inhabit mosses and liverworts, which ordinarily grow most luxuriantly in the spring. Adult flies emerge at two times each year, in most species and in most parts of the range of the genus, there being flight periods in the spring and again in the late summer or early fall. Those larvae which give rise to the spring generation of adults feed in the autumn, become quiescent in winter and feed further in early spring, while those developing into adults of the fall generation, being progeny of the spring adults, feed only during a period of about two and a half to three months in the summer, when their moss habitat is usually drier than at other times of the year. That the fall generation adults in all species are of smaller size than those of the spring generation appears to reflect this variation in larval habitat. This is further suggested by the fact that spring generation adults reared in the laboratory, where they are subjected to various inadequacies of environment, are smaller than flies of that generation occurring in nature. Of course, there is size variation among flies of a single generation, which may result from fluctuations in larval environ- ment or may be due to other, less understood causes. Seasonal variation is well illustrated by Dolichopeza americana. On the assumption that wing length is a reliable index of over-all size of the fly, wings of ten males and ten females representing each generation were measured and compared (Table 1). All spring generation flies were captured between 10 and 25 June and those of the autumn generation all on 4 August; and all specimens were collected within the same general habitat, within a radius of 200 yards, at a locality in west central Indiana. The flies were taken THE CRANE Fiy GENuS DOLICHOPEZA 705: TABLE 1.—Wing Length in Dolichopeza americana ( Measurements in Millimeters ) Generation Mean Range Median Nallesa¢:. 5) ws: Spring 10.8 92-11-11 10.4 Fall 8.4 7.2- 9.1 8.7 Females........ Spring 11.4 11.0-12.9 le Fall 9.4 8.9-10.1 9.2 at random from specimens stored in opaque paper envelopes. In this instance, the means in both males and females are widely separated, and it may be seen from the table that the ranges of wing length of spring and fall flies do not even overlap. This is a rather extreme example, but it illustrates the tendency seen in the other species. Color variation—Color of preserved specimens is greatly affected by their age, exposure to sunlight, and method of preservation. In specimens only a year or two old, as well as in living material, there is sufficient variation in coloration to indicate that ordinarily no reliance can be placed on slight differences of shade or color for taxonomic purposes. Where I have referred to color in describing species, especially in the key to adult females, I have tried to limit its use to color patterns or to general coloration, used in conjunc- tion with other characters. It is very difficult to communicate one’s estimate of color to another person, and because the colors in Dolichopeza are not fixed enough to be referred to a standard color guide, it has seemed best to omit discussion of color details. There are, however, a few examples of significant intra-specific variation in color in North American Dolichopeza, the most out- standing of which is that found in D. walleyi. This species in gen- eral appearance rather closely resembles another, namely D. sayji. In northeastern United States and southeastern Canada, where the ranges of the two species broadly overlap, they are distinctly col- ored. D. sayi has darkly spotted thoracic pleura, nearly black stripes on the prescutum, and darkly sclerotized hypovalves in the ovipositor of the female; in contrast, D. walleyi has pale thoracic pleura, reddish brown prescutal stripes, and paler hypovalves in the female. Specimens of D. walleyi from localities a short dis- tance outside the range of D. sayi have coloration suggestive of some sort of mixture of those described, and populations of walleyi from Florida, Iowa and South Dakota (that is, from localities fur- thest from the range of D. sayi) are very like sayi in the intensity of their markings. This phenomenon has been described as char- acter displacement (Brown and Wilson, 1956). 706 THE UNIVERSITY SCIENCE BULLETIN Growth anomalies—I have occasionally discovered an individual having a very unusual deformity, which by its rarity and nature appears to be the result of defective development rather than the effect of an unusual gene or combination of genes. In Dolichopeza tridenticulata, for example, the posterior margin of the ninth ter- gum of the male abdomen bears a three-pointed projection such as that shown in Figure 59; but the presence of a foreign particle in the tissue of the tergum resulted in the deformity illustrated in Figure 36. Another growth anomaly is the extraordinary enlarge- ment and sclerotization of the apex of one of the inner dististyles of Dolichopeza polita, compared with the normal structure of the opposite appendage in the same individual (Fig. 37). Growth abnormalities may affect any stage in the life history. For ex- ample, a pupa of D. obscura lacked one lateral lobe of the series of four ordinarily occurring in a transverse row on the eighth sternum (Fig. 38). Above the spiracular region of the larva of Dolichopeza, there are four subconical, fleshy lobes, the middle two of which are shorter than the others and situated close together, one at either side of the midline. However, among many hundreds of larvae of all species examined, there was found one specimen in which these two median dorsal lobes had grown together (Fig. 39). This individual, incidentally, showed the normal structure in the next earlier instar. All anomalies of the sort here described have been seen only once among the many thousands of flies examined. Inherited differences—Other morphological differences seem to have their basis in the genetic makeup of the insects. Among these are several abnormalities that recur with varying frequencies, either throughout the range of a particular species or scattered throughout the genus. The habitats of Dolichopeza are discontinuous in most of the range of the genus in North America, and as a result of this there is probably a great deal of local inbreeding. This, at least, would account for the fact that certain abnormalities are locally quite common. Perhaps the most readily observed of such charac- ters is wing venation, the usual pattern of which in subgenus Oropeza is presented, in part, in Figure 40. Now, throughout most of its wide range, Dolichopeza tridenticulata has the normal wing venation of this subgenus, but in central and southern Indiana, local populations of this species may have as high as 40 to 55 percent incidence of deformity of the medial field, that is, of the branches of the media. Figures 41 through 47 illustrate wings of D. tridenticulata in which deformity of the discal cell, a shift of THE CRANE Fry GENUS DOLICHOPEZA 46 i aati Fic. 36. Abnormal development of ninth tergum of male Dolicho- peza (Oropeza) tridenticulata due to presence of foreign body. Fic. 37. Inner dististyles of male Dolichopeza (Oropeza) polita ssp., showing abnormal development of apex of left dististyle. Fic. 38. Eighth sternum of pupa of Dolichopeza (Oropeza) obscura, showing abnormal development of spinous processes. Fic. 39. Spiracular disc of larva of Dolichopeza (Oropeza) sayi, showing abnormal (fused) dorsal lobes. Fics. 40-48. Variations in wing venation in Dolichopeza (Oropeza) tri- denticulata; 40—normal venation, 41-47—aberrations of the medial field in a population from central Indiana, 48—spurious cross-vein in cell Rg in a population from southern Minnesota. 707 708 Tue UNtversiry SCIENCE BULLETIN the base of the vein M, from M, to M,, and other abnormalities may be seen. Figure 43 is particularly interesting in that the discal cell is completely absent, and the m-cu cross-vein has shifted so as to join the media before its first branching, a combination of charac- ters typical of the venation of the subgenus Dolichopeza! Among specimens of D. subalbipes from the southern part of the range of the species, there are many similar deformities of the medial field. Some of these, one even resulting in an extra closed cell beyond the discal cell, are shown in Figures 49 through 51. A further, rather remarkable example of this kind of intraspecific variation is shown in Figure 48. A spurious cross-vein in cell R, of the wing is extremely uncommon anywhere in the genus; yet, in one locality in Minnesota, ten Dolichopeza tridenticulata in a sam- ple of thirty specimens had such a cross-vein in one or both wings. The venation of these flies was in other respects normal. Venational aberrations of sorts similar to those described here have been ob- served in the mosquito, Culex pipiens, in which species they have been found to be controlled by single mutant genes (Laven, 1957: 452 ff.). The occurrence of macrotrichia in the apical cells of the wing is similarly uncommon throughout the North American species of Dolichopeza, but in D. walleyi this variation has appeared in a few populations in scattered parts of the range (Fig. 52). In contrast to such striking departures from the usual condition, there are several morphological variations of lesser degree that are so often found in almost any population of a species that I think they may be considered as normal variation, just as we recognize certain differences in detail of facial features among normally ap- pearing human beings. These slight modifications of form are probably genetically controlled, although no study of this has been made. Examples are found in all species of Dolichopeza, and within each species such characteristics seem to be distributed at random, or nearly so. That is to say, in any population throughout the species’ range, all known variations of the particular character may occur. Such differences include minor shifts in the positions of various wing veins with respect to the discal cell (Figs. 53 through 57) and slight modifications of the caudal margin of the ninth tergum or the tergal arms in males (Figs. 58 through 66). Many other examples could be cited, but these will serve to illus- trate the point. Dolichopeza tridenticulata takes its name from the three-toothed projection on the ninth tergum of the male, a structure ordinarily THE CRANE FLY GENUS DOLICHOPEZA S| De 58 59 60 6| 62 63 ES lle ips oo ea Fics. 49-51. Aberrations of the medial field in wings of Dolicho- peza (Oropeza) subalbipes; 49—specimen from Florida, 50—Georgia, 51—Alabama. Fic. 52. Wing of Dolichopeza (Oropeza) walleyi from Michigan, showing macrotrichia in apical cells. Fics. 53-57. Varia- tions in shape of discal cell (cell 1st M2 occurring commonly in Doli- chopeza (Oropeza) spp. Fics. 58-60. Normal variation in medio- posterior margin of ninth tergum of male Dolichopeza (Oropeza) tridenticulata. Fics. 61-63. Normal variation in medio-posterior mar- gin of ninth tergum of male Dolichopeza (Oropeza) obscura. Fics. 64-66. Normal variation in lateral arms of ninth tergum of male Doli- chopeza (Oropeza) walleyi. 709 710 THE University SCIENCE BULLETIN having a shape like that shown in Figure 59. However, in any sample of several males of this species, one may expect to find the projection more slender, as in Figure 58, or broader and with the points more rounded, as in Figure 60. Likewise, in D. obscura, the median posterior teeth of the ninth tergum, most often having the form of Figure 62, may be variously spaced and deflected, as in- dicated in Figures 61 and 63. Shapes of the apices of the tergal arms in D. walleyi may vary as indicated in Figures 64 through 66, not only from one individual to the next but from side to side in the same fly. In all species of the subgenus Oropeza, slight variations in the positions of veins about the discal cell are commonly seen. For example, the m-cu cross-vein may join M,,, (Fig. 56) or M, alone (Fig. 57) instead of intersecting the junction of these veins as in Figure 55. A common venational variation in all species of Oro- peza is the presence of the short segment of the vein M, between the discal cell (cell lst M,) and the junction with the m-cu cross- vein (Fig. 57). This short length of M, has been called the m-cu cross-vein in the nomenclature of Comstock and Needham (see Needham, 1908), who regard the vein M, as an anterior branch of the cubitus. Following this system, Johnson often refers, in his descriptions of new species of Oropeza, to the presence or absence of the short m-cu cross-vein. Within the obscura group, the cross- vein m may cross rather perpendicularly or diagonally from M,,, to M, (compare Figs. 53 and 55), and it may be quite short (Fig. 54). Since these characters may be expected in any sample of several individuals from any locality within the range of a species, and since they appear to have no correlation with geographic distribu- tion and fall easily within the limits of variation of the species, they must be considered as having no taxonomic significance. In the case of coloration in Dolichopeza walleyi already men- tioned, the change from one kind to the other is clinal in nature. Not only would it be difficult to distinguish clearly two geographic races, in this instance, but naming them would serve no useful purpose. Variation may not be such a continuous gradation be- tween extremes, however, but rather a regional preponderance of a particular phenotype, or phase. In Dolichopeza subalbipes, for example, the lateral arms of the ninth tergum of males are of two types, one with a very slightly expanded tip (as figured by Alex- ander, 1942: 213) and the other with a greatly enlarged, or inflated, bulbous tip. In the southern United States, the latter form is domi- THE CRANE FLy GENUS DOLICHOPEZA val nant in all populations, while the narrow tipped form is almost a rarity. In the northernmost states and in Canada, however, the slender tergal arm is much more common. In any one population, both types are likely to occur, so that this cannot constitute either a cline or a subspecific difference. I have seen only a few specimens that seem to be intergrades between the two extremes. On the other hand, if it is possible to demonstrate that the grouping of in- dividuals of a species around two or more means is somehow corre- lated with geographic distribution, then subspecies may be indi- cated. In Dolichopeza polita, for example, characteristics of col- oration, male genitalial structures, and other features of adult and immature forms are generally grouped around three different means, allopatric in distribution but with zones of intergradation at the borders where they meet. These seem to fulfill the requirements of subspecies, as will be discussed later in detail, and it has been found useful in this case to apply names to the three groups. ADULTS—NATURAL HISTORY Longevity.—The adult life of Dolichopeza is quite brief. Rogers (1933: 29) estimated a maximum duration of adult life of tipulines to be three weeks, but I would judge the maximum for Dolichopeza to be little more than half that. Flies kept in the laboratory without food lived from two to four days following emergence, while those that were provided with sugar-water remained alive in some in- stances as long as six days. The cooler temperatures of natural environments would probably decrease the metabolic rate, allow- ing the flies to live somewhat longer, and it seems likely that natural sources of nourishment for both the larva and the resulting adult would be a factor in longevity. Distribution.—Practically the only activities of adult Dolichopeza are those related to reproduction. These are discussed in detail in the sections on mating behavior and oviposition which follow. But the geographical, ecological and seasonal distribution of species of Dolichopeza present certain problems possibly bearing on re- productive activities. It seems not unreasonable to expect that if species are very closely related there might be occasional flow of genes between them, unless they are somehow wholly isolated from each other. Most species of North American Dolichopeza fall to- gether with one or more other species into groups with very close interrelationship, the species in some instances so nearly identical as to be distinguishable only by microscopic examination. This being the case, it becomes important to know if there is cross- 12 THe UNIverRsIry SCIENCE BULLETIN mating between any two recognized species, either regularly, occa- sionally, or at all. If there is any exchange of genetic material of one kind of Dolichopeza with any of its relatives, it must of course take place in the adult stage. Consequently, it is worthwhile to know what opportunities these flies have for cross-mating; that is, where do the adults of the different forms meet and intermingle geographically, ecologically and seasonally? That there is some ecological separation of the species has already been indicated (p. 672). Darkly-colored species are usually taken in deeply-shaded situations, and those having various patterns of dark on buffy yellow are usually found in the more open shade of forest and marsh vegetation. The general environment itself can be divided into three categories, insofar as Dolichopeza is concerned; these are rocky gorges and ravines, forests, and marshes, bogs and swamps. Even where these broad types of habitats meet, as where rocky gorges occur in forests, the species of Dolichopeza demon- strate a marked degree of ecological separation. It is, of course, the microhabitat that determines the presence or absence of a species. It is nevertheless possible to divide the North American species into three groups, based upon the three major kinds of habi‘ats named, for within these habitats occur the micro- environments that meet the particular needs of the species. Char- acteristic of rocky gorges and ravines of the Appalachian Mountains are Dolichopeza americana, carolus, johnsonella, obscura, polita, tridenticulata, walleyi, and, depending on the part of the moun- tains, subvenosa or venosa. Of these, carolus and walleyi are more specifically associated with leafy vegetation along the sides of the ravines, while subvenosa or venosa are found either in such vegeta- tion or together with the other species in nearly any deeply-shaded crevice or cranny. The same association of species is found west of the Appalachians in places where irregularities in the terrain afford conditions similar to those of mountainside ravines, except that subvenosa is replaced by venosa. In cool, mesic forests, the favorite haunts of Dolichopeza obscura, walleyi and sometimes tridenticulata and (within its range) sub- venosa are hollows in standing or fallen trees, cavities beneath out- cropping rocks, undercut banks, or about the roots of windthrown trees, but walleyi may occur also out among the lower plants, where subvenosa, venosa and carolus are sometimes found. Those species encountered most often in bogs, swamps and marsh borders include Dolichopeza dorsalis, sayi, similis and subal- bipes, although such habitats are commonly occupied also by THe CRANE Fiy GENUS DOLICHOPEZA 713 obscura (again seeking out the deepest shade) and sometimes by americana. Data on seasonal distribution of adults are based mostly upon dates of capture. I have not often worked in one locality long enough to learn precisely the order of appearance of each species and how long it was present as adults. Observations made in 1953 at a particular cavity in a rock cliff in Turkey Run State Park, Parke County, Indiana, showed Dolichopeza americana to be the earliest species, appearing on 18 May and reaching a spring peak of population about 26 May. Dolichopeza polita ssp. first appeared on 23 May and reached a relatively high peak, exceeding the num- bers of all other species together, on the 29th and 30th of May. On the 28th of May, D. walleyi and obscura were found with the others, but in much smaller numbers. Four days later, tridenticu- lata appeared, making five species represented in the sample of 142 flies taken on 1 June from that one rather small cavity, which had perhaps ten or twelve square feet of roof surface from which the flies were suspended. All these were taken with a single sweep of the net, and at other times it was possible to take a sixth species in the same place, together with the five named. It will possibly occur to the reader that this is an unusual situa- tion and that localities in which several species of Dolichopeza exist side by side as adults are really uncommon. But the fact is that what I have described for the Indiana locality can be dupli- cated in many parts of the North American range of the genus. (And in Korea, in a rather similar environment, I found four species of Dolichopeza together in crannies of a cliff in early June.) The following data from my own collections will bear out this statement: Ohio, Portage County, Nelson Ledges State Park, 24-25 June 1953: americana, carolus, johnsonella, polita ssp., subalbipes, tridenticu- lata, venosa and walleyi. On another trip to this place I took also obscura. Total: 9 species. Virginia, Rockingham County, Dry River, 6 July 1952: ameri- cana, carolus, obscura, tridenticulata and walleyi. Total: 5 species, in a sample of 12 specimens taken in about ten minutes of collecting just before a rainstorm. North Carolina, Burke County, Linville Falls, 14 June 1958: americana, carolus, obscura, polita ssp., suwbvenosa, tridenticulata and walleyi. Total: 7 species among 27 specimens collected. Georgia, Union County, Neel’s Gap, 10 June 1958: americana, carolus, johnsonella, obscura, subalbipes, subvenosa and tridentic- 714 Tue UNIversIry SCIENCE BULLETIN ulata. In a collection made at this place on 28 June 1952, I took also polita ssp. and walleyi. Total: 9 species. Virginia, Giles County, near Mountain Lake Biological Station, 21 June 1958: americana, carolus, obscura, subalbipes, subvenosa, tridenticulata and walleyi. Other collections at this locality yielded also johnsonella and polita ssp. Total: 9 species. Maine, Cumberland County, near Bridgton, 1 July 1953: ameri- cana, obscura, sayi, subalbipes and walleyi. Total: 5 species among 7 flies of the genus found. West Virginia, Pocahontas County, Droop Mountain Battlefield State Park, 23 June 1958: americana, johnsonella, obscura, triden- ticulata and walleyi all together among the rafters of one picnic shelter. In the same state and county, a few miles away in Watoga State Park, 5 July 1952: americana, carolus, obscura, polita ssp., subalbipes and tridenticulata. Pennsylvania, Luzerne County, Ricketts Glen, 10 July 1952: americana, carolus, johnsonella, obscura, polita ssp. and tridenticu- lata. In four different collections, here, in three different years, I have never found walleyi, but I am confident it will be added to the list. Ohio, Hocking County, “Neotoma” (a small valley near Rock- bridge belonging to Dr. Edward S. Thomas of the Ohio State Mu- seum), 30 May 1952: americana, carolus, obscura, polita ssp. and tridenticulata. On other occasions, I have also taken within this limited area dorsalis, johnsonella, subalbipes, venosa and wal- leyi, making a total of 10 species. Great Smoky Mountains National Park (collections in Sevier County, Tennessee, and Swain County, North Carolina), 30 June 1952: americana, carolus, obscura, polita ssp., subalbipes, sub- venosa and tridenticulata, Also taken in the park are dorsalis, john- sonella and walleyi. Total: 10 species. Michigan, Washtenaw County, Mud Lake Bog (near Whitmore Lake), various dates: americana, dorsalis, obscura, sayi, similis, subalbipes and walleyi, often four or five species in any one collec- tion. Michigan, Livingston County, Edwin S. George Reserve (a natural area belonging to the University of Michigan, near Pinck- ney), various dates: the same species list as for Mud Lake Bog, except that venosa is added. Total: 8 species. Wisconsin, Juneau County, Rocky Arbor Park (near Wisconsin Dells ), 6 July 1950: carolus, obscura, polita ssp., subalbipes, triden- ticulata and venosa. Total: 6 species. THe CRANE FLY GENuS DOLICHOPEZA (ls Kansas, Douglas County, 15 miles south of Lawrence, 30 August 1957: obscura, polita ssp. and walleyi. These three species are from a locality that is probably near the western limit for the genus, at this latitude. A great many more collection records similar to those presented here, particularly from the Appalachian region, could be men- tioned, but these suffice to illustrate the point that where one species of Dolichopeza occurs, one or more others are likely to be found with it. With two exceptions (similis and venosa), all the species of Dolichopeza that occur in eastern North America appear to have two generations per year in that part of their range lying roughly between the latitudes of northern Florida and New England. In Florida and southernmost Georgia, there are records of some species for nearly all months of the year, suggesting that the life cycle is repeated as often as biologically possible, with perhaps a dormant period in January and February of some years. While there seems to be no particular peak of abundance of individuals of any species at any season, in the Florida region, most of the collection records indicate that early summer (June) is a time when the numbers of most of the species are increased. From Georgia northward, par- ticularly at higher elevations, there occurs a marked two-generation cycle, annually. In general, the peaks of emergence of adults come in late May and early June and again in August, but early and late emergence of individuals give considerable slope to these peaks. Available collection records from the northern edges of the range of the genus are insufficient to give a clear picture, but they suggest that a single mid-summer generation is the usual thing. Further collecting is needed especially in Canada to verify this. Dolicho- peza similis and D. venosa are northern species, found only in the northeastern states and Canada, with a flight period from late May to early August, nearly all the records being for June and the later records being for the more northern parts of the range. More specific data on dates and localities of collections of adults will be found listed in the sections on the species concerned. Activities Certainly the most spectacular activity of adults of Dolichopeza is their peculiar dancing flight. This is most easily observed in species of the obscura group because, somewhat re- stricted to certain niches by their reaction to the intensity of out- side light, they will leave their darkened shelters by day only when greatly alarmed. Dropping from its resting position, the fly moves 716 THe UNIVERSITY SCIENCE BULLETIN in a rather elliptical path, first downward and backward, then rising and forward, keeping the head oriented more or less toward the original point of suspension. This cycle is repeated with such rapidity that the fly seems almost to fade from view, although the dance is limited to about a three-inch ellipse. The dance of Doli- chopeza carolus is more conspicuous, mostly, I suppose, because of the paler colors of that species, which contrast somewhat with its leafy habitat. When disturbed from a distance, carolus moves in a more vertical ellipse, often six inches in length; but if ap- proached closely, it usually moves away rather directly and hur- riedly. It is my impression that the other species that are found most commonly in vegetation also take readily to flight when dis- turbed. This is particularly true of dorsalis and sayi, and I have never observed the elliptical flight in either of these species. Among the species of the deeply shaded, rock gorge habitat, those of the subgenus Oropeza seem much more restless than D. americana; they are more easily alarmed to the point of dancing and are slower to settle back to rest, and in the same respects, males are more rest- less than females. Spontaneous dancing during twilight formation of swarms, possibly related to mating activities, has been noted in Dolichopeza johnsonella. The dance of this species in the open air was observed to be nearly straight up and down, along a fairly constant path of about twenty inches in length, and the flight was markedly slower than that of the diurnal dance beneath a rock ledge. During the daylight hours, the primary occupation of adults of Dolichopeza is resting in such shaded and protected haunts as are available within the general habitat. In repose, the flies almost invariably hang from some overhead support, such as the bottom of an outcropping rock, an undercut bank, the roof of a darkened drain tile or culvert, stems or leaves of low plants, or the eaves of buildings situated among trees. It has been observed that the number of legs used in suspension varies. Alexander (1919: 929- 930) states that a point of distinction between the subgenera is that “Oropeza hangs . . . with only the fore legs attached Dolichopeza, on the contrary, has the four anterior legs on the support, the hind legs dangling. . . .” I think this was a slip of the pen, for while one may observe considerable variation in the resting postures of both subgenera, it is Dolichopeza ameri- cana that is nearly always seen suspended by the two front legs only, with the four other legs held out to the sides, perhaps dangling THe CRANE Fiy GENUS DOLICHOPEZA walid) a bit (Fig. 67). Species of Oropeza, on the other hand, usually hang by both the prothoracic and mesothoracic legs (Figs. 68-70). Variations involving three to six legs are not uncommon in both subgenera. In fact, among more than a hundred Dolichopeza (O.) polita ssp. observed in an Indiana locality on two days in late May, if 70 69 Fic. 67. Resting posture of Dolichopeza (Dolichopeza) americana. Fic. 68. Resting posture of Dolichopeza (Oropeza) polita ssp. (also of johnsonella, obscura and tridenticulata). Fic. 69. Resting posture of Dolichopeza (Oropeza) carolus (also of subalbipes, subvenosa and ve- nosa). Fic. 70. Resting posture of Dolichopeza (Oropeza) sayi (also of dorsalis). 718 THe UNIVERSITY SCIENCE BULLETIN nearly all were suspended by all six legs, whether from an upside- down horizontal surface or on a more or less vertical surface, and whether single individuals or mating pairs. Commenting on this behavioral difference, Alexander (1920:982) says “the resting positions of Dolichopeza are described . . . and the striking dissimilarities to Oropeza noted. It may be that Oropeza is not so close to Dolichopeza as has been believed.” Certainly these two subgenera represent distinct natural groups, but the difference in behavior appears to be no more than a consequence of a morpho- logical dissimilarity that is much less remarkable than the difference in genitalia or wing venation. Comparing several specimens each of D. (D.) americana and D. (O.) tridenticulata, species of about equal size, I found the ratio of the length of the prothoracic leg to that of the mesothoracic leg to be greater in the former species. In both these species (as also in other species of Oropeza), the prothoracic legs slightly exceed the length of the second pair, but the difference in total length is perhaps significant. The ratios for americana and tridenticulata were 30.5:27.9 mm. and 380.7:29.0 mm., respectively, there being a greater difference in americana that might explain its peculiar posture. The position of the wings in repose is of interest, as there is some correlation with the grouping of species on the basis of col- oration, noted earlier. It has been stated that Dolichopeza ameri- cana rests with its wings outspread and that species of the subgenus Oropeza fold their wings over the body when at rest. But this is true only in part. Dolichopeza americana (Fig. 67) does rest with its wings apart, a posture that is unique among those species pre- ferring the deep shade, but it is my observation that the swamp inhabiting D. dorsalis and sayi nearly always rest with their wings outspread and tilted somewhat forward, as they hang among the grasses and stems (Fig. 70). Dolichopeza johnsonella, obscura, polita sspp., and tridenticulata, flies which seek the deepest shade by day, all fold their wings over their backs and rest with the hind legs dangling out to the sides (Fig. 68). D. walleyi is also often seen in this position, especially when concealed in dark recesses, but when out among low plants beneath the trees, it often rests with the wings outspread, a peculiar correlation of variable be- havior with different environment. D. similis, a swamp dweller usually found among low plants, also exhibits considerable varia- tion in wing position when at rest but usually assumes a resting posture like that shown in Figure 68. A most unusual situation in regard to resting posture is that of Dolichopeza carolus, subal- THE CRANE FLY GENUS DOLICHOPEZA 719 bipes, subvenosa and venosa. Morphologically, subalbipes belongs to the obscura group and the other three species to the sayi group, yet in general coloration and size they are very similar. The com- monest environment of each is leafy vegetation in damp, shaded places, although usually no more than two of the four species occur together in the same habitat. The resting postures of these very similar species are identical and at the same time different from those of all other North American species of Dolichopeza. They hang by the front two pairs of legs, with the wings folded over the body and the hind legs held almost straight downward and quite close together (Fig. 69), a position that conceivably has some camouflaging effect, judging from its occurrence throughout this group of convergent species. Reaction of resting flies to changes in light intensity has been noted a few times. Rather regular observations were made of adult Dolichopeza congregated in a recess in a sandstone cliff at Turkey Run State Park, Parke County, Indiana. In late May, this hole is well shaded through most of the day and often contains as many as three hundred flies of various species. About five o’clock in the afternoon, however, sunlight strikes the sandy floor of the cavity and by reflection somewhat illuminates the roof, causing the resting flies after a time to move. Under such circumstances, they usually retreat deeper into their crannies and in this case crowded into a small adjoining hole that was only about eight inches high, with perhaps two square feet of roof. In this small but dark retreat I once found about two hundred flies, that is, nearly a hundred per square foot of area available for suspension. At other times, I found that when I reflected sunlight into resting sites, species of Oropeza reacted readily, while D. americana responded slowly or not at all. When the sky becomes dark with clouds, in the day- time, many Dolichopeza leave their shaded hiding places and fly about among the low plants of the woods, but I have also noticed that at such times those remaining at rest are less easily disturbed by the presence of humans than at other times. For example, on one dark, cloudy day resting flies showed no reaction to the presence of my hand held only half an inch from them or slowly circled around them. They took flight immediately, however, in response to a slight touch. Although Dolichopezas rest quietly during the day, they range rather widely by night. Along McCormick’s Creek, in Indiana, I have watched johnsonella and other species of the obscura group, on a June evening, moving from beneath the limestone ledges out 720 Tue UNtversiry SCIENCE BULLETIN into nearby trees. In the gathering dusk, they could be seen dancing in mid-air among the trees in the ravine and high up on the face of the cliff. At other times and places, I have observed D. americana and polita ssp. still at rest a few minutes after dark- ness, but there is every indication that most Dolichopeza are away from their daytime haunts during most of the hours of darkness. It is a common thing to find crannies where hundreds of them rest by day to be quite empty after nightfall. At daybreak, along forest trails, one may find the flies making their way back to darkened shelters. In northern Michigan, I have waited beside an empty day- time resting place of tridenticulata, in the early morning, and watched the flies returning from all directions as the sun rose. Further evidence of nocturnal wandering by Dolichopeza is found in captures of occasional specimens at light traps (notwithstanding their tendency to avoid the brighter light of the sun), or the finding of flies in temporary structures affording shade from the rising sun (such as tents in a campground), in places far from their diurnal haunts. It is possible that during such excursions away from the protection of daytime habitats, individuals are caught up in high winds and carried to other localities, there to establish new popula- tions of the species. Although the fairly well developed digestive system of the adult Dolichopeza suggests that the flies do some feeding during their short existence, | have only once seen one so occupied. This was Dolichopeza americana, seen drinking at a shallow pool of seepage water on a mossy, horizontal ledge of sandstone. The fly stooped, somewhat in the manner of a drinking giraffe, with its head down and tail end elevated, its legs rather evenly outspread. I have no idea what natural foods are utilized, but in the laboratory the flies fed upon thin sugar syrup, which was held in shreds of paper tissue hung upon the sides of the cage. Speaking of crane flies generally, Rogers (1933: 30) says, “Nearly all species that have been reared were able to mate and oviposit without food, and, although feeding practically doubles the life of the adults and tends to prolong some- what the period of oviposition, lack of food does not appear to de- crease the number of fertile eggs oviposited.” While my records do not show a doubling of adult life time among flies given food, an increased life span is indicated; and I find no effect of food or lack of it on the fertility of eggs or the period of oviposition by female flies in captivity. Predators and parasites—There are no specific records, so far as I am aware, of predation on Dolichopeza by a vertebrate or large invertebrate, although Alexander (1920: 721 ff.) presents a for- THe CRANE Fiy GENUS DOLICHOPEZA 721 midable list of known enemies of crane flies generally, many of which are certainly potential and probably actual predators of Dolichopeza. While probing the darkened recesses occupied by these flies, I have occasionally encountered a small salamander or a toad; and it is not an uncommon sight to see a large dragon fly working along a cliff, now diving into a shaded niche, now backing out and flying to another. I once found an ant attacking an emerging Dolichopeza obscura that was only partly out of its pupal skin. Swamp species of Dolichopeza are undoubtedly picked up by foraging birds, especially at times when they constitute a large fraction of the flying insects present. It has been remarked that some species of Dolichopeza have a notable predilection for resting on spiders’ webs.” I would rather say that they often find their chosen resting places occupied also by spiders, from whose webs they are able to hang without often becoming entangled. That Dolichopezas do fall prey to web-spinning spiders, as well as to cursorial ones that stalk them in their crannies, is shown by the following list of spiders collected with their prey: << Dolomedes sp. (immature) with Dolichopeza (O.) obscura; Michigan, Che- boygan Co., Douglas Lake, 6 July 1949. Philodromus marxii Keyserling with Dolichopeza (O.) polita; Wlinois, La- Salle Co., Starved Rock State Park, 7 July 1951. Meta menardi Latreille (immature) with Dolichopeza (O.) tridenticulata; Indiana, Owen Co., McCormick’s Creek State Park, 24 June 1950. Theridion tepidariorum C. Koch with Dolichopeza (O.) polita; Michigan, Eaton Co., Grand Ledge, 16 August 1951 (three flies in one web). Theridion tepidariorum with Dolichopeza (O.) tridenticulata; Mlinois, La- Salle Co., Starved Rock State Park, 7 July 1951. Theridion tepidariorum (immature) with Dolichopeza (O.) johnsoneila; Indiana, Owen Co., McCormick’s Creek State Park, 20 July 1951. Theridiosoma radiosum McCook with Dolichopeza (O.) tridenticulata; In- diana, Owen Co., McCormick’s Creek State Park, 12 June 1951. Theridiosoma radiosum with Dolichopeza (O.) obscura; Indiana, Parke Co., Turkey Run State Park, 19 June 1950. Theridiosoma radiosum with Dolichopeza (O.) polita ssp. and Dolichopeza (O.) sp.; Indiana, Parke Co., Turkey Run State Park, 11 June 1951 (three records ). Uloborus americanus Walckenaer with Dolichopeza (O.) tridenticulata; Iowa, Clayton Co., near Guttenberg, 8 July 1951. The following portion of a letter from Dr. Willis J. Gertsch, who identified all the spiders, is of interest: “The Dolomedes is a three- clawed vagrant spider which ordinarily runs over the ground or low vegetation, most often near water. The Philodromus is also a vagrant and may be seen on the ground, but is rather more fre- quently swept from herbs and shrubs. The remaining species are all sedentary forms. The Theridion is the very common cosmo- 722 Tue UNIversity SCIENCE BULLETIN politan house spider. The Theridiosma spins an unusual orb web and usually is associated with shaded woods. The same is true of Meta, also an orb weaver which favors dark woods and is even partial to caves. Uloborus is one of the cribellate orb weavers and makes a beautiful horizontal web.” It so happens that all the records above involve flies of the obscura group that frequent rocky crevices and crannies. This fact is re- lated partly to the greater ease with which predation is observed in such habitats and partly to my confessed preference for working in these more comfortable places. I have no doubt that the swamp- land species of Dolichopeza have many spider predators also. It should be added that americana, carolus, walleyi and similis have been found in the webs of spiders, but in these instances I was unable to find the spiders. While external parasites probably seldom kill a crane fly, they do sometimes cause their host to become shriveled and undersized (Rogers, 1942: 57). All species of Dolichopeza are occasionally afflicted with mites, and as many as eight or ten on a fly is a com- mon thing. The heaviest infestation I have found was a group of 98 mites clustered mostly on the abdomen of a male of Dolicho- peza venosa, taken in Eaton County, Michigan. The mites usually attach to some part of the thorax, often to the abdomen, and rarely to proximal parts of the legs, but I have never found them attached to the head or wings. Two kinds of mites are the most frequent ectoparasites of Dolichopeza. One is a plump, reddish or orange colored erythraeid (Acarina: Erythraeidae) of undetermined genus. The other is a somewhat flattened stigmaeid, identified by Dr. Joseph Camin as belonging to the genus Ledermiilleria (Acarina: Stig- maeidae). Actually, there may be two species of this genus para- sitic on Dolichopeza, as those on some flies are of a rusty orange color, while others on other flies are mottled tan or gray. Their body form is depressed, the dorsum broadly curved and the ventral surface slightly concave. Only one instance of nematode endoparasitism of adult Doli- chopeza has come to my attention. This was a male of Dolichopeza tridenticulata, collected in Shenandoah National Park, Virginia, on 28 June 1958. Although the total body length of the fly was only 11 mm., the nematode measured almost 35 mm. in length. The parasite was looped from end to end of the fly's abdomen four complete lengths, and its head was inserted far into the thorax of its host. In spite of this condition, the fly was taken on the wing, THe CRANE Fiy GENuS DOLICHOPEZA 723 and its digestive and reproductive organs, on dissection, appeared intact and capable of functioning. On two occasions, I have found larval Tachinidae endoparasitic in adult Dolichopeza. Dissection of the abdomen of a male Doli- chopeza carolus taken in Neel Gap, Union County, Georgia, 28 June 1952, revealed two small dipterous larvae identified as tach- inids. On 6 June 1960, at Cumberland Falls State Park, Whitley County, Kentucky, I captured a female Dolichopeza tridenticulata with a peculiarly bulging abdomen. Unfortunately, I did not notice this enlargement until after the fly had been killed in the cyanide jar and the 3.5 mm. long tachinid parasite was forcing its way out of the abdominal cavity through the pleural membrane. Dissec- tion showed the vital organs of the host intact, so it seems likely that the tachinid had fed chiefly on the fat tissue and perhaps the haemolymph. MATING BEHAVIOR It is probable that in most species of Dolichopeza mating usually occurs during the night, although in certain species pairs are not uncommonly found still in copulation well into the following day- light hours; and occasionally matings are commenced during the day, especially if the light intensity is low, as when the sky grows dark before a summer thunderstorm. As suggested earlier, late evening swarming noted in a few species may be a preliminary to mating, but it is not possible at this time to say what pre-mating activities are typical for the genus under natural conditions. Ob- servation of diurnally mating pairs and those artificially illuminated in the laboratory provides the only exact information on mating be- havior, the basis of the following descriptive comments. Attracted to the same shaded daytime resting sites, males and females are brought into close proximity but appear to ignore one another unless actual physical contact, usually of tarsi, occurs. Now and then, when a resting group of flies is disturbed and is set into dancing motion, a male may attempt copulation with one or more females before settling to rest again. Unreceptive females, in the case of Dolichopeza polita ssp., have been seen to respond to these attentions by fluttering their wings and kicking their hind legs over their backs. If a female is receptive, however, she moves her wings slightly apart and remains quietly suspended, while the male hovers over her, curves his abdomen down alongside hers, and reaches from below or behind to gain a grip upon the ovipositor with his dististyles (Fig. 71). I have observed these reactions, with minor variations, also in Dolichopeza americana, sayi, tridenticulata 724 THE UNIVERSITY SCIENCE BULLETIN and walleyi and suspect that they are rather uniform throughout the genus. During copulation, the female ordinarily supports the weight of both flies, as the male hangs head downward, held fast by the grip of his genital appendages (Fig. 72). Seen from the side, the legs of the flies are not arranged generally in one plane as they might seem from the figure. The fore and hind legs of the male are held out somewhat dorsally with respect to the fly’s body and the meso- thoracic legs somewhat ventrally. The hind legs of the female (and the middle legs, in the case of americana) are deflected in a similar manner. In americana, the wings of both partners are held out to the side, but in Oropeza spp. the female normally keeps her wings folded over her back, while the male holds his wings out- spread. Because of the manner of coupling, suspended mating pairs are so oriented that the dorsal surface of the male and the ventral surface of the female face the same direction. If both members of the pair are clinging to some support, as often happens, their abdomens are variously revolved to maintain the proper clasp. Each may rotate the abdomen 90 degrees, or one may twist 180 degrees, etc. The following is the usual arrangement of the copulatory struc- tures in mating: The male grasps the female genital segments from below or behind, his outer dististyles getting a loose hold about the tenth tergum. When the hypovalves of the female have been maneuvered into the male’s genital chamber, he closes the tips of the inner dististyles into sclerotized folds on the dorsal or inner surfaces of the hypovalves. This provides the firm grasp that is so strong as to allow mating pairs to take to the wing, when alarmed, and not become separated. It may be seen in Figure 73 that the pressure of the tergal arms on the edges of the hypovalves serves to brace them securely against the underside of the ninth tergum of the male, while the gonapophyses force the tenth tergum and cerci apart from the hypovalves. This wide separation of upper and lower elements of the ovipositor exposes the opening to the bursa copulatrix, into which the penis is inserted, as indicated by dotted lines in Figure 73. In Dolichopeza americana, the absence of tergal arms seems compensated by a shift dorsad in the position of the inner dististyles, which hold the hypovalves well apart from the cerci. The specific configuration of the medio-posterior margin of the ninth tergum of the male does not seem to have any functional significance. Although I have never observed males waiting beside female pupae, as described by Alexander (1919: 881) for several other THE CRANE FLY GENUS DOLICHOPEZA Fic. 71. Position of male (shaded) and female (not shaded) Dolichopeza (Oropeza) sp. preliminary to mating. Fic. 72. Same as 71 but showing position in mating. Fic. 73. Detail of hypopygia of male and female Dolichopeza (Oropeza) sp. in mating. Fic. Stages in development of egg of Dolichopeza (Oropeza) sp.; a-c—im- mature eggs, showing cap of cells from which terminal filament de- velops, d—mature egg, showing coiled terminal filament. Fic. 75. Egg of Dolichopeza (Oropeza) sp. with terminal filament uncoiled. Fic. 76 same as 75, enlarged, showing micropyle. Fic. 77. Egg of Dolichopeza (Oropeza) polita polita, showing vestigial terminal fila- ment. Fic. 78. Hatching of egg of Dolichopeza (Oropeza) sp. (scale pertains to Figs. 78 and 79 only). Fic. 79. Larva of Dolichopeza (Oropeza) obscura, first instar, 48 hours after emergence from egg. 725 726 THE UNIVERSITY SCIENCE BULLETIN genera, I do believe that females are somehow found by males and mated with very soon after emergence. Mating pairs frequently involve teneral females whose abdomens are still soft and show the greenish color of the immature stages. As emergence almost always occurs during the hours of night, it is somewhat of a problem to understand just how males and females come together (compare with above description of daytime mating activities). To learn more about nocturnal behavior of these flies, I have made many trips to known larval and pupal habitats, and, using an electric lan- tern, watched for emergence and subsequent mating. In no case did I find mating to take place near the pupal site, so I assume that some amount of flight is involved in bringing the two sexes together. Mating appears to take place usually at some distance from diurnal resting sites, for these are nearly always vacant at night. In north- ern Michigan and in Indiana, however, I have found many mating pairs of Dolichopeza tridenticulata beneath the eaves of sheds, at about midnight, where earlier in the day only single flies could be found. In dimly-lit rock gorges and other similar places, one may find mating pairs, here and there, nearly any time of day, in season; but almost all pairs of the majority of species were observed to separate by mid-morming. Mating pairs of americana and polita sspp., how- ever, are often found during the daylight hours. In Clifty Ravine, Jefferson County, Indiana, on 3 August, I found hundreds of mating pairs of americana, at nine o'clock in the morning, when the hazy sunlight had not yet become bright through the trees. Almost no unmated individuals were seen. This same situation was found again, about noon of the same day, in Muscatatuck State Park, Jen- nings County, Indiana. At Nelson Ledges State Park, Portage County, Ohio, dozens of mating pairs of polita sspp. were collected in the hour just before noon on 14 July. The duration of mating is not known, but the pairs of americana and polita mentioned seemed to have been formed during the pre- vious night, that is to say at least six to eight hours before the time of observation. In the laboratory, pairs of walleyi and sayji re- mained in copulation for as long as eight hours, and under natural conditions a pair of tridenticulata was observed together at sunrise and for three hours thereafter. It is probable that mating takes place throughout the egg-laying period, which lasts about three or four days, perhaps longer in some females. When flies were confined to close quarters in the laboratory, mating sometimes occurred on three successive nights. THE CRANE Fiy GENUS DOLICHOPEZA Fig Females from mating pairs collected in the field frequently are found to contain only a few eggs, suggesting at least one prior mating. It would seem that among closely-related species living so near to one another that there might occur frequent matings or attempted matings in which the partners are of different species. However, I have never observed this to happen. Among many hundreds of mating pairs observed or collected, not one was found to involve two species. A male of Dolichopeza carolus was once found ap- parently attempting to gain a copulatory hold on a female of ameri- cana that was already mated with a male of her own species; and I have seen, in dense, dancing swarms of several species, males very briefly approaching females of another species but never going so far as to secure a grasp with the hypopygium. It is understandable that size differences would serve as a barrier to cross-mating, in some cases, as between males of polita and females of americana often found sharing the same resting sites. But among adults of many other species that come into actual contact or near contact there are no such obvious barriers to cross-mating. The genitalial clasp is general in nature, far from the complex “lock and key” interrelationship described for some insects, and appears to be essentially the same among many species observed. But it is more significant that cross-matings seem in general not even to be attempted. Accordingly, it must be that by means of some sense as yet unknown the flies are attracted to mates of their own kind and either repelled or at least not attracted by other species, no matter how closely the two may apparently be related. OVIPOSITION In the laboratory, oviposition occurred almost without exception during the night and was never observed by day. Neither have I ever seen females of Dolichopeza in their natural surroundings be- having during daylight hours as if they were laying eggs. How- ever, on a few occasions females of polita, tridenticulata and walleyi were seen ovipositing during late evening, just before the dark of night. These females were flitting up and down cliff faces, flying very close to the thin carpet of mosses that covered the rock. After a time of seemingly random movement (it is possible certain mosses were being sought), a given female fly would commence a series of bouncing motions over an area of moss perhaps four inches in di- ameter. The bouncing action is accomplished by rapidly alighting on the tips of the tarsi, flexing the legs slightly, and pushing off into 728 Tue UNIversiry SCIENCE BULLETIN flight once more, the wings vibrating all the while. As the fly thus bounces, at about 90 times per minute, it occasionally and quite suddenly thrusts the tip of the abdomen deeply into the moss, with- drawing it almost immediately. The fly may repeat this motion as many as 25 or 30 times before departing to another place. The activity associated with oviposition in these three species was quite uniform, and it therefore seems that it may be rather similar in the other species, allowing of course for some modification due to differences in the kind of habitat, as for example whether the moss surfaces are vertical or horizontal. In the mosses, the eggs are found to be laid singly but several in a small area; that is, while several may be found within a space an inch in diameter, they do not occur in any sort of cluster, such as has been described for some other Tipulinae. In the laboratory, also, eggs are deposited singly, except when the available substrate is so limited as to force females to lay their eggs in groups. EGGS Matured eggs of all species of North American Dolichopeza have the same general appearance: oblong, slightly narrower at the posterior end, and with a chorion of intense black color, with a bluish or purplish, almost metallic sheen. The chorion is so smooth that at more than 200 magnification there is no indication what- soever of any surface sculpture. The position of the micropyle is indicated by an indentation of the chorion on the ventral surface, near the anterior end (Fig. 76). Orientation of the egg is based upon the usual position of the larva at the time of hatching. Within the ovarioles, the developing eggs appear as buff colored, granular, oblong bodies. The smooth chorion is found only on those eggs which have moved into the lateral oviducts or are in the ex- treme lower ends of the ovarioles. Pale colored at the time of emer- gence of the adult female, the eggs soon darken, so that in a female only a few hours old the lateral oviducts are bulging with black masses of fully-formed eggs, all lying more or less horizontally, with their posterior ends directed towards the gonopore. Eggs of some species of Dolichopeza bear a terminal filament of great length, the function of which seems to be retention of the egg in the moss where it is laid. This structure originates from a whit- ish cap of cells upon the end of the developing egg (Fig. 74) and by the time of oviposition has the form of a springlike, broadly conical, tight coil. Drawn out to its full length, it is seen to be from Tue CRANE FLy GENuS DOLICHOPEZA 729 ten to fifteen times the length of the egg itself. I have not been able to correlate the presence or absence of the terminal filament with either the type of larval habitat or the phylogenetic affinities of the species. It is present in Dolichopeza americana, in johnson- ella, obscura and tridenticulata of the obscura group, and in sayi of the sayi group. It is totally absent in subalbipes of the obscura group, and in carolus, dorsalis, similis, subvenosa, venosa and walleyi of the sayi group. Curiously, in all three races of polita it is present in a very rudimentary form, although recognizably different in each subspecies. In the typical subspecies (Fig. 77) the filament usually forms one open coil or loop; in the western subspecies it has the shape of a hook and never forms a complete coil; and in the central race it forms one or two irregular coils in a plane perpendicular to the basal part of the filament. In attempting to find out what numbers of eggs are laid (or are probably laid) by the various species, I have dissected females that were collected into alcohol at a time when they seemed to be still rather teneral and quite filled with eggs, even into the first abdominal segment. In such females, it is difficult to imagine where more eggs would be placed if there were any more, so the count of matured and nearly matured eggs present was taken as the num- ber which the fly would probably have oviposited. In nearly-spent females, perhaps only one or two matured eggs may be found in the oviducts, while the very much undeveloped eggs in the ovari- oles are still in place. Thus, there seems not to be a continuing development of eggs throughout the period of oviposition, but rather a brief initial period of maturing of all those eggs which are going to be laid at all. Egg counts in females of several species from different parts of the country are surprisingly constant, vary- ing in the vicinity of 100 to 140, depending on the species. Rogers (1933: 32) gives the number of eggs actually laid as 250 to 350 for the Tipulinae in general, and, in an unpublished paper, the num- bers 200 and 250, more or less, for two individuals of Dolichopeza obscura from Florida. However, my examination of another Florida female of obscura, in which all eggs seemed to be still in place, revealed 104 eggs in advanced stages of maturity and no more than three dozen countable eggs in the ovarioles. A very teneral female tridenticulata from Jefferson County, Indiana, contained 90 fairly darkened eggs and about 50 in various developmental stages. From a female americana, 106 matured eggs were taken, and there were not many others that looked as if they might eventually have 25—5840 730 Tue University SCIENCE BULLETIN been laid. From all the indications available to me, I would esti- mate the potential egg production of an average female of Dolicho- peza to be about 120 eggs. Dimensions of the eggs vary somewhat proportionally with the body size of the individual female and only very generally with the average body size of the species. The smallest eggs measured were of the small species, Dolichopeza americana and. tridenticulata, while the largest eggs found were those of similis, the largest Doli- chopeza. Average dimensions of eggs from average sized females were: americana—.60 x .25 mm. similis—.92 x .35 mm. carolus—.79 x .37 mm. subalbipes—.83 x .36 mm. dorsalis—.76 x .31 mm. subvenosa—.75 x .31 mm. johnsonella—.74 x .29 mm. tridenticulata—.65 x .30 mm. obscura—.73 x .40 mm. venosa—.79 x .32 mm. polita ssp.—.75 x .30 mm. walleyi—.76 x .32 mm. sayi—.72 x .31 mm. Measurements were made with a camera lucida scale from eggs mounted on slides. Because the eggs are likely to be somewhat flattened when so mounted, the widths given here are subject to slight error. Duration of the egg stage in Dolichopeza has been reported only once (Rogers, 1933: 32), the time given being 13 to 16 days, for D. obscura at outdoor temperatures in March and April, in Florida. However, among several groups of eggs kept at both room and out- door temperatures, in Indiana, Michigan and Minnesota, I have found the time to be closer to eight days. My observations are summarized in Table 2. TABLE 2.—Duration of the Egg Stage in Some Species of North American Dolichopeza Date of Duration of Species hatching egg stage Where kept WINCHICONG LE oe. ot 19 Aug. 7-8 days laboratory CATOWUSHA PE. PS TY 24 June 9 days outdoors (Indiana ) dorsalisy Perrin awa I. 22 July 9 days outdoors ( Minnesota ) johnsonella ......... . 29 June 8-9 days outdoors (Indiana ) OUSCUTO Mas Pn Even 2 Aug.? 10 days? — laboratory DOIG USSU as at eee: 5 24 Aug. 7-8 days laboratory DOUME SSDs 0 3. age 6 Sept. 7-8 days laboratory SAGA I ee ih tht moe 2 Cr Sept: 6-7 days laboratory CO ae oe ce ... 26 Sept. 7 days laboratory RIES Petra t: ligty docs. VA's 6 25 June 7 days laboratory PMI WE os ite as ak Lisa 8 June 7 days laboratory TOUS Mee rest en es 9 June 7-8 days laboratory THe CRANE Fry GENUS DOLICHOPEZA 731 Durations are given from time of oviposition to date of abundant hatching. Where the length of the egg stage is indicated as uncer- tain, the date of oviposition of the first eggs to hatch is not known but is one of two known dates. It will be noted that outdoor times are in all cases the longer, except for obscura, which was in an un- heated building in the woods of northern Michigan. Laboratory or room temperatures seem to accelerate development by about one day. LARVAE—MORPHOLOGY First instar larva —As it comes from the egg, the larva of Dolicho- peza seems to be mostly head capsule and caudal setae (Fig. 78); the body is extremely short and transparent, and only the tips of the mandibles are at all strongly sclerotized. Within three or four hours after hatching, however, the larva stretches out to somewhat less than two millimeters in length, exclusive of the caudal setae, and the head capsule, darkened to a grayish color, becomes dis- cernible in some detail. At first nearly rectangular in outline when seen from above, the head assumes its more oval shape in about a day. The integument of the very young larva becomes a dirty yellowish color, yet is transparent enough that the viscera show clearly through it; and it bears an inconspicuous covering of ex- tremely minute hairs, in addition to the weak bristles indicated in Figure 79. There are several details in which the head of the first instar larva (Fig. 80) differs markedly from that of later instars. Perhaps the most conspicuous of these is the antennal structure. In later instars, the antennae are cylindrical, with an apical sensory peg, but in the first instar they are setiform. The fronto-clypeal plate is rela- tively larger, and the backward extensions of the lateral plates that parallel it in later instars are absent in the very young larva. Al- though not studied in more than a few species, the mandibles, mentum and hypopharynx generally have fewer and more rounded teeth than their counterparts in the more mature larva. In the heavily sclerotized head capsules of older larvae, I had long over- looked the eyes, but their position in the first instar larva is clearly indicated by concentrated spots of violet-purple pigment similar in color to that associated with the compound eyes of the adult flies. The spiracular disc of the first instar larva contrasts even more remarkably with those of all subsequent instars; in fact, except for the presence of spiracles, it has almost no characteristics in common with them. Below the spiracular openings are four sclerotized 732 Tue Universiry SCIENCE BULLETIN ant E € 8 dsgl a sgl oes tr © py gc 4 vn aw ic Si ° rct Fic. 80. Head capsule of first instar larva of Dolichopeza (Oro- peza) sp.; ant—antenna. Fic. 81. Spiracular disc and associated structures of first instar larva of Dolichopeza (Oropeza) obscura. Fic. 82. Gross internal anatomy, digestive and respiratory systems only, of fourth instar larva of Dolichopeza (Oropeza) obscura; dsgl— duct of salivary gland, ge—gastric caecum, ic—intestinal caecum, li— large intestine, mp—Malpighian tubule, oes—oesophagus, py—pro- ventriculus, ret-—rectum, sgl—salivary gland, si—small intestine, tr— trachea, vn—ventriculus. Fic. 83. Fourth instar larva of Dolichopeza (Oropeza) obscura, indicating thoracic and abdominal segmentation; th—thorax, abd—abdomen. THE CRANE FLY GENUS DOLICHOPEZA 733 plates, a dorsal pair narrowly fused at the midline and a ventral pair of broader plates, the medial borders of which are indistinct. Each of these plates bears a long, slender caudal seta on its distal margin (Figs. 79 and 81). Across the top of the spiracular disc are four broadly-rounded, low protuberances, each with a tuft of bristles at its apex. Two similar tufts of bristles are borne on each of the ventral plates, in the positions shown in Figure 81. It is diffi- cult to homologize these various structures with those seen on the spiracular disc of the older larva. However, it seems not unlikely that the four dorsal protuberances correspond to the four conical dorsal lobes found in the same position in the later instars. Either pair of sclerotized plates may be homologous with sclerotized, bristle-bearing areas of relatively much smaller extent on the ven- tral lobes of the spiracular disc of the more mature larva; but I am more inclined to believe that it is the lower pair, which means the upper pair has no counterpart in the older larva. The gen- eral similarity of the spiracular disc of the very young Dolichopeza larva to that of the later stage larvae of some Limoniinae (such as Limnophila and Pseudolimnophila) is striking. The very long caudal setae and tufts of bristles possibly serve to hold the respiratory openings above the water surface in case the larval habitat becomes saturated. Similar structures are found in the first instar as well as later stage larvae of the more aquatic rela- tives of Dolichopeza, such as Megistocera (Rogers, 1949:9); there- fore, it may be that these bristles and setae have no function in Dolichopeza but are merely rudiments inherited from ancestors that lived in wetter habitats. I have observed that some larvae lose part or all of one or more caudal setae and apparently are none the worse for it. With the first molt, the tufts of bristles and the sclerotized plates and their setae are all lost. Fourth instar larva—Larvae in the second, third and fourth instars differ mostly in size and are so similar in structure that a description of the last will suffice for all. The body is elongate, nearly terete but slightly depressed, thickest at mid-length, tapering rather evenly and gradually toward both ends, 12 to 18 millimeters in length and about two millimeters in breadth, depending on the species. The skin is thin and soft but tough. It is covered by ex- tremely small patches and rows of microscopic hairs, variously ar- ranged according to the species. In most species, there are also transverse ridges of larger hairs on both the dorsal and ventral surfaces but more distinctly developed on the dorsum. 734 THe UNIVERSITY SCIENCE BULLETIN The body of the larva consists of a strongly sclerotized, oval, depressed head capsule, three “thoracic” segments and ten “abdom- inal” segments, of which all are rather easily distinguishable ex- cept the last two (Fig. 83). Heretofore, the region I have des- ignated the ninth and tenth abdominal segments has been con- sidered a single segment (Wardle, 1926:29) but may comprise the ninth through eleventh morphological segments, while that which I have indicated as the prothorax has been regarded as two seg- ments, probably because it bears two rows of bristles instead of the single row found on most other segments. The interpretation of segmentation presented here is based upon studies of the internal as well as external anatomy of the larva and upon correlation of larval structures with those of the pupal and adult stages. Head and mouthparts—Since virtually the only larval struc- tures that are strongly sclerotized, hence of fixed size and shape, are associated with the head, this part of the larva has been studied in considerable detail by earlier investigators. Thus, Alexander (1920:983) gives a rather detailed descriptive paragraph on the head and mouthparts of the larva of Dolichopeza tridenticulata (as Oropeza obscura), the only North American species concerning which there are any published morphological data. The head has the characteristic tipuline appearance. Its sides are formed by the large lateral plates, which by their oval outline, striated pattern and concavo-convex shape somewhat suggest deep clam shells. Cook (1949:8) regarded these lateral plates as belonging to the maxillary segment of the head, in Holorusia, but I cannot ac- cept such an interpretation in the case of Dolichopeza. That the muscles of the mandibles and maxillae originate on the lateral plates suggests rather that these plates are homologous with the vertex of the ocular segment, as in so many other kinds of larvae. Ventrally, the lateral plates are separated by a somewhat tri- angular space (Fig. 84), but dorsally they fuse with the narrowly triangular fronto-clypeus (Fig. 85, felp) along much of its length. From the dorsal margin of each lateral plate, a bladelike projection extends backward alongside the frontoclypeus, slightly exceeding the posterior end of the lateral plate in species of Oropeza but con- spicuously longer in the subgenus Dolichopeza. The sclerite I have designated as the fronto-clypeus has been called the epicranial plate by Comstock and Kellogg (1902) and the clypeus by Cook (1949), who so identified it by the fact that the muscles of the cibarium originate on it. However, the dorsal muscles of the THE CRANE FiLy GENUS DOLICHOPEZA 7(3t5) ant eye 90 Fic. 84. Head capsule of fourth instar larva of Dolichopeza (Oropeza) obscura, ventral aspect; ant—antenna, Ip—lateral plate, max—maxilla, ment—mentum. Fic. 85. Same as 84, dorsal aspect; as—line of attach- ment of skin, eph—epipharynx, felp—tfronto- clypeus, lbr—labrum, md— mandible. Fic. 86. Same as 84, left lateral aspect. Fic. 87. Right man- dible of fourth instar larva of Dolichopeza (Oropeza) similis, mesial as- pect; tnd—tendon of adductor muscles of mandible. Fic. 88. Right maxilla of fourth instar larva of Dolichopeza (Oropeza) obscura, ventro- lateral aspect; cd—cardo, gal—galea, lac—lacinia, mxp—maxillary palp, sti—stipes. Fic. 89. Mentum of Dolichopeza (Oropeza) similis, ventral aspect. Fic. 90. Hypopharynx of Dolichopeza (Oropeza) dorsalis, ven- tral aspect. 736 Tue UNIversiry SCIENCE BULLETIN pharynx, which typically have their origin on the front, also originate on this same sclerite, which is why I regard it as the fronto-clypeus. A zone of fine wrinkles girdles the entire head at the line of attach- ment of the skin. The fronto-clypeus joins the labrum almost im- perceptibly. A large bristle, subequal in length to the antenna, is borne near the middle of each side of the labrum, and a narrow, medial projection from the anterior margin of the labrum appears to bisect the broad, brushy epipharynx, to which it is immovably joined. The antennae consist of a cylindrical basal segment about four times as long as thick and an apical sensory peg. They are attached to membranous areas at either side of the labrum. Below the labrum-epipharynx are the mandibles, each of which articulates on a blunt dorsal condyle at the junction of the lateral edge of the labrum with the lateral plate and an elongate ventral one on the mid-anterior edge of the lateral plate (Figs. 86, 87). The mandible has four marginal teeth (rarely five) and a broad, inner molar pro- tuberance having somewhat the form of a ridge across the man- dible. Behind this molar ridge is an excavation within which is attached the tendon of the adductor muscle of the mandible and at the distal end of which is situated a small tuft of bristles. This tuft has been called the brustia; its function is unknown to me. Attached below the mandibles but overlapping them apically are the subrectangular maxillae. The maxilla (Fig. 88) comprises a narrowly triangular cardo bearing three long setae, an irregularly shaped and very heavily sclerotized stipes, a blunt maxillary palp, and the indistinctly differentiated, apically hairy galea and lacinia, the latter more ventral in position. The mouth cavity is closed ventrally by the mentum, a structure formed by the fusion of for- ward extensions of the lateral plates (Fig. 84). In some species, particularly in the obscura group, the mentum usually has only five rather blunt teeth, but in others (as D. similis, Fig. 89) they number seven and are more acute. In all species, there is in addi- tion a small lateral tubercle at each side of the base of the mentum. Just dorsal to the mentum, within the mouth cavity, is the hypo- pharynx, into which the salivary duct empties. Its upper part is weakly sclerotized and covered with hairs, but the lower portion is strongly sclerotized and bears usually three blunt teeth (Fig. 90). The shapes of the hypopharynx and mentum, specifically of their toothed margins, have been found of limited use in the recognition of species. Since larvae of Dolichopeza are sensitive to light, it seems only reasonable to suspect that they have eyes of some sort, which they THe CRANE Fiy GENus DOLICHOPEZA 131 do. The eyes consist externally of a single, circular, only slightly protruding ocellus. They are situated one at either side of the head, in the foremost part of the lateral plate, just behind the mandible (Fig. 86). As the pigmented portion of the eye behind the hyaline lens is dark, the eyes are well concealed in the densely sclerotized walls of the head. Probably all crane fly larvae have eyes, but references to their existence in the larvae of the subfamily Tipulinae are rare. Peterson (1951: 256) probably was referring to species of the Limoniinae when he wrote “ocelli . . . may occur adjacent to the mandibles.” At any rate, his figure C, page 279, shows no eyes in Tipula abdominalis. Cook (1949), in a de- tailed study of the head capsule of Holorusia rubiginosa, noted the presence of eyes in that species; but most authors, less at- tentive to detail, have generally overlooked them throughout the Tipulinae, Comstock and Kellogg having failed to observe them even in Holorusia rubiginosa, a species they described in detail. In the less strongly sclerotized head capsules of some Limoniinae, such as Pedicia and Dicranota, the eyes are more easily seen but still have been referred to only cautiously as “eye spots” (Miall, 1893: 237 and plate I; Alexander, 1920: 1092, plate LXI). Caudal segments—On comparing genera of larval crane flies, one’s attention is almost invariably drawn first to the spiracular disc, which is the area surrounding the two large spiracles on the ninth abdominal segment. This ordinarily well-exposed surface usually exhibits great differences from genus to genus and often easily noticeable differences among species within a large genus. Within groups of closely related species such as North American Dolichopeza, however, there is considerable uniformity in the ap- pearance of the spiracular disc and its surrounding structures. In larvae of Dolichopeza, the spiracular disc is generally devoid of hairs but has three pairs of small, darkly sclerotized plates disposed around the spiracles as shown in Figure 92. Just below the disc are two broad, rounded lobes, each having a sclerotized apex from which extends a weak bristle. Other bristles occur on the sides of these lobes as shown. Above the disc are the more or less conical, fleshy lobes to which reference has already been made in the dis- cussion of the first instar larva. These four lobes are, in one form or another, characteristic of all species of Tipulinae, although the belief that there were only three dorsal lobes in Dolichopeza s. s., as first described by Beling (1886: 190), has persisted almost to the present day (for example, see Hennig, 1950: 396). Beling 738 Tue University SCIENCE BULLETIN stated that, of the usual four lobes above the spiracular disc in larval tipulines, the middle two in Dolichopeza albipes of Europe are grown together. Since Beling’s time, several authors have com- mented on this condition, it being unique among all Tipulinae. When I first found the larva of Dolichopeza americana, in the winter of 1949, I noted that the apparently single median lobe was actually formed of two lobes closely uppressed and that the two separated easily at the light touch of a needle. It seemed that the same might be true of D. albipes, but it was not until 1955 that I was able to verify this by examination of specimens in the British Museum. Only once among several hundred larvae examined have I found an instance of true complete fusion of the middle lobes. This was in a fourth instar larva of D. sayi, which in the third instar had been quite normal (see Fig. 39). This unusual specimen was pre- served because of its singularity. Situated dorsolaterally on each side of the eighth abdominal seg- ment of larvae of the subgenus Oropeza is a subconical lobe ( Fig. 92), similar in texture and often in size to the outer lobes above the spiracular disc. These lobes are characteristic of Oropeza but are also found in at least one other subgenus, Trichodolichopeza, or at least in the only two species, hirtipennis and flavifrons of South Africa, known in the larval stage (Wood, 1952: 88). Absence of such lobes on the sides of the eighth abdominal segment in the subgenus Dolichopeza (Fig. 91) gives those larvae somewhat the appearance of larvae of Tipula, but the green color and peculiar markings of the dorsum (Fig. 97) are different from any Tipula known to me and should make differentiation easy. The tenth abdominal segment, or anal segment, is located ven- trally with respect to the ninth and comprises four blunt, pale, mem- branous lobes surrounding the anus. There is no doubt that these lobes are homologous with the anal gills developed in many species of Tipulinae as well as rather generally in the family; however, as they appear on dissection to have no respiratory function (as dis- cussed below), I shall refer to them merely as anal lobes. There are a supra-anal or posterior pair and a sub-anal pair. Integument.—The soft but tough larval skin bears two size cate- gories of microscopic hairs, in addition to the macroscopic setae (which are the same in all species). As the hair patterns of the dorsum and venter are in general similar, although much less well developed on the underside of the body, a description applying in detail only to the dorsum follows. Seen by the unaided eye or only slightly magnified, the larva of Dolichopeza appears to have many THE CRANE FLY GENUS DOLICHOPEZA 739 darkened, transverse ridges on its back. In some species, these occur six or seven (rarely eight) per body segment, except where they are obscurely merged on the thorax and on the eighth abdomi- nal segment; but in other species it is difficult to discern separate ridges at all, due to their irregular distribution. Using magnifica- tions near 30 or 40 , the transverse ridges are seen to be formed of rows of hairs (Fig. 94). These, the larger of the two sizes of microscopic hairs, are many times as long as they are in diameter. The other size referred to are extremely short and appear to be no more than tiny spots, except at very high magnifications. At medium magnifications, rows of these minute hairs look like short, dark lines in the areas between transverse ridges and oriented more or less parallel to the ridges; but at high magnifications (60 to 80 x ) they may be seen as individual, short hairs, arranged in short or long rows (Figs. 95, 96). On the pleura the minute hairs may be arranged in rather circular patches or may occur singly, while the longer hairs are absent. As the pattern of distribution of these two kinds of integumental hairs seems to be fairly constant for a species, I have made extensive use of it in the key that follows this section. Internal anatomy.—The internal structure of larvae of Dolicho- peza was examined only in a general way, in an attempt to find specific or group characters for taxonomic use, but the anatomy of all species studied was found to be rather uniform. Only the di- gestive, tracheal and nervous systems are discussed here. The sketch of the internal organs (Fig. 82) was made from a larva killed by immersion in boiling water, which accounts for the out- stretched condition of all the viscera. In living larvae or those killed by means other than heat, the organs are shorter, broader and not so linearly arranged. The body cavity, especially pos- teriorly from about mid-length of the ventriculus, is largely occu- pied by a whitish or light gray perforated sheet of adipose tissue. Digestive system.—Most conspicuous of the internal structures are the organs of the digestive tract. The oesophagus is a muscu- lar tube, often enlarged at its posterior end. At either side of it lie the whitish-hyaline salivary glands, shown in Figure 82 with a single curve at about their mid-length, although they sometimes are doubly convoluted, Each salivary gland has a narrow lumen sur- rounded by a secretory portion made up of a single layer of large, cuboidal cells which have correspondingly large central nuclei con- taining long and much intertwined chromosomes. The chromo- somes were not examined in detail. The salivary glands are par- 740 THE UNIvERSITY SCIENCE BULLETIN tially invested with adipose tissue, especially apically, and are drained by slender salivary ducts which unite beneath the oesopha- gus just within the head capsule. Separated from the oesophagus by a slight constriction is the proventriculus, which in turn empties into the thick-walled, muscular ventriculus. Two short, glistening, pocket-like gastric caeca are appended to each side of the anterior end of the ventriculus, and its posterior end is marked by the attachment of four Malpighian tubules, beaded in appearance except for their nearly-transparent proximal ends, and similarly arranged two at each side of the ventriculus. A short, slender small intestine connects the ventriculus with the sac-like anterior end of the large intestine, to which is also joined a large, thick-walled di- verticulum, the intestinal caecum. This caecum is always found, in freshly killed larvae, to be filled with particles of food and count- less bacilliform bacteria which are probably symbiotically related to the larva. The large intestine tapers toward the rectum, the junc- tion between these two regions being poorly defined. A comparison of the relatively simple digestive system of a carnivorous tipulid larva such as Dicranota (Miall, 1893) with that of moss-eating Dolichopeza, with its voluminous divisions and many caeca, recalls similar differences between carnivorous and herbivorous mammals. Tracheal system—Two main tracheal trunks extend from the spiracles on the ninth abdominal segment to the anterior part of the mesothorax. They are situated close beneath the dorsolateral body wall, through which they may readily be seen in living larvae. From a bulb-like thickening at the mesothoracic terminus of each trunk, seven branch tracheae are given off to the brain, various parts of the inside of the head, and to the body wall of the prothorax and mesothorax. A metathoracic branch at about mid-length in that segment appears to supply oxygen to the adjacent body wall only. In each abdominal segment from the first through the seventh, there is a pair of ventrolateral tracheal branches, each dividing in turn into a short branch to the muscles and skin of the body wall and a longer visceral branch (Fig. 82); and in each seg- ment there is a slender tracheal connection between the two trunks, passing near the dorsal vessel and probably oxygenating it. Only the bases of these dorsal commissures are indicated in the sketch. The visceral tracheae are longest in the mid-region of the body, where maximum displacement occurs, during locomotion, between the digestive organs and the relatively stationary tracheal trunks. The point of branching from the trunk and the general THE CrANE FiLy GENusS DOLICHOPEZA 741 visceral destination of the abdominal tracheal branches of one side are summarized as follows: First abdominal branch: from posterior part of first segment or anterior part of second; to posterior oesophagus and anterior proventriculus. Second branch: from anterior part of third segment; to proven- triculus. Third branch: from anterior part of fourth segment; to proven- triculus and ventriculus in region of caeca. Fourth branch: from near junction of fourth and fifth segments; to ventriculus. Fifth branch: from posterior part of fifth segment; to small intestine and posterior part of ventriculus. Sixth branch: from posterior part of sixth segment; to large in- testine. Seventh branch: from mid-region of seventh segment; to rectum. The eighth and ninth abdominal segments have no dorsal com- missures or visceral branches but are oxygenated by short, much branched tracheae given off just anterior to the spiracle. A few branches from this same source reach the anal lobes, but the tracheation of these is so slight that it seems unlikely that they could serve a respiratory function. Central nervous system.—As in other crane fly larvae, the brain of Dolichopeza is located in the prothorax, outside the head cap- sule. It is composed of two subspherical lobes, narrowly joined and giving off three anterior branches, which I have not traced but which are presumed to innervate the eyes, antennae and labrum (see Cook, 1949: 39). The brain lies just below and before the open anterior end of the dorsal vessel, or aorta. Circumoesopha- geal commissures connect the brain with the suboesophageal ganglion, from the anterior surface of which arise two large nerves, each of which divides into three branches innervating certain mouthparts. From the posterior surface of the suboesophageal ganglion, two stout nerve cords converge to connect with the first of a series of four large, rounded and depressed ganglia lying against the floor of the mesothorax and anteriormost metathorax. The first three of these large ganglia supposedly belong to the three thoracic segments, while the last possibly represents the fusion of two or more abdominal ganglia. Ganglia of somewhat smaller size occur in the posterior part of the first abdominal seg- ment, anterior part of the third, near the middle of the fourth, fifth and sixth, and in the anterior half of the seventh abdominal seg- 742 Tue UNIvERSITY SCIENCE BULLETIN ments. Some tipulid larvae have one or two additional ganglia, but the equivalents of these in the larva of Dolichopeza are unknown. Behind the last ganglion, the two main strands of the ventral nerve cord diverge and bend posterodorsally, giving off several branches. None of the branches of the central nervous system was traced in detail. Larval key characters——As stated earlier, most keys to crane fly larvae have made use of characters of the spiracular disc and the head, particularly the mouthparts. Among species of Dolichopeza, however, there is such similarity of apperance of the spiracular disc and the lobes surrounding it that I have so far been able to make only limited use of these structures in species recognition; and to utilize details of the head capsule in taxonomy almost always re- quires dissection of the larva. For life history studies or other reasons, it is often desirable to segregate larvae before preserva- tion. Accordingly, taxonomic characters were sought that would allow species identification of living larvae. A satisfactory means of species separation has been found in the pattern of distribution of the two sizes of microscopic hairs of the integument discussed ear- lier in this section, used in conjunction with the few reliably con- sistent differences in the caudal segments. The key that follows has certain limitations. Mature larvae of Dolichopeza carolus, johnsonella, subvenosa and two new species described herein are still unknown; and I have reared certain of the other species only a few times and from only a very few locali- ties within their total ranges. The key, however, is reliable to the extent of the material examined. Comments on the species as yet unknown will be found under the respective species headings in the systematic account. Although prepared on the basis of fourth instar larvae, this key may prove adequate for second and third instar larvae as well. Magnifications of 30 to 80 >< have been used in describing char- acters, but some observations have been verified by use of the com- pound microscope. TENTATIVE KEY TO FOURTH INSTAR LARVAE OF NORTH AMERICAN DOLICHOPEZA 1. Body coloration in life green with two series of irregular brown to black spots on dorsum (Fig. 97); median lobes above spiracular disc closely appressed; no conical projections from dorsolateral sur- faces of eighth abdominal segment (Fig. 91) (subgenus Dolicho- CST Ce eo ie ns ee eee ee CR americana Body coloration in life greenish, brownish or some blending of these, lacking spots or other markings, except for transverse ridges of small THE CRANE FLY GENUS DOLICHOPEZA = — * ———— : .2yee- Bla er 3 Minute microscopic hairs of pleura grouped in small, circular patches, the individual hairs not thickened and not seen as separate hairs except at high magnifications; those of dorsum arranged in transverse rows or very rarely single; transverse ridges of larger microscopic hairs evident on most body segments but may be weak on second through: sixth\.abdominal: segments ... +... =... / .. Seat 4 3. Minute microscopic hairs of pleura of eighth abdominal segment as numerous, dense or evenly distributed as those of seventh and other REPENS Goce ct 2 ee Rona oo eS ecg kee polita sspp. Minute microscopic hairs of pleura of eighth abdominal segment sparse or irregularly distributed, leaving bare areas ............ tridenticulata 4, Larger microscopic hairs of dorsum arranged in dense transverse ridges that are variously interrupted, deflected or staggered (Fig. 93), so that it is difficult to count the number of ridges on a typical ab- dominalesepment ©. 72... ....-... .. ke ae ee 5 Larger microscopic hairs of dorsum arranged in subparallel transverse ridges, usually numbering six or seven per segment (Fig. 94); ridges may be weak on abdominal segments two through six ...... 6 5. Minute microscopic hairs in short rows of varying lengths, the rows distinct, clearly separated’. ...<.:...4200%.-+. 24 MO ee venosa Minute microscopic hairs in long, indefinite rows with poorly defined terminations, or not clearly arranged in rows (Fig. 93) .... subalbipes 6. Minute microscopic hairs in short, distinct rows with clearly defined terminations, the rows clearly separated (Fig. 95) .............. if Minute microscopic hairs in long, indefinite rows with poorly defined terminations, or not clearly arranged in rows (Fig. 96) .......... 8 7. Dorsolateral lobes of eighth abdominal segment nearly conical, of about same shape as lateral lobes above spiracular disc and approximately two-taresvasmiong asthe latter... 2...) @8e0.. olde a walleyi Dorsolateral lobes of eighth abdominal segment short and blunt, only half or less than half as long as lateral lobes above spiracular GIS AEs ys thr ss Sons cf nisin od a pene! Eppa oe similis 8. Transverse ridges of larger microscopic hairs about equally well marked Onvallgabdomimal segments .... 2.0 a .ese eee eee obscura Transverse ridges of larger microscopic hairs more weakly developed on second through fifth (or sixth) abdominal segments than on tho- racic, first and seventh abdominal segments ................... 9 9. Minute microscopic hairs clearly in rows, the rows crowded together near transverse ridges of larger microscopic hairs, leaving a narrow zone without nas (Fig. 96) 20°) 2). 22 2 eee sayi Minute microscopic hairs usually in poorly defined rows, the rows about evenly distributed between transverse ridges of larger microscopic hairs; hairs long and dense on eighth abdominal segment ..... . dorsalis THE CRANE FLY GENUS DOLICHOPEZA 745 LARVAE—NATURAL HISTORY It is in the larval state that an individual of Dolichopeza passes most of its existence, from only a few days after the egg is laid until about a week before the adult is on the wing. In regions where there are two annual generations, this span of time may be about 280 days in the overwintering generation but only about 50 days in the summer generation (Fig. 99). Larval life comprises four stadia, the last much the longest in duration. Eggs about to hatch pulsate slightly and irregularly for as much Fic. 97. Fourth instar larva of Dolichopeza (Dolichopeza) ameri- cana, dorsal aspect. Fic. 98. Fourth instar larva of Dolichopeza (Oropeza) obscura just before emergence of the pupa; note position of head capsule, pupal thoracic respiratory horn visible through larval skin, and contracted lobes of spiracular disc. 746 THE UNIVERSITY SCIENCE BULLETIN as three hours before the chorion ruptures. At hatching, the chorion splits nearly the full length of the egg, mostly along the dorsal side but somewhat on the ventral side, and the head of the young larva appears, mandibles foremost. Grasping at plant fragments or other surrounding objects with its jaws, the larva pulls itself free of the egg covering in a few minutes. Within a few hours the relative proportions of the body of the larva change markedly, as described in the preceding section and illustrated in Figures 78 and 79. Once freed from the egg, the larva sets out almost immediately in search of food, clamping its mandibles firmly into whatever it encounters.* For this reason, newly-hatched larvae in the laboratory must be provided with food, if several are kept together, or they will com- mence biting each other. Very young larvae are extremely sensitive to instabilities of en- vironment, such as abrupt changes in temperature or moisture. As mentioned earlier, they drown readily in small excesses of moisture. They perish equally rapidly when subjected to even slight desicca- tion. Mortality of first instar larvae in the laboratory was very high in all species, possibly because no special equipment was employed to stabilize either humidity or temperature. Although young larvae appear able by the remarkable gape of their jaws to partake of parts of moss leaves such as are eaten by more mature larvae, their earliest diet seems to consist primarily of material scraped from leaf surfaces and plant scraps which collect about moss stems. Within a day of their emergence, they have gorged themselves so that the alimentary canal may be clearly seen as a bright green streak from one end to the other of the nearly transparent body. Confined to small dishes in the laboratory, very young larvae frequently devoured others that had died, and some- how opened and consumed the contents of unhatched, presumably unfertilized eggs, discarding the chorion. I do not take it from this that cannibalism is the rule in Dolichopeza but merely assume that available sources of protein are utilized whenever they are found by young larvae, just as they are by older larvae, which have been observed to feed on dead pupae and larvae in laboratory rearing dishes. Duration of the first stadium is not constant, even among larvae of the same brood hatched about the same time. It usually varied from eight to fifteen days, indoors, although a few laggard individ- i This tactile response of the mandibles is a convenience in picking up the tiny larvae, which might be damaged by other means. A hair or bristle of a brush placed before the larva is soon seized between the jaws, and in the time it takes the larva to discover it has not got hold of food, it may be transferred to the desired place. 747 THE CRANE FLY GENUS DOLICHOPEZA ‘opel of ype tu ‘ednd [ ‘RAI 1e}SUL YWNOF 97x] — ‘BAILT 1vSUL YAMOF ‘stOJVoIpuT YYMoIS—I-I ‘VAIV] IBISUT YZANOT ATIVO—Y ‘eAILT IVJsSUT PITUI—B “eae I Tees LL yee aa I Toy [ rake! I Ievysul PUuOd9S—]J-9 “BAIe] IeJSUT JSI—p ‘“SS9 Jo Suryoyeyu—o ‘S80—q ‘uontsodtao Jo our B SURSIYOI uleyynos UL tins (pzado1c—Q) pradoyoyog JO UOVIOUES IOWWUINS OY} UT YYMOoIS pur ojoA0 ot aa ai | bY Yod I I Ft ol 2! b/ 9! ‘ww eg] La a ae) yaa ——— LS MONG. o9 os Ov of ao) 66 ol 2! bl 9! gywuw 748 THe UNIVERSITY SCIENCE BULLETIN uals required three weeks to pass to the next instar. Larvae reach a length of about four millimeters before the first ecdysis. Size change at molting is not noticeable, but some remarkable changes take place in the appearance of the larva. The four long caudal setae as well as the tufts of bristles (Fig. 81) are shed with the first larval skin. These structures are characteristic of the first instar larva of all species in which this instar was observed (americana, carolus, dorsalis, johnsonella, obscura, polita sspp., sayi, similis, venosa and walleyi); accordingly, their absence is taken as evidence that the first molt has occurred. In the second instar, the larva of Dolichopeza resembles the later instars in general appearance. Although low and broad, the char- acteristic six lobes surrounding the spiracular disc are present and plainly visible, as are also the dorsolateral lobes of the eighth abdominal segment. At the first molt the larva also acquires the coating of microscopic hairs that gives the body the roughened surface so useful in locomotion. (There are hairs on the skin of the first instar larva, but they are so short and slender as to be scarcely visible at 200 magnification.) Following the first ecdysis, the larva feeds more actively and grows rapidly, reaching the end of the second stadium in about five or six days. Third and fourth instar larvae are almost identical in appear- ance, differing only in size. The third stadium may be as short as five days, so that some larvae may attain the fourth instar only about three weeks after hatching from the egg. Thus, the fourth stadium lasts as long as the other three combined, in the summer generation, and is much longer in the overwintering generation. Size increase is especially rapid during the first few days following a molt, in all instars (Fig. 99), and thereafter it proceeds more slowly. An explanation for this may be found in the fact that the larva ceases to eat for a short time preceding ecdysis, but once the mouthparts regain their strength and rigidity after the molt it commences to eat voraciously. Of those larvae that survived, in a rearing dish of Dolichopeza sayi, all that hatched on or about 26 September reached the fourth instar between 15 October and 5 November, indoors. This was in southern Michigan, where sunny days are common enough into late October that most larvae pro- duced from late summer matings are able to attain the fourth, or at least the late third, instar before cold weather forces cessation of feeding and the winter quiescence. THE CRANE FLY GENuS DOLICHOPEZA 749 Molting.—Several hours before shedding its skin, the larva of Dolichopeza ceases feeding and other activities, and the body slowly becomes turgid, assuming a swollen appearance like that sometimes seen in larvae that have been dead several hours. The caudal lobes become tumid and blunt, losing their nearly-conical shape, and the spiracular disc bulges outward. The head is fully extended, while the thorax close behind it is swollen with the skin darkened, almost black. In this inactive condition, the larva rests until the separation of old cuticle from new has been completed. When movement is resumed, shortly before the actual molt begins, a perfect duplication of bristles, microscopic hairs and all other ex- ternal features of the larva may be seen through the old integu- ment. The larva suddenly begins to jerk the anterior one-third or one-half of its body from side to side in frantic and spasmodic motions, the head capsule being bent further and further ventrad all the while. About two minutes after the onset of this activity, the skin splits along the dorsal mid-line of the thorax, just behind the old head capsule, and the head of the larva is thrust out. Only another sixty seconds or so are required for the virtual completion of sloughing off the old skin, as the pale and feeble larva crawls among the moss stems, working the old skin off by repeated body undulations and by scraping against its surroundings. Attached at the spiracular disc by the linings of the tracheal trunks, the crumpled molted skin is dragged about for a while and is nearly turned inside out before the tracheal linings pull out and the larva is at last free of its old skin. Each ecdysis seems to produce a double crisis in the life of the larva, first in the struggle to throw off the old skin and second in surviving the periods of comparative helplessness before and after the molt. In fact, I find it somewhat fatiguing merely to observe the process! Each of the four larval instars of a given species may be recog- nized by the dimensions of the head capsule, which changes size abruptly at molting while the softer parts of the body increase gradually and continuously. Increase in breadth of the larval head by a more or less fixed factor at each ecdysis has been ob- served in many kinds of insects and has been summarized as Dyar’s law. In Dolichopeza (O.) sayi, this fixed factor of increase was found to be 1.5. Thus, from the head capsule width of the first instar larva measured as 0.25 mm., one estimates 0.38 mm. as the width of the second, 0.57 mm. for the third, and 0.86 mm. for the fourth instar head capsule. Observed widths were 0.25 mm., 0.40 750 THe UNIvERSITY SCIENCE BULLETIN mm., 0.60 mm. and 0.85 mm., respectively, measured by means of a camera lucida scale. Head capsule length in this same series of specimens averaged 0.40 mm., 0.65 mm., 1.00 mm. and 1.50 mm. for the four larval instars, showing that length increase is not ex- actly proportional to width increase. In a single instance, a larva seemed, on the basis of the size of head capsule recovered from the pupal burrow, to have pupated at the end of the third stadium. No indication of a fifth instar was ever found. Feeding—Very young Dolichopeza larvae are able to obtain food by scraping the surfaces of moss leaves and may also eat or- ganic debris of various sorts, including their own dead. From the fact that often only the head capsule could be recovered from a rearing dish soon after a larva had molted and that I once found a larva with its head directed toward its freshly-cast skin, I suspect that the larva may often eat its recent skin for its first meal follow- ing ecdysis. By far the greatest item in the diet of the second, third and fourth instar larvae, however, is leaves of certain mosses and leafy liverworts. The particular species of such food plants are listed later in this section, under ecological distribution. Most feeding is done at or near the surface of the moss habitat, where the leaves are green and fresh. When isolated bits of moss are offered to larvae in rearing dishes, the leaves are largely or entirely consumed, while the stems and other parts are ordinarily rejected. There is an interesting difference in feeding behavior between larvae of Dolichopeza americana and those of species of subgenus Oropeza. Larvae of all species of Oropeza are rather uniformly colored, brownish, greenish or some subtle blending of these (the color in any case largely dependent upon recent diet) and in gen- eral lack any pattern. Transverse ridges of microscopic hairs might be regarded as a pattern, but this is not of the obliterative sort that draws attention away from the shape of the insect. Therefore, Oropeza larvae are conspicuous against the background of greenish highlights and dark shadows characteristic of most mossy surfaces on which they feed. Dolichopeza americana, on the other hand, has a green color and is irregularly marked with dark lines and blotches (Fig. 97) which serve remarkably well to break up the visual outline of the insect. And while larvae of Oropeza conceal themselves well down in the moss by day and feed almost wholly at night, americana often browses at the surface in full daylight. Larvae of americana may also be found more deeply in the moss by day, but it is a rare occurrence to find an Oropeza larva exposed THE CRANE FLY GENUS DOLICHOPEZA 751 to view.* Occasionally the irregularly-shaped, green and brown larvae of Liogma nodicornis (Tipulidae: Cylindrotiminae) are found on the same moss with Dolichopeza americana, and a sim- ilarly-colored and patterned moth caterpillar (Noctuidae—my iden- tification) was three times taken together with these two species of tipulid larvae, all providing striking examples of protective resemblance of insect to environment. It seems clear to me that there is a correlation of color pattern in Dolichopeza larvae with the time of active feeding. Locomotion.—In the field the collector will probably find larvae of Dolichopeza in an inactive state in which the anal lobes are withdrawn and the head is retracted within the second and/or third thoracic segments. This is the attitude assumed by the larva in reaction to tactile or light stimuli, as it shrinks from both. In this position, the larva may remain quite motionless for many min- utes. When the irritation has eased or ceased, the larva puts out its head, protrudes the anal lobes and proceeds to a place of con- cealment. Lacking any sort of creeping welts such as are found on larvae of some other genera of Tipulidae, the larva of Dolicho- peza accomplishes locomotion by three other means. It may move slowly by extension and contraction of the body, using the rough skin and large bristles as braces against the surrounding medium. More rapid locomotion is achieved by the pushing action of the anal lobes in combination with bodily undulations. Sometimes the head is extended forward and a moss stem or other support is grasped between the mandibles, so that the body may be pulled forward by contraction. This method of getting around is occa- sionally used alone but more often simultaneously with one or both other means described. Preparation for pupation.—Larvae not feeding at the moss surface will ordinarily be found just beneath the surface in fairly well formed tubes. Close examination of such a tube shows that it is constructed not by the larva’s having eaten or forced its way through firm material or dense moss but rather by caulking larger, pre-existing cracks with a frass of digested moss fragments and other debris. Larvae do not penetrate more than a few millimeters into wood or soil upon which their moss habitat occurs, and they have never been found in any but the most spongy, soft wood in advanced stages of decay. A tube thus formed is of course usually * To collect larvae from mosses brought into the laboratory, I have spread the mosses in a pan, leafy surface upward, and covered them with a few sheets of dampened news- paper. After a few hours in this artificial darkness, larvae may be found feeding at the surface of the moss. Whey THe UNIversIry SCIENCE BULLETIN closed behind the larva and is shifted from place to place according to available food and appropriate sites. But in the last days of its fourth stadium, the larva prepares a more compactly walled tube, oriented roughly parallel to the moss surface and with an open anterior end curving up to the surface. It is within this tube that pupation occurs and the pupal stage is spent. The pupal burrow is usually about one and a half times as long as the pupa but has been found as short as one and a quarter times the length of the pupa in Dolichopeza subalbipes and as long as twice the pupal length in venosa. Pupation.—As the pupa takes shape within the skin of the fourth instar larva, the larval head capsule is pushed slowly forward and downward. In an individual about to undergo the transformation from larva to pupa (that is, the actual molt), the pupal structures become visible through the larval cuticle only a few hours before ecdysis. The head capsule becomes sharply deflected downward, about two hours prior to the molt, until its longitudinal axis forms an acute angle with that of the body generally (Fig. 98), stretch- ing the skin of the prothoracic dorsum taut. The lobes surround- ing the spiracular disc are drawn together in such a way as nearly to conceal the spiracles. A period of about fifteen minutes of relative inactivity was noted to precede the molt, following which occurred an abrupt resumption of movement, terminating after only a few minutes in the splitting of the old head capsule and prothoracic skin and the rapid emergence of the pupa. Worked off backwards by bodily undulations of the pupa, the old larval skin is pushed into the back of the burrow, in the form of a small, black- ish, wrinkled wad, scarcely recognizable from surrounding frass and other debris, except for the head capsule. With some knowl- edge of the usual length of the pupal burrow, one ordinarily may recover the shed last larval skin rather easily by watching for the head capsule. If the skin is not recovered and preserved within a couple of days after pupation, however, decomposition sets in, and various organisms, chiefly mites and Collembola, begin to feed on it, destroying all but the strongly sclerotized head. Seasonal distribution—It has been my repeated observation that not only is the late summer generation of adults composed of indi- viduals smaller than those of the spring generation, for reasons pre- sented earlier, but also that there are fewer flies in the late summer generation. Several explanations for this are possible, such as the likelihood of a greater amount of parasitism or predation on larvae THE CRANE FLY GENUS DOLICHOPEZA 753 during the summer, or starvation or desiccation of larvae due to summer drouth conditions. It seems not improbable, however, that it may be due, at least in part, to differential rates of larval development; that is to say, some larvae in favorable habitats com- plete all their growth stadia, pupate and produce the fall genera- tion adults, while others in more marginal habitats possibly attain the fourth larval instar but are prevented from growing rapidly enough to pupate in late summer and as a result overwinter as larvae. The result of this would be a spring generation of adults derived from larvae some of which are approximately nine months old and others of which are nearly a year old. Intensive field ob- servation of larvae at a particular locality through an entire year would probably provide an answer to this problem. The seasons for collecting larvae in various instars have been discussed under the heading of collection and preparation of material for study. Exploration of larval habitats during the spring, summer and autumn shows that the larvae seldom penetrate more than a few millimeters, perhaps a centimeter, below the surface of their bryo- phyte food supplies and then only into soft substrata such as rotten wood or soil. I was curious to know where they spent the winter, whether still near the surface or deep beneath the moss, burrow- ing downward like the immature stages of some other kinds of insects. Plainly, those larvae that lived in the shallow growth of moss covering boulders or rock cliffs could not go very far from the surface to escape winter’s cold. And there are not many places in the northern parts of the range of the genus where larvae could go below the frost line. In order to answer this question, I marked several microenvironments (that is, particular patches or hum- mocks of moss) found to contain numerous larvae in late summer and fall and revisited these in midwinter. While some of the larvae had gone a short distance below the surface, most were in about the same places where they had been active during warmer weather. From moss solidly frozen in a cake of ice that was chopped from the floor of a marsh in Michigan on 31 December, I succeeded in rearing Dolichopeza sayi. A larva thawed from the ice slowly extended its contracted body as it warmed in the labora- tory, and after a few hours it was actively feeding. By 30 January, this larva had developed into an adult fly. By remaining at or near the surface, larvae are able to take ad- vantage of occasional warm days, in late autumn and early spring, to feed. Furthermore, drainage of the microenvironment is suffi- 754 THe UNIvERSITY SCIENCE BULLETIN cient in most instances to protect the larvae from becoming frozen within a formation of ice; but even if so frozen, they are able, at least some of the time, to survive. Physiological adaptations, such as supercooling, reduction of body water with consequently higher salt concentration in the haemolymph, adsorption of water on pro- tein molecules, an so on, surely enable the larvae to withstand many of the low temperatures to which they are subjected. It is not in- conceivable that in their contracted state they are able to conserve at least for a while minute amounts of radiant heat received at times when there is no snow cover. In spite of all this, I believe winter weather does take an appreciable toll, for the spring popu- lations of larvae are much reduced from those of the autumn, and there is not likely to have been extensive predation in winter. Ecological distribution —lIt is possible to describe the general habitats of the various species of Dolichopeza in rather broad terms, such as swamps and marshes, rocky ravines, mesic wood- lands, and the like; but what is actually described in these cases is the kind of place in which adults of the species may be found. These habitats, for the various species and species groups of Dolichopeza, have already been discussed under the natural his- tory of adults. Within or adjacent to these broad and not always well-defined types of habitats, the microhabitats of the larvae may be widely distributed or narrowly restricted. Larval microhabitats have been discovered for all North Ameri- can species of Dolichopeza, except the two new species described herein, and in all cases the food substance was a moss or liverwort, although larvae may occasionally be found in the various substrata upon which the bryophyte is growing (cf. Alexander, 1920: 981). Larvae of most species of Dolichopeza have been observed to feed on a variety of bryophytes, a variety that will surely be found vastly increased as observations are continued. It might therefore appear that the distribution of a particular species is controlled by the distribution of the bryophytes upon which its larvae can subsist. This, however, is only partly true. Whatever mosses or liverworts are utilized by Dolichopeza as larval habitats are those the microenvironmental characteristics of which are to a large meas- ure conditioned by the general environment in which they are growing. While it is true that most bryophytes have fairly limited ranges of ecological tolerance, some grow under various conditions. Therefore, it does not follow that, if larvae of a species of Dolicho- peza have been found feeding on a particular kind of moss, where- ever that moss occurs there will be found also the fly. For ex- THE CRANE FLY GENuS DOLICHOPEZA (eb) ample, Leucobryum glaucum growing loosely and luxuriantly on hummocks in a birch-maple swamp is a known, albeit uncommon, larval habitat for Dolichopeza obscura; but I would not expect to find obscura larvae in that same species of moss growing, as it does much more often, in compact, dry cushions on the floor of a hilltop oak-hickory woods. Hedwigia albicans, growing usually on ex- posed rocks, is a dry, blackish moss that seems an unlikely habitat for any crane fly larva. But if the rocks were in a wooded area, shaded by trees, the moss could retain enough moisture to support larvae of Dolichopeza tridenticulata, a species much more tolerant of dry conditions than any other in the genus (see Alexander, 1920: 983, under Oropeza obscura). That it is the conditioning effect of the general environment on the microhabitat and not the general environment itself that gov- erns the distribution of larvae of Dolichopeza is shown by the fact that where appropriate microhabitats obtain in contrasting kinds of general environments the crane flies may be found. Dolichopeza sayi, for example, seems to find the ecological conditions required by its immature stages rather generally available along the shaded borders of northern swamps and marshes, and its larvae are found widely scattered there in the mosses of the swamp floor and hum- mocks. By contrast, in an area of rocky ravines, where sayi is present in much smaller numbers and where adults were seen only in a limited area of vegetation in the bottom of a ravine, larvae were found only around one small patch of swampy soil, only a few square feet in extent, below a permanent seepage from the cliff. It is thus not the swamp type of general environment that limits the distribution of the species. As a further example, larvae of Dolichopeza tridenticulata occur in great numbers in thin growths of moss on boulders and rocky walls of gorges in southern Indiana and Ohio, but the species also finds a vastly more limited yet some- how satisfactory habitat in a sparse growth of moss in crevices of tree bark in a woodland on comparatively level ground, in southern Michigan. Precisely what the larvae need from their surroundings has not been studied in detail. Their reactions in the laboratory to excess moisture and to desiccation have supplied some information, how- ever, about their ecological requirements. Very young larvae are easily drowned in even a small amount of water, when they are unable to extricate themselves from it. If its moss habitat becomes flooded, a larva in an advanced stage of development first closes together the upper and lower lobes of its spiracular disc, then ex- 756 THE UNIVERSITY SCIENCE BULLETIN tends its caudal end up to the surface, exposing the spiracles to the open air. The surface of the disc is very smooth and apparently repels water, while the lobes surrounding it so effectively hold back the surface film of the water that the disc may actually be with- drawn very slightly below the surrounding water level without be- ing wetted. A submerged larva also extends its head, making the body fully outstretched.* If the water does not soon recede, the larva begins to crawl toward a drier place, periodically pausing and putting its spiracles to the surface for air. If a larva is removed from damp moss and placed in an empty dish or on a tabletop, it will wander about aimlessly until the cuti- cle dries. At this point, the larva retracts its head and anal lobes, decreases the exposure of its spiracular disc by drawing the lobes closer together, and contracts its entire body, thus reducing the evaporation surface. Movement ceases, and unless restored to a moist environment the larva will continue to lose body fluids through the skin and will shrivel and die on that spot. In view of these reactions to drying and excessive wetting, one may make some generalizations about the bryophyte habitats of larvae of Dolichopeza. It seems unlikely, for example, that pri- marily aquatic mosses, such as Fontinalis or some species of Dre- panocladus or Hygrohypnum would be utilized. The same could be said of mosses that ordinarily grow in dry places, such as Grim- mia, Ceratodon and Hedwigia, although it has been shown that such mosses may, when their usual ecological situation is modified, support these larvae. It of course follows from this that whatever bryophytes are utilized must not at any time during the larval part of the crane fly life cycle become inundated or desiccated for a prolonged period, if the larvae are to survive. On the other hand, it seems not unlikely that larvae do perish as a result of oviposition in such fluctuating, marginal microhabitats that happened to have been in a favorable condition during the season for oviposition. Larvae of Dolichopeza were sought in all kinds of mosses and liverworts available, as well as in soil, decayed wood, fungi and other plants, leaf litter, and the like. They were found in a wide variety of bryophytes, which demonstrates a certain latitude in their ecological tolerance, but some generalizations about their habitat preferences may be gained also, I believe, from the bryo- phytes in which they were not found. Aside from never having been found in aquatic or in very dry mosses, as discussed above, they were never collected either from loosely-growing, bushy mosses * This initial reaction to inundation is useful in obtaining over-all measurements of living larvae. THE CRANE FLy GENUS DOLICHOPEZA (57 or from those that grow in tightly-arranged tufts or patches. I suspect that the very compactly-growing mosses, such as Bryum argenteum or Ceratodon purpureus, even if growing in the appro- priate environment, would offer too much resistance to the tunnel- ing larvae, while the loosely-growing mosses, such as Climacium, Rhodobryum, Polytrichum and most Mnium cuspidatum, present too diffuse a framework for the construction of burrows by the caulking method described earlier. Not only the growth habit of the bryophyte but also its leaf texture affects the crane fly’s selection of microhabitat, or at least its choice of food. Coarse-leaved mosses like Polytrichum were always rejected, when offered to larvae in laboratory rearing dishes, and apparently are not eaten under natural conditions. Bryophytes with thick leaves, such as the thallose liverworts, seem not to be eaten either. In a mixed growth of leafy Mnium punctatum moss and the thallose liverwort, Pellia epiphylla, larvae of Dolichopeza americana fed exclusively, as far as could be determined by field observation, on the Mnium. It scarcely need be mentioned that mosses which do not occur anywhere about the various general environments of Dolichopeza are not likely larval habitats. These would include some moss species that usually grow on sand or clay in open fields, in sea- shore habitats, and other unforested places. Many such mosses were collected and examined, however, during the early stages of this investigation. Where several species of Dolichopeza occur within one general environment, as is usually the case, each species has its characteristic larval microhabitat, although these often overlap broadly, resulting at times in the occupation of a single clump of moss by larvae of as many as three species. In a swarm of adults under a rock ledge in a wooded ravine, I may find together adults of Dolichopeza obscura, tridenticulata, polita, walleyi and americana. A search for the immature stages might show, however, that obscura came from Hypnum moss growing on a rotten log, tridenticulata from Dicranella on the rather dry upper surface of a large boulder, polita from mixed Tetraphis moss and powdery lichen on the moist undercut surface of the rock outcrop, walleyi from Plagiothecium growing as a mat on the rich soil of the floor of the ravine, and americana from Mnium growing on the damp rock at the base of the cliff. More than fifty species of bryophytes have been found associated 758 THe UNIVERSITY SCIENCE BULLETIN with larvae of Dolichopeza, ranging from rather dry Dicranella on boulders to decidedly wet Amblystegium of swamps, from calcare- ous, marl-forming Gymnostomum to acidic Sphagnum, and from fairly compact cushions of Tetraphis to the loose mats of certain liverworts or the low, leafy clumps of Mnium. In general, these microhabitat plants are those through which the larvae can move freely but which are at the same time compact enough to permit burrow formation * and which are not excluded for any of the reasons stated above. Certain species of Dolichopeza have so far been found usually in mosses that were only damp, while other species have always been collected from wet mosses. A correlation between species of crane fly and the amount of moisture in its larval habitat can be indicated by a rather arbitrary arrangement set forth in Table 3. In this table, known food plants are arranged from the driest at the top to the wettest at the bottom. As many mosses will grow under diverse conditions, they are entered in the table according to their relative water content at the time of collection. The species of Dolichopeza are arranged from left to right across the top of the table in the order which I believe expresses their increasing affinity for moisture. Mosses in particular but liverworts as well are often found growing in close associations of two or more species. The following mosses and liverwort were collected together with food plants listed in the table and are possible or probable food sources for the larvae, although the relationship has not been established: _Anomodon attenuatus, Anomodon rostratus, Aula- comnium palustre, Brachythecium sp., Dicranella sp., Dicranum flagellare, Drepanocladus aduncus, Lophcolea heterophylla, Mnium cuspidatum, Thuidium delicatulum and Thuidium recognitum. Although several larval microhabitats have been discovered, there must be a great many more still unknown. When I compare, in the field, the number of adults on the wing with the number of pupal skins found (or the number of larvae that could have been found at an earlier date), it sometimes seems to me that I have scarcely touched upon their major moss habitats, that I might have over- looked the most important places. This of course is largely a result of the diffuse distribution of the immature stages and the concen- tration of adults in places not always near the larval habitats. * So often do the larvae follow lines of least resistance in moving through moss that a convenient way to collect them from fairly compact cushions is to bend the leafy outer surface of the moss into a convex curve until it opens along naturally formed cracks. In these cracks will be found nearly all the larvae which inhabit the particular bit of moss. THE CRANE FLY GENUS DOLICHOPEZA 759 A systematic arrangement, by family, of the habitat bryophytes listed in Table 3 illustrates something of the range of plants of this phylum that are utilized by larvae of Dolichopeza. Eleven of the twenty-seven families of mosses known from North America are represented, and of those not appearing here many are specialized families including only a few species. The family Hypnaceae, by far the largest in this continent, is not disproportionately represented by the sixteen listed species. In contrast to the broad distribution of food plants among the mosses, all five species of liverworts fall into a single family, the Jungermanniaceae, or leafy liverworts, which are not unlike mosses in leaf texture. The other five families are so far unknown as Dolichopeza habitats, and it seems that most of the liverwort habitats yet to be discovered will also be species of the Jungermanniaceae. MOSSES Sphagnaceae: Sphagnum palustre, Sphagnum sp. Georgiaceae: Tetraphis pellucida Polytrichaceae: Atrichum macmillani, Atrichum undulatum Fissidentaceae: Fissidens taxifolius Dicranaceae: Dicranella heteromalla, Dicranum scoparium, Leucobryum glaucum Grimmiaceae: Hedwigia albicans Tortulaceae: | Desmatodon obtusifolius, Didymodon tophaceus, Gymno- stomum calcareum Orthotrichaceae: Orthotrichum sordidum Bryaceae: Leptobryum pyriforme, Mnium affine, Mnium punctatum, Mnium sp. Leskeaceae: Myurella careyana Hypnaceae: Amblystegium riparium, Amblystegium varium, Brachythe- cium salebrosum, Bryhnia graminicolor, Campylium chrysophyllum, Entodon cladorrhizans, Eurhynchium pulchellum, Eurhynchium serrulatum, Heterophyl- lium haldanianum, Hypum curvifolium, Hypnum imponens, Hypnum lind- bergii, Hypnum sp., Plagiothecium denticulatum, Plagiothecium deplanatum, Plagiothecium roeseanum, Platygyrium repens. LIVERWORTS Jungermanniaceae: Calypogeia trichomanis, Chiloscyphus pallescens, Geo- calyx graveolens, Plagiochila asplenioides, Scapania nemorosa. In examining bryophytes for the larvae of Dolichopeza, one re- peatedly comes across certain other organisms, mostly arthropods but, if the plants are from wet habitats, molluscs as well. Those most regularly found in association with Dolichopeza larvae are small pyralidoid caterpillars (identified by Mr. H. W. Capps), various stratiomyid larvae (unidentified, occurring in wet mosses especially), several species of Collembola (Entomobryidae, Po- THE UNIVERSITY SCIENCE BULLETIN 760 ge ere 9 ae Jo euoyspurs Jo aseq Iwou “D))]DWOJLajaYy D)JaUD.LIUCT Ba shee sok SpOOM UL Yo auoyspuLsS Jo oseq Ivou “WwunanD)b wnhaqoonaT Sate) Calc ee. hp oy 4Qro"lo 6.90.6 yo euoyspues uo “upyjrUmonU wnYyoyy ge ae eae a Yl vuoyspurs popeys uo ‘srzupwoyor.) diabodhjny “YI ouoyspues popeys ‘usyor] ArapMod yyIM “vpronjjad srydv.ya], ee ee SpooM UT ‘YB dL} JO SAOTAGIO UI “WNpIpsOS WNYDYOYIAC) x aE te OS x x || 2 xx I 2s 2K |) 2s x x ees || 2x x x x elelelelelelalzlelalelzle Sr Sh er tS RT Neto Sek i) Sites lS} || Sy Wes Te eee mM —e Saersees ieee ide. ies le 1 a |B) ee ee na ro} jy |p Jat et 2 - ® Cae) 5 at wm Nae 2 f nzadoyoyog jo satwedg O10 6 oS AD f ‘ ’ y SpOoM ut Jepynoq jo vovjans zeddn uo “n77MWoLajay D)jaUDLIUT Sl eWueles: sa(elso/s) cimepiol, 6-9: (9) laule, Soh) w uphie: \anie (ea se Weite, («Sema ae een SYOOL uo ‘supo2q]D pibunpa HT LVLIAV]] ALAHAOAUG yU9}UOD, aIN}sSIO 0} Surpioooy pasuvuy syyqeyosr eyydodkig ut pradoyoyog uvoweury YWON jo saoedg fo avaiv'yT jo uonnqrnsiq—sg ATav], 761 THE CRANE FLY GENUS DOLICHOPEZA BASEN aE? iene diem pa iar Sutids aptseq yuRq uo “wnonnj]b wnhaqoonaT digas sabe alon eens Ao yurq urvoarys Ios AvpO uo ‘sunztyL.opp)o UOpOJUA RCE a abt SpooM ul yuBq WRIIys UO “WnjDUD dap wnIdaYI0D] J GASP ae eke abe sk a li ea syoouumny durems uo ‘swapoaanib xfijpI004) SIRS Sia gE aS gE RCE syoouumny durvas uo ‘ppronjjad srydp.ya J, eae Aton s ab a halioey Yo ouoyspuRs Jom uO ‘snzyofisnjgo Uopo}MUsacT eS i Oe Bia yy ouoyspurs yom uo “wnppfiydoshayo wniyfiduvg) SAG eee ASS oe spoom ut ‘yoorq Jo yuvq ‘ros dup uo “ds wnupy ati iieiaarediates iieies Ylo ouoyspuvs yom Jo osvq ye ‘ausofihid wnhsqojdaT © eka AE se aS BEN Sp al adoys []Q [eIoxys uo ‘suaosayjpd snydfiasor.y) Sta aRatysinte TALS, tessa tests asia ag adoys [1 [VIor[s uo “wnpoynpun wnyoy py Sy Reena eae R Yo uoyspuvs durep jo aseq ye “wnjzojound wniu py pecan ibe Hakata tras Geese YOO shyB} ouojspuRs UO ‘DsoLowau DiundDIg Da eee yy ouoyspurs ‘ros poddesy uo ‘wnpnpnisas wnryouliyungy b Riast ses ST ASR GS og Yl euoyspurs uo ‘saprorvuajdsp vjryoorbn) J wi lsiinsie Colter sie lale! lela /ale (elie) uliehielieMegialisyielisy alia! yyo auoyspuvs uo ‘nunhawno ppjainh yy Meee: deo ne spoompaey ut fer} UaT[ey JO Yavq uo ‘suadas wnisabhjo) J 26—5840 Tue UNtversiry SCIENCE BULLETIN 762 5 : ) ae SP eS ee) Geral) See Stl genaa aaa eI =H es Dn - =. (o) © i= a =} fo) i ie) ak > rs) =e = — 2 D | a ie} po = B = ior || SS e fe) 5 i) a = ae n hae he 5S 5 S 5 ise, fe mn | © = 2 LVLIAV]] ALAHAOAUGT D 2 — 2 © i) pzadoyoog jo sawedg papnjauoy —ju9}U0D + aIM}SIOJY 0} SUIpiosoy pasuvity sjeyqeyorolY 9}AydoAig ur pzadoyoyoqd uvotieury YyWON jo soroadg jo avaiey jo uonqmysiq—¢ aATaV 763 THE CRANE FLY GENUS DOLICHOPEZA | x msn See Nao a a oo) ysaeur Jo 100y uo “wna unibashiquy Ca AR aS tan at Cs “AOD auoz dal} JO adpa ye ‘Boq “ds wnubpydg x 28 B00 0 Oo Deo 0 ONG, Ow et D0 OOD Oth 30q jo oUuoZz popoom ‘ujsryod wnuboydy Pegs ek Se ESS aS dures Jo o8po 4B [10s yom uO “22bvaqpuy wnudhy eS itnak es ae ae [los uo ‘ulsivur dures “wnsougayps wnwayjhyooig Px |x| x|x ee academe ce [los uo ‘dureas jo ursavur ye unpjayojnd wniyouhyungy Page. aes? Sa ae ae eS aUI[II}VM 4B ‘aUOYSpuRS UO ‘snaazDydo7? UopowsiprT O00 00 0 Olowpetta nO O16 0 O.0 OOo OG ato jood SpooM jo UIs1BUL 4B ‘guyf’p UN YA 764 Tue Universiry SCIENCE BULLETIN duridae and Sminthuridae), and numerous species of Acarina, notably the “beetle mites” (Oribatoidea) and parasitoid mites. Of these, the lepidopterous larvae are worthy of further comment, here, for their trails through the mosses may often be mistaken for those of Dolichopeza. These caterpillars usually had reached a length of about 5 to 8 mm. by the time I first noticed them, and I saw some over a centimeter long, which I took to be later instars. The caterpillars are of a reddish brown color, sparsely haired, with large, glossy brown heads. They are extremely active creatures and ravenous eaters of moss. Their tunnels in the moss are fre- quently lined with webby silk fibers and are always littered with fecal pellets, which is what makes them resemble the trails of Dolichopeza. 1 feel fairly certain that these common pyralidoid caterpillars are the immatures of small, light brown or light grayish moths that are often taken in the rearing dishes. Mr. Sherman Moore, of Detroit, has identified several of the moths as Crambus alboclevellus Zell. (from Mnium cuspidatum), C. elegans Clem. (from Aulacomnium palustre), C. albellus Clem. (from Amblyste- gium varium), and Crambus sp. (from Dicranum scoparium). Mosses and liverworts while harboring many kinds of organisms appear to furnish food directly to only a few, such as the moth lar- vae just mentioned and certain kinds of Tipulidae. From the rock gorge habitats, I have reared Tipula (Oreomyza) ignobilis, Tipula (O.) fragilis, Liogma nodicornis, Limonia (Dicranomyia) sp., Eri- optera (Symplecta) cana and Gonomyia sp. From mosses of north- ern marsh borders, Tipula (Yamatotipula) sulphurea, Tipula (Tri- chotipula) oropezoides and Limonia (Dicranomyia) immodesta were reared, together with Dolichopeza sayji. In a discussion of the ecological relationships of the larvae of Dolichopeza, it seems appropriate to mention associations of lar- vae of various species within the genus, for it often happens that the larvae, like the adults, keep close company. Within an area of Tetraphis moss no more than three inches in diameter, for ex- ample, may be found larvae of americana, polita and tridenticulata. Larvae of obscura and subalbipes often occur together in the mosses of swamp hummocks. Other such associations are suggested in Table 3. The important thing about these close interspecies asso- ciations is that there appears to be no competition whatsoever. There seems to be always enough moss to feed all the larvae and to keep them comfortably apart from one another. Rarely does a natural growth of moss contain so many larvae that noticeable de- foliation results; in fact, I have observed only one such occurrence, THE CRANE Fry GENUS DOLICHOPEZA 765 in which several larvae of Dolichopeza americana and Liogma nodi- cornis feeding on a thin growth of Mnium punctatum on a sand- stone cliff had conspicuously damaged an area perhaps four inches in diameter. If larvae do encounter each other occasionally during their tunneling, it seems unlikely that competition in the form of fighting would result. Even when several larvae are confined in a dish without moss, they do not show any aggressive behavior (such as is known among some other tipuline larvae) and could be de- scribed in general as lethargic. In view of these facts, it is diffi- cult to imagine what form interspecies competition among larvae might take. Parasites and predators.—Probably the most important difficulties with which the life of the larva is beset, except for the rigors of weather already mentioned, are predation and disease (including parasitism). There would be no point to enumerating those organ- isms that might prey on Dolichopeza larvae. The survey of preda- tors on the Tipulidae in general given by Alexander (1920:721 ff. ) will suggest several possible ones. The only time I have actually witnessed predation on a Dolichopeza larva was an instance of D. walleyi attacked by a larva of a small species of Tabanus * in the moss Hypnum imponens growing at the edge of a woodland pool in southern Michigan. Evidence of parasitism is as rare as that of predation, and there are no established records of parasites in or on the larvae of Dolicho- peza. Certain hymenopterous parasites that become evident only during the pupal stage of the crane fly host (see section on pupal life history ) must occur within the larva throughout its life but do not hinder its activities and allow it to grow and metamorphose. In rearing dishes I have found the braconid wasp, Macrocentrus reticulatus Muesebeck, and an ichneumonid, Mesoleptus sp. + Con- cerning these, Dr. Muesebeck wrote: “The Macrocentrus must be a parasite of some lepidopterous larvae although no host seems yet to have been recorded for M. reticulatus. For species of Meso- leptus we have little host information. There are some records of the rearing of specimens of this genus from Diptera but I do not know how authentic these are.” Parasitism in certain crane flies other than Dolichopeza, which were observed, may be relevant here, inasmuch as the larvae in- volved were taken in microhabitats of Dolichopeza. Perhaps the * Identified by Dr. W. W. Wirth, U. S. National Museum. + Dr. Muesebeck identified the braconid, which was the first male of the species to come to his attention. Miss L. M. Walkley determined the ichneumonid. 766 THe UNIversIty SCIENCE BULLETIN most spectacular case was that of a nematode infestation of a larva of Tipula, found in a moribund condition. Two large, whitish nema- todes, one nearly 20 mm. long and the other over 30 mm. in length, were subsequently found to have pierced the body wall of their 16 mm. host and just completed their escape. Postmortem examina- tion of the larva showed the internal organs to be intact but the usual sheet of fat absent. The skin around the two exit holes was thickened and blackish, and the larva was left quite limp and somewhat flattened. This is the first instance of nematode infesta- tion of a crane fly larva to come to my attention, although it must be a common occurrence in nature. A bacterial infection causing eversion of a portion of the caudal end of the digestive tube has been observed in three species of tipuline larvae. Internal symptoms include nodular growths on all major portions of the alimentary canal. From these whitish cysts, smears were prepared. The contents were found to consist of many small, fat-like globules and highly motile, short rod bacilli, which proved to be gram negative. Colonies of the bacilli on blood agar were gray-buff in color, with a definitely cleared zone eight millimeters in width surrounding them. While this disease kills larvae within two days of appearance of first symptoms, I was not successful in inoculating larvae of Dolichopeza with the bacilli. PUPAE—MORPHOLOGY Size of the pupa depends upon both species and environment and is directly proportional to and smaller than the size of the fourth instar larva, just before pupation (Fig. 99). Pupae of the smaller species (americana, dorsalis and tridenticulata) are commonly only 10 or 11 mm. long, while those of larger species (such as similis ) reach lengths of 16 mm. or more. Measurement of both whole pupae and cast pupal skins is subject to the same sort of error as occurs in measurement of the other life history stages, due to the position of the insect—whether outstretched, contracted, how curved, and so on. There is always a certain amount of guesswork involved, to which experience may lend some reliability. The ac- companying illustrations of the pupa (Figs. 100, 101) were pre- pared from a specimen killed in hot water and thus slightly more than normally extended. No structures on the pupal head appear to be characteristic of Dolichopeza alone. The shape and position of the eyes and an- tennal sheaths, the rather triangular labrum, and the generally- flattened aspect of the entire head are typical of pupae of the sub- THe CRANE Fiy GENuS DOLICHOPEZA 767 Fic. 100. Pupa of Dolichopeza (Oropeza) walleyi, ventral aspect, indicating abdominal segmentation; abd—abdominal segment. Fc. 101. Same as 100, left lateral aspect; pn—pronotum. Fic. 102. Same as 100, dorsal aspect of thorax. Fic. 103. Mesothoracic respiratory horn of Dolichopeza (Oropeza) sp. Fic. 104. Cauda of pupa of Doli- chopeza (Oropeza) sp., female, left lateral aspect; cr—cercus, hy— hypovalve, 8s—eighth sternum. Fic. 105. Same as 104, male; gon— gonapophysis, id—inner dististyle, od—outer dististyle, ta—tergal arm, 8s—eighth sternum. 768 THe UNIvEeRSITY SCIENCE BULLETIN family Tipulinae. The cephalic crests are small, as in most species of Tipula. In having the tips of the sheaths of the maxillary palpi recurved, Dolichopeza differs from its nearest North American rela- tives, Brachypremna and Megistocera, and from more distantly related Tanyptera and Longurio, but resembles Nephrotoma and most forms of Tipula. Aside from the respiratory horns, the most conspicuous features of the thoracic dorsum of the pupa are two rough-surfaced, reticu- lated areas (Fig. 102) which are useful in recognition of the genus. However, it should be pointed out that certain species of Tipula possess similar reticulated areas on the pupal mesonotum (cf. Alex- ander, 1920: 977). Tipula ignobilis, for example, a species that often occurs in mosses containing pupae of some kinds of Dolicho- peza, is so marked; but in this species, as well as in the other species of Tipula pupae I have seen having such a reticulation on the thor- acic dorsum, the pattern of wrinkles extends anteriorly around the bases of the respiratory horns and onto the pronotum. In Dolicho- peza, on the other hand, the reticulation is confined to the mesonotal cuticle posterior to the pseudosuture, where it occurs as roughly triangular areas in the subgenus Oropeza (but weakly developed in subvenosa) or more oblong patches in Dolichopeza americana. The dorsal mid-line of the mesonotum is characterized by a low, anterior carina, behind which are a series of transverse ridges or folds crossing the mid-line in the region between the reticulated areas. There are half a dozen pairs of exteremly small spines on the mesonotum, disposed approximately as shown in Figure 102. Taka- hashi (1958: 121) has found the number of pairs of these spinules useful in distinguishing pupae of certain Japanese species of Oro- peza, but in the North American species their taxonomic use is impracticable. While at the outermost position, in Dolichopeza subalbipes, there are ordinarily two spines together, the pattern for the other species closely resembles that figured. Occasionally a spine may be broken off or missing altogther, or there may be only one where in the corresponding position on the other side two occur together. The metathorax is narrow and, where exposed dorsally, rather closely resembles the first abdominal segment. Only the proximal ends of the sheaths of the halteres are exposed, in the subgenus Oropeza, the remainders being normally concealed beneath the wing sheaths. In americana, however, the sheaths of the halteres THe CrANE FLY GENuS DOLICHOPEZA 769 are fully exposed but closely applied to the posterior edges of the wing sheaths. Close behind the head is the very narrow prothorax, which in this genus is distinctly separated from the mesothorax. The respira- tory horns, which are of rather uniform shape throughout all the species and which are ordinarily referred to in literature on the Tipulidae as “pronotal breathing horns,” are not situated upon this prothoracic segment. Alexander (1920: 752), describing the tipu- lid pupa in general, said: “Immediately behind the head on the pronotum are the two breathing horns. ” In the taxonomy of pupae of Tipulidae, the term “pronotal breathing horns” persists even to the present time (see, for example, Wood, 1952: 81, etc.). The respiratory horns, however, are situated on the pupal mesono- tum, between its anterior margin and the pseudosutures (which in the adult are indicated by the pseudosutural foveae, or humeral pits). Dissection of the pupa when the developing adult within is fairly well formed clearly shows the relationship of parts of the imaginal thorax to the areas of the pupal thorax as here identified. Furthermore, the tracheal tube leading from the respiratory hor connects to the mesothoracic spiracle of the developing adult. A most peculiar difference exists between pupae of Dolichopeza polita and all the other North American species of the genus in the appearance of the tracheal connection between the respiratory horn and the spiracle. In most species, this tube is rather straight, making a direct connection, while in polita it is greatly convoluted and folded back upon itself, its length easily twice the distance from the base of the breathing horn and the spiracle (Fig. 107). This condition is found in all three races of polita. The only other occurrence known to me is in the pupa of Dolichopeza (Nesopeza) geniculata Alexander, which I found projecting from a growth of liverwort on the slope of the volcanic peak, Halla San, on Cheju Island, Korea, in the fall of 1954. I know of few other characteris- tics in any stage of Dolichopeza that so completely set apart one species from all others in a continental fauna. The first abdominal segment of the pupa is short and indistinctly divided, while the next six segments are formed of broader, well- marked basal and caudal, or anterior and posterior portions. The basal rings lack spinous processes on the dorsal and ventral surfaces but have these present on the pleural folds. The posterior rings bear, in addition to the pleural spinous processes, transverse rows of more or less conical projections, in varying numbers, on the dorsal 770 THe UNIVERSITY SCIENCE BULLETIN surfaces of all segments and the ventral surfaces after the third or fourth segment. The eighth abdominal segment is not differen- tiated into rings. It bears four * large, subconical, spine-tipped pro- jections on its sternal surface and four slightly smaller ones dorsally, and it has the pleural spinous process present only in the “posterior ring” position. On the dorsolateral surfaces of the eighth abdominal segment, near its anterior edge, there are two low, blunt lobes, one at each side. Unlike the other projections of the pupal abdomen, these lack spinous tips and apparently are homologous with the fleshy lobes occurring in a similar position in the larva. This sup- posed homology seems correct, inasmuch as the lobes described are characteristic of pupae of species of Oropeza but do not occur in Dolichopeza americana. These lobes are lacking also in other genera of the Tipulinae, so that the same feature, in a sense, can be used to separate both larvae and pupae of Oropeza spp. from those of Tipula, Nephrotoma and allied genera. Sheaths of the various hypopygial elements of the adult com- prise the greater part of the caudal segment of the pupa. In the female these (sometimes termed the acidotheca) rather closely re- semble the terminal abdominal structures of the adult (Fig. 104). In the male pupa (Fig. 105) the sheaths bear less resemblance to imaginal structures, although the relationship of corresponding parts is clear. Comparison of Figures 111 and 112 shows a con- spicuous difference between male pupal terminal structures of Dolichopeza americana and species of Oropeza. The base of the sheath of the outer dististyle is adjacent to that of the tergal arm in Oropeza spp. (Fig. 111), but these two structures are distinctly separated basally in americana (Fig. 112). The differences in spinous tips of the caudal sheaths and projections of the eighth seg- ment apparent in these two figures are not reliable for subgeneric recognition, as there is too great an amount of variation in these parts. Recognition of species in the pupal stage of Dolichopeza would seem, from a cursory survey of structures, to be a fairly easy under- taking. The unfortunate truth is, however, that very little can be salvaged for taxonomic use at the species level from the described array of bristles, spinous projections and other apparently taxonomi- cally useful features. With only a scattering of species or a small number of specimens at hand, one might soon find some character in which the available pupae differ. Complications ordinarily arise, however, when larger numbers of specimens are examined, * Rarely will this number vary, but see Figure 38. THe CRANE FiLy GENUS DOLICHOPEZA Tira 107 108 NO Fic. 106. Spiracular yoke of pupa of Dolichopeza (Oropeza) simi- lis; is_—inner secondary lobe, ly—lobe of yoke, osl—outer secondary lobe, tr—larval tracheal trunk. Fic. 107. Mesothoracic respiratory horn and tracheal connection to spiracle in Dolichopeza (Oropeza) polita sspp. Fic. 108. Same as 107, Dolichopeza (Oropeza) triden- ticulata. Fic. 109. Arrangement of spinous processes of eighth ab- dominal sternum of pupa of Dolichopeza (Oropeza) sayi group. Fic. 110. Same as 109, Dolichopeza (Oropeza) obscura group. Fic. 111. Cauda of male pupa of Dolichopeza (Oropeza) sp., left lateral aspect. Fic. 112. Same as 111, Dolichopeza (Dolichopeza) americana. 772 THe UNIVERSITY SCIENCE BULLETIN especially when these represent species of a very uniform group. In some characters, like the bristles of the mesonotum already men- tioned, there is too much uniformity throughout the genus, while in others there is as much variation within one species as there is among species. An example of the latter is the number of spine tipped projections on the fifth, sixth and seventh abdominal sterna. The count of these conical points may vary from species to species and from row to row; but on one species, the count of projections in a certain row was found to vary from 8 to 13, complicated by partial fusions, branching, irregular spacing and unusual sizes of projections. As a further example, the spinous projections of the eighth abdominal segment may be single or multiple tipped and may be hooked and variously provided with bristles; but an ex- amination of variation in the tip of a particular projection in a single species showed that the tip varied from two to four branched and that bristles were either present or absent. One species group within the genus can, however, be separated on the basis of characteristics of the pleural spinous processes. In Dolichopeza carolus, dorsalis, obscura and subalbipes, these proc- esses are rather blunt-tipped, with an apical bristle that nearly equals the length of the subconical, basal part of the process. In the remaining species, these processes have bifid tips, one of the branches usually short, the other longer and very sharp, with the apical bristle arising in the notch between the two points. Species in the obscura group usually have the two medial projections of the eighth sternum originating from a common base or with their bases contiguous (Fig. 110), while in the sayi group these projec- tions are ordinarily spaced apart, their bases not touching (Fig. 109). This character is not wholly reliable, exceptions having been seen, but it applies in general. Another character that is nearly always reliable is the presence or absence of spinous projections on the posterior ring of the fourth abdominal sternum. In certain species, these are present and from half to nearly equal the size of those on the fifth segment, whereas in the other species they are absent altogether or represented only by bristles. One additional character, of a completely different sort, permits separation of species when used in conjunction with the group char- acters just noted. To my knowledge, this character has never been observed prior to this study and consequently is new to the classi- fication of crane fly pupae. Just posterior to the median projections of the eighth abdominal tergum, there is an infolding of the pupal skin, a sort of pouch or pocket, flattened dorsoventrally, situated THE CRANE FLY GENUS DOLICHOPEZA 773 with its closed end directed cephalad in such a way as to be rather readily visible through the eighth tergum, especially in cast pupal skins immersed in alcohol. Microscopic examination of this struc- ture reveals that tiny “stress lines” in the surrounding pupal cuticle converge into the pocket and that the remains of the larval spiracles are attached to the strongly sclerotized arch formed by the con- verging folds of cuticle (Fig. 106). Although I have not studied the development of this pouch, it seems to be intimately associated with, perhaps a result of, the withdrawal of the larval spiracles at the time of pupation, when the tracheal outlet shifts from a caudal to thoracic position. Since the entire structure seems to me to act as a rigid bar connecting the two spiracles and bringing about their coincident displacement, I have suggested the name “spiracular yoke” for the pouch and related cuticular thickenings at its sides (Byers, 1958: 136). The spiracular yoke in both males and females has a characteristic shape, within limits, for each species, and a study of many specimens indicates that certain features of the yoke are reliably constant. Figures 113 through 124 show something of the variation of the spiracular yoke from species to species. It is not practicable to attempt to describe the variation within each species, but it should be pointed out that the illustrations provided are com- posites, each based upon several pupae. In the key, reference is made only to the general form of the yoke, such as whether its lobes are elongated or short and whether there are inner, secondary lobes (Fig. 106, isl). I have observed the spiracular yoke also in other genera (Brachypremna, Megistocera and several species of Tipula), where it is again distinct for the species examined. Per- haps it will prove to be a character of general application in the classification of the pupae of Tipulidae. TENTATIVE KEY TO PUPAE OF NORTH AMERICAN DOLICHOPEZA Pupae are most easily studied in alcohol. Pupal skins preserved dry should be soaked a few minutes in alcohol, then cleaned of adhering moss particles, soil and other debris before being studied. Skins thus prepared may be flattened with a blunt needle so that the spiracular yoke is seen in the proper profile. Whole pupae may require removal of the eighth and ninth abdominal segments, which separate easily from the imaginal tissue after the fourth day of the pupal stage. Adult genitalia thus exposed may identify the species; the spiracular yoke, at any rate, is made more readily visible. Live pupae, given a few days, will provide their own identification. 774 Tue UNIVERSITY SCIENCE BULLETIN P| 4 3 114 15 Ju, DR, Sek 116 117 118 Awk Sank Donk Ng l20 I2| Bat, Sk, J, laa 123 124 Fics. 113-124. Spiracular yokes of pupae; 113—Dolichopeza (Doli- chopeza) americana, 114—Dolichopeza (Oropeza) carolus, 115—Dolicho- peza (Oropeza) dorsalis, 116—Dolichopeza (Oropeza) obscura, 117— Dolichopeza (Oropeza) polita, 118—Dolichopeza (Oropeza) sayi, 119— Dolichopeza (Oropeza) similis, 120—Dolichopeza (Oropeza) subalbipes, 121—Dolichopeza (Oropeza) subvenosa, 122—Dolichopeza (Oropeza) tri- denticulata, 123—Dolichopeza (Oropeza) venosa, 124—Dolichopeza (Oro- peza) walleyi. THe CRANE FiLy GENuS DOLICHOPEZA fads) The following key is limited in several respects. The pupal instar of Dolichopeza johnsonella, as well as that of the two new species described herein, is unknown. Probably johnsonella will “key out” near subalbipes. Pupal skins identified as those of carolus and sub- venosa were not obtained from reared specimens but were identi- fied on the basis of circumstantial evidence, as explained later un- der the respective species headings. Only four female pupal skins thought to be of swhvenosa and three male and three female skins regarded as of carolus were found. The pupae of dorsalis and similis are known from only about half a dozen specimens each, but of the remaining species several to many pupae have been found and studied. 1. Reticulated areas of mesonotum oblong in shape; no blunt lateral lobes on anterior portion of eighth abdominal tergum; spiracular yoke roughly triangular in shape and weakly sclerotized except BEE CLO ES a CIS uO) uk | prs ag a ae eR americana Reticulated areas of mesonotum triangular in shape; a blunt lateral lobe anteriorly on each side of eighth abdominal tergum; spiracular yoke variously shaped, usually with two conspicuous lobes, Strongly sclerotized sthroughout 9.2 Acls)... abies See ne eke oe 2. Tracheal tube connecting respiratory horn with adult spiracle strongly CONVO MTEC Me ee eee UN Arce le aN i a hd polita sspp. Mrachealmtupemstraroht: yy boep ate Wee, ie ute eee ein ge 8 8 8. Pleural spinous processes of eecond dinonrah seventh abdominal seg- ments blunt-tipped, bearing a single apical bristle about as long as basal portion of process Pleural spinous processes of second through seventh abdominal seg- ments mostly bifid at tip, with at least one sharply-pointed branch and with apical bristle arising from notch between branches . . 4. Middle projections of eighth abdominal sternum set apart so that the four are about evenly spaced (Fig. 109); spiracular yoke shallowly and irregularly emarginate, the lateral lobes shorter than their WAGEeA EN DASCm (HIG a MIND. Ange Soars A Ceo Ie eee i _. dorsalis Middle projections of eighth abdominal sternum set closer together than distance from either to outer projections, or divergent from a low base common to both (Fig. 110); spiracular yoke relatively deeply emarginate, the lateral lobes longer than their width at base, 5 5. Lobes of spiracular yoke only about one and a quarter to one and a half times as long as their width at base, their tips variously rounded; emargination between lobes usually smoothly curved (CLESSEF “LUNGS mete eee MER OAPEE Eee Meck = beak | the Re ee Rn obscura Lobes of spiracular yoke from one and a half to three times as long as their width at base, their tips irregularly truncated, often ex- panded and notched; emargination between lobes more angular Cir Ghani BT Qos Aa 5 ee ea eee oe. hs en oe er 6 6. Basal part of spiracular yoke deep, or thick, the emargination between lobes irregular by reason of inner secondary lobes (Fig. 114); middle projections of eighth abdominal sternum clearly separated bo ~l 776 THe UNIVERSITY SCIENCE BULLETIN at base; spinous projections on fourth abdominal sternum present, about one-third to one-half as long as those on fifth sternum; projections on fifth through eighth sterna very darkly tipped. . .carolus Basal part of spiracular yoke usually shallow, or slender, the emar- gination between lobes not interrupted by secondary lobes (Fig. 120); middle projections of eighth abdominal sternum set very close together or arising from low base common to both (Fig. 110); usually no spinous projections on fourth abdominal sternum, those of fifth through eighth sterna not all darkly tipped... . subalbipes 7. Middle projections of eighth abdominal sternum set close together or divergent from a low base common to both (Fig. 110); spiracu- lar yoke broadly and shallowly emarginate, the lobes shorter than their width at base, sometimes with an apical papilla (Fig. 122), tridenticulata Middle projections of eighth abdominal sternum set apart so that the four are about evenly spaced; lobes of spiracular yoke slightly to much longer than their width ‘at base ’".>...'. 2. 2 8 8. Spinous projections on fourth abdominal sternum usually conspicuous, about one-half as long as those on fifth sternum, or longer ...... 9 Spinous projections on fourth abdominal sternum usually inconspic- uous or absent, sometimes represented by bristles .............. Le | 9. Lobes of spiracular yoke only about one and a half times as long as their width at base, often bluntly rounded at tips; emargination between lobes rather smoothly curved (Fig. 124) ........... walleyi Lobes of spiracular yoke about twice as long as their width at base, their tips usually slightly expanded and notched; emargination between lobes more angular, often irregular due to presence of inner secondary ‘lobes’ 9: . 25) 0. ens es oo ue 10 10. Reticulation on mesonotum distinct, occurring in roughly triangular areas: spitacular yoke. as. im) Figure’ 123°... > 22) ae venosa Reticulation on mesonotum obscure, mostly in form of low, transverse wrinkles or folds similar to those along mesonotal mid-line; Spiracular yoke! as in) igure Ii 2s.) en subvenosa 11. Lobes of spiracular yoke about three times as long as their width at base; emargination between lobes deep and angular, with small inner secondary lobes; basal portion of yoke slender, not thick (Glctitcha ( S ) alenee eea Nee Rena Bate AAMONMIR a ES similis Lobes of spiracular yoke one and a half to two times as long as their width at base; emargination between lobes angular, some- times irregular due to inner secondary lobes, but not deep; basal portion .of ‘yoke thick (Figs, 718) 225... 24 ac: ole sayi PUPAE—NATURAL HISTORY In nearly all rearing experiments, the length of the pupal stadium in Dolichopeza was six days, which is about average for the Tipuli- dae generally. Occasionally the period was found to be seven days, and one female of Dolichopeza venosa spent eight days in the pupal condition, indoors, but did not emerge normally and died in the process. In the laboratory, the time of emergence was artificially THE CRANE FLY GENUS DOLICHOPEZA 777 altered, that is to say postponed, by exposure of pupae to bright light or to chilling. It therefore seems reasonable to believe that naturally occurring fluctuations in environment could effect changes in the length of time that the fully formed adult insect within the pupal skin would remain there. It would be interesting to know, for example, whether Dolichopeza adults emerge at the end of a six-day pupal stadium, if that time happens to coincide with an all-night rainstorm. Although it is essentially in a resting state, in which many larval structures disintegrate and the adult anatomy takes shape, the pupa is capable of considerable activity at nearly any time during its existence. Vital physiological activities of the insect must, of course, proceed without interruption; and even as it frees itself of the last larval skin the pupa may be seen to have already a fully-formed and functioning tracheal system and a regularly-pul- sating dorsal vessel, both visible through the smooth, pale integu- ment. Pupal activities in ecdysis show that at least a part of the muscular system also is already in operation. In the absence of artificial stimulation, however, the pupa usually remains quiescent, only occasionally wriggling to the opening of its tube and perhaps projecting slightly from it. This movement is accomplished by dorso-ventral undulations of the abdominal segments behind the tips of the wing sheaths, which by their stout spinous projections gain a firm hold in the moss. Since the pupa lives in the same microhabitat as the larva, little comment needs to be added concerning its general ecology. Each pupa lives in isolation within a burrow often scarcely more than half again as long as the pupa itself. Where moss habitats have been favorable for the development of several larvae, the pupal burrows may be numerous and often crowded, as for example at one place where I found three pupal skins of Dolichopeza polita ssp. projecting from one square centimeter of moss surface. Predation and parasitism—tThe relatively defenseless pupae would seem to fall easy prey to several kinds of predators that live in mosses or feed rather generally over the surface of the ground. Only once, however, did I actually find evidence of predation on a Dolichopeza pupa. As I was examining a rearing dish, I saw a pupa of Dolichopeza similis come wriggling up out of the moss in a most hurried manner. When I saw that its two posterior segments had been nearly torn away, I probed its tube in the moss and dis- covered a twenty millimeter larva of some species of Elateridae.* * Alexander (1920:729) reports elaterid larvae feeding on a dead larva of Tipula trivittata. 778 THE UNIvERSITY SCIENCE BULLETIN Larvae of Bittacus apicalis (Mecoptera) found foraging in moss containing pupae of Dolichopeza americana and polita were prob- ably seeking only dead pupae or other such decomposing material, as they fed readily on dead pupae and refused living ones when both were offered as food. Several instances of parasitism of pupae by chalcidoid wasps were noted in the course of rearing experiments. At first, the infested pupae were found dead, broken open at one or more places, with half a dozen or more tiny, spindle shaped larvae busily consuming both the pupal contents and the necrotic edges of the openings in the skin. These larvae were about two millimeters in length, pale at the ends, with a broad, pale spotted band of tan showing through the smooth, transparent skin of the middle region. It appeared that these larvae were only scavengers, there being no direct evidence of parasitism. Yet I wondered why the number of them feeding on any one pupa was always about the same and why they were never found feeding on dead larvae in the same rearing dish. Attempts to rear these supposed parasites invariably failed, until on 11 June 1952 I found in one of my dishes a newly-formed pupa of Dolichopeza walleyi, through the pale integument of which I could see the familiar larvae at work. Inspection showed that the pupal skin was intact, making it clear that the larvae were indeed parasitic. It seems that the parasites must remain very small throughout the larval life of their host and then develop rapidly as soon as the pupa is formed, for, although there is no interference with the process of pupation, by the time the pupa is only a day old it is largely hollowed out by the parasites. By 14 June, activity within the pupal skin had ceased, and two days later the whitish pupae of the wasps could be seen. On 17 June, the reddish eye pigment of the parasites became visible, and between evening of 19 June and morning of the next day there occurred an abrupt change in color of the parasites, the eyes turning deep red and the bodies nearly black. Between 10:00 a.m. and 3:30 p.m. on 23 June, about a dozen tiny wasps emerged through a small hole in the side of the fly pupa. These were identified by Dr. B. D. Burks of the U. S. National Museum as belonging to an undescribed species of Tetrastichus (Hymenoptera: Eulophidae). Similar wasps were subsequently reared from a pupa of Dolicho- peza americana, in which instance, however, there were only seven parasites. Other species of Dolichopeza parasitized by Tetrastichus sp. include obscura and subalbipes. All cases of infestation by this THE CrANE FLY GENus DOLICHOPEZA 779 parasite were observed in May and June, but there is as yet no reason to suspect that pupae of the fall generation are not also parasitized. Rogers (1933: 35) reports braconid parasites in the pupa of Dolichopeza walleyi (as Oropeza sayi); however, as none of these wasps was preserved, there is no way to be sure whether what he actually observed were also Tetrastichus. It seems probable that the mites often found on adults of Doli- chopeza congregate on the pupa and attach themselves to the body of the imago as it emerges from the pupal skin. Pupal development.—The newly-formed pupa is pale tan in color, often with a greenish tinge showing through, especially from the abdominal contents. The dorsum of the thorax is usually a slightly darker tan or brown than is the rest of the pupa, and the thoracic respiratory horns are quite dark brown from the outset of pupal life. During the first few hours of the pupal stadium, the eyes are pale, but the violet-purple pigment of the developing om- matidia within slowly diffuses across, beneath the pupal eye cover- ing, so that at the end of the first day the pupal “eyes” are fully colored and have a brick red appearance. As the individual lenses of the developing adult eye are clearly visible through the pupal skin, it may be seen that the compound eye does not initially form directly beneath the pupal eye cover but is displaced slightly an- terolaterally. Within the wing sheaths, the veins sometimes glisten as if filled with air, and the membrane appears as a dark, strongly- convoluted, compactly-folded mass. As the pupa grows older, its colors become darker, the greenish hue of the abdomen being replaced by brown in most species of Oropeza but remaining in americana and to a lesser extent in dor- salis and subalbipes. The thoracic breathing horns become very dark brown with black tips, and the low median carina of the mesonotum takes on a black color in many species. As the adult body takes shape, some parts of the pupal skin are left apparently empty, or nearly so, by the contraction of the initially rather amor- phous tissue of the young pupa into the smaller and more compact adult structures. The hypopygial elements largely pull out of their formative sheaths, and the rostrum, maxillary palpi, antennae and other parts loosen from their sheaths, so that by the end of about four days of the pupal stadium the developing adult is separated almost completely from its pupal covering. Very marked changes in over-all appearance are undergone by the pupa shortly before the adult is to emerge. When the pupal skin grows rapidly darker brown in color and becomes dry and 780 THe UNIverRsITy SCIENCE BULLETIN shiny, emergence of the adult may be expected within the follow- ing eight to twelve hours. At this time, the coloration especially of the head, thorax and wing and leg sheaths becomes noticeably darker, somewhat resembling dark amber-brown glass. Eye color changes from brick red to almost black in only a few hours. On the surface of the wing sheaths, the characteristic pattern of venation of the adult wing, in miniature, can easily be made out in reflected light. In the laboratory, pupae often moved one-third to half out of their tubes at the onset of this drying and darkening period, but I have not been able to verify that they do so under natural con- ditions. Emergence of adult——Emergence of the adult is a rapid process, requiring only a few minutes. I have twice had the good fortune to observe emergence from its earliest moments, when the pupal skin first began to split, and have witnessed by far the greater part of the process on several other occasions. In preparation for emer- gence, the pupa moves out of its burrow far enough to expose the anterior one-third to one-half its length. After a time, the pupal skin parts in two places: a transverse split between the antennal sheaths and the pronotum, from the mid-line behind the head down to the sheaths of the front legs, and a longitudinal split along the dorsal mid-line, across the occipital portion of the head, the pro- thorax and the mesothorax almost to the metathorax.* Then, for a few seconds, the adult insect, still fully encased, swallows air in order to expand and lengthen its body. Air is taken into the ab- dominal part of the ventriculus but much more extensively into the membranous crop, which in one fly killed a few seconds after emer- gence was found distended so as to fill the abdomen back to the anterior edge of the fifth segment. As a consequence of this infla- tion, the adult body rises from the pupal skin rapidly and without apparent exertion, until the wings are drawn free of their sheaths, the femora are fully exposed, and approximately the anterior half of the abdomen has emerged (Figs. 125, 126). Although the com- pressed neck quickly extends, the head remains bowed, with the antennae and maxillary palps directed downward. The wings are not withdrawn as crumpled structures but, somewhat in the manner of an opening parachute, spread to their full expanse as fold after fold of the membrane is drawn taut by the pull from outside. Dur- ing the minute or less that is required for emergence to the point of freeing the wings, almost the only motion noticeable is the slight tugging action of the coxae. * The coverings of head and antennae are left attached to the rest of the pupal hull by only a thin, membranous skin and sometimes break off as a unit, a sort of mask. Tue CRANE FLY GENuS DOLICHOPEZA 781 128 Fics. 125-128. Successive stages in the emergence of the adult of Dolichopeza (Oropeza) polita ssp. trom the pupal skin. With the greater part of its body free of the pupal skin, the fly begins a rocking motion, combining dorso-ventral movements and extensions and contractions of the abdomen. The legs are with- drawn with somewhat less exertion, the joints between femora and tibiae flexed first out to the sides (Fig. 127), then high overhead, as the long tarsi are pulled free. As the legs are freed, the tip of the abdomen remains fast in the pupal skin, the wings are kept folded over the back, and the antennae are directed upward (Fig. 128). Then, with its feet braced against the moss and its wings fluttering weakly, the fly suddenly breaks completely away from its 782 THe UNIVERSITY SCIENCE BULLETIN pupal hull, moves a short distance over the moss surface, and pauses to rest, with legs outspread and the weight of the thorax and abdo- men directly on the moss.* The soft and almost colorless sclerites begin to harden and take on slight color in a matter of minutes, while the body size reduces by discharge of swallowed air and of the last of the copious green- ish meconium, the fluid wastes resulting from pupal metabolism. Most of the meconium is expelled during emergence, leaving the pupal skin almost a quarter full of the greenish fluid, unlike the condition found in most Tipulinae, which leave their pupal skins empty. It may be a quarter of an hour before the fly attempts to take to the wing, and it seems that in its natural surroundings the newly-emerged fly must be easy prey for all manner of predators that are active at night.| Within two or three hours, the fly is well colored and hardened and is capable of sustained flight. Its full coloration develops in about six to eight hours following emer- gence, although because of callowness of body the fly may be con- sidered teneral for a longer period of time. With emergence, the life cycle begins anew. Figure 99 summarizes the stages of develop- ment, indicates the approximate rate of growth, and illustrates the changes in body size as the insect passes from egg through the lar- val instars, pupation and emergence as an adult fly. KEY TO ADULT MALES OF NORTH AMERICAN DOLICHOPEZA 1. Wings without discal cell (cell lst Me); cross-vein m-cu joining media before its first branching (Fig. 2); ninth tergum without lateral arms (Fig. 129) ......... subgenus Dolichopeza; note that the male of D. (D.) borealis new species is so far unknown; see discussion of ‘thatspecies: ..1 9... .. 615: = eee americana Wings with discal cell normally present; cross-vein m-cu joining the media beyond its first branching (Fig. 1); ninth tergum with tera Sarins aete. eee en Pee er en cae ae subgenus Oropeza 2 2. Gonapophyses shaped like small knobs, bearing decurved, stout black spines and bristles (Fig. 6) ............... obscura group 3 Gonapophyses with tips flattened and blade-like or slender and pointed, sometimes with hairs but never with stout black spines, sayi group 10 3. Tarsi white and femora and tibiae pale with narrowly darkened tips; body coloration yellowish marked with dark brown or black; wings with intense brown stigmal spot; ninth tergum as Figure 208 OT FQOS A) tens 2S Ral oat dee oe Sree ee eae ee oe subalbipes *JIn instances of emergence from pupae in mosses on the undersides of rock ledges or on vertical surfaces, some variation of the described procedure is of course necessary. Such conditions, however, were not duplicated in the laboratory. + Emergence in the laboratory usually occurred in the early hours of darkness, between 9:00 p.m. and midnight. _ 10. iE THe CRANE FiLy GENUS DOLICHOPEZA 783 Tarsi dusky, brown, yellowish brown or yellowish white, but tips of femora and tibiae not contrastingly darkened; body coloration brownish with dark brown markings ................... 4 . Apical two-thirds of tarsi yellowish white, grading into light Meow on basal third; medio-posterior margin of ninth tergum with two broad, rounded lobes and a slightly longer central tooth (Fig. 137) australis Tarsi uniformly colored throughout, dusky, brown or yellowish brown and not conspicuously, palerthan degs*...2 8 3.2.2 2.54. ALS 5 Medio-posterior margin of ninth tergum undulating, with two broad! rounded lobes; tergal arms slender and not much expanded at tips (Fig. 157); a median brush of black hairs on eighth sternum (CUET a2, TH 6D) 9 Den PN VR Se PoP [ONSET Seed RTT ee eee eee johnsonella Medio-posterior margin of ninth tergum with projecting teeth; tergal arms various; no brush of hairs on eighth sternum ........ 6 Tergal arms widely flared and emarginate at tips; teeth of ninth tergum not set close together (Fig. 164) ................. obscura Tergal arms not flared or emarginate at tips; teeth of ninth tergum set close together, usually on a common basal projection . . Outer dististyles blunt-tipped and not expanded at bases; median Ihe of tergal margin subrectangular in shape, bearing three teeth of neanlypequal lengthy (Mig DOL) nie ok | nes pees seach tridenticulata Outer dististyles with pointed tips and with basal portion usually enlarged; median lobe of tergal margin various but usually with three teeth, the central one the longest ......... polita subspecies 8 Thoracic dorsum shiny or “polished”; basal enlargement of outer dististyle darkened around the edge; median lobe of tergal margin usuallyaslender but mayevary (Fig; I71) ........cs.b.2- polita polita Thoracic dorsum dull and opaque, rarely slightly shiny in certain lights; basal enlargement of outer dististyle pale; median lobe of tergal margin usually distinctly three-toothed .......... 9 Three teeth of median lobe of nearly equal length, the Snide one slightly the longest, and all set close together; tergal arms flared about mid-length and tapering to slender tips (Fig. 188), laf ‘ polita cornuta Three teeth of median lobe variable in length, middle tooth the long- est, but teeth not always set close together; tergal arms only slightly expanded at tips but widest there (Fig. 180) ..... polita pratti Outer dististyles pale, shorter than inner dististyles, not conspicu- ously projecting (Fig. 145); tarsi white and femora and tibiae pale Witwenarowly: Garkened: tips .+ af omeaeee so Se te. ol. G _ carolus Outer dististyles dark or yellowish, longer than inner dististyles and conspicuously projecting; tarsi yellowish white, tan, gray or | SED i 3) 5 le ea On ane a and Ne CRI REMC Ey gh) Sere Cs, Sake at a ee iti Gonapophyses narrow, tapering to blackened, heavily sclerotized points; tergal arms broadly expanded and spatulate at tips; medio- posterior margin of ninth tergum with a broad, irregular pro- TKGIIIOVT sa oe ee agin ee ade eg Acetate Porn! Pe ee a ee ee ee ee 12 Gonapophyses not narrow and tapering to points; tergal arms and margins various but not as above 784 12. 13. 14. 15. THE UNIvERSITY SCIENCE BULLETIN Adminiculum with a prominent, subapical, posterior (or ventral) Splinteg BAAN Oi scke eG os cues art Metsad eects, eeu ea he subvenosa Adminiculum without a subapical spine (Fig. 234) ............ venosa Ninth tergum very shallowly emarginate, with undulating margin; tergal arms not expanded at tips (Fig. 149); gonapophyses short, terminating in an abrupt, spinous tip (Fig. 152) .......... dorsalis Ninth tergum slightly to moderately emarginate and toothed; tergal arms slightly to markedly expanded at tips; gonapophyses longer than broad, variously expanded at tips 22:2. 2. )..2 =e 14 Gonapophyses slightly widened toward the tips, the margins entire; tergal arms only slightly expanded at tips (Fig. 195) ........... sayi Gonapophyses expanded at tips, the margins irregularly toothed; tergal arms markedly expanded at tips ......:./.0/5933 15 No prominent projections from inner faces of gonapophyses; ad- miniculum without spines at tip (Fig. 241)" ....... 9.292203 walleyi A long, spinous projection directed dorso-cephalad from the inner (anterior or dorsal) face of each gonapophysis; adminiculum with two small apical spines (Fig: 204) >.<. 2). 2.7 See similis KEY TO ADULT FEMALES OF NORTH AMERICAN DOLICHOPEZA 1 Wings without discal cell (cell Ist M2); cross-vein m-cu joining media before its first branching (Fig. 2) .. subgenus Dolichopeza 2 Wings with discal cell normally present; cross-vein m-cu joining the media beyond its first branching (Fig. 1) ..subgenus Oropeza 3 Hypovalves of ovipositor each with a heavily sclerotized apical tooth (Fig. 135); tarsi white from mid-length of first segment to apex OL fourth! -., S-(peteseess Bop Oh ARG Rete Cee eee americana Hypovalves of ovipositor tapering evenly to tip (Fig. 136); only about apical one-sixth of first segment and all of second segment of tarsi pale (white in hind tarsi only), third segment dusky, fourthy and. fitth, dark... .... .s00 so... Soe an borealis Tarsi white; tibiae and femora pale in color, their tips conspicuously darker. and sharply contrasting... .. = <..3,:,420 40k Gee 4 Tarsi dusky, yellowish brown, or brown; or, if whitish, without dark- ened, contrasting tips on tibiae and femora...................- 5 A heavily sclerotized spot, conspicuously darker than the tip or any other part of the hypovalve, present on each side of eighth sternum (Fig. 148), approximately below the ninth or base of the tenth (igh Ae ee Re carolus No such spot present, although entire eighth sternum and hypovalves may be uniformly strongly sclerotized; or spot may show weakly, in rare instances, never as strongly sclerotized as tip of hypo- BW ese Ano es si vadvduse SS cu oo er IRN eo subalbipes Dorsum of thorax contrastingly colored or striped, the darker portions much darker than the pleura; if striped, the stripes distinct; body coloration generally yellowish brown, the abdomen annulated with darker brown or black; legs yellowish, yellow-brown, yellow- ish svhite: or brown... 02). ..5 5, 2/6 ids Gey. UE ee 6 10. 1 THE CRANE Fiy GENUS DOLICHOPEZA 785 Dorsum of thorax not contrastingly colored or striped and not con- spicuously different in color from pleura; if striped, the stripes indistinct and not much darker than the ground color; general body coloration dusky brown, dark yellowish brown, or light brown, annulated with darker brown in each case; legs dusky, gray-black, dark brown, brown, or brown with pale, yellowish brown tarsi. . 12 Thoracic pleura pale and unmarked (except for slight darkening of the pre-episternum in occasional specimens................... a Thoracic pleura distinctly marked, especially on the anepisternum, ventral part of pre-episternum and mesothoracic meron (see Fig. 14) Dorsum of thorax reddish brown in color, sometimes indistinctly striped, with a narrow, cocoa-brown median stripe usually present but not clearly differentiated; flies of bogs, marsh borders and Swalnplanadssespeciall yume: seein 2. CAO! Pe LAM AES dorsalis Dorsum of thorax marked by three distinct, reddish brown stripes on a yellowish brown ground color; principally flies of mesic wood- lands, in northern and eastern North America.......... walleyi (part) Cubital vein scarcely, if at all, more conspicuous than other wing veins (use no magnification); stigmal spot usually not very sharply defined and not intensely colored; flies of mesic woodlands generally south of Virginia, Kentucky and southern Indiana, and west of Illinois, Wisconsin, etc. (cf. 7, above)......... walleyi (part, Cubital vein clearly more conspicuous than any other (use no magni- fication) by reason of a dark seam of color along the vein; stigmal spot intense, well defined and much darker than the Sak i color OfathemwinCase ee. eee Tenth tergum more strongly sclerotized than cerci or hypovalves, although dark spots may be present on hypovalves and eighth sternum (Fig. 207); stripes of thoracic dorsum brown to dark brown, tending to merge behind middle of prescutum; large flies of yellowish coloration, the wings with a gold-amber tinge. ... . |: similis Strongest sclerotization of ovipositor is on hypovalves; stripes on thoracic dorsum dark brown, distinctly separated; wings tinged Withelightaorayishubrowils... seta: mee hos TL eres 10 Legs dark grayish brown, the tarsi light brown; body coloration yel- lowish with narrow annulations at incisures of abdomen; dark coloration on hypovalves neither very intense nor widespread ( Fig. 200); flies of marsh borders and swamps, especially............ sayi Legs pale, the tarsi whitish, yellowish white, or very pale tan; body coloration tan with broad, dark brown annulations at incisures of abdomen; dark coloration of hypovalves intense and generally dis- tributed behind eighth sternum, the posterior margin abruptly paler (Figs. 220, 237); flies of rocky areas and stream borders... 11 Flies of the Appalachian Mountain region, from northern West Vir- ginia southward through Virginia and North Carolina to northern (GEOL CLARE ee ee Ey ee in eran 5 bac ae. wos ae ates subvenosa Flies of northeastern United States and Canada, from northern West Virginia northeastward to Nova Scotia and northwestward to Min- nesota,, Alberta and, Yukond ees. 6 sedis tte oS PA ee venosa 786 13. 14. 15. 16. 17 THE UNIversiry SCIENCE BULLETIN Legs light brown, tarsi fading to pale yellowish brown or yellowish VELLeR Me eee, bb. carrie - aisind « saa Ben eee australis Legs and tarsi of about the same color, brown, dark brown, dusky or blackish). 4 Savas sa toad as Made e cleo d eee eeee 13 Pronotum, prescutum and pleura of thorax markedly lighter in color than occiput of head; antennae with the two basal segments abruptly paler’ than: flagellum }.... +24. In 4. al ee - eae ee 14 Pronotum, prescutum and pleura of thorax not much, if at all, lighter in color than occiput of head; antennae more unicolorous, or, if basal segments are pale, the flagellar segments gradually becoming darker toward the tip of the antenna... ./..... 2...) 5 eee 17 Wing tinged with golden brown, the stigmal spot darker brown, with lighter areas both before and beyond it; the larger veins also dark brown, with a seam of brown color along veins Cu and m-cu; prescutal stripes usually evident, although poorly defined; abdom- inal annulations distinct; an intensely sclerotized spot medially on eighth sternum at base of hypovalves (Fig. 163)...... johnsonella Wing faintly tinged with light gray or grayish brown, the stigmal spot only slightly darker than the ground color; veins gray to brown, lacking seam of color along Cu or m-cu; prescutum uni- colorous or nearly so; abdominal annulations indistinct; no intensely sclerotized spot near base of hypovalves, although hypovalves may be in part strongly sclerotized.............. (polita subspecies) 15 Thoracic dorsum: shiny, or, polished ..4-. 92% 4a eee polita polita Thoracic dorsum usually dull and opaque, light brown or light red- dish, brown: in Colom s.{54%5 ..s0)27. get etna} ae, ee 16 Flies of Minnesota, western Wisconsin, Iowa, western Illinois, Mis- souni, .Kansas,.,Arkansas, jete:ies: icy i fats Sues Se polita pratti Flies of Indiana, eastern Illinois, Ohio, Kentucky, southern Michigan, western New: York net@sawees-14 oe. erie be ee polita cornuta Tenth tergum more densely sclerotized than any other part of ovi- positor (Fig. 170); thoracic coloration usually as dark as occiput of head; stigmal spot grayish brown to brown.............. obscura Subterminal area of hypovalves more densely sclerotized than any other part of ovipositor (Fig. 229); thoracic coloration usu- ally slightly paler than that of occiput; stigmal spot grayish Brow otieks he ..cehersge ee oe ee, ee tridenticulata Subgenus Dolichopeza Dolichopeza (Dolichopeza) americana Needham Literature references.—Dolichopeza americana Needham. Need- ham, 1908: 210-211, pl. 16 (wing); Johnson, 1909: 117, pl. 15 (wing); Alexander, 1919: 929, pl. 43 (wing); Alexander, 1920: 981- 982; Dietz, 1921: 259; Alexander, 1922b: 61; Alexander, 1924: 59; Alexander, 1925: 172; Johnson, 1925: 82; Alexander, 1926: 239; Pierre, 1926: 9; Leonard, 1928: 698; Alexander, 1929a: 236; Alex- ander, 1929b: 25; Johnson, 1929: 180; Alexander, 1930a: 272; Rogers, 1930: 22; Alexander, 193la: 138. THE CrANE FLy GENuS DOLICHOPEZA 787 Dolichopeza (Dolichopeza) americana Needham. Alexander, 1936: 279; Alexander, 1940: 618; Alexander, 1941a: 295; Alexander, 1942: 210, fig. 24A (wing); Rogers, 1942: 118; Foote, 1956: 221. Original description —‘Its expanse of wing is 21 mm. Its color is brownish, paler ventrally. Its antennae are of moderate length, with the brown flagellum consisting of ten segments, slowly di- minishing in length toward the tip and beset with a few stout, black hairs. The wings are of pale brown, with venation as shown in the figure just cited (fig. 5 of pl. 16), the halteres are infuscated at tips. The legs are of the usual excessive length; femora and tibiae are brown, with white bases, and all the tarsus is white except the basal half of the first segment and the apical half of the fifth seg- ment.” Types.—No holotype or syntypical series was specifically desig- nated by Dr. Needham, who wrote to me, in August 1951: “The specimens of Dolichopeza americana went to the New York State Museum at the conclusion of my third report. Whether they have been preserved or not I do not know; some other material has been lost.” No specimens of this species are now in the collection at Albany, so I must assume that the type series is lost or destroyed. However, because americana is readily recognized, it seems un- necessary at this time to designate a neotype. Diagnostic characteristics—This is the most easily recognized species of Dolichopeza, for because of its dusky coloration, long, slender legs and white feet it is not likely to be confused with any other North American crane fly. Although it somewhat resembles Hexatoma albitarsis (O.S.), a species that sometimes occurs in the same general habitats, americana may be readily recognized in the field by its much longer and more slender legs. Its usual resting posture—hanging from overhead support by the prothoracic legs only, with the wings outspread—is also an aid in field recognition. From the only other known American species of the subgenus Dolichopeza (that is, borealis), americana may be distinguished by its almost completely white tarsi, as well as by certain other color and structural differences discussed under that species. Descriptive comments.—At a distance, americana appears to be colored black with white feet, but at close range it is seen to be of a dark brown color dorsally, paler beneath, at the tip of the abdo- men, and on the thoracic pleura, which are marked by dark spots on the mesothoracic meron, pre-episternum and anepisternum. The prescutum is indistinctly marked by three longitudinal stripes of a 788 THE UNIvERSITY SCIENCE BULLETIN rich, dark brown color, while the scutellum and postscutellum are abruptly paler than all other dorsal surfaces, being a glossy grayish tan. These and other color markings of the species are frequently obscure on pinned specimens of much age. Membranous areas of the abdomens of newly-emerged females often have a greenish color, which carries over from the larval stage; this is only tran- sitory, however, and gives way in a few days to the characteristic brown. Venation in americana is extremely constant throughout the spe- cies’ range, only a few abnormalities having been found among more than a thousand specimens examined. Presence of the vein Sc, has been noted two times, absence of a section of M, has been found twice, and an indication of a vein, in the form of a row of macrotrichia, projecting into cell M, has been seen in one specimen. I collected a specimen in North Carolina that had a spurious cross- vein in cell R, and in West Virginia one that had a distally directed spur vein from the m-cu cross-vein. One instance of a forked M, vein has been seen. Males vary in over-all length of body from slighty over 7 mm. to about 10 mm., their wings from about 8 mm. to approximately 11 mm., nearly always just a bit longer than the body. Females average somewhat larger, the body 8.5 mm. to 12 mm., wing 8.5 mm. to 12.5 mm. All the smaller specimens were collected in August, while the larger ones were taken in June, mostly from the northern states or high in the Appalachians. In its hypopygial structure, the male americana is markedly different from the species of Oropeza, although except for the tergal arms each part in the one finds its counterpart in the other. The ninth tergum is intensely sclerotized, emarginate medially but with a narrow, divided backward projection along the dorsum of the tenth segment (Fig. 129). Into the two deep concavities of the ninth tergum fit the strongly recurved tips of the inner dis- tistyles. There is no Oropeza in which the apical portion of the inner dististyles is so reflexed or fits into the sculptured margin of the ninth tergum. The outer dististyles are shorter than the broad, basal portion of the inner dististyles, are pale straw yellow in color, and are covered with fine hairs. Viewed from above, they are slender, but they are somewhat flattened, so as to appear broader when seen from the side (Fig. 131). The gonapophyses are yellowish and are partially fused to form a trough-like backward extension of the ninth sternum (Fig. 133). Adminiculum and penis are pale yellow, less strongly sclerotized than these same struc- THE CRANE FLY GENUS DOLICHOPEZA 789 oO LOmm. SESE ee | oO 0.5 1.Omm. l ; ; : ! | | SCALE, FIGS. 135-136 SCALE, FIGS. 129-134 Fics. 129-135. Dolichopeza (Dolichopeza) americana; 129—ninth tergum of male, with tenth tergum projecting from beneath, 130—left inner dististyle of male, dorsal aspect, 131—left outer dististyle, lateral aspect, 132—vesica and its apodemes, dorsal aspect, 133—vesica, penis, adminiculum and gonapophyses, left lateral aspect, 134—hypovalves of female, ventral aspect, 135—terminal abdominal segments of female, right lateral aspect. Fic. 136. Terminal abdominal segments of fe- male Dolichopeza (Dolichopeza) borealis new species, holotype, right lateral aspect. 790 THe UNIVERSITY SCIENCE BULLETIN tures in any of the species of Oropeza. In comparison with the eighth sternum, the ninth in americana is much smaller than it is in any Oropeza, suggesting a closer relationship with certain species of subgenus Nesopeza in which the ninth sternum is nearly con- cealed by the eighth. The ovipositor of the female americana is unlike that of any other species of the genus in North America, in that the hypovalves each terminate in a blackened tooth that projects beyond the otherwise truncate end (Fig. 135). Also, the ninth tergum is somewhat longer in this species than in species of Oropeza of comparable size. As described by Needham, the number of antennal segments is twelve. I cannot find any evidence of fusion of the last two seg- ments, in microscope slide preparations of the antennae, but one slide (Haywood County, North Carolina, 30 June 1924, J. S. Rogers ) clearly shows thirteen segments, a small apical segment present on each antenna, as is the case in species of Oropeza. Alongside the lower edge of each eye, there is a conspicuous black spot. Such a spot occurs also in Dolichopeza (D.) albipes of Europe but is lack- ing in the only other American species of this subgenus, as well as in Oropeza. The significance of these spots is not known to me. Geographical distribution —At the outset of this study, the known range of this species included only eastern Labrador and the Ap- palachian Mountain region southward to northern Georgia, with a westward extension from New York to Michigan. However, americana proved to have been well named. It has the most ex- tensive geographic range of any species of Dolichopeza in the world, covering over two and a half million square miles of North America, from the range described above westward into the Black Hills of South Dakota and northwestward to Alberta and the central valley of Alaska. There are, however, some wide gaps in this pat- tern of distribution, due at least in part to insufficient exploration. Just how the Black Hills habitat connects with the rest, if indeed it does, is not clear; this problem is discussed in the conclusions. In the Ozark-Ouachita Mountain region, there are what I regard as favorable habitats for this species, here and there, and perhaps in time the species will be recorded from that area. Habitats—Adults of americana are most often associated witi rocky habitats, such as the limestone or sandstone gorges of In- diana and Ohio, the “rock houses” of the Appalachians and their foothills (Rogers, 1930: 5), and undercut or broken rock along streams in most elevated parts of eastern North America. In the THE CRANE FLy GENuS DOLICHOPEZA 791 Map 1. Range of Dolichopeza (Dolichopeza) americana Needham. Each spot represents one or more collections within a county (United States) or at a locality. northern part of its range, americana seems to find the tempera- ture, dampness and deep shade of its southern haunts matched in the cool forests and wet woodlands, although I have never found it to be so numerous in such habitats. Perhaps the reverse is true: that the rocky Appalachian ravines provide the “northern” micro- climate in which americana can prosper. Individual flies are often found concentrated in small niches where suitable resting condi- tions prevail. For example, all individuals of americana taken on Harney Peak in the Black Hills were found beneath one outcrop- ping rock, and an all-day search of the nearby forest revealed no additional flies. Again, in a southern [Illinois locality, a few flies of this species were discovered in a darkened cranny, while a care- ful inspection of a wide surrounding area proved fruitless. Seasonal distribution—In those regions where I have watched and collected americana, I have found it the first species of the genus on the wing, both in the spring and late summer generations. In most of Ohio, Indiana and southern Michigan, it may be ex- pected about two or three weeks after the average date of last 792 THE UNIVERSITY SCIENCE BULLETIN killing frost, in the spring. This would place the peak of spring emergence in the areas mentioned at about the first to third weeks of May, depending upon local average climate, and, of course, depending upon the weather of the year. In Washtenaw County, Michigan, a pupa was collected on 13 May. Prolonged observa- tions in a limited area in Turkey Run State Park, Parke County, Indiana, in 1953, showed americana present about a week before the first appearance of any species of Oropeza. On 18 May, ten single males and two mating pairs were observed; and on succeed- ing days the numbers of mating pairs increased steadily, reaching a peak of twenty-one pairs on 24 May and thereafter declining steadily. Between 18 and 29 May, only one unmated female was seen, but after 29 May, as the numbers of males diminished, the numbers of single females increased. By 9 June, americana had all but disappeared from the particular cranny where these obser- vations were made and from the nearby woods. The spring popu- lation generally begins to wane, in this region, in late May to mid- June, or about ten days to two weeks after strong emergence begins, although stragglers may be netted almost any time through- out the summer, and mating pairs may be encountered, now and then, late into June. From the first generation to the second is a matter of only about ten weeks, and a second peak of emergence occurs in the early days of August, continuing for one to two weeks. For data on other parts of the species’ range, see the notes on dis- tribution, below. Immature stages—Eggs of Dolichopeza (D.) americana meas- ure about .60 mm.x.25 mm., and a single female lays about 110 eggs. The duration of the egg stage is seven days, after which the larvae may reach the fourth instar within twenty-three days. The larva of americana is the most colorful of all North Ameri- can species of the genus. Its green ground color, brownish black markings and bright green spiracular disc are distinctive character- istics (Fig. 97). Lacking the dorsolateral lobes of the eighth ab- dominal segment, the larva of americana more closely resembles certain species of Tipula than it does its nearer relatives in Oropeza. It is very similar to the larva of Dolichopeza (D.) albipes of Eu- rope, the only other described larva in this subgenus (see Beling, 1886: 189-191); and it may be that the closely appressed median lobes of the spiracular disc and the peculiar markings of the dorsum are characteristic of the subgenus. Numerous larval habitats of this species have been discovered. THE CRANE FLY GENUS DOLICHOPEZA 793 These vary from rather dry Tetraphis pellucida growing together with a powdery lichen on the undercut surface of a sandstone cliff to decidedly damp Hypnum imponens growing on rich forest soil beside a small stream. The ability of americana to survive in this broad range of habitats probably accounts for its wide distribution and leads me to expect it in many areas where it has so far not been found, such as in the Ozark Mountains. Known larval habi- tats of this species include, in addition to those mentioned above, Atrichum macmillani, Leucobryum glaucum, Dicranella_ hetero- malla, Bryhnia graminicolor, Mnium punctatum, and the liverwort Plagiochila asplenioides. All these bryophytes were growing as thin coverings over sandstone cliffs, except the Bryhnia, which was collected from a deeply shaded clay-till bank. Feeding habits of the larva of americana as compared with those of Oropeza larvae are discussed under larval natural history. In the pupal stage, the color pattern is altered but still very different from any of species of Oropeza. The green coloration of the abdominal segments and to some extent the zig-zag dorsal mark- ings are carried over from the larva. The leg sheaths, wing pads, thorax and head pass through shades of tan and brown, eventually becoming nearly black, shortly before emergence of the adult. The cast pupal skin often shows an irregular, marbled pattern of dark brown in the region of the thoracic dorsum; this pattern does not show clearly in the older pupa, although it can be made out when the thorax is paler in color. Morphological characteristics of the pupa of americana which distinguish it from pupae of species of Oropeza are covered in the section on pupal morphology. Notes on distribution —Ataska—58 miles southeast of Fairbanks (Alaska Highway Milepost 1465), 10 July. AvLBerra—Banff Na- tional Park, Banff, 4 July; 39 miles south-southeast of Valley View, 10 July. Connecricur—Hartford County, 8 to 12 June; Litchfield County, 9 to 12 June; Tolland County, 14 June. Grorcra—Bibb County, 9 June; DeKalb County, 19 August; Fulton County, 25 August to 3 September; Lumpkin County, 23 May; Union County, 10 to 28 June. I_trmors—Pope County, 14-15 July. Inpiana— Jefferson County, 22 July to 27 August; Jennings County, 3 August; Parke County, 10 to 25 June, 15 July, and 11 to 30 August. Iowa— Boone County, 26 May. Krenrucky—Letcher County, 3 July. Laprapor—Rigolet, 16 July. Mare—Cumberland County, 1 July; Hancock County, 16 June; Oxford County, 11 July; Penobscot County, 14 June. MaryLanp—Garrett County, 26 June. Massa- 27—5840 794 THE UNIvERSITY SCIENCE BULLETIN cHusETts—Berkshire County, 13 to 20 June and 19 July; Middlesex County, 11 July; Norfolk County, 11 June; Suffolk County, 11 July. MicuicAN—Cheboygan County, 2 July; Eaton County, 31 May; Keweenaw County, 25 June; Livingston County, 30 May; Mar- quette County, 16 July; Oscoda County, 17 June; Presque Isle County, 29 June; Washtenaw County, 12 June. New Hampsurre— Carroll County, 4 July; Cheshire County, 16, 19 June; Coos County, July; Grafton County, 29 June and 8 July; Hillsboro County, no date. Nrw JersrEy—Bergen County, 5 July; Essex County, June and July. New Yorx—Albany County, 18 June; Cattaraugus County, 30 June; Essex County, 19 June; Fulton County, 18 June; Hamilton County, 17 to 23 June; Herkimer County, August; Queens County, 22 June. NorrH Carotrya—Buncombe County, 30 May; Burke County, 14, 21 June and 1 July; Haywood County, 30 June and 29-30 July; Macon County, 11 June and 23 August to 2 Sep- tember; Swain County, 11 to 19 June and 9 August; Transylvania County, 10 June; Yancey County, 9 to 15 June. Nova Scotia— Guysborough County, 30 June; Halifax County, 29 June; Inverness County, 3 July; Yarmouth County, 27 June. Onro—Delaware County, May; Harrison County, 5 August; Hocking County, 20 May to 8 June; Portage County, 25 June and 16 August; Washington County, 19 June. Onrarro—Kenora District, Sioux Lookout, 6 July; Orillia, 10 June. PrNNsyLvANria—Centre County, 25 June and 5 July; Columbia County, 15 July; Luzerne County, 4, 11 to 19, and 27 June and 5 to 10 July. Qurepec—Bradore Bay (north shore St. Lawrence River), 13 July; Gaspé Peninsula (Mt. Lyell and north shore), 3 July; Knowlton, 26 June. SourH Carotina— Greenville County, 10 June and 16 July to 30 August; Pickens County, 29 June. Soura Daxora—Custer County (Harney Peak), 12-13 and 24 July; Pennington County (Horsethief Lake, in Black Hills ), 16 and 24 July. TENNEssrE—Blount County, 12 June; Cum- berland County, 25 June; Fentress County, 30 May to 24 June, and July and August without dates; Sevier County, 7 May to 13 June, 30 June and 16 September; Scott County, 30 May and 24 July. VeRMOoNT—Chittenden County, 15 to 24 June; Washington County, July. Vimcmia—Giles County, 5 June to 18 July and 11 to 28 August; Nelson County, 14 August; Rockingham County, 6 July; Smyth County, 20 June; Washington County, 18 August; Wise County, 2 July and 16 August. Wesr Vircinra—Marion County, 5 August; Pocahontas County, 23 June and 5 July; Preston County, 25 June; Randolph County, 5 July; Tucker County, 24, 25 June; Wetzel County, 5 August. THE CRANE Fiy GENUS DOLICHOPEZA 795 Dolichopeza (Dolichopeza) borealis, new species The following description is based upon a single female, pinned. The specific name refers to the far northern range of the species. Description.—General coloration dark grayish brown, similar to that of americana. Prescutum uniformly dark, without indication of stripes. Scutellum grayish buff as in americana but postscutellum dark, not pale as in americana. Anepisternum and pre-episternum of mesothorax evenly darkened throughout; that is, without the central pale area found in americana. Eyes black. Antennae thirteen-segmented, the terminal segment elongate, about half as long as the penultimate segment; scape and pedicel pale, the flagel- lum dusky. Wings strongly tinged with grayish brown, the stigmal spot and heavier veins very dark brown. Femora and tibiae dark grayish brown. Approximately the apical one-sixth of each basi- tarsus is pale, the remainder similar in color to the tibia; the second tarsal segment is pale, white in the hind tarsi but only pale tan to yellowish white (as in venosa or australis) in the fore and middle tarsi; the third segment darkening and the fourth and fifth very dark, like the legs; claws simple. Abdomen dark grayish brown, lacking distinct annulations. The tenth tergum is strongly sclero- tized, its surface glossy. Cerci narrowed in apical one-third and darkened at their tips (Fig. 136); the hypovalves taper rather evenly to a slender tip, instead of being truncate with a projecting apical tooth as in americana. Body length of holotype female 15.0 mm.; wing 12.9 mm. Type.—Holotype female; about 30 miles north of Seward, Kenai Peninsula, Alaska (60°31'N, 149°32’W), 30 June 1957, George W. Byers (Field Catalogue Number 22); in the Snow Entomological Museum, University of Kansas, Lawrence, Kansas. Descriptive comments.—It is anticipated that discovery of the male of this species will provide additional characters to separate it from Dolichopeza americana. At present, the larger size of borealis, its peculiar tarsal coloration (which is presumed to be the same in both sexes), and details of structure of the ovipositor will serve to distinguish the two species. The tarsi in borealis are proportionately shorter than in most species of the genus. Thus, while the tarsi of americana exceed the combined lengths of femur and tibia, those of borealis (assum- ing the type is representative) are noticeably shorter, as shown in Table 4. 796 THE UNIveRSITY SCIENCE BULLETIN TABLE 4.—Measurements of Leg Segments in Dolichopeza (D.) borealis (Measurements in Millimeters ) Femur Tibia Tarsus ( Basitarsus ) Total Horeslegy 2:fi> gc 6.1 6.0 9.9 (7.2) 22.0 Middle leg ........ 6.6 6.8 9.7 (6.7) 23.1 Findlee i. Cell 7.0 10.0 (Hell), 24.1 Geographical distribution —Dolichopeza (D.) borealis is so far known only from the type locality on the Kenai Peninsula, and it seems likely that its range does not extend far northward or east- ward from there. It may, however, occur in the forested lowlands of southwestern Alaska and, if the shade provided by trees is not a limiting factor in its distribution, in the Aleutian Islands. Habitats—The holotype was taken in fairly deep shade at the edge of a clump of spruce trees in a hummocky bog, in which species of Sphagnum were the predominant mosses. There was much fallen timber in the bog and a widespread undergrowth of brushy, moose-browsed willow. In the open areas among the willows, grasses, sedges and wild rose were growing in the mosses, and there were scattered, small areas of open water. Seasonal distribution—The collection date in late June makes it difficult to say whether there are one or two annual generations, but on the basis of knowledge of the other species of the genus it seems reasonable to expect only one summer generation at the latitude of the type locality. Immature stages—None of the immature stages of this species is so far known. Subgenus Oropeza Dolichopeza (Oropeza) australis, new species Among specimens of Dolichopeza subalbipes from various lo- calities on the southeastern coastal plain, I have found several flies resembling that species in some respects but differing in details of coloration and structure. The following description is based upon 22 males and eight females from Georgia (type series) and two males and two females from other areas. The specific name refers to the fact that this species has the southernmost over-all range of any North American Dolichopeza. Description —General body coloration light brown with dark brown markings, the tarsi pale. Prescutum dark brown, a pattern of three dark stripes sometimes faintly indicated but usually ob- scured altogether. (There is, in some specimens, a slightly paler THe CRANE FiLy GENUS DOLICHOPEZA 797 (e) 0.5 1.Omm. fo} 1LOomm. (Liwtiie eisai ee es ee eee (eee es) SCALE, FIGS. 137-141 SCALE, FIG.142 Fics. 137-142. Dolichopeza (Oropeza) australis new species; 137— ninth tergum of male, 138—left inner dististyle of male, dorsal aspect, 139—left outer dististyle, dorsal aspect, 140—gonapophysis, dorsal as- pect, 141—vesica and penis, 142—terminal abdominal segments of fe- male, right lateral aspect. median line, but I believe this is a post-mortem effect.) Meso- thoracic anepisternum brownish, only a little darker than adjacent portion of pre-episternum. Eyes black. Antennae with twelve seg- ments, the scape and pedicel slightly paler than the dusky flagellum. Wings tinged lightly with brown, the stigma darker brown. Fem- ora and tibiae brown, this color extending over approximately the basal two-thirds of each basitarsus and grading through yellowish brown to pale yellowish white on the second through fifth seg- ments. Abdomen light brown with distinct darker brown annula- tions. The medio-posterior margin of the ninth tergum of the male has a central, pointed tooth with a broader, rounded lobe on either side, the central tooth the longest; and the tergal arms are slender, their tips flattened and slightly widened (Fig. 187). Gonapophyses 798 THE UNIVERSITY SCIENCE BULLETIN set with black bristles, as in other species of the obscura group. Inner dististyles with an angular protuberance on the outer sur- face, the elevated area bearing a few strong hairs (Fig. 188). Hypovalves of female ovipositor moderately darkened subapically (Fig. 142). Body length of males is 8.0 mm. to 9.8 mm. (holotype 9.5 mm.); of females 9.8 mm. to 11.6 mm. (allotype 9.8 mm.). Wing length in males varies from 10.4 mm. to 11.8 mm. (holotype 10.9 mm.); in females, from 10.2 mm. to 12.8 mm. (allotype 10.6 mm. ). Types.—Holotype male, Mossy Pond (Emory University Field Station, near Newton), Baker County, Georgia, 27 January 1949, R. E. Bellamy (Catalogue Number 2936). Allotype, same locality and collector, 31 March 1949 (Catalogue Number 3007 ). Paratypes, 21 males and 7 females from the type locality, as follows: 27 Au- gust 1948, 1 g ; 27 January 1949, 1 9° , taken in copulation with the holotype but too damaged to be made the allotype; 9 February 1949, 1 g; 17 February 1949, 1g, 19; 24 February 1949, 1g; 24 March 1949, 7 ¢, 49; 381 March 1949, 6 g, 19; 7 April 1949, 1g,19; 21 April 1949, 3 g; all collected by Dr. R. E. Bellamy. Holotype, allotype and most of the paratypes are in the University of Michigan Museum of Zoology, Ann Arbor, Michigan; one male paratype has been sent to the United States National Museum, Washington. The few specimens seen from areas other than the type locality have not been designated paratypes. Descriptive comments.—Dolichopeza australis is most nearly related to subalbipes and johnsonella, as indicated by the hypo- pygial structure of the male, especially the inner dististyles and the posterior margin of the ninth tergum. In body coloration, australis resembles johnsonella, the thoracic pleura lacking the contrast in shade and color along the border between anepisternum and pre-epi- sternum, as found in subalbipes, and the prescutum being obscurely marked. However, the coloration of the legs and details of the external reproductive structures of both sexes will easily separate the two species. In subalbipes, the tarsi are white and the femora and tibiae are pale with darkened tips, and a part of the last tarsal segment is darker than the remainder of the tarsus. In contrast. the leg segments of australis are neither pale nor dark tipped, and the tarsi are pale throughout. The characteristic basal enlargement of the penis in subalbipes is only faintly indicated in australis (Fig. 141), and details of the external genitalia as well will serve to dis- tinguish these two species. Certain male flies assigned to subal- THE CRANE FLY GENuS DOLICHOPEZA 799 bipes have slender tergal arms similar to those of australis, but the two species may be differentiated by the shape of the margin of the ninth tergum and by the color differences already mentioned. Of the type series of flies, about one quarter have venational ab- normalities in one or both wings. Most of these anomalies consist of absence of a closed cell 1st M, (discal cell) by loss of the medial cross-vein. Geographical distribution —Dolichopeza australis is so far known from only three localities, one of which is the type locality in southwestern Georgia. The other two, Washington County in southwestern Alabama and Alachua County in northern peninsular Florida, suggest a rather widespread occurrence of the species on the Gulf Coastal Plain. Habitats similar to those in which australis has been found also occur in southern Mississippi and Louisiana. Habitats —All known habitats of this species are swamps—that is, wooded areas with lush undergrowth and considerable amounts of standing water. Mossy Pond, the type locality, is an isolated cypress swamp of a few acres extent, surrounded by rather dry land that is sparsely wooded with pines. The cypresses are fes- tooned with Spanish moss, while the water below supports a lux- uriant growth of floating and emergent vegetation. Mosses grow- ing above the water line on the cypress boles and at the edge of the pond are probably the larval habitat of australis and of the other species of Dolichopeza (obscura and subalbipes) found to- gether with it. Seasonal distribution —It has been mentioned earlier that in the southernmost part of the range of the genus in North America, Dolichopeza seems to reproduce and complete its life cycle as rapidly as local conditions will permit; that is, there is no limitation to one or two annual generations. Date records for australis include January, February, March, April and August at the type locality, with the largest numbers having been taken in the last half of March. Records from Alachua County, Florida, are for 12 and 14 November and 16 January, and from Washington County, Ala- bama, 14 June. Immature stages —Eggs of this species average .71 mm. in length and .31 mm. in greatest diameter. They do not have a terminal filament. Larvae and pupae of australis have not yet been found, as I have never had the opportunity to study this species in the field. During my brief visit to the type locality, in April, only Dolichopeza obscura was collected, and I have never seen even the adults of australis alive. 800 THe UNIversiry SCIENCE BULLETIN Dolichopeza (Oropeza) carolus Alexander Literature references—Oropeza albipes Johnson. Johnson, 1909: 121, pl. 15 (hypopygium); Johnson, 1910: 708; Alexander, 1919: 930; Alexander and McAtee, 1920: 393; Dietz, 1921: 259; Alexander, 1924: 60; Alexander, 1925: 172; Johnson, 1925: 32; Pierre, 1926: 11; Leonard, 1928: 698; Alexander, 1929a: 236; Alexander, 1929c: 297; Alexander, 1930c: 113; Rogers, 1930: 22; Rogers, 1933: 48. Dolichopeza (Oropeza) carolus Alexander, new name for Oropeza albipes Johnson, preoccupied in Dolichopeza by D. albipes (Strém). Alexander, 1940: 618; Alexander, 194la: 295; Alexander, 1942: 211- 212, fig. 26A (hypopygium); Whittaker, 1952: 32; Foote, 1956: 221. Original description—*Face yellow, rostrum and vertex brown, palpi brown. Antennae yellow, the three basal joints lighter than the others. Thorax dark brown, showing three rather distinct stripes; plurae, scutellum, and metanotum light yellow, translucent, with livid spots below the base of the wings, between the coxae, at the base of the halteres, and on the posterior margin of the metanotum. Abdomen yellow, the black bands connected on the dorsal line leaving a row of large spots on the sides as in O. venosa. In the female the bands are more distinct. Genitalia yellow, ap- pendages light yellow, very short or rudimentary, style red, base black, appendages at base of style short and tipped with a slightly curved spine, ventral margin deeply emarginate. Ovipositor reddish brown. Femora yellow, tibiae and tarsi white, tipped with brown. Halteres white, knobs brown. Wings brownish hyaline, veins light brown, stigma dark brown, median cubital cross-vein * present or wanting. Length, male, 10 mm.; female, 13 mm.” Types—Holotype male, Cohasset (Norfolk County), Massa- chusetts, 1 July 1907, Owen Bryant. Allotype, Dummerston (Wind- ham County), Vermont, 14 July 1908, C. W. Johnson. Paratypes, six males and four females, together with holotype and allotype in the collection of the ‘Museum of Comparative Zoology, Harvard University; one male and one female in the Academy of Natural Sciences of Philadelphia; and one male in the Alexander collection, at Amherst, Massachusetts. Johnson mentions having twenty-two specimens; however, it is possible that he did not make types of all of them. Diagnostic characteristics —Among the species of Oropeza in North America, only two possess white tarsi and pale legs, with the tips of femora and tibiae darkened. These are carolus and * See page 710. ai THE CRANE Fiy GENUS DOLICHOPEZA 80L SCALE, FIG. 148 Oo 1.0mm. (Va. Cee es a et | SCALE, FIGS. 143-147 0 0.5 fone! 148 a SS EEE EEE Fics. 143-148. Dolichopeza (Oropeza) carolus; 143—ninth tergum of male, 144—left inner dististyle of male, dorsal aspect, 145—left outer dististyle, dorsal aspect, 146—gonapophyses, dorsal aspect, 147—-vesica and penis, 148—terminal abdominal segments of female, left lateral aspect. 802 THE UNIVERSITY SCIENCE BULLETIN subalbipes. The similarity of coloration and size in these two species is striking, involving nearly every detail. In the field, habitat is an indication of species identity, carolus being more often found in rocky places and moist to mesic forests and sub- albipes in wetter woodlands, along stream banks, and in swamps and marshes. Males of carolus may be identified with certainty by their short, pale outer dististyles. Females may be separated from those of subalbipes by the presence of a dark, well-defined spot on the basal portion of the hypovalve, on each side, approxi- mately below the ninth tergum (Fig. 148). The absence of sharp contrast of color between the mesothoracic anepisternum and pre- episternum in carolus is also an aid in distinguishing this species from subalbipes. All these features are visible to the unaided eye, so that field determination of collected specimens is not difficult. Descriptive comments.—The similarity of coloration between carolus and subalbipes extends to virtually every feature except those few enumerated above, but the darker markings are more intense in subalbipes, in specimens that are not faded with age. While carolus is a fly of contrasting colors, these are, as in all of the genus, usually more subtle than the paint-box assortment in- dicated in Johnson’s description. There is a degree of brown in all the colors, so that in general appearance carolus is light tawny with dark brown markings. Against a dark background, the tarsi are pale enough to be called white, without any chance of confusion. Old or sun-faded specimens take on the appearance of Dolichopeza venosa, which species is easily recognized in both sexes by the genitalia. Wing venation in carolus deviates only rarely from the subgeneric scheme. An occasional absence of the medial cross-vein or failure of some branch of the media to reach the wing margin are the most commonly seen anomalies. In a collection of 40 males and 20 females from a Georgia locality, only four venational abnormalities were found: one incompletely closed discal cell and three instances of a short spur vein projecting distally from the medial cross- vein. The latter anomaly has also been seen from a few other localities. Body length of males is from slightly less than 10 mm. to some- what over 12 mm.; and the wings vary within the same limits. Females measure about 11 to 15 mm., their wings from 10 to 14 mm. The smaller specimens are from August collections, and the larger ones are spring records from Indiana, Michigan and moun- tainous western Virginia. THE CRANE FLY GENUS DOLICHOPEZA §03 Variation in details of the male hypopygium is slight and is found mostly in the shapes of the ends of the tergal arms and of the inner dististyles. The widened, flattened tips of the tergal arms may be elongate, rounded or somewhat pointed, and it is often the case that the two arms of the same fly are unlike (Fig. 143). Curvature of the inner dististyles varies, as does the depth of the apical cleft. It should be mentioned that the apparent shape of such flattened or partially flattened structures as the tergal arms and (in this species) the outer dististyles is considerably affected by a shift in the point of view. Thus, in dorsal aspect (Fig. 145) the outer disti- style seems somewhat narrowed in the basal half, while viewed from the side it is of almost equal width throughout. In the emar- gination of the ninth tergum and the shape of the tergal arms, carolus most nearly resembles walleyi and sayi, but the gona- pophyses are rather like those of dorsalis. The size and configuration of the vesica and its apodemes and the diameter and apical curvature of the penis suggest a relation- ship of this species with Dolichopeza venosa (compare Figs. 147 and 234). The seminal duct in carolus is only about one-third as long as that of polita, a species of comparable size, and the bases of the accessory glands are enlarged, resembling the common vas deferens at the point where all three ducts enter the seminal vesicle. Geographical distribution—The known range of Dolichopeza carolus extends from Maine westward to Wisconsin and southward through the Appalachian Mountains. Following cool ravines along, Map 2. Range of Dolichopeza (Oropeza) carolus Alexander. Each spot represents one or more collections within a county (United States) or at a locality. §04 THe UNIVERSITY SCIENCE BULLETIN the Chattahoochee-Apalachicola drainage, carolus, apparently an essentially northern species, has reached western Florida, where it is a common member of the tipulid fauna of the Apalachicola River valley, an area unlike the rest of the state. I did not find in the Ozark Mountains what I consider adequate carolus habitats, but I would expect to find this species in parts of Minnesota, western Ontario, and other northern areas from which it is not known at this time. There is a published record of the occurrence of this species in New Brunswick (Alexander, 1942: 211), but I have not been able to get more specific data, for which reason the record does not appear on the map. Habitats—Adults of Dolichopeza carolus are most commonly taken in leafy vegetation, such as is found along small streams through deciduous forests or in rocky ravines where leafy plants and ferns abound. I have never found them far from water or cool shade, although by night or on damp days they may fly a hundred yards or so from their usual streamside haunts. They are numerous in Michigan woodlands in low vegetation in the wetter areas. I found them particularly abundant in the sandstone ravines of Indiana and Ohio, where they rested among ferns and other low plants along the edges of outcropping strata or suspended them- selves in the deeper shade of a cranny beneath an outcrop. Follow- ing small brooks upstream, I encountered carolus nearly as far as there were trees providing shade. Dietz (1921) records carolus as “. . . common, June and July, in damp, open woods.” On the (emubenand Plateau, Rogers (1930) found them “* equally characteristic of wet ‘rock-houses, of the shaded base a dripping cliffs, of upland stream-margin thickets, and of the mossy banks of shaded brooks.” In the Florida area mentioned earlier, carolus was found in “. . . moist recesses beneath overhanging banks and within rank herbage on shaded stream banks” (Rogers, 1933). Foote (1956) records the species from moist woodland ravines and wet cliff faces. I have taken carolus together with subalbipes in deep grasses and broad-leafed weeds along a stream in Virginia, among ferns along a mountain brook in Maine and in a swamp in Wisconsin. In the brushy margin of a mountain bog in Pennsylvania, I took it to- gether with venosa, while along a rocky brook course a mile down the mountain it was found together with americana. It is my be- lief that the species does not range far from its larval habitats. which are wet or damp mosses usually occurring near small streams. On one occasion, however, several teneral flies were captured on THE CRANE FLY GENUS DOLICHOPEZA 805 a mossy but not wet cliff face at 6:30 in the evening, but when I returned to the same spot an hour after nightfall, these and many other adults seen earlier were nowhere to be found. Seasonal distribution—From southern Michigan and New Eng- land to Georgia, carolus seems to have two main periods of emer- gence annually: one in late May and June, the other in late July through August, the period in any case governed in part by the local weather conditions. Nearly all western Florida records are for March and April, although later summer months are repre- sented. In Michigan and in Wisconsin, carolus is on the wing throughout July but mostly early in the month. In Turkey Run State Park, Parke County, Indiana, in 1953, the first individuals of carolus were observed on 1 June, two weeks after the start of the emergence of americana and nine days after polita first appeared. Only a few females were found as late as 10 July, although one mating pair was collected on 12 July. In Neel Gap, Union County, Georgia, in 1952, carolus had reached peak emergence on 28 June. Six years later, at the same place, equivalent numbers were found on 10 June. Immature stages—Eggs of this species are comparatively wide for their length, measuring around .79 mm. long and .37 mm. wide. Not many more than a hundred eggs appear to be matured and laid by one female. Eggs laid on 15 June, in Indiana, hatched nine days later at outdoor temperatures. In connection with oviposition, there is an interesting observation to relate: while all other species of Dolichopeza induced to lay eggs in captivity laid them on or near wet paper pads in the bottoms of the breeding jars, all females of carolus (several individuals, in two years) deposited their eggs upon the sides of the glass jars, as if avoiding the paper. Now, if the tip of the ovipositor is sensitive to the appropriate larval en- vironment (see account of oviposition), then this peculiar behavior may have some bearing on the kind of habitat in which the eggs are laid under natural conditions. The first instar larvae cannot be distinguished from those of other species, at this time, and I have not succeeded in rearing them to the later instars or in locating the latter in their natural habitats. I would expect the larvae of carolus to resemble rather closely those of subalbipes and venosa. Not until the summer of 1958 was I able to find pupae of this species, and I hasten to add that these have been identified only by inference, on the basis of comparison with known pupae and by close association with adult flies. At a Penn- 806 THe UNIVERSITY SCIENCE BULLETIN sylvania locality, in a season when adults of americana and carolus only were present, both male and female pupal skins of a species of Oropeza were found projecting from mats of the mosses Tetra- phis pellucida and Leucobryum glaucum. The mosses were grow- ing on root-bound soil on the bank of a small brook below a spring. Beneath wood ferns and low broad-leafed plants only a few inches from the pupal site, adults of carolus were resting. The pupae most nearly resembled those of subalbipes but appeared to belong to a species of the sayi group—a combination one would almost expect on the basis of the adults. Their size, the configuration of their spiracular yokes, and other characters (see key to pupae, page 773) make it virtually certain that these are the pupae of carolus. Notes on distribution—Connecticut—Fairfield County, 23 Au- gust; Hartford County, 8 June and 9 July; Litchfield County, 9 July and 23 July to 12 September; Windham County, 30 June. FLorma —Gadsden County, 19 April; Leon County, 16 March to 25 May; Liberty County, 27 March to 26 April and 15 to 27 July. Grorcra— Dawson County, 9 to 27 June; Hall County, 6 June; Lumpkin County, 8 June; Rabun County, 20 May to 12 June and 20 August; Union County, 10 and 28 June. INpriANna—Jefferson County, 3 June and 23 August; Monroe County, 23 June; Montgomery County, 28 June; Owen County, 18 to 27 June; Parke County, 9 June to 15 July and 30 August. Kenrucky—Bell County, 18-19 June; Ed- monson County, June (?); Letcher County, 3 July; Pike County, 3 July; Whitley County, 24 June. Marve—Hancock County, July; York County, 1 July. Maryitanp—Baltimore County, 17 June; Dis- trict of Columbia, 30 May and 25 August; Montgomery County, 4 to 15 June, July, and 29 August to 5 September; Prince George’s County, 29 June. Massacnuserrs—Berkshire County, 12 to 30 June; Hampshire County, 14 July and 9 August; Middlesex County, 7 June; Norfolk County, 1 and 10 July. Micuican—Berrien County, 8 to 17 July; Cheboygan County, July; Saginaw County, 6 June. New Hampsuire—Grafton County, 4 July; Rockingham County, no date. New Jersey—Bergen County, 21 June; Burlington County, 3 to 11 August; Morris County, 17 June; Warren County, 11 July. New Yorx—Cattaraugus County, 4 July; Columbia County, mid- August; Cortland County, 20 July; Erie County, 3 to 10 July; Fulton County, 28 June; Herkimer County, 20 June; Oneida County, July; Orange County, 25 June to 28 July and 27-28 August; Suffolk County, 25 June and 9 August; Tompkins County, 14 July; Ulster County, 28 June; Westchester County, 9 June. Norra Carotina— THe CRANE Fiy GENUS DOLICHOPEZA 807 Avery County, 14 June; Buncombe County, 15 June; Burke County, 14 June and 9 July; Haywood County, 27 July to 2 August; Macon County, 8 to 22 June; McDowell County, 13 June; North- hampton County, 7 June; Transylvania County, 9-10 June; Yancey County, 10-11 and 22 June. Onto—Cuyahoga County, 24 June; Delaware County, June; Fairfield County, 8 June; Hocking County, 30 May and 6 to 8 June; Portage County, 24-25 June and 14 July. Onrarto—Jordan (near Niagara Falls), 1 August; Toronto, 4 July. PENNSYLVANIA—Allegheny County, 14 June; Carbon County, 3 July; Centre County, 13 and 25 June, 2 to 9 July and 5 August; Luzerne County, 13 to 27 June and 5 to 24 July; Mercer County, 25 June; Monroe County, 6 August; Philadelphia County, 31 July, 2 August and 2 September. Quresec—Montreal, June. Ruopr IsLanD—Kent County, 21 June. SourH CaroLtina—Greenville County, 7 June and 1 September; Pickens County, 29 June. TEN- NESSEE—Bledsoe County, 26 June; Cumberland County, 25 June; Fentress County, 6 June to 16 July and 16 August; Morgan County, 12 June and 5 August; Scott County, 28 to 30 May; Sevier County 10 to 12 and 30 June. VeRMont—Orange County, 11 July; Wind- ham County, 14-15 July. Vircry1a—Arlington County, 6 June to 4 July and 11 to 25 August; Augusta County, 28 June; Bedford County (locality uncertain), 30 June; Fairfax County, 29 May to 9 June, 11 to 25 July, and 22 August to 5 September; Giles County, 3 June to 13 July and 24 July to 29 August; Rockingham County, 6 July; Shenandoah National Park, 6 July; Smyth County, 20 June; Washington County, 2 July; Wise County, 2 July. Wersr Vircin1a— Mingo County, 3 July; Pendleton County, 26 June; Pocahontas County, 5 July and 7 August; Randolph County, 5 July. Wisconsin —Juneau County, 6 July; Sauk County, 5 July. Dolichopeza (Oropeza) dorsalis (Johnson) Literature references.——Oropeza_ dorsalis Johnson. Johnson, 1909: 119-120, pl. 15 (hypopygium); Alexander, 1919: 930; Alex- ander and McAtee, 1920: 393 (possibly misidentified); Johnson, 1925: 33; Pierre, 1926: 11; Alexander, 1928: 57, Alexander, 193la: 138. Oropeza rogersi Alexander. Alexander, 1922a: 6-7; Pierre, 1926: 12) Oropeza dorsalis rogersi Alexander. Rogers, 1933: 49. Dolichopeza (Oropeza) dorsalis (Johnson). Alexander, 1941a: 295; Alexander, 1942: 212, fig. 26B (hypopygium ); Rogers, 1942: 59. Dolichopeza (Oropeza) sessilis Alexander. Alexander, 1941a: 296-297, fig. 1 (wing) (new synonymy ). 808 Tue UNIVERSITY SCIENCE BULLETIN Dolichopeza (Oropeza) dorsalis rogersi (Alexander). Alexander, 1942: 212 (new synonymy ). Original description—Face yellow, rostrum and vertex dark brown, palpi blackish; antennae with three basal joints yellow, the remainder fuscous. Thorax dark brown, the three stripes indistinct, the brown extending over the scutellum and metanotum as a broad stripe, leaving only a narrow lateral margin of yellow; pleurae yel- lowish white, subtranslucent. Abdomen yellow with a dorsal stripe of brown, spreading over the third to the seventh segments and somewhat obscuring the black bands. Genitalia yellowish, append- ages black, style short, ventral margin deeply emarginate. Oviposi- tor brown. Halteres yellow, knobs dark brown. Legs brown, coxae and basal portion of the femora yellow. Wings light smoky brown, veins and stigma slightly darker, the petiole between the discal and second posterior cell very short, the median cubital cross-vein* present. Length, male, 9 mm.; female, 10.5 mm.” Types.—Holotype male, Capens (on Deer Island, Moosehead Lake, Piscataquis County), Maine, 14 July 1907, C. W. Johnson. Allotype, same locality and collector, 15 July 1907. Both are in the collection of the Museum of Comparative Zoology at Harvard University. Diagnostic characteristics—The contrasting dark dorsum and pale sides and underparts of the thorax of dorsalis make it readily identifiable on sight, both in the field and laboratory, sight recogni- tion being easier, however, when the fly is alive, the contrast then seeming greater. In its natural habitat, this species is most likely to be confused with Dolichopeza walleyi or sayi, which often occur together with it. From sayi, it differs in details of the external re- productive structures in both sexes and in coloration, especially of the thoracic pleura, pale in dorsalis and darkly colored on the anep- isternum and ventral pre-episternum in sayi. Preserved specimens closely resemble the form of walleyi that has pale thoracic pleura, particularly if the prescutal stripes are at all distinct. Males of these two species are easily differentiated by hypopygial characters. The faintness of abdominal annulations and the shorter, more heavily sclerotized cerci of females of dorsalis will aid in distinguish- ing them from walleyi females having pale pleura, generally yellowish coloration, and the stigmal spot of the wing not deeply colored. Also, the knobs of the halteres are very dark gray-brown in dorsalis but are only slightly infuscated in walleyi. * See page 710. THE CRANE FLy GENus DOLICHOPEZA 809 155 156 SCALE, FIGS. 149-153 Fics. 149-156. Dolichopeza (Oropeza) dorsalis; 149—ninth tergum of male, 150—left inner dististyle of male, dorsal aspect, 151—left outer dististyle, dorsal aspect, 152—gonapophyses, dorsal aspect, 153—vesica, penis and adminiculum, left lateral aspect, 154—terminal abdominal segments of female, right lateral aspect, 155—apex of wing, showing cell M, with short petiole, 156—apex of wing showing sessile cell My. Descriptive comments.—Dolichopeza dorsalis is a small, tawny- yellow fly with a dark, brownish thoracic dorsum and with amber- tinged wings that are relatively wider for their length than in other North American species of the genus. The coloration of the dorsal sclerites of the thorax varies from a rich cinnamon brown to very dark brown. Females, and less often males, may not have the dark coloration extending uninterruptedly along the abdominal terga, as described by Johnson. An outstanding and very common variation in wing venation in this species is the shortening of that part of the vein M,,, which forms the petiole of cell M,, that cell being correspondingly elon- §10 Tue UNIvERSITY SCIENCE BULLETIN gated. This is mentioned by Johnson in the original description and again in the description of Oropeza rogersi (Alexander, 1922a: 6). In every large sample of a population of dorsalis, I have found a few flies in which this petiole was extremely short (as Fig. 155) or absent completely (as Fig. 156). In fact, several individuals have been seen in which the cell M, was petiolate in one wing and sessile in the other! In view of this, I have placed Dolichopeza (Oropeza) sessilis, distinguished from dorsalis by having the cells M, sessile, in synonymy with D. (O.) dorsalis. Dolichopeza dorsalis is the smallest North American representa- tive of the genus, although some males of americana and very rarely of other species will be smaller than the average dorsalis. Males range in length from about 7 to 10 mm., most being close to 9 mm. Their wings vary from 8.5 to 11 mm. and usually slightly exceed the length of the body. Females measure 9.5 to 12 mm. in body length and 9 to 11 mm. in length of wing. Throughout the range of the species, quite small individuals may be found. On the basis of certain of these, Alexander (1922a: 6) described Oropeza rogersi, which was stated to differ from dorsalis only in size. In a later review of the genus, Alexander (1942: 212) regarded this small form as a subspecies of dorsalis, having a range from southern Indiana eastward to Virginia and southward to Florida. Rogers, nine years earlier, had also considered rogersi a subspecies of dor- salis (Rogers, 1933: 49). However, among the specimens available to me there are individuals from Michigan and British Columbia that are of small enough size to fit the description of the southern race, rogersi, while many specimens from within the described range of rogersi, including some from the type locality, are as large as any northern dorsalis. Because of this situation, I consider Do- lichopeza (Oropeza) dorsalis rogersi a synonym of dorsalis. In its general profile, the ninth tergum of the male of dorsalis suggests that of johnsonella, although the lateral shoulders are less prominent and the tergal arms are proportionately shorter and more rounded at the tips (Fig. 149). The inner dististyles (Fig. 150) are more like those of walleyi or sayi, while the gonapophyses most nearly resemble those of carolus (Fig. 152). The edge of the gonapophysis that bears the smooth, spinous tip may be dorso- lateral in position or may be slightly rotated into a dorsal position. The outer dististyles are very dark brown, except at their bases, which are pale like the rest of the hypopygium. The adminiculum is short, conical and strongly sclerotized. THE CrANE FLY GENusS DOLICHOPEZA 811 Except for the venational variation already mentioned, the wings show few departures from the subgeneric pattern. A small closed cell at the distal end of the discal cell connecting it with the cell M, has been seen once. Dolichopeza dorsalis rests with wings outspread and tilted slightly forward. I have noticed that this species sometimes, instead of hanging from overhead support, stands upon vegetation, especially the bent-over leaves of grass. It is a very alert fly, which takes to the wing at the slightest alarm. Geographical distribution—Dolichopeza dorsalis is the second most widespread member of its genus in the world. Apparently without subspeciation, it ranges across three thousand miles of dorsalis Map 3. Range of Dolichopeza (Oropeza) dorsalis (Johnson). Each spot represents one or more collections within a county (United States) or at a locality. North America, from Florida and New England northwestward to the Yukon. Its distribution within this vast area (see map) is very poorly known, and extensive further collecting is necessary, especially in the provinces of Canada and the middle Atlantic states, to bridge existing gaps in the range. Only recently (summer of 1959) this species was discovered in an apparently isolated habi- 812 Tue UNtversiry SCIENCE BULLETIN tat in the Black Hills of southwestern South Dakota. Its occur- rence there is discussed more fully in the conclusions. Habitats—This species is usually found associated with bogs, swamps and brushy marsh borders. In Michigan, dorsalis is “com- mon in shrub and shrub-sedge marshes and near the shrub bordered margins of the sedge-fern-grass marshes. Numerous but local in a variety of other shaded wet situations: tamarack- sumac swamps, birch-maple swamps, and shaded seepage areas” (Rogers, 1942: 59). The holotype of Dolichopeza sessilis was col- lected “. . . in a small accessory swampy patch just off the main stream. . . . at 2200 feet elevation, in the Great Smoky Mountains, North Carolina (Alexander, 1941a: 296). Rogers (1933: 49) reports dorsalis collected in “rank herbage of swampy or marshy rills,” in Jackson and Liberty counties, western Florida. The South Dakota habitat was a marsh of grasses, Carex spp., Osmunda fern, low willow brush, etc., below a spring—a habitat similar in many details to places in which I have collected dorsalis in Michi- gan, Minnesota and elsewhere. Dolichopeza sayi often occurs to- gether with dorsalis in such environments, but in the Black Hills that species was replaced by the dark form of walleyi. Seasonal distribution—There is evidence of two annual genera- ations in the collection records from southern Michigan, southern Indiana and North Carolina. Most of the Florida specimens were taken in April, and all were collected in spring. Further north, the first emergence is in June and the second in August. Rogers (1942: 59) estimated the emergence periods, in Livingston County, Michi- gan, to be from 4 June to 14 July and from 11 to 30 August, in the years 1936 through 1938. The records for farther north in Michigan and for Minnesota, Ontario, and Manitoba, as well as for South Dakota, are all very late June or July. In these regions, it is prob- able that there is but one generation per year. It is not clear from the collection records in Alberta and northern British Columbia whether there are one or two generations per year in those areas. The latitude would lead me to expect but one; however, at Liard Hot Springs, northern British Columbia, on 13 June 1957, while en- route to Alaska, I found only males of dorsalis, suggesting that the flight period was just beginning. Returning on 8 July, I found only females. If the period of emergence of adults had progressed as rapidly as indicated by these collections, it is possible that a second generation might have been produced before the onset of freezing weather. THe CRANE Fiy GENUS DOLICHOPEZA §13 Immature stages.—The eggs of dorsalis average .76 by .31 mm., which is rather large, considering the size of the adult female. | have no counts of the quantity laid by one female but would es- timate the number to be ninety. Larvae began to hatch on 22 July from eggs that had been laid on 12 July, by females taken at Lake Itasca, Clearwater County, Minnesota. The first instar larvae of dorsalis are of a more buffy color than are those of the other species. Larvae of dorsalis are bright greenish, like the background color of americana larvae. Since the green color is not partly concealed by transverse ridges of microscopic hairs on most of the dorsum, as it is in some other species, the larva has by comparison a rather striking appearance. The greenish color is dulled, however, in the thoracic and eighth abdominal segments, by the presence of micro- scopic hairs that are so long and dense that they give a “woolly” appearance at high magnifications. The color of the body of the larva probably results from and changes according to the foods chosen by the larva. To see whether this is the case, I offered leaves of Sphagnum rubellum, which are dull red in color, to a larva that had been feeding on a green species of Sphagnum, probably S. palustre. During the sev- eral weeks that both mosses were offered as food, however, the larva never fed on the red moss. Dolichopeza dorsalis is the only species the larva of which is known to feed on Sphagnum, but it is not limited to that genus, having been taken also in Amblystegium va- rium, and possibly in Plagiothecium denticulatum growing together with the Amblystegium on the floor of a marsh. From these same mosses were reared Tipula sulphurea, Erioptera uliginosa, E. need- hami, Pseudolimnophila noveboracensis, and Dolichopeza sayji. The pupa has the structure characteristic of its subgenus, com- bining with this those details by which it is identified in the key to pupae. The spiracular yoke is illustrated in Figure 115. In the laboratory, the pupal stadium was six days. Thorax and wing sheaths were pale tawny when the pupa was two or three days old but darkened to gray about 36 hours before emergence. About 12 hours prior to emergence of the adult, the entire pupal skin darkened, as observed in other species. Throughout the pupal life, the green coloration persists in the abdominal region. The larva of dorsalis does not seem to construct a well-organized burrow for pupation, but the tube is evident enough to allow the cast larval skin to be recovered without much searching. 814 THE UNIVERSITY SCIENCE BULLETIN Notes on distribution—ALBERTA—Bilby (about 30 miles west of Edmonton), 20 July; Lesser Slave Lake, 21 June. BririsH Co- LuMBIA—Liard Hot Springs (milepost 496.5, Alaska Highway), 18 and 25 June and 8 July. Connecricutr—New Haven County, 3 and 14 July. FLorma—Jackson County, 6 May; Leon County, 31 March to 24 April; Liberty County, 25-26 April. Grorcra—Hall County, 5-6 June. Inprana—Jefferson County, 2 to 16 June and 18 to 23 August. Marne—Hancock County, 11 June to 17 July; Piscataquis County, 14-15 July; Washington County, 25 July. Manrropa— Aweme (20 miles southeast of Brandon), 25 June; Victoria Beach (southeast shore of Lake Winnipeg), 19 July. MaryLanp—Dis- trict of Columbia (in error?), 17 May and 30 August. MicaicAn— Alger County, 29 June; Allegan County, no date; Antrim County, 2 July; Berrien County, 11 July; Cheboygan County, 18 July; Lake County, 8 July; Livingston County, 4 June to 14 July and 11 to 30 August; Roscommon County, 1 July; Washtenaw County, 28 May to 24 June and 23 August. MiInNresora—Clearwater County, 10-11 July; Ramsey County, 6 July. New Hampsuire—Grafton County, 12 July. Norra Carotrysa—Haywood County, 3 August; Macon County, 11 to 15 June; Swain County, 30 June; Transylvania County, 9 June and 8 September; Yancey County, 26 May and 1 August. Nova Scotra—Halifax County, 26 June. Onto—Hocking County, S June. Onrarro—Burke Falls, 13 July. QuEBEc—Knowlton, 12-13 July. Sourm Daxora—Pennington County (Black Hills), 11 July. ‘TENNESSEE—Campbell County, 10 June. Dolichopeza (Oropeza) johnsonella ( Alexander ) Literature references—As Tipula annulata Say. Wiedemann, 1828: 54. Oropeza johnsonella Alexander. Alexander, 1930b: 279. Dolichopeza (Oropeza) johnsonella (Alexander). Alexander, 1941a: 295; Alexander, 1942: 212, fig. 26C (hypopygium). Original description —‘Size small (wing, male, under 10 mm.); mesonotum reddish brown, the brown praescutal stripes relatively indistinct; halteres dusky; legs pale brown, the tarsi a little paler; male hypopygium with the inner dististyle a flattened blade, the apex subtruncate, on outer margin near base with a small setiferous tubercle; gonapophyses recurved, tipped with acute spines. “Male. Length about 8 mm.; wing 9.8 mm. “Frontal prolongation of head and palpi dark brown. Antennae (male) relatively elongate, if bent backward extending to beyond the base of abdomen; scape honey-yellow; first flagellar segment THE CraNnrE Fiy GENus DOLICHOPEZA 815 short, the remaining segments passing into brown. Head dark brown. “Mesonotum reddish brown, the praescutum with three indistinct darker brown stripes. Pleura lighter brown, with vaguely indi- cated darker areas on anepisternum, ventral sternopleurite and ven- tral pleurotergite. Halteres dusky, the base of stem yellow. Legs with the coxae yellow, infuscated at base; trochanters yellow; a single (posterior) leg remains, pale brown, the tarsi a trifle paler, more yellowish brown. Wings tinged with brown, the stigma darker brown; veins brown. Venation: Cell M, about one-half longer than its petiole. “Abdominal segments ringed with brown and yellow, the apices of the segments paler than the bases. Male hypopygium with the lateral portions of the tergite produced into conspicuous setiferous shoulders, the intermediate margin very gently crenulate; ventro- lateral arms of tergite strongly curved, slender, not expanded out- wardly, the apex acute or subacute. Outer dististyle a little longer than the inner dististyle, cylindrical, not dilated at base. Inner dististyle a flattened blade, near base on outer margin with a small tubercle set with conspicuous setae; apex of style subtruncate. Gonapophyses recurved, setiferous, the tips set with several acute spines. ¥ In its small size and general appearance, Oropeza john- sonella agrees most closely with O. rogersi Alexander, differing in the structure of the male hypopygium, especially the inner dis- tistyle, which bears a setiferous tubercle on outer margin beyond base.” Type.—Holotype male, Riverton (northwest Burlington County ), New Jersey, August 1911, C. W. Johnson. This specimen, in the Museum of Comparative Zoology, Harvard University, now con- sists of the pinned thorax, part of the abdomen, and head with part of one antenna. One broken wing is attached to the thorax. The hypopygium has been mounted in a drop of balsam on a celluloid square attached to the pin; but, due to curling of the celluloid, the balsam has become wrinkled and cracked, as a result of which the hypopygium is scarcely visible and thus worthless. Although the type is of no use for comparison and, in addition, is scarcely typical of the species (as will be explained later), it seems unnecessary at the present to designate a substitute type. Diagnostic characteristics—As one of the dusky colored species of the obscura group, johnsonella is most easily confused with 816 Tue UNrversiry SCIENCE BULLETIN obscura, polita and tridenticulata, and it is difficult to identify even males without first capturing them. The golden-brown tinge and well-defined, oval stigmal spot of the wing and the presence of discernible stripes on the prescutum and spots on the thoracic pleura all aid in the recognition of this species. A good characteristic for field determination of males is the darkness of the eighth abdominal segment and contrasting paleness of the ninth, especially the ster- num, coupled with the presence of a usually conspicuous tuft of black hairs at the ventro-posterior apex of the eighth sternum. This latter feature, overlooked at the time of original description of the species and found only in johnsonella, is shown in Figure 162. Preserved males may be easily distinguished from all other species of the genus by the structure of the hypopygium. Of the obscura group, only johnsonella and subalbipes lack the toothed medio-pos- terior margin of the ninth tergum. Some specimens of subalbipes found in collections were labelled johnsonella, and attention was drawn to the white tarsi. Even legless specimens of subalbipes of the form having slender tergal arms may be told from johnsonella by the fact that the medio-posterior region of the ninth tergum, while it may be only two lobed, is produced caudad, in contrast to the broadly but shallowly emarginate, rather undulating margin of the ninth tergum of johnsonella (Fig. 157). Occasionally there is a minute median tooth, suggesting a similarity to australis; how- ever, the ninth tergum of the latter has the median tooth projecting beyond the broad lobes. Differences between these two species are discussed more fully under australis. Female specimens, some- times mistaken for faded Dolichopeza venosa, are distinguished from that species by the hypovalves: in venosa, there is dense sclerotization of the hypovalve from its dorsal margin to the mid- ventral line, beginning beneath about the mid-length of the tenth tergum and extending very nearly to the tip, which is abruptly paler by contrast; in johnsonella, the dense sclerotization is more limited to the ventro-lateral surface and is especially intense below the tenth tergum, growing less dense toward the tip of the hypo- valve (Fig. 163). Also, even in faded specimens, the golden-brown tinge of the wing will usually still be present and distinguish john- sonella. This wing coloration rather resembles that of similis, females of which may be recognized by the characters stated in the key. Descriptive comments.—The original description is fairly ade- quate as far as coloration is concerned. It might be well, however, THE CRANE FLY GENUS DOLICHOPEZA 817 oO 0.5 LOmm. SCALE, FIGS. 157-161 SCALE, FIGS. 162-163 Fics. 157-163. Dolichopeza (Oropeza) johnsonella; 157—ninth ter- gum of male, 158—left inner dististyle of male, dorsal aspect, 159— left outer dististyle, dorsal aspect, 160—gonapophyses, dorsal aspect, 161—-vesica and penis, 162—terminal abdominal segments of male, left lateral aspect, 163—terminal abdominal segments of female, left lateral aspect. §18 THE UNIversiry SCIENCE BULLETIN to mention that considerable variation should be expected. For example, prescutal stripes may sometimes be plainly visible, while in other specimens they may be obliterated completely; the legs may be dark brown as well as light brown; and under some condi- tions of preservation, body coloration of females may be so pale as to cause confusion of this species with Dolichopeza walleyi, until close examination of prescutum, wings and abdominal annulation is made. Venational variations found in johnsonella include loss of the proximal half of the vein M., presence of Sc,, presence of an extra cross-vein in cell R,, branching of M, from M,,, beyond the discal cell, loss of the medial cross-vein resulting in absence of a closed discal cell, and presence of a basal fragment of R,,,. All these variations are uncommon, and the venation of the species in general is quite constant. The cell M, is not always half again as long as its petiole, as described. Because the holotype was an autumn fly, affected by environ- ment as described earlier in this report, the size given by Alex- ander is not precisely characteristic of the species. The fact is that johnsonella is one of the larger species of the genus, some females reaching 14 mm. in body length, with wings of equal length. Males measure from 7 to 10 mm., the wing length varying from 9 to 12 mm. In all cases the small individuals were collected in August and September. Females of the late summer generation may be as small as 9 mm. in length, with wings 10 mm. long. The difference in mean size between spring and late summer generations in this species is marked, although individuals as small as the holotype are unusual. The characteristic structure of the male hypopygium of johnson- ella has already been described. Variation in the ninth tergum consists of slight differences in breadth of the low, rounded lobes and of the concavity between them, minor differences in length and shape of the tip of the tergal arms, and, as already mentioned, occasional presence of a tiny median projection. The setae on the outer surface of the inner dististyle (Fig. 158) are not always con- centrated upon a tubercle (compare Alexander, 1942: 213, fig. 26C) but may be more sparsely distributed over the angular bend of the dististyle. Geographical distribution Previously published records indicate that the range of Dolichopeza johnsonella is the Appalachian re- gion from New Jersey to South Carolina, but recent exploration THe CrANE Fity GENus DOLICHOPEZA 819 Map 4. Range of Dolichopeza (Oropeza) johnsonella ( Alexander ). Each spot represents one or more collections within a county ( United States) or at a locality. has shown that this area is actually only a small part of the species’ range. I found it common in three Indiana localities and in smaller numbers in southern Illinois and the Ozark-Ouachita Mountains region, as far west as the Arkansas-Oklahoma border. It also occurs as far north as New England and Quebec. Still, the distribution of this species is only sketchily known. It may be seen, by com- parison of the distribution maps of johnsonella and other species of the obscura group, that the known range of johnsonella is very unlike that of any of its near relatives. However, I have seen what I consider appropriate habitats for this species in Missouri, northern Illinois, Wisconsin and Minnesota, and I believe that further collect- ing will establish that johnsonella ranges much more widely than present evidence indicates. Habitats—Dolichopeza johnsonella appears to be a denizen of well-shaded situations that are in most cases rocky and often drier than habitats of most species of the genus. In the single published reference to habitats of this species, Alexander (194la: 295) lists as its haunts in western North Carolina “. along Neal’s Creek, beneath culverts . . . in small rock caverns below the (Con- estee) Falls, associated with D. (D.) americana Needham and D. (O.) dorsalis (Johnson) . . . along Hughes Ridge trail above Smokemont, 2500 ft., under darkened overhanging banks, swarm- ing in hollows thus formed. ” In Indiana, I took johnson- ella only in shaded crevices or beneath rock outcrops in limestone 820 Tue UnNrversiry SCIENCE BULLETIN and sandstone ravines; that is, never in equally dense shade in the woods nearby. My Ohio records are for similar habitats. Col- lections in the Ozark Mountains and Ouachita Mountains were made along rocky stream beds or on rocky hillsides. In western Florida, the habitats are not rocky but are rather like the pied- mont region in many respects, both physically and_ biologically (Rogers, 1933: 24-25). Localized aggregations of individuals are sometimes found, as for example along McCormick’s Creek, Owen County, Indiana, where I was unable to find johnsonella among swarms of resting Dolichopeza species for great distances above and below a water- fall, yet beneath certain outcropping ledges of limestone along about a fifty-foot length of the cliff near the fall this species was almost the only one to be found and was present in great numbers. It was this aggregation, incidentally, that led to discovery of a larval habitat for the species. Less pronounced localization of johnsonella was noticed at other localities. Seasonal distribution —Except in the Florida portion of its range, johnsonella has two annual generations, one in June and the other in August. Most Florida specimens were taken in the early days of April. Even the northernmost records seem to represent a June generation, but there have been too few collections of this species in New England and Canada to make this clear. It is my impres- sion, based on concentrated collecting in Indiana, that johnsonella reaches its peak of emergence a short while after all other species have passed their peaks. In Pope County, Illinois, in mid-July, it seemed that the local population of johnsonella was dwindling at the same time the second generation of americana and polita was getting underway, but I believe this situation is somewhat atypical for that latitude, especially insofar as the date is concerned. In Indiana, the spring peak of abundance of adults of johnsonella comes about the third week of June and the peak of the late sum- mer generation in late August. Immature stages.—Dissection of gravid females indicates one may lay as many as 120 eggs, which measure, on the average, .74 by .29mm. These are measurements and counts on June specimens. Eggs laid on the nights of 21 and 22 June began to hatch on the night of 28 June. In the near vicinity of the concentrations of johnsonella described above, the only bryophyte growing in any quantity was the moss Gymnostomum calcareum. It therefore seemed likely that this THe CRANE FLY GENUS DOLICHOPEZA 821 moss was the larval habitat of the crane fly, and a large amount of the moss and its marl substrate was brought into the laboratory and placed in a terrarium. From it, there emerged during the fol- lowing two weeks or so several adults of Dolichopeza polita cor- nuta and one adult of johnsonella; however, as I was away from the laboratory making further collections and field studies, I was unable to find the pupal skin of johnsonella at the time of emer- gence, and it had been destroyed by the time of my return. Thus, while the larval habitat is known, neither late stage larva nor pupa of johnsonella has yet been found. Having made a careful exami- nation of the Gymnostomum moss before placing it in the terra- rium and having found no larvae, I believe they must find daytime shelter in the porous marl deposited by the moss about its lower stems. It will be seen in Table 3 that johnsonella has been placed nearer the “dry habitat” end of the scale than would appear war- ranted on the basis of this one known habitat, which was quite wet. This was done because the species is so often taken as adults in situations where much drier conditions prevail. It is expected that the larva and pupa of johnsonella will somewhat resemble subal- bipes, perhaps having some characteristics of obscura. Notes on distribution—Arxansas—Garland County, 31 July; Polk County, 30 July; Washington County, 30 July. FLorma— Gadsden County, 28-29 March; Leon County, 18 March and 16 to 24 April; Liberty County, 3 to 18 April. Grorcra—De Kalb County, 19 August; Fulton County, 25 August and 24 September; Hall County, 6 June; Lumpkin County, 7 June; Oconee County, 2] Au- gust; Rabun County, 12 June; Union County, 10 and 28 June. ILtrnors—Jackson County, 4 June; Pope County, 14-15 July. Iv- pIANA—Montgomery County, 28 June; Owen County, 8 and 18 to 26 June; Parke County, 10 to 28 June and 30 August. MaryLanp— Baltimore County, 14 to 17 June; District of Columbia, 11 June and 9 July; Garrett County, 26 June; Montgomery County, 7 June and 18 August; Prince George’s County, 29 June. MassAcHUSETTS —Middlesex County, 18 June. NEw JersEy—Bergen County, | to 15 June; Burlington County, August; Essex County, June. NEw York—Nassau County, 27 June; Orange County, 26 August. NorrH Carotina—Cherokee County, 31 August; Haywood County, 29 July; Macon County, 8 to 12 June and 4 to 7 September; Swain County, 20 June; Transylvania County, 8-9 June and 8 September; Yancey County, 6 to 14 June. Onrto—Hocking County, 7 June; Portage County, 25 June; Summit County, 19 June. PENNSYLVANIA 822 THe UNIversiry SCIENCE BULLETIN —Huntington County, 9 July; Luzerne County, 10 July. QuEBEc —Knowlton, 20 June. SourH Carotina—Greenville County, 20 May; Pickens County, 29 June. TENNEssEE—Fentress County, 31 May to 18 June and 11 August to 3 September; Morgan County, 12 June; Scott County, 30 May. Vircinra—Arlington County, 11 to 25 August; Augusta County, 28 June; Fairfax County, 30 May, 5 and 16 June, and 30 August to 2 September; Giles County, 11 June to 7 July and 12 August; Wise County, 2 July. Wesr Vircrnta—Poca- hontas County, 23 June; Tucker County, 24 June. Dolichopeza (Oropeza) obscura (Johnson ) Literature references.—Oropeza obscura Johnson. Johnson, 1909: 122, pl. 15 (hypopygium; wing figured as this species is actually of tridenticulata); Johnson, 1910: 708; Alexander, 1919: 930, pl. 43 (wing); Alexander, 1920: 983-984, pl. 86 (larva figured is triden- ticulata; pupa is polita); Alexander and McAtee, 1920: 393; Alex- ander, 1925: 172; Johnson, 1925: 33; Pierre, 1926: 11; Alexander, 1928: 57; Leonard, 1928: 698; Alexander, 1929a: 236; Alexander, 1930a: 272; Alexander, 1930c: 1138; Rogers, 1930: 22-23 (includes polita and tridenticulata); Alexander, 193la: 138; Dickinson, 1932: 212, Fig. 114 (wing) (apparently refers to tridenticulata and polita pratti only ); Rogers, 1933: 49. Dolichopeza (Oropeza) obscura (Johnson). Alexander, 1936: 280; Alexander, 1940: 620; Alexander, 1941a: 295; Alexander, 1942: 212, 214, fig. 26D (hypopygium); Rogers, 1942: 59, 121; Whittaker, 1952: 36. Dolichopezia (Oropeza) obscura (Johnson). Rogers, 1949: 12 (typographical error ). Note: Dolichopeza obscura Brunetti, 1912, became a secondary homonym of Oropeza obscura Johnson, 1909, when the two genera were combined in 1931. Brunetti’s species, from India, was also recently placed in the subgenus Oropeza, according to Dr. Alex- ander, who is renaming it. Original description—*Head and thorax dark brown, opaque; palpi black, antennae fuscous, the two basal joints yellow; abdomen light brown somewhat shining, the black bands at the incisures not connected on the dorsal line. Genitalia yellowish, appendages blackish; style yellow, long and very slender, curved and often extending to the base of the penultimate segment, appendages at the base of style short and armed with small black spines, margin but slightly emarginate. Ovipositor brown, cerci yellow. Halteres THE CRANE FLy GENUS DOLICHOPEZA §23 yellow, knobs brown. Legs brownish yellow. Wings smoky brown, the veins and stigma a slightly darker brown, the median cubital cross-vein wanting. Length, male, 8 mm.; female, 10 mm.” Types.—Holotype male, North Adams ( Berkshire County ), Mas- sachusetts, 19 June 1906, C. W. Johnson. Allotype, Hammond's Pond, near Brookline (Norfolk County), Massachusetts, 18 June 1908. Five male and one female paratypes are together with the holotype and allotype, in the Museum of Comparative Zoology at Harvard University. Of these, the males labelled Auburndale, Mas- sachusetts, 11 July 1905, Squam Lake, New Hampshire, 14 July 1907, Norwich, Vermont, 7 July 1908, and North Adams, Massachu- setts, 19 June 1906, are actually Dolichopeza tridenticulata Alex- ander. The allotype and one female paratype may also be this species, but I could not be certain at the time I examined these specimens. Another male and female, both paratypes and both obscura, are in the collection of Dr. Alexander, at Amherst, Mas- sachusetts. One male bearing the same label data as the holotype, thus probably one of the original paratypes although not so labelled, is in the Snow Entomoloical Museum, Lawrence, Kansas. Two fe- male paratypes are in the collection of the Academy of Natural Sci- ences of Philadelphia. Johnson based his description of obscura on thirty-four specimens; the present whereabouts of the twenty-one not accounted for here is unknown. Diagnostic characteristics—Males of obscura may be at once recognized by the shape of the lateral arms of the ninth tergum. In no other species are these structures widely flared apically, then emarginate at the tips (Fig. 164). The profile of the medio- posterior margin of the ninth tergum is also a very useful character, having three distinct and rather widely separated teeth, the median tooth usually shorter than the others. Even after considerable experience with this species in the field, it is difficult and not reliable to make sight identifications. Doli- chopeza obscura too closely resembles tridenticulata for field recognition without capturing specimens. In most places where I have found the two species together, obscura has been the more darkly colored, although, particularly in the northern parts of the range, the coloration of both is nearly alike. The same might be said of their relative sizes: obscura in most parts of the range is slightly larger than tridenticulata, but in the northern states and Canada and at many localities in the highlands of the Appalachians, tridenticulata often reaches the size of obscura. The wing colora- 824 THE UNIversITy SCIENCE BULLETIN tion of obscura is more intense than that of tridenticulata, but it is almost necessary to see the two species together, knowing which is which, in order to appreciate this comparison. Female specimens may be distinguished from polita by having distinct annulations on the abdomen and the prescutum dark brown and never shining. Lack of stripes on the thoracic dorsum and differences in coloration of the hypovalves of the ovipositor will separate obscura from johnsonella. Separation of obscura and tri- denticulata females is very difficult. The details of coloration of the hypovalves mentioned in the key to adult females is the most reli- able character I have found and is based upon study of a large num- ber of females of known identity, as those taken in copula. A pecu- liarity noted in many preserved specimens is that while the cerci of obscura remain in alignment with the rest of the abdomen, those of tridenticulata are often deflected ventrad, so that the tips of the hypovalves are concealed between them. I have not discovered any anatomical reason for this difference of position of the cerci in killed flies. The color differences between obscura and tri- denticulata described for males apply also to females. Associated males are a help in identifying either obscura or tri- denticulata females, as they provide at least standards for compari- son of color. Habitat also, within certain limits, will aid in identi- fication. Descriptive comments.—Dolichopeza obscura is the darkest of the species of Oropeza, a very dusky-brown fly that is nearly invis- ible in the deeply shaded recesses beneath trees and rocks where it hides by day. The annulated pattern of abdominal coloration is distinct. The occiput of the head is a very dark grayish brown but scarcely darker than the deep brown of the pronotum and prescutum. Johnson’s description of the wing coloration as “smoky brown” is apt, for the ground color does indeed suggest smoked glass. The veins and stigmal spot are dark brown, usually con- trasting more with the ground color than is the case in triden- ticulata. Although irregularities in wing venation are not at all uncommon in obscura, I have not found any strong local concentrations of par- ticular variations, such as described earlier for tridenticulata. Ab- sence of the medial cross-vein is the most frequently observed abnormality of venation. Loss of sections of veins in the medial field is not unusual, and the presence of the vein Sc, has been noted many times. Junction of the m-cu cross-vein near the mid-length THE CRANE Fiy GENUS DOLICHOPEZA 825 fe) 0.5 1LOmm. SCALE, FIGS. 164-169 Fics. 164-170. Dolichopeza (Oropeza) obscura; 164—ninth tergum of male, 165—left inner dististyle of male, dorsal aspect, 166—left outer dististyle, dorsal aspect, 167—gonapophyses, adminiculum and adminicular rods, dorsal aspect, 168—vesica and penis, 169—medio- posterior margin of ninth tergum of male holotype, 170—terminal abdominal segments of female, left lateral aspect. 28—5840 §26 Tue UNrversiry SCIENCE BULLETIN of the short M,,, is quite common. One specimen was seen in which the cell M, was very short, its petiole correspondingly elon- gated, and one other had an extremely short discal cell in both wings. A short spur directed proximally from about mid-length of the m-cu cross-vein has been seen once. Body length of males varies from 7.5 to 10 mm., the wing length from 9 to 12.5 mm. Females measure from slightly under 10 mm. to 12 mm. in body length, their wings from 10 to 13 mm. Most of the larger specimens I have seen were from the northern states and from high elevations in the central Appalachian Mountains (North Carolina and Virginia), in July and June, respectively. Smaller individuals were found in October and March collections from northern Florida and southern Georgia (Baker County). Variation in the ninth tergum of the male consists mainly of differences in shape, length and spacing of the three acute teeth of the margin, and of fluctuations in the width, curvature and apical outline of the tergal arms. In the holotype, for example, the central tooth is deflected to one side (Fig. 169); in a specimen from Whit- ley County, Kentucky, there are four teeth, the usual median tooth apparently having become divided; and in a male from Westmore- land County, Pennsylvania, one of the lateral teeth has an apical bristle, while the central tooth is represented only by a bristle and the entire medio-posterior margin is atypical in form. Breadth of the basal portion of the inner dististyle varies noticeably, although the over-all shape of this structure never approaches that of any other species of Dolichopeza. The inner dististyle is narrower, compared to its length, in obscura than in the other species (Fig. 165). In males, the antennal length is nearly equal to half the length of the body. Scape and pedicel are usually of a yellowish gray color, the flagellum darkening from that to deep fuscous at the tip. Geographical distribution —Collection records indicate that ob- scura is the commonest Dolichopeza in North America. It has been found from Nova Scotia westward to Alberta, southward to Florida and southwestward to Arkansas. I expect it will subsequently be found in swampy areas of Louisiana and Mississippi, in the Ozark Mountains of Missouri and in the rocky ravines of Iowa. The dis- tribution map of this species shows the range fairly well outlined in eastern North America, and although I anticipated obscura would occur somewhat westward from Manitoba and Minnesota, I was rather surprised to find it in a forest in western Alberta, in 1957. THE CRANE FLY GENUS DOLICHOPEZA 827 Map 5. Range of Dolichopeza (Oropeza) obscura (Johnson). Each spot sooresents one or more collections within a county (United States) or at a locality. Collections such as this and my earlier discoveries of obscura in Kansas and Arkansas serve to emphasize how little still is known of the geographical distribution of many of the North American crane flies. Habitats —The relative commonness of obscura in collections is very probably a reflection of its ability to flourish in a wide variety of general environments. This is the only species of Dolichopeza that seems equally at home in rock gorges, in moist to wet wood- lands, and swamps, bogs and border vegetation of marshes. In southern Michigan, Rogers (1942: 59) found obscura “numerous to occasionally common in the most densely shaded birch-maple-elm and tamarack-sumac swamps. Numerous but very local in oak- hickory woods, where they were occasionally taken from the dark interior of a hollow stump or log.” In his ecological study of the Tipulidae of northern Florida (Rogers, 1933), he records this species as occurring “. . . within hollow trees and dark, cool recesses of banks and cliffs.” Habitats of obscura on the Cumber- land Plateau are described as “. . . shaded, rocky, upper, talus slope brooks, ‘rock-houses, beneath overhanging banks of the upland brooks, and . . . large hollows in standing trees. Rare 828 THe UNIVERSITY SCIENCE BULLETIN in the stream-margin thickets” (Rogers, 1930: 23). In central In- diana and Ohio, I found obscura in sandstone ravines and rarely in limestone ravines, which seem to be drier, but more often the species was taken in upland woods. In numbers of individuals, obscura never equals polita or tridenticulata in the rock gorge habi- tats. In northern Michigan, obscura is extremely numerous in mesic and wet woodlands, especially in areas of deciduous forest. In the white cedar swamps and drier coniferous forests, it was always found comparatively less abundant. In the spruce forests along the crest of the Appalachians, however, obscura is the common- est Dolichopeza. It is my impression that Dolichopeza obscura is the most strongly negatively phototropic member of the genus. Where this species was observed resting together with americana and polita cornuta, in Parke County, Indiana, it was in the darkest recesses of cavities shared by the three species. In other localities, also, obscura is found concentrated in the most intense shade available, and the flies are less readily driven from their hiding places than is the case with other species. In view of this reaction to light, it might seem surprising that obscura has been more often taken in light traps than any other species of its genus. On the University of Florida Conservation Reserve, at Welaka, Putnam County, Florida, Dr. R. E. Bellamy operated a light trap on most nights of the year 1946, taking six individuals of Dolichopeza obscura in April, three in May, seventy-five in June, five in July, two in August, none in September, and seven in October. He also obtained March, Sep- tember and November records, that year, by other means. The only other Dolichopeza Bellamy found was subalbipes, which came to the light trap in much smaller numbers. It should be mentioned that Bellamy’s light was very near the daytime resting places of the crane flies. It seems likely that obscura was present in the vicinity of his trap in great numbers and that because the species is active at night a certain number of flies randomly wandered to the light. It is, of course, possible that the reaction of Dolichopeza obscura (as well as other species) to light is like that of many moths, which are repelled by the bright light of the sun by day but are attracted to lights of lower intensity at night. In Cheboy- gan County, Michigan, I operated a light trap on a dry hill top, a few hundred yards distant from any known diurnal haunts of obscura, and still trapped, during July and August, half a dozen specimens. This suggests that obscura does not vigorously avoid light under all circumstances. THe CRANE FiLy GENUS DOLICHOPEZA §29 Seasonal distribution—There is evidence, locally through much of the range, that there are two generations per year, except in Florida and in the northernmost parts of the range in the United States and Canada. In southern Michigan and in Indiana, I have found two well-marked annual generations, the first from about mid-June to early July and the second from mid-August to early September. In 1953, however, the first male of obscura seen at Turkey Run State Park, Parke County, Indiana, appeared on 28 May, about two weeks after the emergence of americana began; but there were only one or two individuals of obscura observed on any day until 9 June, when an abrupt increase in numbers was noted. There are Florida records for nearly every month of the year, al- though peaks of abundance occur in March and June. Scattered collection records indicate that obscura is present in greater or smaller numbers during June, July and August, in virtually any part of the known range. It is only in areas of concentrated col- lecting that the two-generation pattern becomes evident. The July-August trend in numbers of individuals and ratio of males to females in the northern Michigan populations suggests that there is only one, generally midsummer generation each year in that area. Immature stages—Eggs of Dolichopeza obscura from Michigan averaged .73 by .40 mm., which is unusually wide for the length. Smaller females from Florida and Georgia were found to contain shorter eggs which were comparatively more slender, measuring 65 by .28 mm. Duration of the egg stage appears to vary with temperature, for I found it to be about ten days, in northern Michi- gan, in late July, while Rogers (1933: 32) reports 13 to 16 days, in Florida, in March and April. Based upon dissection of teneral, gravid females, the number of eggs produced by a single fly is close to 120, perhaps occasionally more than that but probably more often fewer, I believe, for that average count includes late- stage unmatured eggs in the ovarioles, which I have reason to suspect may not be matured and laid. A previously published account of the larva of obscura (Alex- ander, 1920: 983) described the habitat as dry moss, Hedwigia albicans; however, in later investigations, Rogers (1933: 49) dis- covered the immatures of obscura in “wet to moderately dry moss clumps; pupation in short tunnels between moss and wood.” In my own studies, I repeatedly reared obscura from damp mosses, never from dry ones. I then discovered, from morphological studies of the larvae, that Alexander’s notes actually applied to tridenticu- $30 THE UNIversiry SCIENCE BULLETIN lata, which species had then not yet been recognized as distinct from obscura. In hardwood swamps in southern Michigan, I found larvae of obscura numerous in the moss, Tetraphis pellucida, grow- ing on decayed stumps and roots on hummocks; and less com- monly they were found in mats of Heterophyllium haldanianum moss growing on sodden, rotten logs. Other larval habitats on the low hummocks in swampwoods were the mosses Hypnum curvi- folium and Leucobryum glaucum, the latter growing much more luxuriantly there than it does more commonly on drier soil in up- land woods. Although mosses growing on decayed wood are the commonest larval habitats of obscura wherever the species occurs, I have also reared it from bryophytes growing on soil, such as the moss, Fissidens taxifolius, and the liverwort, Chiloscyphus pal- lescens, both found on glacial till on well-shaded slopes. The fourth instar larva may be identified by the presence of conspicuous transverse ridges of microscopic hairs on the dorsum of the thoracic and all abdominal segments, with the intervening rows of minute microscopic hairs long, irregular, and with their ends not clearly defined. On a typical abdominal segment, there are from front to rear four ridges about equally spaced, then a wider space, then a fifth well-marked ridge, followed by two or some- times three fainter, more irregularly spaced ridges. The larva con- structs a distinct burrow, in most mosses, in which to pupate. Pupae of Dolichopeza obscura frequently occur together with those of subalbipes in a single mat or cushion of moss, and the pupae are very much alike. In both, the pleural spines of the sec- ond through seventh abdominal segments bear a single, long bristle, and in both the middle projections of the eighth abdominal sternum are set very close together or upon a common base. The spiracular yoke of obscura (Fig. 116) is, however, unlike that of subalbipes, as described in the key to pupae. Comparing pupae of the two species during their development, I noted that by the fourth day the pupa of obscura had an over-all grayish brown color, the wing sheaths somewhat paler and the thoracic dorsum brown, while the pupa of subalbipes was decidedly more greenish—rather a mixture of brown, gray and green difficult of description—the wing sheaths again paler but the thoracic dorsum grayish green. Notes on distribution —A.Berta—89 miles south-southeast of Valley View, 10 July. Arkansas—Garland County, 31 July; Wash- ington County, 30 July. Connrcricutr—Litchfield County, 9 and 23-24 July. FLorma—Alachua County, 7 to 29 March, 1 to 25 THe CRANE FiLy GENUS DOLICHOPEZA 831 April, 3 May, 10 June, 4 and 29 July, 16 October and 14 November; Escambia County, 7 April; Gadsden County, 29 March to 19 April; Hernando County, 21 March; Jackson County, 31 March, 12-13 April and 6 May; Leon County, 31 March, 14 to 24 April and 13 June; Liberty County, 27 March and 3 to 13 April; Madison County, 1 September; Marion County, 24 March to 4 April; Putnam County, 26 date records from February to November, with peak abundance in early June; Washington County, 8 June. GErorci1a—Baker County, 19 date records for January through May, July, August and October; Chattahoochee County, 11 October; DeKalb County, 4 October; Lumpkin County, 23 May; Union County, 10 and 28 June; Ware County, 1 July. Ittrmors—LaSalle County, 7 July; Ogle County, 19 June. Inprana—Allen County, 10 July; Jefferson County, 7 June and 1 August; Monroe County, 23 June; Mont- gomery County, 28 June; Parke County, 28 May to 29 June, 15 July and I and 30 August. Kansas—Douglas County, 30 August. KEN- tTucKy — Barren County, 2 August; Whitley County, 24 June. Matne—Cumberland County, 1 July; Hancock County, 16 August; Oxford County, 11 July; Piscataquis County, 17 July; Washington County, July. Manrropa—West Hawk Lake, near Rennie, 4 Au- gust. MaryLanp—Baltimore County, 14 and 17 June; District of Columbia, 29 August (reported 17 May to 30 August, but these records are not supported by available specimens ); Garrett County, 26 June; Montgomery County, 30 May; Prince George’s County, 18 June. Massacnuserrs—Berkshire County, 19 June and 10-11 Au- gust; Hampden County, 3 August; Norfolk County, no date; Suffolk County, 8 August. MicuicAn—Antrim County, 2 July; Cheboygan County, 22 June to 8 August; Chippewa County, 13 to 15 July; Emmet County, 20 July to 4 August; Gogebic County, 6 August; Huron County, 20 June; Iosco County, 22 July; Lake County, 22 June to 12 July; Livingston County, 27 May to 9 July, 31 July, and 11 August to 9 September; Mackinac County, 27 July to 7 August; Marquette County, 13 to 17 July; Oceana County, 26 July; Oscoda County, 23 to 26 June; Otsego County, 1 to 4 July; Presque Isle County, 28 July; Washtenaw County, 12 June to 3 July and 10 to 15 August. Minnesora—Carlton County, 5-6 August; Hennepin County, 12 June; Winona County, 7 July. New Brunswick—Fred- erickton, 26 August. New Hampsurre—Cheshire County, 16 June; Coos County, 2 to 11 July; Grafton County, no date; Merrimack County, July (?). New Jersey—Bergen County, 1 to 15 June; Burlington County, 11 August. New Yorx—Albany County, 17 June to 3 July; Cattaraugus County, 29 June to 4 July; Chenango 832 Tue UNtversiry SCIENCE BULLETIN County, 21 July; Cortland County, 20 July; Erie County, 16 June to 10 July; Fulton County, 15 June to 20 August; Greene County, 1 July; Hamilton County, 12 and 30 July; Herkimer County, 3 July and 12 August; Oneida County, 20 June and July; Orange County, 26 to 28 August; Suffolk County, 5 July; Tompkins County, 6 July; Warren County, 26 July. Norra CaroLtina—Avery County, 14 to 16 June; Buncombe County, 29 May and 13 June; Burke County, 14 June and 1 July; Haywood County, 30 July; Macon County, 8 to 15 and 29 June, and 24 August to 3 September; Mitchell County, 16 June; Swain County, 11 to 30 June; Transylvania County, 8 to 13 June; Yancey County, 26 May, 2 to 22 June and 1 July. Nova ScoTtra—Guysborough County, 30 June; Victoria County, 1 July. Ounr1o—Hocking County, 30 May and 7 June; Portage County, 24 June. Onrarro—Algonquin Park, 23 June to 11 July, and 21 Au- gust; Burke Falls, 14 July; Gull Lake, Muskoka District, June; Lake of the Woods, Kenora District, 4 August; Lyn (20 miles northeast of Gananoque), 10 August; Niagara Glen, 30 June; Normandale (on Lake Erie, about 10 miles south of Simcoe), 28 June; Pte. au Baril, Georgian Bay, 6 August; Severn (Lake Simcoe), 16 June; Thunder Bay (Lake Superior), 8 July. PENNsyLvANIA—Bedford County, 8 July; Centre County, 25 June; Columbia County, 15 July: Huntington County, 9 July; Luzerne County, 17 to 28 June, 5 and 10 July and 15 August; Mercer County, 25 June; Sullivan County, 27 June and 10 July; Wayne County, 18 July; Westmoreland County, no date. QurBEc—Knowlton, 29 June to 12 July; Rigaud, 25 June. RuHopE Istanp—Washington County, 18 June and 10 August. SoutH Carotina—Greenville County, 24 August. TENNESSEE— Fentress County, 11 to 19 June, 2 and 16 July and 13 August; Hay- wood County, 25 May; Knox County, 29 May; Morgan County, 12 June; Scott County, 30 May; Sevier County, 20 May to 12 June and 30 June. VERMont—Chittenden County, 15 to 24 June; Essex County, 2 July; Orange County, 11 July; Rutland County, 12 July; Washington County, July. Vimcrra—Arlington County, 11 June and 11 August; Fairfax County, 2 and 16 June and 5 September; Giles County, 18 June to 22 July, 2 and 15 to 26 August; New Kent County, 31 May; Page County, 15 July; Rockingham County, 6 July; Shenandoah National Park, 29 June and 6 July; Washington County, 2 July; Wise County, 2 July. Wesr Vircinta—Greenbrier County, 4 July; Hardy County, 2-3 July; Pocahontas County, 23 June and 5 July; Preston County, 25 June and 5 August; Randolph County, 5 July and 6 August. Wusconstn—Juneau County, 6 July; Trempealeau County, 7 July. THE CRANE FLY GENUS DOLICHOPEZA 833 Dolichopeza (Oropeza) polita polita (Johnson ) Note: Two similar forms of Dolichopeza, regarded heretofore as full species, have been described as polita and pratti. Closely related to these is a third form, described below. The reasons for considering these three forms as subspecies, here, are discussed in detail in the conclusions. Literature references—Oropeza obscura var. polita Johnson. Johnson, 1909: 122-123, pl. 15 (wing); Alexander, 1919: 930; John- son, 1925: 33; Alexander, 1929c: 297; Rogers, 1930: 22 (as part of obscura Johnson). Oropeza polita Johnson. Alexander, 193la: 138-139 (the variety redescribed as a full species ). Dolichopeza (Oropeza) polita (Johnson). Alexander, 1941la: 296; Alexander, 1942: 214, fig. 26E (hypopygium). Original description —‘Distinguished from the typical form by having the entire dorsum highly polished. One specimen has the vein forming the anterior side of the discal cell wanting or indicated by a stub. Length, male, 9 mm.; female, 11 mm.” Redescription by Alexander (1931a):—“Generally similar to O. obscura Johns., differing especially in the short antennae and struc- ture of the male hypopygium. “Antennae much shorter than in obscura, if bent backward scarcely attaining the root of the haltere. Mesonotum dark brown, nitidous, without stripes. Knobs of halteres darkened. Tarsi more evidently darkened. Wings with the stigmal area paler, not con- trasting strongly with the ground-color. Venation: cell Ist M, narrow at base. Abdominal tergites almost uniformly darkened, not conspicuously bicolorous as in several related species, the outer segment and hypopygium almost black; basal sternites a_ little brighter but not conspicuously dimidiate. Male hypopygium with the median region of the tergite produced into a quadrate plate that is further produced into a sharp median point; incurved lateral arms of tergite elongate, at tips dilated into spatulate dusky blades, the margins smooth. Outer dististyle black, sinuous, at base dilated and expanded, at tips nearly acute. Inner dististyle much more expanded than in obscura, the blade approximately as wide as long.” Types.—Holotype male, North Adams (Berkshire County), Mas- sachusetts, 8 August 1907, Owen Bryant. Allotype, same data. One male and two female paratypes, together with the holotype $34 THE UNIVERSITY SCIENCE BULLETIN and allotype, in the collection of the Museum of Comparative Zoology, Harvard University. Diagnostic characteristics —Both sexes of this race of polita may be identified in the field by their dark grayish brown coloration, large size and polished appearance, the last especially evident on the thoracic dorsum, which has a somewhat paler color and higher gloss than the abdomen. Males may be recognized by the struc- ture of the hypopygium: the tergal arms are expanded at the tips and very strongly sclerotized, nearly black, and the bulbous en- largement on the mesal surface of the base of each outer disti- style is conspicuous and usually darkened (Figs. 171 and 173). Lack of distinct abdominal annulations and the low contrast in color of the stigmal spot of the wing as compared to the ground color are characters that will aid in distinguishing polita polita from other species but that are found also, in greater or lesser degree, in all three forms of the species polita. Females of this subspecies have the subapical sclerotization of the hypovalves more intense than that of the other two races. Descriptive comments.—The highly polished appearance of this fly, while pronounced enough to identify the subspecies, is actually a relative quality. There is a glossiness, in certain lights, in the exoskeleton of obscura, sometimes in tridenticulata, and in the other races of polita. Rogers (1930: 22) concluded that some de- gree of the polished effect is related to “ post-mortem changes, associated with slow and imperfect drying.” I have seen a few specimens of the other forms of polita in which some thoracic lustre seems to have resulted from the internal separation of the flight muscles from the sclerites, but the polished appearance thus produced never equals that of the typical form of polita, in which it is very evident in the living flies. It is stated by Rogers (1930: 22) that “among the large series (about 100) taken of O. obscura are a number of specimens that now have the appearance of O. obscura polita with the dorsum of the thorax highly polished.” Actually, this series does contain, in addition to obscura, several specimens of the typical race of polita, some of polita cornuta, and some ap- parently intergrades (see distribution map for this species). The general coloration of this subspecies is darker than that of the other two, the abdomen especially being strongly tinged with dark green, particularly in females. In the races pratti and cornuta, the abdomen is more often of the grayish brown color found in 835 THe CRANE Fiy GENuS DOLICHOPEZA niga [e) 0.5mm. l a=) SCALE, FIGS.171-178 oO 1LOmm. rerbisee J 2ty Fics. 171-179. Dolichopeza (Oropeza) polita polita; 171—ninth ter- gum of male, 172—left inner dististyle of male, dorsal aspect, 173—left outer dististyle, dorsal aspect, 174—-gonapophyses and adminiculum, dor- sal aspect, 175—vesica and penis, 176-178—variations in medio-posterior margin of ninth tergum of male, 179—terminal abdominal segments of female, left lateral aspect. 836 THE UNIVERSITY SCIENCE BULLETIN tridenticulata, so that in males there are annulations of darker color often in evidence. Venational variations observed in Dolichopeza polita polita in- clude loss of the medial cross-vein with the discal cell remaining open and absence of this cross-vein due to closure of the distal end of the discal cell by temporary contact of the veins M,,, and M,. Absence of portions of M, and M, are not uncommon, but the ab- normality figured by Johnson (1909: 122 and plate 15, as mentioned in the original description quoted above) is very unusual. Slight displacements of the m-cu cross-vein with relation to the discal cell, as well as minor fluctuations in the positions of other veins, are often seen. Body length of males varies from § to 11 mm.; wing length from 10.5 to 14 mm. Females measure 8 to 12 mm.; wing length 9 to 13 mm. Throughout this species, that is, in all three races, the females are rather generally smaller than the males, although the range of body length given here does not indicate this. In all other species of North American Dolichopeza, the reverse is true, as it is also among Diptera in general. The size difference between spring generation individuals and those of the late summer generation is pronounced, all the larger specimens seen having been collected in June or early July. Most variation in the male hypopygium occurs in the structure of the ninth tergum. The tergal arms vary only slightly, always within the limits of the description by Alexander, that is, with the tips widened, flattened and entire. The profile of the medio- posterior margin of the ninth tergum, in contrast, varies greatly, as indicated in Figures 176 to 178. Specimens having three acute teeth in this position are not common, the form most often seen being that figured in the illustration of the typical tergum (Fig. 171). Specimens approaching either polita pratti or polita cornuta in this character are quite uncommon. While other species of the obscura group have comparatively long antennae, in the males, the antennae of this and the other races of polita are short for the over-all body size, as emphasized by Alexander. Geographical distribution —Although ranging southward to Georgia along the eastern slopes of the Appalachian Mountains, this race is, | believe, the most northern of the three in the character of its distribution. The central Michigan and west-central Wiscon- sin localities (see distribution map) seem out of place and are in- THE CRANE FLY GENUS DOLICHOPEZA 837 subspecies of polita Map 6. Range of Dolichopeza (Oropeza) polita sspp. Squares—po- lita polita (Johnson); solid circles—polita cornuta new subspecies; hollow circles—areas of intergradation between p. polita and p. cornuta; solid tri- angles—polita pratti Alexander; hollow triangle—area of intergradation be- tween p. cornuta and p. pratti. Each spot represents one or more collections within a county (United States) or at a locality. deed difficult to explain. As polita polita seems less bound to rocky ravine habitats in the northernmost parts of its eastern range than is either of the other two races, I believe it may extend westward from New England, across southeastern Canada, where such habi- tats are rare, to connect with the Michigan and Wisconsin popu- lations. Further collecting in the northern Great Lakes area is necessary to show the true relationships of the races in the northern parts of their ranges. Where this form occurs together with polita cornuta (hollow circles on the map ), intergrades are usually found, as already indicated for the eastern Tennessee locality. In Hocking County, Ohio (the southern of two hollow symbols shown in that state), only two specimens from a sample of 101 were of the polita polita type, while in the northern part of Ohio (Portage County ) the two forms may occur in almost equal numbers, without inter- grades. This problem is discussed in detail in the conclusions. Habitats —Through the greatest part of its range, polita polita is found in rock gorges, in the shade of undercut sandstone or lime- stone cliffs, and in darkened crevices between or beneath boulders in the mountainous areas. In the northeastern part of the United States, however, it occurs in much smaller crannies, such as among §38 THE UNIVERSITY SCIENCE BULLETIN boulders of rather small size on wooded slopes and along mountain brooks. While this race seems to have somewhat higher moisture requirements than either of the other races of polita, I have taken it occasionally in shaded niches having no evident moisture and located a hundred yards or more from any source of water. Alexander (1929c: 297) described habitats of this fly in Taconic Park, Columbia County, New York, very similar to those in which I have collected it. Seasonal distribution—In only a few localities is there any evi- dence of two annual generations. In Eaton County, Michigan, I found the first emergence of adults to take place around the mid- dle of June and the second from mid-August to early September. In the eastern states, the collection records are scattered, so that there is no evident pattern of seasonal distribution. There are records for late May to the end of August, with the greatest num- ber of specimens having been taken in July. The earliest dates are for southern Ohio and the southern Appalachian Mountains (30 May and 5 June, respectively ), but there are records for only a few days later as far north as Quebec. Immature stages —Eggs of Dolichopeza polita polita laid by fe- males of the late summer generation, in Eaton County, Michigan, averaged .75 by .30 mm. in size, with very short, slightly curled terminal filaments. A single teneral female contains about 110 eggs that are matured or nearly so. Of eggs laid on 17 August, the first began to hatch on 24 August, and many were hatched by the next day. The first molt came about 17 days later. Larval habitats of this subspecies include the mosses Campylium chrysophyllum, Desmatodon obtusifolius, Gymnostomum calcareum and Myurella careyana and the hepatic Scapania nemorosa. All these bryophytes were growing on slightly moist to damp, shaded sandstone, except the Gymnostomum, which was wet. Pupal skins of polita polita and/or polita cornuta were found in the mosses Eurhynchium serrulatum and Hypnum sp. growing on damp, trapped soil on a sandstone cliff in Portage County, Ohio; and in Eaton County, Michigan, larvae were found in the moss Plagiothe- cium roeseanum growing under similar conditions. Larvae of polita polita closely resemble those of tridenticulata but are larger, often attaining 18 mm. in length. I am so far unable to distinguish the larvae of the three races of polita, but separation of any of these from tridenticulata may be accomplished by com- THE CRANE FLY GENUS DOLICHOPEZA 839 parison of the pleural regions of the eighth abdominal segments, as stated in the key to larvae. The pupa may be recognized by having the tracheal connection between the thoracic breathing horn and the mesothoracic spiracle convoluted, as well as by the characteristic form of the spiracular yoke (Figs. 107 and 117). Before the final darkening just prior to emergence of the adult, the pupa has a dark grayish green color, the thorax and wing sheaths light brown. Notes on distribution —Connecticut—Litchfield County, 24 July and 19 August. Grorcra—DeKalb County, 19 August; Lumpkin County, 5 June; Union County, 28 June. KeNtrucky—Letcher County, 3 July (includes intergrades with ssp. cornuta). Matwe— Oxford County, 11 July. Maryitanp—District of Columbia, 25 Au- gust; Garrett County, 26 June (also intergrades with ssp. cornuta). MassacuusETts—Berkshire County, 8 and 11 August; Norfolk County, June; Worcester County, no date. MicHicAn—Eaton County, 3 to 14 June and 16 to 30 August. New Hampsuire—Coos County, 2 and 20 July; Grafton County, 5 July. New JersEy—Essex County, June. New Yorx—Broome County, 13 July; Columbia County, mid-August; Hamilton County, 12 July; Herkimer County, 3 July. Norra Carotina—Burke County, 14 and 21 June and 1 July; Haywood County, 25 to 30 July; Macon County, 11 to 13 June and 23 August; Swain County, 30 June; Transylvania County, 14 June. Onio—Hocking County, 30 May, 6-7 June and 31 August (also intergrades with ssp. cornuta); Portage County, 24-25 June and 14 July (ssp. cornuta also here). Ontrarro—Gananoque, 8 July. PENNSYLVANIA—Luzerne County, 27 June and 5 to 10 July; Wyoming County, 15 July. Qursec—Knowlton, 12 June, 12 and 29 July. SoutH Carotina—Greenville County, 7 June. TENNEssEE—Fentress County, 22 July to 14 August (includes intergrades with ssp. cor- nuta); Sevier County, 30 June. Vircrnta—Arlington County, 25 August; Giles County, 21 June to 13 July and 1 August (includes intergrades with ssp. cornuta); Grayson County, 29 June and 7 Au- gust; Shenandoah National Park, 6 July; Washington County, 2 July; Wise County, 2 July (includes intergrades with ssp. cornuta). Wrst Vircinta—Pocahontas County, 5 July (includes intergrades with ssp. cornuta); Randolph County, 5 July. Wisconstn—Juneau County, 6 July. 840 THe UNIVERSITY SCIENCE BULLETIN Dolichopeza (Oropeza) polita pratti Alexander (new combination ) Literature references—Dolichopeza (Oropeza) pratti Alexander. Alexander, 1941b: 192-193. As Oropeza obscura Johnson. Dickinson, 1932: 212 (part). Original description —*Belongs to the obscura group; general col- oration of mesonotum opaque brown, without clearly defined stripes; legs dark; wings with a brownish tinge, the oval stigma a little darker brown; vein Sc, preserved; abdominal segments bicolored; male hypopygium with median area of tergite narrowly produced into a tridentate lobe; lateral tergal arms appearing as narrow spatu- late blades; outer dististyle a little dilated on basal portion, the apex a short, spinous point; inner dististyle deep, its rostral prolonga- tion long; aedeagus simple, unarmed. “Male. Length about 8-9 mm.; wing 10-10.5 mm.; antenna about 2.8 mm. “Frontal prolongation of head brownish black; palpi dark brown. Antennae with scape brownish yellow; pedicel light yellow; flagel- lum black; verticils of flagellar segments coarse. Head dark gray. “Mesonontum brown, the surface of praescutum opaque, the posterior sclerites more nitidous; in some cases the praescutum with faint indications of lighter stripes. Pleura paler brown. Halteres dusky. “Legs with the coxae pale brown; trochanters obscure yellow; remainder of legs brown, including the tarsi. “Wings with a brownish tinge, the oval stigma a little darker brown; prearcular field a very little brightened; veins brown. Vena- tion: Sc, preserved, Sc, ending opposite or just beyond the origin of Rs; petiole of cell M, exceeding m. “Abdominal tergites obscure brownish yellow to testaceous yel- low, the lateral margins and incisures darkened, on the outer seg- ments and hypopygium the dark color including all of the seg- ments; basal sternites yellow, the incisures narrowly darkened, the outer segments more generally suffused. Male hypopygium with the median area of tergite produced into a narrow lobe, the apex of which is further toothed, usually tridentate, with the central point longest; lateral tergal arms with outer blades expanded into weak spatulae, in some cases these only a little wider than the arms. Outer dististyle a little dilated on basal portion, the apex a short spinous point. Inner dististyle with the blade deep, the rostrum THE CRANE FLy GENUS DOLICHOPEZA §41 long-produced, its apex weakly bidentate. Aedeagus simple, un- armed.” Types.—Holotype male, St. Paul (Ramsey County), Minne- sota, 14 September 1940, H. D. Pratt. Five male paratypes with same data. All are in the collection of Dr. C. P. Alexander at Amherst, Massachusetts. Diagnostic characteristics —This subspecies may be recognized as belonging to the polita complex by the structure of the male hypopygium and, in the female, by the coloration, specifically the weak or absent abdominal annulations and the low contrast be- tween stigmal spot and ground color of the wing. From the typical race, it differs in lacking the polished appearance in both sexes; furthermore the tergal arms are less densely sclerotized and the basal portion of the outer dististyles less expanded and darkened than in polita polita. The configuration of the tergal arms will distinguish polita pratti from polita cornuta, for while these struc- tures are widely flared at about mid-length and narrowed at the tip in polita cornuta they are nearly always widest at the tip in polita pratti. The acute apex and expanded base of the outer dististyles will at once separate this fly from tridenticulata, the only species with which it might be confused on the basis of other characters of the male hypopygium; polita pratti is also noticeably the larger of these two in the areas where they are found together. Descriptive comments.—Although the abdominal segments are darker at the edges, as described (and this applies to males, very rarely to females), the annulation as in the other races of polita is indistinct, never as in other species of the obscura group. The brownish color of the thoracic dorsum seems to contain a trace of red or yellow, the prescutum more nearly resembling that of tri- denticulata than of obscura, which is brown. The tinge of the wings of polita pratti is grayish brown, fainter than the smoky brown of the wing of obscura. Preservation of the vein Sc, mentioned in the original description is an individual variation. Most specimens seen lack this vein and have wing venation characteristic of their subgenus. Other vena- tional variations observed include loss of the medial cross-vein, the vein R, incomplete apically, minor displacements and interrup- tions of the branches of the media, and a very short cell M, with correspondingly lengthened petiole, M, ,.. Body length of males is about 8 to 11 mm., and the wing length varies from 10 to somewhat over 13 mm. in the specimens seen. §42 THe UNIverSITY SCIENCE BULLETIN Females are slightly smaller than males, on the average (this is the case in more than a dozen mating pairs examined ), measuring 9 to 11 mm. in body length and 9.5 to 12 mm. in length of wing. The smallest measurements recorded were those of the types, which, it will be noted, were taken in September. Late August indi- viduals collected in Kansas were smaller than the June and July specimens but larger than the types. Variation in the shape of the medio-posterior margin of the ninth tergum of the male is extreme, exceeding even that found in the typical race. In this characteristic, polita pratti is much like polita polita, although except for the fact that the tergal arms are widest at the tip this subspecies otherwise more nearly resembles polita cornuta in hypopygial characters. Among the six types (all males) there are no less than four distinct shapes of the toothed medio-posterior margin of the ninth tergum. In Trempealeau County, Wisconsin, I collected a group including forty males, among which there were twelve major shapes of this structure, some ex- tremes of which are shown in Figures 184-186. Within this same group of flies were found six kinds of tergal arms and three main classes of outer dististyles, the last on the basis of the degree of pro- trusion of the bulbous basal enlargement, which sometimes ap- proaches the condition found in polita polita (compare with Fig. 182). Assigning letters to each different shape of these three struc- tures, I set down a formula describing each of the forty flies. There were only six duplications in the resulting three-letter formulas; that is, there were thirty-four combinations on the basis of variation in these three characters only! Geographical distribution —Dolichopeza polita pratti is known from western Wisconsin, southern Minnesota, Iowa, northwestern Illinois, the Ozark-Ouachita Mountains region of Missouri and Arkansas, and eastern Kansas. Specimens regarded as intergrades with polita cornuta were found in a large collection including both races from Starved Rock State Park, La Salle County, Illinois. It is interesting that the transition from polita polita in Juneau County, Wisconsin, to polita pratti in nearby Monroe County should be so abrupt. The two localities represented by these collections are only 36 miles apart, and I sampled both habitats on the same day. It seems likely that northeastern Minnesota is within the range of the species polita, but of which subspecies I am not certain. I am confident that polita pratti will be found in eastern Oklahoma, and it may reach western Iowa and parts of Nebraska and South THe CrANE FLY GENus DOLICHOPEZA 843 SCALE, FIG. 187 SCALE, FIGS. 180-186 Fics. 180-187. Dolichopeza (Oropeza) polita pratti; 180—ninth ter- gum of male, 181—left inner dististyle of male, dorsal aspect, 182—left outer dististyle, dorsal aspect, 183—vesica, penis, gonapophysis and ad- miniculum, left lateral aspect, 184-186—variations in medio-posterior margin of ninth tergum of male, 187—terminal abdominal segments of female, left lateral aspect. 844 THE UNIVERSITY SCIENCE BULLETIN Dakota along the Missouri River. Surely also it occurs much more widely in the states from which it is already recorded. Habitats—All the known habitats are characterized by out- cropping rock or large blocks of broken rock, in nearly all instances along streams, although the particular crannies in which the flies were taken were often well removed from the water. In fact, these cavities were in some cases apparently dry, but they were in- variably deeply shaded so that any available moisture in the air would be conserved. Some of the flies were found beneath cul- verts, in cavities under earthen banks and in various other types of niches, where these occurred within the general rocky ravine or rocky hillside habitat. Seasonal distribution —Little is known about the seasonal oc- currence of Dolichopeza in the western parts of the range of the genus, for too few collections have been made. In two successive years I obtained fair numbers of polita pratti in the upper Mississippi valley region in early July, and it was found numerous in central Iowa during the latter part of June by Dr. Jean Laffoon. Therefore, the capture of the type series in the same general area in September strongly suggests the usual two generations per year cycle. Dr. Rogers obtained a few specimens of this form in Taney County, Missouri, in mid-June, and I found them in the Ozark-Ouachita region in small numbers at the end of July and in Kansas late in August. The collection of both sexes in fair numbers in Kansas on 30 August indicated a period of emergence well underway. Ac- cordingly, there would appear to be two annual generations of polita pratti in the southern part of its range, as well. Immature stages —Eggs taken from preserved females show the very short terminal filament characteristic of the species polita, the filament only slightly curved in this race. Larvae of this form have not been seen but probably closely resemble those of the other sub- species. Pupal skins were found projecting from a thin mat of the moss Tetraphis pellucida and a powdery lichen on the underside of a sandstone outcrop. Protected from rainfall and from seepage, this larval-pupal habitat was rather dry at the time of collection of the pupal skins. Of 17 skins recovered, four females had the middle spinous processes of the eighth abdominal sternum slightly separated at the base, that is, not divergent from a common base, and’ one male lacked one of these projections. Only one specimen had any spinous processes on the fourth abdominal sternum, and this was on one side THE CRANE FiLy GENUS DOLICHOPEZA 845 only. Among 11 female skins and 6 of males, there was uniformity of shape of the spiracular yoke, as Figure 117. Notes on distribution—Arkansas—Garland County, 31 July; Washington County, 30 July. I~trvors—Carroll County, 7 July; La Salle County, 7 July. Iowa—Boone County, 21 to 25 June; Jack- son County, 8 July; Webster County, 30 June. Kansas—Douglas County, 30 August. Mrinnesora—Blue Earth County, 8 July; Ram- sey County, 14 September; Winona County, 7 July. Mu1ssourr— Barry County, 29 July; Carter County, 6 June; Taney County, 10 to 20 June. Wisconstn—Monroe County, 6 July; Trempealeau County, 7 July. Dolichopeza (Oropeza) polita cornuta, new subspecies Literature references.—As Dolichopeza (Oropeza) polita (John- son). Foote, 1956: 221. The name cornuta is selected to describe the lateral arms of the ninth tergum, which, being widened near the base and narrowed at the tip, suggest horns. Description —Dolichopeza (Oropeza) polita cornuta is olive- brown in general coloration and, depending upon preservation, may appear light to rather dark olive-brown, although not as dark as polita polita. Thoracic dorsum, especially the prescutum, more reddish brown, opaque and usually dull. Abdomen faintly annu- lated with fuscous. Occiput of head, the maxillary palps, and flagellar portion of the antennae gray-brown to dark gray. Scape yellowish brown and pedicel straw yellow. Legs brown. Wings tinged with pale grayish brown, the stigmal spot only slightly darker than the ground color. Lateral arms of ninth tergum of male flattened and expanded at about their mid-length but narrowed apically to blunt, horn-like tips. From the median lobe of the ninth tergum, three slender, acute teeth, the center tooth slightly the longest, project caudad, nearly parallel. Bases of the outer disti- styles are expanded, forming a bulbous enlargement on the mesal surface of each, the enlargement abruptly paler than the rest of the dististyle. Tenth tergum and most of hypovalve of female ovipositor about equally heavily sclerotized, the cerci and tips of the hypo- valves paler by comparison. Body length of holotype male 10.5 mm.; wing 12.5 mm. Allotype length 12 mm.; wing 13.5 mm. Flies of the late summer emergence may be smaller, as indicated below. Types.—Holotype male, Turkey Run State Park, Parke County (Field Catalogue Number 6), Indiana, 20 June 1950, G. W. Byers. 846 Tue UNIveERSITY SCIENCE BULLETIN Allotype, same data as holotype. Paratypes, forty males and ten females from the type locality, collected during the spring period of emergence. The holotype, allotype and most of the paratypes are in the University of Michigan Museum of Zoology, Ann Arbor, Michigan. Two paratypes have been sent to the United States National Museum, Washington, D. C. Because I have hundreds of specimens from several localities and in varying states of preser- vation, it seems reasonable to limit the paratypic series in some way; hence the selection of fifty topotypes in good condition and from the same seasonal generation as the holotype. Descriptive comments.—Venational abnormalities are common in this race, although most of these are limited to the medial field. Loss of the medial cross-vein is not unusual, and I have seen many different patterns of fragmentation of the branches of the media. Rarely, wings having a well-developed Sc, or R,,, have been found. In one male, the m-cu cross-vein has shifted so as to inter- sect the media at its first division. As in the other subspecies of polita, females of polita cornuta are usually slightly smaller than males of the same generation in the same vicinity. Including both spring and late summer genera- tions, measurements of males range from 9 to 11 mm., wing 11.2 to 13.8 mm.; females, from 9 to 12 mm., wing 9.5 to 13.5 mm. In male hypopygial characteristics, polita cornuta is the most uniform of the three races. There is some variation in the length of the teeth on the ninth tergum, and the central tooth may be more or less depressed below the plane of the other two. The expanded portion of the outer dististyle is somewhat less developed than in the typical race but usually exceeds that of polita pratti. It is never darkly colored, as in polita polita. Often the mesal surface of the outer dististyle just beyond the bulbous basal en- largement lacks hairs, a condition that has also been seen in the other subspecies. The shape of the tergal arms, which most readily identifies this subspecies, is extremely constant, although the width of the flared portion may vary slightly. The antennae are short, as in the other two races. Geographical distribution—Dolichopeza polita cornuta ranges from the upper “thumb” region of the Lower Peninsula of Michi- gan and from western New York state southwestward to eastern Illinois, southern Kentucky, and the Appalachian foothills of east- ern Tennessee and western slopes of the Appalachians in Virginia and West Virginia. This area is rather well surrounded by the THE CRANE FLY GENUS DOLICHOPEZA SCALE, FIG.I94 oO 0.5 LOmm. SCALE, FIGS. 188-193 Fics. 188-194. Dolichopeza (Oropeza) polita cornuta new subspe- cies; 188—ninth tergum of male, 189—left inner dististyle of male, dorsal aspect, 190—left outer dististyle, dorsal aspect, 191—gona- pophyses, dorsal aspect, 192—vesica and penis, 193—vesica and its apodemes, dorsal aspect, 194—terminal abdominal segments of female, left lateral aspect. 847 848 Tue University SCIENCE BULLETIN ranges of polita polita and polita pratti and by absence of apparently suitable habitats to the soutwest. I therefore believe that the range as indicated on the map is very close to the actual range in nature. The relationship of polita cornuta to the eastern sub- species is confusing at several localities, and it will be interesting to see how the ranges of the two will meet or overlap in such places as western Pennsylvania, where further collecting must be done. I have seen a specimen from Luzerne County, Pennsylvania, that is in many ways like polita cornuta, but because there are also certain dissimilarities and because the specimen is only one from a very large sample I have regarded it as an atypical indi- vidual and have not indicated it on the map. Habitats——Localities from which polita cornuta is known are characterized by rock outcrops or rock-walled ravines, providing deep shade for the adults and retaining moisture sufficient to grow mats or clumps of mosses for the immature stages. Adults are some- times found in great numbers in darkened recesses beneath project- ing ledges of rock. Where the space between outcropping rock and the ground below is great—say, more than ten feet, this and other species of Dolichopeza occurring with it are not likely to be found. But when the sheltering rock is close above the ground, providing deeper shade and protecting the flies from desiccation, they are likely to congregate. In places where I have most often collected this fly, it only rarely ventures up out of the cool ravines, although it sometimes may be found in parts of the ravines where no mosses could be seen. Seasonal distribution—There is abundant evidence of a two-gen- eration annual cycle. In Indiana, where most intensive studies of this subspecies were carried on, the start of the spring period of emergence of adults closely follows that of americana, the peak being reached about the end of May. In 1953, for example, a par- ticular cranny was visited every day so that comparative abundance of the various species could be estimated, although occurrence was noted throughout the area. First individuals of polita cornuta ap- peared on 23 May, when ten flies were counted. This was about a week after the onset of emergence of americana. Thereafter. population build-up was rapid, and by 28 May I estimated num- bers in excess of 150 flies of this form. On 30 May, peak numbers were present—probably well over 200 individuals of polita cornuta in the one small cranny mentioned. A single sweep of the collect- ing net yielded 61 males and 52 females, as well as 29 flies of other THE CRANE FLY GENUS DOLICHOPEZA 849 species of Dolichopeza, on 1 June; by 9 June only 15 polita cornuta could be found at this site, and a few individuals of the spring gen- eration were still present on 11 July. The late summer peak, in Indiana, falls in the first two weeks of August, polita cornuta again appearing shortly after americana and before any other species of Oropeza. On 3 August, great numbers of these flies were collected in Jefferson and Jennings counties, in southern Indiana; earlier col- lecting in that vicinity clearly showed these to belong to a second annual generation. Immature stages —Eggs of polita cornuta are about the same size as those of polita polita, being .75 by .30 mm., on the average, depending on the size of the female fly. There is a very short termi- nal filament, which projects slightly and then undergoes about one complete, circular coil; the filament is thus a little longer and more coiled than that in either of the other subspecies. Eggs from females of the spring generation were placed in a rearing dish on 24 June, and second generation adults appeared on 17 August. The larvae are ordinarily green in color by reason of the lack of any dense coating of microscopic hairs to obscure the pigment in the body contents. Most of the minute hairs on the body are extremely short and occur singly, a characteristic found elsewhere only in tridenticulata and the other races of polita. On the eighth abdominal segment of the larva, the dorsal microscopic hairs are abruptly longer than any others on the body, which gives a some- what darker appearance to the hind part of the larva and a sharp contrast to the paler, bright greenish spiracular disc. Fourth instar larvae may attain a length of nearly 18 mm., just before pupation. The pupa is characterized by the convoluted tracheal connection between the thoracic respiratory horn and the mesothoracic spiracle of the developing adult within; also by the spiracular yoke, which may sometimes lack the small apical projections, leaving the lobes broadly rounded or irregularly truncate. Habitats of the immature stages include the mosses Atrichum macmillani, Leucobryum glaucum, Dicranella heteromalla, and Mnium punctatum and the hepatic Calypogeia trichomanis, all growing on sandstone cliffs. The Leucobryum was only about half an inch deep and was more moist than it ordinairly is in its more usual hilltop environment, where it grows in thick cushions; and the Dicranella was less compact than usual, probably a new growth. The commonest larval habitat of polita cornuta is the thin coating of Tetraphis pellucida moss mixed with greenish white, powdery §50 THE UNIVERSITY SCIENCE BULLETIN lichen that grows on the sheltered lower surface of outcropping rock ledges. Since water from above falls from the brink of the outcropping ledge and seepage resumes only at the contact with the ground or other rock below, this habitat is relatively dry, even at times of heavy rainfall. In contrast to this common habitat is the wet, marl-forming moss, Gymnostomum calcareum, from which polita cornuta and johnsonella were reared. Notes on distribution.—ILLinois—La Salle County, 7 July (includ- ing a few intergrades with ssp. pratti); Pope County, 15 July; Ver- milion County, 13 June. INpraNna—Jefferson County, 8-9 June and 3 August; Jennings County, 3 August; Montgomery County, 28 June; Owen County, 22 June, 3 and 28 August; Parke County, 23 May to 3 June, 9 to 28 June, 10 to 15 July, and 11 to 30 August. Kentucky—Barren County, 2 August; Edmonson County, June (?); Letcher County, 3 July (includes intergrades with ssp. polita); Trimble County, 3 August. MrcarcAn—Huron County, 22 July. New Yorx—Cattaraugus County, 31 July. Osnto—Delaware County, 14 June; Geauga County, 17 July; Hocking County, 20 May to 7 June (includes two intergrades with ssp. polita among 101 speci- mens); Medina County, 4 July; Portage County, 24-25 June and 14 July (ssp. polita also here) and 16 August (ssp. polita not taken, this date); Summit County, 19 June. Onrarro—Niagara Glen, 30 June. TENNESSEE—Fentress County, 22 July to 14 August (includes intergrades with ssp. polita). Wircrs1a—Giles County, 21 June to 13 July and 21 August (includes intergrades with ssp. polita); Wise County, 2 July (includes intergrades with ssp. polita). Wrst Vir- cintAa—Pocahontas County, 5 July (includes intergrades with ssp. polita); Wetzel County, 5 August. Dolichopeza (Oropeza) sayi (Johnson) Literature references—Tipula annulata Say. Say, 1823: 25-26. Dolichopeza annulata (Say). Osten Sacken, 1878: 40. Oropeza annulata (Say). Needham, 1908: 210-211, pl. 16 (wing, mistakenly labelled “Oropeza annularis” ); Pierre, 1926: 11. Oropeza sayi Johnson (new name for Tipula annulata Say, pre- occupied by T. annulata Linnaeus). Johnson, 1909: 118-119, pl. 15 (hypopygium and wing, possibly of walleyi Alexander, due to mixing of the two forms, as indicated below); Johnson, 1910: 708 (walleyi?); Alexander, 1919: 930; Alexander and McAtee, 1920: 393 (at least in part walleyi); Johnson, 1925: 32; Leonard, 1928: 698; Rogers, 1930: 23 (walleyi); Rogers, 1933: 35 and 49 (walleyi). THE CRANE FLy GENUS DOLICHOPEZA 851 Dolichopeza (Oropeza) sayi (Johnson). Alexander, 1942: 214, fig. 26F (hypopygium ); Rogers, 1942: 59; Rogers, 1949: 12; Foote, 1956: 221. Original description—*“A dark brown stigma; abdomen pale, annulate with black. Inhabits Pennsylvania. Antennae fuscous, first and second joints whitish; rostrum, and lower portion of the front whitish; vertex and occiput dusky; palpi fuscous; thorax yel- lowish-brown, the indented lines paler; metathorax light livid; wings with a brown stigmata, nervures brown, arranged like those of Meigen’s fig. 9, pl. 6; feet dusky-brownish; abdomen _yellowish- white, incisures and their margins black, forming annulations com- plete. Length two-fifths of an inch.” Types.—No type specimen or series of syntypes has ever been specifically designated for this species, so far as I have been able to determine. If Say had any types, these are assumed to have been destroyed, as explained in the historical review. Needham did not clearly fix the identity of the species when he made Tipula annulata the type species of Oropeza in 1908. His illustration of the wing could apply to any species of Oropeza, and since his speci- mens, collected near Old Forge, New York, have been lost, it is impossible to say with assurance what species he actually had. In the collection of the Museum of Comparative Zoology, beside a label reading “O. sayi—Johnson (T. annulata Say),” there are some of the twenty-five specimens that Johnson had before him when he renamed the species as Oropeza sayi in 1909. These are two males and two females of Dolichopeza walleyi. The males are those recorded (Johnson, 1909: 119) from Niagara Falls, New York, 23 June, and the females are those from Hanover, New Hamp- shire, 6 July 1908, and Acquia Creek, Virginia, 24 May 1896. In addition to these four, there is one male (Montpelier, Vermont, 25 June 1906) that was sent to Dr. Alexander, in trade. This specimen is the form now known as sayi, and it was by comparison with it that Alexander distinguished his new species, walleyi. The twenty specimens of Johnson’s series not accounted for here were not avail- able to me for study; possibly they are in storage in a municipal museum in Boston or in Boston University. There is little in the original description of this species that is helpful in establishing its identity. Say’s wording is so general as to be equally applicable to several species of Tipula or Nephrotoma that occur in Pennsylvania, as well as to certain species of Doli- chopeza. His reference to “ Meigen’s fig. 9, pl. 6” pertains 852 THE UNIversIty SCIENCE BULLETIN to a figure in the first volume of J. W. Meigen’s Systematische Beschreibung der bekannten Europdischen zweifliigeligen Insekten (Meigen, 1818). This figure clearly represents a species of Nephro- toma and had been sent to Meigen under the name Nephrotoma imperialis; however, he placed it in Tipula. As pointed out in the historical review, Wiedemann identified specimens of Dolichopeza johnsonella and of Tipula as Say’s Tipula annulata, and Osten Sacken had in his collection certain specimens he regarded as be- longing to this species. Actually, Osten Sacken’s specimens are one male of Dolichopeza tridenticulata, one male of walleyi, one female of similis (also pin-labelled “Pachyrhine?” ) and four females of the obscura group. Of these, only the male walleyi and the female similis fit at all Say’s description. This chaos of poor descriptions and lost specimens, of mixed and misidentified species, seems to me a proper occasion for designation of a neotype. Accordingly, I have selected a male specimen from New York as neotype and a female from Pennsylvania as neallotype. (At the time of this selection, I had not seen the three males of sayi from Luzerne County, Pennsylvania, which are in the Dietz collection in the Academy of Natural Sciences, Philadelphia; how- ever, the neotype chosen is in better condition than any of these. ) The neotype agrees with the species Dolichopeza sayi as distin- guished from Dolichopeza walleyi by Alexander (1942: 214, fig. 26F) and as illustrated in my Figures 195 through 199. The neallotype agrees with my Figure 200, and both sexes conform to the descriptive comments given below. Label data on these types are as follows: Neotype male, Ulster County (Field Catalogue Number 1), New York, 28 June 1953, G. W. Byers. Neallotype, Berks County (Field Catalogue Number 1), Pennsylvania, 15 August 1956, G. W. Byers. Locality for the neotype is Esopus Creek, 1.3 miles west of Shan- daken; the specimen is in the collection of the University of Michigan Museum of Zoology. The neallotype was taken in French Creek State Park, about 11 miles southeast of Reading; the specimen is in the Snow Entomological Museum, University of Kansas. Diagnostic characteristics—In the field, where sayi often occurs together with dorsalis, it may be distinguished from the latter by its dark pleural markings and the much darker stigmal spot of its wing. Dolichopeza sayi most closely resembles walleyi, but the entire margin of the gonapophyses in sayi (Fig. 198) will readily separate males of this species from walleyi, in which the edges of THE CRANE FLY GENUS DOLICHOPEZA 853 IC 19:9 1LOmm. [ee ar 200 SCALE, FIG. 200 oO 0.5 1.0mm. ee Sg pet er Ee ee a ee ee) SCALE, FIGS. 195-199 Fics. 195-200. Dolichopeza (Oropeza) sayi; 195—ninth tergum of male, 196—left inner dististyle of male, dorsal aspect, 197—left outer dististyle, dorsal aspect, 198—gonapophyses, dorsal aspect, 199—vesica and penis, 200—terminal abdominal segments of female, left lateral aspect. the gonapophyses are irregularly toothed. Furthermore, in the areas where these two species occur together, or where their ranges over- lap, sayi may be recognized by the deep gray pleural spots on the mesothoracic anepisternum, ventral portion of pre-episternum and meron, for in this part of its range walleyi has the pleural surfaces of the thorax pale, almost as in dorsalis. Recognition of females is difficult, but they are separable on the basis of the color markings stated, when attention is given to their geographic locality. A fur- ther point of coloration useful is distinguishing sayi from walleyi is that the prescutal markings of the former are dark grayish brown (often approaching black in recently collected specimens), while those of walleyi are ordinarily a lighter reddish brown. Females of 854 Tue UNIvERsITY SCIENCE BULLETIN sayi also may be easily confused with those of similis. In the latter, however, the abdominal annulations are brownish, are of rather uniform breadth around the body, and are about as wide as the yellowish spaces between annulations. In sayi, the annulations are more grayish brown, are broadest on the dorsum and taper laterally, and are often narrow or indistinct. Wing length of females of these two species overlaps in the range of 12.0 to 13.6 mm., a measurement of less than 12.0 mm. indicating the specimen is sayi (when other descriptive details fit, of course), and a measurement of more than 13.6 mm. indicating similis. Descriptive comments.—This species may be briefly redescribed as follows: Dolichopeza sayi is in general brownish yellow, marked with dark grayish brown. Occiput gray, rostrum and frons below anten- nal bases yellowish; scape, pedicel and basal part of first flagellar segment yellowish, the flagellum grayish brown. Prescutum brown- ish yellow, marked with three longitudinal stripes of very dark grayish brown, the middle one extending farthest cephalad; scutum with two elongate, dark spots, divergent anteriorly, sometimes each divided into two; scutellum somewhat paler than prescutal ground color; mesothoracic meron and ventral part of the pre-episternum nearly as dark as prescutal stripes, anepisternum a little less dark- ened, the remaining thoracic pleural areas paler by contrast. Wings with a grayish tinge in living flies, the color becoming grayish amber in older specimens; stigmal spot very dark; a narrow, conspicuous band of grayish brown along the cubitus; halteres dusky. Legs dark grayish brown paling to light brown on the tarsi; coxae pale. Ab- domen brownish yellow, often with the dark annulations widened dorsally, resulting in a continuous darkened stripe along the dorsal mid-line, especially in males. Annulations tapering laterally, espe- cially on anterior segments, often incomplete and in females some- times difficult to discern. Gonapophyses of male thick at base, flattened and slightly widened apically, the inner angle upturned, dorsal surface heavily sclerotized and a little concave, ventral sur- face convex, pale and sparsely pubescent (compare Figs. 7 and 198). Outer dististyles (Fig. 197) yellowish to pale yellowish brown throughout, rarely darker. Inner dististyles (Fig. 196) with an elongate, glabrous, narrowed apical portion. Ninth tergum in- tensely sclerotized medially, with a small central tooth flanked by two blackened lobes; tergal arms short, their distal ends only slightly expanded (Fig. 195). Female ovipositor narrowly darkened near tips of hypovalves but this area not intensely sclerotized (Fig. 200). THE CRANE Fiy GENUS DOLICHOPEZA 855 Body length of males varies from § to slightly over 10 mm. Wings of males are nearly always of greater length than the over-all body measurement, varying from about 10 to 12.5 mm. In females, on the other hand, wing length is often less even than the length of the abdomen, hence much less than over-all body length. Wing lengths of females measuring from 10 to 14 mm. ranged from 9.5 to 13 mm. Maximum expected length of wing computed from measurements of 77 females is 13.6 mm. Venation is very constant. I have found a few instances of par- tial loss the the medial cross-vein, one example of a partial fusion of the subcosta with the costa, and one specimen in which the radial sector was of unusual length, suggesting that of some species of Tipula. In coloration, there may be an over-all difference in intensity among, specimens, so that one will appear paler in all respects than another taken at the same locality on the same date. Apparent variation in the shape of the gonapophyses of the male may result from their partial rotation, but a few specimens have been seen in which their latero-caudal corners were rounded. The outer dististyles, though usually yellowish, occasionally have a light grayish or brownish color. A great amount of variation has been noted in the shape of the tips of the tergal arms, some suggesting those of similis, walleyi or johnsonella. As in some other species Map 7. Range of Dolichopeza (Oropeza) sayi (Johnson). Each spot teeth one or more collections within a county (United States) or at a ocality. 856 THe UNIVERSITY SCIENCE BULLETIN of Dolichopeza, it is not uncommon to find one tergal arm quite unlike the other, on one fly. Geographical distribution—Dolichopeza sayi is known from New Brunswick westward to northern Minnesota and southwest- ward to Virginia. Although it is essentially a northern species, in the character of its distribution, I would expect it to occur much more widely in Pennsylvania, Ohio and West Virginia than present records indicate. Suitable habitats for this species are also thought to occur in Canada much farther northward and westward than the now known range. Many distributional records for sayi published prior to 1940 will be found to apply to walleyi. Some of these are indicated in the literature references. In other cases, the specimens have not been seen by me. Habitats—Habitats in which I have collected this species are all in one way or another marshy or swampy, although in general appearance they may be quite diverse. Low, shaded vegetation of marsh borders, shrubby margins of lakes and ponds, and swampy woods are typical habitats. In the eastern part of the range, small areas of marshy terrain isolated among otherwise rocky, mesic forest lands have proved to be habitats of sayi. The occurrence of sayi in the rock gorge type of general environment, as described earlier under the natural history of adult flies, was found to be conditioned by the same microenvironmental features as those of a marsh bor- der but present on much smaller scale. Shade sufficient to hold the temperature in the lower vegetation several degrees below that generally prevailing in the sunny parts of open marsh seems to be characteristic of sayi habitats. Therefore, where a pond is sur- rounded only by a zone of cattail (Typha), grasses and Carex, sayi is not likely to be found. Foote (1956: 221) reports collection of two males of sayi in “moist upland woods” in Delaware County, Ohio. I have not seen these specimens; the described habitat seems more appropriate for walleyi. Seasonal distribution—Where there has been enough collecting done within a limited area to give any accurate indication of the extent of the flight periods of this species, there is strong evidence of two well-marked generations per year. Rogers (1942: 59) found, in a three-year study in southern Michigan, that peaks of emergence for sayi came in June and August, with individuals present in May,* July and September in smaller numbers. I have * I have been unable to verify the 1 May date given by Dr. Rogers, either by specimens or his field notes. It is not likely that sayi was on the wing at such an early date. Earliest established records for this area are in the last week of May. Tue CRANE Fry GENuS DOLICHOPEZA 857 repeated this observation in three southern Michigan localities. Evidence from the northern portion of the range suggests that only one main period of emergence, in July, is the rule. The life cycle, indicating seasons of emergence of adults, is presented in Figure 99. Immature stages—Eggs of an average-sized female of Dolicho- peza sayi from southern Michigan measured .72 by .31 mm. and showed a well-developed terminal filament. The egg stage is seven days’ duration, in the laboratory. Of a large number of larvae hatched, nearly all reached the first molt within two weeks, and thereafter the molts came at about one-week intervals, until the fourth instar was reached. The last instar larva very closely re- sembles that of Dolichopeza walleyi in having the ridges of micro- scopic hairs most pronounced on the thoracic, first, seventh and eighth abdominal segments, while the ridges are fainter on the intervening segments. Larvae just before pupation may exceed 16 mm. in length. In the laboratory, pupation did not always take place within a tube, and the six- to seven-day pupal period was, in such cases, spent with the pupa lying on its side in the bottom of the dish. At the time of emergence of the adult, however, these pupae invariably maneuvered themselves into an upright position. This was probably to prevent contact of the adult with water, for although the immature stages of sayi live in a semi-aquatic environ- ment the adults are easily trapped in water and drowned. The pupa is structurally like that of walleyi except for details of the spiracular yoke and the absence of projections from the posterior ring of the fourth abdominal segment. I have found the immature stages of sayi in wet mosses growing flat on the muddy soil in or adjacent to swamps and marsh borders. These mosses are Amblystegium varium, Brachythecium salebro- sum, Eurhynchium pulchellum and Hypnum lindbergii. Larvae were also found in Heterophyllium haldanianum moss on a sodden, decayed log in a swamp, and in the moss Didymodon tophaceous, where it was growing on a rock at the bank of a brook, just below a swampy, shaded area that drained into the brook. In all cases, the larval and pupal habitat was found to be quite wet and always well shaded. Notes on distribution—Connecticut—Litchfield County, 12-13 June; New Haven County (?), no date; Windham County, 14 June. Maine—Cumberland County, 1 July; Hancock County, 12 July; Piscataquis County, 17 July. MaryLanp—District of Columbia (?), 29—5840 858 THE UNIVERSITY SCIENCE BULLETIN 29 August. Massacnusetrs—Berkshire County, 19 June; Dukes County, 17 July; Hampden County, 14 July; Norfolk County, 11 June and 25 August. Micuican—Antrim County, 2 July; Cheboy- gan County, 2-3 July; Eaton County, 30 August; Lake County, 26-27 June; Livingston County, 24 May to 9 July and 28 July to 3 September; Midland County, 10 June; Montmorency County, 16 July; Newaygo County, 21 June; Oscoda County, 14 to 26 June; Otsego County, 3 July; St. Joseph County, 30 May and 12 August; Schoolcraft County, July; Washtenaw County, 5 to 17 June and 13 to 19 August. Mrxnesora—Clearwater County, 11 July. NEw Brunswick—Oromocto, 9 July. NEw Hampsurre—Coos County, no date; Grafton County, 6 July. NEw Jersey—Camden County (?), 6 June; Gloucester County, 6 June; Morris County, July. New York—Erie County (?), 27 August; Herkimer County (?), August; Tompkins County (?), 6 July; Ulster County, 28 June. Onto— Delaware County (?), 13 June. Onrarro—Algonquin Park, 3 July and October; Gull Lake, Muskoka District, June; Simcoe, 9 June. PENNsYLvANIA—Berks County, 15 August; Luzerne County, 28 June and 8 July. VerMont—Caledonia County, July; Chittenden County, 24 June; Washington County, 25 June; Windham County, July; Windsor County, 7 July. Vircrnta—Stafford County (locality uncertain ), 24 May; Russell County, 15 August. Note: County records are queried when there is reason to suspect that there may be confusion of this species with walleyi and speci- mens are not available for confirmation. Dolichopeza (Oropeza) similis (Johnson) Literature references——Oropeza similis Johnson. Johnson, 1909: 119, pl. 15 (hypopygium); Alexander, 1919: 930; Johnson, 1925: 32; Pierre, 1926: 12; Dickinson, 1932: 212, fig. 113 (in error; this pertains to walleyi). Dolichopeza (Oropeza) similis (Johnson). Alexander, 1936; 280; Alexander, 1942: 214-215, fig. 24B (wing), fig. 26G (hypo- pygium ); Rogers, 1942: 60, 121. Original description —“Head yellow, vertex brown, palpi yellow, antennae yellow becoming fuscous toward the tips. Thorax yellow- ish, with three wide and poorly defined black stripes covering the dorsum; scutellum, metanotum, and plurae light yellow, with livid spots between the coxae, on the plurae, and at the end of the metanotum. Abdomen yellow, with blackish rings at the margins of the segments. Genitalia brown, appendages yellow, style black THE CRANE FLY GENUS DOLICHOPEZA 859 and slightly forked, appendages at base of style irregular, curved, and hamate. Ventral margin broadly emarginate. Ovipositor yel- low. Halteres and legs yellow. Wings yellowish hyaline, veins and stigma dark brown. Length, male, 10 mm.; female, 13 mm.” Types.—Holotype male, Ricketts, North Mountain (Sullivan County), Pennsylvania, 8 June 1898, C. W. Johnson. Allotype, Auburndale (about 1.5 miles northwest of Newton, Middlesex County ), Massachusetts, 4 June , C. W. Johnson. Together with the holotype and allotype in the collection of the Museum of Comparative Zoology, Harvard University, there is one female paratype. A fourth specimen mentioned by Johnson (1909: 119) is not in this collection. Although the site of the village of Ricketts, on North Mountain, is still indicated on some recent maps of Pennsylvania, the village itself has long since disappeared. Once a thriving lumbering town of several hundreds population, Ricketts vanished with the removal of all the valuable timber from the sur- rounding hills. Development of second-growth forest is restoring much of the area to enough of its former condition to provide many good habitats for species of Dolichopeza. Diagnostic characteristics—Although this is the largest species of North American Dolichopeza, it is not sufficiently larger than the similar walleyi or some other species occurring in the same habitats to be recognized in the field on the basis of size alone. In its natural swampland habitat, similis may be recognized by the com- bination of its large size with the golden-brown tinged wings with a dark seam along the cubitus, the yellowish body coloration and the nearly complete fusion of the prescutal stripes. On close examination, male specimens are readily distinguishable by their unique gonapophyses (Fig. 204). In no other species is the dorsolateral angle of the gonapophysis produced into a thick, spine-like structure, projecting dorsocephalad. Also, the adminic- ulum is unlike that of the other species in that it bears a short spine on either side, at the apex. The female of similis is most likely to be confused with sayi and walleyi. Separation of these from sayi females is discussed in detail under that species. From walleyi females, these differ in having the tenth tergum more densely sclerotized than any other part of the ovipositor, the abdominal annulations of more uniform width throughout, and the prescutal stripes less distinct. Specimens of considerable age have some re- semblance to females of johnsonella because of the indistinctness of the prescutal markings, the amber tinge of the wings and the uni- 860 THE UNIVERSITY SCIENCE BULLETIN formly wide abdominal annulations; however, in similis females the pleural markings are more distinct, and the intense sclerotization of the hypovalve below the tenth tergum found in johnsonella is absent, although darkened areas somewhat like those in carolus females are occasionally seen (Fig. 207). Descriptive comments.——Concerning coloration, the thoracic stripes, described by Johnson (1909: 119) as black, are a dull, red- dish brown to brown, in specimens of a few years’ age, grayish brown in some more recent specimens. In general aspect, similis has a dark, tawny-yellow color, clearly annulated on the abdomen and contrastingly marked with prescutal stripes and pleural spots on the thorax. The wings are tinged with golden brown, and the stigmal spot is long oval in most specimens. Venation in this species is generally of the subgeneric pattern, but a few abnormalities have been seen. Presence of the veins Sc, and R,,, was noted in a few specimens, and a spurious cross-vein in cell R, was equally rarely seen. The cell M, is occasionally nearly sessile, rarely sessile. Slight variations in position or com- pleteness of the branches of the media are not uncommon. Sizes recorded earlier for similis (Alexander, 1942: 215; Johnson, 1909: 119) seem unduly small. These measurements represent similis as no larger than several other species of Oropeza, although it is in fact the largest, so far as I can compute from the few (67) specimens I have seen. Males range from about 10 to 12 mm. in over-all body length, their wings from slightly less than 12 to about 14.5 mm. Females vary in length of body from 11 to 14.5 mm., and in wing length from 12.7 to slightly less than 15 mm. Compu- tation of expected range of wing length (a range of three standard deviations either side of the mean of the sample) indicates the wings of females may vary from 12 to 15.6 mm. Being irregularly toothed at their margins, the gonapophyses show considerable variation, but they have been sufficiently uni- form in all specimens seen to identify the species. It should be noted that these structures in males of similis (as also in walleyi, which is apparently the most closely related species) may be bent backward and downward from their normal position, in such a way that their dorsal or inner surfaces are exposed and the strong spines point caudad. While the usual profile of the ninth tergum is as shown in Figure 201, the median tooth, never very prominent, may be lacking. The inner dististyle (Fig. 202) is rather narrow throughout, with an abruptly more slender apical portion, some- THE CRANE FLY GENUS DOLICHOPEZA 861 1.Omm. SCALE, FIG. 207 205 Oo 0.5 1LOmm. SE EE | SCALE, FIGS. 201-204, 206 oO 0.5mm. SCALE, FIG. 205 Fics. 201-207. Dolichopeza (Oropeza) similis; 201—ninth tergum of male, 202—left inner dististyle of male, dorsal aspect, 203—left outer dististyle, dorsal aspect, 204—gonapophyses and adminiculum, dorsal as- pect, 205—hypopygium of male, left lateral aspect, 206—vesica and penis, 207—terminal abdominal segments of female. left lateral aspect. 862 THe UNIVERSITY SCIENCE BULLETIN what suggesting the condition found in sayi. The basal portion of the penis is unusually enlarged (Fig. 206), more so than in subalbipes, the only other species in which this development has been found. This enlargement of the tube seems to be no more than a continuation of the lumen of the vesica; its particular func- tion is unknown. One specimen has been seen in which the ad- miniculum bears a short, blunt spine on its ventral surface, resem- bling subvenosa, but the spine less prominent. The shape of the distal ends of the tergal arms varies a little but is usually as Figure 201. Geographical distribution—The range as presented by Alex- ander (1942:215) nineteen years ago has not been materially extended. In fact, I cannot verify his published record for north- Mar 8. Range of Dolichopeza (Oropeza) similis (Johnson). Each spot represents one or more collections within a county (United States) or at a locality. ern Indiana, and Dickinson’s Wisconsin record (Dickinson, 1932: 113) is actually based on walleyi, so the range of similis is dimin- ished to this extent. On the basis of specimens I have examined, the range extends from Maine westward to the Lake of the Woods region (Kenora District) of Ontario and southwestward to Mary- land. There is a reliable record for Isle Royale, Michigan, in the Michigan collection, but the record from far western Ontario is based on a single female that I found dead in a spider web. This damaged female is darker than any other I have seen but keys to THe CRANE Fry GENUS DOLICHOPEZA S63 similis without question. It seems, from knowledge of the ecologi- cal distribution of this species, that it must range much further north in the eastern part of North America and westward across Manitoba, perhaps to northern Alberta. Although the species is northern in the character of its distribution, it probably ranges much farther south in the Appalachian Mountains than present collection records indicate. Habitats—I have collected similis only a few times, always in well-shaded situations where moisture is abundant, such as a hard- woods of red maple, yellow birch and American elm, in southern Michigan. This woods is flooded every spring, remains wet or shallowly flooded until early summer and dries by late summer or early fall. Adults of similis were taken among ferns and other low vegetation when the water still stood three or more inches deep in most of the woods. Rogers (1942: 60, 121) took similis in a very similar hardwood swamp and in a tamarack-sumac swamp. He records “. . . scores of individuals . . . on the ceilings of the dimly-lit, miniature cavelike recesses that are formed where the mossy roots or root-bound platforms project over the water or the semisuspended silt margins of the swamp pools. Such low, dank recesses were sometimes shared by obscura, sayi, or subal- bipes, but similis was the most abundant species here and was prac- tically confined to these spots.” Alexander (1936: 280) took similis along Tuckerman Trail on Mount Washington, an area much dif- ferent from those already described but in which similar micro- habitats may be found. Seasonal distribution—There is no evidence of more than one generation per year. In southern Michigan, Rogers (1942: 60, 121) made a careful study of an area where similis occurs, finding the flight period to be 26 May to 22 June, in the years 1936 through 1941. Late summer emergence was never found to occur. Throughout the species’ range, there are date records from late May to early August, but the preponderance of specimens were taken in June. In any event, it appears that locally the flight period is quite brief, which would possibly explain the dearth of specimens in collections. Immature stages—The eggs of Dolichopeza similis are the largest of any North American species of the genus, measuring on the average .92 by .35 mm. There is no terminal filament. The first instar larvae are proportionately large, and the fourth instar 864 THE UNIVERSITY SCIENCE BULLETIN larva is the largest Dolichopeza larva I have found, in North Amer- ica or abroad, reaching a length of 19 mm. Oddly enough, adults being hard to find, the larvae and pupae of similis were among the first immature stages of Dolichopeza encountered in this study. The larva resembles that of walleyi in having the minute micro- scopic hairs in short, definite, well-separated rows, and resembles sayi in the character of the dorsolateral lobes of the eighth ab- dominal segment. The transverse ridges of larger microscopic hairs are usually faintly indicated on the mid-region of the larva of similis but are a little more distinct in walleyi. The pupa may be recognized by the spiracular yoke, which is rather like that of subalbipes, and the bifid pleural spines (single-bristled in subal- bipes). The only known larval habitats of similis are the mosses Hetero- phyllium haldanianum, Plagiothecium denticulatum and Tetraphis pellucida. The Plagiothecium was growing on living trees, at their bases just above the water line, in a flooded hardwoods. The other mosses were growing on well decayed, wet stumps and fallen logs. All the habitats were well shaded. As they have a yellowish brown color, the large larvae are rather conspicuous in the green mosses. Notes on distribution—Connecticut—Hartford County, 12 June; Litchfield County, 13 June. Marme—Hancock County, 20 June. Mary_anp—Garrett County, 28 June. MassacHusETTs— Essex County, 12 June; Middlesex County, 4 June. Micaican— Keweenaw County (Isle Royale, Lake Superior), 5 August; Liv- ingston County, 26 May to 22 June; Oscoda County, 24 to 26 June; Washtenaw County, 5 to 22 June. New Hampsure—Cheshire County, 18 June; Coos County, no date; Grafton County, 2 to 8 July. New Yorx—Erie County, 7 to 14 June and 20 August; Rens- selaer County, 7 June. Onrarro—Algonquin Park, 3 to 23 June; Burke Falls, 9 July; Kenora District (Lake of the Woods), 4 Au- gust (specimen found dead). PENNsyLvaNni1a—Luzerne County, 11 June; Sullivan County, 8 June. Qursec—Rigaud, 25 June. Dolichopeza (Oropeza) subalbipes (Johnson ) Literature references—Oropeza subalbipes Johnson. Johnson, 1909: 121-122, pl. 15 (hypopygium ); Johnson, 1910: 708; Alexander, 1919: 930; Alexander and McAtee, 1920: 393; Dietz, 1921: 260; Johnson, 1925: 32; Pierre, 1926: 12; Alexander, 1928: 57; Leonard, 1928: 698; Rogers, 1930: 23; Rogers, 1933: 49. THE CRANE FLY GENUS DOLICHOPEZA 865. Dolichopeza (Oropeza) subalbipes (Johnson). Alexander, 1940: 620; Alexander, 1941a: 297; Alexander, 1942: 215, fig. 26H (hypo- pygium ); Rogers, 1942: 60, 121; Rogers, 1949: 12. Dolichopeza subalbipes (Johnson). Judd, 1958: 624-625. Original description—‘Similar to O. albipes, but readily sep- arated by the genitalia. Tibiae more or less yellowish, tarsi entirely white. Genitalia yellow, appendages moderately long, yellow at the base, fuscous toward the tip; style yellow, appendages at the base short and tipped with small black spines, ventral margin slightly emarginate. Length, male, 9 mm.; female, 13 mm.” Types.—Holotype male, Clementon (Camden County), New Jersey, 3 June 1897, C. W. Johnson. Allotype, same locality as holotype, 8 August 1897, C. W. Johnson. Two male paratypes, those listed by Johnson from Long Branch, New Jersey, and Auburn- dale, Massachusetts, are together with the holotype and allotype in the collection of the Museum of Comparative Zoology, Harvard University. One male paratype (Long Branch, New Jersey) is in the Academy of Natural Sciences of Philadelphia. I have not been able to account for the other two specimens that Johnson had: the ones from Westville and Riverton, New Jersey. Diagnostic characteristics—Because of the appearance of the tarsi, tibiae and femora and the general coloration of the body, subalbipes will probably not be confused in the field with any species other than carolus, which is not only similarly colored but which also has the same resting posture. Dolichopeza venosa and subvenosa somewhat resemble subalbipes in color but lack the darkened tips on the leg segments. In differentiating suwbalbipes and carolus in the field, habitat is a useful but not always reliable indicator, as discussed under carolus. Hypopygial structure identifies swbalbipes males, for this is the only species of the obscura group in which the median region of the ninth tergum is broadly extended caudad, except australis, in which there is a central, pointed tooth flanked by two lower rounded lobes. The coloration of swbalbipes is also unique in the obscura group. Usually, the tips of the tergal arms are bulbous and inflated (that is, hollow), a feature found in no other species of the genus. There are, however, specimens in which the tergal arms are slender, either of rather uniform thickness throughout or only slightly en- larged at the tips, often resembling those of johnsonella or aus- tralis. In johnsonella, the dorsal profile of the ninth tergum is 866 THe UNIverRSITY SCIENCE BULLETIN undulating (Fig. 157), while in subalbipes (Figs. 208, 209) and australis (Fig. 137) there is a definitely projecting, broad extension or lobe. Females of subalbipes may be distinguished from those of carolus by their lacking the darkly sclerotized spot, near the base of each hypovalve, characteristic of the latter species. In subalbipes, also, there is a greater contrast between the dark mesothoracic anepi- sternum and the pale dorsal portion of the pre-episternum than is found in carolus. Descriptive comments.—The close similarity of color detail be- tween subalbipes and carolus is covered in the discussion of the latter species. It might be added that both species have, in gen- eral, the same range of variation in coloration, except that there appears to me to be a tendency in northern populations of swbal- bipes to have the darker areas of the body ( prescutal stripes, pleural spots, abdominal annulations, etc.) relatively darker than the corresponding areas on individuals of carolus from the same locality. For example, where both species were collected together in Wis- consin, carolus was the paler in over-all appearance. Locally the incidence of venational variation is very high, and, as in other species of the obscura group especially, this variation is concentrated in the branches of the media. In a Florida population, several specimens had a short vein projecting into the cell M, from the medial cross-vein; and in a population from southern Georgia there was an unusual amount of fragmentation of the distal ends of the branches of the media. These and an extraordinary varia- tion in a specimen from Alabama are illustrated in Figures 49-51. Size variation in Dolichopeza subalbipes is not conspicuous through most of the range, an average male of the spring generation measuring about 9.5 mm. in body length, with an 11.5 mm. wing. Late summer males are ordinarily slightly smaller in all dimensions than those of the spring period of emergence. However, I have seen a few males from northern peninsular Florida, all collected between late October and early March, that are remarkably di- minutive, the smallest having a body length of only 7 mm. and a wing length of 7.5 mm. Another very small male, like these, is from southern Georgia but was taken on the tenth of June. These males all differ structurally from most subalbipes of the area, as described below. Largest males seen were from Michigan and from the Appalachian Mountains in Giles County, Virginia. These flies reach a body length of slightly over 10 mm., their wings around THE CRANE FLy GENuS DOLICHOPEZA 867 12 mm. Females attain greater lengths (11.2 to 15 mm.), and their wings are nearly always shorter than the over-all body length (10 to 12.5 mm.). The most outstanding variation noted in this species is that of the hypopygium of the male, particularly of the shape of the tergal te) 1.Omm. ——————— SCALE, FIG. 214 214 0.5 1.Omm. SCALE, FIGS. 208- 213 Fics. 208-214. Dolichopeza (Oropeza) subalbipes; 208—ninth ter- gum of male, showing expanded tergal arms, 209—ninth tergum of male with slender tergal arms, 210—left inner dististyle of male, dorsal aspect 211—left outer dististyle, dorsal aspect, 212—gonapophyses and ad. miniculum, dorsal aspect, 213—vesica and penis, 214—terminal abdomi- nal segments of female, left lateral aspect. 868 THE UNIVERSITY SCIENCE BULLETIN arm, as discussed earlier in the section on intraspecific variation. One form has tergal arms the tips of which are bulbous and knob- like (Fig. 208). The holotype of swbalbipes belongs to this form. In Florida, the type of tergal arm with bulbous or inflated tip is by far the most common, and all specimens having the slender-tipped tergal arm are the small, winter males mentioned above. Small samples of subalbipes from Clearwater County, Minnesota, Juneau County, Wisconsin, and Washtenaw County, Michigan, contain males having only the type of tergal arm that is of rather uniform width or thickness throughout (Fig. 209; and Alexander, 1942: 213, fig. 26H). A sample of 15 males from Livingston County, Michi- gan, contained seven with the inflated tip and eight with the slender tip, with an indication that the slender tip form is somewhat more abundant in the spring generation. In Cheboygan County, Michi- gan (northern Lower Peninsula), Giles County, Virginia, and Fen- tress County, Tennessee, areas from which large samples of subal- bipes are available, both types of tergal arm are found, the slender type commoner in the more northern localities. Very few speci- mens that could possibly be regarded as intergrades have been seen, an example being a fly from eastern Tennessee in which the tergal arms have inflated tips but are reduced in over-all size. Less conspicuous but much more variable in shape is the broad, projecting median lobe of the ninth tergum. This structure may bear three obtuse peaks or, less often, two; it may project only slightly or may extend well caudad from the adjacent margin of the tergum; and it may be relatively either wide or narrow. Two com- monly observed profiles of the ninth tergum are illustrated in Figures 208 and 209. In certain specimens there is, as in Dolichopeza similis, a basal enlargement of the penis, although in subalbipes this is of a smaller diameter and somewhat greater length (Fig. 213). This character has been examined in only a few specimens that were dissected or cleared and mounted on microscope slides, but among these per- haps two dozen males there was close correlation between the presence of the basal enlargement of the penis and the inflated tip of the tergal arm. In males having tergal arms with slender tips, the penis is thick at the base but not more than twice its di- ameter at mid-length and without the abrupt reduction in diameter shown in Figure 213. The descriptive data just presented lead one to question whether the nominal species subalbipes is not in reality two species. The two forms of males seem to exist together throughout the range, and THE CRANE FLY GENUS DOLICHOPEZA 869 I have so far been unable to distinguish any two corresponding types of females. Variation in the shape of the ninth tergum of males cuts across the line drawn on the basis of the tergal arm and the basal enlargement of the penis, as does variation in the shape of the inner dististyle (which is sometimes wider near its apex than illustrated in Figure 210). In view of these things, I regard the described differences between males as polymorphic forms within one species. Geographical distribution—Dolichopeza subalbipes ranges far- ther southwest than any other species, as far as is now known, one specimen having been collected in Natchitoches County, Louisiana. Map 9. Range of Dolichopeza (Oropeza) subalbipes (Johnson). Each spot represents one or more collections within a county (United States) or at a locality. I have not seen this fly, a female, which was reported to me as carolus, but I feel confident that it is not that species, and it seems unlikely that it would have been confused with any species other than subalbipes. Alexander (1942: 215) has also regarded this record as subalbipes. He reports the species from Indiana, which I am unable to verify by specimens or collection records, although I have no doubt it occurs there. The range of Dolichopeza sub- albipes extends from New Brunswick westward to northern Min- nesota and southward as far as central Florida and the Gulf Coastal Plain as far west as Louisiana. I believe there are suitable habitats in Kentucky, Indiana, Illinois, eastern Arkansas, western Tennessee 870 THe UNIVERSITY SCIENCE BULLETIN and eastern Texas. Also, the natural range probably extends much farther north and northwestward than records now available in- dicate. Habitats—Adults of Dolichopeza subalbipes have nearly always been taken in close association with low, leafy vegetation of swampy or marshy places, or along shaded streams. It was mentioned earlier that subalbipes had been collected in a rocky ravine, in Wisconsin. This was an unusual locality (Rocky Arbor State Park), however, in that while rock gorge habitats could be found along its sandstone walls, the flat floor, some 200 yards in width, was covered with a typical northern swamp. Only once, in Portage County, Ohio, have I taken this species in a rocky ravine type of gen- eral habitat where even localized swampy conditions were absent al- together. In Minnesota and northern Michigan, subalbipes was taken in cool, well-shaded, swampy lowlands; in southern Michigan, the habitats where I found this species were swampy hardwoods and the seasonally flooded hardwood zone around a bog. I have netted subalbipes among dense ferns and deep grasses along forest brooks, in the Appalachian Mountains from northern Georgia to Maine, and in Sphagnum bogs (where other mosses also were abundant) in West Virginia and Michigan. On the Cumberland Plateau of eastern Tennessee, Rogers (1930: 23) found subalbipes “more confined to the upland” than carolus and “taken from the stream-margin thickets, sphagnum-huckleberry bogs, and from the banks of both cleared and wooded brooks. A few records are from the talus slope brooks and from the river bank of Clear Fork, but it was not collected from the ‘rock houses’ or the base of the wet rim rock.” In Pennsylvania, Dietz (1921: 260) observed sub- albipes to be an inhabitant of “swampy places.” Northern Florida habitats of this species are described as “ beneath luxuriant ferns and herbage that overhang wooded rills and small brooks” (Rogers, 1933: 49). All habitats described by other authors are places in which damp mosses are likely to have been growing. Seasonal distribution—There is evidence of rather continuous emergence of adults, through most of the year, in the Florida region. North of Florida, however, the indication is that two annual gen- erations occur, except in the northernmost parts of the United States and in Canada, where collection records are so few as to render the situation obscure. There are far northern records for June through August, but most of these are for July and suggest a single, mid- summer generation for this region. Rogers (1942: 60, 121) found THE CRANE FLY GENUS DOLICHOPEZA 871 the adults of subalbipes on the wing in southern Michigan from 2 June to 14 July and from 14 to 28 August, the June generation the largest in numbers. He records the species as “abundant in the birch-maple swamp on June 5, and spread for 150 feet or more into the adjacent oak-hickory woods. . . . Apparently a huge ‘hatch’ had taken place on June 3, 4, or 5, for only a few teneral males could be found in the swamp on June 3.” In eastern Ten- nessee, another region in which suwbalbipes has been collected extensively enough to indicate its seasonal distribution, Rogers (1930: 23) found it in May to early July and again in August. Immature stages—Egegs of an average sized female of Dolicho- peza subalbipes from southern Michigan measured .83 by .36 mm., and those of a small female from Florida (one taken in February with the small males noted above) averaged only about .72 by .34 mm. In each case, the number of eggs matured appears to be around 100. There is no indication of a terminal filament. Larvae of the first two instars are not known, but the third instar very closely resembles the fourth. The fourth instar larva is one of the easiest to identify of all species of Oropeza. The larger micro- scopic hairs on the dorsum of all segments are longer than in any other species except venosa and are arranged in broken and irregular transverse ridges, again as in venosa. The minute microscopic hairs are in long, indefinite rows or sometimes not arranged in rows, while in venosa they are in shorter, well-defined rows. Larval habitats are only poorly known. Rogers (1933: 49) found the larvae in “saturated mosses and liverworts, on wet earth banks, on rocks and (more rarely) sodden logs.” He observed that they fed on the growing portion of the habitat plants, mainly at night. His collection of these larvae was found in include both obscura and subalbipes, however, and the bryophytes were not identified. My few collections of the larvae and pupae of sub- albipes, made in southern Michigan, were from the mosses Tetra- phis pellucida and Dicranum scoparium and the hepatic, Geocalyx graveolens. These bryophytes were all found on extremely de- composed and sodden wood, in a birch-maple-elm swamp, but the plants themselves were not saturated with water, as were those described by Rogers. I have never found subalbipes larvae in mosses as wet as those in which the immatures of sayi or dorsalis occur. Accordingly, I question the inference by Judd (1958: 624 ff. ) that subalbipes adults emerged from water about two feet deep. It is possible that larvae were in some fragment of moss accidentally 872 Tue UNIversiry SCIENCE BULLETIN adhered to his tent trap at or above the water line, but that they might have been in submerged mosses seems unlikely to me, in- asmuch as my laboratory experiments indicate the larvae of Doli- chopeza spp. drown when wholly covered with water. The pupa of swbalbipes may be recognized by its slender lobed spiracular yoke and the fact that the pleural spinous processes are tipped each by a single bristle. Color comparison of the pupa of this species with that of obscura is made in the discussion of the pupa of the latter. The pupal stadium lasts six or seven days, under laboratory conditions. Notes on distribution —ALABAMA—Washington County, 14 June. Connecticut—Hartford County, 14 June; Litchfield County, 31 May; New Haven County, no date; Windham County, 14-15 June. FLorma—Alachua County, 16 January, 17 to 28 February, 11 to 21 March, 1 to 16 April, 3 June, 25 July, 14 October, and 12 to 25 November; Escambia County, 7 April; Gadsden County, 28 to 31 March and 5 June; Hernando County, 21 March; Jackson County, 31 March to 29 April; Jefferson County, 31 March; Leon County, 18 March and 31 March to 13 June; Liberty County, 3 March to 26 April, and 15 to 27 July; Marion County, 4 April; Nassau County, 27 April; Putnam County, records for every month except Decem- ber and February, with peak abundance in June (23 date records). Grorcia—Baker County, 27 January, 9 to 24 February, 24 to 31 March, 7 and 21 April, 26 May, 27 August, and 14 October; Clay County, July; Cook County, 10 June; Hall County, 6 June; Lump- kin County, 2 to 8 June; Rabun County, 12 June; Union County, 23 May and 10 and 28 June. Kenrucky—Edmonson County, June (?). Lovurstana—Natchitoches County, no date. Matne—Cum- berland County, 1 July; Penobscot County, 18 June; York County, 1 July. MaryLtanp—District of Columbia, 29 August; Anne Arun- del County, 20 June; Baltimore County, 17 June; Prince George’s County, 9 June. Massacnuuserrs—Middlesex County, 10 to 16 June. MicuicAn—Cheboygan County, 3 and 31 July; Iosco County, 22 July; Lake County, 27 June; Livingston County, 2 June to 8 (14?) July and 14 to 28 August; Manistee County, 21 June; Wash- tenaw County, 12 and 20 June. Mrinnesora—Clearwater County, 11 July; Cook County, 5 July. Mrsstssrppr—Itawamba County, 14 July. New Brunswicx—Charlotte County, 5 July. New JersEy— Bergen County, 1 to 15 June; Burlington County, 11 August; Cam- den County, 3 June and 8 August; Gloucester County, 6 June; Mon- mouth County, 12 June. New Yorx—Erie County, 9 July; West- THe CRANE FLY GENUS DOLICHOPEZA 873 chester County, 9 June. Norra Carotrya—Buncombe County, 12 June; Macon County, 11 to 13 June and 26 August to 4 Sep- tember; Mitchell County, 16 June; Onslow County, 9 July; Swain County, 20 June; Transylvania County, 8-9 June; Yancey County, 7 June. Onto—Hocking County, 6 June; Portage County, 25 June. Ontrario—Blackburn (5 miles east of Ottawa), 6 August; Kearney (about 20 miles west of Algonquin Park), 2 July; Byron (4 miles west of London), 16 June. PENNsyLvANtA—Berks County, 15 Au- gust; Centre County, 25-26 June and 4 to 9 July; Luzerne County, 4 and 14 to 19 June; Sullivan County, 10 July. Quesec—Montreal, 14 July. Ruopr IsLtanp—Kent County, 15 to 21 June. Sourn Caro- LiInA—Greenville County, 22 May to 7 June and 20 July to 11 August. TENNESSEE—F entress County, 6 to 18 June, 5 to 9 July, and 8 to 13 August; Morgan County, 12 June; Scott County, 29-30 May; Sevier County, 15 and 30 June. Vircrm1a—Arlington County, 31 May; Augusta County, 28 June; Giles County, 31 May to 24 June; Nor- folk County, 10 August. Wesr Vircinta—Pocahontas County, 23 June and 5 July; Tucker County, 24 June. Wusconsrn—Juneau County, 6 July. Dolichopeza (Oropeza) subvenosa Alexander Literature references—Dolichopeza (Oropeza) subvenosa Alex- ander. Alexander, 1940: 618-620, fig. 15 (wing), fig 18 (hypo- pygium); Alexander, 1941la: 297. Original description—‘Allied to venosa; mesonotal praescutum grayish yellow with three conspicuous blackish stripes; pleura gray- ish yellow, variegated with brown; tips of tibiae and the tarsi paling to yellow, femoral tips narrowly brightened; wings with a strong brownish tinge; stigma darker brown, with conspicuous cream-col- ored post-stigmal areas; male hypopygium with the inner dististyle broad, especially opposite its outer end; aedeagus with a strong spine on ventral face near apex. Male. Length, about 9-11 mm.; wing, 11-12.5 mm.; antenna, about 3-3.5 mm. Female. Length, about 11 mm.; wing, 11 mm. “Frontal prolongation of head pale testaceous; palpi brown. Antennae with scape and pedicel pale yellow; base of first flagellar segment pale, the remainder black; segments relatively long, sub- cylindrical, much exceeding the erect verticils; pubescence of segments short. Head brownish gray, darker in central portion. “Pronotum brown, blackened medially. Mesonotal praescutum grayish yellow, with three conspicuous blackish stripes, the pos- terior interspaces obscured; scutal lobes brownish black, the median 874 Tue UNIversiry SCIENCE BULLETIN area very restrictedly paler; posterior sclerites not black or brown- ish black. Pleura grayish yellow, variegated with brown on the anepisternum, ventral sternopleurite, meron and extreme ventral edge of pleurotergite. Halteres elongate, stem yellow, knob darkened. Legs with coxae pale, the fore pair darker; trochanters yellow; femora blackened, the tips narrowly pale; tibiae dark brown, paling to obscure yellow on distal portion; tarsi obscure yellow. Wings with a strong brownish tinge; stigma oval, dark brown; a conspicuous, paler brown seam along vein Cu; membrane adjoining the stigma, especially in cells beyond the stigma, cream- yellow; a restricted brightening across the fork of M; veins brown. Venation: Sc, ending opposite or just beyond origin of Rs. “Abdominal segments blackened medially and on basal rings of segments, leaving extensive yellow areas on sides of posterior rings; outer sternites and tergites, including hypopygium, more uniformly blackened. Male hypopygium with the lateral arms of tergite blackened, the apices dilated into oval spatulate blades, the margins smooth or without conspicuous angulations; median area of tergite trilobed, the central lobe longer and more spinous. Outer dististyle uniformly blackened, cylindrical, a little exceeding the inner style. Inner dististyle much deeper than in venosa, ele- vated just above the short apical beak. Gonapophyses much as in venosa, appearing as blackened spines, their bases with conspicuous setae. Aedeagus close to apex on ventral face with a strong, erect spine, this variable in length but always strongly developed. In venosa, the lateral arms of tergite are more angular, the median area of the caudal margin with the central lobe low or lacking; inner dististyle narrower, especially above and before the apical beak; aedeagus without spine.” Types.—Holotype male, Anakeesta Ridge, 4500 ft. (Great Smoky Mountains National Park), Tennessee, 12 June 1939, C. P. Alex- ander. Allotype, same locality as holotype, 4000 ft., 5 June 1939, C. P. Alexander. Eight male paratypes, of which six are from the Great Smoky Mountains National Park, all above 2500 feet eleva- tion, various dates in June and various collectors; the other two from nearby Mt. Mitchell, Yancey County, North Carolina, above 4000 feet, June, collected by C. P. Alexander. Holotype, allotype and five paratypes are in the collection of Dr. Alexander, at Am- herst, Massachusetts. One paratype is in the collection of the University of Michigan Museum of Zoology, and one is in the United States National Museum. THe CRANE FLY GENUS DOLICHOPEZA 875 Diagnostic characteristics—Dolichopeza subvenosa is differen- tiated from carolus and subalbipes by its over-all darker color, by having yellowish instead of white tarsi, and by lacking the con- trastingly darkened tips on the femora and tibiae. It very closely resembles the more northern Dolichopeza venosa, and I have not yet discovered any reliable means of distinguishing the females of the two forms. The most readily visible characters for recog- nition of males of swhvenosa from those of venosa are the presence of a conspicuous spine on the ventral or posterior surface of the adminiculum (aedeagus, in Alexander’s terminology) and the more 0.5 1.0mm. SCALE, FIG.220 SCALE, FIGS. 215-219 Fics. 215-220. Dolichopeza (Oropeza) subvenosa; 215—ninth ter- gum of male, 216—left inner dististyle of male, dorsal aspect, 217—left outer dististyle, dorsal aspect, 218—medio-posterior margin of ninth tergum of male holotype, 219—vesica, penis, adminiculum, adminicular rod and gonapophysis, left lateral aspect, 220—terminal abdominal seg- ments of female, left lateral aspect. S76 Tue UNIversIty SCIENCE BULLETIN rounded tips of the lateral arms of the ninth tergum (Figs. 219, 215). Although the greater subapical width of the inner dististyle in subvenosa is a good indication of the species (as compared to venosa), it is not easy to observe except in alcoholic material or specimens mounted on microscopic slides, as the tips of these structures are usually more or less drawn down into the genital chamber. Descriptive comments.—A tawny-yellow fly with dark brown markings, swbvenosa is slightly paler in average, over-all coloration than venosa, partly by reason of the narrower abdominal annulations and partly because of the less intense color of these and the other markings. A few specimens from Giles County, Virginia, have darker tarsi than usual, these being light brown to the tips. Wing venation in the relatively small number of specimens ex- amined (138) was found to adhere closely to the subgeneric pat- tern. Slight variations in the medial field, such as the intersection of the m-cu cross-vein with the short M, ,, or with M, beyond its junction with M, are not uncommon. One male was seen in which the discal cell was closed distally by a temporary fusion of M, and M, . ., in the absence of the medial cross-vein. Body size of males ranges from 9 to nearly 11.5 mm., with the wings varying from 10.5 to 13 mm. Females are slightly larger, their over-all body length being from 11 to 12.5 mm. and their wings from 11 to 13 mm. There is so much variation in the male hypopygial structures used in differentiating suwbvenosa and venosa that for a time I re- garded these two forms as only subspecifically distinct. The spine on the adminiculum, for example, which is the most outstanding feature of swbvenosa, is normally long, acute and conspicuous, as in Figure 219; however, there are occasional specimens in which it is reduced to a low, obtuse bump, even less conspicuous than the spinous development found on the adminiculum in many speci- mens of venosa. The more rounded edges of the flattened, apical portion of the tergal arm are an ordinarily reliable means of rec- ognition of swhvenosa, but this character also varies across the boundary between species. Males with angularly spatulate tergal arms but otherwise typically subvenosa have been seen from Virginia and North Carolina, while the rounded form has been observed in venosa males from New York and Michigan. A more detailed analysis of variation in hypopygial structures will be found in the discussion of Dolichopeza venosa. The intensely sclerotized, THe CRANE Fry GENUS DOLICHOPEZA 877 sharply pointed gonapophyses and darkly colored outer dististyles appear to me to be the same in both species. The pronounced hump in the outer curvature of the inner dististyle just before the short apical prolongation (Fig. 216) is a character of suwbvenosa that is only very rarely seen in venosa, the inner dististyle in the latter species ordinarily being much more slender near the tip (Fig. 281). Geographical distribution—For eighteen years following its de- scription, subvenosa was known only from the central Appalachian Mountains of Tennessee, North Carolina, and nearby Georgia and South Carolina. Dr. Rogers collected many specimens also in southwestern Virginia, but the range of the species remained widely separated from the known range of venosa, to the north. Records of venosa in South Carolina (Alexander, 1942: 215) pertain to this form. Intensive collecting in West Virginia, south- ern Ohio and Pennsylvania and western Virginia in 1958 did much to clarify the respective ranges of subvenosa and venosa, the gap between the two finally being closed in Preston County, northern West Virginia, and subsequently in Garrett County, western Mary- land, with the discovery of the presence of both forms, there. These ranges, so far wholly allopatric, are conveniently mapped together, the distribution map included with the discussion of venosa. As mentioned earlier, I had considered suwbvenosa and venosa sub- species of one species, at one time; however, because there is no clear zone of intergradation between the two and because their identifying characteristics are sufficient to distinguish the average specimen (on the basis of males), I have abandoned that opinion and recognize two species. Habitats—Most collection records for subvenosa are for high elevations in the Appalachians, such as the 2500 to 4500 foot eleva- tions of the type series. I have taken the species at elevations near 1000 feet, which is approximately that of the Greenville, South Carolina, locality, and at various elevations up to around 6600 feet, near the summit of Mt. Mitchell, Yancey County, North Carolina. The species thus has a considerable vertical distribution, yet not as great as either obscura or americana, species I took together with it atop Mt. Mitchell. Within its mountainous and forested range, subvenosa has diurnal resting places similar to those of venosa: shaded, rocky crevices, undercut earthen banks, and other situa- tions where deep shade obtains. Occasionally I found subvenosa suspended from ferns or other leafy vegetation near such darkened 878 THe UNIvERSITY SCIENCE BULLETIN crannies. Along mountain brooks, they are often found very near the water, such as beneath an overhanging, rootbound and mossy bank, where the clearance above the water may be only a few inches. Unpublished field notes by Dr. Rogers indicate similar habitats in the places where he collected, near Mountain Lake, Virginia, and at various localities in North Carolina. Seasonal distribution—All but a few of the records for this species are in June, suggesting that there is but one generation per year, and the numbers of individuals taken on the various dates indicate the peak of emergence in the southern and central Appalachian Mountains comes about the end of the first week of June, perhaps a week later than that in West Virginia. At no time or place, however, have I obtained swbvenosa in large num- bers—never in aggregations and never more numerous than any other species of Dolichopeza occurring together with it. Immature stages—No study has been made of the immature stages of this species. Eggs dissected from the abdomen of a female of average size were found to have nearly the same meas- urements as those of venosa. The average dimensions were .79 by .34 mm., but eggs from a slightly smaller female measured only .75 by .31 mm. There is no terminal filament. No larvae have been found, but I would expect them to resemble rather closely those of venosa. Pupal skins apparently of this species were col- lected from a sparse growth of Mnium sp. on a steep soil bank above a mountain brook in western North Carolina. In these, the spiracular yoke resembles that of venosa except that the outer secondary lobes are more developed; also, the reticulation of the mesonotum is faint, consisting mostly of transverse wrinkles simi- lar to those along the mid-line. There are spinous processes on the fourth abdominal sternum about half as long as those on the fifth. Four female pupae from North Carolina and Georgia ranged in size from 10.2 to 12.0 mm.; males assignable to this species were not found. Notes on distribution—Grorc1a—Rabun County, 5 June; Towns County, 12 June; Union County, 10 June. MaryLanp—Garrett County, 28 June (venosa also here). Norra Carotina—Burke County, 14 June; Haywood County, 28 May; Macon County, 8 to 15 June; Swain County, 11 to 17 and 30 June; Transylvania County, 9 June; Yancey County, 7 to 22 June. Sourn Carotiwa—Greenville County, 3 to 29 June. TENNESsEE—Sevier County, 12 to 19 and 30 June. Vircinia—Giles County, 8 June to 9 July. West Vircinra— THe CRANE FLY GENUS DOLICHOPEZA 879 Pendleton County, 27 June; Preston County, 25 June (venosa also here); Tucker County, 24-25 June. Dolichopeza (Oropeza) tridenticulata Alexander Literature references——Dolichopeza (Oropeza) tridenticulata Al- exander. Alexander, 193lc: 177-178; Alexander, 1940: 620; Alex- ander, 1941la: 297; Alexander, 1942: 215, fig. 261 (hypopygium ); Foote, 1956: 222. As Oropeza obscura Johnson. Johnson, 1909: 122 (part), pl. 15 (wing); Rogers, 1930: 22-23 (part); Dickinson, 1932: 212 (part), fig. 114 (wing). Dolichopeza (Oropeza) tridentata Alexander. Crampton, 1942: 147, fig. 6A (misspelling ). Original description— “Male. Length about 10 mm.; wing 11 mm. Described from alcoholic specimens. Closely related to ob- scura, differing especially in the structure of the male hypopygium. Antennae dark brown. Mesonotum dark reddish brown, the pleura still darker. Legs with the tarsi a little paler than the tibiae. Wings suffused with brown, the oval stigma slightly darker brown; paler areas before and beyond the stigma and across the base of cell 1st M,. Abdominal segments brownish yellow, conspicuously ringed with dark brown on the incisures, on the sternites the bases of the segments more broadly darkened than the apices. Male hy- popygium with the central portion of the tergal margin produced into a small rectangular area that bears three small chitinized points; lateral arms of tergite evenly rounded at tips. Inner disti- style very broad, weakly bidentate at tip, one of the points being a small blackened spine. Outer dististyle and gonapophyses much as in obscura.” Types.—Holotype male, Goshen, Hampshire County, Massachu- setts, 1 July 1931, G. C. Crampton. The holotype, from alcohol, is in the collection of Dr. C. P. Alexander, at Amherst, Massachusetts. Several females taken at the same time as the holotype were not designated paratypes and have apparently been discarded. Diagnostic characteristics—This species may be confused with obscura, polita sspp. and possibly with johnsonella. Jn the field, it may be distinguished from polita by its smaller size and distinct abdominal annulations; from johnsonella it differs in having the wings tinged with grayish brown instead of amber or golden brown. Field differentiation of tridenticulata from obscura should ordinarily not be attempted, for the reasons discussed under the latter species. 880 THE UNIVERSITY SCIENCE BULLETIN Male specimens of tridenticulata may be recognized readily by the configuration of the ninth tergum, with its characteristic three- toothed median lobe and slender tergal arms. In polita cornuta there is a tridentate median lobe, but the central tooth is invariably longer than the others and is often depressed; furthermore, the tergal arms and outer dististyles are conspicuously unlike those of tridenticulata. In female specimens of tridenticulata, the most darkly sclerotized part of the ovipositor is the subapical portion of the hypovalves, while in obscura females the tenth tergum is dark- est. Females of all forms of polita have weaker abdominal annula- tions than tridenticulata or lack these altogether; in addition, the basal segments of the antennae of polita are paler than in tridenticu- lata. Descriptive comments.—Dolichopeza tridenticulata, like obscura, is of a dusky brown color such that the flies are easily concealed in their shaded daytime haunts. However, where I have found the two species together, tridenticulata has usually been perceptibly lighter in over-all coloration. Particularly the thoracic dorsum is paler, being a dark reddish brown in tridenticulata, as described by Alexander, but dark brown in obscura. In the northern part of the range, tridenticulata is colored more nearly like obscura. In alco- holic collections of tridenticulata and polita sspp., flies often col- lected together in rocky ravine habitats, the former species may be recognized by the fact that the pleurotergite and a narrow portion of the other sclerites around the second spiracle are of the same shade of brown as the other sclerites of the thoracic pleura, while in polita these areas above the base of the haltere are darker than the rest of the pleura. This color difference is not evident in flies preserved dry. Venational aberrations, especially those involving the media be- yond its first fork, are widespread and very common in tridenticu- lata. Several of these were illustrated earlier (Figs. 40 through 48 ) in the discussion of intraspecific variation. These were regarded as locally inbred abnormalities. One specimen has been seen in which the m-cu cross-vein is broken and one segment of it strongly deflected (Fig. 228); in another, there is a spurious cross-vein in cell Ist M, (Fig. 225); and a third shows a most unusual abnor- mality of the branches of the radius (Fig. 226). Presence of the vein Sc, has been noted many times, and a few instances of failure of R, to reach the wing margin have been found. Dolichopeza tridenticulata is one of the smaller species of Oro- Tue CrAaNE Fiy GENus DOLICHOPEZA 228 Oo 0.5 1.Omm. SCALE, FIG 229 SCALE, FIGS. 221-224 Fics. 221-229. Dolichopeza (Oropeza) tridenticulata; 221—ninth tergum of male, 222—left inner dististyle of male, dorsal aspect, 223— left outer dististyle, dorsal aspect, 224—vesica, penis, adminiculum, adminicular rod and gonapophysis, left lateral aspect, 225-228—varia- tions in wing venation, 229—terminal abdominal segments of female, left lateral aspect. 881 882 THE UNtversiry SCIENCE BULLETIN peza, nearly as small as Dolichopeza (D.) americana. Males average around 9 mm. in body length, with a 10.5 mm. wing, al- though they range from 7 to 10 mm. in length and have a variation in wing length from 8.8 to slightly over 12 mm. Females range in length of body from about 8 to 12.5 mm., their wings from 9 to 13 mm. The largest individuals of tridenticulata seen were from the Appalachian highlands of North Carolina and Virginia. Variation in the profile of the medio-posterior margin of the ninth tergum of males consists primarily of the three widths of the subrectangular projection, as described earlier, under intra- specific variation. There are few localities in which tridenticulata has been taken where not all three of these forms are present. The commonest form, however, is that shown in Figure 221. The tergal arms are usually slightly flattened at their tips and may be some- what widened (Fig. 221) or, less often, of uniform width through- out. The inner dististyle (Fig. 222) is very similar to that of polita sspp. I have noticed that in many female specimens preserved either dry or in alcohol the cerci are deflected downward, so that the tips of the hypovalves are concealed. Why this happens in triden- ticulata and not in obscura, polita or other species is not known. Antennae of females are about three-fifths as long as those of males taken at the same time and place. Geographical distribution—The geographical range of triden- ticulata very closely approximates that of the three subspecies of polita combined. Collection records show that this species occurs from Maine westward to southeastern Manitoba, southwestward to the Ouachita Mountains, and southward along the Appalachian Mountains to northeastern Georgia. It must be much more wide- spread in southeastern Canada than present records indicate, and collections in western Arkansas, southern Missouri, and Iowa only sketchily outline its distribution along the edge of the plains. Doli- chopeza tridenticulata will probably eventually be found in eastern Oklahoma and Kansas, possibly in Nebraska and eastern North and South Dakota. It may also occur in the Turtle Mountains of North Dakota and in the Black Hills of South Dakota. I would not expect it to range very far northwest across Canada beyond the presently known range. Habitats—In my experience, tridenticulata is primarily a rock gorge species; at least, it is in such habitats that I have found this species in the greatest numbers. Although I found it locally THE CRANE Fiy GENUS DOLICHOPEZA $83 L tridenticulata Map 10. Range of Dolichopeza (Oropeza) tridenticulata Alexander. Each spot represents one or more collections within a county (United States) or at a locality. abundant in certain northern woodlands, it was never as common there as in many rocky Appalachian ravines and rock ledge gorges in Indiana and Ohio. The only published habitat data are from the mountainous region of western North Carolina, where Alex- ander (1941la: 297) found tridenticulata “. . . common under low earthen banks along road.” In the vicinity of rocky ravines and boulder-strewn, forested mountainsides, tridenticulata is often taken away from the actual rocks, in such places as culverts, out- houses and other man-made shelters, open drain tiles, and be- neath exposed tree roots, overhanging banks, or almost any other situation affording deep shade. They were frequent visitors in my tent, in Indiana, even when I had pitched it several hundred yards back from the rocky ravine where the flies were usually found by day. They always arrived during the night, in such cases, and frequently stayed until darkness fell again. In the northern forests, tridenticulata was most often taken in or about small buildings, such as outhouses and picnic shelters. In some localities, it was found among granite boulders, where these were so arranged on a slope as to provide small, darkened crannies, here and there. In other localities, however, there were no rock outcrops seen, and the relief was so low that there could not have been any ravines; in one such place, tridenticulata was reared from a scanty growth of moss on a tree trunk. §84 THe UNIversiry SCIENCE BULLETIN Locally and in season, tridenticulata forms great swarms in shaded niches in the rocks, often in the company of polita and other species. At such times, one may easily capture hundreds of specimens. Be- cause of this dense swarming, it seems to me a little unusual that prior to the beginning of this study only about 200 specimens of tridenticulata had been collected in all of North America, as far as I am able to tell. During periods of heaviest emergence in the gorges of southern Indiana, for example, it is possible to collect 200 individuals easily within ten minutes, and I have often stopped collecting only because I was satisfied that further examples of the species would not alter the characteristics of my sample. Seasonal distribution—Although there are collection records for all the summer months for practically all parts of the range, there is also good evidence of two annual peaks of emergence, in most localities. From Georgia and Arkansas northward to New England and southern Michigan and Minnesota, these peaks fall in June and August, locally about a week after the appearance of polita. In Parke County, Indiana, in 1953, the first individuals of triden- ticulata were collected on 1 June, which was about two weeks after americana began to emerge, nine days after polita cornuta, four days after walleyi and obscura, and the same day as carolus first appeared. Peak numbers are reached about mid-June, in Indiana, and again between mid-August and early September. By 10 July, a few females were still present in the Parke County locality just mentioned. In another year, 28 August seemed almost the end of the late summer emergence of tridenticulata in nearby Owen County, Indiana. As in several other species, there appears to be a rather prolonged, single mid-summer period of emergence of adults of tridenticulata in northernmost United States and south- ern Canada. In northern Michigan, I found the peak of abundance to be about mid-July. Individuals of the late summer generation are not conspicuously smaller than those taken in June. Immature stages—Average measurements of eggs from several females of tridenticulata are .65 by .30 mm. There is a well-devel- oped terminal filament. Early instar larvae are unknown, but the third and fourth instars have been found in various relatively dry mosses. Larvae and many pupal skins were recovered from the moss Dicranella heteromalla, growing on the rather dry, upper sur- faces of blocks of rock in wooded, rocky habitats in Ohio and Mis- souri. Pupae were obtained from Tetraphis pellucida where that ‘Toe CRANE Fiy GENuS DOLICHOPEZA 885 moss was growing together with a powdery lichen, on a shaded sandstone cliff. The only specimens of tridenticulata known from southern Michigan are those reared from larvae collected from sparse tufts of the moss Orthotrichum sordidum that were growing in crevices of the bark of a tree, five feet above the ground. Larvae taken in the dry moss, Hedwigia albicans, and described as those of obscura (Alexander, 1920: 983) were found to be tridenticulata. The larva is very similar to that of the closely related polita, but is smaller and has the pleural hairs of the eighth abdominal segment irregularly arranged, leaving bare areas. These two species stand apart from all others in the genus, in that their larvae have the pleural hairs thickened and not arranged in circular groups. Pupae of tridenticulata may be identified by the spiracular yoke, which is similar to that of polita. The yoke is broadly but shal- lowly emarginate, the emargination lacking the inner, rounded “shoulders” seen in polita and thus appearing very broadly V- shaped. In a sample of 16 pupae from Carter County, Missouri, all but three had a spiracular yoke as shown in Figure 122, the three having a shallower emargination. In this same group of pupae, 11 had no spinous processes on the fourth abdominal ster- num and 5 had these developed to some degree. It was noted that in a few specimens the middle pair of spinous processes of the eighth sternum were barely close enough together to be re- garded as of the usual form for the obscura group; such specimens, however, cause no difficulty in use of the key to pupae. The pupal stadium, under laboratory conditions, was six days. Notes on distribution—ALABAMA—Etowah County, 26 June; Madison County, 8 June. Arkansas—Polk County, 30 July; Wash- ington County, 30 July. Connecricut—Litchfield County, 9 July; New Haven County, 4 to 18 July. Grorcra—Lumpkin County, 7 June; Stephens County, June; Towns County, 12 June; Union County, 23 May, 9-10 and 28 June. I_~rnors—Carroll County, 7 July; Jack- son County, 4 June; La Salle County, 7 July; Vermilion County, 18 June. Inprana—Allen County, 10 July; Cass County, 8 June; Jef- ferson County, 2 to 23 June and 20 August; Montgomery County, 28 June; Monroe County, 23 June; Owen County, 4 to 7 and 18 to 27 June and 28 August; Parke County, 1 and 9 to 28 June, 10 July, and 11 to 30 August; Tippecanoe County, 20 June. Iowa—Boone County, 21 to 29 June; Clayton County, 8 July; Jackson County, 8 July; Linn County, 8 July; Woodbury County, 11 June. KEn- Tucky—Barren County, 2 August; Bell County, 18 June; Edmonson 886 THe UNIversiry SCIENCE BULLETIN County, June (?); Franklin County, 23 June; Garrard County, 23 June; Letcher County, 3 July; Pike County, 3 July; Whitley County, 24 June. Maine—Cumberland County, 13 August; Oxford County, 11 July; Penobscot County, 25 June to 5 July; York County, 1 July. Maniropa—West Hawk Lake (near Rennie), 3-4 August. Mary- LAND—Garrett County, 26 June; Montgomery County, 24 May to 28 June. Massacuuserrs—Berkshire County, 19 June; Franklin County, no date; Hampshire County, 1 July; Middlesex County, 11 July; Norfolk Co., 18 June. Micuican—Cheboygan County, 5 to 30 July; Chippewa County, 13 July; Marquette County, 13 to 15 July; Washtenaw County, 14 May. Mrnnesora—Blue Earth County, 8 July; Carlton County, 6 August; Redwood County, 8 July; Winona County, 7 July. Missourr—Carter County, 6 June; Taney County, 10 to 20 June. New Hampsuire—Coos County, 11 July; Grafton County, 14 July. New Yorx—Essex Co., 1 July; Hamilton County, 12 July; Oneida County, 20 June and July; Ulster County, 28 June; Warren County, 2 July. Norra Carotina—Bun- combe County, 30 May; Burke County, 14 and 21 June, 1 July; Hay- wood County, 28 July; Macon County, 11 to 20 June; Swain County, 30 June; Transylvania County, 9 to 14 June; Yancey County, 26 May to 20 June, and 7 July. Ontro—Delaware County, May and 6 to 14 June; Geauga County, 17 July; Hocking County, 30 May and 6-7 June; Medina County, 4 July; Portage County, 24-25 June, 14 July, and 16 August. Onrarro—Algonquin Park, 23 June and 10 August; Gananoque, 8 July; Go-Home Bay (Georgian Bay ), 22 June; Kenora District (Lake of the Woods), 4 August; Pte. au Baril (Georgian Bay ), 6 to 16 July and 15 August. PrENNsyLvAN1Ia—Centre County, 25-26 June and 9 July; Columbia County, 15 July; Huntington County, 9 July; Luzerne County, 27 June and 5 to 10 July; West- moreland County, no date. QurBec—Rigaud, 25 June. SouTH Carotina—Greenville County, 7 June and 1 September; Pickens County, 29 June. TrENNEssEE—Bledsoe County, 26 June; Blount County, 20 May; Cumberland County, 25 June; Fentress County, 31 May to 20 July; Morgan County, 12 June; Sevier County, 7 to 23 June. VERMont—Windsor County, 7 July. Vimcinta—Fairfax County, 30 June and 5 September; Giles County, 1 June to 7 July and 20 July to 1 August; Rockingham County, 6 July; Shenandoah Na- tional Park, 28-29 June and 6 July; Smyth County, 20 June; Washing- ton County, 20 June and 2 July; Wise County, 2 July. West Vir- cinta—Greenbrier County, 4 July; Pendleton County, 26-27 June; Tue CRANE Fiy GENusS DOLICHOPEZA 887 Pocahontas County, 23 June and 5 July; Preston County, 25 June and 5 August; Randolph County, 5 July; Tucker County, 24-25 June and 6 August. Wisconsin—Iron County, 6 August; Jefferson County, 21 June and 21 July; Juneau County, 6 July; Monroe County, 6 July; Sauk County, 5 July; Trempealeau County, June and 7 July. Dolichopeza (Oropeza) venosa (Johnson ) Literature references.—Oropeza venosa Johnson. Johnson, 1909: 120, pl. 15 (hypopygium); Alexander, 1919: 930; Alexander, 1922b: 61; Alexander, 1924: 59-60; Alexander, 1925: 172; Johnson, 1925: 33; Alexander, 1928: 57; Leonard, 1928: 698; Alexander, 1929a: 236; Alexander, 1929b: 25; Alexander, 1930a: 212; Alexander, 1931a: 189. Dolichopeza (Oropeza) venosa (Johnson). Alexander, 1936: 280; Alexander, 1942: 215, fig. 26] (hypopygium ). Original description—‘“Front and rostrum yellow, vertex and occiput dark brown, palpi brown, antennae yellow, becoming fus- cous beyond the first joint of the flagellum. Thorax yellowish with three broad, dark brown stripes, the dorsal stripe ending at the suture, the lateral stripes abbreviated anteriorly and interrupted at the suture; plurae subtranslucent; collar, scutellum, metanotum, a large spot on the center of the plurae, smaller spots at the base of the halteres and between the coxae dark brown. The black bands of the abdomen are united along the dorsal line, leaving a large yellow spot on the side of each segment. Halteres long, yellow; knobs dark brown. Legs light yellow, the tarsi yellowish white. Wings brownish hyaline; stigma and veins dark brown, the radial and cubital veins noticeably prominent, and the short median cubital cross-vein wanting. Genitalia brown, appendages black, style long, reaching the end of the penultimate segment, red- dish, base black; appendages at the base of style acute, brown, tipped with black, margin deeply emarginate. Length, 10 mm.” Types.—Holotype male, Mt. Greylock (Berkshire County), Mas- sachusetts, 15 June 1906, C. W. Johnson. Five paratypes, all males, from Brookline, Massachusetts, Moosehead Lake, Maine, Bartlett and Hanover, New Hampshire, and St. Johnsbury, Vermont. These six specimens are in the collection of the Museum of Comparative Zoology, Harvard University. Another male in this collection, from the type locality, 11 June 1906, but without a paratype label, is probably a paratype and the seventh specimen mentioned by Johnson (1909: 120). 888 THE UNIVERSITY SCIENCE BULLETIN Diagnostic characteristics —Dolichopeza venosa rather closely resembles both carolus and subalbipes, not only in general colora- tion but also in resting posture, but it differs from both in having yellowish instead of white tarsi and in lacking darkened tips on the femora and tibiae. Darker than carolus in over-all aspect, although like that species colored generally dark brown on tawny or dark buff, venosa is less contrastingly colored than subalbipes, which is more of a brownish black on yellowish buff. This colora- tion is particularly evident in the abdominal region. The prescutal stripes of venosa are more distinct than those of either carolus or subalbipes. This species is somewhat darker than the closely re- lated subvenosa, a difference useful only in comparison of the two species. Having long, acutely tipped, blackened gonapophyses (Fig. 233), males of venosa may be quickly separated from all other species except subvenosa. From that species, they may be distinguished by characteristics of the tergal arms and the adminiculum. The mar- gin of the spatulate tip of the tergal arm in venosa is ordinarly angular (Fig. 230), but it is rather evenly curved in suwbvenosa; and venosa lacks the well-developed ventral spine found on the adminiculum of subvenosa. The inner dististyles of venosa (Fig. 231) are usually slender subapically but broadened in subvenosa. Females of venosa may be recognized by the combination of the color characteristics described above with the intense subapical sclerotization of the hypovalves. They are separable from females of swhvenosa on the basis of geographical range and, less reliably, by their darker color. Descriptive comments.—Color variation in this species is par- ticularly noticeable in the intensity of over-all coloration, certain individuals (usually males) appearing nearly as dark as johnson- ella or even obscura, when seen from a distance of several inches. Coloration of the legs, especially of the femora, may be brown to yellowish brown, and the tarsi may vary from a very pale tan through yellowish white to almost white. The extent of the stigmal spot of the wing varies, being somewhat smaller in specimens from Wisconsin and Minnesota than elsewhere. Venation in Dolichopeza venosa is very stable, as compared to some species in the obscura group. I have seen only three instances of an open cell Ist M, (discal cell) by loss of the medial cross-vein. One of these is a paratype, in which the abnormality was noted by Johnson. Half a dozen specimens seen had the veins M,,, and M, THe CRANE Fiy GENUS DOLICHOPEZA 889 pe A (os fa = liar LAINE NLA i J | i SCALE, FIG. 237 O 0.5 1.0mm. th f pee ase or, : 2 6 SCALE, FIGS. 230- 236 Fics. 230-237. Dolichopeza (Oropeza) venosa; 230—ninth tergum of male, 231—left inner dististyle of male, dorsal aspect, 232—left outer dististyle, dorsal aspect, 233—gonapophyses, ventral aspect, 234—-vesica, penis, adminiculum and adminicular rod, left lateral aspect, 235—varia- tion in medio-posterior margin of ninth tergum of male, 236—adminicu- lum of male holotype, left lateral aspect, 237—terminal abdominal seg- ments of female, left lateral aspect. 30—5840 890 Tue Unrversiry SCIENCE BULLETIN both joined with M, at the m-cu cross-vein; that is, the short length of M,,, was missing. It is not unusual for the vein Sc, to join R, well beyond the origin of the radial sector. Johnson described the m-cu cross-vein as lacking; he referred to a short length of M, that results from the intersection of m-cu with M, instead of the junction of M, and M,, but actually this disposition of the veins is not at ali uncommon. A long, sloping medial cross-vein has been seen in two specimens, and a distally spurred cross-vein in only one. Body length of males varies from 10 to 11.5 mm.; wings from 11.5 to 13 mm. Females measure 10 to 13 mm., their wings also from 10 to 138 mm. The smaller specimens are from Wisconsin and Minnesota, and the larger ones are from the central and northern Lower Peninsula of Michigan. There is a great amount of variation in the profile of the ninth tergum of the males, although the medio-posterior projection ordi- narity consists of a broad, three-lobed or three-pointed promi- nence. As in swbalbipes, the central point or lobe is sometimes lost (Fig. 235). In a collection of 38 males of venosa from Eaton County, Michigan, I found 14 shapes of this part of the ninth tergum. Eighteen specimens resembled the form shown in Figure 230, two had two small median points, four lacked the central point but varied in the depth of the emargination, one had an accessory lateral lobe, and the other types were variations of the relative lengths of the central and lateral lobes. In this same group of 38 males, 27 had irregularly angular apical portions of the tergal arms, three had these rounded as in subvenosa, and eight were classed as intermediates. Rounded margins of the apical part of the tergal arm have also been seen in specimens from other parts of the range. Among the 88 males mentioned, 27 had inner dististyles somewhat suggestive of those of subvenosa, 11 as illustrated in Figure 231; 26 had no vestige of a spine on the ad- miniculum, but one had a spine about half as long as that found in the typical form of subvenosa, while 11 were of intermediate form. In specimens from New York, Connecticut and Massachu- setts, I have seen various developments approaching the adminicu- lar spine; in fact, the holotype of venosa has a low, rounded pro- tuberance in the same position as that occupied by the spine in subvenosa (Fig. 236). Geographical distribution—Dolichopeza venosa occurs from Nova Scotia and the Gaspé Peninsula westward to Minnesota, northwestward to British Columbia and the Yukon, and southward Tue CRANE Fiy GENuS DOLICHOPEZA 891 ¥) venosa and subvenosa Map ll. Range of Dolichopeza (Oropeza) venosa (Johnson) and Doli- chopeza (Oropeza) subvenosa Alexander. Solid circles—venosa; solid tri- angles—subvenosa; circle within triangle—area where both forms occur. Each spot represents one or more collections within a county ( United States ) or at a locality. to Ohio and Maryland. I have not seen the specimens from west- ern Canada, which were collected by Dr. Alexander on his Alaskan trip in the summer of 1952, but I had rather expected that venosa ranged well northwest of Minnesota and northern Michigan, on the basis of the ecological distribution of the flies. The wide gap between the ranges of venosa and subvenosa that obscured their spatial relationships for many years was closed in the summer of 1958 with the discovery of both forms in northern West Virginia. My reasons for not regarding these two forms as subspecies have already been presented in the discussion of swhvenosa. Habitats —This species has been found associated with the rocky ravine type of environment in some localities and elsewhere with cool, moist northern forests. Often, these habitats are combined. Deeply shaded crannies in a northern hardwood forest in Cheboy- gan County, Michigan, occupied mostly by obscura, were shared by venosa; and in Juneau County, Wisconsin, venosa was taken to- gether with polita, under sandstone ledges. The venosa-polita asso- 892, THe UNIveRSITY SCIENCE BULLETIN ciation in a rocky habitat was found again in Eaton County, Michi- gan, where large numbers of both species were taken on the same date. In Centre County, Pennsylvania, venosa was swept from deep grasses at the edge of a hemlock-hardwood forest, where the forest opened rather abruptly into a mountaintop marsh. The habitat at Lake Itasca, Clearwater County, Minnesota, was in many ways similar to this. In general, forested areas, especially where these are also rocky, appear to be the most suitable environment for venosa, and much of the northeastern part of the United States and southeastern Canada comprises such an environment. Within this widespread habitat, venosa seems to be nowhere very abundant. Only once, in Eaton County, Michigan, have I seen aggregations of this species. Seasonal distribution—There are a few records for May, and most of the specimens are males, females appearing in the last days of that month. Most of the July records, on the other hand, are for females and are sporadic after the first week of the month. Records of June collections are by far the most numerous and point to a single early summer period of emergence, the peak coming in the last two weeks of June in most vicinities shown on the map, but several days earlier in the southern Ohio (Hocking County) locality. Immature stages—Eggs of this species, taken from a female of average size, measured about .79 by .32 mm. They have no ter- minal filament. Eggs laid on the nights of 31 May through 2 June began to hatch on 8 June, indicating the egg stadium to be about seven days. Four females, caged in the laboratory, produced over 300 eggs, each fly retaining several; accordingly, a female probably matures around 100 eggs. The larvae bear a striking resemblance to those of swbalbipes, in that both species have the larger micro- scopic hairs of the dorsum arranged in irregular, interrupted and staggered or deflected transverse rows, which are short with the hairs long and densely set. This arrangement makes it virtually impossible to count the number of transverse ridges on any seg- ment. Larvae were found in rather dry cushions of the moss Plagiothecium roeseanum that was growing on small amounts of soil trapped on a shaded sandstone cliff. It seems certain that this moss microhabitat was periodically well wetted by surface runoff and retained this moisture for many days. Nearby, on moist sand at the base of a cliff over which seepage water flowed, the moss Leptobryum pyriforme was growing, and there also larvae of venosa THE CRANE FLY GENUS DOLICHOPEZA §93 were found. The pupa is characterized by having conspicuous spinous processes on the posterior ring of the fourth abdominal sternum and by the spiracular yoke, the lobes of which are about twice as long as their basal width. It has been noted that the burrows prepared by the last instar larvae for pupation are un- usually long, being about one and three-fourths to twice as long as the pupa. The duration of the pupal stadium is ordinarily six days; one female emerged on the eighth day but was abnormal and died before becoming completely freed of the pupal skin. Notes on distribution—British Cotumpia—47 miles south of Fort Nelson, Alaska Highway milepost 253, 25 June. CONNECTICUT —Hartford County, 8 to 10 June; Litchfield County, 9 to 13 June; New Haven County, 29 June; Tolland County, 14 June; Windham County, 14 June. Marwe—Cumberland County, 1 July; Hancock County, 15 to 22 June; Oxford County, 11 July; Penobscot County, 25 June to 5 July; Piscataquis County, 11 July; Washington County, 15 July. MaryLanp—Garrett County, 28 June (swbvenosa also here). Massacnuserrs—Berkshire County, 15 to 28 June; Hamp- shire County, 15 June; Norfolk County, 30 May and 18 June. Micu1cAN—Cheyboygan County, 22 to 30 June; Eaton County, 28 to 31 May; Livingston County, 11 June; Ogemaw County, 20 June; Oscoda County, 16 to 26 June; Washtenaw County, 4 June. Mrn- NESOTA—Clearwater County, 16 June. New Brunswick—Freder- icton, 20 June. New Hampsnrre—Carroll County, 2 July; Chesh- ire County, 18 June; Coos County, 2 to 22 July; Grafton County, 5 July. New Jersey—Bergen County, 1 to 15 June; Essex County, June; Morris County, July. New Yorx—Albany County, 18 June; Cattaraugus County, 30 June; Cortland County, 21 July; Erie County, 4 to 23 June; Essex County, 16 June; Fulton County, 15 to 25 June; Greene County, June; Hamilton County, 17 to 23 June; Herkimer County, 9 June; Rensselaer County, mid-June; Tompkins County, 5 June. Nova Scorra—Cape Breton County, 4 July; Rich- mond County, 1 July; Victoria County, 2 July. Onio—Hocking County, 20 May; Portage County, 24 June. Onrarro—Algonquin Park, 3 to 29 June; Burke Falls, 9 to 13 July; Gull Lake (Muskoka District ), June; Orillia, 26 June to 2 July; Sand Lake (7 miles north of Gananoque, called “Sand Bay” on some maps), 2 July; Toronto, 17 June. PENNsyLVANtA—Centre County, 25 June; Luzerne County, 2 to 7 and 29 June. QuEBEc—Knowlton, 26 June; north shore Gaspé Peninsula, July. VeRmMont—Caledonia County, 27 June; Chittenden County, 15 to 24 June; Washington County, 17 June. West Vir- 894 THE UNIvERSITY SCIENCE BULLETIN GInt1A—Preston County, 25 June (subvenosa also here). Wisconsin —Juneau County, 6 July. YuKon—Watson Lake, Alaska Highway milepost 632, 28-29 June. Dolichopeza (Oropeza) walleyi (Alexander ) Literature references—Oropeza walleyi Alexander. Alexander, 193la: 139-140. Dolichopeza (Oropeza) walleyi (Alexander). Alexander, 1936: 280; Alexander, 1940: 618; Alexander, 1941a: 297; Alexander, 1942: 216, fig. 26K (hypopygium); Rogers, 1942: 60; Rogers, 1949: 12; Whittaker, 1952: 34; Foote, 1956: 222. As Oropeza sayi Johnson. Johnson, 1909: 118-119 (part); Alex- ander and McAtee, 1920: 393 (at least part); Rogers, 1930: 23; Rogers, 1933: 35, 49. As Oropeza similis Johnson. Dickinson, 1932: 212, fig. 113 (wing ). As Dolichopeza (Oropeza) dakota Alexander. Alexander, 1944: 241-243 (new synonymy ). Original description—“General coloration brownish yellow, the praescutum and scutum with clearly-defined brown areas; head gray; pleura yellowish white, without distinct markings; halteres with slightly infuscated knobs; legs pale brownish yellow; wings brownish yellow, the stigma brown; abdominal tergites obscure yellow with a brown median stripe, the lateral margins not dark- ened; male hypopygium with the gonapophyses large and con- spicuous, the margins irregularly dentate. Male. Length, about 9-10 mm.; wing, 11-11.5 mm. Female. Length, 11-12 mm.; wing, 12 mm. “Frontal prolongation of head pale yellow; palpi pale, the ter- minal segment suddenly blackened. Antennae (male) elongate, the basal three segments (male) or two segments (female) yellow, the remaining segments passing into brown, the basal enlarge- ments a trifle darker. Head gray, with a dark median and posterior border, the occiput paler. “Mesonotal praescutum brownish yellow, with three very distinct and clearly defined brown stripes; scutum similar, each lobe with two confluent dark brown areas; scutellum and postnotum pale brownish testaceous. Pleura yellowish white, without distinct dark markings, only the sternopleurite a little darkened. Halteres yellow at base, darkened outwardly, the knobs slightly infuscated. Legs with the coxae and trochanters yellow; remainder of legs Tue CRANE FLY GENUS DOLICHOPEZA 895 pale brownish yellow. Wings brownish yellow, the costal region deeper yellow; stigma oval, conspicuous, brown; obliterative areas before the stigma and across the base of cell Ist M,; veins brown. “Abdominal tergites obscure yellow, with a dorso-median brown stripe, the lateral margins pale; sternites yellow, with a dark spot at the incisures, the outer segments more uniformly darkened. Male hypopygium with the caudal margin of the tergite with a broad V-shaped notch that is extended into a flattened flange bear- ing a small slender spine at base of notch; lateral arms of tergite expanded at tips into obtuse blades. Outer dististyle long and slender, the base not enlarged. Inner dististyle dilated, produced into a blackened beak that is unequally bidentate. Gonapophyses very large and conspicuous, yellow, the margins irregularly dentate.” Types.—Holotype male, Knowlton (Brome County), Quebec, 4 July 1929, G. S. Walley. Allotype, same locality and collector as holotype, 12 July 1929. Alexander (1931la: 140) listed in addi- tion to the holotype and allotype “paratopotypes, 3 males and females, June 29-July 12, 1929 (G. S. Walley); paratypes, 2 males, Brookview, Rensselaer County, New York, June 14-21, 1923 (C. P. Alexander ).” The holotype and one paratype of each sex are in the Canadian National Collection, in the care of the Department of Agriculture, Ottawa, Ontario. The allotype, one male para- topotype and a female specimen from Brookview, New York, 14 June 1923, labelled paratype, are in the collection of Dr. Alexander, in Amherst, Massachusetts. The present location of the other specimen or specimens from the type series is unknown to me. Diagnostic characteristics —Field recognition of walleyi is usually possible on the basis of a combination of appearance of the fly and the type of environment. When it rests by day in deeply shaded crannies beneath rock outcrops, together with species of the obscura group, walleyi may be easily recognized by its con- spicuously paler thoracic pleura. From carolus, subalbipes, venosa and subvenosa, species with which it often occurs in leafy, low vegetation, walleyi is distinguished by its darker tarsi, which are never as pale as the darkest in those species. Occasionally, I have taken walleyi together with its nearer relatives, sayi and dorsalis, in swamp margins, and similis, in swampy woods. In such cases, sight recognition is difficult and not reliable; specimens should be captured before identification is attempted. Males of walleyi may be readily identified by their gonapophyses, which are yellowish in color, flattened and flared at their tips and 896 Tue UNIVERSITY SCIENCE BULLETIN have irregularly toothed margins. Although in caudal, or ventral, aspect the gonapophyses of walleyi are very like those of similis, they do not have the heavy spine projecting from the inner face of each gonapophysis, as in similis. The gonapophyses in sayi, the only other species in which these structures resemble those of walleyi, are only slightly widened at their tips and have the margins entire. Both males and females of walleyi may be recognized by their coloration, the yellowish body, yellowish brown legs and darkened thoracic dorsum together resembling, in a general way, only sayi, dorsalis and similis. The prescutal stripes in walleyi are, through most of the species’ range, reddish brown, as compared with the nearly black stripes of sayi and the more brown but nearly obliter- ated stripes of similis and dorsalis. Where walleyi overlaps the range of sayi, the thoracic pleura are unmarked, or nearly so. There may be a slight darkening of the ventral part of the pre-episternum and mesothoracic meron, but this is never intense enough to be confused with the dark gray pleural markings of sayi. Outside the range of sayi, both sexes of walleyi take on color characteristics similar to those of sayi, so that females are most reliably identified by association with males. Descriptive comments.—As pointed out earlier, in the discussion of intraspecific variation, the transition from unmarked to strongly marked thoracic pleura and the correlated transition from reddish brown to grayish brown thoracic stripes in Dolichopeza walleyi is not abrupt. Moving southward and westward from the center of the range of sayi, one first encounters noticeably marked or darkened populations of walleyi in Georgia, Tennessee, Indiana, Illinois, southern Minnesota and Saskatchewan. Specimens of walleyi collected south and west of the line indicated are increas- ingly strongly pigmented as their locality is farther from the line, so that the outermost populations, such as those from Florida, southern Missouri, eastern Kansas, western Iowa and the Black Hills of South Dakota, greatly resemble sayi in most details of coloration. The single female specimen on which Dolichopeza (Oropeza) dakota is based (Alexander, 1944:241-243) has distinct grayish brown thoracic stripes and dark pleural markings; it has a seam of dark color along the cubitus and a rather darkly colored stigmal spot; and the subapical sclerotization of the hypovalves is quite dark. In all, this fly much more nearly resembles eastern sayi than Tue CRANE FLy GEeNusS DOLICHOPEZA 897 it does the paler eastern form of walleyi. It was probably the in- tensity of the coloration that led Alexander to regard dakota as nearest to venosa. Having collected a large number of both sexes of the dark form of walleyi in the Black Hills and having compared females from among these with the holotype of dakota, | am con- vinced that dakota should be placed in the synonymy of walleyi. Venational variation in walleyi includes the branching of M, together with M,,. from the first fork of the media, so that M, 7 San Re y NN itttimerntl LK HY ie + & :\ ) 1.0mm. 243 eee he | SCALE, FIG. 243 oO 0.5 1.Omm. SCALE, FIGS. 238-242 Fics, 238-243. Dolichopeza (Oropeza) walleyi; 238—ninth tergum of male, 239—left inner dististyle of male, dorsal aspect, 240—left outer dististyle, dorsal aspect, 241—gonapophyses, adminiculum and adminicular rods, dorsal aspect, 242—vesica and penis, 243—terminal abdominal segments of female, left lateral aspect. 898 Tue UNrversiry SCIENCE BULLETIN has no connection with M,, a fragment of a spurious cross-vein in cell M,, a nearly sessile cell M,, and the vein Sc, sometimes present. Loss of the discal cell by absence of the medial cross-vein is not uncommon; in fact, the holotype of walleyi lacks the medial cross- vein in one wing and has it only faintly indicated in the other. In one specimen, the vein M, is basally interrupted, leaving the discal cell open for about half the distance between the junction of M, with m-cu and the medial cross-vein. In the vicinity of Ann Arbor, Michigan, there is a population in which individuals having macro- trichia in the outer cells of the wings are not uncommon. Occur- rence of this condition in other parts of the range of the species has been noticed, but only in isolated cases. Males of walleyi vary from 9 to slightly over 11 mm. in body length, and their wing measurements range from 9 to 12.5 mm. In the females, the wings are relatively shorter. Body lengths are from 11 to 14 mm., in the specimens examined, while the wings of these measured from about 11 to 13 mm. Larger specimens of both sexes were taken in June, and nearly all the small individuals were from August collections. There seems to be no particular pattern to the toothed margin of the gonapophyses; that it, within a local population there are as many kinds of margins of gonapophyses as there are gona- pophyses. This asymmetry is indicated in Figure 241. In walleyji, as in its near relative similis, the gonapophyses can be bent back- ward and ventrally from their bases, and in many specimens this has happened, apparently as a result of muscular action after the flies were placed in the cyanide jar. I have not seen the gonapoph- yses in such a position in living flies, so I do not know whether this action can be controlled. In the shape of the tergal arms, there is also great variation (Figs. 64-66), and again there may be dissimilarity between the two tergal arms of one fly (Fig. 238). The depth to which the ninth tergum is cleft varies somewhat in living flies but also varies mark- edly in specimens preserved dry or on microscope slides, depending on the degree of compression or depression. Geographical distribution —The known range of walleyi is nearly as extensive as that of dorsalis in over-all length and is greater in breadth. The species ranges from Nova Scotia westward to central Alberta, southward to Florida and southwestward to the Ozark Mountains and eastern Kansas. I expect that walleyi will eventually be collected in eastern Nebraska and Oklahoma, in Alabama, Mis- THE CRANE Fiy GENUS DOLICHOPEZA 899 / Map 12. Range of Dolichopeza (Oropeza) walleyi (Alexander). Each spot represents one or more collections within a county (United States) or at a locality. sissippi and Arkansas, and probably in the Turtle Mountains of North Dakota. Its range in Canada is only sketchily indicated; probably it is widespread in moist forests from the Great Lakes to Hudson Bay and northwestward to the Yukon, perhaps even into Alaska. Its occurrence in the Black Hills, where the popula- tion may be isolated, is discussed in the conclusions. Habitats—Like obscura, this species has appeared in a variety of general environments, and because of this wide range of habitat tolerance, walleyi is the second most numerous species of Doli- chopeza in general collections. Common in the vicinity of rocky ravines and gorges, it is also found in many forest situations, where moisture conditions vary from wet to only slightly damp. In Turkey Run State Park, Parke County, Indiana, I regularly found walleyi together with the dusky-brown species of the 900 Tue Universiry SCIENCE BULLETIN obscura group in darkened crannies beneath outcropping ledges of sandstone. It was almost as commonly seen suspended among ferns and other leafy plants along the rims of the ravines, where carolus was usually found. In narrow ravines through the thick sandstone strata, where there was no accumulation of soil on the ravine floor, hence no trees, few or no low, herbaceous plants, and no mosses (the ravines are rather effectively scoured out by annual spring floods), walleyi was not found, although polita cornuta, tridenticulata, americana and occasionally some other species were present. This seems to be related to the fact that walleyi most often spends its immature stages in mosses that grow on the soil, while the other species mentioned usually pass their larval and pupal stages in mosses that grow on the sandstone cliffs. It is not unusual to collect walleyi in upland woods, es- pecially early in the morning. Alexander (194la: 297), for in- stance, reported it from rather dry woods at an elevation of 5200 feet, in western North Carolina. I found walleyi flitting about in low forest undergrowth atop Spruce Knob, Pendleton County, West Virginia, at approximately 4850 feet elevation, on a very foggy summer morning when the temperature stood at only 49°F. Rogers (1930: 23) records it from “. . . the course of a shaded, rocky, talus slope brook. . . .” in a cove at the edge of the Cumber- land Plateau. In Florida, he reared adults from mosses on de- cayed logs and found other flies of this species among rank herbage on wet, shaded soil (Rogers, 1933: 49). In the southern Appala- chian region, Dolichopeza walleyi has been found in forest and rocky ravine situations from the piedmont level up to nearly 6000 feet. Whittaker (1952: 34) took walleyi in a beech forest; and Rogers (unpublished notes) found it along brooks in wooded localities and about rock outcrops, in the vicinity of Mountain Lake, Giles County, Virginia. Several instances of this fly’s having come to or into buildings near wooded areas have come to my attention; and while certain of these structures were picnic shelters and the like, others were buildings in which electric lights were op- erated at night, such as utility buildings in public campgrounds. Furthermore, there are several records of collection of walleyi in light traps, although rarely more than one specimen in any col- lection. This information, together with repeated daytime ob- servations of walleyi spontaneously on the wing over low vegeta- tion in moist woodlands, such as in floodplain forests, suggests that walleyi is not as strongly repelled by light as are several other species of Dolichopeza. THe CRANE Fiy GENus DOLICHOPEZA 901 Occurrence of walleyi in swampy woods is not common, although I have occasionally taken it in such habitats together with obscura and subalbipes, and Rogers (1942:60) noted that the species was occasionally common in the lower flood-plain woods, hehe maple-elm and tamarack-sumac swamps, and along shaded stream banks; not taken in the marshes where the closely allied D. sayi is common, although the habitats of the two species overlap to some extent in the swamps.” Only rarely have I taken walleyi in marshes; however, in South Dakota, this species was common, together with dorsalis, in a marsh of grasses, Carex, ferns and sparse, low willow scrub. Seasonal distribution—In seasonal distribution, there is again a close similarity between obscura and walleyi, in that it is some- what problematical whether there are two annual generations in the area between the southernmost Gulf coast and Florida and northernmost United States and Canada. In his study of the Tipulidae of a southern Michigan locality, Rogers (1942: 60) states that walleyi was taken from 28 May to 9 July, in the years 1936 through 1938, with a few records for mid-August of 1938. He concluded that “. . . generally there is only a single gen- eration a year in this region.” From my collecting and rearing data, gathered in the same general area, I believe there are two annual generations, the second, however, without as well defined a peak of emergence as the first. In central and southern Indiana, where I have collected a great number of Dolichopeza, I have found rather well-marked June and August generations of walleyi, the spring peak of abundance coming early in June and the late summer peak around the third week of August. At Turkey Run State Park, Parke County, Indiana, in 1953, walleyi males first appeared on 28 May, five days after the first appearance of polita cornuta and about two weeks after the onset of emergence of americana. By | June, the numbers of walleyi were still at a low level, but the rise in population seemed to be steady. By 11 July, only a few females were left. The August emergence in this lo- cality was found, in two different years, to be nearly completed by the last two or three days of that month. The Florida records are for March, April and June. In the vicinity of Washington, D. C., there appears to be a June generation and a second flight period from about the end of the third week of August extending into early September. In the northernmost parts of the United States and in the Canadian part of the range, there is a single 902 THe Untversiry SCIENCE BULLETIN generation annually, all the records being for June, July and early August. Immature stages—Eggs of Dolichopeza walleyi from a female from southern Michigan measured, on the average, .76 by .32 mm. There is no terminal filament. A single female may lay as many as 120 eggs, over a period of three or four nights. In the laboratory one female laid about 110 eggs in two nights but died before laying the few that remained. Hatching began seven days after the first eggs were laid, and many larvae had hatched out by the eighth day. The larva of Dolichopeza walleyi resembles that of sayi and, to a lesser degree, that of similis. All these larvae have the micro- scopic hairs of the dorsum most abundant on the last two thoracic and first abdominal segments and on the seventh and eighth ab- dominal segments, while the transverse ridges are more faintly indicated on the intervening segments. These ridges, in walleyi and sayi, are readily perceptible, but in similis they are extremely faint. The rows of minute microscopic hairs interspersed among the transverse ridges are well defined, about six hairs in length, and about equally spaced in walleyi, while they are less well de- fined, composed of various numbers of hairs, and are crowded toward the transverse ridges in sayi. A larva that I believe is almost certainly this species was de- scribed nearly 65 years ago by Hart (1895: 214-215), as a species of Tipula. The unknown larva, which Hart called Tipula larva “a,” was taken from soil on the floodplain of Spoon River, Illinois. It was regarded as “probably young,” very likely because of its small size as compared to other larvae of Tipula. Hart described it as follows: “Length 16 mm., diameter 1.5 mm.; grayish, covered with microscopic short dark brown pubescence denser on thorax and last segments, in transverse arrangement; folds very distinct, each slightly tranversely carinate and crested with darker pubes- cence, four folds on anterior divisions; each segment with the usual four setae, prothorax with several setae. Last segment with lower stigmatal teeth very small, triangularly black on upper surface; upper teeth slender, pointed, whitish, outer pair nearly twice as long as inner pair; a similar tooth anterior to each vt the outer two, near the anterior margin of the segment; a blackish spot at the base of each upper tooth, and one below each stigmatal plate; anal prominence with four blunt tubercles about the anal open- THe CRANE FiLy GENuS DOLICHOPEZA 903 ing.” This is the same larva later designated as “Tipula sp. 6,” by Malloch (1917: 202, pl. 31, figs. 6,7). In the shape of the spiracular yoke, walleyi pupae rather closely resemble those of obscura, but the two species may be readily distinguished by the characteristics of the pleural spinous proc- esses (see key to pupae). Presence of well-developed spinous processes on the posterior ring of the fourth abdominal sternum will aid in recognition of pupae of walleyi, as these are found only in a few other species, notably venosa, swhvenosa and carolus. Duration of the pupal stadium is six to seven days, under laboratory conditions. The occurrence of eulophid parasites in this species is described under the natural history of the pupae. Bryophytes known to serve as microhabitats of the immature stages of walleyi are several mosses growing under widely differ- ing moisture conditions. Atrichum undulatum is a coarse, often dry moss, usually growing on poor hilltop soil; yet the first im- matures of walleyi seen were found in this moss. Also fairly dry was the thin mat of Platygerium repens, growing on the bark of a fallen tree. In contrast to these were the moist mosses Amblys- tegium riparium and Mnium affine, both on the soil at the margin of a temporary, shallow pool in hardwoods, in spring. Other mosses from which larvae or pupae of walleyi have been recovered are Plagiothecium deplanatum, Entodon cladorrhizans, and Hypnum imponens, all of intermediate degrees of moisture and all growing in rather thin mats on forest soil, near streams. Notes on distribution —ALBERTA—Bilby (about 30 miles west of Edmonton), 28 June; Fawcett (74 miles north-northwest of Ed- monton), 20 June. FLorma—Alachua County, 29 March, April, and 3 June; Gadsden County, 5 April; Jackson County, 28 March to 18 April. Grorcra—Lumpkin Coun'y, 7 June; Union County, 28 June. I_.Linors—Carroll County, 7 July; La Salle County, 7 July; St. Clair County, 13 August; Vermilion County, 13 June. Inprana— Jefferson County, 8-9 June and 30 July to 28 August; Jennings County, 3 August; Montgomery County, 28 June; Owen County, 6-7 and 18 to 26 June, and 3 and 28 August; Parke County, 28 May to 28 June, 11 July, and 28 to 30 August. Iowa—Linn County, 8 July; Winneshiek County, 8 July; Woodbury County, 11 June. Kansas—Anderson County, 28 July; Douglas County, 25 May, 2 June, and 30 August. Krenrucky—Barren County, 2 Augus*; Ed- monson County, June (?); Franklin County, 23 June; Whitley County, 24 June. Marine—Cumberland County, 1 July; Hancock 904 Tue UNrversiry SCIENCE BULLETIN County, 17 July; Oxford County, 11 July; Penobscot County, 25 June to 5 July. Manrropa—Aweme (20 miles southeast of Bran- don), 23 June; West Hawk Lake (near Rennie), 4 August. Mary- LAND—District of Columbia, 29 August; Garrett County, 26 June; Montgomery County, 17 to 19 June and 25 August. Massacnu- sEtrs—Hampshire County, 27 May. MicaicAn—Antrim County, 2 July; Berrien County, 17-18 July; Cheboygan County, 22 June to 15 July; Chippewa County, 13 to 15 July; Eaton County, 30 August; Gogebic County, 27 July to 2 August; Iosco County, 28 August; Lake County, 12 June to 1 July; Livingston County, 28 May to 18 June and 14-15 August; Marquette County, 15 to 17 July; Oscoda County, 14 June; Otsego County, 3 July; Presque Isle County, 29 June; St. Joseph County, 12 August; Washtenaw County, 28 May to 29 June and 15 August. Mrnnesora—Blue Earth County, 8 July; Clearwater County, 11 to 17 July; Redwood County, 8 July; Winona County, 7 July. Mrssourr—Barry County, 29 July; Carter County, 6 June; Stoddard County, 31 May; Taney County, 10 June. New Brunswick—Fredericton, 26 July. NEw HampsHrre—Coos County, 2 to 11 July; Grafton County, 6 July. New Jersey—Ber- gen County, § to 15 June; Essex County, June. New York—Broome County, 18 July; Erie County, 19 July; Essex County, 19 June; Hamilton County, 12 July; Herkimer County, 3 July; Niagara County, 23 to 28 June; Rensselaer County, 14 June. Norra Caro- LinA—Avery County, 14 June; Burke County, 14 June; Haywood County, no date; Macon County, 11 June; Swain County, 20 June; Transylvania County, 14 June; Yancey County, 9 to 22 June. Nova Scotra—Halifax County, 5 August; Yarmouth County, 27 June. Oxnto—Delaware County, 9 to 14 June; Portage County, 24 June and 14 July; Washington County, 19 June. Onrarro—Algonquin Park, 26 June to 3 July; Burke Falls, 13 July; Sand Lake (7 miles north of Gananoque, called “Sand Bay” on some maps), 5 July: Thunder Bay (Lake Superior), 9 July. PENNsyLvANria—Hunting- ton County, 9 July. QuEBec—Knowlton, 29 June to 12 July. Sas- KATCHEWAN—Stony Lake (near Humboldt), 5 June. Sourn Da- KoTA—Pennington County (Black Hills), 11 and 23-24 July. TENNESSEE—Blount County, 15 to 17 June; Cumberland County, 25 June; Fentress County, 13 August; Haywood County, 25 May; Knox County, 25 May; Sevier County, 7 to 18 June: Washington County, 30 August. VERMont—Lamoille County, 17 June; Orange County, 11 July; Rutland County, 12 July. Vieacinta—Fairfax County, 24 June and 5 September; Giles County, 7 to 22 June and THe Crane Fiy Genus DOLICHOPEZA 905 26 August; New Kent County, 31 May; Rappahannock County, 30 June; Rockingham County, 6 July; Washington County, 18 August. Wesr Vircinta—Greenbrier County, 4 July; Mingo County, 3 July; Pendleton County, 27 June; Pocahontas County, 23 June; Preston County, 25 June; Tucker County, 24-25 June. Wusconsin—Jeffer- son County, 21 June; Trempealeau County, 7 July. SUMMARY AND CONCLUSIONS These laboratory and field studies of Dolichopeza point out cer- tain problems involving groups of several species and in addition allow a few generalizations that apply equally to all North American species. It has seemed best to discuss these here, in conclusion, following presentation of the detailed data relating to each species. Taxonomic categories.—It was one of the purposes of this in- vestigation to try to determine what biological species are rep- resented by the various described forms. To do this, it is first necessary to decide what is meant by a species. I visualize a species of Dolichopeza as a series of populations of morphologically similar, actually or potentially interbreeding individuals, effectively reproductively isolated from other such series of populations, having a certain geographic range and ecological distribution, and main- taining these characteristics through a span of time in the evolu- tionary sense. A satisfactory interpretation of the diversity of structure (and other characters) found in North American Dolichopeza requires the examination of large numbers of individuals. Study of one or two specimens here and a few there places undue emphasis on their differences, but when more nearly the full spectrum of varia- tion within a genus can be seen, similarities are emphasized and natural groups can be separated and ranked with perhaps less ease but much more confidence. Individual variants, locally inbred mutations, and geographic races are more likely to fall into their proper taxonomic places when viewed against the background of the entire range of variation of their species. Application of this point of view, I believe, can lead equally to “splitting” or “lump- ing” of older taxonomic categories. In judging where specific boundaries should be drawn in Doli- chopeza (for no matter how ideal his definition of a species, the taxonomist must in the end judge specimens and assign them to specific categories according to his opinion), I have relied heavily on comparative morphology. Individuals having essentially the 31—5840 906 Tue UNtversiry SCIENCE BULLETIN same appearance (allowance being made for sexual differences} are presumed to have very similar genotypes and are further pre- sumed to belong to an interbreeding population, if collected in one locality, or to potentially interbreeding populations, if col- lected in scattered localities. Specific boundaries have been fur- ther clarified by the evidence of reproductive isolation among the species recognized. Of the more than 11,000 flies examined, not one was regarded as a hybrid of any two species. Furthermore, among hundreds of mating pairs observed in the field and in the laboratory, none was ever found to involve male and female of different species. Although species of Dolichopeza repeatedly come into close proximity under natural conditions and were even caged together in the laboratory, no indication has been found of exchange of genetic materials between species. Minor variations having more or less random distribution within a species (such as small differences in shape of ninth tergum or minor shifts in position of wing veins), as well as somewhat more striking variations of localized occurrence (for example, locally inbred venational peculiarities ) and seasonal size differences, have required no taxonomic recognition. Even in certain cases when variation could be correlated with geographical distribution, tax- onomic recognition seemed not to be indicated, such as the instance of Dolichopeza dakota, which was judged to be only an extreme in the clinal distribution of color in D. walleyi. Where it has appeared clear that there is gene flow between morphologically distinct forms that are allopatric except for rela- tively narrow zones of overlap, these forms have been regarded as subspecies, the single example being in Dolichopeza polita. Across the zone of contact between the central and eastern forms of this species, in the central and southern Appalachian Mountains, there is concordant change in all the characters that have been found to vary geographically—structure of hypopygium, degree of luster of the thorax, and details of coloration. Individuals of the respective forms from outside the zone of intergradation are in all cases taxonomically distinct and readily separable, but within that zone both forms may occur, together with varying percentages of intermediates. Similar intergradation occurs at the only known locality where the western and central forms meet. If degree of difference between sympatric species can be taken as a fair indication of the relationship of allopatric forms in the same genus, some support for regarding the subdivisions of polita THE CrANE Fity GENus DOLICHOPEZA 907 as subspecies can be gained from the fact that a series of unusual characters common to these very similar forms are not found in any of the other species. Among these are the reversed relative sizes of males and females, the basal enlargement and pointed tip of the outer dististyles in the males, the coiled tracheal connection in the pupa, and the abbreviated terminal filament of the egg. A series of collections made at Nelson Ledges State Park, Portage County, Ohio, in three different years, included many mating pairs of the eastern and central forms yet never a pair involving both forms and never an individual that could be regarded as an inter- grade between the two. Notwithstanding the fact that these two forms intergrade in a zone extending from nearby points in Ohio to eastern Tennessee, in this one locality they may have achieved a specific level of reproductive isolation. Where the eastern form enters Michigan and Wisconsin it is wholly distinct. no areas of contact with the other races having been discovered in that region. An attempt to explain these relationships follows below, under the heading, species formation. When morphologically distinct forms have allopatric but con- tiguous geographical distributions yet show no clear evidence of intergradation along the line of contact, as in the case of Doli- chopeza venosa and subvenosa, the forms have been regarded as full species. That these two forms do come into contact has only recently been established, and there is so far no indication of inter- breeding, except for the very sporadic occurrence of certain modi- fied characters of each form within the range of the other, as dis- cussed earlier. In the case of suwbalbipes, where there are two forms with regard to one character only (tergal arm), with continuous gradation in other characters, it seems reasonable at this time to recognize only one species. Isolating mechanisms.—In Dolichopeza, closely related species often have nearly identical geographical ranges and are often the most closely associated ecologically. There is some very general environmental segregation, yet adults of as many as half a dozen species may not infrequently occur so close together in their day- time resting places that they come into physical contact, especially when in swarming flight following some disturbance. Larvae of two or three species are occasionally found within a patch of a single species of moss no larger than the palm of one’s hand. Under such circumstances, it is difficult to imagine how these species might 908 Tue UNIveRSITY SCIENCE BULLETIN be separated by any environmental barriers. Of course, isolation in any stage but the adult is irrelevant to the problem of inter- specific isolation; we are concerned here with reproductive isolation, which must be complete in the case of sympatric sibling species in order for them to retain their identity. Reproductive isolation ordinarily occurs because males and females of different species are not attracted to each other and thus do not mate at all. In other instances in which mating between species does occur, either there may be no fertilization or, if fertilization is accomplished, the off- spring may be either inviable or sterile. The composition of mating pairs mentioned earlier suggests that in Dolichopeza the first of these possibilities is the rule. Differences in size of individuals and in relative proportions of genitalic structures could probably account for a certain amount of isolation between species in the same habitat, as between similis and dorsalis, but in most cases I cannot see any structural barriers to interspecific matings. As pointed out earlier, there is no “lock and key” arrangement in copulation. The fact that peaks of emergence of adults do not coincide may provide a degree of reproductive isolation, but there is still such great overlap in the seasons of emergence of the various species in any locality that this would scarcely seem to be an effective isolating mechanism. It has been noted that mating in americana often takes place by day, while species of Oropeza ordinarily mate at night. This, however, fails to explain isolation among the closely related species of Oropeza, which is really the problem at hand. That species of Dolichopeza carry on mating and other activities primarily in the hours of darkness may still have some bearing on the means of isolation. It seems not unlikely that the process by which males and females of one species are attracted or brought together might also serve to repel all other species. That is, each species might be characterized by a particular odor, allowing males to detect females of their own species and distinguish them readily from all others. Flies caged in the laboratory, as well as those congregated in natural resting sites, occasionally exhibit diurnal mating but in general ignore one another, by daylight, un- less they come into bodily contact. It is possible that matings begun in daylight hours are merely the result of reflex reactions following appropriate but accidental contacts and that the stimulus of daylight causing resting behavior ordinarily overcomes any stimulus for other activity short of actual contact. THe CrANE Fiy Genus DOLICHOPEZA 909 Population control_—Potentially, a local population of any species of Dolichopeza could saturate its environment in a short while; how- ever, it is my general impression that, in areas where I have made observations over a period of several years, such local populations have remained about the same. It is, of course, difficult to dis- cover whether populations of various species of Dolichopeza are in the long run increasing or decreasing. It seems not unreasonable to assume that, although each species is constantly striving to in- crease, any particular population fluctuates from year to year ac- cording to environmental factors and that net gains during a fa- vorable year may be wholly or in part cancelled by losses in an unfavorable one. The number of individuals of a species in a local population (and it is both reasonable and convenient to think of species of Dolichopeza in terms of many localized, partially isolated popula- tions) is derived from an interaction of the species’ potential rate of reproduction and the effects of the environment. Each female reaching the adult stage is presumably capable of depositing 100 to 120 eggs. The longevity of ovipositing females needs to be taken into account, although most of the eggs seem to be laid within three or four days of the time of emergence. One must also con- sider, in connection with the species’ reproductive potential, the number of generations per year. Assuming for simplicity that there is but one generation each year and that the population shows no measurable increase from year to year, it may be seen that 98 percent or more of the potential offspring of each generation are lost before the next. What is the fate of all these individuals? Due probably to the method of fertilization, as many as a third of the eggs of one female have been observed, in the laboratory, not to hatch and are assumed to have been infertile. The pro- portion of such losses in nature cannot easily be estimated, but it is dificult to imagine biological perpetuation of such inefficiency. Mortality in the larval stages has been found to result from preda- tion, parasitism, disease, and the effects of severe adverse weather, such as excessive rainfall, drouth, or extreme temperatures. In the laboratory, many individuals died in the pupal stage, from preda- tion, parasitism and unknown causes. Of the adults that survive the hazards of emergence from the pupal skin, certainly many do not reproduce, again by reason of predation, disease, weather or accidents of one sort or another. Other environmental limitations that need to be considered are 910 Tue UNIVERSITY SCIENCE BULLETIN availability and quality of food, space in the habitat, and compe- tition—both among individuals of the same species and among the various species inhabiting the same general environment—for food and space. Competition.—There is no evidence of competition among species of Dolichopeza, either from records of changes in species composi- tion in various localized environments, over the years, or from field observations of many larvae and numerous large aggregations of adults. In habitats where a number of species are present, there seems always to be more than enough of both the bryophyte micro- habitats for the immature stages and shaded resting places for the adults. As mentioned earlier, the moss or liverwort habitat is al- most never fed upon by the larvae sufficiently to cause readily visible damage. Even though larvae of three species may be feed- ing together in one kind of moss, they rarely come into contact with each other because of the abundance of the moss (which is at once their food and shelter) and because individuals were initially spaced within the habitat as a result of the method of oviposition. It might be theorized that, in time, one species or another, by being more exactly adapted to the environment, would so gain in numbers as to force the others out through competition. However, it seems probable that fluctuations in the environment favor first one species and then another and that, because of the irregular geo- graphical distribution of such fluctuations, a species may gain tem- porary predominance in one locality but fall to low numbers in another, in an ever-changing pattern. As Ross (1957: 127) points out, “Because no two species are exactly alike genetically and therefore physiologically, each species should theoretically have an ecological optimum slightly different from all other species. It follows that, in cases of multiple occupancy of a niche [by species of the same genus], each species will tend to become abundant at a time or place different from the other species.” It appears that numbers of Dolichopeza are so reduced by preda- tion, parasitism, disease and rigors of the physical environment that competition, except perhaps at the level of the individual, never comes into play as a force in population control. Absence of competition, either among species of Dolichopeza or between this genus and species of Liogma, Tipula, etc., living in the same habitat, would furthermore seem to be an important factor in the ability of several species of Dolichopeza to occupy one and the same ecological niche. THE CRANE FiLy GENUS DOLICHOPEZA ===) se in appro: eSsTS 3T S ( beech, (a eastern Y FOREST CK OAK - HI OAK PI SOUTHEAS Eeaee [eed = SQ RIVER BO iS, or TUNDRA in far north GRASSLAN Mar 13. Some ecolog 911 peza in North Amer- rage length of frost- y lines are isophenes of aver: ical factors influencing the distribution of Dolicho orested regions; heavy Shading indicates f free period, in number of days indicated. ica. 912 Tue UNtversiry SCIENCE BULLETIN Geographical distribution—In a general way, the pattern of distribution of the genus Dolichopeza in North America corre- sponds with that of the forests in the eastern and subarctic regions of the continent (Map 13). Within this vast area, there are many recognized forest types, and one species of Dolichopeza may range through several of these. Accordingly, the composition of the forest does not appear as important to the distribution of these crane flies as the general effect that the presence of forest trees has upon the microhabitats required by the flies and their larvae. So similar are the ecological needs of species in various groups of Dolichopeza that the geographical ranges of certain of them nearly coincide. In fact, if the known ranges of all species were outlined on one map, it would be seen that the area including roughly the Great Lakes basin and the northern Appalachian Mountains is in- habited by twelve of the fifteen recognized species. A series of zones drawn around this central area would include diminishing numbers of species toward the periphery of the range of the genus. Even near the periphery, however, any habitat suitable for one species is likely to support two or more. Thus, in eastern Kansas, near the western limit of the forests, were found obscura, polita pratti and walleyi; of two specimens of Dolichopeza taken in the Ouachita Mountains on the border between Arkansas and Okla- homa, one was johnsonella and the other tridenticulata; and in west- ern Alberta, near the Continental Divide, over which no species seems to have passed, at that latitude, were found americana, dor- salis, obscura and walleyi. On the other hand, the range of every species must naturally come to an end somewhere, and I certainly do not suppose that, at the actual outer limits of the range of the genus, at least two species will always be found existing in the same habitats. Beyond the region of generally satisfactory habitats, perhaps one species here and another there may be pioneering, as it were, various marginal environments. Each species probably expands its range during years in which its over-all environment has been most nearly optimum, only to be thrown back in unfavor- able years, perhaps here and there holding some of the gained ground in spite of adversity. Although there seem to be altogether favorable habitats for Dolichopeza in the luxuriant forests of the Cascade Range of Oregon and Washington, the genus has apparently been prevented from westward expansion at this latitude by the Great Plains. By the route of the subarctic forest, several species occurring in eastern THE CRANE Fiy GENuS DOLICHOPEZA 913 North America have spread northwestward across Canada toward Alaska, bypassing the treeless plains and even moving a little way southward again, below the eastern slopes of the Rocky Mountains, following an extension of the boreal spruce-fir forest. So far as is known, however, these species have not penetrated the western coniferous forest (lodgepole pine, western yellow pine, etc.) and remain east and north of the Continental Divide. If the plains are such an effective barrier to the spread of Dolichopeza, how does the genus happen to be represented by at least three species in the isolated Black Hills of South Dakota? One might at first suspect a connection to the main eastern range by way of the valleys of the Missouri and Cheyenne rivers, yet there is a 300-mile gap between the Black Hills and the north- western extremity of oak-hickory forest in the Missouri Valley. The species composition of the Black Hills group (americana, dor- salis and walleyi) suggests that the group’s geographic afhnities are with the area to the north; however, once more the continuity of habitats is widely broken. It seems that the Black Hills are in fact wholly isolated and that species of Dolichopeza now living there have no connections whatsoever with their main eastern and northern ranges. In order to account for this distribution, it is necessary to visualize that at the time of maximum extent of the last (Wisconsin) continental glaciation, Dolichopeza occupied a forest that grew all along the ice front, which is to say from the Rocky Mountains along the Missouri and Ohio rivers to the Appa- lachian Mountains. As the ice withdrew northward, followed by the boreal forest, a remnant of that forest, into which elements of the Rocky Mountain forest had become introduced, was left be- hind, where, because of the cooler climate and higher rainfall in the isolated mountains, it withstood the encroachment of drier for- ests and eventually of grasslands. If this is how species of Dolicho- peza came to be in the Black Hills today, what accounts for their not having visibly changed from the parent stocks in their ten thousand years or more of isolation? The species involved all have unusually extensive ranges, span- ning approximately thirty degrees of latitude and including habi- tats from near sea level to more than 6000 feet elevation, distri- buted in a great arc nearly 4000 miles long from southeastern United States to northwestern Canada and Alaska. Not one of them, however, has differentiated into geographic races in response to the great ecological diversity encountered within its range. This suggests that these species have a high degree of adaptability and 914 Tue UNIVERSITY SCIENCE BULLETIN that such changes in genetic constitution as have been necessary to adapt physiology to environment have been so small that they have not become manifest in the phenotypes. It is thus not un- usual that species of such constitution have remained essentially unchanged through ten thousand years in isolation, for surely the environment has changed less in the Black Hills during that span of time than it changes today over the length and breadth of the described range. Seasonal distribution—The geographic range of the genus Doli- chopeza in North America may be divided into three major zones with respect to seasonal distribution. These are: (1) Florida and the extreme southern Gulf Coastal Plain, where emergence of adults may occur in almost any month of the year; (2) a broad middle belt extending from Florida and the Gulf Coast area north- ward to New England, central Michigan and southern Wisconsin and Minnesota, in which there are normally two annual appearances of adults, falling roughly in June and August; and (3) a northern zone, in which one annual flight period, usually during late June and July, is the rule. As this pattern of seasonal occurrence cor- relates in general with latitude and since cold weather is known to inhibit feeding activities of the larvae, tempera’ure appears to be the chief factor involved. Duration of the frost-free season may perhaps be taken as a measure of the time available for larval feeding and growth, hence of the potential production of adults. If contours are drawn upon a map, connecting points of equal average duration of frost-free period, two of these contours very closely approximate the zonal boundaries of seasonal distribution in Dolichopeza. The boundary between the southern and middle zones falls near the 260-day contour (Map 13), and the middle and northern zones are sep- arated more or less along the 150-day contour. On the average, the last frost of spring along the southern contour is in early March and the first frost of fall in late November or early December. Along the northern contour, the last spring frost is usually around mid-May, while the first in autumn comes in early October. Within the broad middle zone, there is thus an average feeding and grow- ing season of somewhat over 200 days. Rearing of Dolichopeza sayi in southern Michigan indicates (Fig. 99) that about 70 days are required for the development of one generation. Accordingly, supposing that occurrence of frost brought larval feeding and growth to a halt, there would be only sufficient frost-free days at THE CRANE FLY GENUS DOLICHOPEZA 915 that latitude to produce two generations of sayi in an average year. It is furthermore not unlikely that when frosts come early in the fall or persist late in spring, only one generation may be produced, as in the northern zone. Depending on local micro-climates, there may result a partial second generation, which would perhaps account for the fact that, in the northern part of the middle zone, the number of adults in the late summer generation is often con- spicuously smaller than the number seen at the same locality during the spring emergence. Species formation.—Dolichopeza is probably a very old genus. The present day tipuline genera Brachypremna, Tipula, Ctenophora and possibly Nephrotoma, as well as nearly twenty modern genera of the Limoniinae, were already differentiated in Oligocene time (see Alexander, 1931d) and are represented as fossils in deposits of that epoch in Europe and North America, especially in the Baltic amber. As Dolichopeza is regarded as a rather primitive genus among the Tipulinae, it quite possibly was differentiated prior to the Oligocene and spread widely into North America from Asia, at a time when a low, forested land connection existed be- tween the two continents in the region of Alaska. There appear to have been three, possibly as many as five, separate introductions: one or two each of Dolichopeza s.s. and Oropeza and one of Megistomastix, a West Indian subgenus that seems not to have been derived from either Oropeza or typical Dolichopeza. Given the forty million years or so, suggested above, to diversify in North America, these originally introduced species could easily have given rise to the present fauna, notwithstanding their apparently slow rate of evolutionary change. Despite the evident ability of species of Dolichopeza to adapt to diverse general environments, it seems likely that the fifteen species present today are all that remain of a sometimes greater or sometimes lesser number of predecessors, many species having arisen and_ later become extinct since the original introduction of the genus into this continent. Species formation has, I believe, resulted from effective spatial isolation. It is not difficult to imagine that during the geological past there were, as there are today, portions of the ranges of species in which populations are partially or altogether isolated. It is plain from the above studies that different genotypes may build up even in various parts of the more or less continuous range of a species, probably due to low vagility of the flies. Depending upon 916 Tue UNIVERSITY SCIENCE BULLETIN the degree of isolation, adaptive and non-adaptive changes that arise by mutation in such populations will be either slowly diffused into the genotype of the entire species or retained locally. Changes mixed into the genetic composition of the species as a whole may become lost completely or may persist as intraspecific variation. If locally occurring changes are advantageous (that is, have some small survival value), they may alter the entire parent species by outward diffusion from the place of origin, or, if long enough iso- lated, they may give rise to new species. Differentiation of most of the existing species of Dolichopeza must have occurred at a time so remote that one could only specu- late in general terms how or where it came about. If, however, one regards geographic races as incipient species, the case of Dolichopeza polita may be instructive. Much of the present range of this species lies in the area occupied by the last Pleistocene con- tinental glacier; therefore, if the species existed at the time of that glacial advance, it must have been forced southward, into the Ozark Mountains, the central and southern Appalachian Mountains and part of the broad intervening lowland. If the habitats of the species have not changed appreciably, then the Ozark population (polita pratti) must have been rather effectively isolated by the lower Mississippi Valley, in which rock outcrops are not common; and the eastern and central populations probably were partially isolated on the eastern and western slopes of the Appalachians, respectively. Following the withdrawal of the glacier, the Ozark and central forms could have slowly made their way northward through the advancing forests, from one rocky or outcrop area to the next, not coming into contact until they reached north central Illinois. The central form (polita cornuta) maintained its subspe- cific relationship with polita polita along the Appalachian ridges but on reaching northern Ohio encountered representatives of the typical race moving westward from Pennsylvania, from which area the central form had long been isolated by the glacier. There the two forms reacted as full species. The eastern form, meanwhile, being less restricted to rocky habitats and proceeding from an unglaciated area much farther north than the other two, moved westward through the Great Lakes region until it encountered polita cornuta in Michigan and polita pratti in Wisconsin, in each case failing to interbreed and thus maintaining a specific degree of isolation. This interpretation is based upon present knowledge of distribution (see Map 6) and, of course, may need to be changed when further data are available. THE CRANE FLY GENuS DOLICHOPEZA 917 DuDIUAaWwD $1/D910q Q SIJIWIS & 2 IKa||OM oe IADS $ s SI/DS1Op = = psouaaqns . DSOUADA 8 Snj|O409 S wel n sadiqjoqgns 6 $1/D4JSND a2) Z D|jauosuyo! 5, a aerate a p4inosqo ap ete pjOjndIjuap!4y + a< Spo ahiees DJIIOd D{I0d NJ NZ DJNUusOD D{IIOd ge > re pe 404d oyI0d 918 Tue UNIversIry SCIENCE BULLETIN Interspecific relationships—It seems likely that within the genus Dolichopeza certain adaptive structures have become well estab- lished and might be thought of as having reached a degree of stability as a result of close correlation of form with function. Stated another way, it may be that the genus must retain these characteristics in order to survive in its particular range of eco- logical niches. These characters common to all species of the genus may be said to be primitive for the genus, although they may be specialized in comparison to common characters of the family Tipulidae. Applying this same principle to the arrangement of species within the genus, widespread characters such as the pattern of wing venation may be regarded as primitive at the subgeneric level, or a certain kind of gonapophyses as primitive for a species group. Thus, the more characters any two species have in com- mon, the more closely related they are judged to be. In attempt- ing to arrange the species according to their most probable rela- tionships (Fig. 244), I have used wherever possible characters least likely to have been directly influenced by environment. Hy- popygial features, for example, have been found most useful, for I cannot perceive that the variations in shape of the ninth tergum, the dististyles or gonapophyses would be affected materially by a change in the species’ environment. In contrast, coloration seems to be closely correlated with the type of habitat. On the basis of coloration, subalbipes would appear more closely related to carolus than to johnsonella, and walleyi more closely akin to dorsalis than to sayi. The various relationships indicated in Figure 244 are dis- cussed in detail in the species accounts and need not be repeated here. Although this figure indicates no time scale, the levels of separation of lines of descent are intended to suggest relative lengths of time. The presence of most species both within and outside the area of the last Pleistocene glaciation suggests they have existed for approximately 10,000 years, at least. Lines of descent termi- nated by dashes are introduced to indicate supposed indirect deriva- tion of one existing line from another. Lastly, considering only those characteristics of larvae and pupae used in this study for differentiating species, it appears that differences in the immature stages are of phylogenetic significance rather than adaptations to environment. A brief hypothetical history of the genus Dolichopeza in North America.—The distribution of North American species of Dolicho- peza indicates that they are a boreal and sub-boreal group, vir- THe CRANE FLy Genus DOLICHOPEZA 919 tually limited to forested areas. In the Oligocene, the genus Dolli- chopeza may have been widely distributed in the northern parts of Europe, Asia and North America, for cool forests are thought to have been widespread in the northern continents, at that time. As a result of subsequent cooling of continental climates, this forest belt was displaced southward, carrying with it its fauna; and in early Miocene time, the genus Dolichopeza possibly occurred throughout much of its present North American range, as well as in much of the western part of the continent. With the renewed uplift of the present Rocky Mountains, during the Pliocene epoch, the western forests were widely obliterated, giving way to grass- lands in the rain shadow area east of the mountains and being somewhat replaced in the far west by a new flora of conifers. Dolichopeza, now confined to the hemlock-hardwood and deciduous forests of eastern North America, was widely separated from the Asiatic stock. Advances and recessions of Pleistocene ice forced some components of the genus as far southward as the Gulf coast and probably caused fractions of the populations of various species to become isolated, giving rise to new forms. With the last retreat of the continental glaciers, most of the populations drifted north- ward into the newly developing forests, but some stayed behind in the more deeply-shaded swamps and cooler valleys of the South. As the land warmed and oak-hickory forests replaced the moister woodlands, fragments of the once widespread populations of Doli- chopeza were left in the cool ravines and valleys of the Mississippi basin, together with relict stands of the hemlock-hardwood forest. Today, the stronghold of Dolichopeza is in the forests of the upper Great Lakes, St. Lawrence and New England regions, and in frag- ments of that forest extending southward along the crests of the Appalachian Mountains. From this area, some species have moved far northwestward through the subarctic spruce-fir forest, finding there environmental conditions adequate for their needs. Outside the spruce-fir and hemlock-hardwood forests, however, the present discontinuous pattern of distribution is, in a sense, a slowly fading ghost of what once was probably the center of concentration of the genus Dolichopeza in North America. THe Untversiry SCIENCE BULLETIN LITERATURE CITED ALEXANDER, CHARLES P. 1919. 1929a. 1929b. 1929c. 1930a. 1930b. 1930c. 1931a. 1931b. 1931c. 1931d. 1936. 1940. The crane-flies of New York. Part I. Distribution and taxonomy of the adult flies. Cornell Univ. Agr. Exp. Sta., Memoir 25:766-993, plates XXX-LV. . The crane-flies of New York. Part II. Biology and phylogeny. Cornell Univ. Agr. Exp. Sta., Memoir 38:691-1133, plates XII- XCVII. . New species of crane-flies from southern Indiana. Occ. Papers Mus. Zool. Univ. Mich., 127:1-7. .The crane-flies of New York: first supplementary list. Bull. Brooklyn Ent. Soc., 17 (2) :58-62. . The crane-flies of New York: second supplementary list. Bull. Brooklyn Ent. Soc., 19(3):57-64. The crane-flies (Tipulidae) of New England: first supplementary list. Occ. Papers Boston Soc. Nat. Hist., 5: 169-174. Records of crane-flies (Tipulidae) from Ontario. Canadian En- tomologist, 58(10):236-240. Records of crane-flies (Tipulidae) from Ontario. Part II. Cana- dian Entomologist, 60(3 ) :54-60. A list of the crane-flies of Quebec. Part I. Canadian Entomologist, 61(11):2381-236, 247-251. The crane-flies of New York: third supplementary list. Bull. Brooklyn Ent. Soc., 24(1):22-29. The crane-flies of New York: fourth supplementary list. Bull. Brooklyn Ent. Soc., 24(5):295-302. The crane-flies (Tipulidae) of New England: third supplementary list. Occ. Papers Boston Soc. Nat. Hist., 5:267-278. New or insufficiently-known crane-flies from the Nearctic Region (Tipulidae, Diptera). Part II. Bull. Brooklyn Ent. Soc., 25(5): 276-282. Tipulidae and Ptychopteridae. In Curran, C. H., Report on the Diptera collected at the Station for the Study of Insects, Harriman Interstate Park, N. Y. Bull. Amer. Mus. Nat. Hist., 61(2):118-115. A list of the crane-flies of Quebec. Part II. Canadian Entomolo- gist, 63(6) :135-147. New or little known Tipulidae from the Philippines (Diptera). Part XI. Philippine Jour. Sci., 46(2):269-304, plates 1-3. New or insufficiently-known crane-flies from the Nearctic Region (Tipulidae, Diptera). Part III. Bull. Brooklyn Ent. Soc., 26(4): 177-184. Crane-flies of the Baltic Amber (Diptera). Bernstein-Forschungen (Amber Studies), Heft 2, ed. by K. Andree. Berlin and Leipzig: Gruyter and Co. 135 pp., illus. The crane-flies (Tipulidae) of New England: fourth supplemen- tary list. Occ. Papers Boston Soc. Nat. Hist., 8:273-292. Records and descriptions of North American crane-flies (Diptera ). Part I. Tipuloidea of the Great Smoky Mountains National Park, Tennessee. Amer. Midland Nat., 24(3):602-644. THe CraNE Fiy Genus DOLICHOPEZA 921 1941a. Records and descriptions of North American crane-flies (Diptera), Part II. Tipuloidea of mountainous western North Carolina. Amer. Midland Nat., 26(2):281-319. 1941b. Undescribed species of crane-flies from the eastern United States and Canada. Part VII. Ent. News, 52(7):192-196. 1942. Family Tipulidae. In The Diptera or true flies of Connecticut. Conn. State Geol. and Nat. Hist. Surv., Bull. 64:196-509. 1944. Undescribed species of crane-flies from the eastern United States and Canada. Part IX. Ent. News, 55(9) :241-247. ALEXANDER, CHARLES P., and W. L. MCATEE. 1920. Diptera of the Superfamily Tipuloidea found in the District of Columbia. Proc. U. S. Nat. Mus., 58:385-435, plate 26. BELING, THEODOR. 1886. Dritter Beitrag zur Naturgeschichte (Metamorphose ) verschiedener Arten aus der Familie der Tipuliden. K.-K. Zool.-Bot. Gesell- schaft, Wien, Verh. 36 Abhandl.:171-214. Brown, W. L., and E. O. WILson. 1956. Character displacement. Syst. Zool., 5(2):49-64. BRUNETTI, E. 1912. Diptera Nematocera (excluding Chironomidae and Culicidae). In The Fauna of British India, as one volume in a series. London: Taylor and Francis. 581 pp., 12 plates. Byers, GEORGE W. 1958. Species recognition in immature crane flies (Diptera: Tipulidae). Proc. Tenth Int. Congress Ent., 1956, 1:131-136. Comstock, JOHN H., and V. L. KELLocc. 1902. The elements of insect anatomy. Fourth edition, revised. Ithaca, New York: Comstock Publ. Co. 145 pp., illus. Cook, Epwin F. 1949. The evolution of the head in the larvae of the Diptera. Microen- tomology, 14(1):1-57. CRAMPTON, G. C. 1942. The external morphology of the Diptera. In The Diptera or true flies of Connecticut. Conn. State Geol. and Nat. Hist. Surv., Bull. 64:10-165. CurtTIs, JOHN. 1825. Dolichopeza sylvicola. In British Entomology, as plate 62 of a series of 770. London: published by the author, 1823-1840; pp. 1-2, plate 62. Dickinson, W. E. 1932. The crane-flies of Wisconsin. Bull. of the Public Museum of the City of Milwaukee, 18(2):139-266. Dietz, WILLIAM G. 1921. A list of the crane-flies taken in the vicinity of Hazleton, Pennsyl- vania. Trans. Amer. Ent. Soc., 47:233-268. Epwarps, F. W. 1938. British short-palped craneflies. Taxonomy of adults. Trans. Soc. British Ent., 5(1):1-168, plates I-V. 922 THe UNIVERSITY SCIENCE BULLETIN Emopen, I. Frirz vAN, and W1L.L1 HENNIG. 1956. Diptera. In Tuxen, S. L., Taxonomist’s glossary of genitalia in insects. Copenhagen: E. Munksgaard. 284 pp., illus. Foote, BENJAMIN A. 1956. A preliminary survey of the crane-flies of Delaware County, Ohio (Diptera: Tipuloidea). Ohio Jour. Sci., 56(4):217-229. Ina (On /a\e 1895. On the entomology of the Illinois River and adjacent waters. First paper. Bull. Illinois State Lab. of Nat. Hist., 4(6):149-273, plates I-XV. HENNIG, WILLI. 1950. Die Larvenformen der Dipteren. Part II. Berlin: Akademie-Ver- lag. 458 pp., illus. HusBELL, THEODORE H. 1936. A monographic revision of the genus Ceuthophilus. Univ. of Fla. Publ. (Biol. Sci. Series), 2(1):1-551, plates I-XXXVIII. JOHNSON, CHARLES W. 1909. New and little known Tipulidae. Proc. Boston Soc. Nat. Hist., 84 (5):115-133, plates 15-16. 1910. The Diptera. In Smith, John B., The insects of New Jersey. Rept. of the New Jersey State Mus.: 703-814. 1925. Fauna of New England: 15. List of the Diptera or two-winged flies. Occ. Papers Boston Soc. Nat. Hist., 7: 1-326. 1929. Diptera of Labrador, Psyche, 36(2):129-146. Jupp, W. W. 1958. Studies of the Byron Bog in southwestern Ontario: IX. Insects trapped as adults emerging from Redmond’s Pond. Canadian Entomologist, 90(10):623-627. KEUCHENIUS, P. E. 1913. The structure of the internal genitalia of some male Diptera. Zeits- chrift fiir wissenschaftliche Zoologie, Band 105:501-536. LAVEN, HANNES. 1957. Vererbung durch Kerngene und das Problem der ausserkaryotischen Vererbung bei Culex pipiens. I. Kernvererbung. Zeitschrift fiir indukt. Abstammungs- und Vererbungslehre, 88:443-477. LEONARD, M. D. 1928. A list of the insects of New York. Cornell Univ. Agr. Exp. Sta., Memoir 101:1-1121. (Tipuloidea, pp. 688-701. ) MA.LiLocn, JOHN R. 1917. A preliminary classification of Diptera, exclusive of Pupipara, based upon larval and pupal characters, with keys to imagines in certain families. Part I. Ball. Illinois State Lab. of Nat. Hist., 12(3): 161-409, plates XXVIII-LVII. MANNHEIMS, BERNHARD. 1951. Tipulidae. In Lindner, Erwin, Die Fliegen der palearktischen Region, Lieferung 167, Band 3(15):1-64 (continued in Lief. 170 and 178, 1952-53). Tue CRANE FLY GENUS DOLICHOPEZA 923 MEIGEN, JOHANN W. 1818. Systematische Beschreibung der bekannten europiischen zweifliige- ligen Insekten, 1:1-269. Mratxi, Louris C. 1893. Dicranota: a carnivorous tipulid larva. Trans. Ent. Soc. of London, Part III: 235-253. NEEDHAM, JAMES G. 1908. Report of the entomologic field station conducted at Old Forge, N. Y., in the summer of 1905. 23d Rept. of the State Entomologist, 1907: 156-248, plates 4-32. OsTEN SACKEN, CHARLES R. 1878. Catalogue of the described Diptera of North America. Second edition. Smithsonian Misc. Publ. 16(270): 1-276. 1886. Studies on Tipulidae. Part I. Review of the published genera of the Tipulidae longipalpi. Berliner Entomo!og. Zeitschrift, 30(2): 153-188. PETERSON, ALVAH. 1916. The head capsule and mouth-parts of Diptera. Illinois Biol. Monogr., 3(2):1-112, plates 1-25. 1951. Larvae of insects. Part II. Coleoptera, Diptera, Neuroptera, Si- phonaptera, Mecoptera, Trichoptera. Columbus, Ohio: published by the author. 416 pp., illus. PIERRE, CLAUDE. 1926. Fam. Tipulidae, Subfam. Tipulinae. In Genera Insectorum, Dip- tera, Fasc. 186: 1-68, plates 1-5. Regs, B. E., and G. F. Ferris. 1939. The morphology of Tipula reesi Alexander (Diptera, Tipulidae). Microentomology, 4(6):143-178, figs. 72-91. Rocers, J. SPEED. 1930. The summer crane-fly fauna of the Cumberland Plateau in Ten- nessee. Occ. Papers Mus. Zool. Univ. Mich., 215:1-50, plates 1-5. 1933. The ecological distribution of the crane-flies of northern Florida. Ecol. Monogr., 3:1-74. 1942. The crane-flies (Tipulidae) of the George Reserve, Michigan. Univ. Mich. Mus. Zool. Misc. Publ. No, 53:1-128, plates 1-8. 1949. The life history of Megistocera longipennis (Macquart) (Tipulidae, Diptera), a member of the neuston fauna. Occ. Papers Mus. Zool. Univ. Mich., 521:1-14, plates 1-2. Ross, HERBERT H. 1957. Principles of natural coexistence indicated by leafhopper popula- tions. Evolution, 11(2):113-129. Say, THOMAS. 1823. Descriptions of dipterous insects of the United States. Jour. Acad. Nat. Sci. Philadelphia, 3:9-54, 73-104. SNopcRASS, ROBERT E. 1903. The terminal abdominal segments of female Tipulidae. Jour. New York Ent. Soc., 11(4):177-183, plates 1-2. 1904. The hypopygium of the Tipulidae. Trans. Amer. Ent. Soc., 30: 179-236, plates 1-11. 924 Tue UNrversiry SCIENCE BULLETIN 1957. A revised interpretation of the external reproductive organs of male insects. Smithsonian Misc. Collections, 1385(6):1-60. TAKAHASHI, MITSUO. 1958. On the pupae of three Japanese species of the genus Oropeza Need- ham (Diptera, Tipulidae). Kontyu, 26(3):118-123. TJEDER, Bo 1948. The Swedish Prionocera. Opuscula Ent., 13:75-99. ToKuNAGA, MASAAKI. 1930. The morphological and biological studies on a new marine cranefly, Limonia (Dicranomyia) Monostromia, from Japan. Mem. Coll. of Agr., Kyoto Imper. Univ., No. 10: 1-93, plates I-XVII. VAN EMDEN—see Emden. WaARDLE, ROBERT A. 1926a. The cephalic skeleton of contrasting types of cranefly larvae. Proc. Zool. Soc. London. Part I, No. 1: 1-28. 1926b. The respiratory system of contrasting types of cranefly larvae. Proc. Zool. Soc. London. Part I, No. 1: 25-48. Weiss, Harry B., and G. M. ZrecLER. 1931. Thomas Say, early American naturalist. Springfield, Illinois: C. M. Thomas Co. 260 pp., illus. Wuittaker, R. H. 1952. A study of the summer foliage insect communities in the Great Smoky Mountains. Ecol. Monogr., 22(1):1-44. WIEDEMANN, C. R. W. 1828. Aussereuropiische zweifliigelige Insekten. Vol. 1. Hamm. 608 pp., 7 plates. Woop, H. G. 1952. The crane-flies of the South-West Cape (Diptera, Tipuloidea). Ann. South African Mus., 39:1-827. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vout. XLII] DECEMBER 29, 1961 [No. 7 Summary of Fossil Microfloral Investigations in the Bevier, Weir-Pittsburg, Lower Williamsburg and Blue Mound Coals of Eastern Kansas BY Sam L. Van LANDINGHAM Axsstract: The following report concerns the identification and description of microflora of the Bevier and Weir-Pittsburg Coals (Des Moinesian Series of Pennsylvanian System) and the Lower Williamsburg and Blue Mound Coals (Virgilian Series of Pennsylvanian System). No attempt is made to correlate any coal beds. This is simply a preliminary report dealing with some of the many types of plant cuticles and other types of plant microfossils present in the Cherokee and Douglas Group Coals. I feel that much more work should be done before any attempt at correlation on a broad scale is attempted. INTRODUCTION The following presentation was done with the aid of a National Science Foundation Research Grant, under the direction of Dr. R. W. Baxter, Department of Botany, University of Kansas, and con- stitutes part of a general study of the coal-age flora of eastern Kansas. The samples studied are described as follows: Sample 19: Bevier Coal from core sample from Mack Colt Oil Company No. 14a Colt-Alexander; 1.95 feet long, extracted from a depth of 994.85 feet; from NW% SW% Sec. 9, T. 23 S., R. 18 E., Anderson County, Kansas; “a” is “<¢9> bottom portion of sample, “b” is middle part and “c” is top. Sample 7: Weir-Pittsburg Coal from NE Sec. 24, T. 32 S., R. 23 W., Cherokee County, Kansas; “a” is lower 2 feet, “b” is middle 1.5 feet and “c” is top 1.3 feet. Sample 6: Weir-Pittsburg Coal from Sec. 7, T. 33 N., R. 33 W., Barton County, Missouri; “a” is lower 10 inches, “b” is middle 10 inches and “c is upper 10 inches. Sample 4: Weir-Pittsburg Coal from Blue Ribbon Mine in Sec. 25, T. 29 S., R. 25 E., Crawford County, Kansas; “a” is lower portion, “b” is middle portion and “c” is upper portion. (925) 32—5840 926 THe University SCIENCE BULLETIN Blue Mound: Blue Mound Coal from NE of NWii of Sec. 28, T. 18 S., R. 20 E., Douglas County, Kansas; site was approximately 100 feet NE of the Coal (Cole) Creek Bridge on the north side of the Creek. (Plate 29.) Lone Star Lake: Lower Williamsburg Coal? from NW corner Sec. 13, T. 14 S., R. 18 E., Douglas County, Kansas; outcrop thickness below the Lone Star Lake Spillway was approximately 18 inches of coal and carbonaceous shale; bottom sample included lowermost 5% inches, middle sample included next 2 inches and upper sample included next 3 inches (see plate 29). The bottom sample was coal, and the middle and upper samples were coal and carbonaceous shale. (Plates 26-28. ) The sample numbers correspond to slide numbers mentioned on plates 1-30. Most of the slides mentioned on plates 1-30 are on file with Dr. R. W. Baxter: Department of Botany, University of Kansas, Lawrence, Kansas. PREPARATION OF SLIDES The coal samples were washed and crushed to particles approxi- mately one cm. in diameter and placed Schulze’s solution (saturated aqueous solution of KCLO, plus two or three parts of cold, con- centrated HNO,). After approximately 24 hours in the Schulze’s solution, the samples were washed and siphoned six times, allowing one and one-half hours for each settling between siphonings. The samples were then placed in a 7% KOH solution and allowed to remain for one hour. They were then washed and siphoned eight times, allowing two hours for each settling between each siphon- ing. The samples were separated through a 625 micron (40 meshes per inch) screen. The smaller fraction (— 625 microns) was placed in successive alcohol baths of 5, 10, 25, 50, 75 and 95%, each for a total of 30 minutes. The treated samples were stored in absolute alcohol. Small amounts of the samples were placed on glass cover slips. Diaphane was then added, and this mixture was spread over the glass slip and allowed to dry. Balsam was then applied to a glass slide, and the glass slip was placed on the balsam. No staining was attempted. The larger fraction (+ 625 microns) was dried and examined for megaspores (mainly Triletes). The large spores were mounted in cardboard microslides. The cuticles were hand picked from the larger fraction and mounted with balsam. Some of the larger pieces of cuticle were hand picked from the samples and treated with household bleach for 154 hours until all of the debris was removed. Fosst MiICROFLORAL INVESTIGATIONS 927 TENTATIVE LIST OF PLANT MATERIALS IN BEVIER, WEIR-PITTSBURG, LOWER WILLIAMSBURG AND BLUE MOUND COALS The following lists are arranged in order of decreasing im- portance. Both frequency of occurrence and total bulk were taken into consideration. Cuticles with unknown affinities are not in- cluded in these listings. In every case the Cordaitales were most important. Bevier Coal, Plates 2-10, 22 1. Cordaitales Cordaites-type fragments (abundant) Florinites (common ) Cordaianthus fragments (rare ) 2. Lepidodendrales Lepidodendron-type fragments (abundant to common) Lycospora (abundant, most frequent microspore ) Triletes (rare) 8. Spores Incertae sedis Punctatosporites (abundant ) Laevigatosporites (common) Latosporites (common ) Verrucososporites (sparce ) Triquitrites (sparse ) Anapiculatisporites (sparse ) Schopfites (rare ) Leiotriletes (rare) Vestispora (rare) 4. Filicineae (ferns) Punctatisporites (abundant ) Granulatisporites (rare ) Raistrickia (rare) 5. Calamitales Calamospora (common ) 6. Medulloseae (and other seed ferns) Monoletes (common ) Medullosa-type fragments (rare) Dolerotheca? (rare, only one fragment was found ) Weir-Pittsburg Coal, Plates 2-5, 9, 10, 23 1. Cordaitales Cordaites-type fragments (abundant ) Florinites (abundant) Endosporites (common to abundant) Cordaianthus fragments (rare) Pityosporites (rare, only two were found) 928 THE UNIVERSITY SCIENCE BULLETIN 2. Lepidodendrales Lepidodendron-type fragments (common to rare ) Lycospora (common) Triletes (common) 8. Filicineae (ferns ) Punctatisporites (abundant ) Granulatisporites (rare ) Raistrickia (rare, only one was found) 4. Medulloseae (and other seed ferns ) Monoletes (common) Medullosa fragments (rare ) 5. Calamitales Calamospora (rare) Lower Williamsburg Coal, Plates 11-21, 24, 25 1. Cordaitales Cordaites-type fragments (abundant ) Cordaianthus fragments (abundant ) Florinites (abundant) Endosporites (abundant ) Wilsonia (rare ) bo Lepidodendrales Triletes (abundant) (some may be related to Sigillariostrobus ) Lepidodendron-type fragments (common ) Lepidocarpon megaspores (common ) Lycospora (common to abundant ) 8. Spores Incertae sedis Triquitrites (common to abundant) Laevigatosporites (common to abundant ) Reticulatisporites (common ) Latosporites (common ) Leiotriletes (common to rare ) Microreticulatisporites (common to rare ) Pustulatisporites (rare) Cyclogranisporites (rare ) Vesicaspora (rare) And etc. 4. Filicineae (ferns) Punctatisporites (abundant) Granulatisporites (common to rare ) Raistrickia (rare ) 5. Calamitales Calamospora (common) 6. Medulloseae (and other seed ferns ) Monoletes (common ) Neuropteris fragments (rare) (mainly N. scheuchzeri) Medullosa fragments (rare) Fosstz MicROFLORAL INVESTIGATIONS 929 Blue Mound Coal, Plates 11-16, 21 1. Cordaitales Cordaites type fragments (abundant ) Florinites (abundant) Endosporites (common to abundant ) Cordaianthus fragments (rare to common ) 2. Lepidodendrales Triletes (abundant) (some may be related to Sigillariostrobus ) Lycospora (common to abundant) Lepidodendron type fragments (common) Lepidocarpon megaspores (rare ) 3. Spores Incertae sedis 4. Filicineae (ferns) Punctatisporites (abundant ) Granulatisporites (common to rare ) Raistrickia (rare) 5. Calamitales Calamospora (rare to common ) 6. Medulloseae (and other seed ferns ) Monoletes (common ) Medullosa type fragments (rare ) CONCLUSION Coal is a natural occurrence, and strata are natural occurrences. Fossils also occur naturally. All plant fossils must be studied from a natural standpoint in order to obtain a natural reconstruction of the past. The artificial method of studying plant microfossils is that method by which spores and other plant fragments are used for correlative purposes with little or no consideration of the plants from which they came. Any artificial approach to the study of plant fossils produces an artificial picture and also an incomplete picture of the past. The artificial method is perhaps the most efficient, but the most efficient method is not always the best method. Much of geology and paleontology is based upon the phrase “The present is the key to the past.” The present is natural, not artificial. Is it not a logical procedure to interpret the natural present in terms of the natural past? This short paper is an attempt at the natural approach to the study of plant microfossils. One good natural procedure is to use the spores to determine stratigraphic position, the associated cuticles and other fragments to correlate these stratigraphic positions, and the cuticles and spores together to determine and recon- 930 THe UNIVERSITY SCIENCE BULLETIN struct the plants from which they came. The reconstructed plant is the key to the environment. A great deal more work must be done before a procedure of this type would be possible. Fossil spores are very useful in determining the stratigraphic positions of coal samples, but in correlating several coal samples, the amounts of the various spore genera and species vary from sample to sample. Plant cuticle types are usually present in fairly consistent amounts from coal samples of the same age. In many cases fossil plant cuticles and other tissues exclusive of spores are much better to do correlative work with than spores. A person needs only to look at Wilson and Hoffmeister’s (1956) cuticle histo- grams on plate 2 of that paper to see this. A better understanding of fossil plant cuticles associated with fossil spores would mean a better, more natural, and more complete understanding of the past than the now popular artificial system gives us. ACKNOWLEDGMENTS I thank Dr. R. W. Baxter for his aid and assistance. Also, thanks to Arthur Cridland and John Morris for critical evaluation. I am also grateful to Al Hornbaker and Dan Habib of the Kansas Geological Survey for their help on the Bevier and Weir-Pittsburg Coals. LITERATURE USED ANDREWS, H. N. 1942. Contributions to our knowledge of American Carboniferous floras, I, Scleropteris gen nov., Mesoxylon and Amyelon. Missouri Botanical Garden Annals, vol. 29, p. 275-282. ARNOLD, C, A. 1931. Cordaitean wood from the Pennsylvanian of Michigan and Ohio. Botanical Gazette, vol. 91, p. 77-87. 1940. Lepidodendron johnsonii, sp. nov., from the Lower Pennsylvanian of central Colorado. Contributions Michigan University Museum Paleontology, vol. 6, p. 21-52. 1947. An Introduction to Paleobotany. McGraw-Hill Book Co., New York, 433 p. BARTLETT, H. H. 1929. Fossils of the Carboniferous coal pebbles of the glacial drift at Ann Arbor. Michigan Academy Science, Arts and Letters Papers, vol. 9, p. 11-28. Baxter, R. W. 1951. Coal Balls—new discoveries in plant petrifactions from Kansas coal beds. Transactions Kansas Academy Science, vol. 54, p. 1-10. BarcuHoorwn, E. S., and Scott, R. A. 1958. Degradation of the plant cell wall and its relation to certain tracheary features of the Lepidodendrales, American Journal of Botany, vol. 45, p. 222-227. Fossa. MicROFLORAL INVESTIGATIONS 931 Bowsuenr, A. L., and JEwetrT, J. M. 1943. Coal resources of the Douglas group in east-central Kansas. Kansas Geological Survey Bulletin 46, 94 p. CHALONER, W. G. 1953. On the megaspores of Sigillaria. Ann. and Magazine of Natural History, vol. 6, p. 881-897. CrRIDLAND, ARTHUR. 1960. Fragmentary remains of a lycopod from the Desmoinesian Series (Pennsylvanian) of Southeastern Kansas. University of Kansas Science Bulletin, vol. 40, p. 3-19. Dixstra, S. J. 1958. On a megaspore-bearing lycopod strobilus. Acta Bot. Neerlandica, vol. 7, p. 217-222. FELIX, C. J. 1952. A study of the arborescent lycopods of Southeastern Kansas, Mis- souri Botanical Garden Annals, vol. 39, p. 263-289. FLorIn, R. 1931. Untersuchungen zur Stammesgeschichte der Coniferales und Cor- daitales. K. Svensk. Vetenskapsakad. Hand., ser. 3, vol. 10, Stockholm. Granp Eury, F. C. 1877. Mémoire sur la flore Carbanifere du department de la Loire et du centre de la France. Academie Science Institution France Mémoire, tome 24, 624 p. GUENNEL, G. K. 1952. Fossil spores of the Alleghenian Coals in Indiana. Indiana Geo- logical Survey Report Progress No. 4, 40 p. Hasis, DANIEL. 1960. Palynological correlation of the Bevier and Wheeler Coals. Uni- versity of Kansas Master’s Thesis. Harris, T. M. 1956. The fossil plant cuticle. Endeavour, vol. 15, p. 210-214. Howe, W. B. 1956. Stratigraphy of Pre-Marmaton Desmoinesian (Cherokee) rocks in Southeastern Kansas. Kansas Geological Survey Bulletin 128, 132 p. JANSSEN, R. E. 1939. Leaves and stems from fossil forests. Illinois State Museum Popu- lar Science Series, vol. 1, 190 p. KANSAS GEOLOGICAL SURVEY. 1957. Guidebook of Second Annual Science and Mathematics Camp Geologic Field Conference between Lawrence and Wyandotte County Park in Northeastern Kansas. Kansas Geological Survey, 39 p. KosAnkE, R. M. 1950. Pennsylvanian spores of Illinois and their use in correlation. Illinois Geological Survey Bulletin 74, 128 p. Lins, T. W. 1950. Origin and environment of the Tonganoxie Sandstone in North- eastern Kansas. Kansas Geological Survey Bulletin 86, part 5, p. 105-140. 932 Tue UNrversiry SCIENCE BULLETIN Moore, R. C., e¢ al. 1951. The Kansas rock column. Kansas Geological Survey Bulletin 89, 182 p. POTONIE, ROBERT. 1954a. Stellung der paliozoischen sportengattungen in natirlichen system. Paliont. Zeitschr., Band 28, Heft 3-4, p. 103-139. 1954b. Les spores des plantes paléozoiques dans les systéme naturel (morphologique). Lejeunia, Revue de Botanique, tome 18, p. 5-20. Potoniz, ROBERT, and KrEMpP, GERHARD. 1954. Die Gattungen der paliozoischen sporae dispersae und ihre strati- graphie. Geol. Jahrb., Band. 69, p. 11-194. 1955. Die sporae dispersae des Ruhrkarbons, ihre morphographie und stratigraphie mit Ausblicken auf Arten anderer Gebiete und Zeitab- schnitte-Teil I. Palaeontographica, Abt. B, Band 98, Lief. 1-3, p. 1-136. 1956. Die sporae dispersae des Ruhrkarbons, ihre morphographie und stratigraphie mit Ausblicken auf Arten anderer Gebiete und Zeitab- schnitte-Teil II. Palaeontographica, Abt. B, Band 99, Lief. 4-6, p. 85-191. REED, F. D. 1941. Coal flora studies: Lepidodendrales. Botanical Gazette, vol. 102, p. 663-683. REED, F. D., and SANDOE, M. T. 1951. Cordaites affinis; a new species of Cordaitean leaf from American coal fields. Torrey Botanical Club Bulletin, vol. 78, p. 449-457. Scuopr, J. M. 1938. Spores from the Herrin (No. 6) Coal bed in Illinois. Illinois Geological Survey Report of Investigations No. 50, 73 p. 1948. Pteridosperm male fructifications: American species of Dolerotheca, with notes regarding certain allied forms. Journal of Paleontology, vol. 22, p. 681-724. Scuopr, J. M., Witson, L. R., and BENTALL, Ray. 1944. An annotated synopsis of Paleozoic fossil spores and the definition of generic groups. Illinois Geological Survey Report of Investiga- tions No. 91, 73 p. Scort, J. H. 1920-1923. Studies in fossil botany, vols. I and II, London. SEwapp, A. C. 1898-1919. Fossil Plants, vols. I-IV, Cambridge. Stopss, M. C. 1903. On the leaf structure of Cordaites. New Phytology, vol. 2, p. 91-98. UNGER, FRANZ 1850. Genera et species plantarum fossilium. Vienna, 627 p. WALTON, JOHN. 1953. An introduction to the study of fossil plants. Adam and Charles Black, London, 201 p. WuitrForp, A. C. 1916. Preserved epidermis from the Carboniferous of Nebraska. Ne- braska Geological Survey, volume 7, p. 93-101. Fossa. MICROFLORAL INVESTIGATIONS 933 Wait ta, R. E. 1940. Coal resources of Kansas. Post-Cherokee deposits. Kansas Geo- logical Survey Bulletin 32, 64 p. WInsLow, M. R. 1959. Mississippian and Pennsylvanian megaspores and other plant micro- fossils from Illinois. Illinois Geological Survey Bulletin 86, 135 p. Wison, L. R., and HorrMeIstEr, W. S. 1956. Plant microfossils of the Croweburg Coal. Oklahoma Geological Survey Circular 32, 57 p. ZERNDT, JAN. 1937. Megasporen aus dem Westfal und Stefan in Bohmem. Bull. Intern. Acad. Polon. Sci., ser. A, p. 583-599. 934 Tue UNIversiry SCIENCE BULLETIN PLATE 1 Fic. 1. Idealized stratigraphic column of Southeastern Kansas showing position of Cherokee Group coals (Des Moinesian Series of Pennsylvanian System). One inch equals approximately 45 feet. Fic. 2. Idealized stratigraphic column in Douglas County, Kansas, show- ing position of Douglas Group coals ( Virgilian Series of Pennsylvanian System). One inch equals approximately 120 feet. Fosstz MICROFLORAL INVESTIGATIONS PLATE 1 Marmaton group Blackjack Cr.- Mulky Breezy Hill gear Cyclothem Bevier Cyclothem Upper Williamsburg coal Bevier pede Lawrence oo Williamsburg Shale coal a = Ardmore ow Cyclothem Bebe Upper Oo Sibley Fleming Fleming coal coal u| Cyclothem Lower — Stranger ey a ae Sib| | Mineral eae Formation OS} Cyclothem | Mineral or coal Blue uJ Mound 5 Scammon | scammon coal Cyclothem coal -------- Lansing 7 Group Cyclothem Pilot” oe Wyene) eir coo Kansas Cyclothem| Weir- i Pittsburg City coal Blue) acket Ss. Group 935 GROUP DOUGLAS 936 Tue UNrversiry SCIENCE BULLETIN PLATE 2 Fic. 1. Longitudinal view of Lepidodendron scalariform tracheid of primary xylem; common in Bevier Coals, especially in upper levels; not common in the Weir-Pittsburg Coal; Slide 19a No. 2, Bevier Coal; 59. Fic. 2. Enlargement of area framed in black on figure 1; note fimbrils (“threads”) extending across pit openings (white arrow); X 271. Fic. 3. Tracheids of Cordaitean affinities; common in both Bevier and Weir-Pittsburg Coals; Slide 19a No. 2, Bevier Coal; x 59. Fic. 4. Enlargement of area framed in black on figure 3; x 271. Fic. 5. Tracheids probably associated with some Cordaitean-type plant; pits are elliptical-shaped and in some of the tracheids there are only two or three rows of pits while in others there are as many as seven rows; common in Weir-Pittsburg Coal samples; Slide 6a No. 4, Weir-Pittsburg Coal; x 59. Fic. 6. Enlargement of area framed in black on figure 5; X 271. 937 Fossi. MICROFLORAL INVESTIGATIONS PLATE 2 938 Tue UNIVERSITY SCIENCE BULLETIN PLATE 3 Fic. 1. Wood fragment from the transition between primary and secondary xylem of Cordaites? tracheids; (Cordaites metaxylem with large spiral tra- cheids?); fairly common in the Weir-Pittsburg samples; Slide 4c No. 36, Weir- Pittsburg Coal; < 271. Fic. 2. Tracheid of Lepidodendron; note fimbrils (white arrow); Slide 19a No. 2, Bevier Coal; 271. Fic. 3. Wood fragments of Cordaitean or Lycopodian affinities; upper left —Cordaites? scalariform tracheids; center—Cordaites? metaxylem with spiral tracheids; common in Weir-Pittsburg and Bevier coals; Slide 4c No. 34, Weir- Pittsburg Coal; 59. Fic. 4. Unidentified wood fragment; Slide 4a No. 26, Weir-Pittsburg Coal; x59: Fic. 5. Same; Slide 19a No. 2, Bevier Coal; 59. Fic. 6. Same; Slide 4a No. 26, Weir-Pittsburg Coal; x 59. Fic. 7. Lepidodendron? tracheids; Slide 7c No. 25, Weir-Pittsburg Coal; < 59. Fic. 8. Unidentified wood fragment; Slide 7a No. 16, Weir-Pittsburg Coal; < 59. Fic. 9. Same; Slide 4c No. 26, Weir-Pittsburg Coal; 59. Fosst. MICROFLORAL INVESTIGATIONS 939 PLATE 3 940 Tue UNIVERSITY SCIENCE BULLETIN PLATE 4 AutTHoR’s Notre: In all cases in the following plate explanations, the word, cell, refers to the imprint or shape of the plant cells which are preserved in the cuticle. Fic. 1. Tracheid of Medullosa?; rare in all coals studied in this report; Slide 7a No. 16, Weir-Pittsburg Coal; 271. Fic. 2. Same; Slide 19b No. 8, Bevier Coal; 271. Fic. 3. Same; Slide 19a No. 2, Bevier Coal; x 271. Fic. 4. Unidentified tracheids that may be Lycopodian; Slide 19a No. 2, Bevier Coal; < 271. Fic. 5. Cortical strands of Lycopodian affinities; Slide 7c No. 22, Weir- Pittsburg Coal; < 271. Fic. 6. Cuticle; diamond-shaped cells averaging 15-20 microns on a side; may be caused by one portion of cuticle with rectangular cells being folded over another portion of the same type of cuticle in a diagonal manner; Slide 19a No. 2, Bevier Coal; x 59. Fic. 7. Same; slide 6a No. 40, Weir-Pittsburg Coal; 271. Fic. 8. Cuticle with polygonal-shaped cells; cell walls are rather thick; may be from a megaspore similar to figure 1, plate 24; Slide 19a No. 2, Bevier Coal; x 271. Fic. 9. Cuticle with wavy, elongated cells averaging approximately 110- 120 microns in length by 10-20 microns in width; found only in the Wier-Pitts- burg, rare; Slide 6a No. 40, Weir-Pittsburg Coal; 59. Fic. 10. Cordaites cuticle; dense, narrow rows of cells; Slide 19a No. 4, Bevier Coal; 59. Fic. 11. Similar to Wilson and Hoffmeister’s (1956) Cuticle type B; film- like and apparently structureless; network of pits is not as apparent as in figure 1, plate 6; approximately one micron thick; usually much folding is present; very similar to figure 2 on plate 6, but no perforations are present in this specimen; sparcely distributed throughout the Bevier and Weir-Pittsburg samples; Slide 19c No. 11, Bevier Coal; x 59. Fic. 12. Similar to Winslow’s (1959) Plate 16, figure 4; resembles cellular pattern from lower epidermis of Cordaites leaf; both lateral and terminal cells are shared by each stomate; stomata occur in rows of approximately 6; lateral cells are broader and shorter than ordinary epidermal cells and are approxi- mately 50-60 microns in length by 25-33 microns in width; terminal cells are nearly always round, but sometimes oval and average 10-18 microns in diam- eter; ordinary epidermal cells are approximately 75-85 microns in length by 15-25 microns in width and are distinctly rectangular; Slide 4a No. 26, Weir- Pittsburg Coal; x 59. Fic. 18. Cordaites cuticle similar to figure 12; Slide 6b No. 42, Weir-Pitts- burg Coal; x 59. Fic. 14. Cordaites cuticle; see figure 1, plate 23 for description; Slide 4a No. 26 Weir-Pittsburg Coal; 59. Fossm. MICROFLORAL INVESTIGATIONS 94] PLATE 4 942, Tue UNtversiry SCIENCE BULLETIN PLATE 5 Fic. 1. Cuticle with perforations 45-50 microns in diameter surrounded by polygonal-shaped cells averaging approximately 25-30 microns on a _ side; Slide 4a No. 26, Weir-Pittsburg Coal; 59. Fic. 2. Similar to Winslow's (1959) plate 16, figure 5; cuticle with open- ings averaging approximately 30-34 microns in diameter and which are sur- rounded by polygonal-shaped cells; these openings may be stomatal in nature; rare; Slide 19a No. 4, Bevier Coal; & 59. Fic. 3. Similar to Winslow’s (1959) plate 16, figure 7; cuticle has dome cells and cuticular crests and resembles some of Bartlett’s (1929) plates; prob- able affinity is Lepidodendron; rare in all coals studied in this report; Slide 7c No. 25, Weir-Pittsburg Coal; « 59. Fic. 4. Cuticle with conspicuous perforations approximately 28-33 microns in diameter, surrounded by polygonal-shaped cells averaging 25-30 microns on a side; may be associated with such genera as Dolerotheca, male fructifica- tion of Whittleseyinae; rare; for more complete description see figure 1, plate 22; Slide 19a No. 2, Bevier Coal; < 271. Fic. 5. Cuticle with elongated cells associated with long strings of variously- sized perforations; cells are approximately 95-100 microns long by 25-30 microns wide; Slide 6a No. 37, Weir-Pittsburg Coal; 271. Fic. 6. Same; Slide 19c No. 11, Bevier Coal; x 59. Fic. 7. Same; Slide 19c No. 11, Bevier Coal; « 59. Fosstt MicROFLORAL INVESTIGATIONS 943 PLATE 5 944 Tue UNIversiry SCIENCE BULLETIN PLATE 6 Fic. 1. This material is probably not of plant origin; similar to Winslow’s (1959) plate 16, figure 10; characterized by openings from 35-50 microns in diameter surrounded by closely-spaced depressions up to 12 microns in di- ameter; Winslow indicates that this may be membranes of animal origin; sparsely distributed throughout Weir-Pittsburg and Bevier Coals; Slide 19c No. 11, Bevier Coal; 59. Fic. 2. Similar to Wilson and Hoffmeister’s (1956) Cuticle type B; struc- tureless and less than one micron thick; folds are common; perforations are approximately 5-6 microns in diameter; Slide 19c No. 10, Bevier Coal; 271. Fic. 3. Similar to figure 2; perforations are about 10 microns in diameter surrounded by small depressions; Slide 19b No. 9, Bevier Coal; 59. Fic. 4. Irregularly-perforated material; perforations are about 20-25 mi- crons in diameter; Slide 7a No. 15, Weir-Pittsburg Coal; « 59. Fic. 5. Probably a small portion of the type of material pictured in figure 7; Slide 19a No. 5, Bevier Coal; 59. Fic. 6. Same; Slide 19a No. 4, Bevier Coal; « 59. Fic. 7. Similar to Winslow’s (1959) plate 16, figure 10; perforations are approximately 45-50 microns in diameter, surrounded by numerous small depressions approximately 6 microns in diameter; Slide 19b No. 9, Bevier Coal; x 59. Fosst MicROFLORAL INVESTIGATIONS 945 PEATE: G 946 Tue UNIVERSITY SCIENCE BULLETIN PLATE 7 Fic. 1. Leaf cuticle; cell walls are very prominent and approximately 2-3 microns thick; cell shapes are variable but mostly polygonal; Slide 19c No. 10, Bevier Coal; x 59. Fic. 2. Similar to Wilson and Hoffmeister’s (1956) Cuticle type D; stom- atal structures may be present; hexagonal to polygonal cells; walls approxi- mately one micron thick; average cell is approximately 45 microns on a side; Slide 19a No. 2, Bevier Coal; 59. Fic. 8. Cuticle with elongated cells, interfingering at ends; Slide 19a No. 2, Bevier Coal; « 59. Fic. 4. Similar to Wilson and Hoffmeister’s (1956) Cuticle type D; most of the cells are hexagonal or polygonal in shape, occasionally rectangular; walls approximately 1-2 microns thick; cells average approximately 45 microns on a side; most abundant form of cuticle in the Bevier Coal and is also plentiful in the Weir-Pittsburg; Slide 19a No. 2, Bevier Coal; « 59. Fic. 5. Cuticle with extremely irregular shaped cells, with varied wall thicknesses; Slide 19a No. 2, Bevier Coal; « 59. Fic. 6. Same as figure 4; Slide 19a No. 2, Bevier Coal; 59. Fic. 7. Upper surface of leaf?; cuticle with occasional perforations 3.5-4 microns in diameter; cells hexagonal to elongated and have fairly thin walls; note the interfingering effect; sparse distribution in both Weir-Pittsburg and Bevier Coals; Slide 19c No. 14, Bevier Coal; & 271. Fossa. MicROFLORAL INVESTIGATIONS 947 948 THe UNrversiry SCIENCE BULLETIN PLATE 8 Fic. 1. Cuticle with long thin cells associated with circular pits; pits (white arrow) are approximately 7 microns in diameter; cells average approxi- mately 55 by 13 microns; rare, found only in Bevier Coal; Slide 19a No. 2, Bevier Coal; x 390. Fic. 2. Cuticle with irregular polygonal-shaped cells; cells are thin-walled, less than 1 micron thick; cells average approximately 50 microns on a side; fairly common in all coals studied; Slide 19a No. 2, Bevier Coal; x 390. Fossa. MicROFLORAL INVESTIGATIONS PLATE 8 949 950 THE UNIversiry SCIENCE BULLETIN PEATE? Fic. 1. Cordaitean cuticle; cells are oblong, with lengths averaging about 45-95 microns in length by 15-35 microns in width; Slide 19a No. 2; Bevier Coal; x 59. Fic. 2. Similar to Wilson and Hoffmeister’s (1956) Cuticle type A; cells are mostly rectangular; Slide 19a No. 2; Bevier Coal; 59. Fic. 3. Similar to Wilson and Hoffmeister’s (1956) Cuticle type A; probably Cordaitean; cells are long and rectangular and average from 12 to 24 microns in width to 40-75 microns in length; cell walls are approximately 1.5-2 microns thick; one of the most abundant types of cuticle found in the Weir- Pittsburg Coal; Slide 19a No. 4, Bevier Coal; 59. Fic. 4. Similar to figure 3 but cells are much smaller and more narrow; walls are fairly thick; almost identical with specimens known to be Cordaites; Slide 6b No. 47, Weir-Pittsburg Coal; > 59. Fic. 5. Similar to figure 3; Cordaitean; Slide 6b No. 44; Weir-Pittsburg Coal; x 271. Fic. 6. Cuticle with cells extremely long and narrow; some are as long as 250 microns and longer; widths average about 20 microns; cell walls are very thin; Slide 19c No. 10; Bevier Coal; 59. Fic. 7. Cuticle similar to figures 5 and 3; Slide 6a No. 38; Weir-Pittsburg Coal; < 271. Fossaz MicROFLORAL INVESTIGATIONS 951 PLATE: 9 952 Tue University SCIENCE BULLETIN PLATE 10 Fic. 1. Cuticle with cells that are typically square- to slightly rectangular- shaped; cell walls are as much as 7 microns thick; cells average approximately 835 by 45 microns; Slide 19a No. 2, Bevier Coal; 59. Fic. 2. Portion of figure 1 enclosed in black lines; X 271. Fic. 8. Compares favorably with Wilson and Hoffmeister’s (1956) Cuticle type D; cells are polygonal- to squared-shaped and average approximately 50 by 35 microns; cell walls are approximately 2 microns thick; Slide 19a No. 2, Bevier Coal; 59. Fic. 4. Portion of figure 3 enclosed in black lines: > 271. Fic. 5. Cuticle with thick-walled polygonal- to hexagonal-shaped cells; walls are apparently approximately 5-6 microns thick; the average cell side is approximately 15 microns; believed to be closely related to cuticle of Lepido- carpon megaspores, see plate 24, figure 1; fairly common in all coals studied; Slide 19c No. 14, Bevier Coal; 59. Fic. 6. Cuticle with cell walls that are very irregular in size and shape and in wall thickness; compares favorably with larger cuticle fragments taken from Cordaites leaf fragments; Slide 4c No. 34, Weir-Pittsburg Coal; x 59. Fossu. MicROFLORAL INVESTIGATIONS 953 PLATE 10 954 Tue UNIversity SCIENCE BULLETIN PLATE 11 Fic. 1. Cuticle of unknown affinities apparently without structures of any kind except thick-walled perforations averaging 10-18 microns in diameter; Slide 51, Bottom of Lower Williamsburg Coal; 85. Fic. 2. Cordaitean tracheids?; characterized by alternating rows of ellip- tical openings, 1.5-2 microns in diameter; common in Blue Mound and Lower Williamsburg Coals; Slide 51, Bottom of Lower Williamsburg Coal; X 625. Fic. 3. Unidentified wood fragment (Lycopodian?); openings average 13 microns in length and approximately 5 microns in width; Slide 57, Bottom of Lower Williamsburg Coal; x 390. Fic. 4. Cordaites secondary tracheid with multiseriate bordered pits; open- ings average about 4 microns in length and approximately one micron in width; these tracheids usually average about 6-8 pits in width; rare to common in Lower Williamsburg Coal; Slide 54, Bottom of Lower Williamsburg Coal; < 625. Fic. 5. Cordaites tracheids; common; Slide 56, Bottom of Lower Williams- burg Coal; 390. Fic. 6. Unidentified fusainized fragment with holes averaging approxi- mately 2.5-4 microns in diameter; fairly common in Middle of Lower Williams- burg Coal; Slide 62, Middle of Lower Williamsburg Coal; x 390. 955 Fosst MIGROFLORAL INVESTIGATIONS PLATE 11 bla Sd : RT oe ie eg << ae pont &é. 3 @ cd ‘ A eeeeetegees eeeeeeeees A ae 956 THE UNIVERSITY SCIENCE BULLETIN PLATE 12 Fic. 1. Unidentified fragment with variable circular openings ranging from 35 to 2 microns in diameter; common to rare in the Lower Williamsburg and Blue Mound Coals; Slide 94, Blue Mound Coal; « 390. Fic. 2. Unidentified wood fragment with variable circular openings rang- ing from approximately 5-10 microns; common to rare in Blue Mound and Lower Williamsburg Coals; Slide 56, Bottom of Lower Williamsbur Coal « 390. Fosstt MicROFLORAL INVESTIGATIONS 957 PEATE, 12 33—5840 958 Tue Universiry SCIENCE BULLETIN PLATE, 13 Fic. 1. Tracheid from transition zone between primary and secondary xylem of Cordaites?; this fragment probably occurred between the protoxylem and scalariform tracheids of a Cordaitean-type plant; fairly common in the Weir-Pittsburg, Lower Williamsburg and Blue Mound Coals; Slide 63, Middle of Lower Williamsburg Coal; * 526. Fic. 2. Leaf cuticle of Cordaites; this is the most abundant type present in the Lower Williamsburg and Blue Mound Coals; occasional pits and open- ings with a diameter of 11-12 microns; cell walls 3-6 microns thick; cells average about 80 by 23 microns and are wedge-shaped at the ends; Slide 97, Blue Mound Coal; 85. Fic. 3. Sac-shaped mass of square cells averaging about 40-50 microns on a side; rare; Slide 52, Bottom of Lower Williamsburg Coal; x 85. Fic. 4. Cuticle characterized by square-shaped cells approximately 12 microns on a side; cell walls are thick and average 2-4 microns in thickness; Slide 75, Bottom of Lower Williamsburg Coal; « 390. Fosst. MiICROFLORAL INVESTIGATIONS PLATE 13 960 THe UNIversIty SCIENCE BULLETIN PLATE 14 Fic. 1. Cuticle with irregular elongated polygonal-shaped cells approxi- mately 100 microns in length; cell walls 2.5-3 microns thick; abundant in Blue Mound and Lower Williamsburg Coals; Slide 97, Blue Mound Coal; x 57. Fic. 2. Enlarged area framed in white in figure 1; * 260. Fic. 3. Cuticle with elongated irregnlar-shaped cells averaging approxi- mately 60 microns in length and 20 microns in width; fairly common in both Blue Mound and Lower Williamsburg Coals; Slide 97, Blue Mound Coal; < 260. Fic. 4. Enlarged portion of figure 3. Fosstt MicROFLORAL INVESTIGATIONS 961 962 Tue UNIVERSITY SCIENCE BULLETIN PLATE 15 Fic. 1. Triletes sp.; note the prominent undulating rays and the coarse rugose network covering the surface; conspicuous “wrinkles” are approximately 50 microns wide and form sharp to rounded ridges; diameter is 2000 microns; rare in the Blue Mound and Lower Williamsburg Coals; this specimen was taken from the Blue Mound Coal; x 42; reflected light. Fic, 2. Opposite side of same specimen. Fossi. MICROFLORAL INVESTIGATIONS PLATE 15 963 964 Tue Untversiry SCIENCE BULLETIN PLATE 16 Fic. 1. Triletes (sectio Auriculati) sp.; proximal view of proximo-distal compression with conspicuous darkened rays and darkened outer margin; rather smooth surtace; ray ridges rather low in relief; diameter is 2100 microns; from Blue Mound Coal; x 42; reflected light. Fic. 2. Distal view of same specimen. Fossa. MicROFLORAL INVESTIGATIONS 965 PLATE 16 966 Tue UNrversiry SCIENCE BULLETIN PLATE 17 Fic. 1. Triletes sp.; proximal view of proximo-distal compression with conspicuous indentions along rays; conspicuous furrow along top of rays; rays decrease in relief toward the indented areas; rare, most frequent in top portion of Lower Williamsburg Coal; diameter 1900 microns; 42; reflected light. Fic. 2. Distal view of same specimen. Fosst. MiCROFLORAL INVESTIGATIONS 967 PLATE 17 968 THe UNIversiry SCIENCE BULLETIN PLATE 18 Fic. 1. Triletes (sectio Auriculati) sp.; proximal view of proximo-distal compression with wide darkened rays; Top of Lower Williamsburg Coal; diameter 1900 microns; x 42, reflected light. Fic. 2. Distal view of same specimen. Fosst, MicROFLORAL INVESTIGATIONS 969 PLATE 18 970 Tue UNIversiry SCIENCE BULLETIN PLATENS Fic. 1. Triletes (sectio Auriculali) sp.; proximo-distal compression with low slightly-undulating rays; rays darkened; surface relatively smooth; Bottom of Lower Williamsburg Coal; 1900 microns in diameter; X 42, reflected light. Fic. 2. Distal view of same specimen. Fossr. MicROFLORAL INVESTIGATIONS 971 PEATE AS 972 THE UNIVERSITY SCIENCE BULLETIN PLATE 20 Fic. 1. Triletes sp.; proximo-distal compression with low, darkened, slightly-undulating rays; Bottom of Lower Williamsburg Coal; 1825 microns in diameter; < 42; reflected light. Fic. 2. Distal view of same specimen; note the network of “wrinkles.” Fosstt MicROFLORAL INVESTIGATIONS 973 PLATE 20 974 THe UNIversITY SCIENCE BULLETIN PLATE 21 Fic. 1. Triletes (sectio Lagenicula) levis (Zerndt) Schopf, Wilson and Bentall, 1944; lateral compression with blunt and conspicuous apical protu- berance; may be associated with the organ genus Sigillariostrobus (see Dijkstra, 1958); common to abundant in Blue Mound and Lower Williamsburg Coals and is most abundant in the Middle part of Lower Williamsburg Coal; 1300 microns in total length; Middle of Lower Williamsburg Coal; 42; reflected light. Fic. 2. Opposite side of same specimen. Fic. 3. Triletes (sectio Lagenicula) levis (Zerndt) Schopf, Wilson and Bentall, 1944; 1050 microns in total length; Middle of Lower Williamsburg Coal; 42; reflected light. Fic. 4. Opposite side of same specimen. Fic. 5. Triletes (sectio Auriculati) sp.; outer margin well-defined; rays inconspicuous and poorly-developed; rays fade away at approximately % the distance to the margin; Blue Mound Coal; approximately 1300 microns in diameter; < 42, reflected light. Fic. 6. Opposite side of same specimen. Fosst. MicROFLORAL INVESTIGATIONS 975 PLATE 21 976 THE UNIVERSITY SCIENCE BULLETIN PLATE 22 Fic. 1. Camera lucida drawing of figure 4, plate 5 of this report; thick cuticle with conspicuous perforations approximately 28-33 microns in diameter surrounded by polygonal-shaped cells; may be associated with Dolerotheca or related genera; closely resembles the cell pattern of the epidermis at the base of dorsal hairs; large perforations correspond to the position occupied by hairs; note heavy thickenings around perforations; rare, found only in the Bevier Coal; from slide 19a No. 2; * 312. Fic. 2. Camera lucida drawing of cuticle shown in figure 2, plate 5; con- spicuous openings averaging 35 microns in diameter; these openings might be stomatal in nature; rare; Slide 19a No. 4, Bevier Coal; 172. Fossiz MiCROFLORAL INVESTIGATIONS PLATE 22 S77 978 Tue Unrversiry SCIENCE BULLETIN PLATE 23 Fic. 1. Camera lucida drawing of figure 14, plate 3; note in both this drawing and the photograph on plate 3 how the terminal subsidiary cells are nearly always darker than the stomatal areas; resembles Cordaites lower epi- dermis cuticle of leaf; the stomata are closely crowded together and both the lateral and terminal cells are shared; the lateral cells are 45-55 microns in length by 15-25 microns in width; terminal cells are oval-shaped and average 20-30 microns in diameter; the lateral cells are broader and shorter than the ordinary epidermal cells; guard cells were not distinguishable; Slide 4a No. 26, Weir-Pittsburg Coal; & 236.5. Fic. 2. Camera lucida drawing of transition zone between primary and secondary xylem of Cordaites tracheids; fairly common in the Weir-Pittsburg samples; 434.7. o79 Fossit Micro PLATE 238 980 Tue UNIverRSITY SCIENCE BULLETIN PLATE 24 Fic. 1 Lepidocarpon megaspore taken from the Top of the Lower Williams- burg Coal; fairly common; this camera lucida drawing is of the largest specimen found during this study; usually these megaspores are only % to % this size, cuticle fragments scattered throughout the Lower Williamsburg, Bevier, Blue Mound and Weir-Pittsburg Coals bear a very close resemblance to this speci- men; cell walls are irregular and thick; * 13.4, inset is & 53.5. The opening at the right end of the megaspore is a tear, not a morphological feature. Fic. 2. Microspore histogram modified from Habib’s (1960) thesis (see references); this histogram was prepared by using the same sample of Bevier Coal that was used for the investigation of cuticles and larger microflora in this report (from Bevier sample 19). 981 Fosstt MicROFLORAL INVESTIGATIONS PLATE 24 ie uy} pain WES yy? At #4 PY iN i AN ANA sitheane wy a f Loe, > Sa, eo yy) nt INO Nata Weare SOMA x rk y) wale EX) PRY 1960 fied from Hab Mod = sz D4IOAdSI{S8A $848//140197 $O41fDOYIS DIYII44{SIOY4 Seslsodsijojnuos9 Sapslsodsojo7 sajisodsijojnoidouy Sagsaginbsay $84140dS040b1Aa07 Sa{lodsosognssa/ SAfUsIO] J Dsodsowonyjo7 saplsodsijojoungd Saslsodsoj,ojoung Ds10dsookq 982 THe UNIverSITY SCIENCE BULLETIN PLATE 25 Fic. 1. Camera lucida drawing of cuticle fragment showing the relation- ship between polygonal-shaped cells and long, rectangular-shaped cells; cuticle of this type is not common but it shows at least a partial association of po- lygonal and rectangular cells; Slide 19b No. 7, Bevier Coal; x 200. Fic. 2. Fragment of Cordaianthus sp. Grand ’Eury, Cordaites inflorescense; fragments of cuticle of this type are common in the Lower Williamsburg and Blue Mound Coals; cells are square to rectangular and sometimes hexagonal or polygonal, usually about 35 microns long in the specimen illustrated here; other specimens sometimes show somewhat smaller cells; cells have conspicu- ous wall thickenings; may correspond with Wilson and Hoffmeister’s (1956) Cuticle type C; sex of the various strobili was undetermined; camera lucida drawing from Top of Lower Williamsburg Coal; 36, inset « 108. 983 Fossiz. MiCROFLORAL INVESTIGATIONS PLATE 25 Ep onsen TOY ees wp f@ ge z = 2 TAS GP ae ay ae 0 we STROBILUS ed —=. —— 2, ===. a ——— 984 THe UNIversiry SCIENCE BULLETIN PLATE 26 Outcrop of Lower Williamsburg Coal in NW corner Sec. 13, Twp. 14 S., Range 18 E., Douglas County, Kansas, approximately 200 yards below Lone Star Lake Spillway. Exposed in this bank is about 14 feet of Lawrence Shale. Good compressions of Neuropteris scheuchzeri Hoffm., Pecopteris, Annularia, etc., may be found here in the Lawrence shale a few inches above the Lower Williamsburg Coal bed. Foss. MicROFLORAL INVESTIGATIONS 985 PLATE 26 986 THe UNIvErRSITY SCIENCE BULLETIN PLATE 27 Outcrop of Lower Williamsburg Coal 75 yards below the Lone Star Lake Spillway in NW corner Sec. 13, Twp. 14 S., Range 18 E., Douglas County, Kansas. The coal occurs in the west bank of a tributary that runs into the Wakarusa River. 987 Fossa. MicROFLORAL INVESTIGATIONS PLATE 27 IIo oe i OL POG OES 988 THe UNiversiry SCIENCE BULLETIN PLATE 28 Close-up of Lower Williamsburg Coal outcrop below Lone Star Lake Spill- way. The top zone is 3 inches thick, middle is 2 inches thick and the bottom is 5% inches thick. This picture is an enlargement of the area directly above the “g” in the word, Williamsburg, on plate 27. 989 Fossi MicROFLORAL INVESTIGATIONS PLATE 28 34—5840 990 THe UNIVERSITY SCIENCE BULLETIN PLATE 29 Outcrop of Blue Mound Coal in NE 20 acres of NW of Sec. 28, Twp. 13 S., Range 30 E., Douglas County, Kansas. The outcrop is along the east side of Coal (sometimes called Cole) Creek. Outcrop thickness of coal was 10-11 inches. Good Compressions of Neuropteris scheuchzeri Hoftm., Alethop- teris, Pecopteris, Alloiopteris, Asterophyllites and Annularia may be found in the Tonganoxie Sandstone, a few feet above the coal bed. 991 Fosst. MicROFLORAL INVESTIGATIONS 29 PLATE THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vout. XLII] DECEMBER 29, 1961 [No. 8 Observations on the Biology and Taxonomy of Flies Found Over Swarm Raids of Army Ants. (Diptera: Tachinidae, Conopidae )' Cari W. RETTENMEYER 2 Department of Entomology, University of Kansas, Lawrence, Kansas Apstract: There are three important genera of flies found over and near swarm raids of army ants (Formicidae: Dorylinae) in Panama. These three genera are Stylogaster (Conopidae), and Calodexia and Androeuryops (Tachini- dae). All three genera are mainly found over the area where the raiding ants are concentrated at the front of the swarm and up to two meters in advance of the swarm front. There is no good evidence that any flies are found fre- quently in the fan area behind the swarm front or over raid or emigration columns. When these flies from the swarm front are found in the fan area, it is usually because a specific host insect has remained there or has fled to the fan area. These flies are found neither within the nests nor running in the ant columns. Stylogaster differs in its behavior from the other two genera by its almost constant hovering. Calodexia and Androeuryops usually rest on low objects and frequently shift position to avoid the ants. No species of Stylogaster has been reared, but it appears that those species found over army ant swarms are parasitic on cockroaches. This is indicated by their darting after cock- roaches driven out by the ants and by the finding of at least one egg on a cockroach. However, their eggs have been found also on Calodexia and Andro- euryops which may possibly be hosts for the conopids. A first instar larva of Stylogaster is described from an egg found on a Calodexia. 1. Contribution number 1094 from the Department of Entomology of the University of Kansas, Lawrence. This paper is part of a dissertation submitted to the University of Kansas in partial fulfillment of the requirements for the degree of Doctor of Philosophy. The National Science Foundation’s support of this project is gratefully acknowledged along with the assistance of Drs. C. D. Michener and R. E. Beer. Drs. G. W. Byers and H. S. Fitch have also read the manuscript and given many helpful suggestions, and F. J. Rohlf provided advice on the statistical techniques. In addition I wish to thank the officials of the Canal Zone Biological Area (James Zetek, C. B. Koford, and Mrs. Robert Gomez) for their help and the following taxonomists for determining part of the insects listed: §S. Camras (Conopidae); C. H. Curran, C. W. Sabrosky, and D. F. Beneway (Tachinidae); H. J. Grant, Jr. (Orthoptera). I am also indebted to Dr. T. C. Schneirla who stimulated my interest in army ants and their guests and provided my first opportunity to study them. The assistance of my wife, Marian E. Rettenmeyer, has been greatly appreciated throughout the field work and up to the present time. 2. Present address: Department of Entomology, Kansas State University, Manhattan, Kansas. (993) 994 THe UNIverRsITY SCIENCE BULLETIN Calodexia males are extremely rare near the ants, but the females are very common, over a thousand accompanying a large swarm raid. Large numbers of Calodexia are attracted to swarms of both Labidus praedator and Eciton burchelli at any time during daylight and twilight. These flies appear in the first ten minutes after the start of a raid, and increase to about 60 flies within a half hour. They probably locate the swarm front by the odor of the ants. Of 12 species of Calodexia compared, six show a significant preference for swarm raids of either L. praedator or E. burchelli. The greater abundance of these species over one or the other species of ant is probably attributable to differences in the raiding behavior of the ants. The dissimilarity in raiding results in quali- tative and quantitative differences in the arthropods driven out by the ants. Calodexia is larviparous and probably deposits larvae on the surface of a host rather than inserting them. Brief descriptions of the three larval instars are given. Larvae of Calodexia were found in crickets (Gryllidae) and cock- roaches (Blattoidea), and a few adults were reared from these hosts. These are the most common large Orthoptera found in the litter of the forest floor. Males of Androeuryops ecitonis are almost twice as abundant as females over the swarm raids of army ants. The behavior of A. ecitonis around the swarms appears to be similar to that of Calodexia, but no host is known for the species. It is oviparous and its ovipositor indicates that the eggs are laid inside the host. Unlike Stylogaster and Calodexia, the numbers of Androeuryops were much higher in the dry season and early rainy season than they were later in the rainy season. All three genera are thought to find their hosts by seeing them as they run or fly to escape the ants. Apparently these flies would be unable to find hosts without the aid of the ants or at least would be much less successful at finding hosts. The combined attack by the ants and these Diptera must kill a large number of orthopterans. Possibly less than ten percent of the potential hosts within the area swept by a swarm raid escape death by either the ants or these flies; indeed, meager data indicate that 50 to 90% of the cockroaches and crickets that escape the ants are parasitized by Calodexia alone. Although this study was done in Panama, identical or related species of flies appear to be present throughout the range of the two ants with which they are associated. These ants, Labidus praedator and Eciton burchelli, range from southern Mexico to southern Brasil in areas of wet tropical forest. A revised key to the species of Calodexia and partial redescriptions are in- cluded. CONTENTS PAGE INTRODUCTIONS stirs ote Sy Ss ek 2 Solio Sol coe Oe eee 995 NEOTROPICAL ARMY ANTS AND THEIR Types OF Rams ................ 996 SUMMARY OF THE BIOLOGY OF THE SWARM RAIDERS WITH EMPHASIS ON EACTORS#INELUENGING (HATDING |..- 7... See este eee 997 STUET ENS WARING MINATDI see cc tos d Gul, shu, Saat eee eo ice eae 998 DIPTERA ASSOCIATED WITH THE SWARM RAs .............-.-+-«:-.-. 1001 (ham BEHAVIORUORZOTYLOGASTER® ....~ Saxos YO GO OG. Ogee I if 9 G Sf) Ee Gea ol Gr iMG ie CE tal yee ee §[8}O.L z I Ti Goeeooon (anon) cnn Ceca ctar Tt lee ales le ee gee aie s]Uv avou JON e I Zz bP velo) Touse! lapvenin) Xefe(ey lesdnd (yo) -b\cenieltie)te}l|\ecingce, apie ba;||\te Net io iene Miao) (ede) (vis I oy emis in. 6) Teale Yohlows: volte) ‘s)) 6 suvbhoa a z (Aan (God on one ne cc (a Ge (ee oe oe (nn (a wunoULDY “A VASiN homme teal I! Lt I I G 6 & if 8 I Bie, | SOx a) cae ee es CEE MEN, SG ze I 61 PAVE tal ey see ees ca Pie le rap es ae ae fF , ¢ I i ; aml (ee ele Peta soywpavid T dee aes co) me} © é é fo) OF HERG slr é On| psojfijs psovads sisuawpund pynurnu VUDLINI isyUDq SS UY ; : ° UVaN NAV] S[eqOL, * dayspbojhijg Jo satoadg | sp1o99ye-UOTDITJOD JoysesoyAyG Jo Arewung—y] ATAV |, Furies Founp OveR SwarRM Rats oF ARMY ANTS 1009 a total of 51 specimens taken over a single swarm of E. burchelli in Mexico. These three species have also been found over burchelli in Panama (with the possible exception of ethiopa where the species of ant was not given). Table 1 shows a total of five specimens taken near Eciton hamatum or E. vagans, both of which raid in columns. The two females taken near E. hamatum were observed on 10 March 1955, hovering under a log next to the statary bivouac of colony 55 H-E. Both hovered for a few seconds only 6 to 12 cm. over the ants on the ground next to the bivouac, then shifted to another spot 10 to 20 cm. away so quickly that I could hardly follow them. The flies were not seen to dart at any of the ants, and no eggs were seen to leave the ovipositor (though they might have been missed because of their small size). Stylogaster was never seen near any other bivouac of E. hamatum even though more than 50 nests of this species were watched for brief periods or for several hours. Furthermore, I had pulled many of these bivouac apart in order to take samples of the brood and myrmecophiles. Following such a disturbance thousands of ants would be milling around the area of the nest, and the odor of the colony would be noticeably stronger. There are no published records of Stylogaster near hamatum. On 9 July 1956, I disturbed a bivouac of E. vagans which had a brood of reproductives approximately at the midpoint of their larval development. Twenty-five minutes were spent thoroughly examining the bivouac and taking the queen and a large sample of the brood and workers. When this was completed, I noticed that about 50 Calodexia and Stylogaster had been attracted to the mass of vagans milling on two or three square meters of ground around the nest. From 10:50 to 11:20 a.m. before I attacked the bivouac, I had watched the raid which had an unusual swarm front of about two square meters quite densely covered with ants. During this time two Calodexia and one other fly were seen, but no Stylogaster. Following the attack on the bivouac, I swept over the ants milling around the site from 11:45 a.m. to 12:15 p.m., and took three Stylogaster and 18 Calodexia. Since normally vagans is a column raider, the unusual swarm at the head of the raid column was possibly due to the fact that this colony was stimulated to unusual activity by the sexual brood. The fact that the distal end of the raid was also only about ten meters from the bivouac may account in part for the swarm. Otherwise such flies have not been found associated with column raiders but only with the swarm raiders burchelli and praedator. 1010 THe UNrversiry SCIENCE BULLETIN These last two cases indicate that Stylogaster apparently finds the ants by their odor and can locate at least four species of Dory- linae in this way. This conopid probably is attracted to many col- onies but stays only with praedator and burchelli which mantain swarm raids. This can be considered additional indirect evidence against the hypothesis that Stylogaster is parasitic or predaceous on the ants. E. vagans and hamatum are as big or bigger than burchelli and L. praedator and should be just as suitable as hosts. THe Lire CycLeE OF STYLOGASTER Although no larvae have been described for any species of Stylo- gaster, the eggs of most Neotropical species have been illustrated and described by Lopes (1937) or Lopes and Monteiro (1959). Lopes was the first to show that the eggs have a number of reliable specific characters which alone are sufficient for determining the species. The most outstanding characters of the egg are the shape of the pointed end, the number of recurrent spines (one to four) and the shape of a bladder-like protuberance near that end, and the pattern of reticulations on the surface (figure 1). Because of their value WF SEE Ficures 1 to 8. Egg of Stylogaster currani Aldrich with larva protruding from blunt end; egg found on base of wing of Calodexia venteris Curran female (2341-C). Ficure 1. Entire egg with anterior end of larva protruding; three ventral bands of small spines and part of larval body wall can be seen within chorion. Ficure 2. Large spine at posterior end of egg showing one large, dark, re- current spine near tip and one smaller recurrent spine near base of large spine. Ficure 8. Buccal armature and anterior end of larva; small structure antero- dorsal to oral hooks is probably antenna; (only right half of larva illustrated). Fires Founp OvER SwaRM Rats oF ARMY ANTS 1011 in determining females which are in poor condition, as well as for determining eggs found on possible hosts, it is hoped that future descriptions of Stylogaster females will include descriptions of the eggs. Stylogaster eggs are most easily removed through the ventral, membranous part of the abdomen while it is still soft shortly after collecting. However, they can also be recovered from dried speci- mens, though it is often impossible to avoid breaking off the abdo- men because of its rather weak attachment to the thorax. The eggs which I examined were partially cleared in a clearing solution modified from that of Nesbitt (1945: 141): 40 g. chloral hydrate, 25 ml. water, 2.5 ml. hydrochloric acid, and 2.5 ml. glycerin. They were then mounted in Hoyer’s medium made by the formula given by Baker and Wharton (1952: 10): 50 g. distilled water, 30 g. clear crystals of gum arabic, 200 g. chloral hydrate, and 20 g. glycerin. Clearing in clove oil, cedar wood oil, or xylene followed by mounting in balsam, diaphane, or permount was more time con- suming, and the greater permanence of the slides is apparently the only advantage. These latter techniques also require more com- plete clearing to show the reticulation of the chorion. All the eggs found in Stylogaster females had completed cho- rions, and eggs in earlier stages of development were not seen. This is in marked contrast to Calodexia in which larvae and eggs in vari- ous stages of development are found in the adult females. The maximum number of eggs in any one female seems to be between 60 and 80. Lopes (1937: 266) reported 60 in S. stylata in Brazil. Neither Lopes nor I found any eggs, either in females or on other insects, showing any sign of larval development except for the one discussed. below. Lopes described an egg of S. stylata which he found inserted be- tween the fourth and fifth terga of an undetermined orthopteran. This orthopteran was escaping from a raid swarm of L. praedator above which Stylogaster had been seen hovering and dashing after cockroaches and other Orthoptera (Lopes, 1937: 259-260, 267-268 ). Later he reported that when examining the collection of cock- roaches at the Instituto de Biologia Vegetal in Rio de Janeiro, Dario Mendes found a Stylogaster egg between terga near the end of the abdomen of an adult Chorisoneura sp. The species of Stylo- gaster to which this egg belongs is not known (Lopes, 1937: 289- 290). Examination of about 50 cockroaches taken on Barro Col- orado Island revealed no Stylogaster eggs; however, only about THE UNIverRSITY SCIENCE BULLETIN 1012 III pure TT R310} u90M40q ‘TB104R'] ouahie cq. r6 j pjidnssaqur “Dosh ees See eds lrccccccccte . q-LLZ1 AI pu® TIT RB10} u90M40q ‘1B19} 8] eireyta eh atrel aor j CEIG 'g Teal tee ae a ae Te a 6 ( 2 eee ps2 ack QO QO EDs || Re PT PE 21D V-6Z8% II] pue J] B810} UsaMgoq ‘feroyeT | oo SPSS. |e ee ee UU UID |e pa Rane ee ie V-GGLT Ill pus II BB104 u90M4oq ‘Te.l0}V'] 0 ean Oe Om Oab j saaup To) n CACAO. OLDE eter? c asyung HPO CMO Cap Od V-I¥&Z OBpPIUTYOR T, dayspbophiig puro 7, uo dd9 Jo UOTYBOO'T jo jo UAAWAN ATA sotwedg sotoadg aepruryoey, Uo punoy sssqy Joyseso[AjG—'Z AIAV |, Fires FouND OvER SwaRM Rats or ARMY ANTS 1013 ten of these cockroaches were found near army ant raids where Stylogaster was present. After our return from Panama, the examination of the tachinids associated with the swarm raids revealed eggs of Stylogaster on several species of Calodexia as well as on Androeuryops ecitonis (Townsend). gc cg COL «Mol ts] teMfelkejie® (eitv re. ieljelte' we) inure ier leh vei.s. DIANLLAVUL 4) 426 £9 ¢8 GL 19 8% UG eae yl een eek tae see Bay DEOUNLS) 1¢9'L EhZ Ze eg 8L Oh Girma [ace Beg ag pypvospf */) 468 ZEIT 461 OL LEI 0€ 1S eermeremn Leeiliat i cane 1h me mage iy soap *\) ‘su 20'0 611 6Z Ee TL 8 eS Ry ge Se ee LO iy) 468° €I gg 6 G 16 OG ake ~ _ SE = ii as ie oe eee n729 Beef a a eae eee z OI Z 0 0 Th 0.0% D0 on ol moO 00 6 Oo Lo sypoidp °@) ‘8 U1Z°0 18z YE 18 %69 1 eon ae aot ae mn eg «Sh syibo *) soloeds SpIBI ULIBMS SpIBl ULIEMS Sone! qyoq Lopppanid snpiqn'T UYJAYILNG UOPIA sdohiunaoipuy 10 DixapojD,) ae) I9AO I9AO JUddIed pus I9AO JUddIed pue Jo sotedg [TFIOL aq uun Ny TOQUIN NT aeprluIyoe], OJ splooY UoNodaToD jo Arewung—'Eg atav |, 1024 THE UNIVERSITY SCIENCE BULLETIN species, and in two cases, dives and panamensis this is especially marked. The biological significance of these results is doubtful. It is possible that some species of Calodexia are attracted to one species of ant more than to the other on the basis of the odor differences of the ants with no survival advantages for the fly. However, for certain species of Calodexia, there may be an advantage attribu- table to the different raiding behavior of the ants. The more sub- terranean Labidus pracdator clearly raids to a larger degree under leaves and other objects which are close to the ground. Because of this difference in raiding, it is probable that praedator drives out some species of arthropods which are not often disturbed by burchelli. The workers of praedator are smaller than those of burchelli, and the former appear to capture more smaller arthropods while perhaps missing more of the larger orthopterans. In addi- tion, burchelli raids more frequently and much higher up trees, going near the tops of the tallest trees more than 30 meters from the ground. It is not known whether any of the flies follow the swarm raids in these trees, but this arboreal raiding must drive out many arthropods not found on the forest floor. Other than the gross behavior of the ants, there is no evidence for a difference in the proportions of potential hosts driven out, since no quantita- tive nor qualitative study of the arthropods either driven out or eaten by the ants has been made. Knowledge of the hosts may eventually support the observed differences in the proportions of the different Calodexia species associated with the two species of army ants. However, at this time no reliable conclusion can be reached because far too few hosts are known. In addition to Calodexia associated with burchelli and praedator, two specimens were found near the unusual swarm raid of Eciton vagans (discussed earlier, under Stylogaster). After the bivouac of this colony was located, and the queen, a large sample of the workers, and part of the sexual brood were taken, it was discovered that Calodexia had been attracted to the ants milling around the bivouac site. About 50 flies were in the area and 18 were taken. In contrast to this experience with a disturbed bivouac of vagans, no Calodexia could be found near the ants milling around more than 20 nests of E. hamatum which had been similarly disturbed. As with all other species of Eciton, the odor from a bivouac of E. hamatum becomes much stronger when the ants are disturbed. Fires Founp Over SwarRM Rats oF ARMY ANTS 1025 THe LirE CycLE OF CALODEXIA Although Calodexia is much more numerous around the swarm front than Stylogaster, there are fewer records of its presence and less speculation about its biology. This is probably due to the un- usual morphology of Stylogaster and the fact that the Tachinidae are poorly known taxonomically. Calodexia is not conspicuously different from hundreds of other genera of higher Diptera, and very little work has been done on the taxonomy of the genus even though it is widely distributed from Mexico to Brazil (and probably Argentina ). The most extensive previous collection was made by Curran, who has also described most of the species of Calodexia. He gave only brief comments on their association with army ants and decided that the flies parasitized cockroaches since they flew after them (Curran, 1934a: 1-2). Curran did not know whether Calodexia deposited eggs or larvae and did not identify any of the ants or cockroaches involved. The only positive information on their life cycle was given by Lopes (1937: 260) who reared “Calo- dexia Pventeris’ from one specimen of Periplaneta americana, ap- parently taken as it was escaping from a swarm raid of Labidus praedator. All females of Calodexia which were examined had more than 25 larvae in their oviducts, and some had over 200 arranged in dense, continuous spirals, showing progressive development from small eggs to actively squirming larvae. It is very easy to slit the mem- brane between the first and second abdominal segments and re- move the posterior segments from either a dried or fresh specimen. Dried abdomens then can be softened in hot water or Nesbitt’s clearing solution and the oviducts and larvae removed. Larvae from dried specimens are readily expanded and cleared in Nesbitt’s solution and can be mounted directly on slides in Hoyer’s medium. (See the section on “The Life Cycle of Stylogaster” for formulas of these solutions.) First instar larvae of eight species were ex- amined; these all showed specific characters. The first instar larvae have an indistinct pseudocephalon which lacks spines and shows no clear separation from the first thoracic segment. (The terminology for the larvae used here follows that of Roback [1951].) The three thoracic segments have very fine spinules in segmental bands and some species have ventral patches of strong spines on the meso- and metathoracic segments. All first instar larvae examined have strong bands of abdominal spines. 1026 THE UNIVERSITY SCIENCE BULLETIN These bands are composed of about five rows of spines of varying lengths and shapes, arranged so that the spines are contiguous for most of the width of the larva. The arrangement and shape of these spines are useful for separating species. There are hooks at the posterior end of the body near the simple posterior spiracles, and these hooks probably are used to hold on to part of host's tracheal system. The buccal armature is thin and long with broad or long, thin, sharp hooks. These larvae appear well-adapted for piercing a host and for moving on most surfaces. Females dropped larvae when they were etherized, killed in cyanide bottles, or if their abdomens were squeezed. These larvae were very active when they came out of the female. Probably they are dropped on the surface of the host and then burrow into it, since the larvi- positor of the female shows no piercing adaptations. Although all instars have excellent specific characters among the few species where more than the first instar is known, there do not seem to be any specific characters which are stable throughout the larval period. Therefore, it is impossible to identify later instars by com- parison with the larvae removed from adult flies. The second instars lose the prominent bands of spines but retain bands of spinules or roughened areas. The oral hooks become much broader and very different in shape, prothoracic spiracles develop, and there is a prominent papilla (antenna?) at each anterolateral angle of the pseudocephalon. The third instar larva has a pair of large posterior spiracular plates with peritremes which are almost closed. A button and three elongate, almost straight, spiracu- lar slits are found on each plate. The oral hooks are shorter and broader than in the second instar, and there are two small holes in the central area of the basal part of each hook. The ventral edges of the hooks may have a series of fine teeth. Only one series of Calodexia was reared to adults from a deter- mined host parasitized in the forest under natural conditions. This series of eight C. interrupta was reared from a gryllid, Ponca venosa Hebard (2260-X), which was taken as it was running from a swarm raid of burchelli. Three to five adults of Calodexia were seen to fly closely after the cricket as it ran about a meter in advance of the ants. This cricket was taken on 7 August 1956, and on 1] August nine puparia were found in the rearing vial. Four males emerged on 21 August and four females on 22 August. No fly emerged from the ninth puparium, and no additional larvae could be found in the cricket. Thus, the larval stage was very short, with a duration Fires FouND OvER SwarRM Rais oF ARMY ANTS 1027 of about four days; and the pupal stage lasted 10 to 11 days. No larvae were found in four additional specimens of Ponca venosa (2231, 2157) which were dissected 4, 6, 16, and 28 days after groups of Calodexia were seen flying after them in front of raid swarms. In addition to this one case where the species of both host and parasite are known, there are several less complete cases. One Calodexia pupa was reared from a cockroach (2259) taken on 1 August as it was running while pursued by L. praedator and Calo- dexia. The cockroach died on 5 August, and the single puparium was found on 6 August, but no adult emerged. The oral hooks from this puparium are very similar to those from a known C. venteris puparium, but positive identification cannot be made until other puparia are available. A gryllid, Eneoptera sp. (2205), taken as it was running away from a swarm raid of Labidus praedator on 27 July, was dissected on the same day. Three first instar Calodexia larvae near agilis or interrupta were found inside near the dorsum of the first abdominal segment. These larvae were not firmly imbedded nor attached to any tissue, but they may have shifted in position after the cricket was killed. A second Enecoptera sp. (2184) taken under similar conditions was dissected 60 hours after the time of suspected larviposition. Three small second instar larvae and nine early third instars, perhaps of the same species, were found in it. In a third specimen of Eneoptera (2231) taken in similar circumstances no larvae could be found. Two cockroaches, Epilampra azteca Saussure (2069-A, 2069-B ), were taken escaping from a praedator swarm on 5 July. One was dissected that day and no larvae were found in it. The second specimen was dissected 24 hours after the time of suspected larviposition, and nine larvae were found in it. All these appear to be second instar Calodexia larvae. These larvae were found close to the abdominal body wall of the cockroach except for one or two which were either close to the gut or were pulled away from the body wall when the cockroach was dissected. A sub- sequent attempt to introduce Calodexia larvae into a specimen of Epilampra azteca (2295) was unsuccessful and when it was dis- sected after eight days, no larvae could be found. A cockroach nymph (2069-C) which could not be identified to genus was taken at the same time as the above mentioned Epilam- pra. When it was running from the ants, it was clearly seen to be pursued by several females of one of the largest species of Calodexia 1028 THe UNrIversiry SCIENCE BULLETIN and by no other flies. A puparium was found on 15 August in the rearing vial with this cockroach, but the larva probably came out of the cockroach several days earlier. No adult emerged from this puparium. In December, 1959, the puparium was dissected and an advanced, but light-colored, female pupa of Calodexia dives was found in it. In spite of the incomplete development this speci- men could be determined since it is the only Calodexia which lacks presutural acrosticals and has one pair of postsutural acrosticals. Although other species of Calodexia were flying nearby, it is note- worthy that only large Calodexia were seen to pursue these last three cockroaches. However, the nymph in which the dives devel- oped was a small host, only about 15 mm. long. On one occasion a single C. dives was taken as it and at least two other smaller species hovered over a cricket (?Ponca venosa) being attacked by the ants. About 12 larvae from Calodexia venteris were introduced into each of two specimens of Eublaberus posticus (Erickson) (2235, 2236) on 30 July 1956. On 5 August four puparia and one mature larva were found outside of the second cockroach, and two males and one female C. venteris emerged from these puparia on 18 August. Although the cockroach appeared weak, it remained alive until 6 August. It was not preserved in alcohol until 18 August, however, since additional larvae were still seen inside it. When the cockroach was dissected, four first or second instar larvae were found which were probably laid on the cockroach after it died. Nine second instar larvae, which might be Calodexia, and the three larger second instar larvae, which were discussed above as possible Stylogaster, were also found. In hopes of rearing specimens of males, which are almost im- possible to find near the ant swarms, Calodexia larvae were in- troduced into additional insects which were readily available at the laboratory. The following insects were tried: prepupae and pupae of Polistes and a sphecid mud dauber, tettigoniids (Chloroscirtus discocercus Rehn, Euceraia sp., Microcentrum stylatum Hebard, M. philammon Rehn, Microcentrum sp., and Phylloptera dimidiata Brunner), gryllids (undetermined nymphs), and _ cockroaches (Eurycotis sp., Periplaneta brunnea [Burm.], and undetermined nymphs). No larvae completed development in any of these “hosts,” although in a few cases they were seen moving under the body wall for at least a day following the insertion of the larvae. Several first instars at least doubled in size, and two C. fumosa Fires Founp OvER SwaRM Ras oF ARMY ANTS 1029 larvae (2242-C) developed to second instars in a tettigoniid which died a few hours after they were inserted. In summary, the known fragments concerning the behavior and biology of Calodexia indicate that the different species are all larvip- arous and internal parasites of Blattoidea and Gryllidae. The larval period is only four or five days and the pupal stage lasts ten or 11 days. These were reared under fluctuating laboratory condi- tions at around 28° C. which is slightly warmer than forest floor temperatures. Even allowing for the possibility that some first instar larvae were overlooked in dissecting the hosts, it is clear that of the cockroaches and crickets which escape the ants, 50% to 90% are subsequently parasitized by the flies accompanying the ants. In most cases this parasitism is fatal to the host. OBSERVATIONS ON THE TAXONOMY OF CALODEXIA All 23 known species of Calodexia are included below in a table of diagnostic characters and a key. Additional records and clarifi- cations of the original descriptions are also given for all species found on Barro Colorado Island. Nine species in the genus have not been examined by the author, and the brief treatment of them is abstracted from the published descriptions. About 170 speci- mens, representing thirteen species, were previously known from Barro Colorado Island. Since most of these species were based on few individuals, examination of 1802 additional specimens from this locality has revealed a few errors and more variation than reported in the original descriptions. Possible geographical varia- tion is not involved here because all the specimens are from the type locality except for the specimens of similis and fumosa. In the generic and specific diagnoses given below, all characters which are not mentioned are considered to agree with the original descriptions. Only additions or major modifications of the de- scriptions are included. Information on specific characters which can be easily presented in table 4 is not repeated in the comments for each species unless some clarification seems desirable. Throughout the present treatment the terms used by Curran (1934a, 1934b) have been used, but in some cases modifications of these terms have been made in an attempt to be more precise. The method of describing the arrangement of bristles follows that used by many dipterists, e. g., acrosticals 1-3 means that there is one pair of presutural and three pairs of postsutural acrostical bristles. The term frontals is used for the row of bristles along the inner edge of the parafrontals and includes both the proclinate 1030 THE UNiversiry SCIENCE BULLETIN and reclinate orbital bristles. The width of the eye is measured per- pendicularly to the median margin of the eye when the head is ob- served perpendicularly to the anterior surface. Abdominal terga are numbered as in Curran, i. e., the “first tergum” is the first obvious segment, but morphologically, it is the second tergum. Other terms are explained here with reference to table 4, and the abbreviations used in that table are given in parentheses. A question mark (?) indicates that the state of a character cannot be determined with certainty from the published description. Where several states are included for the same character, the most common state is italicized; e. g., number of bristles 2-3. This indicates most speci- mens have two bristles, but an occasional fly or possibly as many as ten percent of the specimens have three bristles. Where frac- tions are used, e. g., % (abbreviated as 1), % (2), the abbreviation 1-2 means from # to %. For all thoracic and abdominal characters using patterns of pollen or vittae, the specimens should be viewed with the light striking the specimen at a slightly posterior angle. Small bristles have been included in the counts of bristles when they are only slightly larger than neighboring hairs and in the “correct positions.” When looking perpendicularly at a section of body wall using a microscope with a bright light and 40 & magnifi- cation, “white” hairs usually cannot be seen although they can be seen when the specimen is turned more obliquely to the line of sight. “Black” hairs can be seen in perpendicular view as well as in oblique view and are usually shorter and coarser than white hairs. “Black” hairs in some lights have a brownish or reddish sheen. Throughout this paper the following colors are referred to only by the primary color: orange, usually a dull, pale, slightly brownish- orange (0); yellow, a dull, pale straw color or slightly brownish yellow (y); gold (g); brown (br); black, due to pollen often ap- pears brownish in some lights (b); bluish (bl); white to pale gray- white (w). Combinations of these symbols are also used; e. g., (obr) = orange-brown; (o-br) = orange to brown (where color varies from orange to brown in a single specimen or species). All observations were made with a stereoscopic microscope using 40 « magnification and a fairly bright light (Spencer microscope light). Under lower magnification and dimmer light the colors look darker and closer to the colors as given in Curran (1934a). Notes applicable only to an individual character are given below and are numbered to correspond to the numbers in the left-hand column of table 4. AF 14. 20. 22. Fries Founp Over Swarm Rarps or ARMY ANTS 1031 Color of occipital pollen: Mostly white to yellowish on posterior aspect of head, but dorsal and median black bands of pollen may be present (w); mostly black, but lateral edges may be whitish (b). Specimens should be examined laterally to see maximum amount of white pollen. Downward extent of row of black bristles along outer edge of occiput measured as fraction of length of eye: Fractions %, %5, %, and 44 (ex- tending completely under eye) are abbreviated as 1, 2, 3, and 4. Head is observed laterally to determine length of eye and extent of occipital bristles. Width of palp: Nearly equal throughout its length or enlarged apically to about twice the width of the basal half (e); greatly enlarged near apex to three or four times width of basal half (en). Anterior outer dark vittae of mesothorax: Do not extend anteriorly to reach the position of median sublaterals (ne); extend to level of median sub- laterals (es); extend anteriorly beyond position of median sublaterals (ea). Where outer dark vittae are greatly narrowed but continue as fine lines to anterior surface of thorax, only broad main vittae are considered. Where there is a darker stripe in the middle of an area lacking white pollen, the entire dark area is considered to be the vitta, not only the darker stripe. Median postsutural dark vittae of mesothorax: United for approximately their entire lengths to form a single board vitta, but median section of this vitta may be more grey or have a few light pollen spots (u); separated directly behind suture by light pollen but joined posteriorly for about one- half their lengths (sbs); separated for entire lengths behind suture (sel). Light colored postsutural mesothoracic vittae which separate outer dark vittae from median dark vittae: Absent, so that all postsutural dark vittae are united for at least half their lengths, usually appearing as wide solid black fascia across thorax behind suture (a); narrow, at widest point about half width of an outer dark vittae (n); about equal in width to an outer black vittae (e). Marginal scutellar bristles include lateral bristles and large pair on pos- terior lateral angles of scutellum but not apical marginals. (Townsend [1912: 309] includes posterior lateral marginals with apicals. ) Infrasquamal cilia are sometimes reduced or absent. Posterior femora when seen ventrally: Only proximal one-third to three- fourth orange (po); or color of entire ventral surface except distal ex- tremity (o, br, etc.). Specimens should be turned to view posterior femora from several angles to see maximum amount of light color in ventral view. Ventral yellow cilia are arranged in more distinct rows than majority of hairs on ventral surface of femur, and longest cilia are at least as long as thickness of tibia. Few scattered hairs on inner or outer central margins not considered to be rows unless they extend along at least half length of femur. Color of shining dorsal abdominal pollen (ignoring duller gray-white pollen). Specimens examined with light striking dorsal surface approxi- 1032 THE UNIVERSITY SCIENCE BULLETIN 28. 29. 34. 36. mately perpendicularly to give brightest and most highly colored reflection from pollen. Color of dorsum of first tergum (ignoring pollen): Orange, may have a weak middorsal dark stripe (0); orange with strong middorsal dark stripe about as wide as distance between marginals (os); entirely brown (br); or black, usually with brown pollen (b). Whitish pollen on dorsum of first abdominal tergum: Abundant and dense at any point in dorsal view (dwp); absent or sparse (a). Light pollen of ventral areas of both first and second abdominal terga: Equally white and dense (e); absent or much less white and less dense on first compared with second tergum (a); absent or weak on venters of both terga (aa). Pale pollen on second abdominal tergum: Not interrupted by median longitudinal stripe (ni); interrupted for almost entire length of tergum by narrow dark or pollenless median stripe less than half as wide as distance between distals (ns); interrupted by wide, dark or pollenless stripe at least as broad as distance between distals (ws). Specimens should be observed in posterior dorsal view to show maximum extent of pale pollen. Lateral marginals are same as “lateral bristles” of some authors. Ventral marginals include bristles on tergum ventral to laterals. Pale pollen on third abdominal tergum: Extending posteriorly to margin between median marginals (ex); not extending posteriorly between median marginals, but small marginal spot not connected with pale basal fascia may be present (ne). Dark fascia on 4th tergum: Absent (a); or extending basally from median apical margin of the tergum %, %, %, or 44 distance to basal margin (abbreviated 1, 2, 3, 4 respectively). The entire surface of tergum should be observed from behind to determine maximum contrast between white pollen and dark fascia. Body length is measured from anterior surface of third antennal segment (when it is in normal position close against surface of head) to most posterior part of abdomen (not to end of larvipositor or most posterior part of abdomen in a morphological sense ). Total number of specimens and number of types examined by present author are listed in detail under discussions of each species. The number of specimens is given in parentheses and refers to females unless males are indicated. All specimens were taken on Barro Colorado Island, and details concerning their collection are given in a list of all field numbers at the end of the species discussions. 0 OL 986 G SOL 0 L461 & ELI CEUs alll ieee et a Ae eS a Sa ee peurmexe suowtoads jo 1oquinN *9¢ Sy eA 82 66) OL gma ae ep ONG Peay) Ca) 0'8 Pepe eeeeececeererecessrseses sss 9s ST@QQUlTT[TOr UL yysue, Apog wnuUNIxeW “CE g°9 6°9 ry Le9 6S £011 0°9 GL 8°F DAG el aye ase es uoulloads a[#uls Jo yySuUs] JO SisZOUTTT[IU UT YyZUe_ Apoq winuluUIpyY “PE é FI-9 OI-9 9-F LI-6 +¢ Z1-9 8-2 Z1-8 (()] Gents Joe Ml doiahaaes aes eet cegiee JoquINU UMOTYUN JO MOI IO OQUINU :WINS10} YJINOJ UO S[BosICT “EE é B Z-[-®8 I Z-[-8 ée-1 e I I €-1 “*-quosqe ‘f/f '$/8 '¢/% ‘Z/T i pBseq SuIpue}xe wiNns19} YINOJ UO Blosey YIVC, “CE jo q-4q q-1qo q-1q q iiq q-1q q-o iq-O 41q-o Bae i SS CNS RSs be (uajod Sulous!) wWiNns19} YAINOF JO Baie [BIZUBD JO 1OJOF) “1 E F 6-@ Zz Z Z z 6 z jz Yael aia tae a ahora aga ae ges Tre we UINS194 Pity} UO STBOSIp Jo oquinNY “OF é xo xo xo xo é eu—xo xo-0U Xo UL ap ee see es one UIZIVUL UBIPIUI 0} PapUd}Xxo JOU :popUs}xX9 :WINS10} PAY JO UI][Oq “6% é Gok I I Tosh él if I [ [ee || GE ee uINnd19} PUODIS UO S[VUIZIVUI [VIZUGA PUB [B1OzZB] JO SulVd Jo JoqUINN' “8% é I I I I I I I I [ite [chins hee ae eee MOL B IO ‘sired JO JaqUINU :UINSIa} pUOdGS UO S[jBUISIBU UBIPIT “LZ é 1u su-1u su lu A su—lu tu tu 110 fale [pee ae auou 10 adii4s optm 10 MOIIvU Aq pazydNII9}UL :UINSI0} pUOdeS UO UeT[Od “YZ é a-8 eB a-B SS) é 3) B o-D eS) “-* +" yQ0q UO YUaSGB !puodas 07 [BNba :4yUasqeB : UINFI0} 4YSIY UO Us[Od [ByUBA “GZ é dap B te B {8 ® 2 daMp=p7 ose (8) YBoM 10 guesqe :(dMp) asuep :UINSI04 4Ysiy UO Ud[[Od ozIYM [BSIOC] “FZ 41q q q q q so q q-iqa q (camel ven ee wie aus ey quesaid jt (S) edt4js puw UNI} 4SIy JO UINSIOP JO 10JOI) “ES é cl A-1q A-M A-M é Iq b—-A& A AG: ~ af OD Sheil nal as Saha a 8 eo eye ae Wn ge ane oe OP us[[Od [BUIWIOPg® [BSIOp JO OTOL) “ZS 20 0 0 0 0 I H) I 0 OV A ee ee inure} 1O11aysod uo BITIO MOT[OA [BIZUBA JO SMO JO IOBQUINNY “TZ é q o-od od od 41q iq-o q-4iq q-41q |= je eee S48 Jo[Oo aatjua ‘(od) asuBI0 AT]BUITxXOId : BOUT} 10[10}S0d JO 10[09 [BIJUIA “OZ z G ¢-¢ Zz z I z Z (4 | am || eee MM TRAD SI Ta ae AINUIO} I[PPIUl JO s[PplU Ive So]}SIIq IOeJUB JO JEqUINN, “61 é I I I I I I I I CE) ie en peat oe at BIqI} O[PPrur JO o[PPIW AwoU SopySI4q [BAZUaA Jo IOqUINN “ST é +0S GS-OP GE—-Go OI-s é Gi OS—-06 OS—06 SOL Gila eee ete ee x00 JUOIJ GUO UO SO]}SIIq PUB SITBY YoR[q JO Joquinu [BIOL “LT é q q M M éq ‘a M M+q M+q 7 | ace le OR a ane po (SopJStIq SuIpNypoul YOU) vwxoo yuodZ uO SITBY JO IO[OD) “OT é iq—-0 41qo-—o 1q-o 1q-o 4q-o 1q-o Iq-o Lq-O RT CEO ai seth 2 a ea ee (ua][od SULLOUDI) 9BXOO JUOIJ JO BOVJINS IOLIO}UB JO IOJOT) “CT é q q q q q q q q Cf | Raa ane (A) MoT[ad 3(q) Juesqe AT[BUOISBID0 IO YoRT :¥I[IO [BuUIBNbsvsjyuy “FT é q q AM q-A-—M& M q q q-¥8 ye me OS Cnn aCe eae ea a (A ‘M ‘q) guaseid UsyM 1O]O9 :(B) JUoSqB :i1ey [BINe|dOJON “ET é q q M M Mw M fi-m fi-M AAT aks | iss Settee cle pee Sea Bino[dosaul IOMajuBv IOMO] puw BINGTdOUII}S UO ofId Jo 10JOT) “ZT g € ¥-S ¢ ¢ g 8-G £ € OO ea cea a Sle NN a ae Sa]JSUq IB][9}NoSs [BUISIVUL [B19}BT JO Sued Jo aqUINN “TT 4 M qq qq q4 644 qq qq qq Ra lass ate (qq) pusq xoulq [BSBq YQIM oFTYM SYoRTA ozTyM :UINJ[ayNos Jo wNsI0g, “OT é u o-u u u {a-u u a—-u a—-u ey ROR Sa ae es (9) [enba !(u) MorreU !(B) JUESGB :9Bq4IA YYSTT PBrngnsysod JojNO 6 é jes n n n n n n n 19 Geeta | |e ae uejod yysty Aq pozyeredeas :poyrun :aingns pulyed oB}}IA YIBp UBIPIT “8 é ei) va Ba acts) é Bo Bo Bo BOE” aiigice ies S[Bi9}B[GNs UBIPEUT YIM porBdUIOD 9B94IA Yep JoyNO JO 4Ue}xe IOWoEJUy “LZ g g I [ 2-1 I I I I I s/BoIysO10B [BINjusysod jo sured jo aquINN “9 (6 I I I 2-1 I 0 I I | ee 1 ap aaa ae cn, OPEN SIRT EELS EST s[Bolysoi0B [Bingnsaid jo sured jo JoquinN “¢ 1q iq-O 1qo-o q-iq 1q-—o 40 1q-o 1qo-o 1qo-—o LC O—ONm | iin alias BYSIIB IY} JO UOTJLSUI OY} OF [BYSIP JUGUISOS [BUUOJUB PATY, JO 1OJOT) “fF é a a ue a pi) Ee) 5) cS) OM ees tec (ua) xodv 1v9u padie[ua A[}BoID + (9) [enbs ynoqe :dyed jo yApPIM “Ee é F &- iS Q = = 3 IS Sy & = & g & s S & SUTLOVUVHD, :. 3 a S$ 3 a & i) = 8 Ss 2 : 3 8 2 S s =. & => &% i & panu1ju0 j—vixapoyeyD jo satsadg umouy 10F siopovieyD oysouseiIq—'fF ATav], PPG 0 G 0 G ag 0 0 G 2 ee te a ae ee ae ** pauimexo susuo0ds Jo JoqUINN “9g 419 ¢'8 GL ele Ob adm, £8 DESh Tl” Cobos atl ne ie, ob 6°8 SIOPOUITTUL UT GysuUET Apoq WINWIXBy “CE OF 08 6°9 GL 68 e'¢ 0° OL 9°6 88 fF) pia | SRS OO usuIveds a[ZUIS Jo YYBus] IO SuoyoUTT][IWU UT YBZuUe, Apoq uNuIUITPY “FE FI-Z Z $-6 Fp 9 OL-9 MOL MOI 1-9 SGP Ble tees ee IS(uUNU UMOTYUN JO MOI IO :1IQUINU : UING19} YINOJ UO S[BOSICT “EE €-1 ¥ F é I aa é é €-1 Z-I-¥ | quesqe ‘p/7 'h/e '8/e *3/T :pEseq SuIpuszXo UINA19} YIANOJ UO BlOsBy AVC “SE q-o q-1q q-1q é oO q—1qo i é qiq-o (c+ TE | Neca aaa Me (UaT[od Sulsous!) WINs10} YJANOJ JO Bol [BI}UGO JO OTOL) “[E Z z z Z z z Zz Zz Zz Aw gy ICSE ree On EA OR i ae ge 0.0 LRRe A eas uIndz94 parqy uo s[wostp jo JoquinN OF au é ou é ou xo-9U R é ou-ra Kole ele ee eee UIBIVUL UBIPIU 0} papUds}xe JOU ‘ popUs}xXe :UIND10} PIL} JO UB]]Od °6Z +é-1 él I é I I Ai é I Tete, | By Oo aoe yee UINS19} PUODES UO S[BUISIVUT [BI}UGA PUB [B10}¥] JO SulBd JO LOqUINN' “SZ I [ I I I I I I I 10S a i eee ee ari aS MOL B IO {sured JO JaqUINU ;WINSI9} PUODES UO S[BUIsIBUI UBIPETY °2Z SM—1U é su—lu é iu su—lUu A, é su su “+s s-9u0u JO od143s opId 10 MOLIBU Aq poPdNAJezUL : WINSI9} PUODES UO UIT]Oq *9Z eB éB B 3 B eB é é B eB -**y30q UO yuasqe !puodes 09 yeNba ‘yuasqB :UINF19} ySIy UO UdTTOd [B1ZUBA “CZ e e a é ® B é {8 e he Soe (EA cy (8) YBaM 10 JUesqe :(dMp) esuep : UINS19} YS1y UO usa][od 9}14yM ]BsIO”(] "PZ q-41q q q é o so sa so 80-0 C0 eile Mc sine, ROS quosaid jt (s) odt1js puw WINDIo} 4ySiy JO UINSLOP JO IOjOD “EZ A-—M é Iq é 3 3—A é é AM 14 SOO CHONO DCm POLO Orsi bch O eC Cet aC, OOO Ch Oo ont ua]jod [evurmopqe [Bs10p jo RCo) (0Y@) “2G, 0 0 0 0) I 0 Chal G I 0 POE CASS BN ed sive Inutaf 1o11e}s0d UO BIO MOT[OA [BIZUBA JO SMOI JO IOQUINN “TG od od od 4o-od IqA—-A& AIqA—o (0) od od (oo an ace OS Jo[Oo argue $(od) adueBi0 AT[euITXOId :B10UIE} 1O119}SOd JO JOJO [BIZUDA "OZ G éG a é I jz é é G Fa || ECO SEES Daa aay Cae AINUIB} V[PPIUl JO o[PplU IvIU SoT}SIIq 1Ol1o}UB JO IEGUINN “6T 0 I I I I I I L I I PRP ae ogee Sat 2 BIqI} o]Pplur JO o[PplU ABOU So]}sTIq [BIZUGA Jo JOQUINN “ST SE-Gs é gg-G¢é é OI-S 0€-S é é O&—06 OSES Crs ee tee ae BxOO JUOIJ 9UO UO SaT}SIIq puw SIIVY YoRTq JO Joquinu [BIOL “LT q é q é hm q-fi-« é é q-hi-M QER=M ORS eee ee eee aa (So]JSIIq SUIPN[OUl JOU) vVxoo yuoIZ UO SAVVY JO IOTOC) “OT 1q—o pe) 1qo-o Pao) A o-A& 0 jA 40 o TOO =U | ae ee 9 eibe ae as (ua[jod Furoust) sVxoo yUOoAZ JO VOVJINS IONo}UB JO OTOL) “GT q q q q q q A-M q q q (A) MoTJeA $(q) JUaSGB AT[BUOISBID0 10 YORI :BI[IO [eUTeNbsesjUy “PT q-8 éq q é A-M-@ | q-M—8 i a q ete ll a andl a a ae (A ‘@ ‘q) queserd Way JOTOO $(B) JuasqB rey [RINe[doJON “ET q q q q A-M A-M é é AM AAA Ay (SECS? EON ira Sin Ae BiNo[dosour O119JUB JaMOT PUB BAINe[douIoys UO eIId JO 1OJO) “ZT G G G g (2=f6 € é ¢ ¢ Cae igs ae ee cache ceed SoPJSUG IB][oINOS [BUISIBUT [B10}BT JO Saved jo JoquInN “TT qq qq qq qq qq qq oq 64 qq Cola feu (qq) puvq YouyTq [BSBq YQIM oPTYM -YoRTG soz :wANTPoyNos Jo WINsIOCT “OT u—D 8 ce u u u-¥ ) o-u a-u Uae ce ee ores es (a) yenba !(u) MorBeU :(B) JUASqB :9By4IA 4FYST] yeangnsysod Joyng °6 n n n n n n é gn sqs SQ Shy Near eos uatjod yya1] Aq poyeredeas ‘poyiun :oingns puryod oBzIA YAVp UBIPIT “8 ou {ou ou é ee) Ba A Bi Bo Bo - -gTBIezBIGUS URIpsuT YIM poreduUL0d 9B44TA YAVpP 1o}NO JO JUO}X9 1OWOJUW “LZ ool [ I I | I I I I I ee OTN OO EN? Saat a s[Borysoi0B [Binynsysod jo sured jo taquinN °9 tit Ga! I I [ I I él l I Tl oh Bae ec se a eee s[BolysoroR [eingnseid jo saved jo JoquINN “¢ 41q—o on of o o aqo—fi 1q-O lq Iqo-o ATC O=O Nor eee BISUB IY} JO UOTJAOSUT 9} OF [BISIP JUOUIZOS [RUUOJUB PAY} JO LOOT “Ff ) 69 S) Pe) S) cS) eS) é cS) Olmert War kes aces (ua) xode avou pasiepue AT}BoId :(e) [enbe ynoqe :dyed jo GIPIM €-3 o-1 eI € ZI GI é é z q ors B/ B/g fe/e tZ/T :8a]}SUq [BIdI000 YoRlq JO 4U9}X9 PIVMUMOT, °G mM M M M M M M M M M Pe CG) yoRTG £(M) YSIMOT[OA 07 OFT :UaT[Od [ezId1090 Jo IOJOD “T 6 © é é e é Le © © 6 ~ 8 :: = = g %. % S s S 5 SUALOVUVHY g 3 = 3 5 as} 8 a B" 8 oo 8 g & papnjouoj—vixapo[ey JO sotoadg uMmouy J0F siopovieyy) oNsousvIq—'f AIAV I, 1036 THE UNIVERSITY SCIENCE BULLETIN CALODEXIA VAN DER WULP Calodexia van der Wulp (1891:257). Brauer and Bergenstamm (1891:375- 376, 1893:130-131), diagnosis. Coquillett (1910:517), C. majuscula desig- nated as type species. Townsend (1927:219), key. Curran (1934a:2-3), diagnosis. Curran (in Curran, etal. 1934:505-506), diagnosis. Curran (1934b:432, 439, 461), key, figures of head, wing. Oestrogaster Townsend (1912:310; 1914:28). Townsend (1927:223), key. Aldrich (1929:21). enon Townsend (1915:424). Townsend (1927:223), key. Aldrich Oestrogastrodes Townsend (1915:425). Townsend (1927:225), key. Aldrich (1929:21). The generic diagnosis given by Curran (1934a: 2-3) is considered to be correct except for the following modifications. Occiput with white or yellowish pile and black, white, or yellow- ish pollen; row of black occipital bristles extends downward at least half length of eye. Check at narrowest point %th to Msth as wide as height of eye (= maximum length of eye); ocellars short and weak; outer verticals present in females, may be absent in males; palpus with parallel sides or slightly enlarged at apex and with rounded or bluntly pointed tip, some males with apex of pal- pus swollen to form ovoid club; prementum about twice as long as wide in lateral view. Dorsocentrals 2-3, rarely 3-3; acrosticals usually 1-1, sometimes 0-0, 0-1, 1-3, 2-1, 2-3, 3-1, 3-3, rarely 0-2, 1-2, 2-1, or 2-2, one or more pairs often weak; no presutural intra-alar (= posterior sublateral); anterior intra-alar often weak but rarely absent; presutural present. Middle of propleuron bare; meso- pleuron and sternopleuron with fine white, yellowish, or coarser black hairs. Scutellum with two to four pairs of marginals; first pair strong; last pair, located on posterior lateral angle of scutellum, also strong; intermediate pair(s) weak or absent; in addition, a weak or strong apical pair or rarely two weak apical pairs may be present; pollen of scutellum usually dark basally, light at apex, but sometimes entirely light or dark. Squama white, often with faint brown or pale yellow-brown tinge in center or on most of posterior lobe; infrasquamal setulae usually present and_ black, sometimes absent or yellow. Middle femur with two, occasionally one or three, anterior bristles near middle. Males often with dense row of yellow hairs along inner ventral margin of hind femur, sometimes with rows on both inner and outer margins or no distinct row. Middle tibia usually with one ventral bristle near middle, sometimes a second shorter bristle or no ventral bristle near middle; anterior tibia with one posterior bristle near middle. Abdomen Fires Founp OveR SwarRM Raps oF ARMY ANTS 1037 black, often partially yellow, orange, or reddish and with pale pol- len; second and third terga of most species each with basal band of whitish pollen which widens medially; first and second ab- dominal terga each with a pair of median marginals and a pair of lateral marginals, sometimes with weak row of lateral marginals; third tergum usually with one pair of discals, occasionally one row of four or more; fourth tergum without discals or with one or two rows Or a pair. 4 (8). BACi): KEY TO SPECIES OF CALODEXIA Modified from that in Curran (1934a: 3-4) Occipital “pollen mostly, black) 000) 424. e. Yl es don Occipital pollen mostly white or yellowish ................. Acrosticals 3-1 or 2-1; pale pollinose fascia on second tergum broadly interrupted in middle; one pair of median marginals and one or more pairs of lateral and ventral marginals on SeConGlitergumits Asche. path tc eee fumosa (Townsend) Acrosticals 1-1; pale pollinose fascia on second tergum entire; continuous row of marginals along entire edge of second terpuiiades. voter wets caw peo ess aldrichi Curran Without acrostical bristles Atleast one pairtot .acrostical! bristles#e2 0s). 5.3585). ...0.. Posterior femora orange basally; median postsutural dark meso- thoracic vittae separated by light vitta (seen with light strik- ing the thorax from posterior angle) ...... fulvibasis Curran Posterior femora entirely brownish-black; median dark vittae united behind suture by dark pollen ...... townsendi Curran At least one pair of presutural acrosticals Without presutural acrosticals Three pairs of postsutural acrosticals One pair of postsutural acrosticals (occasionally a second weak TORU): | Re a ee ae ak eae ee Acrosticals 2-3 or 3-3; median dark vittae united behind suture or dark fascia across center of entire mesoscutum except just abOVeEUWwINeS 22. .%. = c0< Acrosticals 1-3; median dark postsutural vittae separated for en- tire weneth by, light ppollen faces ety tan ee varia Curran Third tergum with four discals; acrosticals 2-3, mexicana (Townsend ) Third tergum with two discals; acrosticals 2-3 or 3-3 Outer dark vittae fused with median dark vitta behind suture; acrosticals 3-3; body length about 8 mm. ....valera Curran Outer dark vittae separated from median dark vitta by outer light vittae behind suture; acrosticals 2-3, rarely 3-3; body length 4 to 5.5 mmi., rarely. as large as ‘7 mm: ......-........- at 10 1038 10 (9, 2 16): 19( Tay. 13(12). 14(13). 15(14). 16(14). 17(16). 18(16). 19(18). Tue University SCIENCE BULLETIN 3). Scutellum with reddish-brown pollen confined to basal dark band; third antennal segment with basal orange color extend- ing distally beyond insertion of arista at least as far as distance from arista to base of third segment... panamensis Curran Scutellum with reddish-brown pollen over entire dorsal surface except for narrow posterior and lateral margins; third antennal segment orange from base to arista ....... bequaerti Curran Middle tibia with one or two strong ventral bristles near or be- yond the middle ............2.s2:5¢::.0:: oo Middle tibia without ventral bristle near middle. . fasciata Curran Posterior femora, when seen in ventral view, orange on at least basal fourth, abruptly becomimg orange-brown to black distally and with apices usually orange; or entire femora light yellow to orange in ventral view .. .... ..43.<5 ae Posterior femora entirely brown to black with no abrupt change in color, the apices sometimes orange or reddish .......... Iniasquamal, cilia black or absent... ..... 1 &. 14-6 eee Infrasquamal cilia yellow ............ majuscula van der Wulp Median pale mesonotal vitta extends to well behind suture .... . Median light mesonotal vitta absent immediately behind suture (when light strikes thorax from posterior angle) .......... Male with a row of yellow cilia on both inner and outer ventral edges of posterior femur for at least distal half; female PUTT RTAG Ts 2h chs gee ee flavipes (Shiner) Male with a row of yellowish cilia only on inner ventral edge of posterior femur for at least distal half; both sexes with front coxae yellow-orange; sternopleural and lower mesopleural pile light yellowish to brown; female abdomen dorsally black with bluish-white pollen .................... major Curran Pile of sternopleura and lower anterior mesopleura fine and TEE ere cade. hus ae< |e lacy occa 0S USE k aed ee ee do Pile of sternopleura and lower anterior mesopleura rather coarse AMI DAK oes oy cass. 3 at’ ky mola edad. ee ae First abdominal tergum orange with median longitudinal black stripe in female, with weak stripe or no stripe in males, venteris Curran Dorsum of first tergum all black ......... similis (Townsend ) Outer light vittae absent behind suture or no wider than half width of outer dark vittae: fourth tergum with 2 to 4 discals Outer postsutural light vittae as wide as outer dark vittae; fourth tercum with G=to 10 discals. 2. ..5.. 040. agilis Curran Scutellum with three pairs of lateral marginals; fourth tergum with arched row of four discals ........... insolita Curran Scutellum with two pairs of lateral marginals; fourth tergum with pair of discals (occasionally a marginal pair may look like DESCCONMMP AIK) 9. oP ee ate oe eee apicalis Curran 12 13 20 14 15 16 ily 18 19 Fires Founp OvEeR SwarRM Raps oF ARMY ANTS 1039 20(12). Sternopleura and lower mesopleura with fine whitish hair; outer dark vittae separated from median dark vitta by light vittae Dehincagsuture eA ais eH e ey cace ee deetet eviyepten dad Soci: OM Sternopleura and lower mesopleura with coarser black hair; all dark vittae united behind suture or outer dark vittae sepa- rated from median dark vitta by light vittae ............. 23 21(20). Middle femur with two anterior bristles near middle; dorsum of first tergum black, may be yellow-brown dorsolaterally in TAT Cl GWA ors Reed at cle CASAS TERE APMP Sek pose ied: SAS SMS: 2 92 Middle femur with one anterior bristle near middle; first tergum orange with median dark stripe; female unknown, caudata Curran 22(21). Anterior coxae with total of about 35 (range 20 to 50) coarse black hairs and bristles and a few or no fine white hairs; third abdominal tergum of female with pale pollen extending to margin between marginal bristles .... interrupta Curran Anterior coxae with total of about 12 (range 10 to 18) black hairs and bristles and numerous fine white hairs; third abdominal tergum of female with pale pollen not extending to margin between marginal bristles but small marginal spot of pale pollen, may be present... =". 27.222: %5 2 continua Curran 23(20). All dark vittae united behind suture .......... bella Curran Outer light vittae present behind suture (will key out above through couplet seven if all postsutural acrosticals are siatcrll oC) 8 .4) effects of factors on variables. Data for this graph have been taken from table 3. Correlations aoene factors (primary axes) are not shown since they were quite low (see table 4). 37—5840 1086 THe UNIvERSITY SCIENCE BULLETIN summarize the nature of the covariation in the study by a discussion of the factors. The heart beat of the roach (1) varied from 65 beats per minute to 155 beats per minute in a reasonable approximation to normality, although some evidence of bimodality might be detected. The variable was not transformed. It appears to be affected only by factor I which, as will be seen later, seems to be the factor repre- senting the general energy level of the environment. This is a reasonable relationship based on the knowledge of this character from previous studies (Beard, 1953). The rate of heart beat was doubtlessly affected by the immobilization of the roach and its resultant struggles for freedom. In looking over the louse characters we find that gutbeat (2) varied from 4.5 beats per minute to 58.5 beats per minute and was considerably skewed to the right. A square root transformation when adopted spread the data out considerably, although it still could not be considered to have produced the normal frequency distribution. Gutbeat is affected by factor I, the energy factor, which again reflects the increase of metabolic activity with an in- crease of temperature. This relationship was probably lowered by the fact that the louse was tethered and also since there was varia- tion among individuals in the extent of the inflation of the gut. The rate of locomotion (8) of the louse varied from 1 cm. per minute to 19cm. per minute, with some indication of a skew to the right. The data were not transformed. The rate of locomotion is affected by the energy factor (I) as well as by factor V which influences mostly the louse, but also one aspect of our beetle behavior. An increase in factor V appears to raise the rate of locomotion of the louse. The louse’s angle to light (9) varied from 76 degrees to 176 de- grees. Only one louse reacted at less than 90 degrees which means that essentially all were repelled by light. Of all the insects studied the lice had the most circuitous path as will be seen in the excursion index. The data were not transformed. The distribution of the excursion index in the louse (17), like that of all the other excursion indices in the study, is strongly skewed to the left. The excursion index really represents a percentage. It was found, therefore, that the probit transformation gave an adequate transformation for this as well as all other excursion indexes. The range in values was from .23 to .98. Both the angle to light as well as the excursion index are positively affected by factor V. We may therefore sum- marize the relations of the louse variables in the following manner: An APPLICATION OF FacTror ANALYSIS 1087 gutbeat and the rate of locomotion both appear to be affected by the energy factor (I). It should be noted that the effect of factor I on the louse emerged, although variables (2) and (8) were meas- ured on different individuals. Rate of locomotion is also affected by factor V, which factor will make the louse not only move fastex, but also move away from the light and along a more direct path. Only three flour beetle variables were measured. The rate of locomotion (4) was skewed to the right and ranged from 3 to 78 cm. per minute. A square root transformation managed to make the data more bell shaped. Flour beetle rate was one of those char- acters that did not possess more than a minute amount of variance shared with other variables and therefore no common factor appears to have a high loading on it. We have to conclude that some un- known cause, unrelated to the physical variables measured by us, affected the locomotory rate. The range of temperatures offered, which varied principally from 72° F. to 84° F., did not affect the rate of locomotion. The angles to light (5) varied from 36 degrees to 176 degrees, but this range is misleading since all but two of the animals exhibited an angle of larger than 90 degrees toward light, showing that the beetles were negatively phototactic. The data were not transformed. The excursion index of the flour beetle (15) was relatively high showing approximate straight line movement by the beetle. The index ranged from .05 to .95 but with most data massed between .75 and .95. It was transformed by the probit transforma- tion. When discussing the effect of factors on (5) and (15) we are faced with a situation which is very difficult to interpret. There appears to be no common factor at all affecting flour beetle varia- bles. Variable (5) is negatively affected by factor II, while variable (15) is positively affected by factor V. Factors I] and V seem to be responsible largely for flour moth and louse variables, respectively, and the relations of these factors to the flour beetle are obscure. The five variables measured on the flour moth larva fall into two distinct groups owing to their relations with common factors. The rate of locomotion (10) and the number of undulations per minute (12) were both affected by factor III. The heartbeat (3), angle to light (11) and excursion index (18) were affected by factor II. The rate of locomotion ranged from 1 cm. to 27 cm. per minute. The fre- quency distribution was skewed to the right as well as bimodal. Transformation to the square root of the reciprocal of the rate pro- vided a somewhat more acceptable frequency distribution. The number of undulations per minute had a distribution similar to that of the previous variable. It ranged from 5 to 115 undulations per 1088 THe UNIversiry SCIENCE BULLETIN minute. The same transformation was employed. Factor III has extremely high loadings on the number of undulations and mod- erately high loadings on the rate of locomotion. It appears that this factor is none other than the rate of undulations itself. It is fairly obvious that the rate of locomotion will depend very largely on the undulation rate. It is interesting to note that this relation is not a perfect one, i. e., that there must be factors other than the number of undulations determining rate of locomotion. What caused dif- ferences in undulation rate we cannot learn, except to say that it does not appear to be related to any of the other factors in the study. The heartbeat of the flour moth larva ranged from 6 to 126 per minute and was reasonably evenly dispersed over the various fre- quency classes. The variable was not transformed. Angle to light ranged from 53 to 173 degrees with, however, all but one reading 98 degrees or larger. The readings were evenly dispersed and the data were not transformed. The excursion index of the flour moth larva ranged from .53 to .98 and was transformed by the use of probits. All three variables are affected by factor II. This factor appears to reduce the heartbeat of the flour moth at the same time making it go straighter and farther away from the light. It should be pointed out that the heartbeat was measured on a different indi- vidual from that which performed on the glass plate. An interpre- tation of these findings is given in the discussion of factor II below. Three wasp characters were measured. The rate of locomotion of the wasps (6) ranged from 13 centimeters per minute to 78 centi- meters per minute. It was somewhat bimodal but was not trans- formed. The rate of locomotion was affected by factor I, the general energy factor, as well as by a special wasp direction factor (factor IV), which will be discussed later. Variable 7, the angle to light of the wasp, was one of the variables having no correlations what- soever with any other ones and therefore exhibiting no common variance. It ranged from 6° to 176° and was randomly distributed to each side of 90°. We may therefore state that the reaction of Habrobracon to direction of light was essentially random. What- ever factors did determine this response were not apparent from our study and did not seem to affect any of the other variables. The excursion index ranged from .18 to .98 and was transformed by the probit transformation. It is affected very highly (.94) by factor IV which appears to be a directed movement factor in the wasp. In the housefly larva we studied rate of locomotion (13) which ranged from .$cm. to 18.8cm. per minute in a more or less bell AN APPLICATION OF Factor ANALYSIS 1089 shaped fashion. This particular variable had no common variance as did variable 19, the excursion index of the housefly larva, which ranged from .53 to .98. Variable 14, the angle to light of the housefly larva, ranged from 6° to 176° with most readings in the range above 90°. Its distribution was extremely skewed to the left, most larvae maintaining an angle of 166° or more with the light. Successive trials at transformation failed to remove this skewness to our satis- faction. The final transformation adopted, the reciprocal of the square, while better than the original scale, was still not adequate. It appears that the very skewness of the variable biased the computa- tion of the correlation coefficients in such a manner as to produce spurious correlations. While other housefly variables did not exhibit the effects of common factors, the angle to light was affected by factor III, which otherwise appears to be a moth larva undulation factor. When the moth variables were studied in connection with variable 14, it could be shown that the resulting scattergram rein- forced the bias of the two non-normal distributions. We believe therefore that the effect of factor III on the housefly angle to light is spurious. The physical, so-called causal, variables were the following: tem- perature (20) ranged from 66° F. to 88° F.; most of the temperature readings lay between 72° F. and 83° F. The distribution was rea- sonably bell shaped and no transformation was attempted. Tem- perature is strongly affected by factor I, the solar energy factor. Relative humidity ranged from 28% to 76%, with a distribution similar to that of temperature. The data were not transformed. Relative humidity was affected both by the solar energy factor (I) and the time series factor (VI). Pressure ranged from 28.82 to 29.36 inches. While the distribution was anything but normal it was of such indefinite shape that transformation was not at- tempted. A separate series of light readings had to be taken for each variable as was explained in the Materials and Procedures section. Nearly all light intensities ranged from 25 to 475 footcandles, were strongly skewed to the right and truncated at the left with an ac- cumulation of values at the lower end of the scale. A square root transformation of the data improved the appearance of the fre- quency distribution but generally accentuated a tendency toward bimodality which was present in the data. This bimodality may have been induced by passing clouds. Light was affected both by 1090 Tue UnNrversiry SCIENCE BULLETIN the solar energy factor (I) and by the time series factor (VI). The day of the month (24) was more or less evenly distributed over the 21 days with an average of three readings per day. The hours of the day (25) ranged from 0800 to 1800 hours and showed peaks just before 1200 and just after 1300 hours. The last two variables were not transformed since they could not be expected to have normal distributions. We shall now summarize the effects of the individual factors and speculate upon their nature. The factors, The physical factors will be discussed first followed by the apparently biological factors. The numerical code of the factors is entirely arbitrary when first assigned, but in order to make the factors consistent with a factor analysis of the sub-matrices (see Sokal, Daly & Rohlf, 1961) we adopted the system given here. Factors are therefore not necessarily discussed in numerical order. Factor I is most probably an indication of the variation in solar energy reaching the experimental animal directly through the heating of the atmosphere by the rays of the sun or indirectly through a general insolation of the building in which the experi- mental room was located. Factor I had a very high loading (.98) on temperature and might therefore be thought to represent the temperature factor. From a philosophical point of view this would be an incorrect identification. Those animals responding to this factor react to a change in the thermal energy level of their im- mediate environment. A temperature reading is merely another response, measured by a non-living system, to this same energy level and, under the circumstances of our experiment, is apparently the most sensitive indication of the environmental thermal energy level. Light and relative humidity, which we had hoped would be the other two determining factors of our study, appear from the analysis to have no independent effect on the response variables. Their correlations with the response variables can be entirely ex- plained by the effects of the solar energy factor on light and humidity. The loadings of factor I on these two variables are not very high. Under circumstances other than those of our experiment light and temperature might be more closely correlated, since they are measures of the same source of energy. Under the natural con- ditions of weather affecting the environment in our room a passing cloud may have drastically lowered light intensity, but the warm air and exposed surfaces maintain their temperatures relatively An APPLICATION OF Factor ANALYSIS 1091 constant over a period of time. Furthermore, since the organisms were never exposed to incident sunlight the correlation between temperature and light or between the energy level and the amount of light prevalent in the room was not as high as it might have been. Similar considerations guide us in explaining the low loading of factor I on relative humidity. In addition it should be pointed out that changes in the absolute water content of the air would of course affect the relative humidity and further reduce the correlation of the energy level with relative humidity. The latter variable was also affected by the time series factor (VI) to be discussed below, which indicated a progressive decrease in absolute water content of the air over the period of the study. There was an appreciable loading from factor I on the hour of the day. This can be explained, since in general the later the hour of the day the higher the temperature in the room, except for read- ings performed in the very late afternoon, which were not too numerous. Had such readings not been performed at all the relation between hour of the day and general energy level would have been considerably higher. Two groups of biological variables were affected by factor I. These were the rates of pulsation of heart and gut and the rates of locomotion. Among the first there were strong effects on roach heartbeat and louse gutbeat while the second group included rela- tively low effects on wasp and louse locomotion rate. These effects seem eminently reasonable in view of the vast body of literature dealing with the effects of temperature on poikilothermic animals. The interesting aspects of this situation are the four variables in the study falling into these two classes which were not affected by factor I. Two possible explanations are that strong responses to other environmental stimuli, including those introduced by the ex- perimenter in handling and measurement, reduced the effect of the energy factor or that the variables were not linearly related to the energy factor. The former assumption is appropriate in our situa- tion since scattergrams of the variables concerned did not indicate curvilinear relations with temperature. This group of variables in- cludes the heartbeat of flour moth larvae which was moderately affected by factor II, but only to a slight extent by factor I. Since the highest correlation among factors was between I and II (—.30) and since the effect of I on moth heartbeat was of the magnitude of .24, it might be argued that there is at least a part of factor I which determines this variable. However, these relations are of such a low 1092 THe UNIversiry SCIENCE BULLETIN order of magnitude that we feel hesitant about ascribing any signifi- cance to them. The beetle and the housefly rate of locomotion did not appear to have any significant covariation with other variables in the study. The moth rate, on the other hand, had an independent factor unrelated to the energy level determining its variation. These are surprising and disconcerting results. It occurred to us that the wasp and the louse were animals exposed in the experiment to suboptimal conditions compared to those under which they had been reared for many generations. They also had been used to constant temperatures. Contrariwise the beetles and the moths had been accustomed to temperatures fluctuating more or less over the range to which they were exposed during the experiment. We thought that these latter species might therefore be more homeo- static with reference to temperature changes, while the former ones would respond more readily to temperature fluctuations. A dif_fi- culty of this interpretation is to explain the absence of response in the housefly larva which was habituated to a reasonably constant temperature of 30° C. One might suppose that the housefly larva should have been the most irritated animal in the study, being accustomed to very high humidities and the protection of medium around it, conditions quite different from those of the experimental glass plate. One may therefore speculate that the rate of locomotion in the housefly larva was a function of its general irritation which overrode the effect of variable temperature. Factor VI affects four of the physical variables, but none of the biological variables. It is thus an independent dimension of the physical environment. While it is therefore of little interest in interpreting the correlations among our response variables, it does demonstrate the efficacy of factor analysis in isolating the causes of correlation, The factor appears to be a time series factor or trend and can be identified as the passage of time. It informs us that as the days progressed the barometric pressure became higher during the 22 days span of the experiment. Factor VI also has effects on relative humidity and light intensity. The weather became dryer and brighter during the later days of the study. This trend is related to climatic changes which occurred during the period in the Law- rence, Kansas, area. An examination of weather records revealed that there were more cloudy days and more precipitation in the earlier part of the period than in the latter part. We now turn to a discussion of the biological or behavioral factors revealed in the study. Factor II appears to be a factor primarily An APPLICATION OF FACTOR ANALYSIS 1093 affecting the moth larva. The greater this factor, the more directly away from light a given moth larva will travel and also the straighter its path will be. On the other hand an increase in this factor will reduce the rate of heartbeat of an immobilized larva. It should be kept in mind that the heartbeat was measured on an individual different from the one on which excursion index and angle to light were observed. We do not know the rate of heartbeat in the animal orienting on the glass plate. It is quite probable that immobilization affected the rate that we measured, perhaps by the induction of a pseudocataleptic state while the heartbeat of the moving animal might have increased. The factor analysis, however, shows covari- ance between the two individuals. The interpretation of factor IT is complicated by a negative loading on the angle to light of the flour beetle. This loading, while not of very great magnitude, seems to be the exact opposite of the response of the flour moth larva, i. e., they respond to light in different directions. This is, however, only a relative difference. Both flour moths as well as flour beetles were negatively phototactic under the conditions of the experiment. Factor II determines only their relative aversion to light, which be- came greater with an increase in factor II for the flour moth and less with an increase of factor II for the flour beetle. Since factor II (and factor V discussed below ) affect the animals’ reaction to light, the reader may wonder why under such circum- stances the light intensity itself did not have an independent effect on the organism. It is possible that a correlation of response to light intensity is not detected because of a non-linear relationship such as would be produced by a threshold of response in the lower light values. Our scattergrams showed no such curvilinear trends. There- fore, we interpret this phenomenon along the following lines: the organism is stimulated not primarily by the light but by another environmental or behavioral dimension. Its response to the stimulus of factor II is a movement which orients itself by the light. The light is therefore not the critical eliciting factor but is a variable which serves to guide the response of the organism. Factor III appears to be a factor having to do with the locomotion of the flour moth larva. It is clearly independent of factor II, which also affects the flour moth larva. The number of undulations are primarily responsible for the rate of locomotion, as shown by the high loading on the former variable. The rate of locomotion is, however, not totally determined by the undulations because the flour moths used in the study varied to some degree around the 1094 Tue UNIversITY SCIENCE BULLETIN desired length of 1 cm. There is the disturbing occurrence in this factor of the housefly angle to light. We have stated previously that we believe this to be a spurious correlation. Factor IV can be interpreted primarily as a wasp factor. The straighter the wasp moves, the faster it will move. This factor also does not seem to correspond to any physical factor measured by us. As previously mentioned wasp rate is also affected by the solar energy factor (I). Factor V can be interpreted in a manner similar to factor II. It appears to be a factor affecting the louse. As the stimulus increases the louse will move faster, straighter and more directly away from light. Louse rate is also affected by energy level (factor I). Another flour beetle measurement tends to confuse the interpretation of this factor. Factor V will make the beetle go straighter regardless of direction. We should point out that rate of locomotion and the excursion index are logically entirely independent dimensions of an insect’s behavior. Thus a certain species of insect could move either slow or fast and either directly or circuitously, in any of the four possible combinations. On the other hand it might well be that certain in- sects when moving fast tend to move straighter or tend to move in a more diffuse manner. From the evidence of our analysis it would appear that the wasp and the louse, as seen in factors IV and V, respectively, tend to correlate rate with excursion index, i. e., the faster they move the straighter they move. On the other hand the flour moth appears to have entirely independent dimensions for these two characteristics. In this species there is a positive correla- tion between directness and angle of locomotion, but not between directness and rate of travel. We might add that the house fly larva and flour beetle behaved similar to the flour moth larva in that their rate and excursion index were independent properties. d, Discussion and Conclusions The specific problem of the experiment has been discussed suf- ficiently and does not warrant further treatment here. However, a few words are in order to evaluate the results in terms of the original aim of the study, namely the demonstration of the value of factor analysis as a tool in experimental biological research. Of the six factors uncovered in our analysis only one turned out to be a physical factor apparently behind some correlations among the nineteen biological response variables. This factor evidently An APPLICATION OF FACTOR ANALYSIS 1095 represents fluctuations in the solar energy level during the experi- ment. In terms of the conventional physical variables referred to in such studies it might be labeled “temperature.” While a second physical factor emerged it had no effect on the biological variables and the other four factors appeared to be largely species specific or situation specific. These results are due mostly to a poor choice of experimental biological variables, as well as an inadequate meas- urement of particular features of the environment introduced by the experimenters. Had we chosen appropriate biological response variables it is our conviction, based on the clearly defined structure of present analysis, that the several independent physical dimensions would have shown up quite definitely as factors affecting the cor- relations. The analysis of our data had the advantage of revealing and evaluating unidentified sources of common variation which would otherwise not have been suspected. Since these sources of stimuli may well be produced by the experimenter’s technique or the experimental environment, factor analysis can indicate the need for further control. We are therefore firmly convinced of the value of factor analysis in discovering causal structure in correlated variables and in describing the effects of common factors on re- sponse variables. In view of our present experience we might ask ourselves how an experiment, such as was done by us, might be improved in order that clearer and more interpretable results might be obtained. From the point of view of demonstrating the efficacy of factor analysis it would be far preferable to use one, or at least fewer, species so that species specific factors would be reduced in num- ber if not entirely removed. Some doubt remains in our mind on this point since a study conducted on only one species (but many characters thereof) would very likely exhibit organ specific or organ-system specific factors which again might complicate the interpretation of effects of the physical variables on the biological variables studied. On the other hand, while such a complication muddles the interpretation of the covariation in the study, it is only a reflection of the complicated relationships which are actually found in nature and which belie the arbitrary simplification of biological systems. The application of better measuring and recording equipment, preferably continuously recording equipment, would be likely to lead to much better and higher correlations among the response variables observed. The use of such instruments as minute ther- 1096 THe UNIVERSITY SCIENCE BULLETIN mocouples and kymographs applied to the insect in a physiological investigation would surely result in higher correlations, conse- quently higher factor loadings and presumably in a clearer simple structure, When adapting factor analytic techniques to a field situation the nature of the problem would have to be taken into account before a correct setup is devised. If we are dealing with a study largely concerned with the physiological effects of the environmental vari- ables on the activity of the insects, then a relatively short term but intensive investigation is called for. Activity variables, such as numbers of insects active, types of activity and rates of activity should be studied in a limited environmental situation together with a considerable number of environmental variables meas- ured simultaneously in that situation. If we are studying factors affecting population growth then the study will have to extend over a considerably longer period of time. It would have to involve climatic records of various sorts, population counts and perhaps such variables as age distributions, It is very likely that in situa- tions such as diapause and predator-prey relationships we would have to compute lagged correlations in order to obtain a high and meaningful correlation matrix. Persons familiar with factor analysis will be interested in our evaluation of the suitability of a simple structure constellation for biological investigations. We are not prepared at this stage to claim superiority of simple structure over other constellations and feel that a much larger body of data will have to be accrued from a variety of biological fields before even tentative conclusions can be drawn. On the basis of a priori consideration each of the customary types of factor constellations, such as principal axes, bipolar factors and simple structure can be defended and situations visualized where these constellations would represent the “true” relationship among variables in nature. It might be added that in the experience of the senior author simple structure has given satisfactory and interpretable solutions in each of the several biological cases in which he has applied it. This problem is discussed in greater detail in the technical companion paper (Sokal, Daly & Rohlf, 1961). At least two reservations to the use of factor analysis should be added at this point. The problem of the significance of the number of factors extracted, as well as the levels of significance for factor loadings is one that is still quite controversial in factor analysis circles. Whenever a factor matrix is interpreted a real difficulty AN APPLICATION OF Factor ANALYSIS 1097 develops concerning the level at which a factor loading should be considered significant and an interpretation of it should be at- tempted. Also there are usually shadowy minor factors which may or may not be significant. Much thought may sometimes be spent on interpreting factors of this nature while considerable doubt will remain regarding their significance. It should also be kept in mind that the relations between factors and variables in the factor matrix are strictly linear. Where the relation among variables and between factors and variables is other than linear one should not expect very meaningful factor resolution. Such variables will have to be trans- formed into a scale which will be linearly related in order to get meaningful factor structure. Postscript: During proofreading it was brought to our attention that Habrobracon juglandis (Ashmead) as well as Ephestia kiihniella (Zeller) have both been renamed Bracon hebetor Say and Anagasta kiihniella (Zeller). We have not changed names in this paper since, it would first of all have required much re-editing of the manuscript and since we have doubts on principle about the renaming of organisms well known to the general biological literature. LITERATURE CITED BEARD, R. L. 1953. Circulation. In Roeder, K. D., Insect Physiology. John Wiley & Sons, New York, pp. 232-272. CATTELL, R. B. 1952. Factor Analysis. Harper & Brothers, New York, 280 pp. FRUCHTER, B. 1954. Factor Analysis. Van Nostrand, New York, 280 pp. Hammonp, W. H. 1957. The constancy of physical types as determined by factorial analysis. Human Biol., vol. 29, pp. 40-61. Howe Lts, W. W. 1953. Correlations of brothers in factor scores. Amer. J. Phys. Anthrop., n. s. vol. 11, pp. 121-140. SoKAL, R. R. 1958. Thurstone’s analytical method for simple structure and a mass modi- fication thereof. Psychometrika, vol. 23, pp. 237-257. 1958a. Quantification of systematic relationships and of phylogenetic trends. Proc. Xth Int. Cong. Ent., Montreal, vol. 1, pp. 409-415. SoxaL, R. R., Daty, H. V., and Routr, F. J. 1961. Factor analytical procedures in a biological model. Kans. Univ. Sci. Bull., vol. 42, pp. 1099-1121. Stroup, C. P. 1953. An application of factor analysis to the systematics of Kalotermes. Syst. Zool., vol. 2, pp. 76-92. Tuompson, G. H. 1951. The factorial analysis of human ability. (5th ed.). Houghton- Mifflin Co., Boston, 383 pp. THURSTONE, L. L. 1947. Multiple-factor analysis. Univ. Chicago Press, Chicago, 535 pp. - im ¢) Aaa = 4 rm iPe , 4 7\ eile ; a a. . eae a ee ta ay fe ? 4 wal! ‘pnihies he v <\ oy hue Ay a7 a Pag) stant wf eee ef: ow 4h 9 (ey Wh. satatiwe cer be mat | ‘ ve Ba AN 4 ae ott eae AVA cae Lae oc jai ( he y HARE, MOPED nT t- if Be Fe i ai res itt hin yt =4 ‘A 4; giv ae na : 4 \ ae 4 tin f iz " 0 ‘ oe ed || ee AP y-cnc) - * wer - % mate Gh ets * _ 1%, aes s . f =f nce “eo 4) THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vou. XLIT] DECEMBER 29, 1961 [No. 10 Factor Analytical Procedures in a Biological Model * * BY Rosert R. SoKAL, HOWELL V. Day? and F. James ROHLF * AsstRACT: This study attempts to demonstrate the value of factor analysis in a dynamic as contrasted with a static biological model. It also wishes to investigate the efficacy of a simple structure solution, the maintenance of fac- torial stability in separate and joint matrices and the inclusion of causal variables in the original matrix. Nineteen biological variables were taken from six species of insects. Six physical variables were measured concurrently with the biologi- cal variables. Fifty-five replicate readings were taken. Correlation coefficients were computed among the 25 biological and physical variables. Simple struc- ture solutions by the mass modification of Thurstone’s analytical method (MTAM) are given for the 19x 19 response variable matrix, the 6x 6 causal variable matrix and the 25 x 25 joint correlation matrix. Adequate simple struc- tures are produced by MTAM. Since the reference vectors were largely orthogonal, the varimax solution (Kaiser) and elementary linkage analysis (Mc- Quitty ) of the 25 x 25 variable matrix gave essentially identical solutions. Bio- logical implications of the factor analysis are summarized, but details are discussed in a companion paper (Sokal and Daly, 1961). The article concludes with an account of some recent experiences with MTAM, a discussion of the applicability of simple structure to biological matrices, some mention of the stability of the factor solution in separate and joint matrices and our conclusion that including causal variables in the original matrix is a useful technique for identifying causal factors. INTRODUCTION Aims of the Study The present paper discusses certain statistical problems pertinent to the application of factor analysis to a biological model. It origi- nated in an attempt to set up a model with which to validate factor analysis as a tool for discerning cause and effect relationships. Simi- 1. Contribution number 1040 from the Department of Entomology, University of Kansas. 2. This work was carried out under grants from the General Research Fund, University of pes ae the National Institute of Mental Health, U.S. Public Health Service, Contract o. M-2356. 3. Present address of H. V. Daly: Department of Entomology, University of California, Berkeley, California. 4. The authors acknowledge with thanks the help of Mrs. Maxine L. Howe and Mrs. Ann Schlager with the computations. The senior author is indebted to Professor Raymond B. Cattell of the University of Illinois, who has encouraged him in his studies of factor analysis and has spent much time discussing plans for the study. (1099) 1100 THe UNIVERSITY SCIENCE BULLETIN lar models have been attempted in the past; e. g., the classical anal- ysis of the dimensions of boxes by Thurstone (1947) or the anala- gous study of the dimensions of bottles by Barlow and Burt (1954). A study by Dickman of the factors determining the amount of bouncing in balls of different sizes and materials has not as yet been published. The two published studies suffer from the same draw- back in that they describe essentially static, geometric situations. Some psychometric theories consider the mind in a static, multidi- mensional manner. Whereas the authors in their ignorance of psy- chological theory are not prepared to contest this view, it is our considered opinion that static, geometric models would not repro- duce reality in terms of biological processes and therefore would be of little interest for experimental biologists. Since it is also our view that in the biological sciences it should be considerably easier than in psychology to reify factors, it became important to us to con- struct a model in which known and dynamic factors would be oper- ating. The aim therefore, was to find a series of response variables which would be in part determined by a few, preferably three, causal variables which were well known to the experimenter and which exhibited, in general, relations compatible with the concept of simple structure, i. e., that every factor does not affect every re sponse variable and that any one response variable is not affected by all three factors. The study was undertaken on insect material, this being easiest obtainable to the authors and the material with which they are most familiar. It was hoped that a study of the response variables (over a series of observations sufficiently numerous to inspire confidence in the resulting coefficients) would yield a correlation matrix which, when subjected to factor analysis, would yield in turn the variables known to be causing covariation in the study. We anticipated that such a demonstration would lend support to the practice of factor analysis and would serve as a stimulus to experimental biologists to consider factor analysis as a complementary or supplementary technique in the analysis of physiological and ethological phe- nomena. It was expected that a number of points of interest to factor analysts would emerge as byproducts of the study, as follows: (1) The further application of an analytical method for finding simple structure, recently developed by the senior author (Sokal, 1958); (2) The applicability of the concept of simple structure to the case under consideration; (3) The maintenance of factorial stability when correlation matrices. are analyzed separately, and Facror ANALYTICAL PROCEDURES 1101 subsequently joined together; (4) The inclusion of causal or inde- pendent variables in the original correlation matrix. This latter idea has been proposed by Cattell (1952), but has not, so far as is known to the authors, received wide attention or application. The experimental setup and the biological findings are detailed in a companion paper (Sokal and Daly, 1961). Their mention and discussion in the present paper is incidental to its major purpose, namely, a review and discussion of the statistical aspects of the problem and specifically the four points mentioned above. Limitations of the Study After considerable experimental effort with a number of insect species, it was found that the desired experimental design was im- possible of achievement. A series of variables whose causation was known and in part determined by three major causal variables could not be found. We had to settle for variables whose causation was only partly known, if at all, and whose mutual intercorrelations were not predictable. Furthermore, it was not known whether the relation between causal and response variables would follow a sim- ple structure constellation. When the data were analyzed and in- terpreted it was found that four of the “causal” variables appeared to share variance with the response variables. Many of the response variables appear to be correlated and determined by causal vari- ables whose nature was unknown and which differed from the ones ini’ially considered. In the sense of having developed a known dynamic model and then re-established it by means of factor analysis our experiment was thus only partially successful. We feel justified in bringing our findings to the attention of the public interested in factor analysis because of the information we obtained on points 1 through 4 above. As will be seen, our results suggest that with better equipment and more effort a similar model could be designed and reproduced by factor analysis. MATERIALS AND METHODS Experimental Materials and Procedures Table | lists the kind of insect and the stage studied as well as the biological variables measured and their code numbers (1-19). The nonbiological variables were the following: (20) temperature in °F; (21) percent relative humidity; (22) atmospheric pressure in inches of mercury; (23) light in foot-candles; (24) the day of the test; and (25) the hour of the day. 1102 THE UNIVERSITY SCIENCE BULLETIN TABLE 1.—List of Insect Species Used and Characters Measured SCIENTIFIC AND ComMon NAME; CopE NUMBER Stace or INSECT AND VARIABLE Periplaneta americana L. 1. Heartbeat rate American roach, adult. Pediculus humanus DeGeer 2. Gutbeat rate 8. Rate of locomotion 9. Angle to light 7 Human louse, adult or large nymph. Excursion index Rate of locomotion Tribolium confusum Duval 4. 5. Angle to light 5 Confused flour beetle, adult. 15. Excursion index Ephestia kiihniella Zeller 3. Heartbeat rate 10. Rate of locomotion Mediterranean flour moth, larva, 11. Angle to light 1 cm. in length. 12. Undulation rate 18. Excursion index Habrobracon juglandis (Ashmead) 6. Rate of locomotion Parasitic wasp, adult, no more than 5 7. Angle to light days since ecdysis. Females only. 16. Excursion index Musca domestica L. 13. Rate of locomotion 14. Angle to light House fly, 5-day-old larva. 19. Excursion index The experiments were conducted in a room illuminated by natural light from a window. The insects were brought to the room and prepared for the test as follows: One of each species, except the roach, was put into a separate vial and a roach, a louse, and a moth larva were immobilized on pieces of glass. Then they were placed for at least an hour in a special container which admitted a current of air from a fan, but excluded all light. After this period of ac- climatization, the insects were tested on a horizontal stage of frosted glass opposite the window. Before each insect’s performance the temperature, relative humidity and light were measured. Averages of the readings for the temperature and for relative humidity were used as indicative of these variables. Separate light readings were employed since these varied widely during the hour of testing. The pulsations of the dorsal vessel (heart) of the roach and the Facror ANALYTICAL PROCEDURES 1103 moth larva, and the peristaltic waves of the gut of the louse were observed with a binocular microscope in the immobilized insects. They were recorded as beats per minute. The insects in the vials were now released on the stage in a standard sequence and the route of escape of all but the house fly larva was traced in ink on a pane of glass suspended one half inch over the frosted pane. The house fly larva traced its route on the frosted glass with an aqueous solution of methylene blue. Equal units of time were signaled by an as- sistant and marked along the insects’ tracks. Ten of the tracks were less than 5 cm. and were discarded; the remaining routes were traced on paper. Distance between time marks, as measured by a planimeter, gave the rate in cm. per minute. Units of 5 cm. were then marked along the track. The excursion index (linear distance from point of origin to last 5 cm. mark divided by the total distance of the insect’s journey to that mark) gave an indication of the extent to which the track deviated from linearity. The 5 cm. marks were now connected by straight lines along the track and the angle measured between each straight segment and a line perpendicular to the window. An average of these angles gave an average angle to light for the track. For insects attracted to light the acute angle was recorded, while for those moving away, the obtuse angle was recorded. An additional variable, the number of locomotory un- dulations, was recorded for the moth larva. No individual insect was used more than once and there were 55 replications over 22 days. Further details on materials and methods may be obtained from the companion paper (Sokal and Daly, 1961). Statistical Treatment of the Data Two minor statistical problems had to be overcome before and during the computation of the Pearsonian product-moment correla- tion coefficient. The variables had to be transformed in order to approximate normal or at least symmetric distributions. A good many of the variables were extremely skewed and/or bimodal in the scale in which they were recorded. A number of different transformations were tried on each variable where transformation seemed advisable. While we were guided by empirical results of the transformations, logical transformation scales such as _ re- ciprocals for rates and square roots for counts were preferred. The following variables required square root transformation: 2, 4, and 23. The excursion indices (variables 15 through 19), which were initially recorded as a proportion and strongly skewed to the left, 1104 THE UNiversiry SCIENCE BULLETIN improved considerably when transformed into. the probit scale. Variables 10 and 12 (both rate phenomena) appeared to be best transformed by taking the square root of the reciprocal of these variables. A positive loading on variables 10 and 12 really cor- responds to a deceleration of the rate, while a negative loading achieves exactly the opposite. Variable 14 was transformed by taking the reciprocal of its square. Variables 24 and 25, days and hours respectively, were not trans- formed in spite of their non-normal] distributions since normal dis- tribution of these variables was hardly to be expected. The computation of the correlation coefficient was also compli- cated by the absence of observations for certain variables at given times. This necessitated a particular program to be written for the IBM 650 digital computer which computed correlation coefficients of a correlation matrix with missing observations. The details of this method do not concern us here except to state that the correla- tions were, of course, based only upon observation pairs and the lower number of observations for any pair of variables determined the number of pairs on which the correlation was based. The mini- mum number of observations in any one variable was 34 (in vari- able 17). Only very few variables had that few replications; most variables had around 45, while the maximum number of readings was 55. In computing the significance of correlation coefficients in the resulting matrix the different sample sizes were taken into con-_ sideration. RESULTS The Correlation Matrix In order to save space the correlation matrix among the 25 vari- ables is not shown here. For those who are interested it is figured in the companion paper already referred to (Sokal and Daly, 1961). The generally low correlations among the variables observed were a disappointing feature of the study. The average absolute correla- tion was 0.15. There were 44 significant correlations out of the 300 correlation coefficients in the matrix. This is, however, considerably above the 5 percent level of significance. Our assumption that sig- nificant correlation existed in the matrix was borne out by a study of the frequency distribution of the correlation coefficients: which turned out to be mildly leptokurtic (g; = .544, .1>P>.05). Vari- ables 7 and 19 had no significant correlations with any other vari- able, but were included in the study because it was felt that it would be of interest to observe their behavior during rotation to simple structure. Facror ANALYTICAL PROCEDURES 1105 All computations in correlation work, as well as all subsequent computations, were carried out to 8-place accuracy and were per- formed on the University’s IBM 650 digital computer. Simple Structure in the 19x 19 Response Variable Matrix In line with the aims of the study, the first analysis was carried out solely on the 19x19 submatrix of response variables. The matrix was subjected to principal axis factor analysis with reduced communalities. The analysis was repeated three times until initial communality estimates agreed with the computed communalities below a difference criterion of .05. At this point communality esti- mates were considered stable. The problem of the number of fac- tors to be extracted was attacked by the method of testing for sig- nificance of partial residuals as described by Sokal (1959). By this method 5 factors were indicated. The fifth residual matrix showed only three significant partial correlations out of a possible total of 171. This is considerably below the accepted 5 percent level of significance. The findings were borne out by a study of the latent roots of the correlation matrix. An impressive decrease in magni- tude appeared between the fifth and sixth latent roots. Subsequent to the original decision a computer program called RESITEST-I was developed by one of the authors (F. J. R.) which enabled us to test the significance of the residuals very rapidly by a variety of methods. The results of these tests for all three correlation mat- rices are shown in Table 2. TaBLeE 2.—Summary of the Decisions on Number of Factors in the Three Correlation Matrices Reached by the Various Methods Matrices Test EMPLOYED 6x 6 19x 19 | 25 x 25 siurekerzsreniterlonews tee eset ico. eee le 2 Sie Live th Die ate MoNemarsvcriterionea. 2.54 ne sesso 2 5 6 . . | ior SAUNG ELS aCriceMOnen eal ere ea ae >4 i 9 i Sz | Individualiresidualsi eee eee ae 2, 5 6 Emp reyes Ulery eve se eee eter ces es oe 2 1 3 Number of factors decided upon........ D, 5 6 1106 THe Universiry SCIENCE BULLETIN The F, matrix was rotated to simple structure using the mass modification of Thurstone’s analytical method (MTAM) as de- veloped by Sokal (1958). The simple structure obtained by the MTAM method was reiterated three successive times. The aim in this undertaking was to see whether increasingly improved simple structure could be obtained on successive reiteration by the MTAM method. Surprisingly it turned out that fewer variables were in the hyperplanes at the later than at the earlier iterations. Also, various test vectors began to become correlated during the later iterations. These findings seem to contradict the senior author's earlier sup- position (Sokal, 1958) that on successive iterations the simple struc- ture would be improved. We therefore used the first of the itera- TaBLE 3.—MTAM Simple Structure and CR Matrices of 19x 19 Correlation Matrix Reference vectors Variable number I II III IV Vv TRGYEVEIM OVEE NS 6.2) doece NSP Re 1 71 16 —16 04 —02 ousckbeatian Aston fae he 2 64 47 13 24 00 Hphestiaibeat.. ct. dese nesi. 4. 3 42 —38 03 21 -33 AMnlloohivinn igh, ood do kone sees 4 04 —21 -31 10 -17 Mripoliumyancleye ests. ae. 5 04 | -39 00 05 36 Habrobraconmater.... 0+... .. 6 39 10 —31 —22 —10 Habrobracon angle........... 7 —03 | -10 O7 | -21 —06 WGousSemate ees ce ces iS se ee aes 8 48 02 09 —12 41 WKousevanclewereaee oka eae S 9. -11 -03 | -06 06 63 Biplestia UW /Tabe. jee a. sy ae eles 10. 08 06 56 | -04 | -21 Bphestis angles. if2%).c.c0 0s». is 20 67 | -00 09 00 Ephestia 1/undulations....... 12. —11 —03 89 | -11 06 INiISCamra beeen tens. sos oo 13. 10 08 —22 47 —02 Miuscanli/anplenras ict sc Naas: .: 14. —08 06 44 -34 | -27 Striboliummndexe)...:... +... 15. —O1 -O1 —05 08 45 Habrobracon index........... 16. 10 -12 —28 —70 ~15 Wousenmcd exe yar depen: sicideaecs fhe 05 03 20 —09 iss Ephestia index. 2... 2st 18. 03 44 | —40 O4 02 Miuscatindexes.-.-. 0s. soe 19. —28 15 —09 08 —07 CR I II III IV V I XG 30 —10 04 05 II 30 xX 07 02 —07 III —10 07 xX -19 —21 IV 04 02 -19 xX —02 V 05 —07 —21 —02 X Decimal points are omitted before coefficients in this and following tables. Facror ANALYTICAL PROCEDURES 1107 tions, which gave by far the best simple structure. The simple structure matrix is reproduced in Table 3, which also shows the correlation between the reference vectors (CR matrix). The table shows that three variables (4, 7, and 19) had no loading larger than 35. This was to be expected since these three variables had very low communalities. A discussion of the significance of this matrix will be postponed until a later section. Simple Structure in the 6x6 “Causal” Variable Matrix The 6x6 correlation matrix of the so-called causal variables was subjected to a principal axis factor analysis with reduced com- munalities in the diagonal. The procedure followed was very similar to that described in the previous section for the 19x 19 response variable matrix. It took three iterations to stabilize the communalities within the criterion of .05. A number of tests of completeness of factor extraction were performed which in sum- mary indicated that two common factors sufficed to account for the covariance in the matrix (see Table 2). The first cycle of the MTAM procedure resulted in a very satisfactory simple structure TasLe 4.—MTAM Simple Structure and CR Matrices of 6x6 Correlation Matrix Reference vectors | Variable 2, number I (ye SCTHPERADITER Pea sete he nw has et 20 —04 96 Relative Mmumidityie< 5). gi... se eos ve de ee 21 —49 —71 Atmospheric pressure); .:.. so o .35). The correlations among reference vectors shown in the same table indicate generally orthogonal relations with Facror ANALYTICAL PROCEDURES 1109 TaBLE 5.—MTAM Simple Structure and CR Matrices of 25x25 Correla- tion Matrix Reference vectors Variable number I II Ill IV Vv VI TR@gveln IBIIB wk baal a 1 2 05 —05 14 04 -03 MouserGuti Bee yee. oo 2 61 33 21 —20 09 -17 Hiphestianhe Bes. jc. 3 23 —57 05 —07 —34 —O1 Tribolium rate........ 4 08 —23 —32 06 —25 —12 Tribolium angle....... 5 10 44 02 04 33 —05 Habrobracon rate...... 6 38 O07 —24 48 —04 —l|1 Habrobracon angle..... 7 -12 | -11 06 32 | -01 22 Louse rate............ 8 40 —04 lat 11 44 O07 housevanglet= ayes 9. —06 10 | -05 | -03 56 04 Ephestia 1/rate....... 10. 14 | -01 58 | -09 | -17 08 Ephestia angle... ..... yk 27 62 03 01 113% O1 Ephestia 1/undulation Ze 00 | -02 91 —28 05 | -02 Miuscarategs.2).--0.. 13. 21 O7 14 31 13 09 Musca l/angle........ 14. —05 12 48 183) S283) 1G Tribolium index....... 115}. 20 O08 10 04 49 02 Habrobracon index... . 16. O07 03 —32 90 —06 -15 Louse index.......... Ne 09 10 17 —09 a —O1 Ephestia index........ 18. 06 40 —33 09 05 03 Musca index.......... 19. —20 24 06 05 14 15 Temperature.......... 20. 92 20 Ol O07 —04 O7 1815 (ane ae 21. —60 01 02 08 O8 —53 IPTESSUTO eS cee kee ade. DOE 10 7 —24 —O1 21 63 DEA VGa dG oo ae ae oe PB 53 —16 O01 05 —23 41 Day of month........ 24. 03 —O1 08 -13 02 83 Hounofiday 22.42... . 25. 48 07 17 06 15 27 CR I II Ill IV V VI I ax By 11 09 05 03 II 32 me |) ale! O08 13 —12 Ill 11 14 xX DO, 15 Ol IV 09 08 —22 xi 14 —O1 V 05 13 -15 14 xX —O1 VI 03 12 O01 Ol Ol xX the possible exception of reference vectors I and II, correlated to the extent of .32. A remarkable feature of the analysis became apparent, when it was noted that the factor loadings observed in the 25 x 25 matrix were nothing but composites of the factor loadings observed in the 19x 19 and 6x6 matrix. Factor I of the 19 x 19 matrix and factor II of the 6 x 6 matrix re-emerged as factor I of the 25 x 25 matrix. Fac- 1110 Tue University SCIENCE BULLETIN tors II through V of the 25 x 25 matrix were numbered to correspond with factors II through V of the 19x 19 matrix, which they match. Factor I of the 6x 6 matrix is factor VI of the 25x 25 matrix. The relative invariance of the results from the three matrices lends en- couragement and support to our belief that the MTAM method does produce a reasonably unequivocal simple structure and that the simple structure solution itself is a relatively meaningful position for the data shown here. The latter belief, however, will need further justification, which we hope to provide in the following sections. The Varimax Solution and Centroid Linkage Analysis An alternative method of taking the data to simple structure is the Varimax method developed by Kaiser (1958). Through the courtesy of Dr. Kaiser we were able to process our 25 x 25 matrix on the ILLIAC computer using the Varimax routine. Table 6 shows TaBLE 6.—Varimax Simple Structure of 25x 25 Correlation Matrix Reference vectors Variable | number) i | vou Tt | vi 7) ees | A B C D E F Roachw#HeBieanon os. - 1 76 O1 -06 | -02 05 04 ouse:Gutplssere 42 ene « Dis 57 —09 17 10 —20 -30 Ephestia H.B......... 3. 34 31 04 00 62 | -03 Tribolium rate........ 4. 16 30 | -—33 10 20 00 Tribolium angle....... 5S. —O1 —38 | -02 06 47 01 Habrobracon rate...... 6. 43 15 —22 11 —06 38 Habrobracon angle..... rhe -11 02 10 —17 O08 38 Wouse maleic see ea-5 « 8. 43 —44 12 —08 | 13 09 Louse angle........... 9. O7 57 05 —04 —09 —04 Ephestia 1/rate....... | 10. 08 09 58 | —06 10 | -0l Ephestia angle........ le 15 | -07 03 | -04 | -61 | —08 Ephestia 1/undulation Ue |) 0s zit 90 05 16 | -14 IN MOE H TEI) o.4 w og occ o}- Akg 15 -17 —16 —07 —36 Musca 1/angle........ 14. —09 18 50 ie —O7 19 Triboleum index...... . | Nfs). -18 51 10 01 06 | -02 Habrobracon index... . | 16. 15 Paik || 5) 22 | —08 84 Wouseimdexes sc. «1. bese 08 81 17 Ol —02 —08 Ephestia index........ Seen | OL 04 | -32 | -06 | -46 | -Ol Miuscajindex.........- 19. —27 15 06 14 31 —04 Temperature.......... 20. 92 11 oO | -14 | -08 | -06 [Roel GG), Oi tata eect eee Pale —56 —09 02 59 —04 10 IPTeSSIILG weeteedte: coe, < | Pe Ol 17 21 65 —26 00 ight tem te ss. fe coor 53) 1) 95 | s085 etn 19 | 05 Dsysot ammount. «2+ | 24. —08 | -05 12 | -83 | -06 | -03 lsloiiy Or CEnio so oot noe 25. 55 19 —20 21 15 —17 Reference vectors are identified by capital letters. The Roman numerals refer to the reference vectors from Table 5 which the Varimax vectors resemble. Facror ANALYTICAL PROCEDURES julal the Varimax simple structure solution of the 25x 25 correlation matrix. Readers will be struck by the similarity of this solution with that shown in Table 5 by the MTAM method. The Varimax solu- tion is an orthogonal solution forcing the reference vectors to remain uncorrelated. It was to be expected that this solution would be similar to the one found by MTAM since in this particular case the reference vectors were not very correlated, the highest correlation coefficient being .32 between reference vectors I and II. It must not be assumed, however, that in cases where the correlation be- tween reference vectors will be high Varimax and MTAM are likely to give the same solution. We also undertook an elementary linkage analysis (McQuitty, 1957) of the 25 x 25 correlation matrix. This yielded five clusters which contained the same variables as five of the six factors which eventually resulted from the simple structure analysis. A similar result was obtained when centroid linkage analysis (McQuitty, 1957) was applied to the 6x6 matrix. When factor loadings were estimated from the centroid linkage analysis, using the known com- munalities of the 6 x 6 matrix, the typal relevancies agreed to within .04 with the factor loadings obtained by MTAM. It is not surprising that these various methods of cluster and linkage analysis yield results similar to Thurstone’s method for simple structure, since as was mentioned previously, the correlations among the reference vectors in this study were relatively low. Whenever the factors are discrete and have very little correlation among themselves, separate non-overlapping clusters are formed in the correlation matrix and are often quite easily discernible by simple methods of cluster analysis. This is not to recommend cluster analysis in place of factor analysis, because in cases where factors would have large areas of overlap or be highly correlated with each other cluster analysis would be of no particular value. DISCUSSION Biological Implications Since a suitable test of significance for factor loadings is lacking, we have arbitrarily selected a level of .4 as the lower limit for important loadings. The six factors can be clearly separated into two kinds: physical and biological. These two categories of factors affect only a few variables in common. Factor (VI) did not affect biological variables. It loaded highly on “day of experiment,” indicating a time series trend over the span RE THE UNtversiry SCIENCE BULLETIN of the experimental period. During this period pressure increased, the atmosphere became drier and average light intensities during the observations rose. The last two variables, as well as “hour of the day,” are affected by the second physical factor (1) which is closely related if not identical with temperature. This factor is more properly regarded as a measure of the general solar energy of the environment. This energy factor was the only physical factor with appreciable loadings on biological responses, i. e., rates of pulsa- tions of the roach heart and louse gut. These rates were probably also affected by the immobilization required for observation as well as by physiological condition of the insect. The predominant in- fluence, however, is that of the energy level which is a relationship to be expected in cold-blooded animals. Two rates on locomotion, that of the wasp and that of the louse, are also affected by Factor I. These two variables load on additional factors. A wasp factor (IV) shows loadings on the above rate as well as high loadings on wasp excursion. This means that wasps tend to run much straighter with an increase in rate of travel. The remaining wasp variables, angle to light, shows no covariance with the other variables in the study. In a similar manner, the louse rate is affected not only by the energy factor, but also by a factor (V) which influences all the remaining louse characters. An increase in this special factor means an increased rate of locomotion and a much straighter route directed more away from light. Factor V also has loadings on the excursion index of the beetle. This would indicate a common source of stimu- lation not measured as a physical factor. Such a source may have been provided by the experimenter in handling and measurement of the insects, by irregular vibrations in the building, or by visual patterns in the room which influenced certain features of the escape routes of the two species. The beetle’s angle to light is affected by a factor (II) which otherwise shows loadings only on characters of the flour moth larva. This is the second factor in which the source of stimulation is un- known. The rate of locomotion of the beetle showed no covariance with other variables in the study. An increase in factor II gives a route more toward the light for the beetle while for the moth larva, it gives a straighter route more away from the light. The different orientations to light are only relative since both species were re- pelled by light. The heartbeat of the immobilized flour moth larva is reduced by an increase in this factor. It should be remembered Facror ANALYTICAL PROCEDURES 1118 that the heartbeat and the features of the route were measured on different individuals of the larval form. While the analysis shows covariance between these characters we suspect that the heartbeat of the larva was affected by the confinement and probably different from the unmeasured heartbeat of the larva orienting on the glass plate. Finally, the rate of locomotion and the number of undulations of the flour moth larva are affected by Factor III. In this case it is obvious that the undulations are the primary locomotory move- ments, but this relation is not absolute. The loading of the house fly larva apears to arise from a spurious correlation. The relationships shown by the analysis are reasonable in most cases. The complications are found in those factors which lack loadings on physical variables, but influence more than one species of insect. The mixture of species affected by Factor III is probablv the result of spurious correlation. The mixtures in Factors II and V seem to indicate common stimuli whose sources are unknown. It is quite possible that the source is provided by the experimenter’s activities, treatment of the insects, and the experimental apparatus. If this is the case, the analysis has revealed and evaluated appreci- able components of variation which would have been routinely lumped as experimental error. It is not unusual that the experi- menter represents a dynamic influence in the experimental environ- ment. A more detailed discussion of the factors can be found in Sokal and Daly (1961). Factor Analytic Implications (a) Further experiences with the mass modification of Thurstone’s analytical method The method for arriving at simple structure employed in this paper was developed by the senior author (Sokal, 1958). The ref- erence just cited included the application of the method to four matrices from various biological sciences which yielded satisfactory simple structures by the method abbreviated as MTAM. Since the above work was done, two of the present authors (R.R.S. and F.J.R.) have had considerable experience in applying MTAM to other correlation matrices. In addition to the three matrices dis- cussed in the present paper a Q-type matrix involving correlations among 23 species (the matrix used as an example in Sokal and Mich- ener, 1958) was rotated to simple structure, as well as five smaller matrices of dimensions from 9 to 14 based on a variety of biological 1114 THE UNIveRSITY SCIENCE BULLETIN situations. All of these yielded simple structures which by the cri- teria stated appear to be satisfactory. In carrying out the steps for an MTAM analysis we became con- cerned whether the empirical procedure suggested by the senior author (Sokal, 1958, pp. 251) would bias the simple structure along either of the two criteria used to establish it. Test vectors are chosen both by the number of variables which each reference vector carries in its hyperplane as well as by the number of low correla- tions with other reference vectors. Exclusive emphasis on the for- mer criterion would yield simple structures with a maximum num- ber of variables in the hyperplanes, in extreme cases with almost specific factors, while these factors could potentially be quite highly correlated. On the other hand overemphasis on the second criterion could result in an essentially orthogonal factor matrix but with not clearly defined factors which might have intermediate size loadings on many variables. It has been our experience with these various matrices that we were rarely, if ever, put before the difficult choice of evaluating one criterion against the other. If a variable was satisfactory by one criterion it was in general satisfactory by the other criterion. In cases where the two criteria did not match up it has been our con- sistent policy to give somewhat greater weight to the criterion of number of variables in the hyperplane, rather than that of low correlations, since we feel that such requirements are more in the spirit of the simple structure. The inclusion of a criterion prefer- ring orthogonal factors (all other things being equal) resulted in quite low CR matrices in our study, as can be seen in Tables 3, 4, and 5. It should be pointed out that MTAM does not require or- thogonality of reference vectors per se. It chooses reference vectors which are orthogonal to many other reference vectors but which are not necessarily orthogonal to specific vectors chosen for our analysis. While the reader may get the impression that orthogonality is a principal goal when he examines the Cr matrices in this paper, Cr matrices in other studies showed appreciable high correlation co- efficients. It has been our general experience that the number of clusters of reference vectors which are established as part of the MTAM pro- cedure is usually close if not identical to the number of factors that have been extracted. Thus in the 25x25 matrix, where six fac- tors were extracted we found six reasonably clear clusters of refer- ence vectors in the Cp matrix. Similarly in the 19 x 19 matrix, where Facror ANALYTICAL PROCEDURES 10s five factors were extracted, five clusters emerged and in the 6x6 matrix only two clusters appeared, matching two extracted factors. These findings are also borne out by similar studies on other mat rices. By way of an experiment we re-analyzed the 25 x 25 correiation matrix by Thurstone’s complete centroid method, but only extracted three factors, knowing this number to be too low. We iterated to stability on three factors and then subjected the resulting factor matrix to an MTAM analysis. Three, possibly four clusters of ref- erence vectors resulted. One of these clusters appeared to represent factor I as established by the previous analysis; the others could not readily be identified with any of the previous factors and did not appear to provide more interpretable solutions. Underextraction thus appears to lead to unclear results. While one or more factors may be identical to those which might result from MTAM analysis of a complete extraction, most factors from the underextracted matrix would probably differ and be less desir- able. MTAM, however, does not provide an easy check demon- strating underextraction of the correlation matrix. It is our expectation, which we have not yet been able to corrob- orate, that overextraction of the matrix by the centroid method and subsequent MTAM treatment would bring out the fact that too many factors had been taken out. This might come about by find- ing fewer clusters of reference vectors in the Cg matrix. If as many clusters appeared as extracted factors the supernumerary vectors might have low and confused loadings on the variables. We have already stated previously that the iteration of the Vj, does not provide increasingly clear simple structure although the senior author had earlier supposed that it would (Sokal, 1958). Extensive corroboration of this point will have to await improved computational methods as the re-iteration of a sizeable matrix is still a tedious undertaking. Since the first publication of the MTAM method the computation has been considerably simplified by having most of the steps pro- grammed for the IBM 650 digital computer with indexing accumu- lators and floating-decimal arithmetic by one of the authors (F.J.R.). The program is entitled Matrix IV and uses the output of a complete centroid extraction as input. It will perform almost all the necessary computations except for the weighting, which is carried out by another program called MATRIX WEIGHT, and the matrix inversion which requires the program entitled INVERSE 1116 THe UNIveEerRSITY SCIENCE BULLETIN SYMMETRIC MATRIX (by U. W. Hochstrasser modified by Rohlf). MATRIX IV as written now requires the repeated passing of the data through the digital computer. Ideally the data should be subjected to MTAM with one pass through the machine, but this cannot be done on the limited memory (2000 words) available on an IBM 650 without accessory storage units. (b) The applicability of simple structure to biological matrices The use of the simple structure constellation has been a some- what controversial one in the psychometric literature. Different schools of thought have developed regarding it both in this country and in England. The authors are not prepared within the context of the present paper to review the history of the controversy nor even to outline the main positions of the protagonists. On the other hand it seems to us that some justification is necessary for rotation to simple structure in a biological matrix. Simple structure is a hypothesis of simple factors in nature repre- senting the action of few forces in relatively simple patterns. This concept, epitomized by natura est simplex, which was in vogue from the beginning of experimental biology in the late 19th century until the first two decades of the present century, has lately gone out of fashion. However simple and elegant a biological concept may be and however necessary it may be to simplify initial biological hypotheses in order to be able to prove their existence, subsequent study and penetration of the problem invariably results in relations considerably more complex than they were initially believed to be. This is true even for genetics, the biological science most often held up as a model of exactness and simplicity. Modern genetics honors the Mendelian laws by their breach rather than by their observance. Initially simple “diagrammatic” phenomena such as sex determina- tion have been shown, in any organism in which they were carefully examined, to be quite complex, depending on equilibrium condi- tions of a whole series of factors. Whatever branch of biology one chooses, be it physiology, endocrinology, developmental mechanics or evolutionary mechanisms, one finds therein consistently a multi- plicity of factors interacting in a most complex manner to produce the results which are observed and studied. Is it therefore plausible that we should simplify complex phenomena of covariation among variables by reducing them to a few dimensions, each of which af- fects only a very few of the total array of characters? A second doubt arises from the problem of the number of factors to be extracted from a correlation matrix. Classical factor analytic Facror ANALYTICAL PROCEDURES Py theory has tried to relate the number of factors to the rank of the matrix. Attempts have been made to test for the significance of the residuals left after a number of factors have been extracted. In a provocative recent article Cattell (1958) re-examined this entire problem of number of factors to be extracted. He reasons, and very cogently so, that undoubtedly each correlation matrix is caused by many more factors than there are variables in the study. There- fore the question of the number of significant factors which can be extracted from a correlation matrix, which is customarily assumed to be smaller than the dimension of the correlation matrix, is really meaningless. Thus the entire matter of testing for the significance of residuals, which has never been in a very satisfactory state, is reopened. We bring up Cattell’s opinion largely to show that it may be erroneous to suppose that only a few factors determine covariation in a matrix. It may well be that the factors which we extract are purely mathematical constructs, which conveniently summarize the information given by the correlation matrix but do not at all repre- sent reality. Had each of these many factors roughly equal and low influence on all the variables in the study, a vague, perhaps inde- terminate and certainly meaningless structure would result during a factor analysis. On the other hand, if factors are unequal in their importance and unequal in their effects on different variables, it may well be that a structure somewhat similar to the simple structure pattern may be found. We may question whether we have a con- tinuum of factor loadings from very high to negligibly low loadings or whether the distribution of the magnitude of these loadings is bimodal or multimodal so that a series of definite, partly overlap- ping clusters is formed by the effects of the factors on the variables. Simple structure solutions as practiced today usually show bi- modality of loading magnitudes. As a matter of fact some of the analytical solutions proposed (see Saunders, 1953; Neuhaus and Wrigley, 1954; Kaiser, 1958) in effect force bimodality upon the loadings. It seems difficult to see any a priori reason for bimodality. One would rather suppose that factors had an entirely continuous gradient of effects on variables. When the cause-and-effect system is looked at from the point of the characters or variables, we are led to ask how many factors affect each variable. Simple structure would indicate that only few of the factors in the study affect any given character. However, it may quite legitimately be claimed that to some degree all of the factors 38—5840 1118 Tue UNrversiry SCIENCE BULLETIN affect a given character. Genetic mechanisms presumably underly many of the biological phenomena studied. An appreciable number of geneticists would tend to support the idea that every gene affects every character. There must then be an entire spectrum of gene action from the most drastic and immediate kind of effect to the most subtle pleiotropic effect of a gene upon another character. Philosophical reasons may exist for preferring one or the other of these contrasting views, i. e., for preferring bimodality of loadings vs. a gradient of factor effects. Empirical evidence from a variety of situations is required to resolve this issue. A study of many simple structure matrices is not likely to yield such empirical in- formation because the simple structure constellation deliberately rotates (and possibly distorts ) the relationships between factors and variables into bimodality. The senior author has had serious misgivings about the validity of simple structure since he first became interested in factor analysis. As may be evident from the above discussion he still harbors grave doubts about the method. However, actual results in a series of different researches have demonstrated that simple structure has invariably given a more meaningful interpretation to the data than did either a straight centroid extraction or even a principal com- ponent analysis. While the simple structure solution in the present study yielded several factors which were difficult to interpret, in- spection of the unrotated centroid and principal axes factor matrices showed these to be considerably less clear than the rotated con- stellation. Until further insights and work are brought to bear upon this problem it will tend to remain in its present state of uncertainty. (c) The stability of the factor solution Brief mention should be made of the stability of the factor solu- tion, when the separate 19x19 and 6x6 matrices were analyzed jointly in the 25 x 25 matrix. It has already been pointed out (and the reader can again confirm it by referring to Tables 3, 4, and 5) that the loadings obtained by a separate analysis of these matrices are very similar. We are thus able to recognize easily in the 25 x 25 matrix the factors which were isolated separately in the 19 x 19 and 6x6 matrices. This situation recalls factorial invariances but cannot properly be so called since we are not dealing with an independent selection of variables. We can, however, demonstrate that the shifts in the center of gravity of a matrix resulting from the addition of the new variables has not been such as to result in a different simple structure solution. Factor ANALYTICAL PROCEDURES 1119 (d) The inclusion of causal or independent variables in the original matrix One of the functions of factor analysis in the exploratory stage is to delineate the nexus of cause and effect variables and to attempt to arrange these in order of causation. This may be quite difficult as Cattell (1952) has pointed out, in that many social and biological cause-and-effect relationships are really circular interactions. At the biological level which we are studying and in the particular type of data analyzed in this paper circular interaction is not likely to have taken place, at least not between the physical and the biologi- cal variables. The temperature of the room is not likely to have been materially altered by the activities of the insects therein, although it might have been affected by the activities of the investigators. Similarly, light and relative humidity were probably unaffected by the responses of the insects. The other “causal” variables were evidently quite independent of the biological response variables. Cattell (1952) supposed that among the variables loaded by a given factor the one that is loaded the highest is likely to be the cause or factor itself. While Cattell does not expostulate on this point it seems to us that only in cases where the loading between the factor and a given variable is very high is it likely this variable is the cause of some of the other correlated variables affected by the same factor. In any case the inclusion of possibly causal vari- ables, as in our study, does provide the opportunity of reifying some of the factors. When we examine our data we find that two of the factors, namely factor I (the solar energy factor) and factor VI (the time series factor ) have high loadings on variables which we had thought likely to be common causes in our study. Solar energy has a load- ing of .98 on temperature which makes the two practically synony- mous. However, an interesting philosophical point requires discussion. While temperature is the common measure for solar en- ergy the latter cannot strictly be identified with temperature. Tem- perature as measured by us and by most investigators refers only to a state of molecular excitation exhibited by a column of mercury in a special device (a thermometer ) or pertains to a property of a physical element of some other type of temperature-measuring ap- paratus, such as change of resistance in a thermocouple. The amount of solar energy received by a given surface would be diffi- cult to measure and would vary with the type of surface and other environmental conditions. However, the biologically significant 1120 THe UNIVERSITY SCIENCE BULLETIN temperature concerns only the amount of heat received by a given organism, or better still, by certain specific portions of a given organism performing a specified metabolic task. In studying the effect of temperature on the heartbeat of the roach, for instance, we are concerned with the energy level of those portions of the insect'’s anatomy which participate in and affect the heartbeat. Thus the temperature of the room is only indirectly causal to the energy level at which the insect performs its function. This rather obvious point has been made in great detail in order to show that certain variables which commonly are thought of as causal might not in the strict sense be so considered. Thus we might have an elusive, some- what abstract, factor called solar energy which affects a number of these variables such as temperature, roach heartbeat, louse gut- beat and others without any one variable being causal to any other sensu strictu. On the other hand, we may wish to use the terms “cause” and “effect” more loosely, partly with a view to an economy of language, and may still use the criterion of the largest loading to identify causal factors. We might then state that temperature is the cause which affects the variables loaded by factor I. In a simpler sense, and probably more justifiably so, days (which refer to the passage of time) are the cause of the time series factor (VI). It is interesting to speculate whether high loading variables on the other four biological or situation specific factors are also causal. In the case of factor III the answer is surely in the affirmative. Moth undulation appears to be nothing but the basic locomotion of the. moth, which is not completely correlated with the rate of locomotion because some of the undulation movements are not very productive by way of moving the larva along. In IV and V, where wasp excursion index and louse rate, respectively, are the highest loading variable, the interpretation becomes difficult. It is our firm opinion that had there been more causal variables in the study, these would have shown up in the same variable clusters with the biological variables dependent on them. Thus the inclusion of possible causal variables is to be highly recommended in explora- tory studies. Postscript: During proofreading it was brought to our attention that Habrobracon juglandis (Ashmead) as well as Ephestia kiihniella (Zeller) have both been renamed Bracon hebetor Say and Anagasta kiihniella (Zeller). We have not changed names in this paper since, it would first of all have required much re-editing of the manuscript and since we have doubts on principle about the renaming of organisms well known to the general biological literature. - Facror ANALYTICAL PROCEDURES IH DAL LITERATURE CITED Bartow, J. A., and C. Burr 1954. The identification of factors from different experiments. Brit. J. Stat. Psychol. vol. 7, pp. 52-53. Carrenn, hh.) B: 1952. Factor Analysis. Harper & Brothers, New York, pp. xiii + 462. 1958. Extracting the correct number of factors in factor analysis. Educ. Psychol. Measmt. vol. 18, pp. 791-838. Kaiser, H. F. 1958. The Varimax criterion for analytic rotation in factor analysis. Psy- chometrika, vol. 23, pp. 187-200. McOurmeny,. Ii.) 1: 1957. Elementary linkage analysis for isolating orthogonal and oblique types and typal relevancies. Educ. Psychol. Measmt. vol. 17, pp. 207-229. Neunaus, J. O., and C. WRIGLEY 1954. The quartimax method: an analytical approach to orthogonal simple structure. Brit. J. Stat. Psychol. vol. 7, pp. 88-92. SAUNDERS, D. R. 1953. An analytical method for rotation to orthogonal simple structure. Educational Testing Service Bull., RB-53-10 (Multilithed), (also Amer. Psychologist, vol. 8, p. 428 ( Abstract) ). Soka, R. R. 1958. Thurstone’s analytical method for simple structure and a mass modification thereof. Psychometrika, vol. 23, pp. 237-257. 1959. A comparison of five tests for completeness of factor extraction. Trans. Kansas Acad. Sci, vol. 62, pp. 141-152. Sowa, Rh. RB: and EH: V. Dary 1959. An application of factor analysis to insect behavior. Univ. Kans. Sci. Bull., vol. 42, pp. 1067-1097. Soka, R. R., and C. D. MicHENER 1958. A statistical method for evaluating systematic relationships. Univ. Kansas Sci. Bull., vol. 38, pp. 1409-1438. TuuRSTONE, L. L. 1947. Multiple factor analysis. Univ. Chicago Press, Chicago, pp. xiv + DS. 1 * re / = es ' ha ¥ A] ‘ J Fa " Te ve ‘ = 45 a] wren Ad - e ° “yw (A ee Lobe | PARA) Sie) q i D4 weet : i. WEA j a ia ae A j t UTE ie bog r ~ We kee , ' ie ‘ | a ‘ oh nr. : a i wa , ‘ 7 ’ Bie! 2 = | : : edie 1 = ; wate, ot * s 7 ‘ ‘ pes 1 Se 4 *y ‘ , * 1 A : re ar Wn "a A: yo. she ‘ me iS Tiree we @ *» An ooh Gel Te: Aa ip ‘fai Al, THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vou. XLIT] DECEMBER 29, 1961 [No. 11 The Bionomics of a Primitively Social Bee, Lasioglossum inconspicuum' BY CHARLES D. MICHENER and ALVARO WILLE ” Apstract: Lasioglossum inconspicuum is a common small halictid bee in eastern North America. It nests solitarily or in loose aggregations. The nests are in the ground, in sunny areas of bare soil. The burrows are deep, more or less vertical, with subhorizontal cells along them (Figures 5 to 16). This species visits a wide variety of flowers (Table I) but largely ignores the extensive flora of yellow autumnal Compositae. Afternoon activity on flowers is markedly less than that in the morning, and a much smaller percentage of the bees that are afield collect pollen in the afternoon than in the morning. A brief midday drop in activity is indicated by the observations (Figure 1). Overwintering occurs as fecundated queens in the nests used the previous season; these queens begin spring activities in late March or April. Most of them establish new nests as lone individuals but some remain in the overwinter- ing nests. Sometimes as many as six queens or potential queens may jointly occupy an old nest. There is evidence that in such cases some of them some- times have a workerlike rather than egg-laying function. The mortality of colonies is very high throughout the season (Figure 4) but the reasons for the high mortality are not clear. The queen nests, whose cells are provisioned by queens in the spring, become closed in May and are reopened by emerging workers at the end of May or in early June. The nests are then enlarged, a process which goes on (if the colony survives) through the summer. The queen lays more eggs as cells are com- pleted and provisioned during the summer. During this time the queens do not leave the nests with the frequency of workers, but possibly do so less fre- quently to feed. They do not collect pollen. Their mandibles, however, be- come progressively more worn, indicating that they work in the nest. Some queens die during the summer and are seemingly replaced by young females. Such replacement queens do not survive the following winter. Workers, in contrast to queens, are quite short lived, surviving as adults for perhaps three weeks. Some workers have one or even two ovarioles enlarged and presumably lay a few eggs (Figure 28). Such egg-laying workers collect 1. Contribution number 1088 from the Department of Entomology, The University of Kansas, Lawrence. 2. Present address: Universidad de Costa Rica, San José, Costa Rica. (1123) 1124 Tue University SCIENCE BULLETIN pollen freely and seem to act otherwise like typical workers. Young workers act as guards and become lost if removed from the nest. Older workers do the foraging except in spring when the queens are establishing their colonies. The number of foraging trips per unit time varies greatly (Table II), as does the duration of such trips (Figure 3). Males are absent in the first broods produced by overwintered queens but appear in small numbers in June, much larger numbers in August and early September (Figure 22). The average number of females in a colony varies seasonally as shown in Figure 21; there is very wide individual variation in colony size as shown in Table VII. Workers average smaller than queens (Figure 31), but in some nests the largest worker is larger than the queen. Egg-laying workers average larger than those with slender ovaries, but smaller than queens. The mean size ot workers in July is smaller than in June or August. A similar seasonal fluctuation in size of males was noted (Figure 23). There is no evidence of communication among individuals of a colony. The return of a forager does not seem to constitute a stimulus for other foragers to leave the nest. The guards seem quite unable to distinguish individuals of their own colony from those from other colonies. Young workers that have not yet been afield and learned the location of their own nest can be transferred to another nest where they soon function as guards and live out their lives, their alien origin apparently not recognized. Workers that are older, if transferred to another nest, soon leave it and return to their own nest. CONTENTS PAGE INTRODUCTION hit e os fe nb sage no eee he Ce ee 1125 GENERAL ACCOUNT OF LASIOGLOSSUM INCONSPICUUM .................. 1127 DIstnibWONea ce sales acs oe Sede oe ce wee 2 OO ee 1127 [Digs OBO) sy nh a MEETS i 1128 GEASOHAIIGVCIE Se os ck rhe te tee eee 1128 INaturalEmemies 6 6... see nc Peas et eee ee 1129 Activimy, OUTSIDETHE NESTS) ; . 22... 22!) | hs nscale eee 1130 Conditious sor Outside Activity <)" .. 7), eee 1130 Hours ot Activity ....°.... 020 S00 ee ee eee 1130 Flowers Visited EE I Mee ie ee 1132 Methods OF POrdging: 0. oe: OS eee ee eee 1133 Activity Aroung the*Nests .. 0:5. 55.) eo oe eee 1134 Duration of Trips and Periods Between Them ............. AP re ‘1137 Probable lack of Social Facilitation .....:->.....00 “eee 1140 CAGE OLIGO 5 elt ss le ee oe a eee 1143 NESTS seem eer artes te CAR EY ee A 11438 Pisisioution-of, Nests. 5. oy CO I ee eee 1143 Commion Features of Nesting Areas ... ....’. >. °). 9 ee 1147 Differences in Activities Among Nesting Areas ..................... 1148 Survival of Nesting Aggregations ». 0)... i See 1149 BRIM ALCCEMNOSES |. (5s. 2 «1 ius euciicahoed hs ange carats RABE N Oed ee ee 1150 Nest Structure . Se eyplistie sg & bd gitar et Spy cnn ty. eee er 1151 Nest Construction | . wlth sah biete oe ae oe Gt aS ee 1161 Provisions arid Immature Stages .............. 0-04 1163 BIONOMICS OF A PRIMITIVELY SOCIAL BEE LDS PAGE IINDIVIDWAT- SAND: SOCTAT HDEHAVIOR UNE miei ek te cies as cca cles oe SAE 1165 SOT ee IONS HOle NESESI A ci) Pinte ee ten de Ad Fisiwiaie Ses on gduis Canoe 1165 NESE GMULALIONS Stee eet Me PIE Cin foe tds chordie Coes eee 1166 IME cael rs ba ie DR he coat a Te ca} ook Na Ae 1167 Variation in Size Among Field-collected Females ................... 1171 GASteHIDIFETeENCES) He et iee era tea te epee ye SA ee ea ae Ey 1171 SeasonalaDifterences: Inu Size aes Sree ine hes eta Gees 1179 (ONMERTG 15's does 2 5 Grole wong cae Oona eeaa clea het nc OED CAE ener a oe 1182 UOT KCESmmMrN Pyar cea nyt tim. SAN eM Tecate Sommer ue ed atte Me, 1187 BehavioLormOueens: 1s) e eee eee a en nk, ae! as ne 1190 Division of Labor Among Queens in Polygynous Nests .............. 1192 BetantOrzOlaV OLkersauue trex er eek in A PT ee eh eT Akad 1193 Guarding and Other Activities at the Nest Entrance ........ ee eer sul Od: (o-operative: Activity Among Workers’ 4.0.55 2.05 4 26 6 en te eee 1198 INEEDBEORSMURDHERUS LUDY.) 1 ance een es ee easy Se ie ee 1199 JELPERATURE) GIVE eget ar ees eee Sete tee, oe se eA area | 1201 INTRODUCTION This paper provides an account of the life cycle and behavior of one of the primitively social halictine bees, Lasioglossum (Chlora- lictus) inconspicuum (Smith). This study is part of an investigation of comparative halictine behavior intended to shed light on the origin and evolution of social behavior and castes. Observations were made over a period of eight years in the vicinity of Lawrence, Kansas. During several of those years efforts were made to follow the life cycle as completely as possible; during other years our efforts were directed toward elucidation of certain details of the life cycle or behavior. The methods used are largely those de- scribed by Linsley, MacSwain and Smith (1952) and by Michener, Cross, Daly, Rettenmeyer and Wille (1955), although many of them are further explained in appropriate places below. Over 200 nests were excavated and 929 females were dissected to determine the condition of the ovaries, spermatheca, and crop, examined to determine mandibular wear and measured to determine size. Hundreds of other observations on living marked and un- marked bees were made, so that for example, we have about 125 records of the duration of pollen collecting flights of individual fe- males, Probably a majority of the species of the enormous subfamily Halictinae are social to a greater or lesser degree. The several bees inhabiting a single nest are said to constitute a colony; groups of nests placed close to another another constitute an aggregation. Since social behavior evidently arose repeatedly and independently 1126 Tue University SCIENCE BULLETIN in various evolutionary lines within the subfamily, this group offers unparalleled opportunities to investigate incipient and primitive societies. A synthesis of the available information on_halictine behavior and evolution will be presented later. It will suffice here to give a brief account of previous works on the biology of the sub- genus Chloralictus!. L. opacum (Moure) from Brazil appears to be a solitary bee without strong aggregative tendencies and without a worker caste (Michener and Lange, 1958). Presumably males and females appear in roughly equal numbers throughout the season. In L. rhytidophorum (Moure) (see Michener and Lange, 1958) and probably in L. seabrai (Moure) and guaruvae (Moure) (see Michener and Seabra, 1959), small colonies exist in which there are weakly differentiated queens and workers. At least in rhytido- phorum the queens, of which there are usually two or more per nest, are rather frequently replaced and males are produced throughout the reproductive season, although in smaller numbers in summer than in spring and fall. In L. inconspicuum the colonies are larger, there is more often only one queen in a nest, more work- ers are produced per queen, workers and queens are a little better differentiated, and queens probably often survive for most of a year. Males are not or scarcely produced early in the reproductive season (spring and early summer), appear only in moderate num- bers in midsummer, but become abundant in the autumn. The nest architecture is known for more species of Chloralictus. In addition to those listed above it is known for L. pruinosum (Rob- ertson) (see Melander and Brues, 1903), rohweri (Ellis) (see Sakagami and Michener, in press), smilacinae (Robertson) (see Brittain, 1933), versatum (Robertson) (see Sakagami and Michener, in press ) zephyrum (Smith) (see Sakagami and Michener, in press ) and for perhaps half a dozen unidentified species. In all of these the nests are similar, each consisting of a branching main burrow from which diverge subhorizontal, bilaterally symmetrical, wax lined cells connected to the main burrow without or by very short lateral burrows. Only in L. aricense (Schrottky) (= Halictus gla- briventris Friese) and herbstiellum (Friese) (see Claude-Joseph, 1926) from Chile are the basic features of the nest architecture dif- ferent from that of the above species. Possibly these forms are not properly included in Chloralictus, or perhaps they were misidenti- fied by Claude-Joseph. A small amount of information on nests and behavior of L. in- _ 1, Australian species included in Chloralictus by Rayment in various publications treat- ing of biology are in reality not related to Chloralictus. BIONOMICS OF A PRIMITIVELY SOCIAL BEE ub) Beye conspicuum is found in the papers by Michener, Cross, Daly, Ret- tenmeyer and Wille (1955), Michener (1958), and Michener and Lange (1958). In these publications the synonymous name, stultwm Cresson, is used for this species. The meager conclusions on this species presented in those papers appear in greater detail in the present work, We wish to acknowledge grants from the National Science Foun- dation which made the study possible. We wish also to thank several persons who helped with the field observations. This has been no light task, as those who have spent long hours on the ground watching or digging bees’ nests in the hot summer sun can testify. The persons concerned are Earle A. Cross (now of North- western State College, Natchitoches, Louisiana), Howell V. Daly (now of the University of California, Berkeley, California), Wal- lace E. La Berge (now of the University of Nebraska, Lincoln, Ne- braska), Ellen Ordway of the University of Kansas, Carl W. Ret- tenmeyer (now of Kansas State University, Manhattan, Kansas), and Alvin F. Shinn (now of Stephen F. Austin State College, Nacog- doches, Texas). Dr. David S. Simonett of the Department of Geography, The University of Kansas, kindly visited the main nesting sites and provided the data on soils. The statistical analyses were done by Miss Ellen Ordway with guidance from Gunther Schlager and F. James Rohlf. GENERAL ACCOUNT OF LASIOGLOSSUM INCONSPICUUM Distribution: This minute greenish black bee, nearly 4 to nearly 5 mm. in length, is widespread over eastern North America, occurring from Quebec to Georgia, westward to Wisconsin, New Mexico, and Texas (Michener, 1951). Over much of this area it is one of the commonest native bees, although so inconspicuous that it is often unnoticed even by entomologists. The species also occurs in the vicinity of Riverside, California, where it was presumably intro- duced from the eastern or central part of the United States. In eastern Kansas, where our observations were made, the spe- cies is generally distributed. It is sometimes very common and can be taken at least occasionally on almost any clump of suitable flow- ers. In this area trees and bushes cover most of the noncultivated terrain and the average rainfall is relatively high (35 inches an- nually at Lawrence). In central Kansas (area of Caldwell, Hutch- inson, and Salina), however, the rainfall is less and uncultivated land is largely treeless except in stream valleys. In this area L. in- 1128 THe UNIvEerRSITY SCIENCE BULLETIN conspicuum is largely confined to stream valleys although other species of its subgenus are common in the drier uplands. Although Lasioglossum inconspicuum is found in an area most of which was once deciduous forest, it is not a bee of forests them- selves. It usually nests in exposed bare soil, as will be discussed in detail in the section on “Distribution of Nests,” and we have never found it burrowing in the forest floor. Originally the forest mar- gins and perhaps occasional bare streamside areas must have been its chief habitats. Destruction of forests, overgrazing of pastures, erosion of soil and development of brushy or weedy wastelands must have enabled the species to become much more abundant and possibly more widespread than in primeval times. Life History: A brief account of the life history of L. inconspic- uum is given here to provide background for subsequent sections of the paper, where most of the matters mentioned will be treated in greater detail. The bee is social to the extent that its nests, which are burrows in the soil, are usually occupied by several females. One or more of these is a queen, the others workers, but these castes are very similar and overlap broadly in size. Without information as to season, young females can be placed as to caste only if unus- ually small (workers) or large (queens). The caste of older indi- viduals (with worn mandibles) can be determined more readily when ovarian development and spermathecal contents are con- sidered. The species overwinters as adult fertilized queens in the nests. In the spring these queens provision cells and lay eggs, either in new nests or in the old ones. The cells are mass provisioned and usually closed after the eggs are laid, as in most other halictines. After preparing about five cells, the queen ceases her activities until these progeny reach maturity. All of them are workers. They deepen the nest and provision other cells in each of which the queen lays an egg. From this time on the nest is usually continually ac- tive, there being no separated broods as in the European L. mala- churum. The short-lived workers are replaced as they die so that several of them are ordinarily present in the nest. The queens probably are sometimes replaced during the summer, and some males are produced during that season. In fall queens and males are produced in the nests, and it is these young queens which pass the winter. Seasonal Cycle’: Overwintered queens were first seen in the 1. Numbers in parentheses in this section indicate the number of years when appro- priate observations were made. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1129 spring at Lawrence, Kansas, on various dates from March 25 to April 16(6). Three of the six early spring records are in the first week of April. Each of the early records is based on an observa- tion of bees flying about or alighting on the ground, but without any evidence of excavation of cells or other work in the nest. Tumuli (piles of excavated earth) at the nest entrances or regular going in and out of well rounded nest entrances were first noted on dates ranging from April 3 to 26(6). The first queen seen carrying a pollen load into the nest (i. e., the first evidence of provisioning of cells) was observed on dates ranging from April 13 to May 1(7). The first bees recognized as workers, either by dissection or by augmentation of nest populations, were noted on dates ranging from May 31 to June 10(5). Males were first seen, always in small numbers (usually only one or two in hours or days of observation ), on dates ranging from June | to June 18(5). Because males might be produced as a sort of abnormality in nests that had lost their queens, and because the records in the first half of June may repre- sent the parasitic species, Lasioglossum cephalotes, it seems worth noting that our earliest excavations of male pupae from nests which also contained female pupae were from June 20 to July 8(3). Since we dug relatively few nests, about 40 during June, compared to the total numbers of nests in the areas studied, it is not surprising that males from one or another nest would be seen in flight before any male pupae were found in nests. We therefore cannot be sure of the significance of the late dates for the male pupae. We did dig one nest containing male but no female pupae on June 17. The last females, presumably workers, were seen carrying pollen on dates ranging from August 28 to September 13(4) and very few cells are provisioned as late as these dates. The last overwintering queens were seen in flight from October 1 to 22(4); the last males from September 26 to October 12(3). Overwintering queens were seen guarding nest entrances as late as October 10 to 22(3). Natural Enemies: Lasioglossum (Chloralictus) inconspicuum is frequently parasitized by L. (Paralictus) cephalotes (Dalla Torre). A later paper will concern the biology of this parasitic species and its relations to its various hosts, all of which are in the subgenus Chloralictus. A small mutillid, Pseudomethoca frigida frigida (Smith) (det. K. V. Krombein) is common in and about the nests of L. incon- spicuum. It has not actually been reared from cells of this bee. [We have also taken it in the nests of Lasioglossum zephyrum 1130 Tue Unrversiry SCIENCE BULLETIN (Smith), rohweri (Ellis), and versatum (Robertson) and in nests of Augochlorella striata (Provancher ).] We have reared males of a minute tiphiid, Myromosula parvula (Fox) (det. K. V. Krombein), from their cocoons in cells of L. inconspicuum and have captured a few females in the nests of the Lasioglossum. Female (but not male) Strepsiptera (Halictoxenos) are occasion- ally found in abdomens of L. inconspicuum, as are larvae which appear to belong to the Conopidae (Diptera). Except for the forms listed above, natural enemies in nest of L. inconspicuum are not noticeable. Some cells are destroyed by mold, but there is always a question as to whether this is a primary enemy or merely attacks already dead eggs or larvae, thence spreading onto the pollen mass. Outside the nests, of course the adults are subject to the usual predators. Activity OuTsmE THE NESTS Conditions for Outside Activity: Detailed data on temperature, light, wind, and other factors associated with flights from the nest have not been obtained, as other data seemed more important. However, the meager information obtained is perhaps worth re- cording. Except as otherwise indicated, temperatures were taken in the shade of the observer’s body 30 cm. above the ground. At 50° F. no bees were seen. At 62° F. guards were seen in some nests, but no flights were made. Pollen-collecting flights were observed at temperatures ranging from 70° to 105° F. Of course temperatures at the soil surface are much higher in direct sun light, and as will be explained in more detail in the section on “Guards,” the guards withdraw from unshaded nest entrances when the temperature ob- tained by laying a thermometer on the ground surface in direct sun light reaches about 125° F. Of course strong wind usually results in reduced activity, but pollen collecting flights sometimes occur when a warm wind is blow- ing so strongly that the bees are rolled over and over on the ground if they alight in approaching the nest. At such times bees may cling to weeds and leaves, and several may then enter a nest in quick suc- cession when there is a lull in the wind. Hours of Activity: The time of day when activity begins varies with the season. On various clear days in June, the first flights from nests in the morning were observed from 6:40 a. m. to 9:00 a. m.; on BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1131 cloudy days the first flights were later. In the afternoon, a bee was seen to leave the nest as late as 4:55 p. m. It is a common observation of persons collecting bees about flowers that, except on the hottest days, collecting is best in late morning hours. As a biproduct of observations of the activities of marked bees, we obtained numerical verification of this collector’s impression. Figure 1 shows the number of worker bees entering at different times of the day into nests that were under observations. The maximum was reached during the period from 8:30 to 11:30. Our data were largely gathered during hot summer days; this may explain the relatively early maximum. Afternoon activity was about half of that in the morning. Separate lines for bees returning to the ENTRIES PER HALF HOUR vt, Pollen wine Ge if 8- 9- 1o- ue i2- \- 2 3- 4- 6:30 7:30 8:30 9:30 10:30 1:30 12:30 1:30 2:30 3:30 4:30 HALF HOUR PERIODS FicurE 1. Graph showing the average number of times per half hour period that females (probably all workers) entered their nests after trips afield. All observations were made during the months of June, July, and August. The figures across the top of the graph show the number of half-hour periods during which nests were under observation; irregularities in the early morning are doubtless due to the small number of observations. The solid line indicates the total average number of entries. The nests studied were usually the most active ones available, so that the figures are strongly biased upward. The broken line shows the average number of entries of bees with pollen loads on the legs. The dotted line shows the average number of entries of bees without pollen loads. The total is sometimes higher than the sum of the other two, since bees some- times got into their nests without our determining whether or not they were carrying pollen. nest with and without pollen loads on the scopa show pollen col- lecting to be the dominant outside activity in mornings but flights presumably for nectar gathering or probably merely for feeding of adults increase through the day (except for a midday lull) and are the dominant outside activity in late afternoon. We know from observations of bees marked for individual recognition that at least 1132 Tue UNrversiry SCIENCE BULLETIN some of the afternoon flights from which bees returned with no pollen loads were made by individuals which had collected pollen earlier in the day and did so also the next day. There is no reason to believe that the reduced afternoon activity is due to poorer weather in the afternoon. The observations were made during 50 different days scattered through three summers. The reduced activity may be related to reduced availability of pollen because it is utilized by bees during the day, but some individuals were able to obtain pollen loads in short periods of time in the afternoon and the average duration of pollen-collecting trips was not longer in the afternoons than in the mornings. It seems likely that an innate feature common to the activity of most bees is the tendency, in fine weather, to do more foraging in the mornings than in the afternoon. Scattered data on other species suggests that the afternoon increase in trips for purposes other than collecting pollen is also a common tendency at least among halictines. Flowers Visited: Like most halictines, Lasioglossum inconspicuum is polylectic, that is, it visits and even collects pollen from a wide variety of flowers. Polylecty is a usual characteristic of social bees; since such bees are active through a long season, no one flower is ordinarily available to them at all seasons. No serious effort was made to obtain a comprehensive list of kinds of flowers visited by L. inconspicuum, since it was felt that almost any flower visited by any bees in our area would be visited at least occasionally by this species. Table I was based upon data on specimens in the Snow Entomological Museum as well as on observations made in the course of the study of this species. Except as otherwise indicated, females or both sexes were found on flowers in the list. Examination of Table I shows that the large yellow-flowered Compositae (e. g., Helianthus, Silphium) so conspicuous in late summer in Kansas are little visited by L. inconspicuum, although many other bees, including other species of the subgenus Chlora- lictus, regularly visit these flowers. The selection of flowers is similar to, although more catholic than, that of Megachile brevis Cresson (see Michener, 1953) which also uses primarily blue to whitish flowers. L. inconspicuum sometimes collects pollen from yellow flowers of Taraxacum officinale in the early spring, possibly because there is sometimes a shortage of other flowers at that season. It is apparent that even in a species of seemingly thoroughly polylectic bee, by no means all of the potential pollen and nectar sources are used. Expressed in different terms, the bees seem to BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1133 TABLE I.—List of Flowers on which Lasioglossum inconspicuum Has Been Taken in Eastern Kansas. P indicates that females were collecting pollen, S that they were sucking nectar but not collecting pollen. We have no notes concerning activities of bees on flowers not marked either P or S. Wilincéae Malvaceae Allium sp. Callirhoe digitata Gramineae Lythraceae Unidentified P Lythrum alatum P Cruciferae Umbelliferae ; sf 9 Sisymbrium sp.! Dn Pen Roseceae ; 2 Asclepiadaceae Pa a gaara : Asclepias tuberosa S Fragaria virginiana Labiatae Malus pumila Monarda fistulosa IP Leguminosae Plantaginaceae Petalostemon candidum Plantago rugeli P Psoralea floribunda (o) Compositae Melilotus officinalis . S Taraxacum officinale 12 Melilotus alba Ji ai = te : Medibago sativa P Silphium speciosum S aes ae abe Pp Silphium perfoliatum Solidago sp. S Amorpha canescens P : Amorpha fruticosa a Se kes e Vernonia baldwinii 12 Anacardiaceae Erigeron philadelphicus — (o) Rhus aromatica Rudbeckia hirta Rhus copallina P Echinacea pallida Rhus glabra Ratibida pinnata have preferences for certain sorts of flowers. Such preferences doubtless occur among most polylectic bees and are responsible for the observation that at a given time different polylectic species often use different flowers in the same area. Such observations are very easily made with various species of Trigona and Melipona (see Michener, 1946). Methods of Foraging: Methods of gathering pollen vary accord- ing to the flower being utilized as a food source. When working about over the masses of small flowers of Rhus copallina, the females can be rather easily seen to brush the anthers with the fore tarsi, to transfer the pollen to the middle legs, and thence to the scopa on the rear legs and probably to the hairs of the abdominal sterna. There is no regular timing of the movements, nor is there regular alternation of the movements; that is, sometimes pollen is trans- 1134 THE UNIvERSITY SCIENCE BULLETIN ferred backward twice in succession on one side of the body before the same action occurs on the other side. It seems that basically similar actions are used for collecting pollen on other flowers, but the modifications resulting from dif- ferent flower structure are great. For example, to gather pollen from the flowers of Plantago rugelii, the bee alights on the long stamens, and as she hangs downward she clasps several of them together with the hind legs, then brushes pollen off of the anthers with the forelegs, transferring it backward to the scopa by the process already described. The bee flies from place to place on the flower mass; walking among the stamens looks to be impractical. It appears that pollen gathering must vary greatly in efficiency on different flowers. | Observations of pollen collecting bees in areas of mixed flowers showed that on any given trip a bee usually gathered pollen from a single kind of flower instead of going to various kinds and ob- taining mixed pollen loads. Observations of marked bees going in and out of a single nest often showed that they were collecting from different flowers. For example, one might bring in loads of yellow pollen while another was bringing white. Moreover, a single bee might change the type of pollen it was gathering, so that after a yellow load, she might bring a white one. Sometimes bees collecting pollen may suck nectar from another flower, then go on collecting pollen from the first. For example, a bee was seen collecting pollen of Plantago. She then flew to Melilotus alba, sucked some nectar, and returned to collecting pollen on Plantago. Melilotus alba is sometimes used as a pollen source, although perhaps not when Plantago pollen is also available. Sometimes bees are seen to return to the nest with only a light dusting of pollen on the scopa. Presumably such bees were gather- ing nectar or feeding and in brushing among anthers, some pollen adhered to their hairs. What becomes of such pollen is quite un- known. In late fall, after most of the flowers have disappeared, over- wintering queens sometimes feed on honeydew. Details of this will be given in the section “Behavior of Queens.” Activity Around the Nests: Most of the activity of female bees around their nests has to do with coming and going of the bees. Bees leaving the nests often fly directly and promptly away from the nest entrance. Sometimes, especially in cool weather or in the early morning, bees come out of the hole very slowly, turning the a — BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1185 head jerkily from side to side, often lingering half exposed for some seconds, and may crawl as much as a centimeter from the hole be- fore flying. In flight, the departing bee may dart away with very little “orientation flight” over the nesting area. This is particularly characteristic of departures occurring during the heat of summer days. On cooler days and in early morning hours, it is more com- mon to see irregular zig zag or sometimes spiral “orientation flights” as bees leave. Such a flight requires up to 15 seconds, and during it the bee rises to a height of some four centimeters above the ground surface, flying about over an area of up to 30 cm. in diameter, al- though usually smaller, before making off. Figure 2 illustrates some flight patterns of bees leaving nests. 30 cm ee | Ficure 2. Flight patterns of females leaving nests. The lower one repre- sents an unusually extensive “orientation flight.” 1136 Tue University SCIENCE BULLETIN The more elaborate and prolonged forms of departure mentioned in the preceding paragraph are characteristic, often in exaggerated form, of young adults. We noted this on several occasions, when activities of marked pollen collectors were being observed day after day. As unmarked bees began making trips from such a nest, they almost always made more elaborate departures than the pre- sumably older, marked, foraging bees leaving the nest at the same time. A few days later the unmarked bees were seen collecting pollen and making the usual quick departures. Bees accustomed to the nest usually fly almost directly to it, zig zagging only slightly in the last few centimeters of the approach to the nest. They alight almost at the threshold of the hole and quickly crawl in. Queens going into their nests in spring, and young females at other seasons, often seem to have difficulty find- ing their nests. They may alight on the ground near the nest, and after walking about, even starting to dig, they fly again, often re- peating the performance for 15 or 30 minutes until they ultimately find the nest entrance. It seems apparent that such bees have little difficulty in getting within about half a meter of the nest, but the exact location of the nest seems often to be discovered by chance. Of course the number of such disoriented bees is greatly increased by any disturbance of the soil surface, such as trampling by cattle or persons, which may occur while bees are afield. Occasionally disoriented bees enter the wrong nest, but usually soon leave again. The probability exists, however, that they may sometimes become incorporated into the populations of other nests as suggested by the results of the experiments on transferral of young workers from nest to nest, as described in the section on “Behavior of Workers.” Collections of disoriented or lost bees in locations where no sur- face disturbance could account for disorientation revealed not only young bees with unworn mandibles, as would be expected from the preceding paragraphs, but also some bees with much worn or very much worn mandibles. Presumably these are senile bees; the few dissected have all been workers. Like the unworn ones they give every impression of looking for their nests, examining every little hole, sometimes starting to dig, but soon walking or flying on. Such lost bees sometimes go away from the nesting area for half an hour or more, then return and resume their searching. We have marked such bees with paint and found them still searching the next day. We presume that some senile bees lose their ability to BIONOMICS OF A PRIMITIVELY SOCIAL BEE iD Y¢ return to their nests. In two cases which we observed, lost bees were individuals that had been previously marked and their activi- ties observed in certain nests. In both cases the nests were seem- ingly normal, and exhibited normal activity, while the disoriented senile bees searched within a meter or two of their nests. Duration of Trips and Periods Between Them: The data on this subject were obtained by means of bees marked for individual recognition. Most of the data were gathered from marked work- ers, but queens in the spring, when they are provisioning cells, make trips of similar duration and similar variabilty in duration. The in- pollen collecting trips in rest between trips NUMBER OF BEES 2 4 © 8 '9 '2 '4 '6 'a 20 22 24 26 28 20 32 34 35 38 40 42 44 46 48 50 52 54 56 Sa 80 &2 Oe Be Oa 2 4 MINUTES FicurE 3. Histograms showing (above) durations in minutes of trips away from nests, black areas indicating that the bees returned to the nests with a pollen load on the scopa, and (below) durations of periods that bees spent in the nests between pollen-collecting trips. formation on queens has therefore been lumped with that obtained by observing workers, and is presented in Figure 3. Because pollen-collecting bees usually enter and leave their nests rapidly, it was found difficult to sit beside one or two nests and re- cord every entry or departure. Sometimes bees would be missed altogether; more often they would be seen but their color markings could not be recorded with certainty. To slow down the bees en- tering the nests, we often put a straw or grass blade across the nest entrance. The resultant delay enabled us to record the color com- binations of entering bees, but did not help with the departing in- dividuals which would crawl from beneath the obstruction and immediately take wing. Therefore we have over one hundred rec- ords of the period from a time when a bee entered her nest until the next time that she entered her nest. These data support the picture indicated by Figure 3. In two cases the period from one entrance to the next was over 200 minutes. We have records of trips from which bees returned with pollen loads on the legs ranging from 3 to 105 minutes in duration. As shown in Figure 3 the bulk of such trips take from 8 to 40 minutes, 11388 THe UNIvErsSITY SCIENCE BULLETIN with the maximum number of trips requiring about 16 minutes. The great variation in duration of trips may be due to variations in the ease with which pollen is obtained. There is some tendency for a series of successive trips by a single bee to be of about the same duration, as shown by this example (time given in minutes, P signifies return with pollen load): 13P, 24P, 13P, 15P; another example, 8P, 9P, 9P. However, durations of successive trips may vary strikingly, as shown by the following series: 36, 40P, 15P, 11P and 15P, 14P, 8P, 5P. Successive periods spent in the nest between trips often vary considerably, as shown by the following five series (given in min- utes); 4,5: 5: G, 2, ?, 6: 3, 2, 4, 9, 7; 8, 31, 9° 6, G16; 872 Asishowm by Figure 3, the range of variation is far less than for trips afield. In the case of the last series shown above, the 37 minute stay in the nest was due to the bee spending a considerable time as a guard before leaving on the last trip. These data in general show rapid trips and in theory a bee ought to be able to make many trips per day. As shown in Figure 1, however, outside activity of all sorts and pollen collecting trips par- ticularly, diminish in the afternoons. The largest number of pollen collecting trips that we have ever seen a bee make in a day is seven. We often watched nests until activity seemed to have ceased for the day, and we are confident that more than seven trips per day would be unusual. By no means all of the bees in a nest that are apparently able to collect pollen actually do so actively at any one time. This is shown in Table II, which shows the number of pollen collecting trips made by various workers during periods of three hours. As indicated by the zeros, a considerable number of workers made no trips what- ever. Some of these were presumably young adults not yet ready to undertake foraging activities but others had been seen returning with pollen on previous days. It is also evident from Table II that many make only one or two trips during the time that others make five or six. The trips of the bees making few trips take no more time, on the average, than those made by bees making more trips. Rather, the bees making few trips seem to start late, stop early, or both, thus spending much more time in their nests. In connection with the above paragraphs it should be mentioned that the data were collected from some of the most active nests and on days of great activity. The majority were less active, and nests like nest 1, Table Il, which showed no activity in three hours, were common. 1139 BIONOMICS OF A PRIMITIVELY SOCIAL BEE 0 6 Gj 9 8 8 g i 9 L G al 8 Slot ee ae a speo] uarfod 7870], 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 I 0 0 0 0 0 0 0 0 0 I 0 I 0 0 0 0 0 I 0 0 0 0 I I 0 G 0 I 0 I 0 z 0 I I I I z I ¢ (MOpeg PAT}OR 4SwoT 0 G I G i G G I I 6 I F Gj G ‘QAOGB SLOBVIOF DATJOR SOT) 0 9 I € P 6 6 G F € G 4 g Os ell a Fr See ToyIOM Yoo IO} SpBoy uoTjod jo zoquin Ny ¥ F gq g g 9 9 9 L l 1 8 By Eales sh 7 SIOYIOM JO JAQuUINN I y { I Yq 5} J i ra) p p 9 q iaje > |i og 010 6.4 Bosc OOS a.0"ola.o O.0\o-o's qSON ‘ysnsny Jo ‘Ajn{ ‘ounf jo syyuout oy} SurInp “Ure (CO'ZT 0} ‘UL'R 00'6 JO ‘Ul'e (QO: TT 0} “We g “s'a ‘sIMOY SuIUIOLWU 9Yy} SULINp potoyyes oIoM eJep [TY “SUUUNTOD [ROnAVA potd}e| Ul poyeiedeas ore SAep JUSIOPIpP IO S}sou JUOIOHIP Wo ved IO} ‘poyeorpul st ,spvo[ UsTjod jo Joquiny,, Mo[Iq puke s}soddQ ‘simoy se1Y} OF ATsNnONuTUOI PaArasqoO sPA\ yt UsYM Avp ay} UO jsou oY} UI yUaSaId 9q 0} UMOUY (UOIUSOOOI [RNPIAIPUL OF PaYTVUL) SIOYIOM Jo JoquINU 9Y} Po}eOIpUr ST ,STOYIOAA JO JOqUINN,, aysoddg 9 ‘UuoneiInd simof{ se1y J, JO Spoleg Sulnq seeg [enptAtpuy Aq s}soN OUT PotueD Speo'T UsT[og Jo JoqunN—T]] ATAV ‘JSoU oY} OJUT pated ays zeYQ SpRoy Jo JoquuMU 9} “IaxIOM Yove 1140 THe UNrversiry SCIENCE BULLETIN Variability in activity among nests is the rule rather than the ex- ception. For example on July 3, 1952, three nests within 10 cm. of one another at Engle’s Place were observed all day. In one, six different marked bees were seen at the entrance or going in and out, some of them busily carrying pollen. In another nest, two bees were seen, one of which made one trip afield. In the third nest, only a guard was seen, and she never left the nest. Such di- versity is associated with varying nest populations resulting in part from variations in the number and activity of queens. In a popula- tion in which many colonies are dying throughout the year, one must expect to find many nests which seem inactive because of dwindling populations. However, we often also found relatively inactive colonies which later became active, presumably because of emergence of additional workers. Probable Lack of Social Facilitation: Observers watching nests of L. inconspicuum (as well as other colonial halictines) are often im- pressed by the fact that there seems to be a tendency for several individuals to leave a nest or return to it in a short space of time. It sometimes happens that after watching a nest for many minutes and seeing no activity, an observer sees several bees leave within a period of five or ten minutes. To determine whether there was social facilitation so that de- parture of one individual tended to be associated with departure of others, departures as well as arrivals during nineteen morning periods of good weather ranging in length from 60 to 190 minutes were tabulated. The observations were made at various nests and during all the summer months. The periods of observation were divided into ten minute units. There were 179 such units during which bees leaving the nests had been noted and 162 when bees entering nest had been recorded. During about 150 of the periods both arrivals and departures were recorded; during the remainder only departures or only arrivals were noted for one reason or an- other. The results of this study are shown in Table III. It can be seen that for both entering and departing bees, the frequencies of two or more per ten minute period exceed the frequencies expected from the Poisson distribution. The differences from the Poisson distributions are not satistically significant, however, for departures and are scarcely so (.02 < X* < .05) for arrivals. The consistency of the differences for the two sorts of data (arrivals and departures ) suggests that they may not be mere sampling errors. They might result from external factors (a cloud over the sun or the presence BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1141 Tas_e III.—Numbers of Females Entering or Leaving Nests During Ten Minute Periods. Most of those entering carried pollen loads on their legs. Observed and expected frequencies Bees per ten minute period Entering Departing Observed | Expected | Observed | Expected O) 3° .8i a SSS SEAS ce 60 59.60 70 72.78 PES Sa ee ee ee es eee 54 59.60 64 65.50 DON PBS “Ck Nh eh Ae nia a 31 29.79 30 29.48 Bao snips ocd ot AS ee ee ee a 12 9.93 9 8.84 dL suey errestom bile Bech oe en eRe ee ear ar 3 2.48 4 1.99 Ys Gt ayo. Eee Re ae ae En 2 50 1 36 Gere reer Mem oaliy Gariitah Larios decile a ane eet 1 05 MOtAISME cro ty seus Aer NG ZARA ar eaters: LOM sie svete s chek | of the observer) holding up departure or return of foragers for a time until several are ready to leave or return at about the same time. Alternatively, guards may sometimes retard departures until several bees are ready to go. Since the study reported in Table III seemed inconclusive, and since it seemed desirable to relate arrivals at a nest with the depar- tures immediately following them,! another study was made, again based on morning observations made during good weather. Be- cause the units of time used were five instead of ten minutes, in length, it seemed reasonable to include briefer periods of observa- tion in the study; twenty-five periods ranging from 40 to 190 minutes in duration were used. Both arrivals and departures were recorded during these periods. The average number of departures for five minutes was determined for each period of observation (Table IV, column 2). If departures were associated with arrivals, then the average number of departures per five minutes following an ar- rival should exceed the general average number per five minute period. Column 3, Table IV, shows the average number of de- 1. Arrivals of successful forages are related to departures of other foragers in many social insects, such as Apis, Trigona, Melipona, Bombus, Polistes, Vespula, and many ants. 1142 THE UNIVERSITY SCIENCE BULLETIN TABLE IV.—Average Number of Females Leaving Nests in Consecutive Five- minute Periods (Column 2) and in Five-minute Periods Starting with the Return of a Bee. (1) (2) (3) (4) Mean number : Mean number of departures Duration of Observational : : : seraal of departures in five minutes observations per five minutes | following arrivals (minutes) 1 15 0 100 2 wiley 0 90 3 | 20 0 50 4 525 0 40 5 PH 5338) ifs 6 .28 B25 70 7 .29 .67 70 8 | 30 09 185 9 Sl .70 145 10 .oL .83 145 11 .36 .40 140 12 .36 fo 100 13 .36 a22 100 14 .36 67 95 15 .39 .58 | 190 16 A4 .80 | 45 17 49 ste 185 18 .00 1.00 110 19 .58 .50 155 20 .60 86 50 21 .76 .33 75 22 .78 .67 45 23 .85 23 65 24 .88 2.00 40 25 1.13 1.00 75 Mean .46 .56 partures in five minute periods, each of which starts at the moment of arrival of a bee (almost always a pollen-laden forager ). Examination of Table IV, in which the observational periods are arranged in order of increasing mean departure rate, shows that when there is little activity, the few departures seem to have little relation to the arrivals. At low departure rates, assuming no rela- tion between an arrival and the ensuing departure, it is not sur- prising that the departures commonly fall in the long intervals BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1148 between arrivals. When there is more activity in nests, the number of departures per five minutes following arrivals clearly exceeds, in most instances, the mean number of departures per five minutes. At first thought this would seem to indicate some sort of social facilitation in which the return of a forager stimulates the departure of others, acting either through the guards or directly upon other foragers. This is apparently not true, however, for Figure 3 shows that some 28 percent of the periods between trips by a given forager are five minutes or less. Excluding first trips for the day, this per- centage of the departures is in reality associated with arrivals for purposes of the five-minute units used in this study. Therefore it is not surprising that the excess of the mean of Column 3, Table IV over that of Column 2 is 22 percent. In any event this excess clearly does not show an influence of returning foragers on different de- parting ones. Lack of social facilitation of foraging trips is hardly surprising in a primitively social form such as L. inconspicuum with the nests branched and cells dispersed. Length of Flights: In early spring when flowers are scarce, one sometimes finds nests being provisioned a considerable distance from any flowers. We have seen such nests as much as 50 meters from the nearest flowers (Prunus) that we could find. Presumably the bees fly at least that far. During the rest of the active season flowers are so generally available that it was never far to them from the nesting sites that we found. We experimented with carrying marked pollen collect- ing bees away from their nest and releasing them. We assumed that if they found their way back quickly, they were probably familiar with the landmarks at the place where they were liberated. Bees returned promptly from points 75 meters distant from their nest. Unfortunately the experiment was not continued to the point where some of the bees were lost, but it does indicate that in spite of their minute size, these bees fly moderate distances. NESTS Distribution of Nests: Nests have been found in the soil of a wide variety of sunny or partly sunny locations. Undoubtedly many of the nests are isolated. Because of their inconspicuousness, solitary nests are difficult to find. We came upon them occasionally, however, in overgrazed pastures, parkings or partly bare areas on the University campus, along an abandoned road, etc. Those that 1144 THe UNIversiry SCIENCE BULLETIN we studied or excavated seemed quite comparable to nests in ag- gregations, It was convenient for us to make our studies in places where we could find numerous nests in the same vicinity. Fortunately sev- eral such locations were found, after much and often fruitless searching, in the vicinity of Lawrence, Kansas. The accounts of these places below give some idea of habitats in which Lasioglossum inconspicuum nests. The locations are sufficiently variable that we have never been able to learn how to recognize a nesting area from a distance. This is because of the great amount of ground seemingly similar to that of the nesting aggregations, and yet with- out nests or with only widely scattered, isolated ones. Even in the aggregations, the nests were not closely placed. In late spring, when the number of nests is at its maximum, there were occasionally found places having perhaps ten nests in an area of one fourth square meter. These were exceptional local concentrations within larger aggregations. By midsummer or fall most of the nests established in spring are dead, and often there was only about one nest per square meter in the aggregations. The aggregations are loose and usually lack definite margins because with careful search- ing scattered nests can often be found in the surrounding areas. The nine aggregations from which we obtained nearly all of our data on Lasioglossum inconspicuum are briefly described as follows: 1. Potter's Lake. This is on the campus of the University of Kansas. The nests were found along a slightly sloping path which received considerable use. Although a few isolated nests were found from time to time elsewhere along this path, most of them were in an area about 10 meters long and 2 meters wide. Scat- tered plants of bluegrass (Poa) and plantain (Plantago rugelii) grew in and especially along the sides of the path. The nests were often found among these plants as well as in areas of bare soil. Large trees (elm and hackberry) shaded the path except during four or five midday hours. In one area where new nests were es- tablished each spring before leaves grew on the trees, shade con- tinued nearly all day when the trees were fully leaved. In this shady place nests died out in early summer. The mortality of nests elsewhere was so high, however, that one could not be certain of the effect of shading on survival of the nests. The soil was hard and rather crumbly when dry, with considerable amounts of clay. The surface was often dusty due to foot travel along the path. It BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1145 was in a low area, however, and remained moist longer than soil of the hill tops. Among nests of L. inconspicuum at Potter's Lake were numerous nests of Andrena erythronii Robertson (area II of Michener and Rettenmeyer, 1956) and during the earlier years of observation, Calliopsis andreniformis Smith. Other bees nested in the area only occasionally. 2. Cooper's Place. There were two sites grouped together under this name in a little used pasture about nine miles southeast of Lawrence. They were on nearly flat but well drained ground, and were in small areas of short sparse grass where the soil surface was exposed to the sun. Each of the areas was about five meters across. There were no nearby trees or large bushes to provide shade. The main vegetation outside the immediate areas where the nests were found consisted of Symphoricarpus orbiculatus and Rhus copallina. The nests were in surface soil of Dennis Fine Sandy Loam, developed on sandstone. In one of the two areas in the pasture, enormous numbers of nests of Lasioglossum (Chloralictus) rohweri (Ellis) occurred in 1951, but not in subsequent years. A few nests of Augochlorella aurata (Smith) and A, striata (Provancher ) also were found in the same area. In the spring of 1957 the pasture was plowed and sown, destroy- ing it for the time being as a nesting place. 3. Engle’s Place. The nesting area called by this name consisted of some nearly bare areas sparsely inhabited by L. inconspicuum in an overgrazed pasture about eight miles southeast of Lawrence. The ground was gently sloping, well drained, and high enough to be quite dry. The vegetation was sparse; aside from scattered trees of hedge apple (Maclura pomifera) which did not shade the nesting areas, the principal vegetation was weeds such as Vernonia baldwini. The soil was like that of Cooper’s Place. In the spring of 1957 the area was plowed, destroying it for use by L. inconspicuum as long as cultivation is continued. 4. Intersection (Figure 40). This area was on a dry, rocky hill about nine miles southeast of Lawrence. The ground in the nesting area was flat. It was protected from grazing and had never been cultivated, but had been much disturbed because it was on the former site of a schoolhouse. Perhaps because of its rockiness, vegetation was not everywhere dense; there was an area about 1146 THe UNIVERSITY SCIENCE BULLETIN six meters square largely occupied by sparse vegetation so that the soil surface was largely exposed. There was no shading by bushes or trees. In this area the principal plants were tall grasses and Verbena stricta. Outside of the nesting area the same plants, to- gether with Rhus copallina and others, formed dense masses which made the ground there unsuitable for nesting of L. inconspicuum. Because of removal of the surface soil at some time in the past, the nests were in subsoil of Dennis Fine Sandy Loam, developed on sandstone. A few nests of Halictus confusus Smith were found in the same area, 5. State Property. The area called by this name was about two miles southwest of Lawrence, on a gentle slope. Although not cul- tivated for many years, if ever, native vegetation was entirely gone and replaced by weeds. Hedge apples (Maclura pomifera) were nearby trees, but did not shade the nesting area. The nesting area itself was sparsely covered with grass, unlike much of the adjacent ground which was densely weedy or grassy. The soil was Newtonia Silty Clay Loam developed on Oread Limestone. The nests oc- curred in a space of about four square meters and were associated with burrows of Augochlorella striata (Provancher) and A. aurata (Smith), After 1955 no nests of Lasioglossum were found in this area, probably because the grass became dense in the nesting place. 6. Sycamore Slope (Figure 38). This area, on the University of Kansas Natural History Reservation, was in ground cultivated at one time, but since 1946 it has been sown with native prairie grasses and left strictly alone. The nests were on the sides and summits of some small bare ridges resulting from erosion. The nests were found in an area about three meters on a side. There was no shade. Nests of Augochlorella aurata (Smith) and striata (Provancher ) were associated with those of Lasioglossum. Another area in which we found some nests was along the side of the road near the main sycamore slope area. Both of these places are upland areas of severely eroded Pawnee Silty Clay Loam developed on glacial drift. 7. Prairie Road (Figure 39). This was on another part of the University of Kansas Natural History Reservation. An abandoned road (merely a pair of wheel tracks) crosses few acres of upland prairie that has never been cultivated. In one section of this road, about seven meters long, nests were found in considerable numbers BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1147 both in the exposed soil and along the edges of the tracks where they were largely shaded by overhanging prairie plants. Else- where along the same road across similar soil, there were only oc- casional isolated nests. Vegetation consisted of typical prairie plants with some intrusion of Rhus copallina. The soil at this lo- cation was a Newtonia-like Silty Loam developed on limestone. 8. Petefish’s Place. The nests at this location were in a level area about three meters square of mixed sand and clay along a stream five miles southwest of Lawrence. The location was so low that it must have been flooded most years; flooding was shown by the bent grass and drift wood at the time of our work in this area. The surface was bare except for scattered clumps of grass and of sweet clover, Melilotus alba. The area was exposed to the sun most of the day but was shaded in the later parts of the afternoons by elms, hackberries, and other trees. The soil was a sandy loam, quite loose and obviously more sandy than at any other location where this species was studied. 9. County Line: This was an overgrazed upland pasture along the Douglas County-Franklin County line south of Lawrence. There was sparse grass and abundant weeds, such as especially Vernonia baldwini. The nests were widely scattered in small bare areas in the pasture. There was no shade except that afforded by the grass and weeds. Common Features of Nesting Areas: From the descriptions of nesting areas given above it is possible to extract some common features which are presumably necessary for nesting by this species. The soil surface is flat or gently sloping in all cases. It must be exposed; no nests were found where the soil was densely covered by vegetation (as on undisturbed portions of prairies or in densely weedy areas) or by fallen leaves (as on the forest floor). The nesting site is typically exposed to the sun all or most of the day. Nests started in the spring under deciduous trees disappeared after the foliage cast its shade. The soil was in all cases loam, usually surface soil but sometimes subsoil. The soil may be derived from sandstone, limestone, or glacial drift. Nesting may occur in prairie soil disturbed only by removal of vegetation or in areas that were once plowed and have been severely eroded or even in streamside sandy loam deposits. In spite of all this variation in soils used, the material is always compact and often hard; nests never occur in loose sandy soils. Nesting may occur in stony situations where nests must curve around the rocks or penetrate them if not too 1148 THe Unrversiry SCIENCE BULLETIN hard. Nests may be found on hilltops, slopes, or in stream valleys. As explained in the section on “Survival of Nests” below, we think that lack of moisture in soil sometimes results in death of colonies of the bees. Because of destruction of natural forests and prairies, heavy graz- ing, and sheet erosion, very numerous areas provide the require- ments for nestling noted above. Some possible reasons for de- velopment of aggregations will be discussed in the section on “Distribution of Nests.” Differences in Activities Among Nesting Areas: Differences in activities among the various nesting areas were often noted. Thus in late May and early June, 1952, we noted that some nests at Cooper's Place were still being provisioned by the founding queens, that there were no smooth areas around the nest entrances nor were there guards, and that most of the nests were closed and unrec- ognizable. At the same time at Potter's Lake many of the nests had already produced workers which guarded the well-formed entrances and made smooth areas around them; queens were prob- ably rarely seen at the surface of the ground. Not until June 5 did these features appear in the Cooper's Place population. Later we noted that guards in the nests at Cooper's Place ap- peared to be timid and disappeared into their nests at the slightest disturbance; only rarely would they turn and block the entrance with the abdomen and then only well down below the surface of the soil. At Potter's Lake the guards were much less easily dis- turbed, and when disturbed they readily turned and blocked the entrance with the abdomen, doing so close enough to the surface that they could easily be marked with paint for individual recogni- tion while in this position, On July 5, 1952, we noted that colonies at Potter's Lake and Cooper's Place had been inactive for several days, possibly because of prolonged dry, hot weather (reaching 39° to 41° C. daily in shade), but colonies at Petefish’s Place remained active, possibly because of the low, cool location near a stream. In examining nests in different locations, we found that there is a strong tendency for nests in any one location to be similar in depth and form. This is probably because of differences in soil (including soil moisture) in different places. For example, in midsummer, nests at Cooper's Place tended to be deep (usually over 65 cm.), and little branched except at the lower levels where there were cells, At the same season nests from Potter’s Lake were rarely over BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1149 60 cm. in depth, usually shallower, and often rather broadly branched even at the upper levels. These examples illustrate the danger of generalizing from nests observed in only one location or for only a brief period. Survival of Nesting Aggregations: As already indicated, several of the areas described above were occupied by nests of L. incon- spicuum over a period of several years. The site at Potter's Lake had nests in it every year of our study. Sometimes in late summer there were only one or two nests known to us in the area, but each spring a new lot of overwintered queens established nests there. At Cooper’s Place, one of the areas had nests for only two years but the other was occupied each year until it was plowed in 1957. Sometimes in late summer or fall not one nest could be found in the area, but in spring it was always reoccupied. Similar observations were made at other areas of aggregation. There was consistently a relatively enormous mortality of nests during each season of activity. Sometimes, as at the roadside aggregation at Sycamore Slope, although numerous nests were seen in April and May, not one could be found my mid-June. More often a few survived for the entire year. During dry summers we sometimes got the impression that soil humidity was an important factor in nest survival. At the Intersection site, which often became very dry, excavation of nests in August showed that the only ones surviving were those that ac- tually penetrated relatively moist sandstone rock. At Potter’s Lake also, one August when but few nests survived, we found that those which we dug penetrated chunks of moist sandstone scattered deep in the soil. These observations, combined with the fact that nests are dug deeper and deeper during the summer at the same time that the soil is becoming dryer, led to our surmise that lack of soil hu- midity was an important factor in mortality of nests. During the drouth years of 1953 to 56, some of the nest aggregations became entirely extinct, not even being replenished in spring. This oc- curred at the Intersection, an obviously dry hilltop; at Petefish’s Place where the soil became very dry in appearance, probably be- cause of its sandiness, in spite of its low lying location, and probably at Engle’s Place. We only have knowledge of Petefish’s Place being occupied one summer; it is the most abnormal of the sites found. During years of adequate and high rainfall many nests also die. We were not, in fact, able to study any nest aggregation for any 39—5840 1150 THE UNIVERSITY SCIENCE BULLETIN season without being impressed at the high rate of mortality for which we have no satisfactory explanation. The question of the source of the overwintered queens which establish nests in the spring is interesting. There is a surprising number of such bees considering the late spring and summer mor- tality of nests. Presumably they come from the nests that did sur- vive the summer in the areas of aggregation as well as from the isolated nests. The latter are doubtless far more numerous than out data suggest, for the minute holes in the soil are difficult to find. Among nests excavated in the fall and early winter, the aver- age production of young queens was between six and seven, with a maximum of 18 observed on one occasion. Most of these bees disperse in the spring from their parental nests and establish new ones. Survival of Nests: Numerous individual nests were marked with numbered nails pushed into the soil a few centimeters from each nest. Each group so marked and used for the present aspect of the study consisted of nests in a single aggregation. The number of marked nests in each group is indicated in Figure 4. The nests were re-examined at daily, weekly, or occasionally less frequent in- tervals. As will be explained later, nests are often closed tem- porarily. Colonies were considered as dead only when repeated visits showed a nest to be permanently closed. Excavation of a considerable sample of such nests showed no surviving bees of any 100 75 50 25 PERCENTAGE OF SURVIVING NESTS — MAY JUNE % JULY AUGUST SEPTEMBER Ficure 4. Survivorship curves for colonies in six aggrega- tions. A colony was considered dead when its nest became per- manently closed. The number of nests in each group is indi- cated, after the year, along each curve. Each group consisted of marked nests in a single aggregation. BIONOMICS OF A PRIMITIVELY SOCIAL BEE LUST age. Survivorship curves for each six groups of marked nests are shown in Figure 4. It can be seen that in each group, nests succumbed for a time in such numbers that complete extermination before the end of summer seemed likely. As indicated above, this sometimes hap- pens as aggregations of nests do sometimes become extinct. How- ever, each of the curves shown in Figure 4 flattens out as the season progresses, showing that some nests survived through the summer (or at least until the end of the observations) in each case. Pre- sumably this indicates that a relatively few nests, fortunate in lo- cation or in some other attribute, have a good chance of survival while the others die. This means that once a nest survives for a certain time, it has a good chance for longer existence, The group studied in 1953 reached a lower level earlier in the season than any of the others. It was at the Intersection, which as already explained was a dry location from which bees ultimately disappeared completely. The groups for 1951, 1952, and the first for 1954 (marked in April) were all at Potter’s Lake while that marked in July, 1954 and that marked in 1955 were at Engle’s Place. For the great majority of nests, death of the population was not due to any of the obvious natural enemies (mutillids, Paralictus, mold attacking pollen masses). Probably death of the queen and failure to replace her, followed by death of the short-lived workers, was the common cause of extinction. Nest Structure: The nests consist of burrows which extend essen- tially straight downward into the soil, but often meander consid- erably (Figures 5 to 16). In soil full of stones, there are often long horizontal sections following the surfaces of the stones until they reach places where they can go downward. Sometimes burrows, after passing through the soil for some distance, enter sandstone and perhaps other similarly soft rocks; as already suggested such nests sometimes survive better, possibly because of moisture in the rock, than their neighbors entirely in soil. The burrows range from 1.8 to 2.5 mm. in diameter, rarely reaching 3 mm. in diameter in certain parts of the nest. At the surface of the ground the burrows are narrowed to 1.3 to 1.5 mm. in diameter. Nests made by single queens in the spring have no specialized structure at the surface and the minute holes are difficult to find. Holes occupied by more than one queen in the spring, and those occupied by queens and workers in summer, are ordinarily sur- rounded by a depressed smooth, shining area (Figure 41) which P52 Tue Universiry SCIENCE BULLETIN appears to be thinly covered by a layer of material of perhaps salivary origin. These depressed areas are 3 to 4 mm. in diameter and the surface is usually horizontal. When the soil is dry they may be poorly kept and may even disappear due to pulverization of the soil but after a rain each active nest will again be provided with such an area. It is only when quite dry that the salivary layer cover- ing the smooth area can be easily detected, for then it pulls away from the soil in places and therefore has a whitish appearance. Rarely in Kansas, we have observed such areas sloping inward, form- = ie Ficures 5 to 7, Diagrams of nests. For these figures as well as Figures 8 to 16, several abbreviations and symbols are used, Cells indicated by broken lines are earth-filled. There were often more of these than indi- cated; and such cells are often omitted even though some were present, for they are hard to recognize unless very freshly filled with earth. The following letter symbols are used in connection with cells: e, egg; E, empty; EF, empty except for larval feces; L, larva of moderate size; LL, large larva and prepupa; M, cell contents destroyed by mold (fungus); P, pupa (all females unless marked male); PB, pollen ball (but no egg or larva); SL, small larva. When diagrams have been broken to conserve space, connecting points are indicated by the same letters. Figure 5. Two young queen nests opened on April 27. The one at the left has a fresh tumulus and a single empty cell. Each nest was occupied by a single queen. FicurE 6. A queen nest opened on May 6 showing sequence of the brood from oldest above to youngest below, The nest was occupied by a single queen. Ficure 7. An old nest opened on April 24 and found to be occupied by a group of four queens which were reusing the burrows of the previous season. Earthfilled burrows extending to deeper levels are indicated by broken lines. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1153 ing small funnels around the entrance holes. One of us (C.D. M.) noted the entrances regularly funnel shaped, sometimes steeply so, at nests of this species in the garden of Mr. P. H. Timberlake at Riverside, California, where L. inconspicuum has been introduced from the eastern or central United States. Probably the form is in some way related to the texture of the soil. When bees are digging, the excavated earth is pushed up to form a tumulus which spreads irregularly around the entrance and reaches diameters of 12 to 18 mm. The material of the tumulus is granular —~ 8 Pp P Pp P P Ficures 8 to 11. Diagrams of nests. For symbols see explanation of Figures 5 to 7. Ficure 8. Closed queen nest opened for study on May 28. The nest was occupied by a single queen. Ficure 9. Nest opened on June 3. It was occu- pied by a single queen and two adult workers. Ex- cavation beyond that of a queen nest is shown by the cells containing eggs, by the lower empty cells, and by a partially excavated cell lower than any completed cell, FicurE 10. Nest opened on July 16. It contained a queen, two workers, and one bee which escaped. Ficure 11. Nest opened on July 16. It contained only one queen and one worker. Obviously the queen had laid no eggs recently. 1154 Tue UNIversiry SCIENCE BULLETIN but often tends to cling together; the bees find their way through it in irregular passages, there is not a uniform passageway through it as is found in some species. The tumulus may cover the smooth area completely. Tumuli, however, are ephemeral structures which wash away in rains or are blown away by winds, leaving the en- trances open and surrounded by obvious smooth areas. We have not noted evidence of burrows being lined with dirt from deep in the nest, as is so often the case among other halictine bees [e. g., Augochloropsis diversipennis (Lepeletier ), see Michener and Lange, 1959; and Lasioglossum malachurum (Kirby ), Bonelli, 1948]. The burrows are round in cross-section, not especially smooth walled. Queen nests, established in the spring by single queens, range from Ficures 12 to 14. Diagrams of nests. For symbols see explanation of Figures 5 to 7. Ficure 12. Nest opened on July 1. It contained one old queen, five worn workers, and eleven unwom females almost all of which were certainly workers. Ficure 13. Nest opened on July 2. It contained one queen, five worn workers, and three unworn females that almost certainly were workers. Ficure 14, Nest opened on June 26. It contained one little worn queen, eight worm workers, and two unworn females that almost cer- tainly were workers. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 155 Ficures 15 to 16. Diagrams of nests. For symbols see explanation of Figures 5 to 7. FicurE 15. Nest opened on August 4. It contained two queens, twenty workers, and two females of doubtful caste. FicureE 16. Nest opened on September 6. It contained one probable worn queen, eight worn workers, and two young females that might well have been overwintering queens. 9.5 to 24 cm. deep and the burrows are little branched (Figures 5, 6, 8, and 9). Nests occupied by two or more queens, usually or always using the burrows of the previous year in which the queens presumably hibernated, are often much deeper, and more fully branched (Figure 7); the deepest that we excavated was 55 cm. deep, and abandoned burrows filled with earth extended even deeper. Lower parts of such reused nests are always partially filled with soil in spring. With production of workers the nests are extended deeper into the soil, and as shown in Figure 17 and Table V, maximum depths, on the average, are attained late in the summer. It is obvious that burrows are deepened during the summer. It is possible, however, there is differential survival of colonies that have made deeper nests, for as the soil dries during the summer, it may happen that only 1156 Tue UNIVERSITY SCIENCE BULLETIN deep nests reach levels where the soil is damp enough. If this is true, selection would probably favor colonies using nests of the previous year established by more than one queen. The enormous mortality of colonies noted in a preceding section was so far as known mostly in nests started by single queens. Colonies started each spring by several queens may occupy the same burrows for several years in succession and perhaps deepen the burrows some- what each year. The lower parts become filled with dirt (often loose) during spring and early summer, but this may be cleaned out or new branches made during the summer. The deepest burrow number of branch burrows “Siena APRIL- JUNE JULY AUGUST- MAY SEPTEMBER FicurE 17. Graph showing mean nest depth, cell depths, and mean number of branch burrows. For April-May, broken lines also show depths of newly established nests and of old nests of the previous year that are being re-used. Other statistics concerning these matters appear in Table V. BIONOMICS OF A PRIMITIVELY SOCIAL BEE LI57 found (in August) was 93 cm. deep, much branched after the man- ner of Figure 14. The shallowest, late summer or fall, burrow was only 14 cm. deep. As shown in Figure 17 and Table V, the number of branch burrows is much greater in the large nests late in the season than in the small queen nests. Nonetheless the number of branches is not as well correlated with burrow depth as might be supposed. The correla- tion coefficient (r) of nest depth with number of branch burrows ranges from .45 to .50 during the summer months. Figure 14 shows a nest with about the maximum number of branch burrows, while Figures 13 and 16 show rather deep nests with few branches. Of course the level of occupied cells descends with the deepening of the burrows (Figure 17 and Table V). In the queen nests in the spring the uppermost cell is quite regularly about 8 cm. below the TaBLE V.—Nest Depths (in centimeters) and Number of Branches. For total depths, statistics on new nests and on nests of the previous year are segregated in April and May. Later in the season it becomes impossible to distinguish new from old nests. mn=—=number of nests examined; X = mean; SE = standard error of mean. Nest depth | Depth of upper cell n x SE range | n x | SE range April-May....... Ag \re54e| 15389: 55 ae 7-74) Sse (ain tit MEO Wineries = 35 | 13.50 .452 | 9.5-24 Ole shee ae 11 | 30.41 | 4.540 | 17-55 | VUNG eae casts 32 S104 31 | 27.56 | 1.960 | 10-52 | 30 | 12.50 .951 | 6.5-30 Aol pion Scien Hates ee ae 24 | 32.06 | 2.170 | 18-67 | 23 | 15.50 .951 9-27 August-September | 26 | 61.23 | 4.656 | 14-93 | 21 | 41.02 | 4.515 | 7-63 Depth of lower cell Number of branches | n x SE range | n x SE range April-May....... Pe |) Mal 572) B55 7-18 | 29 48 n1G2 0-3 JUNC Mk 30 | 24.20 | 1.790 9-48 | 30 1.60 568 0-11 Jur livg entre ts eices 22 | 26.82 | 2.188 | 13-55 | 24 Dealt 528 0-12 August-September | 21 | 57.95 | 4.65 9-84 | 24 | 5.75 | 1.029 | 0-16 1158 Tue UNrversiry SCIENCE BULLETIN surface of the soil, although depths ranging from 3.5 to 11 cm. have been recorded. After workers mature and the burrows are deep- ened, new cells are constructed at deeper and deeper levels and at increasingly variable levels, as shown in Table V. Cells from which bees have emerged are filled with dirt, usually very soon after emergence, so that the number of used cells that have not been filled with earth in any one nest is always very small except occa- sionally when the adult population is for some reason much reduced. The cells themselves are subhorizontal, sloping downward slightly to their posterior ends, bilaterally symmetrical, the lower surface being flatter than the others, very smooth on the inner surfaces which are lined with a thin coating of waxlike material (Figure 42). Measurements of cells range from 5 to 6.6 mm. in length and from 2.8 to 3.2 mm. in diameter. The variation in length is partly due to variation in the exceedingly short lateral burrows connecting cells to the main burrows. The laterals may be said to be so short as to be absent. Ordinarily after an egg is laid in a cell, it is closed by means of a plug of loose, friable soil. Sometimes when plaster of Paris is poured into a nest, it enters a few of the cells. We do not know if such cells were really open; perhaps the plaster merely pushed aside the delicate plugs. It is clear that plugs are normally present; it is also clear that plugs are absent or destroyed with equal frequency in cells of all ages, including those containing pupae. It is therefore most unlikely that there is any progressive feeding of larvae, since if that were the case one would expect to find plugs more frequently absent in cells containing larvae of certain ages. Occasionally a nest was found in which plaster entered nearly every cell. It should be remembered that in L. malachurum (Kirby) and other species in Europe some authors have insisted that the cells are open while others have found them closed. Perhaps there is variability in the strength of the plugs or even in their presence. Sakagami and Michener (in press) have discussed the various statements in the literature concerning this matter, As shown in the figures, the cells are often grouped more or less close together, but do not form distinct clusters as in some other halictine bees. (See discussion of halictine architecture by Sakagami and Michener, in press ). The numbers of cells in queen nests that had probably reached the age where no additional cells would be made until emergence of workers are shown in Figure 18. It is evident that nests made by a single queen usually contain about five cells, although there is BIONOMICS OF A PRIMITIVELY SOCIAL BEE NUMBER OF NESTS -N Wb OD NUMBER OF CELLS FicurE 18. Histogram showing numbers of cells in queen nests. Queen nests containing only eggs and young larvae, or which for any other reason seemed to be still grow- ing, were excluded from consideration. Exceptions are nests containing 28 and 32 cells which were included because of the large numbers of cells (and queens) even though they might have produced even more cells. The numbers in the squares indicate the number of queens found in each nest. It seems probable that two queens were in nests having 10 and 12 cells, but that one disappeared or was lost in two cases. NUMBER OF NESTS 47 32 20 13 10 MEAN NUMBER PER NEST ) Sea JUNE JULY AUGUST SEPTEMBER Ficure 19. Cells and their contents. The excess of the total cells over eggs + larvae + pupae is due to cells under construction and being provisioned as well as old abandoned cells that were not filled with earth. The dip in cell number in July compared to June and August may be a result of small sample sizes or baised samp- ling resulting from comparing in this figure data obtained in dif- ferent nesting places and in different years; or it may represent biological fact. However, the only year (1952) for which we have good records for both June and July from the same nesting area suggests that the dip in July did not occur in that place in that ie a eee moist soil [June x = 12.3 (n= 22); July x= 15.6 i — F 1159 1160 Tue University SCIENCE BULLETIN considerable variation, Nests occupied by more than one queen contained correspondingly more cells. The average number of cells per nest was 10.1 (This mean is higher than shown in Figure 19 and Table VI because incomplete nests are included in the latter figure and table). After emergence of the workers, the number of cells increases, as shown in Figure 19 and Table VI. The maximum number found in any one nest was 90 (Figure 14), found near the end of June, but nests with over 50 cells were found at various times until early September. During each of the summer months nests with only 4 to 7 cells were also found, and even two or three nests without cells. The nests without cells were not new ones, but old ones with earth-filled remnants of cells, but without new ones. The above figures for cell numbers, and those used in preparing Figure 19 and Table VI, exclude old, used cells filled with earth and in- TasLe VI.—Number of Cells and of Eggs, Larvae and Pupae Per Nest. n= number of nests; x = mean; SE = standard error of mean. Under the head- ing “%” are shown the percentage of the cells that contain eggs, larvae, and pupae. Number of cells Number of eggs n x SE range x SE % | range April-May...| 47 5.45 867 0-28 | 1.81 .574 | 33.2 | 0-24 JUNG ae 32 18.56 4.025 1-90 | 2.78 .891 | 15.0 | 0-19 Dulye canes 20 13.75 2.385 4-47 | 2.65 .956 | 19.3 | 0-14 INU EATERS 5 oo 6 oc 13 21.62 5.281 0-67 62 sO 2.8 | 0-4 September. ..| 10 16.30 4.690 0-52 .70 .422 | 4.3 | 0-4 Number of larvae Number of pupae n x SE % | range x SE % | range April-May...| 47 | 1.92 .3f0 | 35-2 | 0-10 06 .064 | 1.1 | 0-3 JULIe Ses 32 | 5.19 | 1.493 | 28.0 | 0-34 | 5.84 | 1.327 | 31.5 | 0-34 Jialyya rete 20 | 5.40 | 1.125 |} 39.3 | 0-21 | 2.45 .541 | 17.8 | 0-7 August:: .)." 13 | 6.92 | 1.806 | 32.2 | 0-23 | 7.69 | 2.379 | 35.6 | 0-23 September...| 10 | 3.50 | 1.609 | 21.5 | 0-12 | 6.40 | 3.140 | 39.3 | 0-30 BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1161 clude only those containing immature stages of bees and those in the process of being constructed or provisioned and such old empty cells as exist. According to the system of Sakagami and Michener (in press) nests of L. inconspicuum are to be classified as follows: OCh"B. Nest construction: The starting of a new nest by a single queen bee was observed several times in the spring. At this season many of the overwintered queens spend much time crawling over the surface of the ground, flying short distances, alighting here and there, and crawling more, often biting at the ground with their mandibles. Not infrequently they start to dig, but usually after making a hole one or two millimeters deep, they abandon it and continue the activity which looks like searching. This can be ob- served most readily in areas of nest aggregations but with patience can also be seen in many other patches of bare ground. Some- times several such bees were seen at the same time in an area where a nest was never subsequently found. Occasionally, perhaps most commonly in the afternoons, such a bee continues its digging, disap- pearing into the soil in as little as 20 minutes if the soil is moist and easily worked. The digging is done with the mandibles; as the hole becomes deeper earth is pushed out with the posterior-dorsal surface of the abdomen and forms a tumulus which spreads ir- regularly all around the entrance of the burrow. As explained previously, dirt excavated in digging new cells or extending the burrows in established nests is often put in old aban- doned cells (which rarely stay open long) or in short branch bur- rows, rather than being carried all the way to the surface. We replaced the upper 10 to 20 cm. of certain nests with glass tubes which had cell-like evaginations blown in them. The bees never used these evaginations as cells but soon filled them with earth brought up as a result of work below the level of the glass tubes. Nonetheless tumuli appear at irregular intervals throughout the summer at the entrance of every nest containing an active colony of bees, for there is never enough space below the surface for all the dirt that is dug out by the bees. The smooth areas around the nest entrances appear to be made by the bees which act as guards; such areas are absent at the en- trances of nests inhabited by only a single bee, and guarding is not observed at such nests. A bee at the nest entrance can often be seen mouthing the smooth area with its labium, and also rubbing it with its front tarsi, which are bent mesad under the head. Special 1162 Tue Universiry SCIENCE BULLETIN attention is paid to the rim of the entrance hole, i. e., to the inner margin of the smooth area. The postero-dorsal extremity of the abdomen is often used to shape this region, by patting motions when the soil is moist. Similar patting was sometimes observed to shape the outer margin of the smooth area, as is shown by the following notes made soon after a heavy rain. The smooth area had already been fairly well reformed, but the guard bee was still at work on it. She reversed herself inside the nest and backed out of the nest so that the whole abdomen was exposed and strongly arched so that the tip was directed downward. Supported within the entrance by the outside surfaces of the legs thrust against the walls, the bee tamped the smooth area with the apex of the ab- domen, the whole abdomen being moved up and down. Small changes in position allowed the abdomen to tamp different parts of the smooth area. The parts close to the nest entrance were tamped as the bee slowly moved into the nest, and similar tamping continued inside of the nest, smoothing and making more firm the inside wall of the burrow. It would seem that the diameter of the smooth area is a reflection of how far the bee can reach with the abdomen. Probably the size of various other structures of the nest are also related to dimensions of the bees, the entrance being of a size to admit only one bee at a time, the rest of the burrow being large enough to allow bees to pass one another readily. When the soil surface is very dry and the smooth area completely gone due to crumbling of the soil, a little water allowed to soak into the soil about the nest entrance is sufficient to cause the guard to construct a smooth area. Sometimes two or three bees will work at and just inside the nest entrance simultaneously, patting with their abdomens to shape the entrance and its smooth area. Because of the relatively uniformly colored soil in the areas of our study, we did not readily detect the working of the soil forming the lining of the burrow. Soil from deep in the nest lines the bur- row to the surface in many halictine bees and L. inconspicuum is probably not an exception. At least this species has the ability to construct burrow linings of soil, for if we replaced the upper part of a nest with a glass tube slightly too large in diameter, or if we installed a glass window in the side of a burrow, the bees promptly lined the tube or window with soil, patting it into place with the dorso-apical part of the abdomen. This covering was much less BIONOMICS OF A PRIMITIVELY SOCIAL BEE HUGS likely when the glass was covered (except at the times of our ob- servations ) to prevent light from entering the burrow, If the constriction of the burrow at the surface of the ground was destroyed, then the bees would narrow the entrance again by bringing soil from deep in the nest and forming a lining around the entrance of the burrow. Through the glass windows and tubes, as well as by observations of undisturbed nest entrances, it was possible to see that the bees bring dirt up from deep in the nest either by backing upward pushing the dirt with the dorsal apical part of the abdomen, or by moving upward head foremost, pushing the dirt with the face, es- pecially the clypeus. In queen nests occupied by single individuals, the uppermost cell is usually the first provisioned, and provisioning continues downward. The result is that as in Figure 6, there is usually a se- quence in ages of the immature stages from oldest above to youngest below. This is not invariable, however, for sometimes a pupa will be found below a larva, for example. In nests occupied by several queens this order may persist, but more often does not. Thus in the upper right-hand branch of the nest shown in Figure 7, the mixture of eggs and small larvae is quite complete. Because of the progressive deepening of the burrow and of the level where new cells are made, the same sequence, from older above to younger below, occurs in most nests in summer and fall, when the colony contains workers. However, these nests are usually full of ir- regularities in the order of the cells, as can be seen in Figure 16 and, to a lesser extent, in several other illustrations. Most nests in which cells are being constructed and provisioned have one or two complete cells ready to be provisioned, as indi- cated by the letter “E” on the diagrams of nests. This is true even of small insects, such as queen nests made by single queens in the spring. Contrary to many bees which make, provision, and seal a cell before starting to make the nest, L. inconspicuum, while provisioning one cell, seems to be making one or two others. This is true of most Halictinae (see Sakagami and Michener, in press). Provisions and Immature Stages: In opening nests one occa- sionally finds fresh cells containing a mass of loose pollen on the floor (Figure 20, a). Sometimes a small lump has been moistened, presumably by honey, and hence forms an irregular firm mass. When the provisions are complete, however, there is virtually no dry pollen in the cell; all of it is worked into a smooth mass. This 1164 Tue UNrversiry SCIENCE BULLETIN mass of provisions has the form of a flattened sphere (Figure 20, b), 1.9 to 2.5 mm. in horizontal diameter and 1.4 to 2.0 mm. in vertical diameter. The egg is white, curved, 1.33 to 1.50 mm. long about .25 mm, thick (Figure 20, b), and is supported by its two ends on top of the mass of provisions. The larva feeds down into the upper anterior* part of the pollen mass (Figure 20, c, d, e), finally turns over with what is left of the pollen mass (Figure 20, f). After the larva has eaten all the food it lies with its head toward the anterior end of the cell. The feces are then voided as soft pellets pushed against the upper posterior portion of the cell, where they form a firm layer (Figure 20, h; Figure 43). Voiding of the feces took —» Ce Ce Le ca ce Ficure 20. Diagrammatic saavital sections of cells of Lasioglossum inconspicuum. a, loose pollen placed on floor of cell: b, mass of provisions with egg on top; c, d, e, f, larva in different growing stages; g, prepupa; h, pupa. a day and a half for specimens in the laboratory. The larva, at this stage, straightens out and becomes a prepupa, and ultimately a pupa. We have not done much work on the duration of the develop- mental stages. We have found it difficult to rear the larvae in the laboratory. We were repeatedly successful in rearing prepupae through the pupal stage to adults but the younger larvae died. The data given below are based entirely on females (probably workers). The duration of the pupal stage varies enormously with temperature; probably that of the other stages varies equally. Pre- pupae taken on May 10 pupated on May 12; the eyes of these pupae were black 10 days later and adults emerged after another 8 to 9 days; thus the entire pupal stage at room temperatures of 70° to 75° F. required 18 or 19 days. By contrast, another group of prepupae pupated on June 12, and developed at a time when the room tem- perature in which they were kept ranged from 90° to 95° F. These pupae had black eyes three days after pupation and adults emerged * Here and elsewhere the words anterior and posterior are applied to cells and structures therein with reference to the orientation of the egg. Probably in all bee cells the anterior end of the egg is toward the oriface of the cell, i. e., toward plug that is made from the outside by the mother bee as she closes the cell. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1165 three days later, so that the total pupal stage required only six days, or about one third of the time required by the group men- tioned previously. Obviously the rate of development will vary widely according to weather and the depth of the cells. When pupae were reared at temperatures of 81°-83° F., this stage required eight or nine days; this is probably a reasonable estimate of the duration of the pupal stage in summer. Among 70 nests fully excavated during June, July, and August, we found totals of 159 eggs, 354 larvae (including prepupae), 324 pupae and 31 teneral adults with milky wings, still in their natal cells. By omitting vernal and autumnal nests from this tabulation, we have avoided the spring nests that have not yet had time to produce pupae and the autumn nests that lack eggs and young larvae while containing many pupae. The total numbers of in- dividuals of the various stages given above ought to be proportional to the duration of the stages. If this be so, the egg stage is a little less than half as long as the pupal stage or perhaps four days, the larval stage is perhaps slightly longer than the pupal stage, say nine days, and the teneral adult remains in its cell for a period less than one tenth as long as the pupal stage, that is somewhat less than one day. This would give us a total of about 21 days from egg laying until emergence of the adult from its cell under summer conditions. The interval between the average date when pollen collecting was first seen in the spring and the average first date when workers were seen at the nest entrance is 43 days. This period must be roughly comparable to the egg laying to emergence period discussed in the preceding paragraph although pollen is collected and a cell provisioned before the egg is laid; the 43-day period is about twice the estimated summer developmental period of 21 days. This corresponds nicely to the observed duration of the pupal stage in May (18 or 19 days), which was about twice the observed duration of this stage at 80° to 83° F. (8 or 9 days). As indicated in a later section on “Behavior of Queens,” observations of individual nests in the field suggest an egg-laying to adult period of about 30 days in the late spring. INDIVIDUAL AND SOCIAL BEHAVIOR Aggregations of Nests: As stated previously under “Distribution of Nests,” many of the nests are isolated. On the other hand, many are in loose aggregations. Nests were found in relatively few of the apparently suitable areas and aggregations in even fewer. 1166 THe UNtversiry SCIENCE BULLETIN Usually nesting aggregations do not occupy all of the apparently suitable areas of soil in which they occur. It is not clear why the aggregations occur, in view of the apparent success of isolated nests elsewhere and the abundance of seemingly suitable soil. The areas of concentration are not necessarily areas of successful nest- ing; we have seen whole aggregations disappear, probably due to death of bees. The same sites are often attractive spring after spring, and many queens establish nests in areas of nest aggrega- tions, even though few nests may have survived the summer in those places. Possibly the odor of nests or bees attracts others to the same vicinity; perhaps a factor contributing to development of aggregations is a tendency of the young queens to return to the vicinity of birth to make new nests. Nest Populations: The number of adult female bees in one colony (i. @., one nest) varies from one to 25. Most queen nests in the NUMBER OF NESTS 47 32 40 17 3 MEAN NUMBER PER NEST ie} on JUNE JULY AUGUST SEPTEMBER Ficure 21. Number of adult females per nest. Egg-laying queens (groups A and B) were recognized by dissection, The ex- cess of the total over workers (groups C and E) plus queens is due to individuals that were doubtful as to caste as well as, par- ticularly in September, to young queens (group D), The meaning of the lettered groups is explained later in the text. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1167 spring contain only a lone queen, although we have studied nests at that season inhabited by as many as six queens (Figure 18). With the production of workers, the population of adults in the nest rises, reaching a peak in August (Figure 21, Table VII) of over eight females per nest, on the average. This average total, of course, includes numerous small colonies, for throughout the season many colonies die out, and during the process they often reach the stage of having only one or two bees each, TaBLE VII.—Number of Adult Females Per Nest. n= number of nests; xX = mean; SE = standard error of mean. Queens were fertilized egg-laying individuals; workers were usually unfertilized and had not over one or two enlarged oocytes; other females were doubtful as to caste or were young queens. The meaning of groups A, B, C, D, and E, is explained in the text later. The number of nests shown here is more than indicated in Table VI and related materials because of nests whose populations were preserved but for which data on cells, etc., was not recorded, Queens Workers Ca wat : (groups A and B) (groups C and E) aieouted) x SE | range x SE | range x SE | range April- May.) 47. 1221 | .161 0-6 .09 | .053 | 0-2 199) 2.072) | O=2 June....| 32 | .72 | 138 | 0-3 | 3.00 | .803 | 0-16 0 July....| 40 .70 | .090 0-2 | 4.48 | .475 | 0-14 .18 | .062 | 0-2 August..| 17 | 1.18 | .182 | 0-3 | 6.65 |1.654 | 0-22 .24 | .106 | 0-1 Septem- ber...| 9 to > a a ~J 7 —_ _— ~I e) =I oO ‘S fr O 4.11 |1.505 | 0-15 Males: The occurrence of males during the season is indicated in the section on the seasonal cycle. Males leave their nests soon after emerging from their cells; therefore the few males found in the nests do not give a correct idea of the proportion of males produced. Figure 22 shows the percentages of the young produced at various seasons that are males. In contrast to L. rhytidophorum, (see Michener and Lange, 1958) there is no clear peak of male production in spring. Instead, in May and early June no males seem ordinarily to be produced. Males have been recorded, as indicated in the section on the “Seasonal Cycle,” as early as June 1 (two different years) and June 10. Unfortunately these early males have not been preserved and it is now apparent that they might 1168 THe UNtversiry SCIENCE BULLETIN 8 45 164 30 30 72 7 40 29 50 30 (PUPAE AND ADULTS IN CELLS) 20 % MALES MAY JUNE JULY AUGUST SEPTEMBER FicurE 22. Graph showing percentage of pupae and of young adults (in closed cells) that are males. Each month is represented by two points, one for the first half, one for the second half. The number of pupae and young adults examined during each period is shown at the top of the graph. have belonged to Lasioglossum (Paralictus) cephalotes, a parasite in the nests of L. inconspicuum. Most of the males indicated in Figure 22 were produced in nests which were also producing females. However, two nests were found which were producing only males. On July 15, 1951, a nest was opened containing five male pupae but no eggs nor larvae. The nest had been marked on June 6, at which time it contained at least two foraging workers. By July 15 it has been long closed by rains and was unrecognizable at the surface of the ground. It seems possible that the founding queen died, that her unferti- lized workers laid at least five eggs, and that after the workers died these males continued their development. The other nest which produced only males was opened on June 21, 1952. It contained seven male pupae as well as three larvae of unknown sex and three adult females. Two of the latter were workers, both with very slender ovaries while one was a queen; unfortunately her spermatheca was lost and we do not know if she was mated or not. However, she was probably the mother of the BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1169 workers. She contained no very large oocytes (longest .91 mm.), not a surprising finding since the nest contained no eggs or small larvae. The reason for this nest producing males at this season is unknown. Males visit flowers for nectar but are most often seen flying about the nesting areas. At a nesting area like that at Potter’s Lake where the vegetation (except for large trees) is very low, the males zig zag over the soil or small plants, usually flying from two to 12 cm. above the ground. Where there is higher vegetation, they fly higher, often flying in numbers around the weeds at altitudes as high as one meter. In Riverside, California, where L. inconspicuum has be introduced, around the home of Mr. P. H. Timberlake, males were noted in immense numbers flying rapidly about small trees in the nesting area; although there were more at heights of one or two meters, some could be seen around foliage five meters above the ground. When males are scarce, individuals often make these flights alone but when they are numerous they usually dance in loose groups, each individual zigzagging with great rapidity. Often one or a few will alight on the ground or foliage over which they are flying, only to resume flight in a few seconds or minutes. Flights such as this, in the vicinity of nests, are usually in the sun when the temperature is moderate but in the middle of the day on hot days, when the temperature in the shade rises above about 100° F., the flights move into partially shaded areas. At open nesting sites where there is no shade from weeds or nearby trees, the flights cease and the males disappear during the warmest hours of hot days. We are not sure whether we have observed mating or not. Males, sometimes two to four of them simultaneously, often pounce on females. Usually they are repulsed immediately, often only after rolling over on the ground, but sometimes contact lasts for about 10 seconds; this may be copulation. Males pounce upon any halic- tine of about the right size. Young females and old senile females that fly about the nesting area unable to find their nests are most commonly pounced upon by males, but foragers at nearby flowers, males of other species of Chloralictus, and females of the parasite, L. (Paralictus) cephalotes, are all often attacked. Since such activity has been noted on the ground, on foliage, and on flowers; we as- sume that mating may occur at any of these places. We believe that it usually occurs near the nesting site rather than on distant foliage or flowers. 1170 Tue UNrversiry SCIENCE BULLETIN It is significant that males collected on flowers varied in mean size according to the season. It has been shown elsewhere (Mich- ener and Lange, 1958, 1959) that wing length, face width, and other body measurements in halictine bees are highly and positively correlated so that any one can serve as a fair index of size. Facial width and wing length were measured for series of specimens col- lected in different months as shown in Figure 23. For both meas- urements the mean for July males was below means for June and September (difference significant at the 5% level; Q test, Snedecor, 1956, p. 251); differences between the June and September means were not significant. Only small numbers of specimens were avail- able for August; they were intermediate in mean size between the July and September series. As will be shown later, the seasonal °o [o) fo) fe) ° wo [o) 0 rn) nu >) Q PERCENT SEPTEMBER n=39 x= 100.87 SE=.568 CV= 35l ry 63 6 67 S94 73 %5 77 7% 8) 8B % 92 94 %6 98 |Og Io !04 !06 !0g WING LENGTH (MICROMETER FACE WIDTH (MICROMETER DIVISIONS) DIVISIONS) Ficure 23. Histograms showing wing lengths (left) and head widths (right) of males of Lasioglossum inconspicuum collected on flowers in the months of June, July, and September. Statistics are shown in micrometer divisions with equivalents in millimeters across the top. The means are marked by black triangles at the bases of the histograms, Abbreviations; n—=number of individuals measured, iS mean, SE = standard error of mean, CV = coefficient of varia- ion. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1171 size variation of the males is positively correlated with that of fe- males. Variation in Size Among Field-collected Females: Females, like males, collected on flowers varied in mean size according to the season, as shown in Figures 24 and 25. Minimum size occurred in July as determined by wing length, in August as judged by head width. To judge by observations of nests, the April and May in- dividuals should all be queens, the June, July and August individ- uals should nearly all be workers, while those taken in September should be a mixture of workers and young queens, The relatively low means for September (as compared to April and May), and the high coefficients of variation for that month, must be due in part to the mixture of workers and young overwintering queens on the flowers during that month. The young queens that actually over- winter are the same individuals that visit flowers in April and May and should, of course, be of the same size unless there is differential mortality of smaller individuals. The means shown in Figure 24 are significantly different at the 5 percent level (a posteriori Q tests among means, Table 29, upper 5 percent, Pearson and Hartley, 1956) except that September is not significantly different from June and August, and June is not signifi- cantly different from April and May. The means shown in Figure 25 are significantly different at the 5 percent level by the same method except that July and August are not significantly different from one another or from September, and June is not significantly different from April and May. Caste Differences: The interpretation suggested in preceding sec- tions indicates the presence of more or less different female castes (queen, worker). In order to study the castes, entire populations from over 150 nests were captured and dissected. Occasionally a bee escaped, but insofar as possible complete nest populations were obtained (1) by excavating in cool cloudy weather when no bees were afield; or (2) by plugging nests that were to be opened either at night or during other hours when all the bees were presumably in the nest; or (3) by watching a nest for one and one half hours or more before excavating it, in order to capture returning bees, mean- while preventing the departure of all individuals from the nest. All individuals from such nests were dissected to determine ovarian development, presence of pollen in the crop, and of sperm cells in the spermatheca. 1172 THE UNIveRSITY SCIENCE BULLETIN Ficure 24. Histograms showing wing lengths of females of Lasio- glossum inconspicuum collected on flowers in various months. Statistics are shown in micrometer divisions with equivalents in millimeters across the top. The means are marked by black triangles at the bases of the histograms. These measurements cannot be compared directly with wing measurements in subsequent tables for the wings were measured by a different person and in a different manner. Abbreviations: n= number of individuals measured, x = mean, SE = standard error of mean, cv = coefficient of variation. PERCENT fe) 6, BIONOMICS OF A PRIMITIVELY SOCIAL BEE 63 65 67 Sg 7 WING LENGTH JULY n=|37 %=-69.83 SE=.345 CV=5.77 AUGUST n=64 XR=71.64 SE=.397 CV= 4.43 SEPTEMBER n=20 x=73.50 SE=1.169 73 75 77 % 8 83 85 87 (MICROMETER DIVISIONS) 1173 1174 THe UNIversiry SCIENCE BULLETIN On the basis of ovarian development and presence or absence of sperm cells in the spermatheca, females can be divided into sev- eral groups as follows: A—Fertilized, with ovaries much swollen, usually to the point that enlargement of the posterior portions of one or both ovaries often forces the anterior portions of one or both into sharp bends or convolutions (Figure 26). Such ovaries seemed more swollen than in the corresponding class of L. opacum and rhytidophorum (Michener and Lange, 1958), as would be expected since the first is a nonsocial species and the second has shorter- lived queens than those of inconspicuum. No doubt inconspicuum queens produce more eggs than those of either of the aforementioned species of Chloralictus. The ovaries of none of these species are nearly as large, however, as those of the species with perennial nests and large colonies (Plateaux-Quénu, 1959). B—Fertilized, with ovaries swollen but not so much as in Group A, not sharply bent, usually one or two ovarioles in each ovary not swollen (Figure 27). MILLIMETERS {0} te ey 1) oN DO fF aaaa 1,050 1.075 1.100 1,125 1.150 1175 1.300 1.325 1.350 1.375 1.400 1.425 1.450 1.475 Te) nu 2 A PERCENT SEPTEMBER n=20 <0 x= 100.65 SE=1.606 CV= 7.13 fo) 82 84 86 8g % 92 % % %B log !oO5 log !Og !Og Ng No Ng Ng Ig FACE WIDTH (MICROMETER DIVISIONS) Ficure 25. Histograms showing head widths of females collected on flowers in various months, BIONOMICS OF A PRIMITIVELY SOCIAL BEE IL (5) C—Unfertilized, with ovaries slender except for one enlarged oocyte in one ovary, or rarely one in each (Figure 28). D—Fertilized, with slender ovaries. E—Unfertilized, with slender ovaries, Because so many bees fell in this group, it was decided to divide it into two arbitrary subgroups for separate analysis. These were E’ with very slender ovaries (Figure 30) and E” with merely slender ovaries (Figure 29). OVO 9 ¢ Ficures 26 to 30. Ovaries. 26, group A (queen); 27, group B (queen); 28, group C (worker); 29, Group E” (worker); 30, group E’ (worker). From the standpoint of ovarian size, the groups merge. All females emerge from their pupae as Group E’. Some remain in that condition while others develop into one or another of the other groups. It is therefore to be expected that intermediates would occur. There is also excellent evidence that ovarian size may be reduced, so that progress from A to B may occur but probably not as far as D. If a group C female lays an egg she would revert to group E. Peculiarly-shaped ovaries with irregular swellings or with a developing oocyte that is much shorter than normal occur occa- sionally. Such individuals were placed rather arbitrarily in the classification indicated above. In a certain percentage of individuals dissected the spermatheca was missed, or lost before it could be examined for the presence of sperm cells. In some cases such specimens were included in the analyses which follow. For example, since virtually every individual taken in April and May is fertilized, such specimens were included in group A or B even if proof of fertilization was not obtained. Figure 31 shows the sizes (wing lengths) of females segregated into the groups listed above. In view of the seasonal cycle of the species and the differentiation of castes explained below, and to make the groupings more biologically significant, group D is limited for purposes of Figure 31 and related discussion to individuals obtained from October to May (overwintering or overwintered queens) while the other groups are limited to individuals taken in June, July, or August. In this species there are very few fertilized 1176 THe UNrversiry SCIENCE BULLETIN Ficure 31. Histograms showing wing lengths in millimeters of females belonging to groups A to E. These are groups based on ovarian development and presence or absence of sperm cells in the spermatheca, as indicated in the text. The histogram for group D is based on specimens taken from October to May; the other histograms are based on specimens taken from June through August. Means are shown by triangles on the base lines. The vertical broken lines are only to facilitate comparison and represent the minimum mean (group E’) and the maximum mean (group A). PERCENT oO = BIONOMICS OF A PRIMITIVELY SOCIAL BEE 2g 29 So 3) 32 33 34 35 36 37 3g 39 40 44 42 WING LENGTH (MM) 1p wee 1178 THe UNIverSITY SCIENCE BULLETIN workers; therefore there is no large number of group D individuals in summer, as in L. rhytidophorum. It is apparent that, correlated with the functional differences re- sponsible for segregation of individuals into groups A to E, there exist certain size differences, as indicated by measurements of wing lengths. There are not, however, external morphological differences between the female castes as in bees like Trigona and Apis. No difference in size exists between groups E’ and E”. These con- stitute the typical workers. Group C consists of workers which may lay one or more eggs. Their mean size is larger than that of work- ers with slender ovaries. The difference is significant at the one percent level (Q test, a posteriori, among means, Table 29, upper one percent, Pearson and Hartley, 1956). The remaining groups con- sist of queens. Group D, consisting almost entirely of overwintering females, should represent queens as a whole; its mean size (as meas- ured by wing length) is significantly different from that of group C at the one percent level. During the active season queens are divided into two groups, A and B, according to ovarian size. Those falling in group A have a mean wing length slightly but not signifi- cantly larger than those falling in Group B. The mean wing length of groups A and B are not significantly different at the 5 percent level (Q test, a posteriori, among means, Table 29, upper 5 percent, Pearson and Hartley, 1956) from that of group D. The mean wing length of group B is not, but that of group A is significantly different from that of group C at the 5 percent level. Since queens average larger than workers, one might suspect that in any one nest the queen would be larger than the workers which must usually be daughters of the queen. While this is more often true than not, nests were commonly found in which one or more of the workers was larger than the queen or queens. In June 77.7 per- cent of the nests studied had the queen (or at least one of the queens in nests containing more than one) larger than any of the workers. In July and August comparable figures were 65.2 percent and 58.3 percent, respectively. That the queens average larger in June than in July and August, probably because of replacement of some or many overwintered individuals, is suggested in the section on “Sea- sonal Differences in Size,” and would explain the progressive reduc- tion in the percentages shown above. There is no correlation between worker size and colony size in this species. It is noteworthy that in this species the smallest workers are BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1179 smaller than the smallest queens. This is not true of L. rhytido- phorum in which the entire range of size variation of the workers overlaps that of the queens, although the largest queens are larger than the largest workers (see Michener and Lange, 1958). Thus it seems that the differentiation of the castes from the standpoint of size is more complete in inconspicuum than in rhytidophorum, a conclusion that is not surprising in view of the greater differentia- tion of the castes from the viewpoints of longevity and ovarian de- velopment. That there is no necessary correlation among various kinds of caste differences is shown, however, by Plateaux-Quénu’s (1959) study of Lasioglossum marginatum, a species with queens and workers well differentiated in behavior, longevity, etc., but equal in size, Unfortunately suitable statistical data are not avail- able for such well-studied species as L. malachurum. Seasonal Differences in Size: It is evident that part of the sea- sonal variation in size shown in specimens caught on flowers (Fig- ures 24 and 25) is due to the caste differences correlated with ovarian size and mating. Overwintering individuals are all queens (group D). In April and May they are the bees making and pro- visioning new nests. In the fall new overwintering individuals are produced. In the months of June through August, most of the individuals on flowers are workers (groups C and E). It is there- fore not surprising that spring and fall individuals taken on flowers are larger than those taken during the summer. However, during the summer months differences were also noted among the monthly mean wing lengths of field caught specimens and some of the monthly means were significantly different from others. In order to elucidate such variation during the summer, seasonal size variation of bees removed from nests and segregated as to group was investigated. The results are shown in Table XIII. Group B is excluded since it is represented by so few individuals; group D does not appear in the table since virtually all bees of that group (all those utilized in Figure 31) were overwintering or overwintered queens. It is evident from Table XIII that the June individuals average larger, July and August ones smaller. This fact is particularly noteworthy among queens (group A), where the August individuals average larger than those taken in July, and among workers with one or two large oocytes (group C), where the August individuals average smaller than those taken in July. The differences between monthly means for group C are statistically significant at the one percent level and the difference 1180 Tue UNiversiry SCIENCE BULLETIN between June and July means for group E are significant at the five percent level. The consistently larger size in all lettered groups for June than for July and August, considered with like fluctuations in field caught females (Figures 24 and 25) and males (Figure 23), makes reasonably clear that June individuals of each lettered group do average larger than those taken in July and August. The mean sizes of individuals with unworn mandibles (pre- sumably young adults) (last column, Table VIII) practically do not vary from month to month. Presumably most of these indi- viduals would become group E workers. Note the similarity of TaBLe VIII—Wing Length (mean and their standard errors) of Females Belonging to Various Groups (explained in text). Not all were taken from nests. Comparable data on unworn and presumably young individuals ob- tained from nests during the summer months are given in the last column. The numbers of individuals examined are shown in parentheses. E (G A Unworn Junes:2>: 3.36 + .028(60) | 3.56+ .033(15) | 3.65+ .051(15) | 3.31 + .045(28) July.....}| 3.30+.010(265)| 3.45 + .029(23) | 3.50+ .045(18) | 3.29+ .016(105) August...| 3.31 + .010(210)| 3.39 + .024(35) | 3.58+ .051(12) | 3.80+ .018(59) their mean wing lengths to those of group E workers shown in Fig- ure 31. There is evidence presented below (section on “Behavior of Workers”) showing that such workers are short lived; it is therefore not surprising, since they not only constitute the bulk of the pop- ulation but also must constantly be replaced, that the bulk of the summer production is of such workers. Evidence presented below shows that queens and perhaps workers with one or two enlarged oocytes (group C) live longer than group E workers. To appear in summer collections from nests, therefore, such individuals would not need to be produced in very large numbers. The group C workers (plus perhaps various potential group C individuals, as well as group C workers that have just laid an egg and hence look like group E, and intergrades between groups C and E) visit flowers, although queens rarely do so in summer. It seems likely that the monthly fluctuation in mean size of field caught individuals results largely from such workers with reproductive tendencies, which, to- gether with the queens, seem to average distinctly larger in June than in July or August if Table VIII is meaningful. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1181 Yuh Ficure 32. Mandibles of females showing typi- cal examples of the five degrees of wear utilized in the analysis of activity. The indices of wear used are 1, unwom (below); 2, slightly wom; 3, well wom; 4, much wom; 5, very much worn (above). 40—5840 1182 THe UNtversiry SCIENCE BULLETIN The large size of the June queens (group A, Table VIII) com- pared to the overwintering queens (group D, Figure 31) may indicate differential survival of the largest queens. The smaller mean size of the July and August queens (Table VIII) may indi- cate that replacement queens of average size for queens (group D, Figure 31) have been produced in at least some of the nests during the summer. Queens: The percentage of females belonging to each group at various seasons of the year is shown in Table IX. This table is based entirely on material from excavated nests. It is apparent that overwintering occurs almost entirely as fertilized females (queens) with slender ovaries (group D). The only exceptions found to this statement were three unfertilized individuals (group C) that must have overwintered and were found in different nests in April, 1955. Because the wings are rarely nicked in this tiny species, the in- dex of wear, unlike that used for other halictines in previous pub- lications, is based entirely on the mandibles (1, unworn; 2, slightly worn; 3, well worn; 4, much worn; 5, very much worn; see Figure 32). The index of wear therefore reflects the amount of digging in the ground done by the bees. As shown in Table X, the over- wintering bees are unworn or nearly so. TasLe IX.—Percentages of Females Belonging to Groups A to E During Var- ious Months. All specimens were taken from nests (data for April were obtained in the last half of the month). Percent in each group Number of bees | \ B D C E | PM oy AE os dai sae 35 42.9 af. 1 11.4 8.6 4.3 eres Misty oa, steers. cae 45 42.2 33.3 17+: ) bce 6.7 INNES arses as 103 18.5 3A9) *|. aes 12.6 65.1 Rye ws ee. | 218 9.0 4.7 1.4 5.7 a5 ARIBUISG Ss, oct ike hecho 9.2 (| 3.8 6.1 74.8 September........ 59 er ENTS 57.6 LZ 37-3 October si). sae TZ Wa hc eds toe Ge tee TOOT O! "sac | ee November-— December...... 14, Re eee ee 10020) ||... b.5 eee BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1183 By the second half of April, 80 percent of the individuals obtained had enlarged ovaries (groups A and B) as shown in Table IX. The time of emergence from overwintering quarters appears to be highly variable, for we have seen new nests established from early April to mid-May. Nests seen and excavated by us were nearly all open ones; we would have missed any closed ones still containing hiber- nating bees. Therefore it is probable that less than 80 percent of the total female population develops enlarged ovaries by the second half of April. This belief is supported by the relatively high per- centage of bees still having slender ovaries in May (group D, Table IX) and by the not or little worn mandibles of many bees in May (Table X). The sharp reduction in the percentage of queens (groups A and B) in the population in June and subsequent months (Table IX) is due to the large worker populations (groups C and E). The number of queens per nest decreases only slightly and gradually through the summer months, as shown in Table VII and Figure 21, which are based for the most part on data from the same nests as Tables IX and X. Table X shows that the average indices of wear for queens dur- ing the summer are high, probably in group A progressively higher as the summer months pass. Table XI, giving the raw data for group A and B queens lumped together, shows this trend more clearly. This trend seems to indicate that some of the overwintered queens live through much of the summer. It was our impression as specimens were dissected that at least through June and probably well into July and perhaps August, the queens that we found were mostly overwintered individuals. Survival of queens through most of the summer would also be expected since most males appear in late summer and fall, they mate with young queens that will overwinter. If replacement of the queens in nests were a general or universal event, for example in midsummer, males would be expected to be as abundant in midsummer as in late summer and autumn, Table XI shows that scattered unworn and slightly worn queens were found in nests through the summer months. Table X shows that most but not all of these few queens were of group B; it also shows that a few not or little worn, fertilized bees with slender ovaries (group D) were found in July and August. Such bees can only be regarded as young queens produced in summer, perhaps to replace ineffective or dead overwintered queens. Figure 33 shows 1184 THe UNrversiry SCIENCE BULLETIN TaBLE X.—Indices of Wear of Females Belonging to Groups A to E During Various Months. All specimens were taken from nests (data for April were obtained in the last half of the month). Each entry consists of the mean index of wear, the extremes in parentheses, and the number of speci- mens examined. From May to September data are given for each half month instead of on a monthly basis. Mean (and extremes) indices of wear, followed by N A B D C E April 2.9(2-4)15 | 2.8(2-4)13 | 2.5(2-3) 4 | 3.3(2-4) 3 a4 male ss May 1.5(1-3)16 | 2.1(1-3) 9 | 1.6(1-4) 8 3) (@) Sa BeayOA 3) 8 eG) ae 2 -2) 2 x —— 3.3(2-5) 8 | 2.0(1-3) 2 1.5(1-3)22 BLOCH lil || aby 2.2(1-5)13 | 1.8(1-5)45 Tuly 3.2(2—4) 9 | 4.2(4-5) 5 3.0(2—4) 5 | 1.6(1-5)112 3.6(1-5)10 | 3.2(2-5) 5 | 2 (1-8) 3 | 3.4(1-5) 7 | 2.2(1-4)58 Angast 4.2(3-5) 6 | 2.8(1-4) 5/1 (1) 3 | 2.8(1-4) 4 | 1.9(11-5)73 4.7(4-5) 6 | 4.0(8-5) 3 3.2(2-5) 5 | 2.1(1-4)25 September & @) il 2.9@=5) 730 (3) 1 | 227G=5) ita (4) 1 |} 1.1(01-5)27 i (GD) October 1.1(1-2)17 November-— December 1G) lS the increasing percentage of queenless nests as the season advances. No doubt sometimes the death of a queen results in death of the colony; presumably, however, she is sometimes if not normally re- placed, as is regularly the case in nests of L. rhytidophorum (Miche- ner and Lange, 1958). Efforts to shed light on these problems by marking queens in the spring when they are active as pollen collectors, in the hope of later digging them from the same nests and thus gathering data on their longevity, were failures year after year. This failure was due, in different instances, to loss of marking paint, the normal high colony mortality (Figure 4) and perhaps in other factors. The possibility exists that the queens are regularly shorter lived than we have indi- cated above. However, as shown in Figure 22, males are not pro- BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1185 TasLE XI.—Indices of Wear of Queens of Lasioglossum inconspicuum with Enlarged Ovaries. Except for April, data are given for each half month instead of on a monthly basis. The data for April were obtained in the last half of the month. Number of queens (groups A and B) Index of wear April May June July | August | September RE a Ab vai tees. oon liane. aucune Sets aict 82, 2 Ihe" il 2 5 1 AMEN eS ed ae a SP ry Aha. £s: 2% 8 7 4 3 1 ae A Saccaehag Pe ae 7 4 3 4 1] a2 oT) Leal ate es Soho DE ASTS CAE Ee ne 13 lO 2 Dee oe a bal fire eA ot ecg ae Ly eR AE Reeesd aster td 11 eZ i 1 ee sere Stee i duced in numbers until the last half of June. Therefore it seems clear that new queens to replace overwintered individuals would not be likely to function until late June or early July. Most likely they arise as occasion demands from late June on through the summer. Queenless nests are not rare in summer as shown in Figure 33. Many queenless colonies are certainly dying but others probably NUMBER OF NESTS 10 25 33 28 19 4 40 ” E ” w z 30 ” wn mr) —s z Ww Ww > 20. Co x 10. io} APRIL MAY JUNE JULY AUGUST SEPTEMBER Ficure 33. Graph showing the percentage of queenless nests in vari- ous months. 1186 Tue UNrversiry SCIENCE BULLETIN develop new queens. More specific data bearing on replacements of queens are as follows: On June 19, 1952, a nest (A) was opened that contained no queen except an unworn one of group B, with no eggs nearly ready to lay. There were neither eggs nor young larvae in the nest but there were one large larva, one prepupa, 15 female pupae and young adults still in their cells, and six unworn bees (group E) presumed to be workers. (Their wing lengths ranged from 3.04 mm. to 3.13 mm.; the queen’s wing length was 3.27 mm.) During the following six days five other nests were opened; their queens were all old (indices of wear 4 and 5), of group A, and all but one of the nests had brood of all ages. It seems apparent that the original queen in nest A had been replaced, although the replace- ment had not yet laid her first egg, while in the other nests the over- wintering queen was still active. From the above example one might suspect that group B queens are young ones developing to- ward the group A condition. Of course this is sometimes true but as Table X shows, group B individuals are often old and badly wom too. Another interesting nest (B) was excavated on July 23, 1952. It contained brood of all ages and a probably overwintered, much worn (index of wear 4) queen of group A. It also contained another queen of group A and one of group B, both only slightly worn (index 2). Probably these were queens produced during the sum- mer; possibly they replaced dead overwintered queens, but equally possibly they were additional queens. The nest also contained one fertilized worker (group D) (captured and marked as it left the nest on July 20, suggesting worker-like activities but not proving them) having one oocyte, .5 mm. long. Its index of wear was 3. In the nest also were two workers of group C and nine of group E. On August 13, 1952, a nest (C) was opened. It contained two queens (group A) with indices of wear of 5. Possibly they were overwintered. In addition there was an individual of group D which, however, showed slight signs of general ovarial enlargement. Probably this was a young queen. In addition their were two anomalous individuals, among the half dozen summer specimens successfully dissected which were not placed in any one of the groups (A to E). They are probably best called fertilized workers but both had very irregularly enlarged ovaries. The enlargement involved chiefly one ovariole in each ovary, but the others were slightly enlarged; no oocyte approached the size or shape of a mature egg. One of these had an index of wear of 1, the other of 3. There were also 18 workers of group E. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1187 In nest D excavated on June 24, 1954, which contained brood of all ages and three queens of group A (indices of wear 2, 4, and 4), there were also three unmated individuals included in group C. All were not or slightly worn (indices of wear 1, 1, and 2). They dif- fered from ordinary individuals of group C in having all the ovarioles slightly enlarged, no oocyte greatly enlarged (largest .7 mm. long). They were larger than average workers (wing lengths 3.53 to 3.73 mm.). Possibly they were potential replacement queens. In addi- tion there were thirteen probable workers of groups C and E (wing lengths 3.32 to 3.69 mm.). The largest of these was unworn and might have developed toward queenhood but one with a wing length of 3.64 mm. had an index of wear of 3. Nest E, excavated on August 15, 1958, contained brood of various ages and one queen of group A (index of wear, 4). In addition it contained five workers. The latter had wing lengths ranging from 3.16 to 3.39 mm., while the queen was the same size as the largest worker. An additional bee was unworn with very slender ovaries; it was fertilized and therefore fell in group D. Its most remarkable feature was its size, only 2.84 mm. in wing length. It is perhaps significant that most of the anomalous individuals discussed in the preceding paragraph occurred in nests containing relatively large populations and more than one queen. Perhaps whatever circumstances control the fate of a female function with less certainty in such nests, so that the anomalies discussed occur largely in such nests. It is obvious that such individuals are not characteristic of nests where the queen has died, leaving a group of workers some of which might in theory replace the queen func- tionally. Not a single caste anomaly was found in a queenless nest. Workers: The three unworn young workers recorded for the month of May in Tables IX and X were all obtained on May 31. The number of the female bees that were classified as workers (groups C and E) is shown to fall between 77 and 86 percent for all of the summer months in Table IX. As is indicated in Table X, many of these bees which were called workers were unworn indi- viduals that might in theory have developed into queens. That not many of those classified as group C or E in the summer months would have done so is indicated (1) by the small percentage of queens present in summer (2) by the fact that indices of wear for individuals of group E that have wing lengths less than 3.20 mm. and would therefore almost never become queens (Figure 31) aver- age about the same as those of larger individuals that might become queens. If any large percentage of the larger individuals were being 1188 THe UNrversiry SCIENCE BULLETIN removed from the worker class to become queens (so that they would be counted as workers only when young and unworn), the index of wear of the smallar ones classified as group E should aver- age higher than that of the larger. From the above it can be seen that classes A to E, while roughly comparable to similarly lettered classes of L. rhytidophorum (Michener and Lange, 1958), are not all identical in content. Classes A, B, and C are essentially alike for the two species. Class D as used in the present study includes overwintering females which were specifically excluded from that group in the study of rhytido- phorum. They would have fallen in class D, however, had they not been dealt with in another way. Class E was limited by exclusion of unworn individuals in the study of rhytidophorus. This was necessary in order to get satisfactory data on workers because of the large percentage of queens and their presumably frequent replace- ment in rhytidophorum. Unworn individuals placed in group E in September and October are no doubt mostly or all unmated queens that would become group D on mating. Fertilized workers, while constituting a considerable percentage (nearly 8%) of the summer population of females of L. rhytido- phorum, are nearly absent in inconspicuum during the summer. They would, of course, fall in group D. One such individual is dis- cussed in the paragraph on nest B in the preceding section on “Queens,” and one other was tentatively recognized during the sum- mer. Of course such workers could be distinguished from young queens only if their activities were noted or if mandibular wear indi- cated considerable activity and age in a fertilized bee whose ovaries remained slender. Such bees were not found during the summer. In spring, however, some such bees were found as indicated in the section on “Division of Labor Among Queens in Polygynous Nests.” In September, however, fertilized workers appear in some num- bers, as suggested by the high indices of wear of some individuals of group D, Table X. In a queenless nest excavated on September 17, 1952, four of eight workers were fertilized; one was unworn and might have overwintered as a queen but the others, having indices of wear of 2, 3, and 5, would not have overwintered, since over- wintering bees are ordinarily unworn. On September 5, 1954, among specimens captured on flowers was one worker (out of six dissected ) that was fertilized; it was collecting pollen like any field bee, and had an index of wear of 2. On September 15, 1954, five BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1189 workers (out of seven taken on flowers) were fertilized; they had indices of wear of 2 and 3. From these data one may summarize that in the late summer or fall, at the time that overwintering queens are being produced or probably shortly before, workers are produced which have certain queenlike characteristics, specifically willingness to mate. It is not surprising to find a physiological or behavioral continuum between workers and queens, since there is no clear differentiation in size or other characters either. Two unfertilized workers were dissected and found to be anoma- lous for their unusual ovarian development. An individual which had ovaries like those of a group B queen but lacked sperm cells in the spermatheca was taken collecting pollen on flowers on August 5, 1954. It was classified as a group C worker. It was well worn (index of wear, 3) and quite small (wing length 3.39 mm. ). On August 23, 1954, a nest was excavated which contained brood of all ages except eggs. There were two queens of group A (indices of wear, 5) which had perhaps overwintered, and twelve workers (three group C, nine E). One of the workers (classified as group C) was peculiar in having ovaries like those of a queen of group A. Yet it was clearly unfertilized and was captured while entering the nest with a pollen load, indicating workerlike activities. This was the largest worker in the nest (wing length 3.55 mm.; others ranged from 3.21 to 3.42; the queens’ wings measured 3.83 and 3.85 mm.) and had an index of wear of 4. Since egg laying by workers is known to occur in queenless nests of Apis and Bombus, the question naturally arises as to whether workers of group C, and more especially those very unusual ones with much swollen ovaries such as described in the last two para- graphs, arise in nests which have lost their queens. The only un- fertilized individual with much swollen ovaries taken from a nest was with two presumably functional queens (see preceding para- graph). As to ordinary workers of group C, one or more was found in each of 15 nests containing queens (groups A or B), and in only four nests lacking queens. It therefore seems clear that ovarian development of workers is not a result of lack of queens in the nest. The importance of unfertilized egg-laying workers has been studied for halictine bees only by Noll (1931) who believed that most or all of the males develop from eggs of such workers. His evidence was very meager, however. No significant evidence on this matter was obtained in the present study. 1190 Tue Universiry SCIENCE BULLETIN Behavior of queens: Since most of the queens are produced in late summer or fall, we will discuss first the autumnal behavior of the young queens. Workers vanish entirely during September leay- ing the nests occupied by young queens. They do not maintain polished areas around the nest entrances, even if rain moistens the soil so that it would be easily worked. They commonly do maintain the nest entrances in their neat, constricted form, and on warm days at least as late as October 22 they are often to be seen guarding the nest with the head filling the hole, just as do guards at other seasons. We have not seen guarding in the few nests that were later found to contain only a single young queen. During September and October both sexes may be seen sucking nectar from various flowers such as Solidago and especially Aster. Although males become less abundant on flowers as the season progresses, the young queens are sometimes very abundant on flowers in mid-October. From October 19 to 22, 1952, by which time most of the flowers were dead, females of L. inconspicuum as well as other species of Lasioglossum were feeding on honeydew from the aphid, Anoecia corni Koch (determined by E. O. Essig). The bees were lapping the honeydew from the leaf surfaces of Cornus asperifolia on which the aphids were feeding. This was the only time that bees were seen feeding except from flowers. By late October most of the nests become closed by rains and the bees remain inside. No nests were seen open in November. Excava- tion of nests in mid-November showed the overwintering queens at depths of 22 cm. and deeper. They were either in the main burrow, in short branches or in empty cells. As indicated in the section on “Seasonal Cycle,” young queens first reappear at the surface in late March or April. For a time many, at least, of them fly about and alight here and there on the soil as though looking for nesting places as described in the section on “Nest Construction.” Such behavior may be seen as late as May 19, probably indicating diversity in the time when queens start their spring activity. As already indicated, some of the queens remain in their over- wintering nests and re-use them, while others establish new nests. ( Observations on the manner of establishing new nests are described in the section on “Nest Construction.”) Those which establish new nests do so as lone individuals while those remaining in old nests often are associated there with their sisters. BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1191 Lone queens do not make the smooth areas around their nest entrances which otherwise characterize nests of L. inconspicuum. Neither do lone queens appear as guards at their nest entrances. However, disturbance of the entrance with a hair or grass blade does sometimes cause a lone queen to come to the entrance, bite at the offending object, and even turn and plug the entrance with her abdomen as guards typically do (see section on “Guarding and Other Activities at the Nest Entrances”). One such queen was seen to keep a small mutillid out of her nest for an hour with her ab- domen. Nests occupied in spring by two or more queens are often provided with smooth areas around the entrances, just as with the summer matrifilial colonies. In such cases guarding of the entrance occurs regularly. Workerlike activities of some queens in polygy- nous nests are suggested in the section on “Division of Labor among Queens in Polygynous Nests.” Because of variable (and often rainy) spring weather, the dura- tion of the various phases of the spring activity of queens is variable. However, we have many data comparable to the following: Nest A, excavation started April 26, pollen still actively being collected on May 8, the nest closed and not recognizable on May 13, 18, and 25, reopened by one or more workers on June 1; Nest B, excavation started May 8, provisioning observed on May 13 and 15, nest closed and not recognizable on May 25 and June 8, open and at least one worker present on June 12. Such data show that after provisioning the last cell of the queen nest, the queen remains inactive in her nest for two or three weeks until the emergence of the first workers. This is an approach to the condition found in some Eurasian species in which the queen produces her offspring in discrete broods during the spring and summer (Noll 1931; Bonelli, 1948) or annually in the case of queens that live for several years (Plateaux-Quénu, 1959). In L. inconspicuum there is no evident cessation of activity of the queen after the appearance of the first workers in June until her summer or autumn senility. After emergence of the workers and resumption of her activities, the queen never, so far as we know, acts as guard, nor does she collect pollen. Queens regularly have pollen as well as nectar in their crops, showing that they eat freely, as would be necessary con- sidering their continued production of eggs. We have taken two queens (group A), one in July, one in early August, on flowers; both were apparently sucking nectar. We captured one queen (group A) as she was returning to her nest on July 21. From these data 1192 Tue Universiry SCIENCE BULLETIN we suspect that the queens leave the nests to feed. The paucity of such records (3 egg-laying queens out of 180 females captured in the field in June, July, and August and dissected) shows that the queens do not spend much time away from their nests and, further- more, suggests that they may feed sometimes if not regularly on food brought to the nest by the workers. One group D female (probably a replacement queen ) was taken on flowers in July. That queens even after there are workers in their nests, continue to work with their mandibles, presumably in cell or burrow excava- tion, is shown by the increasing average indices of wear of queens as summer progresses (Tables X and XI). This increase is also one of the best evidences of longevity of many queens, although there is also evidence of production of new queens during the summer. That much of the mandibular wear results from construction of cells and perhaps from the construction of the fine smooth cell lining is suggested by the fact that queens taken in April and May from old nests whose burrows were already made the previous year have mandibles just as worn as do queens that started new nests and had to dig their own burrows. Similarly for Augochloropsis sparsilis (Lepeletier) it was deduced by Michener and Lange (1959) that much of the mandibular wear results from cell construction. If this is true, the queens of L. inconspicuum probably are active in cell construction even when workers are present in the nest, since the queens mandibles become more and more worn. Division of Labor Among Queens in Polygynous Nests: In sum- mer, when workers are present in the nests, we have no evidence of division of labor among queens in nests containing more than one queen. The degree of ovarian development of one queen in a nest has no obvious relation to that of another so that there may be, for example, in a digynous nest, two queens of group A, two of group B, or one A and one B. In spring, when there are no workers, the queens in polygynous nests sometimes all have enlarged ovaries and, as Figure 18 indicates, all presumably lay eggs. Figure 18, however, is based on nests opened in May and especially in the later half of May; i. e., on nests in which the production of new cells by the queens has stopped. Earlier in the spring most of the polygynous nests contain one or more “queens” that have slender ovaries (group D) or occasionally ovaries suggestive of those of workers of group C. The fate of these overwintered individuals which are presumably potential queens is not clear. Presumably at least some of them BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1193 attain full ovarian development later. Some of them may leave their overwintering nests to establish their own new nests. However, we have repeated records of such bees collecting pollen and carrying pollen loads to their nest. Clearly in this case they are, for the time at least, functioning like workers in that they are provisioning cells in which other bees (queens of groups A and B) will lay eggs. It is probable also that it is such bees that guard the nests and make the smooth areas around the entrances since these are activities of work- ers in summer. It is therefore quite possible that some of the worker- like but fertilized overwintered females are short-lived like workers. Therefore their disappearance in late spring may result either from their death or from further ovarian development. To give more completely our data on the ovarian conditions of females in polygynous nests opened in April and the first half of May, the conditions of females from such nests are indicated below in terms of groups (as defined in the section entitled “Caste Differ- ences’), those from each nest being enclosed in separate paren- theses: (AB) (AAB) (BB) (ABD) (ABD) (ABD) (AAD) (AABBDD) (BDD) (BD) (ADD) (AAAAD) (AAAD). Lone females were also occasionally in group D but the great majority of lone individuals were in groups A or B, so that most of the indi- viduals of group D found in the spring were in the relatively few polygynous nests. Behavior of Workers: As can be seen from Table X, the average index of wear of workers is much less than that of queens. This indicates that they do much less excavation or cell making per indi- vidual than do queens. Workers of group C have higher average indices of wear than those of group E, showing that they work more or longer in the nest than do those of group E. Individual workers of either group, however, have indices of wear as high as any queen, indicating that occasional workers do much work in the nests. In the course of studying the duration of pollen-collecting trips by marked workers during the summer, data were also obtained on survival and other behavioral characteristics of these workers. In addition to marking foraging bees, we aspirated other bees from the nest entrances and marked them. We learned that, unlike foragers, such bees, when released, often could not find their way back to their nest. We concluded that they probably had not been out of the nest and had not learned the landmarks that would permit them to relocate their nest, and we developed a technique for re- introducing them to their nest through a glass tube after marking. 1194 Tue UNrversiry SCIENCE BULLETIN After all the bees that we could get from a nest had been marked, new ones usually appeared on succeeding days. These new workers were consistently found at the nest entrance or were sucked from the nest with an aspirator, but were not foragers. From this we concluded that the workers pass through a preforaging stage when they do not go out of the nest. During this stage they act as guards or move about in the burrow below the entrance. Bees examined in this stage have unworn mandibles, indicating that they do little or no excavating. The duration of the preforaging stage doubtless varies with temperature and other conditions. We have one record of a marked bee remaining in this stage for six days, but most bees marked as preforagers were foraging within four or five days or less; of course there is no certainty that they were captured and marked as soon as they might have been. However, in artificial glass nests in which pupae were placed, workers acted as guards during the same day that they left their natal cells. In two experiments involving five bees, presumably young guards were transferred artificially to other nests and lived the rest of their lives there. The marked transferred preforagers were first seen as guards, later becoming foragers, and behaved throughout exactly as though they were in the proper nests. Foragers transferred to the wrong nests are not attacked so far as we know but leave the new nest and return to the proper one in a few minutes. Several observations suggest that when the preforagers first leave the nest, they make feeding, not pollen-collecting, trips of long dura- tion which are preceded by unusually long and elaborate orientation flights. We have numerous records of the pollen-collecting activities of individual foragers lasting for eleven and twelve days, and one for fifteen days. Mandibular wear occurs during this stage, indicating that nest construction and foraging are more or less synchronous. The data presented above suggest that after the worker bee leaves its natal cell it may live for about three weeks. We do have a record of one individual that was still in its nest 32 days after marking, but for at least the last ten days of this period it was the only worker in the nest and so far as we know did not leave the nest. It might well have lived longer than more active workers. Guarding and Other Activities at the Nest Entrance: One of the noteworthy activities of worker bees (and queens in polygynous nests in the spring) is guarding. Such activity is characteristic of most or all halictine bees in which more than one individual occurs BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1195 in the same nest and is well developed in L. inconspicuum. Except at night and on unusually rainy days, each nest during the summer months is plugged most of the time by the head of a worker bee, which fits rather neatly into the constriction of the nest entrance. Disturbance of the guard by a fiber or grass blade, or by natural enemies such as mutillids or Paralictus, usually causes the guard to strike out repeatedly with open jaws at the offending object. If the disturbance continues, the guard bee turns and plugs the nest en- trance firmly with the posterior two-thirds of the dorsum of the abdomen. The guard may remain in this position for a long time if the disturbance continues, for example, if a mutillid is trying to dislodge the guard. The guard braces herself very firmly in this position and one of the convenient ways to mark preforaging bees is to paint the marks on their abdomens when they hold this position. Of course when the guard is in this position, foraging bees can neither enter nor leave the nest. Ordinarily, however, when the guard is in the usual position with head up and face flush with the soil surface, a bee flying toward the nest entrance appears to provide a stimulus that causes the guard to retreat, just before the returning bee lands, into the broader part of the burrow below the nest entrance. If two or more nests are close together, returning bees sometimes enter the wrong nest. So far as we know, no fighting occurs in such cases and the guard seems to show no ability to exclude foreign individuals of L. inconspicuum, which however soon leave the nest, as though they recognized their error. Guards have been seen to fight off Paralictus, possibly be- cause they approached hesitatingly rather than rapidly like the L. inconspicuum. Guards also withdraw into the wider part of the burrow to allow departure of bees leaving the nest. The aggressiveness of guards in defending the entrance clearly depends in part on their age. Young guards (e. g., unmarked ones that appeared in nests after all other workers had been marked ) are timid and often descend into the nest without blocking it with the abdomen. Guards with a day or two of experience are much more effective. Occasionally after repeated disturbance a guard pushes earth up from beneath with the end of her abdomen and plugs the entrance. We noted particularly that this often happens when heavy rain falls on the nest. During the middle parts of hot sunny days, guards are often not evident. A small disturbance at the nest entrances will bring them 1196 THe University SCIENCE BULLETIN to the surface at such times, indicating that they are only a few millimeters below the entrances. Temporary shading of the nests by clouds or artificial means will also cause them to resume their positions at the surface on such days. It was found that if a ther- mometer resting on the soil surface registered above 125° F., the guards would not be visible at the nest entrances, but that if clouds or other shadows caused the temperature to drop to 120° F., the guards reappeared with their faces flush with the ground surface. Guarding, as indicated previously, is largely an activity of pre- foraging workers although we have numerous records of a foraging bee spending some time as a guard, and as indicated previously, even queens may sometimes do so in the spring. Even a lone queen sometimes comes to the nest entrance at a small disturbance and may then turn and block the entrance with her abdomen, although such a queen never otherwise takes a position at the entrance as a guard. In nests containing several preforaging workers, each may take a turn as guard. We have seen such guards (marked for indi- vidual recognition ) remaining at the nest entrance for periods rang- ing from 30 seconds to four hours, but perhaps one half to one hour is normal, after which the guard is replaced by another individual. We were sometimes able to glimpse another bee below the guard when the latter allowed a forager to pass and were often able to suck with an aspirator a surprising number of bees from the nest entrance. To investigate the activities of bees below the nest en- trances, we on various occasions and at various seasons dug a large excavation at the side of a nest and either put a piece of glass in such a position that it formed a window in the side of the upper two to four centimeters of the burrow or replaced the upper two to twenty centimeters of the burrow with a glass tube of the same diameter as the burrow. The bees came to and went from such modified nests in a normal manner providing the nest entrance proper was not destroyed and the soil on one side of the entrance was undisturbed. The observations described below were made in such situations but we have many bits of evidence indicating that they represent normal behavior. The bees in such nests were marked with colored paint for individual recognition. By making our excavation large enough to contain the observer's body, it was possible to place the eye or a lens very close to the bees themselves. Light was kept out of the glass walls of the burrows except when observations were being made. At a nest entrance when only one bee was acting as guard, the BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1197 return of a bee from the field caused the activities shown in Figure 34. The departure of a bee to the field caused the activities shown in Figure 85. The turning of the guard shown in Figure 35 is invariable, so far as we know; bees always pass one another in the aa xe (RAR PA aA AAA 37 Ficure 34. Diagrams showing how a guard draws into the burrow, behind the narrow entrance, to allow a returning bee to pass. The guard is marked with black on the abdomen. ee FicureE 35. Diagrams showing how a foraging bee’ passes the guard when leaving the nest. Note that the guard backs into the burrow, away from the narrow entrance, and then turns oyer so that the bees face one another when passing. The guard is marked with black on the abdomen. Ficure 36. Diagrams showing how two guard bees exchange positions. One of them is marked with black on the abdomen. FicurE 37. Diagrams showing activities of three guard bees ( differentially marked for easy recognition) during five minutes of observation. No foragers left or returned during this period. Note that the bees sometimes turn over for unknown reasons other than to allow passing. 41—5840 1198 Tue UNrversiry SCIENCE BULLETIN burrow when facing toward one another and not when facing in the same direction. In most active nests which we studied in summer there were two or more bees spending much of their time at or near the nest en- trance. If the guard were removed artificially, another bee usually replaced it immediately, for it had only to ascend to the nest en- trance from a few millimeters below. The same thing can be repeated several times in nests with large populations of young workers; after the supply of young workers is exhausted it seems to take longer for the older workers to find the unguarded nest entrance and take up positions as guards. In view of the above it was not surprising to find that several workers (usually young ones) often are in the upper part of the nest burrow. They sometimes exchange positions as guards, as indicated in Figure 36; again they pass only when facing one another. Often there is much moving up and down the burrow by bees below the guard, and a bee often turns over and over, summersault fashion, even when not passing another bee. Figure 37 shows the activities of three bees in the upper part of the nest burrow; the entire series of sketches represents a period of about five minutes of observation. Of course when a bee that is leaving the nest comes up the burrow, each of the bees above it must turn over in order to pass, as is shown for a single bee in Figure 35. Co-operative Activity Among Workers: Division of labor among workers of various ages has already been described, younger indi- viduals remaining in the nest and often acting as guards, older individuals serving as foragers, and sometimes also as guards. The present section presents evidence on co-operative activity among foraging workers. It is, of course, possible that each worker makes and provisions its own cells; illustrations such as Figure 14 show that each worker does not make its own branch burrow and group of cells, for there are not enough such groups. Alternatively, the workers might co-operate in making and provisioning cells, several working on one cell simultaneously. Tentatively, we believe that such co-operative activity occurs, at least in provisioning. The number of pollen-collecting bees in a nest appears to consistently exceed the number of cells being provisioned. For example, one nest, watched continually in the morning hours, yielded four pollen collectors; on excavation in the afternoon of the same day only one partly-provisioned cell was found, plus two empty cells which the pollen collectors might theoretically have started to BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1199 provision with the pollen that they were carrying when captured. Another nest yielded seven pollen collectors; when excavated the same afternoon three partially-provisioned and four empty cells were found. A third nest yielded twelve pollen collectors; when excavated the same day two partially provisioned cells, five empty cells and two complete pollen balls in open cells without eggs were found. Such data seems to leave no doubt of co-operative activity in provisioning cells, but unfortunately we often felt that because of the extent of the nest, small size of the burrows, and often poor soil consistency, we might have missed a few cells. We do not believe that such was regularly the case, however. A further effort to prove more decisively that workers co-operate in provisioning cells was made by marking pollen loads of returning foragers with colored powders, using different colors on different foragers which were marked for individual recognition. This proved to be a difficult technique for the foragers were much disturbed if captured; instead we blew the powder onto them just before they entered the burrow. We did find differently colored particles mixed in single pollen masses but always in such proportion that one greatly predominated and the others might have been introduced acci- dentally by bees that became contaminated in the burrow with powder lost from the bodies of other bees. We believe that the method may have some value for use with other species but regard it as inconclusive for such small and inconvenient forms as L. incon- spicuum. NEED FOR FURTHER STUDY To the authors an impressive part of this study has been the enormous areas of investigation in which little or nothing is known (for this bee or for any other) in spite of a fairly intensive study. Some of these matters can be solved by additional field investiga- tions. For example, whatever combination of temperature, light intensity, wind, and perhaps other factors determines when bees fly and when they do not could be elucidated by more field data, as could flight ranges, soil moisture requirements for survival of a colony, etc. However, for investigation of many of the more basic aspects of social organization, our techniques have enabled us to do little more than hint at possible solutions to the problems. It is urgent that techniques be developed so that actual behavior in the nests can be seen rather than merely inferred. We need to know more about integration of members of the colony, the food sources of the queens 1200 Tue UNIversiry SCIENCE BULLETIN in matrifilial colonies (her own trips to the field, or food brought in by workers), the longevity of queens, the interactions among work- ers which are jointly provisioning cells, etc. Obviously many interesting ethological observations could also be made if laboratory colonies could be established. Another field in which more information is needed concerns the question of why some females become long-lived, sexually functional queens while others become short-lived workers. In the better known social insects the answers to this and related questions are partly known, but in primitively social insects such as L. incon- Ficures 38 to 40. Nesting places. Figure 38, Sycamore Slope; Figure 39, Prairie Road; Figure 40, Intersection. Ficure 41. Nest entrance, showing smooth depressed area. Ficure 42. Horizontal section of a cell. The scale at the bottom of the figure is in millimeters. Ficure 43. Pupae in cells; feces (pale) on upper distal surfaces of cells. Ficure 44. Prepupae in cells, The cut through both cells is oblique so that the feces scarcely show and the cell pl fed nly a littl loose dirt at the right-hand end of each cell ieee ts ko 92 BIONOMICS OF A PRIMITIVELY SOCIAL BEE 1201 spicuum, they are quite unknown. Since caste is related to size, and size seems to be related to season (perhaps to temperature, prob- ably not to kind of food as various acceptable pollen sources are available at all seasons), it would be interesting to rear bees under various temperature conditions (as well as with various amounts of food) and observe. the behavior of the resulting adults. On the other hand, since workerlike activity can arise among presumably potential queens in polygynous spring nests, it would appear that caste determination can occur in the adult stage (as has been sug- gested also for L. rhytidophorum (Michener and Lange, 1958). If this is true, an interaction of some sort among females must be postu- lated, with pheromones or some other factor determining ovarian development, mating, longevity, etc. This seems unlikely among insects which seem to have so little contact with one another as members of a colony of Lasioglossum. Obviously an experimental approach to such problems is needed. We have good hopes that it will soon be possible to colonize this or other halictine bees in the laboratory and that having achieved this, some of the questions enumerated above can be answered. LITERATURE CITED BONELLI, BRUNO 1948. Osservazioni biologiche sull’ “Halictus malachurus” K. Boll. In- stituto Entom, Univ. Bologna, 17:22-42, BritTAIN, W. H. 1933. Apple pollination studies in the Annapolis Valley, N.S., Canada, 1928-1932. Dominion of Canada Dept. of Agric. Bull., 162:1-198. CLAUDE-JOSEPH, F. 1926. Recherches biologique sur les Hyménoptéres du Chili. Ann. Sci. Nat., Zool., (10) 9:113-268. LinsLey, E. G., J. W. MacSwarn, and R. F. SmirH 1952. Outline for ecological life histories of solitary and semi-social bees. Ecology, 33:558-567. MELANDER, AXEL LEONARD, and CHARLES THOMAS BRUES 1903. Guests and parasites of the burrowing bee Halictus. Biol. Bull., 5:1-27. MICHENER, CHARLES D. 1946. Notes on the habits of some Panamanian stingless bees (Hymen- optera, Apidae). Jour. New York Ent. Soc., 54:179-197. 1951. Halictidae, pp. 1104-1134, in Muesebeck, C. F. W., Karl V. Krom- bein and Henry K. Townes, Hymenoptera of America North of Mexico. U.S. Dept. Agr. Monograph 2:1-1420. 1953. The biology of a leafcutter bee (Megachile brevis) and its asso- ciates. Univ. Kansas Sci. Bull., 85:1659-1748. 1958. The evolution of social behavior in bees. Proc. Tenth Internat. Congress Entom., 2:441-447. 1202 Tue UNIVERSITY SCIENCE BULLETIN MICHENER, CHARLES D., Earte A. Cross, HoweLtt V. Daty, Cart W. RETTENMEYER and ALVARO WILLIE 1955. Additional techniques for studying the behavior of wild bees. In- sectes Sociaux, 2:237-246. MicHENER, CHARLEs D., and Cart W. RETTENMEYER 1956. The ethology of Andrena erythronii with comparative data on other species (Hymenoptera, Andrenidae). Univ. Kansas Sci. Bull., 37: 645-684. MICHENER, CHARLES D., and Rupotr B. LANGE 1958. Observations on the behavior of Brasilian halictid bees, V, Chlora- lictus. Insectes Sociaux, 5:379-407. 1959. Observations on the behavior of Brazilian halictid bees (Hymen- optera, Apoidea) IV Augochloropsis, with notes on extralimital forms. Amer. Mus. Novitates, 1924:1-41. MICHENER, CHARLEs D., and C. A.C, SEABRA 1959. Observations on the behavior of Brasilian halictid bees, VI, tropical species. Jour, Kansas Ent. Soc., 43:19-24. Noi, JosEF 1931. Untersuchungen iiber die Zeiigung und Staatenbildung des Halictus malachurus Kirby. Zeitschr. f. Morph. u. Okol d. Tiere, 23:285-367, pls. I-III. Pearson, E. E., and H. O. HartTLey 1956. Biometrika tables for statisticians. Vol. 1, xiv-+- 238 pp., Cambridge Univ. Press. PLATEAUX-QUENU, CECILE 1959. Un nouveau type de société d’insectes: Halictus marginatus Brullé (Hym., Apoidea). Année Biol., 35:325-444, pls. I-IX. SAKAGAMI, SHOIcHI F., and CHARLEs D. MICHENER In press. The nest architecture of halictine bees (Hymenoptera, Apoidea). SAKAGAMI, SHOr1cHr F., and Kazuo HayAsHIpA In press. Biology of the primitive social bee, Halictus duplex Dalla Torre, II, Nest structure and immature stages. SNEDECOR, GEORGE W. 1956. Statistical methods, xiii+ 534 pp., 6th ed., Iowa State College Press, Ames, Iowa. Lib WINNT i S WHSE 04548 —_ Se - - — = ae _— ~* SE IONS FEES AL oes : = i - a one Ds a Se Tonge om - “oe a = 0 Oates EE) = ne em — ene Se rr — = = = —_ eet Sg enter OL EP PP POL