O •" fw I HARVARD UNIVERSITY m m LIBRARY OF THE Museum of Comparative Zoology UNIVERSITY OF KANSAS SCIENCE BULLETIN mi. CCMP. ZOQL LSSRARY SEP 2 4 1953 UNIVERSITY OF KANSAS PUBLICATIONS University of Kansas Science Bulletin - Vol. XXXV - Part II September 10, 1953 Lawrence, Kansas ANNOUNCEMENT The University of Kansas Science Bulletin (continuation of the Kansas University Quarterly) is issued in parts at irregular inter- vals. Each volume contains from 600 to 1,000 pages of reading matter, with necessary illustrations. Exchanges with other institu- tions and learned societies everywhere are solicited. All exchanges should be addressed to The University of Kansas Science Bulletin, Library of the University of Kansas, Lawrence, Kan. Editor Edward H. Taylor Editorial Board Paul G. Roofe, Chairman Frank E. Hoecker H. B. HUNGERFORD Arthur J. Mix PUBLICATION DATES The actual date of publication ( i. e., mailing date ) of many of the volumes of the University of Kansas Science Bulletin differs so markedly from the dates bourne on the covers of the publication or on the covers of the separata that it seems wise to offer a corrected list showing the mailing date. The editor has been unable to verify mailing dates earlier than 1932. Separata were issued at the same time as the whole volume. Vol. XX— October 1, 1932. Vol. XXI— November 27, 1934. Vol. XXII— November 15, 1935. Vol. XXIII— August 15, 1936. Vol. XXIV— February 16, 1938. Vol. XXV— July 10, 1939. Vol. XXVI— November 27, 1940. Vol. XXVII, Pt. I— Dec. 30, 1941. Vol. XXVIII, Pt. I— May 15, 1942; Vol. XXVIII, Pt. II— Nov. 12, 1942. Vol. XXIX, Pt. I— July 15, 1943; Vol. XXIX, Pt. II Vol. XXX, Pt. I- Vol. XXX, Pt. II- Vol. XXXI, Pt. I- Vol. XXXI, Pt. II Vol. XXXII— Nov. Vol. XXXIII, Pt. I— Vol. XXXIII, Pt. II Vol. XXXIV, Pt. I- Vol. XXXIV, Pt. II Vol. XXXV, Pt. I- — Oct. 15, 1943. -June 12, 1944. -June 15, 1945. •May 1, 1946. -Nov. 1, 1947. 25, 1948. April 20, 1949. -Mar. 20, 1950. Oct. 1, 1951. -Feb. 15, 1952. July 1, 1952 UNIVERSITY OF KANSAS SCIENCE BULLETIN DEVOTED TO THE PUBLICATION OF THE RESULTS OF RESEARCH BY MEMBERS OF THE UNIVERSITY OF KANSAS Volume XXXV, Part II University of Kansas Purlications Lawrence, Septemrer 10, 1953 PRINTED BY FERD VOILAND. JR.. STATE PRINTER TOPEKA. KANSAS 1953 24-6490 S- 4 U Contents of Volume XXXV, Part II No. PAGE 6. The Effect of Radiation from Intraperitoneally Injected Radium Chloride upon the Hemopoietic Activity of the Bone Marrow in Albino Rats. Plates LXVIII to LXXII. Paul G. Roofe and Hal G. Bingham, 945 7. Bees of the Genus Perdita in the Collection of the Uni- versity of Kansas ( Hymenoptera, Apoidea). P. H. Timherlake, 961 8. Comparative Morphological and Systematic Studies of Bee Larvae with a Key to the Families of Hymenopterous Larvae. Text figs. 1 to 287 Charles D. Michener, 987 9. A Revision of the North American Species of Typhlocyba and its Allies ( Homoptera, Cicadellidae ) . Plates LXXIII to XCII Paul G. Christian, 1103 10. The Ambrysus of Mexico (Hemiptera, Naucoridae). Plates XCIII to C Ira La Rivers, 1279 11. A Revision of the Genus Buenoa (Hemiptera, Notonec- tidae). Plates CI to CXVII Fred S. Truxal, 1351 12. A Review of the Lizards of Ceylon . Edward H. Taylor, 1525 13. Fourth Contribution to the Herpetology of San Luis Potosi. Plates CXVIII-CXIX, map Edward H. Taylor, 1587 14. Report on a Collection of Ceylonese Serpents. Edward H. Taylor, 1615 15. Frogs of the Family Centrolenidae from Brazil. Edward H. Taylor and Doris M. Cochran, 1625 ^sTcoMpTzoon f : LIBRARY \*W 2 4 1953 THE UNIVERSITY OP KANSAS SCIENCE BULLETIN Vol. XXXV, Pt. II] September 10, 1953 [No. 6 The Effect of Radiation from Intraperitoneally Injected Radium Chloride Upon the Hemopoietic Activity of the Bone Marrow in Albino Rats * Paul G. Roofe and Hal G. Bingham f Department of Anatomy, University of Kansas, Lawrence INTRODUCTION In the major portion of experiments in the past, radiation of the bone marrow has been accomplished by an external source, X-rays or gamma-rays in most cases. The literature relating to the effect of roentgen rays on the blood-forming organs is extensive and con- tradictory. The lack of unanimity among various reports may be attributed in part to differences in the experimental conditions in regard to radiation technique, the experimental animals used, and the time interval between radiation and examination of tissues. Previous investigations of the effect of ionizing radiation on bone marrow, using radium as the internal source of radiation, were con- ducted over varying periods of time from a minimum of a few months to a maximum of several years duration. The results ob- tained after a few months of low dosage internal radiation are gen- eral hyperplasia of the erythrocytic tissue (Sabin, Doan, and Fork- ner, '32). Continued exposure over a longer period of time produces hyperplasia of granulocytic tissue and hypoplasia of the erythrocytic tissue ( Rosenthal and Grace, '36 ) . After an even longer time, however, radium causes the marrow to lose its ability to pro- duce either erythrogenic or granulogenic tissue (Martland, '31). The present experiment is an investigation of the effects of radium chloride on the hemopoietic activity of the cells of the bone marrow after exposure of short duration. In addition the effects of varying dosage are reported. ° This investigation was carried out under contract with the Office of the Surgeon Gen- eral, U. S. Army, Research Contract, No. W49-007-MD 466. f Paul G. Roofe, Hal G. Bingham, Department of Anatomy, University of Kansas, Law- rence, Kansas. (945) 946 The University Science Bulletin Following intraperitoneal injection of radium, elimination is rapid during the first three days. The rate decreases and becomes es- sentially zero after fifteen to twenty days. Since the alpha particle penetrates about 25[x in tissue, the maximum damage occurs during the first two or three days when the radium is still found in the circulating blood. Approximately thirty percent of the injected radium chloride is permanently retained in the bone matrix proper (Hoecker and Roofe, '49). When radioactive material is fixed in the bone, it is eliminated very slowly and the major portion remains as a source of alpha particle radiation. METHODS AND MATERIALS Thirty-two Sprague-Dawley male albino rats were used. All animals used were young adults and approximately the same weight. Seven of these were untreated controls to establish the normal bone marrow picture. Twenty-five animals were divided into five groups of five. The first group received lOjxg of intraperitoneally injected RaCl2 and each successive group received an increase of l(fy.g, so that groups receiving 10, 20, 30, 40, and 50j;.g of RaCL were ob- tained. An animal from each group was sacrificed on the third day after injection; and one from each group was sacrificed at three-day intervals thereafter. After the rats were anesthetized with sodium nembutal the femurs were quickly removed. The method of Mayer and Ruzicka ('45) was found to yield well-preserved pencils of marrow without pre- vious decalcification. The marrow was fixed in Zenker-formol, im- bedded in paraffin and sectioned serially at 8 micra. Sections were stained with hematoxylin-eosin-azuren • The bone marrow cell count was made with the aid of a camera lucida which projected a calibrated area onto especially printed sheets for recording the various cell types. The results of these counts were converted into number of cells per mm3 to make com- parison possible with other hemopoietic organs and peripheral blood. Twelve sections were chosen from each tissue so that the entire bone marrow was adequately sampled. Six fields or foci of cell concentration were counted on each of these bone marrow sec- tions. An eosinophil field, a neutrophil field, an erythroid field, and two myeloiderythroid fields 50 X 50 micra were counted under oil immersion. A megakaryocyte field 500 X 500 micra was counted under low power. In most cases the Maximow and Bloom ('42) terminology and method of cell identification were used. The various cell types in the neutrophil and eosinophil series were readily identified by the Effect of Radium on Bone Marrow 947 presence of their characteristic cytoplasmic granules. In addition the myeloid and erythroid series were not difficult to distinguish. However, the maturing cell forms within these series were difficult to distinguish from each other. To avoid this problem the pro- myelocyte and myelocyte forms of the myeloid series were grouped together as immature cells and the metamyelocyte and adult cells were counted as separate cell types. In the erythroid series two cell forms were distinguished, the erythroblast and the normoblast. The term erythroblast includes all the immature erythroid elements up to the cell form containing hemoglobin in its cytoplasm. The term normoblast includes all those cells with hemoglobin in their cytoplasm and maturing cell forms up to the adult red blood cell. The other bone marrow cells were also identified and named by the Maximow and Bloom ('42) terminology. However, plasma cells, lymphocytes and basophils were found with such low fre- quency that they proved to be of little consequence in this study and were not included in the results. In order to determine what the effects of the radiation were on the peripheral blood of the experimental animals, counts were made on the peripheral blood by other workers in this laboratory. The blood for the peripheral count was taken from the tail of the rats just before sacrifice and standard methods of counting were carried out. RESULTS Table I shows the mean and range of each hemopoietic cell type in mm3 found in both the non-injected control animals and the 25 experimental. The 25 experimental animals are represented in the table as five groups of five animals. The mean and the range of each group pertain to those five animals sacrificed on the same day. Each animal in the group of five received a different dose of RaCl2 ranging from 10[xg to 50>g. Statistical analysis of the material in Table I is not included in this paper. Superimposed on the variation due to the smallness of the sample and due to the difficulty in identifying the cells, was the great variability of the individual animals in response to the treat- ment. As a consequence, the range of the cell counts is great in many cases. The data were plotted in the following manner: In each graph the ordinate represents the number of cells and the absissa repre- sents time. The points on the vertical lines represent counts for individual animals with different doses killed on the same day. The curve represents the mean cell count of each day. 948 The University Science Bulletin A. Granulocyte Series As can be seen in Table I, the counts of the immature neutrophils for the experimental animals fall within the normal range. Because of the extreme variation in both control and experimental animals, it is not possible to tell if any change has occurred. On the other hand, the metamyelocyte and adult cell forms exhibit an incidence above normal with increase in time ( Figures 1 and 2 ) . The trend of the count of the three cell types in the eosinophil series is to increase above normal as is shown in figures 3, 4, and 5. One of the 12 day animals gave exceptionally high counts for each of the cell types. These three values are indicated on the graphs, and were also included in the calculations of the means. All eosino- phil cell types demonstrate a trend similar to that of the develop- ing neutrophil cells in that they generally increase in number with time. Furthermore the adult forms of both series appear to reach a maximum at about the 12th day. B. Erythroid Series The cell counts for both erythroblasts and normoblasts drop be- low normal initially, then rise back to normal by the 14th day ( Fig- ures 6 and 7). C. Other Bone Marrow Cells Reticular cells. The cell count shows an initial rise, with a return to normal by the 9th to 12th days (Fig. 10). On the 14th day the count is again above normal. Hemocytoblasts. It can be seen in figure 11, that the cell count is generally above normal and tends to increase with longer time. Megakaryocytes. The cell number decreases reaching minimum values on the 9th and 12th days but returns to normal by the 14th day (Fig. 12). D. Peripheral Blood Megakaryocytes are compared with peripheral platelets in figure 12. The platelets show the same trend as the megakaryocytes but demonstrate the initial decrease and subsequent increase in count later than the megakaryocytes. It appears then that the platelet curve shows about a 3 day time lag when compared with the mega- karyocyte curve. This observation has been published in a paper by Roofe, Bingham, and Comer ('51) and is discussed more ex- tensively there. The peripheral polymorphonuclear leukocyte count is consider- ably decreased from normal and remains that way for the duration of the experiment. Effect of Radium on Bone Marrow 949 The peripheral red blood cell count did not show much variation from normal. E. Dosage Effects In general the effects with increasing dose were comparable to the changes obtained with increase in time. However, it was evi- dent that in most cases time was the important factor in determining the response to the injected radium. The one apparent exception to this was the case of the erythro- blasts and normoblasts. As figures 8 and 9 show, the number of erythroblasts was lowest with the smaller dose, while with increas- ing dose the counts tended to be more in the normal range. How- ever, the normoblast count decreases with increase in dose. (The normoblast curve is included to show the more expected trend of cell count with increased dose. ) DISCUSSION In the neutrophil series an increase in the metamyelocyte and adult cell forms was observed. This observation suggests that the rate of maturation of the earlier forms to the more mature forms may be accelerated. Lingley, Gall, and Hilcken ('40) also ob- served a similar trend in their studies with roentgen radiation. However, Martland ( '31 ) and Rosenthal and Grace ( '36 ) , in experi- ments of longer duration than those discussed here, found that with radiation from radium the immature cells constituted the majority of the neutrophils. As has already been shown it is impossible to determine what, if any, trend has been established in the immature cells in this experiment. In the case of the eosinophils, an accelerated maturation of im- mature cells forms similar to that in the neutrophil series is sug- gested by the results. Also an increased proliferation of the im- mature cell forms probably occurred, since all cell forms of this series increased to above normal incidence with increase in time and in dosage. Isaacs ('32) also found an increase in the number of eosinophils following roentgen radiation, as did Rosenthal and Grace ('36) with radium. Martland ('31) found that the only cell of the granulocytic series which proceeded to maturation was the eosinophilic myelocytes. The observed increase in the number of immature cells in the eosinophil series is in disagreement with the observations of Osgood ('42) and Osgood, Albersold, Erf, and Packham ( '42 ) , who found that roentgen radiation inhibited mitotic division of all progranulocytes, which would be expected to de- crease the production of all cell forms. However in this case larger dosages and longer duration were involved. 950 The University Science Bulletin The peripheral polymorphonuclear leucocyte count dropped con- siderably below normal and remained there throughout the experi- ment. This suggests that some mechanism associated with irradia- tion causes the mature granulocytes to be retained in the bone marrow. Dunlop ( '42 ) , who worked with roentgen rays, also found an increased bone marrow neutrophil count with a concomittant decreased peripheral polymorphonuclear leucocyte count. Mart- land ('31) and Jacobson, Marks, and Lorenz ('49), using radium, observed a reduction in leucocytes in the peripheral blood and Martland ('31) also reported a decrease in the bone marrow neu- trophil count. Both the erythroblast and normoblast show an initial decrease with a return to the normal range at a later time. Rosenthal and Grace ( '36 ) found that continued action of radium inhibited matu- ration of the erythroblast as evidenced by a slight increase in num- ber of erythroblasts despite a decided decrease in the total marrow erythrocytic elements. In the X-ray studies of Lingley, Gall, and Hilcken ('40), the lowered number of normoblasts was explained as due to an accelerated maturation of this cell into adult red blood cells. The lack of appreciable variation of the peripheral red blood cell count from normal is in agreement with the radon studies of Nemenow and Gurewitsch ( '34 ) . On the other hand Dunlop ( '42 ) using roentgen rays and Jacobson, Marks and Lorenz ('49) using RaCl2 found a peripheral anemic condition after a longer period of irradiation. In regard to dosage the erythroid series showed a different kind of response than did the other bone marrow cells. Although the normoblast cell count is lower in animals that received greater doses, this was not the case with the erythroblasts. The greatest decrease in the number of erythroblasts was found in the animals receiving the lowest dose of radium. Interpretation of this phenomen is be- yond the scope of the present experiment. In the literature the myeloid-erythroid ratio is frequently used as an index of change occurring in the cells of the bone marrow. The value for normal animals found in this experiment was 0.739, which falls in the normal range reported in the literature, 0.500 (Brown, '51)-1.940 (Endicott and Ott, '45). The myeloid-erythroid ratio for the experimental animals was increased to 0.910 on the third day, and to 1.020 on the fourteenth day. Cellular processes of prolifera- tion and maturation were increased more in the myeloid tissue, as has already been shown, leading to a cell count above normal for both the eosinophil and neutrophil series. In the erythroid tissue, Effect of Radium on Bone Marrow 951 however, there was an initial decrease in both cell types with a return to normal with longer time. This observation is not in agree- ment either with Bauer ('40) or Chien-Liang and Ma ('40) who both worked with roentgen radiation. They found a hyperplasia of the erythroid cells and a hypoplasia of the myeloid cells. However, in both of these cases a larger dose of radiation was employed. Both the hemocytoblast and reticular cell counts were increased, which is in agreement with the findings of Lingley, Gall and Hilcken ( '40 ) . They found that primitive cells in general proliferated more rapidly in animals subjected to roentgen radiation. Warren, Mac- Millian, and Dixon ('50) using P32 as the internal source of radia- tion found that the stem cells were not injured at any time during their experiment. The number of megakaryocytes was found to be below normal and remained that way throughout the duration of the experiment. Dunlop ('42) working with X-rays and Warren, MacMillian and Dixon ('50) using P32 also found that megakaryocytes were reduced in number. The relationship between the megakaryocyte and peri- pheral platelet curves is in agreement with the theory that the mega- karyocytes fragment to become platelets. Three days pass before the peripheral platelets show any change in number resulting from the initial megakaryocyte decrease, indicating that they possibly survive this length of time in the peripheral blood after entering it. The newly produced megakaryocytes apparently do not fragment into platelets for approximately three days after they are produced because the peripheral platelet curve does not show a similar in- crease in number until this period of time has passed. Jacob- son, Marks, and Lorenz ('49) also noted a reduction in the peri- pheral platelet count after injecting parenterally 0.94tj.g of radium chloride, but did not correlate this with the megakaryocyte changes in the bone marrow. The sum total of the various cell responses to the initial irradia- tion results in a decreased total cell count for the bone marrow and a normal cell count for the peripheral blood. Some bone marrow cells increase while others decrease in number and this is probably due to a change in the rate of proliferation and/or maturation. With increased time the bone marrow cells appear to proliferate and ma- ture rapidly but not to pass into the peripheral circulation, which tends to increase the marrow count and to slightly decrease the peripheral blood count. In contrast to these results, Lingley, Gall, and Hilcken ('40), using roentgen irradiation, observed a secondary or regenerative return of the bone marrow to normal cellular con- 952 The University Science Bulletin tent between the second to fourth weeks, while Warren, MacMillian and Dixon ('50) found that maximum damage occurred at about 10 days with return of myelopoiesis 10 to 15 days after injection of P32. SUMMARY Radium chloride emitting chiefly alpha particles was injected intraperitoneally into the albino rats. The RaCL is eliminated rap- idly from the animals during the first three days after injection so it is mainly found in the circulatory system during that period. Eventually, approximately 30 percent of the injected RaCL becomes deposited in the bone tissue. Both the dosage and, more important, the duration of exposure determined the kind of effect obtained. In general, however, dosage effects and duration effects caused the same type of cell responses. These effects are a change in the rate of proliferation of the immature cells and a change in maturation time of the more mature cells. Considering the effects on the dif- ferent marrow cells separately: 1. Neutrophils and eosinophils. The more mature neutrophils show an acceleration of maturation, but there is neither a decrease nor an increase in the rate of proliferation of the immature forms. The eosinophils follow this same tendency except that their im- mature cells increase in the rate of proliferation. 2. Erythroid cells. Both the normoblast and erythroblast num- ber decreases initially with a return to the normal range at a later time. However, increase in dosage caused the normoblast to de- crease, while the erythroblast decreased initially, it later returned to normal. 3. Myeloid-erythroid ratio. This ratio is increased because the myeloid cells increase above normal and the erythroid cells decrease below normal. 4. Stem cells. The reticular cells and hemocytoblasts show an increased proliferation. 5. Megakaryocytes and platelets. The two curves of these cells are very similar in shape but the platelet curve shows a time lag of about 3 days to the changes occurring in the megakaryocyte curve. This supports the theory that platelets are formed from fragmenting megakaryocytes . 6. Granulocytes and peripheral polymorphonuclear leucocytes, also erythroid cells and peripheral red blood cells. The peripheral blood picture does not indicate adequately the processes occurring in the marrow. The decreased peripheral polymorphonuclear leuco- Effect of Radium on Bone Marrow 953 cyte count gives no indication of the increased granulocyte produc- tion in the marrow, nor does the apparently normal peripheral red blood cell count reflect the initial decrease with later return to nor- mal of the erythroid production in the bone marrow. 7. Total bone marrow and total peripheral blood count. Initially the sum total of the various cell responses to the irradiation causes a total cell decrease in the bone marrow but the total peripheral blood count remains normal. With longer exposure to irradiation the peripheral blood count tends to decrease but the bone marrow count increases. This suggests a greater proliferation and matura- tion of the marrow cells and a mechanism which prevents the pas- sage of these marrow cells into the periphery. 954 The University Science Bulletin C > t~ ■# a> 03 1-14* ■*>ra rno t^o CM05 OS CM COOS 00 ■ 0-f CO -4 osd P ■CO ■ 4*1 • i-4 ■CD •in — 1 co I ■O 1 O 1 ^ 1 O) 1 CD 1 •* 1 (^ i CD 1 co i 4-> JU -* CO -4 (NO o OS coo hn O -4 00 coo 050 -"f 1-4 1—1 -4 co It) CD iO CM co CD OS >* ■•-> a CM —4 d CM —4 —4 ■* C O -4 CD Si 0) F-! S- CO IO o CD m co m Tt< o 00 o 00 o >> a> lO in O U V o3 osco cot~ -4 CM oo CO-* ood COlO CD • oco om CM • 4-1 D ■ —4 •CO • i — 1 ■ a ■•* ■IN ■CM ■CO ■CO • —4 •t~ a co I in 1 CO 1 •a i OS 1 CO 1 oo 1 ■* 1 CO 1 -4 1 m i ntrol' me i: CM CO CM 00 CM-* oo O TfO -4T)< -4 00 CO ,-4TfH CO OS CM HO m OSO co o —4 co 1 — 1 ■"* -4 P -42 O der erent "* O t^ oo CO CM CD o co >> ■tf O ■* -f co OS CM om CO • cot^ 00 CO CD CM CO -4 m • Or- O -f OS ■ 03 ■ -4 •CM •as ■T)< ■CM t> •4f •m eta Q O 1 co I in I O 1 -4 | 00 1 t~ 1 ■* I -4 | t~ 1 CO 1 ©CO CM CO 00 co oo -4 -4 <* -4 00 CD (NO 00O 00 P;a Ol ■"f o> .-4 CM a> m O CM 00 o CM T3 •o -4 CM -4 (M tU-C 4-1 y .2 CO ^h m o o 2 « J-3 00 lO oo m CO o CO OS co CO 00 m >1 COCO 1-4 ■ •*r^ co oo r^cM (MOT co • oo 0-4 r- • c3 ■ -4 ■CM ■in •lO • —4 ■ -4 ■m ■oo ■ — •■* Ct3 Q < — I 1 —4 | Tf 1 00 1 -4 | 00 1 lO 1 CO 1 O 1 oo I co I om CM 00 in CM to ooo <* CO 1-1 __; co CM CO CD T)J ^J ,_; oo 00 CO t~ f-H -4 -4 CD CO o CO n*' i rfCM coco ■* • CM CM -4^*1 0-4 Oco t~ O-H 0-4 CO P~e S g cd ■CM ■CD ■CO ■in - -4 ■ -4 r~ Q CO 1 O 1 lO 1 co I —4 | t^ 1 ■* 1 ■* i O 1 OS 1 m i 1— 1 CO oco CM 00 CD ooo -4 co CO CO -4 4*< CM CO co ISO CO t^o CM OS CO w t~ •-4 -4 -4 -4 ■* >J pq W Gee .y cmh 4J CD O poie sev nts o a = r* O O C^ o Si « — -G 4-> G-, C H o 4^ >> CJ 4* O CO CO X y. CO 43 X to CD ■""' w g CO X CO o >> co a H 3 CO o "cj >> CO s 3 s co -3 3 O i3 *c3 CJ CO 3 44 >. o O In Table and r mean -— ■*- °^ l-~C -5 4* cu S3 a a 5 4- cc E E CO o 4- a 4- K E £ si 1 -*- i i t- 0 z a pt C 4S > c e E a d ) > u a a t a % 1 0 Effect of Radium on Bone Marrow 955 BIBLIOGRAPHY Bauer, von Robert. 1940. Effects of variously administered roentgen rays on marrow and its cell elements. Strahlentherapie, 67: 424-501. Brown, J. Wm. 1951. The effect of protein deficiency on the hemopoietic organs and the blood of the albino rat. Unpublished manuscript. Chien-Liang, Hsu, and W. C. Ma. 1940. Direct and indirect effects of roentgen radiation on the blood-forming organs of rats. Am. J. Cancer 39: 319-333. Dunlop, C. E. 1942. Effects of radiation on the blood and the hemopoietic tissues. Arch, of Path., 34: 592. Endicott, K. M., and M. Ott. 1945. The normal myelogram in albino rats. Anat. Rec, 92: 61-69. Hoecker, F. E., and P. G. Roofe. 1949. Structural differences in bone matrix associated with metabolized radium. Radiology, 52: 856-864. Isaacs, R. 1932. Maturing effect of roentgen rays on blood-forming cells. Arch. Int. Med., 50: 836-842. Jacobson, L. O., E. K. Marks, and E. Lorenz. 1949. The hematological effects of ionizing radiations. Radiology, 52: 371-395. Lingley, Gall, and Hilcken. 1940. Comparative experimental studies of 200 kilovolt and 1000 kilovolt roentgen rays; biologic effects on bone marrow of albino rats. Am. J. Rath., 16: 845-854. Martland, H. S. 1931. The occurrence of malignancy in radioactive persons Am. J. Cancer, 15: 2435. Maximow, A. A., and W. Bloom. 1942. Testbook of Histology. W. B. Saunders Co., Philadelphia and London, 4th Ed. Mayer, E., and A. Z. Ruzicka. 1945. A method for studying numerical and topographic problems in the whole femoral marrow of rats and guinea pigs, with the use of uncalcified sections. Anat. Rec, 93: 213-231. Nemenow, N., and Gurewitsch. 1934. Effects of intravenous injection of radon on blood. Strahlentherapie, 50: 693-704. Osgood, E. E. 1942. Investigation as to whether or not the action of roentgen rays are direct or indirect on human marrow culture. Am. J. Roentgenol., 35: 786-789. Osgood, E., P. Albersold, L. A. Erf, and E. Packham. 1942. Effects of million volt roentgen rays, 200 kilovolt roentgen rays, radioactive phos- phorus, and neutron rays by marrow culture technique. Am. J. M. Sc, 204: 372-381. Roofe, P. G, H. G. Bingham, and R. Comer. 1951. Effect of radiation from intraperitoneally injected radium chloride upon megakaryocyte and blood platelet production in albino rats. Kans. Acad, of Sci., 54: 391-394. Rosenthal, M., and E. J. Grace. 1936. Bone marrow and lymph node changes in rabbits produced by oral administration of radium sulfate. Am. J. M. Sc, 191: 607. Sabin, F. R., C. A. Doan, and C. E. Forkner. 1932. The production of osteogenic sarcomata and the effects on lymph nodes and bone marrow of intravenous injection of radium chloride and mesothorium in rabbits. J. Ex. Med., 56: 267-289. Warren, S., J. C. MacMdllian, and F. J. Dixon. 1950. Effects of internal irradiation of mice with P32; spleen, lymph nodes, thymus, bone marrow. Radiology, 55: 375-389. 956 The University Science Bulletin PLATE LXVIII NEUTROPHILS Celb/ /mn 160 * ~4 1 n * IO FIGURE 1 ADULTS > 150 - < 140 - i 130 1 m _ / i ^***« 120 i i * * 1 > 1 / ' < no : > ./ ■ 100 90 • • > < i 80 i 1 ■ 70 « " 60 - 50 - < > 401- 30 £0 10 FIGURE I METAMYELOCYTES i i i C 3 6 TIME IN 9 \Z DAYS 14 Effect of Radium on Bone Marrow PLATE LXIX EOSINOPHILS 957 Ceils 3 4- mm* io FIGURE 3 ADULTS v 90.0 60- 55- 50- 45 40 35- 30- £5- 10- 15- 10 2.0 15 10 5 15 10 5 h 0 FIGURE 4 METAMYELOCYTES ■$44.& •• FIGURE 5 IMMATURE5 ® £8.0 i i i i i i C 3 6 9 \Z 14 TIME IN DAYS 958 The University Science Bulletin PLATE LXX ERYTHROID CELLS Cells mri^o5 FIGURE 6 £6(- NORMObLASl Cells mnfxio5 FIGURE 8 ££|- NORMOBLAST 8H 6 4 Z O FIGURE 7 ERYTHR05LAST 8- 6- 4 i i i i i i C 3 6 9 12. 14 TIME IN DAYS FIGURE 9 ERYTHRO&LAST M L I I I I I I C IO LO 30 40 50 DOSAGE IN MICROGRAMS Effect of Radium on Bone Marrow PLATE LXXI STEM CELLS 959 Cells wrNu? FIGURE lO RETICULAR CELLS 113 60(- 70 60 50 40 30 2.0 IO O i I60h 140 l£0- 100- 60 GO 40h to oL FIGURE : li HEM0CYT05LASTS • < i i > < 1 ( I 1 « i 1 1 • i , > i >• i ■ i i i I > i i 1 I < ■' m 1 « i c I 3 31—6490 TIME i i 6 9 IN DAYS i 12. 14 960 The University Science Bulletin PLATE LXXII Celb/mtr? NO — FIGURE 12. ■ • MEGAKARYOCYTf AND PLATELET i ( x ) ( ) t 100 i i i t >• i 1 < i i > i 1 90 > 1 1 pg 60 < < < I i • < 1 ! • : i \ \ i: X < : f \ V 70 ( i • » \ ! c / „ X < i \ / i±J : : \ \ \ \ i / >- 60 - ; l \ i ; 7 ' w ?H \ \ : if •/ K QCLU lULd 50 - \ ) \ A • i i 10 10 i 3 6 TIME IN 9 12. DAYS 14- THE UNIVERSITY OF KANSAS SGIENGE BULLETIN Vol. XXXV, Pt. II] September 10, 1953 [No. 7 Bees of the* Genus Perdita in the Collection of the Uni- versity of Kansas ( Hymenoptera, Apoidea)* BY P. H. TlMBERLAKE, University of California Citrus Experiment Station Riverside, California CONTENTS PAGE Subgenus Xerophasma Coekerell 962 Perdita bequaertiana Cockerel! 962 Subgenus Cockerellula Strand 963 Perdita bidenticauda sp. nov 965 Male 965 Female 966 Perdita laticauda sp. nov 967 Male 967 Female 967 Perdita lobata sp. nov 967 Male 967 Subgenus Macroteropsis Ashmead 968 Perdita latior Coekerell 968 Perdita arcuata dinognatha Coekerell 968 Unnamed group 969 Perdita trifasciata sp. nov 969 Male 969 Unnamed group 970 Perdita ainsliei Crawford 970 Perdita peculiaris sp. nov 970 Male 970 Subgenus Cockerellia Ashmead 971 Key to the species of Cockerellia, females 971 Perdita utahensis Coekerell 974 Male 974 Perdita baileyae Coekerell 975 ° Paper No. 758, University of California Citrus Experiment Station, Riverside, Cali- fornia. (961) 962 The University Science Bulletin PAGE Perdita perpulchra flavidior subsp. nov 975 Female 975 Perdita bequaerti indianensis Cockerell 976 Perdita verbesinae Cockerell 976 Perdita fracticincta sp. nov 976 Female 976 Perdita lacteipennis lacteipennis Swenk and Cockerell 977 Perdita lacteipennis pallidipennis Graenicher 978 Perdita tricincta sp. nov 978 Female 978 Male * 979 Perdita albipennis Cresson 980 Perdita scopata sp. nov 980 Female 981 Male 981 Perdita lepachidis lepachidis Cockerell 982 Female 982 Perdita lepachidis levifrons subsp. nov 983 Female 983 Male 983 Perdita coreopsidis coreopsidis Cockerell 983 Perdita coreopsidis kansensis subsp. nov 984 Female 984 Male 984 Perdita coreopsidis obscurior subsp. nov 984 Female 985 Male 985 Literature cited 985 Abstract: This article is a revisionary study of the bees of the genus Perdita Smith, subgenera Xerophasma, Cockerellula, Macroteropsis, Cock- erellia, and two other un-named groups, in the collection of the University of Kansas. Twenty-one species are treated; of these, eight species are new. Four new subspecies are also described. This is a first report on a large collection of bees of the genus Perdita submitted to the author in the fall of 1950 by Charles Dun- can Michener, and collected mostly in the states of Kansas and Texas. Types of new forms will be found in the collection of the University of Kansas, at Lawrence, Kansas. Subgenus XEROPHASMA Cockerell Perdita bequaertiana Cockerell Xerophasma bequaerti Cockerell, Amer. Mus. Novitates vol. 66, 1923, p. 2, 2 (not Perdita bequaerti Viereck, 1917). Perdita bequaertiana Cockerell, in Muesebeck, Krombein, and Townes, U. S. Dept. Agri. Monog. No. 2, 1951, p. 1087. One male, 11 females, Big Springs, Howard Co., Texas, June 20, 1947 (R. H. Beamer). Timberlake: Bees of the Genus Perdita 963 No flower record accompanies these specimens, but several of the females bear the coarse cobwebby pollen of Oenothera. This is a large pallid bee ( very large for the genus Perdita ) , and in the female the ocellar region, usually including the face above level of antennae and the vertex behind ocelli more faintly, may be slightly infuscated. On the abdcmen a pale brownish, or dilute fuscous, preapical crossband, not quite reaching lateral margins, is usually present in tergites 1 to 5. The structural characters have been described rather fully by Cockerell, but one feature not men- tioned by him is the fact that the antennae are inserted unusually high on the face or above the middle of the eyes, in which respect members of Xerophasma differ from all other Perdita. The mouth parts are moderately elongate. Maxillary palpi are long and six- jointed, the first five joints subequal, the sixth somewhat shorter. Labial palpi are about equal to the maxillary pair, the first joint very long, the last three joints each about equaling the apical joint of the maxillary pair. The male resembles the female except in the primary sexual characters. The only specimen at hand lacks the abdominal band- ing, and the dark cloud on the head is restricted to the area between the ocelli. Aedeagus composed of a rather large quadrate capsule, about one half longer than wide and somewhat wider than high. Stipites produced ventrally on each side in a slender lobe, which is abruptly bent inward and upward at the extreme tip. Sagittae moderately stout, tapering toward apex, which is narrow and thinly compressed dorsoventrally. Volsellae very large, the inferior pair reaching middle of sagittae and provided with numerous minute black tubercles along the dorsal margin; dorsal volsellae reaching middle of inferior pair. Length of male, about 9 mm.; anterior wing, 7 mm. This species was described from a female collected about 10 o. m., July 9, 1917, at Fabens, El Paso Co., Texas. Subgenus COCKERELLULA Strand In 1922 (Amer. Mus. Novitates, vol. 47, pp. 1-4) Cockerell de- scribed Perdita opuntiae and erected the subgenus Ltitziella for its reception. Because Ltitziella Cockerell was preoccupied by Lutzi- ella Enderlein, Strand, in 1932 (Folia Zool. Hydrobiol. Riga, vol. 4, p. 196), proposed to rename it Cocker ellula. The group Cockerelhda is similar to Macrotera Smith, but differs in having the tongue short and the abdomen of the male modified in 964 The University Science Bulletin various ways, either at the apex or on the ventral surface. It is quite possible that discovery of additional species may reveal that these characters are gradational and only of specific importance, since members of both groups seem to be largely, although not com- pletely, dependent on the pollen of various species of Opuntia. The several species belonging in Cockerelhda may be distin- guished as follows: 1. Males 2 Females 5 2. Basal ventrite without a median fold 3 Basal ventrite strongly puckered out in middle of apical margin to form a strong fold, the following ventral segments unmodi- fied 4 3. Second ventrite with a broad thickening on middle of apical margin, this thickening strongly dentate at outer ends; apical tergite very broad, with a slender, almost fingerlike lobe at the apical corners, and evenly truncate between the lobes; head and thorax black or blackish brown, with pale-yellow face marks; abdomen ferruginous; length, about 5.5 mm.; Colorado (White Rocks near Boulder) and South Dakota ( Pine Ridge ) opuntiae Cockerell Second ventrite only slightly thickened at middle of apex, the third ventrite with a strong thickening which is unarmed at outer ends and presents a transverse disk as viewed from be- hind; apical tergite quadrate, with apical corners slightly dentate, the disk a little concave and produced medially in a rounded angle a little beyond and below the dentiform lateral angles; head and thorax dark blue-green, the face below antennae pale yellow; abdomen ferruginous; length about 4-4.5 mm.; Texas (Big Bend Park) . . bidenticauda sp. nov. 4. Tergite 7 extremely broad, transversely quadrate, not at all nar- rowed to apex, which is truncate and at most with a small median notch; head and thorax dark blue-green, the face be- low antennae yellowish white, abdomen ferruginous; length, about 4 mm.; Texas (15 miles southeast of Dryden), laticauda sp. nov. Tergite 7 similar to that of laticauda but more convex, distinctly grooved medially and emarginate at apex, thus forming a broad, slightly oblique lobe on each side; head and prothorax brown, the face below antennae pale yellow; thorax otherwise dark blue-green, more or less washed with brown on pleura, sternum, and apical truncation of propodeum; abdomen fer- ruginous; length, 4-4.5 mm.; Texas ( 15 miles northwest of Mission) lobata sp. nov. 5. Head and thorax dark green; smaller species 6 Head and thorax black, the abdomen ferruginous; mesoscutum shining, delicately tessellate and with minute close punctures; length, about 6 mm opuntiae Cockerell Timberlake: Bees of the Genus Perdita 965 6. Mesoscutum tessellate but shining, finely punctured, the punc- tures about two to three puncture widths apart; abdomen ferruginous; length, about 4-4.5 mm bidenticauda sp. nov. Similar to preceding and without good distinguishing characters in the female sex laticauda sp. nov. Perdita bidenticauda sp. nov. Allied to P. opuntiae Cockerell but distinguished by its smaller size and dark-green color, and by the very small lateroapical teeth of the apical tergite of male. Male. — Head and thorax dark green, slightly bluer on meta- notum and propodeum and gradually changing to brown on under surface of head and at anterior end of cheeks. Narrow inferior orbits near base of mandibles, mandibles, except the rufous tips, and entire face below level of antennae, pale yellow, with the broad upper margin of lateral marks oblique and reaching upper level of antennal sockets and intruding slightly between the fovea and eye margin on the orbits. Tubercles and anterior margin of pronotum brownish. Abdomen ferruginous red, with only the lateral foveae of second tergite black. Legs black, more brownish at base, the front tarsi and the front tibae, except a brownish line behind, dull yellow. Antennae almost uniformly pale brown. Tegulae pale testaceous yellow. Wings nearly clear hyaline, the nervures brown, the subcosta more fuscous. Head but little enlarged, broader than long, with the inner orbits slightly diverging below. Mandibles nearly straight, robust, and tapering to the blunt apex which has a small inner tooth. Labrum twice as broad as long. Subantennal plates very small, the lateral plates broad and becoming much broader at anterior ends. Clypeus short and transverse. Frontal foveae about three times as long as wide and well impressed. Cheeks simple and moderately wide. Abdomen moderately broad and depressed. Apex of last tergite rather broad and truncate, the apical corners a little dentiform, the disk slightly concave and medially produced a little beyond and well below level of the dentiform angles; as seen from behind and beneath, the lower apical margin is carinate and forms a broad ogival curve. First segment of venter normal, the second segment slightly, and the third strongly, swollen in middle of apical margin; this swelling elongate-oval as seen from behind, declivous, and with outer ends acute but not at all prominent or dentiform. Marginal cell about one fourth smaller than first submarginal cell, broad and ob- lique at apex, with the apendiculation of radius distinct and elongate. 966 The University Science Bulletin First recurrent nervure received near apex of first submarginal cell. Stigma small and very narrow. Head and thorax minutely tessellate and dullish, the disk of mesoscutum more shining than other parts, the mesopleura duller, with the sculpture more lineolate and the basal middle of propodeum comparatively coarsely tessellate. Punctures of frons and mesoscutum minute and close, those of vertex sparser, and those of disk of clypeus coarser and shallow. Mesopleura impunctate except toward the sternum. Abdomen moderately shining. Pubescence whitish, short and sparse, that of abdomen extremely fine and subappressed, but a row of coarse, stiff hairs present across disk of tergite 6, and similar hairs on outer side of hind tibiae. Length, 3.45-4 mm.; anterior wing, 3 mm. Female. — Dark green, or blue-green, the metanotum and pro- podeum more bluish, the abdomen ferruginous red. Clypeus, supraclypeal area extending in a triangle well above antennae, and narrow orbital margins of face, black. Labrum brownish. Mandi- bles testaceous yellow, with rufous tips. Antennae rather dark brown, the flagellum dull yellowish beneath. Legs dark brown, the tarsi paler, the front knees with a small yellowish spot, the spurs reddish. Tegulae brown at base and pale testaceous on outer margin. Wings as in male. Head as wide as long, the clypeus much larger and more pro- duced than in male. Inner orbits of eyes parallel. Mandibles simple at apex. Frontal foveae about five times longer than wide and deeply impressed. Pygidium of abdomen acute at apex. Hind knee plates distinct. Spurs of middle and hind tibiae slightly hooked at apex and minutely serrate beneath. Claws with a small inner tooth just beyond the middle. Sculpture as in male except as follows: labrum, clypeus, and supraclypeal area with coarse, shallow punctures; frons and vertex with well-separated punctures that are much sparser than those of mesoscutum; abdomen duller, more strongly lineolate than in male, and impunctate. Pubescence as in male, the apical fimbria of abdomen tinged with fuscous. Scopal hairs of hind tibiae sparse and moderately long; those of dorsal margin with minute, close-set, short branches. In comparison with opuntiae, the scopal hairs are definitely shorter and sparser. Length about 3.45-4.5 mm.; anterior wing, 3-3.2 mm. Holotype male, allotype, and 6 males, 2 females (paratypes) at flowers of Opuntia, Cooper's store, Big Bend Park, Texas, April 11, 1947 (Michener and Beamer). Two of the females bear the extremely coarse pollen from Opuntia, and the third a much finer pollen from some other flower. Timberlake: Bees of the Genus Perdita 967 Perdita laticauda sp. nov. Allied to P. bidenticauda, but the male has distinctive differences in the ventral segments and apical tergite of the abdomen. The females of the two species are apparently indistinguishable. Male. — Remarkably similar to the male of bidenticauda except in regard to the special characters of abdomen. Color the same, with the yellow mark on cheeks near base of mandibles either large or evanescent. First ventrite of abdomen strongly folded in middle of the apical margin, thus producing a large projecting pucker. Following ventrites little modified, although the third has a trans- verse preapical band of short dense hair, occupying the middle third of the width. Apical tergite extremely large and broad, more than twice as broad as long, and broadly truncate at apex, the trunca- tion with a small median notch. Length, 3.5-4 mm.; anterior wing, 2.8-3 mm. Female. — Like the female of bidenticauda. First tergite some- times infuscated on basal half. Length, 3.45-4.5 mm.; anterior wing, 2.8-3 mm. Holotype male and allotype, 15 miles southwest of Dryden, Ter- rell County, Texas, at flowers of Gilia acerosa, April 13, 1949 ( Mich- ener and Beamer). Also the following paratypes taken at the same time and place: 10 males, 8 females on Gilia, and 2 males, 6 females on flowers of Chamaesaracha conioides. The females bear a moderately fine-grained yellow pollen, which probably was derived from the Gilia flowers, as it seems to be slightly coarser and less pale yellow than the pollen borne by other species of Perdita that have been taken at flowers of Chamaesaracha. Perdita lobata sp. nov. This species is similar to P. laticauda but has the apical tergite even broader, more convex from side to side, rather strongly grooved medially, and broadly and shallowly emarginate in middle to form two bread, truncate apical lobes. Male. — Head enlarged, nearly as bulky as thorax, the temples and cheeks broad. Inner orbits distinctly diverging anteriorly. Mandibles with a weak inner tooth at apex. Frontal foveae well impressed, about five times as long as wide. Abdomen depressed and much broader than thorax. First ventrite with a strong fold in middle of apical margin, the fold producing a puckerlike process. A band of dense short hair on the third ventrite. Apical tergite as described above. Structural characters, sculpture and pubescence otherwise about as in laticauda and bidenticauda. 968 The University Science Bulletin Thorax dark bluish green, a little suffused with brown, especially below the wings and on the sides and apex of propodeum. Entire prothorax and the head reddish brown, the frons with a slight green- ish tinge. Mandibles, except piceous tips, labrum, face below level of antennae, and mark at anterior end of cheeks, pale yellow. The yellow of face gradually merges into the brown and extends slightly above the level of antennae on each side. Abdomen ferruginous red, nearlv the same color as the head. Foveae on lateral margins of tergite 2 very short and black. Legs brown, the front tibiae on anterior side and front tarsi dull yellow. Antennae nearly uniformly pale brown. Tegulae pale testaceous. Wings hyaline, the nervures brown, the subcosta darker brown. Length, 4-4.5 mm.; anterior wing, 3 mm. Two males (holotype and paratype), 15 miles northwest of Mis- sion, Hidalgo County, Texas, on Opuntia, March 30, 1946 (C. D. Michener ) . Subgenus MACROTEROPSIS Ashmead Perdita lotior Cockerell Sixteen males, 5 females from the following localities in Arizona: 6 miles east of Douglas, Aug. 11, 1940; Portal, Aug. 12, 1940; Cave Creek, Chiricahua Mts., Aug. 12, 1940 (Michener); and Rustler's Park, Chiricahua Mts., July 5^ 1940 (D. E. Hardy). This is a common species in New Mexico and Arizona at flowers of Sphaeralcea. One specimen of a new species of this group was also taken by Michener near Portal, Arizona, at flowers of Sphaeralcea. This will be described elsewhere. Perdita arcuata dinognatha Cockerell Many of both sexes from Mazourka Canyon, 6,000 feet, Inyo Mts., California, on Sphaeralcea ambigua, May 23, 1937 (Michener). Also taken at Mountain Springs Canyon, Argus Mts., 5,000 feet, Inyo County, May 22, 1937; Andreas Canyon near Palm Springs, April 6, 1939; Ribbonwood, San Jacinto Mts., May 21, 1940 (Michener); Pifion Flat, San Jacinto Mts., May 18, 1939 (E. G. Linsley); San Bernardino County, May 1. 1920 (W. Benedict); and Westgard Pass Plateau, Inyo County, May 27, 1937. This form differs from P. arcuata arcuata Fox from Baja Cali- fornia merely in having the very short transverse clypeus of the male yellow. Timberlake: Bees of the Genus Perdita 969 Perdita trifasciata sp. nov. This little species belongs to a new group which will be accorded subgeneric standing in a subsequent paper. In this group belong five species from the Colorado Desert of California, which are usually found at flowers of Coldenia. The New Mexico species differs from all except one of the California species in having the abdomen dark, and from the exception in having the face marks reduced and the pale abdominal bands very narrow and only three in number. Male. — Head and thorax dark olive-green, with yellowish-white markings as follows: basal half mandibles (these gradually chang- ing to red at apex), labrum, clypeus, except usual dots and two broad brown stripes ( somewhat farther apart than their own width ) on disk, transversely linear lateral marks not reaching above level of clypeus, and a small spot on tubercles and at posterior corners of pronotum. Middle of supraclypeal area brownish around a very small whitish spot. Antennae nearly uniformly brown above and testaceous yellow beneath. Legs blackish, the front and middle knees, tibiae and tarsi, base of hind tibiae, and base of both hind tibiae and tarsi, pale yellow; front and middle tibiae blotched with brown behind. Abdomen black with a very narrow subapical yellowish-white band on tergites 1 to 3, the band on tergites 1 and 2 emarginate behind on each side, and that on tergite 3 broadly inter- rupted on each side. Tegulae testaceous brown. Wings somewhat dusky, the nervures and margins of stigma brown. Head quadrate, as broad as long, the face below antennae promi- nent, the temples moderately wide. Disk of clypeus about one half as long as wide and shorter than the transverse length of the lateral extensions. Stigma narrowly lanceolate. Substigmatal and post- stigmatal portions of marginal cell equal. Claws bifid. Sculpture of head and thorax microscopically tessellate, imparting a soft, dull, satiny appearance. Abdomen considerably smoother and more shining than thorax. Pubescence white, rather short, sparse, and coarse; that on mesoscutum subappressed. Length, about 2.5-3 mm.; anterior wing, 2.1-2.5 mm. Two males (holotype and paratype), White Sands, New Mexico, June 27, 1940 (R. H. Beamer). In Cockerell's table ( 1896 ) this species does not run easily, but probably goes best to P. austini Cockerell, which is a somewhat larger, shining species, with the lateral face marks reaching above level of the antennae. 970 The University Science Bulletin Perdita ainsliei Crawford One male, Sioux City, Iowa, July 26, 1924 (C. N. Ainslie). The ainsliei group, which contains the following and some eight other undescribed species from the Southwest, apparently deserves subgeneric recognition. Perdita peculiaris sp. nov. A dark blue-green species, differing from P. ainsliei Crawford in the black abdomen and in having the light markings restricted to the mandibles and a narrow anterior band on disk of clypeus. ( In ainsliei the abdomen is ferruginous and the clypeus and quadrate lateral marks are yellowish white.) In Cockerell's table (1896) this new species runs to a female variety of P. phymatae Cockerell which is only superficially similar. Male. — Head and thorax dark blue-green, the abdomen black. Mandibles, except rufous tips, and a band on anterior margin of disk of clypeus, yellowish white. Remainder of clypeus black. Labrum brown. Antennae blackish, the flagellum more brownish and becoming broadly rather light brown beneath. Legs black, the small joints of tarsi brown. Tegulae testaceous brown. Wings somewhat dusky hyaline, the nervures and margins of stigma sepia brown. Head about as wide as long, rather thick fronto-occipitally, the face below antennae strongly convex. Clypeus prominent, its anterior margin rounded out and well elevated. Cheeks much narrowed anteriorly and simple. Subantennal plates about as broad as high. Inner anterior corners of lateral plates of face rounded and somewhat tumescent next to disk of clypeus. Abdomen broadly ovate, convex above, a little broader than thorax. Stigma of fore wing narrowly lanceolate. Marginal cell broadly and obliquely truncate at apex, with the portions beneath and beyond the stigma equal. Outer nervure of second discoidal cell obsolescent. Claws bifid. Mandibles moderately long, curved, and simple. Maxillary palpi rather long and six-jointed. Head and thorax minutely tes- sellate and moderately shining, the metanotum and propodeum strongly tessellate and more dullish than other parts. Puncturation of mesonotum very fine and close, and that of frons moderately sparse. Abdomen also dullish from a minute dense puncturation (in which it differs from ainsliei as well as almost all other species of Perdita) except on the depressed apical margin of the tergites. Pubescence fine, sparse, and inconspicuous. Length, about 3-4 mm. (the abdomen much recurved at apex); anterior wing, 2.5-3 mm. Timberlake: Bees of the Genus Perdita 971 Aedeagus of the type of P. ainsliei; the long slender apical lobes of stipites abruptly thickened and bent downward at apex (acumi- nate and very sharp at apex in ainsliei). Sagittae forming a fusi- form body, tapering to the acute apex, as seen from side, but the tip of each rod expanding acutely outward, as seen from above, and armed with a slendar upward-directed spine at the inner apical corner (in ainsliei rather similar, but without the apical spine and clothed beneath with a loose fascicle of hairs close to the apex). Volsellae reduced to one pair, which are rather long and acuminate, as seen from above ( two pairs of volsellae in ainsliei ) . Two males (holotype and paratype), 15 miles southeast of Dry- den, Terrell Co., Texas, on Chamaesaracha conioides, April 13, 1949 ( Michener and Beamer ) . Subgenus COCKERELLIA Ashmead The type of Cocker ellia is P. hyalina Cresson, which has been usually considered the male of P. albipennis Cresson, although the synonomy is doubtful. Excepting Xerophasma, the species of Cockerellia are the largest in the genus Perdita and are found almost exclusively at flowers of Compositae, mostly of the genera Helianthus, Rudbeckia, Ratibida, Raileya, Verbesina, Coreopsis, Geraea, and Gaillardia, but with records also from Erigeron, Hetero- theca, Pedis, and Prionopsis. Key to the Species of Cockerellia Females 1. Head and thorax mainly green 2 Mainly bright yellow, the lower half of pleura black; clypeal dots, facial foveae and flagellum above, also blackish; mesono- tum with short, dense, erect yellow hair heata Cockerell 2. Clypeus yellow or whitish, the lateral marks triangular and reach- ing level of antennae or beyond 3 Face marks restricted to spots on clypeus and transverse lateral marks, the latter not extending above level of clypeus 6 3. Mesonotum minutely, densely punctate, the punctures not much more than their own diameter apart; supraclypeal area en- tirely dark 4 Mesonotum polished, finely punctured, the punctures although very numerous, several times their own diameter apart; clypeus, triangular lateral marks, and a vestigial supraclypeal mark, yellowish white; prothorax, legs in large part and ab- domen yellow, the abdomen with four black bands, utahensis Cockerell 4. Head and thorax brassy green, the mesonotum a little tessellate and dullish between the punctures 5 972 The University Science Bulletin Head and thorax dark blue-green; mesonotum shining, the dense punctures very minute; base of mandibles, clypeus, lateral marks, prothorax and legs in large part, and abdomen, yel- lowish white, the abdomen with five black bands; pubescence white baileyae Cockerell 5. Clypeus and lateral marks white; prothorax, legs in large part, and abdomen, pale yellow or yellowish white, the abdomen with five rather broad black bands, and also black on middle of basal declivity of tergite 1 perpulchra perpulchra Cockerell Similar, but light parts of thorax, legs and abdomen bright yellow, the first tergite entirely yellow except for a narrow apical black band perpulchra flavidior subsp. nov. 6. Frons dull or opaque, obscurely punctate or impunctate 7 Frons at most dullish, distinctly although minutely punctate 12 7. Wings slightly dusky hyaline, the nervures and stigma testaceous or pale brown; hair of hind tibiae more or less fuscous 8 Wings whitish or milky hyaline, the nervures pale yellow or whitish, the stigma frequently yellow or pale orange; hair of hind tibiae usually but not always light 9 8. Wings rather distinctly dusky, the nervures and stigma pale brown; clypeus, except two dark stripes which are abbreviated anteriorly and more or less widened behind, transverse lateral marks, and a broad band on tergites 2 to 5, narrowly inter- rupted medially on 2 and 3, yellow ... bequaerti bequaerti Viereck Wings more faintly dusky, the nervures more yellowish; dark stripes on clypeus often enclosing a median yellow stripe, which expands anteriorly bequaerti indianensis Cockerell 9. Frons dull but not opaque, its punctures obscure 10 Frons opaque and virtually impunctate 11 10. Clypeus strongly produced, the width of the anterior truncation a little less than length of the oblique sides to base of mandi- bles, the face below antennae prominent in profile; immacu- lately dark green, with the abdomen blackish, or spot or streak on middle of clypeus, slender transverse lateral marks, two spots on pronotum, and interrupted band on tergites 2 to 4, yellow verbesinae Cockerell Clypeus less produced, the anterior truncation broader than length of oblique sides, the face below antennae little prom- inent in profile; more brassy green than verbesinae; median streak on clypeus, slender, transversely curved lateral marks, scape of antennae beneath, cuneate marks on pronotum, tubercles, and a broad, interrupted band on tergites 2 to 4, yellow fracticincta sp. nov. 11. Large, robust species, about 9 mm. long; face entirely dark, or, typically, with median stripe on clypeus and transverse lateral marks, yellow; tubercles at apex, two spots on pronotum, and broad band, usually reaching lateral margins on tergites 2 to 5, also yellow; hair on hind tibiae whitish to pale brownish lacteipennis lacteipennis Swenk and Cockerell TlMBERLAKE: BEES OF THE GENUS PERDITA 973 Slightly smaller; abdominal bands rather narrow, not quite reach- ing lateral margins and narrowly interrupted on tergites 2 and 3; marks on pronotum and tubercles large; face marks large, the clypeus yellow except for two triangular dark areas on disk, with the yellow extending triangularly upward be- tween them; hair of hind tibiae sometimes pale fuscous; length about 7-8 mm lacteipennis pallidipennis Graenicher 12. Clypeus more prominent, and more produced, the disk fully as long or a little longer than wide 13 Clypeus less produced and prominent, the disk somewhat broader than high; head therefore appears to be slightly broader than long, with face below antennae much less prominent in profile, 15 13. Frons minutely tessellate, shining, and minutely punctate; yellow bands of abdomen usually broad, and a yellow mark present on each side of disk of tergite 1 14 Frons rather strongly tessellate, more dullish and much more strongly punctate than in albipennis; a rather narrow yel- low band, not reaching lateral margins on tergites 2 to 4; tergites 1 and 5 immaculate, or 5 with two yellow spots which are usually small or faint tricincta sp. nov. 14. Head and thorax dark olive-green; hair of legs light, the scopa of hind tibiae pale yellowish, yellow abdominal bands usually reaching lateral margins; face marks pale yellow, the upper part of clypeus disk with two dark blotches which usually outline a yellow "T" mark albipennis Cresson Head and thorax brassy green; hair of legs fuscous to blackish, especially on middle and hind legs; abdominal bands broad and even but failing to reach lateral margins; clypeus usually black, with a medium pale-yellow line scopata sp. nov. 15. Abdomen always dark, the yellow bands nearly or quite reaching lateral margins on tergites 2 to 5, and tergite 1 dark with a yellow mark on each side of disk; median yellow mark on upper part of clypeus not much narrowed above and often T-shaped 16 Abdomen sometimes yellow, with dark subapical and basal spots on each side of segments; or sometimes dark, with enclosed yellow band on tergites 2 to 5, broadest at middle and nar- rowed toward sides (in which case tergite 1 lacks yellow lateral marks on disk, but may be more or less yellow at base ) ; dark blotches on upper part of clypeus large, the yellow intru- sion between them pointed above 17 16. Frons more or less distinctly tessellate and dullish, the punctures minute; yellow bands on tergites 2 to 5 rather even and entire, narrow on tergite 2 and gradually wider on successive seg- ments, that on tergite 5 subabbreviated, lepachidis lepachidis Cockerell Frons almost polished, the punctures consequently very dis- tinct; markings as in lepachidis, but bands on tergites 2 and 3 generally more or less notched medially behind, and that on 2 sometimes distinctly interrupted. . lepachidis levifrons subsp. nov. 974 The University Science Bulletin 17. Abdomen and legs not almost entirely yellow; the abdomen with dark markings, at least on tergites 1 to 4 18 Abdomen and legs almost entirely yellow; four small fuscous spots on disk of tergite 1, and two small subapical spots on tergite 2; femora of front and middle legs with a pale-brown blotch on inner side coreopsidis kansensis subsp. nov. 18. Abdomen yellow with a dark band at junction of tergites 1-2 to 4-5, more or less reduced to a basal spot on lateral margins and a subapical spot on each side; tergite 1 also with a dark blotch on each side at summit of basal declivity, coreopsidis coreopsidis Cockerell Abdomen dark above with an enclosed yellow band on tergites 2 to 5, broadest in middle and narrowed at outer ends; ter- gite 1 also more or less yellow at base, coreopsidis obscurior subsp. nov. Perdita utahensis Cockerell Five males, 2 females, 10 miles south of Tucson, Arizona, Aug. 7, 1940 (C. D. Michener); 5 males, 1 female collected by the writer at the same place and time were found at flowers of Verbesina encelioides. Cockerell described only the female of this species from a specimen collected in southwestern Utah. Male. — Head and thorax dark blue-green. Spot at anterior end of cheeks, mandibles, except rufous tips, labrum, clypeus, large tri- angular lateral marks reaching level of antennae, first three joints of antennae broadly beneath, collar and posterior band on pronotum, narrowed mesad, and the tubercles, bright yellow. Sometimes upper margin of subantennal plates yellow, and this even confluent with lateral marks. Legs blackish, the trochanters in part, the femora broadly at apex and an anterior stripe on front and middle pair, the tibiae except behind, and the tarsi, yellow. Flagellum orange ferruginous, becoming dusky above; the pedical and apex of scape fuscous above. Abdomen shining blackish, the apical tergite testaceous yellow, and the apical depression of tergites 1 to 6 whitish subhyaline. Preceding the depression is a yellowish- white band on tergites 1 to 6, this band obliterated except on lateral margins on tergite 1, dilated on each side and covering reflexed sides on tergites 2 to 6, and with remaining part of band broadly interrupted, or almost so, sublaterally, and notched or subinter- rupted medially on tergites 3 to 5, but the same part on tergite 2 generally obliterated. Tegulae yellow at base and almost hyaline on outer margin. Wings whitish, with a slight dusky-yellowish opacity. Nervures and stigma yellowish, the margins of stigma generally slightly brownish. TlMBERLAKE: BEES OF THE GENUS PERDITA 975 Head quadrate, the cheeks broad. Face below antennae polished, impunctate. Vertex and frons minutely granular-tessellate, the lat- ter moderately shiny and minutely punctate. Mesonotum nearly polished and set with minute, moderately close punctures. Pubes- cence white and moderately dense. Length, about 6-7 mm.; an- terior wing, 4.3-4.9 mm. Perdita baileyae Cockerell One male, Rice Co., Kansas, July 3, 1923 (C. H. Martin). Cockerell described the male of this species from Mesilla, New Mexico. I have seen a female from San Xavier Mission, Arizona (E. P. Van Duzee). Perdita perpulchra flavidior subsp. nov. Cockerell recorded perpulchra from Las Cruces, New Mexico, and described the type as having the abdomen white with black bands, the black on the basal tergite extending triangularly to the base, with the white on each side enclosing a black subapical spot. A female taken 15 miles west of Holbrook, Arizona, at flowers of Erigeron, Sept. 3, 1930 (Timberlake) agrees well with Cockerell's description, except that the abdomen is more yellowish. In flavidior the light parts, except those of face, are bright yellow, and the basal tergite is black only at apex. Female. — Head and thorax dark olive-green. Basal half of man- dibles, clypeus, and triangular lateral marks reaching level of an- tennae, pale yellow. Scape of antennae, except spot at apex above, under side of pedicel and of first joint of flagellum, prothorax and legs in large part, and abdomen, bright yellow. Coxae at base, under side of front and middle trochanters and femora, spot at apex of hind femora above, posterior surface of front and middle tibiae, hind tibiae entirely and all tarsi, except basal joint of front pair, dark brown or blackish. A black band at juncture of tergites 1-2 to 4-5, with the broad intervening yellow bands distinctly notched on each side behind by sublateral swellings of the black, especially on ter- gites 2 and 3. Tergite 5 with a narrow, preapical blackish band. Venter pale brownish yellow at base; segments 2 and 3 each with two fuscous patches, and the apical segments extensively fuscous. Tegulae pale testaceous, with a yellow spot at base. Wings some- what whitish hyaline, slightly tinged with dusky yellowish. Ner- vures and stigma testaceous yellow. Head slightly broader than long, the clypeus little produced and prominent. Frons slightly duller, with somewhat closer and fainter 32—6490 976 The University Science Bulletin punctures than in albipennis. Mesonotum minutely and very closely punctate and covered with fine very short, erect hair. This rather dense pubescence of mesonotum and that of vertex distinctly ochreous. Wings, in comparison with albipennis, small and slightly dusky. Length, 8 mm.; anterior wing, 4.8 mm. One female (holotype), Great Bend, Barton Co., Kansas, on Heterotheca subaxillaris, Sept. 9, 1949 (Michener and Beamer). Perdita bequaerti indianensis Cockerell In preparing the preceding key to the species of Cockerellia, I noticed that a specimen from Gary, Indiana, that had been de- termined by Cockerell as P. pallidipennis Graenicher, had some- what dusky wings, a character which would relegate it to P. be- quaerti Viereck. I conclude that indianensis that was based on a male from Lafayette, Indiana, must be the same as the Gary speci- men, since the margins of stigma and end of marginal cell could not be described as fuscous unless the wings were slightly dusky, rather than milky hyaline as in pallidipennis. This is a rather weakly developed race of bequaerti, and was described as a race of pallidipennis. It and P. wickhami Cockerell were published in the same paper and must be the sexes of prac- tically the same thing (new synonymy). Perdita verbesinae Cockerell Five males, 9 females, Portal, Arizona, on Verbesina encelioides, Aug. 12, 1940 (C. D. Michener). Specimens of this species, even those of a series taken at the same spot, vary greatly in the extent of the yellow markings, and some- times lack the markings altogether. Such immaculate specimens were named var. nigior by Cockerell in his original description. Perdita fracticincta sp. nov. This species is allied to P. lepachidis Cockerell and P. coreopsidis Cockerell in the shape of the head, but in the strongly tessellate and dull frons, the greatly reduced face marks, and the interrupted ab- dominal bands it strongly resembles verbesinae. Female. — Brassy green, the labrum and clypeus black, the latter green, however, on upper margin of disk. Small longitudinal median spot on clypeus, a small transverse spot on each lateral extension of clypeus, and slender to transversely arcuate lateral marks, yellow. The lateral marks bordering anterior margin of lateral plates of face are slightly interrupted on one side and have TlMBERLAKE: BEES OF THE GENUS PERDITA 977 a short extension upward on orbits. Interrupted band on hind margin of pronotum, and the tubercles, yellow. Abdomen blackish, a small spot on lateral margins of tergite 1 (mostly on the reflexed ventral portions) and interrupted band on tergites 2 to 4, yellow; band on tergite 2 reaching the lateral foveae, and that on 3 having a faint, slender posterior extension to lateral margins. Mandibles and pygidium rufo-testaceous, the former more rufous at apex. Antennae blackish above, the scape yellow, the flagellum brown beneath. Legs blackish, a small spot on front knees and the front tibiae on anterior side, yellow. Tegulae pale testaceous. Wings milky hyaline, nervures nearly colorless, the stigma pale yellow. Head hardly longer than wide, the clypeus little produced and prominent. Frons and vertex minutely granular-tessellate and rather dull, the minute shallow punctures of frons only moderately distinct. Face below antennae polished, with sparse minute punc- tures. Mesoscutum polished on disk, although minutely tessellate anteriorly, and provided with rather close minute punctures. In some ways, including shape of head, pubescence, size, etc., agreeing closely with coreopsidis. Length, 7 mm.; anterior wing, 4.9 mm. One female (holotype) 29 miles south of Sarita, Willacy Co., Texas, on Coreopsis, Apr. 14, 1950 ( Michener, Beamer, Stephen and Bozen ) . Perdita lacteipennis lacteipennis Swenk and Cockerell Unquestionably a common species in western Kansas at flowers of Helianthus, as there is a series of 127 females and 145 males in the University of Kansas collection from various localities in the Counties of Chase, Beno, Sheridan, Ottawa, Hamilton, Greeley, Meade, Ford, Finney, Pawnee, Osborne, Barton, Pratt, Russell, Rawlins, Saline, Ellsworth, Comanche, Stafford, Clark, Decatur, Rice, and Rooks. Nearly all the specimens that are labeled with the flower record were taken on Helianthus petiolaris, but 6 males from Chase and Hutchinson were on Prionopsis ciliata. The dates of capture range from June 14 to Sept. 6. Also 79 specimens from La Junta, Colorado, two of them labeled as from Helianthus annuus; 2 males, Lamar, Colorado (Snow); 1 male, Artesia, Moffat Co., Colorado, on Helianthus petiolaris, July 22, 1950 (Michener); and 1 male each from Leonard, North Dakota, July 25, 1937, and Lusk, Wyoming, July 14, 1937 (C. L. Johnston). Specimens from western Kansas are mostly typical lacteipennis, although about 10 per cent of the females have the face marks very small or absent and thus agree with the form canadensis Crawford. 978 The University Science Bulletin Of 28 females from La Junta, Colorado, 6 verge toward canadensis (face marks obsolescent), 6 verge toward pallidipennis (clypeus with a yellow spot on each side besides the median mark), and 16 are typical lacteipennis. Perdita lacteipennis pallidipennis Graenicher Thirty-two females, 15 males, Douglas Co., Kansas, on Hclianthus annuus, Sept. 9, 1949 (Michener and Beamer); 3 females, Topeka, Kansas, Sept. 17, 1910 (E. G. Titus). These specimens are nearly typical pallidipennis, the female of which has the clypeus yellow, with a dark blotch above on each side of disk, or the dark color frequently reduced to a short strip on each side. Males have the basal half of mandibles, labrum, and anterior border of face yellow. Specimens from Clay Co., Kansas (J. C. Bridwell), average somewhat closer to pallidipennis than to typical lacteipennis but probably should be classed as intermediates. In the original description of lacteipennis it is stated on the au- thority of Prof. Myron Swenk that the species does not occur in eastern Nebraska. However, the species does occur there, as I have a female of pallidipennis from Nebraska City on Helianthus, Sept. 14, 1901 (M. A. Carriker). The form pallidipennis has the face markings much as in albipennis and hence must have been confused with albipennis by Swenk. The dull frons of lacteipennis and the shining punctate frons of albipennis easily distinguish the two species. Perdita tricincta sp. no v. Nearest P. albipennis Cresson, but distinguished by the duller, more tessellate, and more coarsely punctate frons, and by the three narrow bands on abdomen (the yellow on the fifth tergite restricted to two small spots or absent). The male has a dark, shining abdo- men and transverse yellow face marks. Female. — Head and thorax dark green, the mesonotum some- what brassy. Labrum and clypeus blackish, the latter with a broad, pale-yellow, median stripe and a suffused yellowish mark on each side. Lateral marks pale yellow, rounded at inner ends and gen- erally somewhat less than twice as broad as high. Mandibles testaceous brown, the tips dark rufous. Collar of pronotum ob- scurely yellowish, the hind margin and tubercles bright yellow. Abdomen blackish, with a rather narrow yellow band on tergites 2 to 4, that on 2 ending at the lateral fovea, the others not quite TlMBERLAKE: BEES OF THE GENUS PeRDITA 979 reaching lateral margins, all slightly notched medially, those on tergites 2 and 3 behind and that on tergite 4 in front. Tergites 1 and 5 entirely black, or the latter frequently with two submedian yellow spots. Pygidium ferruginous brown. Legs blackish, the front tibiae yellowish on anterior side. Antennae blackish, the scape narrowly clear yellow and the flagellum brown beneath. Tegulae pale testaceous. Wings milky hyaline, the nervures nearly colorless, the stigma pale yellow. Head about as broad as long, with the clypeus produced and prominent. Facial foveae about one third as long as eyes. Sculp- ture as in albipennis, except the frons duller, with a stronger tessella- tion and coarser punctures. Pubescence of vertex and mesonotum more ochreous than in albipennis. Scopa of hind tibiae a pale yellowish brown. Length, 7.5-9 mm.; anterior wing, 5-5.2 mm. Male. — Head and thorax dark green. Base of mandibles, an- terior band on clypeus (more testaceous far laterad), sometimes a median streak on clypeus to summit, and lateral face marks, about twice as wide as high, yellow. Mandibles testaceous in middle and rufous at tips. Labrum testaceous. Antennae orange-brown, a little dusky on apical half of flagellum above, becoming darker toward base of flagellum, and blackish on pedicel and on the scape except at base, with underside of scape and of first joint of flagellum yellow. Collar of pronotum yellowish, especially toward the mid- dle, a small cuneate spot on posterior corners of pronotum clear yellow, but tubercles dark. Abdomen shining blackish, the apical tergite testaceous, the apical depression of preceding segments sub- hyaline. In some specimens, including the allotype, there is a trace of yellow just in front of the apical depression on tergites 5 and 6 and on the reflexed sides of the preceding segments. Legs blackish; all tarsi, front tibiae broadly and middle tibiae narrowly on anterior side, and front and middle knees, yellow. Tegulae and wings as in female. Head quadrate, the cheeks broad. Sculpture as in female, the pubescence white. Length, 5.3-8 mm.; anterior wing, 4-5 mm. Holotype female and allotype, Southmost, Cameron Co., Texas, on Helianthus annum, Apr. 13, 1950 (Beamer, Michener, Stephen, and Rozen). Paratypes as follows: 30 females, 40 males taken with the types; 1 male, 5 miles east of Brownsville, Apr. 13, 1950; 3 females, 4 males, Progresso, Hidalgo Co., on Helianthus annuus, Apr. 12, 1950; 1 male, 6 miles east of Rio Grande, Starr Co., Texas, Apr. 12, 1950 ( all, Michener, Beamer, Stephen, and Rozen ) . 980 The University Science Bulletin Perdita albipennis Cresson The material recorded below belongs almost entirely, in respect to the females, to the variety lingualis Cockerell, which is the best- known and the most common form of albipennis in collections. This more highly colored form is, acording to Cockerell, possibly a valid race. The male variety helianthi Cockerell, with yellow markings on abdomen and a more or less blackish flagellum, flies in about equal numbers with males having the flagellum orange brown or ferruginous and the abdominal markings obsolete, and both fly with lingualis. The P. hyalina Cresson, without face or abdominal markings indicated in the original description, seems distinct from albipennis. The variety pasonis Cockerell, with a dull frons, is more likely allied to P. lacteipennis than to albipennis. Altogether, therefore, it is necessary to have a much fuller collection of albi- pennis in order to understand its variation and possible subspecia- tion. The following material of albipennis has been examined: Kansas: 14 females, 17 males, Clark Co., 1,950 feet, Aug. 23, 1911; 2 females, 1 male, Rawlins Co., 2,850 feet; 1 female, Norton Co., 2,270 feet, Aug. 12, 1912; 2 females, Smith Co., 1,800 feet, Sept. 4, 1912; 2 females, 1 male, Stevens Co., 2,700 feet, Aug. 10, 1911; 1 male, Sherman Co., 3,690 feet; 1 male, Ellis Co., 2,000 feet, July 18, 1912; 1 male, Decatur Co., 2,560 feet (all F. X. Williams); 1 male, Ness Co.; 10 females, 4 males, Sharon Springs, Wallace Co., on Helianthus annuus, Aug. 23, 1931 ( Timberlake ) ; 1 male, Lakin, Kearny Co., on Helianthus petiolaris, July 23, 1950 (C. D. Mich- ener ) . Colorado: 4 males, La Junta, July 3, 1949 (R. H. Beamer and J. R. White). Wyoming: 1 male, Lingle, Goshen Co., July 14, 1937 (C. L. Johnston ) . New Mexico: 1 female (probably belonging here, but it cannot be identified positively as the head is missing), Cuervo, Guadalupe Co., June 23, 1940 (E. E. Kenaga). Arizona: 1 male, Fort Apache, Navajo Co., Aug. 26, 1897 (Snow, No. 2113). Perdita scopata sp. no v. This species is much like P. albipennis, but the hair of middle and hind legs and that of front tarsi more or less blackish, a character more faintly indicated in some of the other species of Cockerellia, such as P. bequaerti and P. lacteipennis pallidipennis, which, how- Timberlake: Bees of the Genus Perdita 981 ever, have the frons dull and impunctate. The male of scopata has the face markings greatly reduced or absent and the abdomen im- maculate, opaque black. Female. — Head and thorax dark green, the frons, vertex, and mesonotum with a strong brassy luster, the labrum and clypeus black. Line on under side of scapes, median line on clypeus, not reaching anterior margin, and slightly transverse lateral marks, whitish. The lateral marks, rounded at inner ends, extend about half way to level of antennae. Sometimes a transverse whitish spot on lateral margins of disk of clypeus. Mandibles black at base and rufous on apical third or half. Antennae blackish, the flagellum slightly rufescent beneath. Tubercles and hind margin of pronotum pale yellow, the anterior margin narrowly yellowish. Abdomen black, the pygidial plate dark rufous. Tergites 2 to 5 each with a broad yellow band, these bands more even than in albipennis and falling distinctly short of the lateral margins. A slender, oblique, more or less sinuate and broken yellow line on each side of the base of the disk of tergite 1. Legs blackish, the front knees and a line on anterior side of front tibiae suffused with yellow. Tegulae tes- taceous. Wings whitish hyaline, the nervures and stigma pale tes- taceous yellow (in comparison with albipennis the wings have a barely perceptible dusky tinge and the nervures are less pallid). Structure, sculpture, and pubescence as in albipennis, with the fol- lowing differences: cheeks less broadened and clypeus less prom- inent; punctures of mesonotum slightly closer; abdomen dull, with an excessively minute tessellation; pygidium very broad and broadly subtruncate at apex; basal middle of propodeum minutely rugose; pubescence flavo-ochraceous on top of head and on mesonotum, more or less dark fuscous on middle and hind legs and front basi- tarsi, and dilute fuscous at apex of abdomen. Length, 7-9 mm.; anterior wing, 5-5.5 mm. Male. — Dark green, the mesonotum with a brassy luster, the base of mandibles, labrum and clypeus shining black. Abdomen opaque black, the apical depression of tergites more shining but not palles- cent; tergite 7 becoming brownish testaceous on apical half. No light facial thoracic or abdominal markings, except rarely a small pale-yellow spot on each side of disk of clypeus and another opposite anterior corners of eyes. Mandibles rufous at apex. Antennae blackish, the scape pale testaceous at extreme base, the flagellum brown beneath. Legs blackish, a small spot on front and middle knees pale yellowish, the front and middle tarsi testaceous yellow, and the hind tarsi brownish. Tegulae testaceous. Wings milky 982 The University Science Bulletin hyaline, with pale-yellow nervures and stigma (wings distinctly whiter than in female). Structure, sculpture, and pubescence simi- lar to condition in male of albipennis, except the frons slightly duller, with less distinct punctures, mesonotum with somewhat closer punctures, base of propodeum with a small, triangular, finely rugose area, and abdomen dull from an excessively minute thimble- like tessellation. Length, 6-8 mm., anterior wing, 5-5.5 mm. Holotype female and allotype, 29 miles south of Sarita, Willacy Co., Texas, on Coreopsis, Apr. 14, 1950 (Beamer, Michener, Stephen, and Rozen). Paratypes as follows: 25 females, 61 males, taken with the types; 10 females, 36 males, same locality and date, at flowers of Helianthus annuus; 7 females, 4 males, Southmost, Cameron Co., on Helianthus annuus, Apr. 13, 1950; 1 male, 5 miles east of Brownsville, Apr. 13, 1950 (Beamer, Michener, Stephen, and Rozen ) ; 1 female, 20 miles south of Sarita, on yellow composite, April 2, 1946 (C. D. Michener); 1 female, Hidalgo Co., Texas, May 5, 1929 (H. B. Parks); 1 female, 1 male, Brownsville, June (F. H. Snow, No. 2428); and 1 female, 1 male, Galveston, May (Snow, No. 1824). The Brownsville specimens taken in June ( Snow ) differ in having considerably larger face marks than usual. Perdita lepachidis lepachidis Cockerell This species was described in 1896 from Santa Fe and Socorro, New Mexico. Only the male was known to Cockerell, who col- lected it at flowers of Ratibida tagetes (the plant was then called Lepachys tagetes ) . Female. — Brassy green, much as in coreopsidis, but the median pale-yellow area of clypeus briefly crossed at the top like a "T" and expanding below, or sometimes isolated and Indian club-shaped (in coreopsidis triangular and pointed above). Anterior and pos- terior margin of pronotum and tubercles bright yellow. Abdomen blackish, with a spot on each side of tergite 1 and a crossband on tergites 2 to 5, bright yellow; these bands reaching lateral margins or nearly so, and that on tergite 2 rather narrow, but those on tergites 4 and 5 broad and emarginate behind on each side. Legs dark, the anterior side of front and middle tibiae and a small spot on knees, yellow. Scape of antennae yellow except at apex above, the flagellum brown above and yellowish brown beneath but be- coming yellow beneath on the basal joint and on the pedicel. Tegulae pale testaceous. Wings milky hyaline, the nervures and stigma pale yellow. Length, about 7 mm.; anterior wing, 5-5.3 mm. Timberlake: Bees of the Genus Perdita 983 Four females, Lakin, Kearny Co., Kansas, on Ratibida columnaris, July 23, 1950 ( C. D. Michener ) ; and 1 female, Sheridan Co., Kansas, 2,650 feet (F. X. Williams). The specimen from Sheridan Co. differs in having the clypeus yellow with two dark stripes which are irregular, that on the left side forming a figure 7 and that on the other side confluent with the usual lateral dot. I have examined, also, a pair from Alpine, Texas, July 8, 1942 (H. A. Scullen), one female from Sterling, Logan Co., Colorado, May, 1923 (Grace Sandhouse), and two males from Santa Fe and Socorro, New Mexico (Cockerell). Perdita lepachidis levifrons subsp. nov. This race differs from typical lepachidis in the smoother, more shining frons and in the more restricted facial and abdominal markings. Female. — Brassy green. Lateral face marks and clypeus pale yellow, the disk of latter with a blackish blotch (becoming green on outer margin above) on each side, isolating a rather slender median yellow streak, which is briefly crossed at summit to form a "T." Yellow abdominal bands not quite reaching lateral margins, that on tergite 5 generally much abbreviated, those on tergites 2 and 3 notched medially behind, and that on tergite 2 frequently narrowly interrupted in middle. Flagellum more ferruginous and the dark part of legs, especially on tibiae, more brownish than in lepachidis. Frons so delicately tessellate as to be almost polished, its punctures minute but distinct. Length, 6.5-7.5 mm.; anterior wing, 4.8-5.1 mm. Male. — Not appreciably differing from male of lepachidis from New Mexico, except that frons is nearly polished and the wings are smaller. Length, 6-7 mm.; anterior wing, 4.3-4.6 mm. (length of wing in New Mexico male, 5 mm. ) . Nine females, 4 males (holotype female, allotype, and paratype), Brownsville, Texas, June (F. H. Snow, Nos. 2416 and 2417). Three of the females lack the abdomen. Perdita coreopsidis coreopsidis Cockerell This species was described from specimens collected at Cotulla, La Salle Co., Texas, at flowers of Coreopsis, but it apparently prefers Gaillardia to any other flower. Texas: 2 females, 1 male, Catarina, Dimmit Co., on Monarda punctata coryi, Apr. 11, 1950 (Beamer, Michener, Stephen, and Rozen ) . 984 The University Science Bulletin Kansas: 9 males, Dodge City, Ford Co., on Gaillardia, June 15, 1949 (Michener and Beamer); 1 female, Satanta, Haskell Co., on Gaillardia, June 16, 1949 (Michener and Beamer); 1 female, 5 miles north of Quinter, Sheridan Co., on Gaillardia, June 17, 1950 (Mich- ener); 1 male, Ashland, Clark Co., on Gaillardia, June 12, 1949 (Michener and Beamer); 4 females, 4 males, Lamed, Pawnee Co., on Gaillardia, June 14, 1949 (Michener and Beamer); 2 females, Clark Co., 1,962 feet, May and June (F. H. Snow, Nos. 1160 and 1161); 2 females, Gove Co., 2,813 feet (F. X. Williams). Perdita coreopsidis kansensis subsp. nov. This form flies with the typical subspecies on the northwestern border of the known range of coreopsidis, but farther south and west it apparently attains the status of a race. It differs from typical coreopsidis in having the legs and abdomen of the female nearly all yellow, and the abdomen of male yellow with dark bands. Female. — Like typical coreopsidis except in color of legs and abdomen. Legs yellow, the anterior and posterior surface of front femora beneath except at apex, and anterior side of middle femora beneath except apex, brown or fuscous, and the hind margin of middle and hind tibiae and basitarsi slightly darkened. Abdomen yellow, with four fuscous spots on tergite 1 and sometimes two spots on tergite 2, these spots being situated as in the spotted form of coreopsidis, but small. Male. — Yellow face markings similar to those of female ( obsolete or nearly so in typical male). Apex of femora, the tibiae and tarsi nearly all yellow (the yellow mainly restricted to anterior side of front tibiae in typical male). Tergite 1, except a narrow apical light crossband, and a crossband at base of tergites 2 to 5, black, the rest of abdomen yellow, except for a small transverse preapical dark spot on each side of tergites 2 and 3. Apical depression of tergites whitish subhyaline as in the typical male. Two females, 1 male (holotype female, allotype, and paratype), Gove Co., Kansas, at 813 feet (F. X. Williams); 1 male (paratype), Dodge City, Ford Co., Kansas, on Gaillardia, June 15, 1949 ( Mich- ener and Beamer); and 1 male (paratype), Van Horn, Culberson Co., Texas, May 23, 1932 (E. G. Linsley). Perdita coreopsidis obscurior subsp. nov. This race of coreopsidis appears to be restricted to the coast of Texas, from the mouth of the Rio Grande to Galveston. It differs Timberlake: Bees of the Genus Perdita 985 from the typical form in the less extent of yellow markings, and in the slightly sparser puncturation of the mesoscutum. Female. — Like coreopsidis except in color of abdomen. Tergite 1 nearly all dark except at base, and the following tergites fuscous or blackish around the margins to enclose a broad yellow band that is more or less narrowed and acute at outer ends. On tergites 4 and 5 the dark margin sometimes broadly interrupted in middle apically and more rarely also at the sides, thus more or less isolating two dark spots before the apical depression. Length, 6-7.5 mm.; anterior wing, 4.1-4.6 mm. Male. — Mandibles except tips and lateral extensions of clypeus, yellow; disk of clypeus black, with the yellow sometimes intruding slightly on each side anteriorly. Abdomen with the yellow mark- ings evanescent or absent and never forming more than a thin line interpolated between the dark base and the subhyaline apical de- pression. Punctures of mesoscutum moderately close. Length, 4.8- 6.5 mm.; anterior wing, 3.9-4.4 mm. Sixteen females, 59 males (holotype female, allotype, and para- types), Boca Chica, Cameron Co., Texas, on Gaillardia, Apr. 13, 1950 ( Beamer, Michener, Stephen, and Rozen ) . Also the following paratypes: 1 female, Brownsville, July 3, 1938 (R. H. Beamer); 2 females, 12 males, Brownsville, Apr. 13, 1950 (Beamer); 6 fe- males, 2 males, 18 miles north of Harlingen, Cameron Co., on Gaillardia, Apr. 2, 1938 (C. D. Michener); 1 female, 1 male, Gal- veston, May (F. H. Snow, Nos. 1825, 1826); and 3 females, 1 male, Corpus Christi, June 28, 1942 (E. S. Ross). LITERATURE CITED COCKERELL, T. D. A. 1896. The bees of the genus Perdita F. Smith. Proc. Acad. Nat. Sci. Philadelphia, vol. 48, 1896, pp. 25-107. THE UNIVERSITY OF KANSAS SCIENCE BULLETIN Vol. XXXV, Pt. II] September 10, 1953 [No. 8 Comparative Morphological and Systematic Studies of Bee Larvae With a Key to the Families of Hymenopterous Larvae lp 2 By Charles D. Michener CONTENTS PAGE Introduction 988 Acknowledgments 988 A brief account of hymenopterous larvae 989 A key to some groups of hymenopterous larvae 993 Evolutionary trends 995 Relative specialization of principal types of bee larvae 997 Reversibility in evolution 999 An account of evolution among bee larvae 1000 Some general problems of larval evolution 1005 Techniques 1006 External morphology of Anthophora stanfordiana 1007 Descriptive accounts 1015 Colletidae 1015 Halictidae 1023 Andrenidae 1033 Megachilidae 1040 Xylocopinae 1060 Anthophorinae 1066 Apinae 1084 Abstract: This paper is a study of known bee larvae from all parts of the world. It contains a key to the larvae of the major groups of Hymenoptera and an account of evolutionary trends among bee larvae. A detailed study of the external morphology of Anthophora is presented, ( 1 ) as a contribution to insect morphology, ( 2 ) as a basis for the establishment of a suitable terminology for the structures, and (3), as a basis for comparison of other genera. De- scriptive comments on known bee larvae are presented, together with fuller descriptions and figures of forms from the Western Hemisphere. 1. Prepared with the aid of grants from the University of Kansas General Research Fund and the Kansas Industrial Fellowship Fund. 2. Contribution number 802 from the Department of Entomology of the University of Kansas. (987) 988 The University Science Bulletin INTRODUCTION The purposes of this paper are to shed light on the evolutionary trends and phylogeny of the bees as exemplified by larval character- istics, to interpret insofar as possible problems of larval evolution generally, and to make possible identification of certain bee larvae when not accompanied by associated adults. The scope of the work is necessarily limited by the relatively few species which have been available for study. It is a deplorable fact that while many fine papers on bee biologies have been written, particularly in Europe, and while larvae of many bees have been seen, often sketched or photographed, most of these larvae have never found their way to museums nor have they been described or illustrated in detail. The magnificently illustrated works of Guido Grandi provide the most important exception; his descriptions and illustrations are virtually as useful as specimens themselves. The scope of the study is also limited almost exclusively to mature larvae. The earlier stages are even less available in collections than are mature larvae. Except as otherwise indicated, all descriptions and figures in this paper are based upon mature larvae. ACKNOWLEDGMENTS Special thanks are due to Dr. J. W. MacSwain of the University of California who assembled so many of the larvae used in this study that the work would have been impossible except for his generous loan of specimens. Others who have provided specimens of interest from their own collections or from collections of which they have charge include Drs. E. G. Linsley and P. D. Hurd of the University of California; Dr. B. D. Burks of the Division of Insect Identification, U. S. Department of Agriculture; Dr. G. E. Bohart of the Bureau of Entomology and Plant Quarantine, Logan, Utah; Dr. P. O. Ritcher of Oregon State College; and Dr. Warwick E. Kerr of the University of Sao Paulo. Mention should be made of several associates at the University of Kansas who have made observations on bee habits and obtained some larvae: Jimmie R. White, Alvaro Wille, Wallace E. LaBerge, and Roland L. Fischer. Special thanks are due to those who have helped with the prep- aration of the illustrations, Barbara L. Rozen, Julie Sokal, and Barbara Trotter. Michener: Comparative Studies of Bee Larvae 989 A BRIEF ACCOUNT OF HYMENOPTEROUS LARVAE In order to establish which of the characteristics of the various bee larvae are primitive and which are specialized, a general survey of hymenopterous larvae seems appropriate. As Peterson (1948) says, probably less is known about the larvae of Hymenoptera than about those of any of the other great orders. This applies par- ticularly to the Clistogastra, many students having been impressed by the fact that larvae of this group all look more or less alike as a result of reduction and loss of sclerotization and structures. No doubt this reduction is correlated with their habitats; none of the larvae of Clistogastra are free-living, all are inhabitants of plant or animal tissues or live in protected places (nests) constructed by their parents or their hosts. A recent author remarks on the sim- plified soft bodies of these insects and even goes so far as to say that sclerotized mouthparts are usually wanting. This statement no doubt results from an examination of the larva of Apis, whose mouthparts, even the mandibles, are relatively unsclerotized. A casual examination of figures of the mandibles of other bees will show its inaccuracy. No attempt has been made to give an exhaustive treatment of hymenopterous groups or to make a full survey of the literature other than that on bees. The following remarks, pertaining to mature larvae, are based on the principal literature and on the specimens at hand, special emphasis being placed on characters that vary among bee larvae because of the light which this infor- mation sheds on the relative specialization of such characters among bees. The larvae of the suborder Chalastogastra usually have thoracic legs and often abdominal prolegs as well; the body usually bears setae, and there are often sclerotic projections at the end of the abdomen. Free-living forms have eyes above the antennae. The antennae and maxillary and labial palpi are usually several seg- mented. The cardo is distinct from the stipes, while a lacinia and galea are recognizable. The salivary opening is single, not guarded by lips or sclerotic projections. The mandibles are robust, with the inner surfaces concave and with several stout apical teeth. Some works describing and illustrating these larvae are Peterson (1948), Yuasa (1922), MacGillivray (1914), Short (1952) and Tokunaga, Tsujita, and Yamazaki ( 1951 ) . The larvae of the suborder Clistogastra differ from those of the Chalastogastra in lacking legs, in being unsclerotized and pale 990 The University Science Bulletin in color except for the spiracles and portions of the head capsule and mouthparts, and in lacking eyes. The antennae and maxillary and labial palpi have at most one projecting segment. The laciniae are absent or not recognizable. The wasps have larvae more like those of Chalastogastra than do any other Clistogastra. In the Vespidae the body often bears a few setae and the antennae are reduced to mere convexities each of which bears a slightly more convex central area which is pro- vided with a few sensillae. The maxillae and labium are distinct projecting structures but the maxillary and labial palpi are one- segmented, papillalike. The former are similar to the galeae, so that each maxilla appears to bear two papillae. The cardo is dis- tinct from the stipes, while the lacinia is unrecognizable. The salivary opening is a transverse slit guarded by thin, slightly scle- rotized lips. The mandibles bear several large teeth at their apices, or the apices may be attenuate, bidentate or simple. Vespid larvae are well described and illustrated by several authors ( Grandi, 1928a, 1930, 1934, 1934a, 1935, 1937; Maneval, 1936, 1939; Micheli, 1930, 1934; Parker, 1943; Reid, 1942; Short, 1952; Soika, 1933, 1934). Reid has given an extensive treatment with a key to the subfamilies. The larvae of Scoliidae (Grandi, 1940), Formicidae (Athias- Henriot, 1947; Menozzi, 1936; Starke, 1949; W. M. Wheeler, 1918; G. C. Wheeler, 1928, 1935, 1943, 1948, 1950; G. C. and J. Wheeler, 1951, 1952), Pompilidae (Grandi, 1926, 1939; Maneval, 1936, 1939), Chrysididae (Maneval, 1936; Soika, 1934), and Sphecidae (Grandi, 1926, 1928, 1928a, 1929, 1930, 1931, 1934a; Maneval, 1936, 1939; Micheli, 1929, 1930, 1933, 1937; Soika, 1932, 1934) are essentially similar to those of the Vespidae. The body setae are present in some ( e. g. Sphecidae ) . The antennae are frequently a little better de- veloped than in the Vespidae with a distinct projecting papilla bearing the sensillae. The salivary opening is similar to that of the vespids in all of the above mentioned wasps groups except that in many Sphecidae it is divided into two openings, each of which may project as a long process. All intergrades exist within the Sphecidae from the single slitlike opening found in the Sphecinae (Ammo- phila, Sceliphron) and Astatinae, to the two small separate open- ings on projections characteristic of most other sphecids (see Michener, 1952). The maxillary galea is reduced in size in some sphecids (Cerceris, Grandi, 1926, 1928a; Philanthus, Grandi, 1931; and Miscophus, Maneval, 1939) and is in fact entirely absent in one minute sphecid, Ammoplaniis (Maneval, 1939). While absence of this papilla (galea) is ordinarily a character of the Apoidea, the Michener: Comparative Studies of Bee Larvae 991 Ammoplanus larva does not show other apoid characteristics, for it has the several apical mandibular teeth and the paired salivary openings characteristic of related sphecids. Larvae of the Evaniidae (Genieys, 1924) and Gasteruptionidae, subfamily Gasteruptioninae (Short, 1952) differ from the wasps in lacking maxillary galeae. In this respect they resemble the Apoidea, from which they differ in having broader, tridentate mandibles. Larvae of the Gasteruptionidae, subfamily Aulacinae ( Short, 1952 ) , have retained the galeae as distinct papillae and have paired salivary openings as in most subfamilies of Sphecidae. The Apoidea or bees, like the Evaniidae and the Gasteruptioninae, differ from the wasps principally in lacking the maxillary galeae so that there is only one papilla (the palpus) on each maxilla. Occa- sionally, as in Centris ( ■= Eulaema ) there is a projection in the position which would be occupied by the galea, but it is obviously a mere projection and not a distinct papilla, as is the galea in wasps. The bees also differ from wasps, ants, and the like in lacking the maxillary cardines as distinct sclerites, although in some groups, especially the megachilids and some halictids, they are represented by swellings. Moreover the mandibles are simple or bidentate at the apices, frequently with small teeth or serrations on the margins and inner surfaces. This is in contrast to wasps which usually have several large apical teeth and rarely have small teeth or serrations. The salivary opening is a slit, with sclerotic lips, or the lips may be absent; more rarely it is a round or oval opening, sometimes very inconspicuous. It is never bifid as in many sphecids. Setae are present on the body principally in the Megachilidae. The Oryssidae, which on the basis of adult characters are merely parasitic Chalastogastra, have larvae with mouthparts more re- duced than in the bees. The mandibles are tridentate with the median tooth longest, the whole mandible remarkably nearly bi- laterally symmetrical, as shown by Rohwer and Cushman (1917). The labium is separated from the maxillae by grooves which do not reach far back on the sides of the head. The maxillary struc- ture is difficult to interpret in the light of Rohwer and Cushman's illustration but there appears to be a basal piece (possibly cardo) at the level of the mandibular bases, not back toward the rear margin of the head. If this is the case, there is a great difference between the Oryssus structure and that of any other hymenopteron. Thanks to Dr. B. D. Burks and the authorities of the United States National Museum, I have been able to study some of the material used by Rohwer and Cushman, but it is now mounted on slides and 33—6490 992 The University Science Bulletin difficult to interpret. However, it does appear that the maxillae were more or less correctly illustrated by Rohwer and Cushman. The Chalcidoidea and Cynipoidea are characterized by the re- duction of all mouthpart structures except the mandibles; the latter are commonly acute and edentate although they are sometimes toothed apically, particularly in phytophagous cynipoids. The labium and maxillae are usually (in Chalcidoidea) indistinguishably fused with the head and scarcely recognizable; rarely in chalcidoids, more commonly in cynipoids, they are separated by grooves; at their apices the palpi are absent or reduced to minute papillae or the maxillary palpi are distinct tubercles in cynipoids; the galeae are absent, and the cardines unrecognizable. The salivary opening, when recognizable, is small, single, not guarded by lips, sometimes a transverse slit. The body is often provided with setae. Larvae of many species of chalcidoids are described and illus- trated by Buscalioni and Grandi (1938), Parker ( 1924 ) , Salt (1931), and Short (1952), while those of cynipids are described and illus- trated by Short (1952), and Yasumatsu (1943). The Ichneumonidae, Braconidae and Agriotypidae have remark- able mouthparts in which the rather small, usually acute and edentate, mandibles are the principal movable structures. The other appendages and sclerites of the front of the head are reduced to flat or slightly convex areas. The conspicuous structures in this region (except in Collyria) are not these appendages and sclerites, but the lines between them, which are strongly sclerotized. Sug- gestions of such sclerotization may be seen in bees and wasps, where they may indeed be conspicuous; however, in these forms the maxillae, labium, etc., are always projecting lobes. In the ichneumonoids, on the other hand, there are no projecting lobes or but feeble ones. The sclerotized lines are as follows: (1) one following the course of the epistomal suture, (2) one on each side extending between the anterior and posterior mandibular articula- tions ( pleurostomal thickening), (3) one extending from each pos- terior mandibular articulation to the posterior tentorial pit (hypo- stomal thickening), (4) one on each side separating the maxillae from the labium ( labio-maxillary thickening or maxillary sclerome of Vance and Smith, 1933 ) , ( 5 ) one dividing each maxilla between cardo and stipes (stipital sclerome of Vance and Smith), and (6) one separating the prementum from the postmentum ( labiostipital sclerome of Vance and Smith). In these insects there is only one maxillary papilla (presumably the palpus) and the salivary open- ing is slitlike, without distinct lips, or is apparently absent. Michener: Comparative Studies of Bee Larvae 993 Some excellent descriptive and illustrative work on larvae of ichneumonoids is provided by Beirne (1941), Salt (1931), Short ( 1952 ) , Thorpe ( 1930 ) , and Vance and Smith ( 1933 ) . Beirne and Short give keys for the separation of families and subfamilies. A Key to Some Groups of Hymenopterous Larvae 1 ( Based on mature larvae ) The following key to some groups of Hymenopterous larvae summarizes some of the information contained in the preceding paragraphs. In preparing it, I have borrowed freely from the keys of Peterson ( 1948) and Soika ( 1934). It should be remembered that in the Clistogastra only an exceedingly small percentage of the species have been studied. Therefore many exceptions will probably be found to the statements of the key; it is merely an outline to be improved as investigators obtain more information. 1. Antennae and maxillary and labial palpi distinctly several-seg- mented (if one-segmented, then apex of abdomen bears scle- rotic processes); lacinia distinct; eye spot usually present; thoracic legs usually present and segmented; abdominal pro- legs usually present ( suborder Chalastogastra ) 2 Antennae and maxillary and labial palpi one-segmented to ab- sent (apex of abdomen not sclerotized ) ; lacinia indistinct or absent; eye-spot absent; legs absent. ( Suborders Clistogastra and Idiogastra) 13 2. Thoracic legs with claws 3 Thoracic legs without claws 10 3. No segmented subanal appendages on the tenth abdominal seg- ment, if setiferous subanal knobs are present see couplet 8 (Xyelidae); thoracic legs usually stout, short and irregular in shape; prolegs usually present 4 Paired, segmented, subanal appendages present on the sternum of the tenth abdominal segment; thoracic legs slender, elon- gated and straight; prolegs absent Pamphiliidae 4. Claws on thoracic legs usually conspicuous and with a distinct pad or divergent lobe arising from the base of each tarsal (claw) segment 5 Claws on thoracic legs usually small and without a pad arising from the base of the tarsal (claw) segment. (If pads occur they arise from the tibia ) 6 5. Suckerlike lateral protuberances on abdominal segments 2 to 4 or 5 and 8 Acorduleceridae No suckerlike lateral protuberances on abdominal segments Argidae 6. Antennae with one (or two) segment(s) 7 Antennae with more than two segments 8 1. Since this paper was in proof the larva of an additional hymenopterous family, the Rhopalosomatidae, has been described (Gurney, Ashley B., 1953, Proc. U. S. Nat. Mus., vol. 103, pp. 19-34). The mature larva would run to couplet 14 in this key, agreeing with the Vespidae, etc., in having two papillae ( galea and palpus ) on each maxilla. It differs from these families in the small rather than traverse and slit-shaped salivary opening and in the presence of a long spine on either side of the salivary opening. The cardo isi not described and may not be recognizable. The structures which Gurney considered as eye-spots seem certainly to be something else; possibly they are pigmented spots at the anterior tentorial pits. The preceding larval stage has wholly different and much reduced mouthparts. 994 The University Science Bulletin 7. Mid-abdominal segments with seven annulets; spiracles definitely winged Cimbicidae Mid-abdominal segments with one to four annulets; spiracles not definitely winged (Fenusinae) Tenthredinidae 8. Antennae never with more than five segments 9 Antennae with six or seven segments. ( Prolegs on all abdominal segments; setiferous knobs or protuberances on the subanal areas of the caudal segment) Xyelidae 9. Antennae with three segments, the third peglike and the first and second incomplete, crescentic, flattened areas; prolegs present on abdominal segments 2 to 8 and 10 . . ( Diprioninae ) Diprionidae Antennae with four or five segments, if only three, all are com- plete or the third is short and not peglike; prolegs present on abdominal segments 2 to 6, 2 to 7, or 2 to 8 and 10, Tenthredinidae 10. Thoracic legs mammalike or absent; caudal segment with median sclerotized process 11 Thoracic legs indistinctly four segmented but without claws; caudal segment without median sclerotized process. (Prolegs vestigial on segments 2 to 8 and 10, caudal pair united, form- ing a single protuberance) ( Phyllotominae ) Tenthredinidae 11. No subanal appendages present; eye-spot not pigmented 12 Small and short subanal appendages present; eye-spot pig- mented. ( Antennae with four or five segments ) Cephidae 12. Antennae one-segmented; labial palpi one- or two-segmented; metathoracic spiracles conspicuous Siricidae Antennae with three or four segments; labial palpi three-seg- mented; metathoracic spiracles vestigial Xiphydriidae 13. Salivary opening paired, each opening on a projecting process. Gasteruptionidae ( subfamily Aulacinae ) and Sphecidae ( part ) Salivary opening single (slit-shaped, oval, round) or absent 14 14. Maxillae each with two papillae (galea and palpus). (Cardo separated from stipes by sclerotic line or at least by a fold ex- cept in some ants) . . Vespidae, Scoliidae, Pompilidae, Formicidae,1 Chrysididae, and Sphecidae (part) Maxillae with a single papilla (palpus) or without papillae 15 15. Most conspicuous structures of head, other than the mandibles, the heavy framework of sclerotic rods which lie on the lines between the primary structures of mouthparts and front of head; labium and maxillae not produced; cardo and stipes separated by sclerotic rod, the former very large 21 Lines between mouthparts and parts of head capsule not or feebly sclerotic; cardo and stipes fused or nearly so, without a con- spicuous sclerotic rod between them 16 16. Mandibles and head structures not sclerotized, only visible with staining Ichneumonidae ( subfamily Collyriinae ) Mandibles and usually other parts of head and mouthparts rec- ognizably sclerotic 17 1. In ants the galea and maxillary palpus are often represented by mere clusters of sensillae, or the palpus is represented by a single sensilla in Atta. The peculiar larvae of the ant Leptanilla (see Wheeler, 1928) does not run properly in this key. Michener: Comparative Studies of Bee Larvae 995 17. Maxillae and labium projecting as distinct separate lobes, usually distinguishable to their bases; salivary opening usually a slit. . 20 Maxillae and labium not projecting as separate lobes, either com- pletely fused and indistinguishable or partly so; salivary open- ing sometimes a slit, more often oval, round, or absent 18 18. Mandible symmetrical about longitudinal axis, apex tridentate, middle tooth longest Oryssidae Mandible asymmetrical, lower or outer tooth usually longer than others 19 19. Maxillae and labium reduced and usually indistinguishably fused; maxillary and labial palpi usually unrecognizable. . . . Chalcidoidea Maxillae and labium distinct apically; maxillary and labial palpi usually minute Cynipoidea 20. Mandibles with apices simple or bidentate Apoidea Mandibles with apices tridentate, Evaniidae, Gasteruptionidae (subfamily Gasteruptioninae ) 21. Last abdominal segment ending in two slender hooks; mandibles large Agriotypidae Last abdominal segment without hooks; mandibles small 21 22. Accessory longitudinal tracheal commissure present in thorax, Ichneumonidae ( most ) Accessory longitudinal tracheal commissure of thorax absent, Braconidae EVOLUTIONARY TRENDS With the aid of the data assembled in the preceding sections, it is possible to determine with some certainty whether some of the larval characters of bees are primitive or specialized. Thus, since setae are present on the bodies of the larva of primitive Hymen- optera, an obvious conclusion would be that setae on the bodies of bee larvae are primitive. Long and conspicuous antennal papil- lae and maxillary and labial palpi would be considered primitive compared to short ones or compared to the absence of these papillae and palpi, since these structures are several-segmented in the Chalastogastra and are represented by long and conspicuous papil- lae in most wasps. Among bees the antennae range from scarcely detectable convexities on the head capsule to distinct convex areas each bearing a cylindrical papilla, which may be as much as three times as long as broad. The palpi may be similarly long, or may be entirely absent. The salivary opening, in most wasps, is a transverse slit at the apex of the labial lobe guarded by a pair of thin, sclerotized lips. Such a structure is found in vespids, pompilids, chrysidids, and in some sphecids. In other sphecids the salivary opening is divided, so that there are actually two such openings, each at the apex of a 996 The University Science Bulletin projection. This is a specialization of certain sphecids, a condition not repeated elsewhere among the Hymenoptera except in the Aulacinae and having no counterpart among the bees. Among bees, therefore, the slitlike opening guarded by distinct lips would seem to be the primitive condition. Reduction of the lips, and reduc- tion of the slit to a straight or crescentic mark or to a small round hole must be specializations. The mandibles provide some of the most distinctive characteris- tics of the various groups of bee larvae. The mandibles of most sphecoid and vespoid wasps, chrysidids, evaniids, gasteruptionids, and for that matter those of the Chalastogastra as well, have sev- eral large sharp apical teeth. The greatest number of such teeth found in mandibles of bee larvae is two and it seems likely that acute, rounded, or truncate mandibular apices are derived from the bidentate type. The bees are peculiar in having, in many cases, numerous small teeth along the margins, those of the upper margin forming a band which curves onto the inner surface forming a "cusp". For bees this condition is considered primitive since it seems readily derivable from that found in certain sphecoid wasps in which there is a tooth on the inner surface of the mandible. In bees the inner surface of the mandible beyond the cusp is often hollowed out and the hollow margined by a more or less distinct ridge. This hollow can be seen with the cusp still present in Nomia, Augochlora, and Melissodes, although the cusp is much re- duced in the last. In most forms in which this concavity is well formed the cusp is gone, as are also most or all of the small teeth of the mandible. Although the concavity is a specialized character, it is obviously lost, a further specialization, in Trigona and Apis so that here its absence represents a greater specialization. In most bees the apices of the mandibles are much narrower than the bases, often acutely pointed, yet the mandibles are not drawn out into long attenuate apices. In some, however, the apices are much attenuate. This may occur among relatives of bees with a cusp (e.g., Colletes) or as an obvious derivation from forms with a concavity and no cusp (e. g., Trigona). Table 1 shows a list of eleven trends with the supposed primitive alternatives on the right, specialized alternatives on the left. Inter- mediate conditions occur in all cases, indicating intermediate de- grees of specialization. All of these trends involve characters which, from our present meager knowledge, appear to be of generic or greater importance at least somewhere among the bees. Michener: Comparative Studies of Bee Larvae 997 Table 1 Some trends involved in evolution of bee larvae, with primitive alterna- tives at the left, specialized, at the right. Characters are judged as to primi- tiveness on the basis of the Hymenoptera as a whole. Primitive 1. Setae present on body 2. Antennal papilla long 3. Maxillary and labial palpi long 4. Maxillae and labium distinct 5. Salivary opening a large slit, with lips 6. Prementum and postmentum clearly separate 7. Mandibles with large apical teeth 8. Mandibles with small teeth forming a cusp 9. Mandible without defined inner concavity 10. Apex of mandible rather broad 11. Peritreme present 1. 2. 3. 4. 5. 8. 9. Specialized Setae absent on body Antennal papilla absent Maxillary and labial palpi absent Maxillae and labium fused Salivary opening small, round, without lips Prementum and postmentum indistinguishably fused Mandible acute, rounded or truncate Mandible without cusp Mandible with clearly defined inner apical concavity 10. Apex of mandible attenuate 11. Peritreme absent RELATIVE SPECIALIZATION OF PRINCIPAL TYPES OF BEE LARVAE Considering the small number of bee larvae which are adequately known, it seems futile to attempt to establish a detailed scheme of relationships or phylogenetic tree based upon their characters. However, some general conclusions as to the light which the larval characters shed on relationships can be reached. No known bee larva agrees with all of the primitive characteristics listed in table 1, nor is there any which agrees with all of the spe- cialized features. In general bee larvae may be divided into two large and inter- grading groups, as shown below: Group II antennal papilla distinct mandible without cusp, but with apical inner concavity salivary opening a long transverse slit guarded by lips Group I antennal papilla reduced or absent mandible with a multidentate area or cusp on inner surface salivary opening reduced, without lips In the first group fall the Colletidae, Halictidae, Andrenidae, Melittidae, and the genera Neopasites, Nomada, Epeolus, and Triepeohis of the Apidae. From adult characters it seems likely that Exomalopsis and its relatives would fall here too, were their larvae known. (Colletes, Hylaeus, Euryglossa, Epeolus, Triepeohis and one species of Nomadopsis lack the mandibular cusps but the mandibles are attenuate, showing no resemblance to those of group II. Other larval characters show the obvious relationship of Epeolus 998 The University Science Bulletin and Triepeolus to Nomada, which has cusps, and of the Nomadopsis to another species of the genus which has them.) In the second group fall the Megachilidae and the Apidae (sense of Michener, 1944) except for the genera Neopasites, Nomada, Epeolus, and Triepeolus. (Antennal tubercles are reduced or ab- sent in Anthophora, although not in the similar larva of Melecta, and in Melipona, Trigona and Apis, although not in the similar Bombus. The mandibular concavity is absent in Stelis, although not in the similar larvae of other megachilids, in Ceratina, although not in the similar Xylocopa, and in Trigona and Apis although not in the simi- lar Melipona. The salivary opening is reduced in Ceratina and Xylocopa. If table 1 is correct, as it appears to be from a survey of the characters of the order as a whole, it is obvious that neither the group I larvae nor the group II larvae can be established as primi- tive, since group I combines primitive mandibles with specialized antennae and salivary opening while group II combines specialized mandibles, with primitive antennae and salivary opening. Making use of other characters, one finds that labio-maxillary fusion, an obvious specialization, occurs principally in group I (Colletes, Panurginae, Nomada, Triepeolus, to a lesser extent in Hesperapis and the Halictidae ) . Setae on the body surface, presumably primi- tive, occur only in group II (Megachilidae, AUodape, a few in Bombus). Large apical mandibular teeth, another primitive char- acter, are found in both groups but principally in group II. It is evident that the larvae of Megachilidae exhibit more sup- posedly primitive characters than those of any other bees, agreeing with the primitive alternatives of all trends listed in table 1 except for numbers 8 and 9, which might almost be considered a single trend. If this conclusion is accepted, then the other bees having group II larvae are more specialized in larval characters than the Megachilidae, while the group I larvae have even more specialized features, some of them agreeing with the specialized alternatives of most of the trends except numbers 8 and 9 in table 1. It is scarcely necessary to mention that the combining of primi- tive and specialized traits is very common, for probably no living animal is primitive in every character. One may speak of a certain structure process, or function as primitive in relation to another structure, process, or function with some degree of safety, but when one speaks of a primitive organism it must be with the understand- ing that it has some specialized characters. Michener: Comparative Studies of Bee Larvae 999 REVERSIBILITY IN EVOLUTION A comparison of the foregoing with what is known of bee phy- togeny, based on adult characters (see Michener, 1944) reveals the curious fact that the group II larvae, i. e. those having a number of primitive characters, are associated not with the adults which have numerous primitive characters but with long-tongued adults of specialized groups. The phytogeny of bees, in its general outlines ( e. g. evolution of the proboscis from a short one like that of wasps to a much elongated one ) , is so clear and is supported by so many adult characters that it can scarcely be doubted. Yet the summa- tion of larval characters suggests an opposite evolutionary direction. A further study of the larval characters themselves, however, shows that some of them do fit into the general picture of bee phytogeny. Those that do so are characters 8 and 9 of table 1. The way in which these characters have probably evolved will be shown below. The remaining characters are all ones in which specialization in- volves reduction or loss; all structures involved are relatively simple, for example, papillae and setae. The specialized condition of these characters in primitive bees could be explained as a series of parallel developments. Thus, body setae may have been present in all of the ancestors along the phyletic line leading to the Megachilidae, and such setae lost independently in each of the stocks which branched from this phyletic line. The same might have been true of other structures, such as the antennal papillae. For at least some of these characters, a different explanation seems probable, namely that the structures were lost (or reduced) in very early bee evolution and later regained. The explanation seems more probable because it avoids the hypothesis of very numerous parallelisms. Although the "law of irreversibility" in evolution or Dollo's "law" is well known, there are many excep- tions to it. Muller (1939) pointed out that there is no genetic basis for it. Most structures are controlled by many genes so that exact reversion is most unlikely but it is obvious that where only one or a few genes are concerned, as might be the case where mere length of a papilla or palpus is concerned, exact reversion might occur as a result of reverse mutations. Another possible explana- tion for the apparent reversions in larval characters of bees may be that the structures are lost morphologically as a result of inhibiting factors or complexes but the genetic determinants for the structures retained. Later, if the inhibition were removed, the structures 1000 The University Science Bulletin would reappear unless the loci which govern them were altered in function during the time when the structures were not being pro- duced. Thus it seems that while the relatively complicated mandibular structures do offer characters of easily understood phylogenetic value, the other structures considered, probably because of rever- sions, show a reversal of expected phylogenetic significance. If the hypotheses presented are correct, the primitive and specialized alternatives for these characters shown in table I should be reversed for the bees. AN ACCOUNT OF EVOLUTION AMONG BEE LABVAE As indicated previously, the larvae of sphecid wasps usually have mandibles with several large apical teeth. Presumably in asso- ciation with the change of food from insects or spiders to pollen and honey, the mandibles of bee larvae became smaller and no known form retains more than two large apical teeth. The mandi- bles acquired many small teeth, some of them grouped to form a cusp on the inner surface some distance from the apex, so that there is a concave surface between the cusp and the apex. Larvae having such mandibles are found among the short-tongued families of bees and as already indicated, in certain genera of Apidae. Presumably this mandibular structure is in some way correlated with handling the pollen food for, with some exceptions, all bee larvae have such a concave space, although it is quite different in nature in some. The most primitive of the bees would undoubtedly be placed in the Colletidae. In this family the mother bee applies to the inner surfaces of the cells a thin transparent or translucent substance. This seems to take the place of the cocoon spun by the larva in wasps, for colletids, unlike their presumed ancestors (Sphecidae) spin no cocoons. It seems likely that the cell lining of the colletids and the cocoon of the wasps serve for water conservation. It is noteworthy that among bee larvae as a whole, those which do spin cocoons have larger antennal papillae and palpi than those that do not. Of course they also have a large salivary opening which takes the form of a transverse slit guarded by lips, while those which do not spin usually have the salivary opening reduced to a curved slit, an oval, or a circle, and while it is sometimes pro- duced, it does not have lips. It seems reasonable to suppose that the antennae and palpi bear the sensillae that enable the larva to apply its silk properly in cocoon construction. Among forms which do not spin the palpi are often not so completely reduced as the Michener: Comparative Studies of Bee Larvae 1001 antennal papillae. Therefore it seems not unlikely that the palpi also bear sensillae utilized in some other activity, perhaps eating. Considering the above it is not surprising that the larvae of the Colletidae have reduced antennal tubercles and palpi and a re- duced salivary opening ( i. e. they are group I larvae ) , for they construct no cocoons. In the absence of the need for a cocoon, there would be mutation pressures against the organs making cocoon spinning possible with no corresponding selection pres- sures maintaining these organs. Indeed, because of competition among the organs for the nutrient materials available, the indi- viduals without the organs necessary for spinning should be at an advantage in a mixed population. Among the Colletidae three of the genera studied (Hylaeus, Euryglossa and Colletes, apparently also Lonchopria) have the apices of the mandibles reduced and attenuate, the cusp absent. Similar attenuation can be found among some of the other bees; the reason is unknown but a suggestion is that it occurs in groups whose provisions are usually liquid. Perhaps with such provisions mandibles are not very effective. That there is no universal cor- relation between mandibular structure and food consistency is shown, however, by the fact that a cuckoo bee larva often has quite different mandibles from the host whose provisions it eats (com- pare, for example, Triepeolus and Melissodes) . The Halictidae, Andrenidae and Melittidae, like the Colletidae, have group I larvae with mandibular cusps and with reduced antennal papillae, palpi, and salivary openings. The cells are lined with a very thin layer of wax, so that cocoons are presumably un- necessary. Systropha, however, is reported to spin a cocoon. It may be that, like the higher bees, it has reacquired (or retained?) the necessary organs. The genus Neopasites has larvae which, in head characters, are amazingly like certain halictids. Nomada likewise has larvae with mandibular cusps, reduced salivary opening, etc. Epeolus and Triepeolus are similar to No?nada except that their mandibles are attenuate, the cusps gone. All of these genera are parasitic, but it may be that their pollen-collecting ancestors among the Exo- malopsini have similar group I larvae. Unfortunately larvae of Exomalopsini are virtually unknown. Claude-Joseph (1926) briefly describes and figures the larva of Exomalopsis. Details of mandib- ular structure are not shown, but the small salivary opening sug- gests a group I larva. The antennal papillae and palpi seem elon- gate, however, as in group II larvae. At least some of the more 1002 The University Science Bulletin ANTHOPHORA Fig. 1. Inner views of mandibles of various bee larvae showing the develop- ment of mandibular structure. The vertical series illustrates reduction and loss of the multidentate cusp and development of the carina surrounding the inner apical concavity of the mandible. The offshoots to the right illustrate attenua- tion and reduction of the mandibular apices. Michener: Comparative Studies of Bee Larvae 1003 primitive Apidae ( sense of Michener, 1944 ) have larvae much more like those of the short-tongued families than those of the other Apidae and the Megachilidae. This means that, unless the man- dibular structure of the type II larva arose twice, the Megachilidae arose from apid ancestors rather than from preapid ancestors as indicated by me in 1944. It suggests, furthermore, that Neopasites, Nomada, Epeolus, etc., together with the Exomalopsini might con- stitute a subfamily (or family) division co-ordinate with the Xylo- copinae, Anthophorinae, Apinae, and Fideliinae. This division would be called the Nomadinae. Its limits are unknown at present. Further study may indicate adult characters by which it can be recognized. An alternative explanation of the occurrence of group I larvae among the parasitic Apidae (Neopasites, Nomada, Epeolus, Triepeolus) is that these larvae have mandibles, antennae, and palpi adapted to the type of nest and provisions supplied by their hosts. This suggestion could explain the characteristics of the first three genera listed for they are parasites of bees having group I larvae. Triepeolus, however, as already stated, is a parasite of Melissodes which has fairly typical group II larvae although remnants of man- dibular cusps persist. The larva of Nomada is so similar to those of Epeolus and Trie- peolus as to lend weight to a conclusion reached from a study of adults of South American forms, namely that the Nomadini, Osirini, and Epeolini had a common parasitic ancestor. Details of this will be published elsewhere. In certain group I larvae ( e. g. Augochlora, in the Halictidae ) , the space between the cusp and the apex of the mandible is strongly concave and the concavity is bordered by a sharp line or ridge cutting obliquely across the mandible. In group II larvae (Mega- chilidae and Apidae except Neopasites, Nomada, Epeolus, and their relatives) the cusp, and usually the small teeth of the man- dibular margins, are gone but a distinct inner apical concavity exists, margined basally by a transverse or oblique line or ridge. In Melissodes a small remnant of a cusp is retained and the line margining the concavity is oblique; in others (e.g. Emphor), the cusp is entirely gone, the line oblique; in most group II larvae, however, the line is transverse. As among the group I larvae, there are group II forms in which the mandible becomes attenuate and loses its characteristics group features. This is true of Ceratina, Trigona, Apis, and some species of Stelis. In Apis the mandible is so reduced as to not even be attenuate, as there is no longer a pro- jecting apex. 1004 The University Science Bulletin More or less concurrently in phyletic history with the develop- ment of the type of mandible usual for type II larvae came the return to cocoon spinning habits. This may be associated with the fact that many of the bees of this lineage fail to line their cells with wax or other waterproof material. The Megachilidae, for example, do not do so. In any event the cocoon spinners among the bees have, as already indicated, larger antennae and palpi than others and large salivary openings with lips. It is interesting that, although obviously a group II larva, cocoon spinning has been abandoned in Anthophora. Its antennal papillae are reduced, as are its palpi. Most interestingly, its salivary opening, while of the shape of that of a typical group II larva, is much reduced in size, moved dorsally and no longer at the apex of the labium, and is provided with small nonsclerotic lips. The larvae of the Xylocopinae are an exception to the rule that forms with reduced salivary openings also have reduced antennae, for the antennal papillae are of moderate size in spite of extreme reduction of salivary openings. With regard to the setae on the bodies of larvae, it is remarkable that they are present only on a few of the most specialized of the groups of bees. Thus, scattered setae occur in the social genus Bombus, while setae are more numerous on the bodies of at least some groups of subsocial AUodope, and on megachilids which are among the most specialized of the solitary bees. Some of the best specific characters found among bee larvae occur among the spiracles. Thus Ritcher (1933) was able to dis- tinguish every species of Bombus larva available to him by means of the spiracles, and often these were the only specific characters detected. Although they may provide generic characters, the spi- racles do not provide much help in the recognition of higher cate- gories. Their various features suggest an endless series of recom- binations of a few characteristics. The spines of the atrium, the spines of the subatrium, the peritreme, the collar, etc., may be large, small, or absent, each apparently without regard to the condition of the other structures. The functional significance, if any, of the spiracular spines and other spiracular ornamentation is unknown. Possibly these struc- tures are of value in keeping pollen grains out of the openings. If so, the variability might be correlated to some extent with vari- ability in the kinds of pollen utilized by the bees. It is possibly significant that the only bee larva known to me in which important differences exist between anterior and posterior spiracles is the first Michener: Comparative Studies of Bee Larvae 1005 stage larva of Coelioxys, in which the posterior spiracles are much larger than the others. The first stage larva lives within the rather liquid provisions of Megachile, not on the surface as does the Mega- chile and as does the Coelioxys also after the first stage. The elevations or tubercles of the larval body also vary, like the spiracular characters, in an apparently meaningless manner. In general the tubercles seem higher and more consistently present in group I larvae than in group II larvae, but this is only a tendency. As is well known (Brauns, 1926; Bayment, 1951) the larvae of certain Ceratinini (some species of AUodape, Exoneura) have peculiarly elongate and modified ventrolateral projections on the anterior part of the body which are used in holding the small quan- tities of pollen offered the larvae by these bees, which feed the larvae progressively. Other species of AUodape are provided with numerous body tubercles arranged in more than the usual series (dorsolateral and ventrolateral) found in other bee larvae. Ap- parently these serve to anchor the larvae in groups on the insides of hollow stems where they are fed progressively. A long known peculiarity of some parasitic bee larvae is the great size of the head and mandibles of the first stage larva. It is in this stage that larvae of these cockoo bees destroy the egg or larva of their host. This condition occurs in quite unrelated parasites, for example, Coelioxys (Megachilidae), Triepeohis (Apidae), and Melecta (Apidae). SOME GENEBAL PBOBLEMS OF LABVAL EVOLUTION Certain general problems concerning larval forms which have been more or less evident to entomologists for many years are per- haps worth brief consideration here. The problems involved are incorporated into the following question: How does it happen that larvae can evolve in quite different directions and at different rates from adults, adapting themselves to different food require- ments, habitats, and the like? As to rates, it must of course be remembered that what is meant is rates of phenotypic evolution; there can be no differences in rates of true or genotypic evolution between larvae and adults. It should be pointed out that mutations affect processes, and only indirectly, characters (see Beadle, 1945, 1946). A given process may influence only early development, only late develop- ment, or both. A mutation that affects a process influencing only a certain portion of the development might affect principally larval 1006 The University Science Bulletin characters, since the cells which will develop into the adult struc- tures are segregated into primordia which do not develop much until larval life is well along, or it might affect only adult characters if its influence is not felt until late in development. A mutation may affect a process influencing characters of selective value in one state ( e. g. the larva ) and other characters of little selective value in another stage. It may influence characters visible phenotypically and externally in one stage, but not in the other. It is thus evident that the basis exists for the appearance and subsequent selection of characters of one stage in the life history, with little effect on other stages. Of course since larva and adult are the same individual, selection will depend upon success in both stages; it cannot act independently on either. Fundamentally the situation is little different from that which makes it possible for the young of animals without metamorphosis to have distinctive fea- tures ( e. g. immature plumages of birds ) different from those of the adult and subject to their own partially independent evolution- ary development. The segregation of larval from adult characters is certainly exaggerated in holometabolous insects. Possibly this is due in part to the fact that, at least in Drosophila and Tineola, larval and adult structures are determined at different times during development. Thus there is a time when minor injuries to the egg will cause larval but not adult defects, a later time when they will cause adult but not larval defects. ( For a brief summary of our knowledge in this connection, see Wigglesworth, 1950. ) Rapid evolution of one stage, without comparable change in the other, could easily result and is not fundamentally different from another common phenomenon, namely rapid evolution of one struc- ture or group of structures while the others remain little changed. In this light it is easy to see that larval characters are merely more characters, of no more systematic value and of no less, than adult characters. TECHNIQUES All studies were based upon larvae preserved in alcohol. This is best done by dropping live larvae into nearly boiling water for a few minutes before they are transferred to 70% alcohol. The ma- terial available, however, having been preserved by various col- lectors and at different times and in different ways, is in many different states of preservation and the drawings of whole larvae are undoubtedly adversely affected by this, for some specimens were Michexer: Comparative Studies of Bee Larvae 1007 badly shriveled. It should be pointed out also that last stage larvae vary greatly in shape, size of the tubercles, and the like according to their age. Two drawings of Emphor larvae are presented to illustrate a perhaps extreme example of this point. For most species only a few last stage larvae and these of unknown age, were avail- able. There is no doubt, however, that in most instances the hiber- nating form is the one illustrated. In order to study the mandibles in detail they may be removed from the head and cleared with caustic, or the whole head treated in the same manner. Spiracles were studied by removing them with a bit of adjoining integument and boiling in caustic. The parts so treated can be stained with acid fuchsin if necessary, and were examined in drops of glycerin so that, even under high magnification they could be moved about for study. For preservation such parts were placed in alcohol in small vials within the larger vial contain- ing the rest of the larva. Illustrations of heads were prepared from a study of the whole head both before and after clearing in caustic. Illustrations of the mandibles were prepared by removing a man- dible from the head and examining it in glycerin. Orientation was by means of the mandibular apodemes and articulations. In the inner view the adductor apodeme is as nearly as possible directly above the abductor while in the ventral (or dorsal) view the points of articulation are one above the other and the apodemes are at lateral extremities of the mandibular base. Illustrations of spiracles were prepared with the aid of a com- pound microscope. Because of the need for frequent reorientation of the specimens, oil immersion was rarely used. Lateral views of spiracles are optical sections, while facial views show structures visible at all levels. EXTERNAL MORPHOLOGY OF ANTHOPHORA STANFORDIANA COCKERELL The following morphological account, based on the mature larva, is presented in order to provide a basis for the terminology used elsewhere in this paper and as a contribution to the general fund of knowledge on insect morphology. Anthophora was selected, not only because its larvae were available in numbers but because it is the genus utilized for a comparable study of adult morphology (see Michener, 1944). The larva is grublike, robust, curved, with no separation between the thorax and the abdomen (figure 7). The color is yellowish 34—6490 1008 The University Science Bulletin white, the spiracular peritremes and the harder parts of the head brownish, the apices of the mandibles black. The head (figures 8 to 11) is rounded, that part above (not in- cluding) the maxillae slightly sclerotic, this sclerotized portion which may be called the head capsule thickened and slightly in- flexed along posterior margin. At each lower posterior angle of the head capsule there is a deep invagination into the posterior tentorial pit. The thickening along the posterior margin of the head capsule (termed postocciput by Vance and Smith, 1933) is directly connected with the posterior tentorial arms and continues anteriorly as the hypostomal thickening along the lateral margins of the head capsule to the bases of the mandibles. [These thickenings are termed habenae by Starke (1949).] From each lower (i.e. posterior) mandibular articulation the marginal thickening of the head capsule bends dorsally as the pleurostomal thickening, ex- tending around the posterior margin of the mandibular corium to the anterior mandibular articulation where it meets with the thick- ening which is the internal evidence of the epistomal suture. This entire thickening (including hypostomal and pleurostomal thicken- ings) may be called the marginal thickening of the head capsule. The cleavage lines (see Snodgrass, 1947) or epicranial suture are weak, A-shaped, the unpaired portion appearing rather thick, the arms feeble, appearing as scarcely recognizable creases. Each antenna is a blisterlike convexity with a darkened center. The parietal bands are ill defined, slightly differentiated areas of cuticle (termed optic plate by Nelson, 1924, temporal fossa by Vance and Smith, 1933, and bandarella temporale by Grandi, 1934, etc.). The epistomal suture, separating the clypeus x from the frons, arches between the anterior mandibular articulations, which are in deep pockets or precoilae (Ritcher, 1933) beneath the sides of the cly- peus. The epistomal suture is feeble medially and is marked in- ternally by a ridge (likewise weak medially) which is connected laterally to the marginal thickening of the head capsule. The anterior tentorial arms arise from this internal ridge above and mesad of the precoilae. The labroclupeal suture is a mere fold. The lahrum is a thick projection from the lower margin of the clypeus which bears apically a pair of tubercles, the lahral tuber- cles. Beneath the labrum is a transverse slit, the mouth, leading 1. Du Porte (1946) considers this structure as the frontoclypeus. However, I see no reason for disregarding the evidence provided by muscular attachments emphasized by Snodgrass (1947), and have therefore followed Snodgrass in the interpretation of the face. Since this paper was in proof a subsequent paper (Du Porte, E. Melville and R. S. Bige- low, 1953, Canadian Jour. Zool., vol. 31, pp. 20-29) greatly modifies Du Porte's earlier ( 1946) views. Michener: Comparative Studies of Bee Larvae 1009 into the foregut, the floor of which, a short distance behind the mouth, is marked by a transverse fold. Below the mouth is the small convex hypopharynx. The salivary opening is a transverse slit, shorter than the mouth, between two thin sclerotized projecting lips. The labium, hypopharynx, and to some extent the maxillae form a single complex of fused or partially fused parts. For morphological purposes it is assumed that the salivary open- ing is a suitable landmark lying between the hypopharynx and the prementum. From a study of Anthophora alone this would not be questioned, since in this genus the lips of the salivary opening are in a feeble transverse groove which seems to be the premental- hypopharyngeal line. In most other bee larvae studied a more distinct transverse groove exists in about the same position; how- ever, the salivary opening is more apical in position on the labio- hypopharyngeal lobe and not in or near the groove. The groove, here called the hypopharyngeal groove, cannot be the premental- hypopharyngeal line if the salivary opening is the true landmark. Nonetheless, for descriptive purposes, the area basal to this groove is called the hypopharynx, that distal to it the prementum. This is done because the hypopharynx in this sense is frequently distinc- tively spiculate while the area beyond the hypopharyngeal groove is usually like the prementum in surface characteristics. Dobrovsky (1951) calls the entire region the labium-hypopharynx without attempting to differentiate the two components. The simpler ex- pression "labial lobe" has been used in the descriptions which follow. The mandibles are distinctly sclerotic, truncate, the inner apical surface concave, this concavity limited basally by an arcuate carina. The broad membranous mandibular corium lies between the man- dibular articulations. The maxillae are fleshy, with no separation of cardines and stipites. [Some bees (some megachilids, halictids, and Bombus) have a small faintly sclerotic area on the outer surface of the maxilla at its base ( posterior end ) which represents the cardo, and extend- ing distad from this a weakly sclerotic ribbon described below as the labio-maxillary rod.] Near the apex of each maxilla is a project- ing, slightly sclerotic papilla, the maxillary palpus. Various authors ( e. g. Kellogg, 1902, and Nelson, 1924 ) have cor- rectly interpreted this structure. Ritcher (1933) in an important paper on bumblebee larvae, challenges this view and on the basis of ontogenetic evidence states that the papilla is not the palpus but 1010 The University Science Bulletin polpus stipes palpus - lacinio cordo palpus- palpus galea palpus lacinia — galea locinia Figs. 2-6. 2. Diagram of maxilla of larva of Tenthredinidae. 3. Diagram of maxilla of larva of Sphecidae. 4. Diagram of maxilla of larva of Antho- phorinae. 5. Ventral view of right maxilla of larva Bombus americanorum, showing by broken lines the developing pupal maxilla within. 6. Ventral view of right maxilla of an individual of Bombus americanorum as it is becoming a pupa. is the distal part of the galea. It is clear that he misinterpreted the evidence before him, however, largely due to the fact that in the adult and pupal Bombus the galea is a very large structure and the palpus very small. Grandi (1934a) shows clearly, by a series of developmental stages, that in Polistes, which has two maxillary papillae, the outer one is the palpus. In those sphecoid wasps in which one of the two papillae is reduced in size, it is the inner one; thus one might presume that the remaining one in the bees would be the outer one, or palpus. A re-examination of the developmental stages in Bombus clearly establishes this surmise. As time for pupa- Michener: Comparative Studies of Bee Larvae 1011 tion approaches, the larval maxilla is seen to contain two projecting processes, the longest and most slender reaching the larval palpus, the other shorter (fig. 5 ) and reaching toward the tip of the maxilla but not toward the palpus. It was perhaps natural for Ritcher to conclude that the longer process must represent the galea, which is so long in the pupa and adult. This leaves the inner process un- explained, however. The obvious conclusion is that the outer process, in spite of its length, represents the palpus, the inner the galea, and a weak convexity on the inner surface of the developing pupal maxilla the lacinia. That this interpretation is correct is shown in figure 6, a diagram of the maxilla of a pupa which had recently emerged from the last larval skin. In this individual the processes can readily be homologized with adult structures, and almost equally readily, with the processes of the pupal maxilla seen within the larval maxilla. The phylogenetic series of larval galeae also supports this interpretation ( see figures 2 to 4 ) . The upper margin of the maxilla, beneath the hypostomal thick- ening of the head capsule, is curved inward to form a furrow. Nel- son (1924) calls this the lateral furrow but it is more accurately termed the hypostomal furrow. In many aculeate Hymenoptera there is a sclerotic rod or thickening marking the union between the maxilla and the postmentum (see Snodgrass, 1935, and Ritcher, 1933). The upper or posterior end of this rod is connected with the hypostomal thickening. This rod is very feeble in Anthophora, but is termed the labiomaxillary rod or thickening. This structure is present in the tenthredinids, as a fragmentum of the stipes which supports the lacinia, or perhaps as a basal prolongation of the lacinia itself so that the latter contacts the cardo. Its reduction is clearly shown in the series of maxillae ( figures 2 to 4 ) . The labium is divided by a weak transverse furrow into two parts which, following Snodgrass (1935), are termed the postmentum and prementum. The prementum bears a pair of apical tubercles which appear to represent labial palpi (see Ritcher, 1933). The tentorium in larvae which are not preparing to molt consists of posterior tentorial arms, united by a tentorial bridge, connected to the anterior tentorial arms which give rise to the slender dorsal tentorial arms. As shown in figure 14, the bridge, the dorsal arms, and the posterior halves of the anterior arms disappear before the pupal molt, probably before every molt since molting would seem to be impossible with the tentorium intact. The portions which do not disappear are more heavily sclerotic than the remainder of the tentorium. 1012 The University Science Bulletin caudal annulet cephalic annulet spiracle anal opening abductor apodeme of mandible dorsal tentorial arm 8 lobrum labral tubercle mouth mandible hypopharynx salivary opening anterior tentorial arm ' I ' epistomal ridge ] ' antenna / V X I I \ \ \ clypeus \ posterior thickening of head capsule adductor apodeme of mondible foregut tentoriol bridge solivOTium maxtllo hypostomal thickening premenlum postmentum abductor opodeme of mandible anterior tentoriol arm / W*!* b°nd . , . . . „rm / ' dorsal tentorial arm frons * i / I / \ epistomal suture antenna | / clypeus \ \ labroclypeal suture labrum-^. ~~ labral tubercle mandibular corium — '/ adductor apodeme of mandible mandible- salivary opening - posterior thickening of head capsule —posterior tentorial orm posterior tentorial pit / / / moxillory palpus j premenlum / \ postmentum salivcrium hypostomal thickening pleurostomal thickening Figs. 7-9. Anthophora stanfordiana, mature larva; 7, lateral view; 8, head cut on saggital line to show internal skeletal structures; 9, lateral view of head. Michener: Comparative Studies of Bee Larvae 1013 clevage imo 10 frons cdductor opodeme of mandible s epistomol suture \ \ \ \ ciypeus^ obductor opodeme of mandible ^ lobroclypeol suture labrum — _ labrol tubercle . labium 7f po/ielol bond ( '*) ~~J antenna . — -onterlor tentorial arm "7 _- — ' anterior tentorial pit i^- — pleurostomol thickening mandibular corium ■ — mandible / moxillo maxillory palpus \ / labial palpus solivory opening II adductor opodeme of mandible — obductor opodeme of mandible — mandibular corium — maxilla maxillary pa maxillary palpus \iHl;li